Professional Documents
Culture Documents
Pragmatics of Computer-Mediated Communication
Pragmatics of Computer-Mediated Communication
Pragmatics of Computer-Mediated Communication
HoPs 9
Handbooks of Pragmatics
Editors
Wolfram Bublitz
Andreas H. Jucker
Klaus P. Schneider
Volume 9
De Gruyter Mouton
Pragmatics of
Computer-Mediated
Communication
Edited by
Susan C. Herring
Dieter Stein
Tuija Virtanen
De Gruyter Mouton
ISBN 978-3-11-021445-1
e-ISBN 978-3-11-021446-8
orientation to the user of these handbooks. The series specifically pursues the fol-
lowing aims:
The nine volumes are arranged according to the following principles. The first
three volumes are dedicated to the foundations of pragmatics with a focus on micro
and macro units: Foundations must be at the beginning (volume 1), followed by
the core concepts in pragmatics, speech actions (micro level in volume 2) and dis-
course (macro level in volume 3). The following three volumes provide cognitive
(volume 4), societal (volume 5) and interactional (volume 6) perspectives. The
remaining three volumes discuss variability from a cultural and contrastive (vol-
ume 7), a diachronic (volume 8) and a medial perspective (volume 9):
1. Foundations of pragmatics
Wolfram Bublitz and Neal R. Norrick
2. Pragmatics of speech actions
Marina Sbisà and Ken Turner
3. Pragmatics of discourse
Klaus P. Schneider and Anne Barron
4. Cognitive pragmatics
Hans-Jörg Schmid
5. Pragmatics of society
Gisle Andersen and Karin Aijmer
6. Interpersonal pragmatics
Miriam A. Locher and Sage L. Graham
7. Pragmatics across languages and cultures
Anna Trosborg
8. Historical pragmatics
Andreas H. Jucker and Irma Taavitsainen
9. Pragmatics of computer-mediated communication
Susan C. Herring, Dieter Stein and Tuija Virtanen
2. Email communication
Christa Dürscheid and Carmen Frehner . . . . . . . . . . . . . . . 35
4. Blogging
Cornelius Puschmann . . . . . . . . . . . . . . . . . . . . . . . . . 83
5. Real-time chat
John C. Paolillo and Asta Zelenkauskaite . . . . . . . . . . . . . . . 109
6. Instant messaging
Naomi S. Baron . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
7. Text messaging
Crispin Thurlow and Michele Poff . . . . . . . . . . . . . . . . . . . 163
V. Broader perspectives
by earlier work on speech and writing (e.g., Collot and Belmore 1996, with refer-
ence to Biber 1988).4
Until recently, the vast majority of language-oriented CMC research had as its
subject matter English-language CMC – a reflection, in part, of the origins of the
Internet in the United States. This, too, has been changing, especially as regards
discourse analysis and sociolinguistic research (see, e.g., the papers in Danet and
Herring 2007), in accordance with the rapid diffusion of the Internet to other coun-
tries starting in the mid-1990s. Native-language traditions of research into com-
puter-mediated language have now become established in Germany, France, and
the Nordic countries, and are starting to emerge in Japan, China, Spain, Italy, and
Greece.5 Cross-linguistic research has identified both similarities and cultural spe-
cificities as regards language use online.
While most of the topics mentioned above can be subsumed under the broad
heading of pragmatics, studies focusing explicitly on pragmatic issues are a more
recent phenomenon. When the first author, Susan Herring, was putting together a
keynote lecture for the 2007 International Pragmatics Conference on “The Prag-
matics of Computer-Mediated Communication: Prospectus for an Emerging Re-
search Agenda”, she found few high quality published works to review – or rather,
few that could be classified as pragmatic first and foremost,6 as opposed to the
many studies that could also be classified as discourse analysis, conversation
analysis, sociolinguistics, and the like. In that lecture, she argued that it is nonethe-
less useful to distinguish a “pragmatics of CMC” from other language-focused
approaches to CMC, inasmuch as it can benefit from drawing on pragmatic theory,
as well as methods of analysis developed within the tradition of linguistic prag-
matics, to provide potentially unique perspectives. Specifically, she recommended
that such an approach focus on three kinds of phenomena: 1) classical core prag-
matic phenomena (e.g., implicature, presupposition, relevance, speech acts, polite-
ness) in CMC, 2) CMC-specific phenomena (e.g., emoticons, nicknames, “Net-
speak”), and 3) CMC genres or modes (blogs, SMS, wikis, chat, etc.). That
proposal informs the broad organization of the present volume. However, in the
end it proved difficult to separate out discourse-pragmatic from “core” phenomena,
with the result that the volume includes a section on the discourse pragmatics of
CMC as well.
In short, the “pragmatics of CMC” is still in flux. Yet a handbook traditionally
brings together comprehensive and canonized knowledge, blessed by the passage
of time. Strictly speaking, given the speed of knowledge creation in the field of
CMC pragmatics and its relative youth, this handbook should not be possible.
There are gaps in its coverage, and the research presented in some chapters is new.
For these reasons, the collection might perhaps better be viewed as presenting the
“state of the art” in an emergent field rather than as a distillation of time-honored
knowledge. One notable gap is in the coverage of Web 2.0 phenomena such as
wikis, microblogging, and social network sites, about which significant bodies of
2. Perspectives on pragmatics
characterized by a broader perspective that takes into account the complex inter-
weaving between language use and its technologically-mediated forms.
Another way to situate the approach represented here is suggested by Ariel’s
(2010) overview of how the field of pragmatics might be internally structured.
Ariel makes a useful distinction between two approaches to pragmatics. “Border-
seekers” focus on the distinction between semantics and pragmatics, restricting
the list of topics that belong to each. For border-seekers, pragmatics is predomi-
nantly the domain of a particular set of topics that are defined as pragmatic, rather
than syntactic or semantic. This orientation facilitates the description of grammar
along relatively simple lines, as the “messy” stuff is moved out to a domain of its
own. The theoretical approach adopted, however, imposes severe restrictions as to
what can be included in pragmatics, resulting in core topics that can be accounted
for using particular linguistic tools, such as Gricean pragmatics, neo-Gricean
pragmatics, and Relevance Theory. Border-seekers are an Anglo-American tradi-
tion.
In contrast, “problem-solvers” start out from linguistic puzzles that cannot be
solved in terms of grammar, or “the grammar” in a Chomskyan sense. For prob-
lem-solvers the focus is on the identification of a problem for empirical study, and
pragmatics is a matter of adopting a particular perspective on the object of study.
The number or nature of topics to be explored in pragmatic terms is not limited.
Problem-solvers are particularly popular in the European Continental tradition.
The classical areas in the sense of Levinson (1983), the more narrow-scope ap-
proach to pragmatics, illustrate the “code” approach to pragmatics. They tend to
concern aspects of language that have structure-external context “encoded” in
them, or are closely related or “signaled” in relatively systematic or even gram-
maticalized ways by language. The point of departure for these approaches is
basically the individual expression, whereas in problem-solving pragmatics the
point of departure tends to be a certain mode or genre or phenomenon of externally
defined language use, such as code-switching, in which language is embedded and
receives its conversational, social, and interactional meanings.
While narrow-scope, code-based approaches are covered in this handbook, a
majority of the contributions, following contemporaneous interest in the field of
pragmatics, adopt problem-solving approaches. Among these are discourse- and
conversation-analytical approaches, approaches that consider various facets of the
notion of “context” in language use. Some contributions address narrative and
genre analysis, and more purely descriptive approaches to CMC-specific language
phenomena are also represented. Excluded, in contrast, are approaches that have
some connection to pragmatics but primarily belong to another domain, such as
applied linguistics, corpus/computational linguistics, logic/formal semantics, cog-
nitive linguistics/psycholinguistics, and variationist sociolinguistics. Including
them, or trying to do justice to all subfields represented by such a wide approach,
would have meant producing several volumes.
impersonal and efficient short email message. Parts of Hössjer’s (this volume)
findings in a study of professional communication that includes media switches
can be explained by reference to this type of strategic medium choice.
The second line of discussion, not unconnected to the first, is the notion that
CMC is deficient compared to spoken and written language. This view dominated
early research and application in CMC (e.g., Daft and Lengal 1984; Kiesler et al.
1984); it emphasizes the absence of a number of signal types, such as non-verbal,
auditory, olfactory, gustatory, and tactile. “The problem”, Brennan (1998: 1) wrote,
“is that electronic contexts are often impoverished ones”. CMC scholars who
subscribe to this view often focus on “compensational” features such as emoticons,
graphical devices, repetitions, and deletions, as discussed in the contribution by
Bieswanger (this volume), and performative action words such as *waves* (as dis-
cussed by Virtanen in this volume). These can be interpreted as replacing paralin-
guistic and non-verbal cues that are absent from the written repertoire.
An alternative explanation for these common features of CMC – so-called
“Netspeak” features, as described by Crystal (2001) – is that they are playful and
represent the inherently ludic character of language use on the Internet (e.g., Danet
2001; Danet et al. 1997). The written, persistent nature of CMC makes language
more available for metalinguistic reflection than in the case of speech, and this, to-
gether with a tendency towards loose cross-turn relatedness in multiparticipant
CMC, encourages language play (Herring 1999, this volume). According to this
view, CMC is not so much impoverished relative to speech and writing as different
in nature from them.
Medium effects are addressed in most chapters in this collection in one way or
another. For example, Simpson, citing Cherny (1999), notes the easier construction
of multiple floors in synchronous CMC. Harrison and Allton (this volume) note
that email is more economical in that the interaction is accelerated by combining
several conversational moves into a single message (see also Condon and Čech
2010). At the same time, the rapid exchange of messages in multiparty chat inter-
faces that display messages in the order received can result in disrupted turn adja-
cency, loosened norms of relevance, and decreased conversational coherence, as
discussed in the chapter by Herring.
Finally, several of the chapters shed yet another light on the impoverishment
question. A recent development in research on Internet language, manifested in
several contributions in this volume, involves identifying emergent pragmatic
functions of phenomena that were hitherto predominantly considered socio-
linguistic, such as address terms (de Oliveira this volume) and code-switching (An-
droutsopoulos this volume), as well as the analysis of emoticons as “illocutionary
force indicating devices” analogous to punctuation (Dresner and Herring 2010).
These are examples of how existing forms can become endowed with new prag-
matic meanings, a process of sign genesis that countervails ideas of “impoverish-
ment” of Internet language. This approach holds that new expressive needs and
forms arise from and adapt to specific conditions of the new medium. Such new
functions result from the specific “faceting” structure (Herring 2007) – the affor-
dances and the communicative situation – of the medium, together with an en-
hanced metapragmatic awareness arising from the textual nature of most CMC
(Herring 1999; Thurlow and Poff this volume). However, unlike in technological
determinism, these effects are variable rather than categorical, manifesting differ-
ently in different languages and cultures.
The fact that there is an emphasis on local negotiation of norms raises the very
question of the rise and the status of norms in Internet communications. Even at the
level of individual linguistic structures, creativity and ludic deviations from tradi-
tional norms of standard writing in CMC led early on to negative comments about
the (further) decay of language (Thurlow 2006). Norms at the interactional level,
as well, are different from those for traditional genres – less rigid and more open to
development and local definition, at least in the early phases of their formation.
This may change as CMC genres become more entrenched, but rapid technological
development of modes and affordances, a hallmark of the Internet as against other
language media, could militate against solidification of norms.
A corollary of the discussions around technical determinism is the expectation
that technological forces will result in a convergence of practices in interactional
norms, as predicted several decades ago for email by Baron (1984). This point is
taken up in several chapters, including by Thurlow and Poff in their chapter on text
messaging, a mode with a much higher degree of technical constraints on the for-
mal shape of messages than email. Even there, a comparison of research into sev-
eral cultural contexts shows surprising divergence in usage, or at least much less
convergence than one might have expected. There is even a tendency to establish
regiolects (“regiolectal spellings”) in text messaging within a cultural area, such as
the United States, an observation that is supported by recent work by Eisenstein,
O’Connor, Smith, and Xing (2010).
The issue of convergence versus divergence taps into another general theme
that figures prominently among a number of contributions: innovation and devel-
opment. Notions of linguistic change, at least in more traditional versions, have
tended to focus on the development of individual linguistic expressions, a topic of
speculation especially in earlier CMC research (see Herring 2001 for a brief sur-
vey). Linguistic innovation at the micro level is discussed in the chapters in this
collection by Bieswanger (for textual CMC as a whole) and by Dürscheid and
Frehner (for email). Several chapters also consider the emergence of larger units of
discourse such as genres, as discussed above. There is evidence that CMC genre
characteristics are not directly generated by external communicative conditions so
much as by communicants re-articulating pre-existing forms in new media. Gil-
trow (this volume) cites evidence that points to the manifold ancestries of blogs,
and Georgakopoulou notes that some of the features she identifies as characteristic
of Internet stories have been previously described for conversational narratives. At
the same time, pre-existing qualities are differently combined, emphasized, and
articulated, resulting in an end product that is qualitatively different, with the
potential to express new meanings and functions (Herring in press).
A related theme is the extent to which communicants carry over strategies and
practices from oral communication to the new medium, such as described by
Markman (this volume) for topical coherence. Just as oral residue was a feature of
early written language, so oral conversation management strategies appear to per-
sist in coherence management techniques in electronic team meetings via syn-
chronous computer chat.
Finally, chat is another example of variation and change in affordances: A
range of modes all go by the name of chat. Paolillo and Zelenkauskaite (this
volume) note that multiparticipant text chat is increasingly migrating into other
modes, such as online games. In such cases of media convergence, an interesting
point raised by the authors is: Where does game chat get its pragmatic features
from – traditional chat, such as Internet Relay Chat, or its new host context? This
suggests a line of future research that can be carried out as an in vitro study of
change in process: an opportunity afforded to a previously unprecedented extent by
communication on the Internet.
4. Web 2.0
From controversial beginnings, the term Web 2.0 has become associated with a
fairly well-defined set of popular web-based platforms characterized by social in-
teraction and user-generated content. The World Wide Web was pitched as a con-
cept by physicist Tim Berners-Lee to his employers at CERN in 1990, imple-
mented by 1991, and attracted widespread attention after the first graphical
browser, Mosaic, was launched in 1993.7 The early websites of the mid-1990s
tended to be single-authored, fairly static documents, and included personal home-
pages, lists of frequently-asked questions (FAQs), and e-commerce sites. The late
1990s saw a shift towards more dynamic, interactive websites, however, including,
notably, blogs (Herring et al. 2005) and online newssites (Kutz and Herring 2005),
the content of which could be – and often was – updated frequently and which
allowed users to leave comments on the site. These sites foreshadowed what later
came to be called Web 2.0.
The term itself was first used in 2004 when Tim O’Reilly, a web entrepreneur in
California, decided to call a conference he was organizing for “leaders of the In-
ternet Economy [to] gather to debate and determine business strategy” the
“Web 2.0 Conference” (Battelle and O’Reilly 2010; O’Reilly 2005). At the time,
the meaning of the term was vague, more aspirational and inspirational than de-
scriptive. As a business strategy, “Web 2.0” was supposed to involve viral market-
ing rather than advertising and a focus on services over products. One of O’Reilly’s
mantras is “Applications get better the more people use them” (Linden 2006).
Today the term refers, according to Wikipedia (2011b) and other online sources, to
changing trends in, and new uses of, web technology and web design, especially
involving participatory information sharing; user-generated content; an ethic of
collaboration; and use of the web as a social platform. The term may also refer to
the types of sites that manifest such uses, e.g., blogs, wikis, social network sites,
and media-sharing sites.
From the outset, the notion of “Web 2.0” was controversial. Critics claimed
that it was just a marketing buzzword, or perhaps a meme – an idea that was passed
electronically from one Internet user to another –, rather than a true revolution in
web content and use as its proponents claimed. They questioned whether the web
was qualitatively different in recent years than it had been before, and whether the
applications grouped under the label Web 2.0 – including such diverse phenomena
as online auction sites, photo-sharing sites, collaboratively-authored encyclo-
pedias, social bookmarking sites, news aggregators, and microblogs – formed a co-
herent set. Tim Berners-Lee’s answers to these questions was “no” – for the inven-
tor of the web, the term suffered from excessive hype and lack of definition
(Anderson 2006).
In response to such criticisms, O’Reilly (2005) provided a chart to illustrate the
differences between Web 2.0 and what he retroactively labeled “Web 1.0”. This is
shown in modified and simplified form in Figure 1. The phenomena in the second
column are intended to be the Web 2.0 analogs of the phenomena in the first col-
umn.
Web 1.0 Web 2.0
Personal websites Blogging
Publishing Participation
Britannica online Wikipedia
Content management systems Wikis
Stickiness Syndication
Directories (taxonomies) Tagging (folksonomies)
Figure 1. Web 1.0 vs. Web 2.0 phenomena (adapted from O’Reilly 2005)
Language use in Web 2.0 environments raises many issues for pragmatic analy-
sis. There are new types of content to be analyzed: status updates, text annotations
on video, tags on social bookmarking sites, edits on wikis, etc. New contexts must
also be considered – for example, social network sites based on geographic lo-
cation – as well as new (mass media) audiences, including in other languages and
cultures. (Facebook, for example, now exists in “localized” versions in well over
100 languages [Lenihan, 2011].) Web 2.0 manifests new usage patterns, as well,
such as media co-activity, or near-simultaneous multiple activities on a single plat-
form (e.g., Herring, Kutz, Paolillo, and Zelenkauskaite 2009) and multi-author-
ship, or joint discourse production (e.g., Androutsopoulos 2011; Nishimura 2011).
The above reflect, in part, new affordances made available by new communication
technologies: text chat in multiplayer online games (MOGs); collaboratively edi-
table environments such as wikis; friending and the “walled gardens” of Facebook;
social tagging/recommending; and so forth. Last but not least, Web 2.0 discourse
includes user adaptations to circumvent the constraints of Web 2.0 environments:
for example, interactive uses of @ and #, as well as retweeting, on Twitter (e.g.,
boyd, Golder, and Lotan 2010; Honeycutt and Herring 2009) and performed inter-
activity on what are, in essence, monologic blogs (e.g., Peterson, 2011; Pusch-
mann, this volume). Each of these issues deserves attention, and some are starting
to be addressed, but on a case-by-case, rather than a systemic, basis.
Herring (in press) recently proposed a three-part classification of Web 2.0 dis-
course phenomena: phenomena familiar from older computer-mediated modes
such as email, chat, and discussion forums that appear to carry over into Web 2.0
environments with minimal differences; discourse phenomena that adapt to and are
reconfigured by Web 2.0 environments; and new or emergent phenomena that did
not exist – or if they did exist, did not rise to the level of public awareness – prior to
the era of Web 2.0. She argues that this three-way classification can provide insight
into why particular language phenomena persist, adapt, or arise anew in techno-
logically-mediated environments over time. In so doing, she invokes technological
factors such as multimodality and media convergence, social factors at both the
situational and cultural levels, and inherent differences among linguistic phenom-
ena that make them variably sensitive to technological and social effects.
Web 2.0 phenomena in the first category are well represented in this volume.
Examples of familiar Web 2.0 discourse phenomena include non-standard ty-
pography and orthography (Bieswanger), code switching (Androutsopoulos), ad-
dressivity (de Oliveira), flaming (Danet), and email hoaxes/scams (Gill; Heyd), as
well as the continued predominance of text as a channel of communication among
web users, whether it be in blogs, microblogs, wikis, comments on news sites, or
web discussion forums. The latter, in particular, remain very popular, and illustrate
many of the same kinds of phenomena as did asynchronous online forums in the
1990s. Examples of reconfigured Web 2.0 discourse phenomena include personal
status updates (Lee 2011; cf. Cherny 1999), quoting/retweeting (boyd et al. 2010;
cf. Severinson Eklundh and Macdonald 1994), “small stories” (Georgakopoulou,
this volume; cf. Lindholm 2010), and blogging (Myers 2010; Puschmann, this vol-
ume; cf. Herring et al. 2005), which might on the surface appear new but have
traceable online antecedents, as well as reconfigurations of such familiar phenom-
ena as topical coherence, threading, turn-taking, and repair (cf. the chapters in this
volume by Simpson, Markman, Herring, and Jacobs and Garcia).
Less studied (as yet) are what Herring (in press) calls “emergent” Web 2.0 dis-
course phenomena – although since much of what has been claimed to be unprece-
dented on the web has been found, upon deeper examination, to have online and/or
offline antecedents, caution must be exercised in asserting that any phenomenon is
entirely new. Phenomena that can tentatively be identified as emergent and un-
precedented, at least as common practices, include the dynamic collaborative dis-
course that takes place on wikis (e.g., Emigh and Herring 2005; Myers 2010), as
well as conversational video exchanges (Pihlaja 2011), conversational exchanges
via “image texts” (MacDonald 2007), and multimodal conversation more generally.
Aside from the works cited above, however, little language-focused analysis of
these new phenomena has yet been published. A general challenge for emergent
media environments is that they need to be analyzed descriptively first before more
sophisticated, theoretically-informed analyses can be carried out; this often results
in a lag between the emergence of new environments and scholarly understanding
of them.
The coverage of this volume reflects the inherent interdisciplinarity of the prag-
matics of CMC. Contributors approach the topic from the perspectives of classic
pragmatics, discourse pragmatics, interactional linguistics, sociolinguistics, com-
munication, rhetoric, and other (sub)disciplines in the academic study of language.
Methods vary hand-in-hand with approaches, as is expedient in a new field of
study.
In what follows, the individual chapters of this handbook are encapsulated in
terms of the five themes around which they cluster: (1) pragmatic characteristics of
technologically-based CMC modes; (2) classic pragmatic phenomena in CMC en-
vironments; (3) pragmatic innovations emerging from the affordances and prac-
tices of CMC; (4) interactional phenomena in CMC; and (5) broader metaprag-
matic issues, including code choice, narrativity, and genre dynamics.
Another early mode of CMC is the discussion list. Adopting a broad, functional
perspective on pragmatics, Helmut Gruber discusses listserv communication as a
special kind of email communication, contrasting it with other forms of many-to-
many asynchronous computer-mediated discourse. The research reviewed covers
technological affordances, register and language choice, discourse coherence,
genres, and interactive norms and practices, as well as community formation. Ana-
lyses of mailing list communication benefit from models that succeed in captur-
ing and differentiating technical and socio-situational factors and showing causal
relationships between them and the discourse pragmatics of listserv communi-
cation.
Cornelius Puschmann’s concern is with blogging. Blogs differ from other
modes of CMC in terms of the control that the blogger exerts on possible interac-
tion. While a blogging option is made available by various public and semi-public
hosts for professional purposes, blogs are mainly used for personal purposes, in
particular for self-expression in view of a conceptually restricted, familiar audi-
ence. The purposes of blogging, as indicated by the discourse, appear to influence
audience design, style, and content; studies also indicate relations between these
aspects and age and gender. The chapter reviews research on deixis, givenness, ad-
dressivity, and politeness in blog language and presents two prototypical blogging
styles: topic-centric and author-centric.
John Paolillo and Asta Zelenkauskaite survey research on real-time chat.
They discuss the social nature of technologically-differentiated chat modes such as
MUDs/MOOs, AOL, IRC, and web-based chat, and identify four major domains of
applications: recreational, educational, institutional, and business. The chapter
describes great variation across the affordances of various chat modes and appli-
cations of chat, as well as across users and their communicative and social moti-
vations, all of which are likely to be reflected in the linguistic landscapes of chat.
The pragmatic phenomena in focus include interaction management, international
and intercultural contexts, orthography, and participant gender.
Naomi Baron surveys the history and evolution of Instant Messaging (IM).
She points out that “the questions of language, context, and interpretation that IM
initially generated are no less relevant today to students of CMC” (155). Such
questions include user profile creation, as well as investigations of contexts and
ways of use. Particular attention is paid to IM “away” messages, which reveal prag-
matic information; the structuring of IM conversations; and gendered patterns of
use. The chapter also reports user attitudes to IM as compared to Facebook and tex-
ting, and concludes by singling out a number of central pragmatic phenomena in
need of further investigation.
Crispin Thurlow and Michele Poff approach text messaging (texting, SMS)
from the perspectives of cross-cultural, interactional, pragmalinguistic, and meta-
linguistic contexts. Texting “presents itself in the broadest terms as a social tech-
nology par excellence” (174), predominantly mediating “phatic communion”
persons familiar to them and in anonymous online spaces. The chapter also deals
with sets of linguistic features that are expected to reveal deception, in keeping
with a shift in the study of deception from non-verbal cues to verbal elements. It is
crucial to consider degrees of recordability, synchronicity, and co-presence of in-
terlocutors in the naturalistic settings offered by CMC for the study of the prag-
matic aspects of deceptive discourse.
presents a case study of virtual student team meetings that reveals an orien-
tation towards discourse topics as a means of designing and structuring turns on-
line. Interactional coherence is examined in the light of conversational threads as
well as speakership roles and turn allocation, while also paying attention to the
size of the group, the technical features of the chat environment, and the purpose
of the talk. The chapter argues for the difference between turns and chat mess-
ages, and presents a useful transcription method designed for the analysis of chat
interaction.
Jennifer Jacobs and Angela Garcia examine another central concept in CA,
repair. Their study of educational chat in an intranet context of composition classes
in the U.S. makes use of an innovative transcription system based on videotapes of
participants’ computer screens to allow the analyst to keep track of what is visible
to each student at a given point in the interaction and what keystroke actions they
make in their postings. The case study focuses on avoidance and self/other repair
of troubles in “quasi-synchronous” chat, as well as the technological solutions
adopted by users for these purposes, such as short or split messages, delays in post-
ing, addressivity, and format tying.
Karianne Skovholt and Jan Svennevig investigate norms related to responses
and non-responses in workplace email in Norwegian contexts. Informed by CA
and classic speech act theory, the study relies on participant display of communi-
cative actions. While a response is found to be conditionally relevant after ques-
tions and requests, non-response to requests for comments and corrections to a pro-
posal signals acceptance of the proposal. However, users may volunteer responses
that have an interpersonal function. The emerging norms of responding or not in
workplace email are compared to earlier studies of other kinds of CMC dialogue
and to the literature on offline conversations.
Focusing on email use in Swedish workplace communication, Amelie
Hössjer addresses the relationship between small talk and politeness in two edi-
torial contexts: a physically-based and a digitally-based community of practice. In
the former, email is one of several communication options, and its use is character-
ized by simple, routine matters. Small talk primarily appears in face-to-face con-
tacts, reserved for complex or delicate issues. While not a “lean medium” in itself,
email is used as such. In contrast, the digitally-based community of practice
makes use of email as its primary form of communication. For these users it is a
“rich medium”, allowing them to construct and maintain social relationships
through the discourse strategies of “framing” and “chaining” in work-related inter-
action.
The late Brenda Danet’s concern, in contrast, is with linguistic impoliteness
and rudeness. The CMC phenomenon of “flaming” is provided with a historical
context and related to (im)politeness theories, as well as issues of multilingualism,
gender, and culture. The chapter shows that flaming can sometimes be used to ex-
press solidarity, perhaps jointly with hostility. A contextualized case study is pres-
The contents of the volume can be summarized concisely as follows. Part 1 shows
important variation across CMC modes, Part 2 forcefully demonstrates how
analyses of classic pragmatic phenomena in CMC data suggest developments to
pragmatic theory, and Part 3 identifies a small number of unevenly distributed
pragmatic phenomena that may be labeled CMC-specific, even though they, too,
can be shown to have roots in offline communication. Part 4 raises the issue of the
applicability to CMC data of models devised for the analysis of spontaneous face-
to-face communication, and Part 5 addresses the broader metapragmatic issues of
code alternation and genre in CMC.
Notes
1. Noteworthy among these are Baron (1984), Murray (1985), and Severinson Eklundh
(1986).
2. For a useful (albeit now somewhat dated) overview of sociolinguistic research on CMC,
see Androutsopoulos (2006).
3. See, e.g., Reid (1991) and Werry (1996) for IRC; Reid (1994) and Cherny (1999) for
MUDs; Baron (1998) and Gains (1999) for email; Döring (2002) for German SMS; Thur-
low and Brown (2003) for English SMS; Stein (2006) for business websites; and Herring
et al. (2005) for blogs.
4. See also Weininger and Shield (2004) and Stein (2005) with reference to Koch and Ös-
terreicher (1994); Herring (2007) with reference to Hymes (1974).
5. In keeping with the rise in English as the lingua franca of scholarship, CMC scholarship
from other countries published in English is also on the rise, especially on the web (Cal-
lahan and Herring 2012).
6. A notable, and early, exception is Yus (2001). An English version of this work became
available in 2010.
7. On the creation of the World Wide Web, see Wikipedia (2011a).
8. In the literature on computer-supported cooperative work, however, deixis has been ad-
dressed in just such environments, e.g., by Hindmarsh and Heath (2000) and Suthers, Gir-
ardeau, and Hundhausen (2003), and grounding has been addressed, e.g., by Clark and
Brennan (1991) and Brennan (1998).
References
Anderson, Nate
2006 Tim Berners-Lee on Web 2.0: “[N]obody even knows what it means”. Ars
Technica website. http://arstechnica.com/business/news/2006/09/7650.ars [ac-
cessed December 3, 2011]
Androutsopoulos, Jannis K.
2006 Introduction: Sociolinguistics and computer-mediated communication. Jour-
nal of Sociolinguistics 10(4): 419–438.
Androutsopoulos, Jannis K.
2011 From variation to heteroglossia in the study of computer-mediated discourse.
In: Crispin Thurlow and Kristine Mroczek (eds.), Digital Discourse: Lan-
guage in the New Media, 277–298. New York: Oxford University Press.
Androutsopoulos, Jannis K. and Volker Hinnenkamp
2001 Code-switching in der bilingualen Chat-Kommunikation: ein explorativer
Blick auf #hellas und #turks. In: Michael Beißwenger (ed.), Chat-Kommuni-
kation. Sprache, Interaktion, Sozialität & Identität in synchroner computerver-
mittelter Kommunikation: Perspektiven auf ein interdisziplinäres Forschungs-
feld, 367–401. Stuttgart: Ibidem.
Androutsopoulos, Jannis K. and Evelyn Ziegler
2003 Sprachvariation und Internet: Regionalismen in einer Chat-Gemeinschaft.
In: Jannis K. Androutsopoulos and Evelyn Ziegler (eds.), ‘Standardfragen’:
Soziolinguistische Perspektiven auf Sprachgeschichte, Sprachkontakt und
Sprachvariation, 251–280. Frankfurt am Main: Peter Lang.
Ariel, Mira
2010 Defining Pragmatics. Cambridge, UK: Cambridge University Press.
Baron, Naomi S.
1984 Computer mediated communication as a force in language change. Visible
Language 18(2): 118–141.
Baron, Naomi S.
1998 Letters by phone or speech by other means: The linguistics of email. Language
and Communication 18: 133–170.
Battelle, John and Tim O’Reilly
2010 Web 2.0 Summit is a wrap. Web 2.0 Summit website.
http://www.web2summit.com/web2010 [accessed December 3, 2011]
Biber, Douglas
1988 Variation across Speech and Writing. Cambridge, UK: Cambridge University
Press.
boyd, danah, Scott Golder, and Gilad Lotan
2010 Tweet tweet retweet: Conversational aspects of retweeting on Twitter. In: Pro-
ceedings of the Forty-Third Hawai’i International Conference on System
Sciences. Los Alamitos, CA: IEEE Press.
http://csdl.computer.org/dl/proceedings/hicss/2010/3869/00/03-06-04.pdf
Brennan, Susan
1998 The grounding problem in conversation with and through computers. In: Susan
R. Fussell and Roger J. Kreuz (eds.), Social and Cognitive Psychological Ap-
proaches to Interpersonal Communication, 201–225. Hillsdale, NJ: Lawrence
Erlbaum.
Callahan, Ewa and Susan C. Herring
2012 Language choice on university websites: Longitudinal trends. International
Journal of Communication 6: 322–355.
Cherny, Lynn
1999 Conversation and Community: Chat in a Virtual World. Stanford: CSLI.
Herring, Susan C., Daniel O. Kutz, John C. Paolillo, and Asta Zelenkauskaite
2009 Fast talking, fast shooting: Text chat in an online first-person game. In: Pro-
ceedings of the Forty-Second Hawai’i International Conference on System
Sciences. Los Alamitos, CA: IEEE Press.
http://csdl.computer.org/dl/proceedings/hicss/2009/3450/00/03-05-04.pdf
Herring, Susan C., Lois Ann Scheidt, Sabrina Bonus, and Elijah Wright
2005 Weblogs as a bridging genre. Information, Technology & People 18(2):
142–171.
Herring, Susan C. and Asta Zelenkauskaite
2009 Symbolic capital in a virtual heterosexual market: Abbreviation and insertion
in Italian iTV SMS. Written Communication 26(1): 5–31.
Hindmarsh, Jon and Christian Heath
2000 Embodied reference: A study of deixis in workplace interaction. Journal of
Pragmatics 32(12): 1855–1878.
Honeycutt, Courtenay and Susan C. Herring
2009 Beyond microblogging: Conversation and collaboration via Twitter. In: Pro-
ceedings of the Forty-Second Hawai’i International Conference on System
Sciences. Los Alamitos, CA: IEEE Press.
http://csdl.computer.org/dl/proceedings/hicss/2009/3450/00/03-05-05.pdf
Hymes, Dell
1974 Foundations in Sociolinguistics: An Ethnographic Approach. Philadelphia:
University of Pennsylvania Press.
Jucker, Andreas H.
2010 “Audacious, brilliant!! What a strike!” Live text commentaries on the Internet
as real-time narratives. In: Christian Hoffmann (ed.), Narrative Revisited: Tell-
ing a Story in the Age of New Media, 57–77. Amsterdam/ Philadelphia: John
Benjamins.
Jucker, Andreas H. and Christa Dürscheid
in press The linguistics of keyboard-to-screen communication: A new terminological
framework. Linguistik Online.
Kiesler, Sara, Jane Siegel, and Timothy McGuire
1984 Social psychological aspects of computer-mediated communication. American
Psychologist 39: 1123–1134.
Kim, M-S. and Narayanan Raja
1991 Verbal aggression and self-disclosure on computer bulletin boards. Paper pres-
ented at the Annual Meeting of the International Communication Association,
Chicago, IL, May. ED334620
Koch, Peter and Wulf Österreicher
1994 Schriftlichkeit und Sprache. In: Hartmut Günther and Otto Ludwig (eds.),
Schrift und Schriftlichkeit. Ein interdisziplinäres Handbuch internationaler
Forschung, HSK 10.1, 587–604. Berlin/New York: Walter de Gruyter.
Kutz, Daniel O. and Susan C. Herring
2005 Micro-longitudinal analysis of Web news updates. In: Proceedings of the Thir-
ty-Eighth Hawai’i International Conference on System Sciences. Los Alami-
tos, CA: IEEE Press.
http://csdl.computer.org/dl/proceedings/hicss/2005/2268/04/22680102a.pdf
Lee, Carmen
2011 Texts and practices of micro-blogging: Status updates on Facebook. In: Cris-
pin Thurlow and Kristine Mroczek (eds.), Digital Discourse: Language in the
New Media, 110–128. New York: Oxford University Press.
Lenihan, Aoife
2011 “Join our community of translators”: Language ideologies and Facebook. In:
Crispin Thurlow and Kristine Mroczek (eds.), Digital Discourse: Language in
the New Media, 48–64. New York: Oxford University Press.
Levinson, Stephen
1983 Pragmatics. Cambridge, UK: Cambridge University Press.
Linden, Greg
2006 Tim O’Reilly and defining Web 2.0. Geeking with Greg [blog], May 14.
http://glinden.blogspot.com/2006/05/tim-oreilly-and-defining-web-20.html
Lindholm, Loukia
2010 “A little story, for food for thought …”: Narratives in advice discourse. In
Sanna-Kaisa Tanskanen, Marja-Liisa Helasvuo, Marjut Johansson, and Mia
Raitaniemi (eds.), Discourses in Interaction, 223–236. Amsterdam/Philadel-
phia: John Benjamins.
Ling, Richard and Naomi S. Baron
2007 Text messaging and IM: Linguistic comparison of American college data.
Journal of Language and Social Psychology 26: 291–98.
Malinowski, Bronisław
1923 The problem of meaning in primitive languages. In: C. K. Ogden and I. A. Ri-
chards (eds.), The Meaning of Meaning, 296–336. London: Routledge &
Kegan Paul.
Maynor, Natalie
1994 The language of electronic mail: Written speech? In: Greta D. Little and Mi-
chael Montgomery (eds.), Centennial Usage Studies, 48–54. Tuscaloosa, Uni-
versity of Alabama.
McDonald, David
2007 Visual conversation styles in web communities. In: Proceedings of the Fortieth
Hawaii International Conference on System Sciences. Los Alamitos, CA:
IEEE Press.
http://csdl.computer.org/dl/proceedings/hicss/2007/2755/00/27550076a.pdf
Murray, Denise E.
1985 Composition as conversation: The computer terminal as medium of communi-
cation. In: Lee Odell and Dixie Goswami (eds.), Writing in Nonacademic Set-
tings, 205–229. New York: Guilford.
Murray, Denise E.
1990 CmC. English Today 6: 42–26.
Myers, Greg
2010 The Discourse of Blogs and Wikis. London: Continuum.
Nishimura, Yukiko
2011 Japanese keitai novels and ideologies of literacy. In: Crispin Thurlow and Kris-
tine Mroczek (eds.), Digital Discourse: Language in the New Media, 86–109.
New York: Oxford University Press.
O’Reilly, Tim
2005 What is Web 2.0? Design patterns and business models for the next generation
of software. O’Reilly Network website, September 30.
http://oreilly.com/web2/archive/what-is-web-20.html
Paolillo, John C.
1996 Language choice on soc.culture.punjab. Electronic Journal of Communication
6(3). http://www.cios.org/EJCPUBLIC/006/3/006312.HTML
Paolillo, John C.
1999 The virtual speech community: Social network and language variation on IRC.
Journal of Computer-Mediated Communication 4(4).
http://jcmc.indiana.edu/vol4/issue4/paolillo.html
Peterson, Eric
2011 How conversational are weblogs? Language@Internet 8, article 8.
http://www.languageatinternet.org/articles/2011/Peterson
Pihlaja, Stephen
2011 Cops, popes, and garbage collectors: Metaphor and antagonism in an atheist/
Christian YouTube video thread. Language@Internet 8, article 1.
http://www.languageatinternet.org/articles/2011/Pihlaja
Reid, Elizabeth M.
1991 Electropolis: Communication and community on Internet Relay Chat. Senior
Honours thesis, University of Melbourne, Australia.
http://www.ee.mu.oz.au/papers/emr/index.html
Reid, Elizabeth M.
1994 Cultural formations in text-based virtual realities. Master’s thesis, University
of Melbourne, Australia. http://www.ee.mu.oz.au/papers/emr/index.html
Rintel, Sean and Jeffery Pittam
1997 Strangers in a strange land: Interaction management on Internet Relay Chat.
Human Communication Research 23(4): 507–534.
Severinson Eklundh, Kerstin
1986 Dialogue Processes in Computer-Mediated Communication: A Study of
Letters in the COM System. Linköping: Linköping University, Sweden.
Severinson Eklundh, Kerstin and Clare Macdonald
1994 The use of quoting to preserve context in electronic mail dialogues. IEEE
Transactions on Professional Communication 37(4): 97–202.
Siebenhaar, Beat
2003 Sprachgeographische Aspekte der Morphologie und Verschriftung in schwei-
zerdeutschen Chats. Linguistik Online 15: 125–139.
Sperber, Dan and Deirdre Wilson
1986 Relevance: Communication and Cognition. Oxford: Blackwell.
Stein, Dieter
2005 Distanz und Nähe in Internetkommunikation. In: Angelika Storrer and Michael
Beißwenger (eds.), Chat-Kommunikation in Beruf, Bildung und Medien: Kon-
zepte-Werkzeuge-Anwendungsfelder, 339–348. Stuttgart: Ibidem.
Stein, Dieter
2006 The website as a domain-specific genre. Language@Internet 3, article 6.
http://www.languageatinternet.org/articles/2006/374
1. Introduction
With millions of emails sent every day, email has become a common mode of com-
munication that is used by young and old. To have an email address – or even sev-
eral – is something that is taken for granted; email has become as natural a com-
munication channel as the telephone. Thus, one may assume that most readers are
familiar with sending emails and with their structural elements, which include the
header, the body, and an optional signature. Compared to other modes of computer-
mediated communication (henceforth CMC), email is considered an old mode that
is gradually losing popularity as new, competing modes have sprung up. Neverthe-
less, email is still the most important CMC application because it is the only one
with which the average Internet user is familiar. Thus, when filling in forms, for in-
stance, it is the person’s email address that is asked for and not their Skype user ID
or Twitter account – a practice that is likely to persist for a while. Emails are used
for various purposes, e.g., to exchange information, to submit greetings and invi-
tations, or to send an Internet link or some digital data (ranging from simple word
documents to photos and videos). Being less personal than a letter, it is a relatively
unobtrusive form of communication and also encourages people who would not
send letters otherwise to communicate in writing. What is more, its swift trans-
mission makes it a preferred medium to communicate with people who are far
away. While a letter would take much longer to be delivered, an email reaches its
recipients in no time, no matter where they are located.
The history of email dates back to the 1970s (see Baron 2000: 223–226). Until
the late 1980s, email was used primarily in governmental, business, and computer
science circles; subsequently, it gained widespread popularity with the rise of In-
ternet Service Providers such as CompuServe and AOL. It has replaced telephone
calls and letter mails to a certain extent and has further created new communication
niches: People send each other emails in situations in which they would not have
addressed each other earlier on, and so it has become much easier to approach an-
other person when needing assistance. Teachers, for instance, frequently receive
emails from their students who ask questions about homework assignments, up-
coming exams, or personal matters, inquiring about problems for which they
would have found a solution without the teacher’s help in earlier times. In this con-
text, Baron (2003: 86 and 2008: 165) reports on a student who asked for further in-
formation on his research project while he was preparing his master’s thesis: “Ap-
parently his library did not have many useful sources. After presenting me a long
list of questions, he closed with ‘OK NAOMI … I really need your information as
soon as possible’”. This example clearly illustrates that the inhibition threshold to
write somebody has been lowered, but it also points to another phenomenon that
seems to be typical of email, namely the tendency to use a rather informal style (see
section 3).
Whereas previous studies have often discussed issues dealing with the question
of why people use email, this very question must be modified to inquire about the
reasons behind the fact that some people still send letters, while in other situations
they prefer email to letters. Indeed, there are some advantages of a letter over
email: A letter can transport an object; the receiver might pay more attention to a
letter than to an email; a letter can also be received by people who do not use the In-
ternet. How, then, is it that people often choose email to make an appointment or
discuss problems? Why would they not phone in such situations? An email does
not disturb the recipients at work and permits the senders to think over their words
and modify their sentences. Thus, it is not astonishing that people who work in the
same office, for instance, frequently communicate via email. Another advantage is
that “[email] enables people to communicate speedily the same information to
many others”, as Waldvogel (2007: n.p.) points out. The interactants even get a rec-
ord of the communication, which is not the case when discussing a topic over the
phone.1 In fact, email has become so popular within the business sector that people
often spend hours reading and answering emails, with the result that some em-
ployers have started to reduce the time their employees are allowed to spend with
email in order to save working time.
In section 2, we briefly sketch the history of email research from the 1980s until
the present. In section 3, the features generally claimed to be typical of email are
listed and critically discussed. We argue that there is no reason to assume a lan-
guage that is unique to email and show that it is rather inaccurate to use terms such
as “Netspeak”. Nevertheless, there are new communication practices when it
comes to emailing – practices that influence people’s social and communicative
behaviour. This particularly applies to email dialogues, which can be regarded as a
typical characteristic of CMC. In this context, we also refer to the numerous style
guides on the composition of emails. Section 4 focuses on theoretical approaches
to CMC in general and email communication in particular: On the one hand, we
describe a model which is well known in the work of German and Romance
scholars, namely the orality-literacy model of Peter Koch and Wulf Oesterreicher
(1994). On the other hand, we refer to Susan Herring’s (2001, 2007) discourse
analysis approach to CMC. The last section provides an outlook on the future of
email; we advance the thesis that email will persist in certain areas only, while it
will be replaced by more synchronous means of (online) communication in many
others.
As Stein (2003) points out, many linguistic and communication studies of Internet
language in German are available these days. Accordingly, it is important to focus
not merely on linguistic studies that have been carried out in the Anglophone area
but also on available German research material. The following overview therefore
provides both an overview of English studies and a description of the major Ger-
man works.2
Even though linguistic research on CMC in the Anglophone area dates back
to the beginning of the 1980s, it was only around 10 years later that it received
serious attention by linguists and scholars, and so “CMC was still a novelty topic
of research in the early 1990s” (Herring 2008: xxxv). The same is true for the
German-speaking research community. The earliest CMC studies that came to be
known by a broader public were published in the late 1990s; among these are
Weingarten’s volume Sprachwandel durch Computer (‘Computer-Driven Lan-
guage Change’) (1997) and Runkehl, Schlobinski, and Siever’s frequently cited
book Sprache und Kommunikation im Internet (‘Language and Communication on
the Internet’) (1998). To start with the Anglophone linguistic email research, three
major names must be mentioned in this context, these being Naomi Baron, Susan
Herring, and David Crystal.
a) Naomi Baron published an essay on “Computer mediated communication
as a force in language change” as early as 1984, and she studied the linguistics of
email more closely in her article “Letters by phone or speech by other means: The
linguistics of email”, which appeared in 1998. In the latter article, she also dis-
cussed the difference between writing and speech and considered email against the
background of this difference, concluding that while some features “distribute
themselves neatly on the dichotomous writing/speech spectrum”, others “have
mixed profiles” (Baron 1998: 155). Two years later, Baron came out with a rather
comprehensive study on the history of writing technologies, a book called Alpha-
bet to Email: How Written English Evolved and Where It’s Heading (2000). Always
On (2008) is another important study worth noting in this context. The question
Baron approaches in both books is how technology is changing the way we write.
She claims that “[i]n the fast-moving world of email, content is far more important
than spelling and punctuation” and finds that “the line between the spoken and
written language continues to fade” (Baron 2000: 259). Similarly, in her article
“Why email looks like speech. Proofreading, pedagogy and public face”, which
was published three years later, Baron (2003: 92) argues that “[e]mail resembles
speech”, thereby, however, ignoring the fact that there is actually a wide array of
text types, ranging from informal to formal, as well as a great variety of situational
factors that have to be considered, both of which have a significant influence on
email style. Baron (2003: 92) then points out that “writing in general has become
more speech-like, thanks in part to conscious pedagogical decisions and in part to
guistic features and also defines the uniqueness of email. Crystal has become well
known for coining the term “Netspeak” – a term which appears in many of his
works, such as in the second edition of his volume The English Language: A
Guided Tour of the Language (1988, 2002). This term is critically discussed in
Dürscheid (2004), who finds that “Netspeak” as such does not exist. She argues
that the different kinds of text types written online make it impossible to draw any
generalising conclusions by subsuming the various linguistic features under a
single terminus technicus. Another book which deals with the topic is entitled The
Language Revolution, in which Crystal (2004: 64) points out that “[t]he public ac-
quisition of the Internet was [besides the emergence of a global language and the
phenomenon of language endangerment] the third element contributing to the rev-
olutionary linguistic character of the 1990s, and the one where the epithet ‘revol-
utionary’ is easiest to justify”. It is worth mentioning that Crystal (2001: 52–59)
was one of the first authors to analyse CMC in terms of Grice’s conversational
maxims (see also the chapters by Herring, Heyd, and Lindholm in this volume). He
argues, for instance, that Grice’s maxim of manner (“avoid obscurity of ex-
pression”) may be violated in CMC: “[t]yping, not a natural behaviour, imposes a
strong pressure on the sender to be selective in what is said […]. And selectivity
in expression must lead to all kinds of inclarity” (Crystal 2001: 57–58). Indeed,
it is interesting to apply the maxims to CMC in the way Crystal has done. Accord-
ing to Stein (2003: 160), this part is “one of the strongest and […] most innovative
facets of the book”.
There have been a number of other research articles, among which “Writing in
cyberspace: A study of the uses, style and content of email” by Helen Petrie (1999)
must particularly be mentioned. This study is one of the most extensive email
studies quantitatively, comprising analysis of 38,000 emails. Further important
works on the linguistics of email are Natalie Maynor’s article “The language of
electronic mail: Written speech?” (1994), Judith Yaross Lee’s article “Charting the
codes of cyberspace: A rhetoric of electronic mail” (2003), and Carmen Frehner’s
volume Email – SMS – MMS. The Linguistic Creativity of Asynchronous Discourse
in the New Media Age (2008). Frehner, like Crystal (2001), refers to Grice’s con-
versation maxims in her chapter about “netiquette” (2008: 41–43).
German research on email was launched by two major works, namely Janich’s
(1994) “Electronic Mail – eine betriebsinterne Kommunikationsform” (‘Electronic
mail – a company internal communication form’) and Günther and Wyss’s (1996)
“E-Mail-Briefe – eine neue Textsorte zwischen Mündlichkeit und Schriftlichkeit”
(‘Email letters – a new text type between orality and literacy’). Janich analysed
emails from within the business sector and focused on the characteristic features of
email correspondence in offices, while Günther and Wyss also paid attention to pri-
vate email communication. As the title of the latter work suggests, the authors as-
sume that email corresponds to one single text type. This might have been true at
the time the article was drafted; however, email has long since become multifunc-
Identifying language use that is typical for email communication proves to be dif-
ficult “simply because the vast diversity of settings and purposes of [its] use out-
weigh any common linguistic features” (Androutsopoulos 2006: 420). Nonethe-
less, the following sections attempt to outline the nature of email and its
characteristics. Section 3.1 discusses the terms “emailism” and “Netspeak”. With
structural features at its core, section 3.2 focuses on the graphical, lexical, and
grammatical levels of emails, while the pragmatic features in terms of email dia-
logues and the relation between interactants is addressed in section 3.3.
can be repeated or capitalised, and whole words can be reduplicated or else be put
between asterisks or underlined spaces. The creative use of punctuation may also
add emphasis or indicate loudness, silence, rising intonation, or even emotion.
Another typical email feature is the so-called thread. In email, a thread is a
technical feature, being automatically generated by the mailing programme once
the user replies to a received message by pressing the reply button.4 Nevertheless,
such a thread still has a certain influence on the pragmatic level of emailing (see
section 3.3). It consists of all the sent and received messages on a topic that was
named and inserted into the subject line by the sender of the initial message. It is
thus the subject line which creates a tie among these messages and establishes
some text-external coherence. Having older messages included in one’s answer
allows the respondent to refer to some previously mentioned issue by, for instance,
some anaphoric pronoun only (e.g., Oooh D, that’s not on our agenda for a while
you know, worth trying though […]). In this way, a quasi-dialogue is performed (cf.
Severinson Eklundh 2010), and for this reason, it is also justified to consider email
as a target of discourse analysis.
Apart from this, there are qualitatively no new features that were not familiar to
us before the advent of email. What seems to have changed is their quantity. While
each of the linguistic characteristics described above had been present before email
became popular, there are now more of these elements at once. Smiling faces, for
instance, have been around since 1963 when Harvey Ball, a graphic artist, invented
the yellow smiley face upon being given the task by the State Mutual Life Assur-
ance Company “to design a logo that would uplift its employees after a company
merger had hurt company morale […]. Thinking about what would inspire em-
ployees to smile, Harvey decided the most simple and direct symbol would be a
smile itself and that is what he drew” (Cates 2003: n.p.). Abbreviations have also
been around for many decades, and emulated prosody is very similar to what can
be found in comics, where the same or similar features exist. As for the lack of con-
ventional punctuation marks, even this feature is not new: Telegrams, for instance,
used to have an unusual typeface, often lacking any punctuation marks. What com-
puter language has certainly encouraged is the great variety of abbreviations – just
as it has led to a huge variety of smileys – but in the end, the way in which words
are abbreviated is not new from a qualitive perspective; quite on the contrary, it is a
quantitative change.
commonly believed. Frehner’s empirical study (2008) reveals that in a corpus con-
sisting of 342 emails, there were only 3.16 occurrences of smileys per 1,000 words,
whereas in a corpus of 983 single (fewer than 160 signs) and linked (more than 160
signs) text messages (SMS), emoticons appear 4.88 and 3.98 times per 1,000
words, respectively. No studies so far have explained why emoticons are not so fre-
quent in emails, especially since they are undoubtedly native to CMC. Androutso-
poulos (2006: 425) reports on “emoticons being more often used by females in
Witmer and Katzman (1997), and by teenage males in Huffaker and Calvert (2005)”.
Beutner (2002: 78), in contrast, assumes that it is mainly newbies (i.e., inexperienced
users) who make use of emoticons. This would imply that newbies (i.e., inexperi-
enced users) first overuse emoticons and then, the more accustomed they get to the
communication mode, the less they use them to modify their statements.
At this point, it is worthwhile to take a closer look at the “etiquette” of email, be-
cause it reveals characteristic pragmatic features compared to the etiquette of other
CMC modes, on the one hand, and to that of offline (written) communication, on the
other hand. We do not imagine that people are actually aware of these guidelines, but
their regulations nevertheless allow for conclusions about everyday email use. If
email etiquette tells us, for instance, to avoid certain elements (e.g., the use of words
spelled in all capitals), we can deduce that these have already become a feature of
emails. Another common piece of advice is to not send an email when feeling emo-
tional. As emails are easily composed and dispatched, email writers are likely to
send messages written in the heat of the moment – messages that would never have
been written or sent if they had to be taken to the post office.6 The first set of guide-
lines, titled “Towards an ethics and etiquette for electronic mail”, dates from 1985
and is still available. The authors, Norman Shapiro and Robert Anderson, based
their recommendations on personal observations of inappropriate and counterpro-
ductive uses of email. Since then, numerous guidelines have been composed dealing
with email use. They can be found on the Internet, codified in FAQ documents or
Netiquette guidelines, but also in print newspaper articles and in print books on
Netiquette. Indirectly, these guidelines postulate the cooperative principle and
Grice’s (1975) four conversational maxims: the maxims of quantity, quality, rel-
evance, and manner. Some guidelines recommend, for instance, not sending emails
with large attachments – a rule that concerns the maxim of quantity (“Do not make
your information more informative than is required”). However, there is a differ-
ence between the theoretical status of Grice’s conversational maxims and the status
of the Netiquette guidelines: The maxims describe presumptions about utterances,
i.e., they express the ideal ways in which cooperative interactants should communi-
cate. The guidelines, in contrast, have a strong prescriptive character.
The following instruction concerning the feature of quoting in emails is based
on the maxims of relevance and quantity (e.g., say things related to the current
topic of the conversation; do not say more than is needed): “Only quote the needed
parts (deleting the remainder) and reply directly after the item you wish to respond
to” (Johns 1996: n.p.). The maxim of manner is reflected in the following advice:
“Please remember that you are sending a text-based communication to possible
strangers. They may not know your sarcasm or witty sense of humor like your
family and close friends do” (Johns 1996: n.p.). Other recommendations concern
non-verbal behaviour in CMC. For example, people are advised not to forward an
email without the agreement of the sender, send a carbon copy to several recipients
in which everybody’s address is indicated, or send a large attachment file without
announcing it beforehand.7 These rules clearly demonstrate that there are socio-
cultural factors associated with interaction in CMC that have to be taken into con-
sideration.
4. Theoretical approaches
This section focuses on two approaches that have already been mentioned, these
being Koch and Oesterreicher’s (1994) orality-literacy model,8 which was devel-
oped in the 1980s before the rise of email communication, and Herring’s (2001,
2007) discourse approach to CMC. Koch and Oesterreicher distinguish between
medium and conception, explaining that while a piece of writing can be written
from a medial perspective, the same piece of writing can be oral from a conceptual
perspective and vice versa. While there is a dichotomy within the medial dimen-
sion (language is either phonic or graphic), there is a continuum within the concep-
tual dimension (see also Biber 1988). The two poles of the continuum are called
“language of immediacy” (which is conceptually oral) and “language of distance”
(which is conceptually written). The language of immediacy is typically associated
with private settings (see Landert and Jucker 2011) in which there is a high degree
of familiarity as well as a lack of emotional distance between the interactants; it is
further set in a dialogic situation and characterised by unplanned discourse, while
the opposite is true of the language of distance. Note that labelling these poles as
“oral” and “written” may be misleading even if they are qualified by the attribute
“conceptually”. One must always keep in mind that this dimension is logically in-
dependent of the medial dimension, although there might be a prototypical corre-
spondence between the two. Figure 1 illustrates the model.
In the prototypical case, there is a correspondence between the choice of lin-
guistic features and the medial dimension. This means that written language is
more typically used in situations of distance and when a formal style is required. A
legal text, for instance, is classified as written both from a conceptual and a medial
perspective (position A), while private talk within one’s family or among friends is
spoken and tends to use an informal style (position B). At the same time, some cor-
respondences do not follow this regularity and must be located somewhere be-
tween the conceptual and the medial dimension. Accordingly, a church sermon is
spoken, but tends to be stylistically formal, whereas a greeting postcard is written,
Figure 1. Koch and Oesterreicher’s model; A= legal text, B = private talk, C= private
email
but is of a more colloquial style. For this reason, the model is highly interesting for
CMC research: In CMC, we are often faced with texts that do not meet our expec-
tations concerning the relationship between medium and conception. Private
emails, for instance, are written from the medial point of view, but may be situated
next to the immediacy pole from the conceptual point of view (position C).
It is impossible to generalise about where email communication is located
along the conceptual continuum due to the fact that a wide range of text types is
realised in emails, each of them being associated with its own characteristic lin-
guistic features. A business email, for instance, is usually less conceptually oral
than a private email to a friend. This means that the model is suitable for the clas-
sification of email (or other) text types, but not for the classification of communi-
cation modes as a whole. However, the approach provides a precise terminology
for CMC research, enabling researchers to describe a message’s closeness to
spoken or written language. When Baron (2000: 258) states that “[t]he line be-
tween spoken and written language continues to fade in America”, she is referring
to the conceptual dimension only. Within the medial dimension, the line cannot
fade – despite the fact that on a screen, for instance, phonic and graphic signs may
be combined. Email is not speech; email is exclusively text-based. There may be
oral features, but these features are situated on the conceptual level and not on the
medial one.
All in all, Koch and Oesterreicher’s model is a suitable approach to situate
written interactions such as business and private email messages along the con-
tinuum of communicative immediacy and communicative distance. However, it
does not offer a tool to analyse the context in which the interactions are embedded.
To analyse dialogical situations, it is useful to consider the computer-mediated dis-
course analysis approach (CMDA) presented by Herring (2001). Research studies
that are situated within this frame focus on communication purposes, situational
factors (such as one-to-one, one-to-many, or many-to-many communication), and
the role of demographics (e.g., social class, race, and ethnicity). Indeed, it does
make a difference whether an email is sent in a one-to-many-communication sys-
tem (e.g., within a newsgroup) or in a one-to-one-communication system, within a
business (formal) or private (informal) context. All these factors must be taken into
consideration. Herring (2007) presented a classification scheme that brings to-
gether the relevant aspects of the technical and the social contexts that influence
discourse usage in CMC (see section 2). In this article, she clearly points out that
two main dimensions, medium and situation, jointly influence language use.
Among the medium dimensions, one factor that conditions CMC is one-way or
two-way message transmission, and another is synchronicity of participation (Her-
ring 2007). Concerning the latter, it is useful to check the degree of (a-)synchro-
nicity of a communication mode because this may influence language use. In fact,
there is a continuum between asynchronous computer-mediated discourse (CMD),
which occupies a position closer to writing (i.e., conceptually written in Koch and
Oesterreicher’s terms), and synchronous CMD, which occupies a position closer to
speaking (i.e., conceptually oral). However, this factor does not only depend on the
medium, it also depends on the situation, i.e., the interactants’ use of the medium
(Androutsopoulos 2007). The more synchronous the communication is (i.e., the
shorter the delay between messages), the more likely it is to be conceptually oral.
The future of email research is closely linked to the future of email communication
itself. Although the first email message was sent as early as 1971 (Bryant 2011),
email did not become a widely used means of communication by the public until
towards the end of the 20th century, at which time it quickly advanced to become
the most frequently used Internet application. Until recently, figures for email
usage have risen consistently ever since email became available to the general pub-
lic in the early 1990s. Yet it seems that the quick rise in email usage might be fol-
lowed by an equally quick fall: Newer and more synchronous services such as the
various forms of Instant Messaging,9 as well as the numerous social network sites
among which Facebook and Twitter are the most popular, have started to compete
with email. According to a study in the U.S. by compete.com, Facebook is cur-
rently the most visited social network site, with 700,000,000 estimated unique
monthly visitors in November 2011.10 Compared to email, social network sites are
more personal because one can read the profiles of the addressees, check their status
updates (see Lee 2011), flip through their photo albums, and learn about their per-
sonal lives. In other words, they provide many features other than just email ad-
dresses and messages through which users can represent themselves.
Social network sites have another advantage in comparison to email: As is gen-
erally known, one of the major problems with email is the amount of spam (or un-
solicited messages) that is sent to email accounts. According to Pingdom, a service
provider that checks the availability of online services for major companies, spam
accounted for 89.1 % of all email traffic in 2010.11 Similar problems do not (yet)
exist with Instant Messaging services or with social network sites, which makes
these services even more popular. Accordingly, a telephone survey conducted by
the Pew Internet & American Life Project as early as 2005 stated that “email may
be at the beginning of a slow decline as online teens begin to express a preference
for instant messaging” (Lenhart, Hitlin, and Madden 2005: 5). In a newer Pew
Internet Project study, a 9th grade boy is quoted as saying: “[You go to MySpace12]
when all of your friends have gone to MySpace and they aren’t emailing anymore”
(Lenhart, Arafeh, Smith, and Macgill 2008: 61). The figures presented in the report
show a clear tendency: “Email remains the least popular choice for daily communi-
cation: [J]ust 16 % of teens send emails to their friends on a daily basis” (Lenhart et
al. 2008: 53). It may be assumed that at the time of this writing, the shift is even
more pronounced: Email is no longer the privileged mode of online communi-
cation among young people.
However, another point must be taken into consideration that partly speaks in
favour of the future of emailing: While email has been linked to stationary com-
puters for a long time, technology trends are towards increasing mobility. Es-
pecially with the diffusion of Apple’s iPhone and its high-speed data connection,
emails can be received almost as easily as text messages, which may lead to a new
boost in email figures. Be that as it may, information technology experts regard
email as an “old-fashioned” way of communication, and some people predict that
email use will die off. Lorenz (2007), for instance, published an article on “The
death of e-mail” in which he summarised the decline of email as follows: “Those of
us older than 25 can’t imagine a life without e-mail. For the Facebook generation,
it’s hard to imagine a life of only e-mail […]. As mobile phones and sites like
Twitter and Facebook have become more popular, those old Yahoo! and Hotmail
accounts increasingly lie dormant” (n.p.).
These reports seem to predict doom for email communication. Yet they mainly
concern private email communication. It is possible that the situation is different in
the business sector, where traditional email communication serves important func-
tions and may persist. There are also email-based systems such as web forums and
discussion lists that are active at present and might continue to be used. It can be
concluded that email research will have a future as long as email as a communi-
cation mode finds its niches.
Notes
1. For further details about this kind of “multi-party interaction” see Skovholt and Sven-
nevig (2006).
2. Due to limitations of space, this chapter focuses on these two languages only.
3. In contrast to one-way systems, there exist only a few two-way systems in the written
mode (Google Wave, for instance, offered real time communication, but Google discon-
tinued its development in 2010, only one year after it was released). Yet it is important to
mention this distinction, as – in contrast to Anglophone studies – there is a general
agreement among the German research community that online chat is not synchronous.
Some consider it “quasi-synchronous” (e.g., Dürscheid 2004: 154), while others (e.g.,
Beißwenger 2007: 37) distinguish between simultaneity (i.e., the participants are able to
see the message as it is produced) and synchronicity (i.e., sender and addressee are
logged in at the same time).
4. This is not the case for online chat, however. On cross-turn coherence in chat, see the
chapters in this volume by Herring and Markman.
5. For an empirical study on this issue (although only public emails are analysed), see Her-
ring (1996b).
6. It would be interesting to investigate whether the sociocultural factors that underlie
these rules are universal or whether they vary across cultures.
7. As the speed of data transfer has increasingly become faster in recent years and people
often have high speed Internet connections, this rule may no longer be relevant. Even
large data files can be downloaded quickly and easily these days.
8. This is our translation from the German “Mündlichkeits-/Schriftlichkeitsmodell”.
9. In contrast to email, Instant Messaging is a more synchronous communication mode:
The interlocutors have to be online at the same time to receive each other’s messages.
10. http://www.ebizmba.com/articles/social-networking-websites [accessed 4/30/2012]
11. http://royal.pingdom.com/2011/01/12/internet-2010-in-numbers/
[accessed 11/14/2011]
12. Until Facebook took over in popularity several years ago, MySpace had been the most
popular social network site.
References
Androutsopoulos, Jannis K.
2006 Introduction: Sociolinguistics and computer-mediated communication. Jour-
nal of Sociolinguistics 10(4): 419–438.
Androutsopoulos, Jannis K.
2007 Neue Medien – neue Schriftlichkeit? Mitteilungen des Germanistenverbandes
54(1): 72–97.
Baron, Naomi S.
1984 Computer mediated communication as a force in language change. Visible
Language 18(2): 118–141.
Baron, Naomi S.
1998 Letters by phone or speech by other means: The linguistics of email. Language
and Communication 18: 133–170.
Baron, Naomi S.
2000 Alphabet to Email: How Written English Evolved and Where It’s Heading.
New York/London: Routledge.
Baron, Naomi S.
2003 Why email looks like speech. Proofreading, pedagogy, and public face. In:
Jean Aitchison and Diana M. Lewis (eds.), New Media Language, 85–94. New
York/London: Routledge.
Baron, Naomi S.
2008 Always On: Language in an Online and Mobile World. Oxford, UK: Oxford
University Press.
Beißwenger, Michael
2007 Sprachhandlungskoordination in der Chat-Kommunikation. Berlin/New York:
Mouton de Gruyter.
Beutner, Yvonne
2002 E-Mail Kommunikation. Eine Analyse. Stuttgart: ibidem Verlag.
Biber, Douglas
1988 Variation across Speech and Writing. Cambridge, UK: Cambridge University
Press.
Bryant, Martin
2011 The first email message was sent 40 years ago this month. The Next Web,
October 8. http://thenextweb.com/insider/2011/10/08/the-first-email-was-sent-
40-years-ago-this-month
Cates, Ken
The Harvey Ball smiley face: A short history on the 40th anniversary – 2003.
The Smiley Store. http://www.smileystore.com/smiley-face-history
Cherny, Lynn
1999 Conversation and Community: Chat in a Virtual World. Stanford, CA: CSLI
Publications.
Cho, Thomas
2010 Linguistic features of electronic mail in the workplace: A comparison with
memoranda. Language@Internet 7, article 4. http://www.languageatinternet.
org/articles/2010/2728
Crystal, David
2001 Language and the Internet. Cambridge, UK: Cambridge University Press.
Crystal, David
2002 The English Language: A Guided Tour of the Language, 2nd Edition. London:
Penguin Books Ltd.
Crystal, David
2004 The Language Revolution. Cambridge, UK: Polity Press.
Dresner, Eli and Susan C. Herring
2010 Functions of the non-verbal in CMC: Emoticons and illocutionary force. Com-
munication Theory 20: 249–268.
Dürscheid, Christa and Arne Ziegler (eds.)
2002 Kommunikationsform E-Mail. Tübingen: Stauffenburg.
Dürscheid, Christa
2004 Netzsprache – ein neuer Mythos. In: Michael Beißwenger, Ludger Hoffmann,
and Angelika Storrer (eds.), Internetbasierte Kommunikation. Osnabrücker
Beiträge zur Sprachtheorie 68, 141–157. Duisburg: Gilles & Francke.
Dürscheid, Christa
2005 E-Mail – verändert sie das Schreiben? In: Torsten Siever, Peter Schlobinski,
and Jens Runkehl (eds.), Websprache.net. Sprache und Kommunikation im In-
ternet, 85–97. Berlin/New York: Mouton de Gruyter.
Dürscheid, Christa
2009 E-Mail: eine neue Kommunikationsform? In: Sandro Moraldo (ed.), Inter-
net.kom. Neue Sprach- und Kommunikationsformen im World Wide Web,
39–71. Rome: Aracne Edititrice.
Elspaß, Stephan
2002 Alter Wein und neue Schläuche? Briefe der Wende zum 20. Jahrhundert und
Texte der neuen Medien – ein Vergleich. In: Ulrich Schmitz and Eva Lia Wyss
(eds.), Briefkultur im 20. Jahrhundert. Osnabrücker Beiträge zur Sprachthe-
orie 64, 7–31. Duisburg: Gilles & Francke.
Frehner, Carmen
2008 Email – SMS – MMS. The Linguistic Creativity of Asynchronous Discourse in
the New Media Age. Bern: Peter Lang.
Grice, H. Paul
1975 Logic and conversation. In: Peter Cole and Jerry L. Morgan (eds.), Syntax and
Semantics 3: Speech Acts, 41–48. New York: Academic Press.
Günther, Ulla and Eva Lia Wyss
1996 E-mail-Briefe – eine neue Textsorte zwischen Mündlichkeit und Schriftlich-
keit. In: Ernest W. B. Hess-Lüttich, Werner Holly, and Ulrich Püschel (eds.),
Textstrukturen im Medienwandel, 61–86. Frankfurt am Main: Peter Lang.
Herring, Susan C. (ed.)
1996a Computer-Mediated Communication: Linguistic, Social, and Cross-Cultural
Perspectives. Amsterdam/Philadelphia: John Benjamins.
Herring, Susan C.
1996b Two variants of an electronic message schema. In: Susan C. Herring (ed.),
81–108.
Herring, Susan C.
2001 Computer-mediated discourse. In: Deborah Tannen, Deborah Schiffrin, and
Heidi Hamilton (eds.), Handbook of Discourse Analysis, 612–634. Oxford,
UK: Blackwell.
Herring, Susan C.
2007 A faceted classification scheme for computer-mediated discourse.
Language@Internet 4, article 1.
http://www.languageatinternet.org/articles/2007/761
Herring, Susan C.
2008 Foreword. In: Sigrid Kelsey and Kirk St. Amant (eds.), Handbook of Research
on Computer-Mediated Communication, Volume 1, xxxv–xxxvi. Hershey, PA:
Information Science Reference.
Höflich, Joachim and Julian Gebhard (eds.)
2003 Vermittlungskulturen im Wandel: Brief, Email, SMS. Frankfurt am Main: Peter
Lang.
Huffaker, David A. and Sandra L. Calvert
2005 Gender, identity, and language use in teenage blogs. Journal of Computer-
Mediated Communication 10(2), article 1.
http://jcmc.indiana.edu/vol10/issue2/huffaker.html
Janich, Nina
1994 Electronic Mail – eine betriebsinterne Kommunikationsform. Muttersprache
3: 248–259.
Johns, Kass
1996 Basic electronic mail netiquette (Network etiquette). http://www.kassj.com/
netiquette/netiquette.html
Jucker, Andreas
2006 Live text commentaries: Read about it while it happens. In: Jannis K. Androut-
sopoulos, Jens Runkehl, Peter Schlobinski, and Torsten Siever (eds.), Neuere
Entwicklungen in der linguistischen Internetforschung, 113–131. Hildesheim/
Zürich/New York: Georg Olms Verlag.
Koch, Peter and Wulf Oesterreicher
1994 Schriftlichkeit und Sprache. In: Hartmut Günther and Otto Ludwig (eds.),
Schrift und Schriftlichkeit/Writing and its Use. Ein interdisziplinäres Hand-
buch internationaler Forschung/An Interdisciplinary Handbook of Inter-
national Research, 587–604. Berlin/New York: Mouton de Gruyter.
Landert, Daniela and Andreas H. Jucker
2011 Private and public in mass media communication. From letters to the editor to
online commentaries. Journal of Pragmatics 43: 1422–1434.
Lee, Judith Yaross
2003 Charting the codes of cyberspace: A rhetoric of electronic mail. In: Lance
Strate, Ron L. Jacobson, and Stephanie B. Gibson (eds.), Communication and
Cyberspace, 2nd Edition, 307–328. Cresskill, NJ: Hampton Press.
Lee, Carmen K. M.
2011 Micro-blogging and status updates on Facebook: Texts and practices. In:
Crispin Thurlow and Kristine Mroczek (eds.), Digital Discourse: Lan-
guage in the New Media, 110–128. Oxford/New York: Oxford University
Press.
Lenhart, Amanda, Paul Hitlin, and Mary Madden
2005 Teens and technology. Pew Internet & American Life Project.
http://www.pewinternet.org/Reports/2005/Teens-and-Technology.asp x?r=1
Lenhart, Amanda, Sousan Arafeh, Aaron Smith, and Alexandra Macgill
2008 Writing, technology and teens. Pew Internet & American Life Project.
http://pewinternet.org/Reports/2008/Writing-Technology-and-Teens.aspx
Lorenz, Chad
2007 The death of e-mail. Teenagers are abandoning their Yahoo! and Hotmail ac-
counts. Do the rest of us have to? Slate Magazine, November 14.
http://www.slate.com/id/2177969
Maynor, Natalie
1994 The language of electronic mail: Written speech? In: Greta D. Little and
Michael Montgomery (eds.), Centennial Usage Studies, 48–54. Tuscaloosa,
AL: University of Alabama.
Orlikowski, Wanda and Joanne Yates
1993 From memo to dialogue: Enacting genres of communication in electronic
media. Technical Report TR 139–3525, MIT Sloan School.
http://dspace.mit.edu/bitstream/handle/1721.1/2451/SWP-3525-27731982.pdf
Pansegrau, Petra
1997 Dialogizität und Degrammatikalisierung in E-mails. In: Rüdiger Weingarten
(ed.), Sprachwandel durch Computer, 86–104. Opladen/Wiesbaden: Westdeut-
scher Verlag.
Petrie, Helen
1999 Writing in cyberspace: A study of the uses, style and content of email.
Unpublished paper. University of Hertfordshire/Hatfield.
Runkehl, Jens, Peter Schlobinski, and Torsten Siever
1998 Sprache und Kommunikation im Internet. Opladen/Wiesbaden: Westdeutscher
Verlag.
Schmitz, Ulrich
2002 E-Mails kommen in die Jahre. Telefonbriefe auf dem Weg zu sprachlicher Nor-
malität. In: Christa Dürscheid and Arne Ziegler (eds.), Kommunikationsform
E-Mail, 33–56. Tübingen: Stauffenburg.
Severinson Eklundh, Kerstin
2010 To quote or not to quote: Setting the context for computer-mediated dialogues.
Language@Internet 7, article 5.
http://www.languageatinternet.org/articles/2010/2665
Shapiro, Norman Z. and Robert H. Anderson
1985 Towards an ethics and etiquette for electronic mail. http://www.rand.org/pubs/
reports/R3283.html
Skovholt, Karianne and Jan Svennevig
2006 Email copies in workplace interaction. Journal of Computer-Mediated Com-
munication 12(1), article 3.
http://jcmc.indiana.edu/vol12/issue1/skovholt.html
Stein, Dieter
2003 Book review of David Crystal: Language and the Internet. Linguistics 41(1):
158–163.
Thurlow, Crispin
2001 Language and the Internet. In: Rajend Mesthrie and Ronald E. Asher (eds.),
The Concise Encyclopedia of Sociolinguistics, 287–289. Oxford, UK: Elsevier
Science Ltd.
Urban Dictionary
n.d. http://www.urbandictionary.com [Retrieved 4/30/2012]
Waldvogel, Joan
2007 Greetings and closings in workplace email. Journal of Computer-Mediated
Communication 12(2), article 6.
http://jcmc.indiana.edu/vol12/issue2/waldvogel.html
Weingarten, Rüdiger (ed.)
1997 Sprachwandel durch Computer. Opladen/Wiesbaden: Westdeutscher Verlag.
Witmer, Diane F. and Sandra Lee Katzman
1997 On-line smiles: Does gender make a difference in the use of graphic accents?
Journal of Computer-Mediated Communication 2(4).
http://jcmc.indiana.edu/vol2/issue4/witmer1.html
1. Preliminary remarks
The first mailing list was established in the mid-1970s on the precursor of the
Internet, the ARPANET. The ARPANET was set up as a computer network for
advanced collaborative space research by the United States military (Hauben
1998–1999, 2000) in 1969. It connected governmental and non-governmental
sites of research and grew rapidly. Soon a new format for communicative ex-
changes among users of this network was developed that was called “electronic
mail” or “e-mail” (Hauben 1998–1999). In 1975, the first mailing list on the AR-
PANET was established, the “MsgGroup”. Net Manager Steve Walker’s message
from June 7, 1975 to this group in which he proposed a discussion that would be
based solely on the MsgGroup distribution list facility rather than on regular com-
mittee meetings is viewed as the starting point for email discussion lists (Hauben
1998–1999).
In 1986, L-soft International, Inc. launched the first version of LISTSERV, its
software for managing email distribution lists. It has since become the most widely
used software for maintaining various kinds of automatized email distribution. Be-
cause of this widespread use, the term “listserv” is sometimes used as a cover term
for all kinds of software-supported, public mailing lists, although other products
for managing mailing lists like Majordomo, Procmail, and LISTPROC are also
available. A listserv automatically distributes emails that are sent by one sub-
scriber to all other subscribers of a list. The software also archives postings in an
online searchable database. In order to become a member of a mailing list, pros-
pective subscribers must send an email containing a subscription command to the
server running the listserv software. Their email address is then either automati-
cally included in the distribution list (in the case of public lists) or is manually
added to the list by the list owner (in the case of private lists). Mailing lists are set
up for communication among geographically dispersed persons who are interested
in specialized topics. For this reason, mailing lists have been widely used in vari-
ous academic disciplines, in professional contexts, and in distance education.
Communication on mailing lists may be either moderated or unmoderated. In
the case of moderated lists, a moderator – often the list owner – monitors all in-
coming emails (often supported by specific software which is set up to detect cer-
tain keywords in the mails) and selects those which do not conform to certain stan-
dards of the list (in terms of theme, tone, etc.) and suppresses their further
distribution. All other incoming mails are then distributed to the subscribers of the
list. In the case of unmoderated lists, all incoming mails are automatically dis-
tributed to all subscribers. Messages are distributed in ASCII2 text format by older
versions of the program; newer versions also allow the distribution of messages in
HTML3 format.
Mailing lists differ from newsgroups in several respects. Newsgroups were
originally set up for n:n email message distribution on Usenet, which was created
in 1979 by graduate students at Duke University and the University of North Ca-
rolina and went online in 1980. Originally, it provided a network for users who had
access to Unix systems and who could not connect to the ARPANET for technical
reasons (Hauben 2000). Unlike mailing list software, which runs on a single server,
newsgroups are maintained on clusters of servers which exchange incoming mess-
ages and which are hosted by Internet service providers (ISPs). Until 1995, users
who wanted to become newsgroup members needed an ISP to access Usenet
groups or other newsgroups,4 along with newsreader software. In 1995, Dejanews
(which in 2001 was bought by Google) bought Usenet; since then Usenet groups
are available on the web and can be accessed via standard web browsers (and re-
semble web forums; see section 8 below).
In most newsreaders, newsgroup messages appear organized by “threads”,
which are tree-like structures connecting an initiating message and all follow-up
messages (answers). Threads thus provide a visual representation of the thematic
and interactional relations between messages. In contrast, mailing list messages
appear in the inboxes of subscribers’ email programs, and although their topical
coherence is organized by subject headers, the relation between messages on a cer-
tain topic is not visually represented.
Furthermore, in most cases newsgroups do not have administrators (and are
therefore unmoderated). Newsgroups exist on a vast variety of themes/interests,
and users often participate for recreational purposes; hence the social contexts of
use differ between mailing lists and newsgroups, although there are also news-
groups on academic/scientific topics. Since newsgroups do not have any threshold
of access (see below), they tend not to be considered serious discussion forums by
most professionals in the disciplines.
Mailing list communication is an asynchronous “many-to-many” mode of
communication which is email based. Email communication, however, cannot be
viewed as a single “genre” (or communicative mode); rather, it is a form of com-
munication, and mailing list communication is one of its sub-forms (or a media
class, in the terms of December 1996). Forms of communication (as opposed to
communicative modes, channels, or genres) are defined by characteristic combi-
nations of communicative possibilities in three general dimensions: (a) the specific
sign types (e.g., spoken vs. written) they can process; (b) the direction of com-
munication they permit; (c) the possibilities of transmission or storage of data they
allow (for details of the concept of communicative forms, see Holly 1997; for a
detailed characterization of various forms of communication in the new media, see
Gruber 2008). The characteristic combinations of the relevant variables for mail-
ing list communication are shown in Table 1:
Many of the above features are determined by technological affordances of the dis-
tribution software and other hardware and software components that are involved
in the production and dissemination of email messages on distribution lists. These
features, however, also have wider consequences regarding the social contexts in
which mailing list communication is used and with regard to which groups of users
mainly use this communicative form. Thus technology, communicative features,
contexts of use, and characteristics of user groups are intertwined, and investi-
gations of mailing list communication are therefore always investigations of com-
municative practices under a pragmatic perspective. For this reason, before pro-
ceeding to an overview of the relevant literature on mailing list communication, a
short definition of how “pragmatics” is understood for the purposes of this chapter
is provided.
3. Defining pragmatics
ena under consideration are investigated by taking into account relevant features of
the contexts in which they occur and how these factors contribute to meaning-mak-
ing practices of interlocutors. This conception of pragmatics views language users
as making continuous communicative choices among available options. Language
users are thus flexible in both constructing and interpreting utterances, but the
probability of making certain choices is systematically skewed according to con-
textual features. This view entails that communicative choices are neither fixed op-
tions nor are they irrevocably tied to certain contexts: As contexts change, com-
municative choices also change and adapt. Patterns of language use, of course, also
contribute to the creation of contexts, thus the relations between language use and
the contexts of its use are dynamic and dialectical (for a more detailed outline of
this conception of pragmatics, see Verschueren 1999; Verschueren, Östman, and
Blommaert 1995).
For the present chapter, this conception of pragmatics has the advantage that all
linguistic features of mailing list communication which have been investigated so
far can be systematically related to technological and social characteristics of the
contexts of their use for meaning-making practices (provided that the relevant de-
tails are given in the respective studies). Levels of linguistic description and fea-
tures of context provide a conceptual grid for the organization of the rest of this
chapter: I start with investigations of the word/morpheme level and the question of
language use in mailing list communication (register and language choices), and
then proceed to investigations of features of discourse organisation (interactional
coherence). Subsequently, I turn to the level of generic structures, and finally I ad-
dress interactive norms and practices of mailing list communication.
In the early days of CMD research, the term “email communication” often served
as a cover term for all kinds of messages transmitted via computer connections that
receivers found in the inbox of their email programs. In most cases, however, these
early investigations of “email communication” were based on corpora of listserv or
bulletin board communications,6 as this kind of data was easily accessible for com-
munication, and language scholars were most often university based and therefore
themselves taking part in these early forms of CMD.
The first discourse feature that attracted the interest of language scholars was
the obvious “sloppiness” of many emails. In one of the first empirical investigations
of a corpus of email and chat messages, Maynor (1994) showed that her data dis-
played features of spoken as well as of written language. They frequently lacked
subject pronouns, copula verbs, and articles, and showed a high frequency of ab-
breviations, contractions, and lexical peculiarities. Maynor dubbed these features
“e-style”, implicating that they would characterise all kinds of CMD. In a study of
postings to two scholarly discussion lists, however, Herring (1996b) found that
message characteristics varied with gender of discussion participants, i.e., that
male and female list subscribers used different message forms.
The hybrid mode characteristic of email language seems to be one of the most ro-
bust findings in CMD research. Linguistic features associated with the character-
istics of spoken language include the use of 1st and 3rd person pronouns, discourse
particles, modal elements, hedges (Du Bartell 1995; Gruber 1997a, 2000a; Uhlirova
1994; Yates 1996), and the number of prepositional phrases that are used (Gruber
1997a). The written language characteristics of email communication include a high
type/token ratio and high lexical density, a high proportion of subordinate clauses,
and a lower number of coordinated clauses (Gruber 1997a; Yates 1996).
In a study of messages from an international BBS, Collot and Belmore (1996)
also found that their text corpus combined features of written and spoken language
in terms of Biber’s (1988) register model. Danet (1997) discussed the “orality”
of different communicative genres in written CMD and the wider cultural impli-
cations of the change from printing to digital culture.
Although almost all of these early studies investigated formal, grammatical
features of public email communications and focused neither on the interactional
dimension of CMD nor on the specific interactional practices that asynchronous
public email communication facilitates, all of the authors interpreted their results
within a functional and hence pragmatic perspective: The hybrid mixture of spoken
and written language features in public email messages was seen as a result of the
special contextual and situational frame of CMD. These pragmatically relevant
situational features include the absence of paralinguistic and prosodic features of
language; a lack of direct contact among discussants; the informality of the ex-
change situation, as well as the participants’ shared background knowledge; their
common interest in discussing certain topics; and their awareness of the public na-
ture of the communication situation.
A second formal feature of public email communication that was investigated
results from the above-mentioned limitations of the ASCII character set used for
composing messages. Email users overcame these shortcomings very quickly in a
rather creative way. In an email message posted on 19 September 1982 (11:44) to
Carnegie Mellon Universities’ computer science general board, Scott Fahlman
proposed to use two “emoticons” – combinations of ASCII characters, namely :-)
and :-( – to convey meta-communicative meanings (“joking” and “non-joking”
speaker intentions, respectively) in email messages (Fahlman n.d.-a, n.d.-b). Soon,
the use of these emoticons spread to the ARPANET and Usenet, and many new
emoticons have been invented since then (Argyle and Shields 1996; Baym 1995).
Although most modern email programs allow users to produce emails in HTML
format (and thus offer much richer design opportunities), ASCII emoticons are still
used for conveying meta-communicative meanings (Dresner and Herring 2010),
even in non-email texts.
gations of language choice in public email discussion started much later. One of the
reasons for this might be that in the 1980s through the mid-1990s, the Internet was
perceived as an English-speaking medium. This was especially notable in listserv
discussion groups that were set up in academic contexts for international groups of
scholars who shared an interest in certain topics and who also shared one language,
in most cases English.
One of the first investigations of language choice in public email discussions
was Paolillo’s (1996) study of the use of English and Punjabi in a Usenet news-
group for expatriate Punjabis. Paolillo’s results showed that although the news-
group was set up for communication among Punjabis, the Punjabi language was
used only rarely, while the use of English prevailed. Paolillo discussed four factors
which in his view might be responsible for this: 1) Intergenerational language shift
is responsible for a limited language competence in second generation Punjabis
who live in English speaking countries and who might be (or at least feel) excluded
if Punjabi were to be predominantly used in the newsgroup; 2) cultural ambiva-
lence partly results from this first factor, as many expatriates feel that they can only
preserve their culture in an English-speaking world if their cultural values are
transmitted in English to second- and third-generation group members (i.e., the
language that these persons speak in most situations of their lives); 3) the prestige
status of English in South Asia as a macro-sociological influencing factor is prob-
ably relevant in many post-colonial multilingual countries and/or communities in
which English is the lingua franca of an educated elite (see below). While the in-
fluence of Usenet language norms (Paolillo’s fourth influencing factor) has prob-
ably changed since Paolillo conducted his study, it surely holds true that a domi-
nating language (and culture) will influence language choice in discussion groups
that are established for minority members.
In a study of language choice of young professional Egyptian Internet users,
Warschauer, Said, and Zohry (2002) reported results that resemble Paolillo’s in
some respects. Among the subjects of their study, English was also the predomi-
nantly used language in asynchronous and synchronous CMD, although all sub-
jects were native speakers of Egyptian Arabic (and also fluent in Classical
Arabic). On the one hand, the authors explain this result with reference to the
business environment of the young professionals, who almost exclusively worked
in software- or hardware-related businesses in which English was the established
language of communication. On the other hand, the authors also mention that (at
least at the time when the study was conducted) Arabic script was not available
for the programs the subjects used for Internet communication. Thus, when they
wanted to communicate in Egyptian Arabic, they had to write Arabic in Roman
script, which is rarely done in other contexts. A third influencing factor which –
according to the authors – was responsible for language choice in this context is
the prestige of English as the language of an educated, non-Islamist elite in
Egypt. This factor is reminiscent of Paolillo’s argument that the prestige status
of English in South Asia was an influencing factor for language choice in his
study.
In the context of multilingual Switzerland, Durham (2003) investigated lan-
guage choice on a mailing list for Swiss medical students over a period of three
years. Subscribers to the list came from three linguistic backgrounds (which also
represent three of the four Swiss national languages): Swiss-German, French, and
Italian. Durham’s results show a shift towards the dominant use of English during
the research period. Although French was used mostly at the beginning (as most of
the initial subscribers came from the Geneva area where French is the dominant
language), list communication switched mainly to English after about six months,
even though English was not the mother tongue of any group of list subscribers.
Durham attributes this shift to two main reasons. (1) Sometime after the list had
been established, it was joined by a group of Italian speakers who had difficulties
communicating in both German and French. Using English put everybody in the
group at an equal disadvantage, but it also ensured that everybody would under-
stand messages without requiring additional translations (as was the case with all
three other languages). (2) English is a high prestige language in Switzerland, es-
pecially in educational and business settings, where many international Swiss busi-
nesses have switched to English as their company language.
The studies reviewed in this section show that a variety of pragmatic factors in-
fluences language characteristics of mailing list communication, as well as the
choice of language in mixed language groups. While features of the communication
technology (the lack of visual, paralinguistic, and non-verbal cues) and of the com-
municative setting (shared background knowledge, theme-centred discussions) are
responsible for register characteristics of mailing list communication, language
choice is influenced by local factors such as the availability of certain scripts in
email software, as well as by factors such as social identities, mutual intelligibility,
and language prestige. While these latter factors often result in the choice of Eng-
lish as the lingua franca in multilingual discussion groups, this does not necessarily
imply that linguistic imperialism is at work in all these cases. Empirical evidence
from other multilingual settings (such as the states of the former Soviet Union,
where Russian was the lingua franca for almost 50 years) would be needed in order
to arrive at a closer understanding of the pragmatics of language choice in such set-
tings. Furthermore, both register features and language choice contribute to the
pragmatics of interpersonal meaning making in mailing list communication.
CMD. As Herring (1999) notes, there are two basic problems of interactional co-
herence in asynchronous (as well as in synchronous) CMD: (1) the lack of back-
channel signals, which is caused by the reduced audio-visual cues and the impossi-
bility of message overlap, and (2) disrupted turn adjacency, which is caused by the
fact that messages appear in the order in which they are received and distributed by
the listserv. One of the only ways of establishing topical coherence is thus the use
of the same topic words in the message headers of a discussion.
Two problems (on a more theoretical level) with the organisational features of
CMD have less often been discussed, namely (1) whether we are justified in using
the metaphor of “discussion” for mailing list exchanges and (2) whether theoretical
frameworks that were developed for modelling face-to-face interaction are ad-
equate for the analysis of mailing list “discussions”, or if we need an alternative
(or at least modified) framework. In the following, I review studies that deal with
several aspects of the interactional organisation of mailing list exchanges from a
pragmatic perspective. In closing, I deal briefly with the latter-mentioned theoreti-
cal problems and ask whether the conceptual frameworks of conversation analysis
and discourse analysis that have been used to analyse mailing list communication
so far in fact provide appropriate conceptual and analytical tools for analysing
these interactions.
A first and basic problem of discourse organisation in mailing list interactions
is the question of what constitutes a turn (in conversation analytic terms) or a move
(in Sinclair and Coulthard’s discourse analytic terms) in a multiparty email ex-
change (Severinson-Eklundh 2010). As some authors (Gruber 1998; Harrison
1998; Herring 1999) have noticed, reactive contributions to email discussions
often refer to more than one previous message, and discussion-initiating postings
also often pose more than one question to stimulate further discussion (Gruber
1998). Thus, many mailing list contributions consist of several moves or speech
acts, because in asynchronous CMD the right to “speak” is not contested, i.e., any
participant can post as many messages as he/she wants (a similar point is made by
Condon and Čech 2001 with regard to synchronous CMD). Mailing list contri-
butions therefore often show characteristics of speech-act-chains (“Sprechhand-
lungsverkettungen”, Ehlich and Rehbein 1986), which in traditional interactions
characterise written monologue texts. Under a conversation analytic perspective,
these multi-move contributions – although they often contain explicit next-speaker
selections (Gousseva 1998; Gruber 1998) – may be viewed as offering several tran-
sition relevance places, since every question in a multi-question contribution can
be (and often is) interpreted as an invitation for an answer to this single question
only, as the following extract from a lengthy posting to the LINGUIST list shows
(author details omitted):
Example 1.
Date: Thu, 10 Aug 1995 10:21:41–0500
Reply-To: The Linguist List <[log in to unmask]>
Sender: The LINGUIST Discussion List <[log in to unmask]>
From: The Linguist List <[log in to unmask]>
Subject: 6.1070, Disc: Sex/Lang, Re: 1023
[…]
I have three questions, each of which I provide some of my own
views about.
1) What are some of the parameters of and who (authors) do you
look to for your idea that all history has been male dominated?
[…]
2) Where does your concept of linguistic markedness come from,
and on what basis do you establish a cause-effect relationship
between patriarchy and markedness in pronouns? […]
3) Finally, you propose that in all past history they’ve had
“no concept of sex equality”. […] If we take into account
considerations of class, age, history, geography, etc., I
would argue that on a DE FACTO level, women have had, and
continue to have, COLLECTIVELY, an advantage over men. I know
that saying so will upset a few of us, even give some of us a
little gas and heartburn, my being the direct precipitation of
which I apologize for in advance.
[…]
Each of these questions could (but, in fact, did not) receive an answer, and thus
three adjacency pairs could have been established, each of which would have con-
tributed to the interactional coherence in the LINGUIST list discussion on “Sex/
Lang”. Thus, the excerpt in example 1 above consists of three possible (or virtual)
turns which could be turned into actual turns by subsequent answers. This example
also shows that the conversation analytic concept of conditional relevance seems
to be operative only in a very limited sense in asynchronous multiparty ex-
changes, in which many “dead-end” contributions occur (see Gruber 1997b; Her-
ring 2010; Joyce and Kraut 2006).
So far, I have discussed the properties of first pair parts of (virtual) adjacency
pairs in mailing list discussion; in the following, I discuss the properties of second
pair parts or reactive contributions. Like first pair parts (or initiating moves), reac-
tive contributions in mailing list discussions may (and often do) consist of reac-
tions to a variety of previous contributions (Gruber 1998; Wilkins 1991), as the
extract in example 2 (again from the LINGUIST list discussion on “Sex/Lang”)
shows (author references removed):
Example 2.
Date: Sat, 12 Aug 1995 12:59:37 EDT
From: [log in to unmask]
Subject: Disc: Sex/Lang, Re: 1079
Subject: Re: Disc: Sex/Lang
A writes:
<<<
In our focus on the epicene pronoun that doubles as the
masculine pronoun, and the generic noun (e.g. Man) that doubles
as the masculine form, we sometimes let logic, and the theory
of markedness, overwhelm psychology. The evidence, I believe,
strongly supports the claim that even in clearly non-specific
contexts these items trigger masculine prototypes in
receivers.
>>>
I agree with this generally for the present state of standard
English. […]
B writes
<<<
Question raised in answer to C who writes:
>SO what it all boils down to, again, is that I maintain that
>it makes no sense whatever to discuss the origin of the
>epicene he phenomenon in the context of the story of
>English prescriptive grammar, but only in the context of
>the way in which perceptions of sex roles have informed
>the structure of language (as of any other institution).
Have perceptions of sex roles informed us on the structure of
language, or does the structure of language enlighten us on
socially acquired yet still subconscious sexist behavior of
today’s Homo Sapiens? B/
[Affiliation details of B]
>>>
The “contructivist” view of sexuality is widely held in
academia, but the view has the hauteur of an “in” religious
tenet, and has no well-reasoned position, and only endures
because its opposite can’t “really” be “proved”. […]
In example 2, the author of the actual contribution refers to two previous postings
(contributed by A and B) by providing the exact quotations on which he comments
in his own posting. In terms of interactional coherence, this example shows that
mailing list discussions do not progress in a linear way (neither in time nor in any
Herring (2007) argues for using flexible classification grids (“faceted classifi-
cation schemes”) which combine feature sets that capture different technological
and social aspects of CMD together with “mode”-based classifications in order to
arrive at a flexible categorization of different kinds of CMD. From a linguistic
pragmatic perspective, this classification scheme can be viewed as a complement
to studies of generic features of CMD and mailing list communication respectively,
as the features Herring proposes provide a valuable conceptual matrix for classify-
ing relevant pragmatic factors that could in turn be accompanied by an investi-
gation of the discursive features of certain message types. The obvious hybridity of
mailing list genres, as well as their connection to certain groups of users (“online
communities”; see below), seems to indicate that the pragmatics of mailing list
genres contributes to interpersonal as well as to textual meaning-making processes.
In general, although existing research seems to indicate that different mailing
list communities develop group-specific variants of the discussion-list-posting
genre, which vary in accordance with pragmatic factors such as familiarity with
communication technology and narrow vs. wide scope of legitimate discussion
topics, there is a need for further linguistic pragmatic studies in this area.
1. Steady participation over a defined time interval and the possibility to identify
core participants;
2. A shared history of the group, which leads to a group culture that can be ident-
ified through the investigation of group jargon, abbreviations, etc., and through
the emergence of group norms;
3. The display of group solidarity among community members through various
communicative practices;
4. Group-specific patterns of conflict and communicative conflict resolution;
5. A group’s display of self-awareness;
6. Evidence of roles and hierarchies in a group, which also relates to participation
patterns and norms.
In an illustrative analysis, Herring (2004) shows the degrees to which two online
discussion groups display characteristics of virtual communities in the above sense
and how pragmatic computer-mediated discourse analysis (CMDA) can be used to
identify these relevant characteristics.
Various other studies have investigated some of the above features in order to
address the question of community formation in discussion groups. Gruber (1998)
developed a “reciprocity” measure that allows for computing the perceived rel-
evance of list postings in order to differentiate between subscribers who post many
contributions but receive only a few reactions and those whose postings have a
high relevance for a discussion. This quantitative measure provides a differentiated
view of group roles, hierarchies, and participation patterns.
The importance of a moderator and of core members of a discussion group for
establishing and continuing interaction and becoming a community is pointed out
by Fayard and DeSanctis (2005). Cubbison (1999) discusses how list owners can
influence various aspects of group structure and community formation by choosing
specific configurations of listserv software settings. Although Cubbison’s article is
not linguistically oriented, her paper is interesting in the current context, as it
shows how software settings which, in most cases, are invisible to subscribers, set
a technological – yet pragmatically relevant – frame for communicative practices
and thus for the possibilities of community formation.
The emergence of group norms was investigated as early as the mid-1990s
when McLaughlin, Osborne, and Smith (1995) inductively arrived at a “taxonomy
of reproachable conduct on Usenet”. The taxonomy was based on the analysis of a
corpus of messages from various Usenet groups in which other users of the respect-
ive groups were criticised. McLaughlin, Osborne, and Smith grouped their data
into seven categories that, in many respects, resemble the norms for CMC com-
munication later formulated in many “netiquette” guides. In a qualitative study,
Johnson-Eiola and Selber (1996) investigated one discussion group’s emerging
definitions of an appropriate discussion topic. They concluded that a topic’s appro-
priateness on this list was defined in a framework that combined personal/profes-
sional interests of users and power relations in the workplace settings in which the
contributors were located. Topics that were collectively considered as inappro-
priate for the list were taken to other discussion forums. This marginalisation of
(potentially threatening) topics represented – according to the authors – a techno-
logically facilitated way of fragmenting a community into a majority who – by
means of a “democratic majority vote” – forced “dissidents” to leave the commu-
nity instead of dealing with potentially explosive topics. These studies of the
emergence of groups norms (and of exclusion practices) in discussion groups are
interesting insofar as they challenge the widely-held view of Internet communities
as “egalitarian” groups in which each member has equal rights.
Conflict communication in discussion groups has often been investigated in
terms of “flaming”. “Flaming” was one of the first communicative practices to be
observed and described in newsgroup (but also in discussion list) interaction (Kim
and Raja 1991). The term “flaming” covers all forms of communicative behaviour
in CMC through which discussants confront others in rude and hostile ways (Lea et
al. 1992). For some scholars, the cause of flaming is the “reduced cues” character-
istic of CMC (Kiesler, Siegel, and McGuire 1984). They claim that it is easier to be
rude to one’s interlocutor if one cannot see or hear him/her or, as in most cases, if
one does not even personally know him/her. Herring (1996a, 1996b), however,
claims that flaming is a typical male discussion list behaviour; she found women in
general to be more supportive of others. Depending on the gender majority of list
subscribers, discussion lists “exhibit an overall aligned or opposed orientation”
(Herring 1996a: 104). This finding also shows that discussion groups are commu-
nities whose values are established/influenced by the majority of group members.
Furthermore, Herring’s results caution against a “techno-pragmatic” view of CMD
that would claim that technological affordances of computer-mediated communi-
cation have more influence on communicative practices than users’ characteristics.
Several discursive features of a group’s self awareness were studied by Cassell
and Tversky (2005) in their investigation of the online community formation of
an intercultural, international youth forum that was set up to prepare an inter-
national youth summit in 1998. Their results show that the choice of English as a
common language, the use of personal pronouns (“I” vs. “we”), and the develop-
ment of conversation patterns (from providing information to giving feedback and
responses to each other and planning future activities) reflect the emergence of a
sense of community among the young adolescents. However, as the studied group
was spread over four continents and several cultural areas, there also remained
some characteristics of diversity. Cassell and Tversky’s study and the studies on
language choice in discussion groups (see section 4) show that in multilingual
groups language choice is also an indicator of community formation. As in the
studies reviewed in section 4, Cassell and Tversky found that the young adoles-
cents ultimately communicated exclusively in English in order to speed up com-
munication.
All of the studies reported here show how technical and social pragmatic fea-
tures of mailing list communication can foster community formation among per-
sons who in many cases do not know each other personally and whose only means
of inter-communication is email. It would be interesting, however, to see how per-
sistent these communities are, since mailing list communication is mainly used for
professional and educational purposes, and the resulting communities are only
relevant for their members as long as they belong to their respective professional or
educational fields. This aspect differentiates most mailing list communities from
newsgroups, which are joined by members who share an interest in certain rec-
reational topics. Therefore, newsgroup communities may exist over longer periods
of time (Baym 1995). Additionally, as Lamy and Hampel (2007) in their review of
possible uses of CMC in the language learning classroom point out, there are also
features of (the educational use of) asynchronous CMC that may hamper commu-
nity formation, like the solitariness of users, pressure (exerted by instructors) to
respond, or unnoticed silence (“lurking”7). These factors may lead to the formation
of online communities that exclude certain potential members without this being
noticed by others. Investigations of community formation on discussion lists
should thus take into account a wide range of pragmatic factors such as voluntary
vs. involuntary participation in community activities, intrinsic vs. extrinsic moti-
vations of participants, and other relevant contextual aspects.
8. Closing remarks
beginning providing only a limited character set for composing messages and in-
itially a purely textual communication channel, email programs now support the
Unicode character set (which allows the use of 90 different scripts) as well as
HTML text, which allows the integration of visual and other non-textual elements
into email messages. Furthermore, listserv technology has increasingly been re-
placed by Internet forums that are integrated into websites and that support full
HTML standards for messaging. Some of the studies reviewed above revealed how
users dealt with the limitations of certain “techno-pragmatic” factors. For example,
the invention and frequent use of “emoticons” in email communication shows
clearly how users overcame the early limitations of email programs by inventing a
set of new symbols (based on the available ASCII character set) that compensated
for the narrowness of the communication channel and allowed users to add a layer
of interpersonal meaning to their messages.
Another relevant techno-pragmatic feature of email (and hence mailing list)
communication is quoting. In contrast to the limiting effects of email communi-
cation, quoting is an “enabling” characteristic that allows users to establish refer-
ences between messages easily. Especially in mailing list communication, quoting
provides a straightforward way of establishing thematic connections between fol-
low-up postings in multiparty discussions and thus contributes to establishing
“textual meanings” that help interlocutors make sense of the foregrounding and
backgrounding of information, as well as of the relevant connections between parts
of a communicative exchange.
In terms of social and situational pragmatic factors (Herring 2007) that are rel-
evant in mailing list communication, the review of studies has shown that mailing
lists are used for many-to-many, public interactions, hence their characterisation
as “polylogue”. Although not all groups that communicate via mailing lists be-
come virtual communities, many of the linguistic characteristics of mailing list
communication can be viewed as corollaries of the group situation. The hybrid
“written-like-spoken” register characteristics as well as the frequent use of acro-
nyms in email communication can be viewed as users’ display of in-group identity
markers. Language use in group discussions also plays a relevant role. As studies
of language choice in multilingual discussion groups have indicated, group
members may choose a common language that puts all discussants on an equal
level of disadvantage (i.e., which is nobody’s mother tongue) or simply because
they want to practise a certain language. Both results can also be viewed as con-
tributing to interpersonal meaning making in mailing list discussions.
Gender of discussion participants has been shown to be a relevant pragmatic
factor that impacts group discussion styles (and hence implicit group norms). Dif-
ferent age groups of participants and their communicative practices in discussions
have so far barely been investigated since, until the late 1990s, Internet communi-
cation was mainly the “playground” for a white middle-class population of edu-
cated males between 20 and 40+ years of age. With CMD now becoming an almost
Notes
1. There are two possible perspectives on the communication mode of mailing list and Use-
net groups: From the perspective of the single user it is a 1:n mode, as one user sends his/
her message to a group of recipients. Viewed from the group’s perspective, however, the
communicative mode is n:n, as any users can interact with any others.
2. ASCII stands for “American Standard Code for Information Interchange”, a character en-
coding scheme which has been used for coding computer characters since 1960 and
which includes 94 printable and 33 non-printable characters (plus the “invisible” space
character).
3. HTML stands for “Hyper Text Markup Language”, the predominant markup language for
web pages, which allows a structural description of headings, font types, etc. as well as
the integration of images, sounds, etc. in texts.
4. An example of this is fidonet, which still is in use in areas where most computer users ac-
cess the Internet via telephone modems, such as in Russia and other countries of the
former Soviet Union.
5. This dimension draws upon a distinction introduced by Koch and Österreicher (1994),
who assume that the conception of a communicative product as “spoken” or “written” is
independent of its realisation in the oral or written mode. Thus a speech delivered orally
in front of an audience may have been composed in written mode very carefully by the
speaker, having in mind the face-to-face communicative situation of the speech, resulting
in a conceptually oral, i.e., a “written to be spoken” (Gregory and Carroll 1978: 47), text.
The oral delivery of a pronouncement in court, in contrast, is a typical instance of the oral
presentation of a conceptually written text. The conceptually literate pole of communi-
cation is associated with (and establishes) interpersonal distance, whereas the concep-
tually oral pole is associated with (and establishes) closeness between communication
partners.
6. Bulletin boards (BBSs) became popular in the late 1970s and early 1980s. They consisted
of a central server which was connected to several home computers of registered users. In
the early days of computer communication, users connected to the server by telephone
modems and then could upload and download files to and from the server. BBSs were
used for communication among users, exchange of files, and for online gaming. Since the
mid-1990s web forums have largely replaced BBSs.
7. Viewed from the standpoint of those members of an online community who actively par-
ticipate in online discussions, lurking is often seen as undesirable behavior, as lurkers do
not contribute to the ongoing discussions. Form the lurkers’ perspective, however, lurk-
ing is often viewed as a kind of participation, and lurkers may experience a sense of com-
munity with those persons who contribute to online discussions, even if they do not par-
ticipate in the discussions for various reasons (Nonnecke and Preece 2000).
References
Androutsopoulos, Jannis
2006 Introduction: Sociolinguistics and computer-mediated communication. Jour-
nal of Sociolinguistics 10(4): 419–438.
Argyle, Katie and Rob Shields
1996 Is there a body in the net? In: Rob Shields (ed.), Cultures of Internet: Virtual
Spaces, Real Histories, Living Bodies, 58–69. London: Sage.
Bateman, John A.
2008 Multimodality and Genre: A Foundation for the Systematic Analysis of Multi-
modal Documents. Basingstoke, UK: Palgrave.
Baym, Nancy
1995 The emergence of community in computer-mediated communication. In:
Steven G. Jones (ed.), Cybersociety: Computer-Mediated Communication and
Community, 138–164). London: Sage.
Berners-Lee, Tim and Daniel Connolly
1993 Hypertext Markup Language (HTML). Internet Draft.
http://www.w3.org/MarkUp/draft-ietf-iiir-html-01.txt
Biber, Douglas
1988 Variation Across Speech and Writing. Cambridge, UK: Cambridge University
Press.
Cassell, Justine and Dona Tversky
2005 The language of online internet community formation. Journal of Computer-
Mediated Communication 10(2), article 2.
http://jcmc.indiana.edu/vol10/issue2/cassell.html
Cherny, Lynn
1999 Conversation and Community: Chat in a Virtual World. Stanford, CA: CSLI
Publications.
Collot, Milena and Nancy Belmore
1996 Electronic Language: A new variety of English. In: Susan C. Herring (ed.),
Computer-Mediated Communication: Linguistic, Social, and Cross-Cultural
Perspectives, 13–29. Amsterdam/Philadelphia: John Benjamins.
Condon, Sherri L. and Claude G. Čech
2001 Profiling turns in interaction: Discourse structure and function. In: Proceed-
ings of the 34th Hawai International Conference on System Science, Vol. 4,
4034–4044. Los Alamitos, CA: IEEE Press.
Crystal, David
2001 Language and the Internet. Cambridge, UK: Cambridge University Press.
Cubbison, Laurie
1999 Configuring LISTSERV, configuring discourse. Computers and Composition:
An International Journal for Teachers of Writing 16: 371–381.
Danet, Brenda
1997 Books, letters, documents. The changing aesthetics of texts in late print cul-
ture. Journal of Material Culture 2: 5–39.
December, John
1996 Units of analysis for internet communication. Journal of Computer-Mediated
Communication 1(4). http://jcmc.indiana.edu/vol1/issue4/december.html
Gruber, Helmut
2000a Scholarly email discussion list postings: A single new genre of academic
communication? In: Lyn Pemberton and Simon Shurville (eds.), Words on
the Web: Computer-Mediated Communication, 36–44. Exeter, UK: Intellect
Books.
Gruber, Helmut
2000b Theme and intertextuality in scholarly e-mail messages. Functions of Lan-
guage 7(1): 79–115.
Gruber, Helmut
2008 Specific genre features of new mass media. In: Ruth Wodak and Veronika
Koller (eds.), Handbook of Communication in the Public Sphere, 363–383.
Berlin: de Gruyter.
Günter, Ulrike and Elisabeth L. Wyss
1996 E-mail Briefe – eine neue Textsorte zwischen Mündlichkeit und Schriftlich-
keit. In: Ernest W. B. Hess-Luttich, Werner Holly, and Ulrich Püschel (eds.),
Textstrukturen im Medienwandel, 61–87. Frankfurt am Main: Lang.
Haase, Martin, Michael Huber, Alexander Krumeich, and Georg Rehm
1997 Internetkommunikation und Sprachwandel. In: Rüdiger Weingarten (ed.),
Sprachwandel durch Computer, 51–86. Opladen: Westdeutscher Verlag.
Halliday, M.A.K.
1994 An Introduction to Functional Grammar, 2nd Ed. London: Edward Arnold.
Harrison, Sandra
1998 E-mail discussions as conversation: Moves and acts in a sample from a listserv
discussion. Linguistik Online 1(1).
http://www.linguistik-online.de/harrison.htm
Hauben, Ronda
1998– ARPANET mailing lists and Usent newsgroups: Creating an open and scien-
1999 tific process for technology development and diffusion. Part I. The Amateur
Computerist 9(1). http://www.ais.org/~jrh/acn/Back_Issues/Back_Issues%
5b1998-2002%5d/ACN9-1.txt
Hauben, Ronda
2000 ARPANET mailing lists and Usenet newsgroups: Creating an open and scien-
tific process for technology development and diffusion. Part III. The Amateur
Computerist 10(1). http://www.ais.org/~jrh/acn/Back_Issues/Back_Issues%
5b1998-2002%5d/ACn10-1.txt
Herring, Susan C.
1996a Posting in a different voice: Gender and ethics in CMC. In: Charles Ess (ed.),
Philosophical Perspectives on Computer-Mediated Communication, 115–147.
Albany, NY: State University of New York Press.
Herring, Susan C.
1996b Two variants of an electronic message schema. In: Susan C. Herring (ed.),
Computer-Mediated Communication: Linguistic, Social, and Cross-Cultural
Perspectives, 91–109. Amsterdam/Philadelphia: John Benjamins.
Herring, Susan C.
1999 Interactional coherence in CMC. Journal of Computer-Mediated Communi-
cation 4(4). http://jcmc.indiana.edu/vol4/issue4/herring.html
Herring, Susan C.
2001 Computer-mediated discourse. In: Deborah Schiffrin, Deborah Tannen, and
Heidi E. Hamilton (eds.), The Handbook of Discourse Analysis, 612–634. Ox-
ford, UK: Blackwell Publishers.
Herring, Susan C.
2004 Computer-mediated discourse analysis: An approach to researching online be-
havior. In: Sasha A. Barab, Rob Kling, and James H. Gray (eds.), Designing for
Virtual Communities in the Service of Learning, 338–377. Cambridge, UK:
Cambridge University Press.
Herring, Susan C.
2007 A faceted classification scheme for computer-mediated discourse. Lan-
guage@Internet 4, article 1.
http://www.languageatinternet.org/articles/2007/761/
Herring, Susan C.
2010 Who’s got the floor in computer-mediated conversation? Edelsky’s gender pat-
terns revisited. Language@Internet 7, article 8.
http://www.languageatinternet.org/articles/2010/2857/
Holly, Werner
1997 Zur Rolle von Sprache in Medien. Semiotische und kommunikationsstruk-
turelle Grundlagen. Muttersprache 107: 64–75.
Johnson-Eilola, Johndan and Stuart A. Selber
1996 Policing ourselves: Defining the boundaries of appropriate discussion in online
forums. Computers and Composition: An International Journal for Teachers
of Writing 13: 269–291.
Joyce, Elisabeth and Robert E. Kraut
2006 Predicting continued participation in newsgroups. Journal of Computer-Me-
diated Communication 11(3), article 3.
http://jcmc.indiana.edu/vol11/issue3/joyce.html
Kiesler, Sara, Jane Siegel, and Timothy McGuire
1984 Social psychological aspects of computer-mediated communication. American
Psychologist 39: 1123–1134.
Kim, Min-Sun and Narayan S. Raja
1991 Verbal aggression and self-disclosure in computer bulletin boards. Paper pres-
ented at the annual meeting of the International Communication Association,
Chicago, IL. Washington, D.C.: ERIC Clearinghouse on Languages and Lin-
guistics. Document no. ED334620.
Koch, Peter and Wulf Oesterreicher
1994 Schriftlichkeit und Sprache. In: Hartmut Günther and Otto Ludwig (eds.),
Schrift und Schriftlichkeit: ein interdisziplinäres Handbuch internationaler
Forschung, 587–604. Berlin: de Gruyter.
Kress, Gunther
2010 Multimodality: A Social Semiotic Approach to Contemporary Communication.
London: Routledge.
Lamy, Marie-Noelle and Regine Hampel
2007 Online Communication in Language Learning and Teaching. Houndsmill,
Basingstoke: Palgrave.
Wilkins, Harriet
1991 Computer talk: Long-distance conversations by computer. Written Communi-
cation 8(1): 56–79.
Yates, Simeon J.
1996 Oral and written linguistic aspects of computer conferencing: A corpus based
study. In: Susan C. Herring (ed.), Computer-Mediated Communication: Lin-
guistic, Social, and Cross-Cultural Perspectives, 29–47. Amsterdam/Phila-
delphia: John Benjamins.
even newer forms of CMC such as status updates on social networking sites like
Facebook (boyd 2006; Joinson 2008) and microblogging services such as Twitter
(boyd, Golder, and Lotan 2010; Honeycutt and Herring 2009; Java et al. 2007),
they are increasingly defined in terms of themselves and without reference to pre-
cursors.
Blood (2000) traces the origins of blogs to the practice of link sharing on the
early World Wide Web. She emphasizes that the earliest practitioner definitions of
what constituted a blog were based on the presence of dated entries containing
links, commentary, and thoughts on a personal website:
The original weblogs were link-driven sites. Each was a mixture in unique proportions
of links, commentary, and personal thoughts and essays. Weblogs could only be created
by people who already knew how to make a website. A weblog editor had either taught
herself to code HTML for fun, or, after working all day creating commercial websites,
spent several off-work hours every day surfing the web and posting to her site. These
were web enthusiasts. (para 5)
The function of sharing links and publishing information in a web feed, then-
unique features likely to have spurred the popularity of blogs, are today integrated
into virtually any component of the so-called social web. A variety of media con-
tents (photos, music, video clips, etc.) can easily be embedded in blog entries or
other hypertext-based services, while at the same time portals and multimedia ap-
plications integrate blog-like functions. What remains unchanged is that blogs
structure digital content sequentially and that they are more frequently maintained
by individuals than institutions or companies (McLean 2009).
Definitions based on the structural characteristics of blogs are popular among
researchers; for example, blogs are described as “frequently updated webpages
with a series of archived posts, typically in reverse-chronological order” (Nardi,
Schiano, and Gumbrecht 2004: 222) and “modified web pages in which dated en-
tries are listed in reverse chronological sequence” (Herring, Scheidt, et al. 2004:
1). This approach to definition can be interpreted in light of the extreme topical,
functional, and participatory variance of blogs, i.e., the fact that subject matter,
purpose, and community are highly variable from one blog to another and, inter-
nally, from a single post to the next.
Researchers from linguistics, rhetoric, and literary studies, but also from a var-
iety of other disciplines, frequently apply the term genre to blogs, relating it to
forms, such as the personal diary, which they propose as plausible ancestors (e.g.,
Karlsson 2006: 1, who calls the paper diary the blog’s “evident offline antecedent”).
This approach is not without its conceptual difficulties, because while technical con-
straints and certain stylistic patterns can be considered somewhat stable (Pusch-
mann 2010a: 109–114), the purpose and community of a blog are not, a problem that
poses a challenge to contemporary genre theory (see Askehave and Nielsen [2005]
and Giltrow [this volume] for a discussion of the complex relationship of medium
and genre, and Lomborg [2009], Scheidt [2009], and Miller and Shepherd [2009]
for analyses of blogs from the perspective of genre theory). Blogs share character-
istics with genres that are author-centric in terms of mode and sequential in terms of
text organization, such as the diary and the personal letter, and these common char-
acteristics are sometimes suggested to have been inherited by the blog in a genea-
logical sense. However, this assertion seems questionable in light of the historical
development of blogs. In those instances where they share traits with pre-digital
antecedents, this may partly be a result of conscious imitation or genre migration
and partly the symptom of a psychological desire for mediated self-expression that
predates the implied antecedents (see Gurak and Antonijevic [2008] and Nowson,
Oberlander, and Gill [2005] for psychological approaches to blogging).
Blogs and blogging have attracted a growing body of scholarship since the popu-
larization of the form in the late 1990s and early 2000s. While the use of blogs has
spread to a variety of contexts, such as academic research (Ewins 2005; Suzuki
2004), business (Puschmann 2010b; Sprague 2007), and education (Armstrong,
Berry and Lamshed 2004), studies suggest that the prototypical use of blogs as a
medium for personal publishing by private non-professionals still dominates over
other scenarios (McLean 2009; White and Winn 2009). Blogging has been studied
with a range of methodologies and from a variety of disciplinary perspectives,
among them ethnography (boyd 2006; Brake 2007; Gumbrecht 2004; Nardi,
Schiano, and Gumbrecht 2004; Schiano, Nardi, Gumbrecht, and Swartz 2004),
mass communication (Kelleher and Miller 2006; Schmidt 2007; Stefanone and
Jang 2007), political science (Adamic and Glance 2005; Drezner and Farrell 2004;
Trammell 2006), sociology (Ali-Hasan and Adamic 2007; Efimova, Hendrick, and
Anjewierden 2005), geography (Gopal 2007), computational linguistics (Arga-
mon, Koppel, Pennebaker, and Schler 2007; Mishne 2005; Schler, Koppel, Arga-
mon, and Pennebaker 2006), linguistics (Herring, Scheidt, Bonus, and Wright
2005; Puschmann 2009), rhetoric (McNeill 2005; Miller and Shepherd 2004;
Myers 2010), psychology (Gurak and Antonijevic 2008; Nowson, Oberlander and
Gill 2005), and organizational studies (Efimova and Fiedler, 2004; Efimova and
Grudin, 2007; Jackson, Yates, and Orlikowski 2007). Figure 1 shows an array of
methodologies used by scholars from a variety of fields. Dependent on the field,
different questions are formulated and different data are used to answer them.
Studies from disciplines such as communication, sociology, and ethnography
regard blogging as a practice and process; accordingly, they tend to focus on the
relations and motivations of bloggers, how they characterize their activity, and on
their interaction with others in sociocommunicative networks, investigating these
aspects by either quantitative or qualitative means. In contrast, studies from con-
tent-centric and technical fields tend to focus on the analysis and classification of
What defines a “personal voice”, and how is “being open for dialogue” stylis-
tically signaled? Is it even legitimate to associate the use of certain linguistic fea-
tures with being conversational? (See Peterson 2011 for a critical discussion of this
question in relation to blogs.) Linguistic research can contribute to the study of In-
In the course of the last decade, blogging has changed from a new practice into
something mainstream, in many regards becoming the first mass instantiation of
user-generated content (UGC) on the web (Schiano et al. 2004; White and Winn
2009). A variety of blogging services such as Blogger, Wordpress, and LiveJournal
make it possible for anyone to run a blog without the need for a personal website. In
addition, companies, schools, universities, and other organizations increasingly
offer blog hosting to their constituents, while the mainstream news media also make
extensive use of blogs integrated into conventional websites. Precise statistics on
the number of blogs worldwide are difficult to assemble, and their relevance should
not be overstated, given that many blogs are abandoned after a short time. Services
such as Technorati tracked more than 70 million individual sources in Spring 2007,
and as of 2009, estimates for the number of blogs in the United States and globally
varied between 100 and 180 million (Kutchera 2008). The number of Internet users
reporting having read a blog at least once or doing so on a regular basis also appears
to be on the rise (Project for Excellence in Journalism 2008).
Technorati’s 2009 State of the Blogosphere report (McLean 2009) found that
the majority of the 2,828 U.S. bloggers surveyed were male, between 18 and 44
years old, and had a relatively high level of income and educational qualifications.
However, user demographics vary considerably by country, region, language, and
individual usage community, making it essential to investigate blogging practices
within their respective cultural contexts (see Schlobinski and Siever [2005], as
well as Russell and Echchaibi [2009], for publications on blogging from an inter-
national perspective). While blogging can be a source of revenue, and bloggers
may be hired as ghostwriters, the most commonly cited reason for blogging given
by those surveyed by Technorati was “personal satisfaction” (76 % of the respon-
dents), and the most popular self-description of the blog’s content was “personal
musings” (53 %). An even higher percentage (83 %) described the content of their
blog as “personal musings” in a smaller study by Viégas (2005). Self-expression
thus historically appears to be a core motivation for the majority of bloggers, not
withstanding phenomena such as “A-list” blogging and professional journalism
via blogs (see also Lenhart and Fox 2006). A possible shift in user demographics
that may also impact how blogs are written was observed in a recent Pew Internet &
American Life study: Fewer teenagers and an increasing number of adults are read-
ings and writing blogs (Lenhart et al. 2010).
The discourse on blogging and bloggers in the mass media is ambivalent. It
highlights both the perceived value and potential of blogs as sources of free ex-
pression and public discussion and their dependence on secondary sources and lack
of accountability (Baltatzis 2006; Niles 2007). Personal blogging is also often mis-
characterized as vain or egocentric when critics apply expectations from perceived
antecedents such as print journalism to blogs in an undifferentiated fashion. Fur-
thermore, the public discourse often implicitly takes highly visible topical blogs
and casts them as typical or model instantiations, effectively imposing prescriptive
norms for how to write a “good” blog and creating a skewed perception of the
mainstream, a problem emphasized by Herring, Kouper, Scheidt, and Wright
(2004).
Stylistically, blogs are a highly variable form of self-expression. Depending on
a blogger’s conceptualization of the format, a blog entry can be speech-like or
written-like, colloquial or formal, and can relate private or public (in the sense of
addressing established topical areas of the mainstream media, such as politics,
sports, entertainment, or technology) issues. It can target a wide audience or a
small, select readership and be written for personal or professional reasons. The
virtual context in which posts are presented and the metadata that is associated
with them (title, author, time of writing, tags, categories) allow for a situatedness
that makes blogs compelling to a wide range of users. This is in part because the
metadata facilitates the use of first-person perspective and deictic expressions,
making writing a blog relatively easy even for an inexperienced author. A range of
stylistic and topical options are at the blogger’s disposal, whereas established and
institutionalized forms of publishing are often more restrictive in terms of style and
content. The flexible communicative situation of blogs (Who is assumed to be
among the readers? Who, if anyone, is addressed? Who might be eavesdropping?)
also shapes their style and influences their authors’ rhetorical choices.
Content-based approaches to blogs have often focused on the question of how
to categorize them on the basis of language and the use of specific features such as
linking, quoting, frequency of comments, or other structural criteria of analysis. In
an attempt to combine the facets of form and function, Krishnamurthy (2002) sug-
gests such a categorization along the scales personal – topical and individual –
community for his analysis of blog discourse (Figure 2). His approach implicitly
uses basic categories from genre analysis (communicative purpose and discourse
community; cf. Swales [1990]), which are descriptively inclusive but also difficult
to operationalize.
Herring et al. (2005) extend the categories suggested by Blood (2002), who
refers to the types filters, personal journals, and notebooks, and propose a catego-
rization of blogs into the following types:
This classification is based on purpose inferred from content (Herring et al. 2005:
147) via manual coding. The categories are defined a priori, and then the studied
blogs are classified according to which category they are seen as fitting in best.
This approach to categorization can be criticized on the grounds that the array of
purposes is limited and based on Blood’s somewhat prescriptive categories, which
take her personal view of blogging as a starting point.
Another language-based approach is taken by Argamon et al. (2007), who in-
vestigate the relationship of style and content to age and gender of bloggers based
on stylometric analysis. Via a process they label meaning extraction (Chung and
Pennebaker 2007), the authors isolate 20 thematic lexical factors that are then re-
lated to the dimensions of age and gender (see Argamon et al. Table 3 for details).
They summarize the findings of their large-scale textual analysis of 140 million
words of running text by remarking:
We find that older bloggers tend to write about externally-focused topics, while younger
bloggers tend to write about more personally-focused topics; changes in writing style
with age are closely related. Perhaps surprisingly, similar patterns also characterize
gender-linked differences in language style. In fact, the linguistic factors that increase in
use with age are just those used more by males of any age, and conversely, those that de-
crease in use with age are those used more by females of any age. (Argamon et al. 2007:
para 4)
Argamon et al. approach blogs inductively and based on automation, while the ap-
proaches used by Herring et al. (2004, 2005) classify the content according to func-
tionally-derived categories and manual classification. Both approaches are based
on the style and content of blog entries and make inferences regarding the purposes
associated with blogging on the basis of the data.
