Professional Documents
Culture Documents
Successful Spoken English
Successful Spoken English
Christian Jones is Senior Lecturer Applied Linguistics and TESOL at the Uni-
versity of Liverpool, UK.
Editorial Panel: IVACS (Inter-Varietal Applied Corpus Studies Group), based at Mary
Immaculate College, University of Limerick, is an international research network
linking corpus linguistic researchers interested in exploring and comparing language
in different contexts of use.
Typeset in Sabon
by Apex CoVantage, LLC
Contents
2 Linguistic competence 33
2.1 Introduction 33
2.2 Definitions of linguistic competence 33
2.3 Previous studies 34
2.4 Methods of analysis 36
2.4.1 Frequency profiles 36
2.4.2 Frequency lists 37
vi Contents
3 Strategic competence 78
3.1 Introduction 78
3.2 Definitions of strategic competence and communication
strategies 78
3.3 Previous studies 80
3.4 Methods of analysis 83
3.5 Strategic competence at B1-C1 levels 85
3.5.1 Preliminary analysis of CEFR strategies in B1,
B2 and C1 speech 87
3.5.2 Comparison of CEFR strategy realisation in B1,
B2 and C1 speech 92
3.5.3 Production strategies: correction 94
3.5.4 Interaction strategies: inviting others into the
discussion and seeking clarification 99
3.6 Conclusion 105
6 Conclusion 159
6.1 Introduction 159
6.2 Summary of findings 159
6.2.1 Linguistic competence 159
6.2.2 Strategic competence 160
6.2.3 Discourse competence 162
6.2.4 Pragmatic competence 163
6.2.5 Summary 163
6.3 Implications for research 166
6.4 Implications for teaching 167
6.5 Final thoughts 169
Index 173
Figures
We would like to thank colleagues past and present for help, inspiration and
support:
1.1 Introduction
The aim of this chapter is to explore and attempt to define the concept
of a successful speaker of English and lay the groundwork for the chap-
ters which will follow. In order to work towards this definition, we seek to
explore the notion of a successful speaker within one main framework: the
notion of communicative competence, first developed by Hymes (1972). We
will argue that this is an appropriate manner in which to explore successful
speakers and their interlanguage (Selinker 1972). A successful speaker is,
we believe, one who can demonstrate all facets of communicative compe-
tence at their level and as appropriate for the purpose of the discourse they
are taking part in. We will suggest that the different aspects of a successful
speaker can therefore be analysed by exploring their linguistic, strategic,
discourse and pragmatic competence, as outlined by Hymes (1972), Canale
and Swain (1980), Canale (1983) and Bachman and Palmer (1996) and that
these competences will vary according to a learner’s level of English. This
model is not new but has, we believe, had a large impact upon the develop-
ment of communicative language teaching and testing and was one of the
theoretical bases of the Common European Framework of References for
Languages (Council of Europe 2001), so we feel it is still highly relevant. For
instance, although much learner corpus research focuses on errors or devia-
tions following comparisons of native speaker language use, communicative
competence models still promote the notion that assessments of language
should attend both to language knowledge and how a speaker uses this
knowledge in actual communication. After 30 years, this fundamental view
still resonates in the CEFR, since in viewing learners as social agents with
varying target language needs, development across levels is tracked accord-
ing to what students can do with their language when communicating across
contexts, according to different functions and with changing audiences. In
basing its own treatment of communicative competence on the interplay
between linguistic, pragmatic, sociolinguistic and existential competences,
the CEFR clearly adopts many of the original aspects of communicative
2 What is a successful speaker of English?
thus stems from the sense that the NS target set by learners, and indeed
sometimes by practitioners (see Canagarajah 2007; Kramsch 2003; Timmis
2002), forms the basis of a learner’s assessment of their own achievement:
success equates to attaining native-like proficiency whereas other proficien-
cies equate to failed or incomplete attempts (Birdsong 2004). In contrast
to this, we wish to suggest that learners can be successful at different levels
and not only at a high level of competence. In other words, communicative
competence (as we will move on to discuss) is an attainable goal at different
language levels and it is the ability to be communicatively competent which
we can view as a measure of success. This argument is in opposition to the
notion that the ‘native speaker’ is the only model of success. We suggest that
this can create a ‘deficit relationship’ in which NSs have ‘the upper hand’
(Prodromou 1997: 439) and rather than underlining a learner’s success in
using an L2, they can heighten their sense of failure towards L2 development
and the NS model itself (Cook 2002; 2008). As Naiman et al., (1978: 2)
remark:
The continual pressure to replicate the NS not only creates ‘stereotypes that
die hard’ (Nayar 1994: 4), but it potentially discourages learners from per-
severing with the acquisition of English. It also overlooks the potential that
could be unlocked by viewing learners not as ‘failed native speakers’ but
instead as ‘successful multicompetent speakers’ of more than one language
(Cook 1999: 204).
Of vital importance to this study, however, is the barrier created by the
NS model. By focussing on the overall goal of L2 development, little assis-
tance is offered to learners whose goals for success may simply be to develop
their interlanguage to become more proficient in English than they currently
are. For instance, the NS model cannot be fully relevant to their needs since
in representing the ‘finished article’, it cannot demonstrate the nuances in
interlanguage between the beginner, pre-intermediate, intermediate, upper-
intermediate and advanced learner levels. While this realisation has prompted
some writers to propose alternative models of greater relevance to L2 learners’
multicompetences and language learning strategies (see Alptekin 2002; Cook
2008; Coperías Aguilar 2008; Edge 1988; Medgyes 1992; Modiano 1999
and Preston 1981 for alternative models; see also Cook 1992; Cook 1999;
Coppieters 1987; Galambos and Golin-Meadow 1990 for multicompetence
and O’Malley and Chamot 1990; Oxford 1990; Oxford and Nyikos 1989
for language learning strategies), relatively few studies have sought to inves-
tigate what makes learners within them successful in their own right. By
focussing on learners, this book therefore intends to address the imbalance
4 What is a successful speaker of English?
We have then to account for the fact that a normal child acquires knowl-
edge of sentences not only as grammatical but also as appropriate. He
or she acquires competence as to when to speak, when not, and as to
what to talk about with whom, when, where, in what matter. In short,
a child becomes able to accomplish a repertoire of speech acts, to take
part in speech events, and to evaluate their accomplishment by others.
(Hymes 1972: 277)
The key argument in this theory, therefore, is that linguistic and sociocul-
tural competences are not separate entities with the latter ‘grafted’ onto the
former, nor is sociocultural competence irrelevant to a language user’s over-
all language competence. Similarly, competence no longer remains a static
state of knowledge due to its involvement in the process of creating and
comprehending meaning in a range of communicative contexts (Ellis 1994).
Competence instead embodies the ‘capabilities of a person’, dependent both
on declarative knowledge (knowledge about language) and the ‘ability for
use’ (Hymes 1972: 282). This definition incorporates four distinct qualities:
that language is formally possible, in terms of grammar, culture or com-
munication; feasible, regarding the implementation of language according
to psycholinguistic and cultural factors; appropriate, involving awareness
of contextual features and tacit knowledge of sentences and situations; and
done in the sense that the forms do actually occur in a language (see Hymes
1972: 281–286).
Hymes’s (1972) theory did much to extend the concept of competence in
communication and its applicability to second language learning and suc-
cess. For instance, as a theory fundamental to social interaction (Paulston
1992), it stipulates that grammatical accuracy alone is insufficient in the
learning and use of a language. It advocates that language users need ‘knowl-
edge of both the L2 grammar and of how this system is put to use in actual
communication’ (Ellis 2008: 6). For adult L2 learners, this means, for exam-
ple, that the learning of grammar rules alone would not result in success-
ful communicative competence, they have to also be capable of adapting
language to suit its mode, audience, genre and context. Ultimately, Hymes’s
theory asserts that learners possessing high grammatical accuracy and low
6 What is a successful speaker of English?
Just as Hymes (1972) was able to say that there are rules of grammar
that would be useless without rules of language use, so we feel that there
are rules of language use that would be useless without rules of grammar.
(Canale and Swain 1980: 5)
Lexis
Morphology
Grammatical
competence Syntax
Sentence-grammar semantics
Phonology
Communicative Sociolinguistic
Sociocultural rules of use
competence competence
Discourse
(see Shohamy 1996), despite their overt reluctance to do so (see Canale and
Swain 1980: 7) Canale (1983: 5) acknowledges that their theory referred
alternatively to a communicative competence incorporating ‘underlying
systems of knowledge and skill required for communication’. The knowl-
edge fundamental to communicative competence, or likewise the declarative
knowledge of ‘knowing about’ language is therefore accompanied by skill,
the procedural knowledge concerning the extent to which knowledge ‘can be
performed’ or put to use in ‘actual communication’: the new term created for
performance in order to avoid confusion with Chomsky’s 1965 definition
(Canale 1983; McNamara 1995). Transitioning from knowledge-oriented
approaches of language teaching, Canale asserts that adopting a more skill-
oriented approach is a much–needed shift if students are to learn how to
employ such knowledge adequately:
This model is helpful, as it clarifies some issues which arose from previous
models As the next section will also reveal, the model emphases renewed
attempts by theorists to describe more explicitly how knowledge is utilised
in communication.
As a theory useful for considering and planning test design, Bachman and
Palmer’s model broadens the scope of language usage to incorporate char-
acteristics of the speaker and the task and setting. As figure 1.2 shows, this
change was attributed to features other than language knowledge – personal
characteristics, topical knowledge and affect – playing an important role
in someone’s ability to communicate. For example, in a test situation, a
learner’s display of their language knowledge may be impaired by factors
affecting the individual on the day of the exam, and this could lead to a
potential reluctance to produce language. The figure demonstrates, there-
fore, the characteristics held by an individual (in the bold circle), bearing
some resemblance to the CEFR’s existential competence (Council of Europe
2001), and how this relates to the external task or setting in which they
interact. Although individual differences will not be elaborated upon here, it
is evident that factors such as mood, tiredness, personality, topic knowledge
and willingness to attempt or adapt language use, to name but a few, will
influence a speaker’s overall performance.
In terms of its composition, Bachman and Palmer’s (1996) theory of lan-
guage use and performance rests ultimately on Bachman’s (1990) theory of
language ability. They posit that the combination of two elements, language
knowledge and strategic competence, bestow language users with ‘the ability,
or capacity, to create and interpret discourse, either in responding to tasks
on language tests or in non – test language use’ (Bachman and Palmer 1996:
67). In contrast to previous studies of communicative competence, these two
factors are thus no longer seen as uniform in their impact. For instance, whilst
Canale and Swain (1980) stress that grammatical competence is neither more
nor less important than sociolinguistic or strategic competence, the latter in
this theory is given a renewed role. In Bachman and Palmer’s model, stra-
tegic competence is made central to interactions between task and setting
12 What is a successful speaker of English?
Language
knowledge
Topical Personal
knowledge characteristics
Affect
Strategic
competence
Characteristics of the
language use or test
task and setting
Language
competence
Organisational Pragmatic
competence competence
Sensit. to Cultural
Rhet Ideat. Manip. Heur. Imag. Sensit. to Sensit. to Refs. &
Voc. Morph. Synt. PhonGraph. Cohes. Dial. or
Org. Functs. Functs. Functs. Functs. Reg. Nat. Figs. of
Variety
Speech
Figure 1.2 Bachman and Palmer’s (1996: 63) language use and performance with Bachman’s
(1990: 87) components of language ability
language which can extend meaning beyond the message conveyed at utter-
ance level. When combined with sociolinguistic aspects such as sensitivity to
dialect or variety, register differences, naturalness of language and cultural
references, it becomes easier to see how Bachman and Palmer’s (1996) model
of language use and performance is more thorough than previous theories in
terms of what, exactly, comprises communicative language ability and what
makes it accurate and appropriate.
This theory is of significance to learner success in a number of ways.
Through the use of arrows, a much more tangible attempt has been made to
explore the types of knowledge required by L2 learners and the interactions
occurring between them. It demonstrates how the form, or organisation,
of language is of equal importance to the meanings inferred by the speaker
or interpreted by the interlocutor as a result of pragmatic competence and
the interactive setting. The notion that appropriacy as well as accuracy is
integral to communication is thus maintained. Of particular interest to this
book is the elevated importance given to strategic competence. In a study
of success, it would be easy to confine this competence to compensation or
communication breakdowns; the implication from Bachmann and Palmer is
that learners require strategic competence in a broader sense.
