Professional Documents
Culture Documents
Discourse Particles in Corpus Text
Discourse Particles in Corpus Text
Discourse Particles in Corpus Text
PHOENIX W. Y. LAM
The Open University of Hong Kong, Hong Kong
INTRODUCTION
Discourse particles such as okay, so and well are syntactically optional linguistic
items which have no or little propositional value but serve important prag-
matic functions. Apart from the term ‘discourse particle’ (see, for example,
Aijmer 2002; Fischer 2006), there is no shortage of contenders which are
used in similar and often partly overlapping ways in research,1 including ‘dis-
course marker’ (see, for example, Schiffrin 1987; Schourup 2001) and ‘prag-
matic marker’ (see, for example, Watts 1988; Brinton 1996). Because of their
importance as ‘sharing devices and intimacy signals in our everyday talk’
(Quirk 1955: 179), they are indispensable in spoken discourse. In the pedagog-
ical setting, however, discourse particles are often dismissed as a sign of dys-
fluency and their use discouraged (Erman 1987). Although in recent years an
increasing number of studies have pointed out that the use of discourse par-
ticles contributes to the pragmatic and communicative competence of speakers
(Wierzbicka 1991; Müller 2005) and hence is an essential aspect of language
that learners should master, very few have actually looked at the coverage of
discourse particles in teaching materials. Previous studies of discourse particles
related to language learning, if any, rest on reporting the misuse by learners,
judged by the comparison with native data from outside the learners’ commu-
nity. The pedagogical significance of discourse particles in the foreign language
classroom is therefore severely understudied (Fung and Carter 2007). As an
P. W. Y. LAM 261
attempt to bridge this gap, the present study looks at how the discourse particle
well is used by expert speakers in an intercultural spoken corpus in Hong Kong
and how it is described and presented in textbooks designed for learners within
the same community. In particular, the frequency of occurrence, positional
preference, and discourse functions of the particle in the two data sources
will be compared both qualitatively and quantitatively to investigate the
extent to which textbook data reflects real-world usage.
A typical example of discourse particles, well is one of the most frequently
occurring items and probably the most thoroughly researched (Schourup
challenged (Davies 2003, 2004), particularly following the rise of the global
role of English (Graddol 1997, 2006). The prevalence of intercultural encoun-
ters in English has also generated the heated debate on the ownership of the
language. In the field of English language teaching, there have been rising
reservations about whether EFL learners should follow native English
models as ‘standards’ given the possible historical, cultural and contextual
differences concerned (see, for example, Kachru 1996; Seidlhofer 2002), not
to mention the implications of language imperialism and cultural hegemony
(Phillipson 1992; Pennycook 1998). In order to compare the actual perfor-
a direct comparison of the use of well between invented texts in textbooks and
naturally-occurring data in the HKCSE. While the present study is concerned
with the use of well as a discourse particle (D-use), the propositional use of
the word (P-use) will also be briefly discussed to demonstrate how typically the
word achieves discourse functions.
FINDINGS
Frequency of occurrence
One of the most striking differences between naturally-occurring data in the
HKCSE and sample texts in the textbook database in relation to the discourse
use of well is the rate at which the particle appears in discourse. Table 1 com-
pares the use of well in the two datasets under investigation.
A number of observations can be made from Table 1. First, there is a con-
siderable difference in terms of the total number of well in the two sample
text types in textbooks. While the total number of words in textbook discus-
sions (n = 13,032) is only 20% more than that in textbook presentations
(n = 10,817), the number of well in discussion texts is twelve times more
than that in presentations (n = 117 versus n = 9). This indicates that the word
occurs much more frequently in textbook discussions. In addition, there is
a huge discrepancy in the discourse-function ratio of well (D-use/total use as
Table 1: The comparison of well in the textbook sample section and in the
HKCSE
Textbook Textbook HKCSE
presentations discussions (N = 311)
(N = 38) (N = 11)
expressed in per cent) between the two sample text types. While only 11.11%
of instances of well are used as a discourse particle in textbook presentations,
the percentage rises to 91.45% in textbook discussion texts. This shows that
when the word is used it is most likely to be a particle in textbook discussions
and vice versa in presentations. In consequence, the textbooks give a strong
impression that well as a discourse particle is highly common in discussions
but almost nonexistent in presentations.