However, Aragmon et al.’s strong claims regarding the influence of age and
gender on style are problematic in several ways. First, they fail to take blog type
into account. In an automated analysis of a corpus of blog entries balanced by
gender (male and female bloggers) and blog type (filter and personal journal), Her-
ring and Paolillo (2006: 453–455) found that blog type trumps gender as an in-
fluencing factor. Their analysis suggests that the gender differences observed by
Argamon et al. are a result of failing to differentiate between different types of blogs
based on purpose and audience design. Herring, Kouper, et al. (2004)’s finding that
teen bloggers are more likely to write personal journals and adults are more likely to
write filter blogs similarly problematizes Argomon et al.’s age results. There ap-
pears to be a broad consensus that in blogs there is a dynamic relationship between
style, genre, age, and gender of the blogger, and behaviors such as linking and quot-
ing, but clear and unambiguous causal relations between these dimensions have not
been established. Males, females, and bloggers of different age groups may all be
motivated by different perceived benefits of blogging. It seems likely that the pur-
pose a blogger associates with his or her blog will have a strong impact on how it is
written, and that at the same time purpose depends partly on age and gender, rather
than there being an immediate relationship between these criteria and language use.
Despite the overall variability of blog style and content, some linguistic properties
of blogs are highly stable. The core cohesive element of a blog is time. In contrast
to the syntagmatic structure of wikis (elements are linked by relation), blog entries
are paradigmatically linked by chronology. This chronology can be omitted, for
example when looking at entries on the basis of categories or tags, or when an in-
dividual post is found via a web search, but it acts as the governing organizational
principle for information in blogs.
According to Winer (2001), blog posts canonically encode the following in-
formation:
1. Title
2. Text
3. Tags/Categories
4. Author
5. Time of publication
6. URL
Author, time of publication, and the URL of the post differ from the other fields by
constituting extra-textual information that is automatically associated with the
situation and not manually assigned by the blogger. This information makes deictic
language possible, i.e., use of the first person pronoun (always referring to the
blogger-publisher credited with the post) and use of temporal adverbs (relating to
post publication date as the point of reference) and spatial adverbs (either relating
to the blogger’s location at the time of publishing or conceptualizing the blog or the
Internet as a whole as a space). Furthermore, older blog entries function as co-text
for newer ones, allowing certain information to be presented as given to the (hy-
pothetical) consistent reader.
Example 1, taken from the blog of Robert Scoble (http://scobleizer.com), a
technology expert and former Microsoft employee, illustrates the rhetorical poten-
tial of blogs. Despite being relatively informal in style, it is also clearly well
planned and informative. It employs standard spelling, proper capitalization, and
no use of emoticons or abbreviations typically found in synchronous CMC, al-
though technical acronyms abound (e.g., HDTV, PVR, OPML).1
Example 1.
Trial of Origin (band first linked here) beats 600 others
This is cool, (hyperlink)Trial of Origin(/hyperlink) just won a BBC contest where they
beat 600 other bands. I was one of the first blogs to link to this band and hope they go all
the way!
Speaking of music, last night Maryam and I watched American Idol on our new soo-
peerr-doopeer 60-inch Sony HDTV screen.
Damn.
That’s all I can think while watching that screen. We have to rip ourselves away from the
TV to go to sleep.
Blogging? Forget it. HDTV wins. It is just so stunning.
Oh, funny aside. Remember my post about visiting Yellowstone National Park? So,
what was on my PVR when I got home? That’s right. Sunrise show of Yellowstone’s
Geysers. They look better on HDTV than they do in real life! They don’t smell, though.
Question: I wanna watch the World Cup. Turns out most of the games are on ESPN2.
That’s not in HD on Comcast. So, what’s the best way to get ESPN2?
Since we’re talking about media, I like (hyperlink)the Podcast Readout of Share Your
OPML site(/hyperlink). Is your favorite Podcast there?
Example 1 can be characterized as a relatively typical blog post as regards the use
of short paragraphs, sentence fragments, interjections, deictic expressions, and
first person point of view. The reader is addressed multiple times, a strategy regu-
larly used by Scoble for involvement.
Two items referred to in example 1 are set off by the use of hyperlinks (Trial of
Origin, the Podcast Readout of Share Your OPML site). A referenced point in time
(last night) is retrievable via the contextual (temporal) information automatically
added by the blogging software, while some information (for example, Maryam,
the name of the blogger’s wife) is treated as given based on an assumed familiarity
of the reader with either the co-text (older blog entries in which she is introduced)
or situational context (for those readers personally acquainted with the Scobles).
The availability of deixis, hyperlinks, and a retrievable co-text partly combines the
situatedness of speech with the permanence and stability of writing. The encoding
of author and time of publication as metadata with each blog entry is the basis for
the frequent analogy of blogging with interactive discourse (Puschmann 2010a –
but cf. Peterson 2011 for a different characterization). The meta-information fixes
a deictic center that allows reference to time and speaker/addressee relative to the
point specified by the blog post.
The previous section argued for the importance of metadata in relation to the com-
position and reception of blogs. While the blogger’s identity and the reference of
spatial and temporal expressions are retrievable by the reader through this meta-
data, the identity of the intended recipients of a blog post is frequently much less
clear. Bloggers are likely to conceptualize blog entries on a continuum between
monologue and dialogue, as they can project their writing onto a variable audience.
This fuzzy communicative setting is discussed in an ethnographic study by Brake
(2007), who observes:
To refer to ‘a-communicative’ practices built around a communications medium seems
counter-intuitive, but the interviews clearly revealed a wide variety of motivations to
start and to continue weblogging that had only a tenuous connection to communication.
(15)
The term “audience” in this context does not, as Brake’s remark underlines,
imply a specific, large, or external readership; functions such as using a blog to
store information for personal use (personal knowledge management or PKM) or
to record personal thoughts in the style of a diary are two uses where the blogger
may have a generic, sympathetic listener, or even herself in mind. Qian and Scott
(2007) and Viégas (2005) studied how bloggers conceptualize their readership and
resolve potential conflicts related to the Internet’s opaque participative setting via
surveys. Qian and Scott argue:
The target audience is related to how much anonymity bloggers perceive themselves to
have. Specifically, the bloggers in our study feel more identifiable if the audience in-
cludes people they know offline. At the same time, target audience also influences the
way posts are written and what information is made available (Schiano et al 2004:
1144). When a blog is for people one knows offline (e.g., family/friends), the goal may
be to identify oneself for them and to gain recognition for one’s ideas from others whose
opinions matter to the blogger. (2007: para 37)
Both studies agree that bloggers plan their writing with a specific (if potentially
very small) audience in mind and that they consider the impact of their writing. In
addition to this conceptualized audience (what Viégas calls the the “core”), a
broader actual audience exists, which includes overhearers not anticipated by the
blogger. As the conceptualized audience is what the blogger relies on when writ-
ing, problems can occur when a conflict of interest exists between conceptualized
and intended audience (see Puschmann 2010a: 74 for an example). An alternative
approach is to rely on the topic in order to get an idea of the audience, since knowl-
edge of and interest in certain topics constrains a blog’s potential readership. This
approach is used in example 1. A filter-style, topical focus makes sense because it
acts as a domain-specific anchoring point for the reader, whereas an author-centric
approach assumes that readers are at least somewhat familiar with the blogger.
Nonetheless, bloggers’ expectations towards their audience frequently go
beyond a mutual interest in similar topics. In her study of adolescent bloggers,
Scheidt (2006) maps the style of blog entries onto projected readers in different
roles:
1. A witness testifying to the experience
2. A therapist unconditionally supporting emotions
3. A cultural theorist assessing the contestation of meanings, values, and iden-
tities in the performance
4. A narrative analyst examining genre, truth, or strategy
5. A critic appraising the display of performance knowledge and skill
Scheidt claims that there are gender differences in the extent to which each type of
audience is addressed by teen writers, noting that the female adolescent bloggers
she studied showed “a tendency toward seeking support for their individual
struggles with appeals to the envisioned audience” (2006: 18), whereas boys show
a tendency toward asking the audience to witness the event and notice their skill in
performance.
From these studies, it can be observed that audience design plays an important
role in blogging and that bloggers integrate their conceptualization of the reader-
ship into their style. It is a widely held assumption that most bloggers address mass
audiences, yet the findings of Brake, Qian and Scott, Viégas, and Scheidt all appear
to suggest otherwise. Brake’s description of blogs as “a-communicative” and Qian
and Scott’s characterization of diary blogs as secret or “near-secret” writings,
aimed at very small audiences or even the self, particularly serve to emphasize this
discrepancy between popular perception and everyday use.
In less personal and more goal-oriented blogs, audience design and addressiv-
ity may be part of an elaborate rhetorical strategy. The use of explicit second-per-
son address, for example, allows bloggers to negotiate complex relationships by
shifting pronominal focus, in extreme cases from one clause to the next. Jonathan
Schwartz, former CEO of Sun Microsystems, addresses both employees and exter-
nal stakeholders in his blog in Example 2, illustrating this potential (personal pro-
nouns and references to the company Sun are in bold print to highlight their preva-
lence).
Example 2.
We Think We Can
Paraphrasing Henry Ford, “You think you can, or you think you can’t – either way,
you’re right.” That quote struck me as the perfect summary of our fiscal year 2007 per-
formance. We did what we said we’d do a year ago.
As you may have seen, we’ve announced our fourth quarter and full fiscal year results
(our fiscal year ends with the school year, in June). We exceeded the commitments
made a year ago, to restore Sun to 4 % operating profitability in Q4, and did so by de-
livering our single best operational quarter since 2001. On an annual basis, we im-
proved Sun’s profitability by over a billion dollars. A billion. We grew revenue, ex-
panded gross margins, streamlined our operating expenses – and closed the year with an
8 % operating profit in Q4, more than double what some thought to be an aggressive tar-
get a year ago.
We did this while driving significant product transitions, going after new markets and
product areas, and best of all, while aggressively moving the whole company to open
source software (leading me to hope we can officially put to rest the question, “how will
you make money?”).
And we’re not done – not by any stretch of the imagination. We have more streamlining
to do, more commitments to meet, more customers to serve and developers to attract. But
it’s evident we’ve got the right foundation for growing Sun – with real innovation the
market values, as shown by Q4’s 47 % gross margins, the highest on record in five years.
I’ll be with a variety of external audiences most of this week – and I’ll summarize their
questions and comments in a few days. In the interim … to our customers, partners, and
most of all, our amazing global employee base – thank you for thinking we could.
You were right.
Keep thinking that way. You ain’t seen nothin’ yet.
Example 2 is clearly “more written” in style than example 1. Sentences and para-
graphs are longer and more complex, and there are fewer interjections, less ellipsis,
and limited, strategic use of orality markers (You ain’t seen nothin’ yet). As in
example 1, first person point of view and deictic expressions (a year ago, next
week) are used, but in contrast to Scoble, Schwartz addresses his readers more di-
rectly via second person pronouns.
Schwartz uses we with rhetorical extension to refer to all employees of the
company, including himself. But use of we in this function does not prevent him
from using I to present himself as distinct from everyone at Sun or to use the full
noun Sun as denoting something other than we. Instead, an intricate interplay of
discourse roles is developed (see also Puschmann 2010b):
– We, used as subject with a range of dynamic verbs of action and movement and
semantically as agent, denotes the employees of Sun Microsystems (or, in a
more restrictive interpretation, the management team).
– I is used in contexts where agency of the institutional we would not result in a
6. Politeness
The observations made above about how bloggers conceptualize their audience
and negotiate their relationship with it gain additional traction in relation to the
concepts of face (Goffman 1955) and politeness (Brown and Levinson 1987), i.e.,
the desire to construct, uphold, and reinforce a positive image of the self, while at
the same time being aware that our positive and negative face work brings us into
conflict with the face work of others, leading to face-threatening acts (FTAs).
Politeness is a central strategy for minimizing FTAs (Mills 2003: 6) and thus for
demonstrating awareness and understanding of the need of others to construct
themselves socially. The role of politeness in blogs has been investigated by
Garcia-Gomez (2009), who studied the conceptualization of gender roles as ex-
pressed in the blogs of British and Spanish teenage girls, and by Kouper (2010),
who explored the pragmatics of advice-giving, which she characterizes as an FTA,
in a LiveJournal community.
While a blog can serve as an effective tool for active face work from the per-
spective of the blogger, the capacity for FTAs is somewhat constrained by the fact
that blogger and audience are relatively removed from each other. Inside the space
of her own post, the blogger is in charge – she cannot be interrupted, pressured,
threatened, or interrogated except via comments, over which she has a high degree
of control. Likewise, no one person can be assumed with absolute certainty to be
among the readers, therefore no reader can plausibly be interrupted, pressured,
threatened, or interrogated by the blogger (but see example 4 for a case where as-
sumed readers are addressed directly). Thus face threat understood in a way that
presumes parties who are aware of each other’s presence does not really apply to
blog posts, which may serve to explain their popularity as a personal space that is
both visible to others and at the same time protected from undesired intervention.
While bloggers quote and link to each other and engage in other activities that
make FTAs possible, one’s own blog offers a certain degree of safety compared to
forms of CMC that mandate interaction.
As a result of what Brake (2007) refers to as the “a-communicative” approach
to blogging, authors have the opportunity to express what could otherwise be in-
terpreted as face threatening. Accounts of how blogs are conceptualized in an in-
teractional context support this notion, as a 2004 study of bloggers’ expectations
by Gumbrecht suggests:
In face-to-face conversation or IM, responses are expected immediately or close to it.
As a result, conversational partners may feel ill at ease when trying to broach a sensitive
issue in these media. Lara said she would never tell people, “I’m really sad” in IM, yet
she would have no qualms about stating it in her blog. Why? In blogs, people can choose
to respond to a post or not – it is up to the reader. Over IM, conversational partners might
feel obligated to respond – especially when the subject matter is as heavy as a depressed
emotional state. In this way, the lack of both cotemporality and simultaneity factor into a
blogger’s choice of communicative medium. (2004: 3)
The risk of a social penalty still looms when reacting in a way that can be inter-
preted as face-threatening, but the relative temporal and spatial distance between
blogger and reader (in the virtual sense of not perceiving who else is “present”)
makes certain types of FTAs implausible. In Gumbrecht’s study, this is anticipated
by Lara, who prefers to articulate personal feelings in her blog rather than in chat.
Those of her friends who do not want to articulate a reaction to Lara’s feelings can
refrain from doing so without seeming insensitive. Blog readers can lurk without
indicating that they have even received a message, and the blogger can likewise
claim not to have addressed anyone in particular with her thoughts. Consequently,
the kind of conversational offering that happens in blogs effectively removes the
social pressure from the communicative partners that accumulates in a push-type
mode of CMC such as email, where sending a message places pressure on the re-
cipient to respond.
As has been argued in the previous sections, blogs combine a number of medial
characteristics previously associated with distinct forms of public and private dis-
course. Findings such as those of Argamon et al. (2007) regarding the stylistic
properties of blogs are suggestive of changes over the lifespan, as bloggers’ needs
for different forms of expression change and expectations towards readers shift.
With relatively few exceptions, a blog is a controlled discourse environment be-
longing to an individual and shaped largely by his or her personal tastes and needs;
therefore, the needs a blog fulfills are more individually shaped that in most other
genres of public expression. While readers can participate via comments and track-
backs, this occurs only if the blogger allows it. What is recorded in a blog is per-
manent and (again, with few exceptions) public, while at the same time being di-
rected entirely by the blogger. These situational characteristics can be considered
in concert with stylistic properties to make an argument for two distinct approaches
to blogging: a topic-centric style and an author-centric style. These styles follow
the basic distinction between diary-style and filter-style content proposed by Her-
ring et al. (2005), but they attempt to capture purpose via the relationship between
blogger and reader that is implicit in the text, rather than directly via content. A dis-
cussion of two examples illustrates this.
Example 3. (topic-centric blogging style)
“Cuil it?” I Don’t Think So.
When I teach trademark law classes, I always advise that students select strong protect-
able marks, and the class invariably balks because they want to select marks that suggest
or connote something about the goods or services at issue. That, I tell them, is the touch-
stone of a weak mark, and for examples I look to the Internet space: Google, Yahoo!,
Zillow, and so on are perfect trademarks because they say nothing about the goods or
services with which they are associated.
And now along comes cuil.com (pronounced “cool”), the much-ballyhooed Google-
killer. Great mark, right? “Cuil” says nothing about “Internet search engine,” and is in
fact apparently an old Irish or Gaelic word for “knowledge.” But here’s the rub:
“Google” is becoming a verb in the lexicon very quickly, which is typically anathema to
a trademark, but there’s not much Google can do to stop everyone from saying, e.g., “Go
Google that.” But can you say, e.g., “I am going to ‘cuil’ it?” You could, but people
would hear you say, “I am going to cool it,” and the meaning is lost.
Moral – a great trademark has to be both non-descriptive AND sound cool (pun in-
tended) and distinctive. Now let’s just see if Google goes the way of “escalator” and be-
comes generic for Internet search services …
My birthday
Well, it was my first birthday away from the family. It was interesting … teasin! it wasnt
as bad as i thought and i practiced for my mission. I woke up and texted sis because she
was born on my b-day, for those of you who forogt. Teasin sis! Umm i talked to mom
and dad and the rest of the family throughout the day. Thanks for the calls everyone.
Uhh … i went to breakfast (the good ol Galley) with Jane and Mike and afterwards ran
home to find a package sent from Mom and the famoly in cali. I went to a basketball
game and we won. Then i went out to dinner with a bunch of friends and ate hecka food.
Afterwards we came back to Jane’s and opened presents from Jane and Paul. Paul gave
me a journal his mom gave him … that was special and i needed one so im thankful for
it. Jane gave me the Spanish Book of Mormon. I was pretty excited cause now im gonna
be fluent in Espanol! She also gave me an awesome CTR ring because I have been wear-
ing Paul’s old one for the last year. Umm it also spins so it is way tight.
Then I came back to my dorm and Paul and I opened the rest of my presents. The
Johnsons gave me some nice notes and hecka food which is gettin me fat. haha thanks!
O and thanks Sue for the Head and Shoulder bottle. You’ll be gettin it back sometime
haha! And then I opened the fam in Cali’s package. I got a sweet shirt that had my
Mission on the back. Linda gave me the hymn book in Spanish, its pocket sized. And
then a CD and hecka more food. Grandma also sent a letter. Shes awesome! And that
was my b-day! This might be boring sorry!
Examples 3 and 4, both taken from personal blogs with small audiences (rather
than “pro” blogs), share certain properties: first-person voice and deictic ex-
pressions are used in both posts, in combination with particles and connectives (dis-
course markers) typically used in spoken language or in writing that consciously
imitates speech. Both involve the reader directly via use of direct address (example
3: let’s, you; example 4: you, everyone) and indirectly by using oral style markers.
Example 3 is more conservative and follows stylistic norms for written language
more closely, while example 4 is more progressive in its use of expressions typi-
cally found in online speech production, such as fillers, reductions, and phatic ex-
clamatives (uhh, umm, hecka, teasin, gettin, b-day, haha). The use of fillers in
example 4 runs counter to the claim that blog entries are spontaneous and un-
planned (Crystal 2006: 15), as there is no cognitive need in a blog entry to buy time
while formulating an utterance. Rather, the fillers in this example make the text ap-
pear more speechlike and authentic, which can be assumed to conform with the
blogger’s intentions and allows inferences regarding his conceptualized audience.
The examples differ from one another in that example 3 is written in the style of
an exposition: It presents an argument from the point of view of the blogger and
offers his informed opinion on an aspect of trademark law. Example 4, by contrast,
is written in the style of a personal oral narrative, relating a series of events to the
reader. Crucially, the names and events in example 4 are not readily interpretable to
a random reader, but only to those who are at least somewhat familiar with the
author. The author of example 3 consciously provides contextual information in
order to make his writing more accessible to a non-familiar audience – for
example, by beginning his post with the relative clause When I teach trademark
law classes instead of the main clause I always advise that students select strong
protectable marks. His intended audience is therefore likely to extend beyond his
students (to whom he refers in the third person) to a general readership interested in
trademark law, while the conceptualized readership in example 4 seems more spe-
cific, including the blogger’s sister (Teasin sis!) and those people who called to
congratulate him for his birthday (Thanks for the calls everyone). While the
opinions articulated in example 3 are personally associated with the blogger, they
are relatively independent of any single event, making them useful to a generic
reader (assuming the reader is interested in the subject matter) regardless of when
exactly the post is read or whether blogger and reader are personally acquainted. In
contrast, the event related in example 4 is more specified not only in terms of audi-
ence, but also in terms of the time of its anticipated reception. It is important to
note that the relevance of both examples depends entirely on the identity of the
reader – example 4 is without doubt important to the intended audience of family
and friends, which could well be larger than the readership of the trademark blog.
Examples 3 and 4 represent two individual styles among countless possible ap-
proaches to blogging. Not only do the styles vary considerably across authors, but
the style of a single blogger can vary over longer periods of time, from one post to
the next, and even within a single entry. An author-centric blog post might conform
to standard orthography and be stylistically in line with writing prescriptions,
while a topic-centric blog one might not. The choice of topic and degree of cooper-
ation with the reader (in terms of propositional explicitness; cf. Grice 1989) indi-
cate the tendencies of topic-centric and author-centric blogging, but this is far from
a black and white distinction. Table 1 summarizes the prototypical characteristics
of the two styles.
Fluidity of purpose (and, resulting from this, fluidity of style and content) is en-
demic to blogs, and therefore a wide area of intermediate forms of use lies between
the two extremes. The distinction between topic-centric and author-centric styles is
not intended as a clear-cut system of categorization, but rather as a way of system-
atizing the different audiences and intentions that bloggers associate with their
activity. The distinction can be seen as an extension of the categories proposed
by Blood (2002) and Herring et al. (2005), but with emphasis on the relationship
of blogger, assumed reader, and purpose, rather than on developing a system of
content classification. Findings on gender and age (Herring, Kouper, et al. 2004;
Huffaker and Calvert 2005) can be associated with the distinction: older males tend
to blog more topic-centricly, while females and teenagers of both genders more fre-
quently choose an author-centric style. Rather than concluding that age and gender
simplistically impact language use, it instead seems sensible to argue that users of
different genders and age groups pick styles appropriate to their envisioned target
audience (ranging from the self to a broad public) and their communicative goals
(ranging from self-expression to the accumulation of social capital) and then write
accordingly.
8. Outlook
Weblogs have moved from a niche format to a well-known and thoroughly inter-
national genre of computer-mediated communication in less than a decade. They
continue to thrive as personal publishing platforms, while newer, even more rapid
forms of “lifelogging” such as Facebook status messages and microblogging ser-
vices such as Twitter may be overtaking them as tools for self-documentation.
Multimedia blogging on platforms such as YouTube continues to thrive, showing
that a need for mediated self-expression extends well beyond writing. Blogs are
written in an ever-increasing number of languages, calling for more research on
blogs in languages other than English (cf. Russell and Echchaibi 2009; Schlobinski
and Siever 2005). Linguistic analysis of blogs reveals a wide variety of uses and
contexts, with potential for research that, while using language as the diagnostic, is
able to address questions that go beyond structural linguistics. A challenge for re-
searchers investigating blogs continues to be striking the right balance between
text and practice, especially since large-scale analyses based on content are ever
easier to conduct in the age of cheap and ubiquitous computational resources. Such
imbalances can be avoided by incorporating findings from neighboring disciplines
such as communication, ethnography, sociology, and social psychology, using
them to contextualize language-based approaches.
This chapter has sought to characterize blogs as a form of mediated language
use that affords its users a range of communicative options. Blogs dynamically
combine characteristics of speech and writing due to their format, mode of produc-
tion, and the communicative situation they create via the encoding of metadata.
This makes them a versatile tool for expression and communication that is likely to
continue to interest scholars from a variety of disciplines in the future.
Note
1. HD = High Definition, HDTV = High Definition Television, PVR = Personal Video Re-
corder, OPML = Outline Processor Markup Language. BBC and ESPN2 are British and
American TV stations.
References
Medier. Livet mellom skjermene (Personal Media. Life Between the Screens),
141–163. Oslo: Gyldendal.
Brown, Penelope and Stephen C. Levinson
1987 Politeness: Some Universals in Language Usage. Studies in Interactional So-
ciolinguistics 4. Cambridge/New York: Cambridge University Press.
Bruns, Axel and Joanne Jacobs
2006 Uses of Blogs. New York: Peter Lang.
Chung, Cindy and James W. Pennebaker
2007 The psychological function of function words. In: Klaus Fiedler (ed.), Social
Communication, 343–359. New York: Psychology Press.
Crystal, David
2006 Language and the Internet. London: Cambridge University Press.
Drezner, Daniel and Henry Farrell
2004 The power and politics of blogs. Paper presented at the American Political
Science Association, Chicago, IL, September 2–5.
http://www.utsc.utoronto.ca/~farrell/blogpaperfinal.pdf
Efimova, Lilia and Sebastian Fiedler
2004 Learning webs: Learning in weblog networks. In: Proceedings of the IADIS In-
ternational Conference Web Based Communities, 490–494. Vienna: INSO
Publications.
Efimova, Lilia and Jonathan Grudin
2007 Crossing boundaries: A case study of employee blogging. In: Proceedings of
the 40th Annual Hawaii International Conference on System Sciences, 86–99.
Los Alamitos, CA: IEEE Press.
Efimova, Lilia, Stephanie Hendrick, and Anjo Anjewierden
2005 Finding “the life between buildings”: An approach for defining a weblog com-
munity. In: Proceedings of Internet Research 6.0: Internet Generations.
https://doc.novay.nl/dsweb/Get/File-55092/AOIR_blog_communities.pdf
Ewins, Rory
2005 Who are you? Weblogs and academic identity. E-Learning 2(4): 368–377.
Garcia-Gomez, Antonio
2009 Teenage girls’ personal weblog writing: Truly a new gender discourse? In-
formation, Communication & Society 12(5): 611–638.
Goffman, Erving
1955 On face-work; an analysis of ritual elements in social interaction. Psychiatry
18(3): 213–231.
Gopal, Sucharita
2007 The evolving social geography of blogs. In: Harvey J. Miller (ed.), Societies
and Cities in the Age of Instant Access, 275–293. Berlin: Springer.
Grice, H. Paul
1989 Studies in the Way of Words. Cambridge, MA: Harvard University Press.
Gumbrecht, Michelle
2004 Blogs as “protected space”. In: Proceedings of the WWW 2004 Workshop on
the Weblogging Ecosystem: Aggregation, Analysis and Dynamics. New York.
http://www.blogpulse.com/papers/www2004gumbrecht.pdf
Gurak, Laura J. and Smiljana Antonijevic
2008 The psychology of blogging: You, me, and everyone in between. American Be-
havioral Scientist 52(1): 60–68.
Herring, Susan C., Inna Kouper, Lois A. Scheidt, and Elijah L. Wright
2004 Women and children last: The discursive construction of weblogs. In: Laura J.
Gurak, Smiljana Antonijevic, Laurie Johnson, Clancy Ratliff, and Jessica Rey-
man (eds.), Into the Blogosphere: Rhetoric, Community, and Culture of Web-
logs [Internet volume].
http://blog.lib.umn.edu/blogosphere/women_and_children.html
Herring, Susan C. and John C. Paolillo
2006 Gender and genre variation in weblogs. Journal of Sociolinguistics 10(4):
439–459.
Herring, Susan C., Lois A. Scheidt, Sabrina Bonus, and Elijah L. Wright
2004 Bridging the gap: A genre analysis of weblogs. In: Proceedings of the 37th An-
nual Hawaii International Conference on System Sciences, 101–111. Los
Alamitos, CA: IEEE Press.
Herring, Susan C., Lois A. Scheidt, Sabrina Bonus, and Elijah L. Wright
2005 Weblogs as a bridging genre. Information Technology & People 18(2):
142–171.
Honeycutt, Courtenay and Susan C. Herring
2009 Beyond microblogging: Conversation and collaboration via Twitter. In: Pro-
ceedings of the 42nd Hawaii International Conference on System Sciences,
1–10. Los Alamitos, CA: IEEE Press.
Huffaker, David. A. and Sandra L. Calvert
2005 Gender, identity, and language use in teenage blogs. Journal of Computer-Me-
diated Communication 10(2), article 1.
http://jcmc.indiana.edu/vol10/issue2/huffaker.html
Jackson, Anne, JoAnne Yates, and Wanda Orlikowski
2007 Corporate blogging: Building community through persistent digital talk. In:
Proceedings of the 40th Annual Hawaii International Conference on System
Sciences, 80–97. Los Alamitos, CA: IEEE Press.
Java, Akshay, Xiodan Song, Tim Finin, and Belle Tseng
2007 Why we twitter: Understanding microblogging usage and communities. In:
Proceedings of the 9th WebKDD and 1st SNA-KDD 2007 Workshop on Web
Mining and Social Network Analysis, 56–65. San Jose, CA: ACM.
Joinson, Adam N.
2008 Looking at, looking up or keeping up with people?: Motives and use of face-
book. In: Proceeding of the Twenty-Sixth Annual SIGCHI Conference on
Human Factors in Computing Systems, 1027–1036. Florence, Italy: ACM.
Karlsson, Lena
2006 Acts of reading diary weblogs. Human IT 8(2): 1–59.
Kelleher, Tom and Barbara M. Miller
2006 Organizational blogs and the human voice: Relational strategies and relational
outcomes. Journal of Computer-Mediated Communication 11(2), article 1.
http://jcmc.indiana.edu/vol11/issue2/kelleher.html
Kouper, Inna
2010 The pragmatics of peer advice in a LiveJournal community. Language@Inter-
net 7, article 1.
http://www.languageatinternet.org/articles/2010/2464/index_html/
Krishnamurthy, Sandeep
2002 The multidimensionality of blog conversations: The virtual enactment of Sep-
tember 11. Paper presented at Internet Research 3.0, Maastricht, The Nether-
lands, October.
Kutchera, Joe
2008 Navigate the blogosphere’s biggest ad networks. iMediaConnection.com. Sep-
tember 12. http://www.imediaconnection.com/content/20486.asp
Lenhart, Amanda and Susannah Fox
2006 Bloggers: A portrait of the Internet’s new storytellers. Pew Internet & American
Life Project, July 19. http://www.pewinternet.org/Reports/2006/Bloggers.aspx
Lenhart, Amanda, Kirsten Purcell, Aaron Smith, and Kathryn Zickuhr
2010 Social media and young adults. Pew Internet & American Life
Project, February 3. http://www.pewinternet.org/Reports/2010/Social-Media-
and-Young-Adults.aspx
Lomborg, Stine
2009 Navigating the blogosphere: Towards a genre-based typology of weblogs.
First Monday 14(5). http://firstmonday.org/htbin/cgiwrap/bin/ojs/index.php/
fm/article/viewArticle/2329/2178
McLean, Jennifer
2009 State of the blogosphere 2009. Technorati.com, October 19.
http://technorati.com/blogging/feature/state-of-the-blogosphere-2009/
McNeill, Laurie
2003 Teaching an old genre new tricks: The diary on the Internet. Biography 26(1):
24–47.
McNeill, Laurie
2005 Genre under construction: The diary on the Internet. Language@Internet 2, ar-
ticle 1. http://www.languageatinternet.org/articles/2005/120
Merriam-Webster
2004 Merriam-Webster announces 2004 words of the year. Merriam-Webster On-
line, November.
http://www.merriam-webster.com/info/pr/2004-words-of-year.htm
Miller, Carolyn R. and Dawn Shepherd
2004 Blogging as social action: A genre analysis of the weblog. In: Laura J. Gurak,
Smiljana Antonijevic, Laurie Johnson, Clancy Ratliff, and Jessica Reyman
(eds.), Into the Blogosphere: Rhetoric, Community, and Culture of Weblogs
[Internet volume]. http://blog.lib.umn.edu/blogosphere/blogging_as_social_
action_a_genre_analysis_of_the_weblog.html
Miller, Carolyn R. and Dawn Shepherd
2009 Questions for genre theory from the blogosphere. In: Janet Giltrow and Dieter
Stein (eds.), Genres in the Internet: Issues in the Theory of Genre, 263–290.
Amsterdam/Philadelphia: John Benjamins.
Mills, Sara
2003 Gender and Politeness. Studies in Interactional Sociolinguistics 17. Cam-
bridge/New York: Cambridge University Press.
Mishne, Gilad
2005 Experiments with mood classification in blog posts. In: Proceedings of ACM
SIGIR 2005 Workshop on Stylistic Analysis of Text for Information Access.
http://staff.science.uva.nl/~gilad/pubs/style2005-blogmoods.pdf
Myers, Greg
2010 The Discourse of Blogs and Wikis. London: Continuum.
Nardi, Bonnie, Diane J. Schiano, and Michelle Gumbrecht
2004 Blogging as social activity, or, would you let 900 million people read your
diary? In: Proceedings of the 2004 ACM Conference on Computer Supported
Cooperative Work, 222–231. Chicago, IL: ACM.
Niles, Robert
2007 Are blogs a ‘parasitic’ medium? OJR: Online Journalism Review, March 2.
http://www.ojr.org/ojr/stories/070301niles/
Nilsson, Stephanie
2003 The function of language to facilitate and maintain social networks in research
weblogs. Unpublished manuscript, Umeå Universitet, Engelska lingvistik.
http://www.eng.umu.se/Stephanie/web/LanguageBlogs.pdf
Nowson, Scott, Jon Oberlander, and Alastair J. Gill
2005 Weblogs, genres and individual differences. In: Proceedings of the 27th An-
nual Conference of the Cognitive Science Society, 1666–1671.
Project for Excellence in Journalism
2008 The state of the news media 2008: An annual report on American journalism.
http://stateofthemedia.org/2008/
Peterson, Eric E.
2011 How conversational are weblogs? Language@Internet 8, article 8.
http://www.languageatinternet.org/articles/2011/Peterson
Puschmann, Cornelius
2009 Lies at Wal-Mart: Style and the subversion of genre in the Life at Wal-Mart
blog. In: Janet Giltrow and Dieter Stein (eds.), Genres in the Internet: Issues in
the Theory of Genre, 49–84. Amsterdam/Philadelphia: John Benjamins.
Puschmann, Cornelius
2010a The Corporate Blog as an Emerging Genre of Computer-Mediated Communi-
cation: Features, Constraints, Discourse Situation. Göttinger Schriften zur In-
ternetforschung. Göttingen: Universitätsverlag Göttingen.
Puschmann, Cornelius
2010b “Thank you for thinking we could”? – use and function of interpersonal pro-
nouns in corporate web logs. In: Heidrun Dorgeloh and Anja Wanner (eds.),
Approaches to Syntactic Variation and Genre,167–194. Berlin/New York: De
Gruyter Mouton.
Qian, Hua and Craig R. Scott
2007 Anonymity and self-disclosure on weblogs. Journal of Computer-Mediated
Communication 12(4), article 14. http://jcmc.indiana.edu/vol12/issue4/qian.html
Rosenberg, Scott
2009 Say Everything: How Blogging Began, What It’s Becoming, and Why It
Matters. New York: Crown.
Russell, Adrienne and Nabil Echchaibi
2009 International Blogging: Identity, Politics and Networked Publics. New York:
Peter Lang.
Scheidt, Lois A.
2006 Adolescent diary weblogs and the unseen audience. In: David Buckingham
and Rebekah Willett (eds.), Digital Generations: Children, Young People and
New Media, 193–210. London: Lawrence Erlbaum.
Scheidt, Lois A.
2009 Diary weblogs as genre. Indiana University School of Library & Information
Science doctoral qualifying paper.
http://professional-lurker.com/linked/2008/quals/diary_weblog_genre.pdf
Schiano, Diane J., Bonnie Nardi, Michelle Gumbrecht, and Luke Swartz
2004 Blogging by the rest of us. In: Conference on Human Factors in Computing
Systems, 1143–1146.
http://home.comcast.net/~diane.schiano/CHI04.Blog.pdf
Schler, Jonathan, Moshe Koppel, Shlomo Argamon, and James W. Pennebaker
2006 Effects of age and gender on blogging. In: AAAI 2006 Spring Symposium on
Computational Approaches to Analysing Weblogs (AAAI–CAAW), 199–205.
Schlobinski, Peter and Torsten Siever
2005 Sprachliche und textuelle Aspekte in Weblogs. Ein internationales Projekt.
Networx, 46. http://www.mediensprache.net/de/networx/docs/networx-46.asp
Schmidt, Jan
2007 Blogging practices: An analytical framework. Journal of Computer-Mediated
Communication 12(4), article 13.
http://jcmc.indiana.edu/vol12/issue4/schmidt.html
Sprague, Robert
2007 Business blogs and commercial speech: A new analytical framework for the
21st Century. American Business Law Journal 44(1): 127–159.
Stefanone, Michael A. and Chyn-Yang Jang
2007 Writing for friends and family: The interpersonal nature of blogs. Journal of
Computer-Mediated Communication 13(1), article 7.
http://jcmc.indiana.edu/vol13/issue1/stefanone.html
Suzuki, Renata
2004 Diaries as introspective research tools: From Ashton-Warner to blogs.
TESL-EJ 8(1). http://www-writing.berkeley.edu:16080/TESl-EJ/ej29/int.html
Swales, John
1990 Genre Analysis: English in Academic and Research Settings. London: Cam-
bridge University Press.
Trammell, Kaye D.
2006 Blog offensive: An exploratory analysis of attacks published on campaign blog
posts from a political public relations perspective. Public Relations Review
32(4): 402–406.
Viégas, Fernanda B.
2005 Bloggers’ expectations of privacy and accountability: An initial survey. Jour-
nal of Computer-Mediated Communication 10(3), article 12.
http://jcmc.indiana.edu/vol10/issue3/viegas.html
White, David and Phillip Winn
2009 State of the blogosphere 2008. Technorati.com, August 21.
http://technorati.com/blogging/feature/state-of-the-blogosphere-2008/
Winer, David
2001 The history of weblogs. UserLand Software, September 9.
http://www.userland.com/theHistoryOfWeblogs
1. Introduction
ation ripe for investigation (e.g., Androutsopoulos and Hinnenkamp 2001; Paolillo
2001; Siebenhaar 2006; Zelenkauskaite 2004). Codeswitching commands con-
siderable attention in chat research, as the research contexts are often international,
multilingual, and multicultural.
The linguistics and pragmatics literature often exhibits a somewhat coarse-
grained approach to chat. It is common to find general statements about chat lan-
guage, as if it were one monolithic entity (e.g., Crystal 2001; Ferrara et al. 1991).
However, as we explain in this chapter, chat has many different forms, with differ-
ent technical and social characteristics, which may be distinguished as different
modes of chat (Murray 1988; Herring 2001, 2002), i.e., conventional clusters of
social and technical features around which people orient. Different chat modes have
different pragmatic properties in a way that depends on the technical features of the
mode. The linguistic and other social science literature related to chat revolves
around the technical contributions to the types of interaction observed; indeed, it is
impossible to ignore the issues raised by technical features of chat. Hence, the ex-
position in this chapter foregrounds technical features of different chat modes as a
starting point for the framing of pragmatic research questions about chat.
In this chapter, we survey the current and past literature on chat, aiming to
highlight its pragmatics, to consolidate the evidence explaining the nature of chat,
and to indicate gaps in scholarly understanding needing to be addressed by future
research. We focus exclusively on multiparticipant text-based chat, as opposed to
instant messaging (discussed by Baron in this volume) or multimedia chat, and em-
phasize the central role of IRC as a historical reference point for different chat
modes.
This chapter is organized as follows. In the next section, we consider what chat
CMC modes have in common, following which we discuss specific modes of chat
in historical, social, and technical contexts to illustrate how its pragmatics varies
according to mode. In section three, we describe several pragmatic phenomena
associated with chat. Finally, we synthesize these findings to generalize about the
pragmatics of chat and identify a range of research questions that remain open for
investigation.
the sending user. A chat display thus superficially reads like a dramatic script, a
format that easily represents multiparticipant as well as dyadic exchanges.
Chat is typically organized according to “chat rooms” or “chat channels”,
which delineate the bounds of participation. Users present in the room can partici-
pate; those not present or excluded from the room cannot. Chat rooms can be either
moderated, in which case moderators with special privileges control participation
in the chat room, or unmoderated. Chat rooms typically have an identifying title or
name, possibly indicating a specific geographic region, ethnicity, sub-cultural
group, hobby, fan group, or other affiliation. The name advertises a chat room to
potential participants with appropriate interests. A room may also have a “topic of
the day”, but the existence of a topic does not necessarily imply that users spend
time discussing it, and topics are sometimes used for other purposes, such as teas-
ing (e.g., Erickson 2000).
Chat participants have various options in representing their identities. Because
chat systems seldom require authentication, and users can choose any name for
themselves, users can easily conceal their off-line identities. Often users have
multiple nicknames, thus further obscuring identity. Anonymity may make users
more willing to reveal personal information; other times, play with identity permits
experimentation for personal or merely ludic reasons (Danet 1998, 2001; Danet,
Ruedenberg-Wright, and Rosenbaum-Tamari 1995).
The textual nature of chat means that information otherwise conveyed non-
verbally, such as intonation, facial expressions, and proxemics, must also be repre-
sented as text. Hence, chat uses textual paralinguistic cues, including repetition of
characters for emphasis (e.g., “nooooo!”), emoticons or smileys (e.g., “:-)” or “:)”
representing a rotated smiling face; “O.O” representing a wide-eyed stare), ASCII
art (e.g., “@>--,-’- ->--” and similar “rose” variants, often “given” to others; these
are generally assigned to hotkeys or pasted in from files), as well as specialized ab-
breviations (e.g., “lol” for “laugh out loud”). At times, these strategies are com-
bined: “lolololololololol”, for example, represents a sustained burst of laughter.
These features of chat language were among the earliest to receive scholarly atten-
tion (Danet et al. 1995; Werry 1996).
Chat communication thus encourages informal, spontaneous communication,
while the presence of multiple participants makes it more public than private.
Hence, chat interaction has much in common with casual social interaction at par-
ties or public settings. At the same time, its unusual structure and requirements for
interaction management, and the wide variation in its contexts and content, make
chat pragmatically complex.
Chat’s historical roots go back to the early 1970s, when timesharing computing
systems first became widespread. Such systems generally provided programs for
simultaneously connected users to communicate with one another. Early examples
of these facilities are programs like the Unix command write, which, as its name
suggests, writes a message directly to another user’s screen, and Unix talk and
directed speech 4 ls [to Karen]: Hi. I just walked you here at your request,
since you’re in the car and nowhere near a computer on
the net.
emote 5 Penfold whuggles Karen.
emote 6 Tom eyes Karen warily.
emote 7 lynn eyes Karen and ls warily.
say 8 Tom says, “WHY”
say 9 ls says, “she, uh, thought it would be cool to hang out
with you guys.”
Example (1) illustrates the three most important messages types in chat. Line 1
represents a system message, triggered in response to events beyond any user’s im-
mediate control, as when someone enters of leaves a room. Lines 8 and 9 represent
the “say” command, used for normal communications among users, and lines 2, 3,
5, 6, and 7 represent “emotes”, also called “actions”. Emotes permit users to de-
scribe imagined activities in a third-person perspective, as in written fiction, and
the user’s text is displayed in a form distinct from that of “say”. Says and emotes
have different saliencies, and emotes are involved in many linguistic innovations,
e.g., those affecting the semantics of present tense inflection (Cherny 1999: 199ff).
Most chat systems support these three types of messages, though they may convey
them differently. Line 4 represents a directed speech command found in ElseMOO,
but not supported by all chat systems.
Many commands simplify typing commonly-used text through substitution
of ritualized elements. For example, in an emote, the user types “emote waves” (or
its shorthand equivalent “:waves”); the system substitutes the user’s name for the
command, so that it comes out “lynn waves”, as in line 2. (On the performative na-
ture of this usage, see Virtanen, this volume.) Since all commands produce text
after some suitable computation, MUDs and MOOs can permit extending the com-
mand set via “scripts” written in a computer programming language. Objects are
programmed in much the same way, as repositories of textual descriptions dis-
played using common commands (such as “get” and “drop”) or commands specific
to that object. Consequently, in MUDs and MOOs, but also in other chat environ-
ments, programming skill can be an important social resource, as it is necessary for
customizing the environment. The programmed nature of the text can also have
effects of its own on the nature of the language used in the chat environment
(Cherny 1999).
MUD and MOO culture is highly diverse, and the community has some sharp
social boundaries, even where they are regularly traversed (Cherny 1999: 33–38).
ElseMOO even has an abbreviation “tinflm” (“this is not fucking LambdaMOO”)
that is used to identify such boundaries of acceptable activity (LambdaMOO is a
large MOO originally established as a research project at Xerox PARC; ElseMOO
was formed by participants from LambdaMOO who tired of certain interactions
there, resulting in much sensitivity to boundaries as well as effort for their main-
tenance). In addition to the distinction between gaming and social interaction,
community divisions are found around other issues, such as the kind of fantasy per-
mitted. For example, in ElseMOO, characters with non-human descriptions were
prohibited, but these were common on LambdaMOO (Cherny 1999: 61). Lin-
guistic practices also vary within the community, and strong sanctions may be im-
posed for modes of expression marked as “outside” of whatever local group is con-
cerned (Cherny 1999: 93–95); on ElseMOO these include the abbreviated “r” and
“u” variants of “are” and “you” that are common elsewhere in chat.
While MUDs and MOOs no longer command the popularity that they once did,
most online games (as well as virtual worlds for social and educational purposes)
now support some form of chat as a standard feature, and the nature of that chat, its
relationship to programming, and its various practices draw heavily from chat in
MUDs and MOOs. Use of chat in online games such as World of Warcraft, for in-
stance, is a core gaming competence and an integral part of the gaming experience.
Figure 1. A screen shot of the IRC client mIRC connecting to the #beginner channel on
the Italian server irc2.tin.it (Zelenkauskaite 2004)
participants. IRC networks thus function as enormous social markets. IRC also
permits direct client-to-client connections (DCC) where communications are
handled directly by participants’ computers, without involving the network. DCC
is used for private chat between two participants or for transfers of binary files such
as documents, images, and software. Other features, such as “actions”, resemble
the emote feature of MUDs and MOOs but are less sophisticated.
Unlike in MUDs and MOOs, user identities in IRC were not originally intended
to persist beyond a single user session. When a user connects to a server, s/he is
required to select a globally unique nickname (“nick”). However, nothing requires
a user to have the same nick in a subsequent session or prevents someone else
from using the same nick in a different episode. Users may also change their nicks
in mid-session, and users are even permitted to exchange nicks, sometimes leading
to playful exploitation of identity (Danet 1998). In general, however, users prefer
to maintain only a small number of nicks, in which they tend to become invested,
and impersonation often leads to the offender being banned.
While spatial metaphors and formal games are not a characteristic of IRC, pro-
grammed text and programming skills are common among users, especially since
clients are generally “scriptable” (i.e., programmed using a programming language
interpreter embedded in the client). Scripts can be unsophisticated, e.g., spitting
out random song lyrics or jokes from a file, or highly sophisticated, e.g., respond-
ing to specific events or typed text on a channel. The most sophisticated scripts are
known as “bots”, short for “robots”. Bots usually serve as a proxy for an operator
or a group of operators in their absence, permitting a clique of operators to assert
their presence on a channel round-the-clock (Paolillo 2001). This is often done on
large channels, where constantly-changing membership may lead to confusion
about who deserves to enforce speaking rights. Lack of a negotiated hierarchy of
operators can lead to “op wars”, where multiple contesting groups of operators dis-
rupt a channel’s participation and vie for its control. Hackers are frequent partici-
pants in op wars, sometimes instigating them as an anti-social form of amusement
(Paolillo 2001).
The social environment of IRC is thus hierarchically structured, but governed
by many layers of ad hoc technical and social arrangements. Channels have oper-
ators but exist on servers run by server owners, who have agreements with other
server owners, and who may be responsible to local computer administrators.
While hierarchical, these relations are informal and highly dynamic, sometimes
changing moment-to-moment according to the availability and/or status of partici-
pants. Power thus plays a major role in structuring chat interaction, although it may
arise in different ways. Among operators on IRC channels, mutual support cliques
exercise local power on a channel (Paolillo 2001). In educational settings, the in-
structor, who is institutionally empowered, functions also as moderator (Herring
and Nix 1997; Osman and Herring 2007). Network-wide, other administrators
using various forms of delegation may determine the roles that individual people
play in a chat environment and in what specific ways they are empowered. Some-
times these roles are negotiated within the chat context itself.
When private chat takes place on IRC, it occurs most commonly alongside pub-
lic chat. Each chat room has its own window, and chatters may participate in
multiple channels by switching windows. Consequently, IRC users often experi-
ence multiple levels of interaction while using chat, making “conversational multi-
tasking” a core pragmatic competence of chat (Dresner and Barak 2006). This fea-
ture is commonly carried over into other environments such as online games,
where distinct public, team, administrator and private chat channels are often pro-
vided (e.g., Herring, Kutz, Paolillo, and Zelenkauskaite 2009).
in any given instance of chat use what sort of interface experience a user might
have. The visual presentation of the chat, and what sorts of typographic enhance-
ment or multilingual script support are available, can vary from user to user in un-
known ways. This presents challenges to the analyst in understanding a user’s
chosen mode of expression on a given occasion.
Chat is subject to tremendous technical and social variation, and the pragmatics of
chat interaction vary accordingly. This leads to the question: What determines the
pragmatic features of any instance of chat? Are they based on face-to-face conver-
sation, as suggested by Werry (1996), Reid (1991), and Zhang (2005)? Are they
based primarily on written modes, as suggested by Zitzen and Stein (2004)? Or do
they arise from the social and technical characteristics of the chat mode in some
way (Herring 2007)? While pragmatic phenomena vary in chat, certain ones are
commonly encountered. In this section, we describe some of these phenomena and
attempt to indicate which sorts of technical and social characteristics of the mode
bear on them.
Markman (2005) also observed that disrupted adjacency (what she calls “false
adjacency pairs”, 118) is a common consequence of chat interaction, especially
when multiple participants are involved, and several disjoint conversations can be
interleaved, sometimes with overlapping participation from common participants.
Messages in chat are ordered sequentially, depending on who hit the “send” button
in what order; users do not necessarily intend to interrupt one another when mes-
sages appear out of sequence. Users also must employ context to determine if a
turn was actually completed.
Despite such complexities, it appears that users are not actually confused but
manage to be engaged in several interactions simultaneously. Herring (1999) ex-
plained this contradictory tendency in terms of adaptation to the technological con-
straints or engagement in playful language exchanges. Dresner and Barak (2006)
describe this same tendency as a pragmatic competence for “conversational multi-
3.3. Orthography
Non-standard orthography was one of the earlier features to be recognized as char-
acteristic of chat language; it has been noted in various language contexts: English
(Werry 1996), French (Anis 1999), Spanish (Llisterri 2002), Swedish (Hård af Se-
gerstad 2000), Russian (Lamprecht 2000), Finnish (Kotilainen 2002), and Lithua-
nian, Croatian, and Italian (Zelenkauskaite 2004; Zelenkauskaite and Herring
2006). Three major types of orthographic variation have been found in IRC chat:
deletions or reductions, insertions, and substitutions. Deletion or reduction of char-
acters is often related to economy. Insertions may have an expressive motivation
(such as emphasis), as may substitutions.
Typographic errors are common in chat. These are distinct from other forms of
orthographic modification, in that they are recognized by the users who produce
them as errors, not reflecting their communicative intent. Users sometimes self-
correct by posting a new message, as in the following Lithuanian example (4) from
Zelenkauskaite (2004), in which a metathesis is repaired.6
(4) #alus: 712. <Mefisto> o jau galvoju kad mano meile be tasako …
#alus: 713. <Mefisto> atsako
[#alus: 712. <Mefisto> that my love will be without the ersponse]
[#alus: 713. <Mefisto> response]
3.4. Gender
Gender differences in language are well established in speech (Coates 1993), but
because of the idea that CMC causes a leveling in context cues, potentially making
social distinctions like gender unimportant (e.g., Danet 1998), the case for gender
differences has had to be re-established for CMC (Herring 1992, 1993, 2003). In
MOO chat, Cherny (1994) uncovered a range of gender differences in the use of
emotes and specific actions. Notably, male participants tended to use fewer affec-
tionate actions than females and expressed them more frequently to inanimate and
abstract objects. Males also employed more physically violent imagery. Herring
(2003) observed similar gender differences in IRC, with males being more aggres-
sive and participating more actively, whereas women used more smilies and had
more instances of laughter. Despite some cases of gender-switching that have been
reported in chat environments (e.g., Danet 1998), the assumption in most gender
and CMC research is that online gender usually matches the writer’s offline gender,
on the grounds that it is difficult to fake subtle linguistic cues characteristic of fe-
male or male discourse for any sustained period of time (Herring 2003).
Gender differences in chat are not restricted to English-based chat. In a Thai
cultural setting, it was found that women participate more actively in public chat
and that women more often addressed turns to a specific next speaker and received
more responses by both females and males (Panyametheekul and Herring 2003). It
was also found that females would ask for and offer more email addresses to main-
tain further contacts, while men asked more direct questions and employed flirta-
tious communicative strategies more often.
In a study comparing Lithuanian and Croatian chat, gender differences in
orthography were found for specific diacritic symbols. There is a phonemic differ-
ence between s/z/c and š/ž/č in both Lithuanian and Croatian; however, the IRC
protocol does not provide for the relevant symbols, hence users are forced to adapt.
Zelenkauskaite and Herring (2006) found that users adopted a strategy of using
either digraphs [sh]/[zh]/[ch] or monographs [s]/[z]/[c]; when analyzed by gender,
males used more digraphs than females.
We have illustrated in the above sections that chat is a highly diverse phenomenon,
both in its technical and social contexts, which shape the linguistic manifestations
of its expression. We have illustrated some of the more important linguistic mani-
festations of pragmatic phenomena in chat. Both the contextual aspect of chat and
its linguistic manifestations are rich and substantial areas of research, and valid ap-
proaches can be formulated that focus primarily on one aspect or the other. At the
same time, understanding the pragmatics of chat must go beyond taxonomy. His-
torical, human-computer interface, and linguistic principles all bear on the nature
and use of chat, and the respective roles of these principles need to be explicated in
relation to other modes of discourse. A clear understanding of these principles and
their relation can also help indicate areas in need of development for future re-
search.
An important concept throughout this chapter has been that of chat mode: a
conventional clustering of technical and social characteristics, whose status is
comparable to that of a genre in other forms of discourse (Glitrow and Stein 2009;
Herring 2001, 2002). As chat is multifaceted, there are many such clusters, and dis-
tinct chat modes need to be recognized. At the same time, these modes do not vary
arbitrarily from one another, and the mode is a strong predictor of the kinds of
pragmatic phenomena one may find in any particular instance of chat, as well as
their likely linguistic manifestations.
Chat modes can be divided into two main types, largely on technical grounds:
those which are bounded by a single server, such as MUDs/MOOs, many online
games, and web chat, and those which reside on networks, such as IRC, commer-
cial services such as AOL chat, and the instant messaging family of modes. The so-
cial consequence of this choice concerns the number of potential participants: Net-
worked servers tend to have a much larger pool of participants than isolated
servers. Conversely, isolated servers tend to permit more focalized interaction and
the development or reinforcement of local communities and norms. Thus, among
Internet-based chat modes, we have an analog of the weak-tie/strong-tie distinction
widely recognized in social network analysis (Granovetter 1973), where contacts
on isolated servers resemble strong ties, and those on networks resemble weak ties.
Other technical variables of a mode might also correlate with this distinction. For
example, multimedia modes such as video chat tend to favor connection to spe-
cific individuals or small groups for specific intervals of time, rather than to large,
unbounded and ongoing interactions. The technical reason for this is the extreme
demands of sharing multimedia data in real time, which become increasingly se-
vere when participants are added. Text communication has the same complexity
problems, but because the amount of data associated with text is small, the system
can handle much larger numbers of people before problems of transmission vol-
ume are encountered. Hence, media type can also shape opportunities for social in-
teraction.
Linguistically, there are many consequences of this distinction. On one hand, a
larger pool of potential participants may draw from a broader range of national,
cultural, and linguistic backgrounds; hence the opportunities for linguistic and cul-
tural conflict, and the necessity to negotiate norms of interaction, are greater.
Weak-tie networks form effective mechanisms for the transmission of linguistic
and interactional patterns (Milroy 1992), and so observations of a phenomenon in
one part of the network are not likely to be independent of related events occurring
in other parts. Focalized communities, on the other hand, such as Cherny’s Else-
MOO (1999), are effective structures for the cultivation of distinctive identities, as
expressed through characteristic routines and/or modes of expression that do not
necessarily spread to other locations on the net. Hence, the situation of a chat mode
in the context of a broader network is the first dimension along which chat modes
appear to differ significantly.
Another distinction with similarly large pragmatic consequences is the degree
of choice users can exercise in customizing their chat environment. While custom-
izability in terms of computer programming is relevant and has a variety of effects,
the most important pragmatic aspect of this concerns the selection of potential par-
ticipants by users. Commercial services such as AOL offer the least user control
over participation by fixing in advance the number and type of available chat
rooms and placing power in the hands of corporately-appointed moderators to en-
force policies on appropriate speech. IRC goes largely in the other direction, by
allowing any user to create channels and vesting moderation power in the founding
user, to be delegated as s/he sees fit. At the other extreme is the IM cluster of
modes, where a user is required to select participants from a list of contacts. The
IM cluster acquires from this the characteristics of private interaction, which is not
invaded easily by arbitrary third parties, an opportunity that exists in the more pub-
lic IRC and AOL chat modes.
These two dimensions of variation in chat modes make predictions about what
might be observed in chat contexts that have not yet received extensive study. For
example, online games vary in the degree of user versus corporate control, and
hence, chat in those environments is predicted to be shaped accordingly. World of
Warcraft, a swords-and-sorcery game produced by Blizzard Entertainment, has a
game environment with strict corporate control; its game spaces are structured in
advance, and it employs its own moderators to enforce acceptable speech policies
in the same way as AOL. BZFlag, an open source tank-battle game supported by a
loose federation of independent developers and fans (Herring et al. 2009), is more
like IRC in its delegation of policies and enforcement to individual server owners
and administrators. A prediction from this is that the pragmatics of chat in these
two games, e.g., in the exercise of power in turn-taking and interaction manage-
ment, will differ in ways similar to that found between AOL chat and IRC. This is
one area in which further study of chat modes would be potentially instructive.
Another area for future investigation concerns the ongoing convergence across
media types, represented by the incorporation of chat into various technologies
such as websites, interactive television programming, online games, multimedia
platforms, and telephony applications, among others. Wherever media are chang-
ing, text chat, or something very much like it, seems to appear. To what extent are
the pragmatic characteristics of chat (as found in dedicated applications such as
IRC) carried over into these convergent media? When chat spreads to a convergent
media context, where does it get its pragmatic features from? The studies cited in
this chapter provide some indicators, along with the dimensions of variation
among chat modes described above, but additional research is still needed.
Further research remains to be done in other areas of the pragmatics of chat as
well. More yet needs to be said about medium effects with respect to chat lan-
guage. Existing work, both in design of systems and in cognitive psychological
studies of their effects, are informative, but additional research is needed into the
social effects of the technological medium features, especially beyond local issues
of turn-taking and interaction management. Analyses addressing the global organ-
ization and coherence of conversations in chat, such as those using Dynamic Topic
Analysis (Herring 1999; Herring and Nix 1997; Herring et al. 2009), could do
much to illuminate the ways in which chat functions for actual communication.
The international and intercultural dimensions of chat also deserve more atten-
tion. Today, with the widespread use of IRC and online games, chat is a vehicle of
continuous, large-scale intercultural and cross-linguistic contact, and issues of lan-
guage status, dominance, and conflict have already surfaced in many studies. The
understanding of language contact dynamics stands to gain much from close atten-
tion to chat and its role in the evolving online world. The nature of the social net-
works established in chat and their relationship to off-line and other social net-
works is a major area of potential study that is also related to the foregoing issues.
Its importance to linguistic pragmatics is that the social network offers a different
way of representing and investigating the social context from the approaches of the
past, and so more thorough incorporation of this into research can have an in-
fluence not only on understandings of the pragmatics of chat and CMC, but also on
understandings of social context in linguistic pragmatics more generally.
Notes
1. The game program Adventure (or ADVENT) written for PDP-10 by William Crowther at
BBN in 1975 is generally recognized as the first in the genre of interactive fiction. The
first MUD was written in 1978 by Roy Trubshaw at Essex University, UK.
2. The acronym MOO is recursive, like many other acronyms in the field of computing.
3. MUDs/MOOs are also used for education, and educational MUDs/MOOs are typically
recognized as a third type, distinct from social and gaming MUDs/MOOs.
4. Instant messaging and ICQ are both later developments that were not part of this offering.
5. The language regulation was dropped after one week.
6. In the examples that follow, the first line shows the text as it was originally typed; the
second line shows the same text with standard spelling; and the third line is a free English
translation. In each line, the expressions in focus are in boldface.
References
Androutsopoulos, Jannis K.
2003 Online-Gemeinschaften und Sprachvariation. Soziolinguistische Perspektiven
auf Sprache im Internet. [Online communities and language variation. Socio-
linguistic perspectives on language in the Internet.] Zeitschrift für germanisti-
sche Linguistik 31(2): 173–197.
Androutsopoulos, Jannis K.
2006 Multilingualism, diaspora, and the Internet: Codes and identities on German-
based diaspora websites. Journal of Sociolinguistics 10: 520–547.
Androutsopoulos, Jannis K. and Volker Hinnenkamp
2001 Code-Switching in der bilingualen Chat-Kommunikation: ein explorativer
Blick auf #hellas und #turks. In: Michael Beißwenger (ed.), Chat-Kommuni-
kation: Sprache, Interaktion, Sozialität & Identität in synchroner computer-
vermittelter Kommunikation, 367–402. Stuttgart: Ibidem.
Androutsopoulos, Jannis K. and Evelyn Ziegler
2003 Sprachvariation und Internet: Regionalismen in einer Chat-Gemeinschaft. In:
Jannis Androutsopoulos and Evelyn Ziegler (eds.), Standardfragen: Soziolin-
guistische Perspektiven auf Sprachgeschichte, Sprachkontakt und Sprach-
variation, 251–279. Frankfurt am Main: Peter Lang.
Anis, Jacques
1999 Chats et usages graphiques. In: Jacques Anis (ed.), Internet, communication et
langue française, 74–90. Paris: Hermes.
Barab, Sasha, Michael Thomas, Tyler Dodge, Robert Carteaux, and Hakan Tuzun
2005 Making learning fun: Quest Atlantis, a game without guns. Educational Tech-
nology Research and Development 53(1): 86–107.
Beißwenger, Michael
2007 Sprachhandlungskoordination in der Chat-Kommunikation. Berlin: de Gruyter.
Beißwenger, Michael and Angelika Storrer
2008 Corpora of computer-mediated communication. In: Anke Lüdeling and Merja
Kytö (eds.), Corpus Linguistics: An International Handbook, 292–309. Berlin:
Mouton de Gruyter.
1. Introduction
3. Overview of IM research
Many studies have focused on IM as a medium for social interaction. In this sec-
tion, research is reviewed involving use of IM between peers (especially young
people) and as a tool in the workplace. I also consider the social implications of
using IM while multitasking. Discussion of how IM away messages and conver-
sations are constructed appears in section 4.
As IM tools (and alternative communication media) evolve, it can be antici-
pated that future studies may generate different empirical results. However, vari-
ables identified in existing research (e.g., age, gender, linguistic creativity, speech
acts performed, strategies for controlling conversational flow) are likely to remain
important.
Computer-Based Activities
Web-based activities: 70 %
Computer-based media player: 48 %
Word processing: 39 %
Offline Activities
Face-to-face conversation: 41 %
Eating or drinking: 37 %
Watching television: 29 %
Talking on the telephone: 22 %
4.2. IM conversations
We turn now from IM away messages to IM conversations. Popular discussion of
IM conversations has typically focused on lexical issues (IM “lingo”) such as ab-
breviations, acronyms, emoticons,12 and odd spellings (see Thurlow 2006; Bies-
wanger, this volume). As the ultimate building blocks of conversation, lexical is-
sues are important to study, especially through quantitative corpus analysis.
However, equally relevant are the conversations themselves, including such prag-
matics issues as conversational repair and conversational closings.
Jacobs (2009) reports on a two-year ethnographic study she conducted of IM
usage by an adolescent girl. By videotaping both on-screen IM behavior and other
activity in the physical environment, Jacobs illustrated how young people incor-
porate IM conversations within their larger social worlds (see also Hu, Wood,
Smith, and Westbrook 2004). Focusing on lexical issues, Tagliamonte and Denis
(2008), building on protocols I had earlier developed (see below), collected over a
million words of IM conversations from Canadian teenagers. Comparing the IM
data with speech from some of the same subjects, the authors found that while IM
manifests some distinctive traits, it shows considerable stylistic variation, and it re-
mains rooted in the model of offline language.
In spring 2003, my students and I undertook a small-scale study of the struc-
tural properties of American college student IM conversations.13 Student experi-
menters initiated IM conversations with peers on their AIM buddy lists. After an-
onymizing user screen names, the student experimenters forwarded the IM
conversation files to a project web site. We collected 23 IM conversations, contain-
ing 2,185 distinct transmissions composed of 11,718 words. Nine were conver-
sations between females, 10 between males, and 5 involved male-female pairs.
Our IM analysis explored conversational composition, lexical issues, utterance
breaks, and gender issues.
Figure 2 summarizes our coding system. Names here and in subsequent
examples from the corpus have been anonymized.
Spelling was surprisingly good. Only 171 words were misspelled or lacked
necessary punctuation – averaging one error every 12.8 transmissions. More than
one-third of the errors involved omitting an apostrophe in a contraction (such as
thats) or a possessive (Sams).16 Another third were common spelling mistakes –
adding or omitting letters (e.g., assue for assume), or using the wrong letter (e.g.,
coliege for college). Some of these mistakes probably reflect typing errors, as do
the 21 % of misspellings involving metathesis (e.g., somethign for something). In
9 % of the cases, writers noticed the problem and fixed it (or tried to) in the next
transmission. That is, these writers repaired conversational infelicities, perhaps out
of habit, as a face-saving move or to increase intelligibility.
Conjunctions were the primary device for chunking utterances into multiple
transmissions. Out of 189 break pairs, 112 began the second transmission with a
conjunction. Of these, 89 used conjunctions to introduce sentences (e.g., coordi-
nating conjunction: and she never talks about him; subordinating conjunction: if I
paid my own airfare). The remaining 23 were conjunctions introducing a noun
phrase (or circleville) or verb phrase (and had to pay back the bank). More than
use many one-word utterances; and both can have protracted conversational clos-
ings. As for lexicon, while informal speech uses more contractions than IM, there
are more contractions in IM than in formal writing (where, traditionally, there are
none). Similarly, while not numerous, IM emoticons essentially substitute for
(spoken) prosody. A comparison of the utterance break data with Chafe’s descrip-
tion of utterance breaks in speech (1980) reveals that the two are largely parallel.19
However, when gender considerations are introduced, the profile changes.
Granted, several aspects of female IM usage are congruent with a “spoken” char-
acter, including more protracted conversational closings and more use of emoti-
cons than males. Yet in other important dimensions of IM, females seem to be
adopting a more written model. Females used fewer contractions than males.
Males were more likely to chunk sentences into multiple IM transmissions and to
begin the second member of an utterance break pair with a conjunction, while fe-
males tended to begin the second member with an independent clause. The con-
junction pattern is more common in speech, whereas the independent clause pat-
tern is more characteristic of writing (Chafe and Danielewicz 1987).
Overall, male IM conversations have much in common with face-to-face
speech, while female IMs more closely approximate conventional writing patterns.
Obviously, these generalizations are tempered by individual variation, along with
message function. Moreover, despite the more written quality of female IMs, the
messages remain largely informal.
Why does IM assume some dimensions of more formal, written language? One
explanation is what John Dewey (1922) called habit strength. By the time they
reach college, undergraduates have been typing on computers for years – often for
school work, where acronyms, abbreviations, and contractions are generally un-
welcome. These school-appropriate writing habits may be carrying over into IM,
which is produced on the same keyboard as school compositions – and sometimes
at the same time. According to “The Nation’s Report Card” (National Center for
Educational Statistics 2002), in the K-12 years, girls produce better writing than
boys (measured by assessment of narrative, information, and persuasive writing
tasks). It is therefore not surprising that female IM conversations include fewer
contractions, fewer sentences chunked into multiple transmissions, and fewer sen-
tence breaks involving a conjunction. These findings are consonant with socioling-
uistic research reporting that women’s language generally adheres more closely to
linguistic norms than does men’s (James 1996; Labov 1991).
Is IM speech or writing? It is some of both, but not as much speech as popularly
assumed. What is more, gender matters.
Gender also matters with respect to other IM issues, such as content, word
choice for initiating or closing IM conversations, and tone. Lee (2003) found that
male college students spoke more in their IMs about technology-related topics,
while female conversations involved more emotional subjects; that males tended
to avoid opening greetings or goodbyes, while females used both; and that males
addressed one another with derogatory names and used harsh teasing, while
females displayed a more sympathetic tone and used more emoticons. These find-
ings confirm gender distinctions observed in other forms of CMC (e.g., Herring
2003).
2004 and March 2006, Golder, Wilkinson, and Huberman (2006: 3) found that on
average, users sent less than one message per week. This was the Facebook en-
vironment in which my students and I explored how American university students
juggled IM versus Facebook usage.
By early 2010, the relationship between IM and Facebook had shifted. Anecdotal
evidence from informal observation suggests that at least on American campuses,
traditional IM clients such as AIM have been largely eclipsed by Facebook or the
chat function of Skype (see section 6).
However, not everyone in the study used IM or texting. Thirty percent of Italians
and 47 % of Japanese did no IM, and 12 % of Americans did no texting.
These data were largely consonant with results from another survey question:
Did people spend more time communicating through IM on a computer or through
texting on a mobile phone?
As Table 4 shows, Swedes spent slightly more time on IM, and Americans
spent, slightly more on texting. Since in both Italy and Japan, many subjects did
not use IM, it is not surprising to find the preponderance of both groups using more
texting. In the case of Koreans, although they had been using IM nearly as long as
Americans, their main form of one-to-one communication had become mobile
phones. The same survey revealed that Koreans used both voice functions and tex-
ting on their mobiles more than subjects from the other four countries (Baron
2009a).
Gender was a relevant variable in Sweden, the U.S., Italy, and Korea, where
more females favored texting over IM. One explanation is that females are often
more aware than males of how to manipulate communication media (Baron
2009b). Texting (unlike synchronous IM) enables users to select when to send and
receive messages. Another component of the cross-cultural study lends support to
this notion of control. When asked how important it was to send a text message to a
friend rather than make a voice call because “I want to make my message short, and
talking takes too long”, more females than males in each country responded that
this was a “very important” or “somewhat important” reason. As members of sub-
sequent focus groups explained, with texting, people can convey their own mess-
age without having to listen to other people’s problems.
6. IM in the future
Since its emergence two decades ago, IM has undergone continuing evolution. The
proliferation of personal networked computers in the 1990s made the medium
available to millions of users. New features emerged, allowing individuals to cus-
tomize their self-presentation as well as take advantage of affordances familiar
from email (e.g., messaging people not logged on to the system) and chat rooms
(e.g., conducting multi-person conversations).
The biggest challenges to IM came with the availability of multi-featured so-
cial networking sites (such as Facebook), the popularity of VoIP Systems (such as
Skype) that include typed messaging, and the profusion (most recently in the U.S.)
of inexpensive text messaging on mobile phones. From the perspective of young
users, price is often no longer a consideration: IM, Facebook, and Skype are “free”
(once one has an Internet connection), and texting plans increasingly offer un-
limited messaging, typically after paying a small monthly fee.
In the case of text messaging on mobile phones, portability is a clear advantage
over carrying a computer. IM, Facebook, and Skype can all be done on mobiles, al-
though input tends to be cumbersome. However, input is simplified somewhat by
T9 programs and by full QWERTY keypads, which are especially popular in the
U.S. and may be helping fuel the explosion in American texting.
Social networking and VoIP communication, at least in the U.S., are still pre-
dominantly done on computers. But such sites as Facebook have the advantage
(over traditional IM) of offering multiple additional functions (from photo albums
to event coordination). Applications such as Skype enable users to conduct an audio
(or video) session and simultaneously engage in written chat and/or transmit files.
Whatever name it travels under, IM as a typed synchronous communication
tool linking two (or more) interlocutors will almost certainly persist. Less clear is
whether the distinctive characteristics described in section 4 will survive as well.
As overall use of the Internet continues to grow, and as input devices become in-
creasingly simpler, Dewey’s (1992) notion of habit strength could lead users of
both genders to a single writing style for all compositions produced on computers.
Moreover, now that young people have enjoyed more than a decade of linguistic
creativity in their online (and mobile) communication, they might move on to
novelty in other pursuits, leaving the majority of IM “lingo” to fade.
With regard to pragmatics issues, there are larger questions: How will we man-
age personal interruptions (when an IM arrives), and how will we learn to coordi-
nate turn-taking in multiparty conversations (be they exchanges between friends
or classroom discussions in online education)? We will need to better understand
differences between online and face-to-face issues regarding politeness, deception,
conversational repairs, and conversational closings. We will also need to study
whether meanings are necessarily less clear in mediated communication than in
traditional face-to-face exchange or traditional writing. The latter concern moti-
vated the creation of emoticons in 1982, although it is unclear whether these sym-
bols are themselves semantically unambiguous – or necessary (Baron 2009c).
Since the mid 2000s, IM has increasingly evolved into a function incorporated
into – or replaced by – other mediated communication tools. However, the ques-
tions of language, context, and interpretation that IM initially generated are no less
relevant today to students of CMC.
Notes
1. See http://web.mit.edu/zephyr/doc/usenix/zslides.PS.
2. ICQ, like its predecessors, originally had a split-screen and keystroke-by-keystroke
transmission option, allowing what Herring (2007) calls two-way message transmission.
3. In much of Europe, “chat” refers to what Americans call “instant messaging”.
4. http://blog.facebook.com/blog.php?post=12811122130. See section 5.1 for discussion
of Facebook messaging versus AIM.
5. http://www.cellsigns.com/industry.shtml (accessed July 2, 2012).
6. http://www.multilingual-search.com/forrester-research-inc-data-on-mobile-instant-
messaging-im-in-europ e/25/01/2008 (accessed July 2, 2012).
7. Text messaging is also productively used by hearing-impaired populations (Bakken
2005).
8. See Baron 2008a (Chapter 3) for broader discussion of multitasking issues.
9. Tim Clem and Brian Rabinowitz assisted in this project.
10. For details on the away message study, see Baron, Squires, Tench, and Thompson
(2005).
11. There were also two messages conveying information intended for specific interlocu-
tors, although the messages were viewable by everyone on the sender’s buddy list (e.g.,
“Back in D.C. missing Houston very much … Kate, Erin, and Phil [names anonymized],
thank you for an amazing weekend”). At the time the data were collected, AIM users
could not send messages to people not logged on to the system. The only way to have a
message waiting for them upon their return was through a broadcast away message.
12. Emoticons typically function as commentaries on the preceding utterance. However,
since abbreviations, acronyms, and emoticons are commonly viewed as hallmarks of IM
communication, all are included in this chapter’s discussion of lexical issues.
13. See Baron 2004 for details of the study. While the corpus was small, results are conson-
ant with findings from Tagliamonte and Denis’s larger corpus.
14. This is an operational definition of the term, adopted for purposes of the present analysis.
15. This statistic is from a class project done in 2004 at my university analyzing traits of stu-
dents’ informal speech.
16. There is no evidence whether omission of apostrophes was an unconscious or conscious
act.
17. For details on the analysis, see Baron (2010).
18. For discussion of distinctions between spoken and written language, see Baron (2000),
Biber (1988), Chafe and Danielewicz (1987), Chafe and Tannen (1987), and Crystal
(2001).
19. See Baron (2010) for more detailed analysis.
20. For discussions of the development of Facebook, see Cassidy (2006) and Petersen
(2010). For an early analysis of the social impact of Facebook in college settings, see
Vanden Boogart (2006).
21. Statistics in this paragraph are from http://newsroom.fb.com/ (accessed July 2, 2012).
22. For more information on the Facebook study, see Baron 2008a (Chapter 5). Clare Park
assisted in gathering data.
23. Interviews were conducted by Tamara Brown, Dan Hart, Kathy Rizzo, and Kat Waller.
24. For example, in January 2005, Americans sent 7.25 billion text messages. People in the
UK sent approximately 2.7 billion texts in the same month. However, since the UK had
only one-fifth the population of the U.S. (roughly 60 million versus 296 million), the
UK sent nearly twice as many messages per capita. Three years later, the situation had
changed dramatically: In mid-2008, the UK sent roughly 6 billion texts per month,
while the U.S. sent 75 billion. Adjusting for population, by mid-2008 the U.S. was send-
ing more than twice as many texts as the UK (UK source: http://www.text.it/mediacentre;
U.S. source: http://www.cellsigns.com/industry.shtml (accessed July 2, 2012).
25. For discussion of the larger survey, see Baron (2008b, 2009a).
References
Leopoldina Fortunati, Tomaz Turk, and Eugene Loos (eds.), The Good, the
Bad, and the Challenging: The User and the Future of Information and Com-
munication Technologies, vol. 1, 273–282. COST Action 298 Participation in
the Broadband Society.
Baron, Naomi S.
2009b Control freaks: How online and mobile communication is reshaping social
contact. Keynote address, Language in New Media: Technologies and Ideo-
logies Conference, Seattle, WA, September 6.
Baron, Naomi S.
2009c The myth of impoverished signal: Dispelling the spoken language fallacy for
emoticons in online communication. In: Jane Vincent and Leopoldina Fortu-
nati (eds.), Electronic Emotion: The Mediation of Emotion via Information and
Communication Technologies, 107–135. Oxford, UK: Peter Lang.
Baron, Naomi S.
2010 Discourse structures in instant messaging: The case of utterance breaks. Lan-
guage@Internet 7, article 3.
http://www.languageatinternet.org/articles/2010/2651/index_html/
Baron, Naomi S., Lauren Squires, Sara Tench, and Marshall Thompson
2005 Tethered or mobile? Use of away messages in instant messaging by American
college students. In: Rich Ling and Per E. Pedersen (eds.), Mobile Communi-
cations: Re-Negotiation of the Social Sphere, 293–311. London: Springer.
Baym, Nancy, Yan Bing Zhang, and Mei-Chen Lin
2004 Social interactions across media: Interpersonal communication on the internet,
face-to-face, and the telephone. New Media & Society 6: 299–318.
Biber, Douglas
1988 Variation Across Speech and Writing. Cambridge, UK: Cambridge University
Press.
Bowe, Frank
2002 Deaf and hard of hearing Americans’ instant messaging and e-mail use: A
national survey. American Annals of the Deaf 147(4): 6–10.
Boneva, Bonka, Amy Quinn, Robert Kraut, Sara Kiesler, and Irina Shklovski
2006 Teenage communication in the instant messaging era. In: Robert Kraut, Mal-
colm Brynin, and Sara Kiesler (eds.), Computers, Phones, and the Internet,
201–218. Oxford, UK: Oxford University Press.
Bryant, J. Alison, Ashley Sanders-Jackson, and Amber Smallwood
2006 IMing, text messaging, and adolescent social networks. Journal of Computer-
Mediated Communication 11(2), article 10.
http://jcmc.indiana.edu/vol11/issue2/bryant.html
Cassidy, John
2006 Me media. New Yorker, May 15, 50–59.
Chafe, Wallace L.
1980 The deployment of consciousness in the production of a narrative. In: Wallace
L. Chafe (ed.), The Pear Stories: Cognitive, Cultural, and Linguistic Aspects
of Narrative Production, 9–50. Norwood, NJ: Ablex.
Chafe, Wallace L. and Jane Danielewicz
1987 Properties of spoken and written language. In: Rosalind Horowitz and S. Jay
Samuels (eds.), Comprehending Oral and Written Language, 83–113. San
Diego, CA: Academic Press.
Herring, Susan C.
2007 A faceted classification scheme for computer-mediated discourse. Lan-
guage@Internet 4, article 1.
http://www.languageatinternet.org/articles/2007/761
Hu, Yifeng, Jacqueline Fowler Wood, Vivian Smith, and Nalova Westbrook
2004 Friendship through IM: Examining the relationship between instant messaging
and intimacy. Journal of Computer-Mediated Communication 10(1).
http://jcmc.indiana.edu/vol10/issue1/hu.html
Isaacs, Ellen, Alan Walendowski, Steve Whittaker, Diane Schiano, and Candace Kamm
2002 The character, functions, and styles of instant messaging in the workplace. In:
Proceedings of the 2002 ACM Conference on Computer Supported Cooper-
ative Work, 11–20. New York: ACM Press.
Iqbal, Shamsi, and Eric Horvitz
2007 Disruption and recovery of computing tasks: Field study, analysis, and direc-
tions. In: Proceedings of the 2007 ACM SIGCHI Conference on Human Fac-
tors in Computing Systems, 677–686. New York: ACM Press.
Jacobs, Gloria
2009 Adolescents and Instant Messaging: Literacy, Language, and Identity Devel-
opment in the 21st Century. Saarbrücken: VDM Verlag.
James, Deborah
1996 Women, men, and prestige speech forms: A critical review. In: Victoria L.
Bergvall, Janet M. Bing, and Alice F. Freed (eds.), Rethinking Language and
Gender Research, 95–125. London: Longman.
Kayan, Shipra, Susan Fussell, and Leslie Setlock
2006 Cultural differences in the use of instant messaging in Asia and North
America. In: Proceedings of the 2006 ACM Conference on Computer Sup-
ported Cooperative Work, 525–528. New York: ACM Press.
Labov, William
1991 The intersection of sex and social class in the course of linguistic change. Lan-
guage Variation and Change 2: 205–254.
Lee, Christine
2003 How does instant messaging affect interaction between the genders? The Mer-
cury Project of Instant Messaging Studies, Stanford University.
http://www.stanford.edu/class/pwr3–25/group2/pdfs/IM_Genders.pdf
Lenhart, Amanda, Mary Madden, and Paul Hitlin
2005 Teens and technology: Youth are leading the transition to a fully wired and mo-
bile nation. Pew Internet & American Life Project, July 27.
Lenhart, Amanda, Lee Rainie, and Oliver Lewis
2001 Teenage life online: The rise of the instant-message generation and the inter-
net’s impact on friendships and family relationships. Pew Internet & American
Life Project, June 20.
Leung, Louis
2001 College student motives for chatting on ICQ. New Media & Society 3: 483–500.
Levinson, Stephen
1983 Pragmatics. Cambridge, UK: Cambridge University Press.
Ling, Rich and Naomi S. Baron
2007 Text messaging and IM: A linguistic comparison of American college data.
Journal of Language and Social Psychology 26(3): 291–298.
1. Introduction
The terms “text messaging” and “texting” refer to the brief typed messages sent
using the SMS (‘short message service’) of mobile/cell phones, PDAs (‘personal
digital assistants’), smartphones, or web browsers. Although messages nowadays
often include images, videos, and music (hence the newer term MMS ‘multimedia
message service’), the basic text-based messaging service continues to be enor-
mously popular. Texting was initially developed and released commercially in the
early to mid-1990s and has since seen a huge rise in popularity around the world,
following the rapid spread of mobile telephony in general. (In 2009, the United
Nations reported that more than 60 % of the world’s population – about 4.1 billion
people – had access to a mobile phone.1) Most often used for person-to-person
communication, text messages are also increasingly being used to interact with
automated systems (e.g., buying products, participating in television contests, re-
cruiting voters). One interesting “convergence” phenomenon is the use of short
messaging services with interactive television, which confuses the boundary be-
tween interpersonal and broadcast messaging (Zelenkauskaite and Herring 2008).
As is usually the case, the technology is being transformed continually.
In situating text messaging with reference to computer-mediated communi-
cation (CMC) more generally, it is important to recognize the interplay between
what a technology itself allows (or affords) and what the communicator herself/
himself brings to the technology. Most obviously, in the case of text messaging, the
equipment is small and, eponymously, mobile; it therefore affords most texters an
unobtrusive and relatively inexpensive means of communication. At the same
time, text messaging is also technically and practically restricted, allowing only a
certain number of characters per message. (Set by a worldwide industry standard,
the limit is almost always 160 characters per message.2) Moreover, like text-based
CMC, it is primarily QWERTY-driven – which is to say, reliant on the standard
typewriter keyboard (Anis 2007). Whether or not any mechanical feature of any
technology presents as a communicative constraint or opportunity, however, dep-
ends on the user and on the context of use.
For a technology that only really went “live” in the mid-1990s, it took scholars a
while to attend to texting. Since the early 2000s, however, research from a range of
disciplines and a number of countries has been growing. While much of this work
falls outside the immediate interests of language scholars, it does reveal the in-
creasing importance and application of texting in both scholarly and public con-
texts. This research also demonstrates how much scholarly writing focuses on the
transactional and often commercial uses of texting rather than the relational func-
tion which, as we will suggest, is at the heart of most everyday texting. Texting re-
search spans a wide range of disciplines and topics. From medicine, studies include
the use of texting for patient reminders (e.g., Downer, Meara, Da Costa, and Se-
thuraman 2006; Leong et al. 2006) and for aftercare treatment (e.g., Robinson et al.
2006; Weitzel et al. 2007). In academia, studies include texting as library support
(Herman 2007; Hill, Hill, and Sherman 2007), as a research methodology (Bosnjak
et al. 2008; Cheung 2008; Steeh, Buskirk, and Callegaro 2007), as a pedagogical
tool (Dürscheid 2002a; Naismith 2007), as a recruitment strategy (Maher 2007),
and as a means for reducing school truancy (Allison 2004). Research in environ-
mental development has examined how texting assists Bangladeshi villagers to lo-
cate clean water sources (Opar 2006). Texting research extends to business and
commercial uses (e.g., Bamba and Barnes 2007; Hsu, Wang, and Wen 2006; Ma-
hatanankoon 2007), political campaigning (Prete 2007), and media broadcasting
(Enli 2007). Closer to human communication research, psychologists have looked
at compulsive texting (Rutland, Sheets, and Young 2007) and so-called cyberbul-
lying (e.g., Raskauskas and Stoltz 2007; Smith et al. 2008). What is apparent from
this research is how often the purely informational uses of texting are privileged.
Much other research does address the role of texting as a social-communicative
resource in people’s daily lives. Consider these findings, for example: Thirty-two
percent of adult texters in Malaysia cannot use their mobile phones without texting
(Tanakinjal, Amin, Lajuni, and Bolongkikit 2007); texting is a status symbol with
Hong Kong college students, with texters being predominantly male and having a
high household income (Leung 2007); young adults with lower social skills in
Hong Kong (Leung 2007) and Japan (Ishii 2006) prefer texting to voice communi-
cation; Filipino mothers in the U.S. with children overseas use texting to maintain
real-time relationships with their children (Uy-Tioco 2007); and subtle gender re-
lations are negotiated via texting in Taiwan (Lin and Tong 2007). Lists like this il-
lustrate nicely the ways in which texting is typically embedded in people’s daily
lives.
In terms of language and communication in particular, scholarly interest has
been somewhat slower still to establish itself, and texting continues to be a
relatively under-examined area of research (compared, for example, with other
modes of CMC). This too has been changing, however, and a growing body of
properly sociolinguistic and discourse analytic research attends to texting in Eng-
lish and other national languages. Our brief overview of the literature here, for
example, covers work done in Finland, Sweden, Norway, Denmark, France, Ger-
many, Greece, Italy, South Africa, Nigeria, New Zealand, Kuwait, Malaysia,
Japan, Korea, China, Taiwan, and Hong Kong, as well as the UK and the U.S. Prag-
matically-oriented studies meanwhile have begun to address, among other things,
turn-taking, code-switching, openings and closings, and general communicative
intent. They have also considered, explicitly or not, the pragmatic implications of
message length, textual complexity, grammar and punctuation, orthography, and
the use of emoticons. In every case, studies typically situate pragmalinguistic phe-
nomena with a view to broad cultural and interactional variations, which has im-
portant implications for any gross generalizations about the uniform nature of tex-
ting – a point we return to below.
Korea (Kim et al. 2007), and young people in Japan (Ishii 2006) communicate pre-
dominantly with family or those in their innermost social circles. A study of older
Japanese texters, meanwhile, found texting used more with those in extended social
and even professional circles, in order to preserve respect for the receiver by not
risking interrupting their affairs (Rivière and Licoppe 2005). In a study of Kuwaiti
texters, Haggan (2007) notes the transcription of Arabic texts into English and a ten-
dency towards formality and eloquence, which, she argues, may arise from the value
of the texts more generally. Spagnolli and Gamberini (2007) meanwhile comment
on the way that some Italians send lengthy, elaborate refusals to invitations, which,
the authors argue, reflects particular local norms. Of course, in all these cases, it is
not clear how generalizable the findings are to the rest of the country. Related to this
point, we were unable to find anyone pulling together a large multinational com-
parative study that might offer a more systematic perspective on these types of lin-
guacultural differences, although some steps have been taken in this direction (e.g.,
Dürscheid and Stark 2011; Nickerson, Isaac, and Mak 2008; also see below: Bies-
wanger 2007; Plester, Litteton, et al. 2009; Spilioti 2009, 2011).
ations, writing in either all caps or all lower case, and exchanging longer words for
shorter ones (Hård af Segerstad 2002). In Norway, meanwhile, teenagers do not
use as many spelling alterations as do Swedish texters, with only 6 % of partici-
pants in one study using abbreviations, acronyms or emoticons, and girls largely
responsible for abbreviations and innovative spellings (Ling 2005). German
texters commonly use reduction techniques (Androutsopoulos and Schmidt 2002),
while French texters use phonetic reductions, syllabograms, rebus writing (e.g., as
with the English b4 for ‘before’), logograms which are symbols, acronyms, and
unilateral abbreviations (Anis 2007), and represent spoken forms in writing (Ri-
vière and Licoppe 2005). In the U.S., unambiguous abbreviations (e.g., u for ‘you’;
r for ‘are’), vowel deletions, and lexical shortenings (e.g., Sun for ‘Sunday’) are
common (Ling and Baron 2007). Nigerian English texters employ spelling ma-
nipulations, abbreviations, and phonetic spellings (Chiluwa 2008), and British
texters are also linguistically creative (Tagg 2010). In one of only two African
studies on texting that we are aware of, South Africa provides an interesting case
where texters are found to use abbreviations, paralinguistic restitutions, and non-
standard spellings when texting in English, but not at all when texting in isiXhosa
(Deumert and Masinyana 2008). Capitalization, punctuation, and blank spaces are
often omitted in Swedish text messages (Hård af Segerstad 2002); apostrophes and
sentence-final punctuation are omitted about two-thirds of the time in the U.S.
(Ling and Baron 2007). Emoticons, e.g., smiley faces, are rare but are used in the
U.S. (Ling and Baron 2007) and in Sweden (Hård af Segerstad 2002). Once again,
what is striking about this international research is how much variation it shows.
In addition to examining the lexical and orthographic features of texting, re-
search has also attended to a range of syntactic and textual features (i.e., looking at
the composition, organization, and coherence of messages). In this regard, mess-
age length has received considerable attention; for example, Hård af Segerstad
(2002, 2005a, b) found that, at 14.77 words per message, Swedish text messages
are typically longer than German messages, at 13 words per message (Döring
2002a, b). It is possible that cross-language comparisons are complicated by vari-
able morphemic structures (see Plester et al. 2009). Ling and Baron (2007), mean-
while, found that text messages in the U.S. averaged only 7.7 words each, making
them closer in length to those in Norway, which average 6.95 words per message
for girls and even fewer for boys, 5.54 words per message (Ling 2005). In an ex-
tensive comparison of English and German syntax in texts, Bieswanger (2007)
found that English texts contain on average 91 characters per message, while Ger-
man texts contain 95. The pragmatic significance of message length will become
apparent when we turn to our own case study in a moment.
Research on syntactic features has also investigated message complexity, i.e.,
messages containing multiple clauses. According to this measure, Norwegian teen-
age girls’ messages contain far greater complexity (52 %) than their male counter-
parts (15 %) (Ling 2005). Similar results were found in Finland, where boys prefer
to send one-sentence text messages, while girls prefer longer and more complex
messages (Kasesniemi 2003). These findings are consistent with Ling and Baron’s
(2007) finding that 60 % of their female U.S. university students’ text messages
contained more than one sentence. Along these lines, the omission of auxiliary
verbs, personal pronouns, and function words is common in Germany (Dürscheid
2002a) and Sweden (Hård af Segerstad 2002), where omission of the subject pro-
noun is also the most common syntactic reduction (Hård af Segerstad 2002). In the
UK, analyses of article use and texting language usage more generally are foci of
Tagg’s (2007a, b) work. A common thread across these studies is the syntactic vari-
ation in both gendered and other cross-cultural contexts.
In more specifically pragmatic research, studies show that, for example, open-
ings and closings are frequently dropped by Italian (Spagnolli and Gamberini
2007), Japanese (Ito and Okabe 2005), and German (Dürscheid 2002a, b) text
messagers. Dürscheid views this as a function of the conversational frame of tex-
ting, as texting is often used as a conversation channel – and indeed adheres more
to conversational norms – rather than as a form of communication that adheres to
prescriptions for writing (see also Anis 2007, Ito and Okabe 2005). Kasesniemi
(2003) too has remarked on the increasingly dialogical nature of texting among
young Finns. Other research suggests that turn-taking conventions may be even
stricter in texting than in speech, although this varies by cultural context and topic
content (or communicative intent), and adheres to some highly standardized ex-
change patterns (Androutsopoulos and Schmidt 2002; Spagnolli and Gamberini
2007). Texters in Italy (Spagnolli and Gamberini 2007) and Japan expect reciproc-
ity, and Japanese texters are highly sensitive to the amount of time that passes be-
tween turns, sending a text message prompt for recipients who take too long to
reply (Ito and Okabe 2005). Working with Danish data, Laursen (2005) found ex-
pectations of reciprocity and immediacy, while Harper (2002) argued that in the
U.S., text messages did not necessitate a reply and certainly not immediately.
Some research on code switching has also been done in multilingual settings,
predominantly, although not exclusively, investigating the use of English in com-
bination with another national language. In Kuwait, Haggan (2007) found that
texters use a mixture of Arabic and English in their text messages, while Finnish
teenagers mix Finnish with a medley of foreign language words and expressions,
drawing suitable expressions from any language known by the writer (Kasesniemi
2003), and South African texters blend English with isiXhosa by writing English
nouns with isiXhosa prefixes (Deumert and Masinyana 2008). In contrast, Ni-
gerian texters completely avoid any “Nigerianness” in their succinct messages,
preferring standard (British) English and, even in their personal texts, avoiding Ni-
gerian English and other indigenous languages (Chiluwa 2008). In her study,
Spilioti (2009) provides an account of graphemic representations in Greek texters’
alphabet choices and code-switches. While these and other studies certainly attest
to some important cultural variability, it is telling that so much research still covers
English and other dominant/national European languages and, to some extent, lan-
guages like Japanese and Arabic, to the detriment of the rest of the languages of the
world. This imbalance mirrors patterns of global wealth distribution, as well as the
symbolic marketplace of academia.
One early example of empirical research published on texting was our own study
(Thurlow 2003).3 Based on a corpus of actual text messages, this study presented a
direct challenge to the kinds of largely unfounded claims being made in the media
at the time; it also offered a more properly discourse analytic perspective on tex-
ting (for an overview of discourse analysis and its relation to pragmatics, see Ja-
worski and N. Coupland 2006). A number of the studies reviewed above – es-
pecially those with a specifically (socio)linguistic focus – have used this original
study as a stimulus for extending scholarly understanding of the “language” of tex-
ting. We review the study here in order to demonstrate – and reiterate – the quin-
tessentially pragmatic nature of texting.
What was most noticeable about the non-standard items in our own study was
how so few of them were especially new or especially incomprehensible. (The
media and other commentators often like to play up the “hieroglyphic” unintelligi-
bility of young people’s texting – see Thurlow 2007). In practice, very few of the
text messages we looked at were semantically unrecoverable, even in isolation
from their original, discursive context and even to outsiders such as ourselves.
Much of what texters type in their messages would not be out of place on a
scribbled note left on the fridge door, the dining-room table, or next to the tele-
phone – sites where the same brevity-speed imperative would apply. In this sense,
both academic and lay claims for the impenetrability and exclusivity of texting lan-
guage appear greatly exaggerated and belie the discursive significance of situated
language use. Like the fridge-door note-maker, texters surely recognize the ob-
vious need also for intelligibility – in Gricean terms, for adherence to the maxims
of quantity and manner (Grice 1975; cf. also Lenhart et al. 2008; Plester et al. 2009;
Tagg 2007b). One of the best examples of this, in terms of abbreviation, is the use
of consonant clusters (e.g., THX), which rely on the premise (and metapragmatic
awareness) that consonants in English (as with many other languages) usually con-
vey more semantic detail/value than vowels. Many of the non-conventional spell-
ings found in texting are widespread and pre-date the mobile phone, in any case
(Crystal 2008; Shortis 2007). Examples include the use of z as in girlz and the k in
skool, as well as phonological approximations such as the Americanized forms
gonna, bin, and coz and g-clippings like jumpin, havin.
Our own corpus did require one important caveat regarding the issue of non-
standard orthography. The text messages we looked at revealed only about three
abbreviations per message, which meant that this supposedly defining feature of
texting style accounted for less than 20 % of the overall message content analyzed
(see also Bieswanger 2007). This initial finding certainly appears to run counter to
popular ideas about the unintelligible, highly abbreviated “code” of texting. In the
same vein, relatively few typographic (as opposed to alphabetic) symbols were
found throughout the entire corpus, almost all of which were simply kisses or ex-
clamation marks, usually in multiple sets (e.g., xxxxxx and !!!!!). Emoticons (e.g.,
J), too, were noticeable by their absence. Moreover, in spite of their notoriety in
media reports and despite the title of Crystal’s (2008) book about texting, relatively
few examples of letter-number homophones (e.g., Gr8) were found. Like many of
the paralinguistic and prosodic cues found in older CMC technologies such as IRC
(see Werry 1996), a more frequent type of language play was “accent stylizations”
and “phonological approximations”, such as the regiolectal spelling novern for
‘northern’ (for parallels in languages other than English, see Ling 2005; Hård af
Segerstad 2005b). In addition, we found a range of onomatopoeic, exclamatory
spellings (e.g., haha!, arrrgh!, WOOHOO!, rahh, ahhh) and a handful of other
typographic-cum-linguistic devices for adding prosodic impact (e.g., quick quick,
wakey wakey and yawn) and, therefore, communicative immediacy. Indeed, as re-
searchers like Shortis (e.g., 2007) have shown, the non-standard orthography of
texting almost always expresses the generally creative, playful, and friendly tone
intended by texters (see also Hård af Segerstad 2005a). Herein lies the crux of tex-
ting.
One technologically afforded pragmatic phenomenon in our 2003 data was the
openly sexual tone of many messages. Texting facilitates an interesting mix of inti-
macy and social distance, not unlike various other genres of CMC (e.g., instant
messaging and, to some extent, email); this also complicates traditional boundaries
between private and public (cf. H. Lee 2005; Zelenkauskaite and Herring 2008).
The technical rapidity and ephemerality (it is seldom stored or recorded) of texting
seem to lend them a relative anonymity, even though the sender and receiver are in-
variably known to each other and/or revealed to each other through caller/number
display. This kind of “recognized anonymity” might explain the relative licentious-
ness or “flame” tendencies of some texters (for more on anonymity, see Joinson
2001; Thurlow, Lengel, Tomic 2004).4 The face-saving capacity of this type of
anonymity likewise accounts for texters who send messages to say something they
would ordinarily avoid having to say face-to-face, such as breaking up with a ro-
mantic partner or, in the case of our own study, discussing an unexpected preg-
nancy.
Related to this sense of recognized anonymity, and as another example of
how text messagers capitalize on technological affordances, our corpus revealed
instances where the sender and receiver were apparently within viewing distance
of each other. Texting facilitates this kind of co-present exchange, allowing texters
to interact covertly in an immediate and potentially very intimate form of com-
munication – what some have referred to as the “culture of concealed use” (Ling
and Yttri 2002:164). Texting evidently enhances communication in ways that
allow for multiple or even parallel communicative exchanges (including face-to-
face interaction), offering an attractive combination of mobility, discretion, inti-
macy, and play. This combination of immediacy and intimacy drives the underly-
ing need for sociality and, for the most part, explains the linguistic form, or style,
of texting.
its exaggerated depiction in the media. (In a follow-up to our study, Ling and Baron
2007 found a similar result; see also Hård af Segerstad 2002.)
The notion of standardness in written language is itself a sociocultural conven-
tion and always an abstraction from spoken language (see Cameron 1995; and on
texting in particular, Shortis 2007). In this sense, like the fridge-door note and the
phonetic transcriptions of expert linguists, many of the typographic practices of
texting offer more “correct”, more “authentic” representations of speech to begin
with. As Jaffe (2000: 498) puts it:
The use of non-standard orthography is a powerful expressive resource. … [which] can
graphically capture some of the immediacy, the ‘authenticity’ and ‘flavor’ of the spoken
word in all its diversity. … [and] has the potential to challenge linguistic hierarchies.
In their messages, texters “write it as if saying it” to establish a more informal reg-
ister, which in turn helps to do the kind of small talk and solidary bonding they de-
sire for maximizing sociality. The language texters use is, therefore, invariably ap-
propriate to the context of interaction (see Lenhart et al. 2008; Plester et al. 2009;
Tagg 2007a). Sometimes there is also evidence of texters’ reflexive (often playful)
use of language and their inherent metapragmatic awareness (Verschueren 2004),
as in the following example with its stylized performance of drunkenness and its
tongue-in-cheek misanthropy:
hey babe.T.Drunk.Hate all luv.Have all men.Fuck
them.how r u?We’re ou utery drunk.im changing.
Now.Ruth.xxx. Hate every1
This same metapragmatic awareness may also account for texters’ (admittedly
variable) use of such apparently clichéd forms as letter-number homophones and
emoticons, which can be used with ironic effect and/or self-consciously to enact or
playfully perform “text messaging”. In other words, in a Hallidayan sense (Halli-
day 1969/1997), texting always fulfills both an interpersonal and textual function,
as people send messages not only for relational bonding and social coordination,
but also to be seen to be texting. Texting (and mobile phones) carries cultural capi-
tal in and of itself – as a lifestyle accessory and a ludic resource. Irrespective of
message content, the act of texting has cachet (to follow the cliché, the medium is
also a message) and necessarily communicates something about the sender/user.
And part of buying into the cachet of texting means drawing on – or rejecting al-
together – symbolic resources such as ringtones, keypad covers, and popularized
linguistic markers like initialisms, clippings, and letter-number homophones. It is
precisely this kind of metapragmatic awareness and (life-)stylistic variation that is
typically overlooked in the print media’s own metapragmatic commentary – or
rather complaint – about texting (see Shortis 2007; Thurlow 2006, 2007).
Umberto Eco (2002) notes that we are living in an age where the diminutive,
the brief, and the simple are highly prized in communication. Clearly, texting (and
tweeting) embodies this zeitgeist. And like many earlier communication technol-
ogies, it evokes and/or materializes a range of projected fears and hopes. Indeed,
the history of the development of so-called new communication technologies has
been marked by periods of excessive hype and hysteria about the kinds of cultural,
social, and psychological impacts each new technology is likely to have. This is
not to deny that few people – professional, academic, or lay – could have predicted
the extraordinary rise in popularity of the cell phone and its sister technology tex-
ting. Not surprisingly, public discourse about texting (e.g., in the media) encodes
any number of metapragmatic comments about the nature of both texting and lan-
guage, which are interesting in themselves. If, as Mey (2001: 5) suggests, the field
of pragmatics “is interested in the process of using language and its producers, not
just its end-product, language”, then the kind of everyday, metapragmatic com-
mentary about texting is decidedly a-pragmatic since, for the most part, it fixates
on the structures, forms, and grammars of language – or the perceived lack thereof.
For example, both media discourse and other popular commentary prioritize and
exaggerate the look of text messaging language – its supposedly distinctive lexical
and typographic style. Notwithstanding this, everyday talk about texting does offer
important insights into people’s beliefs (and concerns) about language (and tech-
nology) which, as Pennycook (2004) notes, perfomatively establishes the mean-
ings of language itself.
It is for this reason that the study of texting warrants continued research inter-
est, especially from discourse analysts and other language and communication
scholars. Briefly, this research would do well to focus on situated (or ethnographic)
analyses (i.e., the real, everyday contexts of texting; cf. Thurlow 2009; Thurlow
and Mroczek 2011): to address the use of texting across the lifespan; to pay even
more attention to non-European linguacultures; to explore different practices of
transcultural style- and code-switching;6 to link with other “short messaging” tech-
nologies such as microblogging (e.g., Twitter), which are sustained by text- and in-
stant-messaging as well as by emailing; and, in the same vein, to undertake prop-
erly multimodal discourse analysis (e.g., use of “pxting” in New Zealand, ring
tones, wallpaper). As the technologies of texting are constantly changing, so too
are the practices and meanings of texting; any research on texting needs to be con-
stantly updated. As we have argued before (Thurlow and Bell 2009; cf. also Buck-
ingham 2007: 3), it is also important that scholars lead the way in resisting a super-
ficial fascination with technology – and, in the case of texting, with fleeting
linguistic curiosities – in favour of a deeper engagement with the cultural contexts
and communicative practices that give both technology and language their real
meaning.
Acknowledgements
We are grateful to Susan Herring, Tuija Virtanen, and an anonymous reviewer for
their careful and constructive feedback on an earlier draft of this chapter. Any er-
rors or oversights are of our own doing.
Notes
References
Agar, Michael H.
1994 Language Shock: Understanding the Culture of Conversation. New York: Wil-
liam Morrow and Company.
Allison, Sheila
2004 Text messaging to parents cuts truancy. Youth Studies Australia 23(3): 7.
Androutsopoulous, Jannis K.
2000 Non-standard spellings in media texts: The case of German fanzines. Journal
of Sociolinguistics 4(4): 514–533.
Downer, Sean R., John G. Meara, Annette C. Da Costa, and Kannan Sethuraman
2006 SMS text messaging improves outpatient attendance. Australian Health Re-
view 30(3): 389–396.
Dürscheid, Christa
2002a E-Mail und SMS: ein Vergleich. In: Arne Ziegler and Christa Dürscheid (eds.),
Kommunikationsform E-Mail, 1–22. Tübingen: Stauffenburg.
Dürscheid, Christa
2002b SMS-Schreiben als Gegenstand der Sprachreflexion. Networx 28.
http://www.mediaculture-online.de/fileadmin/bibliothek/duerscheid_sms/
duerscheid_sms.html
Dürscheid, Christa and Elisabeth Stark
2011 sms4science: An international corpus-based texting project and the specific
challenges for multilingual Switzerland. In: Crispin Thurlow and Kristine
Mroczek (eds.), Digital Discourse: Language in the New Media, 299–320.
New York: Oxford University Press.
Eco, Umberto
2002 Diminutive, but perfectly formed. Guardian Newspaper, April 20.
Enli, Gunn Sara
2007 Gate-keeping in the new media age. Javnost-The Public 2: 47–61.
Fairclough, Norman
1989 Language and Power. London: Longman.
Fischer, Claude S.
1988 “Touch someone” – The telephone industry discovers sociability. Technology
and Culture 29: 32–61.
Grice, H. Paul
1975 Logic and conversation. In: Peter Cole and Jerry Morgan (eds.), Syntax and Se-
mantics 3: Speech Acts, 41–58. New York: Academic Press.
Haggan, Madeline
2007 Text messaging in Kuwait. Is the medium the message? Multilingua 26(4):
427–449.
Halliday, M. A. K.
1969/1997 Language in a social perspective. In: Nikolas Coupland and Adam Jaworski
(eds.), Sociolinguistics: A Reader and Coursebook, 31–38. Basingstoke, UK:
Macmillan.
Hård af Segerstad, Ylva
2002 Use and adaptation of written language to the conditions of computer-mediated
communication. Ph.D. dissertation, Department of Linguistics, Göteborg Uni-
versity, Sweden. http://www.ling.gu.se/%7eylvah/dokument/ylva_diss.pdf
Hård af Segerstad, Ylva
2005a Language in SMS – a socio-linguistic view. In: Richard Harper, Leysia Palen,
and Alex Taylor (eds.), The Inside Text: Social, Cultual and Design Perspec-
tives on SMS, 33–52. Dordrecht: Springer.
Hård af Segerstad, Ylva
2005b Language use in Swedish mobile text messaging. In: Rich Ling and Per E. Ped-
ersen (eds.), Mobile Communications: Re-negotiation of the Social Sphere,
313–334. London: Springer.
Harper, Richard
2002 The mobile interface: Old technologies and new arguments. In: Barry Brown,
Nicola Green, and Richard Harper (eds.), Wireless World: Social and Interat-
cional Aspects of the Mobile Age, 207–226. London: Springer.
Herman, Sonia
2007 SMS reference: Keeping up with your clients. The Electronic Library 25(4):
401–408.
Herring, Susan C.
2001 Computer-mediated discourse. In: Deborah Schiffrin, Deborah Tannen, and
Heidi E. Hamilton (eds.), The Handbook of Discourse Analysis, 612–634. Ox-
ford, UK: Blackwell Publishers.
Herring, Susan C. and Asta Zelenkauskaite
2009 Symbolic capital in a virtual heterosexual market: Abbreviation and insertion
in Italian iTV SMS. Written Communication 26(1): 5–31.
Hill, J. B., Cherie Madarash Hill, and Dayne Sherman
2007 Text messaging in an academic library: Integrating SMS into digital reference.
Reference Librarian 47(97): 17–29.
Höflich, Joachim R. and Julian Gebhardt
2005 Changing cultures of written communication: Letter – e-mail – SMS. In: Ri-
chard Harper, Leysia Palen, and Alex Taylor (eds.), The Inside Text: Social,
Cultural and Design Perspectives on SMS, 9–32. Dordrecht: Springer.
Hsu, Tsuen-ho, Yi-sheng Wang and Su-chan Wen
2006 Using the decomposed theory of planned behaviour to analyse consumer be-
havioural intention towards mobile text message coupons. Journal of Target-
ing, Measurement and Analysis for Marketing 14(4): 309–324.
Igarashi, Tasuku, Jiro Takai, and Toshikazu Yoshida
2005 Gender differences in social network development via mobile phone text mes-
sages: A longitudinal study. Journal of Social and Personal Relationships
22(5): 691–713.
Ishii, Kenichi
2006 Implications of mobility: The uses of personal communication media in every-
day life. Journal of Communication 56(2): 346–365.
Ito, Mizuko and Daisuke Okabe
2005 Technosocial situations: Emergent structuring of mobile e-mail use. In: Mi-
zuko Ito, Daisuke Okabe, and Misa Matsuda (eds.), Personal, Portable,
Pedestrian: Mobile Phones in Japanese Life, 257–276. Cambridge, MA:
MIT Press.
Jaffe, Alexandra
2000 Introduction: Non-standard orthography and non-standard speech. Journal of
Sociolinguistics 4(4): 497–513.
Jaworski, Adam
2000 Silence and small talk. In: Justine Coupland (ed.), Small Talk, 110–132. Lon-
don: Longman.
Jaworski, Adam and Nikolas Coupland
2006 Introduction: Perspectives on discourse analysis. In: The Discourse Reader,
2nd Ed., 1–37. London: Routledge.
Joinson, Adam
2001 Self-disclosure in computer-mediated communication: The role of self-aware-
ness and visual anonymity. European Journal of Social Psychology 31(2):
177–192.
Kasesniemi, Eija-Liisa
2003 Mobile Messages: Young People and a New Communication Culture. Tam-
pere: Tampere University Press.
Kim, Hyo, Gwang Jae Kim, Han Woo Park, and Ronald E. Rice
2007 Configurations of relationships in different media: FtF, email, Instant Mess-
enger, mobile phone, and SMS. Journal of Computer-Mediated Communi-
cation 12(4), article 3. http://jcmc.indiana.edu/vol12/issue4/kim.html
Latham, Kevin
2007 Sms, communication, and citizenship in China’s information society. Critical
Asian Studies 39(2): 295–314.
Laursen, Ditte
2005 Please reply! The replying norm in adolescent SMS communication. In: Ri-
chard Harper, Leysia Palen, and Alex Taylor (eds.), The Inside Text: Social,
Cultual and Design Perspectives on SMS, 53–74. Dordrecht: Springer.
Lee, Carmen
2011 Texts and practices of micro-blogging: Status updates on Facebook. In: Crispin
Thurlow and Kristine Mroczek (eds), Digital Discourse: Language in the New
Media, 110–128. New York: Oxford University Press.
Lee, Hangwoo
2005 Implosion, virtuality, and interaction in an internet discussion group. In-
formation, Communication and Society 81(1): 47–63.
Lenhart, Amanda, Sousan Arafeh, Aaron Smith, and Alexandra Macgill
2008 Writing, technology and teenagers. Pew Internet & American Life Project.
http://www.pewinternet.org/Reports/2008/Writing-Technology-and-Teens/
01-Summary-of-Findings.aspx
Leong, Kwok Chi, Wei Seng Chen, Kok Weng Leong, Ismail Mastura, Omar Mimi, Mohd
Amin Sheikh, Abu Hassan Zailinawati, Chirk Jenn Ng, Kai Lit Phua, and
Cheong Lieng Teng
2006 The use of text messaging to improve attendance in primary care: A random-
ized controlled trial. Family Practice 23(6): 699–705.
Leung, Louis
2007 Unwillingness-to-communicate and college students’ motives in SMS mobile
messaging. Telematics & Informatics 24(2): 115–129.
Lin, Angel M. Y. and Avin H. M. Tong
2007 Text-messaging cultures of college girls in Hong Kong: SMS as resources for
achieving intimacy and gift-exchange with multiple functions. Continuum:
Journal of Media & Cultural Studies 21(2): 303–315.
Ling, Richard
2005 The sociolinguistics of SMS: An analysis of SMS use by a random sample of
Norwegians. In: Richard S. Ling and Per E. Pedersen (eds.), Mobile Communi-
cations: Re-negotiation of the Social Sphere, 335–349. London: Springer.
Ling, Richard and Naomi S. Baron
2007 Text messaging and IM: Linguistic comparison of American college data.
Journal of Language and Social Psychology 26(3): 291–298.
1. Introduction
Mobile phones are the world’s most widespread electronic communication tech-
nology. However, unlike landlines, mobiles are multimodal, allowing users to talk,
engage in written communication, and access online information. In addition, the
mobile phone is often a multi-functional device. One can use it to listen to recorded
music, take and send photographs, and keep track of appointments and contact
numbers, along with employing an array of other features, including a radio,
games, a calculator, and sometimes a global positioning system (GPS). There are
phones that can tell people when their house alarm has gone off, remind Muslims
when to pray (Bell 2005), and enable passengers to pay bus fares.
These devices render people continually available to friends, family, and the
world at large (Ling 2008a; Ling and Stald 2010). They help us gather important
information or just chat, arrange a dinner menu with a spouse, or ask children when
they will be coming home. Mobile phones let us coordinate activities, provide a
safety link, connect people to their jobs, and even serve as high-tech fashion acces-
sories.
The technology adds a new set of pragmatic considerations to how people com-
municate. While the issues differ somewhat across cultures, users everywhere must:
– understand differences between the ways in which landlines and mobile phones
work (for instance, mobiles must be turned on to operate; after inputting a
phone number, the user must push “send” for the call to transmit; batteries on
mobile phones must be charged for phones to function)
– learn how to input and retrieve text messages (if they wish to use this modality)
– learn the myriad of additional functions available on many mobile phones (for
example, setting the alarm clock, creating an address book, taking and sending
photographs, downloading information from the Internet – again, if they wish
to use these tools)
– manage the costs (such as keeping track of charges for talking, texting, or using
the Internet; devising strategies to minimize charges)
– set personal usage style and usage boundaries (for instance, choosing to use
abbreviations in text messaging or to write words out in full; deciding if there
are conditions for declining to use the mobile phone – such as at a family dinner
or in a class lecture)
– negotiate new relationships within families and across generations (for example,
adolescents using mobile phones to manage direct parental control)
phone in the previous three months (Zainudeen et al. 2007). This transition has
brought about more access to information and services, greater ability to coordi-
nate everyday activities, and even increased prospects for participating in social
unrest (Paragas 2003; Rafael 2003).
the move more available for telephonic interaction. Because of these issues, mo-
bile phones are increasingly seen as personal devices. They are repositories of in-
dividual lists of friends – and their telephone numbers. Phones store endearing text
messages from significant others and provide a general communication log. To
look at the call register or at the “address book” on other people’s telephones is to
pry into their private lives.
Mobile communication devices affect people’s daily lives. This section explores
three everyday domains: micro-coordination of activities, security and safety, and
integration with the closest sphere of friends and family.
3.1. Micro-coordination
Mobiles are used to orchestrate interactions in nuanced ways. For example, people
use them to coordinate when and where to meet for a beer after work, to keep tabs
on the comings and going of their children, and to manage activities with col-
leagues working on common projects (Ling and Yttri 2002). Norwegian teenagers
in a 2003 focus group describe this process.2
Inger: If you have a mobile phone, you can change plans along the way.
You don’t need to agree on a meeting. You don’t have to agree to
meet either, you can just call whenever actually.
Interviewer: But how do you do agreements then?
Inger: I don’t know, you agree when and where you are going to meet
and if there is some change, then you can say that you will meet
another place, for example, if it is easier.
Arne: I usually make agreements by just calling, “What are you doing
tonight? Ok, then I will just call you.”
Interviewer: Is it like that that you call and ask if you will do something?
Arne: Yeah, like, for example, today when I am here, then I can just
make an agreement with my friends that I will call when I am
done. That is ok instead of planning what you are going to do.
Such micro-coordination is possible because with mobile phones, users are person-
ally addressable and can be reached even when there is no access to a computer or
landline phone. Coordination may take place iteratively, as individuals shift plans
regarding when and where to meet.3 When using landlines, interpersonal coordi-
nation is more rigid. The parties must agree to meet at a specific time and place.
Once they set out for the rendezvous, they cannot conveniently adjust the arrange-
ments. By contrast, mobile phones make this process more dynamic.
instrumental purpose (for instance, arranging a meeting date) but then meandered
off to talk about family, sports, vacation plans, or the like. With the coming of mo-
bile phones, people are keeping in touch through a series of shorter but more fre-
quent interactions. These short communiqués might be voice conversations or text
messages. Such communications might explore a theme of common interest, con-
stitute a sequence of micro-coordinations, or simply be phatic interactions. With
mobile phones, communication among individuals is less a specific event than an
ongoing process, interleaved with other daily activities. Much as one has a continu-
ing thread of interaction with office colleagues that is taken up, dropped, and re-
sumed in the course of the day, the mobile phone facilitates comparable interaction
with physically remote individuals.
In section 1, it was suggested that mobile phones add a new set of pragmatic con-
siderations in how people communicate. In this section, some of the factors at play
are illustrated by looking at cost issues, the emancipation of adolescents through
mobile phones, and the pragmatics of everyday use.
cost per text. Yet for all of A’s friends, receiving so many texts was suddenly an ex-
pensive problem. This situation motivated A’s friends to also get unlimited texting
subscriptions. In recent years, there has been a dramatic increase in texting among
U.S. teens – and in unlimited texting plans. Between 2006 and 2009, the percent of
teens who texted their friends on a daily basis went from 27 % to 54 % (Lenhart et
al. 2010). Hence, payment and subscription structures affect the flow and the form
of communication.
In the developing world, widespread use of pre-paid subscriptions is pushing
increased adoption of mobile phones (Kalba 2008). Employing the system of SIM
cards,6 users can shift cards from one phone to another. Alternatively, a single
individual might have several different subscriptions (i.e., SIM cards) and use only
a single handset. One subscription might have less expensive voice minutes and
another, less expensive texting. Alternatively, it might be cheapest to use one sub-
scription early in the morning, another during the day, and a third in the evening, as
the rates change. In countries such as Ghana, people might have one subscription
for use in larger cities and another that has better coverage when away from core
population areas (Chipchase 2008).
People in the developed world also contrive cost-saving strategies through use
of multiple SIM cards. In Italy, for example, it is common for university students to
have two handsets (each carrying a different SIM card) so they can take advantage
of differential pricing for voice calls and texting from alternative carriers.7 Both
nationally and globally, cost-savings strategies make it difficult to determine
the number of mobile phone users accurately, since single users do not necessarily
have a single subscription.8
Those in the developing world who are too poor to own a phone handset – but
own a SIM card – are exceptions to the model of individual reachability described
earlier as distinguishing mobile phones from landlines. There are, for examples,
places in southern Asia and sub-Saharan Africa where people who own only a SIM
card will borrow the handset of a neighbor and insert their own SIM card to place a
call (Ling and Donner 2009). While this system is somewhat awkward, poorer
populations now have a viable way to check on prices for their crops, arrange
money transfers from children or relatives living in the cities, or maintain social
contact with family members working far away from home (Law and Peng 2006;
Paragas 2008).
Individuals without their own handset or their own SIM card sometimes “rent”
mobile access. This communal function is illustrated by the telephone kiosks op-
erated in Bangladesh by Grameen “telephone ladies” (Singhal, Svenkerud, and
Flydal 2002). Grameen Telephone has lent money to women in rural villages to
purchase simple handsets and solar recharging systems, thereby providing tele-
phone service to local villagers.
Another cost-management strategy involves “beeping” (also known as “flash-
ing” or “missed calls”). The system takes advantage of the caller ID function on
own mobile phones, teenagers can avoid conducting personal social interactions
with friends in the quasi-public space surrounding the family landline phone.
Instead, they can determine when and where the calls (and text messages) are initi-
ated and received. Such personal freedom not only fosters communication with
individuals but also, for many teenagers, facilitates opportunities for strengthening
peer group relationships (Ling and Haddon 2008). At the same time, when adoles-
cents travel outside of the home (as when meeting friends in the evening), parents
gain some peace of mind, knowing that their progeny can call home to report their
whereabouts or in case of emergency.
Another area in which mobile phones allow young people some latitude con-
cerns linguistic usage conventions. In the 1990s, very soon after teenagers began
using text messaging on mobile phones, they started playing with lexicon, syntax,
and punctuation (Thurlow and Poff, this volume). Creation – and sharing – of mo-
bile-based lingo became a badge of group membership, much the way that verbal
slang, clothing, and hair style have traditionally served as markers of social close-
ness for this age cohort (Baron 2008; Ling 2008a). Moreover, such creative usage
signaled a level of security with their language: Adolescents knew the rules but
were able to bend them, while retaining the ability to communicate (Lenhart, Ara-
feh, Smith, and Macgill 2008).
Table 1. Students for whom using voice functions on mobile phones is “always” or
“usually” acceptable
and that they sometimes turn their phones off entirely when they do not want to be
disturbed (for instance, while studying or while watching television at home).10 Yet
despite individual and cultural differences, there is a growing expectation in many
countries that people be available to receive mobile phone calls and text messages.
In the words of James Katz, “You’re a problem for other people if you don’t have a
mobile phone” (Weiner 2007).
A third pragmatic dimension of everyday mobile phone use involves modality
of communication: Does the mobile phone user choose to make a voice call or send
a text message? Sometimes the basis for decision-making is cultural, and some-
times it is individual. In Japan, for example, it is considered rude to be disruptive
of public space. As a result, people tend to modulate their voices when in public –
and, at least until recently, to constrain public mobile phone conversations. By
comparison, in Sweden, while mobile phone users are also relatively quiet in pub-
lic space, they are less hesitant than the Japanese to speak on their phones in public.
Table 1 contrasts Swedish and Japanese findings from a survey conducted in
2007–2008 (Baron and Hård af Segerstad 2010) in which university students aged
18–24 were asked about public situations in which they felt it was “always” or
“usually” acceptable to talk on their mobile phone.
Data analyzed earlier by Misa Matsuda suggest how much Japanese conven-
tions have been liberalized in recent years.11 In her 2001 data, only 49 % of 187
subjects between the ages of 18 and 24 (although not necessarily university stu-
dents) approved of mobile phone voice communication while on the street, com-
pared with Baron and Hård af Segerstad’s recent finding of 74 %.12 The Japanese
sociologist Hidetoshi Kato (2009) suggests that contemporary Japan is experienc-
ing fundamental – and troubling – social change: “We seem to be moving toward
non-edited, non-censored, non-self-restricted society”.
Beyond cultural conditioning factors, users make individual decisions regard-
ing when to talk and when to text. Sometimes the choice is obvious: The person
you want to reach is in a meeting, so you text in order not to disturb him or her. Al-
ternatively, if one’s interlocutor does not use text messaging, the person initiating
the communication calls.
As part of her cross-cultural study of mobile phone use, Baron compared rea-
sons that Swedish, American, Italian, and Japanese university-aged students gave
for choosing to send a text message rather than to make a call.13 Among the vari-
ables explored were: convenience for the message-initiator, convenience for the re-
cipient, desire to keep the message short, and cost. In Sweden, the U.S., and Italy,
personal convenience was judged a “very important” factor in deciding to send a
text message rather than make a voice call. Convenience of the interlocutor was
also deemed “very important”, though slightly less so. In Japan, convenience – of
either the subject or the interlocutor – was seen as somewhat less important. An-
other element in this mix is the age of the person receiving the text. According to
research done by Ling, Norwegians are less willing to send texts to elderly people
than they are to teens. There is the sense that if you want to communicate with an
elderly person, voice is generally the best choice (Ling 2008b).
Keeping a message short was seen as a “very important” reason for Swedes,
Americans, and Italians to text rather than to have a phone conversation. (With a
phone call, one risks needing to spend time engaging in social pleasantries or hav-
ing the interlocutor launch into a new topic.) By contrast, Japanese were only one-
third as likely as their counterparts in other countries to choose texting over talking
to keep the message brief.
Cost considerations proved to be an important deciding factor in choice of mo-
dality. While there are some exceptions (depending upon the type of subscription
plan one has), texting is generally less expensive than talking in Sweden, Italy, and
Japan. In the U.S., the reverse has been true, since most American subscription
plans include a large number of voice minutes but charge additionally for text
messaging, although this picture is changing with the advent of unlimited texting
subscriptions. In the study, Italians were the most cost-conscious, followed by Jap-
anese and then Swedes. That is, Italians were most likely to judge cost to be a very
important reason to choose texting over talking.14
With the introduction of texting on mobile phones in the early 1990s, users ac-
quired a new modality for mobile discourse. While adults have adopted texting to
varying degrees, it is among teenagers and young adults that the distinctive culture
of text messaging lives (Goggin and Crawford 2011; Ling, Bertil, and Sundsøy
2010). To wit: In a 2003 study of teenage mobile communication in Norway,
18-year-old girls sent a mean of more than 20 text messages per day, while their
45-year-old mothers sent only 5.15 Texting has proven especially important to the
younger age demographic both because it is inexpensive and because it provides a
discreet platform for social interaction with peers (Castells et al. 2007; Haddon
2004; Ito and Okabe 2006). In this section a number of these issues are exam-
ined. In addition, the section summarizes findings from empirical studies of text
messaging in Norway and in the U.S.