Bachman and Palmer’s (1996) model is by no means perfect or definitive.
Based principally on language testing, its application to general language use
may be more problematic. For instance, it might be relatively easy to docu-
ment or evaluate aspects of pragmatic or sociolinguistic competence in lan-
guage tests of a familiar design, but in wider, freer communicative settings,
they may be harder to distinguish. Secondly, there is no explanation for how
learners make use of formulaic language in these competences. Finally, though
Bachman and Palmer accentuate the interaction with the task or setting, mod-
els of spoken English in teaching materials may actually interfere with a stu-
dent’s impressions of successful interactive speech (Jones and Horak 2014).
To give an example, one sample goal from the CEFR at B1 level is as follows:
C1 – Effective
A1 – Breakthrough B1 – Threshold Operational
Proficiency
Within the CEFR, language use is not viewed as static, nor is it uniform. Lan-
guage users, such as those in Canale and Swain’s (1980) theory, benefit from
their unique set of competences and strategies which enable them not only
to achieve linguistic and non-linguistic tasks, but also to develop in those
competences. The CEFR, similarly to Canale and Swain (1980), Canale
(1983) and Bachman and Palmer (1996), views communicative competence
as part of a language user’s wider, more general competence. Comprising
knowledge, skills, existential competence and the ability to learn, an almost
cyclic relationship is created: communicative competence is thus a part of
general competence, but it is this competence which rather vaguely is said
to ‘contribute[s] in one way or another’ to learners’ communicative abili-
ties (Council of Europe 2001: 101). Declarative knowledge, concerning the
world, sociocultural and intercultural knowledge, is a product of experience
and formal learning. It relates to one’s language and culture as well as to lan-
guage users’ knowledge of day–to–day life. In a sense, declarative knowledge
is the CEFR’s equivalent to Bachman and Palmer’s (1996) treatment of topic
knowledge and knowledge of different settings. Skills, pertaining to proce-
dural knowledge or ‘know-how’, on the other hand concerns the ‘ability to
carry out procedures’ (Council of Europe 2001: 11). Using similar terminol-
ogy to Canale (1983), they relate to social, living, vocational and leisure skills
which differ in their degree of mastery, and the ease, speed and confidence
with which they can be performed. Existential competence, comparative to
Bachman and Palmer’s (1996) personal characteristics, is composed of per-
sonal attributes unique to language users. Containing attitudes, motivations,
values, beliefs, cognitive styles and personal factors, existential competence
results from acculturation and a person’s readiness to interact socially with
others. Finally, drawing on ‘various types of competence’, the ability to learn
amalgamates the three previous aspects (Council of Europe 2001: 12). Help-
ing language users to ‘deal with the unknown’(Council of Europe 2001: 12),
it involves more than the capability to learn; it suggests that users will dif-
fer in their predisposition, or willingness, to seize or seek opportunities for
exploiting learning potential, hopefully reducing the resistance or threat felt
in learning a new language.
The CEFR’s notion of general competence thus relies on more than
knowledge that something can be done: it highlights the need to know how
it can be done, the willingness to do it, and the ability to add to what is
already known. Regarding communicative competence, comparisons may
be made between this definition and Canale and Swain’s (1980) model as
it contains three similar components: linguistic competences, sociolinguis-
tic competences and pragmatic competences. Interesting to note here is the
What is a successful speaker of English? 19
plurality of the term ‘competence’ as, in the CEFR’s sense once again, com-
petence refers not only to knowledge but to a combination of knowledge,
skills and know–how in each of the three competences. Like the preced-
ing theories, linguistic competences relate closely to previous definitions.
Comprising knowledge of and the ability to use lexical, formulaic, gram-
matical, morphological, syntactical, semantic, phonological, orthographical
and orthoepic competence, lexical competences relates to the construction
and formulation of well–formed, meaningful messages (Council of Europe
2001). Interestingly, in stating that linguistic competence can be scaled, the
CEFR suggests that language is ‘never completely mastered by any of its
users’ (Council of Europe 2001: 109). Perhaps in opposition to previous
NS–learner comparisons, all language users are thus said to always have
something new to learn; their skills and accessibility to the range and qual-
ity of knowledge also differs across individuals. Sociolinguistic competences,
on the other hand, closely assimilate to previous models’ social and cultural
conventions, the particular linguistic customs which may need to be fol-
lowed in different settings. As no truly homogenous settings exist, this set
of competences is said to relate specifically to language use in response to
social contexts surrounding it. Learners thus require to develop their inter-
cultural competence to facilitate this. Despite having knowledge, skills and
know–how in such an area, language users may be uninformed of socio-
linguistic norms: just as Paulston (1992) remarks, language users can be
unaware of cultural rules until the point at which they are broken. Pragmatic
competence, in place of Canale’s (1983) discourse competence, concerns ‘the
functional use of linguistic resources . . . the master of discourse, cohesion
and coherence’ (Council of Europe 2001: 13). Combining the arrangement
of sentences into sequences (discourse competence) as well as the knowledge
of and ability to construct interaction, it conforms closely to Bygate’s (1987)
discussion of ‘routine’ and ‘interactive skills’ to produce spoken conversa-
tion. In short, the view of communicative competence in the CEFR remains
that it constitutes the internalised, yet changeable, state of knowledge, skills
and know–how which are then put to use. It is through performance or
‘observable behaviour’ (Council of Europe 2001: 14) that aspects can be
recognised and activated. However, no specification is offered as to how this
is actually achieved or whether performance is fully able to reflect what is
held in a language learner’s competence.
5 Structures mostly Consistently Use of stress and intonation Consistently makes Sustained interaction in both
accurate for the level demonstrates puts very little strain on extensive, coherent and initiating and responding
with only occasional appropriate and listener and individual relevant contributions which facilitates fluent
minor slips. extensive range of sounds are articulated to the achievement of communication.Very sensitive
lexis for this level. clearly. Utterances are the task. to turn-taking.
consistently understandable.
4.5 More features of band 4 than band 5.
4 Generally Evidence of an Stress and intonation Contributions are Meaningful communication
structurally accurate extensive and patterns may cause generally relevant, is largely achieved through
for the level but appropriate range occasional strain on listener. coherent and of an initiating and responding
some non-impeding of lexis with Individual sounds are appropriate length. effectively. Hesitation is minimal
errors present. occasional lapses. generally articulated clearly. and the norms of turn-taking
are generally applied.
3.5 More features of band 4 than band 3.
3 Reasonable level Lexis is mostly Use of stress and intonation Contributions are Sufficient and appropriate
of structural effective and is sufficiently adequate normally relevant, initiation and response
accuracy but some appropriate for most utterances to coherent and of an generally maintained
impeding errors are although range be comprehensible. Some appropriate length but throughout the discourse
acceptable. and accuracy are intrusive L1 sounds may there may be occasional although there may be some
restricted at times. cause difficulties for the irrelevancies and lack of undue hesitation. Turn-taking
listener. coherence. norms may not be observed.
(Continued)
Table 1.1 (Continued)
use an animated figure within virtual role plays to provide learners with a con-
text and spoken prompt to which they respond and record their answers. For
example, one situation could be ‘You want an extension on your assignment.
You go to your tutor’s office to ask him for extra time’. The tutor then says
‘Hello, you wanted to see me?’ before the learner responds (see figure 1.4).
We acknowledge that this method of data collection cannot replace the
ideal of capturing data in naturally-ocurring settings, but it goes some way to
addressing some of the well-documented drawbacks of traditional written and
oral production tasks such as authenticity of interaction and learner response.
These virtual role plays were selected over face-to-face role plays for being
able to efficiently administer all the tests simultaneously with comparable
groups of students, under controlled conditions. A further advantage of the
CAPT is that the computer-animated characters are able to display a range of
non-verbal signals such as facial expressions and gestures, considered to be as
powerful as verbal cues. Secondly, authentic voice recordings can be uploaded
for the characters which also aim to simulate semi-authentic interaction in the
academic environment within which learners are currently studying. Halenko
(2013: 288) reports the CAPT successfully elicited language which is closer
to what students would actually say in a given situation, rather than what
they might say, when compared to written production tasks. For instance,
Halenko found written discourse completion task responses tended to be
longer, and often included extraneous detail, unlike the more efficient CAPT
responses, as these examples, from the same participant, illustrate;
(lost book borrowed from tutor scenario):
Sorry, sir. I am made a big mistake that I lost a book that you lent me.
I try my best to find where it is but I can’t and I try and find a new one
but I don’t know where I can buy it. I’m really sorry about that. Can
you tell me where I can buy this book? I will buy a new one for you. I
very apologise about that.
(Written production task response)
You:
Figure 1.4 An example of a scenario from the CAPT to elicit spoken data
26 What is a successful speaker of English?
I apologise about that. Last week you lent me a book but now I can’t
find it at home. I’m so sorry. Can I buy you a new one.
(CAPT response)
Table 1.3 Rating scale to evaluate participant request and apology responses.
adapted from Shively and Cohen (2008). The rating scale did not require
attention to the grammatical accuracy of the responses, since the focus was
on their overall effectiveness.
The raters attended a standardisation meeting prior to the actual evalu-
ation stage to explain the project, the instrument, the rating criteria and
procedure. A number of practice items, followed by a comparison of rat-
ings, were completed to achieve a final consensus. A rating of ‘3’ was dis-
cussed as being ‘of minimal satisfaction’ and was included as the cut off
point for a response to be considered appropriate. Where queries were raised
by the raters during the evaluation stage, these were resolved in follow-up
meetings with the researcher. To understand which strategies and language
components were considered most effective for each scenario, all responses
awarded appropriate scores of 3 (appropriate) and 4 (very appropriate) by
the raters were isolated and used to comprise the pragmatics corpus detailed
in this chapter. The corpus also included NS data elicited via the same means.
The NS data were used as a point of comparison and to illustrate alternate
ways success could be achieved. This approach enabled us to measure prag-
matic competency more clearly. Table 1.4 details the learner data in the
pragmatics corpus.
1.8 Conclusion
In this chapter, we have explained the reason for a focus on successful spo-
ken language. In short, there are two reasons for this. Firstly, successful
spoken language, as produced by learners, is worthy of study on its own
terms, not simply as something in deficit to a native- speaker model. We
acknowledge that many learners aspire to reaching a native-speaker stan-
dard but we also argue that this is often unattainable and successful spoken
English can be produced by learners at different levels of competence. Sec-
ondly, a focus on spoken language has been chosen because it is the spoken
skill which many learners aspire to obtaining and yet spoken language is still
relatively under-researched. In this chapter we have also reviewed the theo-
ries of communicative competence which have informed our definition of a
successful speaker of English and how these link to the CEFR levels B1–C2.
To reiterate, the broad definition we are employing is: A successful user of
spoken English is one who is able to use linguistic, strategic, pragmatic and
discourse competences as appropriate for a particular goal, at a particular
language level, as defined by the CEFR. We have also acknowledged that it
is not possible to measure in all contexts and therefore we have chosen to
examine it by looking at speaking data captured from texts and computer-
animated production tests. Such data gives us evidence of the successful lan-
guage students can produce in these contexts but does not, of course, inform
us about language which students only understand. Despite this, we feel it is
a realistic way to examine successful spoken language.
In the chapters which follow, we will use this data to explore successful
spoken language. Clearly, as the definitions in this chapter show, we can also
see that the different aspects of competence are often linked together. Each
chapter which follows this separates the analysis of different competences
which make up successful spoken English and analyses them individually.
We therefore look at linguistic competence, followed by strategic, discourse
and pragmatic competence in separate chapters. This is in order to clarify
the different aspects for the reader but where appropriate, we explicitly
make links between the competences and we acknowledge, as mentioned
above, that the different competences work together to produce communica-
tive competence.
References
Alderson, J.C. 2007. The CEFR and the need for more research. The Modern Lan-
guage Journal, 91(4), 659–663.
Alptekin, C. 2002. Towards intercultural communicative competence in ELT. ELT
Journal, 56(1), 57–64.
Andreou, G. and Galantomos, I. 2009. The native speaker ideal in foreign language
teaching. Electronic Journal of Foreign Language Teaching, 6(2), 200–208.
What is a successful speaker of English? 29
Lyons, J. 1996. On competence and performance and related notions. In: G. Brown,
K. Malmkjӕr and J. Williams, eds. Performance & competence in second language
acquisition. Cambridge: Cambridge University Press, 9–32.