Such a noticeable contrast in the number of D-use well in textbook presen-
tations (n = 1) and discussions (n = 107) is reflected in the discourse rate of well
Position
In the discussion of the position of discourse particles, various units of talk
have been considered, including turn, utterance, tone unit, sentence and
clause-element (see, for example, Stenström 1990; Jucker and Ziv 1998;
Schourup 1999). The term utterance is used in this study as it probably
better accommodates features such as fillers, incomplete structures and over-
lapping speech which are common in spontaneous spoken discourse than
notions such as sentences and clauses which are more geared towards
the written language than the spoken language. Following Stenström (1994:
226), an utterance is defined as ‘anything that a speaker says’. Table 4 com-
pares the distribution of utterance position of D-use well in the HKCSE and
the textbook sample section.
As a discourse particle, well mostly occurs in utterance initial and medial
positions in the HKCSE. Altogether tokens in these two positions constitute
96% of the total number of discourse use of well. Specifically, more than half
of the discourse tokens of well (57.9%) are utterance initial, with a slightly lower
proportion of occurrences (38.1%) embedded in discourse. The use of well as a
The textbooks examined, on the other hand, show a different positional dis-
tribution of the particle. Of the 108 instances of D-use well found in textbook
presentations and discussions, 97 instances occur at the beginning of an utter-
ance, making up 89.8% of the total number of pragmatic use. Example 3
shows an utterance initial well in a textbook discussion:
(3)
Well, look, let’s try to summarize what’s been said so far . . .
(Potter 2003a: 129)
In contrast, only eleven tokens (10.2%) are found medially, making them a
small minority in the textbook database. An utterance medial D-use well is
shown in Example 4:
(4)
O.K. What’s next? An airport . . . well, they’ll be coming by plane, so they’ll see the
new airport when they arrive . . .
(Sutton and Duncan 1999: 154)
In addition, no instances of D-use well are found in final and stand-alone
positions in textbooks, while a small number can be found in the corpus.
Example 5 shows the use of well as a single-word utterance from the HKCSE:
Discourse function
This section first discusses the discourse functions of well as identified from
the HKCSE. This is followed by a qualitative analysis of how the discourse
functions of well are described in the teaching section of the textbooks.
Finally, a quantitative comparison of the functional distribution of D-use
well in the HKCSE and the textbook sample section is made.
In conducting a functional analysis of well, the present study follows a
bottom-up corpus-linguistic approach. In other words, the functional taxon-
omy developed in the analysis is derived from the recurrent patterns observed
in the datasets but not from a pre-existing framework or theory. This involves
rounds of modification before a classification scheme is devised to fully capture
the range of functions observed in the data. In total, 1,913 uses of well as
a discourse particle are examined in the HKCSE, of which 1,889 instances
are functionally categorized. 24 tokens are functionally unclassified owing to
insufficient contextual information. Six major functions of well are identified,
which can be categorized into three different domains: textual, interpersonal
and interactional. Instances of well expressing functions in the textual domain
are concerned with the structuring and organization of discourse. They largely
correspond to the text-oriented organizational unit (OT) in Linear Unit
Grammar (Sinclair and Mauranen 2006). In Halliday’s (1985) terms, they
achieve the textual meta-function. Interpersonal functions are related to the
expression of attitudes, emotions, and personal evaluations whereas interac-
tional functions facilitate processes such as planning and turn management
270 DISCOURSE PARTICLES IN CORPUS DATA AND TEXTBOOKS
Textual function
In the textual functional domain, well serves as a frame (Sinclair and Coulthard
1975) and a link. As noted in the Cambridge Grammar of English, these textual
uses of well ‘organize and monitor an ongoing discourse’ (Carter and McCarthy
2006: 901). The framing use of well to insert a point of division or transition
for easy comprehension is one of the most frequently occurring functions in
the corpus data. In these examples, well acts as a boundary marker in discourse
to signal transitions in topic and discourse stage. At times, it plays a role similar
to punctuation marks in the written language in dividing words into clauses
and sentences. In Example 6, the speaker uses well together with a meta-
linguistic comment let’s talk a little bit about conflict to indicate a topic change:
(6)
. . . but yet at the same time not violate our group harmony (.) yea it can be done (.)
it can be done okay well let’s talk a little bit about conflict why is conflict manage-
ment so important . . .