(the most common) and predictive texting. The multi-tap system utilizes the 12
keys on the mobile phone keypad for inputting each alphanumeric character (or
punctuation mark) in the text. On many mobile phones there is also a predictive
texting function that automatically anticipates later letters in a word, based upon
the initial letters entered through the multi-tap system.
Mobile phone keypads – like the landline dial plates on whose design they were
based – display clusters of three or four letters of the alphabet next to (or below)
each of the numbers “2” through “9”. With the transition from manual to mech-
anized telephone exchanges, letters were added to the (landline) telephone dial
purely to create a mnemonic device to help users in remembering phone numbers
(Ling 2007). Customers were assigned telephone numbers such as “GR 4–2525”,
where the “GR” (for “Greenbelt”) was dialed using the numbers “4” and “7”. The
alphanumeric design made the transition to landline touchtone phones in the
1970s, despite the fact that telephone numbers (at least in the U.S.) had already
eliminated letters in favor of strictly digits. Only with the introduction of mobile
phones were the residual letters once again put into active service, this time for en-
tering text.
Use of the mobile phone keypad for text entry necessitates mastering a new –
and for some cumbersome – writing technique. These keypads are far less efficient
than full computer keyboards, where each letter appears on a separate key. For
example, to write “hello” with the multi-tap system, the mobile phone user presses
the “4” key twice to get an “h”, the “3” key twice to get an “e”, the “5” key three
times to get an “l”, repeats the last sequence to get the second “l”, and finally
presses the “6” key three times to get the “o”. This process often results in jettison-
ing some traditional formalities. In a 2005 analysis of text messages, only about
3.5 % of all the text messages had a salutation (Ling 2005).
According to Silfverberg, MacKenzie, and Korhonen (2000), a proficient and
trained texter using the multi-tap system can attain speeds of 25–27 words per min-
ute.20 By contrast, using predictive texting, proficient texters can attain 46 words
per minute.21 However, in studies of everyday users creating text messages using
both the multi-tap and the predictive texting systems, input speeds were between 5
and 10 words per minute (Curran, Woods, and Riordan 2006; Hård af Segerstad
2003). Since text messages tend to be rather short (usually fewer than 10 words),
the inefficiency of producing the text is somewhat mitigated (Ling 2005).
There is not one prototypical content domain, any more than there is one prototypi-
cal writing style. Nonetheless, to get a flavor of the variety of messages that users
compose, it is instructive to consider the contents of a representative body of text
messages.
Our example here comes from Norwegian data gathered in 2002 (Ling 2005).22
Table 2 summarizes the content distribution (and, in some instances, sentential
form) for all 882 messages produced in the study. One-third of all text messages
in the sample involved coordination. These data confirm the importance of the
coordination function of mobile phones discussed in section 3.1. The next most
frequent category, questions and answers (one-quarter of all messages), illustrates
the use of texting for short bits of information that can appropriately be exchanged
asynchronously. So-called “grooming” or phatic messages were the third most-
frequent content area (18.3 %). With these messages, the actual semantic content
may be less important than the social connection established between two people
by sending a text message.
While the total number of commands and requests (5.0 %) was not high,
requests to “Call me” constituted 1.5 % of all 882 messages. “Call me” requests
demonstrate how users exploit multiple functions on their mobile phones: texting
to get someone’s attention (without interrupting a person at an inopportune mo-
ment), but then having the interlocutor follow up with a voice call, which generally
permits freer give-and-take between interlocutors.
The Norwegian data were also analyzed for demographic distribution, message
length, and punctuation. Women generally wrote longer and often more complex
text messages than men. The mean length of messages generated by males was
5.5 words, compared with 7.0 for women. On the whole, females’ messages in-
volved a larger variety of themes than texts composed by males. Moreover, women
were more likely to use capitalization and punctuation correctly,24 and women used
the letter-like conventions of openings and closings more often than males. At the
same time, women were more likely to use abbreviations than were their male
counterparts.25 As a group, younger users (teenagers and young adults) were the
heaviest users of abbreviations, although the overall frequency of such shortened
forms was relatively low.
Length
mean words per message 7.7 words
mean characters per message 35.0 characters
one-word messages 3.7 % of messages
multi-sentence messages 60.0 % of messages
mean sentences per message 1.8 per message
Emoticons and lexical shortenings
contractions 84.7 % of potential contractions
apostrophes 31.9 % of actual contractions
abbreviations 3.2 % of words in corpus
acronyms <.1 % of words in corpus
emoticons <.1 % of words in corpus
The average text message in the American female college-student corpus was
7.7 words, which closely approximates the 7.0 word average of Norwegian fe-
males. Very few messages were composed of a single word – at least in part be-
cause every text message (at the time) had a price attached to it.27 Rather, 60 % of
all messages contained two or more sentences (although the sentences themselves
were generally brief).
Regarding lexical shortenings, there were a large number of contractions (such
as writing “can’t” instead of “cannot”). Most of the times that subjects could use a
contraction, they did. However, only one-third of the contractions contained apos-
trophes. (In our example, the contracted form was written “cant” more often than
not.) Omission of apostrophes is understandable, given the complexity of inputting
most punctuation marks through multi-taps on mobile phones.
Far less frequent were the kinds of lexical shortenings commonly associated
with text messaging in the popular press. Out of 1473 words, only 47 (that is,
3.2 %) were clear cases of abbreviations (for example, U for ‘you’; R for ‘are’).
Even fewer acronyms (for example, ttyl for ‘talk to you later’) were used – a total
of eight (less than 1 %). As for emoticons, only two were found – both smiley
faces – in the entire corpus.
The paucity of abbreviations, acronyms, and emoticons is consistent with
studies of texting by young people in Canada (Tagliamonte and Denis 2008) and
the UK (Thurlow and Brown 2003). While lexical shortenings appear to be more
common among younger teenagers (especially females), college students often re-
port that they have “outgrown” such contrivances, focusing more on the conversa-
tional substance of their messages than on using special texting language as a form
of social bonding.
It is important, however, not to over-generalize these results to all populations.
In the Norwegian corpus described in section 5.3, roughly 6 % of lexical items
involved abbreviations, acronyms, or emoticons. In the Philippines, where use of
texting has been more prolific than possibly any other place in the world, there is
more intense use of emoticons and abbreviations (Ellwood-Clayton 2003).28 How-
ever, just as it is not possible to generalize about the content of text messages, gen-
eralizations about levels of lexical shortenings cannot be made without specific
empirical data.
What is known is that there is considerable popular apprehension over the
potentially deleterious effects of “texting language” (meaning not just lexical
shortenings, but presumed disregard for grammar, spelling, and punctuation) on
traditional written language (Thurlow 2006). However, empirical research (for in-
stance, Hård af Segerstad and Hashemi 2006; Lenhart et al. 2008; Plester, Wood,
and Bell 2008) continues to indicate that school children are aware of differences
between language they might use in text messages with friends versus the kind of
language appropriate for classroom work.29
6. Concluding remarks
As of 2009, approximately half the people in the world had a mobile phone sub-
scription, rendering mobile telephony the world’s dominant form of telecom-
munication. In some instances, mobiles are augmenting – or replacing – land-
lines or use of the Internet on computers, while for many people (especially in the
developing world), mobiles provide their first access to distance communication.
Citizens of the developed and developing worlds alike encounter new pragmatic
opportunities and challenges with mobile phones. Users devise strategies to
both connect with and distance themselves from potential interlocutors, to mini-
mize cost, and to be creative with language. Mobile phones render us more
independent, while also raising expectations that others will always be able to
reach us.
Every technology evolves. It is difficult to predict the future of mobile phones –
either as technological devices or as linguistic tools for social interaction. Techno-
logically, current trends suggest that mobiles will be used to some degree for func-
tions presently associated with computers, television, or debit cards. Linguisti-
cally, it seems unlikely that texting on mobiles will seriously impact written
language. Much less predictable are the innovative strategies that users from dif-
ferent age groups and diverse cultures will concoct for making mobile phones fill
their communicative needs.
Notes
1. The International Telecommunications Union (ITU) estimated that in 2009 there were 4.7
billion mobile phone subscriptions, compared with 1.8 billion Internet users and 1.2 bil-
lion landlines (ITU 2009).
2. The data were collected in a series of focus groups held in Norway. The focus groups
were conducted in Norwegian, and the material reported here was translated from Nor-
wegian into English.
3. As the number of participants involved increases, so does the complexity of iterative
micro-coordination. Humphreys (2007), for example, discusses the possibilities for
location-based texting when using programs such as Dodgeball, and Shirkey (2008)
describes the quasi-broadcast functions of mobile communication associated with mobi-
lization of social groups in New York City.
4. See Baron (2009). Data are from a free-association task administered as part of a larger
online survey of 18–24 year old university students regarding their use of and attitudes
towards mobile phones. A convenience sample of 2001 subjects was drawn from
Sweden, the U.S., Italy, Japan, and Korea.
5. From interview data collected by Rich Ling.
6. The SIM card (Subscriber Identity Module) is a small chip that is the “subscription”. This
chip is inserted into a handset but is removable. By contrast, in the U.S., the handset and
the subscription are generally inseparable.
7. These reports are from one-hour focus groups that Naomi Baron conducted with univer-
sity students in Italy (March 2008; total participants: 29) as part of a larger cross-cul-
tural study of mobile phone use.
8. In addition, individuals may have multiple phones to use with different interlocutors
(for example, one for business, one for friends and family, and one just for a boyfriend
or girlfriend).
9. In mobile communication, there have been essentially three phases (generations) of de-
velopment. The first generation (1G) was small nationally-based systems that used
analog protocols. 2G refers to the second generation of mobile communication proto-
cols, most notably CDMA in the U.S. and GSM in large portions of the world. This gen-
eration saw the true popularization of mobile communication. 3G refers to third-gener-
ation protocols designed to support data transmission and services such as television.
10. These reports are from participants in one-hour focus groups (or interviews) that Naomi
Baron conducted with university students in Sweden (November – December 2007) and
in Japan (May 2008) as part of a larger cross-cultural study of mobile phone use. Total
participants in Sweden: 10; total participants in Japan: 13.
11. We are grateful to Misa Matsuda for providing these data (email communication to
Naomi Baron dated March 8, 2007), which derive from a larger study of Japanese mo-
bile phone use (Ito, Okabe, and Matsuda 2005).
12. Matsuda’s findings from 2001 regarding use of voice communication on a train or bus
were more similar to the present data – 5 % – due to the ubiquity of signage on Japanese
buses, subways, and commuter trains effectively banning use of voice functions on pub-
lic transport (Baron 2008: Chapter 7).
13. Some of the findings presented in this section are reported in Baron (2010), Baron and
Hård af Segerstad (2010), and Baron and Campbell (2012), while others are unpub-
lished.
14. An explanation offered by several Italian colleagues was that while Italian parents typi-
cally pay most living expenses for children at the university, it is common for parents to
insist that children pay their own mobile phone bill. Hence, their progeny may be es-
pecially keen to economize on phone costs.
15. Findings from data collected by Rich Ling. The text messages were gathered from a
random sample of 2003 Norwegians in May of 2002 by telephone. In addition to beha-
vioral and attitudinal questions associated with mobile and SMS use, respondents were
asked to read (and where necessary, to spell out) the content of the last three messages
they had sent. This procedure resulted in a body of 882 SMS messages. Several other
studies have shown gendered differences in telephone use (Fischer 1992; Moyal 1992;
Rakow1992). Regarding generational use, texting has been found to be a life phase
medium (Axelsson 2010; Helles 2009; Ling 2010).
16. Subjects were 1,000 randomly-selected Norwegians aged 13 and older. The survey was
carried out by Telenor Research and Development (currently unpublished).
17. Focus group data are from one-hour sessions with clusters of university students
(total American participants: 18; total Italian participants: 29) conducted in Spring
2008.
18. See Tseliga (2007) for an example of adaptive strategies for inputting Greek on a com-
puter keyboard built with the Roman alphabet.
19. Construction of text messages in character-based writing systems such as Japanese is
particularly cumbersome, requiring the user to shift back and forth among screens gen-
erating syllabic graphemes representing syllables (hiragana and katakana), the ideo-
graphic characters (kanji), and letters in the Roman alphabet (romaji).
20. Although the paper does not specify, the study appears to have tested English input.
21. By way of comparison, an advanced typist using a keyboard can attain up to 120 words
a minute, while average typists who are composing at the same time they are typing
produce about 19 words per minute (C.-M. Karat, Halverson, Horn, and J. Karat
1999).
22. See note 15 for details of the study design.
23. The Norwegian text messages have been translated into English. Capitalization and
punctuation reflect the Norwegian original.
24. This finding is consonant with other research on gender and non-standard language use
(Chambers 1992; Labov 1991).
25. Herring and Zelenkauskaite (2009) report a similar finding in their study of Italian iTV
SMS.
26. See Baron (2011) for a comparison of the linguistic character of texting versus instant
messaging by American college students. By way of comparison, see Anis (2007) for
linguistic adaptations in French texting.
27. Most subjects in the study had post-paid texting plans that covered an allocated number
of text messages per month at a set fee. However, exceeding the monthly allotment
resulted in a high additional per-message charge.
28. In their study of Italian iTV SMS, Herring and Zelenkauskaite (2009) found many ab-
breviations but very few emoticons.
29. For more extensive discussion of the impact – or lack thereof – of online and mobile
communication on language, see Baron (2011).
References
Anis, Jacques
2007 Neography: Unconventional spelling in French SMS text messages. In: Brenda
Danet and Susan C. Herring (eds.), The Multilingual Internet: Language,
Culture, and Communication Online, 87–116. New York: Oxford University
Press.
Axelsson, Ann-Sofie
2010 Perpetual and personal: Swedish young adults and their use of mobile phones.
New Media & Society 12: 35–54.
Baron, Naomi S.
2008 Always On: Language in an Online and Mobile World. New York: Oxford Uni-
versity Press.
Baron, Naomi S.
2009 Three words about mobile phones: Findings from Sweden, the US, Japan and
Korea. In: Bartolomeo Sapio, Leslie Haddon, Enid Mante-Meijer, Leopoldina
Fortunati, Tomaz Turk, and Eugene Loos (eds.), The Good, the Bad, and the
Challenging: The User and the Future of Information and Communication
Technologies, vol. 1, 273–282. COST Action 298 Participation in the Broad-
band Society. Copenhagen, Denmark.
Baron, Naomi S.
2010 Control freaks: How online and mobile communication is reshaping social
contact. Language at Work 7. http://www.languageatwork.eu/articles.php
Baron, Naomi S.
2011 Assessing the internet’s impact on language. In: Mia Consalvo and Charles
Ess (eds.), The Handbook of Internet Studies, 117–136. Oxford, UK: Wiley-
Blackwell.
Baron, Naomi S. and Elise M. Campbell
2012 Gender and mobile phones in cross-national context. Language Sciences
34(1): 13–27.
Baron, Naomi S. and Ylva Hård af Segerstad
2010 Cross-cultural patterns in mobile phone use: Public space and reachability in
Sweden, the USA and Japan. New Media & Society 12: 13–34.
Baron, Naomi S. and Rich Ling
2007 Emerging patterns of American mobile phone use: Electronically-mediated
communication in transition. In: Gerard Goggin and Larissa Hjorth (eds.),
Mobile Media 2007: Proceedings of an International Conference on Social
and Cultural Aspects of Mobile Phones, Media and Wireless Technologies,
218–230. Sydney, Australia: University of Sydney.
BBC
2006 Mobile misuse causes rescue rise. April 28. http://news.bbc.co.uk/1/hi/wales/
4948182.stm
Bell, Genevieve
2005 The age of the thumb: A cultural reading of mobile technologies from Asia. In:
Peter Glotz, Stefan Bertschi, and Chris Locke (eds.), Thumb Culture: The
Meaning of Mobile Phones for Society, 67–87. New Brunswick, NJ: Transac-
tion Publishers.
Castells, Manuel, Mireia Fernández-Ardèvol, Jack L. Qiu, and Araba Sey
2007 Mobile Communication and Society: A Global Perspective. Cambridge, MA:
MIT Press.
Chambers, J. K.
1992 Linguistic correlates of gender and sex. English World-Wide 13: 173–218.
Chipchase, Jan
2008 Power up: Street charging services, Kampala: How to stay charged without
access to mains power. http://www.janchipchase.org/fp/wp-content/uploads/
presentations/JanChipchase_StreetChargingServices_Kampala_vFinal.pdf
Curran, Kevin, Derek Woods, and Barry O. Riordan
2006 Investigating text input methods for mobile phones. Telematics and In-
formatics 23: 1–21.
Donner, Jonathan
2007 The rules of beeping: Exchanging messages via intentional “missed calls” on
mobile phones. Journal of Computer-Mediated Communication 13(1), article 1.
http://jcmc.indiana.edu/vol13/issue1/donner.html
Dutton, William
2003 The social dynamics of wireless on September 11: Reconfiguring access. In:
A. Michael Noll (ed.), Crisis Communication, 69–82. Lanham, MD: Rowan
and Littlefield.
Ellwood-Clayton, Bella
2003 Virtual strangers: Young love and texting in the Filipino archipelago of cyber-
space. In: Kristóf Nyíri (ed.), Mobile Democracy: Essays on Society, Self and
Politics, 35–45. Vienna: Passagen Verlag.
Fischer, Carl
1992 America Calling: A Social History of the Telephone to 1940. Berkeley, CA:
University of California.
Geirbo, Hanne Cecilia, Per Helmersen, and Kenth Engø-Monsen
2007 Missed call: Messaging for the masses. A study of missed call signaling behavior
in Dhaka. Internal Telenor R&I publication. Fornebu, Norway: Telenor R&I.
Goggin, Gerard and Kate Crawford
2011 Generation disconnections: Youth culture and mobile communication. In: Rich
Ling and Scott Campbell (eds.), Mobile Communication: Bringing Us To-
gether or Tearing Us Apart? The Mobile Communication Research Series,
Volume 2. New Brunswick, Canada: Transaction.
Haddon, Leslie
2004 Information and Communication Technologies in Everyday Life. Oxford, UK:
Berg.
Hård af Segerstad, Ylva
2003 Towards a better system for text prediction in mobile phones. In: John S. Pet-
tersson (ed.), HumanIT, Karlstad University Studies. Karlstad, Sweden: Karl-
stad University Centrum för HumanIT.
Hård af Segerstad, Ylva and Sylvana S. Hashemi
2006 Learning to write in the information age: A case study of schoolchildren’s writ-
ing in Sweden. In: Luuk van Waes, Mariëlle Leijten, and Christine M. Neu-
wirth (eds.), Writing and Digital Media, 49–64. Amsterdam: Elsevier.
Helles, Rasmus
2009 Personlige medier i hverdagslivet. København: Københavns Universitet.
Herring, Susan C. and Asta Zelenkauskaite
2009 Symbolic capital in a virtual heterosexual market: Abbreviation and insertion
in Italian iTV SMS. Written Communication 26: 5–31.
Hillebrand, Friedhelm, Finn Trosby, Kevin Holley, and Ian Harris
2010 Short Message Service: The Creation of Personal Global Text Messaging. Chi-
chester: Wiley.
Humphreys, Lee
2007 Mobile sociality and spatial practice: A qualitative field study of new social
networking technologies. Ph.D. dissertation, Annenberg School for Communi-
cation, University of Pennsylvania.
Ito, Mizuko and Daisuke Okabe
2006 Intimate connections: Contextualizing Japanese youth and mobile messaging
In: Richard Harper, Alex Taylor, and Leysia Palen (eds.), The Inside Text: So-
cial Perspectives on SMS in the Mobile Age, 127–145. London: Klewer.
Ito, Mizuko, Daisuke Okabe, and Misa Matsuda
2005 Personal, Portable, Pedestrian: Mobile Phones in Japanese Life. Cambridge,
MA: MIT Press.
ITU
2009 Market Information and Statistics. http://www.itu.int/ITUD/ict/statistics/
Kalba, Kas
2008 The adoption and diffusion of mobile phones: Nearing the halfway mark. In-
ternational Journal of Communication 2: 631–661.
Karat, Clare-Marie, Christine Halverson, Daniel Horn, and John Karat
1999 Patterns of entry and correction in large vocabulary continuous speech recog-
nition systems. In: Proceedings of the 1999 SIGCHI Conference on Human
Factors in Computing Systems, 568–575. New York: ACM Press.
Kato, Hidetoshi
2009 Personal communication (letter to Naomi S. Baron, August 30).
Katz, James E. and Ron E. Rice
2002 The telephone as an instrument of faith, hope, terror and redemption: America,
9–11. Promethius 20: 247–253.
Labov, William
1991 The intersection of sex and social class in the course of linguistic change. Lan-
guage Variation and Change 2: 205–254.
Lasén, Amparo
2011 “Mobiles are not that personal”: The unexpected consequences of the account-
ability, accessibility and transparency afforded by mobile telephony. In: Rich
Ling and Scott Campbell (eds.), Mobile Communication: Bringing Us Together
or Tearing Us Apart, 83–106. The Mobile Communication Research Series,
Volume 2. New Brunswick: Transaction.
Law, Pui-Lam and Yinni Peng
2006 The use of mobile phones among migrant workers in southern China. In: Pui-
Lam Law, Leopoldina Fortunati, and Shanhua Yang (eds.), New Technologies
in Global Societies, 221–244. London: World Scientific.
Lenhart, Amanda, Sousan Arafeh, Aaron Smith, and Alexandria R. Macgill
2008 Writing, technology and teens. Pew Internet & American Life Project, April 24.
http://www.pewinternet.org/Reports/2008/Writing-Technology-and-Teens.aspx
Lenhart, Amanda, Rich Ling, Scott Campbell, and Kristen Purcell
2010 Teens and mobile phones. Pew Internet & American Life Project, April 20.
http://www.pewinternet.org/~/media//Files/Reports/2010/PIP-Teens-and-
Mobile-2010.pdf
Licoppe, Christian
2004 Connected presence: The emergence of a new repertoire for managing social
relationships in a changing communications technoscape. Environment and
Planning: Society and Space 22: 135–156.
Ling, Rich
1998 “One can talk about common manners!”: The use of mobile telephones in in-
appropriate situations. In: Leslie Haddon (ed.), Communications on the Move:
The Experience of Mobile Telephony in the 1990s. (COST 248 Report). Farsta,
Sweden: Telia. http://richardling.com/papers/1997_One_can_talk_about_
common_manners.pdf
Ling, Rich
1999 “I am happiest by having the best”: The adoption and rejection of mobile tel-
ephony. Telenor R&D report, Kjeller, Norway.
Ling, Rich
2004 The Mobile Connection: The Cell Phone’s Impact on Society. San Francisco,
CA: Morgan Kaufmann.
Ling, Rich
2005 The socio-linguistics of SMS: An analysis of SMS use by a random sample of
Norwegians. In: Rich Ling and Per Pedersen (eds.), Mobile Communications:
Renegotiation of the Social Sphere, 335–349. London: Springer.
Ling, Rich
2007 The length of text messages and use of predictive texting: Who uses it and how
much do they have to say? In: Naomi S. Baron (ed.), American University
TESOL Working Papers Number 4. Washington, DC: American University.
http://www.american.edu/tesol/WorkingPapers04.html
Ling, Rich
2008a New Tech, New Ties: How Mobile Communication is Reshaping Social Cohe-
sion. Cambridge, MA: MIT Press.
Ling, Rich
2008b Exclusion or self-isolation?: Texting and the elderly users. The Information
Society 24: 334–341.
Ling, Rich
2010 Texting as a life phase medium. Journal of Computer-Mediated Communi-
cation 15(2): 277–292. http://www3.interscience.wiley.com/cgi-bin/fulltext/
123248040/PDFS TART
Ling, Rich and Naomi S. Baron
2007 The mechanics of text messaging and instant messaging among American col-
lege students. Journal of Language and Social Psychology 26: 291–298.
Ling, Rich and Jonathan Donner
2009 Mobile Communications. London: Polity Press.
Ling, Rich and Leslie Haddon
2008 Children, youth and the mobile telephone. In: Kristen Dortner and Sonia
Livingstone (eds.), International Handbook of Children, Media and Culture,
137–151. London: Sage.
Ling, Rich, Troels Bertil, and Pål Roe Sundsøy
2010 Texting among same-aged individuals: An analysis of traffic data. Paper pres-
ented at the International Communication Association Pre-Conference Work-
shop: “Mobile Communication”, Singapore, June 20.
Ling, Rich, Tom Julsrud, and Erling Krogh
1998 The goretex principle: The hytte and mobile telephones in Norway. In: Leslie
Haddon (ed.), Communications on the Move: The Experience of Mobile Tel-
ephony in the 1990s (COST 248 Report). Farsta, Sweden: Telia.
Ling, Rich and Gitte Stald
2010 Mobile communities: Are we talking about a village, a clan or a small group?
American Behavioral Scientist 53: 1133–1147.
Ling, Rich and Pål Roe Sundsøy
In press Mobile communication as a mundane medium. In: Kristóf Nyíri (ed.), Mobile
Communication Policy. Vienna: Passagen Verlag.
Ling, Rich and Birgitte Yttri
2002 Hyper-coordination via mobile phones in Norway. In: James E. Katz and
Mark Aakhus (eds.), Perpetual Contact: Mobile Communication, Private
Talk, Public Performance, 139–169. Cambridge, UK: Cambridge University
Press.
Maitland, Donald
1984 The missing link. Report of the Independent Commission for Worldwide Tele-
communications Development. http://www.itu.int/osg/spu/sfo/missinglink/
index.html
Marlatt, Michael
2010 Mobile, at a glance. http://www.ere.net/webinars/mobile-recruitment.asp
Monk, Andrew, Jenni Carroll, Sarah Parker, and Mark Blythe
2004 Why are mobile phones annoying? Behavior and Information Technology 23:
33–41.
Moyal, Ann
1992 The gendered use of the telephone: An Australian case study. Media Culture
and Society 14: 51–72.
Nielsenwire
2008 In US, SMS text messaging tops mobile phone calling. September 22.
http://blog.nielsen.com/nielsenwire/online_mobile/in-us-text-messaging-tops-
mobile-phone-calling/
Paragas, Fernando
2003 Dramatextism: Mobile telephony and people power in the Philippines. In:
Kristóf Nyíri (ed.), Essays on Society, Self and Politics, 259–284. Vienna: Pas-
sagen Verlag.
Paragas, Fernando
2008 Migrant workers and mobile phones: Technological, temporal and spatial sim-
ultaneity. In: Rich Ling and Scott W. Campbell (eds.), The Reconstruction
of Space and Time: Mobile Communication Practices, 39–66. New Brunswick,
NJ: Transaction Press.
Plester, Beverly, Clare Wood, and Victoria Bell
2008 Txt msg n school literacy: Does texting and knowledge of text abbreviations
adversely affect children’s literacy attainment? Literacy 42: 137–144.
Rafael, Vicente L.
2003 The cell phone and the crowd: Messianic politics in the contemporary Philip-
pines. Public Culture 15: 399–425.
Rakow, Lana F.
1992 Gender on the Line. Urbana, IL: University of Illinois.
Sey, Araba
2007 Mobile phones and the pursuit of sustainable livelihoods: Why connectivity
matters. Paper presented at the International Communication Association
Pre-Conference Workshop: “Mobile Communication: Bringing Us Together or
Tearing Us Apart?” San Francisco, CA, May 23–24.
Shirky, Clay
2008 Here Comes Everybody: The Power of Organizing without Organizations.
New York: Penguin Press.
Silfverberg, Miika, Scott MacKenzie, and Panu Korhonen
2000 Predicting text entry speed on mobile phones. In: Proceedings of the 2000
SIGCHI Conference on Human Factors in Computing Systems, 9–16. New
York: ACM Press.
Singhal, Arvind, Peer J. Svenkerud, and Einar Flydal
2002 Multiple bottom lines? Telenor’s mobile telephony operations in Bangladesh.
Telektronikk 2002: 153–160.
1. Introduction
(CA; see Schegloff 1988, 2007), although the studies that are reviewed below use
various theoretical frameworks and methodological approaches, including CA.
The common thread that binds our literature review and analysis of chat data is a
concern to uncover the interactional character of SvCMC. The interactional fea-
tures discussed below include identification-recognition strategies, turn-taking,
floor management, and repair. These features are subsumed under the larger disci-
pline of pragmatics (see Editors’ introduction, this volume).
2. Voice-enabled technologies
Numerous SvCMC technologies exist; some are used for one-to-one communi-
cation, while others accommodate multiple interlocutors. The literature and data
discussed below deal primarily with the latter, although when relevant and necess-
ary, one-to-one communication is discussed, examined, and compared with multi-
party CMC. The reason for doing so is two-fold. First, the interaction-oriented
literature on SvCMC is still in its infancy, and therefore previous observations
and findings that are relevant to this chapter are scarce. Second, the observa-
tions and findings for one-to-one communication, including those for telephones,
are important to a comprehensive understanding of the interactional character of
SvCMC. This section presents an overview of the technologies that are equipped
with voice capabilities that are particularly relevant to our review of the literature
and presentation of data.
Telephone conversations have provided the bulk of existing observations and
findings about voice-only communication. Although telephone technology sup-
ports three-way calling, the more common two-way medium has received the most
attention (see Hutchby 2001). Themes that have been investigated are wide rang-
ing, although most studies have examined issues related to the interactional organ-
ization of talk: sequence organization (e.g., for telephone openings; see Houtkoop-
Steenstra 1991; Schegloff 1972, 1986), turn-taking (e.g., for overlapping talk; see
Jefferson 1984; Schegloff 2000a), and repair (e.g., Schegloff 2000b; Wong 2000).
While some research has investigated telephone communication in the workplace
(e.g., Baker, Emmison, and Firth 2005), the preponderance of research has exam-
ined casual conversations between friends and family. Despite this extensive body
of literature, studies that examine telephone conversations do not typically discuss
the mediating factors involved in voice-only communication (see, however,
Hutchby and Barnett [2005] for a discussion of how telephones mediate conver-
sation openings). Rather, telephone conversations are examined in order to under-
stand dyadic communication in its own right (Schegloff 1979: 24). However, the
telephoning task of communicating in the absence of non-verbal cues is similar in
some ways to SvCMC. These similarities will be identified and discussed in our
presentation of the interactional character of SvCMC.
Mobile phones are similar in many ways to telephones. A main feature of both
devices is the ability to synchronously communicate verbally with people in geo-
graphically different locations. Both devices also support VoIP communication
and possess other communicative modalities (e.g., voice messaging and video
communication). Mobile phones play a central role in the way people communi-
cate in today’s technologically-driven world. This is especially true as more and
more people access online content with their mobile phones. However, very little
work has been done on the interactional character of mobile phone communi-
cation, with the exception of a few conversation analytic studies. Being a relatively
new technology in comparison to the telephone, the interaction-related literature
on mobile phone communication is still very much in its infancy. Existing research
has focussed on how interlocutors open their conversations (e.g., Arminen 2005;
Hutchby and Barnett 2005) and how opening sequences are influenced by caller
identification technology (see section 4). In contrast, extensive research has been
carried out on the interactional character of push-to-talk radios. Push-to-talk
radios are similar to mobile phones in that both devices transmit and receive talk
synchronously (i.e., both devices are two-way radios). However, push-to-talk
radios are different from mobile phones in that communication with the former de-
vice typically involves sending and receiving talk in single intervals (see Woodruff
and Aoki 2004 for an overview). Two-way radios are commonly used in the emerg-
ency and police services, as well as for communication between pilots and air traf-
fic controllers (see Neville 2004 for a book-length study of airline cockpit com-
munication).
Audioconferencing entails two or more interlocutors communicating with tele-
conferencing phones, traditional telephones, and/or computers with headsets. In
some situations, face-to-face communication occurs in conjunction with audiocon-
ferencing. For example, it is not uncommon to have workers in two geographically
displaced locations teleconferencing with each other, as well as communicating
with their colleagues in a physical space (e.g., Wainfan and Davis 2004). Most
studies that have investigated audioconferencing systems are concerned with ex-
ploring business-related topics and/or design issues. Example topics for the former
focus include group decision-making processes (e.g., Hiltz, Johnson, and Turoff
1986) and institutional status in the workplace (e.g., France, Anderson, and
Gardner 2001), while the latter focus is concerned with designing better technol-
ogies for SvCMC (e.g., Rodenstein and Donath 2000). Audioconferencing has also
been investigated somewhat extensively in distance language education settings
(see Lamy and Hampel 2007 for an overview of CMC and language learning).
Much of this research examines the pedagogical benefits of in-house audioconfer-
encing programs (e.g., Lyceum at Open University; see Hampel and Hauck 2004),
although few studies rely on detailed transcripts to investigate interactional issues
(cf. Negretti 1999). However, some interaction-related studies of audioconferenc-
ing exist. For example, Hindus, Ackerman, Mainwaring, and Starr (1996) exam-
cially available programs include, but are not limited to, PalTalk, iVisit, Adobe
Connect, Elluminate Live, and Skype. A common feature in most programs is the
ability to choose from several media (e.g., text, audio, and video). Programs vary
greatly in terms of subscription fees, available features, and intended audience (see
Cziko and Park 2003 for an overview). On the one hand, audio chat programs are
used for meetings at prearranged times and with participants with prior knowledge
of each other. For example, Nolan and Exner (2009) investigated the benefits of
using Adobe Connect for teacher-student meetings (see also Jepson 2005 for an in-
vestigation of student-student interaction). Audio chat rooms used for pre-planned
meetings with acquainted participants in institutional settings are sometimes re-
ferred to as audioconferencing systems (e.g., Thunderwire; see Ackerman, Hindus,
Mainwaring, and Starr 1997). On the other hand, audio chat programs are used for
impromptu conversations with unacquainted interlocutors (e.g., Skypecasts; see
section 3). Interlocutors freely enter and leave these latter types of audio chat
“rooms” in the same way as in text-based chat rooms. Few studies of unacquainted
interlocutors participating in audio chat rooms have been conducted (see, however,
Jenks 2009); this chapter reviews existing observations and findings. Yet another
type of audio chat room investigated in the literature has been used for system de-
sign purposes, where interlocutors communicate in experimental/laboratory en-
vironments. For instance, Aoki et al. (2003) examined an experimental audio chat
room program that allows for multiple, simultaneous conversational floors. One of
the empirical objectives in Aoki et al.’s (2003) study was to design a program that
replicates the conversational schisms that occur in multiparty face-to-face interac-
tion (cf. Egbert 1997). The various audio chat rooms investigated in the literature
are discussed in greater detail when the interactional features of SvCMC are pres-
ented below.
3. Skypecasts
The data that we present to supplement extant observations and findings come
from our recordings of Skypecasts, an audio chat room feature of Skype. Skypecasts
are conducted primarily in the voice medium, although interlocutors can send each
other private text messages. While Skype users can participate in video interac-
tions, Skypecasts do not support videoconferencing. Furthermore, Skypecasts are
used primarily for chatting among unacquainted interlocutors, whereas pre-
planned meetings for institutional purposes are often conducted in Skype. Anyone
with an Internet connection, headset or USB telephone, and a Skype account can
create and host a Skypecast. Public Skypecasts are listed online, with chat room
titles and descriptions.
Skypecasts comprise three aspects of participation: listening, waiting, and talk-
ing. Interlocutors are placed in a “listening area” when they first enter a Skypecast.
In the listening area, they can only listen to the ongoing conversation or upgrade
their status to “waiting”. Moving into the “waiting area” signals to the host that the
interlocutor is ready to take part in the conversation. The host has control over who
moves from the waiting to the “talking area” and when. Once interlocutors are
moved into the talking area, they are free to communicate as they would on the
telephone. Screen names are visible for all interlocutors to see on the user inter-
face. There are no buttons to push to bid for the conversational floor or to ask for
permission to talk.
The Skypecasts that are discussed below were all advertised as English-speak-
ing chat rooms. Many of the interactants have never met before. Consequently, a
large portion of Skypecast talk deals with getting acquainted. The nationalities and
linguistic backgrounds of the interactants are too numerous to list here; most come
from countries where English is not spoken as a first or second language. Data were
recorded using a software program called Pamela. This program recorded all ver-
bal communication, but it was unable to capture text communication. Using voice
and text messages, Pamela notified all Skypecast participants that they were being
recorded. Participants were also told that the recording would stop if they did not
wish to be recorded. Several research assistants collected the data for the purpose
of understanding how Skypecasts work. Over 50 hours of Skypecast interaction
have been analyzed.
4. Identification-recognition strategies
bile phone conversation openings). Extract 1 below illustrates how callers and call
receivers deal with this situation (the punctuation markers used in the examples
below represent intonation and stress; for the transcription conventions, see the
Appendix).
(1)
1 ((summons: i.e. telephone ring))
2 Nancy: H’llo?
3 Hyla: Hi:,
4 Nancy: FHi::.
(Hutchby and Barnett 2005: 153)
In line 1, the ringing of the telephone summons an answer, which Nancy pro-
vides in line 2. Although Nancy’s response is packaged as a greeting token, it also
performs other social actions. The greeting token serves to recognize the sum-
mons, but it also serves indirectly to self-identify Nancy through voice recognition
(Levinson 1983; Schegloff 1979). That is, when the callee answers the phone and
provides a greeting token, the caller is provided with verbal cues that will assist in
identification. Furthermore, unless Nancy was expecting a call from Hyla, Nancy
is not in a position to identify who the caller is. At this point, the only cue Nancy
has is the ringing of the phone. Nancy’s use of rising intonation demonstrates an
orientation to not knowing who is on the other end of the line (as opposed to say-
ing, “Hyla is that you?”). In line 3, Hyla provides her own greeting token. This ut-
terance signals that someone is in fact on the other end of the line, as well as pro-
viding an indirect self-identification cue. In addition, if Nancy is able to use this
cue to identify the caller, Hyla’s acknowledgement and indirect self-identification
in line 3 can function as a first-part greeting. In line 4, Nancy provides a second-
part greeting, demonstrating that she has in fact been able to identify Hyla in the
absence of non-verbal cues. Consequently, Hyla and Nancy open their conver-
sation according to the technological affordances and constraints of the telephone.
Although Hyla and Nancy are able to get on with the business of greeting with
ostensibly little difficulty, greeting tokens that are used to self-identify indirectly
may provide insufficient cues for identification. For example, in Extract 2, the call
receiver (C) is not able to identify the caller (G) with the response given in line 3.
(Numbers in parentheses indicate pause length in seconds.)
(2)
1 ((summons))
2 C: Hello?
3 G: Hello.
4 (1.5)
5 C: Who’s this.
(Schegloff 1979: 39)
As with Extract 1, a telephone rings, summoning the call receiver to answer the
call. In line 2, C provides an answer, and in line 3, G responds with a first-part
greeting. However, C is not able to use the verbal and paraverbal cues provided by
G’s greeting token to identify the caller. Instead of responding with a greeting of
his own, C explicitly requests the caller’s identity. This explicit request for identi-
fication is designed to overcome the fact that C is answering a summons in the ab-
sence of physical cues, and in this particular instance, cannot rely solely on the ver-
bal information provided in a greeting to identify the caller/summoner.
However, troubles in identification can be subverted when additional contex-
tualization cues are available to call receivers. With caller identification technol-
ogy, for example, call receivers can design their turns in a way that creates different
interactional possibilities (Arminen 2005). In Extract 3, SB is using her mobile
phone to call Kisha, who is also on her mobile phone.
(3)
1 ((summons))
2 Kisha: Simone
3 SB: Kisha
(Hutchby and Barnett 2005: 159
Here Kisha is able to answer the call by identifying SB, the caller. In so doing,
Kisha makes use of the caller identification feature on her mobile phone to shape
they way she opens the conversation. By bypassing the need to request identifica-
tion, Kisha demonstrates an ability to exploit what Hutchby (2005) calls “circum-
stantial contingencies”. That is, Kisha designs her turn-at-talk according to the
technology available.
Interlocutors in Skypecasts can also pre-identify with screen names. However,
unlike the one-to-one participatory structure of mobile phone openings, in Skype-
casts there are multiple interlocutors continuously entering and leaving chat
rooms. This “revolving door” feature, coupled with the fact that screen names do
not provide visual cues that signal speakership (cf. Lamy 2004), creates a set of
practices similar to telephone openings that lack caller identification technology.
In Extract 4, three interlocutors are in a Skypecast. No previous talk has oc-
curred before S1 makes his first contribution.
(4)
1 S1: hello?
2 (0.3)
3 S2: hello?=
4 S3: =helloo?
In line 1, S1 provides a greeting token. As with telephone openings, this utterance
does more than the business of greeting. S1’s greeting token also serves to make his
presence known. Much like answering a call after a summons, the rising intonation
in S1’s greeting token also seeks a response that would confirm the presence of any
co-interlocutors. In line 3, S2 provides a similar greeting token with rising inton-
ation. S2’s greeting also serves several functions (i.e., responding to S1’s turn,
allowing S2 to make his presence known, and seeking a response from co-inter-
locutors). In line 4, S3 follows with his own greeting token, this time with greater
stress and more animated intonation. The sequential importance of these greeting
tokens, all of which are delivered with rising intonation, is that they appear to be
necessary first steps in engaging in greetings and further topical talk in a setting
where interlocutors do not know and cannot see each other. The interlocutors are
not only acknowledging their presence with greeting tokens but also creating an in-
teractional space where the next speaker (co-interlocutor), if willing and able to
participate, must provide a receipt token of some sort. In the physical domain, this
type of greeting exchange would be similar to communicating in a cocktail party in
total darkness – as opposed to, say, two friends crossing paths in a corridor.
Although screen names are visible to all Skypecast participants, interlocutors
often seek verbal confirmation from a co-interlocutor, as they may not know who is
actively participating, busy texting another person, or has left the computer unat-
tended. In other words, seeking verbal confirmation from a co-interlocutor is
sometimes necessary in Skypecasts even if everyone in the room knows who is
“logged in”. Extract 5 provides an example of an interlocutor (S6) seeking verbal
confirmation even after two interlocutors have already established their presence
in the room. The extract begins with S5 answering a question posed by S4 (not
shown).
(5)
1 S5: oh you mean where are we from? (0.4) I am from India
2 (1.0)
3 S6: hello?=
4 S7: =y- y- a[-
5 S6: [let’s talk about something
6 (.)
7 S6: eh? (0.5) hello?=
For several lines of interaction before this extract, S5 and S4 have exchanged
greetings and biographical information, and in so doing, they indirectly signal to
the other interactants that they are logged in and communicating. Despite this, S6
enters into the conversation with a greeting token (line 3). This greeting token sig-
nals to the other interlocutors that S6 is in the room and ready to communicate. The
greeting token also serves to seek verbal confirmation from the other interlocutors
that S6’s has been heard, as demonstrated in the use of rising intonation (“hello?”).
After receiving no uptake for his proposal to “talk about something”, S6 uses a tag
question (“eh?”) and second greeting token (“hello?”) to confirm whether the other
interlocutors have heard him.
In SvCMC, moments of silence can lead to doubt about whether other people
are still online. In Extract 6, for example, two co-workers are connected to Thun-
derwire.
(6)
1 John: Is Judy here today?
2 Patty: Judy’s here, … and she was on … earlier. She is either
3 listening or … is not … here with us.
(Ackerman et al. 1997: 52)
In the absence of physical cues, screen names are sometimes used as a resource
to elicit verbal participation. In Extract 7, for instance, a Skypecast participant uses
a screen name to determine whether a co-interlocutor is in fact ready to communi-
cate.
(7)
1 S8: uh huh (0.8) oh we have Jan Lynn Lee eh- ohF (0.6) ((inaudible))
2 (0.6) uh- eh- okay uh- (1.2) Jan Lynn Lee hello?
3 (0.4)
4 S9: hello?
5 (1.2)
6 S8: yes=
7 S10: =hi
The extract begins with S8 realizing that S9 is logged in (“uh huh … oh we have
Jan Lynn Lee”). In line 2, S8 then seeks verbal confirmation (“Jan Lynn Lee
hello?”), which he receives in line 4 (“hello?”).
Yet another way of establishing who is on the other end of the line is to self-
identify. Some interlocutors find self-identification useful, as hearing multiple
voices without any physical cues can make other-identification difficult (e.g.,
matching voices with screen names in a chat room with four or more interlocutors).
In Extract 8, two interlocutors are conversing in a Skypecast with three other par-
ticipants.
(8)
1 S11: hello my name is Jim so (0.9) hello?
2 S12: I know [your
3 S11: [yeah=
4 S12: =name’s Jim we know (0.2) [you
5 S11: [okay
6 S12: said i[t
7 S11: [okay
8 S12: like I don’t know how many times (.) maybe fifteen times (0.4)
9 your names Jim ((begins laughing))
Two interactionally significant events precede this extract. First, S11 has told his co-
interlocutors that he believes it is necessary to self-identify, given the number of par-
ticipants in the chat room. Second, S11 has, on a number of occasions, followed his
own advice. In line 1, S11 designs his turn with a greeting token and self-identification
(“hello my name is Jim …”). After receiving no verbal uptake, S11 seeks confirmation
from his co-interlocutors (“hello?”). Although self-identification is a strategy that is
sometimes used and accepted in text-based chat rooms (Negretti 1999), S12 makes
it clear that he is well aware of S11’s name and does not require him to self identify.
In so doing, S12 demonstrates that he does not accept S11’s turn design strategy.
Similarly, in their study of Thunderwire, Ackerman et al. (1997) show how in-
terlocutors use turn-design strategies to manage speakership and listenership (see
also Hindus et al. 1996). In Ackerman et al.’s study, interlocutors used backchan-
nels to provide recipient uptake while managing tasks in a physical space. The use
of backchannels in Thunderwire showed how interlocutors exploit the technologi-
cal affordances of communicating in a voice-only medium.
(9)
1 Patty: John, are you there?
2 John: Uh-hm.
3 Patty: Thanks for this New York thing.
4 John: Uh-hm.
5 ((Typing, mouse clicks, male clears throat))
(Ackerman et al. 1997: 54)
In line 1 in Extract 9, Patty directly requests a receipt confirmation from John
(“John, are you there?”). Patty’s question is different from the confirmation-seek-
ing moves discussed in Extracts 4 and 5, where the use of “hello” did not specifi-
cally identify a co-interlocutor. According to Ackerman et al. (1997), John re-
sponds with an acknowledgement token that signals disinterest. While the extract
does not provide any information regarding whether Patty has understood John as
disinterested, it is clear from Patty’s response in line 3 that she has interpreted this
acknowledgement token as an indication to continue with her turn-at-talk. In line 4,
John provides the same acknowledgement token, and ostensibly continues with his
“offline” task (line 5). Consequently, John is able to exploit the absence of physical
cues by simultaneously doing disinterest and providing recipient uptake. This situ-
ation can occur in telephone and mobile phone conversations, where interlocutors
can signal “active” listening while engaging in other activities.
This section has identified the strategies interlocutors use to identify and rec-
ognize each other, and how interlocutors make their presence known verbally. The
extracts show how caller identification, screen names, and communicating in a
voice-only medium influence the ways in which interlocutors manage speakership
and listenership in telephone and mobile phone communication, Skypecasts, and
Thunderwire.
5. Turn-taking
In this section, we examine how turns are managed in SvCMC. The importance of
turn-taking in communication has been discussed extensively in the CA literature
(e.g., Sacks et al. 1974). Turns-at-talk serve as an interactional bedrock from which
interlocutors can do a range of social activities, from making their presence known,
as we have seen in the previous section, to reporting on the speech of a fellow in-
terlocutor (e.g., Clift 2006). With regard to voice-only communication, most of the
turn-taking observations that have been made come from telephone conversations
(e.g., Schegloff 1979). In these settings, there is typically one interlocutor speaking
at a time, and any occurrence of overlapping talk is initiated and dealt with in sys-
tematic ways (Schegloff 2000a). Furthermore, topics are proposed, discussed, and
abandoned singly. That is, interlocutors in one-to-one settings generally do not
propose, discuss, and abandon two or more topics simultaneously. We explore
whether these observations are applicable to SvCMC. Our discussion focuses on
topic management and overlapping talk.
Several studies of multiparty communication in face-to-face settings have
observed that one topic often transforms into two or more (e.g., Schegloff 1995).
For example, Speakers A, B, C, and D are discussing Topic 1. Speakers A and C
then turn to each other and begin discussing Topic 2, while Speakers B and D
continue with Topic 1. As a result, two conversational floors are formed in one
group in order to deal with Topics 1 and 2. In face-to-face settings, interlocutors
can use gaze and gesture to determine speakership and listenership (Goodwin
1999, 2003). Often these non-verbal cues are used to initiate and deal with situ-
ations when one topic transforms into two or more (Egbert 1997). In Extract 10,
for example, a group of friends are together having a casual face-to-face conver-
sation.
(10)
Floor A
1 Hans: he does somethi:ng,
2 (1.0) ((Bea shifts torso and gaze to Inge))
3 Bea: tastes goo:d inge. ((Inge shifts gaze to Bea))
Floor B
4 (0.2) (0.2)
5 (0.5) Inge: ye:s?
6 Hans: somehow pro= Inge: [really?
7 Bea: [uh hm
=duct examination (0.5)
8 Stef: wha:t? Bea: tastes yu:mmy.
(Egbert 1997: 9)
The extract begins at some point during Hans’ story-telling. In line 2, Bea shifts her
body and gaze to Inge, and then in line 3 comments on some food or drink. As Bea
nears her turn, Inge shifts her gaze to Bea, and provides a recipient uptake token in
line 5. Bea and Inge then continue communicating with each other, forming a sep-
arate conversational floor. Thus, the meeting of gaze between Bea and Inge repre-
sents the point or turn when one conversational floor transforms into two.
Despite the absence of gaze and gesture, multiple conversational floors have
been reported in some SvCMC environments. In Extract 11, for example, several
interlocutors are participating in Mad Hatter, an experimental audioconferencing
system that monitors communication and adjusts voice amplitude in order to facili-
tate multiple conversational floors. The arrows next to each line number indicate
Topic 1, whereas the line numbers without arrows correspond to Topic 2.
(11)
1f Y: would you rather: (.) have to wear big clown shoes
2 K: four
3 ounces isn’t that-
4 isn’t [that much, I wo]uld say liquid detergent
5f B: [big what shoes?]
6f Y: big clown
7f shoes (.) you know the
8f big [clumsy shoes ] that a clown wears,
9 A: [what was that?]
10 K: I don’t know, I’ll
11 let you choose that one,[since I chose the first] one,
12f B: [okay, I would rather:]
(Aoki et al. 2003: Excerpt I)
In line 1, Y asks a question regarding big clown shoes. Almost in overlap, K in
lines 2–4 follows with another question, this time regarding the weight of liquid
detergent. At this point there are two topics. In line 5, B aligns to Y’s topic by seek-
ing clarification of the clown shoes. After Y clarifies (lines 6–8), A aligns to K’s
topic by seeking clarification of the liquid detergent. In lines 10–11, K provides a
clarification for A, maintaining the liquid detergent topic. The transcript ends in
line 12 when B responds to Y’s initial question.
Although this exchange demonstrates the ability of interlocutors in multiparty
SvCMC to manage two topics simultaneously, each turn is designed according to
one conversational floor. This is most visible in B’s orientation to K’s turns in lines
10–12. Here B positions her overlapping turn after it is transitionally relevant to do
so (Sacks et al. 1974). That is, B waits until K appears to finish clarifying the liquid
detergent topic before responding to the clown shoes topic. It is difficult to say
from the transcript why these interlocutors manage two topics by orienting to the
turn-taking rules of one conversational floor. The ability to adjust body torque,
move spatial positions, and manage gaze are clearly important to communicating
in group talk (Isaacs and Tang 1994), although further investigations are needed to
explore whether an orientation to one conversational floor during the discussion of
two or more topics is the norm in multiparty SvCMC. One explanation is that there
is an implicit expectation in multiparty SvCMC that only one interlocutor should
speak at a time.
Below we examine two extracts that demonstrate orientations to a one-speak-
er-at-a-time rule. In Extract 12, several interlocutors are participating in a Skypecast.
(12)
1 S13: over twenty
2 S14: ah- ah- ah- are you- are you student are you working
3 S13: yeah (0.7) I’m work[ing ((inaudible))]
4 S15: [James ] what are you talking about? I’m
5 still here (.) and I’ll be here too
6 (1.7)
7 S16: excuse me can you hear my voice?
8 (0.5)
9 S14: hello? hello? hello? (0.7) please please wait ’til, we we complete our
10 chat okay?
11 (0.7)
12 S15: I’m sorry yeah go’head
13 (1.0)
14 S14: okay okay
S14 and S13 have for some time been discussing where they are from and exchang-
ing other biographical information. Thus, in the minutes preceding this extract
there has only been one conversational floor. The conversational floor continues in
line 1, where S13 states his age. After receiving this information, S14 in line 2 asks
whether S13 is a student or worker. In line 3, as S13 states that he is a worker, S15
enters the audio space with a series of utterances directed at S16. It is important to
note that while S14 and S13 were exchanging biographical information, S15 and
S16 had been sending each other private text messages. However, for some reason,
in this instance, S15 responds to S16 in the voice-only medium. Consequently, the
text conversation between S15 and S16 spills over into the conversational floor
shared by S14 and S13. In line 7, S16 orients to S15’s audio spill over by asking her
whether she has heard his voice (“excuse me can you hear my voice?”). In so
doing, S16 also enters the ongoing conversational floor. In lines 9–10, S14 re-
quests that the second conversational floor be put on hold until the first one is com-
plete (“please please wait ’til, we we complete our chat okay?”). In line 12, S15
also orients to her audio spill over as a pragmatic error, as she later apologizes to
S14. In this extract, therefore, a second conversational floor is treated as a dis-
ruption.
S14’s explicit reference to the turn-taking practices of the room is just one way
in which interlocutors can orient to a one-floor format. In Extract 13, a group of
different interlocutors are participating in a Skypecast.
(13)
1 S17: •hhh I:: want to::: (.) improve my Engleash and uh:: (.) study, (.) and
2 work (0.6) yes
3 S16: yes (.) how old are you?
4 (1.1)
5 S17: uh:: twenty four (.) no twenty five (.) hh.hh
6 (0.5)
7 S16: and uh: where are you from
8 (0.8)
9 S17: I am from Poland you know (0.3) do you know::=
10 S16: =Poland-
11 (1.7)
12 S17: Pol[and
13 S16: [go on
14 (0.4)
15 S17: do you kn[ow where it is
16 S18: [are you from Poland
17 (0.7)
18 S17: yes
19 S16: Poland of course I know u- e- (.) I am from Romania (0.4) I’m not an
20 American (.) •hh
This extract begins with S17 telling S16 the reason why he wants to visit Australia.
S16 then asks two questions (lines 3 and 7), which S17 answers in lines 5 and 9. As
S17 nears the end his turn in line 9, S16 latches on to the utterance by providing a
confirmation of the country in question (line 10). Although S17 and S16 in lines
9–10 do not necessarily engage in overlapping talk, the onset of contiguous talk
audibly signals that overlapping talk may occur. This signal is oriented to by S17
and S16 in line 11, where both interlocutors pause momentarily. This pause acts as
a yielding device, as the floor is now reset to allow one speaker to regain the floor.
Yet the pause can also act as a source of further overlapping talk, as the interlocu-
tors cannot use physical cues to determine who will regain the floor. This constraint
is evidenced in lines 12–13, where S17 and S16 talk in overlap. In line 12, S17 at-
tempts to regain the conversational floor as he repeats his country of origin,
whereas in line 13, S16 appears to be yielding the floor verbally (“go on”). How-
ever, both utterances occur in overlap. As with the first floor yielding move in line
11, a pause in line 14 follows an instance of overlapping talk. In line 15, S17 con-
tinues with the topic of country origin. However, S18’s attempt to participate in the
discussion results in further overlapping talk, causing yet another pause in line 17.
6. Repair
In this section we discuss how interlocutors do repair work, and examine whether
the repair strategies used in SvCMC are the result of communicating in a voice-
only medium. In their seminal work on repair, Schegloff, Jefferson, and Sacks
(1977) showed that there is an organizational structure to how interlocutors repair
and sometimes correct troubles in communication. Troubles can include anything
from problems in hearing to factually incorrect information. The interactional and
sequential organization of dealing with instances of communicative troubles con-
sists of who has highlighted the trouble source (i.e., the speaker or recipient) and in
what position the repair has occurred (e.g., in the turn following the trouble source
or two turns later). Many studies have revealed – through the process of uncover-
ing the interactional and sequential organization of repair – important insights into
setting-specific communicative norms and expectations (see Drew and Heritage
1992). In language classrooms, for example, teachers adjust their repair strategies
(14)
1 M.M.: uh, Helena, Helena, if I go to Taiwan, which (place) /pli:s/,
2 which /pli:siz/ can I visit Taiwan, Helena?
3 Helena: I beg your pardon me, I didn’t catch you …
4 M.M.: I go to Taiwan, uh, wha uh what /pli:s/ can I visit?
5 Helena: Oh, you mean, sightseeing?
6 M.M.: Uh, what landmarks what landmarks … can I visit in Taiwan?
(Jepson 2005: 89)
(15)
1 Moo-soon: I’m sorry, you are what? You are /wIk/?
2 Junko: Not /wIk/, but weak.
(Jepson 2005: 90)
In both Extracts 14 and 15, the speaker of the trouble source self-repeats. In line 1
of Extract 14, M.M. asks Helena what place she should visit in Taiwan. In line 3,
Helena initiates a repair sequence but does not specifically indicate where the
trouble is (see Drew 1997). M.M. responds to Helena’s other-initiation for repair
by repeating her question (line 4). In line 1 of Extract 15, Moo-soon – the recipient
of Junko’s utterance – other-initiates a repair sequence by specifically identifying
the trouble source. In line 2, Junko self-repeats the pronunciation issue.
While Jepson (2005) reports high instances of self-repetition in audio chat
rooms, it is difficult to say whether these repair examples are a result of the voice-
only medium. It is possible that the repair examples reported occurred because the
interlocutors were language learners, rather than, or in addition to, the medium of
communication. Despite the need for further research to confirm these observa-
tions, Jepson’s (2005) comparison study reveals important CMC differences. For
example, in text-based chat rooms, because interlocutors must sometimes deal
with sudden fluxes of text messages, interlocutors are less likely to scroll back and
highlight troubles in communication. Conversely, Jepson (2005) argues that be-
cause interlocutors in audio chat rooms cannot scroll back to what has been said
previously, troubles in communication are more consequential for ongoing talk.
Furthermore, troubles in communication in audio chat rooms are more likely to be
repaired, as interlocutors generally speak one at a time and follow one conversa-
tional floor.
Similar observations of repair strategies in voice-only media have been made
elsewhere. Aoki et al. (2003) compared two audioconferencing systems: one ex-
perimental, one non-experimental. In their observations of the non-experimental
audioconferencing system, interlocutors had difficulties managing multiple con-
versational floors. This difficulty led to an increased use of other-initiated repair, as
demonstrated in Extract 16.
(16)
1 B: sorry, slow down
2 I- I can’t hear you eh- too well,
3 would you rather: go around carrying the what?
4 CPR dummy?
(Aoki et al. 2003)
In their observations of the experimental audioconferencing system (Mad Hatter),
however, Aoki et al. (2003) noted that there were few problems in hearing. Inter-
locutors seamlessly managed multiple topics, as the experimental audioconferenc-
ing system adjusted the speech volume in each floor to its respective contributors.
Therefore, other-initiated repairs of hearing did not often occur in Mad Hatter.
In Skypecasts, repairs often occur for a different reason. The interlocutors in
Skypecasts are not fully acquainted work colleagues or classmates, as interactants
are in many audioconferencing system settings. Therefore, speaking to an inter-
locutor without referring to his or her screen name can sometimes lead to trouble in
communication. For example, in Extract 17, several interlocutors are getting ac-
quainted in a Skypecast. S18 asks S19 to introduce herself, but he does so without
identifying for whom the question is intended.
(17)
1 S18: so, can you introduce yourself
2 (1.6)
3 S19: eh? (0.4) meF
4 (0.9)
5 S18: yeGahF
6 (0.8)
The extract begins with S18 asking S19 to introduce herself (line 1). The long
pause that follows signals that there may be trouble in communication. In line 3,
S19 other-initiates a repair sequence (“eh?”). The “eh?” reveals that S18’s question
is the source of trouble. After a short pause in line 3, S19 ends her turn by seeking
confirmation whether the question is intended for her (“me?”). Later in line 5, S18
confirms that he was asking S19 to introduce herself. This example of repair shows
how trouble in communication can manifest in an audio chat room with multiple
interlocutors who are not fully acquainted. Although troubles in hearing occur in
Skypecasts, they are less likely a result of the voice-only medium. Rather, troubles
of the type examined in Extract 17 are characteristic of communicating in a multi-
party medium where the interlocutors cannot see each other. Similar observations
have been made for text-based chat rooms (see, for example, Negretti, 1999).
Four types of SvCMC have been discussed in relation to repair: HearMe, Aoki
et al.’s (2003) two audioconferencing systems, and Skypecast. Findings show how
interlocutors adjust or calibrate their speech behavior in order to accommodate the
affordances and constraints of communicating in a voice-only medium. All of the
extracts examined in this section point to the influence visual (or lack thereof) and
auditory cues have on how interlocutors deal with troubles in communication. For
example, interlocutors in Aoki et al.’s (2003) non-experimental audioconferencing
system were unable to manage multiple topics, although this was less of a problem
in the experimental system that attempted to mirror the audio environment of
multiparty face-to-face conversations. Similarly, in face-to-face interaction, gaze
and gesture can be used to identify the recipient of an utterance, but in Skypecasts,
repair may be necessary if explicit identification is not given.
This chapter has examined the interactional character of SvCMC. Our examples of
identification-recognition strategies, turn-taking, and repair have shown that inter-
locutors adapt their behavior according to the affordances and constraints of com-
municating in voice-only media. For example, in the absence of physical cues, in-
terlocutors use greeting tokens to seek the recognition of interlocutors present in
audio chat rooms, while in other situations the turn design strategy of self-identi-
fication is used to help recipients recognize and identify the speaker, although this
strategy can fail if interlocutors do not need this extra cue (see Extract 8). With
regard to turn-taking, interlocutors orient to the turn-taking practices of one con-
versational floor (Extracts 12–13), even when discussing two or more topics
(Extract 11). The preliminary findings of Jepson’s (2005) study revealed that inter-
locutors are more likely to repair troubles when they are communicating in the
voice medium. While Jepson’s (2005) study dealt with language learners, the com-
parison he made between text and voice media highlighted the influence technol-
ogy has on communication. The ephemeral features of audio chat, coupled with the
absence of visual cues, limit what interlocutors can comprehend while communi-
cating with multiple interlocutors (see Extract 16). These observations are similar
to the ones presented in our turn-taking section, where interlocutors demonstrably
avoided overlapping talk in Skypecasts (Extract 13).
Future research will need to examine whether these phenomena exist in other
SvCMC media. In particular, future research will need to continue investigating
the similarities and differences of text and audio chat room communication, es-
pecially as voice and video media are more widely used in today’s highly global-
ized world. As technologies advance and more media options are available, CMC
research must examine why and how interlocutors choose among different media,
and analyze how those choices affect communicative patterns, norms, and expec-
tations. For example, how does simultaneous text and voice communication differ
from voice-only communication? How do interlocutors make the interactional
transition from one medium to another? As Extract 12 shows, interlocutors do not
only communicate in the voice medium in SvCMC environments. In order to de-
velop a more comprehensive understanding of the interactional character of
SvCMC, studies must capture all media that are being used in a system and exam-
ine how each communicative tool is managed (cf. Sindoni, forthcoming). Fur-
thermore, researchers should consider video recording the physical space in which
SvCMC takes place (cf. Beißwenger 2008) in order to understand what happens
“behind the scenes”. For instance, how does the physical space affect voice-only
communication? How do interlocutors use the keyboard and mouse to engage in
the various media available in one system?
Despite many unanswered questions, our review of the literature and analysis
of Skypecast data provide some insight into SvCMC. For example, this chapter has
revealed interactional similarities across SvCMC media, despite each medium pos-
sessing different technological affordances and constraints. Whether one-to-one,
on the telephone, or in a multiparty audio chat room, interlocutors compensate for
the absence of visual cues by designing their turns-at-talk in order to manage
speakership and listenership. This is most visible during those first instances when
two or more interlocutors begin a conversation and/or take part in an ongoing con-
versation. For example, during telephone openings, interlocutors take part in a
series of identification-recognition exchanges before pursuing topical talk in very
much the same way interlocutors in audio chat rooms seek to establish whether co-
interlocutors are ready to communicate. This would suggest that substantial differ-
ences do not exist between one-to-one voice-only media and multiparty SvCMC
with regard to the way interlocutors manage their interaction. Our observations and
findings of turn-taking and repair shed some light on this similarity. Despite the
availability and possibility of communicating with three or more people, interlocu-
tors in multiparty SvCMC orient to the turn-taking practices of one-to-one com-
munication. This is particularly interesting, as chat room communication in text-
based media does not always follow the turn-taking practices of one-to-one com-
munication (Herring 1999). Furthermore, audio chat rooms and audioconferencing
programs are designed to deliver simultaneous communication. Once a topic trans-
forms into two or more, however, the voice-only medium is incapable of providing
the audio-visual requirements for managing multiple conversational floors (Aoki
et al. 2003). While future research must examine these issues in more detail, this
chapter shows how interactional practices such as identification-recognition strat-
egies, turn-taking, and repair are mediated by the affordances and constraints of
communicating in voice-only media.
Appendix
References
Schegloff, Emanuel
2007 Sequence Organization in Interaction: A Primer in Conversation Analysis.
Cambridge, UK: Cambridge University Press.
Schegloff, Emanuel, Gail Jefferson, and Harvey Sacks
1977 The preference for self-correction in the organization of repair in conversation.
Language 53(2): 361–382.
Schullo, Shauna, Amy Hilbelink, Melissa Venable, and Ann Barron
2007 Selecting a virtual classroom system: Elluminate Live vs. Macromedia Breeze
(Adobe Acrobat Connect Professional). Journal of Online Learning and
Teaching 3(4). http://jolt.merlot.org/vol3no4/hilbelink.htm
Seedhouse, Paul
2004 The Interactional Architecture of the Language Classroom: A Conversation
Analysis Perspective. Malden, MA: Blackwell.
Sindoni, Maria Grazia
Forthcoming Through the looking glass: A social semiotic introduction to the study of
video chats. Text & Talk.
Singer, Andrew, Debby Hindus, Lisa Stifelman, and Sean White
1999 Tangible progress: Less is more in Somewire audio spaces. In: Proceedings of
CHI ’99, 104–112. New York: ACM Press.
Tanskanen, Sanna-Kaisa and Johanna Karhukorpi
2008 Concessive repair and negotiation of affiliation in e-mail discourse. Journal of
Pragmatics 40: 1587–1600.
Wainfan, Lynne and Paul Davis
2004 Challenges in Virtual Collaboration: Videoconferencing, Audioconferencing,
and Computer-Mediated Communications. Arlington, VA: RAND Corporation.
Waldvogel, Joan
2007 Greetings and closings in workplace email. Journal of Computer-Mediated
Communication 12(2), article 6.
http://jcmc.indiana.edu/vol12/issue2/waldvogel.html
Wang, Yuping
2004 Supporting synchronous distance language learning with desktop videoconfer-
encing. Language Learning & Technology 8(3): 90–121.
Wong, Jean
2000 Delayed next turn repair initiation in native/non-native speaker English con-
versation. Applied Linguistics 21(1): 244–267.
Woodruff, Allison and Paul Aoki
2004 Push-to-talk social talk. Computer Supported Cooperative Work 13: 409–441.
Yamada, Masanor and Kanji Akahori
2007 Social presence in synchronous CMC-based language learning: How does it
affect the productive performance and consciousness of learning objectives?
Computer Assisted Language Learning 20(1): 37–65.
1. Introduction
2. Background
When addressees are disappointed in their expectations of relevance, they rarely con-
sider as a possible explanation that the communicator is not really trying to be optimally
relevant. It would be tantamount to assuming that the apparent communicator is not
really addressing them, and perhaps not communicating at all. (Sperber and Wilson
1986: 159)
Being relevant also requires effort on the part of the speaker, including main-
taining awareness of the discourse context and the audience’s current knowledge
state, both of which shift as the conversation proceeds. McCann and Higgins
(1992) make explicit the effort being relevant entails in their characterization of
communication as a “game” with “rules”. In addition to being relevant, coherent
and comprehensible, giving neither too much nor too little information, and con-
veying the truth as they see it (Grice’s maxims), communicators should take the
recipient’s characteristics into account and produce a message that is appropriate
to their communicative intent and to the context and the circumstances. Recipi-
ents, in turn, should take the communicator’s characteristics into account, deter-
mine the communicator’s communicative intent or purpose, take the context and
circumstances into account, try to understand the message, and provide feedback,
when possible, to the communicator concerning their understanding of the mess-
age.
The maxims or “rules” of cooperative communication apply especially in task-
focused situations where the goal of communication is to be informative. They
may not be observed in antagonistic or playful situations where cooperation is not
expected (Schwarz 1996). Sperber and Wilson (1986) also note several social situ-
ations in which addressees may relax their expectations that speakers will try hard
to be relevant: informal conversation among friends in a pub after work, teachers
encouraging students to communicate freely or creatively, a master talking to his
servant (although the servant should always make his or her communication to the
master relevant) (160–161). These situations grant a special dispensation to speak-
ers not to be optimally relevant. However, a person who routinely fails to be rel-
evant without such dispensation is generally considered a bore, and others may
avoid his company.
The ideal of relevance in spoken conversation can be represented schemati-
cally as in Figure 1. (A = initiator, B = responder; the solid arrow indicates a mess-
age intended to be optimally relevant.)
The models described above assign to relevance a key role – indeed, a partially-
definitional role – in human communication. Yet none considers modality of com-
munication; rather their claims are presented as universal generalizations, assumed
to be valid regardless of the medium through which the communication takes
place. At the same time, it is clear from the examples given to support both models
that the default “conversation” is produced by means of spoken language. This is
not surprising, given that until the advent of the Internet, there was no readily
available means of exchanging written messages with sufficient speed and reliabil-
ity to earn them the designation “conversations”.1 Now, however, CMC in the form
of email, chat, instant messaging, newsgroups, blogs, microblogs, and the like pro-
vides an opportunity to address “universal” claims about how conversation works,
taking modality into consideration. Yus (2010) suggests that the technological
properties of “cyber-media” affect what counts as relevance, and that “a number of
‘alterations’ of relevance may be produced by the different qualities of these
media” (16). For example, in textual modes of CMC, which lack the contextual
cues present in face-to-face communication, more information gaps have to be fil-
led in inferentially. This chapter is concerned with the norms of relevance in com-
puter-mediated exchanges. It explores how relevance in CMC differs from that in
spoken conversation, explains what might plausibly account for the differences,
and concludes by considering the implications of modality-specific norms of rel-
evance for pragmatic theories of communication.
1999; Herring 1999) – such that simultaneous feedback is lacking as senders type
their messages. Garcia and Jacobs (1999), who make similar observations, charac-
terize real-time chat as “quasi-synchronous” for this reason. It is difficult for con-
tributors to make their contributions relevant to what has gone before when they do
not know what messages will appear before theirs. As a consequence, adjacency
pairs are not infrequently disrupted in 1-way CMC systems by messages from
other conversations taking place in the same system, and threads of conversation
can become intertwined, as shown in example (1).
Disrupted adjacency results in unintended relevance violations, which can
cause online conversations to appear incoherent. For example, it can generate am-
biguity and confusion about which message is being responded to, especially if
multiple threads of discussion are intertwined. A number of compensatory strat-
egies for dealing with the “incoherent” nature of computer-mediated text have
been reported in the literature. McCarthy, Wright, and Monk (1992) observed that
during text-based exchanges participants tended to address the intended listener
more explicitly in an effort to maintain coherence (see also Werry 1996). Lam and
Mackiewicz (2007) identified three strategies that their IM participants in a work-
place setting used in order to maintain cross-turn coherence: short, multiple, and
sequential transmissions; topicalization; and the use of performative verbs (on per-
formatives in CMC, see Virtanen, this volume). Örnberg Berglund (2009) found
that international students using instant messaging in a design course maintained
coherence in cases of disrupted adjacency through lexical repetition and other
forms of cohesion (such as lexical substitutions, anaphora, and explicit linking ex-
pressions). However, even in the absence of cohesive devices, it was clear which
message was being referred to most of the time, either because of the timing or be-
cause, in the case of second-pair parts of adjacency pairs, “the sequential structure
of interaction [wa]s in itself an important clue in coherence creation” (12).2 These
adaptations notwithstanding, coherence remains desirable in task-oriented CMC.
In an experimental study, Ho and Swan (2007) found that student postings to an
asynchronous discussion forum that were low in relevance in relation to immedi-
ately preceding postings received fewer responses, whereas “postings that were
new, personal and relevant received the most responses” (7).
In recreational contexts, in contrast, the incoherence caused by disrupted adja-
cency may have advantages, including fostering playful communication. Much
humor exploits violations of Gricean maxims, especially the maxim of relation
(Attardo 1994; Chiaro 1992; but cf. Yus 20033). Disrupted adjacency creates
maxim violations when adjacent messages are considered side by side; these unin-
tended juxtapositions can suggest humorous interpretations (Herring 1997, 1999).
More generally, there is a tendency in recreational CMC environments for partici-
pants to free associate, sometimes giving rise to chains of associations that digress
rapidly away from the original topic of conversation. Herring (1999) gives an
example from a chat channel in which the conversation topic shifts in rapid se-
quence from blow-up dolls to a bald female singer to pool balls to pool tables to
being under the table (in a drunken stupor). The playful nature of online chat has
also been remarked upon extensively by Danet (2001; Danet, Ruedenberg-Wright,
and Rosenbaum-Tamari 1997), albeit not in connection with the notions of coher-
ence or relevance. The analysis presented in the following sections connects rel-
evance violations to playful behavior even in contexts where disrupted adjacency
does not occur, via what I suggest is an emergent norm of loosened relevance in
recreational CMC.
3.1. Data
Synchronous CMC resembles spoken conversation in many respects, despite the
fact that it is typed on a keyboard and displayed as text on a computer screen: Ex-
changes take place in (near-)real time, messages are typically short, and it is often
used for phatic communion, features which have earned it the label “chat” (Herring
2010). In the case study presented below, relevance is analyzed in extended
samples of interaction from two popular modes of synchronous CMC: IRC (Inter-
net Relay Chat) and MUDs (multi-user dungeons, domains, or dimensions).
IRC was the first CMC protocol that enabled real-time textual exchanges
among large numbers of people at different physical locations.4 Introduced in
1988, the basic protocol has been borrowed by commercial Internet service pro-
viders such as AmericaOnline, web chat forums, and for communication in multi-
participant online games, among other Internet environments (see Paolillo and Ze-
lenkauskaite, this volume). IRC chat spaces are metaphorically partitioned into
channels according (loosely) to the topic of discussion or the kind of people who
are there (#Christians, #thirtysomething, #gaysex, etc.), and users can join more
than one channel simultaneously. The majority of chat users at the time these data
were collected appeared to be teenagers and young adults who used IRC mainly for
informal social interaction and to meet people online.
MUDs originally developed in the late 1970s as networked, multi-participant
role-playing adventure games. The original game MUDs included a real-time chat
feature, which eventually became so popular that purely social MUDs – the best-
known of which is LambdaMOO – started to appear in the late 1980s, and edu-
cational MUDs developed from them in the early 1990s. MUDs are the textual pre-
cursors of graphical social chat environments such as Second Life. The MUD on-
line environment is typically based on a geographic theme – e.g., a fantasy
landscape, a house, a university – textual descriptions of which participants can
view by typing “look”, and through which they can navigate using text-based com-
mands (e.g., “go east”); these features have earned MUDs the designation “text-
based virtual realities”. At the time these data were collected, regular participants
in MUDs tended to have some programming skills, be a little older, and feel more
commitment to the community of users on the MUD than participants in IRC;
moreover, individual MUDs had distinct cultural practices.5
IRC and MUDs share a number of system features that are potentially relevant
to the study of inter-turn coherence. Both are multi-participant, synchronous, text-
only modes of CMC that transmit messages in their entirety, rather than keystroke
by keystroke. This means that interlocutors do not know that a message is being
written to them until it is completed and sent, precluding the possibility for simul-
taneous feedback. Moreover, both IRC and MUD servers distribute messages in
the linear order in which they are received without regard for what they are re-
sponding to, which often results in disrupted adjacency of otherwise logically-re-
lated turns. In both systems, the arrival and departure of each participant is an-
nounced automatically by system messages (which resemble ordinary messages on
the channel or MUD), further interrupting users’ conversations. Finally, conver-
sation in IRC and MUDs is semi-persistent, remaining on users’ screens until new
messages cause it to scroll upward, and accessible for review (by scrolling back)
for as long as the user is logged on. This feature makes it possible for participants
to follow multiple threads of conversation, supplementing their memory with a
textual record of what has been said (Herring 1999). At the same time, depending
on the number of active participants, synchronous CMC can scroll by quickly and
require participants’ full attention to follow along; if one chooses to review the
previous conversation, one does so at the risk of losing track of the current con-
versational thread. It was not uncommon at the time these data were collected for
new users to complain that multi-participant computer chat was “confusing” or
“chaotic”; all of the above-mentioned features contribute to this effect.
The study presented below is based on qualitative, interpretive analysis of two
corpora. The IRC corpus consists of approximately 60 half-hour samples of public
group interaction that I logged at irregular intervals between 1995 and 1998, dur-
ing the peak of IRC’s popularity, as part of a larger ethnographic study of IRC. The
examples discussed below are drawn from four social chat channels: #bayarea,
#chatzone, #france, and #punjab.6 #Chatzone is an open-topic social chat environ-
ments, and the other three are geography/culture based. The MUD corpus includes
both social and educational chat: Two social samples are drawn from logs archived
on the web (not by the author) of interactions from TinyMUD (1991) and from
LambdaMOO (1995), and two educational samples are from online rhetoric
classes taught by instructors in different parts of the U.S. (circa 1995–1996) on
LinguaMOO. Only conversation logs from the MUDs are considered; program-
ming activity, descriptions of the virtual geography, etc. are excluded from this
analysis. In addition, two examples of human interaction with artificial intelligence
agents on MUDs are drawn from published sources (Isbell, Lee, Kearns, Kormann,
Singh, and Stone 2000; Turkle 1995).
3.2. Analysis
The basic claim of this section is that the requirement of optimal relevance is re-
laxed in recreational synchronous CMC in response to constraints imposed by the
medium, resulting in loosened norms of cross-turn relatedness, or loosened rel-
evance. Three kinds of evidence are presented to support this claim. First, dis-
rupted adjacency of logically-related turns is shown to be pervasive in both cor-
pora, due to interpolated system messages and overlapping conversational threads.
Second, evidence is presented that relevance is frequently violated even in non-ad-
jacent logically-related turns, including as a recurrent strategy of some partici-
pants, and that such violations are socially acceptable. Last, it is shown that par-
ticipants interpret contributions as intentionally relevant even when their content is
tenuously related (e.g., because they were produced by a preprogrammed robot in-
capable of intentionality), a further indication that loosened relevance has attained
normative status in these environments.
(5) <Youp> bon, alors, qu’est-ce qu’on disait de bien deja en ce moment?
[<Youp> right, then, what was it we were talking about again here?]
370 Spitfire says, “I feel that my attention is divided into so many different
places that I can’t make any really good points because they will be
lost in the shuffle.”
The lack of coherence caused by disrupted adjacency leads some users to evaluate
computer-mediated conversation in negative terms: as “bad”, a “free-for-all”,
“chaos”, and “not intelligent”.
At first blush, this reaction would seem to support Grice’s and Sperber and Wil-
son’s assertions that conversation without sequential relevance is irrational and/or
non-communicative. Clearly, exchanges that are constantly interrupted by irrel-
evant messages do not obey the principle of sequential relevance; responses in-
tended to be relevant lose their relevant status by appearing too late or with too
many irrelevant turns in between. However, the comments cited above were made
by inexperienced users, who still were operating with expectations of relevance
carried over from spoken conversation, and who found themselves struggling with
the computer medium in an effort to uphold the spoken norm. This situation – what
we might consider the initial stage in the process of “loosening relevance” – is rep-
resented schematically in Figure 2. (Vertical dots indicate that other messages in-
tervene between A and B; jagged lines represent system “noise”.)
Figure 2. The spoken norm of relevance struggling against noise in the CMC medium
However, after about an hour into the class, Spitfire types, “I am sort of starting to
understand this craziness”. Subsequently, she and several other students who had
complained previously start to produce non-serious, off-topic messages them-
selves (the text is presented as typed by the participants, preserving typing errors):
(9) (As the teacher is taking roll call online)
146 Delphi tosses a piece of popcorn high into the air over Pheadrus’s
head.
150 mick sings, “Delphi, Delphi Bo, Belphi, Banana, Fana, Fo, Felphi,
Me, My, Mo, Melphi, Delphi …!”
151 Seinfeld catches the popcorn inhis mouth
152 Kristen asks, “Ryan, I’m at Rensselaer Polytechnic Institute in New
York. Where are you?”
154 Dennis disembowels mick with a demonic GRIN.
155 Zaphod tackles Pheadrus to get teh popcorn
159 Effluvia slices Dennis REAL THIN and tiles the floor with the result.
160 Spitfire asks, “Hey, can I get some popcorn too?”
In line [146], Delphi whimsically tosses a piece of virtual popcorn in the air, and
other students play along with this script in lines [151], [155], and [160]. Mick’s
“singing” in line [150] is a nonsensical non sequitur triggered, presumably, by the
appearance of Delphi’s name, and Kristen’s message directed to Ryan is part of a
personal exchange they are carrying out while the teacher is taking roll.
Far from disapproving of such irrelevant behavior, the teacher participates in it
(see Effluvia’s message in line 159), including producing non sequiturs such as line
[68] below in response to criticism of her roll-taking method:
(10) 63 Vixen wonders when roll call will be finished
Arch’s assertion that he needs a nap in line [23] is presumably related to his yawn-
ing in line [10], although there has been no mention of tiredness in the intervening
exchange (which was on the theme of Auxilary’s response to seeing Arch join the
channel). Lines [25] and [26] are much less plausibly related to what came before,
except inasmuch as they are constructed according to the template “I need X” in-
troduced in [23]. Nor does Auxilary’s question about “tuxedo mask” in line [24] re-
late to anything that has been said in the previous discourse, although since Arch
and Auxilary appear to know one another offline, it may refer to knowledge that
they share. In these messages, Arch and Auxilary appear to be searching for a mu-
tually agreeable topic of conversation, a scenario supported by Arch’s cooperative
response to the “tuxedo mask” question in line [30].
At this point, WendyCA intervenes in line [28] with an apparent relevance vi-
olation that can be considered a flouting, since it can be related to the idea of “need-
ing to get laid” via a sequence of logical inferences. (Assuming that Arch is male
and heterosexual, getting laid involves finding a woman and seducing her; women
are seduced through their emotions (their “heart”); women like to have money
spent on them; people spend money using credit cards.) Auxilary’s response in
[30] is relevant: He has a credit card, therefore he can hope to seduce a woman (im-
plied: WendyCA, who is female). The credit limit on the card is also relevant; the
more money he can spend, the more likely the woman is to have sex with him, ac-
cording to the logic established by WendyCA’s assertion. Her response in [33]
stretches the limits of relevance, however, in that she rejects Auxilary as a potential
sexual partner, despite the fact that he satisfies her previously stated criterion. The
intended implicature appears to be that Auxilary is so sexually undesirable that
even his credit card will not help him. Auxilary’s retort in [35] is also difficult to in-
terpret logically, although it can generally be understood as degrading, to the effect
that WendyCA is not worthy of dating his whole person, and possibly implying that
his right hand is itself degraded, e.g., because it performs acts of masturbation. Fin-
ally, Arch’s remark in line [34] is a non sequitur, related to the previous context
only through the repetition of the word “woman”.
Several points are illustrated by this example. First, all three participants re-
peatedly produce turns that are loosely relevant as a general strategy of interaction.
Loosened relevance is the norm rather than the exception in this sample, which is
typical of the IRC corpus in this respect. Second, loosened relevance can operate
through the mechanism of word association: “the way to a woman’s heart” in [28]
provides the context for “women” getting promotions in the workplace in [34]; “I
need a nap” leads by association to other things that the speaker “needs”. This lexi-
cal repetition is a very simple and superficial form of inter-turn relation, in that it
takes no account of the original speaker’s pragmatic intention. However, it has the
advantage of being easy to achieve under time pressure – one need only scan the
previous messages as they scroll past and respond to whatever word catches one’s
attention. Finally, despite the fact that the conversation jumps around and is some-
what difficult to interpret, no participant complains or expresses frustration with
the interaction. On the contrary, the participants appear to be enjoying themselves,
as evidenced by the use of smiling emoticons at the ends of lines [32] and [33].
This supports the view that loosened relevance is acceptable, and even enjoyable,
in IRC.9
The situation illustrated in example (10) represents the second stage of the
weakening of relevance in synchronous CMC, as represented schematically in Fig-
ure 3. The dotted line indicates weakened relevance.
Figure 4 takes this process even further. Example (11) shows that loosened rel-
evance is preferentially used even when no system messages intervene, in ex-
changes among a small number of participants (hence with fewer overlapping ex-
changes). That is, loosened relevance is found even in the absence of system noise.
The situation in which loosened relevance has become the default for synchronous
CMC interaction is represented in Figure 4.
knowledge, creativity, and “make sense” – that is to say, are contextually rel-
evant.
Remarkably, despite the fact that most AI agents are only capable of very shal-
low (if any) true “understanding”, human CMC users often find them convincing
as interlocutors. This is true even when the agents produce irrelevant or margin-
ally-relevant responses. Consider the following exchange that took place on Tiny-
MUD between Julia Bot and Barry, a human male intent on seducing “her”
(example from Turkle 1995: 92):
(12) [1] Barry says, “Are you still holding out on me?”
[2] Julia says, “I mean, have I convinced you I’m a computer, yet?”
[3] Barry says, “I guess you’re real … but I’d still like to get you in my
DEN to run tests on you.”
[4] Julia says, “I live for your approval, Barry.”
[5] Barry says, “I’m so glad to hear that. Why don’t you come over to my
home.”
[6] Julia, “I want to have a child someday.”
[7] Barry says, “What a coincidence. I too wish to make a baby …
someday.”
[8] Julia says, “The date is Tue Jul 21 23:57:31, Barry.”
Julia’s first two messages work as appropriate responses, despite the fact that
line [2] does not make logical sense (it should be “you mean” instead of “I mean”,
and Julia is programmed to attempt to persuade people that she is NOT a com-
puter, not that she is one). Nonetheless, it successfully generates an appropriate
implicature, namely that Julia is not a “computer”. Line [4] is an example of one of
a number of snappy rejoinders Julia has been programmed to produce in response
to “come-on” lines.11 Presumably, in this case, Julia recognizes the input as a
come-on by virtue of its adherence to a pattern of the type “I’d like to (verb) (prep-
osition) you”. The implicature generated is ironic: Julia does not care whether
Barry approves of her.
In contrast, lines [6] and [8] are much less plausibly relevant. They appear to
have been produced as a result of associative keyword matching, one of the sim-
plest means of creating connection between two utterances. Barry’s use of the word
“home” in line [5] triggers a script for Julia that includes “having a child”; Barry’s
“(some)day” in line [7] triggers Julia’s “time” script. Neither of these meanings are
what Barry appears to have intended, and thus Julia’s responses are non sequiturs.
However, Barry takes them to be meaningfully communicative, drawing from
them contextually appropriate implicatures: in the case of [6], that Julia is flirting
with him by talking about having a child, an activity that involves sexual inter-
course; and in the case of [8], that she is joking again on the theme of being a com-
puter. The conversation continued from this point, with Barry never suspecting that
he was interacting with a bot (Turkle 1995).
This example is revealing for what it suggests about norms of relevance in con-
versational interaction on TinyMUD. All of Julia’s responses, strictly speaking, are
relevance violations (although [2] and [4] can be interpreted as floutings), yet her
interaction strategy passes as normal in the context of the MUD. In this respect,
Julia’s conversation resembles that of Arch, Auxilary, and WendyCA in the #bay-
area example.
The parallels between human and bot conversation in MUDs are even more
evident in the following example. The bot in this case, Cobot, resides in the living
room, or central public area, of LambdaMOO, where visitors to the MUD often in-
teract with “him” (example from Isbell et al. 2000: 5).
(13) [1] Guest [to cobot]: Do you know anyone I should send this home
vasectomy kit to?
[2] Sparklebug peers at Guest suspiciously.
[3] cobot [to Guest]: This is a joke in very poor taste.
[4] Guest [to Hfh]: He’s good.
[5] Guest [to Hfh]: Have you been tinkering with him?
[6] Hfh says, “I wish I could take credit, but I think he’s evolving.”
[7] Guest [to cobot]: You’re evolving?
[8] cobot nods to Guest.
[9] cobot says, “CITIZENS: ACIDHORSE DOES NOT EXIST.”
[10] Guest [to cobot]: he does so, I’m sure I’ve seen him
[11] cobot [to Guest]: What does this prove?
[12] Guest [to cobot]: my hovercraft is full of eels
[13] Guest [to Hfh]: He just made perfect sense. damn him.
[14] cobot [to Guest]: Heavens!
This exchange is notable in several respects. First, it gives the impression of being
an extended, coherent interaction in which the robot never fails to respond in a
plausibly human-like manner. Second, although Cobot produces a non sequitur in
line [9] (“Acidhorse does not exist”), Guest responds to it as if it were perfectly ap-
propriate (although he disagrees with the truth of Cobot’s statement) and produces
a non sequitur of his own in line [12] (“my hovercraft is full of eels”12). Indeed,
Guest’s utterances are arguably of more questionable relevance than Cobot’s – had
the bot produced lines [1] and [12], they would no doubt have been considered
pragmatically strange. What, exactly, is Guest implicating by responding “my
hovercraft is full of eels”? Third, and most tellingly, despite the strangeness and
blatant non sequiturs in the sample, Guest evaluates it as making “perfect sense”.
I submit that it makes sense especially in the context of a social MUD, where loo-
sened relevance is the norm, and where similar human-human conversations take
place on a regular basis.
How does Cobot produce relevant responses, surpassing even some human par-
ticipants in its apparent ability to do so? The bot appears to make use of more soph-
isticated mechanisms than syntactic pattern matching (in order, for example, to de-
termine that the second part of Guest’s utterance in line [10] is intended as “proof”
of the assertion in the first part). However, simple word mapping could account for
how it “knows” that any utterance containing the word “vasectomy” is a joke, and
for why it responds “Heavens!” to the profane expression “damn him”. In this re-
spect, Cobot shares similarities with Julia and with human CMC users who make
use of word association to create the appearance of relevance. Both bots and hu-
mans do so for reasons of simplicity: Word association is easy to program, and it is
easy for human users to manage even when they are cognitively overloaded with
confusing message displays, in that no deeper pragmatic understanding is required
of the intended meaning of the utterance being responded to. Thus bots’ “success”
is relative to humans’ “failure” – bots appear to be relevant, at least in part, because
humans fail to be optimally relevant (cf. Foner 1997). Synchronous, text-based
CMC environments, it would seem, lower the standards for relevance, effectively
narrowing the gap in performance between humans and machines.
4. Discussion
In the previous sections, I argued that a causal link exists between two phenomena
in synchronous CMC that have previously been considered unrelated: (a) a ten-
dency for disrupted adjacency of logically-related messages (Garcia and Jacobs
1999; Herring 1999) and (b) a tendency for users to produce fanciful (and digress-
ive) utterances (e.g., Danet et al. 1997; Herring and Nix 1997). I propose that a lack
of sequential relevance characterizes both phenomena, and that the system con-
straints that cause (a) lead to (b), which is then adopted by synchronous CMC users
in some environments as a communicative norm, as evidenced especially by users’
evaluation of bot conversation as human-like and coherent. Recreational CMC is
especially prone to exploiting loosened relevance for humorous and playful effect,
but as the examples from the LinguaMOO class show, relevance breakdowns occur
and may be exploited for play even in pedagogical chat contexts (see also Colomb
and Simutis 1996). Moreover, while the examples analyzed in this chapter are of
synchronous multiparty chat, disrupted adjacency occurs in other CMC modes that
share the properties of sequential posting and 1-way message transmission, includ-
ing dyadic chat (such as instant messaging) and asynchronous discussion forums
(Herring 1999). However, in contexts where the goals are information focused and
the communicative dynamics are serious, loosened relevance may not be tolerated
or exploited; rather, compensatory strategies may be employed to neutralize its po-
tentially disruptive effects, as the research surveyed in section 2.2 suggests.
Those effects can be attributed directly to properties of the medium. Multi-par-
ticipant, 1-way CMC systems place demands on users’ attention and understand-
ing. In synchronous CMC, in addition, there are time and typing pressures that ad-
versely affect the complexity of message production (Ko 1996). These factors tend
to result in superficial connections across messages, e.g., involving use of simple
cohesive strategies such as question-and-answer adjacency pairs and lexical repeti-
tion. Deep coherence that involves complex inferencing is difficult to sustain,
especially when multiple interactions are going on simultaneously and equally
“loudly” from the perspective of the user. It follows from this that synchronous
text-based CMC is not ideally suited for purposes that involve sustained cross-turn
coherence. Conversely, it appears to be well-suited for creative idea generation, re-
laxation and fun, and for stimulating interaction among people who might not in-
teract otherwise.
If this analysis is correct, it challenges prior characterizations of conversa-
tionalists as, by default, aiming to be “optimally relevant” (Sperber and Wilson
1986). CMC users who produce strings of non sequiturs do not appear to be aiming
to be optimally relevant or even necessarily relevant at all, so much as aiming to
participate in phatic social exchange with (and possibly to be considered pro-
vocative, clever, entertaining, and so forth by) other users. That is, in contrast to
Yus’s (2003) analysis of joking as relevant but not necessarily cooperative, play
with loosened relevance in recreational chat rooms appears to be socially cooper-
ative but not necessarily relevant.
Of course, similar behavior sometimes takes place off the Internet, as for
example in informal, playful social interaction among friends (Chiaro 1992), in
which the requirement to be optimally relevant is relaxed. Dispensations of opti-
mal relevance as mentioned by Sperber and Wilson (1986) – in the pub with
friends, talking to a servant – reflect alternative, context-specific norms of com-
munication. Loosened relevance in recreational CMC can be considered a context-
specific norm that grants a similar dispensation, allowing relevance violations even
when turn adjacency is not disrupted. The difference is that in multi-participant
chat, technological pressures limit interlocutors’ choices. The creative play that
arises from loosened relevance online can be viewed as users’ “making the best” of
what is in essence a limited communication medium – making it, in Walther’s
(1996) word, “hyperpersonal” – or more socially desirable even than face-to-face
communication – by optimizing the behaviors it affords.
Notes
10. In more recent AI research, machine learning algorithms have also been employed to
enable agents to adapt their behavior (“learn”) in response to positive or negative rein-
forcement (e.g., Isbell, Shelton, Kearns, Singh, and Stone 2001).
11. Newer AI agents, such as Apple’s Siri, are similarly programmed with various snappy
retorts to sexual come-ons, which are received often by agents with female personae.
A particularly clever example is the following (retrieved November 4, 2011 from
http://shitthatsirisays.tumblr.com/page/10):
User: Talk dirty to me.
Siri: The carpet needs vacuuming.
Siri’s response is ironically self-referential: It appears to be a non sequitur produced as a
result of an error in lexical mapping of the sort that AI agents often make. In fact it is
pragmatically relevant (via the inference of “rejection” that is generated by flouting the
relevance maxim). This is an example of what Foner (1997: 125) calls “clever but brute-
force programming”, in which an utterance is pre-programmed as a whole in response to
a specific, recurring prompt (e.g., “talk dirty”).
12. This line is from a Monty Python television sketch circa 1970 in which a Hungarian
tourist attempts to use a maliciously-written English phrasebook to make a purchase
from a tobacconist (Wikipedia 2011).
References
Attardo, Salvatore
1994 Linguistic Theories of Humor. Berlin: Mouton de Gruyter.
Brown, Gillian and George Yule
1983 Discourse Analysis. Cambridge, UK: Cambridge University Press.
Cherny, Lynn
1999 Conversation and Community: Chat in a Virtual World. Stanford: CSLI.
Chiaro, Delia
1992 The Language of Jokes: Analysing Verbal Play. London: Routledge.
Colomb, Gregory and Joyce Simutis
1996 Visible conversation and academic inquiry: CMC in a culturally diverse class-
room. In: Susan C. Herring (ed.), Computer-Mediated Communication: Lin-
guistic, Social, and Cross-Cultural Perspectives, 203–222. Amsterdam/Phil-
adelphia: John Benjamins.
Danet, Brenda
2001 Cyberpl@y: Communicating Online. London: Berg.
Danet, Brenda, Lucia Ruedenberg-Wright, and Yehudit Rosenbaum-Tamari
1997 Smoking dope at a virtual party: Writing, play and performance on Internet
Relay Chat. Journal of Computer-Mediated Communication 2(4).
http://onlinelibrary.wiley.com.ezproxy.lib.indiana.edu/doi/10.11 11/j.1083–
6101.1997.tb00195.x/full
Foner, Leonard N.
1997 Entertaining agents: A sociological case study. In: Proceedings of the First In-
ternational Conference on Autonomous Agents.
http://dl.acm.org.ezproxy.lib.indiana.edu/citation.cfm?id=267658.267684
Madden, Mary
2005 New demographic info and activities data. Pew Internet & American Life Pro-
ject. http://www.pewinternet.org/Commentary/2005/May/
New-Demographic-Info-and-Activities-Data.aspx
McCann, C. Douglas and Tory E. Higgins
1992 Personal and contextual factors in communication: A review of the “communi-
cation game”. In: Gün R. Semin and Klaus Fiedler (eds.), Language, Interac-
tion, and Social Cognition, 144–172. Newbury Park, CA: Sage.
McCarthy, John C., Peter C. Wright, and Andrew F. Monk
1992 Coherence in text-based electronic conferencing: Coupling text and content.
Journal of Language and Social Psychology 11: 267–277.
Nunan, David
1993 Introducing Discourse Analysis. Harmondsworth, UK: Penguin.
Örnberg Berglund, Therese
2009 Disrupted turn adjacency and coherence maintenance in Instant Messaging
conversations. Language@Internet 6, article 2.
http://www.languageatinternet.org/articles/2009/2106/Berglund.pdf
Paolillo, John C.
2011 Conversational codeswitching on Usenet and Internet Relay Chat. Lan-
guage@Internet 8, article 3.
http://www.languageatinternet.org/articles/2011/Paolillo
Pew Internet & American Life Project
2010 Trend data for teens. Pew Internet & American Life Project.
http://pewinternet.org/Trend-Data-for-Teens/Usage-Over-Time.aspx
Reid, Elizabeth M.
1991 Electropolis: Communication and community on Internet Relay Chat. Senior
Honours thesis, University of Melbourne, Australia.
http://www.ee.mu.oz.au/papers/emr/index.html
Reid, Elizabeth M.
1994 Cultural formations in text-based virtual realities. Master’s thesis, University
of Melbourne, Australia. http://www.ee.mu.oz.au/papers/emr/index.html
Schwarz, Norbert
1996 Cognition and Communication: Judgmental Biases, Research Methods and the
Logic of Conversation. Hillsdale, NJ: Erlbaum.
Sperber, Dan and Dierdre Wilson
1986 Relevance: Communication and Cognition. Cambridge, MA: Harvard Univer-
sity Press.
Turkle, Sherry
1995 Life on the Screen. New York: Simon & Schuster.
Walther, Joseph B.
1996 Computer-mediated communication: Impersonal, interpersonal, and hyperper-
sonal interaction. Communication Research 23(1): 3–43.
Weizenbaum, Joshua
1966 ELIZA – A computer program for the study of natural language commu-
nication between man and machine. Communications of the ACM 9(1):
36–45.
Werry, Christopher C.
1996 Linguistic and interactional features of Internet Relay Chat. In: Susan C. Her-
ring (ed.), Computer-Mediated Communication: Linguistic, Social and Cross-
Cultural Perspectives, 47–63. Amsterdam/Philadelphia: John Benjamins.
Wikipedia
2011 Dirty Hungarian Phrasebook. http://en.wikipedia.org/wiki/Dirty_Hungarian_
Phrasebook [accessed 7:15 pm EDT, November 3, 2011]
Yus, Francisco
2003 Humor and the search for relevance. Journal of Pragmatics 35: 1295–1331.
Yus, Francisco
2010 Cyberpragmatics: Internet-Mediated Communication in Context. Amsterdam/
Philadelphia: John Benjamins.
1. Introduction
with the existence of several different (virtual) realities to which users adapt and
which they help construct, alter, and/or maintain through CMC.
2. Performatives
Austin (1962) initially distinguished between saying and doing in terms of con-
statives and performatives. While constative utterances, such as The Pepys Library
is in Magdalene College, can be defined simply in terms of their logical properties
as being either true or false, a separate distinction altogether is needed for per-
formative utterances, that of their relative felicity in the communication situation
in which they are put to use. When the chair of a meeting utters the words the meet-
ing is (hereby) adjourned, the adjournment is thereby executed. But if a catering
person bringing coffee, tea, and delicious cakes were to enter the meeting room
uttering these same words, there might indeed be a (happy) silence, but the felicity
conditions would not be right for the utterance to function as a “happy” per-
formative and thus to succeed in (officially) adjourning the meeting. Performatives
thus consist of two simultaneous and interrelated acts: the locutionary act, or the
words being spoken, and the illocutionary act, or the conventional force of those
words in the communicative situation. Austin’s theoretical distinction between
constatives and performatives served its purpose of demonstrating that “every
genuine speech act is both [the locutionary and the illocutionary act]”, and that, as
in the first example above, “stating is only one among very numerous speech acts
of the illocutionary class” (1962: 146).
Austin (1962: 115–116) bound up the illocutionary act with three different
kinds of effects: (i) securing its uptake; (ii) producing conventional effect; and (iii)
inviting a response. Uptake is a precondition for the conventional effect of the illo-
cutionary act to occur. Hence, for a successful warning to have the intended effect,
the audience must hear it and take it to be a warning. In text-based CMC, uptake
may be relatively stable as long as users do not ignore or delete messages without
reading. Second, Austin distinguished the conventional illocutionary act from the
very different perlocutionary act, which is achieved by the producer of the per-
formative additionally or entirely by non-conventional means, perhaps uninten-
tionally. Examples include convincing, persuading, and getting someone to obey
or believe (Austin 1962: 117–118; see also the discussion in Sbisà 2009). Students
of pragmatics have since worried about the status of the indeterminate perlocution,
which may be (un)intentional and essentially resides in the domain of situated dis-
course interpretation; hence, discussions in the literature of perlocutionary acts and
effects may involve two different concepts (for an overview, see, e.g., Sbisà 2007,
2009). Third, Austin stated that performatives may invite responses in the form of
another particular (speech) act, e.g., offering – acceptance. While user intentions
are beyond access, responses to performatives in text-based CMC constitute evi-
(1) I hereby apply for the privilege of admission to Cambridge University Library.
I hereby solemnly promise not to use this privilege to the injury either of the
Library or the University of Cambridge.
Signed: __________________________ Date: __________________________
In this example, as elsewhere, performatives contribute to changing things as a
result of the speech act. Having performed the actions of applying and promising,
the applicant’s reality will no longer be quite the same. The receiving body of these
performatives, the library, will be in for a contextual change, too, through the
uptake of the two speech acts, and a response acknowledging their conventional
effects can be expected.
Despite its early interest in the (imagined) speech situation, speech act theory
has often had to give way to other kinds of approaches to discourse which place
greater weight on dialogicality and the dynamism of text and context. However, the
creation of social and institutional facts through performatives has been the con-
cern of scholars working, for instance, on cross-cultural communication (e.g.,
Blum-Kulka, House, and Kasper 1989), professional discourse (e.g., Kurzon 1986
on legal performatives), and the language of (semi-)public ritual and various cer-
emonies (e.g., Lakoff 2000 on public apology). We can, as Sbisà (2009: 242) ad-
vocates, “force the theory to evolve by confronting it with unforeseen phenom-
ena”. Section 4 is devoted to the truly unforeseen phenomenon of standardized
third-person social action found in CMC of many kinds, raising the issue of the li-
mits of performative theory. And unforeseen or not, the appearance of explicit first-
person performatives including hereby in informal computer-mediated discourse,
in turn, strongly suggests the need for full incorporation of more dimensions into
the theory, as will be argued in section 5. Performativity in CMC is predominantly
a matter of virtual realities, even though performatives may also change things in
users’ “real” world. Section 3 introduces some of the concepts related to the human
ability to create and cope with worlds of different kinds.
3. Etiolated performatives
Analysis of text and discourse takes into account evidence of the construction of
various text worlds and universes of discourse (see, e.g., Enkvist 1989). Interpre-
ting online performativity necessitates investigating interrelated ongoing discur-
sive activities, projection of user identities, and the creation, maintenance, and/or
alteration of virtual realities and communities. An application of performative the-
ory will thus have to grapple with what Austin (1962: 22) called the “doctrine of
the etiolations of language”, that is, language being “in special ways – intelligibly –
used not seriously, but in ways parasitic upon its normal use”. Etiolation, for Aus-
tin, might affect several features of the speech act, such as authority in quotation,
address in soliloquy, authority and address in acting, sense in joking, and reference
in fiction; yet, none of these etiolations would turn the speech act into a non-act.
Sbisà (2007: 471) points out that “an analysis of aetiolation as shifting or weaken-
ing features of the whole speech act might suggest to admit of fictional or virtual
entities, which exist but, so to say, in weakened ways”.
Austin’s discussion of etiolations of language has served as an impetus to the-
ory construction, not only in linguistics and philosophy, but also in the study of lit-
erature, performing arts, and culture. As Parker and Kosofsky Sedwick (2007: 202)
observe in their overview of performance studies, “[m]uch … has long since been
made of Austin’s parasite, which has gone on to enjoy a distinguished career in lit-
erary theory and criticism”. Austin’s own focus was explicitly on “performative ut-
terances, felicitous or not, … issued in ordinary circumstances” (Austin 1962: 22),
and his decision to exclude etiolations of language from the analysis does not mean
that investigations of such matters cannot contribute to theory construction beyond
that work. Quite on the contrary, the CMC evidence presented below warrants an
expansion of performative theory to incorporate other ways of using language, in
addition to what Austin regarded as the default, ordinary one.
At this point, a note on the “parasitic” nature of etiolations of language in CMC
is in order. CMC does not constitute etiolation of language in itself; to users it may
be as “normal”, “ordinary”, and “serious” as much of their use of language offline,
and as such it is in no way parasitic upon an offline host of speech or writing. How-
ever, CMC is a rich source of pragmatic phenomena that can be analysed in terms
of performativity, related to virtual realities under construction. It is these phenom-
ena that may involve situated etiolations of particular aspects of conventional per-
formatives, resulting in “special” use of speech acts.
Students of humour have discussed switches and blends between modes of
“bona fide” and “non-bona fide” communication (Attardo 1994; Raskin 1985).
Bona fide communication refers to “serious” discourse that conforms to Grice’s
(1975) cooperative principle. Non-bona fide modes of communication include, for
instance, lying, play acting, and joke telling, which vary in terms of social accept-
ance and expectations relating to particular situational and socio-cultural contexts.
Danet (2001) characterizes CMC as an inherently playful medium. The CMC
modes explored below manifest playfulness, generally involving mixing of bona
fide and non-bona fide communication, and users also engage in online play.
Applying the notion of etiolations of language helps to account for the conven-
tionalized use of CMC-specific third-person action and the repertoire of verbs that
can be put to performative use in virtual worlds, discussed in section 4, as well as
the highly “parasitic” explicit performatives, incorporating the formal marker
hereby, that occur in informal computer-mediated discourse, which are the concern
of section 5. In the examples presented below, performativity relates to the con-
struction of virtual worlds, even though users’ “real” world may also be affected
occasionally. The predominance of virtual reality is exhibited in the need, in these
CMC environments, for the initialism (i.e., consisting of the initial letters of the ex-
pression) irl ‘in real life’, which constitutes the marked alternative against the un-
marked virtual life. Cherny (1995) shows examples of this abbreviation in contexts
of confusion about the frame of reference, and she discusses the intertwined rela-
tionship of virtual realities and the real world, as constructed and mediated through
online discourse.
Text-based CMC, by its nature, is inherently performative, in that all actions come
into being because they are typed in, and social action is constituted almost entirely
by discourse. Herring (2004: 339) notes that it is quite possible to lose sight of this
fundamental fact; yet, “language is doing, in the truest performative sense, on the
Internet, where physical bodies (and their actions) are technically lacking”. This
section deals with what Herring (in press) calls a “syntactic innovation”: third-per-
son singular present tense utterances through which users perform social action in
text-based CMC environments.
Herring (in press) makes a distinction between two kinds of performative ut-
terances in CMC: “emotes” and “performative predications”. The first of these take
place through a pre-programmed command in online game environments that
allows users to “emote” in the third person, “to perform a social action by logging a
description of that action into a chat window” (Herring, in press). Hence, in (2)
below, from Cherny (1995), lynn waves appears on the screen when the user “lynn”
types /waves. In contrast to such virtual action, action in the real world is rather dif-
ferent: lynn waving in front of the screen is not (necessarily) what is understood to
be the case with emotes such as the one in (2). Herring observes that emotes are
still in use in game environments, and a default prompt inviting similar third-per-
son utterances was, until recently, available for status updates in Facebook. Simp-
son (this volume) presents examples of third-person action and “thinks” bubbles in
synchronous text-based chats.
The popularity of emoting has subsequently given rise in social chat environ-
ments to another type of performative utterance, i.e., performative predications,
which have spread across a number of CMC modes. Unlike emotes, these are typed
out in their entirety by users and often set off by asterisks and other typographic
cues. Herring includes in the category of performative predications a wide reper-
toire of stand-alone utterances: They can be inflected or uninflected, e.g.,
<grin(s)>; represent change not only through virtual actions but also virtual states,
e.g., *confused*; and include vocalizations, such as *sigh(s)* or *lol(s)*, as in-
flected nouns or verbs. In view of the conventionalization process it is of interest to
note the development of the acronym (i.e., a pronounceable initialism) lol ‘laugh-
ing (or: laughs) out loud’, into a noun or verb yielding new spellings (lawl, lulz)
and playful representations of laughter (lololol).
An early study by Cherny (1995) focuses on the use of emotes, third-person
simulated action through the standardized activity type function available to users
in a social MUD. Participants engage in social action in the virtual reality or in the
real world by referring to themselves and others in the third person. In MUDs, the
“emote” mode is distinct from the “say” mode, each of which has a separate com-
mand, as in the following example from Cherny (1995).
(2) lynn says [to Tom], “hi there”
lynn waves.
As evidence of the multifaceted use of the command, Cherny (1995) distin-
guishes five types of emotes in her data: (i) conventional action (lynn waves), (ii)
back-channeling (lynn nods), (iii) joking by-play (lynn hands some prepositions to
Tom), (iv) narration (lynn packs for her trip), and (v) exposition (lynn hated the
film). The first three bring about action in the virtual reality, while the last two con-
vey background information concerning the real world and the users’ mental
states. Apart from the last type of emotes, which manifests tense variation, the
tense is always the simple present. As stated in section 2, classic performative
speech acts employ what Searle (2001: 106) calls the dramatic present, this special
tense marking events that are “to be construed as instantaneous with the utterance”.
The simple present of third-person emotes is no different: The time of the action is
understood to be simultaneous with its appearance on the screen. In contrast, dy-
namic situations viewed as being in progress at the moment of speaking or writing
would be expected to manifest the present progressive, e.g., lynn is waving.
Type (i)-(iii) emotes indicate action in the virtual world that takes place by typ-
ing an utterance in which the verb, in the dramatic present, is both self-referential
and executive (cf. lynn hereby waves/nods/hands some prepositions to Tom) and, it
is assumed, non-cancellable. Kolko (1995) also addresses the issue of non-denia-
bility in her analysis of a harassing performative utterance that the victim tries to
cancel. Herring (2001: 623) argues that online discussions of what particular act
was enacted in a given virtual reality show that such actions can be performative,
as they count as acts solely by virtue of having been typed in.
Unlike in speech, where a warning shouted out to others in a gale may not be
heard, uptake is relatively regular in CMC due to the written record. It is not possible
to access intentionality (unless explicitly displayed by an interlocutor, for instance,
through self-repair), but responses provide analysts with an important source of
information concerning users’ interpretations of the (conventional) illocutionary
force and the situated perlocutionary effects of performatives. Responses – preferred
or dispreferred – are usual in interactional modes of CMC. Yet, they are not neces-
sary for uptake to have occurred; textual silence, too, can be revelatory of the effects
of a particular performative on the virtual world under construction.
The third-person reference is a CMC convention that was embraced early on.
Werry (1996) presents examples of emotes in IRC in which an arrow points back to
the userID. This convention is mimicked in the following performative predi-
cation, from the data to be discussed in section 5.
(5) <----------------- runs to bathroom to look at my stunning new self!!!!
It is of interest to note that utterances may start in the third person and subse-
quently switch to the first person, as in example (5). Herring (in press) relates these
inconsistencies to the systems that prompt third-person emoting, observing that
such utterances are generally “assumed to be unintentional and (at least mildly) in-
felicitous”. The syntactic distance of mentions may be argued to cause inconsist-
ency between the third-person verb and the first-person genitive determiner at the
end of an embedded clause in (5), as well as the clause-final adverbial in *makes
suggestive gestures with my hands and my jackhammer* (from the same source of
data). Such instances seem to point to a fuzzy, or at least flexible, boundary be-
tween the mental configurations of users and user characters.
The phenomenon of third-person action, pre-programmed or typed in by users
as part of their messages, is a fundamental aspect of text-based CMC. Unlike off-
line stage directions in the third person, they are executed upon typing. Users dem-
onstrate a need for emoting by continuing to type in their messages typographically
marked, third-person, present-tense, stand-alone performative predications, across
modes of social CMC, including texting (SMS). Typographic marking can be
conventionalized or elaborate. While what is emoted can also be conventional, it
is, in fact, sometimes quite creative. But popular emotes tend to be used repeatedly,
some abbreviated (e.g., lol), others adapted to the theme of a particular community.
Incidentally, in Baron’s (this volume) Instant Messaging data lol accounds for 90 %
of the abbreviations present. To come to grips with the contributions of emoting to
the construction of virtual realities, performative theory will have to allow for the
third-person reference to a particular user character found in emotes and per-
formative predications.
Furthermore, in addition to acts and activities, broader views of performativity
in CMC include events and states conveyed by emotes and typographically marked
stand-alone predications. These, as well as emoticons (smileys), “supply in-
formation about non-verbal aspects of communication which would have been dis-
cernible if the words had been uttered aloud in a face-to-face encounter” (Danet
2001: 18). However, contextualization cues (Gumperz 1982) such as emoticons,
emotes, and performative predications are generally interpreted as relating to the
states of affairs in a virtual reality under construction, rather than in the physical
world. Studies of emoticons show discrepancies between the indicated thoughts
and feelings and users’ actual facial expressions (Dresner and Herring 2010). Simi-
larly, Cherny (1995) cites evidence that one of the users in her study is not at all
laughing while typing in lol. The need for performative theory to incorporate sev-
eral different dimensions of language use is also apparent in the next section,
where we turn our attention to playful uses of predominantly first-person explicit
performatives in the “say” mode in informal online discussions.
Explicit performatives incorporating the marker hereby are associated with insti-
tutional discourse and public rituals of various kinds. But such constructions may
also be found in informal text-based CMC, warranting investigations of the moti-
vations behind their use. This section summarizes findings from a study of dis-
cussion board materials disclosing playful, etiolated uses of explicit hereby-per-
formatives, which have a multifaceted impact on the pragmatics of the interaction
at hand (Virtanen, in press). They are labelled “mock performatives” because they
deliberately set out to mimic playfully (parts of) established institutional per-
formatives, usually ones involving a high degree of authority. In such usage, the
formal marker hereby is employed to set off the official flavour of the mediated
institutional performative from the otherwise informal discussion. Unlike the play-
ful emotes and performative predications discussed in the previous section, these
mock performatives do not involve a particular “emote mode” or appear as typo-
graphically marked stand-alone units within messages. Instead, they are simply
(part of) the message.
The mock performatives under investigation were detected in September 2006
by searching for the keyword hereby in all messages across the publicly available
English-language discussion boards on a U.S-based website, where (according to
user profiles) predominantly female participants, of all ages, and mostly of Ameri-
can, Canadian, British, and Australian origin, review beauty products and discuss
makeup, fashion, and the like.2 Instances of the performative marker hereby were
found in 29 threads, originating in five different asynchronous discussion boards
on one and the same website. The data were not collected for purposes of quantifi-
cation; instead, the threads were studied using qualitative discourse-analytic
methods. It should be noted that the pace of interaction on the forums is at times
very rapid, a matter of seconds, rather than minutes, and users typically start typing
their messages directly on the subject line (allowing 100 characters, excluding
spaces), to be continued in the message box if necessary.
Example (6) illustrates a mock performative from the forums. User1 (substitute
for a nickname) expresses unhappiness with the unfortunate consequences of a poll
concerning different nail polishes on the discussion board on makeup, to which
User2 responds by typing an explicit performative construction: I hereby recruit
you to the nail board. User1 immediately posts a reply, vigorously protesting
against the recruitment.
(6) why do i feel the need to get every single nail polish mentioned in that poll? bad
bad poll User1 10:14 PM
– I hereby recruit you to the nail board :P User2 10:15 PM
• noooo! lol, more places to spend my $. I’ll never do my fingernails
though, i hate the feeling User1 10:16 PM
The explicit performative construction appearing in User2’s message is
grounded in authority playfully assumed regarding User1. As it is risky for inter-
locutors to assume authority over others, the performative is followed by a miti-
gating element signaling the onset of play: a smiley with a tongue sticking out. It is
important to note that even in playful use the conventional effect of the per-
formative is there. User2’s act of recruiting User1 to the nail board brings about
non-cancellable change in the virtual reality. This is recognized by User1 in her
emphatic reply, protesting against what has already taken place in the virtual
world. The virtual recruitment is a fait accompli that cannot be denied, only pro-
tested against (see Kolko 1995). User1 expresses the undesirability of its conse-
quences in a playful manner by the lengthening of the vowel in noooo, followed by
an exclamation mark, and the acronym lol. Thus the main effect of User2’s playful
performative is one of bonding.
Interlocutors readily respond to mock performatives by producing same-script
or other-script oriented discourse, as in example (7) below. The abbreviation
SO stands for ‘significant other’, and r/o ‘read on’ indicates that the text typed in
the subject box continues in the message box.
(7) On a break with SO – I deserve a mini-haul … right?? ;-) User3 9:43 PM
– [a series of short responses expressing empathy between 9:43–9:49 PM]
– The very best medicine there is! Who needs men when there’s makeup? :P
User4 9:47 PM
• Ok. I, User5, hereby do declare that any and all problems, breakups,
breaks, r/o arguments, or any other ‘situations’ with any SO be it man or
woman, totally warrant, deserve, and indeed require splurges, hauls and
indulgences of any kind. User5 9:51 PM
y
I second the motion! User6 9:52 PM
y thirded. Motion passes. User7 9:55 PM
y here, here! :D User8 9:53 PM
y SO HELP ME GOD. User9 10:07 PM
y You are CORRECT! :) User10 10:21 PM
thirded and declared passed. Meanwhile, the exclamation hear, hear! calls atten-
tion to the declaration, the textual cheering here being accompanied by laughter (a
smiley) to convey approval. Much later, 16 minutes after the declaration, User9
“shouts out” (capital letters) the ritual formula so help me God, customary in sol-
emn oaths and familiar to users from U.S. television series and films cast in cour-
troom settings, where witnesses are expected to take an oath to tell the truth ending
in these words. After another 14 minutes, User10 contributes with a simple ac-
knowledgement of the correctness of the declaration, emphatically and with a
smile.
Mock performatives predominantly function as “discourse transformers”, shift-
ing the discourse into a play mode. This allows users to take time off from their
main concerns, the discussion of beauty and fashion, without abandoning them al-
together. The discourse-transforming potential of mock performatives to instigate
or support a play mode in the discussion helps users produce a different frame in
the interaction, in Goffman’s (1974) terms. The play mode (play frame) thus
opened is distinct from the ongoing blend of bona fide and non-bona fide com-
munication that is characteristic of the inherently playful interaction (see Danet
2001).
One instance of the performative marker hereby seems to suffice per play se-
quence. The rest of the explicit performatives appearing in such contexts rely on
the recognizability of the activated institutional script or constitute short prefabri-
cated phrases that originate in other scripts of (semi-)public rituals and ceremonies
that are assumed to be familiar to users. It can be inferred from users’ engagement
in the play mode that when successful, mock performatives enhance the socially
gratifying nature of the interaction. The activated scripts are not brand new, but
they are jointly and playfully put to novel use in the particular computer-mediated
discourse. Because mock performatives function as discourse transformers, the
dispreferred response to them is textual silence.
The mock performatives under investigation mostly involve authority extend-
ing over (an)other interlocutor(s) in the virtual community, e.g., I hereby sentence
you to …; arrest you for …; appoint you to … In fact, they constitute one of the
principal ways for participants in these forums to assume authority over others,
mitigated by the playfulness of the exercise. Mock performatives are overwhel-
mingly enacted in response to other interlocutors’ contributions. However, owing
to their playful character, they are also sometimes used as a mitigating tool for
attemps to close the discussion. The discourse marker ok preceding the emphatic
explicit performative (I, User5, hereby do declare) in example (7) activates antici-
pations of a conclusive solution to the problems voiced by User3, and the closing
play sequence culminates the discussion.
Users also occasionally assume authority over themselves, as in: I hereby an-
nounce that I’m giving up on cream eyeshadows for good!! =p r/o. Such instances
involve playful vows, bans, declarations, announcements, and promises, and they
appear to relate to real world phenomena. Some refer to the common activities of
the discussion board, such as reviewing products; e.g., ah – yes, I hereby promise
that I will review any and all flavors I try, in response to other users demanding
“flavor recs” of toothpastes. Performatives exercising authority over oneself serve
topic management also by appearing as initiators of new exchanges, as is the case
in the above announcement. They invite endorsement or rejection but not se-
quences of joint play.
Performatives used frequently in replies also appear in abbreviated forms: ita
(‘I totally agree’); 2nd (‘I second the motion’). Such conventionalization takes
place within and across communities, and discussion boards may provide (novice)
users with lists of abbreviations. Agreeing and seconding something are common
discourse activities in CMC (see, e.g., I second this comment in example (8) in
Danet, this volume) and as such are prone to undergo abbreviation through in-
itialism, letter-number combining, and other shorthand techniques. These per-
formatives are an integral part of the informal interaction; they do not trigger play
sequences or incorporate the conspicuous formal marker hereby.
Mock performatives are sometimes accompanied by stand-alone third-person
performative predications, mimicking emotes. These may help users create appro-
priate felicity conditions in a virtual world, as in I hereby banish you from this king-
dom!!!!!!!!!!! *waves septer*, a disagreeing response typed in by an apparently fe-
male user adopting the role of a king waving his sceptre. They may also signal the
empathetic nature of the performative act, as in **hug** you are hereby given a
pass to do nothing today. Further, stand-alone performative predications immedi-
ately following another user’s mock performative may make explicit its virtual ef-
fect through a display of action that conveys a particular mental state; e.g., <– runs
to bathroom to look at my stunning new self!!! in response to the mock per-
formative i hereby delcare you an honorary indian [sic]; *laughs like a hyena* in
response to a controversial declaration; and *places hand on heart* after a pledge
to honour a metaphorical flag. Other performative predications appear as part of a
reply, before or after something that the user is saying, to signal the sincerity and/or
intensity of the virtual emotion or attitude expressed, e.g., *wipes away a tear*;
*rubs hands together*.
Both the CMC-specific third-person actions discussed in section 4 and the pre-
dominantly first-person mock performatives under consideration here contribute to
the construction of virtual realities. Unlike the third-person actions, however,
mock performatives may also appear in informal offline discourse, playfully con-
tributing to the construction of interlocutors’ textual worlds by evoking scripts of a
different kind, even though the activated scripts might not be sustained in quite the
same way. Mock performatives are thus not part of the “garden-variety virtual per-
formatives”3 grown in the emote-mode garden and subsequently spread into the
(message) field while still retaining their garden variety shape (i.e., the third-per-
son orientation and typographic cues of performative predications). Instead, mock
performatives are rather like exotic flowers planted in the natural habitat of infor-
mal discussion, conspicuous because of their origins in the knotted gardens of the
discourse of institutions, yet fully etiolated in comparison to these.
Overall, the mock performatives in the forums contribute to bonding within the
virtual community, mitigation of face threats, identity projection, and the construc-
tion of virtual realities, all of which constitute central aspects of the pragmatics of
CMC. Users readily engage in play when participating in an activated performative
script, contributing to the construction of its virtual felicity conditions, happily
shifting scripts, and/or responding in various other ways to mock performatives.
Like the rest of the interaction included in the threads (which also appears to be so-
cially gratifying), playful episodes display feelings and attitudes of users’ online
personae. Unlike the rest, however, the mediated mock performatives allow for a
shared sense of unpredictability concerning the effects of play on the virtual reality
at hand. The negotiation of this reality through interlocutors’ responses, same-
script or other-script oriented, serves to join them in what Beeman (2010: 121)
calls the performance loop, joint participation in “reinforced feedback”, whereby
“the actions of the performer are … affected and changed”. Beeman argues that
there are neuroscientific grounds for the enhancement of intersubjective emotion
and connectedness through performance.
Users are creative, in a restricted sense of the term, in their choice of, and re-
sponses to, explicit mock performatives, while limiting the innovation so that the
institutionalized contexts of the activated performative scripts remain evident. In
so doing they manifest and assume of their interlocutors a high degree of meta-
pragmatic awareness. Bauman and Briggs (1990), in their discussion of the central
elements of reflexivity involved in discourse as performance, focus on the pro-
cesses of recontextualization which serve to detach discourse from context and an-
chor it to another context of use. Mock performatives succeed in their function as
transformers into play mode largely because “discourse may be fashioned for ease
of detachment from situational context” (Bauman and Briggs 1990: 74). As soon as
the intertextuality of the stereotyped script is considered apparent, users proceed to
a stage of de-authentication of the performative through style-shifting (from for-
mal style back to the default informal style of the interactions), smiley faces,
laughter, and other signals of playfulness.
Mock performatives thus demand authentication work that allows interlocutors
to recognize the mediated institutionalized stereotype, while also signaling detach-
ment from such a context and hence freeing the users from the full implications of
the official or institutionalized discourse. This, in turn, constitutes another avenue
to authenticity in discourse, that of creating computer-mediated discourse that is
situated, unique, personal, and in some sense “genuine” for the interlocutors (Gill
2008, this volume, personal communication). De-authentication can thus be a type
of authentication of the discourse. The outcome of the (de)authentication of mock
performatives in the forums is reminiscent of Coupland’s (2001) analysis of the
6. Conclusion
Unlike performance, in its wide sense of social action performed through dis-
course, performativity in CMC has attracted few studies. Yet, investigations into
the use of performatives in online discourse can help performative theory develop
into a more full-fledged account of human interaction involving an interplay of tex-
tual worlds of different kinds, virtual realities, and real life. CMC provides com-
pelling evidence that performativity is present in the online discourse of many
people. In addition to traditional performatives that have real-world force, such as
the apologies studied by Harrison and Allton (this volume) in email, CMC is teem-
ing with performatives that are routinely used in playful, etiolated ways.