Martyniuk, W. and Noijons, J. 2007. Executive summary of results of a survey on:
The use of the CEFR at national level in the Council of Europe Member States.
Strasbourg: Council of Europe.
McNamara, T.F. 1995. Modelling performance: Opening Pandora’s box. Applied
Linguistics, 16(2), 159–179.
Medgyes, P. 1992. Native or non-native: Who’s worth more? ELT Journal, 46(4),
340–349.
Modiano, M. 1999. International English in the global village. English Today, 15(2),
22–28.
Naiman, N., Frolich, M., Stern, H.H. and Todesco, A. 1978. The good language
learner. Clevedon: Multilingual Matters.
Nayar, P.B. 2004. Whose English is it? TESL-EJ, 1(1), n.p. Available from: <www.
tesl-ej.org/wordpress/issues/volume1/ej01/ej01f1/> [Accessed 27th March 2017].
Nureddeen, F.A. 2008. Cross cultural pragmatics: Apology strategies in Sudanese
Arabic. Journal of Pragmatics, 40, 279–306.
O’Malley, J. and Chamot, A.U. 1990. Learning strategies in second language acquisi-
tion. Cambridge: Cambridge University Press.
Oxford, R. 1990. Language learning strategies: What every teacher should know.
Boston, MA: Wadsworth Publishing.
Oxford, R. and Nyikos, M. 1989. Variables affecting choice of language learning
strategies by university students. The Modern Language Journal, 73(3), 291–300.
Paulston, C.B. 1992. Linguistic and communicative competence. Avon: Multilingual
Matters Ltd.
Piller, I. 2002. Passing for a native speaker: Identity and success in second language
learning. Journal of Sociolinguistics, 6(2), 179–208.
Preston, D.R. 1981. The Ethnography of TESOL. TESOL Quarterly, 15(2), 105–116.
Prodromou, L. 1997. Correspondence. ELT Journal, 63(4), 439–440.
Prodromou, L. 2008. English as a lingua franca: A corpus-based analysis. London:
Continuum.
Rampton, M.B.H. 1990. Displacing the ‘native speaker’: Expertise, affiliation, and
inheritance. ELT Journal, 44(2), 97–101.
Schauer, G.A. 2007. Finding the right words in the study abroad context: The devel-
opment of German learners’ use of external modifiers in English. Intercultural
Pragmatics, 4(2), 193–220.
Selinker, L. 1972. Interlanguage. International Review of Applied Linguistics, 10,
209–231.
Shively, R. and Cohen, A. D. 2008. Development of Spanish requests and apologies
during study abroad. Ikala, revista de language y cultura, 13(20), 57–118.
Shohamy, E. 1996. Competence and performance in language testing. In: G. Brown,
K. Malkjmær and J. Williams, eds. 1996. Performance and competence in second
language acquisition. Cambridge: Cambridge University Press, 138–151.
Taguchi, N. 2006. Analysis of appropriateness in a speech act of request in L2
English. Pragmatics Quarterly Publication of the International Pragmatics Asso-
ciation (IPrA), 16(4), 513–533.
32 What is a successful speaker of English?
Taylor, D.S. 1988. The meaning and use of the term ‘competence’ in linguistics and
applied linguistics. Applied Linguistics, 9(2), 148–168.
Timmis, I. 2002. Native-speaker norms and International English: A classroom view.
ELT Journal, 56(3), 240–249.
Timmis, I., 2003. Corpora, classroom and context: The place of spoken grammar in
English language teaching. Unpublished PhD. University of Nottingham. Avail-
able from: <http://eprints.nottingham.ac.uk/12246/1/397578.pdf> [Accessed
10 December 2015].
Timmis, I. 2005. Towards a framework for teaching spoken grammar. ELT Journal,
59(2), 117–125.
Chapter 2
Linguistic competence
2.1 Introduction
In this chapter we examine the linguistic competence of successful spoken
English at different levels. As mentioned in chapter 1, we take our defini-
tion of this competence from a combination of Hymes (1972), Canale and
Swain (1980), Canale (1983) and Bachman and Palmer (1996) but in doing
so, we are defining linguistic competence as a competence mainly concerned
with lexis. This arises from a belief that the key to developing linguistic
competence is primarily lexical and that learning words, their collocations
and chunks is a process which grammar assists in but is not driven by. As
a result, we will explore the data in terms of word and keyword frequency
and also how keywords collocate and colligate within larger patterns of
language. By doing so, we will explore the patterns of lexis which learners
at different levels use and show how this language varies according to level.
We will also make comparisons with larger reference corpora and show how
these patterns differ or are similar to these corpora.
We have then to account for the fact that a normal child acquires knowl-
edge of sentences not only as grammatical but also as appropriate. He or
she acquires competence as to when to speak, when not, and as to what
to talk about with whom, when, where, in what matter. In short, a child
becomes able to accomplish a repertoire of speech acts, to take part in
speech events, and to evaluate their accomplishment by others . . .
(Hymes 1972: 277)
This definition suggests that when we examine the language which learners
at different levels can use, we must explore what they can use but also how
they use it, i.e. how it functions in discourse. Therefore, in this chapter, we
will explore linguistic competence in terms of the lexis used and how this
functions at utterance and text level to achieve meanings within the English
language tests which make up our corpus. As we will go on to describe in
section 2.4, this has been undertaken through an exploration of the most
frequent words, keywords and chunks used at each level but prior to this, in
section 2.3, we offer a brief review of previous research in this area.
written language (Biber et al., 1999; Erman and Warren 2000), collocations
and lexical chunks cannot be seen as an ‘added extra’. Instead, we can sug-
gest that chunks ‘are extremely frequent, are necessary in discourse and are
fundamental to successful interaction’ (Adolphs and Carter 2013: 36).
Previous studies have managed to shed light on the definitions and func-
tions of formulaic language. In the former, collocation signifies two or
more words in adjacency or proximity of each other (McEnery and Hardie
2012: 123). They co-occur frequently enough that their occurrences cannot
be attributed to chance alone (Greaves and Warren 2010: 212), but they
surpass a threshold of possibility and statistical significance that confirms
their status as a meaningful collocation (Hunston 2002; McEnery and Wil-
son 2001). Collocations are also considered fundamental to the creation of
meaning (Adolphs and Carter 2013; Lewis 1993) despite often being uncon-
nected to pragmatic function (Nattinger and DeCarrico 1992). On the other
hand, lexical chunks, a form of multi-word unit, tend to perform a more
functional role in language (Schmitt 2000) and differ from collocations in
length and role (see De Cock 1998; Nattinger and DeCarrico 1992). In terms
of length, whilst collocations can primarily be associated with word pair-
ings (see Adolphs and Carter 2013: 23) e.g. ‘fussy eater’, ‘tall man’, ‘commit
fraud’, it is clear that lexical chunks extend ‘far beyond’ this level (Schmitt
2000: 400). They can be short, long or ‘anything in between’ (Schmitt and
Carter 2004: 3) but are typically expected to be between two and four words
in length (McCarthy 2010). With regard to function, although they are rel-
evant for conveying meaning or ‘referential’ topic-related information, they
are considered to be of higher importance in realising pragmatic and dis-
course functions (De Cock 1998: 69). However, they cannot be assumed
to be syntactically complete e.g. ‘In the’, ‘top of,’ ‘the end of’ (Adolphs and
Carter 2013; Biber et al., 1999; De Cock 1998), nor are they restricted to
one sole purpose (Erman and Warren 2000). Chunks are examined in rela-
tion to discourse and pragmatic competence in chapters 4 and 5 but the
nature and function of formulaic language in spoken learner language, and
the way it develops or changes across levels, hence remain less clear and
the implications for learner success remain unknown. It is these gaps in the
literature which we will attempt to address in the analysis of learner data
in this chapter.
interviews consisting of three tasks (set topic, free discussion and picture),
it was thought to be particularly comparable to the data in the USTC as
outlined in chapter 1. In preparation for subsequent analyses, a wordlist
composed of learner turns only was constructed from LINDSEI using the
Wordlist function in WordSmith (Scott 2015). Comparisons were also made
to our own NS data and the spoken section of the BYU-BNC corpus (Davies
2004) where we thought this could be illustrative.
Initially, observation of the data was carried out to identify words which
differed greatly in their frequencies. Any words appearing with high fre-
quency across the levels and any words in the USTC data but not the NS
data were immediately earmarked for further investigation. Normalised fre-
quencies were also incorporated into the analysis as these reveal how often a
word can be expected within a set number of tokens (the base of normalisa-
tion), typically, every 10,000 words or 1,000,000 words. They are calculated
by dividing a word type’s total occurrence by the size of the corpus; this fig-
ure is then multiplied by the base of normalisation to allow comparisons to
be made across corpora of differing sizes (McEnery and Hardie 2012). Since
the base of normalisation adapts to corpus size, normalised frequencies in
this study utilised a base of 10,000 for LINDSEI and cross-level compari-
sons of the USTC data. Once a word was selected for further investigation,
qualitative examination of the word in contexts via the use of concordances
allowed us to explore patterns of use. This helped to paint a picture of a
word’s function in learners’ utterances at different levels. To investigate this
further, an additional measure (Juilland’s D) was used to ascertain how a
word was dispersed across the corpus. Whereas frequency identifies how
often a word occurs, dispersion can give more indication of its distribution
and whether it was typical of all texts in a corpus (see Oakes 1998). The
closer the figure is to 1, the higher the dispersion rate. To assist in calculat-
ing this measure, Brezina’s (2014) online toolbox for corpus linguistics was
utilised.
Information was obtained regarding what a particular word occurred
with, the structure it usually occurred in and the purpose it fulfilled. When
performing this qualitative analysis via the use of concordance lines, we
also observed salient collocations. Word co-occurrence is a concept closely
related to concordance lines ‘since the idea of two words occurring in a com-
mon context is similar to that of two words occurring in the same concor-
dance window’ (Oakes 1998: 159) Once a collocation seemed to be salient,
quantitative measures of collocational strength were adopted. These mea-
sures compared the number of appearances of a collocate within a four- or
five-word window (or span) to the left and right of the node (Baker 2006;
McEnery and Hardie 2012). For assessing collocation strength t-scores and
mutual information (MI) scores were used. T-scores are considered more
accurate in the analysis of words with relatively low frequencies due to the
manner of calculation (Baker 2006; Barnbrook 1996) and in this case we
Linguistic competence 39
assumed (following Barnbrook 1996: 98) that any t-score greater than 2
would signify a potentially ‘interesting’ focus and that an MI score of 3
would also be worth examination (Oakes 1998). For the purposes of pre-
senting collocates in this chapter, both t-scores and MI scores will be used
but when looking at the collocates of single words, both figures will have
to satisfy the significance thresholds specified and we have at times only
focused on the most frequent collocates, due to their number. We make this
clear in the text where this occurs. When looking at collocates in the form
of chunks (e.g. collocates of we can), only one figure will have to reach this
threshold due to the lower frequencies of the data.
two-word chunks led us feel that many of these seemed quite fragmentary
in nature and so a decision was made in this chapter to explore chunks of
three and four words in length. Whilst less common than three-word chunks
(Biber et al., 1999), the longer stretches of lexis were also thought interesting
for comparisons of B1, B2 and C1 as many would encapsulate some of the
three-word Ngrams (see Carter and McCarthy 2006). For example, ‘what
do you’ may appear as a three-word chunk and then as a part of the four-
word ‘what do you think? All chunks were included in the analysis, on the
basis of frequency. This was because we believed that their appearance in the
lists still suggests that they are typically retrieved and produced as if they are
single items, even if some might not be considered to be syntactically whole.