(HKCSE, B123)
Apart from segmenting texts, well could be used as a link to introduce explana-
tions and additional information to the preceding discourse. In their work
Cohesion in English, Halliday and Hasan (1976: 269) briefly describe this use
of well as an introducer of ‘an explanatory comment’. Similarly, the use of well
to continue an opinion or an answer is found in Müller (2005). Example 7
shows a speaker providing additional information about her supervisor follow-
ing an utterance medial well:
(7)
. . . so I start er to er write a proposal and I talk to my er supervisor er well he’s
a lecturer . . .
(HKCSE, B082)
Interpersonal function
In the interpersonal domain, the responsive use of well is the most dominant
among all the functions identified. This is in accordance with the observation
P. W. Y. LAM 271
Interactional function
In the interactional domain, well is used for processing purposes and turn
management. This is in accordance with corpus findings from the Longman
Grammar of Spoken and Written English, which suggest that discourse markers
‘facilitate the ongoing interaction’ (Biber et al. 1999: 140). As a processing
272 DISCOURSE PARTICLES IN CORPUS DATA AND TEXTBOOKS
device, well allows speakers to signal to the other participants that some inter-
nal processing is going on and enables them to gain extra time for their turn.
Therefore, it is particularly useful for marking self repair and hesitation (cf.
Carter and McCarthy 2006). In Example 10, the speaker encounters a
word recovery problem when she talks about the element to which deriva-
tional morphemes are added. Notice her use of the filler um and the particle
well as place-holders while she undergoes some internal contemplation before
resting on her choice of the expression word stem among other options:
planning what to say (Potter 2003a: 25). With regard to turn management, well
is associated with the speech function ‘interrupting politely’ (Potter 2003b:
39). This corresponds to the turn-taking sub-type identified in the corpus.
Again, examples given in the textbook teaching section are decontextualized
and it remains unclear how the particle assists in politely taking the conver-
sational floor, as the preceding utterance is not provided. A case in point is
Example 15, which is given in a textbook to illustrate how to interrupt without
being rude:
discourse functions such as framing which are also highly common in natu-
rally-occurring spoken examples.
DISCUSSION
Despite the fact that well is one of the most frequently occurring words in
the spoken language and many of the 15 textbooks collected for the present
study claim to focus on oral skills, none of them devotes a separate section
or paragraph to the description of the particle, not to mention the discussion
in time (McEnery and Wilson 1996) and there are aspects of language which
may not be fully unveiled by corpus evidence (Cook 1998). Yet for a crucial
pragmatic resource like discourse particles which are almost impossible to miss
in any contemporary spoken corpus, it is hard to justify why textbook writers
fail to give a more detailed and accurate description and presentation of how
they are used in real-life situations. Considering the fact that many of the
textbooks examined are specially designed for training English oral skills for
intermediate to advanced learners in Hong Kong in particular, such a mis-
match seems to be at odds with the aim of the oral examination that these
textbook writers now have at their disposal easy access to a large quantity of
naturally-occurring texts to confirm or refute their intuitions. Major publish-
ing houses even have their own in-house corpora on which teaching materials
can be based. In the cases where sizeable corpora from a specific EFL commu-
nity are not available to materials designers, corpus-based dictionaries and
grammars can be consulted (see, for example, Biber et al. 1999; Carter
and McCarthy 2006), with the careful supplement and comparison of data
and studies relevant to the local setting. Findings from the present study
which are generated from a large number of naturally-occurring examples
CONCLUSION
Through the unique examination of the particle well in an intercultural corpus
of spoken English in Hong Kong and its descriptions and presentations in
local textbooks, the present study has provided an example of how language
is taught and how language is actually used in the same community. Findings
from the present study have shown that wide discrepancies are found between
teaching materials and naturally-occurring examples in one of the most fre-
quently used discourse particles in the English language. This raises the issue
to what extent these textbooks examined reflect natural usage and hence
allow learners to be aware of how discourse particles are used. This article
argues that discourse particles are a valuable linguistic resource which learners
have a right to gain access to. Textbooks therefore should at least describe
and present them in the ways they are used in naturally-occurring examples
so that students can have a basic understanding of these items. Whether stu-
dents want to exploit this resource for productive use or only for reception
purposes is at their discretion (cf. Fung and Carter 2007), though they should
be given the choice all the same, especially for language learners at a more
advanced stage. When the issue of critical language awareness has become part
P. W. Y. LAM 279
SUPPLEMENTARY DATA
Supplementary material is available at Applied Linguistics online.