In this chapter performativity in CMC was approached from the perspectives
of two different phenomena that manifest etiolation: (i) the garden-variety vir-
tual performatives of third-person emoting, now established and highly popular in
several modes of CMC; and (ii) mock performatives in discussion-board interac-
tion, de-authenticated from (offline) institutional contexts, to function as conspicu-
ous discourse transformers to a play mode. Emoting has developed from pre-pro-
grammed commands into stand-alone performative predications that users type
directly into their messages, keeping intact the third-person orientation and taking
care to separate them typographically from the rest. On the discussion boards in-
vestigated here, mock performatives manifesting the formal marker hereby were
found to trigger sequences of responses, keeping to the activated script or switch-
ing to other scripts deemed appropriate in the context. Replies and comments in-
vited by virtual performatives testify to their conventional illocutionary force, as
well as the creation of a range of situated, non-conventional effects that can be la-
belled perlocutionary.
CMC of many different kinds manifests virtual performatives, which await
close study that would ideally have as a goal to contribute to an understanding of
the fundamental role of etiolated performatives in social interaction and the con-
struction of virtual realities. In particular, the established nature of virtual third-
person action through emoting warrants further study, focusing on its pragmatic
functions within and across CMC modes. Future investigations of virtual per-
formatives might also focus on the packages of stereotyped ideas tacitly forwarded
through emoting and the use of mock performatives. Despite users’ efforts to de-
authenticate institutional mock performatives for the purposes of particular com-
Notes
1. Performance and performativity are terms with many uses, which have roots in very dif-
ferent research traditions; witness, for instance, the discussions of (one of) these notions
in Bauman and Briggs (1990), Pennycook (2004), and Robinson (2006). For performance
studies, see, e.g., Bial (2007).
2. I am grateful to Loukia Lindholm for collecting the data for this study.
3. Susan Herring (personal communication) has characterized emotes and performative
predications as the “garden-variety virtual performatives”.
4. I am grateful to Martin Gill for discussions of authenticity and the reference to Coup-
land’s work.
References
Attardo, Salvatore
1994 Linguistic Theories of Humor. Berlin/New York: Walter de Gruyter.
Austin, John L.
1962 How to Do Things with Words. Cambridge, MA: Harvard University Press.
Bauman, Richard and Charles L. Briggs
1990 Poetics and performance as critical perspectives on language and social life.
Annual Review of Anthropology 19: 59–88.
Beeman, William O.
2010 Performance pragmatics, neuroscience and evolution. Pragmatics and Society
1(1): 118–137.
Bial, Henry (ed.)
2007 The Performance Studies Reader, 2nd Edition (first published 2004). London:
Routledge.
Blum-Kulka, Shoshana, Juliane House, and Gabriele Kasper (eds.)
1989 Cross-Cultural Pragmatics: Requests and Apologies. Norwood, NJ: Ablex.
Cherny, Lynn
1994 Gender differences in text-based virtual reality. In: Mary Bucholtz, Anita C.
Liang, Laurel Sutton, and Caitlin Hines (eds.), Cultural Performances: Pro-
ceedings of the Third Berkeley Women and Language Conference, 102–115.
Berkeley, CA: Berkeley Women and Language Group.
Cherny, Lynn
1995 The modal complexity of speech events in a social MUD. Electronic Journal of
Communication 5. http://www.cios.org/EJCPUBLIC/005/4/00546.HTML
Coupland, Nikolas
2001 Dialect stylization in radio talk. Language in Society 30(3): 345–375.
Danet, Brenda
2001 Cyberpl@y: Communicating Online. Oxford, UK: Berg.
Dresner, Eli and Susan C. Herring
2010 Functions of the non-verbal in CMC: Emoticons and illocutionary force. Com-
munication Theory 20: 249–268.
Du Bois, John W.
2007 The stance triangle. In: Robert Englebretson (ed.), Stancetaking in Discourse:
Subjectivity, Evaluation, Interaction, 139–182. Amsterdam/Philadelphia: John
Benjamins.
Enkvist, Nils Erik
1989 Connexity, interpretability, universes of discourse, and text worlds. In: Sture
Allén (ed.), Possible Worlds in Humanities, Arts and Sciences, 162–186. Ber-
lin: de Gruyter.
Gill, Martin
2008 Authenticity. In: Jef Verschueren and Jan-Ola Östman (eds.), Handbook of
Pragmatics Online. Amsterdam/Philadelphia: John Benjamins.
http://www.benjamins.com/online/hop [requires subscription]
Goddard, Angela
2003 “Is there anybody out there?” Creative language play and “literariness” in
Internet Relay Chat (IRC). In: Angela Schorr, William Campbell, and Michael
Schenk (eds.), Communication Research and Media Science in Europe,
325–343. Berlin: de Gruyter.
Goffman, Erving
1974 Frame Analysis: An Essay on the Organization of Experience. New York:
Harper & Row.
Grice, H. Paul
1975 Logic and conversation. In: Peter Cole and Jerry L. Morgan (eds.), Syntax and
Semantics 3: Speech Acts, 41–58. New York: Academic Press.
Gumperz, John J.
1982 Discourse Strategies. Cambridge, UK: Cambridge University Press.
Herring, Susan C.
2000 “Doing gender” on the Internet. Guest lecture at Växjö University, Sweden,
10 April, 2000.
Herring, Susan C.
2001 Computer-mediated discourse. In: Deborah Schiffrin, Deborah Tannen, and
Heidi Hamilton (eds.), The Handbook of Discourse Analysis, 612–634. Ox-
ford, UK: Blackwell.
Herring, Susan C.
2003 Gender and power in on-line communication. In: Janet Holmes and Miriam
Meyerhoff (eds.), The Handbook of Language and Gender, 222–228. Oxford,
UK: Blackwell.
Herring, Susan C.
2004 Computer-mediated discourse analysis: An approach to researching online be-
havior. In: Sasha A. Barab, Rob Kling, and James H. Gray (eds.), Designing for
Virtual Communities in the Service of Learning, 338–376. Cambridge, MA:
Cambridge University Press.
Herring, Susan C.
In press Grammar and electronic communication. In: Carol A. Chapelle (ed.), The
Encyclopedia of Applied Linguistics. Malden, MA/Oxford/Chichester: Wiley-
Blackwell.
Herring, Susan C., Lois Ann Scheidt, Sabrina Bonus, and Elijah Wright
2005 Weblogs as a bridging genre. Information Technology & People 18(2): 142–171.
Kolko, Beth
1995 Building a world with words: The narrative reality of virtual communities.
Works and Days 25/26, 13(1/2).
http://bethkolko.com/includes/pdfs/buildingaworldwithwords.pdf
Kurzon, Dennis
1986 It Is Hereby Performed … Explorations in Legal Speech Acts. Amsterdam/
Philadephia: John Benjamins.
Lakoff, Robin Tolmach
2000 The Language War. Berkeley, CA: University of California Press.
Langellier, Kristin M.
2003 Personal narrative, performance, performativity: Two or three things I know
for sure. In: Yvonna S. Lincoln and Norman K. Denzin (eds.), Turning Points
in Qualitative Research, 441–468. Walnut Creek, CA: AltaMira. Reprinted
from Text and Performance Quarterly 19 (1999): 125–144.
Parker, Andrew and Eve Kosofsky Sedgwick
2007 Introduction to Performativity and Performance. In: Henry Bial (ed.), 200–207.
Pennycook, Alastair
2004 Performativity and language studies. Critical Inquiry in Language Studies 1:
1–19.
Peterson, Eric E.
2011 How conversational are weblogs? Language@Internet 8, article 8.
http://www.languageatinternet.org/articles/2011/Peterson
Peterson, Eric E. and Kristin M. Langellier
2006 The performance turn in narrative studies. Narrative Inquiry 16(1): 173–
180.
Raskin, Victor
1985 Semantic Mechanisms of Humor. Dordrecht: Kluwer.
Robinson, Douglas
2006 Introducing Performative Pragmatics. London: Routledge.
Ruedenberg, Lucia, Brenda Danet, and Yehudit Rosenbaum-Tamari
1995 Virtual virtuosos: Play and performance at the computer keyboard. Electronic
Journal of Communication 5.
http://www.cios.org/EJCPUBLIC/005/4/00542.HTML
Sbisà, Marina
2009 Speech act theory. In: Jef Verschueren and Jan-Ola Östman (eds.), Key Notions
for Pragmatics, 229–244. Amsterdam/Philadelphia: John Benjamins.
Sbisà, Marina
2007 How to read Austin. Pragmatics 17: 461–473.
Schank, Roger C. and Robert P. Abelson
1977 Scripts, Plans, Goals and Understanding: An Inquiry into Human Knowledge
Structures. Hillsdale, NJ: Lawrence Erlbaum.
Schlobinski, Peter
2001 *knuddel zurueckknuddel dich ganzdollknuddel*: Inflektive und Inflektiv-
konstruktionen im Deutschen. Zeitschrift für Germanistische Linguistik 29(2):
192–218.
Searle, John R.
1969 Speech Acts: An Essay in the Philosophy of Language. Cambridge, UK: Cam-
bridge University Press.
Searle, John R.
1975 Indirect speech acts. In: Peter Cole and Jerry L. Morgan (eds.), Syntax and
Semantics 3: Speech Acts, 50–82. New York: Academic Press.
Searle, John R.
2001 How performatives work. In: Daniel Vanderveken and Susumu Kubo (eds.),
Essays in Speech Act Theory, 85–107. Amsterdam/Philadelphia: John Benja-
mins. Reprinted from Linguistics and Philosophy 12 (1989): 535–558.
Virtanen, Tuija
In press Mock-performatives in online discussion boards: Toward a discourse-prag-
matic model of computer-mediated communication. In: Deborah Tannen and
Anna Marie Trester (eds.), Discourse 2.0: Language and New Media. Selected
Papers from GURT 2011. Washington, DC: Georgetown University Press.
Werry, Christopher C.
1996 Linguistic and interactional features of Internet Relay Chat. In: Susan C. Her-
ring (ed.), Computer-Mediated Communication: Linguistic, Social, and Cross-
Cultural Perspectives, 47–63. Amsterdam/Philadelphia: John Benjamins.
1. Introduction
Address forms are an indispensable part of the communicative process. They can
serve as indexicals of personal or social identity, highlighting the differences be-
tween interlocutors or stressing their common status or personal qualities. Address
forms can also index features of an unfolding communicative situation, for example
by supporting an interlocutor’s conversational goals through face-threatening
or face-enhancing acts, expressing changes in mood, or addressing an interlocutor
by a different form in the presence of third parties to disguise the closeness of the
relationship or maintain the addressee’s public face. The ability to choose appro-
priate forms of address in a wide variety of communicative contexts and with dif-
ferent interlocutors is part of an individual’s pragmatic competence, whether in
face-to-face or computer-mediated interactions.
The analysis of address form use has not been exclusively, or even primarily,
carried out within the field of pragmatics, however. Address forms shed light on
so many aspects of the communicative process that they have attracted the interest
of researchers from various disciplines within the social and cognitive sciences
(e.g., linguistic pragmatics, sociolinguistics, conversation analysis, sociology of
language, linguistic anthropology, social psychology, cognition). Early on, Roger
Brown and Albert Gilman, who are credited with initiating the systematic study of
address in 1960, noted that “[s]emantic and stylistic analysis of these forms takes
us well into psychology and sociology as well as into linguistics and the study of
literature” (1960: 253).
While every language has pronouns as well as noun phrases (nominal forms)
that an interlocutor may use to identify an intended addressee, great variation exists
among world languages in both the pronominal system and the ways in which nom-
inal forms can be used as forms of address. Standard English has a single pronomi-
nal form you, which may be used in the singular or plural, while French, Danish,
and German have two pronouns in the singular: tu/vous, du/De, and du/Sie, respect-
ively, the second pronoun in each case being a grammatical plural form. Spanish
is more complicated, as there is no single national standard among the countries in
which Spanish is the official language. Portuguese has three forms (tu/você/vós),
the grammatical second-person singular, third-person singular, and second-person
plural pronouns, although the use of vós with a single addressee is becoming
increasingly rare and regionally restricted. At the other end of the spectrum are lan-
guages with honorifics, such as Japanese and Malay, in which the address form sys-
tem is one aspect of a speech “level”, with degree of politeness or formality built
into the grammatical system in a variety of ways.
The syntactic rules governing the use of nominal forms in direct address also
vary from language to language. While in English it is possible to use nominal
forms in direct address as if they were bound pronouns, as in Would Madam [you]
like some tea?, use of such constructions tends to be stylized and limited. In
Portuguese, as in languages with honorifics systems, the equivalent question,
A senhora aceita um chá?, is very common, as are more elaborate forms of nominal
address, such as O senhor professor já deliberou sobre o assunto? (‘Has Mr. Pro-
fessor [have you] thought about the issue?’). In recognition of the fact that in many
languages nominal forms can be grammatically bound, as if they were pronouns of
address, the focus of this chapter and its operational definition of “address forms”
includes both pronominal and nominal forms used in direct address.
Computer-mediated communication (CMC) offers a different investigative mi-
lieu in which to study interaction than does face-to-face communication, as both
the researcher and the interaction itself are place-independent and, in asynchron-
ous CMC, time-independent, as well. A researcher can gain access to enormous
quantities of data that are neither “contaminated” by her presence (cf. the Observ-
er’s Paradox, Labov 1970: 32) nor dependent on it. There is a natural linearity to
the data, and while it may be impossible to know with complete certainly the type
and amount of interaction that occurs between specific interlocutors offline or in
another CMC channel, it is possible to capture the entirety of the interaction in the
channel being studied. Despite these factors and the amount of data that is avail-
able for analysis, the study of address in CMC is still in its infancy.
This chapter provides a historical and theoretical overview of the study of
address, with a focus on contributions from the pragmatics and sociolinguistics
research traditions, which account for the majority of studies. A discussion follows
of existing studies of address usage in CMC. The studies presented fall into three
main categories: greetings, considerations of conversational norms, and studies of
address in educational settings. This discussion is followed by a section that raises
some theoretical and methodological issues associated with CMC-based studies of
address. In recognition of the incipient nature of such research, the chapter con-
cludes with suggestions for future research.
Brown and Gilman’s (1960) model of address opened a new domain of investi-
gation. Their model was founded on the premise that social relationships are redu-
cible to two dimensions, broadly defined: the degree to which one person has
“power over another” in age, socioeconomic status, rank in an organization, edu-
cational level, etc. (1960: 253); and solidarity, as “a set of relations that are sym-
metrical” (1960: 258). Each person was presumed to occupy a particular place on a
hierarchical scale, and an assumption was made that all would be addressed “ap-
propriately” for their status. Brown and Gilman focused on languages with a bi-
nary pronominal system for addressing an individual addressee (e.g., French,
Spanish, German), referring to T and V, so labeled after French tu/vous. T, the pro-
noun used with the second-person singular form of the verb, indexed, according to
their model, not only solidarity per se but intimacy, trust, informality, equal or
lesser rank, similar interests, and so forth. V, indexing power, authority, respect, de-
ference, and higher rank, was so labeled after the second-person plural form in
French, with the recognition that in some other languages the V pronoun is conju-
gated with another form of the verb (e.g., third-person singular, as in Spanish).
Although the study of address is now often considered a sub-field of prag-
matics, and Brown and Gilman’s work was contemporaneous with Austin’s pion-
eering publication on speech acts (1962), most address studies following the pub-
lication of Brown and Gilman (1960) were conceived within the then-emerging
field of sociolinguistics. In fact, studies of address were instrumental in demon-
strating a crucial point of sociolinguistics: that the social context and the social
roles of the speakers can influence language use. Despite the interest in early prag-
matics research in examining how speaker intentions are realized in conversation,
such studies were based on “model” rather than actual speakers (e.g., Lakoff 1973;
Leech 1983). A partial convergence of the two fields occurred with the presenta-
tion of Penelope Brown and Stephen C. Levinson’s work on politeness (1978,
1987), in which power and solidarity, along with social distance, were integrated
into a strategic model of politeness (see Duthler [2006] for a brief description of
Brown and Levinson’s model and its usefulness for CMC studies of politeness; see
Vinagre [2008] for its application to CMC in Spanish). Later, noting the insuffi-
cient treatment of impoliteness in Brown and Levinson’s model, Culpeper (e.g.,
1996) adapted their approach to analyze impoliteness. Perhaps in part a reaction to
Brown and Levinson’s assertion that their strategies of politeness were universal,
studies based on real, as opposed to imagined, language use began to be conducted
around the globe (see, e.g., Arnaiz [2006] for an overview of the application of dif-
ferent theories of politeness to the study of Spanish address forms). Still, despite
some shared concerns (respect and deference, among others), politeness and ad-
dress remain independent systems at the theoretical level.
The central focus of Brown and Levinson’s model is politeness, rather than ad-
dress; the authors only acknowledge the importance of address as an indicator of
in-group status or to show deference (1978: 57, 107). Accordingly, address re-
searchers accustomed to examining natural, as opposed to idealized, conversation
have noted that the role address plays in (im)politeness is underanalyzed in this
model. While more recent studies (e.g., Terkourafi 2004) have recognized that the
choice of address form contributes more to communication than the pioneers in
politeness research envisioned, other scholars continue to view address forms as
markers of rather static identity, ignoring individual speakers’ ability to use the
forms creatively, thereby expanding the forms’ semantic range. In either case, for
politeness theorists, address forms are but a single element in a whole array of fea-
tures (e.g., hedging, degree of imposition, degree of directness) that can define an
interaction as (im)polite.
More recently, the focus on identity as a construction emerging from discourse
has brought pragmatics and sociolinguistic endeavors to another junction (see
Bucholtz and Hall 2005 for an introduction to theoretical perspectives on identity).
Until the 1950s, the view of social scientists was that social identity was stable
and essential in nature. Constructivist views of social class and identity began to
take hold in sociology in the 1950s and 1960s, although in the study of language
variation (sociolinguistics and pragmatics), the transition has been of more recent
vintage. In the mid-1980s, Myers-Scotton (1983) and Medeiros (1985) discussed
the “negotiation” of identity through strategies of address, the former within a
pragmatics framework, the latter within sociolinguistics. However, only after Eck-
ert and McConnell-Ginet’s (1992) application of the communities of practice
(CoP) approach to the study of gender identity did constructivism become main-
stream in those fields (see Wenger 1998 for a general discussion of the CoP ap-
proach). In constructivism identity is not innate; rather, it is constructed “through
social, economic, cultural, historical and political factors”, and thus its interpre-
tation is fluid (D. Smith 2001: 34). In this view, address forms are not merely iden-
tity “tags” or “tokens” whose interpretation is fixed; interlocutors can use them in
ways that do not follow societal norms. Smith (2001) also asserts that essentialists
reduce social behavior to one or two dimensions, while constructivists see a
“multiplicity” of social dimensions. Both the Brown and Gilman and Brown and
Levinson models belong in the essentialist paradigm. In these models, address
forms are viewed as markers of stable identity features (e.g., age, social class,
gender). Many elements are subsumed into the dimensions of power and solidarity
under the premise that the aspects, or factors, associated with power form a natural
group, as do those associated with solidarity. However, while in each case the fac-
tors grouped in each dimension share commonalities, they are not homogeneous.
The paradigm shift to constructivism triggered a major change in the field of
pragmatics. There has been a shift in focus from speaker-focused strategies, a hall-
mark of the Brown and Levinson model, towards “relational work” (e.g., Locher
and Watts 2005), with an emphasis on the collaborative and unfolding process
of communication. For example, Watts’ (2003) discursive model of politeness, de-
scribed below, does not predict specific use, but purports to illustrate where utter-
ances may be open to interpretation.
The need to be able to assess appropriateness in address form use has been
acknowledged in the address form literature for more than three decades. Hymes
(1974), operating within the essentialist paradigm, saw a relationship between the
form and function of each address form:
When the values of the mode of address and the social context match – when both, say,
are formal – then that meaning is of course accomplished, together with the meeting
of expectations. When the values do not match … then a special, or ‘marked’, meaning
is conveyed. (1974: 111)
In my earlier discussions of the negotiation of address form relationships (Medei-
ros 1985; Oliveira 2005), I defined two interactional “planes” – one for conven-
tionalized usage (usage matching societal expectations) and one on which the
speakers determine the address forms appropriate for their relationship, indepen-
dent of outside opinion (negotiated usage), with (non)conformity possible on
either plane. In pragmatics there has been a similar recognition that the interpre-
tation of usage must be differentiated. Watts (e.g., 2003) makes a distinction
between “first-order” and “second-order” politeness. Silverstein (e.g., 2003) refers
to levels of indexicality; Blum-Kulka (1992) and Gal (2008) make a distinction be-
tween public and private use; and Mills (2009) distinguishes between cultural and
individual levels. The latter closely parallels the distinction between conventional-
ized and negotiated in Medeiros (1985) and Oliveira (2005).
More recently, I have been refining a model of address that integrates cogni-
tive, interactional, discursive, socio-psychological, and negotiative/constructivist
elements into a single framework (Oliveira 2006, 2010). The model is “discur-
sive” in that it focuses on the changing address form relationship as it unfolds
through discourse over time. In Stage 1 of the model, S (speaker) and H (hearer)
meet as strangers; their goal is to assess what is a socially-appropriate address
form and use it. In Stage 2, S (or S and H) wishes to signal a change in the rela-
tionship; a new form is negotiated that may or may not conform to societal expec-
tations. It is the interlocutors’ own assessment that governs the choice and the de-
termination of appropriateness; societal expectations are moot. Temporary
changes in use of address form may also occur, brought about by a number of fac-
tors (humor, desire to joke, the presence of certain other people, a situation invol-
ving protocol, etc.). If the temporary change is provoked by the presence of others
(e.g., to disguise a close relationship), then the likely choice of address is a form
that intentionally conforms to societal expectations and can be interpreted accord-
ingly. If the change is due to another reason, then the relevant interpretation is
that of S and H, rather than third parties. A Stage 3 is possible, if S (or S and H)
wishes to signal a further change in the relationship. In that case, the address form
is renegotiated as in Stage 2; this stage may be repeated any number of times. (See
Oliveira 2006, 2010 for sample strategies and a graphical representation of parts
of the model.)
In this discursive model of address, the examination of negotiated use requires
the analysis of messages sent over a period of time so that patterns can be ident-
ified, making possible the identification of usage that is anomalous (marked) for
the interlocutors. At the heart of the model is the notion that there are two spheres
of interaction, one whose notions of markedness are tied to socially conventional-
ized norms of politeness, the other determined by the relationship that speakers
build for and between themselves over time. This negotiated relationship may, but
need not, conform to societal expectations, and the evaluation of outsiders may be
deemed irrelevant by the pair.
Similarly, Watts (2003) envisions two types of politeness analysis: First-order
politeness (politeness1) involves the evaluation of politeness by laypersons, while
second-order politeness (politeness2) refers to the theoretical notion of politeness
used by researchers that “preserve[s] mutually shared consideration for others”
(2003: 277). For Watts, politeness1 behavior can be considered to be “polite” or
“politic”. “Politeness” emerges through practices that “draw on members’ past ex-
periences of previous interaction and the institutionalised, objectified structures of
social reality that have been internalised in the process” (2003: 201), roughly
equivalent to my use of “conventionalized usage”. Politic behavior is defined as
that which “the participants construct as being appropriate to the ongoing social
interaction”, a clear parallel to my concept of “negotiated usage”. Watts’ model
seeks to explain how speakers come to develop behavior “appropriate to the social
interaction type in which [they] are involved” (2003: 143), reflecting a process of
socialization. However, Watts is more concerned with evaluation by others, and
my focus is on how individuals come to find forms that are mutually “satisfying” (a
term my informants have used, Medeiros 1985).
CMC interactions, by their linear nature, lend themselves well to the study of
how social relations are negotiated over time, and the distinction between conven-
tionalized and negotiated usage is relevant online. More CMC-based studies of ad-
dress focus on the former than the latter, however, as is evident from the review of
literature in the following section.
Most studies of address in CMC seem to fall into one of three categories: the role
address plays in greeting behavior and relationship development; considerations of
conversational norms, including the reactions of Internet users to address form
usage that they find objectionable (i.e., that violates conventionalized norms); and
address in collaborative learning projects or in educational institutions generally,
including comparisons between online and offline behavior. Each of these cat-
egories is discussed below.
the speakers simply acknowledge the other’s presence (see Medeiros 1994 for
further discussion). As both greetings and address forms serve to set the conversa-
tional stage (e.g., degree of formality with intended addressee, mood) for the com-
munication, address forms are often included in greetings. Ervin-Tripp (1972)
introduced the term “co-occurrence” to account for linguistic elements of different
types that work in tandem to produce stylistic coherence; one example she pro-
vides joins a greeting and a nominal address form. By her definition, Good morn-
ing, Your Eminence? would be a case of stylistic coherence, while How’s it going,
Your Eminence? violates the rules of co-occurrence (1972: 233). Although address
forms and greetings often co-occur, at least in some languages, both need not be
explicitly present. The degree to which address forms are expected to be spread
throughout a written or oral communication is culturally determined, with honor-
ific languages placing greater importance on the use of explicit nominal forms than
English or the Scandinavian languages. Thus, the omission of an explicit form in a
series of messages in English may escape notice, while in another language a
single message omitting “appropriate” nominal forms of address may be cause for
comment (e.g., Oliveira 2003).
One of the difficulties in analyzing co-occurrences of greetings and address
online is that while the inclusion of an address form in a greeting may signal the
identity of the addressee, the mood of the “speaker”, or something about the situ-
ation (e.g., Waldvogel 2007: 458), the omission of a nominal form (or an explicit
pronoun, in pro-drop languages) may not be significant. The relationship may have
been negotiated offline or in a previous message, with the users finding no further
need to call attention to it; or continual usage of the form may appear to overem-
phasize its significance, actually serving to distort the communication. Finally, it
may simply be that because the CMC format itself provides the name or nick of the
other, the user feels no further need for identification of the parties involved. Her-
ring (1996) invokes the latter explanation to account for why only 13 % of the
messages posted to two asynchronous academic discussion lists she analyzed
started with salutations, most of which took the form “To: [Name]”. She writes:
The relative lack of epistolary conventions can be explained in part by the fact that
a header is added automatically to each message by the electronic mailer, including a
separate line for who the message is “from”, who it is addressed “to” (in these data, the
entire mailing list), and the date and time of posting (87).
made explicit use of the addressee’s name, whereas the 63 % of university emails
with a greeting included 37 % that used the addressee’s name (Gains 1999: 85, 91).
Cho’s (2010) analysis of 197 private email messages exchanged among staff at
an Australian university provides further data from a university setting. Cho found
that names (first, surnames, or nicknames) were sometimes omitted from both greet-
ings and leave-takings, although they were present in approximately two-thirds of
the emails. Thus there is institutional variability even within a single language in the
(non-)use of address forms in (or as) greetings. More generally, these findings pro-
vide evidence that nominal address forms serve functions beyond the simple identi-
fication of speakers, as that task is handled through the email program itself.
Hatipoğlu (2007) compared British and Turkish online calls for academic
papers and found evidence of culturally differentiated usage. While British confer-
ence organizers omitted salutations in 95.4 % of their messages, nearly half
(43.6 %) of the Turkish calls for papers included salutations with address forms
(2007: 766). Hatipoğlu further analyzed the data according to destination – mes-
sages sent to the Linguist List vs. those sent to potential conference attendees.
Little difference was found in the use of postings of international calls for papers
made to the Linguist List, but a significant contrast was found in the second group:
61.3 % of the Turkish messages contained salutations with address forms, as com-
pared to only 8.1 % of the British messages, leading Hatipoğlu to suggest that the
intended receiver of the message is a determining factor in the decision to include
or omit salutations by Turkish speakers (2007: 767).
Waldvogel (2007) also points out differences in the extent to which people in
different countries explicitly acknowledge differences in status. She considered
a group of studies contrasting emails sent to superiors and subordinates vis-à-vis
the inclusion or omission of closings which may, optionally, include an address
form, and noted differences between New Zealand and Malaysia, the latter being
a country where status differences are more often acknowledged. Waldvogel con-
cluded from her data that interpersonal dynamics and cultures are sufficiently dif-
ferent to warrant the design of studies clearly focused on a particular situation
(Waldvogel 2007: 457).
Most of the studies surveyed above focused on asynchronous CMC, especially
email. Rintel, Mulholland, and Pittam (2001) examined greetings and salutations in
(synchronous) Internet Relay Chat (IRC) in English; part of their analysis focused
on “nicks” (the screen names that the users choose to identify themselves, which
can be altered mid-conversation), which were found to be used often as address
forms. Their qualitative study analyzed the different forms nicks could take (plays
on words, altered syllable order, use of punctuation marks, capital letters, etc.).
However, the data presented only included the greetings themselves; as the sur-
rounding context was eliminated, the way these variations played out in interaction
over time was difficult to discern. (For a discussion of nicknames, see also Lind-
holm, this volume.)
ultaneously conveyed subtle gradations of respect and closeness, but they could
alter the mix at any moment. Nishimura further found that the degree of involve-
ment (number of messages authored) as well as length of time of participation on
the BBS affected the mix of the various formality signals.
Evidence of conversational norms as regards address form usage can be in-
ferred from violations of those norms. Kretzenbacher (2005), Oliveira (2003), and
Graham (2007) found that Internet users sometimes react strongly to violations of
address norms, generally the use of T where V is expected, or if several V-forms
exist, use of one denoting greater familiarity than expected. Kretzenbacher (2005:
7) noted that German netiquette often proposes reciprocal address, rather than
stipulating a particular form of address, and recommended that users not be “as-
tonished” by the use of one form or another. Kretzenbacher took as his starting
point the popular notion that “everyone” (i.e., speakers of German) used du on the
Internet. However, in comparing the German-language websites of the IKEA com-
pany in Germany, Switzerland, and Austria, he found significant differences in the
use of the second-person pronouns, thus refuting that notion.
Oliveira (2003) found that on a Portuguese university’s users network, men
posted more messages denouncing a lack of adherence to politeness and grammati-
cal norms than women did. In a conversational thread that was particularly acerbic,
the initial message was a statement by a male professor (Male 1) decrying the uni-
versity newspaper’s disparaging coverage of one of the university’s departments.
An exchange ensued in which address forms played a key role. A woman (Female
1) responded to his message, addressing him as Sr. FN + LN [Mr. First Name +
Last Name]. In Portugal, bachelor’s degree-holders are addressed as Dr., master-
level graduates as Mestre, and Ph.D. graduates as Doutor, preceded by Professor
when they teach at a university. This woman’s omission of academic titles con-
veyed the notion that Male 1 had no university degree at all.
Male 1 responded, referring to her as Colleague (I presume you are one), indi-
cating that he had not bothered to find out her status, and then made an oblique ref-
erence to her not having used the appropriate form: Without any attempt
to diminish you by the address form you chose to address me by …. The female col-
league did not respond, but a “bystander” to the conversation, Male 2, made ex-
plicit what Male 1 conveyed implicitly:
No. ‘Sr. + FN’ [Mr. + First Name] is impossible. It won’t do. The Academic world is
vast, diverse, open and has its fragility: the respect of a certain elegance is not imposed
by law. It’s just as well, but it imposes greater moral duty.
This was followed by a message from a second bystander, Male 3, who joined
Female 1 in denigrating Male 1, whose academic title was Professor Doutor
[Doctor], by referring to him as a certain Dr. + FN + LN [Last Name]. In this ad-
dress, Male 1 was belittled on two fronts, first with the wording “a certain,” a sug-
gestion that Male 1 was an undesirable of some sort, and then by being verbally
stripped of his graduate degrees and status as professor. The finding that men were
serving as the arbiters of politeness ran counter to that of CMC studies in English,
which typically find women exhibiting greater concern with the observance of pol-
iteness norms (cf. Herring 1994; C. Smith, McLaughlin, and Osborne 1997).
Graham (2007) similarly found evidence of a “collaborative” social conscience
in her study of deviations from the politeness norm in a community of practice, an
online church support group. The message that sparked the controversy was a plea
for support from a church member who was having emotional problems with his
male partner. The politeness transgressor was a woman who answered his plea for
help by sending a message that was clearly unsupportive in tone as well as disobey-
ing the list’s instructions for providing informative headers in messages of re-
sponse. After other members wrote to note both her attitude and non-conformance,
she responded in an additional three messages, never once mentioning the name of
the original poster, referring to him as “that gentleman”, as if he had no access to
her messages and/or was of no consequence (2007: 750). In this case, the issue was
the omission of address, as opposed to a form that was poorly chosen.
the age of the Canadian in order to select an appropriate first address form (2005:
62 ff.). The Korean student needed the information in order to be able to determine
their relative status levels. According to the English gloss, the question was asked
three times (age, age in Korean, and year of birth). The Canadian was confused,
both because it was the first utterance from the Korean, but also because of the
multiple formulations of the “same” question. Because of this exchange, the re-
searchers decided to consider the acquisition of the honorifics system generally
and the interactional differences between the two groups at different stages in the
acquisition process. Korean students initially found the English-speakers “rude
and disrespectful”, which hampered communication generally, but the quality and
frequency of cross-cultural interaction increased as the Canadians became profi-
cient in using honorifics (2005: 63).
Belz and Kinginger (2002) set up a collaborative learning project pairing stu-
dents of French and German at an American university with counterparts learning
English in France and Germany. The French sample comprised 300 emails and the
German sample transcriptions of 25 half-hour synchronous chat sessions. Stating
that their goal was to analyze the development of use of T and V in these cross-
national conversations, the researchers inexplicably contaminated the data by tell-
ing the students to use T in their communication. Not all students obeyed, so they
proceeded to analyze use of T and V without considering that the variation they
found might have been a function of how well the students remembered their
instructions, as opposed to previous knowledge regarding the addressing of ac-
quaintances.
While CMC researchers often make references to online versus offline beha-
vior, rarely do studies examine the same population. Do people interact with the
same other people in different ways offline and online? Aarsand’s (2008) study of
Swedish students interacting both through MSN Messenger and face-to-face found
that they did. The students were divided into groups involved in different activities.
Each was connected to Messenger and was able to interact online with anyone in
the room, making it possible to maintain private, in-group conversations concur-
rently with offline discussions. The study focused on collaborative work on school
assignments, so the results cannot be directly extrapolated to other types of dual-
channel communication. Nevertheless, multitasking has become so prevalent, both
online and online/offline, and users often move so quickly from one activity to an-
other, that there may be “slippage” – confusing the addressee and inadvertently
using a form that is marked (unusual for the addressee). Little mention is made of
address forms beyond an instance of pejorative name-calling (160) and an offline
comment from one girl to another that she has a nice name (156), but there is men-
tion of students appropriating the identity tags of others, which then requires them
to identify themselves when they wish to make their own identity clear. Despite the
paucity of information regarding address, the design of the study is interesting and
merits consideration by address researchers.
were responsible for address choice. Related to this are unreflective assertions that
“everyone” uses “informal” address forms on the Internet, as participants them-
selves may claim (e.g., hier im großen internetz, wo sich alle dududuzen ‘here on
the big internets where everyone is on dududu [reciprocal du] terms’, Kretzen-
bacher 2005: 3). Such assertions are easily disproven, as Kretzenbacher himself
went on to do.
A third issue is the underutilization of discursive approaches. Within prag-
matics, discursive models have been shown to be useful to highlight the unfolding
aspect of communication (Haugh 2007; Locher 2006; Watts 2003). In CMC there
is a natural linear aspect to text-based communication, which should favor the use
of analytical techniques that focus on the unfolding nature of the communication.
While these have been employed to some extent (e.g., Garcia and Jacobs 1999;
Graham 2007), the negotiation of address terms over time is rarely the focus of
such studies. An apparent exception is Hongladarom and Hongladarom’s (2005)
analysis of exchanges within a Thai online community. Although they do not indi-
cate the overall scope of the data collected, the authors make reference to conver-
sational threads analyzed, one of which comprised 42 exchanges over four con-
secutive days (2005: 151). By examining entire conversational units, they were
able to identify marked, anomalous usage designed to defuse an argument, noting
various discursive remedies, or “face-balancing acts”, designed to “restore the
state of equilibrium in the course of heated discussions on this online forum” (155).
Types of face-balancing acts included use of joking references, irrelevant material,
and encouraging words. One example they cite of defusing a situation involves
making a joke using the addressee’s user name, “Prostate Gland”:
You know how to play thâanlûukmàak [(‘Mr. Prostate Gland’)]
If we don’t read [your discussion point] carefully, perhaps we’ll feel pain in our prostate
glands.
Hongladarom and Hongladarom (2005: 155)
In this example, joking and off-topic remarks serve as diversionary tactics to redi-
rect the attention not only of “Mr. Prostate Gland”, the principal addressee, but
of the other users as well. In addition to suggesting the value of sequential dis-
course analysis in the study of address online, this example clearly illustrates
a case where conventional notions of conversational politeness must give way to
localized interpretation.
This chapter has shown how the study of address has been shaped over the past five
decades, principally within two research traditions, sociolinguistics and prag-
matics, the latter with a particular focus on politeness research. The past few years
they are addressing. If such slippage occurs, how does it occur, and what is the
reaction of the other interlocutors? Are there any differences in the ways people ne-
gotiate the use of new forms in one channel as opposed to another? In what ways
does the CMC channel itself affect the development of address form relationships,
if it does?
In the section on greetings above, it was noted that either the greeting or the
address form might be omitted. Issues requiring further investigation include the
co-occurrence of greetings and address, the determination as to whether greetings
change over time in tandem with address, and the circumstances under which only
one of the two elements is included. For example, the absence of a nominal form in
a greeting is rarely considered a breach of polite behavior in English; a parallel
study to Cho’s (2010), analyzing data from a language in which the absence
or presence of a nominal form is a marker of politeness level, would be of interest,
as it would indicate the degree to which the “delegation” of status markers can be
left to the CMC system or must be overtly and directly handled by the individual
message sender.
The degree to which “economy of effort” (e.g., Herring 2003: 5) has an impact
on the selection or the presentation (writing, punctuation, etc.) of address forms, as
well as their interpretation by addressees, may be another fruitful area of research.
In European Portuguese, the V-pronoun você is conventionally considered impolite
by many speakers, who generally avoid its use in both oral and written communi-
cation. However, it is frequently used in SMS messages, usually in abbreviated
form (v or vc), among those one does not address by tu, because these forms are
much shorter than the other options normally used (e.g., FN, LN, TITLE + name).
Studies are needed to discover whether these and other abbreviations are used with
anyone to whom one sends a message, or whether social or professional differ-
ences generally impede the use of such economies of effort. Further study
is needed, in a variety of languages, to determine whether there is a tendency to
abbreviate or even omit address forms which are composed of title(s) plus name(s)
in fast-paced synchronous CMC, while still conforming to offline norms in asyn-
chronous CMC. Gottschalk (2010) considers whether abbreviated address forms in
German, such as SgDH for Sehr geehrte Damen und Herren (‘Dear Ladies and
Gentlemen’), are perceived as equally polite as non-abbreviated forms. As ab-
breviated forms are continuously being created in CMC channels, questions such
as these take on increasing importance.
One avenue of research gaining favor in CMC is the communities of practice
(CoP) approach (e.g., Abdelnour Nocera 2002; Murchú and Sorensen 2004; Zhang
and Watts 2008). Lave and Wenger (1991: 98), in their articulation of CoP, empha-
sized the development of relationships over time while engaging in shared activ-
ities and exchanging knowledge. Wenger (2006: 1), however, provides a more nar-
row definition, setting three conditions to establish that a CoP exists: There is a
“shared domain of interest”, a “community” of shared activities, and all members
must be “practitioners” of the field, as well as stressing that the focus is on the
learning of material. In CMC, as in offline environments, the use of address forms,
particularly nominals, in the transformation of an individual’s status from outsider
to insider is particularly interesting, as the choice of forms can index the specific
identity features that are in play. CoP research is ethnographic, emphasizing rela-
tionship management over time; as such, it is well suited to examining this function
of address.
CMC-based research can contribute to the study of address and, more broadly,
to pragmatics. As interlocutors can communicate online without sharing geo-
graphic space, many people have more frequent conversations throughout the day,
even if each conversation is of short duration, than they do in face-to-face com-
munication. Each new conversation provides an opportunity for speakers to re-
negotiate their address form relationship, hence greater frequency of contact trans-
lates into the possibility for an accelerated relation-building process. This process
is transparent to researchers as never before, because of the availability of persist-
ent computer-mediated textual records. The ability to track a relationship means
not only that shifts in address can be traced, but that usage that demonstrates a re-
sistance to change can also be viewed in its proper contextual framework. Also of
interest is the question of whether wide use of certain address forms in CMC event-
ually leads to their greater usage in offline communication.
Conversely, fine-grained analyses of pronominal and nominal address as they
occur in CMC can advance CMC research generally. How people establish and
maintain relationships through CMC, or through a combination of CMC and off-
line communication, is of central interest to CMC researchers, and solid research
on address is an important component of that research goal. Discursive approaches
are particularly well-suited to the study of relationship development and manage-
ment, making it possible for researchers to distinguish between address form
selection designed to move the relationship forward, on the one hand, or to achieve
a specific and immediate purpose (e.g., display mood, achieve a conversational
goal), on the other. Moreover, by not predetermining the degree of formality,
respect, or politeness that the address forms are presumed to convey, research
grounded in discourse can identify usage that differs from offline norms, either
in the form used or in the amount of time elapsed between initial contact and the
development of a stable address form relationship. In short, address in computer-
mediated discourse is a promising area of research deserving closer examination in
the future.
Acknowledgments
The author wishes to thank the volume editors for their close reading and insightful
comments on earlier drafts of this chapter.
References
Duthler, Kirk W.
2006 The politeness of requests made via email and voicemail: Support for the
Hyperpersonal Model. Journal of Computer-Mediated Communication 11(2),
article 6. http://jcmc.indiana.edu/vol11/issue2/duthler.html
Eckert, Penelope and Sally McConnell-Ginet
1992 Think practically and look locally: Language and gender as community-based
practice. Annual Review of Anthropology 21: 461–490.
Eelen, Gino
2001 A Critique of Politeness Theories. Manchester, UK: St. Jerome’s Press.
Eisenchlas, Susana A.
2011 On-line interactions as a resource to raise pragmatic awareness. Journal of
Pragmatics 43(1): 51–61.
Ervin-Tripp, Susan M.
1972 On sociolinguistic rules; alternation and co-occurrence. In: John J. Gumperz
and Dell Hymes (eds.), Directions in Sociolinguistics, 213–250. New York:
Holt, Rinehart, and Winston.
Gains, Jonathan
1999 Electronic mail – a new style of communication or just a new medium? An in-
vestigation into the text features of email. English for Specific Purposes 18(1):
81–101.
Gal, Susan
2008 Language ideologies compared. Journal of Linguistic Anthropology 15(1):
23–37.
García, Angela C. and Jennifer B. Jacobs
1999 The eyes of the beholder: Understanding the turn-taking system in quasi-syn-
chronous computer-mediated communication. Research on Language and
Social Interaction 32(4): 337–367.
Gottschalk, Enrico
2010 Sprachliche Höflichkeit im E-Mail-Verkehr. Munich: GRIN Verlag.
Graham, Sage Lambert
2007 Disagreeing to agree: Conflict, (im)politeness and identity in a computer-
mediated community. Journal of Pragmatics 39: 742–759.
Hastrdlová, Šárka
2009 Language of Internet Relay Chat (reconsidering (im)politeness in the context of
IRC). Ph.D. dissertation, University of Masarykova at Brno (Czech Republic).
http://is.muni.cz/th/56286/ff_d_b2/Dizertacni_prace_09_final_version.doc
Hatipoğlu, Çiler
2007 (Im)politeness, national and professional identities and context: Some evi-
dence from emailed ‘Call for Papers’. Journal of Pragmatics 39: 760–773.
Haugh, Michael
2007 The discursive challenge to politeness research: An interactional approach.
Journal of Politeness Research 3: 295–317.
Herring, Susan C.
1994 Politeness in computer culture: Why women thank and men flame. In: Mary
Bucholtz, Anita Liang, Laurel Sutton, and Caitlin Hines (eds.), Cultural Per-
formances: Proceedings of the Third Berkeley Women and Language Confer-
ence, 278–294. Berkeley, CA: Berkeley Women and Language Group.
Herring, Susan C.
1996 Two variants of an electronic message schema. In: Susan C. Herring (ed.),
Computer-Mediated Communication: Linguistic, Social, and Cross-Cultural
Perspectives, 81–108. Amsterdam/Philadelphia: John Benjamins.
Herring, Susan C.
2003 Computer-mediated discourse. In: Deborah Schiffrin, Deborah Tannen, and
Heidi E. Hamilton (eds.), Handbook of Discourse Analysis, 612–634. Oxford,
UK: Blackwell.
Hess-Lüttich, Ernest W. B.
1997 E-Epistolographie: Briefkultur im Medienwandel. In: Andreas Hepp and
Rainer Winter (eds.), Kultur – Medien – Macht: Cultural Studies und Medien-
analyse, 225–246. Opladen: Westdeutscher Verlag.
Hongladarom, Krisadawan and Soraj Hongladarom
2005 Politeness in Thai computer-mediated communication. In: Robin T. Lakoff and
Sachiko Ide (eds.), Broadening the Horizon of Linguistic Politeness, 145–162.
Amsterdam/Philadelphia: John Benjamins.
Hymes, Dell H.
1974 Foundations of Sociolinguistics: An Ethnographic Approach. London: Routledge.
Kerbrat-Orecchioni, Cathérine
1992 Les interactions verbales. Volume 2. Paris: Armand Colin.
Kinginger, Celeste
2000 Learning the pragmatics of solidarity in the networked foreign language class-
room. In: Joan Kelly Hall and Lorrie Stoops Verplaetse (eds.), Second and
Foreign Language Learning through Classroom Interaction, 23–46. Mahwah,
NJ: Lawrence Erlbaum Associates.
Koike, Dale April
1992 Language and Social Relationship in Brazilian Portuguese: The Pragmatics of
Politeness. Austin, TX: University of Texas Press.
Kretzenbacher, H. Leo
2005 “Hier im großen internetz, wo sich alle dududuzen” [‘Here on the big internets,
where everyone is on dududu terms’]: Internet discourse politeness and Ger-
man address. Paper presented at the 3rd International Conference on Language
Variation in Europe (ICLaVE), Amsterdam.
http://www.rumaccc.unimelb.edu.au/address/Internetaddress.pdf
Labov, William
1970 The study of language in its social context. Studium Generale 23: 30–87.
Lakoff, Robin
1973 The logic of politeness: Or minding your p’s and q’s. Chicago Linguistics So-
ciety 8: 292–305.
Lave, Jean and Étienne Wenger
1991 Situated Learning: Legitimate Peripheral Participation. Cambridge, UK: Uni-
versity of Cambridge Press.
Leech, Geoffrey
1983. Principles of Pragmatics. London: Longman.
LeMond, Malia Michele
2004 Synchronous computer-mediated team-based learning in the Spanish foreign
language classroom. Ph.D. dissertation, University of Texas at Austin. http://
www.lib.utexas.edu/etd/d/2004/lemondm74875/lemondm74875.pdf#page=3
Locher, Miriam A.
2006 Polite behavior within relational work: The discursive approach to politeness.
Multilingua 25: 249–267.
Locher, Miriam A. and Richard J. Watts
2005 Politeness theory and relational work. Journal of Politeness Research 1: 9–33.
Medeiros, Sandi Michele de Oliveira
1985 A model of address form negotiation: A sociolinguistic study of continental
Portuguese. Ph.D. dissertation, University of Texas at Austin.
Medeiros, Sandi Michele de Oliveira
1994 Winning friends and influencing people abroad: Using native speakers’ com-
municative strategies. Intercultural Communication Studies IV(1): 23–44.
Mills, Sara
2009 Impoliteness in a cultural context. Journal of Pragmatics 41: 1047–1060.
Murchú, Daithí Óand Elsebeth Korsgaard Sorensen
2004 Online master communities of practice: Collaborative learning in an intercul-
tural perspective. European Journal of Open and Distance Learning, 2004/I.
http://www.eurodl.org/materials/contrib/2004/Identifying_COPs.html
Murray, Dennis E.
2000 Protean communication: The language of computer-mediated communication.
TESOL Quarterly 34(3): 397–421.
Myers-Scotton, Carol
1983 The negotiation of identities in conversation: A theory of markedness and code
choice. International Journal of the Society of Language 44: 115–136.
Nishimura, Yukiko
2003 Linguistic innovations and interactional features of casual online communi-
cation in Japanese. Journal of Computer-Mediated Communication 9(1).
http://jcmc.indiana.edu/vol9/issue1/nishimura.html
Nyugen, Hanh Thi and Guy Kellogg
2005 Emergent identities in on-line discussions for second language learning. The
Canadian Modern Language Review/La Revue canadienne des langues viv-
antes 62(1): 111–136.
Oliveira, Sandi Michele de
2003 Breaking conversational norms on a Portuguese users network: Men as adjudi-
cators of politeness? Journal of Computer-Mediated Communication 9(1).
http://jcmc.indiana.edu/vol9/issue1/oliveira.html
Oliveira, Sandi Michele de
2005 A retrospective on address in Portugal (1982–2002): Rethinking power and
solidarity. Journal of Historical Pragmatics 6(2): 307–323.
Oliveira, Sandi Michele de
2006 A strategic model of address. In: Irma Taavitsainen, Juhani Härmä, and Jarmo
Korhonen (eds.), Dialogic Language Use, 147–158. Mémoires de la Société
Néophilologique de Helsinki, LXVI. Helsinki: Société Néophilologique.
Oliveira, Sandi Michele de
2009 Negotiating identity, conflict, and cooperation within a strategic model of ad-
dress. In: Ann Denis and Devorah Kalekin-Fishman (eds.), The ISA Handbook
in Contemporary Sociology, 416–432. London: Sage Publication.
1. Introduction
2. Literature review
Two related themes in the literature are of particular relevance to this chapter:
What counts as an apology, and to what extent is sincerity important? This section
explores these themes.
Making comparisons with the legal system, Goffman’s seminal work on
apologies (1967, 1971) considered how speakers deal with lesser offences that
nevertheless need to be addressed in order to maintain social order. Goffman’s
focus was on how offences are remedied, apologies being part of a “Remedial In-
terchange” in which the essential components are the offender, the offended, and
the offence. Important in Goffman’s concept of the apology are evidence of regret
on the part of the offender, as well as evidence of acknowledgement or acceptance
on the part of the offended (1971: 90). Indeed, the “Remedial Interchange” can be a
four-part exchange, with the offended drawing attention to the offence, the of-
fender apologising, the offended accepting the apology, and the offender express-
ing thanks for forgiveness received (1967: 20–22). Goffman (1971: 143) identified
a variety of strategies that can be found in an apology:
expression of embarrassment and chagrin; clarification that one knows what conduct
had been expected and sympathizes with the application of negative sanction; verbal re-
jection, repudiation, and disavowal of the wrong way of behaving along with vilification
of the self that so behaved; espousal of the right way and an avowal henceforth to pursue
that course; performance of penance and volunteering of restitution.
Thus Goffman offers insights into what happens when an apology occurs between
two people in a face-to-face situation to address a significant offence, although he
does acknowledge that there is a ritual aspect to an apology which allows partici-
pants “to act as if they feel that matters are closed and that ritual equilibrium has
been restored” (1971: 173).
Subsequent researchers have grappled with the problem of sincerity and of
apologies made for trivial offences. Owen (1983: 119,121) considers that it is not
essential for the speaker to regret the offence. She suggests distinguishing between
“heartfelt” and “routine” apologies, and discusses one solution, to replace sincerity
with consistency: The apologiser would be expected to refrain from re-offending.
Holmes (1995: 164–5) introduces the concept of “appropriate apology behaviour”
which varies according to “power, social distance, and the seriousness of the of-
fence”. More recently, Deutschmann (2003) attempted to devise categories that
would accommodate all of the apologies in his British National Corpus data, al-
though this necessitated forming judgments about their sincerity, which he ac-
knowledged as problematic. In attempting this he tackles the problem of utterances
that look like apologies but have a different illocutionary force, categorising
apologies as genuine, formulaic, and face attack apologies.
Brown and Levinson discuss apologies as negative politeness strategies used to
mitigate face threatening acts (1987). However, Mills problematises Brown and Le-
vinson’s approach, seeing apologies as “complex negotiations between interactants
over status and over who is seen to be ‘in the right’” (2003: 61). Important factors are
“the right amount of effort and work” and the expression of “sufficient commitment
and sincerity” (112). Lakoff (2001) considers that making an apology is more im-
portant than sincerity, pointing out that there are circumstances where a token
apology is more acceptable than one with all the features of a genuine apology, and
Locher and Watts discuss the concept of facework that is “merely appropriate”
(2005: 9). These recent studies offer approaches to what participants would un-
doubtedly recognise as apologies, but where the offence is trivial, the regret by the
offender is not very great, and the apology receives no acknowledgement.
Although Goffman (1971) claimed that acknowledgement of the apology is es-
sential, later work has questioned this. Owen (1983) found that apologies for very
trivial offences often go without acknowledgement or acceptance, but that for
more significant offences absence of response is equivalent to withholding accept-
ance and confirms the offence. This is in keeping with Holmes’ finding (1995) that
people usually respond by accepting the apology. In contrast, Tannen (1992a)
shows that accepting an apology where the apology is intended to be ritual and ex-
The data used in this study are from email discussion lists with academic or pro-
fessional themes (in contrast to discussion lists with recreational or social network-
ing purposes); participants include academics, practitioners, students, and other in-
terested parties. The advantages of such data are that they are naturally occurring,
and they are not influenced by the presence of the researcher. Although some par-
ticipants may have met each other outside the group or have knowledge about the
work of other participants, the only knowledge readily available to all of the par-
ticipants on a list is that which is found in the message text. Thus the entire dis-
cussion takes place through the medium of email, and the researcher has the same
access to the text as did the participants. Most participants give their first names,
and most of these are gender specific, although in emails it is possible to try to dis-
guise one’s gender by using a gender-neutral name, nickname, or initial, or even
adopting a name of the opposite gender (Baldwin 1996). However, many participants
give their affiliations, and several reference their own publications, so it is clear in
most cases that participants are using their own names. The gender of the sender
was identified in 99 % of the messages and in 99 % of the apologies (see Table 2).
A corpus of problematic incidents was collected by “specimen sampling” (ten
Have 1999: 50–51) from discussion list archives publicly available on the Jiscmail
website (participants are informed on joining a list that their discussions will be
made available in this way). Episodes were selected that contained problematic
incidents such as hostility, miscommunication, or deliberate offences, as well as
non-hostile disagreements. The data used for the apology study are 24 samples
taken from each of eight listserv lists covering a range of topics, including
Geography, Management, and Humour Research. Each sample covers a period of
one month and contains all of the messages sent to that list during the month. Be-
cause of our interest in problematic incidents, each sample includes an episode
such as hostility, miscommunication, a deliberate offence, or non-hostile disagree-
ment, although by including all of the messages for each selected month there is
much that is unrelated to these incidents, including announcements and requests
for information. By using this particular set of data for our study, we hoped to find a
wide range of apology strategies, including apologies related to the problematic
incidents and apologies that occurred during the normal business of the discussion
groups. In the examples below, names have been replaced by FN (first name) and
LN (last name), and any contact details have been removed. The examples retain
the spelling and grammar of the originals.
The total number of messages in the 24 samples is 2,420, and the total word
count, including headers and quotations, is 1,849,397. Because of the size of this
corpus, it was decided to adopt the practice followed by earlier researchers (Blum-
Kulka, House, and Kasper 1989; Deutschmann 2003; Owen 1983) of searching for
apologies using selected word forms, although the authors recognised that this
would exclude any apologies which used other forms or which were expressed in-
directly. The selection of word forms for our study is based on the apology IFIDs
(Illocutionary Force Indicating Devices) listed in the Cross-Cultural Speech Act
Realization Project (CCSARP) coding manual (Blum-Kulka, House, and Kasper
1989: 290): sorry, excuse, apologise, forgive, pardon, regret, and afraid. Prelimi-
nary observations of the data suggested that it would be important to distinguish
the use of “apologies” from “apology”: “Apologies” is used in a distinctive way
and will be discussed in detail below. Thus the search terms used for this study are:
apology, apologise, apologize, apologies, sorry, forgive, forgiveness, excuse,
afraid, regret, pardon. All occurrences of these forms in our data were checked
manually to exclude quotations, reported speech, and words used in a sense other
than the expression of an apology, e.g., “I don’t have to excuse my actions”.
This procedure generated a corpus of 260 apologies. A small number of these
apologies contained more than one apology lexeme, as in “I ask pardon and for-
giveness”. Thus the total number of apology expressions, 278, is slightly higher
than the total of the apologies. In the tables below, the 24 samples are numbered
1–24. Samples from the same list are shown consecutively, and the eight lists are
labelled A-H.
4. Apology forms
The frequency of apology lexemes in our data is shown in Table 1. In the email dis-
cussion data, “apologies” is the most frequently-used form, with 117 occurrences
out of a total of 278 (42.1 %). If we sum all the forms of this lexeme, i.e.,
“apology”, “apologies”, “apologise”, and “apologize”, this total rises to 137 out of
278 (49.3 %). This is in contrast to earlier studies which found that “sorry” was the
most frequently-used form: Owen (1983: 65), who collected data from telephone
calls and recordings in shops, notes that “sorry” is “[b]y far the commonest” form;
Holmes (1995: 161) includes “I’m sorry” in her category “expression of regret”,
which accounts for 53.3 % of female and 42 % of male apologies in her New Zeal-
and workplace corpus; and Deutschmann (2003) finds that “sorry” is used in
59.2 % of cases in his BNC data. Thus “sorry” was found to be the most com-
monly-used form in all three of these widely different spoken corpora. However,
Hatipoğlu (2005) found that whereas in personal (one-to-one) emails, an ex-
pression of regret such as “sorry” was the most frequently-used form (85.7 %), in
professional (one-to-many) emails where the apologiser was apologising in role,
the most common form was an “offer of apology (e.g., I apologise)” (97 %). In our
own data, “sorry” was the second most-frequently used form in the data, account-
ing for 92 instances out of 278 (33.1 %). A variety of other forms were used, albeit
much less frequently.
Owen (1983: 63–65) found that “The use of type (i) [apology, apologies, or
apologise] is rare in spoken English; it is used when either formality or absolute
unambiguity is required”. She had only two occurrences in spoken language, one
very formal and one found in a lecture. The other occurrences in her data were
found in written form. The prevalence of “apologies” in the email discussion data
could therefore be related to the fact that the data are written. However, another
possibility is that use of the noun form enables the writer to avoid the personal pro-
noun, creating a distance between the writer and the responsibility for the offence
(Hatipoğlu 2005).
Considerable variation characterises the lexeme use of different lists. Thus
while “apologies” is the most frequently-found form overall, it does not occur at all
in the samples from list D; and while “forgive” is uncommon, it occurs in all of the
samples from lists E and G.
Because our data samples were selected to contain problematic incidents, which
have been found to be associated with males in online communication (Herring
1993, 1994), we predicted that we would find greater participation (more mess-
ages) by males, and this was indeed the case. As shown in Table 2, males con-
tributed 1719 (71 %) of our messages (total 2420), while females contributed 674
(28 %) (note that totals are less than 100 % because the gender of a very small
number of writers could not be identified). In contrast to the findings on male and
Percentages rounded
*? Signifies gender not identified
female apology behaviour discussed above, the majority of apologies in our data
are made by males, although the overall percentage of apologies by gender is very
similar to the overall percentage of messages sent: males 191 apologies (73 %),
females 66 (25 %). However, the overall pattern masks variation in individual
samples: Looking at samples where the difference is at least 5 %, we see that men
are overrepresented in apologies relative to number of messages in samples 3, 7, 8,
11, 14, 15, 17, 18, 23, 24 (10 samples in total), while women are overrepresented in
apologies in samples 1, 2, 5, 6, 10, 13, 16, 19, 20, 21, 22 (11 samples in total).
The ratio of apologies to turns (i.e., messages) varies widely from one data sample
to another, ranging from one apology in 26 turns (sample 11, ratio of 1:26) to 16
apologies in 76 turns (sample 3, ratio of 1:4.6). Across all samples, the average
ratio of apologies to turns is 1:9.3. For men, the overall ratio is 1:9, and for women,
1:10.2. Thus men make overall slightly more apologies in relation to turns than
women, but the variation across different data logs is so great that this difference is
not significant. Our overall ratio of apologies to turns seems remarkably high;
compare Deutschmann’s finding of 1 apology to 165.6 turns (2003: 51). However,
we would expect to find a higher incidence in data that have been selected to in-
clude problematic incidents. Moreover, messages in email often carry the interac-
tion further than is usual for turns in conversation (Harrison 2007), and each mess-
age can include an accumulation of moves that in spoken conversation would
normally require several turns.
In contrast to spoken conversation, apologies in email discussions are often
made to all of the participants. In our data, 66.5 % (n=173) of apologies were made
to the whole group (e.g., examples 2, 5), 29.6 % (n=77) to individuals (e.g.,
examples 4, 20), and 3.8 % (n=10) to a sub-group as defined by the writer, such as
the following, where the writer is apologising to those for whom the message
would be irrelevant:
(1) Apologies to those of you not going to New Orleans
such as a cross-posting or a long message; (2) retrospective apologies for minor of-
fences, such as sending a blank message or giving an incorrect website address; (3)
retrospective apologies for more serious offences where the tone or the content has
caused offence; and (4) instances in which the form of the apology is subverted and
used for challenges, sarcasm, or joking (see Table 3).
*Sample 4 has one apology, for “length and simplicity”, which realises both categories 1 and
4. This is entered in both columns but is added into the total column once only. See dis-
cussion of example 27.
Our data were selected to include problematic incidents, and we therefore pre-
dicted that there would be some apologies relating to offences that are more serious
than those discussed above. This is indeed the case, although such apologies are in
the minority (14 %, n=37). Apologies for serious offences occur mainly in lists C,
E, F, and G (see Table 3), especially in Samples 6, 14, and 19. There is no typical
offence or form of apology for serious offences – each is unique, and some are
quite complex. These apologies relate to offences such as using words that cause
offence, deliberately sending a private email to other recipients, or posting poten-
Here the vocative and apposition “FN1, my friend”, indicate that it is the behav-
iour, not the person, that is condemned. This prompt is treated seriously by the re-
cipient and generates an acceptance of responsibility, “I see what you mean”, fol-
lowed by an apology. The apology uses the personalised verb form, reinforcing the
acceptance of responsibility, and the apology is intensified with “sincerely”:
(22) Yes, I am going over my posting and I see what you mean. I apologise sin-
cerely for it.
(25) Oh, sorry, forgot to mention, that’s the price the supermarkets pay the farmers.
Here the writer constructs community identity among the email discussion partici-
pants by aiming sarcasm at supermarkets, which are positioned as a common
enemy.
This writer clearly suspects that he has “rambled”. Rather than make an effort to
check and perhaps revise his text, he uses the if clause to question and also to dis-
tance himself from the offence.
Our data also contain examples of if clauses accompanying serious offences
(e.g., examples 14, 19). By questioning whether an offence has occurred, writers
can apologise without fully accepting responsibility and so save face, even though
in each of these examples, the writer would have known from the previous posts
that an offence certainly had occurred. In a face-to-face situation, other partici-
pants might respond, saying whether an offence occurred, but in asynchronous
email discussions the moment passes and the apology is allowed to stand without
comment.
9. Conclusion
The research reported in this chapter has shown that email discussions have dis-
tinctive patterns of apology use that have implications for pragmatics. In particu-
lar, we observe writers adapting their strategies and level of effort in apologising to
match the seriousness of their offences, and we find that minimal unacknowledged
apologies for trivial offences are an established norm.
Although considerable variation was evident among our samples, some overall
patterns emerge. Apologies in our email discussions explicitly refer to the offence:
This is necessary because the communication takes place asynchronously and in
writing. Apologies are frequently addressed to the whole group but may be ad-
dressed to a defined sub-group or to an individual, and they relate to offences that
are internal to the email discussion itself, either to the use of the hardware or soft-
ware or to the content or tone of an email message.
Category (1) apologies differed the most from apologies in spoken language.
They accompanied the offence in the same message, usually at the start of the
message. They were brief and often took the noun form “apologies”, enabling the
apologiser to distance her/himself from the offence. Our findings confirm and ex-
tend the findings of Locher and Watts (2005): The evidence suggests that although
many of the apologies in our data were not genuine expressions of regret for real
offences, the apologies were not out of place. The writers clearly felt that they were
necessary, or desirable, and no objections were made (at least publicly on the mail-
ing lists) to the apologies or to the way in which they were expressed; clearly, these
writers are paying appropriate and relevant attention to their recipients. These find-
ings convince us that even apologies that are “merely appropriate” (Locher and
Watts 2005: 9) and that relate to trivial offences show adherence to group norms
and are important for the smooth running of email discussions.
Apologies in categories (2) and (3) have parallels in spoken language, although
in email discussions they need to refer explicitly to the offence. They occur after
the offence and use a variety of forms to indicate that the apology is genuine: per-
sonal pronouns, verb forms, intensifiers, explanations, and excuses. Acceptance of
apologies is not common and is found only for serious offences. Acceptance seems
to confirm that an offence has taken place, and we have the interesting albeit rare
example in (14) of reciprocal apologies being used to defuse a hostile situation.
Category (4), subverted apologies, also have parallels in spoken language and have
been largely ignored by researchers, except for Deutschmann (2003). This is an
underexplored area that merits future research.
The use of if clauses is much more common in our data than in the spoken
data of Owen (1983) or Deutschmann (2003). It allows apologisers to distance
themselves from the offence, especially in a context where multiple recipients and
lack of immediate feedback can create genuine uncertainty as to whether an of-
fence has taken place. This interesting finding also offers scope for further inves-
tigation.
However, there were marked differences among samples across all the areas of
our investigation, so that although we can make overall statements about the data,
these may not be true for a specific sample. Cross-postings are the most common
offence in these data, but some samples have no apologies for cross-postings.
While the gender of apology senders is approximately in proportion to the gender
of the message senders as a whole, some samples have no apologies from women,
while in others women contribute the majority. Some samples have no apologies
addressed to individuals, whereas others have many. Some samples have only cat-
egory (1) apologies, while others contain a variety of apology forms for a range of
offences. Some samples have no apologies for serious offences, while in others
serious offences abound. These findings suggest that local practices have devel-
oped in different email lists, and it is not only the transfer of strategies from spoken
language modified to some extent by the affordances of the medium, but also these
local practices, that influence apology behaviour.
Notes
1. This work was carried out as part of a project funded by the Arts and Humanities Re-
search Council (UK) entitled “Dealing with Trouble in Email Discussions”.
2. Most participants give first names which are gender specific, and in many cases these are
corroborated by institutional affiliations, but in any case where a male has used a female
first name, this would be counted as female in our statistics.
3. Contrast the data of Davies et al. (2007), where, in corresponding with their tutors, stu-
dents generated email apologies for external offences such as missing a lecture.
References
Baldwin, Beth
1996 Conversations: Computer-mediated dialogue, multilogue, and learning. Ph.D.
Dissertation, University of North Carolina at Greensboro.
Baron, Naomi
1998 Letters by phone or speech by other means: The linguistics of email. Language
and Communication 18: 133–170.
Blum-Kulka, Shoshana, Juliane House, and Gabriele Kasper
1989 Appendix: The CCSARP coding manual. In: Shoshana Blum-Kulka, Juliane
House, and Gabriele Kasper (eds.), Cross-Cultural Pragmatics: Requests and
Apologies, 273–294. Norwood, NJ: Ablex.
Brown, Penelope and Stephen Levinson
1987 Politeness: Some Universals in Language Usage. Cambridge, UK: Cambridge
University Press.
Connery, Brian
1997 IMHO: Authority and egalitarian rhetoric in the virtual coffeehouse. In: David
Porter (ed.), Internet Culture, 161–179. New York: Routledge.
Davies, Bethan, Andrew Merrison, and Angela Goddard
2007 Institutional apologies in UK higher education: Getting back into the black be-
fore going into the red. Journal of Politeness Research 3(1): 39–63.
Deutschmann, Mats
2003 Apologising in British English. (Skrifter från moderna språk 10.) Umeå: Insti-
tutionen för moderna språk, Umeå University.
Goffman, Erving
1967 Interaction Ritual: Essays on Face-to-Face Behaviour. New York: Doubleday
Anchor Books.
Goffman, Erving
1971 Relations in Public. Microstudies of the Public Order. London: Penguin.
Grainger, Karen and Sandra Harris
2007 Apologies: Introduction. Journal of Politeness Research 3(1): 1–9.
Harrison, Sandra
2004 “Reasoned argument” or “gossip”? Metadiscussion in an email discussion
group. In: Linda Allal and Joaquim Dolz (eds.), Proceedings Writing 2004,
CD. Adcom Production.
Harrison, Sandra
2007 Transgressions, miscommunication and flames: Problematic incidents in email
discussions. In: Mia Consalvo and Caroline Haythornthwaite (eds.), AoIR In-
ternet Annual Volume 4, 105–117. New York: Peter Lang.
Hatipoğlu, Çiler
2005 Do apologies in e-mails follow spoken or written norms? Some examples from
British English. Studies about Languages 5: 21–29.
ten Have, Paul
1999 Doing Conversation Analysis: A Practical Guide. London: Sage.
Herring, Susan C.
1993 Gender and democracy in computer-mediated communication. Electronic
Journal of Communication 3(2).
http://www.cios.org/EJCPUBLIC/003/2/00328.HTML
Herring, Susan C.
1994 Politeness in computer culture: Why women thank and men flame. In: Mary
Bucholtz, Anita Liang, Laurel Sutton, and Caitlin Hines (eds.), Cultural Per-
formances: Proceedings of the Third Berkeley Women and Language Confer-
ence, 278–294. Berkeley, CA: Berkeley Women and Language Group.
Herring, Susan C.
1999 Interactional coherence in CMC. Journal of Computer-Mediated Communi-
cation 4(4). http://jcmc.indiana.edu/vol4/issue4/herring.html
Herring, Susan C.
2003 Gender and power in online communication. In: Janet Holmes and Miriam
Meyerhoff (eds.), The Handbook of Language and Gender, 202–228. Oxford,
UK: Blackwell.
Holmes, Janet
1995 Women, Men and Politeness. London: Longman.
Kolb, David
1996 Discourse across links. In: Charles Ess (ed.), Philosophical Perspectives on
Computer-Mediated Communication, 15–26. Albany, NY: State University of
New York Press.
Lakoff, Robin
2001 The Language War. Berkeley, CA: University of California Press.
Locher, Miriam and Richard Watts
2005 Politeness theory and relational work. Journal of Politeness Research 1(1):
9–33.
Mills, Sara
2003 Gender and Politeness. Cambridge, UK: Cambridge University Press.
Owen, Marion
1983 Apologies and Remedial Exchanges. Berlin: Mouton.
RFC 1855
2006 Netiquette guidelines. http://www.dtcc.edu/cs/rfc1855.html#3 [accessed No-
vember 5, 2011]
Smith Christine B., Margaret L. McLaughlin, and Kerry K. Osborne
1997 Conduct control on Usenet. Journal of Computer-Mediated Communication
2(4). http://jcmc.indiana.edu/vol2/issue4/index.html
Tannen, Deborah
1992a That’s Not What I Meant! How Conversational Style Makes or Breaks Your Re-
lations with Others. London: Virago.
Tannen, Deborah
1992b You Just Don’t Understand: Women and Men in Conversation. London: Vir-
ago.
1. Introduction
2. Background on advice-giving
impart the content of advice are influenced by how face-threatening people per-
ceive the act to be. In contexts where advice-seeking and advice-giving are per-
ceived as threatening because of asymmetry between the interactants (Goldsmith
and MacGeorge 2000: 235), mitigation of advice is likely to occur. This can take the
form, for example, of indirectness, hedging, or embedding the advice in further dis-
course (cf. DeCapua and Dunham 2007; Locher 2006a; Morrow 2006). The actual
piece of advice can be realized in many different linguistic forms, from the classic
conditional clauses accompanied by should-instructions, I would do X statements,
and indirect statements realizing suggestions, to imperatives and questions inviting
introspection and action (e.g., DeCapua and Dunham 2007; Hudson 1990; Locher
2006a). Examples of these categories are given in section 4.
Finally, the difference between information-giving and advice-giving needs to
be addressed. Silverman (1997: 154) maintains that non-personalized information
can be interpreted as advice in the appropriate context. In the HIV counseling
sessions that he studies, there are so-called “Advice-as-Information” sequences,
in which the counselor gives general information on HIV and AIDS prevention
without fine-tuning it to the counselee’s particular needs. Because this allows
the advisee to interpret the text as not relevant, or as relevant for others than him-
or herself, this kind of advice is claimed to be less face-threatening to the advisee.
At the same time, the context in which this information is imparted makes an in-
terpretation as advice possible. The context1 of the interaction, or what Waring
(2007: 373) calls its “inferential framework”, thus serves as a backdrop for inter-
actants to work out the pragmatic meaning of the information given as advice.
In addition to sites that are clearly identified as imparting advice, there are
many others in which advice is less explicitly propagated. Many peer support sites,
for example, thrive on the benefits of sharing experience, which can sometimes be
construed as advice for action or non-action (see, e.g., McSeveny et al. 2006 on
posts to an online weight watchers message board; Kouper 2010 on message ex-
changes in a motherhood blog community; Morrow 2012 on divorce advice in a
Japanese online forum; Placencia 2012 on advice in Spanish Yahoo!Respuestas).
Similarly, in the context of the above-mentioned American forum on breast cancer,
the interpretation of an informative sequence or a narrative as advice depends on
the overall frame of the forum, which is meant to be helpful for its readership. That
narratives of personal experience can be interpreted as advice was confirmed in a
study by Harrison and Barlow (2009) of an online arthritis self-management pro-
gram. The participants in this program posted their action plans for self-manage-
ment online and received comments from their group members, in many cases, in
the form of narratives of personal experience. Harrison and Barlow highlight that
their study “demonstrates the importance of communication context for interpre-
tation. In the context of the arthritis workshops, participants are expecting feed-
back on their action plans, and are therefore able to interpret the narratives as ad-
vice” (2009: 108). If readers turn to a site with the aim of finding advice, they are
likely to interpret text as advice even if it is not obviously flagged as such (Waring
2007: 373); this is discussed further in section 2.
Having established that advice can be given online on different interactive plat-
forms and that advice-giving itself is not tied to any single linguistic realization,
the question remains to what extent computer mediation influences advice-giving
and to what extent advice-giving on the Internet differs from advice-giving in face-
to-face interaction or print media. In other words, what are the effects of techno-
logical mediation on advice-giving? A tentative answer concerning Internet advice
columns is advanced below.
In this chapter advice columns are defined as a particular “text type”, i.e., as a
type of written text that is conventionally recognized by people who are familiar
with the print media (Hendley 1977). This recognizable format of advice columns
consists in the exchange of a “problem letter” sent by an advice-seeker and a
“response letter” published by an agony aunt. The term “letter” is used throughout
this chapter, since both the advice-seekers and the advice-givers typically adhere to
the convention of writing a simple letter consisting of an address form (“Dear
Agony Aunt”; “Dear Reader”), followed by the body of the letter and a sign off.
It is important to stress that advice columns are public, while simulating a per-
sonal exchange. The advice-givers gear their responses towards the needs of the
target readership and use the advice-seeker’s problem as an example. To choose
a(n at times fictional) dialogue form to present advice as in advice columns
can provide an incentive for the reader to identify with the acting subject and offer
the reader the opportunity to take on the role of advisee (Franke 1997: 226). The
tension between the public and personal aspect of such texts has garnered research
attention (e.g., Fleischhacker 1987; Locher 2006a, Chapter 7; Mininni 1991).
It manifests itself, for example, in the choice of which “letter” to answer in public
in the first place (usually those questions that are likely to be of relevance to
the larger readership rather than just to the individual advice-seeker) and/or in
the broadening of the scope of the answer in order to best address the needs of the
wider readership. Hutchby (1995), who studied a call-in radio program, found this
latter strategy to be important in his data: The radio host prompted the expert pres-
ent in the studio to answer in more detail than the caller originally asked for, in
order to satisfy what was assumed to be the needs of the wider audience.
In addition, many advice-seekers value the possibility of reading and asking
questions anonymously. This is especially important when the topic of advice is
delicate, such as issues relating to sexuality. Alexander (2003: 548) makes the point
that fears and embarrassment about health issues are the reason men write anony-
mously to the magazine Men’s Health. van Roosmalen (2000: 205) observes that
problem pages are “forums for the unspeakable” for the adolescent women who
convey their experience in letters to an advice column in Teen Magazine. The appeal
of anonymity is certainly an important factor for many online advice sites, as well.
Advice columns have been in use ever since the advent of newspaper print cul-
ture (Hendley 1977), and they are well known in the literate world today (DeCapua
and Dunham 2007: 322). They are a popular feature of many print magazines, and
for this reason it is not surprising that advice columns are also found on the Internet.
According to Hendley (1977: 345, 351), advice columns are “naturally appealing”
because people need advice and are curious and nosy by nature. In other words,
people read advice columns not just to gain knowledge on action alternatives, but
also because they can identify with the advice-seekers and/or are looking for enter-
tainment. Every advice column has a particular flavor that may influence the popu-
larity of the column and that may be in line with its institutional aims and/or the
norms and values of the advice-giver (e.g., Currie 2001; Gough and Talbot 1996;
Locher and Hoffmann 2006; McRobbie 1978, 1991; Mutongi 2000; Talbot 1992,
1995; Thibault 1988, 2002). This flavor, or “voice”, is conveyed by the linguistic
strategies that contribute to the identity7 construction of the advisor persona.
5.1. Background
Professional organizations that deal with health issues are confronted with the
challenge of how best to impart knowledge and how to learn about their “client’s”
needs. One strategy is to use the format of advice columns and the medium of the
Internet to appeal to and reach a large readership. An example of one such column
is “Lucy Answers” (the name has been changed at the specific request of the web-
site), which is run by the health education program of a large university in the
United States. According to information on the website, this advice column is
meant to complement other services offered by the health education program, such
as face-to-face counseling opportunities and printed information material (Lucy
Answers 2004). Topics covered pertain to the fields of “alcohol and other drugs”,
“fitness and nutrition”, “emotional health”, “general health”, “sexuality”, “sexual
health”, and “relationships”. The official aim of the site is to
increase access to, and use of, health information by providing factual, in-depth,
straight-forward, and nonjudgmental information to assist readers’ decision-making
about their physical, sexual, emotional, and spiritual health. (Lucy Answers 2004)
The mission of the site is thus to enable advice-seekers to make up their own
minds about the issues covered without being directive (cf. the discussion of the
ideal of non-directiveness for professional counselors in Sarangi and Clarke 2002 or
Vehviläinen 2003). The site is aimed at college students, while being openly access-
ible to everybody interested. The questioners are entirely anonymous, and no
information is available about their background (Lucy Answers 2004). In 2004,
“Lucy Answers” received nearly 2,000 inquiries per week, although only five were
answered each week (Lucy Answers 2004) – the choice of which question to publish
is determined by the importance of the topic for the public readership. However, this
means that the site is not geared to cater for emergencies, a fact that is stressed by
disclaimers on the site. The “response letters” are written by a team of professional
health educators under the pseudonym “Lucy”. The team thus creates an advisor
persona, an agony aunt, with her own particular voice (Locher and Hoffmann 2006).
In section 5.2, I report on the results of a study of the linguistic realization
of advice in 280 question and response letters (a detailed report can be found in
Locher 2006a). The letters were collected from the online archive in 2002 and 2004.
This brief overview of findings serves as a backdrop for section 5.3, which focuses
on the extent to which the advice column is influenced by being published online.
(1) In the meantime, you can learn ways to be his friend without enabling his
drinking. (alcohol and drugs, “Cocaine versus tequila”)9
(2a) What makes your particular attraction a problem? Is it due to the fact that
we almost always see men romantically involved with women who are
shorter than they are? Or, is it because it can seem strange to come upon the
opposite? (relationships, “Troubled by attraction to tall women”)
(2b) Since this fine financier isn’t your boss or teacher, why not take the letters
A-T-M to stand for Approach That Man. (relationships, “Should I cash in on
hot banker?”)
(3a) Imagine how lonely and frustrating it might feel to always be excluded.
(relationships, “How to tell a nosy roommate to step off”)
(3b) Discuss with him/her whether you might see someone else in Health Services
or whether s/he would be willing to refer you to someone off campus with
whom you could do longer term work. (emotional health, “Disappointed with
therapist?”)
that is created for the advisor persona, i.e., the linguistic strategies that contribute
to the identity construction of “Lucy”, not only displays awareness of potential
face-threats, but also displays a sense of humor (8), as well as empathy and sup-
port (9):
(8) Since this fine financier isn’t your boss or teacher, why not take the letters
A-T-M to stand for Approach That Man. (relationships, “Should I cash in on
hot banker?”)
(9) Lucy understands, however, how frustrating it can be when these habits are
bothersome. (relationships, “No one cares about drunken, messy hallmates, or
do they?”)
Balanced against these strategies that make her appear approachable and infor-
mal, the advisor persona uses a number of linguistic strategies to highlight her
status as a professional expert: “Lucy” gives detailed background information,
quotes facts in numbers and percentages, broadens the scope of the answer, refers
the reader to other sources of information, and refers to herself as a health educator
(Locher 2006a, section 8.2.2). Richardson (2003: 172) points out that credibility
and trust are often at the core of whether an individual’s posting will be accepted or
not: In the Internet newsgroups Richardson studied, users employed what she calls
“warranting strategies”, which are “designed to give fellow participants reasons to
take the information seriously”, for example by referring to (presumably repu-
table) sources (2003: 172). The strategies mentioned above that highlight profes-
sionalism can be claimed to have a similar function in “Lucy Answers”. In addi-
tion, the visual appearance of the column – the placement of the university logo,
the hyperlinks to the health program, and a description of the history and aims of
the service as well as the advisor team behind “Lucy” – situate this practice in a
professional context. In this sense, the site is visually different from those of lay
people who adopt the format of an advice column. Many such people genuinely
wish to help, but they have no further qualifications or training in advising than
their own personal experience and common sense – a fact that many admit quite
frankly in general disclaimers on their sites.13
The strategies of imparting advice discussed in this section on how advice
linguistically manifests itself in “Lucy Answers” can be explained in terms of the
relational work called for by this text type in this particular context. It seems that
advice-giving, which involves the negotiation of asymmetries between interactants
(see section 2), is perceived as face-threatening in “Lucy Answers”, despite the
anonymity of the site. The health advisor team orients itself to the mission state-
ment of the institution and follows a non-directive path in advising. In this way, the
team takes the relational and cultural implications of this text type in this particular
context into account (cf. Leppänen 1998).
S4: Topic or theme “Alcohol and other drugs”, “fitness and nu-
– Of group, e.g., politics, linguistics, trition”, “emotional health”, “general
feminism, soap operas, sex, science health”, “sexuality”, “sexual health”, and
fiction, South Asian culture, medieval “relationships”.
times, pub
In what follows, I comment on those factors that most influence the “Lucy
Answers” discourse. These are the medium factors M3: Persistence of transcript,
M6: Anonymous messaging, and M10: Message format, and the situation factors
S1: Participation structure, S2: Participant characteristics, S3: Purpose, S5: Tone,
and S7: Norms. Since these factors influence each other, the sequence of discussion
will not follow the ordering of factors given in the table, which is for ease of refer-
ence only and does not reflect degree of importance, as Herring (2007) points out.
With respect to S7: Norms of language, the syntactic strategies illustrated in
section 5.2. are not unique to online interaction. They are in fact all-pervasive and
gain their specific meaning through their combination in a particular social prac-
tice. Pragmatic meaning in this case is created by the orientation of the interactants
to the text type of an advice column and to the mission statement of the health
institution. The level of non-directiveness can be explained by the orientation
of the health educators to empowering their readers with knowledge rather than de-
ciding for them. DeCapua and Dunham (2007) found very similar linguistic results
in a study carried out with native speakers of English who had to compose response
letters in the format of an advice column in a discourse completion task, despite the
fact that these native speakers were not professionals in the fields of advice. The
non-native speakers of diverse cultural backgrounds in the same study, however,
generally formulated more direct and less mitigated and embedded responses,
indicating a lack of knowledge of and/or orientation to the socio-pragmatic norms
of the text type (see Bardovi-Harlig and Hartford 1990 for similar findings with
respect to the lack of pragmatic competence in non-native speakers in academic
advising sessions).
DeCapua and Dunham’s (2007) native speakers were all adults or young adults.
The age and experience/expertise factor (S2: Participant characteristics) may ex-
plain the different picture that emerges for peer-to-peer Internet sites covering
similar topics, such as the Studentcenter’s TeenAdvice section (Studentcenter
2004). Here there is a lack of mitigation or seeming awareness of the delicate
nature of advice-giving, as evidenced in the blunt and often face-threatening
manner in which advice is given, often not in the typical letter format. The fact that
the native speakers in DeCapua and Dunham’s (2007) study employed similar
linguistic strategies as the advice-giving team of “Lucy Answers” might indicate
that both groups orient themselves to the face-threatening character of advice-giv-
ing and to the conventions of the text type “advice column” because they are more
experienced in reading and engaging in this particular text type.14
With respect to S3: Purpose, S5: Tone, and S7: Norms of language, I suggest
that the purpose of the advice column has a stronger impact on the choice of lan-
guage (S5 and S7) than the fact that the site belongs to computer-mediated dis-
course. Advice columns are not written in a social vacuum; the ideals and aims
of the institution responsible for a site or of a private individual who runs an advice
column will have an impact on the language used. In the case of “Lucy Answers”,
the clear mission statement leads the team of health educators to use Standard Eng-
lish, lacking computer-mediated language features identified elsewhere (e.g., ab-
breviations or emoticons; see Bieswanger, this volume), to impart knowledge in a
“factual, in-depth, straight-forward, and nonjudgmental” way. At the same time,
the vocabulary employed is not too scientific or formal (i.e., it is straightforward),
and there are displays of humor, support, and empathy, as well as criticism (cf.
Locher 2006a, Chapter 8). In addition, the goal of “Lucy Answers” is to reach and
educate as many individuals of the target group as possible. This public dimension
(S1) also influences the composition of the response letters, as shown in the dual
strategies of choosing a problem to answer that is of relevance to the general public
and within the scope of “Lucy Answers”, and of broadening the scope of the
answer in order to best cover the needs of the wider readership.
The composition of the texts is also influenced by the fact that both the problem
letter writers and the advisor team orient themselves to the text type of established
print “advice columns” that make use of the format of “letters” (address terms,
body of text, in some cases farewell sections, signatures), and by the fact that one
problem letter is followed by one response letter (M10: Message format and S1:
Participation structure). No further possibility for interaction between the advisor
team and the individual advice-seeker is envisaged by the technological set-up of
the site. The advice-seekers cannot be contacted for clarifications of their problem,
as their email addresses are scrambled upon submission, nor can the advice-
seekers reach “Lucy Answers” publicly for a follow up. In this way the technologi-
cal set-up of the site differs, for example, from threads in an advice forum or blog,
where advice-seekers post problems and receive (multiple) answers to which they
can react if they wish. The advice-seekers on “Lucy Answers” can try their luck
and submit a further request to the site, or they can respond in the “reader response
section”, a possibility they rarely take advantage of (see below).15 Because the
problem letters are quite sparing in the amount of information they provide (on
average only 78 words) and the advisor team has no way to ask for clarifications as
in a face-to-face counseling situation, the team often makes transparent the fact
that the advice given is based on insufficient background knowledge – a practice
that mostly occurs in the assessment moves or in advice sentences realized by if
clauses. The latter restrict the advice to a certain group of people, i.e., to those for
whom the condition holds. While the restriction of the participation structure (S1)
to one turn only also holds for print advice columns, “Lucy Answers” readers have
the option of clicking on further links to related questions and answers and reader
responses, so that they are involved in a much wider communication process (see
also the comments below on the archive).
Several researchers have pointed out that anonymity (M6: Anonymous mess-
aging and S1: Participant structure) is a major advantage for online interaction, in
that it allows advice-seekers to enquire about their sometimes delicate problems
without revealing their identity (e.g., Alexander 2003; van Roosmalen 2000;
Wood and Griffiths 2007: 385). Comparing problem-seekers’ formulations with
face-to-face counseling on sexual issues, Harvey et al. (2007: 771) noted that the
teenagers who wrote to the email advice service www.teenagehealthfreak.org “de-
scribed themselves, their anatomy and their identities in meticulous detail, (…)
whereas previous research on sexual health has discovered the use of vague terms
and euphemisms”. Anonymity might be a factor allowing for this detail – or, in the
case of “Lucy Answers”, it might be a reason for turning to the site in the first
place, as opposed to making use of the face-to-face counseling option available on
campus. In “Lucy Answers”, anonymity is guaranteed, in that any submission in-
formation that might lead back to the advice-seeker is destroyed. In the case of
The growing body of letter exchanges – the history of the site – thus influences
the creation of new texts, as well as the process of updating old texts. It is also
possible to argue that the archive influences the creation of the problem letters, on
the assumption that advice-seekers first “lurk” – in the case of “Lucy Answers”,
browse the archive and read numerous exchanges – before they actively submit a
question themselves. In this way, they may be influenced by the way in which the
problem letters are structured, both with respect to the adherence to the general no-
tions of letter writing (address terms, body, signature) and with respect to particu-
lar styles of writing (e.g., problem statements and narratives to describe prob-
lems).
The Participation structure (S1) of the platform also allows readers of the col-
umn to comment on the exchange of problem and response letter. While the “Lucy
Answers” team does not take any responsibility for the content accuracy of these
comments, the reader response texts form a corpus of their own in which support
and advice, as well as criticism, are given. The reader response comments are vis-
ible as hyperlinks directly below the problem-answer exchange. This is the result
of a general layout change to the site that took place in the early 2000s. Before that,
it was difficult to find the comments, and by 2004 the corpus contained only 446
contributions, while the archive of problem and response letters already contained
2,286 exchanges. It is not possible to search the reader response corpus as such.
The comments are thus accessible solely via the original “problem letter” –
“response letter” sequence to which they are linked. That is, the site wishes to give
its readers a more active outlet than before, but does not want the reader responses
to be included in the search for information, since the advisor team cannot guaran-
tee the quality and accuracy of the information contained in the reader responses.
Nevertheless, these texts complement the professional site in an informal way, and
may add to the appeal of the site more generally in that a sense of online commu-
nity is created (cf. Baym 1998). The advice-seeker who turns to the “Lucy
Answers” column thus has more interactive opportunities and more access to ad-
vice than the reader of a print advice column.
Giving and receiving advice has been discussed in studies of language in use from
the perspectives of many different fields (pragmatics, applied linguistics, psychol-
ogy, health related disciplines, etc.). The Internet offers a platform for advice-seek-
ing and advice-giving to both professionally-trained advisors and lay people,
for profit or free of charge. This chapter focused on advice in the context of online
health issues. Advice in this realm has important potential real-world conse-
quences, and professionals need to ask themselves how to impart information
and advice to their target group in such a way as to augment the likelihood that
the advice is heeded (cf. Feng and MacGeorge 2006), at the same time adhering to
institutional constraints, such as for example an ideal of non-directiveness.
DeCapua and Dunham (2007: 319) point out that “the socio-pragmatic rules or
norms governing the appropriate enactment of any given speech act vary greatly
among cultures and languages”. In the case of the advice column “Lucy Answers”,
the texts are composed in such ways that the face-threatening character of advice is
toned down. The mission of the health program to which the site belongs is of ut-
most importance in this context, since it propagates an ideal of non-directiveness.
The practice observed is characterized by the mitigation of advice, the use of
humor, empathy, and support. Using Standard English, “Lucy” engages in strat-
egies that stress her expertise (e.g., presenting well-researched, in-depth in-
formation, quoting, and referring to information sources), while still remaining ap-
proachable.
Outside the context of professional health counseling, however, many advice
columns are not primarily intended to impart quality advice alone, but are there for
entertainment reasons, as well. In those cases, implicit norms of delicacy may be
suspended. Some columnists certainly have made a name for themselves through
their blunt and non-mitigated commentaries, as evidenced in the popular column
“Savage Love”, run by Dan Savage on http://www.thestranger.com/seattle/savage-
love. The tone of his column is very different from that of the institutionally run
“Lucy Answers”. While also quoting relevant literature, Dan Savage uses less
mitigation than “Lucy Answers”, uses profanity frequently, and can at times be
face threatening towards his advice seekers. More research is needed to investigate
how professional and lay people construct their expert identities in order to project
themselves as trustworthy persons to turn to and/or as entertaining advisors: As
noted above, the age of the target readership and the experience of the advisors as
lay persons or professionals, the topic, and the overall aim and ideology of the col-
umn are likely to cause variation in these strategies.
Further research should also tease out what impact different computer-me-
diated forms have on advice-giving. For example, it would be interesting to com-
pare the set-up of advice contexts with advice threads in forums (e.g., Morrow
2006) and blogs (e.g., Kouper 2010), which allow for more follow-up interaction
on the part of advice-seekers and advice-givers. Such studies could contribute to an
understanding of how people employ language differently in different advisory
settings and how they orient to the particular parameters of the context in which
they find themselves (Leppänen 1998: 120). In this way, the pragmatic study of ad-
vice-giving online can add to an understanding of language use in the “complexity
of its cognitive, social, and cultural (i.e., meaningful) functioning in the lives of
human beings” (Verschueren 2009: 19).
Acknowledgements
The author wishes to thank the three editors of this volume, in particular Susan
Herring, and Andreas Jucker as the responsible general editor for their constructive
criticisms, and Brook Bolander for her perceptive comments.
Notes
1. In this chapter, context is defined as “the degree to which the participants share a com-
mon set of cultural expectations with respect to the social activity and the speech events
making up part of that activity” (Watts 1992: 33); this includes “the degree to which the
participants share a common set of assumptions with respect to the information state …
within which the strip of interaction is developed” (Watts 1992: 51). Interactants thus
take prior knowledge of activity types (which form discourses on a larger level) and the
immediate context of interaction into account when interpreting language.
2. http://www.google.com/search?hl=en&q=advice&btnG=Search&cat=gwd%2FT op
[accessed 09/04/2010]
3. http://www.google.com/ [accessed 09/04/2010]
4. http://www.cancer.gov/cancertopics/wyntk/brain/ [accessed 09/04/2010]
5. This statement is based on the pewinternet.org reports (Madden 2003).
6. Site providers in the professional health domain range from international organizations,
such as the World Health Organization (http://www.who.int/en/), national institutions,
such as the Federal Office of Public Health in Switzerland (http://www.bag.admin.ch/
index.html?lang=en) or the National Institutes of Health in the United States (http://www.
nih.gov/), to many institutional organizations in the field of public health more generally.
Examples of the latter are the many official sites of the States/cantons on public health
issues, or organizations dedicated to particular conditions (e.g., http://www.aids.ch/,
www.multiplesklerose.ch/). In addition to these professional sites, many lay sites are dedi-
cated to sharing information and experience about medical conditions as well as support
(e.g., Hamilton 1998, who studied online bone marrow transplantation narratives and
Harrison and Barlow 2009, who analyzed an online arthritis self-management program).
7. In this chapter, the concept of identity is not seen as a fixed given but as “intersubjec-
tively rather than individually produced and interactionally emergent rather than
assigned in an a priori fashion” (Bucholtz and Hall 2005: 587).
8. In rare cases, “Lucy” also offered her own experiences as illustration.
9. The information in parentheses indicates the topic category and the header of the exchange.
10. These numbers add up to 99 % instead of 100 % due to the effect of rounding.
11. The question prompting the response was: “Lucy, If no one in my dorm seems to care if
students come in drunk, creating noise and messes, what can I do?”
12. The question prompting this response was: “Hi Lucy, I wanted to know what I can take
for nausea and what causes it?”
13. An example of a lay advice column is http://ms.liss.tripod.com/UnconventionalWis-
dom/index.html, which also contains a disclaimer.
14. An alternative reading, suggested by Susan Herring, is that the teenagers sought
to index solidarity in their interactions, with the bluntness in the TeenAdvice section
reflecting an absence of social distance.
15. The advice-seekers can, however, make a face-to-face appointment with a health coun-
selor if they belong to the primary target group of the site, i.e., students at the university
that hosts the “Lucy Answers” site.
16. The question prompting this response was: “Dear Lucy, This is sort of embarrassing,
and you touched on it in a previous message, but … how do you decrease penis size? I’m
serious. Is there some sort of surgery? I ask because my girlfriend says mine is too large,
and it sometimes hurts her. Please answer. Signed, Too big”
References
Alexander, Susan M.
2003 Stylish hard bodies: Branded masculinity in Men’s Health magazine. Socio-
logical Perspectives 46(4): 535–554.
Androutsopoulos, Jannis
2006 Introduction: Sociolinguistics and computer-mediated communication. Jour-
nal of Sociolinguistics 10(4): 419–438.
Baym, Nancy K.
1995 The emergence of community in computer-mediated communication. In:
Steven G. Jones (ed.), Cybersociety: Computer-Mediated Communication and
Community, 138–163. Thousand Oaks, CA: Sage.
Baym, Nancy K.
1998 The emergence of on-line community. In: Steven G. Jones (ed.), Cybersociety
2.0: Revisiting Computer-Mediated Communication and Community, 35–68.
Thousand Oaks, CA: Sage.
Bardovi-Harlig, Kathleen and Beverly S. Hartford
1990 Congruence in native and nonnative conversations: Status balance in the aca-
demic advising session. Language Learning 40(4): 467–501.
Bresnahan, Mary
1992 The effects of advisor style on overcoming client resistance in the advising
interview. Discourse Processes 15(2): 229–247.
Brown, Penelope and Stephen C. Levinson
1987 Politeness: Some Universals in Language Usage. Cambridge, UK: Cambridge
University Press.
Bucholtz, Mary and Kira Hall
2005 Identity and interaction: A sociocultural linguistic approach. Discourse
Studies 7(4–5): 584–614.
Collins Concise Dictionary Plus
1989 London/Glasgow: Bath Press.
Currie, Dawn
2001 Dear Abby: Advice pages as a site for the operation of power. Feminist Theory
2(3): 259–281.
DeCapua, Andrea and Joan Findlay Dunham
1993 Strategies in the discourse of advice. Journal of Pragmatics 20(6): 519–
531.
DeCapua, Andrea and Joan Findlay Dunham
2007 The pragmatics of advice giving: Cross-cultural perspectives. Intercultural
Pragmatics 4(3): 319–342.
DeCapua, Andrea and Lisa Huber
1995 “If I were you …”: Advice in American English. Multilingua 14: 117–132.
Erickson, Frederick and Jeffrey Shultz
1982 The Counselor as Gatekeeper: Social Interaction in Interviews. New York:
Academic Press.
Feng, Bo and Erina L. MacGeorge
2006 Predicting receptiveness to advice: Characteristics of the problem, the advice-
giver, and the recipient. Southern Communication Journal 71(1): 67–85.
Fleischhacker, Andrea M.
1987 Zeitschriftenberatung in “Lebensfragen”. Textanalyse unter pragmatischen
Gesichtspunkten. Unpublished Magisterarbeit, Universität Münster.
Franke, Wilhelm
1997 Massenmediale Aufklärung eine sprachwissenschaftliche Untersuchung zu
ratgebenden Beiträgen von elektronischen und Printmedien. Bern: Lang.
Gaik, Frank
1992 Radio talk-show therapy and the pragmatics of possible worlds. In: Alessandro
Duranti and Charles Goodwin (eds.), Rethinking Context: Language as an In-
teractive Phenomenon, 271–290. Cambridge, UK: Cambridge University Press.
Goffman, Erving (ed.)
1967 Interactional Ritual: Essays on Face-to-face Behavior. Garden City, NY: An-
chor Books.
Goldsmith, Daena J. and Erina L. MacGeorge
2000 The impact of politeness and relationship on perceived quality of advice about
a problem. Human Communication Research 26(2): 234–263.
Gough, Val and Mary Talbot
1996 “Guilt over games boys play”: Coherence as a focus for examining the consti-
tution of heterosexual subjectivity on a problem page. In: Carmen Rosa Cal-
das-Coulthard and Malcolm Coulthard (eds.), Texts and Practices: Readings
in Critical Discourse Analysis, 214–230. London: Routledge.
Griffiths, Mark D.
2005 Online therapy for addictive behaviors: Review. Cyberpsychology & Behavior
8(6): 555–561.
Hamilton, Heidi
1998 Reported speech and survivor identity in on-line bone marrow transplantation
narratives. Journal of Sociolinguistics 2: 53–67.
Harvey, Kevin James, Brian Brown, Paul Crawford, Aidan Macfarlane, and Ann McPherson
2007 “Am I normal?” Teenagers, sexual health and the internet. Social Science &
Medicine 65: 771–781.
Harrison, Sandra and Julie Barlow
2009 Politeness strategies and advice-giving in an online arthiritis workshop. Jour-
nal of Politeness Research 5: 93–111.
He, Agnes Weiyun
1994 Withholding academic advice: Institutional context and academic practice.
Discourse Processes 18(3): 297–316.
Hendley, W. Clark
1977 Dear Abby, Miss Lonelyhearts, and the eighteenth century: The origins of the
newspaper advice column. Journal of Popular Culture 11(2): 345–352.
Heritage, John and Sue Sefi
1992 Dilemmas of advice: Aspects of the delivery and reception of advice in inter-
actions between health visitors and first-time mothers. In: Paul Drew and John
Heritage (eds.), Talk at Work: Interaction in Institutional Settings, 359–417.
Cambridge, UK: Cambridge University Press.
Herring, Susan C.
2007 A faceted classification scheme for computer-mediated discourse. Language@
Internet 4, article 1. http://www.languageatinternet.org/articles/2007/761
Hinkel, Eli
1994 Appropriateness of advice as L2 solidarity strategy. RELC Journal 25(2): 71–93.
Hinkel, Eli
1997 Appropriateness of advice: DCT and multiple choice data. Applied Linguistics
18(1): 1–26.
Hudson, Thom
1990 The discourse of advice giving in English – “I wouldn’t feed until spring no
matter what you do”. Language and Communication 10(4): 285–297.
Hutchby, Ian
1995 Aspects of recipient design in expert advice-giving on call-in radio. Discourse
Processes 19(2): 219–238.
Jefferson, Gail and John R. E. Lee
1992 The rejection of advice: Managing the problematic convergence of a “troubles-
telling” and a “service encounter”. In: Paul Drew and John Heritage (eds.),
Talk at Work: Interaction in Institutional Settings, 521–548. Cambridge, UK:
Cambridge University Press.
Kouper, Inna
2010 The pragmatics of peer advice in a LiveJournal community. Language@Inter-
net 7, article 1. http://www.languageatinternet.org/articles/2010/2464
Leppänen, Vesa
1998 The straightforwardness of advice: Advice-giving in interactions between
Swedish district nurses and patients. Research on Language and Social Inter-
action 31(2): 209–239.
Limberg Holger
2011 The Interactional Organization of Academic Talk: Office Hour Consultations.
Amsterdam/Philadelphia: John Benjamins.
Limberg, Holger and Miriam A. Locher (eds.)
2012 Advice in Discourse. Amsterdam/Philadelphia: John Benjamins.
Locher, Miriam A.
2006a Advice Online. Advice-giving in an American Internet Health Column. Am-
sterdam/Philadelphia: John Benjamins.
Locher, Miriam A.
2006b Polite behavior within relational work: The discursive approach to politeness.
Multilingua 25(3): 249–267.
Locher, Miriam A. and Holger Limberg
2012 Introduction to advice in discourse. In: Holger Limberg and Miriam A. Locher
(eds.), Advice in Discourse, 1–27. Amsterdam/Philadelphia: John Benjamins.
Locher, Miriam A. and Sebastian Hoffmann
2006 The emergence of the identity of a fictional expert advice-giver in an American
Internet advice column. Text and Talk 26(1): 67–104.
Locher, Miriam A. and Richard J. Watts
2005 Politeness theory and relational work. Journal of Politeness Research 1(1): 9–33.
Locher, Miriam A. and Richard J. Watts
2008 Relational work and impoliteness: Negotiating norms of linguistic behaviour.
In: Derek Bousfield and Miriam A. Locher (eds.), Impoliteness in Language.
Studies on its Interplay with Power in Theory and Practice, 77–99. Berlin:
Mouton de Gruyter.
Lucy Answers
2004 [for legal reasons, the source cannot be indicated]
Madden, Mary
2003 America’s online pursuits. The changing picture of who’s online and what they
do. Pew Internet & American Life Project. http://www.pewinternet.org/
Reports/2003/Americas-Online-Pursuits.aspx
McRobbie, Angela
1978 Jackie: An Ideology of Adolescent Femininity. Birmingham, UK: The Centre
for Contemporary Cultural Studies.
McRobbie, Angela
1991 Feminism and Youth Culture: From “Jackie” to “Just Seventeen”. Boston,
MA: Unwin Hyman.
McSeveny, Kathy, Karen Grainger, and Kerry Doherty
2006 I Love Weight Watchers: Face management strategies in reports of success on
an online message board. Paper presented at the Appearance Matters 2 confer-
ence, Bath, UK, June.
Miller, John and Kenneth J. Gergen
1998 Life on the line: The therapeutic potentials of computer-mediated conver-
sation. Journal of Marital and Family Therapy 24(2): 189–202.
Mininni, Giuseppe
1991 Diatextual analysis of the advice column. Rassegna Italiana di Linguistica
Applicata 23: 73–81.
Morrow, Phillip R.
2006 Telling about problems and giving advice in an Internet discussion forum:
Some discourse features. Discourse Studies 8(4): 531–548.
Morrow, Phillip R.
2012 Online advice in Japanese: Giving advice in an Internet discussion forum. In:
Holger Limberg and Miriam A. Locher (eds.), Advice in Discourse, 253–279.
Amsterdam/Philadelphia: John Benjamins.
Mutongi, Kenda
2000 “Dear Dolly’s” advice: Representations of youth, courtship, and sexualities in
Africa, 1960–1980. International Journal of African Historical Studies 33(1):
1–24.
Pilnick, Alison
1999 “Patient counseling” by pharmacists: Advice, information, or instruction? So-
ciological Quarterly 40(4): 613–622.
Pilnick, Alison
2001 The interactional organization of pharmacist consultations in a hospital set-
ting: A putative structure. Journal of Pragmatics 33(12): 1927–1945.
Placencia, María Elena
2012 Online peer-to-peer advice in Spanish Yahoo! Respuestas. In: Holger Limberg
and Miriam A. Locher (eds.), Advice in Discourse, 281–305. Amsterdam/Phil-
adelphia: John Benjamins.
Richardson, Kay P.
2003 Health risks on the internet: Establishing credibility on line. Health, Risk and
Society 5(2): 171–184.
Richardson, Kay P.
2005 Internet Discourse and Health Debates. New York: Palgrave, Macmillan.
1. Introduction
Why do people lie? The reasons are as varied as human life itself, but there is al-
most always a reason, and these reasons can be categorized in a variety of ways. St.
Augustine, for example, classified lies into eight types of varying severity. Modern
psychology has produced smaller taxonomies. In the influential self-presentational
framework of deception (DePaulo, Lindsay, et al. 2003; Vrij 2008a), people lie to
enhance or protect their own self-image (self-oriented), their partner’s (partner-
oriented), or some third party (altruistic). According to this view, most everyday
conversational lies are part of an effort to manage interactions with others and
achieve self-presentational goals (Goffman 1959). Other lies, like those associated
with scandals or crime, may be less about self-presentation goals and more about
seeking material rewards or avoiding prosecution, but lies are still a means to ac-
complishing something in the world.
Since lies are useful for accomplishing goals in the world, why not simply lie
all the time? A whole host of factors constrain the use of deception. Some of these
constraints are external, such as the social stigma associated with being caught
lying. Because humans are social animals, it is important to have the trust of other
humans. Getting caught lying can damage that trust (Möllering 2006) and even ul-
timately destroy the ability to communicate with one another (Grice 1989). Other
constraints against lying are internal. Most people view themselves as honest, and
this self-concept of an honest person can shackle how much one is prepared to lie.
In a series of experiments, Ariely and colleagues (Mazar and Ariely 2006) demon-
strated that even when there was no possibility of being caught, participants in
their experiments only lied by a fraction of the full amount they could have lied,
suggesting that the self-concept of honesty prevents people from lying too much.
In this view, the act of lying is a sort of calculus. Lying can be a tool for accom-
plishing things in the world, but it is restricted by internal and external constraints.
The question addressed here is whether communication technologies, like tele-
phones, email, and text messaging, enter into this calculus. And, once the decision
to lie has been made, what are the implications for language?
The present chapter has two primary interests. The first is a concern with how
communication and information technology affect the pragmatic calculus of de-
ception, and what features of these new communication spaces may enter into this
calculus. The focus is on everyday social interactions, which in the present era are
supported primarily by the telephone, email, and various forms of messaging (e.g.,
instant messaging, text messaging via mobile phones), although the chapter also
briefly reviews research examining deception in more anonymous Internet spaces,
including chat rooms and newsgroups. The central argument is that these com-
munication spaces shape the pragmatics of deception by varying communication
features and motivations for deception (Carlson et al. 2004).
The second interest is in the language people use once the decision to lie with
technology has been made. Traditional deception research has examined the verbal
cues of deception (Vrij 2008b), with mixed and sometimes contradictory results.
However, because digital technologies leave traces of language that previously
were evanescent (e.g., the contents of an email or a text message are automatically
recorded), deception research has seen a resurgence of interest in verbal cues. In
combination with the advances in text analysis tools from computer science that
make analyzing and processing textual data extremely fast and efficient, these con-
versational records are providing material for researchers to make new discoveries
about how deception is manifested in language. The second part of this chapter re-
views this research.
In this chapter, digital deception is defined as the intentional control of in-
formation in a technologically-mediated message to create a false belief in the re-
ceiver of the message (Hancock 2007). While this definition is but one of many for
defining deception, it contains two key attributes that are necessary to highlight.
The first is the idea of a false belief, which means that deception leads a conversa-
tional partner to believe something that the speaker does not believe to be true. The
second is the notion of intentionality. False statements must be intentional to be de-
ceptive; otherwise they are simply mistakes.
There are, of course, many ways to deceive, and almost any kind of speech act
can be used deceptively. Indeed, a number of theories and taxonomies lay out how
different speech acts can lead to different types of deception. For example, In-
formation Manipulation Theory (McCornack 1992; McCornack, Levine, Torres, and
Campbell 1992) notes that violations of Grice’s four maxims lead to different types
of deception. The literature on deception and CMC unfortunately has not advanced
to this point, with few exceptions (e.g., Twitchell, Nunamaker, and Burgoon 2004).
Consequently, the chapter limits the analysis and review to the choice of using de-
ception in mediated interactions and the linguistic implications of this decision.
Richness Theory (MRT; Daft and Lengel 1984, 1986), was concerned with how
managers might use different media to accomplish different tasks. The theory
holds that people match the medium to the equivocality of the task they wish to
accomplish. Low equivocality tasks are straightforward and simple, such as ar-
ranging a place and time to meet for lunch. High equivocality tasks are ambiguous
and complex, such as doing a performance review or hiring or firing someone. Ac-
cording to the theory, managers should match low equivocal tasks to lean media
and equivocal tasks to rich media.
Lying is relatively ambiguous and complex given that it is an attempt to mis-
lead, making it an equivocal task (but see Bavelas, Black, Chovil, and Mullett’s
1990 work on equivocality for a more nuanced description of equivocation in com-
munication). Thus, people who have decided to engage in deception should choose
a rich medium over a lean medium, preferring face-to-face to the telephone, and
the telephone to email, for example. This makes sense when one considers that
having a richer medium would allow the liar to track the partner’s responses more
carefully (Buller and Burgoon 1996).
The question of whether people lie more in one medium or another was first di-
rectly tested by psychologist Bella DePaulo and colleagues (DePaulo, Kashy, et al.
1996) in a large diary-based study that had people record all of their lies and social
interactions for seven days. With this method, the researchers counted the number
of times participants reported lying in a conversation and divided that by the total
number of social interactions participants reported. Although the diary method has
a number of limitations, especially the potential for self-report biases, this study
provided the first data that allowed for comparisons of lying rates by the same in-
dividuals across media. Because the study was conducted in the early 1990s, be-
fore email and the web had reached critical mass, the media of interest were limited
to face-to-face, the telephone, and letters (unfortunately there were too few letters
reported to allow for cross-media comparisons).
Of these media, MRT would predict the most lying face-to-face, followed by
the telephone, and lastly letters. However, DePaulo and colleagues had a different
prediction than MRT. They noted that people report feeling temporary discomfort
when telling a lie, presumably because lying is a negative social act and a breech of
trust. They argued that communication technologies that allow for social distance
between the liar and the partner should provide some relief for the discomfort they
experienced when lying.
In fact, this is exactly what DePaulo et al. found when they examined their par-
ticipants’ diaries: People reported lying in a greater proportion of telephone calls
than in face-to-face interactions. While this supported the idea that people lie more
frequently in more socially-distant media, the results were the opposite of the MRT
prediction.
Note that a common assumption of both of these arguments is that media vary
along a single dimension, richness or social distance, which is assumed to affect
Taken together, the components of the Feature-Based Model predict that the
telephone should involve more lies than face-to-face, which in turn should have
more lies than email. Phone conversations are typically not recorded, they are syn-
chronous, and they involve conversation at a distance. Face-to-face conversations
are synchronous and not recorded, but they involve, by definition, communication
in the same physical space. Email is recordable and asynchronous, although not
(necessarily) co-present.
To examine these predictions, we (Hancock, Thom-Santelli, and Ritchie
2004a) replicated the DePaulo et al. (1996) diary study by asking students to record
all of the social interactions and lies over seven days. In almost all respects the pro-
cedure in our study followed that of the DePaulo et al. study, with one important
exception: We had our participants record not only face-to-face interactions and
phone calls, but also emails, instant messaging, chat rooms, multi-user domains
(MUDs) and newsgroups (too few participants recorded any chat room, MUD, or
newsgroup conversations for analysis, so we exclude them here).
Following DePaulo et al., we calculated the lying rate for each medium as the
proportion of lies per number of conversations in that medium for each participant.
Our data replicated the DePaulo et al. finding that phone conversations involved
significantly more lies than face-to-face conversation. The more interesting result,
however, was the lying rate in email. Emails involved fewer lies than any other
medium. This finding contrasted with the idea that social distance should increase
deception. However, the pattern of data fit with the features expected to shape the
pragmatics of deception: recordability, synchronicity, and co-presence.
Obviously, not all lies are the same. In addition to the three orientations of lies
described above (self-oriented, partner-oriented, and altruistic), people can also lie
about different things, and to different people. For instance, in DePaulo et al.’s
(1996) influential deception taxonomy, lies can be about facts, feelings, opinions
and explanations. Also, relationships based on genuine trust should involve fewer
lies than more superficial relationships. Kashy and DePaulo (1998) found this to be
the case: People reported lying less often in conversations with close friends and
romantic partners than with strangers and acquaintances (see also Ennis, Vrij, and
Chance 2008; Whitty and Carville 2008).
In a follow-up study that used the same diary procedure but asked more in-
formation about what and to whom people lie, we (Hancock et al. 2004b) reached
three conclusions. First, the pattern of deception across media, with the most lies
reported on the telephone and the fewest in email, was repeated. Second, tech-
nology affected the content of the lie. Phone conversations involved more action
lies than any other medium, presumably because the phone (at the time of the
study) was the most mobile technology and would allow lies about what people
were doing. Face-to-face conversations had the most lies about one’s feelings,
presumably because synchronous face-to-face conversations are more likely to
involve questions related to one’s opinions (e.g., “What do you think of my new
shirt?”). Finally, email had the most explanation lies. The asynchronous nature of
email allows more time for constructing a sound explanation than synchronous
formats, and this extra time presumably allows the writer to construct a lie that
can withstand the potential of later scrutiny associated with a highly recordable
medium.
Third, the way technology interacted with deception and relationships was
relatively straightforward. Lies to family and friends tended to take place most fre-
quently via a mediated channel, while lies to strangers and acquaintances occurred
primarily face-to-face. The data are consistent with previous research conducted
only in face-to-face settings, suggesting that lies are more likely to be told to
strangers and acquaintances (Ennis, Vrij, and Chance 2008; Kashy and DePaulo
1998). These data add the observation that when people lie to people close to them
they tend to use technology to communicate the deception. Finally, one last inter-
esting effect was that although students did not tell email lies often, the most com-
mon email lie was an explanation lie (e.g., “I couldn’t finish my paper because my
printer died”) to their professors.
The third difference is a statistical one. In the diary studies, lying frequency is
reported as a proportion of the total social interactions, rather than the absolute fre-
quency of lies that take place in a medium, to control for differences in how often
people use different media. To account for these differences across media, as well
as across speakers (some people have many interactions, while others have very
few), the rate of deception is calculated as a function of total social interactions. In
the George and Carlson procedure, only the absolute frequency is available.
results suggested that relationships in online spaces tended to be less involved than
those in physical space.
Taken together, these data suggest that anonymous spaces can reduce involve-
ment in romantic relationships, and that this can increase lying. A good deal of re-
search, however, has indicated that interpersonal relationships can be as intimate
and involved as face-to-face relationships (see Parks and Floyd 1996; Whitty and
Carr 2006). If this is the case, then at least in romantic relationships, the nature of
the relationship may be more important than the anonymous nature of the com-
munication space. This would be similar to face-to-face research that suggests
people are less likely to lie to closer intimates (Kashy and DePaulo 1998).
However, there are many other types of deception in anonymous spaces than
misrepresenting oneself to a romantic partner. In one of the first treatments of de-
ception online, Donath (1999) described several different types of deception, in-
cluding identity concealment, category deception, trolls, and impersonation. Here
we focus on gender switching, when people present a gender online that is different
from their offline gender (Herring 2003).
Gender switching is of particular importance for this chapter because of its lin-
guistic implications. For instance, several studies (reviewed in the next section)
have examined whether people retain characteristics of their gendered language
when acting as the other gender (Herring and Martinson 2004). Here, the question
is whether (and to what extent) men and women take advantage of the lack of vis-
ual and vocal cues in text-based CMC to perform gender deception (Danet 1998).
In self-reports, this type of deception appears to be quite common. For example,
Whitty (2002) found that about one-quarter of her male respondents reported pre-
tending to be a female at some point in public chat rooms. However, as with the
everyday lies described in the previous section, the self-report data and observa-
tional data do not necessarily line up. In her review, Herring (2003) reports on her
examination of Internet Relay Chat (IRC) channels and her observation that gender
deception was not common in this anonymous space. Others have also observed
that gender switching online is relatively rare, perhaps because pretending to be
someone one is not is difficult to maintain over time (Roberts and Parks 1999).
Once the pragmatic decision is made to lie with communication technology, in par-
ticular in text-based CMC (e.g., email, instant messaging, chat rooms, newsgroups,
text messaging, etc.), communicators must make linguistic choices about how to
express their deception. The following section focuses on this issue, laying out re-
cent trends in the research on verbal cues of deception.
ized by more sentences with more words (including verbs and noun phrases). This
finding was contrary to face-to-face studies that show that liars use fewer words
when lying. One reason for this may be the asynchronous nature of the interaction:
Online liars can take their time to craft the lies. Liars also produced less lexical and
content diversity (e.g., a lower type-token ratio), suggesting that their messages
were less semantically complex and less detailed than those of truth tellers. These
findings were consistent with theories that suggest lies tend to be impoverished
relative to truthful accounts. The liars also used more emotional language than truth
tellers, including more positive and negative affect words (e.g., sad, delighted).
Finally, with respect to pronoun usage, liars used more third-person references
than truth tellers, but no differences were found for self-reference.
More recently, Hancock et al. (2008) examined the linguistic patterns of decep-
tion in synchronous CMC. Participants interacted over instant messaging and dis-
cussed four topics. On two of the topics, one of the participants was asked to lie to
his or her partner but to tell the truth on the other topics. Their partners were asked
to simply engage in the conversation and were unaware of the deception manipu-
lation.
The hypotheses in this study were derived from those laid out by Zhou et al.
(2003) described above and from the Newman-Pennebaker (NP) model (Newman,
Pennebaker, Berry, and Richards 2003), an empirically-derived model of deception
that predicts several language features that change relative to truthful language.
According to the NP model, lies should involve fewer first person singular pro-
nouns (“I”), as liars try to distance themselves psychologically from their lie. Sec-
ond, lies should involve fewer exclusive words (“except”, “but”), as lies tend to be
less complex than truthful statements. Third, lies should include more negative
emotion terms (e.g., “sad”, “anxious”, “afraid”), which reflect the guilt and anxiety
related to being deceptive. Last, lies should contain more motion verbs (e.g., “go”,
“fly”), which help move the story along and distract the listener.
In general, the data were supportive of the NP model’s main predictions. The
results revealed that liars, like those in Zhou et al.’s study, produced more words
when lying then when telling the truth, suggesting that CMC liars may use more
words than face-to-face liars to accomplish their deception. Unlike the Zhou et al.
study, liars used fewer self-referencing pronouns and more other-oriented pro-
nouns, consistent with the idea of psychological distancing. Finally, liars tended to
use more causal terms (e.g., “because”, “therefore”, etc.) when lying compared to
telling the truth, suggesting that they were potentially constructing more coherent
and plausible stories.
Unlike most deception studies, Hancock and colleagues also examined the lan-
guage of the partner. Even though the partner was blind to the deception manipu-
lation, the partner’s language differed across deceptive and truthful topics. The dif-
ferences often followed the senders. For instance, both the liar and the partner
increased their use of first person pronouns during deception. This kind of lin-
guistic style matching occurred more frequently during deceptive parts of the con-
versation than during the truthful parts. A number of theories describe why and
how interlocutors in face-to-face communication tend to synchronize their lan-
guage, from syntactic usage (Pickering and Garrod 2004) to coordinating referents
(Brennan and Clark 1995) (for a review, see Burgoon, Stern, and Dillman 1995).
The partners, however, were unable to use this information to improve their decep-
tion detection accuracy: Partners performed at chance when trying to detect the
lies. These data suggest that, while these changes in language patterns did not help
the partner in detecting a lie, the partner’s language patterns may provide some
clues as to whether they are being lied to or not.
While the above-mentioned studies took place in the laboratory, Keila and Skil-
licorn (2005) conducted an analysis of deception in the real world, namely the
Enron email corpus provided by the Enron trial (the corpus contains approximately
500,000 emails relating to a scandal at a large corporation), in which executives
were sued for deliberately misleading investors. The Enron corpus consists of ap-
proximately 500,000 emails, and the researchers applied a subset of the NP model
to the data in an attempt to classify deceptive emails. The dimensions employed
were first person singular, negations, and exclusive words. Using these three di-
mensions, along with statistical techniques for classifying messages along the di-
mensions, Keila and Skillicorn successfully identified the most problematic (i.e.,
deceptive) emails with respect to the court case.
Taken together, a few trends emerge across these three studies. The first is that
when people lie in CMC, they tend to use more words than when they are telling
the truth, a finding that is the opposite of research on face-to-face deception. As
noted, this may be due to the written nature of CMC, which provides more time and
allows for editing. A second trend is changes in pronoun usage. Relative to telling
the truth, lies include fewer self-references but more other-references. Emotional
terms also seem to be different across lies and truths, with more emotional terms,
especially negative ones, observed during deception. Exclusive words (e.g., “ex-
cept”, “however”), which may indicate the complexity of the grammatical struc-
ture, tend to decrease during CMC deception. Last, the linguistic style-matching
data from Hancock et al. (2008) suggest that the partner’s language should also be
considered.
is whether these differences persist when speakers are being deceptive about their
gender. Since gender cues are presumably given off unconsciously (Goffman
1959), they should be difficult for speakers to control.
Several studies have looked at the detection of gendered language (Savicki,
Kelley, and Oesterreich 1999; Thomson, Murachver, and Green 2001). When
Thomson and Murachver presented students with emails and asked them to ident-
ify the gender of the authors, the students were relatively successful in distinguish-
ing female emails from male ones. The authors concluded that a combination of
linguistic features led to the decisions, rather than any one feature alone.
A more recent study examined how people performed gender online when
being deceptive about their actual (offline) gender (Herring and Martinson 2004).
In this study, the authors examined the transcripts from a gender-switching game,
in which some participants were deceptive about their gender and others were not.
The game took place in a synchronous chat space. The researchers found that
participants produced stereotypical content when performing the opposite gender,
but they also gave off stylistic cues to their actual gender. That is, when females
played males they talked about overtly stereotypically male topics or content (e.g.,
cars), while males playing females talked about stereotypically female content
(e.g., shopping). Nonetheless, there were some stylistic cues associated with their
gender that participants were unable to control, such as gendered differences in
pronoun usage, hedges, and rhetorical questions. Judges in the game only guessed
participants’ gender correctly approximately half the time, in part, perhaps, because
they based their judgments primarily on the stereotypical content rather than on
language use.
and Gulikers 1995). This seemingly trivial percentage (less than 0.04 %) of vo-
cabulary accounts for over half of the words in daily speech. Despite the frequency
of their use, function words are the hardest to master when learning a new lan-
guage. Indeed, they are virtually invisible in daily reading and speech (Chung and
Pennebaker 2007).
Recent research shows that the analysis of function words, such as articles,
auxiliary verbs, prepositions, and pronouns, provides important clues about the so-
cial relationship between a speaker and the audience, the psychological state of the
speaker, and information about the speaker’s social role and status in the commu-
nity (Chung and Pennebaker 2007; Herring and Paolillo 2006). Function words
may also be important in the language of deception in CMC, whether it be lies
about gender or some other topic.
cally, the lies generated in most lab studies are mocked up, and liars have little
motivation to succeed in their lies.
Some non-CMC research has begun to examine deception in naturalistic set-
tings. For example, in a project examining deception in political speech, Marko-
witz, Hancock, and Bazarova (2011), compared false statements (e.g., claims that
Iraq had weapons of mass destruction or direct links to al Qaeda) and non-false
statements (e.g., that Saddam Hussein had used gas on his own people) produced
by officials in the Bush administration in the run-up to the 2003 Iraq war. Consist-
ent with the NP model of deception’s predictions, false statements contained sub-
stantially reduced rates of first-person singular (“I”) and exclusive terms (“ex-
cept”, “but”) but more negative emotion terms and action verbs. Using this simple
model, we were able to classify approximately 76 % of the Bush administration’s
statements correctly as either false or not false.
CMC deception seems to be a good fit for naturalistic studies. Like the Enron
analysis by Keila and Skillicorn (2005), future research needs to identify archival
cases in which the ground truth of deception has been established, and the language
records have been preserved. More of this kind of research should lead to major ad-
vances in understanding the pragmatic and linguistic implications of deception in
CMC.
4. Conclusion
The first of this chapter’s central questions was concerned with whether, and if so,
how, technology enters into the calculus of deception. The studies reviewed sug-
gest that the communication spaces created by various technologies can shape
speakers’ decisions to use deception as a pragmatic tool in accomplishing their ob-
jectives. This seems to be the case in both everyday interpersonal interactions, with
technologies such as email, instant messaging, and the telephone, and in more
anonymous online spaces, such as chat rooms and news groups.