To produce the chunk lists, uncoded text files were uploaded into AntConc
(Anthony 2017). The previously discussed approaches to extracting collo-
cational information and analysing concordance lines were also followed,
which allowed us to examine both the language used and also how it func-
tioned in speech. This also meant we again tried to focus our attention on
chunks which occurred with frequency across the levels and were not obvi-
ously frequent in relation to the exam topics. For these reasons, we explore
both a lot of and agree with you.
of 83.26%, B2’s family coverage stood at 83.70% and C1’s family coverage
decreased slightly to 80.62%. It can be inferred from this data that for learn-
ers to be successful and present a solid performance in their speech at the
Threshold, Vantage and Effective Operational stages, the two most frequent
vocabulary bands in English are essential. Figures 2.1–2.3 taken from the
B1, B2 and C1 data demonstrate that successful language at these levels is
indeed comprised of vocabulary mostly from the K-1 and K-2 bands (words
in italics = K-1, words in underlined italics = K-2, words in bold denote a
different frequency band or an off-list word):
Exam Freq. Families Types Tokens Cumul. Total types Total tokens Tokens Standardised
Level Band (%) (%) (%) Token % across all across all per type TTR (Word-
bands bands Smith Tools)
}
B1 K-1 585 878 14762 92.81
(63.59) (65.28) (92.81)
K-2 181 219 553 96.29 1345 15905 11.83 25.00
(19.67) (16.28) (3.48)
}
B2 K-1 671 1011 18706 92.86
(62.89) (66.21) (92.86)
K-2 222 267 884 97.25 1527 20144 13.19 24.90
(20.81) (17.49) (4.39)
}
C1 K-1 675 1136 20303 91.64
(58.39) (62.80) (91.64)
K-2 257 340 1104 96.62 1809 22156 12.25 26.95
(22.23) (18.79) (4.98)
<$19M> Er actually the most popular sport in my country is football er I I like football and er
I’ve found national team. Er actually national team played yesterday last night and er er losed
the cup silver cup. I’m sad today but er the the you know the football it’s a game I think it’s help
the politics to keep the people of the country happy to keep the people in the country you know er
fans watch the TV <$=> it’s </$=> and also the people er happy when they when they watch er
the the foot= the football match.
<$29F> Well they look for higher qualifications and it depends actually on the job if you’re
gonna apply like if it’s for sport they will look for sport section <$=> if it’s for the </$=> it
depends which major you’re gonna you’re gonna apply for not major meaning you’re gonna
apply for the if you’re studying business like I’m studying business now I might t=er go to a
company and manage the company so it depends on the subject you studied and it depends on
the company you’re applying to.
<$29M> To learn to drive I think it must be sixteen or seventeen but to start driving it should be
eighteen because if people just learn to drive in one month or in one session they won’t learn
everything so <$=> they start </$=> two years ago they start having classes some safety some
safety classes about the classes about the cars <$=> can </$=> what they are capable of what
they can do what they can’t do they must tell them everything before they can use the car.
Table 2.2 Types, tokens, means and SDs for K-1 and K-2
B1 K-1 585 163.76 23.76 878 197.59 28.75 14762 868.12 191.63
K-2 181 17.76 6.09 219 19.24 6.87 553 32.53 14.35
B2 K-1 671 186.29 21.98 1011 226.00 31.69 18706 1100.24 286.69
K-2 222 23.82 6.11 267 26.00 6.13 884 52.00 14.93
C1 K-1 675 200.23 25.39 1136 252.12 41.00 20303 1194.18 316.98
K-2 257 29.29 8.38 340 33.94 10.40 1104 64.94 29.30
46.89
60.00
44.63
50.00
40.00
20.27
20.08
20.04
19.49
30.00
15.35
14.17
12.80
12.37
12.17
10.35
9.99
20.00
8.82
7.09
3.73
5.9
4.6
2.06
1.44
0.84
0.76
0.34
0.29
0.29
10.00
0.4
0.00
A1 A2 B1 B2 C1 C2 UNLISTED
B1 Types B2 Types C1 Types NS Types
69.22
67.96
80
70
60
50
40
15.59
13.97
12.73
12.16
30
11.41
11.31
10.75
10.03
20
5.22
5.18
4.29
3.15
2.18
1.63
1.43
1.06
0.54
0.39
0.21
0.16
0.13
0.05
0.05
0.06
10
0
A1 A2 B1 B2 C1 C2 UNLISTED
B1 B2 C1 NS B1 B2 C1 NS
speakers. Though there are small increases across the proficiency levels in B1
and B2 token usage, all speakers in the USTC undeniably used lexis which
was high in frequency and without a great deal of complexity. The Text
Inspector findings thus illustrate that an increase in language level does not
necessarily equate to an increase in usage of what could be considered ‘more
difficult’ vocabulary.
This section has aimed to examine the vocabulary profile of speakers at
B1, B2 and C1. The CEFR states that there should be development in vocab-
ulary range but it does not suggest what this development is in quantitative
terms. Instead it uses vague expressions such as ‘enough language to get
by, with sufficient vocabulary [B1]’ ‘has a sufficient range of language [B2]’
and ‘a broad range of language [C1]’(Council of Europe 2001: 28) to imply
that the amount of vocabulary known by learners should grow in some
way. Whilst vocabulary research can focus on the number of word families,
Linguistic competence 45
Word Freq. % Texts % Word Freq. % Texts % Word Freq. % Texts % Word Freq. % Texts %
1 ER 1,391 8.09 17 100.00 ER 1,312 6.16 17 100.00 THE 1,147 4.97 17 100.00 I 374 4.67 6 100.00
2 I 1,126 6.55 17 100.00 I 926 4.34 17 100.00 I 807 3.50 17 100.00 AND 249 3.11 6 100.00
3 THE 688 4.00 17 100.00 THE 827 3.88 17 100.00 ER 770 3.33 17 100.00 THE 249 3.11 6 100.00
4 AND 523 3.04 17 100.00 AND 591 2.77 17 100.00 AND 665 2.88 17 100.00 LIKE 214 2.67 6 100.00
5 TO 461 2.68 17 100.00 TO 585 2.74 17 100.00 TO 640 2.77 17 100.00 TO 197 2.46 6 100.00
6 ERM 320 1.86 17 100.00 YOU 518 2.43 17 100.00 ERM 460 1.99 17 100.00 YOU 184 2.30 6 100.00
7 IN 296 1.72 17 100.00 YEAH 380 1.78 17 100.00 YEAH 442 1.91 17 100.00 A 172 2.15 6 100.00
8 IS 288 1.67 17 100.00 IN 376 1.76 17 100.00 IS 417 1.81 17 100.00 YEAH 171 2.13 6 100.00
9 SO 288 1.67 16 94.12 A 325 1.52 17 100.00 IN 414 1.79 17 100.00 IT 166 2.07 6 100.00
10 A 269 1.56 17 100.00 IS 316 1.48 16 94.12 YOU 376 1.63 17 100.00 ERM 165 2.06 6 100.00
11 YOU 252 1.47 17 100.00 ERM 308 1.44 17 100.00 A 350 1.52 17 100.00 OF 134 1.67 6 100.00
12 YEAH 250 1.45 16 94.12 LIKE 304 1.43 17 100.00 LIKE 348 1.51 17 100.00 THAT 128 1.60 6 100.00
13 MY 238 1.38 17 100.00 THINK 276 1.29 17 100.00 THINK 342 1.48 17 100.00 THINK 127 1.59 6 100.00
14 THINK 237 1.38 17 100.00 MY 252 1.18 17 100.00 OF 312 1.35 17 100.00 IT’S 100 1.25 6 100.00
15 LIKE 222 1.29 17 100.00 SO 252 1.18 17 100.00 SO 295 1.28 17 100.00 IN 97 1.21 6 100.00
16 BECAUSE 188 1.09 17 100.00 BUT 233 1.09 17 100.00 THEY 285 1.23 16 94.12 SO 79 0.99 6 100.00
17 OF 161 0.94 17 100.00 HAVE 229 1.07 17 100.00 IT’S 276 1.20 17 100.00 JUST 75 0.94 6 100.00
18 IT’S 152 0.88 16 94.12 THEY 226 1.06 15 88.24 IT’S 254 1.10 17 100.00 IF 74 0.92 6 100.00
19 CAN 150 0.87 15 88.24 OF 222 1.04 17 100.00 WE 221 0.96 17 100.00 ON 71 0.89 6 100.00
20 WE 149 0.87 16 94.12 WE 212 0.99 16 94.12 BECAUSE 196 0.85 17 100.00 BUT 69 0.86 6 100.00
Linguistic competence 47
B1 B2 C1
Word List Freq. NF List Freq. NF List Freq. NF List Freq. NF List Freq. NF
pn pn pn pn pn
WE 20 149 86.77 20 212 99.54 19 221 95.74 26 5773 72.80 13 10448 10.45
2.6.1 WE
B1 B2 C1
Collocate Freq. with T-score MI score Collocate Freq. with T-score MI score Collocate Freq. with T-score MI score
node node node
<$0> The most environmentally friendly way to travel? Do you understand that?
<$16F> The most er invently +
<$0> Environmentally friendly.
<$16F> <$O17> Oh </$O17> Yeah I got the meaning.
<$15F> <$O17> Environment </$O17>
<$0> Mm.
<$16F> Mm I answer the question?
<$0> Yeah.
<$16F> In the travel I think the environment friendly it means we go to travel not by a car and it
not er mm we can protect er environment and that is er mm er green green travel yes erm green
travel when we go to travel I think if we use bicycle I think it’s not harmful the er fresh air and
we can’t throw the rubbish anywhere I think it’s er mm friendly environment
usage and repetition of this pronoun helped learners at all levels to formulate
their own turns and to link their own ideas together, within and sometimes
across turns. This is something we will discuss further in chapter 4 but we
can see example of this usage in figure 2.9, which shows an example by B1
level learners.
In terms of more specific functions, we was mainly used by learners to
manage interaction by including the other speaker in the answers to enable
a form of co-construction, confirming what the other speaker said and often
adding to it, as the example in figure 2.9 also shows. This co-construction
and ability to link across speaker turns increased from B1-C1 levels, as we
will discuss in chapter 4.
There was very little evidence of we being used in this manner in NS data,
and a far greater tendency to rely on I to fulfil the same function and to use
other means of linking ideas together, such as the use of synonyms or other
pronouns. Figure 2.10 shows an example of this.
<$1M> First er I think er working for a large international comp= company er company have er
some advantage and disadvantage. I think er +
<$2M> This one er er the advantage is er we can communicate wi= with different people and er
these people er from <$=> different b= <$=> from different place. And er the disadvantage is
the the language is a big problem.
<$1M> If the language is a problem maybe we can talk talk with each other.
<$2M> Yeah. I’m very agree with your opinion and er in addition I would like to say erm er
work in the international company is very erm good and erm er we should er get erm high
income in the company rather than erm interesting in some er work. I mean er in this city erm
the pre= pressure influence on lots of people so most of reason is money so we should get high
income er work to get lots of money for our life.
<$1M> Yeah I agree with you and I think income is er very big thing <$=> and erm erm </$=>
but er we need practise our language if we if we work in UK we need a very high level
<$6F> Is it possible to get a good job without qualifications? It is more important to have a
high salary than an interesting job. Should we start with the first one?
<$5F> Yeah sure. Erm I’d say it is impossible to get a good job without good qualifications. For
example you couldn’t erm be a doctor without good qualifications. Erm you couldn’t be a lawyer
without good qualifications for example.
<$6F> I think it depends like for me like what do you define a good job? Because for example if
I had my own way like a good job for me would to be an actress and you wouldn’t necessarily
need qualifications. You just need to know the right people and talent erm <$=> or </$=> I
guess it is possible I think it’s just if you’re in the right place at the right time.
<$5F> It’s definitely dependent on erm what you like say is a good job.
<$6F> Yeah like what the profession is +
Figure 2.10 Linking ideas with I and synonyms in the native speaker data
52 Linguistic competence
2.6.2 Er/Erm
Er and erm will be further discussed in chapter 4 in relation to discourse
competence but here they are analysed due to their high frequency and con-
tribution in terms of frequency. Er occupied first and third position in the
B1, B2 and C1 frequency lists and erm was 6th most frequent word at B1
and C1, and 11th most frequent at B2. The percentage coverage and fre-
quencies of er gradually fell as proficiency increased but erm showed no
such pattern. It fluctuated between 1.4–2.0% of the B1, B2 and C1 corpora
and its normalised frequencies were comparable, albeit higher, than in the
LINDSEI data (see table 2.7)
T-tests revealed that frequency changes for erm across the levels showed no
statistically significant differences. T-tests revealed that frequency changes
for er, on the other hand, showed key differences. Though there was no
statistically significant increase in usage from B1 to B2, overall test figures
uncovered that B1 (p<0.01) and B2 (p<0.03) learners used er significantly
more than C1 learners with mean figures decreasing from 82 at B1, to 77 at
B2 and to 45 at C1.