NOTES
1 In the literature, there is a lack of is available in the online version of the
a universally accepted term which article.
most people working in the field 3 Notes on transcription:
give consent to. For detailed discus- ‘S1/S2:’ example in the HKCSE with
sions, see Brinton (1996) and Fischer more than one speaker;
(2006). ‘. . .’: an utterance reported only in
2 For the sake of brevity, details of the part;
textbooks used in this study are given ‘’: the beginning of overlapping talk;
in the supplementary appendix, which ‘(.)’: an unfilled pause.
REFERENCES
Aijmer, K. 2002. English Discourse Particles: Brinton, L. J. 1996. Pragmatic Markers in English:
Evidence from a Corpus. John Benjamins. Grammaticalization and Discourse Functions.
Aijmer, K. and A.-M. Simon-Vandenbergen. Mouton de Gruyter.
2003. ‘The discourse particle well and its Carlson, L. 1984. ‘Well’ in Dialogue Games:
equivalents in Swedish and Dutch,’ Linguistics A Discourse Analysis of the Interjection ‘Well’ in
41/6: 1123–61. Idealized Conversation. John Benjamins.
Bhatt, R. M. 2005. ‘Expert discourses, local Carter, R. 1998. ‘Orders of reality: CANCODE,
practices and hybridity: The case of Indian communication, and culture,’ ELT Journal
Englishes’ in A. S. Canagarajah (ed.): 52/1: 43–56.
Reclaiming the Local in Language Policy and Carter, R., R. Hughes, and M. McCarthy.
Practice. Lawrence Erlbaum. 2000. Exploring Grammar in Context. Upper-inter-
Biber, D., S. Johansson, G. Leech, S. Conrad, mediate and Advanced. Cambridge University
and E. Finegan. 1999. Longman Grammar of Press.
Spoken and Written English. Longman. Carter, R. and M. McCarthy. 2006. Cambridge
Breen, M. 1985. ‘Authenticity in the language Grammar of English. Cambridge University
classroom,’ Applied Linguistics 6: 60–70. Press.
280 DISCOURSE PARTICLES IN CORPUS DATA AND TEXTBOOKS
Cheng, W., C. Greaves, and M. Warren. 2005. Hong Kong Examinations and Assessment
‘The creation of prosodically transcribed inter- Authority.
cultural corpus: The Hong Kong Corpus of Johns, T. 1991. ‘Should you be persuaded – two
Spoken English (prosodic),’ ICAME Journal samples of data driven learning materials,’
29: 5–26. English Language Research Journal 4: 1–16.
Cook, G. 1998. ‘The uses of reality: A reply to Jucker, A. H. and Y. Ziv. 1998. ‘Discourse mar-
Ronald Carter,’ ELT Journal 52/1: 57–63. kets: Introduction’ in A. H. Jucker and Y. Ziv
Cook, G. 2001. ‘The philosopher pulled the (eds): Discourse Markers: Descriptions and Theory.
lower jaw of the hen,’ Applied Linguistics 22/3: John Benjamins.
366–87. Kachru, B. 1996. ‘English as lingua franca’
Davies, A. 2003. The Native Speaker: Myth and
Mindt, D. 1997. ‘Corpora and the teaching of Romero Trillo, J. 2002. ‘The pragmatic fossiliza-
English in Germany’ in A. Wichmann, tion of discourse markers in non-native speak-
S. Filgelstone, G. Knowles, and A. McEnery ers of English,’ Journal of Pragmatics 34:
(eds): Teaching and Language Corpora. Longman. 769–84.
Möllering, M. 2004. The Acquisition of German Schiffrin, D. 1987. Discourse Markers. Cambridge
Modal Particles. A Corpus-based Approach. Peter University Press.
Lang. Schourup, L. 1999. ‘Tutorial overview:
Müller, S. 2004. ‘Well you know that type of person: Discourse markers,’ Lingua 107: 227–65.
Functions of well in the speech of American Schourup, L. 2001. ‘Rethinking well,’ Journal of
and German students,’ Journal of Pragmatics Pragmatics 33: 1025–60.
36: 1157–82.