The relationship between deception and technology is clearly not simple, but in
our view, it is systematic. Equally importantly, many more questions remain to be
answered as technologies evolve. For example, new technologies, such as the avat-
ars in graphically-based virtual environments, may make deceptions more sophis-
ticated. As Donath (2007: 53) asks, “as behavioral software becomes more soph-
isticated, are we creating avatars that will be increasingly attractive and seemingly
friendly but are in fact the ideal mask behind which a dishonest or manipulative
person can operate?” Indeed, a recent study suggests that people feel more com-
fortable lying when using an avatar than when lying in text-based communication
(Galanxhi and Fui-Hoon Nah 2007).
Our second question was concerned with the linguistic traces of deception in
CMC. Although this line of research has just begun, it suggests some potential in
being able to identify patterns of deception in language, with liars online using
more words, less self-referencing and more other-referencing language, and pro-
ducing less grammatically complex messages. Whether these patterns stand up to
more ecologically valid and diverse tests remains to be answered. Nonetheless, the
fact that stylistic features (such as function words) tend to be critical in both inter-
personal deception and gender-based deception suggests that this class of words
may be especially important.
This final point has relevance for the question of whether to lie online or not.
Currently, there are no reliable methods for identifying deception in CMC, so this
kind of issue does not enter into the question of whether one should lie online.
However, given the potential to identify deceptive patterns, and the fact that much
of what people say now on the Internet will be around for many years, in the future
it may be necessary to consider even more carefully the question of whether or not
to lie.
References
Hancock, Jeffrey T.
2007 Digital deception: When, where and how people lie online. In: Katelyn
McKenna, Tom Postmes, Ulf-Dietrich Reips, and Adam N. Joinson (eds.), Ox-
ford Handbook of Internet Psychology, 287–301. Oxford, UK: Oxford Univer-
sity Press.
Hancock, Jeffrey T., Lauren E. Curry, Saurabh Goorha, and Michael T. Woodworth
2008 On lying and being lied to: A linguistic analysis of deception. Discourse Pro-
cesses 45(1): 1–23.
Hancock, Jeffrey T., Jennifer Thom-Santelli, and Thompson Ritchie
2004a Deception and design: The impact of communication technologies on lying be-
havior. In: Proceedings of the 2004 ACM Conference on Human Factors in
Computing Systems 6, 130–136. New York: ACM.
Hancock, Jeffrey T., Jennifer Thom-Santelli, and Thompson Ritchie
2004b What lies beneath: The effect of communication medium on deception produc-
tion. Paper presented at the Annual Meeting of the Society for Text and Dis-
course, Chicago, IL.
Herring, Susan C.
1993 Gender and democracy in computer-mediated communication. Electronic
Journal of Communication 3(2).
http://www.cios.org/EJCPUBLIC/003/2/00328.HTML
Herring, Susan C.
2001 Computer-mediated discourse. In: Deborah Schiffrin, Deborah Tannen, and
Heidi Hamilton (eds.), The Handbook of Discourse Analysis, 612–634. Ox-
ford, UK: Blackwell Publishers.
Herring, Susan C.
2003 Gender and power in online communication. In: Janet Holmes and Miriam
Meyerhoff (eds.), The Handbook of Language and Gender, 202–228. Oxford,
UK: Blackwell.
Herring, Susan C.
2007 A faceted classification scheme for computer-mediated discourse. Lan-
guage@Internet 4, article 1.
http://www.languageatinternet.org/articles/2007/761
Herring, Susan C. and Anna Martinson
2004 Assessing gender authenticity in computer mediated language use: Evidence
from an identity game. Journal of Language and Social Psychology 23(4):
424–446.
Herring, Susan C. and John C. Paolillo
2006 Gender and genre variation in weblogs. Journal of Sociolinguistics 10(4):
439–459.
Johnson, Marcia K. and Carol L. Raye
1981 Reality monitoring. Psychological Review 88(1): 67–85.
Kashy, Deborah A. and Bella M. DePaulo
1998 Who lies? Journal of Personality and Social Psychology 70: 1037–1051.
Keila, Parambir S. and David B. Skillicorn
2005 Detecting unusual and deceptive communication in email. IBM Centers for
Advanced Studies Conference, Richmond Hill, Ontario, Canada, October
17–20.
Köhnken, Giinter
1996 Social psychology and the law. In: Gün R. Semin and Klaus Fiedler (eds.), Ap-
plied Social Psychology, 257–282: London: Sage.
Markowitz, David, Hancock, Jeffrey T., and Bazarova, Natalya N.
2011 The language of presidential lies: How words reflect lies about war, personal
scandal and state secrets. Paper presented at the 97th Annual Meeting of the
National Communication Association, New Orleans.
Masip, Jaume, Siegfried L. Sporer, Eugenio Garrido, and Carmen Herrero
2005 The detection of deception with the Reality Monitoring approach: A review of
the empirical evidence. Psychology, Crime, and Law 11: 99–122.
Mazar, Nina and Dan Ariely
2006 Dishonesty in everyday life and its policy implications. Journal of Public Pol-
icy and Marketing 25(1): 1–21.
McCornack, Steven A.
1992 Information Manipulation Theory. Communication Monographs 59(1): 1–16.
McCornack, Steven A., Timothy R. Levine, H. I. Torres, and D. M. Campbell
1992 When the alteration of information is viewed as deception: An empirical test of
information manipulation theory. Communication Monographs 59(1): 17–29.
Möllering, Guido
2006 Trust: Reason, Routine, Reflexivity. Oxford, UK: Elsevier.
Newman, Matthew L., James W. Pennebaker, Diane S. Berry, and Jane M. Richards
2003 Lying words: Predicting deception from linguistic style. Personality and So-
cial Psychology Bulletin 29(5): 665–675.
Parks, Malcolm R. and Kory Floyd
1996 Making friends in cyberspace. Journal of Communication 1(4): 80–96.
Pickering, Martin J. and Simon Garrod
2004 Toward a mechanistic psychology of dialogue. Behavioral and Brain Sciences
27(2): 169–225.
Roberts, Lynne D. and Malcolm R. Parks
1999 The social geography of gender-switching in virtual environments on the In-
ternet. Information, Communication, and Society 2(4): 521–540.
Savicki, Victor, Merle Kelley, and E. Oesterreich
1999 Judgments of gender in computer mediated communication. Computers in
Human Behavior 15(2): 185–194.
Selwyn, Neil
2008 A safe haven for misbehaving? An investigation of online misbehavior among
university students. Social Science Computer Review 26(4): 446–465.
Sproull, Lee and Sara Kiesler
1986 Reducing social context cues: Electronic mail in organizational communi-
cation. Management Science 32(11): 1492–1512.
Tausczik, Yla and James W. Pennebaker
2010 The psychological meaning of words: LIWC and computerized text analysis
methods. Journal of Language and Social Psychology 29(1): 24–54.
Thomson, Rob, Tamar Murachver, and James Green
2001 Where is the gender in gendered language? Psychological Science 12(2):
171–175.
1. Introduction
The Internet, it sometimes seems, is a natural environment for all kinds of decep-
tion. It can be a daunting task even for skilled, experienced users to differentiate
between the sincere and the countless incarnations of trickery, pranks, scams, and
frauds in online communication. Whether the fears about lowered sincerity stan-
dards on the Internet are justified or just a modern form of cultural pessimism, at
times it appears as though Harry Frankfurt’s concept of “bullshit” (Frankfurt 2005)
has come into its own in the digital domain.
Indeed, deceptive behavior is present in virtually all areas and socio-technical
modes of the Internet. Online communities afford the “pseudonymity” of faked
user identities (Heisler and Crabill 2006); online journalism struggles with relia-
bility and quality standards (Borden and Harvey 1998; Hayes, Singer, and Ceppos
2007); teachers and scientific editors lament “mouse-click plagiarism” (e.g., Auer
and Krupar 2001); and even Web 2.0, hailed as a more participatory and open com-
municative environment, has generated its own deceptive phenomena in the form
of “splogs” and “flogs” (spam and fake weblogs; see Kolari, Java, and Finin 2006).
Perhaps most centrally, since the first spamming incident – usually attributed to
Gary Thuerk and his promotional email to 600 ARPANet members in 1978 (see
Hedley 2006 for an overview) – Internet users have become acquainted with spam,
viruses, phishing, and Nigerian scam emails, along with many other genres of de-
ceptive communication that appear in their inboxes.
While digital deception can be visual, as the popular pastime of “photoshop-
ping” pictures strikingly demonstrates (Frank 2004), deceptive behavior online is
first and foremost discourse-based; it is thus a valid and central issue in the study of
computer-mediated communication (CMC). Specifically, deception lends itself
well to description within the framework of pragmatic theory. This chapter thus
has a two-part structure. First, it provides an overview of deceptive behavior in
CMC, describing its most typical instantiations, identifying previous case studies,
and outlining the theoretical pillars of the analysis of non-sincere messages that
will be presented in what follows. The second part of the chapter illustrates how
such approaches can be implemented in actual analysis, through a case study of
a particularly striking deceptive genre of CMC, namely email hoaxing. Email
hoaxes (henceforth EHs), a subgenre of digital folklore, are deceptive messages
that are spread by well-meaning users in digital social networks through the for-
warding function of email programs; examples and a more detailed description are
given below. The analysis, based on a corpus of 150 texts, is grounded in the
Gricean cooperative framework and speech act theory; it highlights the pragmatic
duality of EHs between a cooperative and a non-cooperative status and their de-
pendence for success on uptake patterns in their receivers.
the most established being Relevance Theory (Sperber and Wilson 1986) and clas-
sic politeness theory (Brown and Levinson 1987) – are also strongly influenced by
the cooperative paradigm. A major axis for distinguishing different qualities of CP
violations is the degree to which the speaker is aware of his or her deviant beha-
vior: A spectrum exists from blatant lying to semi-conscious self-deception and the
naïve utterances of irrational speakers (see Heyd 2006). The Gricean model has
also been applied successfully to neighboring phenomena such as humor (e.g., At-
tardo 1993).
Speech act theory is the second major pillar in the pragmatic analysis of decep-
tion. Based on Austin’s classic essay How to Do Things with Words (Austin 1962)
and Searle’s first major study (Searle 1969), speech act theory can be described
very broadly as the analysis of discourse in performance. The main focus has tra-
ditionally been on “illocutionary acts” with relatively well-defined speaker inten-
tions, such as promises, threats, and oaths, and taxonomies of speech act types
have been put forward, such as Bach and Harnish’s (1979) classification into con-
statives, commissives, directives, and acknowledgments. Many analyses focus on
indirect speech acts as they are used to achieve more complex communicative
goals (in particular conveying nonliteral meaning). Deception as a type of speech-
acting has been studied both in applied contexts (e.g., Lee 2000 on the develop-
ment of lying in children) and in more theoretical frameworks (in particular Reboul
1994). Reboul (1994) has emphasized the limitations of speech act theory in ana-
lyzing deception by identifying a clash between illocution (assertion) and perlocu-
tion (lying) that cannot easily be resolved. This critical perspective points to a
more general problem: No comprehensive pragmatic theory of non-cooperation (in
all its instantiations) exists to date. A pragmatic model of deception, possibly
based on evolutionary approaches to communicative behavior such as that pro-
posed by Dunbar (1996), would be an important goal for future studies in prag-
matic theory. For the time being, the conceptual frameworks outlined above can
yield valuable insights both for traditional and digital genres.
Several instantiations of digital deception have already been touched upon in the
introduction. Of course, lying can occur in any form of CMC at any given moment,
just as in non-digital forms of discourse. However, a number of genres can be
loosely grouped together whose central distinctive feature arguably is a certain in-
built insincerity. Some of the better-known genres of deceptive CMC, most of them
forms of email and other modes of asynchronous CMC, are discussed briefly below.
As an initial categorization, deceptive CMC genres can be differentiated along
the axis of “profit” vs. “fun”. Two important phenomena in the first category are
spam (unsolicited bulk email) and phishing (a term derived from hacker jargon for
password theft). The deceptive element in spam is usually based on false product
information, and in phishing, sensitive data such as passwords or credit card details
are elicited through the simulation of a trustworthy identity (e.g., the appropriation
of bank names or popular online services such as Ebay or PayPal). Since both phe-
nomena can have serious financial effects and are a threat to online security, it is
not surprising that many studies focus on means for the automatic detection and fil-
tering of such messages (see, for just two examples among many, Dhamija, Tygar,
and Hearst 2006 for phishing, and Oda and White 2005 for spam). From a more
theoretical perspective, Barron (2006) and Schmückle and Chi (2004) have pub-
lished genre studies of spam.
Another widespread form of fraudulent for-profit communication is the “Ni-
gerian” or “419” email (advance fee fraud). In this form of deceptive email, users
are targeted with false stories – typically elaborated narratives involving deceased
millionaires, usurped dictators, and other sensational elements – promising finan-
cial gain in exchange for a small advance payment. Nigerian emails have been ana-
lyzed in terms of their narrative scope (Cukier, Nesselroth, and Cody 2007), as well
as their writers’ grassroots literacy (Blommaert and Omoniyi 2006); for a dis-
cussion of the concept of “authenticity” in Nigerian emails, see the chapter by Gill
in this volume.
At the other end of the scale, genres exist which cause no financial harm to the
receiver (or gain for the sender) but which may have more complex psychosocial
motivations, such as to demonstrate digital literacy; one of these is email hoaxing.
EHs are particularly interesting from a pragmatic perspective because the com-
municants who forward them usually treat them as trustworthy, non-deceptive
messages. Apart from the framework outlined in this chapter (see Heyd 2008 for
more details), EHs have been studied as digital counterparts of oral traditions such
as urban legends (Fernback 2003) and from a social science perspective in terms of
users’ attitudes toward such messages (Kibby 2005).
Trolling is another socio-technical phenomenon frequently found in online
communities. Trolling is the strategy of provoking other users through blatantly
false, exaggerated, or misleading utterances. It closely resembles email hoaxing
in its psychosocial design, as it strives to elicit responses particularly from inex-
perienced users who engage with the message as if it were bona fide. Discussions
of trolling have been put forward by Donath (1999) and Herring, Job-Sluder,
Scheckler, and Barab (2002).
Finally, websites constitute a more static socio-technical mode that has been
appropriated for deceptive uses. Many fake websites exploit the semiotic power of
URLs by using an evocative lexical item (such as a brand, a proper name, or an-
other well-known cultural artifact) to simulate credibility and sincerity, a tactic
known as “cybersquatting” (see Heyd 2008: 15). Fake websites cover the entire
spectrum of fraud vs. fun: While they are increasingly used as a platform for phish-
ing, they have also been a vehicle for more jocular hoaxes whose primary goal is to
fool the user. As noted in the introduction, Web 2.0 successors of traditional web-
sites such as blogs and social network sites have already seen the rise of mode-spe-
cific deception. These more recent developments indicate that deception is a con-
stant of CMC.
The text above is a typical instance of an email hoax – a deceptive CMC genre
that has been the daily fare of inboxes around the world ever since the early 1990s
(see Boese 2003 and Heyd 2008 for overviews). The specimen from example 1 ap-
peared within days of the September 11, 2001 attacks in the United States and has
been circulating ever since with translations into Spanish and Hungarian, among
other languages. In general, EHs follow a somewhat paradoxical pattern: They
have a high proliferation and spread with remarkable speed, yet they are communi-
cated exclusively from private user to private user; they all contain warnings,
promises, threats, and supposedly relevant information, yet every EH is based on
information that is, at its core, deceptive or at least problematic. It should be clear
from this initial description that EHs constitute complex and intricate speech acts.
To be sure, hoaxing did not originate in the digital medium; a variety of offline
phenomena preceded it, some of which has attracted research activity. There have
been literary hoaxes, such as MacPherson’s Ossian; scientific hoaxes, such as the
fakes by Mark Twain and others; and, more recently, the Sokal hoax that shook the
foundations of poststructuralist academia (Sokal 1996).1 Definitions of hoaxing
have included a “deliberately concocted untruth to masquerade as truth” (Mac-
Dougall 1958: vi) and “a deliberately deceptive act that has succeeded in capturing
the attention (and, ideally, the imagination) of the public” (Boese 2003: 2). It is no-
table that both these descriptions, while not based in a linguistic tradition, hinge on
concepts of pragmatics and discourse analysis. Yet, research on the hoax as a lin-
guistic and pragmatic entity is only just beginning to emerge, both in its pre-digital
(Walsh 2007) and digital (Heyd 2008) instantiations. The following overview out-
lines the distinctive features of email hoaxing as a CMC genre and presents textual
evidence.
of multiple addressing is the header information, where multiple receivers are evi-
dent from long address lists in the “to” and/or “cc” fields. Beyond this purely for-
mal aspect, hoax emails often display an openly encoded conceptual multiplicity
that is most noticeable in openings and closings (example 2).
Dear ALL
Hello Disney fans,
All mobile users pay attention.
range from the plausible (deletion of files, reproduction through the mail pro-
gram’s address book) to the fantastic (destruction of hardware, installation of mal-
ware through plaintext messages). Very often, they contain references to software
or Internet security companies, which are intended to act as credibility markers
(“this information was announced yesterday morning from Microsoft”). Ferbrache
(1992: 17–18) describes the “2400 baud modem virus” EH, which dates as far back
as October 1988. The virus EH, then, is not only the oldest form of email hoaxing;
it is also firmly anchored in the medium it comments upon – it is a meta-medial
statement.
Hi. I am, Michele Cordova, the founder of Bath & Body Works and
I want your business. We are trying a new advertising campaign
through the power of YOU, the consumer. In order for this to
work, you need to send this e-mail to 13 people and I know that
is not a lucky number but that is the number we need in order
for this to work.
Our computer tracking system will keep count of how many
people you send it to so don’t feel like you have to send it to
thirteen people all at once. You may not send it to the same
person more than once unless you internet pals accidently de-
lete the message, we wouldn’t want them to miss out on this
great offer. To compensate you for your hard work, we are going
to send you a $50 dollar gift certificate redeemable in any
store nationwide. This is not a joke, it will be your loss if
you don’t send this to 13 PEOPLE. Thanks again!
Michele Cordova
Founder of Bath & Body Works
So-called giveaway EHs are probably the most homogeneous subgroup, in that
there is very little variance in structure and scope among the different items. The
storyline typically runs like this: A major company (ranging from software and
mobile phone producers to brands of beverage) has started an email initiative
(sometimes for marketing, sometimes for market research reasons). If the user for-
wards the email a certain number of times, or if the message reaches a certain dis-
semination, the participants will be rewarded with money or free goods. The users’
communicative behavior, and the success of the initiative, are monitored by a so-
called “email tracking” system. Needless to say, such claims have no factual basis
for several reasons, the most important being that such a monitoring system was,
and remains, pure fiction. The earliest case of a giveaway EH is probably the 1999
“Outlook Beta Test” which claims to be written by Bill Gates himself. A later vari-
ant, dating from 2003, contains the most exclusive offer: a free case of Veuve Clic-
quot. Giveaway EHs are the most easily analyzable subgroup in terms of com-
municant motivation and behavior: If the deception is believed, the communicant
will perceive a material interest for both himself and the other participants in the
communication chain.
This type is the most precarious ethically, as it plays a prank based on human
suffering that is all too real in many other cases. Such emails usually tell the tale of
a terminally ill or missing child. While most charity EHs are purely fictitious, some
texts do have a factual basis, since real search alerts or charitable projects are oc-
casionally posted on the Internet. However, most of these factual cases could be
subsumed as “problematic” information: Since precise search alerts require a deic-
tic anchoring which is simply not given in the digital medium (statements such as
“went missing last week” are frequent), such messages continue to circulate long
after they have become outdated. Perhaps the most famous case, whose origins
reach back to the pre-digital era, is that of Craig Shergold, a 7-year old American
boy with a brain tumor – while the person in question is real, the boy’s appeal for
breaking the Guinness book record in receiving postcards dates from 1989. What is
fascinating about this case from a genre perspective is that the tale easily made the
transition into the digital medium – a true case of transmedial stability (Zitzen and
Stein 2004).
These emails again put message proliferation above all else and almost never
confront the receiver with material or physical demands (such as donations): All
they ask for is that the message be disseminated, either as the victim’s last will or as
a fundraising system (via the famous “email tracker”) – or, in the case of missing
persons, as an alert for the public. This motivational construct illustrates poign-
antly why Döring (2003: 281) classifies EHs as prosocial behavior: They allow the
communicant to engage in (what the communicant believes is) an altruistic act
without the danger of face loss or physical or material commitment.
Urban legends are tales about spectacular and often sensationalist incidents.
Crucially, the events in question always take place in an everyday setting and
happen to everyday people – a narrative strategy which makes identification easy
for the tale’s addressees. A considerable number of urban legends are true classics,
in that they have been circulating for years or decades (see Brunvand 1981 for a
classic collection of such items, Boese 2003 for a more recent perspective). This in-
cludes tales such as the kidney theft or needles in cinema seats; the ease with which
these items made the transition across media may be an indicator of their import-
ance as communicative and psychosocial tools. Other tales have made their ascent
in and through the digital medium. Two dates have been powerful generators of
myths and fears: The advent of the new millennium, with its apocalyptical associ-
ations, fed many fears, such as the W2K bug, and engendered a number of EHs.
Unsurprisingly, the 9/11 attacks are at the center of an entire maze of myths, alle-
gations, and conspiracy theories that have found their way into a number of EHs.
This typology shows only the core themes of email hoaxing. Other forms of
digital folklore, such as online petitions and “good luck” chain letters, share numer-
ous characteristics with EHs (in particular open plurality and the drive for prolif-
eration). However, the element of deception is less clear-cut in some forms of digi-
tal folklore and nonexistent in others; for this reason, the present analysis has been
restricted to EHs in the narrowest sense.
The claim was disproved within weeks by the major security software com-
panies and other reliable sources; relevant information on the utterance’s truth
status is therefore available online for every email user. Statements that are so bla-
tantly deceptive clearly violate Grice’s Cooperative Principle (CP), most notably
the maxims of quality (by making non-truthful assertions) and quantity (by omit-
ting relevant information) (Grice 1975) – or in a relevance theoretic framework,
the principle of relevance (Sperber and Wilson 1995: 158). Therefore, EHs are in-
herently non-cooperative.
EHs are cooperative: Since EHs are forwarded by their senders with the best of
intentions, they are a prime example of cooperative behavior. The sheer act of for-
warding the message bears evidence of the communicants’ candor. Moreover,
many EH texts contain meta-communicative statements that comment on the mess-
age’s supposed truthfulness. For example, the WTC Survivor EH contains this
closing remark:
This is a serious one.
Others speculate on the message’s truth status, thereby cooperating just the same:
I don’t know how true the following is, I am forwarding it just
in case!
Grice’s maxim of quality merely specifies two requirements: (1) Do not say
what you believe to be false, and (2) do not say that for which you lack adequate
evidence (Grice 1975: 45–46). The communicative behavior of EH forwarders is
not in conflict with these requirements, as the forwarders themselves have been
duped by the deceptive statements. Therefore, EHs are inherently cooperative.
How can this pragmatic paradox be resolved? It can be argued that EHs may in-
deed be ascribed a dual cooperative status. This surprising ascription can be ex-
plained in terms of the multiple sendership of EHs. As noted earlier, hoaxing can
be described as serial communication. EHs engender, if they are successful, a
chain, or network, of communication: The message is transmitted from one sender
s1 to n addressees; out of this set of addressees, a number of senders s2, 3, … n arises
in turn, and so on. Email hoaxing, then, is neither a typical case of one-to-one in-
terpersonal communication, nor can it properly be described as one-to-n mass
communication. Within this communicative network, however, two types of
sender have to be distinguished that are qualitatively quite distinct. The first is s1,
the originator of the message and starting point of the communication chain. Little
is known at present about the genesis of EHs or the motivations of their creators.
Yet it can be speculated that messages that evolve into deceptive utterances with-
out a communicant’s conscious effort, such as certain charity EHs described in the
content-based typology, are in the minority. Most EHs are probably created with
full deceptive intention by communicants whose motivations may be not that dif-
ferent from that of trollers (see above) or hackers (see, e.g., Hruska 1990: chapter
5).3 It is these communicative agents who are responsible for the creation of an EH
and, in a larger sense, for the non-cooperative potential of the message.
Once an EH leaves this initial stage and enters the chain of communication, the
floor is open to an altogether different type of communicant. The forwarders s2, 3, … n
of EHs at this second stage presumably consider the EH a bona fide warning. In
other words, they do not recognize the deception, which is not so surprising. Irony
markers, if present at all, are often quite subtle, and communicants generally have
high expectations about the reliability and cooperativeness of utterances (Grice
1975). To forward an EH, then, is more than just in line with pragmatic rules gov-
erning linguistic behavior: It is usually intended, by its actors, as a downright al-
truistic act.
With the transition from s1 to sn, the initial implicature (violation of the quality
maxim) becomes less apparent. However, its potential remains present in the mess-
age through its iterations. In written contexts, implicatures are arguably more dif-
ficult to spot and process for communicants due to the lack of contextual and non-
verbal cues; at the same time, an implicature’s potential remains activated in a
written utterance even across time and space. This is why EHs may be perceived as
deceptive at any point in the communication chain by a communicant who is sen-
sitive to the EH’s larger context. This multiple sendership, then, is what enables the
dual, and seemingly paradoxical, pragmatic status of EHs as both uncooperative
and cooperative.
This dual cooperativity has an interesting consequence for the receiver of an
EH. Upon receiving an EH, there are a number of possible interaction scenarios for
the addressee, which will be briefly sketched here. Crucially, there is no way of not
cooperating with an EH that would be viable under normal communicative condi-
tions, as all scenarios that actually occur in email transactions are of a cooperative
nature. The central parameter for a communicant’s reaction to an EH is which
frame of reference is activated in terms of sender identity: Sender sn may either at-
tribute the message – with all its pragmatic entailments – to sender sn-1 or to sender
s1. Put more simply: A communicant may either believe or disbelieve the message –
fall for the EH or see through it. Depending on this communicative stance, various
types of interaction are possible:
– Believe the message and forward it: This is the “utterly cooperative” scenario
mentioned above. The communicant receives a message which they take to be
fully cooperative and truthful. Accordingly, they in turn cooperate with the
message, and this includes the demand for message dissemination (see 1.1.3).
This is the interaction pattern that allows EHs to spread. Judging from the over-
whelming success of many EHs, it must be a fairly common scenario.
– Believe the message and ignore it: This option is slightly weaker in terms of co-
operation, yet it still holds as cooperative behavior. While the urge for dissemi-
nation is an important part of the message, its central issue is nevertheless the
deceptive statement at its core. To take the message at face value, then, is the
minimum requirement for cooperativeness. Involvement certainly is a factor in
cooperative behavior, but it is a matter of degree. For example, it violates the
CP not to participate in ritualized verbal exchanges such as adjacency pairs, yet
These strategies are summarized in Figure 1. In the figure, arrows represent the
exchange of messages between communicants. Arrows pointing to the right repre-
sent the forwarding of a message, and arrows pointing to the left represent attribu-
ting a message to a previous sender and/or engaging directly with that sender.
Dotted arrows indicate that a (potentially) longer chain of communication is not
shown in its entirety.
Directives: It is safe to say that every EH contains directives; these are often
portrayed as the tell-tale sign of EHs. The most pervasive type of directive is of a
meta-communicative nature: the exhortation to disseminate the EH. Thus every
EH contains a request for proliferation, very often prominently placed at the end of
the message and very often formulated as an imperative. Here are some examples:
“PLEASE PASS ON AS A LAST REQUEST”
So please start e-mailing and help us build our database.
PLEASE DISTRIBUTE TO AS MANY PEOPLE AS POSSIBLE … THANX
In addition, many EHs contain directives concerning the actual scope of the
message: They give advice and warnings about the (nonexistent) threat established
in their representative part. This is most often the case with virus warnings and
“traditional” urban legends which report shocking incidents. The following is a
particularly elaborated example from a story about HIV-infected needles in ex-
posed places:
IT IS IMPERATIVE TO CAREFULLY CHECK THE HANDLE of the gas pump
each time you use one. LOOK AT EVERY SURFACE YOUR HAND MAY
TOUCH, INCLUDING UNDER THE HANDLE. If you do find a needle af-
fixed to one, immediately contact your local police department
so they can collect the evidence.
Commissives in EHs are a less pervasive and more complex phenomenon. They
occur almost exclusively in one subtype of EH, namely giveaway EHs; for these,
however, they are the most distinctive feature. A giveaway EH promises some ma-
terial gain to its receiver, be it money, free goods, or other type of reward:
you will receive $1000 and a copy of Windows98 at my expense.
we will send a coupon for one six-pack of any of our Miller
Brand beverages
1,300 of the people on the list will receive $5,000, and the
rest will receive a free trip for two to Disney World for one
week during the summer of 1999 at our expense.
Crucially, these commissive speech acts are never to be attributed to sender sn-1,
even in a cooperative framework. They have an additional built-in layer of decep-
tion, usually a fictitious sender s1 (such as a CEO, sales manager, or other repre-
sentative of a well-established company). Thus, the face-threatening potential of
obligation is minimized for the actual sender sn, as the speech act proper is perpetu-
ally relocated to a distant yet powerful source of authority. This is another demon-
stration of what Döring’s (2003) notion of the “prosocial” implies for EHs: a way
of acting altruistically without the need for face loss or commitment, financial or
otherwise. Commissives of this type occasionally migrate into other EH subtypes:
In particular, they appear in some sympathy EHs (most typically those which also
refer to an “email tracker”).
Acknowledgements: This expressive type of speech act occurs frequently
within the interactional discourse setting of email. While the actual distribution of
acknowledgements in EHs would warrant a study in its own right, acknowledge-
ments appear to be surprisingly rare in the message core; this is especially true for
virus warnings, which aim to present a purely informational appearance. When
they occur, acknowledgements tend to cluster around the openings and closings of
the message. This is where the various senders along the communication chain
tend to interact with the text and make their own changes and additions (see Heyd
2008: 56–67 for a structural analysis of such embedding techniques). Messages
may contain a succession of opening and closing formulas, as well as other emot-
ive utterances. The following accumulation is from the closing of a sympathy EH:
So, I know that we can send it to at least 5 or 6. Come on you
guys … and if you’re too selfish to waste 10–15 minutes and
scrolling this and forwarding it to EVERYONE, just think it
could be you one day … and it’s not even your $money$, just your
time. Please help this little girl out guys, I know you can do
it!! I love you guys!
This chapter gives an approximation of the deceptive scope of EHs and their under-
lying pragmatic mechanisms. Our analysis of the dual cooperation pattern charac-
teristic of the genre has shown that pragmatic analysis can yield new and meaning-
ful insights into such deceptive genres. Indeed, the analysis of CMC genres, with
their new socio-technical settings and the easy and ample availability of linguistic
data they provide, may actually stimulate pragmatic theory as a whole and encour-
age the development of new approaches to the study of pragmatics.
While email hoaxes may be a fossilizing genre that will eventually come to be
associated with typical online communication patterns of the 1990s and 2000s, the
broader topic of deception in CMC is persistent and likely to remain a central con-
cern of CMC studies for a long time (see Hancock, this volume). The pragmatic ap-
proach to analyzing deceptive patterns of online interaction described here may
therefore serve for the study of reminiscent CMC genres that coexist with EHs or
that may emerge in the future.
Notes
References
Attardo, Salvatore
1993 Violation of conversational maxims and cooperation: The case of jokes. Jour-
nal of Pragmatics 19(6): 537–558.
Auer, Nicole J. and Ellen M. Krupar
2001 Mouse click plagiarism: The role of technology in plagiarism and the librar-
ian’s role in combating it. Library Trends 49(3): 415–432.
Austin, John
1962 How to Do Things with Words. The William James Lectures Delivered at Har-
vard University in 1955. Edited by J. Urmson. Oxford, UK: Clarendon.
Bach, Kent R. and Robert M. Harnish
1979 Linguistic Communication and Speech Acts. Cambridge, MA: MIT Press.
Barrett, Daniel
1996 Bandits on the Information Superhighway. Sebastopol, CA: O’Reilly Media.
Barron, Anne
2006 Understanding spam: A macro-textual analysis. Journal of Pragmatics 38(6):
880–904.
Blommaert, Jan and Tope Omoniyi
2006 Email fraud: Language, technology, and the indexicals of globalisation. Social
Semiotics 16(4): 573–605.
Boese, Alex
2003 The Museum of Hoaxes: A History of Outrageous Pranks and Deceptions.
New York: Plume.
Grice, H. Paul
1989 Studies in the Way of Words. Cambridge, MA: Harvard University Press.
Hancock, Jeffrey
2007 Digital deception: When, where and how people lie online. In: Adam N. Join-
son, Katelyn McKenna, and Tom Postmes (eds.), Oxford Handbook of Internet
Psychology, 287–301. Oxford, UK: Oxford University Press.
Hayes, Arthur S., Jane B. Singer, and Jerry Ceppos
2007 Shifting roles, enduring values: The credible journalist in a digital age. Journal
of Mass Media Ethics 22(4): 262–279.
Hedley, Steve
2006 A brief history of spam. Information & Communications Technology Law
15(3): 223–238.
Heisler, Jennifer M. and Scott F. Crabill
2006 Who are “stinkybug” and “Packerfan4”? Email pseudonyms and participants’
perceptions of demography, productivity, and personality. Journal of Com-
puter-Mediated Communication 12(1): 114–135.
Herring, Susan C., Kirk Job-Sluder, Rebecca Scheckler, and Sasha Barab
2002 Searching for safety: Managing “trolling” in a feminist discussion forum. The
Information Society 18(5): 371–384.
Heyd, Theresa
2006 Understanding and handling unreliable narratives: A pragmatic model and
method. Semiotica 162: 217–234.
Heyd, Theresa
2008 Email Hoaxes: Form, Function, Genre Ecology. Pragmatics & Beyond New
Series 174. Amsterdam/Philadelphia: John Benjamins.
Hruska, Jan
1990 Computer Viruses and Anti-Virus Warfare. New York: Ellis Horwood.
Kibby, Marjorie
2005 Email forwardables: Folklore in the age of the Internet. New Media & Society
7(6): 770–790.
Kolari, Pranam, Akshay Java, and Tim Finin
2006 Characterizing the splogsophere. In: Proceedings of the 3rd Annual Workshop
on Weblogging Ecosystem: Aggregation, Analysis and Dynamics, 15th World
Wide Web Conference, May 23, Edinburgh.
Lee, Kang
2000 Lying as doing deceptive things with words: A speech act theoretical perspec-
tive. In: Janet Wilde Astington (ed.), Minds in the Making: Essays in Honor of
David R. Olson, 177–196. Oxford, UK: Blackwell.
Leslie, Alan M.
1987 Pretense and representation: The origins of “Theory of Mind”. Psychological
Review 94(4): 412–426.
MacDougall, Curtis D.
1958 Hoaxes. New York: Dover.
Meibauer, Jörg
2005 Lying and falsely implicating. Journal of Pragmatics 37(9): 1373–1399.
Mintz, Anne (ed.)
2002 Web of Deception: Misinformation on the Internet. Medford, NJ: CyberAge
Books.
1. Authentication
1.1. Introduction
Trust is a key issue in computer-mediated communication (CMC). While the diffi-
culty of knowing who and what to believe arises in relation to communication of
all kinds, in CMC these questions are foregrounded. To a greater degree than in
other media, CMC is characterized by relatively high levels of uncertainty about
the identity of interactants, the origin and intentions of messages received, the
presence of lurkers and official surveillance, the likelihood of criminal activity, and
the threat of attack by malicious software. As a result, communication over the In-
ternet is often felt to involve some risk and, in the case of unsolicited email mess-
ages (spam), is likely to raise considerable suspicion.
This suspicion seems justified: Spam now accounts for at least two-thirds of
email communication, and two-thirds of spam messages are estimated to involve de-
ception (Cukier, Cody, and Nesselroth 2006; online deception is also discussed in
the chapter by Hancock and Gonzales in this volume). Much of this material is pro-
motional and, as such, shares generic features with legitimate forms of advertising,
except that its purpose is to lure recipients into buying products which are either
worthless or non-existent (Barron 2006). In the case of another ubiquitous form, the
so-called Nigerian Letter (also known as “419” or “advance fee” fraud; cf. Edelson
2003; Holt and Graves 2007; Smith, Holmes, and Kauffman 1999), the limits of
legality are more or less openly crossed: The ostensible aim is both more personal
and more questionable, usually involving the recipient’s complicity for personal
gain in some tempting but irregular financial arrangement. The real aim is inevitably
to defraud the recipient.
In pragmatics, work has been done on the nature of unsolicited promotional
emails, Nigerian Letters, and other types of online scam (e.g., Barron 2006; Blom-
maert and Omoniyi 2006; Cukier et al. 2006; Heyd 2008, this volume; Tseng 2007,
2010). The focus has been on spam as a genre, its typical moves and move struc-
ture, and the nature of the deceptions perpetrated. This work has pointed to ways in
which these forms differ from conventional text-based genres and exploit com-
puter-specific opportunities for composition and interaction.
Equally important, but as yet less widely studied, are the pragmatic strategies
by which the senders of these messages seek to pass them off as trustworthy, bona
fide communications (for tone discussion, see Tseng 2007). In this respect, Ni-
gerian Letters furnish interesting material. Given the level of suspicion likely to be
encountered, the scale of the task, and the fact that the communication generally
has only one chance of success, it could be expected that senders would go to con-
siderable lengths to allay suspicion and establish recipients’ trust. As we shall see,
this is indeed the case: Writers create a range of apparently trustworthy identities
and attempt sometimes elaborate authenticating work. Yet, despite this, it is their
shortcomings that are likely to strike their First World recipients (albeit clearly not
all of them, to judge from the reports of those successfully conned; see below). It is
then relevant to ask why these efforts should prove futile and, by extension, what
conditions are necessary for authenticating strategies to succeed.
This chapter is concerned with how authenticity is achieved in CMC, and in
particular with the forms and distribution of this authenticating work, which, it will
be argued, amounts to a default requirement in online interaction.1 Drawing on a
corpus of 560 Nigerian Letter emails received by the author between December
2004 and March 2007, it examines the means by which senders negotiate the com-
municative challenges involved in initiating personal contact and establishing
plausible grounds for interaction with their unknown recipients, interaction de-
signed to turn the recipient from detached outsider into would-be accomplice in
serious fraud, and its unwitting victim. The aim is to shed light both on the con-
struction of authenticity in online interaction and on the nature of a genre in which
many normal pragmatic assumptions do not seem to hold.
(a) Consistency
Linguistic and non-linguistic cues are normally expected to reinforce one another
or at least not to conflict. Consistency is unlikely to be noticed when it occurs, but
participants in interaction tend to be sensitive to signs of deviation. Inconsistencies
may be interpreted as deception or duplicity: Suspicion will be aroused, for
example, when a speaker’s words say one thing and body language another; or
when a text lacks coherence or contains marked and unmotivated stylistic contrasts
(Mabry 2001: 249). The Nigerian Letters often fail spectacularly in this respect.
All authenticating work is instantly undone when a sender purporting to be Mr Paul
Smith, a senior official in a London-based bank, begins his message “Hello my
dear” and goes on to commit a string of elementary language and writing errors; or
when a message with the subject line “I AM CRYING FOR HELP” opens with the
greeting “Hi Sweetheart” (see also Blommaert and Omoniyi 2006: 600). Consist-
ency therefore seems to be an overriding requirement.
(b) Quantity
For most purposes we require only the minimum of authenticating work; more than
this will stand out as marked, hence contrived (as in overly fulsome thanks or re-
peated assurances of honesty). If repeated so often that they become formulaic,
cues tend to lose their authenticating effect (cf. Fairclough’s [1989:62] “synthetic
personalization”), although we should distinguish this case from that of such genu-
inely formulaic authenticating expressions as oaths, vows, and other performatives
that have legal force. The cumulative nature of CMC means that this condition
applies not only to a single message but across a collection of messages. In the Ni-
gerian Letter corpus, the potentially authenticating anticipation of the recipient’s
surprise at being approached out of the blue by a stranger is cancelled by its reiter-
ation across a large number of messages, in nearly identical terms (see below).
(c) Spontaneity
As this suggests, behaviour appears to be most authentic when it is most instinc-
tive, in the unguarded moments when our “real” underlying selves are openly on
display. In Goffman’s (1959) terms, such cues are “given off” rather than “given”,
not subject to conscious control. By contrast, evidence of a role being played or of
special effort being made will tend to be disauthenticating, as will overuse of
stereotyped cues (cf. Herring and Martinson 2004: 428). In general, as Goffman
notes, the authenticity of the given – the public face – tends to be measured in re-
lation to the given off – those unconscious traces of behaviour that may leak
around the edges of the impression a speaker wishes to create, with some degree of
self-consistency expected. Ultimately, however much care a speaker takes, total
control will be impossible to achieve, especially over features not normally at-
tended to, and if achieved, would – for that very reason – be likely to appear in-
authentic.
At the same time, certain genres can seem to be naturally self-authenticating:
Confessions and admissions, prayers, expressions of helplessness and victimhood,
impassioned outbursts, private, off-the-record comments, and backstage chat are
all forms in which the naked self may seem to be tellingly revealed. Anonymous
CMC seems conducive to self-revelation of this kind, and many of these genres are
represented in the Nigerian Letter corpus.
dependent of the credibility or otherwise of the message per se: A speaker may
quite plausibly utter implausible things. Appropriateness refers to aspects such as
the choice of a given medium or register. Thus, whatever its plausibility, it hardly
seems appropriate that a UK-based firm of solicitors acting for the estate of the late
Sir Denis Thatcher should choose to use a German freemail service to notify me of
a bequest from him amounting to almost a million pounds.
(e) Engagement
Communication necessarily involves interaction between interlocutors, who there-
fore need to manifest in their messages some awareness of one another, however
minimal. A message sent to a specific recipient can normally be expected to show
signs of being tailored to suit that recipient and to engage in constructing the inter-
action as a joint enterprise. The engagement condition suggests that authenticity
effects should not only be sender-orientated (i.e., concerned with projecting the
sender’s identity) but also relational (i.e., concerned with establishing and further-
ing the interaction as a whole). It is on the basis of this relational work that inter-
locutors form inferences about one another and develop mutual understanding
(Mabry 2001: 6). Engagement may be harder to establish in CMC, with its greater
dependence on text-based cues, and harder still in anonymous interaction, where
there is no accumulation of personal history (Donath 1999). Few of the Nigerian
Letters meet this condition adequately; as will be seen below, ignorance of the re-
cipient is often total, and pragmatic features relevant to sustaining reciprocal in-
volvement are either lacking or purely formulaic.
envelopes, recorded delivery services, etc. Many of the messages in the Nigerian
Letter corpus attempt to reproduce aspects of offline formal letter layout, in par-
ticular by including mailing address and header, but the limited resources of
plain text formatting tend to undermine their effect. Cues such as these, it seems,
cannot simply be transposed from offline to online contexts. Some of the writers
use the hypertext markup language (HTML), giving them greater control over the
appearance of the text on screen, including the use of colour, varied fonts, etc.
However, since such conventions are not standard features of email, they may
tend to foreground the home-made quality of the messages rather than authenti-
cate them.
In certain forms of chat and email, emoticons have developed as a surrogate for
some of the authenticating functions of the physical states and non-verbal behav-
iour of face-to-face interaction, particularly facial expressions, although it is
doubtful to what extent they reproduce physically-embodied authenticity effects
(cf. Dresner and Herring 2010). For one thing, as merely a notation for (rather than
an expression of) intended emotional (etc.) states, emoticons have no necessary
connection with any particular feeling on the part of the sender. Secondly, unlike
much non-verbal behaviour, their use is under conscious control and, as such, not
part of those maximally authenticating aspects of behaviour that are involuntary
(see above). By appearing constructed and strategic, they may even detract from
the sense of authenticity (Walther and d’Addario 2001). Nonetheless, it appears
that in some contexts, such as gaming, by indexing affect, emoticons can also help
promote engagement, in the sense discussed earlier, and the development of rela-
tionships between interlocutors (Utz 2000).
At the same time, in contrast to offline correspondence, email also contains as-
pects that cannot easily be controlled by the sender (the technical “given off”) – for
example, the addition of advertising in freemail services, inclusion of the originat-
ing ISP address, and trace of the path along which the message has been sent –
which in certain cases, including that of Nigerian Letters, are likely to affect the re-
cipient’s perception of its authenticity. Claims as to the physical whereabouts and
identity of the sender are likely to be assessed against the evidence of these fea-
tures.
The same goes for the way in which emails are stored on the recipient’s com-
puter or mail server, making whole archives of messages potentially available for
comparison (cf. the quantity condition, above). It is highly disauthenticating to find
many almost identically worded messages in one’s own inbox. Their sheer volume
undermines the claim that they are one-off approaches to the recipient alone that
will “come as a surprise”, and makes the formulaic routines of their structure and
content all too obvious. This points to the greatest challenge facing online
scammers – one which, in the case of the producers of Nigerian Letters, at least,
they rarely seem to appreciate. That this should be so is perhaps a consequence of
the lack of reciprocity between sender and recipient that characterizes these mess-
ages. In the absence of genuine responses, the senders are unlikely to gain much
understanding of the circumstances in which their messages will be received, read,
or stored.
Type Frequency
Unclaimed deposit 36
Rescue transfer 27
Collusive transfer 16
Charity 8
Transfer 7
Business proposal 5
Bequest 1
Total 100
One other genre – the Lottery win – is in fact the most frequent (236 arrived in my
inbox over the period of this study), but it is excluded from the present discussion
on the grounds that its textual functions are somewhat different from the others. All
of the messages considered here seek to persuade the recipient to act as intermedi-
ary in shifting a huge sum of money abroad by providing a “safe” foreign bank ac-
count and offer a substantial share of it in return.
The types call for a brief comment. In the “unclaimed deposit”, the usual story
is that a businessman has perished, along with all his family, in a road accident or
air crash (the latter usually corroborated by a link to a relevant news report), leav-
ing a huge fortune in a bank account unknown to anyone but the sender and without
traceable next of kin. The “rescue transfer” generally involves a close relative of a
government minister or other senior figure now deceased, who needs help in trans-
ferring the family fortune abroad, out of potential danger. “Collusive transfer”
requires the recipient to assist in blatantly fraudulent activity, often the result of
corrupt management of government contracts, in which the sender has been in-
volved in “over-invoicing”. The “charity” messages come from senders with incur-
able diseases, who now wish to make some (extremely generous) donations to
charitable organizations in other countries. “Transfers” and “business proposals”
are more straightforward. The former concern unclaimed funds, in some cases
written as if the recipient is already involved with the Nigerian government; the
latter propose various lucrative business partnerships. The single example of a “be-
quest” message takes the form of a solicitor’s letter, and again its opening implies
ongoing communication: “Sequel to your non response to our earlier Mail …”.
Senders claim many different origins and identities; they also make use of a
range of freemail addresses in third countries, sometimes unexpected ones (e.g., in
one case, the fan site for a German television actress). The majority expect replies
to be sent to an alternative address. Claimed locations range across Africa, the
Middle East, and Europe, among others, with 10 in London and 34 in Nigeria. (For
further discussion of the email addresses used in Nigerian Letters, see Edelson
2003: 395–396.)
Stylistic features:
Legal/business register
Transactional language (e.g., business arrangements), technical details
“Powerful” language features
Formatting conventions:
Formal layout
Headings
HTML – use of colour and formatting
Personal attributes:
Powerful identity
Status markers; titles
Competence with legal procedures, financial arrangements, etc.
Victimhood – of conflict, personal misfortune, corruption, etc.
Wife/widow/son of powerful person – combining power and victimhood
Concern with moral probity of recipient
Honesty in the face of endemic corruption
Evidence of personal work, effort, risk
Religious sentiments, wishes, blessings; verses from Bible/Qur’an
Brazen dishonesty
Emotional states:
Plight narrative – political crisis, danger, death
Importance/urgency
Need for secrecy
Use of emotive, evaluative language
Helplessness
Decision not lightly taken
Material proof/evidence/corroboration:
Reference to news stories; hyperlinks to online news items, web pages
Familiar stereotypes/images of Africa
Certification/documentation:
Official ID to follow
Contact details (address/fax/phone numbers)
Email address
messages belong to the category of “unclaimed bequests”, in which they are shown
to have acted correctly – making fruitless efforts to contact next of kin, carrying
out research through the bank’s files, etc. In many cases, as noted, they are also
possessors of confidential knowledge: They alone know of the secret account, and
they express repeated concern about the trustworthiness of the recipient. Fur-
thermore, they are individuals of great probity, who are at pains to stress the legal-
ity and risk-free nature of what they are proposing. The following are typical:
I guarantee that this will be executed under legitimate ar-
rangement that will protect you from any breach of the law
as I will use my position as a Bank’s member of group risk to
secure approvals and guarantee the successful execution of
this transaction.
Chris Wreh, Member of the Risk Management Group
I will not fail to bring to your notice that this business is
100 % risk and trouble free and that you should not entertain
any fear as all modalities for fund transfer can be final-
ized within 7 to 9 banking days.
Mr Gipsin Ben, Consulting Auditor
ditions by turning hastily – at just the point where an appeal for help might be ex-
pected – to the “transactional” details, the inevitably huge sums deposited with se-
curity companies (references to which are found in 59 messages in the main study),
the necessary documentation, and the deal for the recipient. Cross-cultural prag-
matic issues also arise. The powerful professional identity that many writers seek
to project, positioning the recipient as a lowly, if favoured, petitioner, may be auth-
enticating in more hierarchically-ordered contexts (cf. Gumperz 1982: 182); for a
European recipient, however, it is likely to be alienating. Similarly, to many recipi-
ents the use of capital letters to mark points of special importance will suggest
shouting. No traces of a “real self” are given off in the form of a distinctive voice;
in fact, the mass of these messages seem to be without any spontaneous features
other than those that point towards their real origins. As this shows, identities are
not collections of cues, and neither the deployment of authenticating strategies nor
the work that goes with them in themselves create the experience of authentic com-
munication.
Example 1.
1 Date: Fri, 26 Aug 2005 04:45:09 From – Fri Feb 17 14:22:43 2006
+0100 (GMT+01:00) Date: Wed, 15 Feb 2006 06:26:15
From: Dr Phil Hodkinson +0100 (GMT+01:00)
<phil_hodkinson@virgilio.it> From: james pual
Reply-To: phil_hodkinson@excite.com <james_pual1@virgilio.it>
Subject: FROM PHIL’S OFFICE Reply-To: ja_pu12@Yahoo.com
Subject: FROM DIRECTOR PETERS
OFFICE
2 FROM PHIL’S OFFICE FROM DIRECTOR PETERS OFFICE
HEAD OF DEPARTMENT BANK OF SCOTLAND
BANK OF SCOTLAND 3RD FLOOR MOUND
UNITED KINGDOM EDINBURGH
EH1 1YZ
UNITED KINDOM,
EMAIL: ja_pu11@yahoo.com
Tel/Fax:+44–7031842574
3 DEAR PARTNER, GREETINGS,
It is immediately striking that the passage of time has not helped to refine the
letter’s authenticity effects. The second message is longer (306 words compared
with 188), but the content and sequence of moves is largely unchanged, and while
the punctuation has been corrected in paragraph 4, many language errors remain
untouched. The major difference in the later message is the addition of a “stan-
dard” move outlining the financial details of the deal and the recipient’s cut of the
proceeds. In paragraph 6 the sum to be transferred has turned from pounds into dol-
lars. Neither message follows the usual pattern by opening with a personal intro-
duction of the sender and anticipation of the recipient’s surprise; instead, both
adopt a brisk, direct tone more appropriate to an already established business rela-
tionship. If anything, the second message seems somewhat less coherent than the
first, since the “surprise” element has been included twice, worded identically, ap-
parently without concern for the most suitable ordering of elements in the text.
Moreover, the unusual surname Pual looks suspiciously like a misspelling of Paul
and is at odds with the heading: “FROM DIRECTOR PETERS OFFICE”.
In a third version (example 2, below), received shortly afterwards, the text has
been adjusted further towards the Nigerian Letter norm, with the inclusion of a first
person introduction (“I am MR. ROBERT PAYNE”) at the beginning (there are 26
occurrences of “I am + title + full name” in the main study, 175 in the full corpus).
Greater emphasis is also placed on gathering details from the recipient. Nonethe-
less, apart from a shift into lower case, the main elements of the message remain
unchanged, new errors have appeared, particularly in the added elements, and the
sum in question is stated as two different amounts, as in the first version. There is
little evidence that potential recipients’ norms and expectations have been used in
any consistent way to guide the changes introduced over this sequence or to en-
hance the authenticity effects created. While the use of lower case removes one
stylistic flaw, capitalization and other written conventions remain highly non-stan-
dard. The writer of example 2 has dropped disauthenticating features such as the
inappropriate heading “FROM DR PHIL’S/DIRECTOR PETERS OFFICE” but
has inserted the equally inappropriate personal introduction, and the text used to
explain the unclaimed deposit has reverted to that of the first message. Overall,
example 2 has a more professional appearance, with the addition of an official-
looking heading: “HOW THE MONEY CAME ABOUT:” and a numbered list of
personal details required at the end; but with respect to its authenticity effects, the
text itself has hardly progressed at all.
Remarkably, despite the higher threshold of acceptability required to pass as an
English-speaking official in the UK communicating with an educated European
audience, these messages show no more sign of serious effort than any of the
others. In the examples above, the sender presents himself as a senior manager of
the Bank of Scotland, based in the UK, yet the text is so manifestly non-standard
that the claim is laughable. In addition, he hardly seems aware of his location: The
identity of John Hughes, a “Foreigner … here in London”, is unconvincing, as is
the urgent need to transfer his money to “a safe Foriegn Account Abroad”. Both
would make much better sense in relation to Nigeria, and, needless to say, this is
the usual case.
Example 2.
Date: Tue, 28 Feb 2006 15:11:35 +0100 (GMT+01:00)
From: “Mr.Robert Payne” <robert_payne@virgilio.it>
Reply-To: payne_robert2006@yahoo.com.au
Subject: CONTACT ME!!!
FROM:MR.ROBERT PAYNE
BANK OF SCOTLAND
UNITED KINGDOM
DATE:
SIR/MADAM,
I am MR. ROBERT PAYNE a British citizen and I am 53years of age.
I am married with three kids, I am the Auditor and computing
Manager of Bank Of Scotland. In Order to Transfer out (FIFTY-
MILLION POUNDS STERLINGS) From our Bank Here in London I Hvae
the courage to look for a reliable person that will be capable
for this important business transaction believing that you
will not let me down Either now or in the Future.
HOW THE MONEY CAME ABOUT:
The owner of this account/money is MR.JOHN HUGHES a Foriegner
and The Manager of PETROL CHEMICAL SERVICE here in London, A
chemeical Engineer by Proffession and He Died since 1995 and
this money has remained unclaimed because the account has no
other Beneficiary and my investigation proved to me as well
that his company does not know anything about this acoount and
the amount involvedis (15,000,000.00)Pounds Sterlings.
I want to Transfer this money into a safe Foriegn Account
Abroad but i don’t know any Foriegner.I know that this Message
will come to you as a Surprise as we don’t know ourselves
before,But be sure that it is Real and a Genuine Business and
as well is 100 % RISK FREE.I Believe in GOD That you will never
let me down in This Investment Opportunity.If you are
Interested in this Business Please the form below and return
back to me to enable me start up with the Necessary Procedures
as to a free and safe transfer of this Fund into the Account
Informations that will be Provided by you as i will be Using my
Position in The Office to influence other Staff.Please treat
this Proposal with Utmost Confidentiality.
YOUR PERSONAL DETAILS AS REQIURED BELOW:
1) FULL NAME
2) AGE
3) NATIONALITY
4) COUNTRY OF RESIDENCE
5) OCCUPATION
6) CONTACT ADDRESS
7) TEL/FAX NUMBER
8) ACCOUNT NO/DETAILS
Yours Sincerely
MR.ROBERT PAYNE.
This, in turn, reflects the fact, to which Blommaert and Omoniyi (2006) draw
attention, that these emails are sent across a great divide of culture, wealth, and ac-
cess to resources, including computer technology. They are sent against the domi-
nant North – South flow of influence, holding out to the already affluent recipient a
vision of an instant fortune to be made from the poorest corner of the world. They
are sent by individuals who may be educated producers of a high status variety of
English in their own locality, “streetwise in the culture of the Internet” (Blommaert
and Omoniyi 2006: 585), alert to the ironies of their post-colonial situation: Some
of them, it appears, are celebrated in Nigeria for striking a blow against the old im-
perialists, and they perhaps knowingly reproduce in their messages a stereotyped
image of Africa as corrupt, violent, poor, and religious, yet rich in resources for
foreigners to exploit (cf. Smith 2008). Yet they show little awareness of how their
English differs, both in form and status, from that of the First World to which their
messages are directed or of how their messages are read in consequence.
2.7. Discussion
Senders of unsolicited email messages face a daunting task. Before they can begin
to persuade potential victims to respond, they must create the conditions in which
their messages will be read as genuine, in a context of high levels of distrust. As
this discussion has illustrated, writers of Nigerian Letters are well aware of the
challenge and able to draw on a repertoire of authenticating strategies, often with
great resourcefulness. Despite their many oddities and linguistic shortcomings, the
messages rarely seem naïve; indeed, viewed as a whole, they give evidence of cal-
culated management of the “given” aspects of self-presentation and technical con-
trol over the medium. They show themselves to be adept CMC users, immersed in
their local context, aware of how Africa is read from the outside and able to exploit
nuances of voice and style to create particular authenticity effects. Their most fa-
voured strategies are those of the managerial/legal/transactional discourse of
power and the “plight” narrative, instantly recognizable to First World news audi-
ences, or a combination of the two. Some of these are strikingly told, some – with
their chests of treasure and secret arrangements – play with familiar folk motifs of
magic wealth.
As any collection of these messages will quickly reveal, however, they are
modular constructions, created according to a variety of templates, in which moves
are combined with no more than minimal concern for their ordering, and whole
texts, textual elements, turns of phrase, names, etc. are constantly (often incoher-
ently) recycled and repeated with variations. As such, it may be wrong to approach
them as fully intentional, human forms of communication at all: Many could
equally be computer generated. Indeed, Holt and Graves (2007: 150) suggest that
standard scripts could be loaded into malicious software to enable the automatic
creation and propagation of these emails, with little intervention by the original
sender.
Moreover, with few exceptions, the overt authenticating work is not relational
or recipient-focused, except in the most formulaic sense (the “will come as a sur-
prise” move, assurances of safety, legality, etc.), and involves no understanding of
the recipient’s circumstances or background. Instead, the effort is sender-focused,
with constant reference to the sender’s attributes, experience, and competences.
There is rarely any evidence that the writers have tried to think themselves into
their role – to imagine the range of competing emotions such a person might feel,
acting clandestinely and at considerable risk to himself – or to project themselves
into the recipient’s position, or even to imagine a likely audience and attempt to
initiate serious interaction. When one sender does so, it stands out; the fact that his
is by far the longest message in the corpus is hardly coincidental.
Thus, the great majority of these writers are adept yet inept. They seem to at-
tend to the need for authenticity within the scope of a given move – for example,
when greeting, introducing themselves, establishing the details of their case, etc. –
but not across whole texts. Despite the care they so obviously take, they frame their
messages carelessly, even when errors will be disauthenticating, and while taking
pains to conceal aspects of the email such as the list of recipients, they rarely show
more than perfunctory concern for its specifically authenticating features such as
the subject line or for the many cues unintentionally given off. Above all, they ne-
glect the aspects of the message that are critical to authentication from the recipi-
ent’s perspective, summarized above as the five conditions of consistency, quan-
tity, spontaneity, plausibility/appropriateness, and engagement.
3. Conclusion
Notes
1. Given the increasing use of email for official purposes and the growth of private transac-
tions online, much of the literature on authentication has been concerned with issues of
“certification”, for example of those accessing virtual communities, online forums, ser-
vices, etc. (e.g., Etzioni and Etzioni 1999) or with the development of technical solutions
to provide secure channels for money, electronic signatures, encryption and the like (see,
e.g., Sasse 2004). Although relevant to authentication in its widest sense, these issues will
not be discussed further here.
References
Barron, Anne
2006 Understanding spam: A macro-textual analysis. Journal of Pragmatics 38:
880–904.
Blommaert, Jan and Tope Omoniyi
2006 Email fraud: Language, technology and the indexicals of globalisation. Social
Semiotics 16(4): 573–605.
Carlson, John, Joey F. George, Judee Burgoon, Mark Adkins, and Cindy White
2004 Deception in computer-mediated communication. Group Decision and Negoti-
ation 13: 5–26.
Coupland, Nikolas
2001 Dialect stylization in radio talk. Language in Society 30: 345–375.
Cukier, Wendy, Susan Cody, and Eva Nesselroth
2006 Genres of spam: Expectations and deceptions. In: Proceedings of the Thirty-
Ninth Hawaii International Conference on System Sciences,
http://csdl.computer.org/comp/proceedings/hicss/2006/2507/03/250730051a.pdf
Danet, Brenda
2002 The language of e-mail. European Union Summer School, University of Rome.
http://www.europhd.psi.uniroma1.it/html/_onda02/04/ss8/pdf_files/lectures/
Danet_email.pdf
Donath, Judith
1999 Identity and deception in the virtual community. In: Marc A. Smith and Peter
Kollock (eds.), Communities in Cyberspace, 29–59. London: Routledge.
Dresner, Eli and Susan C. Herring
2010 Functions of the non-verbal in CMC: Emoticons and illocutionary force. Com-
munication Theory 20: 249–268.
Edelson, Eve
2003 The 419 scam: Information warfare on the spam front and a proposal for local
filtering. Computers and Security 22(5): 392–401.
Ellison, Nicole, Rebecca Heino, and Jennifer Gibbs
2006 Managing impressions online: Self-presentation processes in the online dating
environment. Journal of Computer-Mediated Communication 11: 415–441.
Etzioni, Amitai and Oren Etzioni
1999 Face-to-face and computer-mediated communities, a comparative analysis.
The Information Society 15(4): 241–248.
Fairclough, Norman
1989 Language and Power. London: Longman.
Gill, Martin
2011 Authenticity. In: Jan-Ola Östman and Jef Verschueren (eds.), Pragmatics in
Practice: Handbook of Pragmatics Highlights, Vol. 9, 46-65. Amsterdam/Phil-
adelphia: John Benjamins.
Gimenez, Julio
2000 Business email communication: Some emerging tendencies in register. English
for Specific Purposes 19(3): 237–251.
Goffman, Erving
1959 The Presentation of Self in Everyday Life. Garden City, NY: Doubleday.
Grice, H. Paul
1975 Logic and conversation. In: Peter Cole and Jerry L. Morgan (eds.), Syntax and
Semantics 3: Speech Acts, 41–58. New York: Academic Press.
Gumperz, John
1982 Discourse Strategies. Cambridge, UK: Cambridge University Press.
Henderson, Samantha and Michael Gilding
2004 “I’ve never clicked this much with anyone in my life”: Trust and hyperpersonal
communication in on-line friendships. New Media and Society 6: 487–506.
Herring, Susan C. and Anna Martinson
2004 Assessing gender authenticity in computer-mediated language use: Evidence
from an identity game. Journal of Language and Social Psychology 23(4):
424–446.
Heyd, Theresa
2008 Email Hoaxes: Form, Function, Genre Ecology. Amsterdam/Philadelphia:
John Benjamins.
Holt, Thomas and Danielle Graves
2007 A qualitative analysis of advance fee fraud email schemes. International Jour-
nal of Cyber Criminology 1(1): 137–154.
Joinson, Adam
2001 Self-disclosure in computer-mediated communication: The role of self-aware-
ness and visual anonymity. European Journal of Social Psychology 31:
177–192.
Mabry, Edward
2001 Ambiguous self-identification and sincere communication in CMC. In: Luigi
Anolli, Rita Ciceri, and Giuseppe Riva (eds.), Say Not to Say: New Perspec-
tives on Miscommunication, 247–264. Amsterdam: IOS Press.
Montgomery, Martin
2001 Defining “authentic talk”. Discourse Studies 3(4): 397–405.
Sasse, M. Angela
2004 Usability and trust in information systems.
http://www.dti.gov.uk/files/file15320.pdf
Slater, Don
2002 Social relationships and identity online and offline. In: Leah A. Lievrouw and
Sonia Livingstone (eds.), The Handbook of New Media: Social Shaping and
Consequences of ICTs, 533–546. London: Sage.
Smith, Andrew
2008 Nigerian scam e-mails and the charms of capital. Cultural Studies 2008: 1–21.
http://dx.doi.org/10.1080/09502380802016162
Smith, Russell, Michael Holmes, and Philip Kaufmann
1999 Nigerian advance fee fraud. Trends and Issues in Crime and Criminal Justice.
(No. 121). Canberra: Australian Institute of Criminology.
http://caballonm.com/ti121.pdf
Suler, John
2004 The online disinhibition effect. CyberPsychology & Behavior 7(3): 321–326.
Tseng, Ming-Yu
2007 Telling it as if it were true: Analyzing the language of scams. In: Ming-Yu
Tseng (ed.), Exploring Language at the Interface, 93–125. Taipei: Crane.
Tseng, Ming-Yu
2010 The pragmatic act of fishing for personal details: From choice to performance.
Journal of Pragmatics 42: 1982–1996.
Utz, Sonia
2000. Social information processing in MUDs: The development of friendships in
virtual worlds. Journal of Online Behavior 1(1).
http://www.behavior.net/JOB/v1n1/utz.html
Walsh, Conal
2004 Police swoop on email fraud ringleaders. The Guardian (online edition).
http://www.guardian.co.uk/business/2004/jul/11/crime.uknews
Walther, Joseph B.
1996 Computer-mediated communication: Impersonal, interpersonal, and hyperper-
sonal interaction. Communication Research 23(1): 3–43.
Walther, Joseph B. and Kyle P. d’Addario
2001 The impacts of emoticons on message interpretation in computer-mediated
communication. Social Science Computer Review 19(3): 324–347.
Walther, Joseph B., Tracy Loh, and Laura Granka
2005 Let me count the ways: The interchange of verbal and nonverbal cues in com-
puter-mediated and face-to-face affinity. Journal of Language and Social Psy-
chology 24(1): 35–65.
Zhou, Lina, Judee K. Burgoon, and Douglas P. Twitchell
2003 A longitudinal analysis of language deception behavior in email. In: Hsinchun
Chen, Richard Miranda, Daniel D. Zeng, Chris Demchak, and Therani Madhu-
sudan (eds.), Intelligence and Security Informatics. Lecture Notes in Computer
Science, Vol. 2665, 102–110. Berlin: Springer.
1. Introduction
Online nicknames, or nicks,1 are identity markers that are employed for self-pres-
entation purposes in online interactions. The selection of an online nickname has
been labeled a ritual act, since it is practically impossible to participate in online
interactions without first submitting an online name (Crystal 2001; Tingstad 2003).
Apart from their instrumental nature, online nicknames convey intentionally-
provided information based on how online users wish to present themselves in
cyberspace (cf. Goffman 1959). This is in contrast to real life nicknames, which are
typically out of the bearer’s control in terms of selection (Morgan, O’Neill, and
Harré 1979).
In the absence of physical cues in text-based online interactions, nicknames
can be a rich source of information on how the online self is constructed in both
synchronous and asynchronous computer-mediated communication (CMC) en-
vironments. This chapter focuses on online nicknames as meaning contributions
that fall within the scope of pragmatic analysis. It is argued that applying Grice’s
Cooperative Principle to nicknames helps to account for the communicative efforts
made through the mini-propositions included in them. The findings of an analysis
of two online nickname corpora and online participants’ postings on the construc-
tion, use, and meaning of nicknames show that online nicknames can reflect truth-
ful information, can be made relevant and informative in respect to the topic of dis-
cussions, and can challenge matters of perspicuity, depending on participants’ per-
spectives.
2. Background
have also shown that individuals form impressions of others in cyberspace based
on what nicknames convey (Byrne 1994; Chester and Gwynne 1998; Jacobson
1999).
While private CMC is usually not anonymous, most multiparticipant, public
CMC allows users the possibility to play with their identity by using online nick-
names as textual “carnival masks” (Bechar-Israeli 1995; Danet, Ruedenberg-
Wright, and Rosenbaum-Tamari 1997; Danet 2001; Ruedenberg, Danet and Rosen-
baum-Tamari 1995). In a study of Internet Relay Chat (IRC) nicknames, Bechar-
Israeli (1995) developed a nickname typology that includes seven categories under
which nicknames may be classified: a) real names; b) self-related names; c) names
related to flora, fauna, and objects; d) medium, technologies, and their nature; e)
plays on words or sounds; f) famous people, figures in literature, films, fairytales;
and g) sex and provocation. That study showed that 45 % of participants used self-
related online nicknames, while only 7 % used their real (full or partial) name. Be-
char-Israeli (1995) notes that online communicators may use nicknames not only
to hide their real identity, but also to mask an already established online persona for
the purposes of playful interaction.2
Online users can also draw attention to or divert attention away from their on-
line selves by highlighting or concealing specific aspects of their identity. Subrah-
manyam, Greenfield, and Tynes (2004: 663) found that gender-transparent online
names with sexual connotations were often used in teen chatrooms as “a means of
intimate pairing off with a peer”. In a similar vein, Deul (1996: 133) argues that
participants in text-based virtual worlds use sexually suggestive online names to
encourage potential partners. Conversely, female participants may change their
nicknames and even construct a male persona to avoid receiving unwanted male at-
tention in online interactions (Bruckman 1993; Herring 1999; Kelly, Pomerantz,
and Currie 2006). From this perspective, the role of online nicknames has been ex-
tensively discussed regarding a specific aspect of online identity, gender construc-
tion and performance. Jaffe et al. (1995) found that women are more likely than
men to select nicknames that mask their gender; the authors argued that this is due
to the “implicit social pressure that women feel when interacting in mixed-gender
situations”. This claim is further supported by Pagnucci and Mauriello’s (1999) re-
search in a peer-response discussion forum on student papers, which showed that
female participants selected male nicknames to generate positive responses and
gain credibility in their statements.
In certain online communities, however, participants are expected or even
required to make gender salient through their nicknames. For instance, Hall (1996)
discusses the case of a women-only online discussion group where the use of fe-
male nicknames is demanded as part of a “female-gendered netiquette”. Camp-
bell’s (2004) observations and interviews with participants in gay male IRC chan-
nels revealed a preference for online names that explicitly or implicitly denote a
(gay) male identity. Campbell suggests that this preference reflects the presump-
tion of a gay identity on these channels. Also, it has been shown that women may
use online nicknames that challenge hegemonic perceptions of femininity as vul-
nerable and passive. Kennedy’s (2006) study of female online gamers and their
play practices in a multi-player combat game showed that, in this context, women
select online names that convey empowerment.
Gender references in nicknames are not the only kind of information about the
self that can affect online interaction. Tingstad (2003) found that, apart from
gender, technology-related and popular culture nicknames were most likely to
spark interaction among participants in a children’s chat room. Tynes, Reynolds,
and Greenfield (2004) reported that racial references in online nicknames in a teen
chat room influenced how participants were treated and gave rise to conflict. In
other cases, nicknames may affect interaction in more subtle ways. In a study of
nicknames on an Internet site for people with eating disorders, Stommel (2007)
found that nicknames reflected personal attributes related to femininity and eating
disorders. Stommel’s study suggests that nicknames that allude to the theme of the
forum may influence how participants interact.
Nickname instability is another issue that has been addressed, particularly in
connection with the problem of virtual community presence (Kendall 2002; Liu
1999). The potential malleability of online identities, the argument goes, makes it
difficult for users to keep track of online participation. However, most users do not
change their online nicknames readily, but rather tend to maintain them to ensure
consistency in their online self-presentation (Bechar-Israeli 1995; Mann and Stew-
art 2000; Reid 1991; Rheingold 2000). Unlike offline nicknames, which may be
shared by more than one person, an online nickname must be unique for each user
within a given environment at any given time. Maintaining an online nickname is
so important in a virtual environment that users can become quite possessive over
their nicknames in order to protect their online persona and reputation (Campbell
2004; Curtis 1992).
Moreover, participants in online interactions do not always exploit the possi-
bility of remaining anonymous by changing or selecting nicknames that are differ-
ent from their true identity (Bechar-Israeli 1995; Bruckman 2006). On the contrary,
studies have shown that online self-disclosure is common in CMC environments
such as blogs (Herring et al. 2006; Huffaker and Calvert 2005), newsgroups
(McKenna, Green, and Gleason 2002; Parks 1996), and MOOs (Jacobson 1996). A
plausible explanation for this practice is that online self-presentation is an exten-
sion of the real self in a different social environment (Giese 1998). Wallace (1999:
150–151) points out that self-disclosure is important in establishing rapport and
developing close relationships within online communities.
Online nicknames have also been examined as a strategy for interaction man-
agement in computer-mediated discourse (CMD). Nicknames function as address-
ivity markers that facilitate continuity (Rintel, Mulholland, and Pittam 2001;
Werry 1996) and cross-turn coherence in online interactions (Panyametheekul and
Herring 2003; Simpson 2005). Jacobson (1996: 469) also found that nicknames in
a MOO environment were used as contextualization cues to distinguish between
online and offline contexts; participants used MOO names to evoke the world of
virtual make-believe but switched to real names when they wished to signal a shift
to the “real”, offline world.
3. Illustrative case
3.1. Data
For illustration purposes, two online forums were analyzed: Parent.com and
Photo.com, both featuring publicly-available asynchronous message board discus-
sions.3 The data consist of a corpus of 550 nicknames from each forum, as well as
a collection of discussions in which users explicate and comment on their own
and other users’ online nicknames. Specifically, 24 discussions comprising 392
posted messages were collected from Parent.com, and two discussions comprising
641 messages were collected from Photo.com. The posted messages, or posts, were
retrieved by using the search tool provided in the message boards to find discussion
threads that contained the keywords nickname, screen name, username, or member
name in their titles. The data from Parent.com and Photo.com were collected in
2004 to 2005 and in 2007, respectively, with the researcher acting as a non-partici-
pant observer.
Parent.com is an American-based parenting support forum that is addressed
primarily to mothers and expecting mothers. The forum is profiled as a mom-to-
mom community where members can offer support and advice to each other on
parenting matters. Clusters of message boards generating thousands of messages
3.2. Methodology
This study combines a quantitative analysis of the online nicknames in terms of
their information content with a qualitative analysis of users’ comments on nick-
name meaning and construction.
The nicknames from the two forums were first analyzed quantitatively in terms
of their information content. Adapting Bechar-Israeli’s (1995) typology of IRC
nicknames,4 the nicknames were systematically classified based on their in-
formation content as follows:
1. Real names as nicknames
2. Self-related nicknames
3. Nicknames related to flora, fauna, objects, abstract entities
4. Nicknames related to medium, technology, and lack of identity
5. Nicknames related to figures in literature, films, fairytales, and famous people
6. Unclassifiable nicknames
The four Gricean maxims of conversation were adapted in terms of the nick-
names’ a) usage of apparent real name and/or personal information (quality); b) in-
formative content (quantity); c) relevance to the central, overarching theme of the
forum’s discussions (relevance); and d) clarity of content (manner). The online
nicknames were then coded according to the adapted maxims.
Subsequently, qualitative analysis was conducted of users’ comments on nick-
name meaning and construction. I considered how users discuss the meanings,
uses, and effects of the online nicknames in the two communities in order to gain
insight into what they think the norms of nickname construction in these commu-
nities should be.
4.1. Quality
Following the maxim of quality, a nickname contribution should be one that is true
and not false, as otherwise the maxim is violated. The users in the forum data make
different claims about their identity through their nicknames, which reveal per-
sonal names, gender, affiliation, profession, physical attributes and other self-
related information, e.g., “imahockeymom”, “bikerchic”, “badgewife”, “Cool-
ToolGuy”, “brown_eyes”, “JackieS”, “photo_luver”. In fact, 77 % of the online
nicknames from the parenting forum and 64 % from the photography forum are
self-referential, that is, they pertain directly to the self, with the two most promi-
nent categories being self-related names (44 %, 28 %) and apparent real names
(33 %, 36 %), as Table 1 indicates. The latter category includes first names, (e.g.,
“kevin”), last names (e.g., “mills”), full names (e.g., “maggie_brown”), first names
and initials (e.g., “Mike_R”), and initials and last names (e.g., “j_minard”).
Parent.com Photo.com
Self-referential 423 (77 %) 352 (64 %)
Self-related 242 (44 %) 153 (28 %)
Real names 242 (44 %) 199 (36 %)
Non-self-referential 127 (23 %) 198 (36 %)
4.2. Quantity
The maxim of quantity as applied to nicknames states that they should be as in-
formative as is required for the purposes of the exchange. Apart from self-related
nicknames and real names as nicknames, the analysis of the two nickname corpora
shows that the remaining nicknames can be classified into four additional cat-
egories according to their information content (see Table 2).
Unclassifiable nicknames are those that were impossible to decipher. Such
nicknames are constructed from strings of letters and/or numbers that are obscure
in meaning, sometimes on purpose:
(12) From kwerksly:
I know we did this a while ago on the Expecting Board, but since that time lots
of new people have joined us, and I thought it mihgt be fun to do again. What
is the meaning behind your screen name? Or is there one? Mine has to remain
a bit cryptic, unfortunately, because it comes out of a password …
(Parent7:1)
pecially the case in debate boards in the parenting community, where various is-
sues are intensely discussed. Many debate boards in this community operate on the
basis of dichotomies7 such as for vs. against vaccination parents, stay-at-home vs.
working mothers, and breast-feeding vs. formula-feeding mothers. In the context
of such debates, online nicknames that include some reference to any of these cat-
egories signal right away which group users belong to. Statements about the self in
online names can also cause negative reactions, as shown in example (13):
(13) From Radtech:
(…) there are some sig lines and even usernames I find offensive on a relig-
ious basis.
(Photo1:12)
This raises the question of whether, and if so, how, nicknames can be overin-
formative. Grice (1975: 46) comments that the notion of overinformativeness is
disputable, the only way to define it being with the help of the maxim of relation.
That is, it is only in a certain context that nicknames can be regarded as overin-
formative. In the case of online nicknames, that could be the overall multimodal
context if, for instance, nicknames, signature files, and avatars communicate the
same or overlapping information about a user. Apart from irrelevant information
included in the mini-proposition, users may also consider very long nicknames
overinformative (cf. manner).
Moreover, online nicknames that were sufficiently informative at one time may
not be considered so at a later time. This applies to nicknames that are changed be-
cause they no longer convey accurate (quality) and/or sufficient (quantity) in-
formation about the users and thus are regarded as underinformative and untruth-
ful. In the parenting forum, such changes of information content may occur in
nicknames that indicate parental identity:
(14) From milniks7:
Just wanted to let everyone know that I changed my name from sixfornow to
milniks7 … we obviously are no longer six for now.
(Parent14:1)
(15) From fivekdz4me:
Since I haven’t posted here in awhile first of all I changed my member name
from dragonand3pixies to the present one. I am SAHM to five wonderful kids.
(Parent15:1)
This practice is encouraged by other members of the parenting community, as il-
lustrated in a reply post to a user who has just found out she is pregnant:
(16) From rainbow:
Congratulations! You may have to consider changing your username. ;)
(Parent24:8)
Participants in the parenting forum are well aware of the sensitivity of certain dis-
cussions and often show caution, for example, by selecting online nicknames that
do not express parental identity overtly or at all. This also appears to be a legit-
imate reason within the parenting community for having more than one nickname,
as is exemplified in the message below from a discussion where the issue of having
multiple nicknames was addressed:
Thus, the decision about how much and what kind of information is appropriate to
include in an online nickname may depend not only on what users wish to com-
municate about themselves, but also on online social constraints.
While most forum participants are informative about their personal identity, es-
pecially their parental identity, in their nicknames, others choose to be informative
about other aspects of themselves:
Table 3. Female, male, and gender indeterminate nicknames in the two corpora
4.3. Relation
Grice (1975: 46) observed that it can be difficult to apply the maxim of relation or
relevance due to the different foci of relevance that occur in the course of an inter-
action. In view of the shifting topics of online exchanges, the maxim of relation is
also the least straightforward to apply to nicknames, which remain static after their
first appearance in the interaction. In this section, the relevance of an online nick-
name is considered in relation to the topic of the group as a whole: It is assumed
that the nicknames in the data should relate to parenting or photography.
Since the central theme of the discussions in the mother support forum is par-
enting, the users who align their online nicknames to this overarching topic
broadly observe the maxim of relation. In the corpus, 34 % of the nicknames are
explicitly related to the parenting theme of the site. These nicknames are found
mostly in the first three categories of the nickname typology (Table 2) and make
reference to the parental role through self-description and/or reference to children,
e.g., “momof2petite_filles”, “handymom”, and “my3girls”. In the photography
forum, 9 % of the nicknames are photography related. These nicknames, which are
found in the second, third, and fourth categories of the typology, are self-descrip-
tive or refer to objects and concepts related to photography, e.g., “EoSD30fReAk”,
“KO_300D”, and “35mm”.
Apart from the global adherence to relevance, certain online nicknames in
the two forums seem to reflect local topics: For example, nicknames that were
collected from discussion boards addressing topics about large families emphasize
the number of children, e.g., “mommyoftwinstwice”, “peanutmomof4”, “thekid-
farmx6”, “geminimamato4”, and “momto6qts”. Similarly, some nicknames col-
lected from discussion boards on macro photography reflect this particular interest,
e.g., “macroshooter”, “macrojunkie”, and “snapmacro”. This raises the question of
whether users change their nicknames depending on the topic they contribute to.
No evidence was found to suggest change of nicknames to make them locally rel-
evant. This may be attributed to nickname stability, which is highly valued in the
observed communities in order to ensure continuity and promote trust. As illus-
trated in the previous section, however, participants in the mother support forum
may change their nicknames locally to participate in discussion boards that deal
with sensitive issues or globally to indicate a new addition to the family.
As Levinson (1983: 111) observes, it is difficult to interpret something as irrel-
evant. Interactants always try to find relations between what is stated and what is
implied or might be implied (but cf. Herring, this volume). The central theme of
the two communities prompts users to interpret each other’s online nicknames as
group-relevant. If an online nickname is not explicitly related to parenting or
photography, users do not assume it is irrelevant, but rather attempt to infer its
meaningful relation to the topic of the forum:
(21) From susannahk:
To snifflesally:
After all, I’m going to leap to the conclusion that your family is prone to colds
and that’s how you chose your username.
(Parent8:55)
Indeed, unless users specify the irrelevance of their online nicknames to par-
enting or photography, all nicknames are subject to be interpreted as relevant to the
forum topic. There are users within the two communities who do not opt for par-
enthood or photography-related nicknames and state it clearly in their posts on dis-
cussions about nicknames:
4.4. Manner
Questions of brevity, clarity, and (lack of) ambiguity are particularly important in
the case of online nicknames, because they affect nickname interpretation. The
maxim of manner bears on how an online nickname is made up to convey what it
says. In the data, the nicknames range from simple one-word constructions such as
“sild”, “dana0”, and “duder” to longer and more complex nicknames, such as “BIG-
TUFFGUY”, “CorruptedPhotographer”, “momsurroundedbyboys”, “happyand-
joyfulmom”, and “chase_is_my_sonshine”. Users who opt for phrasal and clausal
constructions in their online nicknames presumably do so for a reason, such as to
provide more information, which discloses a connection between the manner and
the quantity maxims. The following post shows a user providing reasons for
breaching the submaxim of brevity:
(30) From montessorimommysoon:
montessorimommysoon
Long, I know. I teach at a Montessori school (middle school) and I love the
family sensibility that it fosters. So “montessori” … and I can’t wait to be a
“mommy soon”!!
(Parent1:28)
Play with non-standard spelling in nicknames, which has its origins in the prac-
tices of hackers (Bechar-Israeli 1995; Danet 2001: 27), is related to the submaxim of
clarity. Some users in the two discussion forums manipulate spelling in their nick-
names, as in “Grampaw”, “photo_luver”, “whyzzerd”, “cutemom4u”, “mom2kay-
lantristan”, “luvmyboys”, “m_a_r_i_n_a”, and “cambriagurl”. However, non-
standard spelling is found in only 7 % of the nicknames in each forum. The major-
ity of the users do not resort to any language play in constructing their nickname, a
result that resonates with the findings of Bechar-Israeli’s study of IRC nicknames.
Users sometimes provide explanations for non-standard spellings, as shown in the
examples below:
(31) From pumptin:
When dd [‘dear daughter’, LL] was little she couldn’t say pumpkin, she said
pumptin, and we thought it was the cutest thing ever!
(Parent11:9)
(32) From baldylox:
well im fairly bald, have been so for over 10yrs. one of my female friends
thought my chrome dome was sexy as hell and a buddy said since i couldnt
possible have nice locks of hair might as well call me “baldy locks of no hair”
Well it got shortened and changed the spelly for style :D
(Photo2:362)
(33) From karfeef:
Car+Theif. a common misconception about those, who like myself, reside in
Liverpool UK =)
(Photo1:57)
parenting or photography that gave rise to the use of a distinctive vocabulary. Users
transfer elements from this vocabulary to the construction of their nicknames as a
convenient way to say something about who they are. The use of these abbrevi-
ations also demonstrates knowledge of the abbreviation norms of the community,
which can indicate in-group identity.
5. Discussion
The point of departure of this chapter is that online nicknames constitute communi-
cative efforts that can be analyzed in the light of Grice’s Cooperative Principle.
Users treat them as mini-propositions and disclose the existence of shared maxims
of online nicknames. The online nicknames observed in the communities in the
case study were found to orient to all four Gricean maxims to some extent. In terms
of the quality maxim, participants appeared to find truthfulness and nickname sta-
bility important. A possible interpretation of this observation is that a stable and
truthful identity helps to maintain social relationships online. However, as shown
above, nickname changes in the mother-support forum seem to be acceptable under
certain circumstances, e.g., participation in discussions on sensitive topics or a new
addition to the family. In view of the maxim of quantity, participants seem to pro-
vide the amount of information about themselves they think is necessary, which
may include gender, interests, number of children, or photography equipment.
Gender referencing is a crucial component of many online nicknames, manifested
through stereotypical self-descriptions associated with masculinity and femininity.
The informativeness of a nickname may be subject to change if participants feel the
need to adjust the mini-proposition to the particular context.
Regarding the maxim of relation, participants in the mother-support forum
often observed relevance in their selection of topic-related nicknames to showcase
their parental identity. The fact that parenthood is such a central locus of identity
can account for this tendency. In the photography forum, however, photography-
related nicknames are not the preferred choice. With respect to the maxim of
manner, whether or not users succeed in interpreting each other’s nicknames dep-
ends on several factors, including familiarity with abbreviation norms and prac-
tices within the particular community.
The application of the maxims to the propositional content of online nicknames
is not unproblematic, however. Despite the high value placed on sincerity among
members in the two forums, there is always the possibility of users manipulating
their identity. Thus, one cannot be sure whether or not participants observe the
maxim of quality in their nicknames. The maxim of quantity raises the issue of
what is as informative as required and by whom, since participants in different for-
ums may value informativeness differently. Further, the maxim of relation is hard
to apply because nicknames can be regarded as static in interaction. Nonetheless,
nicknames have been shown to vary according to context when users are con-
fronted with sensitive topics. Moreover, participants readily interpret nicknames as
relevant, even when the nicknames themselves do not support such an interpre-
tation. As for the manner maxim, some nicknames remain cryptic unless their bear-
ers explicate them, and what is clear depends on the individual participants’ back-
ground knowledge.
The main insight provided by this case study is that online nicknames have
more than a mere referential function for self-presentation purposes. They can also
have an expressive function, conveying attitudes, feelings, and values; and a poetic
function, revealing a world of imagination and playfulness in the use of metaphors
and language play. Previous studies have singled out the discursive function of
nicknames, identifying ways in which they create coherence in the context in
which they are used. Online nicknames can also be a means of persuasion and a
claim to in-group membership, for example, when users reference a shared inter-
est, such as parenting, or a hobby, such as photography. The mini-propositions in-
cluded in nicknames spark interaction and serve to establish credentials for what
their bearers are saying.
6. Concluding remarks
This chapter investigated the extent to which conversational maxims can be ap-
plied to online nicknames, proposing that nicknames can be seen as mini-proposi-
tions that constitute cooperative efforts. An analysis of online nicknames and par-
ticipants’ comments on two online forums was carried out, concentrating on how
users negotiate meaning through online nicknames and orient to the maxims
through the mini-propositions included in them. The findings show that partici-
pants in the two forums opt for the use of apparently real names, broadly observe
relevance, provide information that they think is necessary to communicate about
themselves, and value clarity in their online names.
The limitations of this study include the fact that only two online forums were
examined. Earlier studies have found that communities can be idiosyncratic as re-
gards their nicknaming conventions. The two forums examined here manifest dif-
ferent conventions related to their topics and populations of users. Comparisons
with the nicknaming practices of other online communities are needed to account
for nickname construction and use. Future work might usefully explore nicknames
in conjunction with signature files and avatars. In a broader context, the study of
online nicknames is an important part of online ethnography. Users value nick-
names as interactional contributions far beyond the ritual act of selecting one. For,
to paraphrase the Roman proverb, sine nomine persona virtualis non est.10
Acknowledgements
I wish to thank the editors and the reviewers for their comments and suggestions. I
am also grateful to Urpo Nikanne and Brita Wårvik for discussion and comments
on an earlier version.
Notes
1. Other terms for online nicknames include pseudonyms, screen names, handles, user-
names, and member names. In this chapter, the term online nickname is used and is de-
fined as a name that a user selects for purposes of online interaction.
2. Bechar-Israeli (1995) gives an example of a nickname game in an IRC discussion on
horses. During the discussion, users changed their nicknames to animal names and
simulated animal noises before switching back to their original nicknames.
3. The names of the two forums have been changed to protect the privacy of the users
whose online names are analyzed. Any similarities to other online forums are uninten-
tional.
4. Language play is not included as a category in the adaptation of the typology but is dis-
cussed separately in the section on manner.
5. The posted messages are reproduced as they appear in the message board discussions.
For the purposes of the present study, only the relevant parts of the posting contributions
are cited in the examples. Each post is coded according to forum of origin, number of
discussion thread in the corpus, and number of post within the thread, e.g., Parent5:20.
6. “Ukifune” is a figure in the novel The Tale of Genji, “Snow White” is a female character
in the fairy tale “Snow White and the Seven Dwarfs”, and “Agent99” is a female char-
acter from the American TV series “Get Smart”.
7. Some boards make this explicit in the wording of the boards’ topics. In other cases, such
dichotomies emerge in the course of a discussion.
8. No similar cases were found in the photography group.
9. Although bif is used primarily to refer to bird-in-flight photography, some participants
use it to indicate photographs of bees in flight.
10. The original Roman legal proverb goes: Sine nomine persona non est (‘no name, no per-
son’). The author’s paraphrase is as follows: ‘No name, no virtual person’.
References
Bays, Hillary
1998 Framing and face in Internet exchanges: A socio-cognitive approach. Linguis-
tik Online 1(1). http://www.linguistik-online.de/bays.htm
Bechar-Israeli, Haya
1995 From <Bonehead> to <cLoNehEAd>: Nicknames, play, and identity on Inter-
net Relay Chat. Journal of Computer-Mediated Communication 1(2).
http://jcmc.indiana.edu/vol1/issue2/bechar.html
Bertills, Yvonne
2003 Beyond identification: Proper names in children’s literature. Ph.D. disser-
tation, Department of Finnish Language, Åbo Akademi University, Finland.
Brown, Gillian and George Yule
1988 Discourse Analysis. Cambridge, UK: Cambridge University Press.
Bruckman, Amy S.
1993 Gender swapping on the Internet. In: Proceedings of INET93.
http://www.cc.gatech.edu/elc/papers/bruckman/gender-swap ping-bruckman.pdf
Bruckman, Amy S.
2006 Teaching students to study online communities ethically. Journal of In-
formation Ethics 15(2): 82–98.
Byrne, Elizabeth
1994 Cyberfusion: The formation of Internet relationships on Internet Relay Chat.
Ph.D. dissertation, Faculty of Humanities and Social Sciences, University of
Western Sydney. http://irchelp.org/irchelp/communication-research/academ ic/
byrne-e-cyberfusion-1993/thesis1.html
Calvert, Sandra L., Brian A. Mahler, Sean M. Zehnder, Abby Jenkins, and Mickey S. Lee
2003 Gender differences in preadolescent children’s online interactions: Symbolic
modes of self-presentation and self-expression. Journal of Applied Develop-
mental Psychology 24(6): 627–644.
Campbell, John E.
2004 Getting it on Online: Cyberspace, Gay Male Sexuality, and Embodied Identity.
New York: Haworth.
Chester, Andrea and Gwynne Gillian
1998 Online teaching: Encouraging collaboration through anonymity. Journal of
Computer-Mediated Communication 4(2).
http://jcmc.indiana.edu/vol4/issue2/chester.html
Crystal, David
2001 Language and the Internet. Cambridge, UK: Cambridge University Press.
Curtis, Pavel
1992 MUDding: Social phenomena in text-based virtual realities. In: Proceedings of
the 1992 Conference on Directions and Implications of Advanced Computing.
ftp://ftp.lambda.moo.mud.org/pub/MOO/papers//DIAC92.txt
Danet, Brenda, Lucia Ruedenberg-Wright, and Yehudit Rosenbaum-Tamari
1997 “Hmmm … Where’s that smoke coming from?” Writing, play and perform-
ance on Internet Relay Chat. Journal of Computer-Mediated Communication
2(4). http://jcmc.indiana.edu/vol2/issue4/danet.html
Danet, Brenda
2001 Cyberpl@y: Communicating Online. Oxford: Berg.
Deul, Nancy
1996 Our passionate response to virtual reality. In: Susan C. Herring (ed.), Com-
puter-Mediated Communication: Linguistic, Social, and Cross-Cultural Per-
spectives, 127–146. Amsterdam/Philadelphia: John Benjamins.
Giese, Mark
1998 Self without body: Textual self-presentation in an electronic community. First
Monday 3(4). http://www.firstmonday.org/issues/issue3_4/giese/
Goffman, Erving
1959 The Presentation of Self in Everyday Life. New York: Doubleday.
Grice, H. Paul
1975 Logic and conversation. In: Peter Cole and Jerry L. Morgan (eds.), Syntax and
Semantics 3: Speech Acts, 41–58. New York: Academic Press.
Hall, Kira
1996 Cyberfeminism. In: Susan C. Herring (ed.), Computer-Mediated Communi-
cation: Linguistic, Social, and Cross-Cultural Perspectives, 147–170. Amster-
dam/Philadelphia: John Benjamins.
Herring, Susan C.
1999 The rhetorical dynamics of gender harassment on-line. The Information So-
ciety 15(3): 151–167.
Herring, Susan C., Lois A. Scheidt, Inna Kouper, and Elijah Wright
2006 A longitudinal content analysis of weblogs: 2003–2004. In: Mark Tremayne
(ed.), Blogging, Citizenship, and the Future of Media, 3–20. London: Rout-
ledge.
Horn, Lawrence R.
2004 Implicature. In: Lawrence R. Horn and Gregory L. Ward (eds.), The Handbook
of Pragmatics, 3–28. Oxford, UK: Blackwell.
Huffaker, David A. and Sandra L. Calvert
2005 Gender, identity, and language use in teenage blogs. Journal of Computer-
Mediated Communication 10(2), article 1.
http://jcmc.indiana.edu/vol10/issue2/huffaker.html
Jacobson, David
1996 Contexts and cues in cyberspace: The pragmatics of naming in text-based vir-
tual realities. Journal of Anthropological Research 52(4): 461–479.
Jacobson, David
1999 Impression formation in cyberspace: Online expectations and offline ex-
periences in text-based virtual communities. Journal of Computer-Mediated
Communication 5(1). http://jcmc.indiana.edu/vol5/issue1/jacobson.html
Jaffe, Michael J., Young-Eum Lee, Li-Ning Huang, and Hayg Oshagan
1995 Gender, pseudonyms, and CMC: Masking identities and baring souls. Paper
presented to the annual conference of the International Communication As-
sociation, Albuquerque, New Mexico.
http://research.haifa.ac.il/~jmjaffe/genderpseudocmc/
Kelly, Deirdre M., Shauna Pomerantz, and Dawn H. Currie
2006 “No boundaries”? Girls’ interactive, online learning about femininities. Youth
& Society 38(1): 3–28.
Kendall, Lori
2002 Hanging Out in the Virtual Pub: Masculinities and Relationships Online. Ber-
keley, CA: University of California Press.
Kennedy, Helen
2006 Illegitimate, monstrous and out there: Female Quake players and inappropriate
pleasures. In: Joanne Harlows and Rachel Moseley (eds.), Feminism in Popu-
lar Culture, 183–201. London: Berg.
Levinson, Stephen C.
1983 Pragmatics. Cambridge, UK: Cambridge University Press.
Liu Geoffrey Z.
1999 Virtual community presence in Internet Relay Chatting. Journal of Computer-
Mediated Communication 5(1). http://jcmc.indiana.edu/vol5/issue1/liu.html
Lyons, John
1977 Semantics, Vol. I. Cambridge, UK: Cambridge University Press.
McKenna, Katelyn Y.A., Amie S. Green, and Marci E. J. Gleason
2002 Relationship formation on the Internet: What’s the big attraction? Journal of
Social Issues 58(1): 9–31.
Mann, Chris and Fiona Stewart
2000 Internet Communication and Qualitative Research: A Handbook for Research-
ing Online. London: Sage.
Morgan, Jane, Christopher O’Neill, and Rom Harré
1979 Nicknames, their Origins and Social Consequences. London: Routledge and
Kegan Paul.
Pagnucci, Gian S. and Nicholas Mauriello
1999 The masquerade: Gender, identity, and writing for the web. Computers and
Composition 16(1): 141–151.
Panyametheekul, Siriporn and Susan C. Herring
2003 Gender and turn allocation in a Thai chat room. Journal of Computer-Mediated
Communication 9(1). http://jcmc.indiana.edu/vol9/issue1/panya_herring.html
Parks, Malcolm R.
1996 Making friends in cyberspace. Journal of Computer-Mediated Communication
1(4). http://jcmc.indiana.edu/vol1/issue4/parks.html
Reid, Elizabeth M.
1991 Electropolis: Communication and community on Internet Relay Chat. Honours
thesis, Department of History, University of Mebourne.
http://www.irchelp.org/irchelp/misc/electropolis.html
Rheingold, Howard
2000 The Virtual Community: Homesteading on the Electronic Frontier. Cambridge,
MA: Massachusetts Institute of Technology Press.
Rintel, Sean E., Joan Mulholland, and Jeffery Pittam
2001 First things first: Internet Relay Chat openings. Journal of Computer-Mediated
Communication 6(3). http://jcmc.indiana.edu/vol6/issue3/rintel.html
Rodino, Michelle
1997 Breaking out of binaries: Reconceptualizing gender and its relationship to lan-
guage in computer-mediated communication. Journal of Computer-Mediated
Communication 3(3). http://jcmc.indiana.edu/vol3/issue3/rodino.html
Ruedenberg, Lucia, Brenda Danet, and Yehudit Rosenbaum-Tamari
1995 Virtual virtuosos: Play and performance at the computer keyboard. Electronic
Journal of Communication 5(4).
http://www.cios.org/EJCPUBLIC/005/4/00542.HTML
Simpson, James
2005 Conversational floors in synchronous text-based CMC discourse. Discourse
Studies 7(3): 337–361.
Sperber, Dan and Deirdre Wilson
1995 Relevance: Communication and Cognition. Oxford, UK: Blackwell.
1. Introduction
use of asterisks for stress (e.g., “please call – we’ve *got* to discuss”), and, more re-
cently, icons (e.g., “>:(”, which shows a sad face when rotated ninety degrees).
1) Emoticons
2) Non-standard spelling and creative use of writing systems
3) Abbreviation
4) Non-standard punctuation
2. Relevant literature
kind of “Internet language”. A further distinction can be made between works fo-
cusing on CMC in one language and studies or edited volumes that deal with CMC
in more than one language.
2.2. Works addressing CMC in one language vs. more than one language
2.2.1. Works addressing CMC in one language
Much of the available scholarly literature on language use in CMC is based on
studies that address CMC in one language only. Particularly in the early days of
linguistic CMC research, scholarly works on CMC written in English often meant
research on CMC in English, often without stating explicitly that English-based
CMC was the focus of attention. For example, Danet et al.’s (1997) study of writ-
ing, play, and performance on IRC does not state in the title, the abstract, or the in-
troductory section that the study is only about English. Danet and Herring (2007b:
5) remark that “[m]ost researchers publishing in English venues have generalized
about the language of computer-mediated communication, whereas in fact they
were describing computer-mediated English, sometimes in a single CMC mode”.
They rightfully complain that much of the CMC research published in English
does not do justice to the fact that the majority of Internet users are not native
speakers of English and that “hundreds of millions of people are already partici-
pating online today in languages other than English, in some form of nonnative
English, or in a mixture of languages, and this trend is projected for the years to
come” (Danet and Herring 2007b: 4).
Although Su (2007) claimed that there is a “relative lack of research on com-
puter-mediated communication (CMC) in non-English-based Internet environ-
ments”, in fact there is a substantial and growing body of research on language use
in non-English-based CMC, much of it published in languages other than English
(see Schlobinski 2001; Stein 2003). For example, the collection of articles in
French edited by Anis (1999a) focuses on language use in French CMC, Swedish
CMC is addressed in Josephson (1997) and Hård af Segerstad (2002), and German
CMC is featured in several volumes: Weingarten (1997), Kallmeyer (2000), Beiß-
wenger (2001), Siever, Schlobinski, and Runkehl (2005), Schlobinski (2006), and
Androutsopoulos, Runkehl, Schlobinski, and Siever 2006).
Unlike many of the studies investigating language use in English-based CMC,
most works on non-English-based CMC published in English explicitly indicate
the language they focus on in the title, such as Anis’ (2007) article “Neography:
Unconventional spelling in French SMS text messages” and Hård af Segerstad’s
(2005) study “Language use in Swedish mobile text messaging” (see also Herring
and Zelenkauskaite 2008; Ling 2005; Nishimura 2007, Siebenhaar 2006; Su 2007;
Tseliga 2007; Yang 2007). However, CMC research has reached a point where par-
ticularly authors publishing in English need to indicate clearly whether they are
“describing computer-mediated English” (Danet and Herring 2007b: 5), whether
they are investigating language use in CMC or modes of CMC in a language or par-
ticular languages other than English, or whether they are attempting to make more
universal claims as to patterns of language use in CMC in all languages.
3.1. Emoticons
The text-based nature of CMC and the resulting paucity of paralinguistic and non-
linguistic cues are generally considered to be the main reasons for the development
of so-called emoticons (Crystal 2001; Danet 2001; Danet et al. 1997; Döring 2003;
Murray 1988; Raymond 1994; Runkehl et al. 1998; Schmitz 2004; Siever 2006;
Witmer and Katzman 1997). As all textual modes of CMC share the lack of certain
channels of communication, it does not come as a surprise that emoticons are a fea-
ture that is “shared across modes” (Danet and Herring 2007b: 12; see also Hård af
Segerstad 2002: 261) and that “recur[s] in languages besides English” (Danet and
Herring 2007b: 27). Emoticons are thus among the micro-level phenomena which
are most “characteristic” of language use in CMC, although their actual shape and
frequency of use differ according to a number of factors. As smiley face icons pre-
date the Internet, and emoticons can now also be found in “snail-mail” letters and
postcards (Östman 2004; Runkehl et al. 1998), it seems more appropriate to clas-
sify them as “characteristic” than to call them “a unique feature of the electronic
language register” (Tseliga 2007: 121; see also Herring 1996) or an “exclusively
digital phenomenon” (Tseliga 2007: 126).
The term “emoticon” is most likely a blend of the nouns “emotion” and “icon”
(Oxford English Dictionary Online 2008), and it is defined as “[a] representation of
a facial expression formed by a short sequence of keyboard characters (usually to
be viewed sideways) […] to convey the sender’s feelings or intended tone” (Ox-
ford English Dictionary Online 2008). Emoticons are also known as “smileys”,
“smiley icons”, or “smiley faces” (Crystal 2001; Danet 2001; Danet and Herring
2007b; Haase et al. 1997; Runkehl et al. 1998; Wolf 2000). Godin (1993) and San-
derson (1993) present early collections of emoticons. Today, some of the clusters
of typographic symbols referred to as emoticons do not represent facial ex-
pressions (Dresner and Hering 2010); for example, many contemporary online dis-
cussion forums offer a selection of symbols, such as or (Bieswanger and
Intemann 2011). It should also be noted that much of the current CMC software
automatically transforms typed sequences such as :) ‘smile’ and :( ‘frown’ into
graphical icons, in this case into the icons K and M respectively, so that emoticons
no longer necessarily have to be read sideways.
In contrast to traditional “Western-style ‘smileys’” (Danet and Herring 2007b:
13), defined as in the quote from the Oxford English Dictionary Online above, “kao-
moji” or “Japanese-style emoticons” (Katsuno and Yano 2007; see also Nishimura
2007; Shirai 2006; Siever 2006), which are also called “Asian style” emoticons
(Lee 2007: 203) or “[v]ertical emoticons” (Lee 2007: 205), have never had to be
read sideways (Nishimura 2007; Shirai 2006). For example, (^_^), (^ ^), (^^), and
(^o^) are among the many different vertical emoticons representing a smile (Kat-
suno and Yano 2007; Schlobinski and Watanabe 2003). Lee (2007: 205) reports that
both Western-style and Asian-style emoticons occur in instant messaging and email
communication in Hong Kong, noting that “[v]ertical emoticons are more common
in Hong Kong CMC than are horizontal ones”. Palfreyman and Al Khalil (2007)
give an example of the use of vertical emoticons in instant messaging in Arabic.
From the point of view of pragmatics and speech act theory, it has recently
been argued that emoticons do not only convey emotions but in fact have multiple
functions, including to “indicate the illocutionary force of the text to which they
are attached” (Dresner and Herring 2010: 250). Dresner and Herring identify
“three ways in which emoticons function: 1) as emotion indicators, mapped di-
rectly onto facial expression; 2) as indicators of non-emotional meanings, mapped
conventionally onto facial expressions, and 3) as illocutionary force indicators that
do not map conventionally onto a facial expression” (250).
Linguistic studies have repeatedly addressed the relationship between the use
of emoticons and sociolinguistic factors, particularly gender, as well as CMC
mode. Witmer and Katzman (1997) found in their study of language use in public
newsgroups and special interest groups that females used more graphic accents
than males. Additionally, they reported that the overall use of graphic accents in
their data was fairly low. Baron (2008: 67) investigated instant messaging com-
munication of American college students and found that “females were far more
likely to use emoticons than males”. In a study of gender-related differences in
emoticon use in online newsgroups, Wolf (2000) also found that females used more
emoticons overall and additionally discovered that males’ use of emoticons was
notably higher in mixed-gender newsgroups than in predominantly male news-
groups. This finding is consistent with the results of Bieswanger and Intemann
(2011), who found that the frequency of emoticons was considerably higher in a
predominantly female forum than in a predominantly male English-language dis-
cussion forum.
Regarding the connection between the use of emoticons and the CMC mode
and/or language employed, studies point to considerable mode-related and lan-
guage-related variation, but more research is necessary before any general con-
clusions can be drawn. There “is a pressing need for systematic cross-linguistic
studies that make use of similar methods in similar contexts involving different
languages” (Danet and Herring 2007b: 28). As regards mode, Lee (2007: 203)
reports that emoticons are “very popular” in Hong Kong CMC, but that in her data
emoticons are much more frequent in instant messaging than in email. In a study of
language use in instant messaging and texting among female American college
students, Baron (2008) found that emoticons were very infrequent in both modes,
making up less than 1 % of all words in each mode. Bieswanger and Intemann
(2011) report that for a corpus of over 82,000 words from three English-based on-
line discussion forums, on average approximately 2 % of the words were emoti-
cons. There are, however, considerable differences across individual discussion
forums in the Bieswanger and Intemann study, with an average of about 0.3 % of
all words being emoticons in one forum and 3.1 % of all words in another. Tseliga’s
(2007: 126) study of four Greek discussion lists and one newsgroup discovered
that “Greeklish users were somewhat more likely to use emoticons” than users of
Greek. Hård af Segerstad (2002: 261) investigated the use of “features character-
istic of CMC” in four different modes, namely email, chat, instant messaging, and
texting, and found that emoticons occurred in each of the modes.
Based on the above review of the available literature on the functions and use
of emoticons in CMC, it seems safe to conclude that emoticons are a characteristic
feature of language use in CMC. However, considerable variation exists related to
the language and script used, the CMC mode employed, the purpose of the inter-
action, and the sociolinguistic backgrounds of the users.
Other recent studies addressing the issue of non-standard spelling and the cre-
ative use of writing in CMC in different languages include Nastri et al. (2006),
Kessler (2008), Anis (2007), and Su (2007). Nastri et al. (2006) found in their
speech act analysis of “away” messages in English instant messaging that “CMC-
based orthography” is very frequently used, with “39 % of away messages contain-
ing some type of CMC orthography”. In her study of instant messaging in Swiss
German and Standard German, Kessler (2008) found that non-standard spelling
and writing were frequently used to imitate spoken language, including via ono-
matopoeic forms, reduplication of letters, and the representation of dialect features
(see also Siebenhaar 2006). The analysis of the data also showed that the com-
munication took place to a large extent in lower case letters only, and that capital
letters were mostly used for emphasis (Kessler 2008). Anis (2007: 97) studied
French texting and developed “a typology of neographic transformations observed
in the French SMS corpus”. In a comparison across modes, Anis (2007: 90) found
that neographic spellings were frequent in synchronous French chat, text messag-
ing, and instant messaging, whereas neography in asynchronous email, newsgroup
postings, and discussion list postings was marginal. Finally, Su (2007: 83–84) in-
vestigated “creative uses of writing” on Taiwanese Bulletin Board Systems (BBS)
and identified “processes of transforming spoken English, Taiwanese, and Taiwan-
ese-accented Mandarin into forms of online language play” that involved, among
other strategies, non-standard mixing of writing systems from different languages.
In short, research has found that various forms of non-standard orthography
and typography as well as creative uses of writing are common in CMC. As with
emoticons, however, there is considerable variation in the use of these forms ac-
cording to the mode of CMC – including (a)synchronicity – the language used, the
writing system(s) available, the purpose and formality of the interactions, the in-
tention of the writers, and the sociolinguistic backgrounds of the users.
3.3. Abbreviation
Abbreviated forms of language have been identified as characteristic of communi-
cation in many modes of CMC and as among the “distinctive CMC features
[which] recur in languages besides English” (Danet and Herring 2007b: 27; see
also Anis 2007; Crystal 2001; Danet 2001; Deumert and Masinyana 2008; Thur-
low and Brown 2003). “Abbreviation” and “shortening” are neutral terms here used
synonymously to refer to all strategies that result in “lexical forms that are made
up by fewer characters than the full form of a word or a combination of words”
(Bieswanger 2007: n.p.). Hård af Segerstad (2002: 71) calls these forms “lexical
reductions”, and Herring and Zelenkauskaite (2008: 82–84) refer to them as “dele-
tions”. Many types of shortenings occur in CMC, for example, initialisms4 such as
lol for ‘laughing out loud’, vowel deletion as in ovr for English ‘over’ or tjrs for
French ‘toujours’ (‘always’), contractions as in don’t for English ‘do not’, and
letter/number homophones such as b for English ‘be’ or 2 for English ‘to’ or ‘too’,
also known as rebus writing (Anis 2007; Baron 2000; Bieswanger 2007; Crystal
2001; Hård af Segerstad 2002; Herring, in press; Kessler 2008; Schlobinski et al.
2001; Werry 1996).
Crystal (2001: 84) writes that “[t]he various types of abbreviation found in Net-
speak have been one of its most remarked features”; for texting, he claims that
“[t]he challenge of the small screen size and its limited character space (about 160
characters), as well as the small keypad, has motivated the evolution of an even
more abbreviated language than emerged in chatgroups and virtual worlds” (229).
Thurlow and Brown (2003, n.p.) expected such a result in their study of English
texting, as “[h]eavily abbreviated language is of course also a generic feature of in-
teractive CMC niches like IRC’s online chat and ICQ’s instant messaging”. They
were surprised, however, to find that shortenings made up less than 20 % of the
content of the messages in their data, casting some doubt on Crystal’s broad claim.
Deumert and Masinyana (2008: 125) give a similar figure of approximately 20 % in
their study of text messaging in South Africa in English and mixed English-isi-
Xhosa messages for “abbreviations, phonological approximations, and non-stan-
dard spellings”.
In a contrastive study of English and German private texting, Bieswanger
(2007) found that the overall number of shortenings used in the German data was
even lower than in the English corpus. His quantitative analysis found an average
of 5.57 shortenings per text message in the English data, as compared to only 0.86
shortenings per message in the German corpus. The frequency of shortenings in the
German data is so low that it calls into question whether shortenings are in fact a
“characteristic” feature of German texting. Kessler’s (2008) study of German in-
stant messaging yielded a similar result: Only 742 of the 45,729 words in her data
were abbreviated forms, i.e., less than 2 % (Kessler 2008). Herring and Zelenkaus-
kaite (2008) adopted the shortening types from Bieswanger (2007) in their study of
Italian iTV SMS – except for the type “contractions”, “because the Italian language
does not have contractions” (84) – and found that the overall frequency of short-
enings was 1.79 per message for female senders and 1.18 for male senders, i.e., the
average number for both genders is markedly lower than the average of 5.57 short-
enings per text message in the English data in Bieswanger (2007). Finally, Deu-
mert and Masinyana (2008: 140) note that “[t]he isiXhosa [text] messages in our
corpus differ markedly from the writers’ English-language messages in that they
contain almost no abbreviated material”. The above findings demonstrate that fea-
tures of language use characteristic of English-language CMC are not always char-
acteristic of CMC in general when viewed from a cross-linguistic perspective.
seems to depend more on the social background of the user and the mode used than
on the language of the interaction.
4. Outlook
This chapter has discussed a number of micro-linguistic structural features that are
frequently considered characteristic of CMC, pointing out variation that occurs ac-
cording to the language and script used, the mode of CMC employed, and other
use- and user-related factors. So far, research on the use of micro-linguistic fea-
tures in CMC has mainly focused on issues such as which features are used, in
which language, in which modes, and by whom. Many of the studies mentioned in
this chapter, however, do not focus solely on the linguistic description of the lan-
guage used in CMC, but also implicitly or explicitly discuss pragmatic aspects
concerning “the process of producing language” (Mey 2001: 5), including address-
ing the various motivations for and functions of the use of the micro-linguistic fea-
tures investigated. Unfortunately, studies of micro-linguistic phenomena with a
specifically pragmatic focus – such as Dresner and Herring’s (2010) investigation
of the association between emoticons and illocutionary force – have so far been
rare. From the point of view of pragmatics, future studies of micro-linguistic fea-
tures of CMC should concentrate on the actual use of these features in different
contexts, including the reasons for their use and the pragmatic meanings their use
conveys.
Finally, although many studies have addressed micro-linguistic structural fea-
tures of CMC, these studies vary in focus and scope as well as in the methodology
used. Danet’s and Herring’s (2007b: 28) observation that “[t]here is a pressing
need for systematic cross-linguistic studies that make use of similar methods in
similar contexts involving different languages” certainly holds true for the lin-
guistic and pragmatic investigation of micro-level features of CMC. Moreover, it is
necessary to broaden the perspective of this exhortation and apply it to other fac-
tors influencing computer-mediated language use, such as mode, purpose, and the
sociolinguistic background of users, as systematic and comparable studies are also
lacking for these factors.
Acknowledgements
Notes
1. See Herring (2007) for a discussion of the terms “CMC mode” and “CMC genre”.
2. Crystal (2006) still uses the term “Netspeak” in the second edition of his much-discussed
book Language and the Internet.
3. I would like to thank Tuija Virtanen for suggesting a number of works in Finnish and
Swedish to be included in this chapter.
4. The term “initialism” is used as a cover term for all shortenings that “consist of the first
letter (or letters) of a combination of more than one word” (Bieswanger 2007: n.p.).
References
Androutsopoulos, Jannis K.
2003 Online-Gemeinschaften und Sprachvariation. Soziolinguistische Perspektiven
auf Sprache im Internet. Zeitschrift für germanistische Linguistik 31: 173–197.
Androutsopoulos, Jannis K.
2006 Introduction: Sociolinguistics and computer-mediated communication. Jour-
nal of Sociolinguistics 10(4): 419–438.
Androutsopoulos, Jannis K., Jens Runkehl, Peter Schlobinski, and Torsten Siever (eds.)
2006 Neuere Entwicklungen in der linguistischen Internetforschung. Hildesheim:
Olms.
Androutsopoulos, Jannis and Gurly Schmidt
2002 SMS-Kommunikation: Ethnografische Gattungsanalyse am Beispiel einer
Kleingruppe. Zeitschrift für Angewandte Linguistik 36: 49–79.
Anis, Jacques (ed.)
1999a Internet communication et langage français. Paris: Hermès.
Anis, Jacques
1999b Chats et usages graphiques du français. In: Jacques Anis (ed.), Internet com-
munication et langage français, 71–90. Paris: Hermès.
Anis, Jacques
2007 Neography: Unconventional spelling in French SMS text messages. In: Brenda
Danet and Susan C. Herring (eds.), The Multilingual Internet: Language, Cul-
ture, and Communication Online, 87–115. New York: Oxford University Press.
Aschwanden, Brigitte
2001 wär wot chätä? Zum Sprachverhalten deutschschweizerischer Chatter. Net-
worx 24. http://www.mediensprache.net/networx/networx-24.pdf
Bader, Jennifer
2002 Schriftlichkeit und Mündlichkeit in der Chat-Kommunkation. Networx 29.
http://www.mediensprache.net/networx/networx-29.pdf
Baron, Naomi S.
1984 Computer-mediated communication as a force in language change. Visible
Language 18(2): 118–141.
Baron, Naomi S.
2000 Alphabet to Email: How Written English Evolved and Where It’s Heading.
London/New York: Routledge.
Baron, Naomi S.
2008 Always On: Language in an Online and Mobile World. New York: Oxford Uni-
versity Press.
Beißwenger, Michael (ed.)
2001 Chat-Kommunikation. Sprache, Interaktion, Sozialität & Identität in syn-
chroner computervermittelter Kommunikation: Perspektiven auf ein interdis-
ziplinäres Forschungsfeld. Stuttgart: ibidem.
Beißwenger, Michael
2007 Sprachhandlungskoordination in der Chat-Kommunikation. Berlin/New York:
de Gruyter.
Bieswanger, Markus
2007 2 abbrevi8 or not 2 abbrevi8: A contrastive analysis of different space- and
time-saving strategies in English and German text messages. In: Simeon
Floyd, Taryne Hallet, Sae Oshima, and Aaron Shield (eds.), Texas Linguistic
Forum 50.
http://studentorgs.utexas.edu/salsa/proceedings/2006/Bieswanger.pdf
Bieswanger, Markus
2010 Gendered language use in computer-mediated communication: Typography in
text messaging. In: Markus Bieswanger, Heiko Motschenbacher, and Susanne
Mühleisen (eds.), Language in Its Socio-Cultural Context: New Explorations
in Gendered, Global and Media Uses, 157–172. Frankfurt: Lang.
Bieswanger, Markus
2011 The sociolinguistics of texting: Methodological considerations and empirical
results. Sprache und Datenverarbeitung: International Journal for Language
data Processing 35(1): 7–24.
Bieswanger, Markus and Frauke Intemann
2011 Patterns and variation in the language use in English-based discussion forums.
In: Martin Luginbühl and Daniel Perrin (eds.), Muster und Variation: Medien-
linguistische Perspektiven auf Textproduktion und Text, 157–187. Bern: Lang.
Crystal, David
2001 [2006] Language and the Internet [2nd Ed.] Cambridge, UK: Cambridge Univer-
sity Press.
Crystal, David
2004a The Language Revolution. Cambridge, UK: Polity.
Crystal, David
2004b A Glossary of Netspeak and Textspeak. Edinburgh, UK: Edinburgh University
Press.
Collot, Milena and Nancy Belmore
1996 Electronic language: A new variety of English. In: Susan C. Herring (ed.),
Computer-Mediated Communication: Linguistic, Social, and Cross-Cultural
Perspectives, 13–28. Amsterdam/Philadelphia: John Benjamins.
Danet, Brenda
2001 Cyberpl@y: Communicating Online. London: Berg.
Danet, Brenda and Susan C. Herring (eds.)
2007a The Multilingual Internet: Language, Culture, and Communication Online.
New York: Oxford University Press.
Herring, Susan C.
1996 Introduction. In: Susan C. Herring (ed.), Computer-Mediated Communication:
Linguistic, Social, and Cross-Cultural Perspectives, 1–10. Amsterdam/Phila-
delphia: John Benjamins.
Herring, Susan C.
2001 Computer-mediated discourse. In: Deborah Schiffrin, Deborah Tannen, and
Heidi Hamilton (eds.), The Handbook of Discourse Analysis, 613–634. Ox-
ford, UK: Blackwell.
Herring, Susan C.
2007 A faceted classification scheme for computer-mediated discourse. Lan-
guage@Internet 4, article 1.
http://www.languageatinternet.org/articles/2007/761/
Herring, Susan C.
in press Grammar and electronic communication. In: Carol Chapelle (ed.), Encyclo-
pedia of Applied Linguistics. Hoboken, NJ: Wiley-Blackwell. Preprint:
http://ella.slis.indiana.edu/~herring/e-grammar.2011.pdf
Herring, Susan C. and Asta Zelenkauskaite
2008 Gendered typography: Abbreviation and insertion in Italian iTV SMS. In:
Jason F. Siegel, Traci C. Nagle, Amandine Lorente-Lapole, and Julie Auger
(eds.), IUWPL7: Gender in Language, Classic Questions, New Contexts,
73–92. Bloomington, IN: IULC Publications.
Herring, Susan C. and Asta Zelenkauskaite
2009 Symbolic capital in a virtual heterosexual market: Abbreviation and insertion
in Italian iTV SMS. Written Communication 26(1): 5–31.
Jaffe, Alexandra
2000 Introduction: Non-standard orthography and non-standard speech. Journal of
Sociolinguistics 4(4): 497–513.
Josephson, Olle (ed.)
1997 Svenskan i IT-samhället. [Swedish in IT Society.] Uppsala: Hallgren & Fall-
gren.
Kallmeyer, Werner (ed.)
2000 Sprache und neue Medien. Jahrbuch des Instituts für deutsche Sprache. Ber-
lin/New York: de Gruyter.
Kasesniemi, Eija-Liisa
2003 Mobile Messages: Young People and a New Communication Culture. Tam-
pere: Tampere University Press.
Katsuno, Hirofumi and Christine Yano
2007 Kaomoji and expressivity in a Japanese housewives’s chat room. In: Brenda
Danet and Susan C. Herring (eds.), The Multilingual Internet: Language, Cul-
ture, and Communication Online, 278–300. New York: Oxford University
Press.
Kessler, Florence
2008 Instant Messaging: Eine neue interpersonale Kommunikationsform. Networx
52. http://www.mediensprache.net/networx/networx-52.pdf
Kiesler, Sara, Jane Siegel, and Timothy W. McGuire
1984 Social psychological aspects of computer-mediated communication. American
Psychologist 39(10): 1123–1134.
Kotilainen, Lari
2002 Moi taas, ai äm päk: lauseet, tilanteet ja englanti suomenkielisessä chat-kes-
kustelussa. [Hello again, I’m back: sentences, situations and English in Finnish
chat.] In: Ilona Herlin, Jyrki Kalliokoski, Lari Kotilainen, and Tiina Onikki-
Rantajääsko (eds.), Äidinkielen merkitykset [The Meanings of the Mother
Tongue], 191–209. Helsinki: SKS.
Lakoff, Robin Tolmach
1982 Some of my favorite writers are literate: The mingling of oral and literate strat-
egies in written communication. In: Deborah Tannen (ed.), Spoken and Written
Language: Exploring Orality and Literacy, 239–260. Norwood, NJ: Ablex.
Lee, Carmen K. M.
2007 Linguistic features of email and ICQ instant messaging in Hong Kong. In:
Brenda Danet and Susan C. Herring (eds.), The Multilingual Internet: Lan-
guage, Culture, and Communication Online, 184–208. New York: Oxford Uni-
versity Press.
Ling, Rich
2005 The socio-linguistics of SMS: An analysis of SMS use by a random sample of
Norwegians. In: Rich Ling and Per E. Pedersen (eds.), Mobile Communi-
cations: Renegotiation of the Social Sphere, 335–349. London: Springer.
Luginbühl, Martin
2003 Streiten im Chat. Linguistik Online 15: 70–87.
http://www.linguistik-online.de/15_03/luginbuehl.pdf
Mey, Jacob
2001 Pragmatics: An Introduction, 2nd Ed. Malden, MA: Blackwell.
Murray, Denise E.
1988 Computer-mediated communication: Implications for ESP. English for Spe-
cific Purposes 7: 3–18.
Murray, Denise E.
1990 CmC. English Today 23: 42–46.
Nastri, Jacqueline, Jorge Peña, and Jeffrey T. Hancock
2006 The construction of away messages: A speech act analysis. Journal of Com-
puter-Mediated Communication 10(2).
http://jcmc.indiana.edu/vol11/issue4/nastri.html
Nishimura, Yukiko
2007 Linguistic innovations and interactional features in Japanese BBS communi-
cation. In: Brenda Danet and Susan C. Herring (eds.), The Multilingual Inter-
net: Language, Culture, and Communication Online, 163–183. New York: Ox-
ford University Press.
Östman, Jan-Ola
2004 The postcard as media. Text 24(3): 423–442.
Oxford English Dictionary Online
2008 Emoticon. http://www.oed.com/
Palfreyman, David and Muhamed al Khalil
2007 “A funky language for teenzz to use;” Representing Gulf Arabic in instant
messaging. In: Brenda Danet and Susan C. Herring (eds.), The Multilingual In-
ternet: Language, Culture, and Communication Online, 43–63. New York:
Oxford University Press.
Pietrini, Daniela
2001 <<X’ 6 :-( ?>>: Gli sms e il trionfo dell’informalità e della scrittura ludica. Ita-
lienisch 46: 92–101.
Raymond, Eric S.
1994 The New Hacker’s Dictionary, 2nd Ed. Cambridge, MA: MIT Press.
Runkehl, Jens, Peter Schlobinski, and Torsten Siever
1998 Sprache und Kommunikation im Internet: Überblick und Analysen. Opladen:
Westdeutscher Verlag.
Sanderson, David W.
1993 Smileys. Sebastopol, CA: O’Reilly.
Schlobinski, Peter
2001 Book review: David Crystal, Language and the Internet.
http://www.mediensprache.net/en/literatur/rezensionen/docs/1452. pdf
Schlobinski, Peter (ed.)
2006 Von “hdl” bis “cul8r”: Sprache und Kommunikation in den neuen Medien.
Mannheim: Dudenverlag.
Schlobinski, Peter, Nadine Fortmann, Olivia Gross, Florian Hogg, Frauke Horstmann, and
Rena Theel
2001 Simsen. Eine Pilotstudie zu sprachlichen und kommunikativen Aspekten in der
SMS-Kommunikation. Networx 22.
http://www.mediensprache.net/networx/networx-22.pdf
Schlobinski, Peter and Thorsten Siever (eds.)