Examination of collocational information also revealed patterns in the use
of er and erm, in particular with coordinating and subordinating conjunc-
tions (see tables 2.8 and 2.9), a point which will be developed further in
chapter four. This was especially true of er and erm at B2 which collocated
significantly with and, but, because, and so.
The combinations of er and erm with different conjunctions at B1, B2 and
C1 may have implications for success in terms of the fluency portrayed by
USTC learners at boundaries where utterances are combined to form chains
of discourse. On one hand, filled pauses such as these can be discouraged in
teaching materials as, rather bluntly, they can make speakers ‘sound stupid’
(Viney and Viney 1996 cited by Hughes 2011: 37). This data, however, ques-
tions whether such a concerted effort is required. Though Biber et al. (1999)
group fillers under dysfluency features, for Götz (2013: 36), they can act as
B1 B2 C1
Word List Pn Freq NF List Pn Freq NF List pn Freq NF List pn Freq NF List pn Freq NF
ER 1 1391 810.09 1 1312 615.99 3 770 333.58 4 23925 301.71 17 88188 88.51
ERM 6 320 186.36 11 308 144.61 6 460 199.28 11 10354 130.57 27 62086 62.31
Table 2.8 Collocates of er
N B1 B2 C1
Word Freq. T-score MI score Word Freq. T-score MI score Word Freq. T-score MI score
N B1 B2 C1
Word Freq. T-score MI score Word Freq. T-score MI score Word Freq. T-score MI score
1 I 231 13.820 3.463 I 152 11.244 3.506 THE 212 12.991 3.214
2 ER 227 13.348 3.132 YOU 73 7.668 3.286 I 156 11.203 3.278
3 THE 118 9.684 3.204 THINK 61 7.300 3.935 ER 143 10.675 3.221
4 AND 86 8.224 3.143 IS 57 6.945 3.642 TO 108 9.165 3.082
5 IS 61 7.124 3.509 A 52 6.560 3.469 THINK 95 9.048 3.801
6 THINK 54 6.748 3.614 LIKE 51 6.526 3.537 YOU 74 7.732 3.304
7 IN 48 6.133 3.123 BUT 43 6.044 3.675 OF 62 7.085 3.318
8 BECAUSE 42 5.941 3.586 CAN 37 5.598 3.650 BECAUSE 54 6.817 3.790
9 LIKE 37 5.404 3.163 MY 33 5.111 3.180 SOME 42 6.087 4.042
10 VERY 31 5.163 3.783 SO 30 4.812 3.042 PEOPLE 39 5.674 3.451
11 CAN 30 4.968 3.426 FOR 28 4.754 3.298 MAYBE 33 5.429 4.186
12 WE 29 4.870 3.387 SOME 26 4.685 3.623 MY 35 5.354 3.395
13 OF 27 4.620 3.172 THEY 28 4.674 3.100 ABOUT 31 5.124 3.649
14 HAVE 21 4.091 3.221 BECAUSE 26 4.569 3.266 THIS 29 4.975 3.713
15 IF 18 3.936 3.788 YES 22 4.271 3.482 KNOW 27 4.805 3.732
16 FOR 19 3.847 3.089 PEOPLE 22 4.265 3.464 CAN 29 4.723 3.024
17 GOOD 17 3.821 3.769 DO 21 4.119 3.305 LOT 23 4.497 4.003
18 WITH 17 3.726 3.376 THIS 19 3.938 3.372 IF 25 4.474 3.249
19 WHEN 17 3.658 3.149 HOW 3.91 17.000 4.246 ARE 25 4.442 3.164
20 ALWAYS 15 3.642 4.070 DON’T 18 3.885 3.568 NOT 24 4.326 3.095
*Top 20 collocates with significant t- and MI scores
Linguistic competence 55
2.7 Keywords
The keywords for B1-C1 levels and the NS corpus are shown in tables
2.10–2.13 in comparison to the LINDSEI corpus.
Analysis using the LINDSEI wordlist as a comparison showed that the
keywords produced tended to be more characteristic of the topics discussed
(see chapter 1 for an overview of the topics used in the test data). This
includes words such as Preston, tourism, football and cinema, which reflect
to a degree the topics discussed in the test based around learners’ lives and
their free time. We do not consider such keywords as central to analysis of
learners’ linguistics competence. Instead, discussion will follow based on two
keywords: think and can. These have been chosen because they appear in the
lists at all levels and also because close examination of the data revealed they
fulfil important functions at all levels and help us to define how successful
learners at different levels express themselves.
Table 2.10 B1 Keywords
N Keyword B1 – LINDSEI
N Keyword B2 – LINDSEI
N Keyword C1 – LINDSEI
N Keyword NS – LINDSEI
2.7.1 Think
Think was of fundamental importance to learners at all levels and was also
a keyword in our NS data. Juilland’s D of approximately 0.9 across all lev-
els also demonstrates its significance in the USTC data. Although we would
expect it to arise as a keyword due to the nature of the corpus data, com-
parison to LINDSEI lists showed that it still occurred more frequently in the
USTC data. In addition, as table 2.14 shows, think increased in frequency
as proficiency developed, and we can see it occurred on average more fre-
quently at C1 level and with a higher keyness factor.
Unsurprisingly, I topped the collocates lists for t-scores and MI scores
across all levels, a finding reflected in the LINDSEI data but colligational
analysis show some interesting patterns in the data. These are displayed in
table 2.15.
The most common colligation contained a complement with an adjec-
tive phrase (this was set apart from comparative and superlative phrases to
enhance analysis). One pattern in particular I [er/erm] think + object + is
very important was prominent in learner utterances:
• B1 = 11 occurrences
• B2 = 12 occurrences
• C1 = 31 occurrences
USTC LINDSEI
Usage B1 B2 C1
city, thing, habit, subject, man, sport, or problem. Such utterances were sim-
ply used to provide an answer to a question briefly or to round off an utter-
ance, as we can see in figures 2.12–2.14.
Since the occurrence of comparative and superlative and relative clause
colligations were low, I think was instead analysed for the collocations
which surrounded it in learner discourse using a window of 5 words left
and right to see if differences arose (see table 2.16).
As the collocations suggest, I think was found to be multifunctional act-
ing as a discourse marker, stance marker and hedging device (see Carter
and McCarthy 2006). It displayed a sequencing function at all levels due to
combinations with and and so; the latter often used to conclude or round
off remarks (see chapter 4 for more in-depth discussion). Conjunctions such
60 Linguistic competence
<$0> Okay. Would you prefer TV without advertising? Would you like +
<$5M> No. No I I don’t agree it because I I have studies two years media and er in the future I
want er go to a TV station. The er the television is erm is a very important part of earn money er
in the of the to the TV er station. So I think it is not a good thing for the TV station.
<$0> Er okay <$15F> what are the major transport problems in your country? Major transport
problems in your country.
<$15F> Mm major transport problems in my country? Okay I think it’s the traffic jams
N B1 B2 C1
Word Freq. T-score MI score Word Freq. T-score MI score Word Freq. T-score MI score
1 ER 101 14.242 11.293 ER 151 12.283 11.260 THE 140 11.828 11.461
2 THE 27 9.051 11.001 THE 95 9.743 11.258 ER 106 10.292 11.634
3 I 22 7.992 9.933 I 72 8.480 10.695 ERM 92 9.590 12.173
4 ERM 29 7.679 11.630 IS 62 7.872 12.030 IS 69 8.304 11.900
5 IS 11 7.414 11.681 ERM 51 7.139 11.785 IT'S 60 7.744 12.293
6 SO 32 6.706 11.392 SO 37 6.081 11.612 AND 52 7.207 10.818
7 IT’S 5 6.402 12.179 IN 37 6.080 11.034 I 45 6.703 10.330
8 A 13 5.828 11.086 IT'S 35 5.915 12.160 SO 43 6.555 11.717
9 IN 14 5.565 10.814 AND 34 5.826 10.260 YEAH 43 6.555 11.133
10 TO 16 5.472 10.128 YEAH 32 5.654 10.810 IN 41 6.400 11.159
11 AND 21 5.472 9.946 A 27 5.193 10.790 THAT 38 6.163 12.165
12 VERY 5 5.290 11.958 TO 27 5.191 9.942 MM 37 6.082 12.762
13 BECAUSE 15 4.794 11.039 BECAUSE 25 4.998 11.477 TO 37 6.078 10.383
14 YOU 6 4.579 10.485 YOU 25 4.995 10.007 IT 30 5.475 11.413
15 THIS 4 4.471 11.868 IT 24 4.897 11.257 OF 29 5.383 11.068
16 MM 11 4.357 11.399 BUT 24 4.897 11.100 A 28 5.289 10.851
17 MY 7 4.356 10.423 FOR 23 4.794 11.281 FOR 27 5.195 11.750
18 GOOD 4 4.242 12.174 THIS 22 4.689 11.850 IMPORTANT 26 5.098 12.495
19 YEAH 12 4.239 10.274 MM 22 4.689 11.630 VERY 25 4.999 11.949
20 FOR 6 3.871 11.070 MAYBE 20 4.471 11.856 BUT 23 4.794 11.652
62 Linguistic competence
speakers to reduce the demands on thinking time during the interaction and
supply learners with a ‘way into’ the discourse. In terms of learner success,
I think often satisfied what the CEFR suggests at B1 levels is the need for
learners to express ideas as a ‘linear sequence of thoughts’, B2 calls for more
‘detailed descriptions’ and for C1 calls for learners to ‘develop particular
points’ as per the productive sustained monologue descriptors (Council of
Europe 2001: 59).
2.7.2 Can
This word was focused upon because it features as a keyword at each level
but not in the NS data. Can peaks in terms of frequency at B2 level and its
keyness is also greatest at this level. This suggests that can has most impor-
tance at this level. These patterns are summarised in table 2.17.
When looking at patterns of collocations, quite distinct differences were
found. The most common pattern at B1 was I can (I can = 70 occurences,
you can = 26, we can =36), while at B2 this was you can (you can = 64
occurences, I can = 40, we can = 22) and C1 this was we can (we can = 51
occurences, I can = 35, You can = 42). Initial observation of the data sug-
gested that this showed that can occupied different functions at different
levels. At B1 level, the most common use was simply to describe that the
speaker has the possibility to do something. This was most commonly
related in some way to the speaker, rather than to descibe a specific ability
to do something e.g. when I stay in my room I can play computer games.
At B2 level, the function was very similar but the use of ‘you’ also allowed
speakers to discuss this general possibility to do something extended
beyond the individual speaker and to a more general theoretical possibil-
ity as a synonym of ‘one’ e.g. If you want to learn a new language er it’s
not difficult you can translate. At C1, can was combined with the general
use of ‘we’ discussed earlier (section 2.6.1) extending to wider groups of
people or society in general and the possibilities they have or do not have,
e.g. we can communication with the others er who are come from other
countries we can know their culture and can make friends. These functions
are shown in figures 2.15–2.17.
USTC LINDSEI
<$0> Thank you. <$2M>. What kind of books do you enjoy reading?
<$2M> Actually er I’m not er like the book <$=> I always er watch </$=> I always er enjoy
the <$G2> because erm I can er learn lots of things about them maybe like fishing.
<$10M> Okay. Necessary rich yes erm I don’t think so. I don’t think so yes but erm rich th= this
words I can’t understand er this this words. That means about your hand and about your money.
Maybe ri= rich er in from my perspective rich that means you have a lot of ideal yes you can
have a erm a strange inspiration maybe if if if er that’s rich that means . . .
<$16F> In the travel I think the environment friendly it means we go to travel not by a car and it
not er mm we can protect er environment and that is er mm er green green travel yes erm green
travel . . .
The use of yeah with other keywords and chunks (as we will discuss further
in chapter 4) allowed successful learners at this level to respond to the turn
of their partner more easily and develop and co-construct ideas. At the B1
and B2 levels, learners showed less evidence of this and instead we can was
used to develop ideas within their own turn. Figures 2.18 and 2.19 show
these different uses.
<$35F> <$O76> Mhm we can bring them </$O76> And normally they have hairdryers so +
<$34F> Yes.
<$35F> or laundry er machines so yeah it’s alright.
<$34F> Yeah if there is amenities i= it’s fine +
<$35F> Yeah.
<$34F> + and it’s good for me +
<$35F> Yeah.
<$34F> + but yeah we can bring it.
<$35F> We can bring it yeah
Figure 2.18 The use of yeah with we can to develop points across speaker turns at C1
Rank B1 B2 C1 NS
Rank B1 B2 C1 NS
It is first striking to see how many chunks at all levels again centre on the
verb think, a word which we have noted is important in the data. Whereas
at B1, emphasis was placed on giving opinions with chunks such er I think,
I think er, I think it’s, at B2 and C1, chunks were employed both to give and
seek opinions: I think it’s, so I think, what do you think, do you think about.
We will not analyse the usage of these chunks containing I think in detail in
this chapter, as we have already discussed the use of I think and such chunks
are looked at again in chapter 4, when we explore discourse competence.
Instead, we will focus on the usage of two other high frequency chunks,
namely a lot of and the various patterns of agree with before a comparison
of chunks used in general across levels.
2.8.1 A lot of
A lot of shared a similar dispersion rate across levels in that Juilland’s D for
B1, B2 and C1 was 0.74, 0.71 and 0.71, respectively. It was therefore high not
only in terms of chunk frequency, but it was utilised rather regularly through-
out the exams at each level. A lot of was explored in terms of its influence on
accuracy and collocation. It is not the intention of this book to analyse errors,
but since gains in accuracy affect judgements of success and since accuracy
can rise with the use of chunks, a brief error analysis was conducted.
Firstly, all uses of a lot of with countable and uncountable nouns were
classed as either correct or incorrect. When turned into a mean percentage,
a lot of had accuracy scores of 75.00% at B1, 71.67% at B2, and 91.30%
at C1. At B1, all uses of uncountable nouns (money, art, food, technology,
and time) were accurate but countable nouns proved more problematic with
the plural ‘s’ often missing, e.g. a lot of experiences, buildings, parks, places,
films, books, kinds and words. At B2, again, countable nouns posed more
difficulties than uncountable nouns with the latter being incorrect on only
three out of 27 occasions. In terms of countable nouns, again, plural ‘s’ was
sometimes omitted: a lot of elements, definitions, films, cultures, traditions,
things, programmes, cultures and sources. At C1, all uses of uncountable
nouns, such as food, frustration, benefit, money, slang and entertainment,
were accurate whilst only five instances of countable nouns, problems,
points, countryies, taxis and expressions were incorrect out of 46.
Statistical significance tests were carried out to establish whether i) there
were increases in usage of a lot of across the three levels or ii) whether there
were changes in accuracy scores. Though no statistically significant out-
comes were found between the levels in both of these respects, comparative
analysis of much and many provided further insight into the worth lexi-
cal chunks have for learner speech. Table 2.21 displays the frequencies and
accuracy scores of many, much and a lot of at B1, B2 and C1.
From this data, it can be argued that chunks such as a lot of have a posi-
tive effect on learner accuracy. Combined scores of much, many and a lot of
Linguistic competence 69
Table 2.21 Comparison of many, much and a lot of across B1, B2 and C1
displayed small drops in accuracy at B2 and C1, but at the lowest level in the
USTC, the chunk increased accuracy by 11%. With general usage of much
and many increasing from B1 to C1, perhaps simultaneously suggesting a
growth in learner confidence with these quantifiers, such data could lend
further support to studies which advocate the benefits of lexical chunks on
learner accuracy and their speech in general.
With regard to collocates with a lot of, a wide variety of items in right posi-
tion was not found due to the small frequencies involved. T-scores were often
lower but MI scores were still able to show that collocations comparable with
LINDSEI did exist. These collocates are shown in tables 2.22 and 2.23.
What is interesting to note here, despite the low frequencies, is the use of
vague language by USTC learners. The MI scores for a lot of things are
comparable to LINDSEI at B2 and C1; its t-score at B1 was not found to be
significant. To be successful, it is necessary that learners become comfortable
and more confident in using vague language. As an element of speech which
can reduce online processing and reduce hesitation, it could also be of bene-
fit to learners who may not always be able to retrieve the precise vocabulary
they require to complete their utterance. The sample in figure 2.20 shows
this chunk in use.
B1 B2 C1
Word Freq. T-score MI score Word Freq. T-score MI score Word Freq. T-score MI score
Word B1 B2 C1 LINDSEI
Freq. T-score MI Freq. T-score MI Freq. T-score MI Freq. T-score MI
score score score score
People 6 1.998 9.712 8 2.643 10.078 12 3.314 10.471 111 9.798 14.770
Things 2 1.413 10.485 9 2.828 12.292 7 2.449 11.621 75 8.366 15.386
Money 3 1.732 12.655 - - - - - - --- 3 1.731 10.654 35 5.657 16.016
Time 5 1.997 9.578 23 4.582 12.202 - - - - - - --- 58 7.000 14.413
Food 2 1.413 10.485 - - - - - - --- 3 1.731 10.352 5 1.732 13.324
Friends 2 1.413 9.860 5 1.999 11.094 - - - - - - --- 16 3.741 13.223
<$0> Er <$27F> how might your culture be different in 100 years’ time?
<$27F> In 100 years’ time in the future. Erm <$E laughs /$E> er I think there is huge huge
change will happen in the future. If er God give me a longer life to see that I will see my my my
my kids to grow in the in that time so I think a lot of things will change. Nobody have s= er the
own or local culture all the culture in the world er the same I think so.
1 Arabic a lot of time when we study. Yeah I agree with you but er I also may use technology
2 How about you ? Er I agree with you but er smartphone at this moment it’s very
3 drive to the city centre well . Yes I agree with you but I think I still enjoys life in er city
4 they have to control this. Yeah Yes I agree with you but er as you know everything it have
5 good teacher. I I agree with you I agree with you but er it depends with the teacher.
6 Mm yeah I + Sorry. + abso= absolutely agree with you but er mm if if cold weather er I
7 er something. Okay er I am totally agree with you but if er the family er make timetable for
8 at the computer. Yeah absolutely I agree with you but er they less communication with
the emphatic use of absolutely to imply explicit agreement with the previous
statement, it is followed by the conjunction but. This was found in a number
of exams at B2 as the example in figure 2.21 shows.
Whilst but was a significant collocate at B2 (t = 3.581, MI = 4.541), it
did not appear in the B1, C1 or LINDSEI collocates lists at all. At C1, the
conjunction but was instead replaced with because (t = 1.946, MI = 5.091).
Though only occurring on four occasions, and once being used to show
72 Linguistic competence
1 environment. No this time I don’t agree with you because it’s not only the cars + Yeah. +
2 erm rubbishes to the bins. Yeah I agree with you because erm people use education to teach
3 cancer tongue cancer well. Yes I agree with you because junk food is also food and also h
4 nment and give us a green life. I agree with you because also the cars er pollute the envi
2.9 Discussion
Overall, the data demonstrated that USTC learners did use formulaic chunks
in their speech. Averaging between 30 and 35 three-word chunks and approx-
imately 12 four-word chunks per exam, formulaic language allowed learners
a degree of success in their interactions. However, correlation between rising
chunk numbers and rising proficiency could not be established. For instance,
though three-word chunk numbers grew by 60 from B1 to C1, statistical sig-
nificance was not reached for either the three- or four-word categories. The
frequency of chunks was therefore not a reliable indicator of growing profi-
ciency in the USTC data. One reason explaining this could be the differences
exhibited in individual exams. Three-word chunk data showed considerable
Linguistic competence 73
2.10 Conclusion
Although we have only described certain aspects of this competence due to
the many aspects of linguistic competence, they still lead us to a number of
74 Linguistic competence
• B1, B2 and C1 learners were comparable in their combined K-1 and K-2
token and family coverages. K-1 and K-2 tokens stood at 97% whereas
family coverage fell between 81–84%.
• Learners did not use less frequent vocabulary, according to K-1 and K-2
bands, as proficiency developed so the 2000 most frequent words are
therefore fundamental to success at each level.
• The number of words used, i.e. tokens, increased significantly from B1
to B2. It was not, however, a distinguishing feature of the changes from
B2 to C1.
• The 20 most frequent words at B1, B2 and C1 comprised approximately
40% of all speech. Knowledge of these words is vital for success.
• We was more frequent in the learner data than the NS data. We was
frequent at each level but the functions changed and broadened as the
levels increased. At B1 and B2 levels, the use tended to be focused on
third parties, whereas at C1 level, the learners were also able to use this
to discuss concepts in a general sense.
• Er and erm were very frequent at all levels but hesitation via er did
reduce as proficiency levels increased, something which is reflected in
the CEFR descriptions of fluency
• Hesitation was often concentrated around conjunctions which linked
utterances and it appeared in many chunks and commonly functioned as
a means of buying time and signalling that the turn was to be continued.
Successful speech requires er and erm to realise discourse competence
(see chapter 4 for further discussion).
• The verb think was a keyword at each level and was used throughout in
a range of patterns, illustrating how communicative routines for giving
opinions could be realised. Analysis revealed a variety of functions. It
enabled learners to successfully sequence utterances, shift focus, express
stance and hedge language.
• Can was also a keyword at each level and was used most often to discuss
general possibility, with the focus shifting as the levels increased. At B1,
this was used to talk from the speaker’s viewpoint regarding what I can
do and at B2 and C1 levels, from a more hypothetical of general view-
point with patterns around you can and we can.
• Success at higher proficiency levels related to the flexibility in functions
an individual word could satisfy. Just because a word is used frequently
does not mean that it is always repetitive or used in the same way.
• Though increases in frequency of lexical chunks were evident across the
levels, there were no statistically significant gains in usage as proficiency
grew.
Linguistic competence 75
References
Adolphs, S. and Carter, R. 2013. Spoken corpus linguistics: From monomodal to
multimodal. Abingdon: Taylor and Francis.
Adolphs, S. and Schmitt, N. 2003. Lexical coverage of spoken discourse. Applied
Linguistics, 24(4), 425–438.
Anthony, L. 2017. AntConc. (Version 3.4.4). [computer software] Tokyo, Japan: Waseda
University. Available from: <www.laurenceanthony.net/> [Accessed 28 March 2017].
Bachman, L.F. and Palmer, A.S. 1996. Language testing in practice. Oxford: Oxford
University Press.
Baker, P. 2006. Using corpora in discourse analysis. London: Continuum.
Barnbrook, G. 1996. Language and computers: A practical introduction to the com-
puter analysis of language. Edinburgh: Edinburgh University Press.
Biber, D., Johansson, S., Leech, G., Conrad, S. and Finegan, E. 1999. Longman gram-
mar of spoken and written English. Harlow: Longman.
Brezina, V. 2014. Statistics in Corpus linguistics: Web resource. [online] Available
from: <http://corpora.lancs.ac.uk/stats> [Accessed 2 June 2016].
Bygate, M. 1987. Speaking. Oxford: Oxford University Press.
Canale, M. 1983. From communicative competence to communicative language
pedagogy. In: J.C. Richards and R.W. Schmidt, eds. Language and communica-
tion. New York: Longman, 2–27.
Canale, M. and Swain, M. 1980. Theoretical bases of communicative approaches to
second language teaching and testing. Applied Linguistics, 1(1), 1–47.
Carter, R. and McCarthy, M. 2006. Cambridge grammar of English: A comprehen-
sive guide: Spoken and written English grammar and usage. Cambridge: Cam-
bridge University Press.
Chujo, K. 2004. Measuring vocabulary levels of English textbooks and tests using
a BNC lemmatised high frequency word list. Language and Computers, 51(1),
231–249.
Cobb, T. 2017. Compleat Lexical Tutor (Lextutor). [Online] Available from: <www.
lextutor.ca.> [Accessed 1 January 2017].
Council of Europe 2001. Common European Framework of Reference for Lan-
guages: Language, teaching, assessment. Cambridge: Cambridge University Press.
76 Linguistic competence
Davies, M. 2004. BYU-BNC (based on the British National Corpus form Oxford
University Press). [Online] Available from: <http://corpus.byu.edu/bnc.> [Accessed
15 February 2017].
De Cock, S. 1998. A recurrent word combination approach to the study of formulae
in the speech of native and non-native speakers of English. International Journal
of Corpus Linguistics, 3(1), 59–80.
English Grammar Profile 2016. [Online] Available from: <www.englishprofile.org/
english-grammar-profile> [Accessed 15 January 2016].
English Vocabulary Profile 2016. [Online] Available from: <www.englishprofile.org/
wordlists> [Accessed 15 January 2016].