2005 Sprachliche und textuelle Merkmale in Weblogs. Ein internationales Projekt.
Networx 46. http://www.mediensprache.net/networx/networx-46.pdf
Schlobinski, Peter and Manabu Watanabe
2003 SMS-Kommunikation – Deutsch/Japanisch kontrastiv. Eine explorative Stu-
die. Networx 31. http://www.mediensprache.net/networx/networx-31.pdf
Schmitz, Ulrich
2004 Sprache in modernen Medien. Einführung in Tatsachen und Theorien, Themen
und Thesen. Berlin: Erich Schmidt Verlag.
Shirai, Hiromi
2006 Kawaicons: Emoticons im japanischen Chat (?_?). Websprache.
http://www.mediensprache.net/de/websprache/chat/emoticons/kawaic ons.asp
Shortis, Tim
2001 The Language of ICT: Information and Communication Technology. London/
New York: Routledge.
Siebenhaar, Beat
2006 Code choice and code switching in Swiss-German Internet Relay Chat rooms.
Journal of Sociolinguistics 10(4): 481–506.
Siever, Thorsten
2006 Emoticons: Gefühle mit Zeichen. Websprache.
http://www.mediensprache.net/de/websprache/chat/emoticons/
Siever, Torsten, Peter Schlobinski, and Jens Runkehl (eds.)
2005 Websprache.net. Sprache und Kommunikation im Internet. Berlin/New York:
de Gruyter.
Stein, Dieter
2003 Book review: David Crystal, Language and the Internet. Linguistics 41:
158–163.
Su, Hsi-Yao
2007 The multilingual and multiorthographic Taiwan-based Internet: Creative uses
of writing systems on college-affiliated BBSs. In: Brenda Danet and Susan C.
Herring (eds.), The Multilingual Internet: Language, Culture, and Communi-
cation Online, 64–86. New York: Oxford University Press.
Thurlow, Crispin
2006 From statistical panic to moral panic: The metadiscursive construction and
popular exaggeration of new media language in the print media. Journal of
Computer-Mediated Communication 11(3), article 1.
http://jcmc.indiana.edu/vol11/issue3/thurlow.html
Thurlow, Crispin, with Alex Brown
2003 Generation Txt? The sociolinguistics of young people’s text-messaging. Dis-
course Analysis Online 1(1).
http://extra.shu.ac.uk/daol/articles/v1/n1/a3/thurlow2002003-paper.html
Tseliga, Theodora
2007 “It’s all Greeklish to me!” Linguistic and sociocultural perspectives on
Roman-alphabet Greek in asynchronous computer-mediated communication.
In: Brenda Danet and Susan C. Herring (eds.), The Multilingual Internet: Lan-
guage, Culture, and Communication Online, 116–141. New York: Oxford Uni-
versity Press.
Waseleski, Carol
2006 Gender and the use of exclamation points in computer-mediated communi-
cation: An analysis of exclamations posted to two electronic discussion lists.
Journal of Computer-Mediated Communication 11(4), article 6.
http://jcmc.indiana.edu/vol11/issue4/waseleski.html
Weingarten, Rüdiger (ed.)
1997 Sprachwandel durch Computer. Opladen: Westdeutscher Verlag.
Werry, Christopher C.
1996 Linguistic and interactional features of Internet Relay Chat. In: Susan C. Her-
ring (ed.), Computer-Mediated Communication: Linguistic, Social, and Cross-
Cultural Perspectives, 47–63. Amsterdam/Philadelphia: John Benjamins.
Witmer, Diane F. and Sandra Lee Katzman
1997 On-line smileys. Does gender make a difference in the use of graphic accents?
Journal of Computer-Mediated Communication 2(4).
http://jcmc.indiana.edu/vol2/issue4/witmer1.html
Wolf, Alecia
2000 Emotional expression online: Gender differences in emoticon use. CyberPsy-
chology and Behavior 3(5): 827–833.
Yang, Chunsheng
2007 Chinese Internet language: A sociolinguistic analysis of adaptations of the
Chinese writing system. Language@Internet 4, article 1. http://www.language
atinternet.org/articles/2007/1142/Yang_Chinese_Internet.pdf
Zelenkauskaite, Asta and Susan C. Herring
2006 Gender encoding of typographical elements in Lithuanian and Croatian IRC.
In: Fay Sudweek and Charles Ess (eds.), Proceedings of Cultural Attitudes To-
wards Technology and Culture 2006, 474–489. Murdoch, Australia: Murdoch
University Press.
1. Introduction
Considerable attention has been paid in pragmatics and related approaches to dis-
course, such as interactional sociolinguistics, conversation analysis, and mediated
discourse analysis, to the role of rhythm and timing in voice-based communi-
cation. In face-to-face interactions and telephone conversation, for example,
people have been found to respond rhythmically to one another in a continuous
manner and are able to perform conversational transactions synchronously through
a shared understanding of timing and transitional signals. Rhythm and timing are
realized paralinguistically through pausing, pacing, and the use of stress and inton-
ation (Couper-Kuhlen and Selting 1996; Erickson 1980) and in face-to-face com-
munication, kinestically, for example through head movement, gesture, and gaze
(Goodwin 2002; Kendon 1974).
Rhythm and timing have been seen to have especially important functions in
the management of turns in conversation. Couper-Kuhlen (1993), for example,
found in a large corpus of British and American data that speech rhythm serves as a
metric for the timing of turn transitions. Goodwin (2002) has shown how prosodic
aspects of speech are used in concert with gesture and posture to create multiple
concurrently relevant temporalities within which participants interactively organ-
ize their interaction. Tannen (1984a, 1989) has observed how rhythmicity and rep-
etition in talk aid in creating coherence in conversations and “involvement” be-
tween participants. Pragmatic studies in cross-cultural communication have shown
that differences among speakers in behaviors such as speech speed, customary tim-
ing of turn-transitional pauses, and timing between periods of talk and periods of
silence can lead to disfluencies and misunderstandings (Hall 1959; Lehtonen and
Sajavaara 1985; Tannen 1984b).
Rhythm and timing are essential elements, not just in the organization and co-
herence of communication that occurs over time (van Leeuwen 2005), but also in
the realization of speech functions and the negotiation of social identities and so-
cial relationships. Studies in interactional sociolinguistics have shown how timing
and rhythm act as contextualization cues, framing utterances in particular ways
(Gumperz 1982), and how they function in meta and phatic communication, pro-
viding information about speakers’ attitudes towards their topics and their inter-
locutors (Tannen 1984a). Timing, rhythm, speed of speech, pausing, stress, and in-
tonation are central ingredients in what Tannen (1984a) has called “conversational
style”.
person is typing, and other innovations are being experimented with. Shankar and
her colleagues (2000), for example, propose a way of representing multiparty chat
that is inspired by musical scores. However, such attempts have not become widely
available, and it is questionable how popular they would be if they were.
This chapter argues that users of computer chat have developed ways to exploit
timing and rhythm for a multitude of communicative purposes within even the
most minimalist framework for mutual monitoring. Text-based CMC, as Ben-
Ze’ev (2004) points out, is both lean and rich: lean in the reduced cues it makes
available and rich in the ways participants use these cues to create the strong sense
of intimacy and solidarity associated with what Walther (1996) calls “hyperper-
sonal communication”. Numerous examinations of computer chat have noted the
importance of timing and rhythm in turn-taking and turn construction (see for
example, Bays 1998; de Siqueira and Herring 2009; Markman 2004; Rintel and
Pittman 1997), as well as in the establishment of particular levels of formality in
interaction and the development of feelings of closeness or “presence” between in-
terlocutors (Ben Ze’ev 2004; Donath 2004). Bays (1998), for example, argues that
the desire to maintain a particular rhythm in relation to their interlocutors is a fac-
tor affecting the linguistic construction of turns by users of IRC (Internet Relay
Chat), and de Siqueira and Herring (2009) found that participants in IM (Instant
Messaging) sessions tend to vary the timing of their responses to harmonize with
their interlocutors.
Timing has been found to be important even in asynchronous forms of CMC
such as email. Walther and Tidwell (1995), for example, argue that email com-
munication often depends on “chronemics” or “time-related messages” to convey
non-verbal cues, and they claim that time is an “intrinsic part” of this type of social
interaction conveying meaning and defining the nature and quality of relationships
with others “across multiple levels” (361). Similarly, Kalman et al. (2006), in their
examination of response latencies in asynchronous CMC, noted that response time
occurred with mathematical regularity, with most responses created quickly and
delays in responding resulting in fewer subsequent responses.
In fact, it can be argued that time is even more important in CMC: Precisely be-
cause of some of the reasons mentioned above – the “narrow bandwidth” and re-
duced cues of computer chat – signals such as language, style, timing, and speed of
writing become particularly salient. According to Lea and Spears (1995: 217),
“even first-time users form impressions of other communicants’ dispositions and
personalities based on their communication style”, which, in part, involves re-
sponse time, rhythm, typing speed, and other temporal factors.
This chapter illustrates these issues with a study of the pragmatic uses of
rhythm and timing in interaction among participants in a Hong Kong based gay
chat room. It seeks to show how timing is used as a semiotic resource in the ac-
complishment of particular concrete social actions within the larger social practice
of “fishing” (searching for sexual partners). The data were collected as part of a
six-month long ethnographic study of this chat room which involved hiring 20
regular users as “participant-researchers” to help in setting goals and plans for the
research, interviewing, and securing consent from their chat partners and monitor-
ing their own computer use. Participant-researchers also took “screen movies” of
their interactions in the chat room, capturing the temporal aspects of their multiple
concurrent interactions with other users.
The chat room itself uses a Java-based chat client, which allows users to enter a
room with a username and profile that can be changed every time one logs in. Upon
entering, one’s username appears on the list of other current users of the room.
Clicking on a particular name opens a private chat window with that user. Users
can also take part in public chat though the main console, but they rarely do so, pre-
ferring to go directly to private conversations with other occupants of the room.
This particular chat client gives no indication of turns in progress. Participants’
contributions are sent only after the user presses “enter” and are received complete
on their interlocutor’s computer screen within seconds, depending on the speed of
the connection. Users frequently carry on chats with multiple users or attend to
other tasks like checking email or looking at pornography while engaged in chat-
ting.
The data were collected using a software program called Spector, which par-
ticipant-researchers installed on their computers. The program took screen shots of
their computers at fixed intervals of three seconds. The advantages of this kind of
data are, first, that it allows the analyst to experience rhythm and timing in a way
similar to that experienced by the user, and second, that it gives the analyst access
to all of the different activities that might be occupying the user’s time, rather than
relying on individual chat logs isolated from their context. Although taking screen-
shots at three-second intervals (the setting determined to interfere least with the
speed of the users’ computers) limits the precision of our analysis when it comes
to timing, because the time between turns tends to be much longer in chat than
in face-to-face communication (see below), this range of data can at least reveal
general patterns of timing and rhythm that exist in this chat room and suggest areas
where later work can focus.
Different media amplify and constrain users’ orientations towards time and
space in different ways, making some ways of interacting and making meaning
more possible and others less possible (van Leeuwen 2005). At the same time, dif-
ferent activities also presuppose particular orientations towards time and space,
what I have previously referred to as “attention structures” (Jones 2005); different
practices have different rhythms. My concern in this chapter is to show, given the
constraints of this medium, how participants cooperate to produce “shared times”
and “shared rhythms” within which this particular social practice is framed, how
they develop conventions about the meaning of different time-related aspects of
their interaction, and how timing is used to create synchrony and feelings of inti-
macy.
One of the most important features of the interaction investigated here is that it is
extremely instrumental, i.e., directed towards the goal of sexual contact. Interac-
tants are sometimes quite efficient in their pursuit of this goal: When asked how
long the interval between making online contact with a potential partner and final-
izing plans for actually meeting up for sex might be, one participant said, “very
short … perhaps as short as a few minutes”. The second important feature of this
kind of interaction is that it is intensely competitive. In the “marketplace” (ten
Have 2000) of the chat room, users constantly compete for “interactional time”
with other users who might want to engage the same target, and, at the same time,
they distribute their own time and attentional resources among a number of differ-
ent targets.
This marketplace, then, operates according to a strict “attention economy”
(Lankshear and Knobel 2002), where the most valuable asset is the attention of
others, and the investment one makes to gain it is paying attention back to them.
The issue of attention goes much deeper here than sexual attraction. Back in 1951,
Reusch and Bateson wrote that the foundation of all social situations is some form
of mutual perception. The form this mutual perception takes determines how we
enact social presence, display our attentional state, and frame the kinds of activities
we are involved in. In face-to-face conversation, both speaker and hearer must ac-
tively respond to what transpires by signalling involvement, either directly through
words or indirectly through gestures or similar non-verbal signals like gaze, pos-
ture, and bodily alignment (Tannen 1989). However, in CMC, in which non-verbal
bodily cues are not available, the “state of talk” becomes more tenuous; only when
one’s message is visible in the scrolling text window does one truly “exist” in the
conversation. Therefore, the only way users have to be present to their interlocu-
tors and communicate attention or involvement is by replying to them within a par-
ticular period of time. Similarly, through interpreting their timing (in particular
the time between turns), users make judgements about their interlocutors’ social
presence, social activity, and attentional state.
(1)
20:02:46 >hey ma n how r u
20:02:50
20:02:53 <muscular bi>fine
20:02:56 <muscular bi>stat?
>r u gam?
20:03:00 <muscular bi>yup
20:03:03
20:03:06
20:03:10 >gwm 38 5’8 80kg stocky fit tanned smooth cut thick vers1
20:03:13 >married?
As can be seen, participants take short, rapid turns with pauses between turns
usually no longer than six seconds. Longer pauses do occur, notably after partici-
pants are asked for their “stats” (self description involving age, weight and height,
and sometimes other relevant information about physical appearance), primarily
because these responses tend to take a bit longer to type, especially if they are elab-
orated upon (as in: “gwm 38 5’8 stocky fit tanned smooth cut thick vers”1). Most
turns consist of one to four words. The string of descriptors: “gwm 38 5’8 stocky fit
tanned smooth cut thick vers” is unusually long, not only creating a slight inter-
ruption in the rhythm of the conversation but also creating implicature, communi-
cating particular involvement, cooperation, and interest.
As the interaction progresses, however, this rhythmic pattern begins to break
down. At 20:30:23, a slight disfluency occurs when both parties ask a question of
each other at the same time. Such an occurrence in itself is not particularly prob-
lematic – users are used to them, and the initiator of the chat immediately com-
pensates by changing his discourse position (from questioner to answerer) and
answering muscular bi’s question before muscular bi finally answers his. What is
notable, though, is that while the first response comes within six seconds of the
question, the response from muscular bi comes more than fifteen seconds after the
question.
Finally, muscular bi does not respond to the question “do u work out” (example
2). After waiting around 20 seconds, his interlocutor makes another attempt to eli-
cit a response by asking a different question, “any pic?”, after which he waits an-
other 33 seconds or so before issuing a more explicit prompt, “hello”, and then
after approximately 30 more seconds finally gives up.
(2)
20:04:00
20:04:04
20:04:07
20:04:10
20:04:14
20:04:17
20:04:20 >any pic?
20:04:24
20:04:27
20:04:31
20:04:34
20:04:37
20:04:41
20:04:44
20:04:47
20:04:51
20:04:54
20:04:58 >hello
20:05:01
20:05:04
20:05:08
20:05:11
20:05:14
20:05:18
20:05:21
20:05:24
20:05:28
20:05:31 (close window)
One might ask why muscular bi does not terminate the conversation himself
if he is not interested. One can never be sure, however, if such pauses indicate a
lack of interest or are due to other extraneous circumstances (the user may be
answering a phone call or attending to chats with other – perhaps more desir-
able – partners). At the same time, this strategy of showing lack of interest is a
particularly prevalent means of rejection. The preference for this means of ter-
mination, rather than the less popular “sorry”, “gotta go”, or “not my type”,
comes from the fundamental difficulty involved in “ending” characteristic of
all communication (Schegloff and Sacks 1973), magnified by the even more
“face threatening” nature of termination of this particular kind of interaction.
In the entire corpus of chats collected, 86% are terminated without any closing
ritual.
At the same time, very short pauses can communicate intense interest. Just as in
face-to-face conversation, people have different “conversational styles” (Tannen
1984a), and one aspect of style has to do with how much (and how) one shows in-
volvement. Showing involvement in computer chat, as in face-to-face conver-
sation, often includes talking a lot and taking shorter pauses between turns. Tannen
(1984a) calls this “high involvement style”, and, aside from shorter pauses, it also
includes features such as more frequent questions and, in face-to-face conver-
sation, paralinguistic, proxemic, and gestural characteristics.
As I have already mentioned, most conversations in this chat room seem to be
characterized by “high involvement”, when they are progressing successfully, that
is; interactions not showing such involvement usually do not last. There are in-
stances, however, in which this involvement is even more pronounced. The main
feature of this “hyper-involvement” style is when a user issues a rapid series of
consecutive turns without waiting for a reply, as in example 3.
(3)
09:20:26 <LeanFit>nice stats
09:20:30 <LeanFit>i like
09:20:34 <LeanFit>i am lean fit here
09:20:38 >ic
<LeanFit>looking for?
09:20:42 >any
09:20:46 <LeanFit>I look for
09:20:50
09:20:55 <LeanFit>regular gay sex buddy
09:20:59 <LeanFit>ok to u?
09:21:03
09:21:07 <LeanFit>but
>sure
>but what
09:21:11 <LeanFit>we can swim together
09:21:15 >haha
<LeamFit>and have sex
09:21:19 <LeanFit>ok to u?
09:21:23 <LeanFit>hei
09:21:27 >why not
09:21:31 <LeanFit>wanna try swim at night in beach with me?
09:21:36
09:21:40 >i wanna to go this afternoon
09:21:44 <LeanFit>sure
09:21:48
09:21:52 <LeanFit>where shall we go?
09:21:56 <LeanFit>pool?
09:22:00 <LeanFit>where u live?
In this example, LeanFit’s contributions come at the rate of one every 5.7 seconds,
and most intervals are three seconds or less. In the space of the one minute 26 sec-
onds of this excerpt, LeanFit takes 15 turns, while his interlocutor takes six. This
strategy not only gives the user a way to highlight his attention towards his inter-
locutor, but also a way of keeping the interlocutor’s attention by continually “re-
freshing” his presence on his interlocutor’s screen. One’s use of such a rhythmic
strategy, as in face-to-face conversation, can affect others’ perception of one’s
character; users who have different conversational styles might interpret such
strategies as “friendly”, or as “pushy” or “desperate” (Tannen 1984a).
This use of rapid, consecutive turns also has the function of maintaining pres-
ence by avoiding the longer pauses that might ensue when longer stretches of in-
formation need to be given, such as “stats”. In example 4, for instance, hot fun
divides his “stats” into five turns.
(4)
20:26:11 > ur stats
20:26:15
20:26:18 <hot fun> 21
20:26:25 <hot fun> 180 cm
20:26:28 <hot fun> 150lbs
20:26:31 <hot fun> 30 w
20:26:35 <hot fun> 41 c
This rhythmic incremental release of information, however, can also serve another
function, that of introducing “suspense” into the process of seduction. Here,
rhythm plays a crucial role in the game of revealing and withholding information
to maximize involvement from one’s interlocutor (van Leewuen 2005). Hot fun’s
incremental release of his statistics becomes a kind of “textual striptease” (Barthes
1975) that heightens desire by heightening anticipation of the next, partial revel-
ation. A similar strategy is used by Naked Sleeper when he describes what he likes
to do in bed (example 5).
(5)
20:38:37 >what do u like to do in sex?
20:38:42
20:38:47
20:38:52 <Naked Sleeper> kissing
20:38:57 <Naked Sleeper> oral
20:39:02 <Naked Sleeper> fucking
3. Conversational management
I do not mean to suggest that long pauses always show lack of interest or result in
interactions being terminated. It depends very much on where these pauses occur
in the conversation. Conversations are not mono rhythmic – they develop through
many changes in rhythm and, as in face-to-face conversations, these changes of
rhythm and shifts in tempo are often meaningful.
In fact, the average time between turns in my data is 30–39 seconds, which
does not seem consistent with the rapid turn-taking in the examples above. This is
because these interactions tend to follow a pattern of “bursts” and “breaks”, in
which chunks of rapid interaction are interspersed with sometimes quite lengthy
breaks. Within these “bursts” of interaction the average time between turns is 3–9
seconds.
This segmentation of the conversation is by no means random. These bursts
and breaks take place in the context of an interactional format which progresses
through a fairly predictable set of stages made up primarily of question/answer ad-
jacency pairs (the exchange of information about physical appearance, sharing of
other information such as location and availability, the exchange of pictures, or the
making of arrangements to meet). Each stage tends to take place in one burst, with
breaks tending to occur between stages. The initial burst usually involves the ex-
change of information about physical appearance, the second stage usually in-
volves further sharing of information (such as location and availability), the third
involves the exchange of pictures, and the next involves arrangements to meet and
often the exchange of telephone numbers. Breaks often serve to mark the transition
to a new topic or new stage in the interaction. When they occur in between two
halves of an adjacency pair or an offer of information and a reaction to it, they can
create implicature that is interpreted sometimes as deliberation, sometimes as lack
of interest, and sometimes as rudeness. Example 6 illustrates a typical pattern. In
the first burst participants share information about what they are seeking and
physical appearance.
(6)
20:20:23 <nice gam> hihi
20:20:27 > hi
20:20:30 <nice gam> u lk for?
20:20:33
20:20:37 >friends or fun
20:20:40 >u?
20:20:43 <nice gam> anything
20:20:47 > great
20:20:50 <nice gam> stats?
20:20:53
20:20:57
20:21:00 > 36/6’/160
20:21:03 <nice gam> nice stats
20:21:07 <nice gam> u gwm
20:21:10 > yup
20:21:13 > ur stats?
20:21:17
20:21:20 <nice gam> 21/5/11/130
20:21:24 > nice
Then after a break of about 20 seconds they talk about where they live (example 7).
This exchange is followed by another break, also around 20 seconds long.
(7)
20:21:27
20:21:34
20:21:37
20:21:40
20:21:44
20:21:47
20:21:50
20:21:54 <nice gam> live alone
20:21:57 > yes
20:22:00 <nice gam> where
20:22:04 > Tai Po
20:22:07 <nice gam> ic
20:22:11 <nice gam> quite close
20:22:14
20:22:17
20:22:21
20:22:24
20:22:27
After this the exchange of pictures is negotiated. The final break occurs after users
exchange picture links, to give participants a chance to visit the respective sites and
evaluate the pictures (example 8).
(8)
20:23:31 <nice gam> pic
20:22:34 > for trade
20:22:37 <nice gam> email or link
20:22:41 > link
20:22:44 >urs?
20:22:47 <nice gam> my link is piclink.com/xxxxx/
20:22:51
20:22:54 > picturetrail.com/YYYYY
20:22:57
20:23:01
20:23:04
20:23:07
20:23:11
20:23:14
20:23:18 <nice gam> nice pic
20:23:21 > thanks
This patterning gives users a way to manage and signal the accomplishment of
these different stages of the interaction. It also helps them to manage multiple in-
teractions strategically, the breaks giving users time to engage with other interlocu-
tors, as well as to prioritize different interactions based on the stages they have
reached and the likelihood of them progressing to subsequent stages.
As noted before, one thing that characterizes this activity is a high degree of poly-
chronicity. Users often engage in conversations with multiple interlocutors in order
to maximize their chances of success. These multiple, simultaneous interactions do
not proceed at the same pace. Some are slower, and some are faster. Each thread
of conversation follows its own rhythm, “yet [they] are perfectly coordinated with
each other as they weave in and out without ever colliding” (van Leeuwen 2005:
194).
Several factors serve to facilitate these multiple concurrent interactions. First,
the alternation between bursts and breaks makes it easier for users to toggle from
one interaction to another. Second, interlocutors tend to synchronize their rhythm
with the person they are talking to. Such “conversational synchrony” has been
widely noted in face-to-face conversation (Hall 1959; Kendon 1974; Scollon 1982).
Erickson (1980) has shown, for example, that in ordinary talk people speak to each
other in a regular meter of regular beats and time their entrances and exits to the
rhythm of these beats. As Capra (1982: 300, 302) notes, “Human communication
[…] takes place to a significant extent through the synchronization and interlocking
of individual rhythms[, … and] opposition, antipathy, and disharmony will arise
when the rhythms of two individuals are out of synchrony”. This also seems to be
the case in computer-mediated interaction; de Siqueira and Herring (2009), for
example, in their study of dyadic IM (Instant Messaging) conversations, found that
some users attempt to achieve conversational synchrony by adjusting the rhythm
and timing of their contributions to fit the temporal styles of their interlocutors.
(9)
LeanFit SportStylish Gymfit built btm tanguy
09:26:20 >with family <tanguy>alone?
>clerial
09:26:24 >u
09:26:28 <LeanFit>icic >what time u <Gymfit built
go today btm>have icq
number?
09:26:32 <LeanFit>bank >no
clerk here
09:26:36 >ic
09:26:41 <LeanFit>hei
09:26:45 <LeanFit>u fit <SportStylish> >yes, but not open
too? what col of now
your speedo?
09:26:49 <SportStylish>
maybe 1100
09:26:53 >medium
09:26:57 <LeanFit>icic
09:27:01 >arrive??
09:27:05 >mind? <SportStylish>
around
09:27:10 >where u live
09:27:14 <LeanFit>u do <tanguy>me no
exercises? 2
09:27:18 >no now
09:27:22 >no money
09:27:26 <SportStylish> >ic
hung hom. u?
It is also important to remember that these interactions are not just coterminous
but also competitive, and not all participants are equally successful in maintaining
the attention of their interlocutor. The relative pace of the interactions, therefore,
can be an indication of how successfully they are progressing and can also tell us
something about which party is more in control of the interaction.
In example 10, the user begins an engagement with an interlocutor called chris-
boi. While he is waiting for chrisboi to respond, he answers the invitation of AC
Reloaded, returns to chrisboi with a reaction to his information, and then answers
AC reloaded’s question before initiating another sequence with chrisboi.
(10)
20:19:06
20:19:09 >8” here
20:19:13
20:19:16
20:19:19 <AC Reloaded>hello
20:19:23 <chrisboi in central>cool
20:19:26 <chrisboi in central>6.5 here >yo
maybe <AC Reloaded>care for a chat?
20:19:29 >sure
>r u gam?
20:19:33 <AC Reloaded>how’s going man?
20:19:36 >lol
20:19:39 >sounds nice
20:19:43
20:19:46 <AC Reloaded>yeah I am
20:19:50 >horny n u
20:19:53
20:19:56 >r u top or btm
20:20:00 <AC Reloaded>so am I
20:20:03 <chrisboi in central>am vers top
20:20:06 <chrisboi in central>u?
As chrisboi begins to be more forthcoming in his responses, the user pays more at-
tention to that conversation and less attention to AC reloaded (example 11).
(11)
20:20:10 >nice i am vers
20:20:13 <chrisboi in central>ur stats btw?
20:20:16
20:20:20
20:20:23
20:20:27 >gwm 28 6’ 78kg 44c 32w 8 cut
thick vers
20:20:30
20:20:33 <AC Reloaded>care to intro?
20:20:37 >mmm
20:20:40
20:20:43
20:20:47 <chrisboi in central>cool … >gwm 28 6’ 78kg 44c 32w 8 cut
sounds hot thick vers
20:20:50 >do u work out
20:20:53 <chrisboi in central>where are u <AC Reloaded>wow
from?
20:20:57
20:21:00
20:21:03 <AC Reloaded>8” that’s a very
nice size
20:21:07 >Melbourne
20:21:10 >u
20:21:13
20:21:17
20:21:20
20:21:24 <chrisboi in central>not really …
so am not mascular or anything …
but what u see in the pics is what u
get
20:21:27 <chrisboi in central>from Tokyo
20:21:34 >ok nice
20:21:37
20:21:40
20:21:44
Later, as the interaction with chrisboy heats up even more, AC reloaded gets
only a minimal response (“tks” means thanks), before being completely ignored.
At the end AC reloaded says, “Anyway I think you must be pretty occupied
at the moment, sorry for the intrusion” and terminates the interaction (example
12).
(12)
20:21:47 >well I dont see al that I will get
20:21:50 >lol
20:21:54 >tks
20:21:57 <chrisboi in central>lol
20:22:00
20:22:04 <AC Reloaded>no prob
20:22:07 >do ulive here or r u visiting
20:22:11
20:22:14
20:22:17 <chrisboi in central>i live here
20:22:21 >same
As can be seen from the examples above, timing and rhythm are not just a matter of
conversational synchrony, but also a matter of interactional roles and social power.
The way timing is managed affects and is affected by the degree of “desirability”
the user is able to project through the relative “market value” of different personal
and physical attributes within this community and how “desirable” he perceives
his interlocutor to be. Some people, as a consequence of their perceived market
value, are less willing to wait, while others who are not so desirable are forced to
wait longer for responses. In fact, it is sometimes by taking longer pauses that users
signal their desirability and power.
In example 13, after try has given his “stats”, the user replies with his own
“stats”, breaking them up into three consecutive turns, not, as in the above
examples, to create suspense and desirability, but rather, because the last bit of in-
formation, the user’s weight, marks him as having less “market value” than his in-
terlocutor.
(13)
20:20:06 <try> I am aged 23, 130 lbs, 5;11“ how about u?
20:20:10
20:20:13 > I am 23 also
20:20:16 >178 cm
20:20:20 >190 lbs
20:20:23
20:20:27
20:20:30
20:20:33
20:20:37
20:20:40
20:20:43
20:20:47
20:20:50
20:20:53
20:20:57
20:21:00
20:21:03 <try> ic
(14)
20:21:07
20:21:10 > do I scare you?
20:21:13
(Three minutes and nine seconds elapse.)
He then waits more than three minutes for try finally to respond again (example
15).
(15)
20:24:04
20:24:08 <try> honestly, yes but I hope you don’t mind
20:24:11 > it’s ok
20:24:14 > at least you are willing to tell me
20:24:18 <try> I think you should join a weight loss program
20:24:21
The way this particular user described in a subsequent interview his expectations
about timing contrasts sharply with the participants quoted above, who considered
10 to 15 seconds between turns reasonable. He explained:
I think I am a patient guy, I am always willing to wait. There is no time limit for me to
wait. I have tried before, there is guy in the chat room talking to me, after knowing my
size, he doesn’t reply, and I just wait for him till I turn off the computer, but still no re-
sponse. But I think it’s ok, as no matter he replies or not, it won’t disturb me. In fact, I
have so much experience of having delays, no matter long or short. Every time, if no re-
plying for a long time, I will type “are you here?”, “busy?”, “do I scare you?”, but most
of the time, they won’t reply.
This example reminds us that conversational style is not simply a matter of per-
sonal psychological disposition or group norms, it is also often a matter of power
and ideology. Different sets of rhythmic expectations either consciously or uncon-
sciously become submerged into the “historical bodies” (Jones 2005) of users
through sustained interaction in a particular kind of social environment and in re-
sponse to the values the environment reflects and the possibilities it makes avail-
able to different kinds of people.
Finally, I would like to suggest that rhythm and timing serve not just to help users
manage their conversations and judge the relative interest of their interlocutors, but
that they also play a fundamental role in the pleasure participants take in this ac-
tivity.
When chatting, users not only synchronize their rhythms to the multiple
rhythms of their interlocutors, but they also build up their own personal rhythms
(see also de Siquera and Herring 2009), characterized by almost constant rhythmic
activity. Even when they are engaged in just one or two interactions, users are
rarely idle: While waiting for responses from others, they fill in the time, either
chatting with someone else, scrolling though the list of names in the chat room for
new possibilities, surfing the web, answering their email, or playing games. Thus,
while the pace of each individual conversation might be much slower than most
face-to-face encounters, this multitasking gives to this activity a constant, rapid,
rhythmic character.
According to Chapple (1982), in face-to-face conversation, people associate
feelings of pleasure and well-being with the rhythmicity of the interaction. Eve
(2004) suggests that computer chat, with its rapid alternation of sending and re-
ceiving messages, also creates similar feelings in users. In interviews, several par-
ticipants noted how “absorbing” and even “addictive” this activity of chatting with
multiple prospective sexual partners could be. “Once you get into it”, one said, “it
can be hard to stop”. Another noted, “I tend to lose [sic] track of time. Two or three
hours can pass, and I don’t even realize it”.
Csikszentmihalyi (1982) refers to this phenomenon of becoming “lost” in an
activity as “flow”, a state in which a person loses “self-consciousness” and a sense
of time, focusing on “the present, blocking out the past and the future” (38). Others
(Trevino and Webster 1992; Webster, Trevino, and Ryan 1993) have evoked Csiks-
zentmihalyi’s theory not just to explain the pleasure people associate with CMC
but also the heightened sense of presence the medium creates, despite the reduced
cues it makes available.
7. Conclusion
The main point of this chapter is that, despite the many extraneous factors that
might affect rhythm in computer-mediated interaction, timing and rhythm can have
important pragmatic functions in this medium, especially for certain communities
of users engaged in focused activities, for functions such as showing involvement
and interest, and for creating conversational implicature, managing and signaling
different stages in the interaction, and negotiating power relations. It may also play
a role in the enjoyment users experience in the activity of chatting.
I do not wish to suggest that timing is equally important or used in the same
way in all CMC or even in all chat-based environments; indeed, because of its im-
portant role in communicating desire, attractiveness, and seduction, timing may
take on a heightened importance in the kinds of interactions I have been describ-
ing – just as in face-to-face interaction, timing is used differently and takes on dif-
ferent meanings in different communicative contexts.
It is both the “bodily” nature of this communicative context, which makes tim-
ing so important, and rhythm and timing that make the experience of the context
more of an “embodied” one for users. Rhythm, van Leewuen (2005) points out,
provides an important link between semiotic articulation and the body; it con-
tributes not just to organizing communicative events, but also to vitalizing them,
allowing, in the case of these conversations, users to reach very high levels of em-
pathy and intimacy in a shorter time as compared to face-to-face communication.
In many ways, social identities online are very much “identities of rhythm” (Capra
1982).
Future work in this area should include more studies of the timing of turn-tak-
ing in computer chat, integrating quantitative and statistical methods (see, e.g., de
Siqueira and Herring 2009; Kalman et al. 2006), as well as pragmatic and conver-
sation analytic studies of the effects of timing on the conduct of conversation and
the negotiation of meaning. It should also focus more on the specific rhythmic pat-
terns that develop in particular genres of CMC and in particular activity types
(comparing, for example, task-oriented interaction with casual interaction). Fin-
ally, more work needs to be done on the psychological and relational effects of
rhythmicity in computer-mediated interaction, not just in chat, but also in other
types of online interaction such as multiplayer games.
Timing is an analytical site where issues of involvement, attention, pleasure,
and power converge. As Prior and Shipka (2003: 230–231) argue, timing is central
in all interactions for “the production of embodied chronotropes, the production of
a lifeworld with a certain tone and feel, populated by a certain people and their
ideas, calibrated to a certain rhythm”. Different media, with the different “attention
structures” (Jones 2003, 2005) they make available to users, affect the kinds of re-
lationships and “lifeworlds” users are able to enter into and the kinds of meanings
that can be made.
Note
References
Barthes, Roland
1975 The Pleasure of the Text, trans. Richard Howard. New York: Hill and Wang.
Bays, Hillary
1998 Framing and face in Internet exchanges: A socio-cognitive approach. Linguis-
tik Online 1. http://www.linguistik-online.de/bays.htm
Ben-Ze’ev, Aaron
2004 Love Online: Emotions on the Internet. Cambridge, UK: Cambridge Univer-
sity Press.
Capra, Fritjof
1982 The Turning Point. New York: Simon and Schuster.
Chapple, Eliot D.
1982 Movement and sound: The musical language of body rhythms in interaction.
In: Martha Davis (ed.), Interaction Rhythms, 28–50. New York: Human
Sciences Press.
Clore, Gerald C. and Karen Gasper
2000 Feeling is believing: Some affective influences on belief. In: Nico H. Frijda,
Antony S. R. Manstead, and Sacha Bem (eds.), Emotions and Beliefs: How
Feelings Influence Thoughts, 10–44. Paris/Cambridge: Editions de la Maison
des Sciences de l’Homme and Cambridge University Press.
Couper-Kuhlen, Elizabeth
1993 English Speech Rhythm: Form and Function in Everyday Verbal Interaction.
Amsterdam/Philadelphia: John Benjamins.
Couper-Kuhlen, Elizabeth and Margaret Selting (eds.)
1996 Prosody in Conversation: Interactional Studies. New York: Cambridge Uni-
versity Press.
Csikszentmihalyi, Mihaly
1982 Beyond Boredom and Anxiety. San Francisco, CA: Jossey and Bass.
de Siqueira, Amaury and Susan C. Herring
2009 Temporal patterns in student-advisor Instant Messaging exchanges: Individual
variation and accommodation. In: Proceedings of the Forty-Second Hawai’i
International Conference on System Sciences. Los Alamitos, CA: IEEE Press.
Donath, Judith
2004 Sociable media. In: William Sims Bainbridge (ed.), The Berkshire Encyclo-
paedia of Human-Computer Interaction, 631–32. Great Barrington, MA:
Berkshire Publishing Group.
Erickson, Frederick
1980 Timing and context in everyday discourse: Implications for the study of refer-
ential and social meaning. Sociolinguistic Working Paper Number 67. Austin,
TX: Southwest Educational Development Laboratory.
Eve, Robin
2004 Rhythm in interaction. Research Journal of the School of Environmental
Studies, Himeji Institute of Technology 6: 191–200.
Garcia, Angela C. and Jennifer B. Jacobs
1998 The interactional organization of computer mediated communication in the
college classroom. Qualitative Sociology 21(3): 299–317.
Garcia, Angela C. and Jennifer B. Jacobs
1999 The eyes of the beholder: Understanding the turn-taking system in quasi-syn-
chronous computer-mediated communication. Research on Language and So-
cial Interaction 32: 337–367.
Goodwin, Charles
2002 Time in action. Current Anthropology 43: S19–S35. (Supplement August–Oc-
tober 2002, Special issue on Repertoires of Timekeeping in Anthropology,
Andre Gingrich, Elinor Ochs, and Alan Swedlund, eds.
Gumperz, John
1982 Discourse Strategies. Cambridge, UK: Cambridge University Press.
Hall, Edward T.
1959 The Silent Language. New York: Doubleday.
Herring, Susan C.
1999 Interactional coherence in CMC. Journal of Computer-Mediated Communi-
cation 4(4). http://jcmc.indiana.edu/vol4/issue4/herring.html
Jacobson, David
1999 Impression formation in cyberspace: Online expectations and offline experi-
ences in text-based virtual communities. Journal of Computer-Mediated Com-
munication 5(1). http://jcmc.indiana.edu/vol5/issue1/jacobson.html
Jones, Rodney
2003 Inter-activity: How new media can help us understand old media. Paper pres-
ented at the International Conference of Historical Linguistics, Copenhagen,
August 11–15.
Jones, Rodney
2005 Sites of engagement as sites of attention: Time, space and culture in electronic
discourse. In: Sigrid Norris and Rodney Jones (eds.), Discourse in Action: In-
troducing Mediated Discourse Analysis, 144–154. London: Routledge.
Kalman, Yoram M., Gilad Ravid, Daphne. R. Raban, and Sheizaf Rafaeli
2006 Pauses and response latencies: A chronemic analysis of asynchronous CMC.
Journal of Computer-Mediated Communication 12(1). http://jcmc.indiana.
edu/vol12/issue1/kalman.html
Kendon, Adam
1974 Movement coordination in social interaction. In: Shirley Weitz (ed.), Nonver-
bal Communication, 150–68. New York: Oxford University Press.
Lankshear, Colin and Michele Knobel
2002 Do we have your attention? New literacies, digital technologies and the edu-
cation of adolescents. In: Donna E. Alverman (ed.), Adolescents and Literacies
in a Digital World, 19–39. New York: Peter Lang.
Lea, Martin and Russell Spears
1995 Love at first byte? Building personal relationships over computer networks. In:
Julia T. Wood and Steve Duck (eds.), Under-studied Relationships: Off the
Beaten Track, 197–233. Thousand Oaks, CA: Sage.
Tannen, Deborah
1989 Talking Voices: Repetition, Dialogue, and Imagery in Conversational Dis-
course. Cambridge, UK: Cambridge University Press.
ten Have, Paul
2000 Computer-mediated chat: Ways of finding chat partners. M/C: A Journal of
Media and Culture 3(4). http://www.api-network.com/mc/0008/partners.php
Trevino, Linda K. and Jane Webster
1992 Flow in computer-mediated communication. Communication Research 19(5):
539–573.
van Leeuwen, Theo
2005 Introducing Social Semiotics. London: Routledge.
Walther, John B.
1996 Computer mediated communication: Impersonal, interpersonal and hyperper-
sonal interaction. Communication Research 23(1): 3–43.
Walther, John B. and Lisa C. Tidwell
1995 Nonverbal cues in computer-mediated communication, and the effects of chro-
nemics on relational communication. Journal of Organizational Computing
5(4): 355–378.
Webster, Jane, Linda K. Trevino, and Lisa Ryan
1993 The dimensionality and correlates of flow in human-computer interactions.
Computers in Human Behavior 9: 411–426.
1. Introduction
the two modes (text-based and spoken) should not be seen as possessing the same
properties. For example, cohesion will be manifest in different ways, and partici-
pants will ascribe coherence in different ways, in SCMC (see the chapters by Her-
ring and Markman, this volume). There are, however, striking similarities between
SCMC and multiparty spoken discourse in particular, and a consideration of floor
in multiparty spoken discourse provides something of a “way in” for analysis of its
computer-mediated equivalent.
The chapter begins with a review of the literature on floor in spoken and written
(i.e., SCMC) conversation, which I relate to the concepts of inter-turn cohesion and
coherence in CMC. I then introduce a case study from which I draw my subsequent
examples. The case study is in two parts. The first part exemplifies the elements
and some types of conversational floor; in the second part, I turn to three key di-
mensions of interaction which influence the development of floors: the role re-
lations of the participants; the communicative action and topic of the discourse
(i.e., the verbal activity); and certain medium-related factors. Their influence is
discussed with reference to an extended sample of SCMC text from the case study,
in which I trace the development of conversational floors.
This review first addresses the question of cohesion in SCMC, its relationship with
coherence, and how it operates in different ways from cohesion in spoken dis-
course. I then draw on the CMC literature to explain why conversational threads as
they relate to certain types of conversational floor are relevant to an SCMC account
of cohesion and coherence in discourse.
its frequent lack of cohesion (Örnberg Berglund 2009); if it were not, it would not
be so popular, as noted by Herring (1999).
In SCMC discourse, coherence is not fundamentally dependent on the coordi-
nation of transfer in turn-taking, as it would be in spoken discourse. Rather,
broader and looser constructs such as the conversational floor, as described below,
are the cohesive “glue” that contribute towards participants’ ascribing coherence
(unity, meaning) to the discourse. One result of the lack of cohesion in turn-taking –
what Herring (1999) calls the lack of sequential coherence – is, as Herring also
notes, disrupted turn adjacency. Example 1 is a typical instance of disrupted turn
adjacency in SCMC, as it occurred in a two-dimensional graphics-enhanced chat
room called The Palace. The extract is from a saved log of the interaction and ap-
pears as it was typed. It shows how, in comparison with spoken conversation,
SCMC displays a reduced sensitivity to coordination of transfer in turn-taking,
which can be viewed as a lack of fine tuning. Here and in subsequent examples,
turns have been numbered for ease of reference.
simply, the presence of the first pair part opens a slot in conversation for an ex-
pected, or conditionally relevant, second pair part. MichaelC’s first pair part
(turn 1) is followed by the second pair part (2) from Ying-Lan. This response, “not
so good”, is a dispreferred response (Heritage 1984: 265–269; Nofsinger 1991:
71–2). That is to say, although the response is expected, or conditionally relevant,
it is not as expected (or preferred) as a response such as “I’m fine, thanks”. Follow-
ing a tendency noted in dispreferred second pair parts, the response is followed by
an elaboration in turn 3. But MichaelC’s next turn (4) seems to be in response to
Ying-Lan’s turn 2, rather than to turn 3. This is a case of disrupted turn adjacency.
The disrupted turn adjacency in example 1 may well be a result of reduced co-
ordination of transfer, in that MichaelC was typing turn 4 at the same time as Ying-
Lan was typing her elaboration following her dispreferred response (turn 3). It hap-
pened that they sent their turns at about the same time, but Ying-Lan sent her turn
fractionally before MichaelC sent his. Thus it appears in the log of the chat, and ap-
peared at the time on the screen, that Ying-Lan answers MichaelC’s question be-
fore he asks it.
On a broader level, these observations on disrupted turn adjacency tend to sup-
port the view that applying models of turn-taking in spoken conversation directly
to SCMC discourse is not profitable (Beißwenger 2008). The chapter now turns to
an alternative approach to cohesion: the conversational floor.
ström’s (1994: 34) definition of the turn as “everything A says before B takes over,
and vice versa” is basic but entirely workable in SCMC, a discourse environment
where turns cannot be co-constructed and where there is no overlap. It is a techni-
cal definition with little ambiguity. Definition of the floor is less clear-cut, depend-
ent as it is upon inferring how participants themselves view the unfolding dis-
course. The definition as provided by Edelsky (1981: 405) in her influential work is
as follows:
The floor is defined as the acknowledged what’s-going-on within a psychological time/
space. What’s going on can be the development of a topic or a function (teasing, solicit-
ing a response, etc.) or an interaction of the two. It can be developed or controlled by
one person at a time or by several simultaneously or in quick succession. It is official or
acknowledged in that, if questioned, participants could describe what’s going on as
“he’s talking about grades” or “she’s making a suggestion” or “we’re all answering her”.
A reading of Edelsky suggests that there are three definable elements to the floor:
(1) the topic, or the “aboutness” of the discourse; (2) the communicative action: how
things are being said in the discourse; and (3) the participants’ sense of what is hap-
pening in the conversation. From the analyst’s point of view, these are each evident
only to the extent to which they can be inferred from the text, a constraint which
should be acknowledged as a caveat in a discussion of floor. Nonetheless, the text of
SCMC allows an analyst to gain a closer participant’s sense of what was going on
than, for example, a transcription of spoken discourse. This is because the chat par-
ticipants themselves are denied the range of visual and aural feedback cues avail-
able in spoken discourse; any ratification must ipso facto appear in the text itself.
“Simply talking, in itself, does not constitute having the floor”, note Shultz,
Florio, and Erikson (1982: 95). “The ‘floor’ is interactionally produced, in that
speakers and hearers must work together at maintaining it”. Thus one can be the
speaker but not hold the floor. In her study of floor and gender patterns in asyn-
chronous CMC, Herring (2010) supports Edelsky’s assertion that to be a floor-
holding turn, a message must be ratified by other participants. In spoken discourse,
such ratification can be done verbally or through non-verbal nods and backchan-
nels. In SCMC, floor ratification can also be done verbally or through responses
that represent non-verbal behaviour.
mal, cooperative ventures” (Edelsky 1981: 416). Herring (2010) found that these
two floor types were evident in her study of asynchronous discourse in two aca-
demic discussion lists.
Missing from Edelsky’s bipartite distinction are cases where two or more floors
of conversation are continuing in parallel. A broader classification deriving from
research into dinner table conversation and classroom discourse by Shultz, Florio,
and Erikson (1982) (also in Erikson and Shultz 1977) posits categories of partici-
pation structure where floors are single or multiple. Although there are further sub-
divisions in this classification, single floors, broadly speaking, correspond to
Edelsky’s F1 and F2: a single speaker with a number of attenders, or a floor which
is more collective or collaborative. Multiple floors, “type IV participation struc-
tures” in the typology of Shultz et al., are described by these authors (1982: 102) as
having “subgroups of the persons present participating in topically distinct simul-
taneous conversations”.
Hayashi’s (1991) summary grouping of floor types draws on the findings of
Shultz et al. and Edelsky. Hayashi also divides floor types into single conversa-
tional floors and multiple conversational floors, and also subdivides the single floor
type into the single person floor and the collaborative floor. Further sub-categori-
sations are described, based on relative levels of interaction. Hayashi’s taxonomy
was adapted by Cherny (1999: 176ff) to describe floor types in SCMC discourse in
a social MOO.
erature on CMC; see Turkle (1995) on the elasticity of roles and identity on-screen,
Cherny (1999) on the rich online life of a MOO community, as well as Smith and
Kollock’s (1999) collection of articles on online communities. Taxonomies of
roles in online groups orient around the type and level of interaction at which par-
ticipants engage, and include such roles as: core participants, newbies, flamers, and
lurkers (Marcoccia 2004; Turner, Smith, Fisher, and Welser 2005; White 2001).
The examples in this chapter come from an online language learning community,
where teacher/learner roles are more flexible than they are in archetypal face-to-
face classrooms. On the subject of the roles of teachers and students face-to-face
and online, see Warschauer (1996, 1999), Kern (1995), Salmon (2004), Godwin-
Jones (2005), and Garrison and Cleveland-Innes (2005).
The illustrative data in this chapter are from a virtual group dedicated to exploring
ways of language learning online, Webheads. This long-standing and expanding
online community of English language learners and teachers was instigated by
Vance Stevens and colleagues in 1998 and has been meeting on the Internet ever
In example 3, BJB ratifies Susanne’s turn with a ‘nod’. This is an action, a turn
sent in the third person to represent non-verbal behaviour (on performative actions
in text-based CMC, see Virtanen, this volume):
Floor ratification by members of the Webheads group has the dual purpose of
signalling both that the participants are paying attention to the floor holder and that
they comprehend what has been written. In her investigation into backchannel re-
sponses in a MOO, Cherny (1999: 194) similarly maintains that “it is difficult if not
impossible to separate affect out from the back channel function in this medium,
since an appropriate emotional response to a turn (e.g., a laugh) indicates both at-
tention and understanding just as well as a nod does”.
Turns are often directed in SCMC by naming the participant to whom they are
addressed. This cohesive device characteristic of SCMC, cross-turn reference (Her-
ring 1999) or addressivity (Werry 1996), is used by Maggi in turn 8 of example 4 to
repair the misunderstanding. In other cases, as with example 5 below, addressivity
is included in the original floor-holding and floor-ratifying turns (turns 1 and 4).
This can be considered a navigation technique in response to the fact that there are
a number of participants and potentially simultaneous multiple threads of conver-
sation in the interaction. Note that turns 2 and 3 belong to a different thread from
turns 1 and 4.
structure across contexts of SCMC use. Three habitually-occurring floor types that
also occur in MOOs are described and illustrated below: the speaker-and-supporter
floor, the collaborative floor, and the multiple conversational floor.
Note that the verbal activity is “instructing” here: Vance is instructing Maggi
and Ying-Lan how to create a web page, step-by-step, and their attention is di-
rected towards him. It is no surprise, therefore, for the floor pattern most associated
with the activity “instructing” is a speaker-and-supporter floor.
Within a multiple conversational floor, as Cherny (1999: 176) notes, there can
be a main floor and side floors, or there can be two or more main floors running in
parallel. In SCMC discourse it is possible for an individual participant to be in-
volved in more than one floor of conversation. In the above example of a multiple
conversational floor, three of the four participants contribute to both floors. This
tendency of the proficient SCMC participant to switch between floors is an echo of
other traits of CMC use. For example, multitasking – attending to a number of dif-
ferent on-screen activities at once – is commonplace (Jones 2002). And in some
SCMC contexts, participants cycle among on-screen identities that they have cre-
ated (Donath 1999; Turkle 1995).
In example 6 above (the speaker-and-supporter floor), one participant was ex-
plaining to others how to do something – in this case, how to build a website. This
is in contrast to the pattern in example 7 (the collaborative floor), where the par-
ticipants could be said to be “chatting”, which is, after all, the prototypical activity
in a chat room. Again it is notable but not surprising that a certain type of com-
municative action (e.g., “chatting”) is associated with certain floor types: collab-
orative or multiple. In the following section, I expand on this point, relating floor
development to the topic and purpose of the conversation (the verbal activity), the
role relationships of the participants, and the computer-mediated nature of the dis-
course.
In the discourse of the virtual community under consideration here, Webheads, the
status of the participants and their various role relations are influential in shaping
16 LianA says, “it said if you burn the scandle from 2 ends, you
will be a busy man.”
17 BJB thinks there are several threads to this conversation
18 VanceS says, “Also you will burn yourself out”
19 LianA says, “welcome vance to wuhan next time to china.”
20 PhilB says, “Normally to ”burn the candle on both ends“
means to work so much you tire yourself out. With ”scandal“
instead of ”candle“ it sounds like Bill Clinton with his hot
interns. <g>”
21 BJB chuckles. Same result, though.
22 VanceS. o O (this is a normal consequence of multitasking)
23 VanceS says, “he must have had too many hot interns in the
fire”
24 LianA giggles
25 PhilB asks, “Hey, I found a new free resource called ”stuf-
fincommon virtual communities“. Anyone heard of it?”
26 VanceS says, “never”
27 Sue [guest] says, “no”
28 LianA says, “no”
29 PhilB asks, “Wanna see?”
30 VanceS says, “sure”
31 Sue [guest] says, “sure”
32 PhilB says, “It has chat, tools, and a neat whiteboard.”
33 LianA says, “yes.”
34 PhilB asks, “I’m going to project. Sue, Lian, do you know
about projections?”
35 LianA says, “yes”
36 Sue [guest] says, “not sure”
Example 9a.
1 VanceS says, “I never knew what chili was exactly before”
3 LianA says, “come to china, then you will know what it is,
vance.”
6 PhilB says, “Vance, there’s a lot of confusion between the
words ”chili“ and ”chile“ (borrowed from Spanish.”
15 VanceS says, “I’ve been to China several times, but never to
Wuhan”
Contextual and temporal aspects of the discourse would suggest that much at-
tention is paid to individual participant’s topic in SCMC. In example 9a above,
within a broad topic framework that could be said to be about chili, LianA’s topic is
China. This also becomes Vance’s topic in turn 15; a topic drift has taken place
within a floor of conversation. Neither the topic of chili nor of China is developed
further in the interaction.
Example 9b is the end of another conversation. BJB has been explaining to Sue
how to open the graphical interface of Tapped In:
Example 9b.
2 BJB. o O (that will open a web window to go with your text
client)
4 BJB. o O (I hope)
BJB is using a device whereby her turn is displayed inside an ASCII “thinks”
bubble. This is done in Tapped In by prefacing the turn with the command /thinks.
One might infer that she uses this technique because the turns are directed towards
only one among many participants. However, it is possible in Tapped In to send a
turn privately to another participant using the /whisper command. That BJB does
not do this suggests, in line with her role at Tapped In, that she feels the in-
formation might be of use to more than one participant.
Example 9c is an exchange of three turns spread over seven turns of the extract.
Example 9c.
7 Sue [guest] asks, “Lian, how much did you take on GRE?”
9 LianA says, “not very high, only 2160”
13 Sue [guest] says, “so hight? i am wondering i can only take
1500”
The two language learners here are discussing an English language test. Although
perhaps not highly proficient in English, they are both adept at SCMC discourse.
Both Sue and LianA participate in more than one conversation in this extract; turns
by LianA appear in four of the six isolated examples highlighted here.
Example 9d is an aside:
Example 9d.
17 BJB thinks there are several threads to this conversation
22 VanceS. o O (this is a normal consequence of multitasking)
The topic here is the conversation itself. In spoken discourse the turns would be
expected to appear together, as an adjacency pair or as the initiation and response
of an exchange. However, in this sample of SCMC they are separated by four
unrelated turns. Also of note is the fact that BJB’s turn is posted in the third per-
son as a metacomment; that is, a comment on the unfolding conversation. Vance’s
turn is also a representation of something other than speech, using, as BJB did ear-
lier, the cartoon “thinks” bubble (see Cherny 1995 on the modal complexity of
This is a collaborative floor, in as much as four participants are involved in its de-
velopment. My contention is that it dominates because the verbal activity is “ex-
plaining about a phrase LianA has read”, which involves the communicative action
“explaining”. The topic, raised quite explicitly by LianA in turn 8, is a phrase that
LianA has presumably read or heard and that she wants help in understanding. As
noted above, LianA is an English language learner, and Webheads is a virtual com-
munity dedicated to language learning. In other words, when a language learner
raises a language learning point, much of the focus redirects towards that particular
floor, the floor becomes collaborative, and the communicative action of the verbal
activity orients towards “explaining”.
In example 9f the floor type can also be attributed directly to the participant and
the verbal activity:
Example 9 f.
25 PhilB asks, “Hey, I found a new free resource called ”stuf-
fincommon virtual communities“. Anyone heard of it?”
26 VanceS says, “never”
This is a single floor with one floor holder being supported by other participants.
In the terminology adopted here from Hayashi (1991) and Cherny (1999), it is a
speaker-and-supporter floor. The communicative action of the verbal activity is
primarily didactic: PhilB is “demonstrating an Internet resource called stuffincom-
mon”. There is a sub-topic in turns 34 to 36: using the “project” command in
Tapped In. The communicative activity remains explanatory.
In analysing this extended example, the concern has been with a limited set of
patterns of participation and floor types. There are undoubtedly many other pat-
terns that relate to the development of other floor types. This notwithstanding, the
conclusion can be drawn from this analysis that floor development – in these data
at least – is related to verbal activity. When the topic of the verbal activity is a lan-
guage point raised by a learner, the floor becomes collaborative. It either develops
as the single collaboratively constructed floor or as the main floor of a multiple
conversational floor. When the topic is related to the technologies of electronic lit-
eracy (for example, how to build a website, or where a particular resource can be
found) the floor develops into a speaker-and-supporter floor. The participant role
is also important. In each case the communicative action (explaining/demonstrat-
ing) is pedagogic. However, when the verbal activity is directly related to the ac-
quisition of the second or foreign language (English), a number of participants
contribute substantive turns. When the verbal activity is related to the development
of the skills of electronic literacy, other participants focus their attention on the
single floor holder.
6. Conclusion
Notes
1. This chapter contains data and analysis first discussed in Simpson, James 2005 Conver-
sational floors in synchronous text-based CMC discourse. Discourse Studies 7(3):
337–361.
2. A MOO is a type of MUD (multi-user domain), a user-extensible text-based virtual
reality environment (see Cherny, 1999).
3. Tapped In: http://www.tappedin.org
4. The Palace: http://www.thepalace.com
5. The TOEFL, or Test of English as a Foreign Language, is an admission requirement for
non-native English speakers at many English-medium universities.
6. The term “expert user” is used in preference to the term “native speaker” because expert
users of a language are not necessarily native speakers. For discussion of this and other
issues surrounding the notion of the “native speaker”, see Rampton (1990).
References
Androutsopoulos, Jannis K.
2006 Introduction: Sociolinguistics and computer-mediated communication. Jour-
nal of Sociolinguistics 10(4): 419–438.
Beißwenger, Michael
2008 Situated chat analysis as a window to the user’s perspective: Aspects of tem-
poral and sequential organization. Language@Internet 5, article 6.
http://www.languageatinternet.org/articles/2008/1532/index_html/
1. Introduction
Jacobs and Garcia, this volume). Schönfeldt and Golato (2003) found that partici-
pants in German IRC channels adapted the same basic repair mechanisms used in
oral conversation to the chat medium. Trouble sources, once identified, caused the
ongoing sequence to be stopped while the trouble was addressed. Schönfeldt and
Golato noted that the preference for self-repair in oral interaction holds for chat, as
well. The most frequent type of repair they found was other-initiated self-repair,
followed by self-initiated self-repair, despite the fact that the sequential locations
normally associated with repair in oral conversation (transition space, third turn)
were not available in the chat medium. Schönfeldt and Golato also identified a
trouble source unique to the chat environment, the non-response (Rintel, Pittam,
and Mulholland 2003): Participants frequently initiated repair when a first-pair
part was felt to have been ignored.
Indeed, frequent silences or gaps between posts are a prominent feature of chat
interaction. In examining such gaps in IRC, Rintel et al. (2003) found that partici-
pants employed specific strategies to help them differentiate among types of am-
biguous non-responses. One of the most common strategies was to clarify that
a non-response was due to user action and not to a technical system problem. To
do this, the IRC users engaged in activities such as re-greeting, reconnecting, or
checking connections with the system. Rintel et al. (2003) also noted that many of
the problems associated with ambiguous non-responses could be eliminated if IRC
or other chat systems adopted monitoring systems. These are now common fea-
tures of various instant messaging systems, whereby the system tells users when an
interactional partner has begun typing a message.
It is clear from the research discussed thus far that although obvious points of
difference between oral conversation and chat exist, chat participants often make
use of (and adapt) strategies for the management of oral conversation to chat en-
vironments. The next section looks at research that specifically examines how chat
conversations are organized with respect to turn management and coherence.
cisely the sequential organization of turns as they are displayed. The result is dis-
rupted turn adjacency; turns that are designed to be adjacent are “interrupted” by
other turns. This problem is especially common when there are large numbers of
participants in a chat room (Herring 1999).
The inevitability of disrupted turn adjacency means that sequential organiz-
ation cannot be relied upon to provide coherence in chat interaction (see also
Jacobs and Garcia, this volume; Simpson, this volume), and in fact, Herring (1999)
points out that violations of sequential coherence are the rule in most forms of
CMC. As a result, participants have adapted other strategies to cope with the dis-
jointed nature of chat conversations, for example, using punctuation or other means
to indicate that a turn is incomplete (Herring 1999). Werry (1996) identified four
conventions typically used in IRC to manage conversations: addressivity, abbrevi-
ation (of turn length and words and phrases), paralinguistic and prosodic cues, and
actions and gestures. Of these, addressivity, or the practice of routinely naming the
intended recipient within each post, has been shown to be important for conversa-
tional coherence in large group chat interactions (Herring 1999; Panyametheekul
and Herring 2003; Simpson 2005). A common convention for addressivity in chat
is to type the recipient’s name followed by a colon before beginning a turn, as in:
<Shaquille> ariadnne: what the hell does that mean?
<ariadnne> shaq: what are you yapping your lips about? (Werry 1996: 52)
In most chat systems, the speaker’s name, here enclosed in angle brackets, is
automatically appended at the beginning of the message, and the message sender
optionally designates the intended recipient of the turn. This is not the only format
for addressivity, as any use of the recipient’s name in a post will suffice; it should
also be noted that turns may be addressed to all participants or to no one in particu-
lar (Werry 1996: 53).
Addressivity also plays a role in turn allocation. In oral interaction, the sequen-
tial organization of talk is provided for by the turn-taking system (Sacks, Scheg-
loff, and Jefferson 1974: 704). Part of the turn-taking system includes provisions
for managing speaker exchange (i.e., allocating turns). Generally, a current speaker
may select the next speaker, who in turn selects the next, and so forth. If the current
speaker does not select a next speaker, then a next speaker may self-select, with the
rights to the turn going to the first starter. If a new speaker does not self-select, then
the current speaker may continue. This process recurs so as to provide for minimal
gap and overlap between speakers.
Turn-taking in chat is a disjointed process, resembling less of an exchange sys-
tem than a system of adding to an evolving conversation. In fact, it is necessary to
distinguish between turns and posts in chat. A post is one complete unit of text sent
to the chat window, after which another post may appear. It is the arrangement of
posts in the chat window that provides a semblance of a turn exchange system. In
contrast, a turn may be defined as a complete idea or thought unit. Participants may
include multiple turns in a single post or spread one turn across multiple posts, al-
though the default rule in chat is one turn per post (Condon and Čech 2001). There-
fore, in such data a turn construction unit (TCU), or a chunk of text that is analy-
zable as a possibly complete turn (Sacks, Schegloff, and Jefferson 1974), may
encompass part of a post, a complete post, or multiple posts. Accordingly, the tran-
sition-relevance place (TRP), which in oral interaction is the first place where a
next speaker can start a turn so as to provide for minimal gap and overlap (Sacks,
Schegloff, and Jefferson: 703–704), is more nebulous in chat.
However, to establish coherence in chat conversations, participants must still
orient to some form of turn allocation. Panyametheekul and Herring (2003) found
that, although the technical design of chat systems favors speaker self-selection,
the turns that were most successful in garnering responses in a Thai web chat room
were those that followed a current-speaker-selects-next strategy, which was gen-
erally achieved through addressivity. In their data, self-selected turns were the least
successful, and turns that were constructed as current speaker continues were only
moderately successful, because they represented the continuation of initially un-
successful turns. Panyametheekul and Herring conclude that current-speaker-se-
lects-next is the preferred turn allocation strategy for promoting coherence in Thai
web chat interactions.
Using data from small group, classroom-based chat interactions, Garcia and
Jacobs (1999) described a turn-taking system for the medium. In contrast to other
research on chat, Garcia and Jacobs did not rely on chat logs for data analysis, but
instead used video recordings of each participant’s computer screen as the primary
data. They described the participant roles in chat as those of message constructor,
message poster, waiter, reader, and worker (when participants were engaged in non-
screen activities). Garcia and Jacobs noted that in chat, participants could play
multiple roles at one time. They also described the differences between the roles of
current and next speaker in face-to-face and chat interaction. They found that the
current-speaker-selects-next technique was not always successful and argued that
in chat, a “current” poster can really only select a “future” poster, as opposed to a se-
quentially next poster. Garcia and Jacobs also argued that the strategy of next poster
self-selects is best understood as future poster self-selects, again because partici-
pants cannot guarantee that their posts will be placed next to the intended referent.
Overall, the findings reviewed here underscore that problems can occur when
participants attempt to rely solely on sequential placement of turns to provide co-
herence (see also Jacobs and Garcia, this volume). Different strategies have been
adapted to cope with temporal displacement and provide coherence. In particular,
previous research has shown that discourse topics can be a source of coherence in
chat (Herring 2003), instant messaging (Herring and Kurtz 2006; Örnberg Ber-
glund 2009), and in text messages sent via interactive television (Zelenkauskaite
and Herring 2008). In the next section, I demonstrate how an orientation towards
discourse topics helps to establish conversational coherence in the context of vir-
tual team meetings.
3. Case study
The data for this illustration are drawn from a larger case study (Markman 2006) of
conversational structures in virtual team meetings. The team members were five
undergraduate students from a large university in the southwestern United States who
were enrolled in a five-and-a-half week summer semester independent study course
designed to teach them how to work in a virtual (i.e., physically dispersed) team. The
students held four virtual meetings using the chat function in the Blackboard course
management system. (They did not receive any specific training on the Blackboard
system or chat interface.) The author also participated in the team meetings.
Because CA demands that analysts study the unfolding of interaction, moment
by moment, it is necessary to study computer chat as it happens, rather than after
the fact, in order to describe precisely how participants manage conversation in
these spaces. To do this, software was installed on each participant’s computer in
order to record the activities visible on each member’s screen during the team
meetings. Unfortunately, due to technical difficulties, screen recordings are not
available for all team members for the first two meetings. The chat conversations
were also logged automatically by the Blackboard system. Both the video record-
ings and the chat logs were used as data for the analysis.
The video recordings from each team member’s computer screen were tran-
scribed using a system of the author’s design based on the principles of conver-
sation analysis. Text was transcribed exactly as entered, timed in one-second inter-
vals and aligned with the time stamps given to each turn in the chat by the server.
Pauses were counted when 10 or more frames of the video recordings indicated no
activity. Symbols were used to indicate cursor movement, deleted text, and the ac-
tion of pressing the “enter” key or clicking “send”. Other screen actions were de-
scribed verbally and enclosed in double parentheses. Each team member’s individ-
ual transcript was then integrated into a single transcript to allow for side-by-side
comparison of the entire team’s interaction. In the transcription examples in this
chapter, the Chat Window column indicates turns posted to the chat, accessible to
all, and the Chat Time column indicates the clock time (in hh:mm:ss) of actions in
the chat window and on the participants’ computer screens. The remaining col-
umns show the individual team members’ screen actions. Abbreviated versions of
the full team transcripts are presented here due to space considerations. A descrip-
tion of the transcription conventions is provided in the appendix.
Cadiz, and Burkhalter 2000 for an example of true threaded chat). However, a
number of scholars have identified conversational threads as a recurring feature of
chat (Greenfield and Subrahmanyam 2003; Holmer 2008; O’Neill and Martin 2003;
Shi et al. 2006; Simpson 2005; Zitzen and Stein 2004), where threads are defined as
a series of related messages (Holmer 2008; Shi et al. 2006) or parallel conversations
(Greenfield and Subrahmanyam 2003; O’Neill and Martin 2003; Simpson 2005,
this volume). What makes a “thread” in chat is not a visual tie between related posts,
but rather the observable contextual relations (O’Neill and Martin 2003: 44) be-
tween messages, most notably the discourse topic. Thus conversational threads,
understood as a series of topically-related messages, serve as a primary source for
coherence in synchronous chat, although in chat rooms with large numbers of par-
ticipants and multiple topics of conversation, topics tend to decay rapidly, making
conversational threads inadequate as a sole source of coherence (Herring 1999).
A detailed examination of the conversation in the chat meetings presented here
shows that participants organize their conversations and maintain coherence pri-
marily through an orientation towards discourse topics in their patterns of turn con-
struction. Specifically, I demonstrate how adjacency sequences, lexical cohesion,
and addressivity are used in turn design to contribute to a coherent discussion. In
all examples, the names of the team members have been changed and other iden-
tifying information has been omitted to protect confidentiality.
(1)
Chat Time Chat Window THADINE REBECA
5:21:17 REBECA: what do you think? is it ((pause))
doable?
5:21:18
(omitted) [omitted]
5:21:49 REBECA: what areas are we all inter-
ested in reearching?
5:21:50
((pause))
5:21:51
5:21:52
5:21:53
5:21:54 ((returns to chat))
5:21:55 ok
5:21:56 I-f
5:21:57 ang,
5:21:58 it’s
5:21:59 ((pause))
5:22:00 on Bl
5:22:01 ackboar
5:22:02 d ((pause))
5:22:03 ((pause)) now
Prior to this excerpt, team member Rebeca had ascertained that the other
members present at the meeting had received some documents she had emailed to
them earlier in the day. She then posted two questions in a single turn at 5:21:17,
designed to elicit an assessment of those documents from her teammates. After a
17-second pause, during which no new turns were posted to the chat, Rebeca began
typing a new turn that was posted at 5:21:49. This turn is also phrased as a ques-
tion, this time introducing a new topic (what the team planned to study for their
project). The next contribution to the chat was provided by Thadine, who had been
complying with a request made by another team member. Thadine began typing as
soon as she returned to the chat window, and her turn at 5:22:04 indicated that the
request had been fulfilled.
After a pause, Thadine turned her attention to Rebeca’s two sets of questions at
5:22:14. As transcript 1 shows, Thadine began by addressing Rebeca’s two turns in
one post, addressing the first question (“is it doable?”) first, then linking this
answer with the conjunction “and” at 5:22:20 to her answer to the second question
(“what areas are we all interested in re[s]earching?”). However, Thadine only gets
as far as the word “interested” before pausing and then, at 5:22:43, deleting the sec-
ond half of her turn. She then immediately sends the first response at 5:22:47 and
without pausing, re-types her second response, which she completes at 5:22:56. By
deciding to split her response into two consecutive turns, Thadine mimics the
structure of Rebeca’s original posts and aligns her responses to their respective
topics.
Rebeca’s turns at 5:21:17 and 5:21:49 and Thadine’s turns at 5:22:47 and
5:22:57 display some of the elements commonly used in these data to construct
turns that are sensitive to topical coherence. Adjacency sequences, such as ques-
tions/answers and greetings, are frequently used. By phrasing her two posts as
questions, Rebeca makes it more likely not only that her turns will receive re-
sponses, but that the responses will be in a form, namely answers, that will be more
easily matched to the appropriate prior turns. However, it is often the case that
other elements are necessary to make the topical connection clear. Thadine pro-
vided an answer to Rebeca’s first post with her turn at 5:22:47 by providing an as-
sessment. Thadine could have simply answered “yes” and achieved the same
meaning. Instead she repeats specific lexical items from Rebeca’s turn, namely “do
think” and “doable”, to make her second pair part cohesive with Rebeca’s first pair
part. “Cohesion is a semantic relation between an element in the text and some
other element that is crucial to the interpretation of it” (Halliday and Hasan 1976:
8). One type of cohesive tie that is found recurrently in these data is that of reiter-
ation (Hasan and Halliday 1976: 288); see also Greenfield and Subrahmanyam
2003). Lexical cohesion is also illustrated by Thadine’s repetition of the word “in-
terested” in her turn at 5:22:57, which links her turn to Rebeca’s post at 5:21:49.
Reiteration can also be accomplished with a synonym, as in Evan’s turn at 5:22:44.
By starting his turn with “I”, Evan indicates his inclusion in “we all” of Rebeca’s
question. Additionally, in Evan’s turn “like” provides a semantic match with “in-
terested”, and “wireless” has a cohesive tie with “areas” through collocation (Hal-
liday and Hasan 1976: 285), as it had been discussed earlier in the meetings.
The analysis of example 1 has thus far demonstrated how the use of lexical
cohesion and adjacency sequences can contribute to a topically-coherent interac-
tion. Before moving on, it is worth noting another element present in this extract.
Previously, the practice of addressivity was identified as a coherence-maintaining
device commonly used in public chat rooms. Interestingly, the use of address terms
is relatively uncommon in the virtual team data. In fact, these data do not contain
any instances of the name+colon convention described by Werry (1996). When ad-
dress terms are used in these meetings, it is usually in greetings (e.g., “hi evan” or
“hi guys”), to express agreement, support for an idea, or thanks (e.g., “yeah I-fang”,
“rebeca that sounds pretty cool …”, and “thanks evan”), or less frequently, as re-
pair initiators (e.g., “what do you mean sid?” and “sorry, that was for Rebeca”). In
example 1 above, Thadine’s post at 5:22:04 (“ok I-fang, it’s on Blackboard now”)
is an example of another use for address terms. Thadine’s turn is specifically ad-
dressed to the team member I-Fang, who approximately five minutes earlier had
made the following request:
5:17:03 I-FANG: can someone post what rebeca sent … coz i cant access my email …
5:17:13 THADINE: it’s kinda long
5:17:22 I-FANG: on blackboard i mean
Thadine’s use of an address term (I-Fang) at 5:22:04 serves to help link her turn to
the correct conversational thread by indicating the intended recipient.
3.2. Speakership roles and turn allocation in the virtual team meeting
Further evidence of the role of discourse topics in interactional coherence becomes
apparent when considering the speakership roles oriented to by the team members.
I use terms such as “speaker” and “conversation” here deliberately to reinforce my
assertion that chat interactions are a type of text-based talk-in-interaction. This
section shows how these speakership roles are related to conversational threads
and provides examples that demonstrate that participants are orienting to two basic
units in the chat conversation system: the post and the turn, as discussed earlier. Al-
though Garcia and Jacobs (1999) identify the transition-relevance point in chat as
following a complete, posted message, there is evidence in these data that the par-
ticipants orient to multiple possible points for beginning a new turn.
In contrast to large public chat rooms such as those studied by Panyametheekul
and Herring (2003), Rintel and Pittam (1997), and Werry (1996), the number and
frequency of posts in these small group meetings are such that participants did not
find it necessary to designate a turn recipient for most turns. Moreover, because
chat does not allow participants to control the sequential order of posts in the chat
window, they are not able to designate a specific, sequentially next speaker; the
next possible speaker is anyone/everyone, and thus there is no need to do much in
the way of formally allocating turns. However, it is sometimes useful to design
turns in such a way that they make particular team members relevant future speak-
ers. Because there is a conceptual difference between a post and a turn in CMC, it
is also possible for chat participants to adopt the role of a continuing current
speaker. Thus, participants in the virtual team meetings may adopt three possible
speakership roles: self-selected new speaker, continuing current speaker, and rel-
evant future speaker. The next sections examine how these roles work in conjunc-
tion with conversational threads.
(3)
Chat Time Chat Window EVAN THADINE
5:11:48 THADINE: the only thing about the
articles Rebeca found (and I didn’t read
them all yet) is that I believe we’re
supposed to be focusing on what’s
going on in [city name]
5:11:49
5:11:50
5:11:51
5:11:52
5:11:53 and ev
5:11:54 en more
5:11:55 specifica
5:11:56 Ye lly
5:11:57 ah, ((pause))
5:11:58 ((pause)) a ((pause))
5:11:59 the r
5:12:00 genera ound camp
5:12:01 THADINE: and even more specifically l idea pus
around campus
5:12:02 is t ((pause))
5:12:03 oSget
5:12:04 oSt
5:12:05 I-FANG: are we? ude
5:12:06 nts per
5:12:07 spective
5:12:08 s…
5:12:09 thou
5:12:10 gh
5:12:11 rweaa
5:12:12 re able
5:12:13 rto exp that’s th
5:12:14 and ou e imp
5:12:15 tside of th res
5:12:15 [KRIS joined the session] ((pause)) sion
5:12:16 at I
5:12:17 ((pause)) got
5:12:17 THADINE: that’s the impression I got
5:12:18 EVAN: Yeah, the general idea is to get
Students perspectives … though we are
able to expand outside of that
(4)
Chat Time Chat Window SIDNEY THADINE
5:28:23 so
5:28:24 li thought
5:28:25 lthe who maybe
5:28:26 le pr ((pause))r , bu
5:28:27 urp ((pause)) t it
5:28:28 ose of all has
5:28:29 othis ((pause)) to
5:28:30 proje ((pause)) be out
5:28:31 ct of class
5:28:32 cwas to f stuff,
5:28:33 ind ((pause)) the in and
5:28:34 ((pause)) nov grad
5:28:35 ators reserrarch
5:28:36 not to isn’t exacy
Whereas one strategy for maintaining the role of current speaker is to continue
a single turn over multiple posts, these data also show that participants may act as
current speakers over multiple turns. In these instances, TCUs may be typed con-
tinuously or near continuously, as seen in the previous examples, or they may be
separated by a brief pause. What marks these posts as continuations, as opposed to
self-selected new turns, is that they are analyzably complete turns that relate to the
same topic. Extract 5 shows a typical example of this phenomenon.
(5)
Chat Time Chat Window EVAN I-FANG
(omitted)
5:32:44
5:32:45 ble ((pause))
i
5:32:46 EVAN: I noticed that we are all bringing ((pause)) d ((pause))
5:32:47 our own definitions of technology to the
5:32:48 table W out
5:32:49 e all of or
5:32:50 assume it dinary
5:32:51 has to do way?
((pause))
5:32:52 I-FANG: cool … but did he do it in a out with com
5:32:53 of ordinary way?
5:32:54 puters
5:32:55 and elec
tronic
5:32:56 EVAN: We all assume it has to do with s
computers and electronics
5:32:57 ((pause))
(omitted)
5:33:01 it
5:33:02 MOOM does
5:33:03 Comp nt have
5:33:04 uters to be
5:33:05 speci … but
5:33:06 fically those
5:33:07 EVAN: Computers specifically are the
5:33:08 But most p
5:33:09 I was opula
5:33:10 in the show r ((pause))
5:33:11 er an
5:33:12 d realiz ones
5:33:13 ed that
At 5:32:45, Evan has completed a self-selected turn that introduces a new topic
into the conversation. After a very brief pause, he begins typing a new turn. Evan’s
new turn is cohesively linked to his first turn specifically by repeating the pronoun
“we” and using the superordinate “computers and electronics” for “technology”
(Halliday and Hasan 1976: 288). The transcript also shows that, although an unre-
lated turn by I-Fang (at 5:32:52) gets posted to the chat before Evan’s second turn,
he began typing his turn before any responses had been received. It is possible that
the silence had motivated him to expand on his original idea, but given the length
of his pause (just under three seconds), it is unlikely that anyone else in the meeting
would have had a chance to respond before he began typing again, thus reinforcing
this as an example of a current speaker attempting to hold the floor. After Evan sent
his post at 5:32:56, he followed a similar pattern, pausing for approximately four
seconds before beginning another turn that further refined his previous post. Evan
posted “Computers specifically” at 5:33:07, then immediately continued typing a
new turn at 5:33:08, which displays a link to the previous set of turns through the
use of the conjunction “but”. What Evan was able to accomplish in extract 5 was a
set of four individual yet related turns that, taken together, expressed and elabor-
ated on a single idea. In doing so, he was also able to establish a new topic and was
successful in moving the conversation in that direction.
Thus far it has been demonstrated how participants in these virtual meetings
adopted speakership roles as a self-selecting new speaker or as a continuing current
speaker. The final segment of data analysis examines the third speakership role
present in these data, that of the relevant future speaker.
Just as questions are commonly used to make general future speakers rel-
evant, as seen in example 1, they are also employed when a response from a
specific person is sought. Example 6 shows a typical instance. Here, Evan re-
quested information about the status of part of the team’s project. Thadine was
the team member responsible, and thus Evan addressed his turn specifically to
her at 5:12:26. Thadine did not have the chat window visible on her computer
screen at the time Evan’s turn posted, but five seconds after she returned to the
chat window she replied to his question, displaying her orientation as a specific
relevant future speaker. Thadine does not use an address turn in her response,
but by recycling the phrase “turni[ng] that in” from Evan’s turn in her post,
Thadine makes a cohesive tie to Evan’s question (Halliday and Hasan 1976).
In general, these data show that questions are the most common way that spe-
cific future speakers are made relevant, but other first pair parts are sometimes
used.
As noted above, general questions and other first pair parts make all partici-
pants generally relevant as future speakers. However, there are a small number of
instances where a specific future speaker is made relevant, but a second individual
treats the first pair part as an opportunity to respond as a general future speaker (see
also Panyametheekul and Herring 2003).
(7)
Chat Time Chat Window SIDNEY
10:12:43 KRIS: if you set your display to 256 colors it will
reduce the file size
(omitted)
10:13:43 I-FANG: how do we set it to 256 colors?
(omitted)
10:13:50
10:13:51
10:13:52 right
10:13:53 click
10:13:54 THADINE: ok, blue bars gone now. sorry about on
that
4. Discussion
The findings described above illustrate that several factors play important roles in
how people organize and maintain coherence in chat conversations. These factors
include the size of the group, the specific technical features of the chat environ-
ment, and the purpose and context of the talk (institutional versus recreational,
etc.).
The size of the chat group is important in a number of respects. With a smaller
number of people in a chat room, there is less competition for the floor. More par-
ticipants generally equals more turns, and therefore potentially increases not only
the instances of disrupted turn adjacency but also the distance between adjacent
turns. To compensate for the increased competition, participants in large public
chat rooms tend to construct relatively short turns (Cherny 1999; Werry 1996).
Posting shorter turns more frequently increases the likelihood that a given turn will
be posted close to its antecedent turn, thereby potentially securing the floor for a
given participant. By contrast, the turns in the data discussed here varied in size,
but they were generally longer than turns observable in data drawn from large pub-
lic chat rooms. It was not uncommon for turns in these data to consist of multiple
clauses or sentences, a characteristic also observable in the small chat groups
studied by Garcia and Jacobs (1999).
In addition, the size of the group appeared to influence participants’ use of
addressivity, which was much less frequent than what has been observed in large
chat rooms (e.g., Werry 1996). Finally, the size of the group, coupled with team
members’ tendency to post long turns, meant that there were fewer turns overall
posted during these meetings, compared to what might be expected in a large, pub-
lic chat room. Features of the chat environment also crucially affect the overall or-
ganization of conversation in this medium. The persistence of talk is an advantage
that text-based conversational systems have over oral interaction. Many chat sys-
tems used in education, such as the one used here and those examined by Simpson
(2005, this volume) and Colomb and Simutis (1996), offer users access to an on-
going log of the interaction, allowing them to scroll back to previous parts of the
conversation. This feature was used recurrently by the team members in this study,
and on some occasions it contributed directly to the formulation of new turns. Al-
though not all chat environments offer a logging feature, Herring (1999) notes that
even in such cases, chat environments are more persistent than in oral interaction,
because the most recent posts are typically available for viewing in the chat
window for some period of time before they scroll up and off the screen. The abil-
ity to read and re-read prior turns gives chat participants additional resources for
turn construction, and by extension, for conversational coherence.
In chat interactions with numerous participants, it could be decidedly disad-
vantageous to act as a continuing current speaker by spreading turns over multiple
posts or by posting multiple, related turns. However, in small group meetings, this
orientation can be an effective strategy for securing attention and increasing the
chances of holding the floor. Multiple, related posts have a visual presence that can
compensate for the egalitarian nature of the technology. Thus while the chat sys-
tem’s self-select feature means that no one is prevented from taking a turn at any
time, the current speaker continuous role gives participants the opportunity to
make their textual “voices” somewhat more prominent, which may encourage
others to yield the floor.
5. Conclusion
Convention/symbol Explanation
Double olleh bob Deleted characters/spaces
strikethrough
Deleted characters are shown in reverse
order, unless the entire word and/or phrase
was selected and deleted in one action, in
which case the deleted information is shown
in normal order, e.g., hello bob
Double ((mouses to chat Transcriber description/comments
parentheses window))
Enter/Send
References
Boden, Deirdre
1994 The Business of Talk: Organizations in Action. Cambridge, MA: Polity Press.
Bolden, Galina B.
2006 Little words that matter: Discourse markers “so” and “oh” and the doing of other-
attentiveness in social interaction. Journal of Communication 56: 661–688.
Cherny, Lynn
1999 Conversation and Community: Chat in a Virtual World. Stanford, CA: CSLI
Publications.
Colomb, Gregory G. and Joyce A. Simutis
1996 Visible conversation and academic inquiry: CMC in a culturally diverse class-
room. In: Susan C. Herring (ed.), Computer-Mediated Communication: Lin-
guistic, Social, and Cross-Cultural Perspectives, 203–222. Amsterdam/Phil-
adelphia: John Benjamins.
Condon, Sherri L. and Claude G. Čech
1996a Functional comparisons of face-to-face and computer-mediated decision mak-
ing interactions. In: Susan C. Herring (ed.), Computer-Mediated Communi-
cation: Linguistic, Social, and Cross-Cultural Perspectives, 65–80. Amster-
dam/Philadelphia: John Benjamins.
1. Introduction
The social use of technology has become an increasingly important area of study as
different forms of technology transform ever more arenas of social life. By chang-
ing the medium through which humans act and construct meaning, computer-
mediated communication (CMC) creates a new interactional environment, which
requires participants to alter the techniques and procedures they use to organize
their actions and interpret the actions of others.
Textual Chat is used in informal interactions as well as in a variety of institu-
tional contexts (Garcia and Jacobs 1998, 1999; Gergle, Millen, Kraut, and Fussell
2004; Golato and Taleghani-Nikazm 2006; Jepson 2005; Ligorio 2001; Markman
2009; Rintel and Pittam 1997; Rohfeld and Hiemstra 1995; Veerman, Andriessen,
and Kanselaar 2000). When the pragmatic features of oral conversation (such as
appropriateness of contribution, speed of response, maintenance of flow, feedback,
and repair; Todman and Alm 2003) are compared to the features of chat room in-
teraction, it is clear that many of these features are not used consistently. This can
be a main source of interactional troubles and a major cause of failures of intersub-
jective understanding in chat.
Research on the pragmatics of repair in interaction generally involves the ap-
plication of conversation analytic techniques and/or linguistic approaches to the
analysis of spoken data. The phenomenon of repair has been explored in a wide
variety of languages, including English (McKellin et al. 2007; Schegloff 1987b),
German (Egbert 2004; Rieger 2003; Schönfeldt and Golato 2003), Italian (Bosco,
Bucciarelli, and Bara 2006), Norwegian (Svennevig 2008), Japanese (Fox,
Hayashi, and Jasperson 1996), Spanish (Mason 2004), Thai (Moerman 1977), and
the Caribbean language of Bequian (Sidnell 2007). Most of the research on repair
in oral communication has been on informal conversation, either face-to-face or
over the telephone.
In oral conversation, error repair procedures enable participants to present their
understanding of the interaction to others and to identify and correct misunder-
standings or mistakes as they arise (Jefferson 1974; Kurhila 2001; Sacks, Scheg-
loff, and Jefferson 1974; Schegloff 1987a; Schegloff, Jefferson, and Sacks 1977).
Kötter (2002) addresses the role of repair in achieving intersubjectivity, which he
discusses in terms of the negotiation of meaning. For example, a question creates a
space for an answer and an expectation that it will be the next thing that happens in
the conversation (Heritage 1984). This structure helps participants interpret the
Garcia and Jacobs (1999), using videotapes of chat participants’ computer screens,
showed that chat turn-taking systems provide for the selection of future, rather than
next, “speakers” (message posters). Because chat software does not allow partici-
pants to determine exactly where in the conversation their message will be posted,
a message may not always be adjacent to its referent, and the practicality of struc-
tures such as adjacency pairs (e.g., question-answer pairs) for achieving intersub-
jectivity is diminished. If the answer to the question is not placed adjacent to the
question, it may not be interpretable as an answer to that question. While these
types of placement problems can sometimes be identified and resolved by partici-
pants, situations often cause each participant to have a different understanding of
what is happening in the conversation.
The message display window and printout, therefore, often do not accurately
represent the actual sequence of turns at talk intended by participants (Garcia and
Jacobs 1998, 1999; Herring 1999; Negretti 1999; Panyametheekul and Herring
2003). However, much of the previous research on turn-taking in chat does not
analyze the message construction process necessary to discern how turn exchange
is managed in chat and thus does not explain the significance of the inability of the
turn-taking system in chat to manage distribution of next turns.
Several studies have noted the interactional confusion that can result in chat
when messages are not posted where participants intended them to be placed. Her-
ring refers to this displacement as “disrupted adjacency” (1999); Garcia and Jacobs
(1998) refer to the misinterpretation of adjacency caused by “phantom responsive-
ness” and “phantom adjacency pairs” in their study of university classroom parti-
cipants. Negretti (1999) analyzed printouts of web chat conversations among
36 graduate students, including both native and ESL learners, and found that over-
lapping turns prevented participants from being able to negotiate starting, finish-
ing, and giving turns as they would in face-to-face interactions. Nonetheless, par-
ticipants “competently managed” (79) the complexity of an interaction in which
“the expectation of a relevant reply to each turn is often unfulfilled or delayed”
(81). Tan and Tan (2006) compared turn-taking patterns in two instructional dis-
cussions between teachers and their classes, one conducted face-to-face and the
other via an online chat discussion program. They found the pattern of turn-taking
to be markedly different in the two conditions, with many more students producing
responses and less topical continuity in the chat condition. Kitade (2000) analyzed
second-language learners, using qualitative observations including discourse
analysis and conversation analysis to study IRC conversations among students and
native speakers in college Japanese classes. She found that the prevalence of
multiple topics overlapping and disrupted adjacency caused frustration among
some participants. One way that participants compensated for the confusion was to
use address terms. Simpson (2005) also analyzed participants practicing second-
language skills using an online website and found that disrupted adjacency was not
a problem for participants, who expected the conversation to be disjointed and
were able to discuss multiple threads in patterned yet intersecting interactions.
However, Nilsen and Mäkitalo’s (2010) study of professionals in an online, in-ser-
vice training course found that participants not only noted that their interaction was
disjointed but also interrupted their activity in order to address the discontinuity
before continuing their conversation. (See also Herring, this volume, on partici-
pants’ reactions to “loosened relevance” in multiparty chat.)
Some studies have specifically addressed repair in chat communication. Many
of these studies address, directly or indirectly, the role of turn organization in the
need for and/or accomplishment of repair. Schönfeldt and Golato’s (2003) study of
German-language web chats noted that, while many messages seemed to be unre-
lated, participants could read the different threads and reconstruct the adjacency of
each topic. They found that initiation and completion of repair followed a pattern
similar to that in oral conversation, with most incidences of repair being self-initi-
ated and self-repaired within the same turn. Also, like Kitade (2000), they found
that participants used address terms to distinguish which prior speaker was being
addressed. Jepson (2005) compared the frequency of repair in text and voice chats
of ESL students and found that more error repair occurred in the voice chat, which
he attributed primarily to pronunciation-related corrections. However, these
studies have all relied on logs or printouts of the conversations.
Zemel and Cakir (2009) analyzed math problem-solving sessions in chat by
using a software tool that allowed them to replay the sessions in real time based on
time-stamped records of participants’ actions. They found that – unlike in oral con-
versation, in which overlapping turns compromise speakers’ comprehensibility –
displaced turns in chat did not affect the “readability” of the text (270). None-
theless, participants oriented to the problem of intelligibility by constructing repair
messages to indicate both that it was a repair and what it was repairing.
More recently, some researchers have also begun to examine video screens,
as well as participants themselves. Beißwenger’s (2008) study of 18 chats using
32 participants analyzed video recordings not only of their computer monitors but
also of the participants themselves in order to record data on “facial expression,
gesturing, posture, and verbal uttering” (7); he found that deletions of text were
often a result of a participant’s having read a new onscreen message. Marcoccia,
Atifi, and Gauducheau (2008) analyzed recordings of four instant messaging dis-
cussions and discovered that the lack of interactivity in instant messaging did not
prevent participants from producing “interactional gestures or facial expressions”
(8), making instant messaging comparable to telephone conversation. Markman
(2010) studied four virtual meetings via the chat function of “Blackboard” that
took place among five undergraduate students and their instructor in a university
course. Computer software was used to record participants’ computer screens, and
both video recordings and chat logs were used for analysis. She found that partici-
pants often engaged in “self-initiated self-repair” by correcting mistakes or typo-
graphical errors in their immediate next turns and commonly used asterisks to in-
dicate when their posts were repairs of prior turns.
In this chapter we show that the turn-taking system of chat is critically conse-
quential for the organization of repair and the avoidance of troubles in talk. With
videotapes of each participant’s computer screen, we have access to what was
typed in the message input boxes as well as the posting box; thus we can determine
where participants intended to place their messages by analyzing when they started
typing in relationship to when others’ messages were posted (see Garcia and Jac-
obs 1998, 1999; Jacobs and Garcia 2002). These data enable us to determine pre-
cisely how participants produce, avoid, and repair troubles in chat.
3. Methodological approach
action, and a search for the ways in which intersubjectivity is achieved. Levinson
(1983) integrates the analytical tools of ethnomethodological conversation analy-
sis (e.g., Sacks et al. 1974) with linguistic approaches.
The study of the basic organizing principles of oral conversation, such as turn-
taking (Sacks et al. 1974), repair (Jefferson 1974), and preference organization
(Schegloff et al. 1977), has proved essential to understanding participants’ actions
in a wide variety of settings. As the research reviewed above has shown, conver-
sation analysis is increasingly being used to understand interaction in chat in both
informal and institutional contexts. In this chapter we use conversation analysis to
extend understanding of repair in chat conversations in a college classroom.
4. Data
Transcript, page 3
1 Lizard: WHASSUP!
2 Frank Sinatra: HELLO.
3 Lizard: HELLO, BLUE EYES … P=0:30
*******
Lizard
Frank Sinatra
B=0:31 W=0:39–0:46, ECUTEP=0:52
WHAT D YOU THINK THE “MASK”
Purple
Transcript, page 4
1 Lizard: WHASSUP!
2 Frank Sinatra: HELLO.
3 Lizard: HELLO, BLUE EYES …
4 Frank Sinatra: CUTEP=0:52
*******
Lizard
B=0.56HE
Frank Sinatra
Purple
B=0.53HELLOP=0.57
The posting box on page 3 of the transcript shows that three lines have al-
ready been posted at this point in the conversation. The most recent posting,
Lizard’s “HELLO, BLUE EYES …”, was posted (P) at 30 seconds into the con-
versation, as indicated by the superscript italics P=0:30. Frank Sinatra began
typing (B) his next message 31 seconds into the conversation (superscript
B=0:31), waited (W) for seven seconds after typing “WHAT D YOU THINK
THE ‘MASK’” (superscript W=0:39–0:46), and then erased (E) his in-progress
message (superscript E and the subscripted text of his message). He then typed
and posted “CUTE” 52 seconds into the conversation (P=0:52). The other par-
ticipants would see only Frank Sinatra’s edited message “CUTE” when it ap-
peared as line 4 in the posting box; they did not know that he had begun to
compose a different message. Page 4 of the transcript then shows the partici-
pants’ actions between the time Frank Sinatra posted line 4 and PURPLE posted
the next message.2
Since chat is a written mode of communication, the types of errors that can be
avoided and the interactional context of their avoidance differ in some respects from
those of oral conversation. While some errors in chat are analogous to errors made
in oral conversation (i.e., typographical errors and slips of the tongue), other types
of errors are either unique to chat or more common in chat than in oral conversation.
For example, in oral conversation, participants generally respond to the most recent
utterance, using misplacement markers such as “by the way” (Schegloff 1987b) to
refer to earlier utterances. In our data, chat participants do not typically mark a
message as out of place even if it does not refer to an immediately prior message.
One reason is that, as they are typing a message, participants do not always know
whether their message will be placed adjacent to its intended referent, because the
textual nature of chat requires a more active role on the part of participants to see,
literally, that others are participating. Unless a participant is actually looking at the
posting box, she may not be aware of whether – or when – others are participating.
Participants in these data respond to postings in two ways: 1) by reading the
posting box as a document and 2) by reading it as a conversation. When reading the
posting box as a document, the reader starts with the first message posted since the
last time she read the posting box and read (and responded) to messages in the
order in which they were posted. When reading the posting box as a conversation,
the reader first reads the last message posted, responds to that, and then works her
way back to respond to previous postings, if necessary. Participants may use the
document reading mode habitually, or the conversational mode habitually, or
switch from one to the other as convenient. These two ways of reading the posting
box make for different types of errors. For example, if one participant is using a
conversational mode and responding to the most recent posting, a reader using the
document reading mode may interpret that message as a response to an earlier
message. Whichever mode participants use, if they read just far enough to find a
message they want to respond to and then type a response before reading other
messages, they run the risk of missing something or confusing another participant.
The locations for error repair in chat differ greatly from those in oral conver-
sation, given that there is no opportunity for self-repair of a posted message until a
future turn or for other-repair until a message is posted. Because chat participants
may be typing messages at the same time, it is common for one participant to post a
message while other participants are typing. Writers sometimes take advantage of
this feature to edit their utterances-in-progress in response to messages that are
posted while they are composing, which may enable the writers to avoid a place-
ment problem or simply provide another choice about what to respond to in the
conversation. For example, in Excerpt 1 the students are discussing a poem they
have read; the most recent posting is Lizard’s line 65 (“I SEE THE AFRICAN
THING AS A VALID INTERPRETATION, BUT I DON’T SEE IT MYSELF”),
which contradicts something PURPLE had posted earlier. PURPLE was engaged
in typing when line 65 was posted (no wait time is indicated in her message entry
box). The underlining of “DOE” in PURPLE’s message entry box shows that
PURPLE had composed that part of her message before line 65 was posted.
PURPLE types as far as “DOES HE WANT US TO HEA”, but apparently she
reads Lizard’s message while typing because she erases her message-in-progress
and immediately types “WHAT DO YOU MEAN YOU DON’T SEE IT YOUR-
SELF”, which she posts in response to Lizard’s line 65.
EXCERPT 1
44 PURPLE: THE AUTHOR IS SARCASTIC BECAUSE HE KNOWS HOW HE MUST HIDE
45 HIS TRUE EMOTIONS
46 PURPLE: DO YOU GUYS AGREE THAT HE IS WRITING FROM AN AFRICAN
47 AMERICAN STANDPOINT
48 Lizard: I WAS HOPING HE WAS SARCASTIC, BECAUSE ENJOYING THIS MASK IS
49 A BAD WAY TO LIVE
50 Lizard: I DONT SEE THE AFRICAN PART
51 PURPLE: YES BUT LOOK AT THE REASONS SURROUONDING WHY “WE” MUST
52 WEAR THE MASK
53 Frank Sinatra: IN LINE THREE, “THIS DEBT WE PAY TO HUMAN GUILE”, HE SEEMS TO
54 MEAN THAT THE MASK IS HARMFUL IN SOME WAY; PERHAPS CAUSING
55 THE “BLEEDING HEARTS OF LINE 4
56 Lizard: IT HURTS TO WEAR IT
57 PURPLE: HIS BACKGROUND, AND THE TIME FRAME IS FROM THE LATE 1800’S
58 AND THE AFFRICAN AMERICAN COULD NOT SPEAK HIS MIND BECAUSE
59 HE RISKED LOSING HIS LIFE, BECAUSE “WE” WERE NOT ALLOWED TO
60 SPEAK HOW WE REALLY FELT
61 Frank Sinatra: WHAT DOES THAT MEAN THOUGH
62 Frank Sinatra: DOES HE WANT US TO WEAR IT OR NOT?
63 PURPLE: OF COURSE IT HURTS TO WEAR IT BECAUSE HOW WOULD YOU LIKE TO
64 BE SHUT OUT FROM YOUR OWN WAY OF EXPRESSION.
65 Lizard: I SEE THE AFRICAN THING AS A VALID INTERPRETATION, BUT I DON’T
66 SEE IT MYSELFP=11:14
**************
Lizard:
Frank Sinatra:
B=11:30LI
NOTE: From 11:14 TO 11:27 Frank Sinatra uses his cursor to scroll through the chat box. He dis-
plays lines 17–25 (only the first two lines of line 25 are displayed)
PURPLE
This excerpt shows a chat participant editing her message in response to an-
other participant’s message, posted during the composition of her own message.
This type of self-repair is private in that other participants do not know that a mess-
age has been edited, but it is interactive because PURPLE clearly changed her
message in response to Lizard’s posting. Goodwin (1995: 117) has noted that
“speakers [in oral conversation] change the emerging structure of sentences even
as they are speaking them in order to maintain the appropriateness of their talk for
the dynamic situation which the sentence both emerges from, and helps to further
constitute”. Clearly, the same is true of participants in chat, who, despite not hav-
ing listener responses to the messages they are currently constructing, can nonethe-
less be responsive to others’ postings (Beißwenger 2008; Garcia and Jacobs 1998,
1999).
While the location for self-repair in chat is still within the writer’s turn, there is
no opportunity for intervention by others, given that the message entry box is pri-
vate. Therefore, chat participants need not worry about whether another can make
sense of their message-in-progress; others will only evaluate the final product.
Within-turn self-repair (editing of the message-in-progress) is, interactionally
speaking, more analogous to what Jefferson (1974) calls error avoidance in oral
conversation than to error repair.
Chat participants can perform self-repair of errors in a future turn position. If
they act quickly and are lucky, the repair may indeed be posted after the message
they are attempting to self-repair; however, other participants’ posts may inter-
vene. Hence, although sequences may be found in chat that superficially resemble
repair sequences in oral conversation, they are only coincidentally similar because
chat participants do not have the same control over placement that participants in
oral conversation do.
Excerpt 2 shows that PURPLE began typing “PLEASE EXPLAIN” in her
message composition box at 11:37 as a self-repair of her post in line 67, which was
posted at 11:31 and seemed like a challenge (“WHAT DO YOU MEAN …”).
PURPLE did not mitigate or delay this disagreement with Lizard (Pomerantz
1984) and used format tying (“ … YOU DON’T SEE IT YOURSELF”), which
Goodwin and Goodwin (1987) identified as an argumentative move. Format tying
is when participants in a dispute tie their oppositional utterances to prior talk by
creating their turn out of components of the prior turn. Her use of “please” (a pol-
iteness marker) in her repair message suggests that she is trying to mitigate or re-
pair the argumentative tone of line 67. PURPLE succeeds in having her “PLEASE
EXPLAIN” message placed adjacent to her complaint by typing and posting it
quickly. However, given that the turn-taking system in chat posts messages in the
order in which they are received and not in the order in which participants begin
typing them, this placement was partly due to luck. Another participant could have
posted a message before PURPLE’s, thus displacing “PLEASE EXPLAIN” and
rendering it impotent as a repair.
EXCERPT 2
46 PURPLE: DO YOU GUYS AGREE THAT HE IS WRITING FROM AN AFRICAN
47 AMERICAN STANDPOINT
48 Lizard: I WAS HOPING HE WAS SARCASTIC, BECAUSE ENJOYING THIS MASK IS
49 A BAD WAY TO LIVE
50 Lizard: I DONT SEE THE AFRICAN PART
51 PURPLE: YES BUT LOOK AT THE REASONS SURROUONDING WHY “WE” MUST
52 WEAR THE MASK
53 Frank Sinatra: IN LINE THREE, “THIS DEBT WE PAY TO HUMAN GUILE”, HE SEEMS TO
54 MEAN THAT THE MASK IS HARMFUL IN SOME WAY; PERHAPS CAUSING
55 THE “BLEEDING HEARTS OF LINE 4
56 Lizard: IT HURTS TO WEAR IT
57 PURPLE: HIS BACKGROUND, AND THE TIME FRAME IS FROM THE LATE 1800’S
58 AND THE AFFRICAN AMERICAN COULD NOT SPEAK HIS MIND BECAUSE
59 HE RISKED LOSING HIS LIFE, BECAUSE “WE” WERE NOT ALLOWED TO
60 SPEAK HOW WE REALLY FELT
61 Frank Sinatra: WHAT DOES THAT MEAN THOUGH
62 Frank Sinatra: DOES HE WANT US TO WEAR IT OR NOT?
63 PURPLE: OF COURSE IT HURTS TO WEAR IT BECAUSE HOW WOULD YOU LIKE TO
64 BE SHUT OUT FROM YOUR OWN WAY OF EXPRESSION.
65 Lizard: I SEE THE AFRICAN THING AS A VALID INTERPRETATION, BUT I DON’T
66 SEE IT MYSELF
67 PURPLE: WHAT DO YOU MEAN YOU DON’T SEE IT YOURSELFP=11:31
**************
Lizard:
B=11:36I DIDN’T
Frank Sinatra:
LIW=11:31–36ZARD, YOU SA
PURPLE:
B=11:37PLEASE EXPLAINP=11:42
EXCERPT 3
10 Lizard: THE WAY WE REACT AND COEXIST WITH THE EXTERNAL WORLD
11 PURPLE: FOR EXAMPLE WHEN AS PERSON IS UPSET AND MAD, HE MAY HIDE
12 IHIS FEELINGS AND SHOW HAPPINESS FOR EVERYONE (P 2:42)
13 Lizard: RIGHT ON (P 2:56)
14 Frank Sinatra: I THINK THAT IT REFERS TO THE WAYS THAT PEOPLE USE TO HIDE
15 THIER FEELINGS, EPECIALLY IN THE INCREASINGLY HOSTILE WORLD
16 IN WHICH WE LIVE. (P 3:01A)
17 PURPLE: I AGREE LIZARD, (P 3:01B)
18 PURPLE: WHAT DO YOU GUYS THINK THE SECOND STANZA MEANS (P 3:32)
19 Frank Sinatra: ANY PARTICULAR LINE? (P 3:46)
20 Lizard: THIS “MASK” KINDA KEEPS A COHESION BETWEEN PEOPLE THROUGH
21 THE UGLINESS OF DAY-TO-DAY (P 3:53)
22 Lizard: SORRY, WE’VE MOVED ON (P 4:08)
23 Lizard: PEOPLE SHOULDN’T GET TO CLOSE (P 5:01)
24 Lizard: KNOW TOO MUCH (P 5:07)
25 Frank Sinatra: THE SECOND STANZA SEEMS TO BE TALKING FROM A DIFFERENT
26 POINT OF VIEW THAN THE OTHER TWO. IT TALKS ABOUT THE MASK AS
27 A GOOD THING WHEREAS THE OTHERS SEEM TO POINT OUT THE BAD
28 PARTS OF IT. (P 5:40A)
29 PURPLE: I BELIEVE IT MEANS THAT WHY SHOULD EVERYONE KNOW EXACTLY
30 HOW A PERSON IS FEELING OR WHAT A PERSON IS THINKING ABOUT.
31 SO THE WRITER IS STATING THAT HID YOU R FEELINGS AND KEEP
THIEM TO YOURSELFP=5:40B
Subsequent pages of the transcript show that PURPLE continues typing without
waiting while Lizard posts two messages, lines 20 and 22, the last one at 4:08 (see
Excerpt 3). Frank Sinatra waits 14 more seconds for a reply from PURPLE to his
repair initiation attempt and then begins to type an analysis of stanza two. Lizard
posts two more short messages (lines 23 and 24) while Frank Sinatra is composing
his message, which he then posts at 5:40 (line 25). At the same time, PURPLE fin-
ally completes her long message (line 29), which is posted virtually simultaneously
at 5:40 and appears in the posting box just after Frank Sinatra’s message.
PURPLE’s message in line 29 provides a retrospective warrant for her failure to
answer Frank Sinatra’s question in line 19: She was engaged in typing a lengthy
message and may not have seen his question prior to posting it. In this way, the in-
teractional organization of chat provides a rationale for PURPLE’s failure to re-
spond to Frank Sinatra’s question.
Although Frank Sinatra’s line 19 was an adjacency pair first pair part (a ques-
tion) and no second pair part was produced, the absence of an answer is not treated
as problematic by the participants. The participants appear to be viewing this “ab-
sence” through a chat lens, in which delayed responses are not unusual. In chat,
participants know that a “silent” participant may be engaged in typing a message,
scrolling up in the posting box, or reading hard documents rather than new post-
ings. Chat participants, therefore, may choose not to repair silence. Frank Sinatra
lets his repair initiation attempt go unanswered and ultimately answers PURPLE’s
line 18 without aid from her as to which line to focus on. PURPLE’s missing re-
sponse to Frank Sinatra’s question is a type of “placement error” in the interac-
tional environment of chat; it is a silence only from the point of view of Frank Sin-
atra and Lizard. From PURPLE’s point of view, there was no silence because she
was engaged in typing and did not become aware of the question Frank Sinatra
asked until later. In sum, this excerpt shows a repair initiation attempt that was un-
successful because of the interactional organization of chat but which was not
problematic for the interaction, in as much as Frank Sinatra displayed an orien-
tation to how chat works and went on with the conversation.
Excerpt 1 (above) shows an instance in which an other-initiation of repair was
successful in eliciting a repair from the person producing the “error”. However, it
took three attempts at other-initiation of repair before the requested repair was
produced. In Excerpt 1 Frank Sinatra initiates an other-repair of Lizard’s line 56
(“IT HURTS TO WEAR IT”). Ten seconds later, Frank Sinatra begins his re-
sponse (“WHAT DOES THAT MEAN THOUGH”), which is posted (line 61)
after an intervening post from PURPLE (see Excerpt 1). Frank Sinatra does not
use completion punctuation at the end of line 61 and quickly types a follow-up
message (“DOES HE WANT US TO WEAR IT OR NOT?”), which he posts at
10:48 (Excerpt 1, line 62). This follow-up message in line 62 could be an attempt
to clarify the referent of his line 61; by using the words “to wear it”, Frank Sinatra
uses format tying (Goodwin and Goodwin 1987) to link his message with Lizard’s
line 56 rather than PURPLE’s line 57, and then he waits for a response to his
messages. Instead of a response from Lizard, a posting from PURPLE appears,
and then another posting from Lizard (line 65) in response to an earlier message
from PURPLE.
Frank Sinatra then begins another message (see Excerpt 4), explicitly ad-
dressed to Lizard (“LIZARD, YOU SAID EARLIER THAT THE AUTHOR
WANTED US TO HIDE OUR FEELINGS, HAS YOUR OPINION
CHANGED?”), in which he revises and reissues his previous repair initiation.
While Lizard was “silent”, Frank Sinatra waited because he could assume that Liz-
ard was composing a response to his (Sinatra’s) repair initiation; but after Lizard
posted line 65, it was clear that he was not. At that point Frank Sinatra begins typ-
ing the elaboration/repair of his repair initiation to make a third attempt to get Liz-
ard to respond. This third attempt at repair initiation is not immediately successful
because the next message Lizard posts is a response to one of PURPLE’s earlier
messages. However, Lizard then types “NO”, which he posts as line 73. This “NO”
is an answer to Frank Sinatra’s repair initiations.
EXCERPT 4
65 Lizard: I SEE THE AFRICAN THING AS A VALID INTERPRETATION, BUT I DON’T
66 SEE IT MYSELF
67 PURPLE: WHAT DO YOU MEAN YOU DON’T SEE IT YOURSELF
68 PURPLE: PLEASE EXPLAIN
69 Frank Sinatra: LIZARD, YOU SAID EARLIER THAT THE AUTHOR WANTED US TO HIDE
70 OUR FEELINGS, HAS YOUR OPINION CHANGED?
71 LIZARD: I DIDN’T READ IT FROM THE STANDPOINT OF AN AFRICAN AMERICAN. I
72 AM WHITE, AND I CAN SEE THIS APPLYING TO MYSELF
73 LIZARD: NOP=12:34
The sequence in Excerpt 4 differs in several ways from repair sequences in oral
conversation. First, the original other-repair initiation was not placed in the turn
after the trouble source, because chat software allows other participants to post
messages at will and does not guarantee placement of a message at a specific point.
The delayed placement of Frank Sinatra’s first other-repair initiation attempt could
have made it difficult for Lizard to interpret (hence, Frank Sinatra’s repair of his re-
pair message in line 62). Also, because Lizard was engaged in responding to
PURPLE’s postings, he was not yet at leisure to respond to Frank Sinatra. Finally,
when he does respond to Frank Sinatra’s third attempt at error repair, his “NO” is
not placed adjacent to the repair initiation and hence may be difficult to interpret.
However, in this case, Frank Sinatra seems able to interpret Lizard’s “NO” as a re-
sponse to his query; he responds to him in his next message (not shown). This re-
pair sequence was eventually completed successfully; however, because of the in-
teractional organization of chat, it took three attempts and was not sequential.
In sum, these examples of error avoidance and self- and other-repair in chat
show that the interactional organization of chat, in particular its turn-taking sys-
tem, leads to a different configuration of error and error repair than exists in oral
conversation. The placement of messages in slots other than those in which the
poster intended or desired them to be placed, or in slots other than those adjacent to
the message they refer to, is a common event in chat, with deep implications for the
organization of repair.
We found that two types of placement problems occur in chat. The first is an actual
displacement of a post, in which a participant’s message is posted after one or more
other messages intervene between it and its referent. The second occurs when a
participant fails to take steps to avoid misunderstandings that can result from dis-
placed posts. This could be thought of as a failure to anticipate potential displace-
ment of a message.
a convenient time to post it in order to minimize the chance that it is placed awk-
wardly in the conversation. If his or her post would inappropriately appear to be a
response to an immediately prior post, the writer can edit the message before post-
ing to avoid comprehension problems.
Even when participants wait to post, however, placement problems can occur,
simply because it is not possible to know what other participants are writing in
their message entry boxes. An example of pausing to avoid a placement problem
can be seen in Excerpt 5.
EXCERPT 5
1 Mr White: Hello, is anybody there? (0:01)
2 Silver: hello everybody (0:08)
3 Mr White: Hi ho, silver (0:24)
4 Silver: hey what’s up (0:40)
5 FRED: YABBA DABBA DO! (1:01)
6 Mr White: hey Fred (1:12)
7 Silver: So who want to go first? (1:17)
8 Mr White: I will (1:23)
9 FRED: HI MR WHITE AND SILVER (1:25)
10 Silver: Okay which paper are you going to revise (1:40)
Although Silver’s post in line 7 appears to follow the opening sequence seam-
lessly, she actually waited approximately 12 seconds after having typed this mess-
age before she posted it. Silver waited to begin the substantive part of their con-
versation until it appeared that all participants had logged on and posted a greeting.
Mr White accepts Silver’s invitation to discuss his paper (“I will” – line 8), but
FRED’s continuation of the greeting sequence (line 9) interrupts the conversation.
While a participant can increase the chance that his or her message will not be
displaced by waiting to post it, other participants’ posts can still intervene. Check-
ing the posting box before posting may help a participant to avoid some placement
problems; however, the longer a poster waits, the more likely it is that other par-
ticipants will post messages, increasing the likelihood that his or her post will be-
come “out of date”.
there is no mechanism for reserving a next slot for a specific poster. Subsequent re-
search has noted that participants in chat can use address terms to minimize the
likelihood of confusion if a chat message ends up not being posted adjacent to its
intended referent (e.g., Herring 1999; Kitade 2000; Kötter 2002). In our data, chat
participants also use address terms to ensure that a post can be understood even if a
placement problem occurs. Address terms can be used to select a specific future
poster (Garcia and Jacobs 1999) and also retrospectively to indicate to which prior
message they are referring, as in Excerpt 6.
EXCERPT 6
9 Purple: I THINK THE MASK IS A TYPE OF FRONT FOR EMOTIONS (1:50)
10 Lizard: THE WAY WE REACT AND COEXIST WITH THE EXTERNAL WORLD (2:39)
11 Purple: FOR EXAMPLE WHEN AS PERSON IS UPSET AND MAD, HE MAY HIDE
12 IHIS FEELINGS AND SHOW HAPPINESS FOR EVERYONE (2:42)
13 Lizard: RIGHT ON (2:56)
14 Frank Sinatra: I THINK THAT IT REFERS TO THE WAYS THAT PEOPLE USE TO HIDE
15 THIER FEELINGS, ESPECIALLY IN THE INCREASINGLY HOSTILE WORLD
16 IN WHICH WE LIVE. (3:01A)
17 PURPLE: I AGREE LIZARD, (3:01B)
18 PURPLE: WHAT DO YOU GUYS THINK THE SECOND STANZA MEANS (3:32)
When Lizard posts line 10, PURPLE is engaged in typing a long message (line 11)
and therefore cannot respond immediately to Lizard’s post. However, after
PURPLE posts line 11, she begins to type a response to Lizard (“I AGREE”).
While she is working on this message, Lizard posts line 13 (“RIGHT ON”) as a re-
sponse to line 11. PURPLE then completes and posts her message (“I AGREE LIZ-
ARD”) virtually simultaneously with Frank Sinatra’s post in line 14; his is posted
first, so hers appears as line 17. Because PURPLE knows that she has posted a
lengthy message (line 11) and that others’ posts might intervene between Lizard’s
line 10 and her agreement with it, she cannot be sure that her “I AGREE” statement
will be understood as agreement with Lizard’s line 10 unless she addresses Lizard
by name. PURPLE’s precaution pays off: Frank Sinatra’s post intervenes, yet
PURPLE’s use of an address term ensures that the participants know with whom
she agrees. However, address terms are not a surefire way of counteracting place-
ment problems. Note that because Lizard says “RIGHT ON” in line 13 (responding
to PURPLE’s line 11), participants could mistake PURPLE’s line 17 as a response
to line 13 instead of to line 10.
EXCERPT 7
77 Frank Sinatra: BUT IF IT HURTS SO MUCH TO DO SO, WHY WOULD HE PUSH US TO DO
78 SO. (13:29)
79 Lizard: IS HE PUSHING US? (13:43A)
80 PURPLE: TRUE, SO YOU THINK HE IS MEANING THIS POEM TO
81 INCLUDE EVERYONE (13:43B)
82 Lizard: YES (13:51)
Lizard has posted several displaced messages throughout the conversation and has
tried several methods to prevent displacement of his posts (including composing
short or split messages, as seen above). In line 79 he repeats the phrase “PUSHING
US” from Frank Sinatra’s line 77. Because Lizard used format tying, his question
will be understood as a response to Frank Sinatra’s question, even if it is displaced.
Had he written simply “IS HE?”, his post would have been vague enough to cause
confusion if another message had intervened. In fact, line 80 is a question to which
“IS HE?” would be a reasonable response (and to which Lizard actually responds
“YES”). If misplaced, Lizard’s line 79 could have created confusion were it not for
format tying.
Differences in how chat and oral conversation are organized have profound impli-
cations for the organization of repair. A fundamental problem for chat participants
is, given that messages may be displaced, how do participants locate a message’s
referent? If the context is unclear, how can the conversation proceed?
The interactional problems regarding repair in chat differ from those in oral
conversation because the resources available to participants for avoiding and re-
pairing errors in timing, production, and comprehension of messages differ in the
two contexts. The frequency and location of errors in chat occur in part because the
chat software does not allow participants to view each other’s message composi-
tion boxes and because it controls placement of the messages, such that simulta-
neous turns are impossible. Therefore, participants have no routine way of know-
ing where a participant intended to place a displaced message; however, messages
may intervene unproblematically between a message and its referent if they will
not appear to be possible antecedents to that message.
Our analysis has shown that chat participants may have difficulty understand-
ing the meaning of a message and may need to re-read several lines or posts
preceding the message in question, and not just the immediately preceding line or
post, in order to understand it. Even in conversations in which all participants are
fluent and skillful in chat, failures of interpretation will occur because the actual
sequence of messages (the order in which they are noticed and read) may differ for
each participant.
Learning to avoid and repair errors in chat involves not only learning to change
one’s lens from an oral to a chat perspective, but also learning to negotiate between
oral and chat approaches within a single conversation. Hence, our participants’
“errors” or misunderstandings are the products sometimes simply of the nature of
chat and its probability of producing displaced messages, sometimes of over-re-
liance on conventions of oral talk in a non-oral medium, and sometimes of incom-
patibility of varying approaches across different participants in the conversation.
Researchers must therefore keep both interactional frames in mind when analyzing
chat conversations.
Skilled chat participants understand that lack of complete comprehensibility of
the conversation resulting from placement errors is the norm, and they do not seek
to understand every message that is displaced. They at times agree to misunder-
stand each other, not to respond to each other’s questions or repair initiations, not
to edit their messages in response to priors, etc., in the interests of keeping up with
the flow of the conversation. Accordingly, placement problems in chat are not
necessarily errors; they are part of the territory, and participants do not always treat
them as troubles in the talk or orient to them as errors. The result is that participants
do not always understand one another, but they understand that they will not
always understand, and this is intersubjectivity of a sort. They learn to negotiate
between oral and chat structures and conventions; they construct and interpret
messages as if they refer to priors and search out possible referents if the message
is not understandable as referring to a prior.
Understanding how the interactional work is done is critical to understanding
who says what to whom and what each is reading/understanding. Understanding
how repair works is an important part of this process. If one goal in using computer
technologies in the classroom is to facilitate collaborative learning (e.g., Kötter
2002), we must begin by understanding how effective those technologies are and
how they might be made more effective. Teachers should be aware of the types of
communication problems that can occur in chat so that they can help students avoid
errors and misunderstandings and help to make chat a more valuable learning tool.
Discovering how participants understand chat as it is happening and how errors
and misunderstandings can arise may help eliminate some of the problems inherent
in those interactions in the classroom, in the workplace, and in personal use by giv-
ing participants a more comprehensive understanding of the interactional impli-
cations of the technology.
Notes
1. A large body of research in workplace studies addresses what participants are doing
bodily with regard to paper documents, gaze, and movements, and how these are coordi-
nated with communication technology (e.g., Crabtree et al. 2000; Heath et al. 2001; Luff
and Heath 2002; Suchman 1987). In this chapter our data are limited to what is on the par-
ticipants’ computer screens. Since students were aware that they were being studied, the
study itself may have affected the types of interactions they had. However, the students
were used to having their chat room interactions observed, because the instructors rou-
tinely “listened in” on their chat room conversations to make sure that students were on
task. If these data were being collected today, screen save programs would have been
used instead of video recorders.
2. The data collection process required students to use television monitors instead of com-
puter monitors so that the videotaping equipment could be connected to the computers.
Because visual clarity was not as sharp as it is on a computer screen, we asked the re-
search participants to type in capital letters to improve visibility.
References
Beißwenger, Michael
2008 Situated chat analysis as a window to the user’s perspective: Aspects of tem-
poral and sequential organization. Language@Internet 5, article 6.
http://www.languageatinternet.org/articles/2008/1532
Bosco, Francesca M., Monica Bucciarelli, and Bruno G. Bara
2006 Recognition and repair of communicative failures: A developmental perspec-
tive. Journal of Pragmatics 38: 1398–1429.
Cooke, Maeve
1998 Introduction. In: Jürgen Habermas, On the Pragmatics of Communication,
1–20. Cambridge, MA: MIT Press.
Crabtree, Andy, David M. Nichols, Jon O’Brien, Mark Rouncefield, and Michael B. Twi-
dale
2000 Ethnomethodologically informed ethnography and information system de-
sign. Journal of the American Society for Information Science 51(7): 666–
682.
Egbert, Maria
2004 Other-initiated repair and membership categorization – Some conversational
events that trigger linguistic and regional membership categorization. Journal
of Pragmatics 36: 1467–1498.
Fox, Barbara, Makoto Hayashi, and Robert Jasperson
1996 Resources and repair: A cross-linguistic study of syntax and repair. In: Elinor
Ochs, Emanuel A. Schegloff, and Sandra A. Thompson (eds.), Interaction and
Grammar, 185–237. Cambridge, UK: Cambridge University Press.
Garcia, Angela Cora and Jennifer Baker Jacobs
1998 The interactional organization of computer mediated communication in the
college classroom. Qualitative Sociology 21(3): 299–317.
Kötter, Markus
2002 Tandem Learning on the Internet: Learner Interactions in Virtual Online En-
vironments (MOOs). Frankfurt: Peter Lang.
Kurhila, Salla
2001 Correction in talk between native and non-native speakers. Journal of Prag-
matics 33: 1083–1110.
Levinson, Stephen C.
1983 Pragmatics. Cambridge, UK: Cambridge University Press.
Ligorio, M. Beatrice
2001 Integrating communication formats: Synchronous versus asynchronous and
text-based versus visual. Computers & Education 37: 103–125.
Luff, Paul and Christian Heath
2002 Broadcast talk: Initiating calls through a computer-mediated technology. Re-
search on Language and Social Interaction 35(3): 337–366.
Marcoccia, Michel, Hassan Atifi, and Nadia Gauducheau
2008 Text-centered versus multimodal analysis of instant messaging conversation.
Language@Internet 5, article 7. http://www.languageatinternet.org/articles/
2008/1621
Markman, Kris M.
2009 So what shall we talk about: Openings and closings in chat-based virtual meet-
ings. Journal of Business Communication 46(1): 150–170.
Markman, Kris M.
2010 Learning to work virtually: Conversational repair as a resource for norm de-
velopment in computer-mediated team meetings. In: J. Park and E. Abels
(eds.), Interpersonal Relations and Social Patterns in Communication Tech-
nologies: Discourse Norms, Language Structures and Cultural Variables,
220–236. Hershey, PA: IGI Global.
Mason, Marianne
2004 Referential choices and the need for repairs in covertly-taped conversations.
Journal of Pragmatics 36: 1139–1156.
McKellin, William H., Kimary Shahin, Murray Hodgson, Janet Jamieson, and Kathleen Pi-
chora-Fuller
2007 Pragmatics of conversation and communication in noisy settings. Journal of
Pragmatics 39: 2159–2184.
Moerman, Michael
1977 The preference for self-correction in a Tai conversational corpus. Language
53(4): 872–882.
Negretti, Raffaella
1999 Web-based activities and SLA: A conversation analysis research approach.
Language Learning & Technology 3(1): 75–87.
Nilsen, Mona and Åsa Mäkitalo
2010 Towards a conversational culture? How participants establish strategies for co-
ordinating chat postings in the context of in-service training. Discourse Studies
12(1): 90–105.
Panyametheekul, Siriporn and Susan C. Herring
2003 Gender and turn allocation in a Thai chat room. Journal of Computer Mediated
Communication 9(1). http://jcmc.indiana.edu/vol9/issue1/panya_herring.html
Pomerantz, Anita
1984 Agreeing and disagreeing with assessments: Some features of preferred/dis-
preferred turn shapes. In: J. Maxwell Atkinson and John Heritage (eds.), Struc-
tures of Social Action: Studies in Conversation Analysis, 57–101. Cambridge,
UK: Cambridge University Press.
Rieger, Caroline L.
2003 Repetitions as self-repair strategies in English and German conversations.
Journal of Pragmatics 35: 47–69.
Rintel, E. Sean and Jeffery Pittam
1997 Strangers in a strange land: Interaction management on Internet Relay Chat.
Human Communication Research 23: 507–534.