Erman, B. and Warren, B. 2000. The idiom principle and the open choice principle.
Text, 20(1), 29–62.
Galaczi, E. and Ffrench, A. 2011. Context validity. In: L. Taylor, ed. Examining
speaking: Research and practice in assessing Second language speaking (studies in
language testing). Cambridge: Cambridge University Press, 112–170.
Gilquin, G. and De Cock, S. 2013. Errors in disfluencies in spoken corpora. Amster-
dam: John Benjamins Publishing.
Gilquin, G., De Cock, S. and Granger, S. 2010. LINDSEI: Louvain International
Database of Spoken English Interlanguage. [CD-ROM]. Louvian: Presses Univer-
sitaires de Louvain.
Götz, S. 2013. Fluency in native and non-native English speech. Amsterdam: John
Benjamins Publishing.
Goulden, R., Nation, P. and Read, J. 1990. How large can a receptive vocabulary be?
Applied Linguistics, 11(4), 341–363.
Greaves, C. and Warren, M. 2010. What can a corpus tell us about multi-word units?
In: A. O’Keeffe and M. McCarthy, eds. The Routledge Handbook of Corpus Lin-
guistics. Abingdon: Routledge, 212–226.
Hoey, M. 2005. Lexical priming: A new theory of words and language. London:
Routledge.
Hughes, R. 2011. Teaching and researching speaking. 2nd ed. London: Pearson Education.
Hunston, S. 2002. Corpora in applied linguistics. Cambridge: Cambridge University
Press.
Hymes, D. 1972. On communicative competence. In: J.B. Pride and J. Holmes, eds.
Sociolinguistics: Selected readings. Harmondsworth: Penguin, 269–293.
Laufer, B. and Nation, P. 1995. Vocabulary size and use: Lexical richness in L2 writ-
ten production. Applied Linguistics, 16(3), 307–322.
Laufer, B. and Nation, P. 1999. A vocabulary-size test of controlled productive ability.
Language Testing, 19(33), 33–51.
Lewis, M. 1993. The lexical approach: The state of ELT and the way forward. Hove:
Language Teaching.
McCarthy, M. 1999. What constitutes a basic vocabulary for spoken communica-
tion? Studies in English Language and Literature, 1, 233–249.
McCarthy, M. 2006. Explorations in corpus linguistics. Cambridge: Cambridge
University Press. Available from: <www.cambridge.org/elt/teacher-support/pdf/
McCarthy-Corpus-Linguistics.pdf> [Accessed 14 June 2016].
McCarthy, M. 2010. Spoken fluency revisited. English Profile Journal, 1(1), 1–15.
McEnery, T. and Hardie, A. 2012. Corpus linguistics. Cambridge: Cambridge Uni-
versity Press.
Linguistic competence 77
McEnery, T. and Wilson, W. 2001. Corpus linguistics. 2nd ed. Edinburgh: Edinburgh
University Press.
Meara, P. 1996. The dimensions of lexical competence. In: G. Brown, K. Malmkjӕr
and J. Williams, eds. Performance and competence in second language acquisition.
Cambridge: Cambridge University Press, 70–88.
Nation, I.S.P. 2001. Learning vocabulary in another language. Cambridge: Cam-
bridge University Press.
Nation, P. and Chung, T. 2009. Teaching and testing vocabulary. In: M.H. Long and
C.J. Doughty, eds. 2009. The handbook of language teaching, Oxford: John Wiley
and Sons, 543–559.
Nation, P. and Waring, R. 1997. Vocabulary size, text coverage and word lists.
Vocabulary: Description, Acquisition and Pedagogy, 14, 6–19.
Nattinger, J.R. and DeCarrico, J.S. 1992. Lexical phrases and language teaching.
Oxford: Oxford University Press.
Oakes, M.P. 1998. Statistics for corpus linguistics. Edinburgh: Edinburgh University
Press.
O’Keeffe, A., McCarthy, M. and Carter, R. 2007. From corpus to classroom: Lan-
guage use and language teaching. Cambridge: Cambridge University Press.
Pawley, A. and Syder, F.H. 1983. Two puzzles for linguistic theory: Nativelike selec-
tion and nativelike fluency. In: J.C. Richards and R.W. Schmidt, eds. Language and
communication. New York: Longman, 191–227.
Schmitt, N. 2000. Key concepts in ELT. ELT Journal, 54(4), 400–401.
Schmitt, N. 2008. Instructed second language vocabulary learning. Language Teach-
ing Research, 12(3), 329–363.
Schmitt, N. and Carter, R. 2004. Formulaic sequences in action: An introduction. In:
N. Schmitt, ed. Formulaic sequences. Amsterdam: John Benjamins, 1–22.
Schonell, F.J., Meddleton, I.G. and Shaw, B.A. 1956. A study of the oral vocabulary
of adults. Brisbane: University of Queensland Press.
Scott, M. 2015. WordSmith Tools.v.6.0.0.252. [Online]. Lexical Analysis Software
Ltd. Available from <http://lexically.net/wordsmith/> [Accessed 29 December 2015].
Scott, M. and Tribble, C. 2006. Textual patterns: Keywords and corpus analysis in
language education. Amsterdam: John Benjamins Publishing.
Sinclair, J.M. 1991. Corpus concordance collocation. Oxford: Oxford University
Press.
Stæhr, L.S. 2008. Vocabulary size and the skills of listening, reading and writing. The
Language Learning Journal, 36(2), 139–152.
Thornbury, S. and Slade, D. 2006. Conversation: From description to pedagogy.
Cambridge: Cambridge University Press.
Van Zeeland, H. and Schmitt, N. 2013. Lexical coverage in L1 and L2 listening
comprehension: The same of different from reading comprehension? Applied Lin-
guistics, 34(4), 457–479.
Viney, P. and Viney, K. 1996. Handshake: A course in communication. Oxford:
Oxford University Press.
WebLingua. 2016. Text Inspector. Available from: <www.textinspector.com/> [Accessed
28 March 2017].
Chapter 3
Strategic competence
3.1 Introduction
Identified as a component of communicative competence by Canale and
Swain (1980), strategic competence is typically associated with the ability
to convey and negotiate meaning successfully despite obstacles in the com-
munication process. Previously neglected in competence theories centring
on the native speaker (NS) (Chomsky 1965; Hymes 1972) and receiving less
attention than linguistic and sociolinguistic competences (Tarone and Yule
1989), it was a shift in focus towards the language learner that saw strate-
gic competence transition from being of equal importance to grammatical
and sociolinguistic competences, to it subsequently being foregrounded as
a central element of all language users’ production (Dörnyei and Thurrell
1991; Bachman and Palmer 1996). This chapter aims to briefly review stra-
tegic competence definitions before exploring how it can be exhibited and
utilised in the speech of learners at different levels. The argument we wish
to reinforce is that strategic competence, and the communication strategies
employed by learners, can be facilitative and indicative of successful speech
and, as such, should not simply be dismissed as a sign of insufficient pro-
ficiency. With earlier research also criticised for emphasising discussion of
strategies rather than their actual usage in communication (Ellis 1985), this
chapter additionally aims to illustrate the common patterns that B1, B2, and
C1 language users display in their interactions and the effects they have on
their overall performance as successful speakers of English.
Strategies are a means the language user exploits to mobilise and balance
his or her resources, to activate skills and procedures, in order to fulfil
the demands of communication in context and successfully complete the
task in question in the most comprehensive or most economical way fea-
sible depending on his or her precise purpose. Communication strategies
should therefore not be viewed simply with a disability model – as a way
of making up for a language deficit or a miscommunication. Native speak-
ers regularly employ communication strategies of all kinds . . . when the
strategy is appropriate to the communicative demands placed upon them.
(Council of Europe 2001: 57)
B1, B2, C1 and C2 spoken can-do statements filtered according to relevance for the
USTC speaking exams. This was completed by the researcher and independent
judges.
Final list of selected statements for all three levels constructed. Statements checked
for overlap or repetition to make sure categories were distinct and operationalisable.
All exam transcripts containing either i) learner and examiner language or ii) only
leaner language mined for examples of language satisfying can-do statements.
Independent judges for each level completed the same process using one exam only.
Selected exams compared to ensure rater-reliability and to reduce subjectivity.
Language analysis of codes: can-do occurrence and language use compared across B1,
B2, C1 and C2.
B1 Strategies Can start again using a different tactic when communication breaks down.
(production) Can correct mix-ups with tenses or expressions that lead to misunderstandings provided the interlocutor indicates there is a problem.
Can define the features of something concrete for which he/she can’t remember the word. Can convey meaning by qualifying a word
meaning something similar (e.g. a truck for people = bus).
Strategies Can summarise the point reached in a discussion and so help focus the talk.
(interaction) Can invite others into the discussion.
Can ask someone to clarify or elaborate what they have just said.
Can repeat back part of what someone has said to confirm mutual understanding and help keep the development of ideas on course.
Can invite others into the discussion.
Can exploit a basic repertoire of language and strategies to help keep a conversation or discussion going.
Can initiate, maintain and close simple, face-to-face conversation on topics that are familiar or of personal interest.
Can intervene in a discussion on a familiar topic, using a suitable phrase to get the floor.
B2 Strategies Can use circumlocution and paraphrase to cover gaps in vocabulary and structure.
(production) Can correct slips and errors if he/she becomes conscious of them or if they have led to misunderstandings.
Can make a note of ‘favourite mistakes’ and consciously monitor speech for it/them.
Strategies Can intervene appropriately in discussion, exploiting appropriate language to do so.
(interaction) Can initiate discourse, take his/her turn when appropriate, and end conversations when he/she needs to, though he/she may not always
do this elegantly.
Can use stock phrases (e.g. ‘That’s a difficult question to answer’) to gain time and keep the turn whilst formulating what to say.
Can help the discussion along on familiar ground, inviting others in etc.
Can ask follow-up questions to check that he/has understood what a speaker intended to say, and get clarification of ambiguous points.
C1 Strategies Can backtrack when he/she encounters a difficulty and reformulate what he/she wants to say without fully interrupting the flow of
(production) speech.
(As B2+) Can use circumlocution and paraphrase to cover gaps in vocabulary and structure
Strategies (As B2) Can ask follow-up questions to check that he/she has understood what a speaker intended to say, and get clarification of
(interaction) ambiguous points.
Can relate own contribution skillfully to those of other speakers.
Can select a suitable phrase from a readily available range of discourse functions to preface his/her remarks appropriately in order to
get the floor, or to gain time and keep the floor whilst thinking.
C2 Strategies Can substitute an equivalent term for a word he/she can’t recall so smoothly that it is scarcely noticeable.
(production) Can backtrack and restructure around a difficulty so smoothly the interlocutor is hardly aware of it.
Strategies (As C1) Can select a suitable phrase from a readily available range of discourse functions to preface his/her remarks appropriately in
(interaction) order to get the floor, or to gain time and keep the floor whilst thinking.
(As C1) Can relate own contribution skillfully to those of other speakers.
(As B2) Can ask follow-up questions to check that he/she has understood what a speaker intended to say, and get clarification of
ambiguous points.
Strategic competence 87
B1 production strategies
0 10 20 30 40 50 60
B1 interaction strategies
46
Can invite others into the discussion. 11.59
12.99
0 10 20 30 40 50 60 70 80 90
B2 production strategies
27
Can use circumlocution and paraphrase to cover gaps in
9.47
vocabulary and structure.
32.14
Production strategy
57
Can correct slips and errors if he/she becomes conscious
20.00
of them or if they have led to misunderstandings.
67.86
0
Can make a note of ‘favourite mistakes’ and consciously
0.00
monitor speech for it/them.
0.00
B2 interaction strategies
C1 production strategies
42
Productive Strategies
0.00 10.00 20.00 30.00 40.00 50.00 60.00 70.00 80.00 90.00
C1 interaction strategies
40
Interaction Strategies
75
(as B2) Can help along the progress of the work by
16.34
inviting others to join in, say what they think, etc.
22.32
Similarly, another distinction across the two levels was revealed in learners’
use of production strategies, in particular those associated with correction.