Rohfeld, Rae Wahl and Roger Hiemstra
1995 Moderating discussions in the electronic classroom. In: Zane L. Berge and
Mauri P. Collins (eds.), Computer-Mediated Communication and the Online
Classroom: Volume III: Distance Learning, 91–104. Cresskill, NJ: Hampton
Press.
Sacks, Harvey, Emanuel A. Schegloff, and Gail Jefferson
1974 A simplest systematics for the organization of turn-taking in conversation.
Language 50: 696–735.
Schegloff, Emanuel. A.
1968 Sequencing in conversational openings. American Anthropologist 70:
1075–1095.
Schegloff, Emanuel A.
1987a Between micro and macro: Contexts and other connections. In: Jeffrey C.
Alexander, Bernhard Giesen, Richard Münch, and Neil J. Smelser (eds.), The
Micro-Macro Link, 207–234. Berkeley, CA: University of California Press.
Schegloff, Emanuel A.
1987b Recycled turn beginnings: A precise repair mechanism in conversation’s turn-
taking organisation. In: Graham Button and John R. E. Lee (eds.), Talk and So-
cial Organisation, 70–85. Clevedon, UK: Multilingual Matters.
Schegloff, Emanuel A. and Harvey Sacks
1973 Opening up closings. Semiotica 7: 289–327.
Schegloff, Emanuel A., Gail Jefferson, and Harvey Sacks
1977 The preference for self-correction in the organization of repair in conversation.
Language 53: 361–382.
Schönfeldt, Juliane and Andrea Golato
2003 Repair in chats: A conversation analytic approach. Research on Language and
Social Interaction 36(3): 241–284.
Sidnell, Jack
2007 Repairing person reference in a small Caribbean community. In: Nick J. En-
field and Tanya Stivers (eds.), Person Reference in Interaction: Linguistic,
Cultural and Social Perspectives, 281–309. Cambridge, UK: Cambridge Uni-
versity Press.
Simpson, James
2005 Conversational floors in synchronous text-based CMC discourse. Discourse
Studies 7(3): 337–361.
Suchman, Lucy A.
1987 Plans and Situated Actions: The Problem of Human Machine Communication.
Cambridge, UK and New York: Cambridge University Press.
Svennevig, Jan
2008 Trying the easiest solution first in other-initiation of repair. Journal of Prag-
matics 40: 333–348.
Tan, Seng-Chee and Aik-Ling Tan
2006 Conversational analysis as an analytical tool for face-to-face and online con-
versations. Educational Media International 43(4): 347–361.
Todman, John and Norman Alm
2003 Modeling conversational pragmatics in communication aids. Journal of Prag-
matics 35: 523–538.
Veerman, Arja L., Jerry E. B. Andriessen, and Gellof Kanselaar
2000 Learning through synchronous electronic discussion. Computers & Education
34(3–4): 269–290.
Werry, Christopher C.
1996 Linguistic and interactional features of Internet Relay Chat. In: Susan C. Her-
ring (ed.), Computer-Mediated Communication: Linguistic, Social, and Cross-
Cultural Perspectives, 47–64. Amsterdam/Philadelphia: John Benjamins.
Zemel, Alan and Murat Cakir
2009 Reading’s work in VMT. In: Gerry Stahl (ed.), Studying Virtual Math Teams,
261–276. New York: Springer Publishing.
1. Introduction
Email has become the central communication tool for most office workers. It is the
medium in which people carry out their daily professional activities and in which
workers and business partners build and maintain professional and interpersonal
relations. In this chapter, we examine the implicit norms of responding in work-
place emails. More specifically, we investigate how employees in a distributed
work group produce email responses, with reference to what is known about giving
response in oral conversations. We investigate to which communicative acts par-
ticipants respond, to which communicative acts they may omit to respond, and to
what extent they produce acknowledging receipts.
The research reported here was inspired by several pragmatic theories. We
draw on speech act theory in order to identify the communicative acts that are re-
sponded to vs. those acts that are not responded to. Furthermore, the research was
conducted within the framework of Conversation Analysis (CA), one of the main
objectives of which is to identify how participants in face-to-face conversation ac-
complish social order by orienting to norms. This research is based on the assump-
tion that users of email, as in oral conversation, orient to some basic interactional
norms.
It has previously been shown that participants trade on mundane expectancies
when communicating via new technologies (Garcia and Jacobs 1998, 1999; Orli-
kowski and Yates 1994). Likewise, in the present study, it is expected that conver-
sational norms constitute a major source for the establishment of new norms in
computer-mediated communication (CMC), a medium that is still in the process of
being shaped.
Although email is a written medium, it has certain features that make it comparable
to oral conversation.1 First, email is interactive. Its speed and ease of delivery en-
able the exchange of messages with minimal time lapse. Second, email messages
may involve extended interchanges, or threads of messages, organized in turns and
sequences. However, an important difference between email and face-to-face con-
versation is that the interlocutors do not share a common temporal and physical
context; that is, while face-to-face conversation is synchronous by nature, email in-
In the quote above, Condon and Čech suggest that an absent response signals
agreement. A similar point is made by Murray (1991), who interprets a delayed re-
sponse in a decision routine as an implicit promise. A promise to comply with a re-
quest is often “unmarked” (Murray 1991: 110), i.e., it occurs without linguistic sig-
nals. “Often the time delay between the original request and the response is hours
or even days. The requestor does not consider this rude or unusual, but interprets
the silence as a promise to try to fulfill the request” (Murray 1991: 110). Interesting
observations have also been made by Tyler and Tang (2003), who interviewed em-
ployees about their perceptions of how they responded to emails and how they
formed expectations of their co-participants’ responses to them. The researchers
found that email rhythms were based much more on relationships than isolated
messages. Furthermore, participants accommodated to the email behaviors of
others and mostly favored quick responses, for instance by sending short messages
which stated when they intended to provide a full reply.
With regard to how participants in CMC deal with non-responses, Rintel et al.
(2003) investigated ambiguous non-responses on Internet Relay Chat (IRC). The
researchers found that when dealing with non-responses, participants tried the ea-
sier solution first. That is, they reconnected and re-greeted before assuming that
there was an interpersonal problem.
Condon and Čech (1996), Baym (1996), Murray (1991), Tyler and Tang (2003)
and Rintel et al. (2003) seem to agree that absent or delayed responses can have
different meanings in CMC than in oral conversation. Inspired by their research, in
this study we ask whether a missing response to a request in an institutional setting
corresponds to an implicit response norm. If so, does the absent response signal
agreement? Do participants’ tacit expectations about institutional routines allow
them not to respond or give acknowledging receipts in certain cases?
There seems to be agreement in some of the previous CMC studies that delayed
or absent responses to emails are in some circumstances in accordance with par-
ticipants’ norms of behavior. Some authors relate this to genre, others to routines.
In interaction studies it has been shown that different types of activities and set-
tings make relevant different norms and speech exchange systems (Arminen 2005;
Drew and Heritage 1992). We expected this to be the case in CMC as well. It is not
necessarily the mediating channel that actualizes the norms of talk. Rather, it is the
different activities and speaker roles involved.
However, none of the studies reviewed above examined how participants ac-
tually respond to email messages or show how missing responses are treated in a
distributed work group. There is a need for systematic analysis of email interac-
tions with the aim of explicitly accounting for their interactional norms. In the
study described below we attempt to do this, and in order to take account of the in-
teractional context we distinguish between responses that occur after questions and
requests, on the one hand, and responses that occur after communicative acts that
do not explicitly request a response, on the other. Our research questions can be
stated as follows:
1. What are the implicit norms of responding to email questions and requests?
a) How and when do participants respond to email requests?
b) How do participants attend to absent or delayed responses?
2. What are the implicit norms of responding to non-requesting emails?
a) What kinds of non-requesting messages are responded to?
b) What kinds of non-requesting messages are not responded to?
4.1. Data
The data analyzed in this study are emails from members of a distributed project
group called “Agenda” in a Norwegian telecommunications company, “Telecom”
(henceforth, TCM).2 The email exchanges were gathered from August 2004 to De-
cember 2004 and total 491 emails. The messages were collected with assistance
from the Agenda members, who during the period continuously forwarded their
exchanges to the first author.3
Agenda’s main task was to edit all confirmation letters that were distributed to
their customers. The members formulated, revised, and standardized letters and
templates, with the purpose of giving all letters a uniform style that communicated
in a clear and polite manner to the customers. The group was frequently contacted
by various divisions in the company with requests to reformulate their letters. In
order to give the letters a unified tone of voice, group members followed a set of
guidelines concerning orthography, sentence structure, and style.
The project group consisted of 12 members, each representing a team from the
different divisions of the company: Telephony: Birgitte Hansen, ADSL: Siri Vogt,
Dial-up: Elisabeth Eide, Mobile: Jon Olavsen, Customer Service: Per Olsen,
Credit: Lise Larsen, and Invoice: Janne Eia. In addition, Agenda includes text de-
signer Maria Monsen, system operator Geir Johnsen, and an observer and system
operator from another division, Tore Strand. The manager of the project group was
Line Myhre, who in turn reported to Arvid Lervik, the Information Director of
TCM.
The senders and recipients of the messages were primarily Agenda members,
most of whom were physically located in different departments and even different
parts of Norway. They had occasional meetings at the main office, but most of their
communication was carried out by email or telephone. The members usually inter-
acted in dyads, sometimes with selected co-members copied in, and seldom ad-
dressed the group as a whole. Some members were more active than others – for in-
stance, text designer Maria Monsen, manager Line Myhre, and the system
operators Geir Johnsen and Tore Strand – yet all of the members are represented in
the data. The members also interacted with external collaborators within and out-
side TCM, such as employees at the print center “Strålfors” and members of
Agenda Business. In addition to the email collection, extensive ethnographic and
interview data were collected in order to contextualize group members’ communi-
cative practices. The emails were written originally in Norwegian and were trans-
lated into English by the authors. The translations are idiomatic, rather than word-
by-word, in order to keep the style as close to the Norwegian original as possible.
4.2. Method
The study applies the methodology of Conversation Analysis (CA), which entails
that the object of study is the participants’ own procedures and norms for interac-
tion and that the approach used to grasp them is sequential analysis, with claims
about individual utterances based on the interpretations made by the co-partici-
pants as displayed in their subsequent responses. In order to identify communi-
cative acts performed in the email messages, the study also applies terms from
speech act analysis (Searle 1969).4
Studying response-giving involves distinguishing between turns (messages)
that make a response conditionally relevant and those that do not. Certain types of
utterances create the expectation of a specific type of response and thus constitute a
first pair part of an adjacency pair. Sacks and Schegloff’s notions of “conditional
relevance” and “accountability” provide a useful resource in order to understand
this phenomenon:
When one utterance (A) is conditionally relevant on another (S), then the occurrence of
S provides for the relevance of the occurrence of A. If A occurs, it occurs (i.e. is pro-
duced and heard) as a “responsive to” S, i.e. in a serial or sequenced relation to it; and if
it does not occur, its occurrence is an event, i.e. it is not only non-occurring (…), it is ab-
sent, or “officially” or “notably” absent. That it is an event can be seen not only from its
“noticeability”, but from its use as legitimate and recognizable grounds for a set of in-
ferences (e.g. about the participant who failed to produce it). (Schegloff 1972: 76)
In CA, a conversational norm may become visible in two ways: first, when a con-
versational feature appears as a structural regularity and second, when a norm is
breached in so-called “deviant cases”:
Only in cases of breach or anticipated breach may the reflexive features of conduct
come to be fleetingly entertained, for on the whole it is in cases of breach that actors may
anticipate being held to account for their actions and, as part of this anticipation, con-
template its being held that they could have known and done differently and therefore
should have done so. (Heritage 1984: 118)
In the following sections, we aim to uncover what sort of email messages provide
for the relevance of a responding message and what sort of response are expected.
To do this, we follow the method of searching for structural regularities as well as
deviant cases, mainly in the form of accounts and prompts oriented to “noticeably
absent” responses.
In the data, messages that elicit a response are primarily directives such as
questions and requests (Searle 1969). Requests for information (e.g., questions)
explicitly ask for a verbal response. Requests for action (such as task assignments)
primarily elicit a non-verbal response, but they also create the expectation of a ver-
bal display of acceptance or rejection (Lindström 1999). Requests for information
and action may take several forms, such as the following:
As Schegloff (1984) and Clayman and Heritage (2002) have pointed out, there is
no one-to-one correspondence between an utterance’s grammatical form and the
speech act it accomplishes. Declaratives can function as questions, and inter-
rogatives can accomplish non-questioning actions. Whether an utterance consti-
tutes a question or not depends on the sequential context.
An additional feature of the context that is of special relevance here is the in-
stitutional activities and identities involved. Certain routinized patterns of action
can influence the interpretation of individual utterances (Condon and Čech 1996,
cited above). The fact that the manager routinely assigns tasks to the project group
members will make her messages more prone to being understood as requests than
messages from other members. For instance, a task may be assigned by the mere
mentioning of the type of work involved, as in the following email message: “Text-
work. Deadline is 17 sept”. This implicit formulation of a request presupposes a
common understanding of the institutional context, including different communi-
cative rights and obligations related to the group members’ respective professional
roles. This implies that we need to look beyond the linguistic form and the sequen-
tial context to see whether there are organizational structures and procedures that
contribute to determining the function of a certain type of utterance.
In this study, every email in the data corpus was transferred into an Excel docu-
ment, given an identifying number, and categorized as a requesting or a non-re-
questing message. The requesting messages were divided into requests for action
and requests for information. Further, each email was coded with respect to the oc-
currence vs. non-occurrence of responses, response latencies, reminders, accounts,
and receipts. Then, a qualitative analysis of the response messages sought to ident-
ify a response pattern with respect to the normative expectations of the interac-
tants, depending on the form of the message, its sequential position, the institu-
tional activity involved, and the participants’ institutional roles.
The quantitative analysis resulted in two major findings. First, almost half – 47 %
(230) – of the total number of messages did not receive a response. Even more sur-
prisingly, 34 % (101) of the requesting messages did not receive a response, as
compared to 65 % (129) of the non-requesting messages that did not receive re-
sponse. Second, the average response latency was found to be 28.01 hours. The
quantitative response patterns are presented in Table 1.
The qualitative analysis produced two general findings. First, the most common
types of exchanges were question/answer sequences and requests-for-action/com-
pliance sequences, or even request-for-action followed by no response or a place-
holder. Second, absent responses to questions and requests were followed by re-
minders from the senders or accounts by the recipients, suggesting that absent
responses are treated as noticeably absent. In what follows, we present the norms we
have identified in more detail and illustrate our findings with examples from our data.
(“What’s the status here?”) seems to make relevant an account for not accomplish-
ing the task. Additionally, the offer of sitting down with MM can be interpreted as a
way of dealing with a problem. It implies that the task has not been completed and
that the recipient may be having trouble with it. In short, by sending MM a rem-
inder, BH treats MM’s absent response as reportable. BH holds MM accountable
for the missing reply, and the appended mail functions as evidence for the missing
response. Twenty minutes later, MM responds with the following message: “I have
moved my Tuesday-Agenda-day to tomorrow […] so I am going to look at this to-
morrow. […]. I hope that’s all right – because I’m soooo stressed this week, […]. I
had a sick child Thursday last week, and when he recovered Friday I discovered
that he had got lice – and then I spent the day eradicating them. That’s why I
haven’t looked at this yet. Best regards, Maria”.
In her reply, MM produces a narrative account that explains why she has not
complied with BH’s request. This lengthy account displays an orientation to the
normative expectation of complying with requests without delay or alternatively,
of sending a placeholder response.
To sum up, email resembles conversation in that requests provide for the con-
ditional relevance of a response in the form of a granting or rejection. In the case of
questions, there seems to be a tolerance for delay ranging up to some days. In re-
quests for action, the tolerance for delay seems greater, especially if the requested
action is extensive. Support for this is suggested by the quantitative results, which
show that 18 % (17) of the requests for information lacked response, compared
with 42 % (84) of the requests for action. However, a practice is available for pre-
empting negative inferences from missing responses, namely placeholder mess-
ages signaling compliance and committing to a future granting of the request. This
seems to be a practice that has developed to deal with expectations based on the
meaning of silence that are transferred from oral conversation to the email
medium. More important, the strong orientation to a norm of complying with re-
quests in a workplace setting, as in this study, must be seen in light of the partici-
pants’ institutional roles. MM’s role as text designer allows her colleagues to make
requests of her, on the one hand, and on the other hand it obliges her to comply with
their requests.
In one message, the system administrator (TS) informs the Agenda group that a
photo of group members (which was taken during the previous meeting) has been
published on their homepage: “Hi everybody, now the picture is published on the
Agenda page. I hope I chose the right picture out of the 3 I received from Line.
Should I make any changes? Tore”.
In this example, TS performs three separate speech acts. First, he informs the
group that the photo has been published; furthermore, he indirectly requests LM to
confirm that he has published the right photo. Finally, he interrogatively requests
the Agenda members to report if he should make changes. Interestingly, no one re-
sponds to this request. The fact that the message includes multiple speech acts
makes it possible for the recipients to ignore the request and to merely attend to the
informing part. The request invites suggestions for changes and provides a possi-
bility of not responding if one does not have a suggestion to make. In such cases,
then, non-response seems to be acceptable after a request and carries the meaning
that one has nothing to report.
In the next example, TS responds to a message from MM in which she sent him
an attached file and (indirectly) requested him to publish it on the Agenda home-
page: “For our pages on the intranet. :) Maria”. One hour later, TS responds to MM,
adding the Agenda leader LM to the address field: “Is this okay on [reference to
website]? Best regards, Tore”.
TS’s message above primarily informs MM and LM that the new letter tem-
plate has been published. However, his utterance is interrogatively designed as a
request for approval by MM and LM. As in his previous message, he gets no re-
sponse, and he seems to treat the silence as approval. The participants appear to ad-
here to a norm which allows them to remain silent if they approve a proposal and to
respond only if they disapprove or have corrections to make.
Our main explanation for this pattern of non-response is the multifunctionality
of the initiating message – the fact that it informs and invites comments simulta-
neously. However, this also seems to be a pattern that is more common in emails
sent to multiple recipients than in messages to individuals. Addressing a larger
group seems to contribute to making it less relevant for the recipients to reply, un-
less one has a substantial contribution to make.
reply message is not always only a response to a previous message; rather, the
sender may use the reply function to initiate a new and unrelated message. TS
posted a purely informative message to his leader LM: “Copy to you Line”. Four
days later, LM replied with a copy to the Information Manager: “Thanks for your
info, Tore. Nice to know how AGB is organized. The pictures of AGP are coming,
I’m going to send them over this evening. rgds Line”.
Here LM performs two actions. She thanks TS for the information he provided,
and she informs (or promises) him that she is going to send over some pictures of
Agenda, which he, as system administrator, is responsible for publishing on Agen-
da’s homepage. These actions are not internally coherent but correspond rather to
two unrelated institutional activities. The message was sent four days after the in-
itial message, using the reply function in the mail program. As a result, the “sub-
ject” heading was inherited from the previous exchange and was not relevant to the
issue of informing about the status of the picture. The thanking may thus be a way
of linking back to the heading and to the participants’ previous history of interac-
tion rather than being a relevant action in itself. Although informing messages do
not require a response, they may be treated as actions that make a specific type of
response relevant. By thanking TS, LM treats the prior message as an offer of a ser-
vice and thus makes it response-worthy.
Non-requesting messages are also responded to by appraisals and supportive
statements. Typical instances are supportive responses from leaders. In the next
example, TS replies to a message from one of the employees in Customer Services
concerning a customer complaint. The request for help was originally addressed to
LM, but since she was not available on the day in question, it was forwarded to
Agenda’s mailbox and handled by TS. TS copied in LM in his reply: “Hi! We in
Telecom have no control over how another sales department has treated this trans-
action! Unfortunately it will most likely be difficult to have full control of how
other sales departments treat transactions. We in Telecom have clear routines for
how the letters from the customer shall be treated. See presentation”. Three days
later, LM responded to TS: “Great support, Tore, you are very organized and a
great reply to Erik! rgds Line”. In contrast to her previous message to TS above
(“thanks for the info”), LM does not elaborate on TS’s message or introduce a new
activity. Rather, it is a purely responsive message providing positive evaluation
and acknowledgement. This is also a volunteered response, occasioned but not
made conditionally relevant by the previous message.
In the data it is typically superiors who give acknowledging and supportive re-
ceipts to non-requesting messages. In one message, LM informs AL of the halt in
progression in Agenda: “Hi Arvid! Refer to mail below concerning new participant
from Dial-up. In addition I want to inform you that I was a little bit strict on the prior
Agenda meeting in cases where I don’t think the work has made any progress. The
templates were not finished – I’ll contact you for a chat. rgds Line”. Twenty-one
minutes later, AL responds: “Ok. Progress is especially important here! Arvid”.
AL first acknowledges receipt of LM’s message (“Ok”) and then goes on to sig-
nal support of LM’s line of action. The use of an intensifier (“especially”) and an
exclamation mark reinforces his display of agreement. The primary role of this re-
sponse is to acknowledge the information and align with LM’s position. This type
of evaluating response from managers enacts the institutional role of a person with
the right and obligation to evaluate the performance of his or her employees.
Participants also respond to non-requesting messages to check their under-
standing of or make corrections to the previous mail. Text designer MM receives a
copy of LM’s mail to EE: “Welcome to AgendaPrivate, Elisabeth. I’ll invite you to
a new meeting with only the two of us, to brief you about your work :-) rgds Lise”.
Fifteen minutes later MM responds: “Since you’re sending me a copy – is Elisa-
beth going to work with text? I have heard that she works with Internet today. Hope
you have had a nice week-end! :) Maria”. In the original message, MM is a secon-
dary recipient (Skovholt and Svennevig 2006) and is not addressed explicitly in the
address heading. In her response to LM, MM shows that the very fact that she has
been copied in gives rise to inferences about relevance, in this case the hypothesis
that EE will be working with text, like herself. MM’s question serves to initiate re-
pair, i.e., to check her understanding of LM’s message. (On the topic of repair, see
also the chapter by Jacobs and Garcia, this volume.)
In sum, we found that participants hold each other accountable for non-re-
sponses to questions and requests. However, for a certain class of messages that are
primarily informative and only secondarily requestive, a response is only condi-
tionally relevant if the recipients have corrections or comments to make. Also, we
found that responses to non-requesting messages are not conditionally relevant but
are occasionally volunteered as more interpersonally oriented responses, e.g., to
provide positive evaluations, appraisal, and support.
tions – for example, how sender and receiver collaboratively construe and identify
the trouble source in reminders and accounts. Which trouble sources do they ad-
dress and what kinds of accounts are provided?
The reminders sometimes provide a potential explanation for the problem and
present it as something beyond the recipient’s control (MM to LH: “I hope you
enjoy the summer, and – if you are not on holiday – that you as quickly as possible
can conjure up a reply to the mail below. I guess that it has drowned in your holiday
mail”). The senders produce reminders with hedges that make the request appear
indirect and unobtrusive (GJ to AL: “I wonder if you could find the signatures”).
The reminders may be based on the assumption that the recipient has already pro-
vided the requested information or carried out the work (JE to BS: “Was there a
solution to this?”). Interestingly, the reminders may also construe the problem as
rooted in the sender’s own insufficiency (AF to GO: “Now I’m really nagging, but
[…]”; LM to KB: “Did you get assistance from Tor on this case? Greetings from
she who collects old mails”).
Accounts from the recipients generally present the problem as time manage-
ment (MM to LM: “Sorry for this late answer – it is rather hectic nowadays”. GO to
AF: “I will do the check Friday this week, other issues have turned up which made
me put this on ice for a while […]”. KB to LM: “No, I had forgotten that […]”).
The accounts may also be presented as simple apologies (LH to MM: “Sorry, I will
come back to you by next Monday”). There are eight instances of accounts in total
in the corpus.
By treating the trouble sources as systemically occasioned (such as by vacation
or a hectic work situation), the participants display that an absent or delayed re-
sponse is not the result of a relational problem. This observation may suggest that
missing responses have other functions in email interaction than they do in face-
to-face interaction. It also supports Pomerantz’s (1984) and Rintel et al.’s (2003)
finding: Participants try the easier solution first, for example by looking for insti-
tutional or technical reasons before they assume that there is an interpersonal prob-
lem.
7. Conclusion
The study presented in this chapter shows that email resembles conversation as
regards response and non-response, in that there is a norm that requests and ques-
tions should in general be followed by responses. However, there are situations in
email where a response to a question or request is not necessarily conditionally rel-
evant, in contrast to conversational norms. This indicates that the participants have
developed alternative norms for email interaction. Our analysis identified three
general interactional norms:
The first norm is that questions and requests should normatively receive a re-
sponse. Absent responses are likely to be followed by reminders from the re-
questers or by accounts and excuses from the recipients. In the case of questions,
there seems to be a tolerance for delay ranging up to several days. In requests for
action, the tolerance for delay appears to be greater, especially if the requested ac-
tion is elaborate. In cases where a request cannot be granted immediately, the par-
ticipants either wait until the task can be carried out to respond, or they send a
“placeholder message”, i.e., a message that signals compliance and commits to a
future granting of the request. The practice of sending placeholder messages is
oriented towards preempting negative inferences from absent responses. This prac-
tice seems to have developed to deal with the potential negative implications of si-
lence derived from conversational interaction. These findings are consistent with
Tyler and Tang’s (2003) interviews, which revealed that participants often sent
short responses to state when they intended to provide a fuller response.
The second norm comprises a class of exceptions to the norm of responding to
requests. When email messages contain invitations to correct or comment on pro-
posals, they perform two actions simultaneously: They present information to the
addressee, and they simultaneously request comments or corrections to the propo-
sal. In these cases, a response is conditionally relevant only if the recipients disap-
prove, disagree, or have corrections to make. This norm may explain why a sur-
prising 34 % of the requesting messages did not receive a response. Not responding
to requests violates traditional norms of adjacency pairs: A first-pair part generally
should be responded to by a second-pair part. However, in workplace meetings or
classroom interaction, silence means that one has nothing to report, and it is not
treated as problematic. Rather, silence conventionally may be understood as con-
veying acceptance and agreement, consistent with the findings of Murray (1991)
and Condon and Čech (1996).
The third norm we identified shows that non-requesting email messages, e.g.,
attachments, FYI messages, and in-copied messages, do not require a response.
Neither do first-pair parts that do not request information or action beyond a
simple acknowledgement, e.g., congratulations and compliments. The lack of re-
sponse in these instances was not treated as problematic or deviant by participants
in this study. This pattern is inconsistent with traditional adjacency pair norms
in oral conversation, where congratulations and compliments typically receive
thanks.
Acknowledgements
We would like to thank Susan Herring and Arja Piirainen-Marsh for their helpful
comments.
Notes
1. For more on the hybrid nature of CMC language, see Skovholt and Svennevig (2006) and
Skovholt (2009).
2. All names of companies and employees have been replaced by pseudonyms.
3. Depending on the employees to forward their emails is a possible error source for a study
of responses, in that certain short acknowledging messages may have been deemed un-
important and thus not forwarded, or they may have been forgotten. In cases where we
had doubts whether all messages in a thread had been forwarded, we contacted certain
members with a request to check their inboxes.
4. For more on the application of CA and speech act analysis, see Skovholt (2009).
References
Arminen, Ilkka
2005 Institutional Interaction: Studies of Talk at Work. Aldershot, UK: Ashgate.
Baym, Nancy K.
1996 Agreement and disagreement in a computer-mediated discussion. Research on
Language and Social Interaction 29(4): 325–345.
Clayman, Stephen and John Heritage
2002 The News Interview: Journalists and Public Figures on the Air. Cambridge,
UK: Cambridge University Press.
Condon, Sherri L. and Claude G. Čech
1996 Discourse management strategies in face-to-face and computer-mediated deci-
sion making interactions. Electronic Journal of Communication/La Revue
électronique de communication 6(3).
http://www.cios.org/EJCPUBLIC/006/3/006314.HTML
Drew, Paul and John Heritage (eds.)
1992 Talk at Work: Interaction in Institutional Settings. Cambridge, UK: Cambridge
University Press.
Garcia, Angela C. and Jennifer B. Jacobs
1998 The interactional organization of computer mediated communication in the
college classroom. Qualitative Sociology 21: 299–317.
Garcia, Angela C. and Jennifer B. Jacobs
1999 The eyes of the beholder: Understanding the turn-taking system in quasi-syn-
chronous computer-mediated communication. Research on Language & So-
cial Interaction 32: 337–367.
Grønning, Annette
2006 Personen bag. Tilstedevær i email som interaktionsform mellem kunde og
medarbejder i dansk forsikringskontekst. [The person behind. Social presence
in email interaction between customer and employee in a Danish insurance
context.] Ph.D. dissertation, Copenhagen Business School.
Heritage, John
1984 Garfinkel and Ethnomethodology. Oxford, UK: Polity Press.
Herring, Susan C.
1999 Interactional coherence in CMC. Journal of Computer-Mediated Communi-
cation 4(4). http://jcmc.indiana.edu/vol4/issue4/herring.html
Herring, Susan C.
2010 Computer-mediated conversation: Introduction and overview. Language@
Internet 7, article 2. http://www.languageatinternet.org/articles/2010/2801/
index_html/
Jones, Quentin, Gilad Ravid, and Sheizaf Rafaeli
2004 Information overload and the message dynamics of online interaction spaces:
A theoretical model and empirical exploration. Information Systems Research
15(2): 194–210.
Kalman, Yorman M., Gilad Ravid, Daphne R. Raban, and Sheizaf Rafaeli
2006 Pauses and response latencies: A chronemic analysis of asynchronous CMC.
Journal of Computer-Mediated Communication 12(1), article 1. http://jcmc.
indiana.edu/vol12/issue1/kalman.html
Kankaanranta, Anne
2005a English as a corporate language: Company-internal e-mail messages written
by Finns and Swedes. In: Britt L. Gunnarsson (ed.), Communication in the
Workplace, 42–59. Uppsala: University of Uppsala.
Kankaanranta, Anne
2005b “Hej Seppo, could you pls comment on this!” Internal email communication in
lingua franca English in a multinational company. Ph.D. dissertation, Centre
for Applied Language Studies, University of Jyväskylä.
Laursen, Ditte
2005 Please reply! The replying norm in adolescent SMS communication. In: Ri-
chard Harper, Leysia Ann Palen, and Alex S. Taylor (eds.), The Inside Text:
Social, Cultural and Design Perspectives on SMS, 53–73. Dordrecht, The
Netherlands: Springer.
Lindström, Anna
1999 Language as Social Action. Grammar, Prosody, and Interaction in Swedish
Conversation. Uppsala University: Department of Nordic Languages.
Mattila, Anna S. and Danliel J. Mount
2003 The impact of selected customer characteristics and response time on e-com-
plaint satisfaction and return intent. International Journal of Hospitality Man-
agement 22(2): 135–145.
Murray, Denise E.
1991 Conversation for Action: The Computer Terminal as Medium of Communi-
cation. Amsterdam/Philadelphia: John Benjamins.
Orlikowski, Wanda J. and JoAnne Yates
1994 Genre repertoire: Examining the structuring of communicative practices in or-
ganizations. Administrative Science Quarterly 39(4): 541–574.
Skovholt, Karianne
2009 Email literacy in the workplace. A study of interaction norms, leadership com-
munication, and social networks in a Norwegian distributed work group. Ph.D.
dissertation, Faculty of Education, University of Oslo.
Skovholt, Karianne and Jan Svennevig
2006 E-mail copies in workplace interaction. Journal of Computer-Mediated Com-
munication 12(1), article 3. http://jcmc.indiana.edu/vol12/issue1/skovholt.html
Strauss, Judy and Donna J. Hill
2001 Consumer complaints by e-mail: An exploratory investigation of corporate re-
sponses and customers reactions. Journal of Interactive Marketing 15(1):
63–73.
Tyler, Joshua R. and John C. Tang
2003 When can I expect an e-mail response? A study of rhythms in e-mail usage. In:
Proceedings of the ECSCW 2003 European Conference on Computer Sup-
ported Cooperative Work, 239–258. New York: Springer-Verlag.
1. Introduction
Email has been characterized as a lean medium (Daft and Lengel 1984), meaning
that it does not allow immediate feedback or the use of subtle signals for interpre-
tation normally conveyed non-verbally in spoken communication via tone of voice
and facial expression. The lack of such interpretive cues has been seen as restrict-
ing the kinds of communication that can take place via email (for overviews, see
Abdullah 2006; Bunz and Campbell 2004; Panteli and Seely 2006; Waldvogel
2007). Important information, it is claimed, is lost regarding what is said and how
it should be perceived, which in turn increases the risk of misunderstanding and
conflict. The use of so-called “flaming” – passages where people become enraged
for little apparent reason and exchange incensed comments – has been observed in
various studies (for an overview, see Turnage 2007; for a case study of flaming, see
Danet, this volume) and taken as evidence that the medium is not especially well-
suited to sensitive subjects of conversation.
This view has been questioned to an increasing extent, however. A number of
studies have attempted to explain how a “lean” medium such as email can nonethe-
less produce rich information (Panteli 2002; Waldvogel 2001). Researchers have
demonstrated how people in using computer-mediated communication (CMC)
tend to develop informal codes, referred to as “emotext” by Persichitte, Young, and
Tharp (1997: 280), in order to compensate for the lack of non-verbal cues and
make typed CMC more expressive and/or speech-like (Baldauf and Stair 2008).
These informal codes include grammatical markers, strategic capitalization, lexi-
cal surrogates, and intentional misspelling (e.g., thaaaaaank youuuuuu), as well as
emoticons, which are claimed to substitute for facial expression (Bieswanger, this
volume; Derks, Bos, and Grumbkow 2008; Derks, Fisher, and Bos 2008; but cf.
Dresner and Herring 2010).
Email can also be socially rich, including the use of politeness. Several studies
have shown how individuals in email use apologies, indirectness, inclusive forms
(“we”, “you”), and greetings and closings in order to create a good work climate
(Job-Sluder and Barab 2004; Kankaanranta 2005; Kaul and Kulkarni 2005; Mor-
and and Ocker 2003; Waldvogel 2007). It has also been discussed how this kind of
behavior in email relates to spoken interaction (Boneva, Kraut, and Frohlich 2001;
Zack 1993) and other forms of mediated communication (Watts 2007); how the in-
teraction is affected by various pragmatic parameters: content of the message (Bie-
senbach-Lucas 2005), the stylistic level used by the writers (Bunz and Campbell
2004), and the relative status of the interlocutors (Biesenbach-Lucas 2007, 2005;
Chen 2006, 2001; Rogers and Lee-Wong 2003; Sherblom 1988; Skovholt and
Svennevig 2006); and how politeness work differs between individuals communi-
cating in their native language and in a second language (Biesenbach-Lucas 2005,
2007; Chen 2001, 2006). Studies of politeness in email have also drawn attention
to different first language/cultural backgrounds of the interlocutors (Biesenbach-
Lucas and Weasenforth 2000; Murphy and Levy 2006). Other studies have targeted
reactions to messages that are perceived as more polite or less polite (Duthler
2006; Jessmer and Anderson 2001). Still others discuss gender differences in how
politeness is performed (Boneva, Kraut, and Frohlich 2001; Herring 1994; Oliveira
2007), with “email” broadly defined to include postings to asynchronous message
forums as well as one-to-one communication. While impressive, this body of re-
search on politeness in email communication has tended to leave the actual situ-
ation out of substantial consideration by not examining email interaction in close
relation to the external contexts in which it occurs.
Studies based on more situationally grounded analysis do exist, however. For
example, Skovholt (2009) examined how leadership was linguistically accom-
plished in a Norwegian distributed work group in the leader’s day-to-day email in-
teraction. The study showed how the leader invited group identification by giving
the group positively loaded nicknames, using different kind of narrating devices,
and providing receipts and supportive acknowledgments. Waldvogel (2007)
studied a manufacturing plant and found that use of greetings and closings in email
messages reflected and constructed open and positive relationships between staff
and management and a direct, friendly, and familial workplace culture. In these
studies, rhetorical and politeness strategies were found to play an important role in
constructing workplace familarity. These strategies were not explicitly related to
the concept of small talk, however. Moreover, although relational content in CMC
has been studied in relation to issues of media choice (e.g, Johnson et al. 2008; Utz
2007) and self-disclosure (e.g, Joinson 2001; McKenna, Green, and Gleason 2002)
in the context of maintaining and creating friendships, such research has not made
an explicit link to small talk. Such a link is needed, in that small talk can facilitate
understanding of how a positive climate is created under varying contextual con-
ditions. Small talk is used here with linguistic politeness theory as an interpretative
platform, with the aim to deepen such an understanding.
sional identities, and the purpose and type of medium in which the interaction is
carried out (Danet and Herring 2003; Herring 2007). Politeness is described as “re-
lational work” (Holmes and Schnurr 2005) – a relational practice (Fletcher 1999)
that creates a special logical, interactive, context-bound efficiency that finds its
meaning in the social commonality in which the interaction takes place (for over-
views, see Bargiela-Chiappini and Harris 2006; Eelen 2001; Holmes and Schnurr
2005). What needs to be further specified, though, is how such commonalities are
related to the use of various media – an important reality in everyday work life in
modern societies – and how this usage is connected to the flow of interaction of
which email messages form a part. This is addressed in this chapter. First, however,
the following section explains the relationship between small talk and the concept
of politeness.
struction formed by the turns the correspondents produce in the ongoing exchange
of messages, where the meaning of various language acts (including general FBAs
and mitigating FBAs) is interpreted based on the definitions that the involved par-
ties ascribe to them through their behavior. This behavior is seen as something so-
cially shared in relation to the context the group represents and the norms devel-
oped there for what is acceptable communicative behavior, rather than simply as
the product of individual interlocutors (Mills 2002: 69). On this basis I use the con-
cept of “community of practice”, defined as “an aggregate of people who come to-
gether around mutual engagement in an endeavour” (Eckert and McConell-Ginet
1992: 464). Politeness is seen “as a set of practices or strategies which commu-
nities of practice develop, affirm, and contest” (Mills 2003: 9), and such commu-
nities have their own norms for what constitutes politeness (Mullany 2006).
I distinguish two basic conditions for how interaction within communities of
practice is connected to the external context. On the one hand, there are physically-
based communities of practice, where the interlocutors share an external physical
setting. The members may be co-located in the same building and may actively
choose among various communication alternatives, such as face-to-face interac-
tion, email, and SMS. Norms (including politeness norms) will potentially be elab-
orated for use of a certain communication alternative as opposed to another, based
on the socially-shared meanings ascribed to the two alternatives (cf. Kim, Kim,
Park, and Rice 2007; Murray 1988). On the other hand, there are digitally-based
communities of practice, where the interlocutors are geographically remote. That
which in a physically-based community of practice may be communicated through
various media will in a digitally-based community of practice typically (although
not necessarily) be realized through one medium: email, a chat room, etc. Norms
(including politeness norms) will develop through dealing with various content
within that single medium, and certain strategies will be used in order to further
common undertakings.
Using these concepts (physically-based community of practice and digitally-
based community of practice) entails two consequences for the analysis. First, the
focus is on understanding the use of politeness in each analyzed setting in relation
to the external conditions surrounding the use of email. In providing contextual in-
formation, I distinguish not only what kinds of content, genres, and exchange pat-
terns are used in messages from different interlocutors but, more fundamentally,
how these patterns relate to the occurrence of communication face-to-face and via
other media during the ongoing email interaction. Second, the study concentrates
on broad structures in the groups studied, rather than on individual differences be-
tween particular correspondents. The overarching issues of interest are to what ex-
tent small talk is employed, where this kind of discourse appears in the messages,
and how it relates to complications that risk threatening the addressee’s face – in
other words, how small talk is used independent of and in connection with FTAs, as
general FBAs and mitigating FBAs.
The case study is based on two corpora. The primary focus is on a corpus from the
digitally-based community of practice: a Swedish print journal where most of the
work is carried out via email. The second corpus is from the physically-based com-
munity of practice, a larger Swedish print newspaper office. Each of these is de-
scribed further below.
dividual editors and the board occasionally arrange face-to-face meetings, and they
make phone calls with some regularity, alongside their email communication.
However, most contacts are by email. Not part of the network are external senders
(companies, etc.) that sometimes circulate information for the journal to the edi-
tor-in-chief.
For this journal, a corpus of 3,200 messages from the editor-in-chief’s email-
box was selected in order to obtain a rich picture of the communication within the
digitally-based community of practice. According to the editor-in-chief, her role
involved the most differentiated communication in the community of practice, in-
cluding both a wide range (in genres and in content) of emails and interaction with
different writers. The email messages were written over a three-year period
(2002–2005) by individuals in the community of practice, where the writers repre-
sent the work categories mentioned above and where the data comprise all mess-
ages that were sent and received during the period through the chief’s email pro-
gram (Eudora). Table 1 gives a breakdown of the correspondence between the
editor-in-chief and different groups of writers. Most of these messages were sent
from one individual to another (n=2176), but a significant portion were also sent as
group messages or were sent to a reciever and cc’ed simultaneously to the editor-
in-chief for her information (n=1024).
Information about the messages was classified by genre, content, exchange pat-
terns (one-to-many vs. “interactive” messages), when in a sequence a certain email
appears, and between which actors (in terms of their role in the office hierarchy)
the email was exchanged. Every email was also reviewed for the occurrence and
type of small talk in relation to general FBAs and mitigating FBAs. These data
were supplemented by four interviews with the editor-in-chief about the settings
for various types of messages and the network’s work routines. The focus in the
analysis is not on quantifying instances in these data but on demonstrating what
some of the frequent types of politeness strategies in the data look like that occur
throughout the period studied.
The following data presentation starts with a characterization of the email interac-
tion in the physically-based community of practice, a context in which alternatives
to email are available. This is followed by a discussion of what happens to elec-
ations without explanations. The phrase “The pregnant fish” in the message is the
title of the article in question in the messages.
This example illustrates an evident reality in this physical workplace. Email is re-
served for circulating information and for simple, tersely formulated exchanges in
which small talk is not included.
A functional divide exists at the office among various communication options
according to how the individuals involved view the communication at hand and the
ways in which the external setting socially frames various work issues. Oral face-
to-face conversation is a site for small talk in corridors; at coffee-makers, desks,
and faxing machines; for discussing more complex issues; and for airing sensitive
issues – matters about which an addressee can be expected to have a different
opinion or be angered or saddened (Hössjer 2002, 2006). Telephone is the pre-
(3) Exchange between editor-in-chief and chairman of the board of the journal
Hi!
Thanks for the greetings. Yes, Christmas always keeps you running in circles.
What’s more, I fell and hurt my foot and I’m having a hard time walking.
There’s hope I’ll survive, but it sure complicates life. Hard to walk down the
church aisle with any dignity on Christmas morning, for instance!
Best wishes
Carin
It would be possible to see this type of everyday anecdote as evidence that the pub-
lishing frequency of this office is lower than at the news office, and thus that there
is more scope for social digression. It should be borne in mind, however, that this
exchange takes place alongside day jobs and that the overall work situation for
everyone concerned is packed to the breaking point, as is commented on in various
messages. Given this, a more reasonable interpretation might be that chatting about
health functions as a way of maintaining social continuity. This type of social fram-
ing creates a common ground that the associates can share. There are sometimes
long time intervals (days, weeks) between certain messages. After these long in-
tervals, small talk could be seen as a way of re-establishing a context that might be
somewhat remote in the participants’ memories, when inserted quotations are not
sufficient to maintain the flow. However, this situation is not the case for the vast
majority of messages, which are sent in the course of one or two days.
What is provided in the vast majority of messages with small talk is an invoked
social space of external events and impressions that can be “seen” and referred to: a
shared external social context that helps frame the ongoing work. There are clear
similarities here with the type of small talk that occurs at the news office in face-
to-face conversations (Hössjer 2006) and that various studies have shown to play a
key role in the sense of community that emerges in a workgroup when individuals
take time to listen and react empathically to non-work-related information (cf. Eelen
2001; Fletcher 1999; Holmes and Schnurr 2005; Mills 2003; Watts 1992, 2003;
Wierzbicka 1999). In this sense, the digitally-based community of practice can be
described as a group created over time by the sustained pursuit of a shared enterprise
and a shared repertoire of ways of doing things (cf. Eckert and McConell-Ginet
1992; Wenger 1999). A sense of community is created within the framework of
shared work and routines. However, a key difference from a work community that
has the form of a physically-based community of practice (e.g., the news office) is
that the creation of commonality in the digitally-based community of practice oc-
curs through more varied kinds of email content that includes small talk. In this di-
gitally-based community, small talk in messages makes the job tolerable by creating
an atmosphere in which monotonous job practices are woven together with stories,
events, dramas, and the rhythms of common life (Wenger 1999: 46).
The occurrence of this kind of social reference in the messages varies de-
pending to some extent on the nature of the matter at hand, who is writing, and
who is receiving. However, two basic situations emerge in how such references
are made that relate to two different conditions for the editorial work. This can be
seen across various genres of communication (e.g., work directives, requests, dis-
cussions of manuscripts, planning of new journal issues) and with interlocutors in
different positions.
The first condition is that messages are produced with the expectation that
duties are being carried out as prescribed by routines. This condition tends to in-
volve messages aiming to provide information, come up with a suggestion, and/or
plan simple things. In such messages, the referencing of a social context takes the
character of “framing” – linking the work context to another, non-work frame
(Goffman 1974, 1981; Tannen 1993, 2007), in more-or-less independent passages.
Small talk introduces or concludes messages, and it does not tie in directly with the
information provided or requested, beyond contextualizing it in a general social
exchange (i.e., it is a general FBA). Such comments might take the form of an
initial/concluding remark, a postscript occurring as a relatively freestanding inser-
tion at the end of a message, or as a single off-topic message within a larger
sequence of mails devoted to work. The latter is the case in example 3, in which
Carin replies to a greeting from the chairman. His greeting was a freestanding
insertion at the end of a message (“Hope you’ll have a healthy and peaceful Christ-
mas”), the rest of the text being devoted to routine economic planning. Carin’s
reply to his greeting (“Hi!”) is followed by the narrative in example 3. In subse-
quent messages, after a short reaction from the chairman in the beginning of the
following message, the interaction returns to the work topic (economic planning)
given in the subject heading of the thread.
The second condition under which a social frame is built up around the edi-
torial work involves complications or uncertainty in the work situation that may
risk threatening the addressee’s social face, and it constitutes a noticeable depar-
ture from the regular rhythm of work. In messages sent under this condition there is
a closer connection between the framed contexts than in example 3. Small talk
serves as explanation or excuse for why certain outcomes cannot be realized or
have not been done so far (this being an FTA). Face in the situation is boosted by
writers giving reasonable justification for this shortcoming (i.e., such acts are miti-
gating FBAs).
In example 4, the social framing functions in this way. The message is about
coordinating calendars ahead of a meeting, and the writer (chairman of the board)
starts by saying that he will be busy during a certain week (work-related context),
thereby hinting that an arrangement for meeting this special week (as earlier pro-
posed by Carin) will not be possible (this statement from the chairman is an FTA).
This is followed by the reason that will be keeping him busy (social context, a
mitigating FBA), which then leads to further discussion about coordinating work
(work-related content, other mitigating FBAs). The overall effect of this framing
of work passages in a social context is to help explain why the writer cannot make
it to a meeting next week. The writer treats the addressee’s request for meeting as
important and thus mitigates the FTA, while boosting the addressee’s face by
using small talk as a mitigating FBA.
(4) Exchange between editor-in-chief and chairman of the board of the journal
During the week Aug. 11–17 I’ll be busy. We’re going to be looking after our
oldest grandchildren that week. The family is moving into their new house in
Helsingborg.
I can make it during the week starting Aug. 18 though. How does that week
look for you?
I’d be grateful for a reply.
Yours
Bo
pite its advanced age. When the legendary mayor of Sigtuna Gustav ( ? ) Dahl
was interviewed ahead of his 100th birthday and asked how was doing, he
answered: “At my age you feel fine, otherwise you wouldn’t be there to talk
about it.”
Yours
Jörgen
In example 5, the FTA entailed by the delay of the article is followed by six pas-
sages that can all more or less be regarded as mitigating FBAs. These FBAs in-
clude an apology for the delay (“I’m sorry …”), an explanation of a general but
non-specified reason for it (“but there are reasons”) in connection with the delivery
of the delayed material (“The last word should be …”), and a request for confirma-
tion that the message has reached the addressee (“Could you let me know …”).
This request can be interpreted both as an FTA, because of the extra work the con-
firmation might cause the addressee, and as an FBA, because of the importance that
the writer is giving the matter in asking for the confirmation. The request is fol-
lowed by a wish for well-being (“Hope all is otherwise well …”, invoking a social
context) and an instance of praise (“It is always a pleasure …”, another social con-
text) in connection with a joke (“At my age you feel fine …”, yet another social
context). All of these FBAs are in different ways directed toward the addressee
(editor Carin and her work with the journal). The first FBAs relate to work,
whereas the latter ones relate to various social contexts. Social harmony is thus re-
established (as confirmed by a subsequent message containing another joke from
Carin), following the initial FTA (the delay), in progressive steps through chaining
of mitigating FBAs by the writer (Jörgen).
The three latter passages in this message can be seen as indirectly enhancing
the writer’s (Jörgen’s) face, as well. Praising and showing an interest in other
people – their deeds and expressions – as Jörgen is doing here is often perceived as
a positive trait in social contacts. This evidence of the co-existence of more and
less direct levels provides a glimpse of the complexity of face management carried
out through the messages.
The framing and chaining evident in these messages is not unique to this com-
munity of practice, nor is the implicitness in the face work that can be seen here.
Substantial research describes how such politeness strategies are used in various
types of (oral) institutional conversations: university tutorials, nurse/doctor-patient
dialogs, midwife dialogs, and operator calls (Benwell and Stokoe 2002; Bergmann
1992; Linell and Bredmar 1996; McEnery, Baker, and Cheepen 2002; Spiers
1998). This previous body of research points out, in addition to indirectness, the
use of strategies such as avoidance and the invoking of routines and external cir-
cumstances in averting interpersonal conflicts. The strategies also evince a similar
orientation in the social interaction that occurs among colleagues at various work-
places – toward favoring a constructive climate for cooperation and the successful
execution of work (e.g., Eelen 2001; Mills 2003; Watts 1992, 2003; Wierzbicka
1999).
In the digitally-based community of practice, as compared to the physically-
based community of practice, the interaction in email messages can be seen as an
adaptation to an external situation in which the writers have no physical contact
and no clear channel for immediate feedback. Special strategies (framing and
chaining) are mustered in the community to protect social face through linguistic
expressions the email writers jointly contribute that do not occur in messages
within the physically-based community of practice. These strategies dynamically
interact with how sensitive the situation is, whether or not it involves compli-
cations that may risk threatening the recipient’s face. The lack of physical periph-
eral information seems to invoke a kind of chronic cautiousness, which interacts
with attempts to establish a shared social frame to compensate for the lack of in-
terpretive cues in the communicative situation. In the physically-based community
of practice, where such a frame exists, these strategies can be excluded from email
messages without threatening the exchange. The use of small talk as a politeness
strategy is thus highly context-bound and related to practices within the commu-
nities of practice. As highlighted in previous research on politeness, it is clear that
the characterization of email as a lean medium is over-simplified; email is ex-
ploited in different ways in different situations within the communities studied
here. Moreover, individuals in the digitally-based community of practice use small
talk as a more dynamic adapted strategy in relation to the external context (small
talk as general FBAs or mitigating FBAs) than has been discussed in the previous
email politeness literature.
6. Conclusion
This chapter discussed the professional use of email at a Swedish print journal
where most of the work is carried out via email. This represents a work situation of
broad relevance in modern life, where it is common for professionals to make use
of CMC as a virtual meeting place. This situation was compared with email com-
munication taking place at a Swedish news office where the workers share a physi-
cal context.
This comparison showed that politeness is a situationally-adapted strategy that
is influenced by the external conditions that hold for email communication within a
community of practice. Interlocutors seem to adapt their politeness behavior to the
external situation. Small talk plays an important role in these professional contexts.
In the physically-based community of practice, small talk is left out of email mess-
ages and carried out face-to-face. In the digitally-based community of practice,
small talk tends to take the form of independent passages when the workflow fol-
lows prescribed routines (general FBAs). In cases of a departure from the regular
rhythm in the work – complications or uncertainty in the situation that may risk
threatening the recipient’s social face – small talk more often takes the form of
mitigating FBAs to mitigate the face threat.
These findings point to a need for a situated view of politeness in email com-
munication in relation to various external conditions, e.g., in workplace environ-
ments. They also suggest a need for more elaborated concepts in politeness theory
in line with those demonstrated here using the concept of community of practice.
Such elaborated concepts ought to take into account the dynamics of individuals’
politeness strategies in different work formations (e.g., according to size and his-
tory, including short term groups), in relation to changing working conditions in
actual situations (e.g., more-or-less routine work in connection with more-or-less
time-critical work conditions), and in connection with individuals’ use of various
modes of CMC. The findings of the study presented in this chapter indicate that
different expectations are in place regarding what constitutes polite behavior in
connection with the use of email within the different communities of practice.
These expectations include, on the one hand, the role the medium as such plays in
relation to other media vis-à-vis what type of communication it is used for. On the
other hand, they include the type of politeness (normatively) associated with the
medium in actual situations, which functions to accomplish this communication in
an efficient manner.
From this perspective, research in different contexts might be extended to ad-
dress how the flow of work issues and small talk are woven into the dynamics of
turn-taking mechanisms in the process of encouraging a sense of togetherness and
successful execution of work. In physically-based communities of practice, this
may include questions about how interaction is forwarded when starting in one
medium and moving into another when dealing with more or less problematic and
time-critical matters. In digitally-based communities of practice, such questions
may be connected with the use of synchronous and asynchronous media: Strategies
such as framing and chaining may undergo media adaptation in relation to ongoing
work.
Acknowledgements
I am grateful to Jonas Carlquist, Else Nygren, and Kerstin Severinson Eklundh for
valuable comments on previous versions of the manuscript, and to Donald Mac-
Queen for help with translating and improving the language.
Note
1. For the sake of anonymity, the operational sphere of this journal is not revealed in this
chapter. In cases where the content of individual emails describes phenomena that would
make it possible to identify the journal, this content has been replaced by a more general
equivalent. The use of fictitious pseudonyms to preserve anonymity also applies to the
news office material.
References
Fukushima, Saeko
2004 Evaluation of politeness: The case of attentiveness. Multilingua 23: 365–387.
Goffman, Erving
1974 Frame Analysis. New York: Harper & Row.
Goffman, Erving
1981 Forms of Talk. Philadelphia, PA: University of Pennsylvania Press.
Goldsmith, Daena J.
2006 Brown and Levinson’s politeness theory. In: Bryan B. Whaley and Wendy
Samitel (eds.), Explaining Communication: Contemporary Theories and
Exemplars, 2919–236. Mahwah, NJ: Lawrence Erlbaum.
Gu, Yueguo
1990 Politeness phenomena in modern Chinese. Journal of Pragmatics 14:
237–257.
Hatipoğlu, Çiler
2007 (Im)politeness, national and professional identities and context: Some evi-
dence from e-mailed ‘Call for Papers’. Journal of Pragmatics 39(4): 760–773.
Hernández-Flores, Nieves
2004 Politeness as face enhancement: An analysis of Spanish conversations between
friends and family. In: Rosina Márquez-Reiter and María E. Placencia (eds.),
Current Trends in Pragmatics of Spanish, 265–286. Amsterdam/Philadelphia:
John Benjamins.
Herring, Susan C.
1994 Politeness in computer culture: Why women thank and men flame. In: Cultural
performances: Proceedings of the Third Berkeley Women and Language Con-
ference, 278–294. Berkeley, CA: Berkeley Women and Language Group.
Herring, Susan C.
2007 A faceted classification scheme for computer-mediated discourse. Language@
Internet 4, article 1. http://www.languageatinternet.org/articles/2007/761
Holmes, Janet
2000 Doing collegiality and keeping control at work: Small talk in government de-
partments: In: Justine Coupland (ed.), Small Talk, 32–61. Harlow, UK Pearson.
Holmes, Janet and Stephanie Schnurr
2005 Politeness, humour and gender in the workplace: Negotiating norms and iden-
tifying contestation. Journal of Politeness Research 1: 121–149.
Hongladarom, Krisadawan and Soraj Hongladarom
2005 Politeness in Thai computer-mediated communication. In: Robin T. Lakoff and
Sachiko Ide (eds.), Broadening the Horizon of Linguistic Politeness, 145–162.
Amsterdam/Philadelphia: John Benjamins Publishing Company.
Hössjer, Amelie
2002 Former för dialog och beslut. Om växling mellan kommunikationsformer och
bruk av media i redaktionellt tidningsarbete. [Forms of dialogues and deci-
sions. On the alternation between forms of communication in news editorial
work.] IPLAb 206. Stockholm: Department of Numerical Analysis and Com-
puter Science TRITA_NA_P0301.
Hössjer, Amelie
2006 Externa brev och intern vardagskommunikation. Om bruket av e-post på en
tidningsredaktion. [External mails and internal everyday communication. The
Lee, Cynthia F. K.
2004 Written requests in emails sent by adult Chinese learners of English. Lan-
guage, Culture and Curriculum 17(1): 58–72.
Linell, Per and Margareta Bredmar
1996 Reconstructing topical sensitivity: Aspects of face-work in talks between mid-
wifes and expectant mothers. Research on Language & Social Interaction 29:
347–379.
Löfgren Nilsson, Monica
1999 On ideals and organizational principles in everyday editorial work. Ph.D. dis-
sertation, Göteborg University Department of Journalism and Mass Communi-
cation.
Malinowski, Bronislav
1923 The problem of meaning in primitive languages. In: C. K. Ogden and I. A. Ri-
chards (eds), The Meaning of Meaning: A Study of Influence of Language Upon
Thought and of the Science of Symbolism, 296–336. New York: Harcourt,
Brace and World.
Matsumoto, Yoshiko
1988 Reexamination of the universality of face: Politeness phenomena in Japanese.
Journal of Pragmatics 12(4): 403–426.
McEnery, Tony, Paul Baker, and Christine Cheepen
2002 Lexis, indirectness and politeness in operator calls. In: Pam Peters, Peter Col-
lins, and Adam Smith (eds.), New frontiers of Corpus Research: Papers from
the Twenty First International Conference on English Language on Computer-
ized Corpora Sydney, 53–70. Amsterdam/New York: Rodopi.
McKenna, Katelyn Y. A., Amie S. Green, and Marci E. J. Gleason
2002 Relationship formation on the Internet: What’s the big attraction? Journal of
Social Issues 58(1): 9–31.
Mills, Sara
2002 Rethinking politeness, impoliteness, and gender identity. In: Lia Litosseliti and
Jane Sunderland (eds.), Gender, Identity and Discourse Analysis, 69–89. Am-
sterdam/Philadelphia: John Benjamins.
Mills, Sara
2003 Gender and Politeness. Cambridge, UK: Cambridge University Press.
Morand, David A. and Rosalie J. Ocker
2003 Politeness theory and computer-mediated communication: A sociolinguistic
approach to analyzing relational messages. In: Proceedings of the Thirty-Sixth
Hawaii International Conference on System Sciences. Los Alamitos, CA:
IEEE Press. http://csdl.computer.org/comp/proceedings/hicss/2003/1874/01/
1874toc.htm
Mullany, Louise
2006 “Girls on tour”: Politeness, small talk, and gender in managerial business
meetings. Journal of Politeness Research 2(1): 55–77.
Murphy, Margaret and Mike Levy
2006 Politeness in intercultural email communication: Australian and Korean Per-
spectives. Journal of Intercultural Communication 12. http://www.immi.se/
intercultural/nr12/murphy.htm
Murray, Denise E.
1988 The context of oral and written language: A framework for mode and medium
switching. Language in Society 17: 351–373.
Oestreicher, Amelie
2000 Editing news texts. A study of how texts are processed at six Swedish dailies.
[Bearbetning av nyhetstext. En studie av texthantering vid sex svenska dag-
stidningar.] Ph.D. disseration, Uppsala University. Publications of the Depart-
ment of Scandinavian Languages, Uppsala University 52.
Oliveira, Sandi M. de
2007 Breaking conversational norms on Portuguese users network: Men as adjudi-
cators of politeness? In: Brenda Danet and Susan C. Herring (eds.), The Multi-
lingual Internet: Language, Culture, and Communication Online, 256–278.
New York: Oxford University Press.
Panteli, Niki,
2002 Richness, power cues and email text. Information & Management 40(2):
75–86.
Panteli, Niki and Monica Seely
2006 The email gender gap. In: Mary Barrett and Marilyn J. Davidson (eds.),
Gender and Communication at Work, 250–259. Hampshire and Burlington:
Ashgate.
Persichitte, Kay A., Suzanne Young, and Donald Tharp
1997 Conducting research on the Internet: Strategies for electronic interviewing. In:
Proceedings of Selected Research and Development Presentations at the 1997
National Convention of the Association for Educational Communications and
Technology, 279–283.
Rogers, Priscilla and Song Mel Lee-Wong
2003 Reconceptualizing politeness to accommodate dynamic tensions in subordi-
nate-to-superior reporting. Journal of Business and Technical Communication
17: 379–412.
Schegloff, Emanuel A.
1988 Goffman and the analysis of conversation. In: Paul Drew and Anthony Woot-
ton (eds.), Erving Goffman: Exploring the Interaction Order, 9–135. Cam-
bridge, UK: Polity Press.
Sherblom, John
1988 Direction, function and signature in electronic mail. Journal of Business Com-
munication 25(4): 39–54.
Skovholt, Karianne
2009 Email literacy in the workplace. A study of interaction norms, leadership com-
munication and social networks in a norwegian distributed work group. Ph.D.
dissertation, University of Oslo.
Skovholt, Karianne and Jan Svennevig
2006 Email copies in workplace interaction. Journal of Computer-Mediated Com-
munication 12(1), article 3.
http://jcmc.indiana.edu/vol12/issue1/skovholt.html
Spiers, Judith A.
1998 The use of face work and politeness theory. Quaitativel Health Research 8(1):
25–47.
Tannen, Deborah
1993 What’s in a frame? Surface evidence for underlying expectations. In: Deborah
Tannen (ed.), Framing in Discourse, 14–50. Oxford, UK: Oxford University
Press.
Tannen, Deborah
2007 Talking Voices: Repetition, Dialogue, and Imagery in Conversational Dis-
course. Cambridge, UK: Cambridge University Press.
Terkourafi, Marina
2001 Politeness in Cypriot Greek: A frame-based approach. Ph.D. disseration, Uni-
versiy of Cambridge.
Turnage, Anna K.
2007 Email flaming behaviors and organizational conflict. Journal of Computer-Me-
diated Communication 13(1), article 3.
http://jcmc.indiana.edu/vol13/issue1/turnage.html
Utz, Sonja
2007 Media use in long-distance friendships. Information, Communication & So-
ciety 10(5): 694–713.
Waldvogel, Joan
2001 Email and workplace communication: A literature review. Language in the
Workplace Occasional Papers 3. http://www.victoria.ac.nz/lals/publications/
lwp.aspx
Waldvogel, Joan
2007 Greetings and closings in workplace email. Journal of Computer-Mediated
Communication 12(2), article 6.
http://jcmc.indiana.edu/vol12/issue2/waldvogel.html
Watts, Richard J.
1992 Relevance and relational work: Linguistic politeness and politic behaviour.
Multilingua 8(2/3): 131–166.
Watts, Richad J.
2003 Politeness. Cambridge, UK: Cambridge University Press.
Watts, Richard J., Sachiko Ide, and Konrad Ehlich (eds.)
2005 Politeness in Language: Studies in Its History, Theory and Practice, 2nd Ed.
Berlin: Mouton de Gruyter.
Watts, Stephanie
2007 Evaluative feedback: Perspectives on media effects. Journal of Computer-
Mediated Communication 12(2): 384–211.
Wenger, Etienne
1999 Communities of Practice: Learning, Meaning, and Identity. Cambridge, UK:
Cambridge University Press
Wierzbicka, Anna
1999 Emotions Across Languages and Cultures. Cambridge, UK: Cambridge Uni-
versity Press.
Zack, Michael H.
1993 Interactivity and communication mode choice in ongoing management groups.
Information Systems Research 4(3): 207–239.
1. Introduction
2. Theoretical background
only individuals interacting within Communities of Practice who will be able to as-
sess whether a particular act is polite or impolite” (Mills 2002: 76; cf. O’Sullivan
and Flanagin 2003). (3) One should focus not only on the impoliteness of individ-
ual contributions but on negotiation regarding impolite behavior over time; and
primarily, (4) on seriously intended, strategic rudeness.
The sequence analyzed here occurred over four days in October–November 2005,
on an electronic mailing list for Israelis in the San Francisco Bay Area in the
United States. Of 800 subscribers, 100 were students at the host university. The list
mandate states that posters may announce events, seek people with common inter-
ests, ask questions, or get recommendations. Posting messages that are defaming,
threatening, or that contain obscene language is prohibited.
Presumably, most subscribers are native-born, highly educated professionals or
their spouses/partners, part of an Israeli diaspora of 40,000 in the San Francisco/
Silicon Valley area (Shakhar 2005). Most use their real names in email addresses
and when signing postings. Subscribers may know one another offline profession-
ally or socially, participate in the same local activities, or have children in the same
schools. Typically, postings are short exchanges about goods and services.
3.2. Participants
Nine males and five females participated in the analyzed sequence. Most had con-
temporary Hebrew first names; three, with Russian first names, may have immi-
grated to the United States directly, or via Israel. To protect participants’ identity,
all names have been changed. Comparable names were assigned: A male with a
typical name for a 20-40-year-old Israeli Jew was assigned a similar name; a fe-
male with a Russian or French name was given another of similar connotation.
– How and why does this flame sequence begin, escalate, and end? What key
speech acts are involved?
– What evidence is there that conflict is occurring?
– What forms of linguistic impoliteness occur?
– What remedial activity, if any, takes place?
– In what ways might Jewish and Israeli communication styles find expression in
this sequence?
– Does gender play a role, and if so, how?
4.1. Metatalk
We know that a conflict is occurring in these data, first of all, because participants
engage in metatalk about it (Table 1), although they never use the term “flaming”
or any Hebrew equivalent.
While “this whole exchange”, “the whole thing”, “the whole event”, and “this
topic” neutralize the trouble, “monster”, “brawl”, “tsunami”, and “storm” connote
extreme conflict. The provocateur who was central in the entire sequence, called
here Gad, produces seven of the 10 labels and is the most verbally creative. Among
other expressions, he represents the conflict as a “wonderful do-krav”6 (‘a wonder-
ful duel’). Romanizing the Hebrew term for ‘duel’, he indicates that he is enjoying
the fight, even if his complaint was initially serious. He also characterizes the tor-
rents of words as “a storm” and a series of “verbal tsunamis”.
Gad’s contrast of efroach (efroax in my transliteration system)7 and man-eating
pre-historic monster resembles the English “to make a mountain out of a molehill”.
The Hebrew equivalent is la’asot pil mizvuv (‘to make an elephant out of a fly’;
Doniach and Kahane 1998; Rosenthal 2005; Zilkha 2002). In another posting Gad
speaks of typing words metaphorically – “everyone put down their guns and key-
boards” – characterizing the list as “a battle field” and “a ship that sailed again in
safe waters”.
4.3. Citations
Citation of previous messages is a common feature of asynchronous CMC modes.
It helps recipients keep track of topics and correspondents, but it also serves other
functions (Marcoccia 2004; Severinson Eklundh 2010; Tanskanen 2001;
Thompsen and Foulger 1996). Analyzing two asynchronous modes, Tanskanen
(2001) suggested that “overt reference by the writers to the viewpoint they are dis-
agreeing with could (…) be a collaborative strategy, motivated by the wish to
avoid conflict or even by a fear of flaming” (238–239; italics added). This interpre-
tation is not applicable, however. The 24 postings, including all citations, fill 25
pages of single-spaced text in Word, over 7,000 words!8 Twelve posters include the
initial incendiary posting as well as up to five others, suggesting that they were re-
sponding emotionally and rapidly.
4.4. Reprimands
Reprimands also indicate trouble. There were brief reprimands by two females and
one male, and lengthier ones by the organization’s president and listowner. Dina,
the president, wrote:
(1) I am ashamed that I cannot be away from my email for even a weekend without
coming back to find my … mailbox littered with irrelevant emails. Tal [list
owner] and I do not need to be your kindergarten teachers.
The latter is a patronizing transfer from Hebrew, Tal v’ani lo tsrixim l’hiot hagan-
anot shelaxem. While this is a perfectly acceptable English sentence, it is much
more probable in the Israeli context. Tal concluded the event as follows:
(2) Enough! Do not send any more replies to this topic. No more clarifications. No
more last words. I sincerely appologize to the list members for missing the 1st
wave of emails and not putting an end to this sooner. I guess the list cannot be
left unattended for a weekend.
Complaining about ostensibly unnecessary postings, Gad claims that Yossi could
find the information by asking a telephone operator or looking in the phone book or
online. The complaint is formulated as a series of directives to avoid the offending
act in the future, followed by two rhetorical questions.
The bald on-record imperative form, in itself, is not necessarily offensive, since
Israelis tend to be direct even when polite (Blum-Kulka 1987; Blum-Kulka, Danet,
and Gherson 1985; Blum-Kulka, House, and Kasper 1989). However, it is not the
job of a subscriber to issue orders to others; contextually, these imperatives are im-
polite. Issuing this challenge in view of all is more insulting than if the challenge
were private.
In its excessive length the posting violates Grice’s (1975) Maxim of Quantity.
Its commanding tone is aggravated by writing-specific means that mimic emotion-
ally heightened speech: dramatic play with typography – the use of all caps,
multiple exclamation points, and the comics-like “DAHHHH!!!!!!!!!!!!!!”. Devi-
ating from brief, unmarked list style, these features are known to occur in other in-
cendiary communication online (McKee 2002). In most online contexts, all caps
are understood as the textual equivalent of shouting (Crystal 2006: 37; Danet 2001:
17–19), and here they simulate shouting during an oral argument (Schiffrin 1984).
That Gad’s text and play with typography were carefully planned and edited is in-
dicated by the numbered listing of items, the three imperatives, and the sudden
switch from all caps to all lower-case writing.
The structure of this challenge contains parallelism and repetition, which “in-
creases the imposition upon the target and/or emphasizes the negative attitude of
the speaker toward the target” (Culpeper, Bousfield, and Wichmann 2003: 1561),
thereby boosting impoliteness. Switching to all lower case changes the footing to
the play frame (Goffman 1979): Gad effectively downgrades the aggressiveness of
his posting. The call for future offenders to buy ice cream for all is absurd, since the
communication is virtual, and nobody knows who “all” are.
Gad has so far directed five ad hominem insults to Yossi:
1) You are lazy.
2) You waste our time, fill our screens with garbage, and may even cost us modem
time and money.
3) You have no imagination.
4) You are stupid.
5) You are so incompetent that you don’t know even how to use a broom.
The last item is particularly insulting, since everybody knows how to use a broom,
as is implied by the rhetorical form of the utterance, and it is a gross violation of
Grice’s (1975) Maxim of Relevance.
6. Escalation
Insults can manifest three possible combinations of illocutionary force and perlo-
cutionary effect (Austin 1975; Searle 1976; Vrooman 2002): (1) an utterance is in-
tended to hurt, and is experienced as such; (2) an utterance is not intended to hurt,
but is experienced as hurtful; or (3) an utterance is intended to hurt, but is not ex-
perienced as hurtful.10 The sequence under analysis is full of type 1 personal insults.
Table 2 lists all the insulters and insultees. The same five individuals are both
insulters and insultees. Of seven insult pairs, three are reciprocal. Gad insults two
other individuals besides Yossi.
Gad f Yossi
Yossi f Natasha
Natasha f Yossi
Irena f Gad
Yoav f Gad
Gad f Yoav
Gad f Irena
Natasha agrees with Gad’s complaint. Typing a portion of her posting in all
caps, she signals her identification with him:
(5) I totally agree and I’m glad someone had the balls to say something. (I didn’t
even see the ice cream question cause I rarely look at the list anymore – BE-
CAUSE THERE ARE TOO MANY STUPID QUESTIONS ON IT) but I DID
notice your post because it was addressed to all, so I paid attention.
Thanks!
Instead of responding directly to Gad, Yossi posts without comment a
630-word explanation from the web about an 809 Area Code scam, in which un-
witting dialers incur exorbitant international charges for ostensibly American
calls. Seemingly helpful, he shares this information with the others. However,
“blasting” the list with this explanation without comment is an indirect, subtly im-
polite way of counter-attacking Gad, communicating not only “You were wrong to
attack me”, but also “You (Gad) flooded me with words; now I am flooding you
with even more words, and in view of all the others – so there”!
Direct evidence for how Yossi feels about Gad’s attack is revealed when Nat-
asha copies into a posting a private email Yossi had sent her:
(6) Thanks for the compliment Yossi. Were you too cowardly to send to the list?
Afraid they might see what kind of a person you are? Well, I’ll send it for you …
maybe you will get booted off now.
-----Original Message-----
From: Yossi
To: Natasha
That’s because you are a stupid cow …
Yossi has called her a “stupid cow” because of her agreement with Gad, revealing
that he was very hurt and angry at Gad. Natasha insults Yossi by violating a general
online norm – not to forward private email to others without the writer’s per-
mission, especially if the content is sensitive. She sends her ironic response to the
list, “thanking” him for the sarcastic “compliment”, calling him a “coward”. Doing
so before an audience compounds the insult.
Another female poster, Irena, lashes out at Gad:
(7) Gad, you are a very sad, pathetic and defective human being. Thank you for
sharing your insignificant opinions and spamming the list with that spew of
filth you pass for as a reply.
I am sure everyone on the list is anxiously waiting for your next idiotic mono-
log.
In attacking Gad in a partially ironic manner (“thank you for sharing”), Irena re-
sponds not only to his attack on Yossi, but to a previous posting before this event,
in which he had attacked her for twice asking the list to recommend a hairdresser.
Yoav, another male, agrees, adding new insults:
(8) I second this comment. Gad seems to always complain about the way the com-
munity chooses to use the list. Gad, why don’t you stop using it and save us
your negative comments and energy. In the same amount of time you took to
criticize the person asking the question, you could have just given the answer
and been a kind human being instead of a rotten one.
I have a solution for you and Natasha … Create a separate folder where all the
[list] emails go into automatically. That way it won’t “load up your inbox” as
you put it …
Do you think you know how to do this, or “do you want us to all come over and
show you how” to use your computer?
Pathetic!
Yoav finds Gad’s comments “negative”; he is “a rotten human being”. Like Irena,
he thinks Gad is “pathetic”. Especially hurtful is the suggestion that Gad is com-
puter illiterate, a violation of Grice’s (1975) Maxims of Relevance and Quality.
This ironic rhetorical question parallels Gad’s question about the broom, as indi-
cated by the quotation marks around these words. Thus Irena and Yoav ally them-
selves against Gad and Natasha.
In the next posting, another male, Avishai, posts new information:
(9) I believe that it is the North American 242 area code. like 1–242-XXX–XXXX
from the US and Canada. But I am not absolutely sure.
Gad playfully rejects this, seemingly abandoning “the facts” and his challenge to
Yossi. He pretends to chastise Avishai for putting a stop to “this wonderful ‘do-
krav’” [‘duel’; quotation marks in original]:
(10) Avishai, are you trying to put a stop to this wonderful “do-krav” [‘duel’] we
are having. Shame on you. Shana Tova Umetuka like your wife’s wonderful
cakes. Gadi
Playfully, he code-mixes again, wishing Avishai Shana Tova Umetuka – ‘A happy
and sweet New Year’. Because the fall Jewish holidays are over, use of this ex-
pression is incongruous and humorous. Picking up on metuka (‘sweet’), he refers
to Avishai’s wife’s cakes as also sweet, another violation of the Maxim of Rel-
evance (Grice 1975). In a final ingratiating move, he signs the posting with his
nickname, “Gadi”, a diminutive of Gad (Blum-Kulka and Katriel 1991). Turning
playful compounds Gad’s original impoliteness: By transforming communication
into a mere game he trivializes the hurt he has caused.
Gad’s posting to Avishai was preceded chronologically by his insulting re-
sponse to Irena:
(12) Irena, I am sad for you that humor is not part of your makeup.
“People who live in glass houses shouldn’t throw stones” as the old proverb
says.
Irena, read my words like mild paprika … feel yours like cayenne pepper. Ex-
cuse me Irena, I have to get a tissue to clean your saliva out of my eyes, I hope
it’s not venom.
Gad’s words were intended as “mild paprika”, not as the strong insult she appar-
ently felt, whereas he feels hers as if they are much stronger “cayenne pepper”, as
well as “saliva in his eyes” from a spitting snake. Ostensibly mitigating his pre-
vious words to her, he newly insults her, insinuating that she is a snake (“I hope
[your words are] not venom”). Verbal inventiveness, a form of fun for him, is in-
separable from expression of hostility.
Eventually, Gad realizes that his attack on Yossi was unjustified. Twice he engages
in remedial activity. Goffman (1967: 20–22) identified four components of a full
remedial interchange: (1) offended person draws attention to the offense; (2) of-
fender apologizes; (3) offended accepts the apology; and (4) offender thanks the
other for forgiveness received. As in previous research on remedial activity online
for serious offenses (see Harrison and Allton, this volume), remedial activity here
is extremely truncated.
Olshtain and Cohen (1983) identified five types of strategies employed in
apologies (Table 3):
Table 3. Five strategies for the performance of an apology (Olshtain and Cohen 1983)
Strategy Type
(1) IFID (illocutionary force indicating device)
(2) Expression of speaker’s responsibility; relates to speaker’s willingness to
admit to fault
(3) Explanation
(4) Offer of repair
(5) Promise of forbearance in the future
The first category includes formulaic, routinized forms of apolog, such as the per-
formative verb “I apologize” and the expression “I’m sorry”. The others are self-
explanatory. An apology may combine two or more of these strategies.
Gad’s initial remedial attempt is primarily category (3), focusing on confusion
between “scheme” and “scam”:
(13) In the message below, which I reacted to, and responded to, you used the word
“scheme”, which does not necessarily have a negative connotation, such as a
color scheme,or any system of correlated things. I took it in a positive sense
such as a layout or a list of phone numbers for the Bahamas …
If you had used the word scam – if that is what you meant – or “would any-
one like to know the area codes of the Bahamas phone scam”? there would
have been no [doubt] in my mind as to your intention and I would not have
responded as I did. And from there the “efroach” [‘chick’] was hatched and
became a man-eating-pre-historic-dinosaur. So lets choose and use our words
correctly. This whole exchange didn’t bring the nicest things out of some
people, you for one, in the eyes of some, me, and of course the cobra who spat
at me, what was her name? I hope she has somewhere to channel her anger.
Shana Tova
Gadi
Even native speakers of English might have confused the two expressions. The
Oxford English Dictionary defines “scheme” as “a plan, design; a programme of
action; the designed scope and method of an undertaking (…); often with unfavor-
able notion, a self-seeking or an underhand project, a plot” [emphasis mine].
“Scam” is “a trick, a ruse; a swindle, a racket”. In Hebrew, too, there are both neu-
tral and negative terms for the equivalent of “scheme”, but only negative terms,
tarmit and hona’a, for “scam” (Levy 1995). Yossi’s confusion between two simi-
lar-sounding lexical items in English and Gad’s misunderstanding were the initial
cause for the conflict. Gad blames Yossi almost entirely for it, conceding only that:
(14) This whole exchange didn’t bring the nicest things out of some people, you
for one, in the eyes of some, me, and of course the cobra who spat at me.
Putting Yossi first in the list of those behaving badly is unfair. Yossi’s only postings
were the initial question and its answer. In fact, Gad dominated the entire sequence,
with four postings before this one and seven in all. In the expression “In the eyes of
some, me” Gad concedes only that others thought his behavior “not nice”, thus
avoiding the first person pronoun “I”, the agent of responsibility for action. He
only uses it in “If you had used the word scam (…) I would not have responded as I
did”, thereby evading having to admit exactly how he responded.
Gad’s contrast between efroax (‘chick’) and “man-eating monster” grossly
minimizes his responsibility. Efroax connotes something small, harmless, cuddly,
adorable, bright yellow like the sun; “man-eating pre-historic dinosaur” connotes a
huge, frightening killer. While “chick-monster” is animate, “storm” and “tsunami”
refer to forces of nature, as if the trouble were a product of forces beyond the con-
trol of humans, that is to say, of these posters, especially himself.
Gad also deploys affiliative, conciliatory strategies to mitigate responsibility.
Again, he signs with his nickname, “Gadi”. Switching strategically to the Hebrew
“Shana Tova”, he wishes Yossi and the others a belated Happy New Year. Here,
overtly agonistic elements are absent, except for Gad’s reference to Irena as a
“cobra”.
9. Summing up
Six sets of factors fostered impoliteness in this sequence, some familiar, others per-
haps newly identified here.
9.3. Gender
Nine males but only five females participated. The primary flamers were both
male – Gad, the challenger, and Yossi, the victim. Three of the five insulters were
male, and a male, Gad, addressed the most insults to others and posted the most
(seven postings). Eighteen of the 24 postings were by males. This pattern is con-
sistent with aggressive, assertive masculinity and female subordination in Israeli
society (Lemish and Lahav 2004) and with findings that men participate more in
CMC and flame more than women (Aiken and Waller 2000; Herring 1994, 2003).
Agonistic/jocular elements of playfulness in Gad’s behavior were consistent with
Israeli macho masculinity. Herring’s assessment of why men flame more than
women and why some women participate fits here, too:
(…) men flame, at least in part, to regulate the social order, as self-appointed vigilantes
on the “virtual frontier”. Such behavior is rationalized within a male system of values
that assigns greater importance to freedom of expression and firmness of verbal action
than to possible consequences to the addressee’s face needs (…) All act under the in-
fluence of a larger culture in which confrontation and aggression are valorized in males
(…) flaming is “contagious” – normally polite participants, including women, can be
pushed to flame back when sufficiently provoked by another’s flames. (Herring 1994:
292)
9.4. Personality
Gad’s personality was a contributing factor. His practice of combining a serious
message with superfluous playful, provocative additions suggests a stronger than
usual tendency to provoke, even for a male.
Notes
1. I thank Shoshana Blum-Kulka, Tamar Katriel, Ayelet Kohn, Motti Neiger, Esther
Schely-Newman, Eli Salzberger, Fania Oz-Salzberger, Carmel Vaisman, the anonymous
reviewers, and the editors for their comments.
2. See Alonzo and Aiken (2004); Culman and Markus (1987); Joinson (1998); Kiesler,
Siegel, and McGuire (1984); Siegel, Dubrovsky, Kiesler, and McGuire (1986); Suler
(2004); Thompsen (1996); Thompsen and Foulger (1996); Turnage (2007).
3. See Baron (1998); Dery (1994); Dubrovsky, Kiesler, and Sethna (1991); Herring
(1994); Kiesler, Siegel, and McGuire (1984); Korenman and Wyatt (1996); Lea, O’Shea,
Fung, and Spears (1992); Parks and Floyd (1995); Thompsen (1996); Thompsen and
Foulger (1996).
4. In colloquial Arabic, dugri means ‘straight’, as in “straight road” (Katriel 1986: 11).
5. There is an English-language daily as well as English editions of some Hebrew papers.
6. The Hebrew term should be du-krav.
7. I follow the transliteration system of Berman and Armon-Lotem (1990) and Blum-
Kulka (1997).
8. Subscribers do not necessarily see all citations; one would have to scroll down to see all
cited material.
9. Gad is referring here to the American “Cold Stone Creamery” ice cream chain.
10. In ritual insults (Labov 1972), the utterance is not intended to hurt and is not experi-
enced as such; the recipient does not contest its truth value, retorting instead with an-
other stylized insult.
References
Abrahams, Roger
1973 Playing the dozens. In: Alan Dundes (ed.), Mother Wit from the Laughing Bar-
rel: Readings in the Interpretation of Afro-American Folklore, 295–309.
Englewood Cliffs, NJ: Prentice-Hall.
Aiken, Milam and Bennie Waller
2000 Flaming among first-time group support system users. Information and Man-
agement 37(2): 95–100.
Alonso, Andoni and Iñaki Arzoz
2006 Basque Cyberculture: From Digital Euskadi to Cybereuskalherria. Reno, NV:
Center for Basque Studies, University of Nevada.
Alonzo, Mei and Milam Aiken
2004 Flaming in electronic communication. Decision Support Systems 36(3):
205–213.
Androutsopoulos, Jannis K.
2007 Language choice and code-switching in German-based diasporic web forums.
In: Brenda Danet and Susan C. Herring (eds.), The Multilingual Internet: Lan-
guage, Culture, and Communication Online, 340–361. New York: Oxford Uni-
versity Press.
Austin, John L.
1975 How To Do Things with Words. Revised, edited by Marina Sbisà and J. O. Urm-
son. Oxford, UK: Clarendon Press.
Avgerinakou, Anthi
2003 Flaming in computer-mediated interaction. In: Colin B. Grant (ed.), Rethinking
Communicative Interaction: New Interdisciplinary Horizons, 273–293. Am-
sterdam/Philadelphia: John Benjamins.
Baron, Naomi S.
1998 Letters by phone or speech by other means: The linguistics of email. Language
and Communication 18: 133–170.
Baroz, Nitza Baharuzi (ed.)
2005 Pashkevilim: Modaot Kir uKrazot Pulmus baRexov haXaredi. [Pashkevilim:
Contentious Wall Posters on the Ultra-orthodox Street.] Tel Aviv/Jerusalem:
Ha’aretz Museum and Yad Ben Zvi.
Bauman, Richard
1992 Performance. In: Richard Bauman (ed.), Folklore, Cultural Performances, and
Popular Entertainments: A Communications-Centered Handbook, 41–49.
New York: Oxford University Press.
Berman, Ruth Aronson and Sharon Armon-Lotem
1990 Transcription and Coding Conventions: Language Acquisition Project. Work-
ing Paper no. 6. Tel Aviv: Tel Aviv University.
Black, Bob and Adam Parfrey (eds.)
1989 Rants and Incendiary Tracts: Voices of Desperate Illumination, 1558 – Pres-
ent. New York: Amok Press.
Blais, Dotan
2001 “Walla chat” – xevra, kehila, zehut: Mabat etnografi al tchat yisraeli. [“Walla
chat” – society, community, identity: An ethnographic view of Israeli chat.]
Kesher 30: 77–92.
Blondheim, Menahem and Shoshana Blum-Kulka
2001 Literacy, orality, television: Mediation and authenticity in Jewish conversa-
tional arguing 1–2000 C.E. The Communication Review 4: 511–540.
Blum-Kulka, Shoshana
1987 Indirectness and politeness in requests: Same or different? Journal of Prag-
matics 11(2): 131–146.
Blum-Kulka, Shoshana
1997 Dinner Talk: Cultural Patterns of Sociability and Socialization in Family Dis-
course. Mahwah, NJ: Lawrence Erlbaum.
Blum-Kulka, Shoshana, Menahem Blondheim, and Gonen Hacohen
2002 Traditions of dispute: From negotiations of Talmudic texts to the arena of
political discourse in the media. Journal of Pragmatics 34: 1569–1594.
Blum-Kulka, Shoshana, Brenda Danet, and Rimona Gherson
1985 The language of requesting in Israeli society. In: Joseph P. Forgas (ed.),
Language and Social Situations, 113–139. New York/Berlin: Springer-
Verlag.
Blum-Kulka, Shoshana, Juliane House, and Gabriele Kasper (eds.)
1989 Cross-cultural Pragmatics: Requests and Apologies. Norwood, NJ: Ablex.
Dewaele, Jean-Marc
2004a Blistering barnacles! What language do multilinguals swear in?! Estudios de
Sociolinguistica 5(1): 83–105.
Dewaele, Jean-Marc
2004b The emotional force of swearwords and taboo words in the speech of multi-
linguals. Journal of Multilingual and Multicultural Development 25(2–3):
204–222.
Doniach, Nakdimon S. and Ahuvia Kahane
1998 The Oxford English-Hebrew Dictionary. New York: Oxford University Press.
Dubrovsky, Vitaly J., Sara Kiesler, and Beheruz N. Sethna
1991 The equalization phenomenon: Status effects in computer-mediated and
face-to-face decision-making groups. Human-Computer Interaction 6(2):
119–146.
Edwards, Viv and Thomas J. Sienkewicz
1990 Oral Cultures Past and Present: Rappin’ and Homer. Oxford, UK: Basil
Blackwell.
Eelen, Gino
2001 A Critique of Politeness Theories. Manchester/Northampton: St. Jerome.
Fialkova, Larisa
2005 Emigrants from the FSU and the Russian-language Internet. Toronto Slavic
Quarterly 12. http://www.utoronto.ca/tsq/12/fialkova12.shtml
Goffman, Erving
1967 Interaction Ritual: Essays on Face-to-face Behavior. Garden City, NY: Anchor
Books.
Goffman, Erving
1971 Relations in Public. Harmondsworth: Penguin.
Goffman, Erving
1974 Frame Analysis: An Essay on the Organization of Experience. New York:
Harper and Row.
Goffman, Erving
1979 Footing. Semiotica 25: 1–29.
Golan, Galia
1997 Militarization and gender: The Israeli experience. Women’s Studies Inter-
national Forum 20(5): 581–586.
Grice, H. Paul
1975 Logic and conversation. In: Peter Cole and Jerry L. Morgan (eds.), Syntax and
Semantics 3: Speech Acts, 41–58. New York: Academic Press.
Hacohen, Gonen
2001 Jewish argumentation: Conversational features of two forms of arguments in
Israel. Unpublished M.A. thesis, The Hebrew University of Jerusalem. [In
Hebrew]
Halama, Alta Cools
1996 Flytes of fancy: Boasting and boasters from Beowulf to gangsta rap. Essays in
Medieval Studies 13: 81–93.
Halperin-Kaddari, Ruth
2004 Women in Israel: A State of Their Own. Philadelphia, PA: University of Penn-
sylvania Press.
Handelman, Don
1976 Play and ritual: Complementary frames of meta-communication. In: Antony J.
Chapman and Hugh C. Foot (eds.), It’s a Funny Thing, Humour, 185–192. Lon-
don: Pergamon.
Harrison, Sandra
2004 Subverting conversational repair in computer-mediated communication:
Pseudo repair and refusal to repair in a hostile email discussion. In: Mike Bay-
nam, Alice Deignan, and Goodith White (eds.), Applied Linguistics at the In-
terface, 63–77. British Studies in Applied Linguistics 19. London: BAAL
Equinox.
Harrison, Sandra
2007 Transgressions, miscommunication, and flames: Problematic incidents in
email discussions. In: Mia Consalvo and Caroline Haythornthwaite (eds.),
AOIR Internet Annual 4: 105–117. New York: Peter Lang.
Helmreich, William B.
1982 The World of the Yeshiva: An Intimate Portrait of Orthodox Jewry. New York/
London: Free Press and Collier Macmillan.
Herring, Susan C.
1994 Politeness in computer culture: Why women thank and men flame. In: Mary
Bucholtz, Anita C. Liang, Laurel Sutton, and Caitlin Hines (eds.), Cultural
Performances: Proceedings of the Third Berkeley Women and Language Con-
ference, 278–294. Berkeley, CA: Berkeley Women and Language Group.
Herring, Susan C.
2003 Gender and power in online communication. In: Janet Holmes and Miriam
Meyerhoff (eds.), The Handbook of Language and Gender, 202–228. Oxford,
UK: Blackwell.
Herzog, Hanna
2004 Women in Israeli society. In: Uzi Rebhun and Chaim I. Waxman (eds.), Jews in
Israel: Contemporary Social and Cultural Patterns, 195–220. Hanover, NH:
Brandeis University Press and University Press of New England.
Hinrichs, Lars
2006 Codeswitching on the Web: English and Jamaican Creole in E-Mail Communi-
cation. Amsterdam/Philadelphia: John Benjamins.
Hoffman, Eva
1989 Lost in Translation: A Life in a New Language. New York: Penguin.
Horowitz, Dan and Moshe Lissak
1990 Metsukot be-Utopiah: Yisrael – Xevrah beOmes-yeter. [Trouble in Utopia: Is-
rael, an Overburdened Society.] Tel Aviv: Am Oved.
Huddersfield University and Lancaster University
2006 Linguistic Impoliteness and Rudeness: Confrontation and Conflict in Dis-
course. Conference abstracts: http://www2.hud.ac.uk/mhm/english/research/
conferences/mhm-eng-conference-downloads/LIAR-2006-Abstracts.pdf
Izraeli, Dafna N.
1992 Culture, policy, and women in dual-earner families in Israel. In: Suzan Lewis,
Dafna N. Izraeli, and Helen M. Hootsmans (eds.), Dual-earner Families: In-
ternational Perspectives, 19–45. Thousand Oaks, CA: Sage.
seliti and Jane Sutherland (eds.), Discourse Analysis and Gender Identity,
69–89. Amsterdam/Philadelphia: John Benjamins.
Myerhoff, Barbara G.
1994 Number Our Days: Culture and Community among Elderly Jews in an Ameri-
can Ghetto. New York: Meridian.
Oliveira, Sandi Michele de
2003 Breaking conversational norms on a Portuguese users network: Men as adjudi-
cators of politeness? Journal of Computer-Mediated Communication 9(1).
http://jcmc.indiana.edu/vol9/issue1/
Olshtain, Elite and Andrew D. Cohen
1983 Apology: A speech act set. In: Nessa Wolfson and Elliot Judd (eds.), Sociol-
inguistics and Language Acquisition,18–35. Rowley, MA: Newbury House.
O’Sullivan, Patrick B. and Andrew J. Flanagin
2003 Reconceptualizing “flaming” and other problematic messages. New Media &
Society 5(1): 69–94.
Owen, Marion
1983 Apologies and Remedial Interchanges: A Study of Language Use in Social In-
teraction. Berlin/New York: Mouton.
Panyametheekul, Siriporn and Susan C. Herring
2003 Gender and turn allocation in a Thai chat room. Journal of Computer-Mediated
Communication 9(1). http://jcmc.indiana.edu/vol9/issue1/panya_herring.html
Paolillo, John C.
1996 Language choice on soc.culture.punjab. Electronic Journal of Communication
6(3). http://www.cios.org/EJCPUBLIC/006/3/006312.HTML
Parks, Malcolm R. and Keri Floyd
1995 Making friends in cyberspace. Journal of Computer-Mediated Communication
1(4). http://jcmc.indiana.edu/vol1/issue4/
Parks, Ward
1986 Flyting, sounding, debate: Three verbal contest genres. Poetics Today 7:
439–458.
Parks, Ward
1990 Verbal Dueling in Heroic Narrative: The Homeric and Old English Traditions.
Princeton, NJ: Princeton University Press.
Pavlenko, Anna
2005 Emotions and Multilingualism. Cambridge/New York: Cambridge University
Press.
Reinig, Bruce A. and Roberto J. Mejias
2004 The effects of national culture and anonymity on flaming and criticalness in
GSS-supported discussions. Small Group Research 35(6): 698–723.
Rosenthal, Ruvik
2005 Milon hasleng hamekif. [The Comprehensive Slang Dictionary.] Tel Aviv:
Keter.
Russell, Mark (ed.)
1997 Out of Character: Rants, Raves, and Monologues from Today’s Top Perform-
ance Artists. New York: Bantam Books.
Salzberger, Eli and Fania Oz-Salzberger
2006 Masoret xofesh habitui b’yisrael. [The tradition of freedom of speech in Israel.]
In: Michael Birnhack (ed.), Sheket, hem medabrim: Hatarbut hamishpatit shel
xofesh habitui b’yisrael [Quiet, Someone is Speaking! The Legal Culture of
Free Speech in Israel], 27–69. Tel Aviv: Ramot Publishers.
Schiffrin, Deborah
1984 Jewish argument as sociability. Language in Society 13: 311–335.
Searle, John
1976 A classification of illocutionary acts. Language in Society 5(1): 1–23.
Severinson Eklundh, Kerstin
2010 To quote or not to quote: Setting the context for computer-mediated dialogues.
Language@Internet 7, article 5.
http://www.languageatinternet.org/articles/2010/2665
Shakhar, Il-Il
2005 Emek haSilicon b’yadenu. [Silicon valley is in our hands.] Maariv, September
22. http://www.nrg.co.il/online/16/ART/987/062.html
Siegel, Jane, V. Dubrovsky, Sara Kiesler, and Timothy W. McGuire
1986 Group processes in computer-mediated communication. Organizational Beha-
vior and Human Decision Processes 37(2): 157–187.
Spolsky, Bernard
1997 Multilingualism in Israel. Annual Review of Applied Linguistics 17: 138–150.
Spolsky, Bernard and Elana G. Shohamy
1999 The Languages of Israel: Policy, Ideology, and Practice. Clevedon/Buffalo:
Multilingual Matters.
Statistics, Central Bureau of
2004 On the 56th anniversary of Israel’s independence the population of Israel
numbers 6.8 million. April 25. http://www.cbs.gov.il/hodaot2004/01_04_98e.htm
Suler, John
2004 The disinhibition effect. CyberPsychology & Behavior 7: 321–326.
Tannen, Deborah
1981 New York Jewish conversational style. International Journal of the Sociology
of Language 30: 133–149.
Tannen, Deborah
1984 Conversational Style: Analyzing Talk among Friends. Norwood, NJ: Ablex.
Tannen, Deborah (ed.)
1993 Framing in Discourse. New York: Oxford University Press.
Tanskanen, Sanna-Kaisa
2001 Avoiding conflict in computer-mediated discussions, or, fear of flaming. In:
Risto Hiltunen, Keith Battarbee, Matti Peikola, and Sanna-Kaisa Tanskanen
(eds.), English in Zigs and Zags, 227–242. Anglicana Turkuensia 23. Turku:
University of Turku.
Thompsen, Philip A.
1996 What’s fueling the flames in cyberspace? A social influence model. In: Lance
Strate, Ron Jacobson, and Stephanie B. Gibson (eds.), Communication and
Cyberspace: Social Interaction in an Electronic Environment, 297–315. Cress-
kill, NJ: Hampton Press.
Thompsen, Philip A. and Davis A. Foulger
1996 Effects of pictographs and quoting on flaming in electronic mail. Computers in
Human Behavior 12(2): 225–243.
Turnage, Anna K.
2007 Email flaming behaviors and organizational conflict. Journal of Computer-
Mediated Communication 13(1), article 3. http://www.blackwell-synergy.com/
doi/full/10.1111/j.1083–6101.2007.00385.x
Vrooman, Steven S.
2002 The art of invective: Performing identity in cyberspace. New Media & Society
4(1): 51–70.
Weimann, Gabriel
2000 Gender differences in Israeli television commercials. Megamot 40(3):
466–485. [In Hebrew]
Wex, Michael
2005 Born to Kvetch: Yiddish Language and Culture in All its Moods. New York:
St. Martin’s Press.
Wright, Sue (ed.)
2004 Multilingualism on the Internet. International Journal on Multicultural So-
cieties 6. http://www.unesco.org/new/en/social-and-human-sciences/resources/
periodicals/diversities/past-issues/vol-6-no-1-2004/
Yishai, Yael
1997 Between the Flag and the Banner: Women in Israeli Politics. Albany, NY:
State University of New York Press.
Zilkha, Avraham
2002 Modern English-Hebrew Dictionary. New Haven, CT: Yale University Press.
1. Introduction
guage/technology relations, and some of its key categories are conceived and best
operate within a monolingual frame, such as the position of computer-mediated
discourse between spoken and written language. However, CS defies easy classifi-
cations based on media factors alone, and it requires an emphasis on the interre-
lation of medium and social/situation factors (Herring 2007). Moreover, CMC as a
discourse field challenges the assumption that spoken face-to-face interaction is
the essential site of code-switching.
Based on research literature on several languages,2 the aim of this chapter is to
organise the available research evidence, identify commonly asked and still un-
tapped questions, and pinpoint limitations of present scholarship. The chapter does
not claim to offer a comprehensive overview of forms and functions of CS on the
Internet; rather, it can only outline how CS has been studied in the languages and
sociolinguistic settings I have been able to identify in the literature (section 4).
Still, the available literature offers ample evidence that CMC is a site for the mean-
ingful use of language alternation, and a critical synthesis of available research can
offer insights into what are promising perspectives for further research, as well as
what methods have been mainly used.
The organisation of this chapter is as follows: After a brief outline of code-
switching research frameworks (section 2), I distinguish between conversational
and non-conversational CS in CMC and delimit both from other patterns of multi-
lingualism online (section 3). The subsequent two sections provide an overview of
research (section 4) and discuss the different types of settings in which CS in CMC
has been studied (section 5). The next section turns to forms and functions of CS
online (sections 6). The chapter concludes with a discussion of CMC as a new do-
main for CS (section 7) and an outlook for future research (section 8).
2. Code-switching frameworks
structural descriptions are not entirely absent from the extant literature, its
main aim has been to understand the pragmatic functions, social purposes, and
interactional dynamics of CS online.
– The correspondence of online written CS to its offline spoken counterpart is a
common concern, but it is also a contested issue, as will become obvious in this
chapter.
In terms of frameworks, the literature contains elements of the “three most in-
fluential contributions to theory in the sociolinguistic branch of CS studies” (Hin-
richs 2006: 28): The first is the markedness model of Carol Myers-Scotton (1993),
in particular her concepts of code-switching as a marked (i.e., unexpected, uncon-
ventional) or unmarked (expected) choice. Second are concepts introduced by John
J. Gumperz (1982), such as the distinction between situational and metaphorical
CS, the distinction between “we-code” and “they-code”, his classification of dis-
course functions of conversational CS, and the notion of CS as a contextualization
cue. Thirdly, researchers also draw on the conversation-analytic framework for the
study of bilingual interaction by Peter Auer (1995, 1998b, 1999, 2000), which
builds on and develops some of Gumperz’ ideas.3 Other repeatedly used concepts
include the syntactic distinction between inter- and intra-sentential CS; concepts
from pragmatics such as politeness, face, and interpersonal alignment (e.g., Geor-
gakopoulou 1997); Myers-Scotton’s notion of matrix language; and Auer’s notion
of base language, which refers to the backdrop against which switches to another
language or dialect become meaningful.
In terms of discourse functions of CS, the classifications by Gumperz (1982)
and Auer are widely used in the literature, e.g., by Androutsopoulos (2006a,
2007a), Sebba (2003), Androutsopoulos and Hinnenkamp (2001), and Paolillo
(1996, 2011). Both conceive of CS as a contextualization cue, i.e., a resource used
by participants to frame their interpretations of what is being said. In a nutshell,
Gumperz’ categories include: switching for reported speech; addressee specifi-
cation; clarification, emphasis, expressivity; message qualification (e.g., separat-
ing facts from comment); and contrasting personal with objective viewpoints.
Auer distinguishes between preference- (or participant-)related and discourse-re-
lated CS. The first comprises switches that suit the speaker’s or addressee’s pref-
erence, as well as instances of language negotiation between the interlocutors.4
Discourse-related switching in a conversational episode “contributes to the organ-
ization of discourse in that particular episode” (Auer 1995: 125). Its subtypes par-
tially overlap with those by Gumperz, one important addition being the focus on
CS as a device for the internal organization of conversational turns. How these cat-
egories have been applied to CS in CMC is discussed in the overview of research
findings below (section 6).
The limitations of a conversation-analytic approach with respect to CMC data
are well discussed in the CMC literature (e.g., Beißwenger 2008; Herring 1999).
CMC technologies rule out one key mechanism of conversational organisation, the
turn-taking system; more generally, the lack of visual channels – and, in asyn-
chronous CMC, the temporal gap between contributions – means that important di-
mensions of the interactional co-construction of meaning are altered or restricted.
However, these restrictions do not rule out the sequential organization of com-
puter-mediated discourse,5 which can be studied with conversation analytic cat-
egories (see, e.g., the chapters in this volume by Jacobs and Garcia, Markman, and
Skovholt and Svennevig). Furthermore, CMC research has established that users
develop creative procedures to cope with these limitations, including the usage
of specific turn-taking signals and linguistic innovations such as emoticons and
laughter acronyms. Related to this, and specifically relevant to CS, is Georgako-
poulou’s (1997) suggestion that the lack of ordinary contextualization cues due to
the absence of the visual channel “results in an increased reliance on code-centered
contextualization cuing, which would be otherwise delegated to different signals”
(158). In other words, CMC interlocutors use code-switching, style shifting, and
other manipulations of written signs in order to accomplish pragmatic work that
would be accomplished by phonological variation, prosody, gaze, posture, and other
cues in ordinary spoken conversation. This establishes a productive theoretical link
between linguistic choices, communicative practices, and media affordances.
These are elements of a basic theoretical vocabulary which CMC researchers
appropriate, articulate, and apply in different ways. For example, Hinrichs (2006)
combines ideas from all three frameworks with categories from Creole linguistics.
Leppänen (2007) draws on the four types of language alternation in Auer’s frame-
work (i.e., insertional switching, insertional mixing, alternational switching, and
alternational mixing) to examine alternation between Finnish and English in a
range of digital genres. Androutsopoulos (2006a, 2007a) defines the discussion
thread of web forums as the equivalent of a conversational episode and therefore as
the level at which to determine the base language of discussion (cf. Auer 2000),
against which the directionality of switches is examined.
Relay Chat, web-based chat systems, or Instant Messaging, which some re-
searchers identify as the closest approximations of spoken conversation. However,
a considerable part of the literature examines CS in asynchronous (dyadic or multi-
party) CMD, in particular email, newsgroups, web forums, and texting. Still other
researchers have located bilingual discourse in edited genres aimed at a reading
audience, such as weblogs and fan fiction. These, in turn, are reminiscent of more
traditional written CS, which is a practice known at least since medieval poetry and
ranges from fictional representations of conversational dialogue (as in novels or
stageplay) to diary writing and newspaper discourse (Androutsopoulos 2007b;
Callahan 2004). The approach advocated here is an inclusive one. I suggest that
there is no a priori reason not to consider blogs or social networking profile pages
as sites of bilingual discourse. Nevertheless, a terminological and analytical dis-
tinction is required between two main types of CS in CMC: For convenience, I
shall call them “conversational” (dyadic or multiparty, synchronous or asynchron-
ous) and “non-conversational” (edited and published by a single author), bearing in
mind the caveats mentioned earlier.
With these distinctions in mind, CS will now be distinguished from other as-
pects of multilingual CMD. The aim in doing so is to determine when the coexist-
ence of more than one language in CMD constitutes CS and to identify the dis-
course units, or equivalents of a “speech exchange”, in which CS can be located.6 I
distinguish CS from four other patterns of multilingualism in CMC:
a) The “multilingual Internet” as a whole
b) The coexistence of different languages on a web page or thread
c) Language choices for emblems
d) Sequential language choices lacking a dialogical interrelation
The first pattern refers to the multilingualism that emerges from the coexist-
ence of different websites, channels, forums, etc. on the web. For instance, Internet
Relay Chat (IRC) as an entire system is massively multilingual, and so are plat-
forms like flickr and blogger.com in their entirety. While this level is sometimes re-
ferred to in discussions of the multilingual Internet (see, e.g., some chapters in
Danet and Herring 2007a), CS must be located on the more concrete level of indi-
vidual web pages, discussion threads, or chat channel sessions. At this level, it is
necessary to single out the multilingualism that emerges from the coexistence of
different discourse units on a single web page. Contemporary websites are
composed from textual units of diverse kinds and origins – editorial content, user-
generated content, advertisements, graphic-designed banners, user comments –
which are adjacent in screen space. Multilingual surfaces emerge when such adjac-
ent units “speak”, or are cast in, different languages or dialects as a result of dif-
ferent purposes, audiences, or production processes. Consider flickr pages, for
example, where headlines, banners tags, and comments are not necessarily in the
same language (Lee and Barton 2011). Consider media-sharing sites, where the
language of posted items is often different from that of comments, and social net-
work profile pages, where different “friends” may contribute wall posts in different
languages. These are multilingual or indeed heteroglossic discourse spaces (An-
droutsopoulos 2011), but they do not automatically constitute instances of CS. To
the extent that their constituents differ in terms of authorship and production pro-
cess, they often cannot be conceived of as part of one “episode”; what holds them
together is their spatial coexistence in product and reception and not their dialogic
orientation to each other. Indeed, the units that make up multilingual web surfaces
are often monolingual in themselves. However, nothing prevents some of these
units from containing CS. Think of an online newspaper where comments to news
items come in different varieties (e.g., standard and dialect) or languages (e.g., the
national and a minority language), thus constituting candidates for CS.
Still at the level of single web pages or threads, the third distinction is between
CS and language choices for emblems, i.e., textual units that identify and represent
individual or institutional actors in CMC. Emblems include website names and
screen names, slogans, user signatures, and navigation bars: They are usually
graphically designed and placed on website layout. There is evidence that emblem-
atic language choices can extend the multilingual make-up of a website or dis-
cussion forum by introducing linguistic resources that are not regularly used in on-
going user discussions or editorial copy (Androutsopoulos 2006a, 2007b). This
strategic allocation of languages leads to a sort of emblematic bilingualism, which
does not challenge the dominant language in terms of informational load but selects
another code as relevant to the identity of an institution or individual. Important as
these processes may be to multilingual discourse on the web, they are not instances
of CS (to the extent that these textual units do not include CS in themselves), even
though they may coexist with instances of CS: Think of a blog that uses emblems in
order to make a minority language visible, while its entries and their comments in-
clude CS between that minority and the respective majority language.
Finally, conversational CMD modes need to be scrutinised in order to assess if
the linguistic diversity they host constitutes CS. Consider IRC, a model case of in-
teractive written discourse. Even if IRC is carried out in a language other than Eng-
lish, the system-generated messages that announce users who are entering or leav-
ing a channel are automatically cast in English (Siebenhaar 2005, 2006). While
they contribute to the multilingual make-up of channels, these automated messages
are not part of participants’ code choices and therefore need to be excluded from
CS analysis (as Siebenhaar explicitly does). In a similar vein, we may ask whether
the lists of comments that respond to a “spectacle” or “prompt” in Web 2.0 environ-
ments (a blog post, photo, video, song, etc.) qualify as instances of CS. Comments
may in principle be posted in any language, and even though the language of the
“prompt” will often make a particular language relevant, it is not uncommon to find
comments in different languages. However, whether these individual code choices
by different commenters are sequentially related to each other needs to be demon-
strated rather than assumed a priori. My observations suggest that comments often
respond to the “prompt” rather than to other posts, and while they all contribute to
the multilingual character of a chain of comments, they are not forcibly instances of
CS (although they can draw on CS, of course, in their own internal organisation).
In conclusion, even though all CS entails a juxtaposition of linguistic codes, as
Gumperz’ seminal definition puts it (see section 1), there is good reason to adapt a
restrictive view of what juxtapositions will constitute CS in CMD. Multilingual
CMD environments are shaped at different levels by contrastive language choices
which are motivated and meaningful, but for these contrasts to qualify as CS, evi-
dence is required that they are in some way dialogically interrelated by responding
to previous, and contextualizing subsequent, contributions.
The third column lists the languages that are (most) relevant to each study.8 A
number of sociolinguistic constellations are evident here, between majority and
minority (heritage, immigrant, community) language, between national language
and English, between a Creole and its lexifier (Hinrichs 2006), standard and dia-
lect, and standard and stylized vernacular speech (e.g., “Ali G language”, Sebba
2003). Code-switching between varieties of one language is reported from the
Greek- and German-speaking areas and studied along the same functional lines as
bilingual CS. For the sake of overview, this table downplays the few documented
instances of “polylingual languaging” (Hinnenkamp 2008; Jørgensen 2008; Tsi-
plakou 2009), which involves the (playful) use of bits and pieces from different
languages, language varieties, or styles (see sections 6, 7).
The user groups and social settings that are examined in each study are catego-
rised in the fourth column. The label “ethnic minority” brings together situations in
which an immigrant or diasporic group uses a minority and a majority language.
Other labels foreground online practices by young people, youth culture groups, or
local communities, in which relations between national language and English, or
standard and dialect, have been examined. Other documented cases involve multi-
lingual academics or professionals, small language communities (Sperlich 2005),
and a message board devoted to a Creole-speaking comedian (Sebba 2003). The
overview does not mean to suggest rigid constraints between language pairs and
social settings. Some language pairs (especially those involving English) are in-
stantiated in different social settings, with distinct CS patterns in each case. For
example, English/Creole CS shows different patterns in emails among Jamaican
students and on forums by diasporic Jamaicans (Hinrichs 2006). Likewise, CS be-
tween national language and English CS is markedly different in youth-culture
contexts and among elite bilingual expatriates (see section 5).
The rightmost column features a broad distinction between qualitative and
quantitative methods of analysis. The “qualitative” label comprises methods from
conversation, discourse, narrative, or style analysis, which have been used for the
study of both conversational and non-conversational CMD. Such research often
involves elements of online ethnography, whereby researchers work with data
from social networks they themselves belong to (as with Georgakopoulou 1997
and Tsiplakou 2009), and focuses on functions rather than structures of CS.
“Quantitative” encompasses quantifications based on questionnaire data (Gold-
barg 2009; Tsiplakou 2009) or coding of textual data (as with Paolillo 2001; Sie-
benhaar 2008). What is remarkable is the frequency of mixed method approaches,
which combine a “bird’s eye view” of the distribution of languages over a large
data set with a detailed view of local processes of switching and mixing. Sieben-
haar (2008) adds an intermediate level, “windowed or moving average analysis”,
in which quantification zooms in to slices of time or channel activity as opposed
to whole channel comparisons. Leppänen and Peuronen (2012) argue that the
transfer of frameworks developed for the study of spoken language and interac-
However, such a bottom line cannot be taken for granted in public CMD. There
is evidence to suggest that public CMD spaces create favourable conditions for the
“functional marginalization” (Paolillo 1996) of minority and heritage languages on
the Internet. Digital networks that explicitly focus on a shared ethnolinguistic iden-
tity may in fact be used by ethnolinguistically heterogeneous individuals, leading
to an orientation to the majority language as a common denominator. The reasons
for such heterogeneity may be diverse: The ethnic communities that use CMD
spaces are often undergoing intergenerational language shift, some of their
members may not be fluent in the heritage language anymore, or users may be eth-
nically mixed, including members of the ethnic majority group. In such cases, the
communicative aim of reaching as many audience members as possible may over-
ride the preference for the heritage language, and at the same time the wish to index
ethnolinguistic identity may lead to patterns of formulaic and emblematic code-
switching from the dominant, majority language into the heritage language, as de-
scribed by Paolillo and Androutsopoulos, among others. However, it seems im-
portant not to generalise this finding, as chat channels with a predominant use of
migrant/ethnic languages have been studied as well (e.g., Androutsopoulos and
Hinnenkamp 2001; Hinnenkamp 2008; Tsaliki 2003).
By contrast, private CMD provides different conditions for recipient design (to
use a term from conversation analysis), as participants can rely on a much greater
inferential potential; i.e., they can count on their code-choices and switches being
understood by virtue of common background knowledge and common practices.
Speakers have more leeway to explore playfully the associative potential of lan-
guage, dialects, and styles in their shared repertoire (Tsiplakou 2009).
quests, slogans, etc.) are inserted into the base national or majority language. Their
choice is often indexical to the groups’ lifestyle orientations, including stylized
representations of vernacular “Englishes” (Androutsopoulos 2004, 2007b).
Research from Finland offers evidence for an even broader range of online bi-
lingual practices involving English in Finnish youth cultures (Leppänen 2007;
Leppänen et al. 2009). One such practice is alternational CS in fan fiction (i.e.,
stories written by fans about popular fiction characters or settings), whereby Finn-
ish and English are used for narrative and reported speech, respectively, so that
neither constitutes the predominant language of discourse. In another digital genre,
diary weblogs, English is the matrix language into which Finnish cultural key-
words and expressions are inserted. In all these cases, there is no evidence that this
usage mirrors corresponding bilingual styles in face-to-face communication.
Rather, it is CMD that enables these bilingual practices in the first place. In addi-
tion to these Finnish cases, there is some evidence of CMD from what one might
call elite bilinguals (i.e., academics, white-collar professionals) with an L1 other
than English, whose private email communication shows complex patterns of
switching and mixing (Georgakopoulou 1997; Tsiplakou 2009, both on L1 speak-
ers of Greek).
The social conditions of CS online may of course be more complex that that.
For instance, the members of fan communities within one nation-state are often of
ethnically diverse origins, and the cultural practices they orient to may originate
beyond the English-speaking world. One anecdotal example is the web forum of a
local network of salsa fans in northern Germany (based on unpublished material),
which includes native speakers of German, Spanish, and various other migrant lan-
guages. Regardless of their origin, forum members orient to Spanish as a symbol of
their common cultural practice, and the usage of Spanish in this forum is often for-
mulaic, resembling the patterns of minimal bilingualism that are often identified
with English. Spanish is also used among Spanish-speaking participants, and so
are migrant languages (Turkish, Polish) among their respective speakers. The out-
come is a complex polylingual space, in which the majority language, the language
linked to the particular cultural practice, and immigrant languages all find their
place.
prefer local IRC rooms and use their dialect as the default code, whereas middle-
aged users prefer a cross-regional flirt chat, in which they use a considerable
amount of standard German as well. Paolillo (2001) and Ziegler (2005) provide
evidence for the impact of institutional roles within Internet culture on CS. Paolillo
finds that operators of the IRC channel #india switch less into Hindi than periph-
eral users. As Paolillo points out, an identity-based hypothesis would predict the
opposite, i.e., that core members of the IRC network use the heritage language to a
higher degree. Ziegler (2005) finds that in the local IRC channel of a German city,
#mannheim, channel operators use higher amounts of the local city dialect, thus CS
from and to standard German, than ordinary chatters. Paolillo offers a pragmatic
explanation of his finding based on an attention-based hypothesis: U.S.-born IRC
users of Indian descent display ethnic affiliation by other means than heritage lan-
guage, and it is peripheral (rather than core) members of the chat community who
code-switch in order to attract attention. Tsiplakou (2009) presents findings from a
questionnaire study that aimed at identifying variables predicting the degree of CS
in email (between Greek and English, among native Greeks), one such variable
being the use of English at home.
6. Code-switching patterns
CS in CMC is not confined to just a few typical patterns of usage. In terms of struc-
ture, reported patterns of usage range from a few formulaic switches to dense,
multilingual code-mixing and “polylingual languaging” (sections 6, 7). A restric-
tion to formulaic CS has been reported for public CMD in some ethnic commu-
nities (see work by Paolillo and Androutsopoulos), for English/Jamaican Creole in
personal emails (Hinrichs 2006), and for national language/English in youth-cul-
tural contexts (see section 5.3). Bilingual code-mixing has been reported for a
number of languages and modes, including personal emails among fellow aca-
demics in Cyprus (Tsiplakou 2009), chatting among youth of Turkish descent in
Germany (Hinnenkamp 2008), and “mixed messages characterized by English-
isiXhosa code-mixing and code-switching” in texting among South African youths
(Deumert and Masinyana 2008: 137). Deumert and Masinyana (2008) claim that
these linguistically mixed messages mirror code-mixing as the unmarked choice in
the spoken vernacular usage of these youths. The percentage of mixed text mess-
ages in their data is 23 % (in a corpus of 312 messages), and they are further dif-
ferentiated by communicative purpose. Code-mixing occurs in text messages on
social arrangements and information exchange, but less so in romantic messages.
In terms of discourse functions, the literature drawing on widely accepted clas-
sifications of conversational CS (see section 2) has produced evidence for a
number of discourse functions of CS in CMC.9 These include:
Some of these functions – especially (a) and (b) – have been found to favor a sus-
tained use of minority or migrant languages, while for others, the pragmatic effect
is created through the situated contrast between the codes involved (Hinrichs 2006
provides an extensive discussion of this distinction).
While the comparability of CS in CMC with general discourse functions of
conversational CS is thus in principle firmly established across languages, modes,
and social settings, individual manifestations of CS in CMD data may be difficult
to categorize, and switching and mixing may co-occur in the discourse of one user
or community. A few examples from my own research on predominantly German-
speaking diasporic forums are presented below to illustrate these points (Androut-
sopoulos 2006a, b, 2007a). The following three examples come from an Indian,
Persian, and Greek web discussion forum, respectively, and exemplify three vari-
ants of CS within a post:10
(1) Excerpt from the Indian forum, theinder.net (base language is German,
English italicized)
im westen ist es auch tradition jungfräulich in die ehe zu treten!!! das tun auch
einige (bsp. spanierinnen, italienerinnen, etc.)! wieso wird immer der westen
für alles verantwortlich gemacht??? is there no gravity in indian brains?
[‘in the west, no sex before marriage is also a tradition!!! And some stick to it
as well (e.g., Spaniards, Italians etc.)! Why is the West always being blamed for
everything??? is there no gravity in indian brains?’]
(2) Excerpt from the Iranian forum, iran-now.de (base language is Persian,
German italicized)
Bare B. juuuuuuuuun, kamelan hagh dari. in marda ham khasisan, ham bima-
zan wa ham gedan!!! Genau das dachte ich mir auch, chon iran-klick waghat
klick mikhado sonst nix, aber neeeeeeeeeeeee A. agha kann uns ja net ghalbesh
az ma beporseh. ghorbuuuuuuuuuuuune harfat khanumi. P. e immer noch
ghamginiani
[‘That’s true, dear B, you’re absolutely right. These men are stingy, boring, and
quite the poor creatures!!! That’s exactly what I thought, because on iran-klick
you just need to click, nothing else, but nooo Mister A. just can’t ask in ad-
vance. I love you for your words, my dear. P is still sad’]
(3) Excerpt from the Greek forum, greex.net (base language is Greek, German
italicized)
edo iparxi pollous ellines apo tin makedonia epidis i wirtschaftliche lage tous
den einai kali … palia i makedoni itane plousioi … eftiaxnan gounes ktlp ala
tora pige i wirtschaft me tis gounes den bach runter
[‘there are many greeks from macedonia here but their financial situation is not
good … macedonians were rich in the past, they were trading with furs, but
now the fur business is going down the drain’]
In these and other forums examined in that study, ethnic minority languages
mainly occur in isolated, insertional switches. In Excerpt 1, taken from a dis-
cussion thread on premarital sex, the concluding switch into English serves to ac-
centuate the writer’s critical conclusion and sets it off from the preceding argumen-
tation. This type of bilingual discourse is quite common in these forums, and
several types of discourse functions listed above are instantiated in this insertional
manner. In Excerpt 2, the first instance of German can be analysed as a switch that
marks a change of perspective from the previous evaluation of “these men” to the
writer’s own views – “That’s exactly what I thought”. However, this line of inter-
pretation becomes increasingly difficult as the post unfolds, and no contextualizing
function can be identified for the penultimate adverbial phrase in German, which is
more typical for code-mixing. In Excerpt (3), whose base language is Greek, the
three German phrases (financial situation, business, down the drain) have neither
any obvious discourse function nor do they serve a referential necessity. In both
cases, a Greek equivalent would presumably have been readily available, and the
writer seems to select German lexical items and idioms for reasons of habit or con-
venience. The last instance of German in this excerpt is an idiom (den Bach run-
tergehen ‘go down the drain’), and the writer expresses the finite verb in Greek and
the idiomatic phrase in German. In Auer’s discourse-functional framework (see
section 2), such cases are also typically classified as code-mixing.
Working with classifications of discourse functions provides an initial over-
view of patterns of CS in a CMC environment and a useful point of entry for
exploratory research. Cumulatively, analyses along these lines offer valuable evi-
dence for the regularity and conventionality of CS online, as well as for its func-
tional similarity to CS in other discourse environments, and thereby contribute to
its normalization. However, too heavy a reliance on classifications also entails the
risk of reducing analysis to a simple “category check”, which disconnects CS from
the conversational activity in which it is embedded and may result in a decontex-
tualized listing of CS instances. That danger can be reduced by ethnographic
knowledge of the population and digital platform under study and by detailed se-
quential analyses that take into account “the place within the interactional episode
in which languages alternate” (Auer 1998b: 3), the way switches align to previous
code choices of other speakers, and the way they index participants’ background
knowledge.
More specifically, the functional analysis of CS in CMC needs to transcend the
level of single turns or posts and examine the sequential organisation of code-
switching within threads of dialogically related posts or messages. CS is embedded
in the “polylogues” (Marcoccia 2004) that unfold on spaces of public, asynchron-
ous, thematically focused discourse, and makes use of the specific affordances pro-
vided by these spaces (see also discussion in Georgakopoulou 1997; Siebenhaar
2006). In the case of the Germany-based diaspora forums, switches within a post
are a resource for responding to different addressees and engaging with different
strands of a discussion thread. Users sometimes start in one language, quote from
previous posts which are cast in a different language, respond to that quote in its
language choice, and then return to their original code choice (Androutsopoulos
2006a, b, 2007a).
In addition, a pragmatically informed micro-analysis of CS in CMC will aim to
examine “how, within frameworks of generic assumptions and expectations, speech
communities draw upon their linguistic resources in order to maximize the effec-
tiveness and functionality of their communication” (Georgakopoulou (1997: 160).
In such an analysis it is possible to identify how different codes in a group’s usage
take on pragmatic functions and identity values, which cannot be assumed a priori
based on the wider cultural associations of these languages. The use of linguistic
heterogeneity to index social identities is a key issue in much of the work reviewed
here. Some researchers draw on Gumperz’ distinction between “we-code” and
“they-code”, which was originally equated with minority and mainstream language,
respectively. This works well when a few salient instances of a code are used as a
means to signal the ethnic identity associated with that code, and such “we-code”
signals reported repeatedly include greetings, openings, closing, slogans, and the
like (Androutsopoulos 2006a; Hinrichs 2006; Paolillo 1996). This approach is com-
plemented by a more fluid and dynamic understanding of language/identity re-
lations, in which “we”/“they” contrasts are locally constructed in discourse. In dias-
pora forum discussions, for example, the “we-code” is not always the minority
language, and the “they-code” is not always the mainstream language. Writers may
switch into their heritage language to index traditionalist views they distance them-
selves from, or, in another case, switch into the majority language to emphasize
their own multicultural outlook (see Androutsopoulos 2006a; Hinrichs 2006).
Based on the preceding discussion, we are now in a position to assess the relation
of CS in CMC to spoken conversational CS and written CS. The reviewed litera-
ture does not offer a generally agreed-upon position on this, and its suggestions
mirror to some extent the wider discussion on spoken and written aspects of lan-
guage on the Internet. While CS in CMC obviously qualifies as written in terms of
the written representation of linguistic signs, it also bears resemblance to spoken
conversational CS, most obviously in terms of its dialogic context and its discourse
functions (see section 6). From the perspective of contemporary language-focused
CMC scholarship, however, both a dichotomy of that sort and the homogenisation
of various CMD genres and practices as “language on the Internet” seem rather
crude. CMC is generally viewed as a heterogeneous domain of discourse, in which
traditional dichotomies between written and spoken, private and public, immediate
and mediated discourse, are blurred.
Theorizing CS in CMC needs to take the specific pragmatic and social condi-
tions of written language use in digital media more systematically into account.
Writing in networked digital media is different from other types of written dis-
course in a number of ways: It is dialogical, i.e., oriented to particular addressees,
and often embedded in multiparty conversational sequences; it also is often ver-
nacular, i.e., located outside of educational, professional, and other institutions;
and it is often simultaneously used together with other semiotic resources. Taking
these properties of digital writing into account when studying CS in CMC will con-
tribute to a deconstruction of spoken/written dichotomies and to a move beyond
the assumption that only spoken conversational CS constitutes “authentic” CS and
therefore sets the benchmark against which CS in CMC ought to be assessed. Such
expectations are familiar from other domains of written CS, notably in fiction,
where CS “authenticity” is assessed in terms of fictional settings, characters, and
the thematic content of a novel (Callahan 2004: 99–111). These criteria no doubt
operate, albeit implicitly, when researchers assess the ethnic or diasporic “authen-
ticity” of a web forum or chat channel. Notions of authenticity are also reproduced
within CMC, for example in the assumption that only synchronous CMC modes
will host “real” CS (see sections 3, 5). An alternative and, I would suggest, more
productive approach would be to ask how CS is used as a pragmatic resource under
the specific conditions of CMD, and how specific conditions of written online dis-
course can give rise to distinct CS practices. Two relevant points are touched upon
here: planning and the semiotics of writing itself.
Planning focuses on the difference between the immediate production of
speech in a spoken conversational setting and the gaps between production and re-
ception that are inherent in CMC, especially its asynchronous modes. The avail-
ability of planning time and the lack of visual cues are at the core of the assumption
that CMC is characterised by code-centered contextualization cues (Georgakopou-
lou 1997, 2003). However, the assumption that a “higher level of consciousness
[…] seems inevitable in producing written CS” (Dorleijn and Nortier 2009: 131)
should not be generalised prematurely. It is well known that CMD is sometimes
produced quickly and spontaneously, while in other cases it may involve extensive
drafting and rewording. It is therefore more useful to expect different levels or de-
grees of conscious organisation of bilingual discourse, depending on mode and cir-
cumstances.
Moreover, it seems necessary to distinguish between various potential conse-
quences of planning for CS. One of these, as already noted (section 6), relates to
the composition of contributions within multiparty conversational exchanges.
Asynchronous modes, especially, enable writers to deploy strategies of partici-
pation in which a single post may compile responses to a number of previous posts;
in multilingual contexts, writers may thereby use post-internal CS as a resource for
distinguishing between addressees or perspectives. What is particular to CS in this
regard is its strategic deployment in a context of discourse organisation that is
uniquely digital.
A second and less well understood impact of planning is on the potential avoid-
ance of or preference for certain types of switching or mixing. While examples like
(3) may reflect the mixed language style that has been reported as typical for the
speech of immigrant background youth (e.g., Hinnenkamp 2008), language mixing
is relatively rare across the forums of that study. I suggested (Androutsopoulos
2007a: 347–348) that this may be due to the combined impact of the conditions of
even script choice) in the production of CS in CMC. Again, different aspects need
to be distinguished preliminarily. One is the impact of standardised orthographies
on the selection of base language in CMD. Hinrichs (2006) reports that unlike
everyday communication in Jamaica, where Creole is the default medium of com-
munication, its lack of a standardised orthography makes it less suitable to that
purpose in CMD, so that the Jamaican students and expatriates in his study subjects
basically draw on English, occasionally switching to Creole. However, in other
settings, writers may stick to their vernacular in CMD despite its lack of a stan-
dardised orthography, for example in German-speaking Switzerland (Siebenhaar
2005, 2006), where the online spelling of local dialects tends to reflect vernacular
pronunciation more accurately than traditional dialect spelling does.
A further issue, again following up on the notion of code-centered contextual-
isation cues, is the role of spelling in signalling CS. It is obvious that CS in CMD is
produced by representing in writing another language (or dialect), but what is less
obvious is that this can be done based on different orthographic conventions. The
switch from one language to another usually co-occurs with, and is indexed by, a
switch between the respective orthographies, albeit not necessarily so. Represent-
ing one language in the orthography of another may sometimes be a matter of
necessity, but it may also be a more or less conscious choice and thereby a source
of pragmatic meaning in its own right. There is also evidence that spelling may
be exploited in its own right, with CMD writers drawing on the contrast with
normative orthography to create pragmatic meaning. A striking example is the de-
liberate “mixing of alphabetic conventions” reported for German-Turkish chatters
who create mixed-language conversations as “German words and even phrases get
a kind of Turkish wrapping”, which consists of Turkish orthography (Hinnenkamp
2008: 262, 266). Spelling Deutsch (in correct German orthography) as Doyc (based
on Turkish orthography) is, as the context makes clear, not a typo or spelling mis-
take but a conscious blend of language and orthography designed to elicit prag-
matic meaning, which ties in with the debate on language and ethnic boundaries
that dominates that chat session. Likewise, although at a different level of written
structure, there is evidence of purposeful “script switching” between native and
Roman script in the Romanised transliteration of different languages on the Inter-
net (see chapters in Danet and Herring 2007a).
Such evidence, I argue, suggests that the study of CS in CMC primarily in
terms of its apparent “authenticity” or correspondence to spoken conversational
CS may be limiting, and that important insights will be gained by theorising the
written digital mode not as a limitation but as a new set of conditions for the de-
ployment of multilingual resources in discourse.
Research has only just started tackling the massive bilingualism and multilingual-
ism that occurs as global multilingual populations increasingly gain access to digi-
tal communications media. Much remains to be done in documenting different
sites and types of CS online, and systematic comparisons among modes, language,
and settings are needed. Most lacking are, first, studies of private, dyadic data; sec-
ond, cross-media and cross-mode comparisons of CS usage based on the same
writer(s); third, multimodal data from social networking and media-sharing web-
sites; and fourth, case studies of multilingual CMD in transnational work teams. In
terms of method, research is moving away from static classifications and towards
ethnographically and pragmatically informed analyses of the local interactional
purposes that CS serves in its generic and sequential context. However, mixed-
methods combinations of qualitative and quantitative techniques, e.g., by means of
questionnaires or language choice analyses, are bound to remain productive and in-
sightful. One limitation of current research is the restriction to single modes, which
are analytically examined in isolation. Motivated by practical necessities as that
may be, it creates an isolationism that runs counter to actual computer-mediated
practices, which are spread across modes and platforms in combinations and rou-
tines that are not yet well understood. There is reason to assume that CS patterns
will often cut across modes, and understanding such code/mode repertoires will
deepen the understanding of the specific properties of CS online.
Two issues that cut across the sections of this chapter are how to theorise the re-
lationship between medium and social/situation factors, on the one hand, and be-
tween online written and offline spoken CS, on the other. Even though it might
seem customary to think of CS on the Internet in terms of its “authenticity” or cor-
respondence to an assumed spoken conversational blueprint in the usage of an in-
dividual or community, I have argued that CMD is unscripted, dynamically unfold-
ing communication in its own right. Taking this into account would cast doubt on
both the necessity and the means of establishing such authenticity. This chapter has
presented a number of cases where bilingual practices are not verbatim reproduc-
tions of face-to-face interaction patterns but, judging from ethnographic and lin-
guistic evidence, specific to CMD. I therefore suggest that rather than examining
CS online in terms of its authenticity or equivalence to offline speech, a more pro-
ductive question to pursue is how CS is used as a resource, under the specific con-
ditions of communication offered by digital media.
Notes
1. Only two volumes to date include more than one contribution dealing, at least in part,
with CS (Danet and Herring 2007a and a 2006 special issue of the Journal of Socioling-
uistics 10[4]), while in other recent books the topic is either entirely absent or marginal
(e.g., Baron 2008; Crystal 2006; Rowe and Wyss 2009).
2. The languages of publication covered are English and German. Literature in other lan-
guages can be traced through the reference list. Literature coverage is through June
2010.
3. The latter is used by Androutsopoulos and Hinnenkamp (2001), Androutsopoulos and
Ziegler (2004), Androutsopoulos (2006a, 2007a), Hinnenkamp (2008), Hinrichs (2006),
Leppänen (2007), Siebenhaar (2005, 2006), and Tsiplakou (2009).
4. Auer points out that “a speaker may simply want to avoid the language in which he or
she feels insecure and speak the one in which he or she has greater competence. Yet
preference-related switching may also be due to a deliberate decision based on political
considerations” (1995: 125).
5. Following Herring (2004), I use computer-mediated discourse (CMD) to indicate a nar-
rower focus on the use of semiotic resources, whereas CMC denotes a broader view on
communication processes facilitated by digital technologies.
6. In earlier work (Androutsopoulos 2006a), I located CS within a quadripartite matrix of
multilingualism on ethnic portals. Its main site is regular user-contributed discourse, es-
pecially in forums and occasionally in journalists’ edited content as well. This is distin-
guished from user and site-specific emblems, such as names, mottos, and slogans. These
distinctions are taken up and recontextualised in the present discussion.
7. This mirrors the limitations of research on multilingualism in CMC in general (see
Danet and Herring 2007b: 24).
8. The label “L1” (first language) was only added if clearly stated in the research, usually
based on controlled socio-demographic data. Other studies either do not control for
this factor or they study public CMD settings where assigning an L1 is not straightfor-
ward.
9. All of the following studies include discussions of these: Androutsopoulos (2006a,
2007a), Androutsopoulos and Hinnenkamp (2001), Dorleijn and Nortier (2009), Geor-
gakopoulou (1997), Hinrichs (2006), Kadende-Kaiser (2000), McClure (2001), Paolillo
(1996, 2011), Sebba (2003), Sperlich (2005), and Tsiplakou (2009).
10. The examples are anonymised, and presentation follows Androutsopoulos (2006a, b).
References
Androutsopoulos, Jannis K.
2004 Non-native English and sub-cultural identities in media discourse. In: Helge
Sandøy (ed.), Den fleirspråklege utfordringa/The Multilingual Challenge,
83–98. Oslo: Novus.
Androutsopoulos, Jannis K.
2006a Multilingualism, diaspora, and the Internet: Codes and identities on German-
based diaspora websites. Journal of Sociolinguistics 10(4): 524–551.
Androutsopoulos, Jannis K.
2006b Mehrsprachigkeit im deutschen Internet: Sprachwahl und Sprachwechsel in
Ethno-Portalen. [Multilingualism on the German Internet: Language choice
and code-switching on ethnic portals.] In: Peter Schlobinski (ed.), Von *hdl*
bis *cul8r*. Sprache und Kommunikation in den Neuen Medien [From *hdl*
to *cul8r*. Language and communication in New Media], 172–196. Thema
Deutsch, Band 7. Mannheim: Dudenverlag.
Androutsopoulos, Jannis K.
2007a Language choice and code-switching in German-based diasporic web forums.
In: Brenda Danet and Susan C. Herring (eds.), 340–361.
Androutsopoulos, Jannis K.
2007b Bilingualism in the mass media and on the Internet. In: Monica Heller (ed.),
Bilingualism: A Social Approach, 207–230. New York: Palgrave Macmillan.
Androutsopoulos, Jannis K.
2011 From variation to heteroglossia in the study of computer-mediated discourse.
In: Crispin Thurlow and Kristine Mroczek (eds.), Digital Discourse: Lan-
guage in the New Media, 277–298. New York: Oxford University Press.
Androutsopoulos, Jannis and Volker Hinnenkamp
2001 Code-Switching in der bilingualen Chat-Kommunikation: ein explorativer
Blick auf #hellas und #turks. [Code-switching in bilingual chat communi-
cation: an exploratory analysis of #hellas and #turks.] In: Michael Beißwenger
(ed.), Chat-Kommunikation, 367–402. Stuttgart: Ibidem.
Androutsopoulos, Jannis and Evelyn Ziegler
2004 Exploring language variation on the Internet: Regional speech in a chat com-
munity. In: Britt-Louise Gunnarsson, Lena Bergström, Gerd Eklund, Staffan
Fridell, Lise H. Hansen, Angela Karstadt, Bengt Nordberg, Eva Sundgren, and
Mats Thelander (eds.), Language Variation in Europe, 99–111. Uppsala:
Uppsala University Press.
Auer, Peter
1995 The pragmatics of code-switching: A sequential approach. In: Lesley Milroy
and Pieter Muysken (eds.), One Speaker, Two Languages, 115–135. Cam-
bridge, UK: Cambridge University Press.
Auer, Peter (ed.)
1998a Code-Switching in Conversation. London: Routledge.
Auer, Peter
1998b Introduction: Bilingual conversation revisited. In: Peter Auer (ed.), 1–24.
Auer, Peter
1999 From codeswitching via language mixing to fused lects: Toward a dynamic
typology of bilingual speech. International Journal of Bilingualism 3(4):
309–332.
Auer, Peter
2000 Why should we and how can we determine the “base language” of a bilingual
conversation? Estudios de Sociolingüística 1(1): 129–144.
Baron, Naomi S.
2008 Always On: Language in an Online and Mobile World. New York: Oxford Uni-
versity Press.
Beißwenger, Michael
2008 Situated chat analysis as a window to the user’s perspective: Aspects of tem-
poral and sequential organization. Language@Internet 5, article 6.
http://www.languageatinternet.org/articles/2008/1532/
Bell, Allan
1984 Language style as audience design. Language in Society 13: 145–204.
Callahan, Laura
2004 Spanish/English Codeswitching in a Written Corpus. Amsterdam/Philadelp-
hia: John Benjamins.
Crystal, David
2006 Language and the Internet, 2nd Ed. Cambridge, UK: Cambridge University
Press
Danet, Brenda and Susan C. Herring (eds.)
2007a The Multilingual Internet: Language, Culture, and Communication Online.
New York: Oxford University Press.
Danet, Brenda and Susan C. Herring
2007b Introduction: Welcome to the multilingual Internet. In: Brenda Danet and
Susan C. Herring (eds.), 3–38.
Deumert, Ana and Oscar Masinyana Sibabalwe
2008 Mobile language choices – The use of English and isiXhosa in text messages
(SMS): Evidence from a bilingual South African sample. English World-Wide
29(2): 117–147.
Dorleijn, Margreet and Jacomine Nortier
2009 Code-switching and the internet. In: Barbara E. Bullock and Almeida J. Tori-
bio (eds.), The Cambridge Handbook of Linguistic Code-Switching, 127–141.
Cambridge, UK: Cambridge University Press.
Ferrara, Kathleen, Hans Brunner, and Greg Whittemore
1991 Interactive written discourse as an emergent register. Written Communication
8(1): 8–34.
Georgakopoulou, Alexandra
1997 Self-presentation and interactional alignments in e-mail discourse: The style-
and code switches of Greek messages. International Journal of Applied Lin-
guistics 7(2): 141–164.
Georgakopoulou, Alexandra
2003 Computer-mediated communication. In: Jef Verschueren, Jan-Ola Östman, Jan
Blommaert, and Chris Bulcaen (eds.), Handbook of Pragmatics (2001 Install-
ment), 1–20. Amsterdam/Philadelphia: John Benjamins.
Georgakopoulou, Alexandra
2004 To tell or not to tell? Email stories between on- and off-line interactions.
Language@Internet 1, article 1.
http://www.languageatinternet.org/articles/2004/36
Goldbarg, Rosalyn Negrón
2009 Spanish-English codeswitching in email communication. Language@Internet 6,
article 3. http://www.languageatinternet.org/articles/2009/2139
Gumperz, John J.
1982 Discourse Strategies. Cambridge, UK: Cambridge University Press.
Siebenhaar, Beat
2006 Code choice and code-switching in Swiss-German Internet Relay Chat rooms.
Journal of Sociolinguistics 10(4): 481–509.
Siebenhaar, Beat
2008 Quantitative approaches to linguistic variation in IRC: Implications for quali-
tative research. Language@Internet 5, article 4.
http://www.languageatinternet.org/articles/2008/1615/index_html/
Sperlich, Wolfgang B.
2005 Will cyberforums save endangered languages? A Niuean case study. Inter-
national Journal of the Sociology of Language 172: 51–77.
Tsaliki, Liza
2003 Globalization and hybridity: The construction of Greekness on the Internet. In:
Karim H. Karim (ed.), The Media of Diaspora: Mapping the Globe, 162–176.
London: Routledge.
Tsiplakou, Stavroula
2009 Doing (bi)lingualism: Language alternation as performative construction of
online identities. Pragmatics 19(3): 361–391.
Warschauer, Mark, Ghada R. El Said, and Ayman Zohry
2007 Language choice online: Globalization and identity in Egypt. In: Brenda Danet
and Susan C. Herring (eds.), 303–318.
Wright, Sue (ed.)
2004 Multilingualism on the Internet. Special issue, International Journal on Multi-
cultural Societies 6(1). UNESCO Social and Human Sciences. http://www.un-
esco.org/shs/ijms
Ziegler, Evelyn
2005 Die Bedeutung von Interaktionsstatus und Interaktionsmodus für die Dialekt-
Standard-Variation in der Chatkommunikation. [Relevance of interactional
status and interactional mode for dialect/standard variation in chat communi-
cation.] In: Eckhardt Eggers, Jürgen Erich Schmidt, and Dieter Stellmacher
(eds.), Neue Dialekte – moderne Dialektologie [New dialects – modern dialec-
tology], 719–745. Stuttgart: Steiner.
1. Introduction
dividuals relate about their past and how they relate them, when prompted to do so
in research interviews, has been a favoured preoccupation and focus of identity
analysis in the social sciences (e.g., Riessman 2008). Clearly, there is a discrep-
ancy between narrative being of focal concern outside CMC research and being
under-represented within it. In light of this, the discussion in this chapter will in
part be programmatic, since this is an area with obvious gaps and avenues for
further research. At the same time, my survey of existing studies of narrative in
CMC will show that various connections are either made by the studies themselves
or can be made with some of the latest tendencies of pragmatic and sociolinguistic
research on narrative that adopt a social-interactional perspective. In this respect, I
will discuss the recent move in narrative analysis towards an opening up of textual
definitions of narrative to include stories that depart in varying degrees from the
narrative canon (i.e., life stories or personal experience past events stories). I will
show how studies of narrative in CMC are providing further evidence in favour of
such a move. Moreover, I will argue that the storytelling forms and functions that
are emerging through such studies as distinctive and characteristic of numerous
CMC environments can and should inform a dialogue with studies of face-to-face
storytelling.
Fascination with oral (vernacular, non-literary) narrative spans a wide range of so-
cial science disciplines – sociology, psychology, social anthropology, etc. – and is
frequently referred to as the “narrative turn” (see articles in Bamberg 2006; Brock-
meier and Carbaugh 2001; Daiute and Lightfoot 2004; De Fina, Schiffrin, and
Bamberg 2006). In these disciplines, narrative is seen as an archetypal, fundamen-
tal mode for making sense of the world and, as such, as a privileged structure, sys-
tem, or mode for tapping into the teller’s social identities and sense of self. To talk
about well-defined and delimited schools of narrative analysis necessitates a crude
process of abstraction that does not do justice to the many different strands of
existing work. Bearing this in mind, it is possible to identify two distinct ways of
doing narrative analysis: one inspired by a conventional and largely canonical
paradigm, the other based on a social-interactional view of narrative. The former
is traceable to Labov’s foundational model of narrative analysis (Labov 1972;
Labov and Waletzky 1967), but it is also informed by the narrative turn in the so-
cial sciences within which narrative is seen as a prime site for the exploration of
people’s identities. The latter encompasses a view of narrative as talk-in-interac-
tion and as social practice, identifying the analysis of interaction as a fundamental
aspect of any study of narrative and the investigation of the intimate links of nar-
rative-interactional processes with larger socio-cultural processes as a prerequisite
for the study of narrative (for a detailed discussion, see De Fina and Georgakopou-
lou 2008a).
In Labov and Waletzky’s model (1967), a narrative text was characterized in
structural terms through the presence of temporal ordering between event clauses
and through its organization into structural components (such as complicating
action, evaluation, and resolution). The model was based on stories told in oral
interviews and, in principle, was meant as a description of narratives of personal
experience, rather than all types of narratives. That said, it has had a profound in-
fluence on narrative research and has paved the way for seeing narrative as a privi-
leged site for the study of a wide range of aspects relevant to the study of language
in society (see contributions in Bamberg 1997). As I have argued elsewhere (Geor-
gakopoulou 2006a: 129–131), Labov’s structural definition of narrative resulted in
a tendency to recognize as narratives only texts that appear to be well organized,
with a beginning, a middle, and an end, that are teller-led and largely monologic,
and that occur as responses to (an interviewer’s) questions. In addition, his focus
on a story’s point and evaluation as the means by which a teller expresses the
story’s tellability has inadvertently privileged the teller as the main producer of
meaning. Despite the critique that Labov’s model has received over the years (see,
e.g., the articles in Bamberg 1997), its long-lasting influence in the study of every-
day, conversational narrative is undeniable. It is commonly the case that the initial
exploration of a narrative data set will rest upon identifying the extent to which the
data conform to or depart from the narrative structure identified by Labov.
As will be discussed below, this has also been the case in existing studies of
narrative genres in CMC. Yet it is also noteworthy that both recent studies of con-
versational storytelling and studies of narrative in CMC have provided ample evi-
dence of an abundance of stories that do not actually conform to Labov’s model, in
as much as they depart from the format of “an active teller, highly tellable account,
relatively detached from surrounding talk and activity, linear temporal and causal
organization, and certain, constant moral stance” (Ochs and Capps 2001: 20). Put
differently, the stories that have been documented as typical of a variety of every-
day conversational and – more recently – CMC contexts include “a gamut of
under-represented and ‘a-typical’ narrative activities, such as tellings of ongoing
events, future or hypothetical events, shared (known) events, but also allusions to
tellings, deferrals of tellings, and refusals to tell” (Georgakopoulou 2006a: 130).
An organized move to put such “non-canonical” or “a-typical” stories on the
map and make them a focal part of narrative analysis is exemplified by recent work
on “small stories” (Bamberg 2004; Bamberg and Georgakopoulou 2008; Georga-
kopoulou 2006a, b, 2007, 2008). Small stories research is in tune with other recent
approaches to narrative both within sociolinguistics (e.g., Ochs and Capps 2001)
and in narratology (e.g., Herman 2009). In these approaches, the widely-held-as-
prototypical definitional criteria of narrative, such as the sequencing of events, are
seen as necessary but not as sufficient. Instead, what counts as a story is connected
with what gets done on particular occasions and in particular settings and what the
local understandings are regarding what a story is. Furthermore, there is room for
flexibility and versatility within each definitional criterion of a story: For instance,
the criterion of event sequencing has tended to privilege the temporal ordering of
past rather than future or hypothetical events. Similarly, the idea that stories build
worlds and that through their tellings the narrators communicate to their audience
“what’s it like” to be part of those worlds (Herman 2009) has routinely prioritized
the “world-disruption” qualities of stories. The assumption has been that, for
stories to be worth telling, they need to relate extra-ordinary events that create
some kind of trouble or complication in the initial state of affairs. This applies to
Labov’s model, too.
What small stories research has argued in relation to such definitional criteria is
that they should be seen as context specific, and that the full continuum that each
allows for should be explored in each case. For instance, world-making rather than
world-disruption may be more important for some stories in certain contexts. In the
same vein, there may be a distinct preference for the sequenced events of a story to
be about the future and not about the past in certain contexts (Georgakopoulou
2007: ch. 2). This should not preclude the identification of such data as stories or
their inclusion in a narrative analysis. A comparable case has been made by Ochs
and Capps (2001), who, rather than identifying a set of distinctive features that al-
ways characterize narrative, stipulate dimensions (tellability, tellership, linearity,
moral stance, etc.) that are relevant to a narrative, even if not elaborately manifest.
Each dimension establishes a range of possibilities that may or may not be realized
in a particular narrative. For example, tellership allows for one main teller but also
for multiple co-tellers. Similarly, a story’s tellability may be high or low.
Overall, social-interactional approaches to narrative, including small stories
research, have shown the importance of attending to the context-specific aspects of
narrative tellings so as to understand how narrative genres shape, as well as are
being shaped by, the norms and social relations of the situational and socio-cultural
contexts in which they occur.1 The last decade or so has also seen social-interac-
tional approaches connect with social science research on narrative as a point of
entry into the teller’s social identities (gender, ethnicity, age, etc.). This study of
identities through emphasis on the communicative how of narrative tellings (e.g.,
De Fina 2003; De Fina, Schiffrin, and Bamberg 2006; Georgakopoulou 2007) has
variously problematized, de-essentialized, or added nuance to the widely-held
view that narrative is a privileged communication mode for making sense of the
self. In section 3, I suggest that there is still much scope for a comparable shift in
studies of narrative in CMC from narratives as texts/genres to narratives as social
practices.
The focus so far in studies of narrative in CMC has been on what types of stories
are told, how, and in which CMC environments. This has brought to the fore recur-
rent findings that allow us to identify certain narrative genres and certain com-
municative choices in them as being prototypically associated with CMC. There
are links that can be made between such “ways of storytelling” (Georgakopoulou
2007) and certain possibilities that communication in the new media affords, as
well as certain social practices with which they are characteristically associated, as
I show below.
The exploration of ways of storytelling in CMC contexts has prioritized “fic-
tional”2 stories3 or “representations”; the turn to personal stories has only recently
been made. This is more or less a reflection of the historical precedence of fictional
storytelling (as in, e.g., computer games) in CMC, which has nonetheless been rad-
ically redefined with the recent explosion of personal storytelling, particularly in
Web 2.0 environments. However, it is instructive to note a convergence of findings
between work on fictional storytelling and work on personal storytelling in CMC.
Literary or fictional storytelling has traditionally been the main object of inquiry of
narratology, and it is thus no accident that the study of fictional, at times artistic,
online stories created by experts and professionals (e.g., experimental electronic
literature, video games) makes use of narratological approaches4 and, to a lesser
extent, sociological ones (e.g., Bell 2010; Ciccorico 2007; Douglas 2000; Hayles
2008; Morris and Swiss 2009; Walker Rettberg 2008). One of the aims of this line
of inquiry has been to examine narrative through a comparative lens across media,
spurred by the recognition that “the comparative study of media as means of ex-
pression lags behind the study of media as channels of communication” (Ryan
2004a: 24).
A central project within this “transmedial” narratology is to study how nar-
rative gets transposed from one medium to another, what kinds of features are
adapted or altered in this process, and how each medium encourages or prohibits
specific ways of narration (Ryan 2004a: 35). Transmedial narratology also seeks to
test the applicability of concepts and analytical modes that have been developed
with regard to the study of narrative in one medium across media. Ryan rightly
stresses that the examination of such questions should “avoid the temptation to at-
tribute features and findings to the medium solely” (34). This danger of technologi-
cal determinism has been widely discussed in studies of CMC in general (see, e.g.,
articles in Androutsopoulos 2006). Ryan also warns against the other extreme, that
is, “media blindness”, which involves an indiscriminate transfer of concepts de-
signed, for example, for the study of narrative in one medium to narratives of
another medium. This concern, too, has been voiced in CMC, particularly with re-
(e.g., Beck 2004; Ritchie and Peters 2001). The term that is mostly employed for
such digitally produced stories is, not unexpectedly, “digital”. Studies of digital
storytelling draw more or less explicitly on insights about the role of narrative
emanating from social science research that is often called narrative inquiry. The
main proponents of narrative inquiry have celebrated narrative as a mode of com-
munication capable of empowering social actors (e.g., Polkinghorne 1988; Sarbin
1986). Lay people telling their stories to an attentive interviewer in interview con-
texts has been seen as a therapeutic exercise that emancipates ordinary people by
letting their voices be heard. This line of inquiry has recently been criticized for
overstating the significance of narrative and for stressing the representational as-
pects of narrative, taking what people tell about themselves at face value, as op-
posed to exploring how such aspects are shaped by contextual and interactional
factors at work (see Atkinson and Delamont 2006). Nonetheless, narrative inquiry
has been a very important qualitative research approach. Many of its epistemologi-
cal assumptions can be seen as regards digital storytelling, too, particularly the be-
lief in the emancipatory power of storytelling, which in the case of digital stories is
argued to be taken to new heights with the possibilities that the media offer for the
co-production and the wide distribution of stories (e.g., Bratteteig 2008). This
latter point resonates across studies of narrative in CMC.
Digital storytelling has figured in debates about media-led uses and new media
participation (Hartley and MacWilliam 2009), the assumption being that it funda-
mentally collapses the distinction between professional and amateur forms of self-
expression of production, at the same time encouraging creativity through the
blending of different semiotic systems. An important initiative on digital storytell-
ing originated in the Center for Digital Storytelling in California, which in the last
15 years or so has digitized 12,000 stories (Lundby 2008). The volume by Harltey
and MacWilliam (2009) documents many similar initiatives in various parts of the
world, where researchers have encouraged and enabled lay people to produce their
stories, however short or amateurishly done (e.g., with off-the-shelf equipment),
and to make them available online.
Studies of digital storytelling have by and large concentrated on stories that
have been produced under the rubrics of cultural institutions rather than on spon-
taneous teller-initiated and -led storytelling, which has recently exploded in Web
2.0 environments such as YouTube. The inclusion of the whole range of self-rep-
resentational media practices in the remit of studies of digital storytelling would
shed further light on the democratization and empowerment potential of digital
stories. More generally, studies of digital storytelling have hardly focused on how
digital stories are actually told, what narrative genres they represent, how they
differ from other narrative genres in other media, and what their main textual and
interactional aspects are. By contrast, an emphasis on narratives in online environ-
ments as genres is emerging as the focus of sociolinguistic research.
Pragmatic and sociolinguistic studies of narrative in CMC have mainly been inter-
ested in exploring what kinds of stories are told and how. This identification of
types of stories tends to be premised on the extent to which they exhibit certain
structural components and the ways in which they realize them. For instance, the
degree and types of evaluation in a story would be important in these approaches.
According to this point of view, both the conventional paradigm of narrative analy-
sis and the latest social-interactional approaches are relevant, inasmuch as the
noted occurrence of non-canonical stories (as described by the conventional para-
digm) in various CMC environments is in tune with the findings of the social-
interactional approaches.
The focus has mainly been on personal stories by lay writers, which are becom-
ing increasingly important in CMC as resources for self-styling and sustaining on-
line communities.5 Although the existing studies present considerable variation in
terms of their methodological and analytical frameworks, there are striking simi-
larities in their findings. In particular, the examined narratives, rather than being
single-authored, offer a distinctive potential for multiple authorship. This is facili-
tated by the possibilities for the wide circulation of a story across CMC sites and
over time, what has been labelled a “distributed story” (Walker 2004). This shap-
ing of a story by different tellers in different environments tends to militate against
a conventional structuring with a beginning, middle, and an end, as for instance
described by Labov. Instead, it frequently results in “fragmented” stories (e.g.,
Walker Rettberg 2008). At the same time, in many CMC environments it makes
little sense to talk about the structure of a story in purely linguistic (verbal) terms
without taking into account the multi-semioticity or multimodality that the
medium offers for its role in the creation of story plots. A case in point is Hoff-
mann’s (2008) and Hoffmann’s and Eisenlauer’s (2010) studies of travel weblog
entries, which show a preponderance of multi-linear, multimodal, collaboratively
produced and distributed narratives. Their analysis stresses the point that CMC en-
ables new forms of narrative interaction and authorship, ranging from the users as
authors who select the telling and design of telling online and (re)shape it, to the
users as “audience” who contribute to online evaluations, selecting the reading
order and serializing their reading engagement with weblog narratives. As noted in
section 3.1 above, this point has also been made in relation to fictional stories.
As a result of medium-enabled features, Hoffmann and Eisenlauer also ob-
served a certain fragmentation of themes and perspectives, a versatile access to the
narration, a lack of closure, and a semiotic flexibility. At the same time, weblog
narratives were not found to form a homogeneous entity but instead to vary in size,
theme, authorship, and forms of interactivity. The idea that narrative tellings on the
web can be transposed (e.g., posted) in new contexts, and that in that process new
meanings can be produced, is emerging as a common theme (see also Walker
2004). For these reasons, the unprecedented potential for co-construction, versatil-
ity, and intertextuality of narratives in online environments is frequently stated as
one of the main findings of studies of narrative in CMC (see also earlier studies of
the web and hypertext in general, e.g., Mitra 1999).
The four features of multi-linearity, fragmentation, multimodality, and interac-
tivity that Hoffmann and Eisenlauer identify as key in their data resonate with
other studies of online storytelling. It is also noteworthy that Hoffmann and Eisen-
lauer’s starting point is Labov’s model of narrative analysis. Their analysis subse-
quently shows that Labov’s model, with its emphasis on a single teller, a begin-
ning, middle, and end, and a linear development of the plot, cannot capture the
multifarious ways in which stories are told, evolve, and are transported across sites
in CMC. To take Labov as the starting point, without connecting with studies of
conversational storytelling that have criticized Labov, is not uncommon in socio-
linguistically-informed studies of narratives in CMC (e.g., Arendholz 2010; Heyd
2009; Jucker 2010).6 However, like the studies that have criticized Labov, the
CMC studies also report the occurrence of stories that depart from Labov’s full-
fledged narratives.7
An example of this is Heyd’s (2009) study of narratives in email hoaxes, defined
as “a typical case of deceptive computer-mediated communication” (1) which con-
sists of messages with false information that are disseminated electronically via the
forwarding function of email programs. Heyd (163–184) argues that the degree of
narrativity and the type of story are a function of the type of email hoax, with urban
legends and charity email hoaxes displaying the strongest tendency towards full-
fledged narratives. Heyd connects her findings with small story research as a frame
of analytical reference.
The context-specificity of the ways of storytelling in CMC is being increas-
ingly documented and, as has been the case in offline environments, there is at-
tested heterogeneity and versatility in what stories are told and how they are told,
not only across CMC environments but also within the same environment. For in-
stance, in the discussion forum of a body-building community, Page (2009) reports
that different types of stories are intimately linked with their placement, i.e., where
exactly in relation to a post they occurred, and their function, i.e., what purpose or
social action they fulfilled. Specifically, stories told in response to a post were often
stories of projected events, offering the original writer advice on how he or she
should act in the future. In another study by the same author (2008) of narratives
in weblogs of bloggers who had been diagnosed and were being treated for cancer,
the different types of stories told were linked to the blogger’s gender.
Similarly, in a study of email messages exchanged between diasporic Greek
friends based in London (Georgakopoulou 2004), it was found that the types of
stories told were connected with the participants’ intimate relationships and actual
geographical proximity, which allowed for numerous offline interactions as well.
Specifically, the stories established links between the current tellings on email and
previous and future face-to-face interactions of the senders and receivers: For in-
stance, certain stories that I called “stories-to-be-told” started out being narrated on
email, but their full telling was deferred and promised for an occasion of face-to-
face interaction. What I called “breaking news” stories (see discussion below) re-
ported events with immediacy, i.e., very recent events (e.g., “this morning”, “just
now”), or as still unfolding, as the story was being constructed. Overall, the study
proposed an account of these types of stories that hinged on the intersection
between aspects of the mediated context at hand and the participants’ intimate re-
lationships. In terms of the former, the tellers’ media-afforded ability to share
events with their friends as they were happening to them, their addressees’ media-
afforded ability to provide feedback on those events very quickly, and the require-
ment (at least in the given group) to keep messages very brief were all relevant fac-
tors. In general, this line of inquiry into how ways of (story)telling in CMC con-
texts are shaped by the medium, on the one hand, and the storytellers’ social roles
and relations, on the other, is still in its early stages.
What existing studies have shown quite definitively is that the canonical nar-
rative of (more or less distant) past events from the teller’s life is far from being the
norm in CMC environments: See, for example, Baym (2000) in relation to the
stories in a soap-opera newsgroup; Georgakopoulou (2004) in relation to the stories
in private email messages; and Page (2009) in relation to status updates on Face-
book. (A detailed discussion of Page’s studies of personal stories in blogs and
Facebook updates can be found in her 2012 book, which was not published and
thus not available to me when writing this chapter.) This shortening of the distance
between the two worlds, that is, the narrating and the narrated world, is emerging
as a distinctive feature not just in personal storytelling in CMC but also in personal
stories facilitated by the new media. This is a finding worth assessing further in the
context of new media affordances for storytelling. In much of the narrative theory
developed so far, the deictic separation between the two worlds of telling and told,
or the so-called double chronology (Chatman 1990), has been a feature of major
importance for the creation of the tellers’ perspective or point of view, as well as
for their ability to reflect on the narrated events (see Freeman 2010). It is therefore
worth scrutinizing the implications of the frequency and salience of the “breaking
news” type of story in online environments for these key features of the narrative
canon.
In particular, the following features of breaking news stories relate to media af-
fordances: the speed with which a story can be shared with others; the potential for
the “audiences” of the story to shape the continuation of the telling with their en-
gagements; and, finally, the potential for distribution of the story. Let us consider
these features in the following examples. The first example comes from a sequence
of two Facebook status updates (SUs) and the comments they received. (Note that
on Facebook, SUs and comments are followed by timestamps appended automati-
cally by the system.)
(1)
Gertie Brown8 is feeling much better with a hole in her leg!
August 27 at 12.19pm
David Martin Got to ask … What! How big??!
August 27 at 3.15pm
Gertie Brown it was about 3 inches! looks like a bullet wound, now about 1 inch.
August 27 at 3:18pm
David Martin how? Why? Is JB shooting at you now?·
August 27 at 3.22pm
Gertie Brown Got in the way of a pigeon …
August 27 at 3.26pm
4. Conclusion
which in these respects depart from the canonical narratives that until recently
monopolized scholarly attention. Both narratological studies of fictional stories
and sociolinguistic studies of personal stories support this finding in a wide range
of CMC environments.
The extent to which such forms of narrative constitute new and/or exclusively
media-shaped genres is, nonetheless, less well understood and agreed upon.
Studies with an emphasis on processes of re-mediation tend to stress the import-
ance of the hypertextual organisation and multimodality of the Internet for the
ways in which narratives are shaped online,13 therefore suggesting more or less ex-
plicitly medium-shaped changes. Studies that draw on research in conversational
(face-to-face) storytelling seem to be more attuned to narrative features such as in-
teractivity, intertextuality, and fragmentation that have recently been attested in
conversational storytelling, as well. The co-construction of stories and their ability
to be transposed across time and place and recontextualized in the process, while at
the same time acquiring new meanings in their new environments, is something
that has been stressed by many a study in relation to conversational storytelling
(e.g., Bauman and Briggs 1990). Yet, as the discussion of breaking news stories at-
tempted to show, there seems to be a difference in the scale of distribution and re-
contextualization enabled by CMC. Also on a different scale is the potential for the
involvement of diverse and in many cases “invisible” to the teller audiences in the
shaping and (re)shaping of a story.
On the basis of this, it can be argued that both the potential for a story’s re-
contextualization and the various modes of audience interaction with it are inti-
mately linked with the new media-enabled potential and speed for telling the story
in the first place, for continuing to develop it in the same or other environments,
and for distributing it widely. To engage more fully with work on conversational
storytelling – for instance, small stories research, which has attested to similar
ways of storytelling – may be one way to push forward the agenda of articulating as
fully as possible what is distinctive about online stories and how that draws on, de-
parts from, or indeed remediates other forms and practices of storytelling in offline
or other media environments.
Further analyses should also connect with the proliferating work on CMC and
identities (see, e.g., chapters in Danet and Herring 2007), as well as with the direc-
tion that the linguistically-informed study of contemporary digital textual practices
is increasingly taking: This direction involves going beyond the exploration of lin-
guistic features of different types of CMC and instead engaging systematically
with their multi-authorship, translocality, and multimodality (e.g., Androutsopou-
los 2010). Finally, unanswered questions still remain about the interrelationships
between online and offline narrative activities, particularly with regard to how new
media as historically and socioculturally shaped resources and tools enable and be-
come oriented to in narrative activities offline.
Acknowledgements
This chapter has benefited greatly from the detailed comments and editorial care of
Susan Herring, Tuija Virtanen, and Dieter Stein, to whom I express my gratitude.
Notes
1. For examples of early and defining work in this respect see Bauman (1986), M. H.
Goodwin (1990), Hymes (1981), and Shuman (1986).
2. The term “virtual” narratives has also been used (Ryan 2004).
3. The terms “narrative” and “story” (storytelling) are used interchangeably in this
chapter. This is not always the case in the literature on narrative, particularly in the nar-
ratological tradition.
4. This discussion cannot possibly do justice to the different schools of narratology as they
have developed over the years nor to the decisive shift in the area from so-called clas-
sical to post-classical narratology (for an overview of the area, see Fludernik 2009).
Suffice it to say that narratological studies of narrative in CMC are mostly situated in the
post-classical narratology, which has increasingly attended to issues of context as well
as opening up the definitions of narrative to include cases beyond literature (e.g., rang-
ing from everyday storytelling to comics and “virtual” stories).
5. For instance, Sack (2003), although interested in story understanding rather than pro-
duction, states that “many online social networks have their basis in the sharing of nar-
ratives”, “a non-trivial portion of social networks are based on discussions of widely
circulated stories”, and “virtual, on-line communities are a result of net-mediated, story-
based relations” (305).
6. Elsewhere (Georgakopoulou 2006b), I have noted a similar trend in studies that have
sought to document the relationships of CMC with spoken and written discourse. I have
specifically suggested that many of them failed to draw on and be informed by previous
sociolinguistic studies of spoken and written discourse that had problematized the rela-
tionship between spoken and written language.
7. Although not related to personal storytelling, Jucker’s (2010) study of online live sports
commentaries stressed the media-afforded (online television) possibility for synchron-
izing the telling time with the event time and for continuously updating the account re-
sults in distinct forms of storytelling that are based on enhanced interactivity.
8. All usernames and person and place names employed in the examples are pseudonyms.
9. A&E is an abbreviation for Accident & Emergency.
10. Jamie Oliver is an English celebrity chef.
11. The readers’ comments on Mail Online articles are rated by other readers positively or
negatively, with the arrows F and G respectively. The number next to the category Rating
refers to the total score of the ratings received. In this particular example, best-rated
comments received overall ratings of approximately 1500.
12. The study was part of an ESRC Identities and Social Action funded project entitled
“Urban Classroom Culture and Interaction” (www. identities.org.uk). The project team
consisted of Ben Rampton (Director), Roxy Harris, Alexandra Georgakopoulou, Con-
stant Leung, Caroline Dover, and Lauren Small.
13. For instance, Fetzer (2010) suggests that political websites fulfill the same persuasive
function as that of comparable material in printed form (e.g., newspapers), but through
the use of embedded and multimodal narratives in hypertextual links.
References
Androutsopoulos, Jannis K.
2006 Introduction: Sociolinguistics and computer-mediated communication. Jour-
nal of Sociolinguistics 10(4): 419–438.
Androutsopoulos, Jannis K.
2010 The study of language and space in media discourse. In: Peter Auer and Jürgen
E. Schmidt (eds.), Language and Space: An International Handbook of Lin-
guistic Variation. Volume I: Theories and Methods, 740–758. Handbooks of
Linguistics and Communication Science 30.1. Berlin/New York: de Gruyter.
Arendholz, Jenny
2010 “Need to put this out there (My Story)” – Narratives in message boards. In:
Christian Hoffmann (ed.), Narrative Revisited, 109–142. Amsterdam/Philad-
elphia: John Benjamins.
Atkinson, Paul and Sarah Delamont
2006 Rescuing narrative from qualitative research. Narrative Inquiry 16: 173–181.
Bamberg, Michael (ed.)
1997 Oral Versions of Personal Experience: Three Decades of Narrative Analysis.
Special issue, Journal of Narrative and Life History 7(1–4).
Bamberg, Michael
2004 Talk, small stories, and adolescent identities. Human Development 47:
331–353.
Bamberg, Michael (ed.)
2006 Narrative – State of the Art. Special issue, Narrative Inquiry 16(1).
Bamberg, Michael and Alexandra Georgakopoulou
2008 Small stories as a new perspective in narrative and identity analysis. Text &
Talk 28: 377–396.
Bauman, Richard and Charles L. Briggs
1990 Poetics and performance as critical perspectives on language and social life.
Annual Review of Anthropology 19: 59–88.
Baym, Nancy K.
2000 Tune in, Log on: Soaps, Fandom, and Online Community. Thousand Oaks, CA:
Sage.
Beck, Dennis
2004 The use of narrative in online education. In: Richard Ferdig et al. (eds.), Pro-
ceedings of Society for Information Technology & Teacher Education Inter-
national Conference 2004, 352–355. Chesapeake, VA: AACE.
Bell, Alice
2010 The Possible Worlds of Hypertext Fiction. Basingstoke, UK: Palgrave Mac-
millan.
Bratteteig, Tone
2008 Does it matter that it is digital? In: Knut Lundby (ed.), Digital Storytelling,
Mediatized Stories: Self-representation in New Media, 271–284. New York:
Peter Lang.
Brockmeier, Jens and Donald Carbaugh (eds.)
2001 Narrative and Identity Studies: Autobiography, Self and Culture. Amsterdam/
Philadelphia: John Benjamins.
Chatman, Seymour
1990 Coming to Terms: The Rhetoric of Narrative in Fiction and Film. Ithaca, NY:
Cornell University Press.
Cherny, Lynn
1999 Conversation and Community: Chat in a Virtual World. Stanford, CA: CSLI.
Ciccoricco, David
2007 Reading Network Fiction. Tuscaloosa, AL: University of Alabama Press.
Daiute, Colette and Cynthia Lightfoot (eds.)
2004 Narrative Analysis: Studying the Development of Individuals in Society. Lon-
don: Sage.
Danet, Brenda
2001 Cyberpl@y: Communicating Online. Oxford, UK: Berg.
Danet, Brenda and Susan C. Herring (eds.)
2003 The Multilingual Internet: Language, Culture, and Communication in Instant
Messaging, Email, and Chat. Special issue, Journal of Computer Mediated
Communication 9(1). http://jcmc.indiana.edu/vol9/issue1/
Danet, Brenda and Susan C. Herring (eds.)
2007 The Multilingual Internet: Language, Culture, and Communication Online.
New York: Oxford University Press.
De Fina, Anna
2003 Identity in Narrative: A Study of Immigrant Discourse. Amsterdam/Philadelp-
hia: John Benjamins.
De Fina, Anna and Alexandra Georgakopoulou
2008 Analysing narratives as practices. Journal of Qualitative Research 8: 379–387.
De Fina, Anna, Deborah Schiffrin, and Michael Bamberg (eds.)
2006 Discourse and Identity. Cambridge, UK: Cambridge University Press.
Douglas, J. Yellowlees
2000 The End of Books – or Books without End? Ann Arbor, MI: University of Mi-
chigan Press.
Ensslin, Astrid
2007 Canonizing Hypertext: Explorations and Constructions. London: Continuum.
Fludernik, Monika
2009 Introduction to Narratology. Abingdon/New York: Routledge.
Freeman, Mark
2010 Hindsight: The Promise and Peril of Looking Backward. New York: Oxford
University Press.
Georgakopoulou, Alexandra
1997 Self-presentation and interactional alignments in e-mail discourse: The style-
and code switches of Greek messages. International Journal of Applied Lin-
guistics 7: 141–164.
Georgakopoulou, Alexandra
2003 Computer-mediated communication. In: Jef Verschueren, Jan-Ola Östman, Jan
Blommaert, and Chris Bulcaen (eds.), Handbook of Pragmatics, 1–20. Am-
sterdam/Philadelphia: Benjamins.
Georgakopoulou, Alexandra
2004 To tell or not to tell?: Email stories between on- and off-line interactions. Lan-
guage@Internet 1, article 1. http://www.languageatinternet.org/articles/2004/36
Georgakopoulou, Alexandra
2006a Small and large identities in narrative (inter)-action. In: Anna De Fina, Debo-
rah Schiffrin, and Michael Bamberg (eds.), Discourse and Identity, 83–102.
Cambridge, UK: Cambridge University Press.
Georgakopoulou, Alexandra
2006b Thinking big with small stories in narrative and identity analysis. Narrative In-
quiry 16: 129–137.
Georgakopoulou, Alexandra
2006c Postscript. Computer-mediated communication in sociolinguistics. Journal of
Sociolinguistics 10(4): 548–557.
Georgakopoulou, Alexandra
2007 Small Stories, Interaction, and Identities. Amsterdam/Philadelphia: John Ben-
jamins.
Georgakopoulou, Alexandra
2008 “On MSN with buff boys”. Self- and other-identity claims in the context of
small stories. Journal of Sociolinguistics 12(5): 597–626.
Goodwin, Marjorie H.
1990 He-said-she-said: Talk as Social Organization among Black Children. Bloom-
ington, IN: Indiana University Press.
Hartley, John and Kelly McWilliam
2009 Story Circle: Digital Storytelling around the World. Malden, MA: Wiley-
Blackwell.
Hayles, Katherine
2008 Electronic Literature: New Horizons for the Literary. Notre Dame, IN: Univer-
sity of Notre Dame Press.
Herman, David
2009 Basic Elements of Narrative. Oxford, UK: Wiley-Blackwell.
Herring, Susan C.
2004 Slouching toward the ordinary: Current trends in computer-mediated com-
munication. New Media & Society 6(1): 26–36.
Heyd, Teresa
2009 Email Hoaxes. Form, Function, Genre Relations. Amsterdam/Philadelphia:
John Benjamins.
Hoffmann, Christian
2008 “Once upon a blog”. Paradigms of narrative interaction. Paper presented to the
Stylistic Workshop – The State of the Art, ISLE 2008, Freiburg, 8–11 October.
Eisenlauer, Volker J. and Christian Hoffmann
2010 Once upon a blog … storytelling in weblogs. In: Christian Hoffmann (ed.),
Narrative Revisited, 79–108. Amsterdam/Philadelphia: John Benjamins.
Hymes, Dell
1981 “In Vain I Tried to Tell You”: Essays in Native American Ethnopoetics. Phil-
adelphia: University of Pennsylvania Press.
Jucker, Andreas
2010 “Audacious, brilliant!! What a strike!” Live text commentaries on the Internet
as real-time narratives. In: Christian Hoffmann (ed.), Narrative Revisited,
57–78. Amsterdam/Philadelphia: John Benjamins.
Kolko, Beth E.
1995 Building a world with words: The narrative reality of virtual communities.
Works and Days 13(1–2): 105–126.
Labov, William
1972 Language in the Inner City. Philadelphia: University of Pennsylvania Press.
Labov, William and Joshua Waletzky
1967 Narrative analysis: Oral versions of personal experience. In: June Helm (ed.),
Essays on the Verbal and Visual Arts, 12–44. Seattle: University of Washington
Press.
Lambert, Joe
2006 Digital Storytelling: Capturing Lives, Creating Community. Berkeley, CA:
Digital Diner Press.
Landow, George
1997 Hypertext 2.0: The Convergence of Contemporary Critical Theory and Tech-
nology. Baltimore, MD: Johns Hopkins University Press.
Lundby, Knut (ed.)
2008 Digital Storytelling, Mediatised Stories: Self-representations in New Media.
New York: Peter Lang.
Mitra, Sanjit
1999 Minimally invasive education for mass computer literacy. CSI Communi-
cations (June): 12–16.
Morris, Adelaide and Thomas Swiss
2009 New Media Poetics – Contexts, Technotexts, and Theories. Cambridge, MA:
MIT Press.
Ochs, Elinor and Lisa Capps
2001 Living Narrative. Cambridge, MA: Harvard University Press.
Page, Ruth E.
2008 Gender and genre revisited: Storyworlds in personal blogs. Genre: Forms of
Discourse and Culture XLI: 151–177.
Page, Ruth E.
2009 Trivia and tellability: Storytelling in status updates. Paper presented at the
Centre for Language, Discourse & Communication Research Seminars, King’s
College London, 14 October.
Page, Ruth E.
2012 Stories and Social Media. New York/London: Routledge.
Polkinghorne, Dan
1988 Narrative Knowing and the Human Sciences. Albany, NY: State University of
New York Press.
Riessman, Cathy
2008 Narrative Research Methods in the Social Sciences. London: Sage.
1. Concepts of genre
Genre is an active term in several areas of language study: not only in pragmatics
but also in corpus studies, in Critical Discourse Analysis, and in areas more peda-
gogically oriented, such as English for Special/Academic Purposes and Sydney
School Systemic-Functional Linguistics. The term also figures importantly in lit-
erary studies, rhetoric, and information and organisational studies. The different
methodologies and research objectives of these fields result in different concep-
tualisations of genre itself. Yet one might say that two principles are shared across
notions of genre. First, genre is a typifying concept: Instances of utterance re-
semble one another and can be classified or recognised thereby. In literary studies,
Cohen (1986) has said that genre is principally a classifying activity and socio-
historically contingent. Second, genre is a phenomenon at the interface of language
and sociality. Even general classifications, such as “academic” or “literary”, point
to speakers’ involvement in a social activity.
Because the emphasis on and application of these principles are uneven in
research practice, genre can have different profiles in different disciplines. Es-
pecially, the level at which genre is named can fluctuate. In some fields, such
as Sydney School Systemic-Functional Linguistics, “narrative” or “recount” can
be designated as a genre, whereas in others, such as recent literary theory, “village
stories, 1820s-1850s” is presented as a genre (Moretti 2005). As this chapter will
show, genre research in the pragmatics of computer-mediated communication
(CMC) has tended to the lower levels. Fluctuations from higher to lower, and even
very low, are still an issue for genre study, however, as Herring (2007) notes, mak-
ing genre classification “imprecise”.
Virtanen (2010) offers relief from this stubborn condition by proposing that
discourse types and text types, on the one hand, and genres, on the other hand, are
categories that address different language phenomena and should be distinguished.
Discourse types (with focus on function) and text types (with focus on form) are
cognitively heuristic: They are “templates” for production and cues for reception.
The same has been said about genres. But text types and discourse types, while
“dynamic” in their interplay, are relatively “abstract” and “prototypical”, accord-
ing to Virtanen, whereas genres are concretely situated. The temptation might be to
say that genre is simply lower on the scale of generality than type (discourse or
text), but Virtanen is asking for a more fundamental re-thinking of this fluctuation –
different dynamics and different cognitive manifestations, not simply more general
vs. more specific. (See Heyd 2009 as discussed below for an application of com-
parable, albeit not identical, principles for managing levels of generality.)
Inseparable from the level at which genre is named is the question of the closed
or open set: Where genre has been named at a very high level, the set is closed, li-
mited to a few timeless universals. When the level of naming is lower, the set opens
to history – as the appearance, and then the disappearance, of the literary genre
“village story” shows as a response to social, economic, and technological change.
Change being a widely perceived feature of CMC, genre research in the pragmatics
of the Internet has tended, in practice, to assume the open set.
Harder to characterise is researchers’ attitude towards genre itself. Tradition-
ally, and perhaps currently, genre has been regarded as what is conventional –
even, as Puschmann states, referring to Swales (1990), what is “fixed” (Puschmann
2009: 81). In literary study, especially, there has been a tradition of imagining
genre as what is to be defied or transcended, when aesthetic aspirations challenge
convention. It is possible, however, to regard genre as what is ceaselessly differ-
entiating, producing not sameness but the variety of “spheres of activity” (Bakhtin
1986): genre, that is, as “difference” (Giltrow 2010). Popular claims for CMC have
often assumed its potential to break away from what is fixed and standard, to over-
turn even regularity itself, and to entertain the unheard-of. The research reported in
this chapter, however, shows that CMC hosts the transfer, emergence, and trans-
formation of genres – not their elimination.
As most recent research in the pragmatics of the Internet has settled to low le-
vels of generality in naming genre, and has both anticipated and recorded an open
set, pragmatics research is most in keeping with rhetorical theorisations of genre,
and these provide a background to this chapter’s review of research. Arguably, rhe-
torical study is the field that has in our time paid the most sustained attention to
genre, developing from U.S. proposals for “situational rhetoric” in the 1970s
(Bitzer 1968, 1980) and elaborated on many fronts since then. Rhetorical theories
of genre bring strong claims for the priority of action over form: Genre is known
principally not by regularities of form but by recurrence of action in response to
socio-cultural situation. People speak in relatively similar ways not to replicate
formal patterns but to participate in the spheres of activity observed by Bakhtin,
these being local to people’s immediate experience. Accordingly, rhetorical con-
ceptions of genre include the notions of exigence (first in Bitzer 1968, 1980) –
people’s mutual recognition of a need or a reason for speaking in a certain way,
here and now – and motive: Amongst those who recognise the exigence, some will
be motivated to speak up. So, for example, many in a certain time and place may
recognise the reasonableness of composing a political blog, while fewer will be
motivated actually to do so. At another time and place, the exigence may be indis-
cernible to all: The utterance may seem a freakish outburst or a quaint archaism. Or
it may simply have no counterpart. Rhetorical genre theory insists thereby on the
phenomenological volatility of the open set, as well as its relative stability.
Rhetorical theorists of genre have argued that genre is a fusion of form and situ-
ation (Campbell and Jamieson 1978; Miller 1984), language features being func-
tionally responsive to features of situation. A strong version of this claim is that
genre is not a linguistic import to a social situation, or a handy instrument for the
accomplishment of goals, but a blend of language and situation: Genre is at once
typical speech1 and recognised situation. For example, an emailed thank you for
dinner-table hospitality is not an instance of the genre thank-you note unless there
has been a dinner party, an accompanying degree of relationship between host and
guest, and so on, all these conditions activating sensations of exigence and motive,
mutually recognised by sender and recipient of the thank-you message.
Close (but not identical) to rhetorical genre theory are Swales’ (1990) propos-
als for “genre analysis”. Whereas rhetorical theory has identified “social action”
(Miller 1984) as the basis for genre status, Swales names communicative purpose
as the principal criterion for “a collection of communicative events” to “[turn] into
a genre” (47). But communicative purpose is not easy to isolate: It can be multiple;
it can disappear altogether (46–47). Swales himself, with Askehave, has revisited
the notion of communicative purpose (Askehave and Swales 2001), urging scepti-
cism of the terms in which both analysts and genre-users identify it. To be sure, so-
cial action is like communicative purpose in being hard – or even harder – to pin
down. But whereas communicative purpose might ask, of a political blogger, for
example, “what is your purpose in saying this?” and hear “to inform”, “to debate”,
“to express an opinion”, social action would ask, “what makes you want to inform,
debate, express?” The answer would be less accessible, more tacitly embedded in
situational exigence, less isolable than communicative purpose as an aspect of
genre.
In his earlier proposals in which he finds it possible to isolate purpose, Swales
also finds it possible to separate form from motive. He defends this formalism on
the basis of the success of international education: Students can practise “discourse
conventions” without experiencing the motives which culturally sanction those
conventions (or, in Swales’ terms, without being “assimilated”); discourse conven-
tions can be employed purely in “an instrumental manner”, that is, without com-
mitment to the sphere of activity in which the exigence is recognised (30–31).
While these are both possibilities, as are deception and pretending, as discussed
below, they may be peripheral rather than core to genre. Nevertheless, as we will
see, Swales’ genre analysis offers important formal measures.
Similarly, Herring’s (2007) proposal for classifying CMC genres offers means
of analysing regularities of occurrence, means lacking in rhetorical theory. In Her-
ring’s proposal, communicative purpose, rather than being genre-definitive, is one
of a range of “situation factors”, joining, for example, participant characteristics,
activity, and norms. Herring’s proposals advance analysis of CMC by methodically
recognising technological features of situation (e.g., synchronicity, persistence of
transcript, anonymity) as pragmatic conditions, thereby recognising technology
itself as a social circumstance and identifying medial parameters which the re-
search reported in this chapter shows to be constitutive of genres. At the same time,
as Herring notes, the results of analysis still need to be reconciled with speakers’
consciousness of the categories discovered by analysis. The research reported
below will show many ways in which speakers’ experience of genres – their sen-
sation of exigence and motive – can be extrapolated from their observable lan-
guage behaviours.
Genre study of CMC may be particularly opportune. At this point in e-history,
citizens of the Internet are still constructing the pragmatic dimensions of genres;
adapting and instituting the conditions of proximity and distance, synchronicity
and sequence, community solidarities and estrangements. Possibly these still-
exposed efforts will retire as newness consolidates into custom. Or they may
persist, owing to CMC’s technological profile. To a degree greater than other
media, including broadcast media, computer media are open to technological
design and re-design in response to techniques of use, and they call for the kind of
faceted analysis Herring proposes.
Genre study of CMC has also exposed an undefeatable sociality. Even as par-
ticipants in CMC experience technical lags and accelerants, they call on rich social
contexts both enduring and evolving, as the research reported below shows. Draw-
ing on these dense contexts, communicative behaviours are typifiable and recog-
nisable to the naked eye. Classification of these types shows not simply that people
develop functional ways of speaking under newly enabled conditions, but that they
construct mutual knowledge of these ways: They not only know and recognise a
way of speaking but also know that others know this speechway. While classifi-
cation – the professional follow-up on folk recognition – is not an end in itself, not
the reason for genre, it is an instrument for tracing the sociality of linguistic con-
sciousness.
Section 2 selects research which, by investigating pragmatic features, dis-
covers digital genres, indicating areas for future inquiry into the pragmatics of
genre in CMC. Section 3 reports research for which CMC genre is itself the object
of inquiry.
2. Genre-related pragmatics
nalistic scene. They are then able to trace the mutation and also the persistence of
the form/function profile as being in step with the mutation or persistence of what
rhetorical genre theory would call the social action of the press release: the filtering
of corporate or organisational messages through the news genres.
In the press releases of two different distributors they find contrasting out-
comes of the migration of the genre to an online environment. One agency issues
press releases whose pragmatic profile keeps some preformulation features but
also adds direct address to the reader, including imperatives, and emphasis by
means of features such as all caps (2). More “selling” than “telling”, these releases
also include links to corporate sites. In addition, they are archived and available to
searches. In other words, they eliminate the “middleman” – the journalist – in their
address to the public and function beyond being today’s or tomorrow’s news. The
other agency, however, honouring its service to news institutions, issues releases in
traditional preformulation packages. Strobbe and Jacobs interpret their findings as
showing that one route for the migrating genre by-passes “the gate-keeping role
traditionally played by journalists in handling paper press releases” (1), while an-
other migration route conveys established practices to a “new location”: “[I]t
seems as if for the time being online news remains – to some extent at least – tied in
with traditional media” (2). As well as showing these different migration paths,
their research also reveals what news genres do, whether online or offline: They
mediate utterances of interested parties (companies, organisations) through the
positions of other institutions (news producers). This is a social action of the news
genres, one that comes to light by means of research in the pragmatics of CMC.
With its historical perspective on a migrating genre, Wyss’s “From the bridal
letter to online flirting: Changes in text type from the nineteenth century to the In-
ternet” (2008) offers much to the study of CMC genres. Wyss notes that study of
the pragmatics of historical text types requires the “reconstruction of the norms
and rules of a sociohistorical context that is different from one’s own contemporary
milieu” (226): in other words, the reconstitution of the exigence and motive which
manifest in a genre. Wyss’s account of the 19th-century Brautbriefe (bridal letter) is
exemplary for revealing the layered complexity of such motives, a complexity far
beyond a compact communicative purpose. Composing the Brautbriefe, engaged
couples “[performed] the bourgeois self” (230), stylistically demonstrating to one
another their eligibility for matrimony. These demonstrations in turn were mean-
ingful through their articulation with contemporary class-indexed concepts of
gender and of language itself. Evidently concentrating many social energies, this
high-performance genre also generated its own “meta-genre” (Giltrow 2002): ad-
visories and rule-giving which coincided imperfectly with actual practice (a mis-
match that could be compared with that which Herring 1996 finds in gender-stereo-
typical commentary and advisories on CMC).
CMC is one amongst several technological amplifications of lovers’ commu-
niqués. The postcard, the telephone, and the telegraph were others, and Wyss notes
that the love letter was and is combined in its use with “other oral and written com-
munication” (233). Whatever the combination, Wyss observes that writers of both
the Brautbriefe and today’s electronic love letters get to know each other through
their correspondence. Crucially, however, the two eras’ correspondents start out
lacking knowledge of each other in different ways, the difference playing out along
both social and technological dimensions. Whereas bridal-letter exchanges ratified
the bourgeois identities of the correspondents, corroborating estimates by a circle
of family and friends, the Internet confronts lovers with (once again) an “enormous
and bewildering lack of context” (244): Who is out there? Accordingly, the “flirt
email” develops, pragmatically “[vague]” to maintain a “non-committal” stance
(244–245) while the correspondents conduct a reconnaissance of the “enormous”
(de)context. Inheriting some forms from “the traditional Western discourse of
love”, the flirt letter is in addition “cryptic”, suggesting that the relevance-weak
patterns other researchers have found in CMC may have many functions and may
be a resource for the development of Internet genres (see Herring’s chapter on rel-
evance in this volume). As well, Wyss asks “whether a flirt letter can be called a
love letter”: Is it a new genre or a variation on, or adaptation of, the old? Her own
analysis of rhetorical situation sets a standard for thoroughness in developing
measures for newness: Has the rhetorical situation changed? To what extent, along
which dimension?
Taking a cross-cultural (Turkish/British) perspective, Hatipoğlu’s (2007) study
of Calls for Papers shows that even seemingly modest differences in socio-techni-
cal aspects of the rhetorical situation can have measurable pragmatic effects in the
politeness features he analyses. Graham (2007) also focuses on (im)politeness, but
in a list situation where possible face threats are more severe than the risk of bur-
dening recipients with uninteresting Calls for Papers. Graham’s research site is
“ChurchList, an unmoderated, tightly-knit email discussion list which provides a
forum for the discussion of issues affecting the Anglican Church” (745) – a context
that is dense enough to engage participants not only in deep affinities but also in
serious disagreements. Graham’s study takes us farther into the socio-communi-
cative environment of lists than any other research discussed in this review, show-
ing that at the heart of ChurchList’s engagement are genres and, in this case, one
genre in particular: the request-for-prayer. List participants experience exigence,
the feeling that a certain kind of utterance is called for, here and now. And they ex-
perience equally the exigence of the up-take (Freadman 2002): fitting, conscien-
tious response. Extending the experience of exigence are politeness constraints,
some of which are technologically conditioned (747): the requirement that the sub-
ject line should specify both the initiating genre and its up-take (this message is for
those ready to respond to a prayer request sympathetically), thereby managing the
audience fluidity of the e-list; and the sanction against “blatting” (copying a private
message into a public arena), thereby mitigating the politeness hazards of Reported
Speech in an online setting. Other politeness constraints are such as would be pres-
ent in offline settings – the conscientiousness of, for example, expression of sup-
port rather than criticism in responding to a troubled co-religionist – and indicate
the transfer of attitudes and interests to the online world. One might speculate as
well that the online prayer request articulates with offline scenes of activity, such as
prayer meetings and Bible-study sessions.
Graham examines a ChurchList episode in which a speaker flouts these norms
by criticising a supplicant, directing her response away from the supplicant by
means of third-person pronouns and, in addition, copying private messages into
her responses. Appearing under the original subject line, the norm-offending up-
take shocks discussants whom the subject line has led to expect a sympathetic
message. Graham speculates that the offending participant may be e-incompetent.
A genre-theoretical explanation would suggest in addition that she does not experi-
ence the same exigence as the other participants. Moreover, Graham’s analysis of
the flurry of meta-generic activity which follows the offending message – rebukes
and attempts to restore the site’s cordiality – shows the capacity of a community to
regard a genre as a pivot of its affinity. The meta-generic discussions implicitly re-
store the request/support pair and explicitly specify rules for subject lines.
In her study of medical spam (offers of products promising, for example,
weight loss, allergy relief, or improved sexual performance), Barron (2006) con-
sults genre theory as practised in English for Special Purposes (citing, e.g., Swales
1990 and Bhatia 1993) and analyses “moves”: discursive sub-goals of a communi-
cative purpose. Barron’s study is a major contribution in, first, comparing the prag-
matic parameters of pre- or non-digital genres – direct mail, promotion letters –
with those of their digital version, spam. While direct mail goes to a wide audience,
spam recipients, owing to techniques for harvesting e-addresses, comprise an even
wider audience, and one relatively less known to the senders. Barron’s analysis
then exposes the feedback sequence of technological innovation and social/legal
responses that has resulted in the spam genre. The audience extension enabled by
address-harvesting has led to recipients’ dislike of these messages (evidently
greater than their dislike of direct mail) and resentment of the cost of receiving
them: in other words, to technologically-induced social responses. As the message
type gets a bad reputation, worthy products desert it. Spam filters (a technological
response) are developed, as are legal constraints (a social response). In turn, the
genre adapts on both technological and social fronts: Anonymity devices are de-
signed to evade prosecution; discursive features develop to camouflage the mess-
age type. The rapidity of this feedback/adaptation cycle may appear to be peculiar
to CMC, but one might point to comparable episodes in pre-digital history, such as
the swift legal-social responses to the spread of pornographic images on postcards
in the early 20th century (Sigel 2000).
While study of spam exposes the shady side of online life, Puschmann’s (2009)
study of Wal-Mart’s corporate blogging investigates another subterfuge: the appro-
priation of the “positive associations” of personal blogging for corporate ends. As
Puschmann points out, the “abusability” of blogging is evidence for the conver-
gence of form and function (51): in other words, the existence of a genre. Formal
features aggregate functionally in a scene of activity and become indications of that
activity – in the case of personal blogging, conversational invitations to interaction
and contact. The site “Life at Wal-Mart” replicates the technologically-formatted
indications of blogging, but then “abuses” them in the service of a corporate social
action: narrowly, dissemination of the message that Wal-Mart is a good employer;
broadly, a response to a complex political, socio-rhetorical situation.
While the technical (or “hard” [58]) forms signal “blog”, Puschmann’s inves-
tigation shows that other (or “soft”) forms are not in keeping with the patterns he
establishes through corpus analysis of pragmatic features of blogs. Use of ‘you’,
for example, is ambiguous in the 52 “Life at Wal-Mart” postings in ways in which
it is not in the larger corpus of blogs Puschmann has studied; Wal-Mart postings
also deviate from those in the larger corpus in favouring past tense forms of verbs.
In addition, they neither quote other blogs nor provide links – but neither do the
personal blogs in Herring, Scheidt, Bonus, and Wright’s sample (2005, discussed
below) match the stereotype of the blog as a genre characterised by links. More
telling for both the Wal-Mart case and for genre study, is the “astounding” (66)
similarity of the Wal-Mart blogs. For Puschmann, the key point is that this simi-
larity shows evidence of editorial oversight, which spoils what he takes to be char-
acteristic about blogs, that is, that they are owned by their authors. For genre study,
the bigger news may be that pronounced similarity actually makes instances sus-
pect and suggests tampering rather than authenticity. Traditional views of genre
emphasise sameness, and Puschmann himself regards genres as “fixed and un-
changing” (81). But the false air of the “astounding” sameness of Wal-Mart blogs
suggests instead that genres, in their real rather than Wal-Mart lives, are characte-
rised by variability rather than uniformity (cf. Devitt 2009 for the view that genres
are, above all, formally variable).
Heyd’s (2009) study extends our vantage on genre’s masquerading capabilities
by adding to faux blogs and evasive spam the “email hoax”. (See also Heyd’s
chapter in this volume.) Neither her findings nor her theoretical proposals aim to
expose deception or mimicry, however. Instead, Heyd is tackling the problem of
newness in genres: If genres are what is accustomed and what activates language-
users’ expectations, how can brand-new genres be explained? (See Miller and
Shepherd 2009 for a rhetorical approach to this question.) As Heyd points out,
while people report experiencing newness in the Internet world, research on the
whole has failed to turn up much that is thoroughly novel, finding instead anteced-
ents for what is felt to be new, or finding that what is new is hybrid or “adapted” in
Crowston and Williams’ (2000) term, rather than entirely new. To account for this
mixed scene of continuity and break-away, at the same time theorising issues in
function-vs.-form debates, Heyd proposes a model that places at a high level of
generality “communicative functions”, which, she says, will be “few and funda-
mental” and thus highly stable (240). Research which focuses on genre-as-func-
tion will therefore not be likely to find wholly new manifestations, and will tend to
uncover likenesses. Difference will appear as the level of generality drops from
these function-defined “super-genres” to their rhetorical instantiation, “branching”
into situations where it is form which differentiates “sub-genres” from one another
(240–241).
Applying this model to the email hoax, Heyd demonstrates that email hoaxes
are not, in fact, direct descendants of pre-digital con games and swindles. Their
genealogy is instead traceable from the super-genre “officelore” through “digital
folklore”: message types which keep open friendly lines of communication
amongst workmates by circulating, for example, ludicrous photos, amusing lists,
or political petitions. Providing a “ready-made” source of “tellable” material and
“[reinforcing] digital social networks” (251–252), digital folklore shares com-
municative function with officelore, but in its execution of this function is distin-
guished by its use of technological affordances, such as address-storing, multiple-
addressing capacities, and the forwarding function. It is also distinguished from of-
ficelore in being more meta-discursive (253). The generic distinction of digital
folklore types is popularly attested, as Heyd notes, by websites devoted to collect-
ing them (254).
Heyd’s proposals are a notable contribution to genre theory and offer an oppor-
tunity to reconcile social action and communicative purpose: Social action may be
the lower level of generality – structured, rhetorical, and open to the immediate ex-
perience of language users – whereas communicative function, at a higher but still
socio-historical level of generality, may be more accessible to analysts than to lan-
guage users.
In Locher’s (2006) book-length study of online advice-giving by health ser-
vices at a U.S. university (see also Locher’s chapter in this volume), antecedence is
a less vexing issue than it is for the study of email hoaxes. The letters issued by
“Lucy”, the persona of health educators who answer questions submitted electroni-
cally to a website, participate in “the well-established tradition of an advice col-
umn” (17) and other non-digital genres: The radio call-in show is a close relative;
more distant is the nurse’s visit to the new mother. All these genres engage the
pragmatic issues involved in mitigating or boosting the force of an expert’s inter-
vention in the advisee’s affairs. Internet advice-giving also shares with radio
call-in shows and print advice columns the border between private and public. Ad-
vice-requesters refer to intimate matters yet are addressed in a wide public arena,
and responses are in each of these genres pragmatically geared not only to the ad-
vice seekers (and their face wants) but also to overhearers. Locher analyses prag-
matic arrangements for overhearers as being aimed at providing them with useful
information.
On the whole, transfer of public advice-giving to the Internet seems to have
been, in the case of the “Lucy” site, a stable one. The audience is digitally extended
beyond the querying/advising pair, but it is not evident that the extension has a rhe-
torical effect different from that of print media, although comparative study based
on Locher’s findings may reveal differences. In genres giving personal advice pub-
licly, the crucial medial distinction is, once again, in the durability of the record.
Lucy’s answers are archived, and information-seekers can search the archive in-
stead of submitting a question or can be directed by Lucy, in the presence of over-
hearers, to search. In turn, while the record persists, it is also revisable. Lucy up-
dates the archive to make the consultation process more efficient: pragmatically
accessible and also refreshed with new medical research.
Emigh and Herring (2005) also analyse a traditional print genre – the encyclo-
paedia entry – on its arrival at online destinations. They want to know what
happens to the genre under different technologically-enabled conditions of produc-
tion: on the one hand, longstanding practices of reference publishing – expert con-
tributions solicited and standardised by editors – continued in the online version of
the Columbia Encyclopedia; on the other hand, the two distinct collaborative on-
line systems of Wikipedia and Everything2. While both Wikipedia and Every-
thing2 are “democratic” in inviting anyone rather than only experts to contribute,
the systems rely on different “quality” mechanisms: All Wikipedia users are poten-
tially authors and can revise others’ entries, whereas Everything2 entries cannot be
revised and depend instead on ratings to monitor quality. Following Biber (1988),
Emigh and Herring analyse entries for formality (noun-formation suffixes) and
informality (personal pronouns and contractions), these features having been
“validated as [indicators] of genres”: Factor analysis has revealed “frequencies of
linguistic features to empirically identify different genres – what [Biber] calls ‘reg-
isters’ – of discourse” (5). Revised and monitored by “good users” (2), the Wiki-
pedia entries self-standardise towards the formality norms of the print encyclo-
paedia entry. In contrast, the Everything2 entries remain variable and less formal,
while most informal are the “discussion” entries on Wikipedia. Emigh and Herring
identify Everything2 as an example of what Crowston and Williams (2000) call an
emergent genre: It is a “hybrid product of the Web” (9), a blend of the online genres
“discussion” and “repository of knowledge”.
Most compelling, however, for students of genre may be the standardisation ef-
fect on Wikipedia, an effect coincident with Wikipedia’s growing credibility. Ask-
ing why the Wikipedia entries end up being so like traditional encyclopaedia en-
tries, despite the democracy of the authoring systems, Emigh and Herring write
that users “appropriate norms and resources” from print traditions, and a small
group of “good users” then enforces those norms (9). Rhetorical research would be
interested in the motivations of the “good users” and also in the exigence experi-
enced by all contributors to either Wikipedia or Everything2. The motives of Wiki-
pedia and Everything2 contributors may differ, as may those of their readers, even
as entries from both sites could be captured under the higher-level category “re-
pository of knowledge”.
The research reviewed in this chapter suggests that pragmatic methodology has
much to offer the study of genre. Once data sets are defined sufficiently to recover a
scene of social action – flirtatious overtures, for example, or offers of medical
products or medical advice, or announcements of conferences, to name just a few
of the actions studied – the analysis of features illuminates not only linguistic
manifestations but the social situations themselves. We have seen that systematic
study of, for example, pronouns or (im)politeness or relevance warrants can ex-
pose local characteristics of the socio-rhetorical situation. In addition, results of
pragmatic study challenge the notion that genre is what is the same, what is for-
mally fixed or even rigid, and what is rule-governed, at best instrumental, at worst
imposing and exacting of conformity. With its capacity to expose the coordinates
of language-users’ social positions and inclinations, pragmatic study can establish
genre as what is versatile and variable, functional because flexible, and subjective
rather than instrumental.
In addition to being productive for genre theory generally, pragmatic study of
CMC genres also uncovers conditions unique to episodes of technological change.
At first, these were characterised as depleting context: jarring departures from co-
presence. But research soon showed users rapidly adapting to technological condi-
tions and reconstituting social context. Technological design, however, still oc-
cupies research attention (for a range of genre-theoretically informed comments on
design issues see, for example, Agre 1998; Beghtol 2000; Hendry and Carlyle
2006; D. Stein 2006; Toms 2001; Watters and Shepherd 1997) and complicates
CMC genre study. To what extent can software designers be considered co-partici-
pants in the life of a genre – ghost-writers or silent partners? What is the signifi-
cance for genre of a pragmatic regularity that is technologically formatted rather
than locally originating? How do these formatted regularities play into partici-
pants’ experience of exigence and motive? By providing a clearer view of the so-
ciality of the technology itself, Herring’s (2007) scheme for classification can help
to address such questions and to further the work of Askehave and Nielsen (2005)
and others who disentangle “hard” from “soft” dimensions of CMC. Drawing on
rhetorical theory and also proposing an “ecology” for CMC, Erickson (2000), for
example, finds that on the experimental “Babble” chat system, conversation
reaches genre status, thanks to users’ learning the system’s capacity. Long separ-
ated in research practice, conversation (“proto-genre”, in Swales’ [1990] term,
“primary” genre in Bakhtin’s [1986]) and genre may draw closer in CMC owing to
the mutual shaping of technologies and their uses.
Genre is a means by which language users know one another, finding familiar-
ity even amongst strangers but also sensing tremors to common ground when
facing the unexpected or unrecognisable. As new technologies have expanded
audiences incalculably and unpredictably, they have introduced unknown speakers
and listeners with unknown motives, bringing about a subduction of rhetorical situ-
ation and sharp sensations of newness. At the same time, designers’ assumptions
about motives for communication may not align immediately with language users’
experience – incurring, again, a feeling that something has changed, some foreign
element has intruded. Both conditions – the inscrutable audience and the misalign-
ment of design and motive – appear to be transitional, however, as the case of
relatively early (2003) weblogs, studied by Herring et al. (2005), suggests. De-
scribing blogging technology as having “bridged” a gap between the multimedia
arena of relatively static HTML (hypertext markup language) websites and the in-
teractivity of other CMC, Herring et al. also find (a) that the type to which most
blogs belong is a descendant of a very old genre, the hand-written diary, now pub-
licised by new technology; and (b) that the uses of the “socio-technical format” of
the blog are so diverse as to make it “no longer meaningful to speak of weblogs as a
single genre” (164).
This is a complex picture. While language users introduce old forms to new
ground, their diverse uses of the new technology also sub-divide the new ground
into socio-rhetorical localities from which arise the differentiations that are genres
themselves. The technology can still name broad domains of use, as when Myers
(2010) writes on The Discourse of Blogs and Wikis while finding manifold differ-
entiations in the discourse. Alongside this broad naming, however, the process of
technical innovation and adoption is evident in a typical series of classifications:
First, communicative activity is named for the technology itself (e.g., email), then
as a sub-type (e.g., political email message), then as a full-fledged genre (e.g.,
e-petition). While the cline of newness proposed by Crowston and Williams (2000)
is confirmed by research, it also needs to be reconciled with the process revealed in
the classification series. At what stage in the process of technological adoption has
the analyst taken the snapshot that exposes a reproduced genre, or an adapted or
emergent one?
Pragmatics research pursuing newness and isolating the criteria for emergent
status may not only shed light on computer-mediated life but also revive early de-
bates in rhetorical genre study about antecedence. For example, when Herring et
al. show that blogs, “rather than deriving from a single source” (2005: 160), de-
rive from many, we can anticipate that the outcome of the multiple input will also
be multiple: local differentiations processed through one technological junction.
Are such congregations and dispersals peculiar to digital genres? Equally, we
could ask of Herring et al.’s finding regarding the blog as occupying a “bridging”
position whether such genre-productive positions can be found in the non-digital
world, or whether the sedimentations of traditional speechways have closed all
the gaps.
So far, the reckoning of newness has tended to be one-way: Do CMC genres
differ from or replicate offline genres? Awaiting pragmatic study is the possibility
that there is transfer in the other direction, a flow-back from online to offline
genres. Equally compelling would be study which thoroughly tests claims for rad-
ical newness against historical episodes. I have mentioned the advent of the post-
card – a technological, social, and legal breakthrough in its time – as a comparison.
There will be others, and Wyss’s study of love letters across two centuries should
be inspiration to look for them (see also Burgess 2009 comparing the commen-
taries surrounding late-18th-century print technologies with those accompanying
late-20th-century CMC).
Where genre is regarded not as protocol-conforming fixities but as versatile
bundles of interactive attitude and orientation, genre study is an effective way of
investigating the newness of CMC. Classification measures the extent and quality
of newness, while analysis of social action reveals users’ experience of exigence,
including their social motives for taking up new technologies and their mutual rec-
ognition of these motives.
Note
References
Agre, Philip
1998 Designing genres for new media: Social, economic, and political contexts. In:
Steven G. Jones (ed.), Cybersociety 2.0: Revisiting Computer-Mediated Com-
munication and Community, 69–90. Thousand Oaks, CA: Sage.
Askehave, Inger and Anne Ellerup Nielsen
2005 Digital genres: A challenge to traditional genre theory. Information Technol-
ogy & People 18(2): 120–141.
Askehave, Inger and John Swales
2001 Genre identification and communicative purpose: A problem and a possible
solution. Applied Linguistics 22(2): 195–212.
Bakhtin, Mikhail M.
1986 The problem of speech genres. In: Michael Holquist and Caryl Emerson (eds.),
Vern McGee (trans.), Speech Genres and Other Late Essays, 62–102. Austin:
University of Texas Press.
Barron, Anne
2006 Understanding spam: A macro-textual analysis. Journal of Pragmatics 38(6):
880–904.
Beghtol, Clare
2000 The concept of genre and its characteristics. Bulletin of the American Society
for Information Science 27(2): 17–20.
Bhatia, Vijay
1993 Analysing Genre: Language Use in Professional Settings. London: Longman.
Biber, Douglas
1988 Variation Across Speech and Writing. Cambridge, UK: Cambridge University
Press.
Bitzer, Lloyd
1968 The rhetorical situation. Philosophy and Rhetoric 1(13): 1–14.
Bitzer, Lloyd
1980 Functional communication: A situational perspective. In: Eugene Edmund
White (ed.), Rhetoric in Transition: Studies in the Nature and Uses of Rhetoric,
21–38. University Park, PA: Pennsylvania State University Press.
Burgess, Miranda
2009 Nation, book, medium: New technologies and their genres. In: Janet Giltrow
and Dieter Stein (eds.), Genres in the Internet: Issues in the Theory of Genre,
193–220. Amsterdam/Philadelphia: John Benjamins.
Campbell, Karlyn Kohrs and Kathleen Hall Jamieson
1978 Form and genre in rhetorical criticism: An introduction. In: Karlyn Kohrs
Campbell and Kathleen Hall Jamieson (eds.), Form and Genre: Shaping Rhe-
torical Action, 9–32. Falls Church, VA: Speech Communication Association.
Clark, Herbert H. and Susan E. Brennan
1991 Grounding in communication. In: Lauren B. Resnick, John M. Levine, and
Stephanie D. Teasley (eds.), Perspectives on Socially Shared Cognition,
127–149. Washington, DC: American Psychological Association.
Cohen, Ralph
1986 History and genre. New Literary History 17(2): 203–219.
Crowston, Kevin and Marie Williams
2000 Reproduced and emergent genres of communication on the World-Wide Web.
The Information Society 16(3): 201–216.
Devitt, Amy J.
2009 Re-fusing form in genre study. In: Janet Giltrow and Dieter Stein (eds.),
Genres in the Internet: Issues in the Theory of Genre, 27–48. Amsterdam/Phil-
adelphia: John Benjamins.
Emigh, William and Susan C. Herring
2005 Collaborative authoring on the Web: A genre analysis of online encyclopedias.
In: Proceedings of the Thirty-Eighth Hawaii International Conference on Sys-
tem Sciences. Los Alamitos, CA: IEEE Computer Society Press.
http://ella.slis.indiana.edu/~herring/wiki.pdf
Erickson, Thomas
2000 Making sense of computer-mediated communication: Conversations as genres,
CMC systems as genre ecologies. In: Proceedings of the Thirty-Third Hawaii In-
ternational Conference on System Sciences. Los Alamitos, CA: IEEE Computer
Society Press. http://www.pliant.org/personal/Tom_Erickson/genreEcologies.
html
Ferizis, George and Peter Bailey
2006 Towards practical genre classification of web documents. In: WWW 2006: Pro-
ceedings of the 15th International Conference on World Wide Web. New York:
Association for Computing Machinery. http://www2006.org/programme/files/
xhtml/p18/xhtml/pp018-ferizis.html
Freadman, Anne
2002 Uptake. In: Richard M. Coe, Lorelei Lingard, and Tatiana Teslenko (eds.), The
Rhetoric and Ideology of Genre: Strategies for Stability and Change, 39–53.
Cresskill, NJ: Hampton Press.
Giltrow, Janet
2002 Meta-genre. In: Richard M. Coe, Lorelei Lingard, and Tatiana Teslenko (eds.),
The Rhetoric and Ideology of Genre: Strategies for Stability and Change,
187–205. Cresskill, NJ: Hampton Press.
Giltrow, Janet
2010 Genre as difference: The sociality of linguistic variation. In: Heidrun Dorgeloh
and Anja Wanner (eds.), Syntactic Variation and Genre, 29–51. Berlin: Mou-
ton de Gruyter.
Graham, Sage Lambert
2007 Disagreeing to agree: Conflict, (im)politeness and identity in a computer-me-
diated community. Journal of Pragmatics 39(4): 742–759.
Hatipoğlu, Çiler
2007 (Im)politeness, national and professional identities and context: Some evi-
dence from e-mailed ‘Call for Papers’. Journal of Pragmatics 39(4): 760–773.
Hendry, David G. and Allyson Carlyle
2006 Hotlist or bibliography? A case of genre on the web. In: Proceedings of the
Thirty-Ninth Annual Hawaii International Conference on System Sciences.
Los Alamitos, CA: IEEE Computer Society Press. http://faculty.washington.
edu/dhendry/docs/hicss_final.pdf
Herring, Susan C.
1996 Posting in a different voice: Gender and ethics in computer-mediated com-
munication. In: Charles Ess (ed.), Philosophical Perspectives on Computer-
Mediated Communication, 115–145. Albany, NY: SUNY Press.
Herring, Susan C.
1999 Interactional coherence in CMC. Journal of Computer-Mediated Communi-
cation 4(4). http://jcmc.indiana.edu/vol4/issue4/herring.html
Herring, Susan C.
2007 A faceted classification scheme for computer-mediated discourse. Language@
Internet 4, article 1. http://www.languageatinternet.org/articles/2007/761/
Herring, Susan C., Lois Ann Scheidt, Sabrina Bonus, and Elijah Wright
2005 Weblogs as a bridging genre. Information Technology & People 18(2):
142–171.
Heyd, Theresa
2009 A model for describing ‘new’ and ‘old’ properties of CMC genres: The case of
Digital Folklore. In: Janet Giltrow and Dieter Stein (eds.), Genres in the Inter-
net: Issues in the Theory of Genre, 239–262. Amsterdam/Philadelphia: John
Benjamins.
Karlgren, Jussi and Douglas Cutting
2005 Recognizing text genres with simple metrics using discriminant analysis. In:
COLING ’94: Proceedings of the 15th International Conference on Computa-
tional Linguistics, Vol. 2. http://dl.acm.org.ezproxy.lib.indiana.edu/citation.
cfm?id=991324
Kwasnik, Barbara H., Kevin Crowston, Michael Nilan, and Dmitri Roussinov
2001 Identifying document genre to improve Web search effectiveness. Bulletin of
the American Society for Information Science 27(2): 23–27.
Locher, Miriam A.
2006 Advice Online: Advice-Giving in an American Internet Health Column. Am-
sterdam/Philadelphia: John Benjamins.
Miller, Carolyn R.
1984 Genre as social action. Quarterly Journal of Speech 70: 151–176.
Miller, Carolyn R. and Dawn Shepherd
2009 Questions for genre theory from the blogosphere. In: Janet Giltrow and Dieter
Stein (eds.), Genres in the Internet: Issues in the Theory of Genre, 263–290.
Amsterdam/Philadelphia: John Benjamins.
Moretti, Franco
2005 Graphs, Maps, Trees: Abstract Models for a Literary History. New York:
Verso.
Myers, Greg
2010 The Discourse of Blogs and Wikis. London: Continuum.
Orlikowski, Wanda J. and JoAnne Yates
1994 Genre repertoire: The structuring of communicative practices in organizations.
Administrative Science Quarterly 39(4): 541–574.
Puschmann, Cornelius
2009 Lies at Wal-Mart: Style and the subversion of genre in the Life at Wal-Mart
blog. In: Janet Giltrow and Dieter Stein (eds.), Genres in the Internet: Issues in
the Theory of Genre, 49–84. Amsterdam/Philadelphia: John Benjamins.
Puschmann, Cornelius
2010 “Thank you for thinking we could”: Use and function of interpersonal pro-
nouns in corporate web logs. In: Heidrun Dorgeloh and Anja Wanner (eds.),
Syntactic Variation and Genre, 167–191. Berlin: Mouton de Gruyter.
Sigel, Lisa Z.
2000 Filth in the wrong people’s hands: Postcards and the expansion of pornography
in Britain and the Atlantic world, 1880–1914. Journal of Social History 33(4):
859–885.
Sperber, Dan and Deirdre Wilson
1986 Relevance: Communication and Cognition. Oxford, UK: Blackwell.
Sproull, Lee and Sara Kiesler
1986 Reducing social context cues: Electronic mail in organizational communi-
cation. Management Science 32(11): 1492–1512.
Stein, Benno
2010 WEGA: Web genre analysis. http://www.uni-weimar.de/cms/index.php?id=
8793
Stein, Dieter
2006 The website as a domain-specific genre. Language@Internet 3, article 4.
http://www.languageatinternet.org/articles/2006/374
Strobbe, Ilse and Geert Jacobs
2005 E-releases: A view from linguistic pragmatics. Public Relations Review 31(2):
289–291.
Swales, John
1990 Genre Analysis: English in Academic and Research Settings. Cambridge, UK:
Cambridge University Press.
Toms, Elaine G.
2001 Recognizing digital genre. American Society for Information Science 27(2):
20–23.
Virtanen, Tuija
2010 Variation across texts and discourses: Theoretical and methodological perspec-
tives on text type and genre. In: Heidrun Dorgeloh and Anja Wanner (eds.),
Syntactic Variation and Genre, 53–84. Berlin: Mouton de Gruyter.
Watters, Carolyn R. and Michael A. Shepherd
1997 The digital broadsheet: An evolving genre. In: Proceedings of the Thirtieth
Hawaii International Conference on System Sciences. Los Alamitos, CA:
IEEE Computer Society Press.
Wyss, Eva L.
2008 From the bridal letter to online flirting: Changes in text type from the nine-
teenth century to the internet era. Journal of Historical Pragmatics 9(2):
225–254.
Yates, JoAnne and Wanda J. Orlikowski
1992 Genres of organizational communication: A structurational approach to study-
ing communication and media. Academy of Management Review 17: 299–326.
Brenda Danet (deceased) did her Ph.D. in Sociology at the University of Chicago,
USA, but spent most of her professional life in Israel. She was a professor at
Hebrew University, where she was appointed Danny Arnold Chair of Communi-
cations in 1998. Following her retirement in 2000, she migrated back to the United
States, where she became a Research Affiliate in Anthropology at Yale. Her most
recent scholarship focused on language, culture, and communication on the Inter-
net. Her publications on CMC include CyberPlay: Communicating Online (Berg
2001) and an edited volume (with Susan Herring) on The Multilingual Internet:
Language, Communication, and Culture Online (Oxford 2007).
Martin Gill is Associate Professor in the department of English Language and Lit-
erature at Åbo Akademi University, Finland, where he teaches sociolinguistics,
applied linguistics, and cultural studies. His current research is concerned with
questions of authenticity/authentication, language, and nativeness. His recent pub-
lications include “Authenticity” in Jef Verschueren and Jan-Ola Östman (eds.),
Pragmatics in Practice (John Benjamins 2011), and “Nativeness, authority, auth-
enticity: The construction of belonging and exclusion in debates about English lan-
guage proficiency and immigration in Britain” in Carol Percy and Mary Ann Da-
vidson (eds.), The Languages of Nation: Attitudes and Norms (Multilingual
Matters 2012).
Janet Giltrow is Professor in the Department of English and, since 2007, Associ-
ate Dean of Arts at the University of British Columbia, Canada. Taking rhetorical
and linguistic approaches to discourse studies, she has published extensively on
literary and non-literary stylistics; genre theory; ideologies of language; and aca-
demic writing, including two textbooks: Academic Writing: Writing and Reading
in the Disciplines, 3rd ed. (Broadview Press 2001) and Academic Writing: An In-
troduction, 2nd ed. (Broadview Press 2010). She co-edited (with Dieter Stein)
Genres in the Internet (John Benjamins 2009).
Subject index
bridal letter 724 f. coherence 3, 8, 12, 14, 16 f., 43, 50, 57,
Bulletin Board System (BBS) 59–61, 76, 59, 63–68, 75, 127, 169, 245, 247–249,
299 f., 467, 472, 474, 477 251, 255, 259, 261–264, 297, 372,
business communication 35 f., 44–49, 413, 430, 439, 441, 456, 489, 515–517,
61–63, 85, 118, 137, 150, 165, 198, 220, 539 ff., 591, 599, 721
417, 589 f., 722 cohesion 20, 248 f., 263, 516–518, 524,
534, 545, 547 f., 550, 555, 557
C collaboration 12–15, 20, 138, 301 f., 424,
chaining 21, 148, 628–631 431 f., 520, 525–527, 529 f., 532 f., 605,
change over time 98, 295 f., 306 645, 702, 708
chat 3 f., 8, 12 f., 16–17, 20 f., 50, 68, college class setting 301, 566 f., 569
109 ff., 117, 127, 136, 217, 220–222, communication mode see also mode 18,
232 f., 235, 245 ff., 250, 269, 274, 47 f., 50, 55–58, 76, 118, 199, 404, 571,
276, 286, 298 f., 305, 438 f., 465 f., 622, 698
472, 475, 489 ff., 515, 517–519, communicative form 57 f.
521–523, 527–530, 539ff., 559, 565ff., communicative function 58, 83, 89, 174,
591, 594, 641 f., 670 f., 674, 676–680, 270, 441, 509, 519, 727 f.
685 communicative purpose 7, 35, 38, 40 f.,
– history see history: chat 48, 57, 61, 68 f., 73, 75, 84, 89, 98,
– modes 12 f., 16 f., 25, 119, 121, 125 f., 111, 124, 247, 411, 444, 469, 491,
134 ff. 500, 534, 680 f., 688, 703, 719, 723 f.,
– rooms 3, 17, 20, 111–114, 117, 154, 728
217, 220–222, 224, 226 f., 232–236, community, online 70–76, 84, 97, 114,
263 f., 367, 370–372, 377 f., 438 f., 125 f., 151, 272, 282–284, 286, 304, 387,
477, 489 ff., 517, 523, 527, 539, 542 f., 390, 442, 434, 438 f., 450, 455 f., 509,
545, 548, 556, 558 f., 565 ff., 641, 521 f., 527, 695, 702, 710
679 f. community of practice 21, 294, 301, 306 f.,
– technical features 21, 109–118, 617–626, 630 f., 640 f.
125–127, 490, 541, 543 computer-mediated discourse (CMD) 9,
Chinese 4, 121 f., 138, 165, 168, 199, 301, 38, 48, 55, 59 f., 62, 64, 67 f., 70–75, 439,
340, 377, 467, 470 f., 474, 674 670–689
chronemics 489, 491 f., 530, 590, 697 f. computer-mediated discourse analysis
citations see quoting (CMDA) 48, 71
classification of CMC 3 f., 9 f., 14, 47 f., conditional relevance 21, 65, 517 f., 596,
68–70, 75, 85–91, 101 f., 168, 190, 598, 600, 603–607
235, 349, 425, 466 f., 544 f., 623, conflict 22, 71 f., 93 f., 96 f., 167, 439, 574,
648–650, 652 ff., 662 f., 673, 695, 719 f., 607, 614, 629, 639 f., 642, 644–653,
731–733 681
code-mixing 22, 641, 650, 654, 667, 670, connection / connected presence 17, 139,
675 f., 679–682, 685 f. 151, 167, 194–196, 207, 499, 508
code-switching 3, 8, 14, 22, 110, 121 f., consonant spelling 42, 173, 475
165, 170, 180, 641, 650, 654, 667 ff. contextualization cues 22, 279, 440, 489,
– conversational 668 f., 680 f., 684 f., 498, 669 f., 685
687 convergence 7, 9, 11 f., 127, 136, 164, 179,
– non-conversational 668, 671, 677 727
cognition 9, 99, 245 ff., 295, 366, 372, conversation analysis (CA) 4, 6, 17, 64,
717 217 f., 291, 489, 515, 518, 540, 544, 565,
– constraints on 171, 245 567–569, 589 f., 596, 675, 678
conversational floor 8, 20, 109, 202, 218, discursive features 59, 70, 72, 75, 178 f.,
221 f., 228–232, 254–256, 399, 515 ff., 726 f.
552, 555, 559 – of blogs 10, 84–86, 89, 101, 702–704,
conversational maxims see maxims 726 f., 730
conversational norms 18, 170, 292, 296, – of chat 112 f., 119–121, 222 ff.,
299–301, 589, 596, 605, 722 250–255, 299 f., 487 ff., 517 ff., 539 ff.,
conversational style 489, 498 f., 508 591
Cooperative Principle 19, 45, 246, – of discussion lists 59 f., 64–70, 75,
273, 372, 388, 398, 413, 437, 440 f., 297, 322 f., 520
455 – of email 35 ff., 69 f., 297 f., 401–405,
cost 17, 44, 152, 191, 196 ff., 208, 220, 589 ff., 675, 721, 723
420, 422, 647, 726 – of instant messaging 141–150, 152,
creativity 11, 60, 67 f., 75, 138, 144, 154, 171, 178 f., 474
169, 171, 174, 199, 207, 260, 263, 279, – of texting 152, 165 f., 168–174,
284, 445, 464, 467, 472–476, 639, 644, 176–179, 203 f., 467, 474
670, 676 f., 686, 695, 701 discussion, online 71, 76, 111, 275, 280,
Creole 674–675, 680, 684–687 301, 438, 443, 463, 466 f., 470 f.
Croatian 122–123, 125 discussion forum, online 14, 72, 75, 249,
cross-cultural 16, 18, 118, 139, 152 f., 262, 280, 282, 301, 304, 341–343, 347,
166–170, 201, 208, 272, 302, 319, 413, 356, 437–439, 441–443, 456, 463, 466 f.,
426, 489, 725 470 f., 473, 561, 667, 670–675, 679,
cues see reduced cues 681–683, 703
culture 9, 13, 21 f., 50, 60, 62, 69, 71, 113, dispute see conflict
121, 124, 175, 180, 191, 195, 201, 207, disrupted adjacency 8, 64, 120, 245,
251, 264, 273, 286, 298, 340 f., 344, 350, 248–255, 262–264, 517 f., 530–533,
355, 377, 431, 615, 639, 641–643, 654, 541 f., 559, 567, 591
674 f., 677–680 distance 15, 46, 48, 56, 76, 97, 207, 590,
CUSeeMe 117 613
Dutch 299
D
Danish 165, 170, 291, 297, 444, 592 E
deception 17–19, 137, 142, 154, 363 ff., economy 42, 122, 306, 472, 476, 493,
387 ff., 393 f., 396, 398, 400 f., 403 ff., 495
411, 413, 419, 719, 727 ellipsis (syntactic) 42, 95
deixis 5, 16, 23 f., 87 f., 91 f., 95, 99, 277, ellipses (repeated dots) 41, 477
424, 431, 704, 730 email 9–11, 14–16, 18 f., 21, 24, 35 ff.,
deletion, syntactic see ellipsis 55 ff., 98, 124, 135 f., 139 f., 142, 148,
demographics, participant 38, 48, 88, 109, 150–154, 163, 175, 178–180, 202, 217,
166, 201, 204, 276, 350, 289, 466, 471, 248, 269, 285, 297–299, 302, 315 ff.,
643 341 f., 349, 353, 363–370, 372, 375 f.,
digital folklore 387, 394, 398, 728 378, 387 ff., 411 ff., 463, 466, 470 f., 474,
digital genres see genre 490–492, 523, 539, 544, 589 ff., 613 ff.,
digital literacy 390, 405, 528, 533 641, 643, 645, 648 f., 652, 667, 671,
discourse analysis 4, 38, 43, 64, 70, 165, 673–677, 679 f., 684, 695, 703–706,
172, 180, 391, 489, 515, 533, 567, 695, 720–723, 725, 727 f., 732
717 email hoaxes 14, 19, 387 ff., 423, 703,
discourse coherence see coherence 727 f.
discourse transformer 282, 284 f. emblems 671 f., 678, 689
emotes 18, 112 f., 116, 124, 269, 274–280, flaming 14, 21, 72, 181, 303, 368, 416, 614,
283, 286 f. 639 ff.
emoticons 3 f., 8, 20, 23, 41, 43–45, 60 f., floor see conversational floor
74, 92, 111, 124, 144, 146, 148–150, 152, flow 20, 138, 197, 286, 508, 539 f., 565,
155, 166, 169, 173, 177, 205 f., 209, 220, 582, 617, 621, 625 f.
258, 279, 281 f., 284, 286, 299, 329, folklore see digital folklore
351 f., 418, 464, 466 f., 469–471, 474, framing/frames of interaction 21, 415, 489,
478, 540, 614, 670 582, 625, 627–631, 639, 684
emphasis 43, 111, 122, 282, 424, 449 f., free association 207, 249
464, 472–474, 477, 647, 669, 681, 683, French 4, 63, 122, 165–167, 169, 171, 209,
724 251, 254, 286, 291, 293, 302, 466–468,
English, use of 4, 24, 37, 39, 61–63, 472, 474 f., 674, 684
72, 114, 121–124, 165, 167, 169–171, future predictions
173, 181, 209, 222, 280, 286, 291 f., – for blogging 102
297–299, 301 f., 306, 320, 340, 351 f., – for chat 127, 236, 263 f.
355, 377, 394, 417, 429, 431, 465–468, – for discussion lists 73
471–477, 522, 528, 531–533, 565, – for email 36, 48 f.
569, 639, 641, 643–645, 651, 654, 670, – for instant messaging 154 f.
672, 674–676, 678–682, 684, 687, – for text messaging 207
689
e-style see also style 59 G
etiolation 272 f., 280, 284 f. games, online 12 f., 76, 109, 112, 114,
expressive function 122, 124, 141–142, 117 f., 123, 125–128, 250, 274, 418
177, 198, 404, 446, 456, 463, 473, 477, gay 20, 438 f., 491
614, 640, 642 gender 3, 16, 21 f., 60, 72, 74, 90 f., 94, 97,
102, 124 f., 136, 138, 141, 144, 148–150,
F 153 f., 165 f., 170, 172, 175, 198, 202,
face 96 f., 175, 284, 291, 304, 316, 324, 204 f., 208 f., 276, 278, 286, 294, 300 f.,
331 f., 334, 340 f., 347 f., 352, 355 f., 397, 317, 322, 335, 350, 371 f., 375–377, 379,
401, 403 f., 497, 614, 616–619, 627–629, 416, 437–439, 443, 446, 449, 455, 466 f.,
654, 669, 681, 725, 728 f. 471, 475, 477, 519 f., 615 f., 623, 639,
Facebook 13 f., 16, 48–50, 84, 102, 136, 641, 644, 653 f., 698, 703, 724
150–152, 154–156, 179, 274, 695 genre 3 f., 6, 9–11, 15–17, 19, 22 f., 57, 60,
Facebook status updates see status updates 68–70, 73, 75, 83–85, 89, 91, 94, 98, 102,
face-boosting act (FBA) 291, 304, 617 f., 125, 128, 172, 175 f., 178, 269, 286,
621, 625, 627–631 387–392, 394, 396, 401 f., 404–406,
faceted classification 9, 48, 70, 340, 411–414, 421, 425, 464, 479, 509,
349–351, 720 592–594, 607, 616, 618, 620, 622, 627,
face-threatening act (FTA) 97, 284, 291, 640, 670 f., 676, 679, 681, 684, 695,
316, 347 f., 352, 355 f., 403, 497, 616, 697–699, 701, 709, 717 ff.
618, 627–629, 630 f., 681, 725 genre change 10 f., 718, 722 f.
farewell formulas see leave-taking formulas German 24, 37, 39 f., 42, 61, 63, 121, 165 f.,
feature-based deception theory 366–370 169–171, 269, 286, 291, 293, 300, 302 f.,
Finnish 122, 165 f., 170, 466 f., 592, 667, 306, 466–469, 474–476, 541, 565, 567,
670, 674, 678 f. 674 f., 678–683, 686 f., 689
first-pair part 541, 606 Google Wave 50
flirting 124, 260, 351, 674, 680, 722, 724 f., Greek 170, 181, 208, 467, 471, 477, 674 f.,
731 679–682, 684, 686, 703
505 f., 508 f., 515–517, 519–521, 523 f., language play 3, 8, 18, 120, 172 f., 176 f.,
530 f., 539–544, 548, 556, 559 f., 179, 245, 247, 249 f., 256, 280–286, 445,
565–569, 571, 573, 576, 578, 581–583, 453, 456 f., 473 f., 476 f., 639 f., 650, 677,
589, 590–594, 596, 598, 599, 602–607, 722
613, 615–622, 625, 627, 629–631, 639, language shift 62, 678
641, 646, 652, 668–670, 676 f., 683, 688, language use 3–8, 13, 38, 40 f., 43, 48, 59,
696, 698, 700–704, 706, 708 f., 721, 727, 74 f., 91, 102, 113, 171, 173, 177, 209,
730 273, 280, 293, 352, 366, 372, 376, 441,
interactional sociolinguistics 489 463–472, 474–476, 478, 539, 668, 684
interaction management 16, 111, 114, languaging 675, 680, 686
119–121, 127, 439, 500–502, 540 f. leave-taking formulas 44, 298, 353, 678,
interactive written discourse 172, 670, 672 681
interactivity 12, 14, 22, 59, 68, 92, 109, lexical repetition 249, 258, 263, 547
112, 117, 127, 164, 179, 315, 342 f., 355, lexical shortening see abbreviation
467, 473, 543, 589, 590 f., 620, 622, 625, lingua franca 24, 62 f., 643
700, 702 f., 707, 709 f., 732 linearity 66 f., 248, 251, 292, 296, 304, 517,
international contexts 16, 110, 121 f. 697 f., 702 f.
Internet Relay Chat (IRC) 12, 16 f., 24, list culture 69–63
109 f., 115–127, 173, 245, 250–252, Listserv 16, 22, 55 f., 59, 62, 64, 68 f., 71,
254–258, 264, 269, 276–279, 286, 298 f., 74, 318, 639 ff.
303, 305, 371, 438, 442, 453, 457, 463 f., Listserv list see also mailing list 16, 56,
466–468, 472, 477, 491, 515, 530, 68 f., 74, 318, 639 ff.
540–542, 567, 594, 671–674, 676, 680 literacy see digital literacy
intertextuality 284, 703, 708 f. Lithuanian 122 f., 125
isiXhosa 161–170, 181, 467, 475, 674, 680 LIWC (Linguistic Inquiry and Word Count)
Italian 4, 63, 115, 122–124, 152 f., 165 f., see text analysis software
170, 201 f., 207–209, 466 f., 473, loosened relevance 8, 17, 245, 250, 252,
475–477, 565 255 f., 258 f., 261–263, 539, 567
love letter 725, 733
J lurking 73, 76, 98, 136, 142 f., 354, 441,
Japanese 4, 13, 152 f., 165, 167, 170 f., 521, 677
199–201, 207 f., 291, 299, 301, 305, 340, lying see also deception 271, 273, 276,
343, 466 f., 470, 472, 477, 565, 567, 639 363, 365–371, 374, 377 f., 388 f.
K M
Korean 152 f., 165, 167, 207, 301 f., 340 mailing list 16, 55 ff., 297, 673 f.
Malay 165, 291, 298
L markedness / marked usage 65 f., 280, 295,
language alternation see code-switching 301, 304, 331, 669
language change see also change over time maxim of manner 39, 45 f., 173, 246, 372,
11, 37, 40 440, 442, 446 f., 452 f., 455–457, 476
language choice / code choice 3, 15 f., 59, maxim of quality 45, 142, 246, 372,
62 f., 72, 74, 641, 671–673, 677 f., 681, 398–401, 440, 442–445, 447, 455, 649,
683, 688 653
language ideology see ideology maxim of quantity 45, 440, 445 f., 455, 647
language of distance see also distance 15, maxim of relation 4, 8, 10, 17 f., 45,
46, 48, 56, 76, 97, 207, 590, 613 245–250, 252, 255 f., 258–265, 398,
language of immediacy 15, 46–48 449–452, 455, 539, 647
media co-activity see also multitasking 13 389 f., 404, 463–469, 471 f., 474,
media richness 7, 364 f., 369 477–479, 490, 516, 523, 534, 539, 571,
medical advice online 18, 339, 342 f., 345, 622, 624, 631, 645, 667, 672–674,
348, 355, 729, 731 676–678, 680 f., 685–688, 698–701,
medium effects see techological determin- 706 f., 709
ism mode-focused approach 465 f.
medium factors 9, 18, 48, 61, 73, 75, 218, moderator / administrator 56–58, 71, 111,
262 f., 349, 351, 354, 463, 490, 516, 520, 114, 116 f., 126 f., 601, 603, 621, 653,
522, 668, 673, 676, 688 722, 725
meetings 12, 21, 220 f., 540, 543–545, motivation 85, 87 f., 91, 93, 122, 138, 350,
548–550, 553, 555, 559 f., 568, 606, 620, 364, 372, 378, 396 f., 399, 466, 473, 476,
726 478, 645, 673, 688, 718, 721, 729
memoranda 44, 722–723 MUDs and MOOs 9, 16 f., 24, 68, 112–114,
message length 141, 165, 169, 172, 178 f., 116–118, 123, 125, 128, 250 f., 254, 256,
204 f., 323, 326, 420, 495, 576, 580 260 f., 264, 269, 275–278, 367, 443, 515,
meta-communication 60 f., 324, 346, 399, 534
403, 489 multilingual Internet 62 f., 72, 74, 118, 121,
metadata 86, 88, 92 f., 102 639, 671–673, 676, 680, 684 f., 688
meta-genre 724, 726 multilingualism 21, 62, 667, 670 f., 689
metalinguistic awareness 3, 8, 16, 121, multimodality 14 f., 22 f., 75, 180, 191,
171, 686 416, 447, 534, 673, 688, 702–703, 709,
metaphor 64, 116, 250, 283, 388, 444 f., 711
453, 456, 552, 645, 669 multiparticipant / multiparty 8, 12, 17, 20,
metapragmatic awareness 9, 15, 163, 173, 50, 63–65, 74, 109–112, 115, 119 f., 154,
177, 180, 284, 286, 686 218, 221, 228–230, 232, 235 f., 245,
metatalk / metadiscourse 277, 324, 351, 250 f., 262, 263 f., 332, 438, 490 f., 515 f.,
531, 644 f., 728 518 f., 524, 567, 590, 670 f., 684 f.
metathesis 122, 147 multiple authorship 13, 22, 702, 709
microblogging 4, 13 f., 23, 84, 102, multitasking see also media co-activity 20,
179–180, 248 117, 120–121, 138–140, 155, 302, 305,
micro-coordination 194, 196, 198, 207 495, 508, 522, 527, 579, 531
minimal bilingualism 678 f. mutual monitoring 491 f., 729
mini-proposition 19, 437, 441, 447, 455 f. MySpace 49 f.
miscommunication 264, 318, 416, 489,
494, 524, 566, 578, 581 f., 614, 651, N
654 n:n communication 16, 48, 55–58, 74, 76,
MMS (multimedia message service) 164 350, 673
mobile phone 5, 15, 17, 49, 135 f., 146, narrative 6, 9–11, 22, 94, 100, 112, 149,
150, 152–155, 164 f., 173 f., 177–179, 277 f., 343, 354, 390, 397, 417, 423, 425,
191 ff., 217, 219 f., 222, 224, 227, 342, 432, 600, 627, 640, 675 f., 679, 684,
364, 367, 395, 463, 468, 613, 708 695 ff.
mobility 17, 49, 175 national language 63, 165, 170 f., 672, 675,
mock performative 18, 280–286 678–680
mode 3 f., 6 f., 9–12, 14–20, 22 f., 33, 35, negotiation 10 f., 19, 23, 94, 96, 116 f., 126,
38, 44 f., 47–50, 55, 57 f., 60, 69, 70, 76, 165, 191, 284, 286, 294–297, 299, 304 f.,
85, 98, 101 f., 109 f., 112, 114, 118 f., 307, 316, 340, 348, 350, 401, 412, 440 f.,
125–127, 149, 165, 178, 195, 199, 248, 456, 489, 490, 509, 540, 565, 567, 582,
250 f., 262, 264, 273–280, 282–286, 641, 669, 708
netiquette 39, 45, 61, 71, 300, 326, 333, openings and closings 165, 205, 393, 404,
438 683
Netspeak 8, 20, 23, 36, 39, 41, 464 f., 476, orality-literacy model 37, 40, 46, 58
479 orthography see also non-standard spelling
newbie see inexperienced user 3, 11, 14, 16, 20, 37, 41 f., 92, 100,
Newman-Pennebaker model 374 f., 378 122 f., 144, 146 f., 165, 168 f., 171–174,
new media 5, 11, 22, 40, 55, 57, 171 f., 177, 206, 275, 299, 318, 327, 417,
178 f., 699, 701, 704, 708 f., 722 428, 453, 464, 467–469, 472–476, 595,
newsgroup 3, 48, 55–57, 61 f., 70, 72 f., 614, 686 f.
248, 348, 364, 367, 370, 372, 439, 471,
474, 592, 639, 671, 676, 695, 704 P
news office 613, 621 f., 625 f., 630, 632 paralinguistic restitution 8, 111, 169, 173,
nicknames, online 4, 19–20, 111, 116, 254, 176, 181, 470, 472
280, 298–299, 318, 416, 437 ff., 492, pauses / pausing 145, 220, 223, 231 f., 234,
650, 652, 654, 672 237, 489–490, 494, 496–500, 506,
Nigerian letters/email/scam 19, 387, 390, 544–547, 549, 551–556, 558, 569, 579
393 f., 411–415, 417–422, 426, 428, performative predications 274, 278–280,
431–433 283, 285–287
Niuean 674 performatives 8, 17 f., 113, 249, 269–287,
noise see system noise 339, 414
non-responses 21, 541, 589 f., 593 f., 598, performativity 10, 18, 269 ff., 532
600–602, 604–606 perlocutionary force 270–271, 275, 285,
non-standard punctuation 14 f., 20, 41–43, 389, 647
179, 199, 464, 467, 469, 472–474, 476 f. Persian 667, 674, 681
non-standard spelling 14, 20, 41, 122 f., persistence of textual CMC 8, 58, 251, 307,
144, 168 f., 171, 173 f., 177, 199, 453, 349, 351, 354, 559, 719, 722
464, 467–469, 472–474 personae, online 18 f., 284, 350, 438 f.,
non-verbal behavior / non-verbal cues 8, 441
19, 46, 61, 63, 111, 218, 220 f., 222–223, personal addressability 192 f.
228, 232, 279, 332, 372 f., 400, 413, 418, personality 22, 491, 494, 639, 654
433, 491, 493, 519, 523 f., 540, 593, 596, person deixis 89, 91 f., 95, 99, 291, 424,
614 431, 730
norms see also conversational norms 8, 11, phatic communion see also small talk 16,
16, 21, 40, 59, 62, 70–72, 74 f., 88, 99, 99, 117, 168, 174, 196, 204, 250, 263,
118, 125 f., 149, 167, 175, 217, 232, 236, 351, 489, 613 ff.
248, 252, 261, 263, 294, 296, 301, 303, phonological approximation 168 f., 173,
306 f., 334, 345, 350–352, 355 f., 412 f., 176 f., 181, 472, 475 f.
416 f., 428, 442, 454 f., 472, 508, 589, planning 22, 46, 72, 92, 99, 101, 118, 613,
590–592, 594, 596, 598, 605–607, 618, 622, 627, 647, 677, 684–686
640, 653, 698, 708, 719, 724, 726, 729 play see language play
Norwegian 17, 21, 165 f., 169, 193–195, Polish 679
201, 203–207, 209, 297, 467, 565, 589, politeness 3, 16, 18, 21, 23, 87, 96 f., 137,
611, 615 140, 154, 292–296, 299–301, 303–307,
315–317, 333, 340, 389, 424, 440, 595,
O 599, 613–618, 621, 629–631, 640 f., 648,
offences 18, 315, 318, 322–324, 326–329, 654, 669, 722, 725
331, 334 f. polychronicity 3, 20, 120, 229, 232, 263,
offensive language see profanity 502–506, 522, 524
script (writing system) 20, 62 f., 74, 118, Spanish 4, 97, 114, 122, 291, 293, 343,
122, 202, 208, 351, 464, 467, 469, 391, 466, 565, 674, 679
472–474, 643, 687 speaker roles 20–21, 71, 94–96, 114,
script-switching 687 116–117, 281, 283, 293, 326, 344,
second-pair part 249, 606 350, 377, 392, 413, 416, 432, 506, 516,
seduction, online see also flirting 260, 499, 520 f., 527 f., 533–534, 542 f., 548 f.,
509 552, 554 f., 559 f., 590, 592, 594, 597,
self-description 449–450, 455 599–600, 604, 620, 652, 673, 680, 700,
self-presentation 18 f., 61, 137, 151, 154, 704
346, 363, 372, 415–417, 419, 432, 439, speech acts 4 f., 17, 19, 21, 23, 64, 137 f.,
441, 456 141 f., 270–275, 293, 299, 339 f., 355,
sequential relatedness 246, 249, 262, 596, 364, 388 f., 391, 393 f., 401–404, 470,
672 474, 589, 591, 596 f., 601, 607 f., 616,
sex 175, 253, 257 f., 260, 265, 341, 344 f., 640, 644, 646, 684
347, 350 f., 353 f., 405 f., 438, 491, 493, speech versus writing 3, 8 f., 37, 99, 102,
498 f., 508, 510, 682 145 f., 148 f., 178, 539
shortening see abbreviation spelling see orthography
signature file 61, 142, 447, 456 spoken interaction 217–237, 541, 544, 615,
sincerity 283, 315–317, 387, 389–390, 668, 670
412 f., 423, 455 standardness 11, 92, 100, 170–171,
Skype 35, 117, 136, 152, 154, 217, 221, 708 176–177, 209, 351–352, 355, 418, 472,
Skypecast 221 f., 224–227, 230–232, 718, 729
234–236 standards 56, 74, 76, 164, 172, 262, 291,
small stories 10, 14, 22, 697 f., 703, 709 387, 687
small talk 7, 21, 174, 177, 613 ff., 700 status, social 101, 165, 292–293, 298,
smiley faces see emoticons 300–302, 306–307, 316–317, 348, 350,
SMS see text messaging 377, 413, 417, 423, 431, 527, 615
social factors 14, 16, 22, 38, 48, 50, 74 f., status updates 13 f., 49, 84, 136, 179, 274,
294, 349–350, 466, 639, 654, 668, 676 695, 704–706
social-interactional approaches 696, 698, stories / storytelling 695 ff.
700, 702 – digital 25, 679, 699–701, 704, 708 f.
sociality 22, 168, 174–179, 717, 720 – fictional 22, 699, 702, 709
social networks 85, 122, 125–127, 154, – life 696, 704
199, 318, 387, 392, 619, 675, 710, 728 – personal experience 22, 695–700,
social network sites 4, 13 f., 16, 23, 40, 702, 704, 709
48–50, 83 f., 102, 136, 150–152, 154, strategic self-presentation 372, 415
156, 179 f., 264, 274, 390, 671, 688, 695, structural features of CMC 3, 8, 15, 20,
704–707 41–45, 58–61, 74, 84, 111–113, 116,
sociocultural factors 22, 46, 50, 273, 639 f., 122–124, 146–148, 152, 165 f., 168–170,
654, 695, 698 173, 176, 205 f., 299, 463 ff.
sociolinguistics 3–6, 8, 15, 24, 61, 149, style 16, 36 f., 39, 40 f., 46 f., 59, 68, 74,
165, 286, 291–294, 299, 304, 466, 471, 83–89, 90–95, 98–102, 143 f., 154,
489, 667–668, 675, 695–697, 700–703, 166–168, 171, 173, 175, 177, 179 f., 191,
708–710 204, 284, 291, 297, 330, 354, 375–377,
solidarity 21, 71, 174, 177, 179, 292–294, 379, 412 f., 416 f., 423, 428, 432 f., 470,
303, 340, 357, 491, 617, 639 f., 720 489, 491, 498 f., 502, 508, 520, 595, 615,
spam 49, 370, 387, 389 f., 392 f., 411, 649, 641–644, 647, 670, 675–679, 684 f., 695,
726 f. 724
Swedish 21, 122, 152 f., 165, 168–170, threading 14, 264, 544, 722
199–201, 207 f., 297, 302, 466–468, 592, timing see also chronemics 20, 249, 317,
613–638 489 ff., 581, 721
synchronous communication 8, 36, 48, 50, topic 3, 6, 16, 20 f., 36, 43, 45, 56–58, 60,
109, 112, 119, 135 f., 154, 217 ff., 245, 62, 64, 67, 70–75, 84, 87–90, 94, 98–102,
250, 252, 259, 263, 274, 341, 433, 437, 111, 115, 121 f., 149, 165, 168, 170, 174,
515, 539, 673 f., 685 201, 219, 225, 228–232, 234–236, 248,
syntax 18, 42, 48, 166, 169 f., 199, 262, 250 f., 256 f., 283, 286, 304, 318, 326,
274, 279, 346, 351, 375, 463, 516, 669, 327, 330, 339, 342, 344 f., 350, 352,
686 354–357, 374, 376 f., 388, 401, 406,
system noise 252, 255, 258, 263 437, 442, 449–450, 455–457, 489, 495,
500, 515–522, 526, 527, 530–534,
T 540–560, 567, 578, 627, 644, 645, 681,
technological determinism see technolo- 722, 730
gical factors T pronoun 291–293, 299–300, 302–304,
technological factors 7–9, 11, 14, 22, 306
59–61, 63 f., 68, 71 f., 74 f., 113, 126 f., transaction, communication as 164, 174,
262 f., 340, 343, 349, 356, 366, 463, 469, 422, 423 ff., 489, 599
475, 520, 534, 539, 553, 639, 653, 676 f., transcription 21, 112, 177, 237, 264, 349,
719 490, 519, 544 f., 560–561, 569 f.
teenagers 17, 49, 72, 88, 94, 102, 138, 144, transmedial narratology 699
152, 166, 169, 170, 191, 194, 196–199, trolling 371, 390, 392, 399, 405
201, 205–207, 344, 357, 592, 675, 679 f., trust 293, 330, 340, 348, 356, 363, 365,
685 367, 389 f., 411 f., 419, 424, 432, 450
teen emancipation 196, 198 f. Turkish 298, 674, 679 f., 687, 725
telegrams 43 f. turn organization 20, 541–543, 560, 567,
telephony 17, 117, 119, 127, 164, 195, 199, 669
207, 217 turn-taking 3, 14, 17, 109, 114, 119 f., 127,
temporal aspects see chronemics 137, 139, 154, 165, 170, 176, 202, 218,
text analysis software 364, 373 220, 228 f., 231 f., 235–237, 264, 517 f.,
text chat see chat 533 f., 542 f., 560, 566–568, 574, 578,
texting see text messaging 580, 590 f., 631, 670, 722
text messaging 4 f., 10 f., 16 f., 24, 39 f., 45, tweeting / retweeting 14, 180, 695, 706 f.
135 f., 143, 148, 150, 152–155, 163 ff., Twitter see also microblogging 14, 35, 40,
191–193, 196 f., 199–209, 220, 225, 264, 48 f., 84, 102, 179–180, 695, 706 f.
279, 306, 342, 363 f., 366, 372, 463, typing effort see economy
465–468, 471, 473–477, 592, 618, 667, typography 18, 117, 172, 280, 463, 467,
671, 674, 680, 708 472–474, 568, 639, 647
text type 37, 39 f., 47, 339, 344, 348–353,
622, 717, 724 U
Thai 124, 304, 543, 565, 641 universalizing approach 465–466
third move 591, 607 UNIX write / talk 38, 57, 111, 135
thread 3, 15, 21, 43, 57, 119, 121, 170, 196, urban legends 390, 397, 403, 703
218, 249, 251 f., 254, 280, 284, 300, 304, Usenet 55–57, 60, 62, 71, 590, 592, 639,
353, 356, 441, 457, 502, 516, 518, 524, 673 f., 676
544 f., 548–551, 556, 567, 589, 602, user adaptations 14, 17, 67, 120, 270, 495
607 f., 627, 643, 670–672, 682 f., 686, user-generated content (UGC) 12, 87, 671
706, 722 utterance breaks 136, 144–149
V W
variation 3, 12, 17, 23, 111, 119, Web 2.0 4 f., 12 f., 14, 24, 217, 387, 390,
122, 126 f., 144, 149, 166, 169, 172, 523, 672, 695, 699, 701
175–177, 275, 277, 286, 291, web chat see also chat 117, 122, 125, 567
294, 298, 302, 320, 322, 334, 356, web forums 49, 57, 76, 670 f.
420, 432, 465–467, 471, 474, Welsh 181
476–478, 515, 607, 670, 676, 686, Wikipedia 12 f., 24, 729 f.
725 wikis 4, 13–15, 91, 523, 732
VAX phone 112 workplace 15, 21, 72, 138 f., 218–220, 258,
video exchanges 15, 117, 126, 219 f. 582 f., 589 f., 592 f., 597, 600, 606, 613,
virtual community see community, online 615, 624, 631, 721
virtual teams 118, 543 ff. World Wide Web, history of 12
voice calls 5, 152 f., 193, 197–201, writing system see script
204
V pronoun 291–293, 299, 300, 302–303, Y
306 YouTube 102, 701