Correction at B2 constituted 68% of production strategy usage and 20%
of all strategy use throughout the discussions; in learner-to-learner commu-
nication, it equated to 56.52% of production strategy use and 17.33% of
all strategy use. Conversely, across all B1 speech, it represented 56% of all
production strategies and 12% of all strategy use but 38.89% of production
strategy usage and 8.24% of all usage in learner-to-learner discussion. This
increase across the two levels could, therefore, to some extent exemplify B2
learners’ ‘new degree of language awareness’ as specified within the CEFR
(Council of Europe 2001: 35). Though no ‘favourite mistakes’ were identi-
fied, learners did correct slips and errors on 57 occasions, a potential indica-
tor of the ‘new degree of language awareness’ which results in their labelling
as ‘vantage’ learners (Council of Europe 2001: 35).
In total, 459 strategies were identified in all C1 speech; 26.80% for pro-
duction, and 73.20% for interaction. Interestingly, despite the fact that this
level had the lowest numbers of selected statements for analysis (B1 = 10,
B2 = 8, C1 = 6 statements), the learners within it produced the highest
amount of statement evidence. Immediately, therefore, it could be argued
that the statements dominating each of the two strategy categories were
fundamental to learners’ overall spoken success. For instance, descriptions
of C1 ability in the CEFR highlight the use of suitable phrases for prefacing
remarks and the ‘controlled use of organisational patterns, connectors and
cohesive devices’ to aid fluency (Council of Europe 2001: 36). With 146
examples of interaction strategy occurrence, nearly a third of that category,
relating to prefacing remarks to get the floor, gain time and keep the floor,
the way that utterances were combined to convey meaning, seemed to take
precedence over the ability to clarify points and invite others into the dis-
cussion. This was replicated in paired learner discussion data which, despite
dramatic falls in prevalence, saw the prefacing statement occupy a major-
ity of 16 (34.78%) interaction strategies at this level. Similarly, with 66%
of production strategy evidence relating to backtracking and reformulat-
ing utterances (57.69% in learner-to-learner speech), once again, the ability
to monitor speech was prominent, as it was in the other two levels. What
is clear from this data, therefore, is that productive strategies do have an
important role to play in successful learner speech in spite of their reduced
occurrence in comparison with interaction strategies.
B1 Can initiate, maintain 80 22.60% Can invite others into the 16 20.51%
and close simple, face- discussion.
to-face conversation
on topics that are of
familiar or of personal
interest.
B2 Can help the discussion 61 30. 35% Can help the discussion 14 18.67%
along on familiar along on familiar ground,
ground, inviting others inviting others in, etc.
in, etc. Can ask follow-up 14 18.67
questions to check that
he/she has understood
what a speaker intended
to say, and get clarification
of ambiguous points.
C1 Can select a suitable 146 43.45% Can select a suitable 16 34.78%
phrase from a readily phrase from a readily
available range of available range of
discourse functions to discourse functions to
preface his/her remarks preface his/her remarks
appropriately in order appropriately in order to
to get the floor, or to get the floor, or to gain
gain time and keep the time and keep the floor
floor whilst thinking. whilst thinking.
94 Strategic competence
What remains to be seen, however, are the factors to which learners lent
their attention. The CEFR beyond B1 level is a little vague in this respect.
B1 descriptors emphasise ‘mix-ups with tense or expressions’, whereas B2
and C1 statements highlight the need to i) repair ‘slips and errors’, or ii)
backtrack upon encountering a ‘difficulty’ (Council of Europe 2001: 65).
Little expansion is offered, therefore, on how such correction or reformula-
tion may represent itself in learners’ speech. To clarify the errors receiving
attention, utterances were assessed for their main corrective focus. Table 3.6
shows, with the help of illustrative learner language, which items were cor-
rected by learners. It also incorporates data from NS speech as a neutral
point for comparison.
The data highlight that B1, B2 and C1 learners were clearly alike in their
ability to correct utterances according to word choice selections. Equating to
approximately half of all learner corrective statements, word choice revisions
comprised 52%, 51% and 47%, respectively, across the levels and 38% in
NS data (again the majority error type). Analysis of successful learner speech
therefore identified that the CEFR’s ambiguous slips and errors related
mostly to the choice of verbs and pronouns (see figures 3.8-3.11), not only
at B2 and C1 but also in B1 and NS speech:
B1 % B2 % C1 % NS %
<$5F> Er I think erm the photographs can be can be important er souvenirs to <$=> our
</$=> to us because
<$29F> First of all if you have a good salary but you don’t like your job you won’t get a you
won’t enjoy it.
<$3F> However <$=> it </$=> they’re two very different things though
<$0> First I’m going to ask you some questions about yourselves. So <$10M> tell me about the
place where you live.
<$10M> Now?
<$0> Yes.
<$10M> Er now I live er I live in I’m living in Preston +
<$31F> + cos some teenagers don’t have like the sense of responsibility +
<$30F> Yeah.
<$31F> + to drive safety so erm <$=> they’ll </$=> they’re thinking about mm hanging out
with their friends so they’re not like with the right sense of safety
<$0> Okay. And final question er <$6F> which job would you least like to do?
<$6F> Erm. I guess one that’s sort of <$=> didn’t doesn’t </$=> doesn’t really go anywhere
or I don’t feel like I’m like utilising my skills
<$37M> I enjoy yeah it’s nice erm I can make new fr= mm new friend I er I mean I erm this
studying er in England in Pr= in Preston I enjoy. I make new friend I spend all my time <$=> I
don’t have time to </$=> sometime I don’t have time to do it.
<$16F> so think when I travel to the American <$=> I can </$=> I also can speak with the
native people and I can speak with British people.
<$20F> Er I want to have my own <$G2> restaurant in the future because I like to eat food and
<$=> I hope I can eat our ah ah wrong word </$=> <$=> so I want my </$=> I want to open
a restaurant in my country.
<$2M> Er I’d agree er like erm I always thought if you know something you can teach it you can
explain it to someone else better <$=> so if you’re just </$=> it’s good enough explaining it to
yourself I suppose.
Table 3.7 CEFR qualitative descriptions of learner accuracy (Council of Europe 2001: 28)
Level Accuracy
speech only occurred once part or all of an utterance had been verbalised (see
figures 3.20 and 3.21). In a sense, the freeing up of online attention resources
which aid more fluent speech at higher levels (see proceduralisation in Tay-
lor 2011), might have been responsible not for quick, minor changes during
such verbalisation, but alternatively, it might have been responsible for the
aforementioned backtracking once the message had been conveyed.
To summarise this section, a closer inspection of production strategy
usage for correction has been used to shed light on the CEFR’s B1, B2 and
C1 statements and how learners can be successful in their speech. Not only
were occurrences of correction seen to rise across the proficiency levels, in
contrast to studies which said strategy use should decrease, but also slips
and backtracking were illustrated by attention to word choice, tense usage,
missing words and occasional rephrasing. Table 3.8 summarises the typi-
cal features reformulated at each level but the data unequivocally does not
associate a higher prevalence of correction to lower achievement or suc-
cess. For instance, the increases in raw frequencies and mean corrections per
exam assert that the ability to self-correct without an interlocutor’s inter-
vention should be seen as a significant indicator of success. Though some
may assume that increasing correction can disrupt the flow of speech, and
therefore perceptions of success, it is a key production strategy in all B1, B2,
C1 and NS speech.
<$3M> I I think that’s a bad thing because <$=> it doesn’t promotes like </$=> even if you
want like your country to erm develop faster you need a good like erm tourism needs to be good
so that erm other countries might want to do like they’d like when they come and they see what
they like they want to do business with your country
<$4F> + well that’s a good point <$O18> television </$O18> or magazine I prefer well I
don’t <$O19> I don’t really </$O19> no <$O20> not magazine </$O20>
Table 3.10 Top three questions used at B2 to invite others into the conversation
Table 3.11 Top three questions used at C1 to invite others into the conversation
What do you 19 25.33 And what do you think about a reasonable price?
think/ reckon Er erm how what do you reckon about this one?
about/of . . .?
How/What 11 14.67 How about the second point?
about . . .? What about finding information on the internet?
Do/Did 7 9.33 Like do you spend time going through it see
you . . .? what other erm other people stayed in the hotel
thinks about the hotel itself or?
Do you think if we visit the places where it
happened er it will still be the same?
102 Strategic competence
Table 3.12 The three most frequent clarification requests at B1, B2 and C1
<$0> <$26M> er what do you like about the place where you live?
<$26M> Mm here or in my country?
<$0> Any.
<$0> Mhm okay and how important do you think cultural awareness is when you’re travelling?
<$10F> Er so you mean cultural awareness?
<$0> Mhm.
<$10F> Mm it means er the people in the other countries tell their feelings?
3.6 Conclusion
The data presented in this chapter have allowed us to reach several conclu-
sions regarding the role of strategic competence in successful learner speech.
Using CEFR spoken strategy statements as a basis for analysis, we have docu-
mented how the use of communication strategies in real-time learner speech
facilitates spoken production by allowing learners to manage discourse and
exploit their linguistic resources in response to the challenges they face.
Though traditionally communication strategies have been viewed in a nega-
tive light as indicators of lacking proficiency, this chapter has demonstrated
that they in fact can be a sign of improving proficiency as the use of CEFR
strategies was found to be highest at C1 level despite a reduced number of
descriptors. The occurrence of different types of strategies also unveiled typi-
cal characteristics that are expected to be seen at each level. In terms of inter-
action, B1 learners were able to maintain discourse and convey their desired
meanings, B2 learners demonstrated an ability to initiate, maintain and close
106 Strategic competence
References
Bachman, L. and Palmer, A. 1996. Language testing in practice: Designing and devel-
oping useful language tests. 3rd ed. New York: Oxford University Press.
Bialystok, E. 1983. Some factors in the selection and implementation of communica-
tion strategies. In: C. Faerch and G. Kasper, eds. 1983. Strategies in interlanguage
communication. London: Longman. 79–99.
Bialystok, E. 1990. Communication strategies. Oxford: Blackwell.
Bialystok, E. and Fröhlich, M. 1980. Oral communication strategies for lexical dif-
ficulties. Interlanguage Studies Bulletin, 5(1), 3–30.
Canale, M. 1983. From communicative competence to communicative language
pedagogy. In: J.C. Richards and R.W. Schmidt, eds. 1983. Language and Com-
munication. New York: Longman, 2–27.
Canale, M. and Swain, M. 1980. Theoretical bases of communicative approaches to
second language teaching and testing. Applied Linguistics, 1(1), 1–47.
Chang, S. and Liu, Y. 2016. From problem-orientedness to goal-orientedness:
Re-conceptualizing communication strategies as forms of intra-mental and inter-
mental mediation. System, 61, 43–54.
Chen, S. 1990. A study of communication strategies in Interlanguage production by
Chinese EFL Learners. Language Learning, 40(2), 155–187.
Chomsky, N. 1965. Aspects of the theory of syntax. Cambridge: The MIT Press.
Cohen, A. 2014. Strategies in learning and using a second language. 2nd ed. Abing-
don: Routledge.
Cohen, L., Manion, L. and Morrison, K. 2011. Research methods in education. 7th ed.
London: Routledge.
Corder, S. 1981. Error analysis and interlanguage. Oxford: Oxford University Press.
Council of Europe 2001. Common European framework of reference for languages.
Cambridge: Cambridge University Press.
Dörnyei, Z. 1995. On the teachability of communication strategies. TESOL Quar-
terly, 29(1), 55–85.
Dörnyei, Z. 2007. Research methods in applied linguistics. 1st ed. Oxford: Oxford
University Press.
Dörnyei, Z. and Scott, M.L. 1995. Communication strategies: An empirical analysis
with retrospection. In: J.S. Turley and K. Lusby, eds. Selected papers from the pro-
ceedings of the 21st annual symposium of the Deseret Language and Linguistics
Society. Provo, UT: Brigham Young University, 155–168.
108 Strategic competence
Discourse competence
4.1 Introduction
This chapter will define discourse competence broadly as the ability of learn-
ers to produce and manage spoken texts with cohesion and coherence. This
is similar to the definition given in the CEFR (2001: 123), which suggests
it can be defined as the way in which messages are ‘organised, structured
and arranged’. This broad definition will be employed to analyse how suc-
cessful speakers use specific features of spoken language to create successful
spoken texts across different levels. After we define terms and review some
related previous studies in sections 4.2 and 4.3, the data will be explored in
two main ways. Firstly, in section 4.4 we will adapt Celce-Murcia, Dornyei
and Thurell’s (1995) more detailed definition of discourse competence and
explore their main categories in terms of the language used to achieve key
aspects of discourse competence at different levels. Secondly, in section 4.5,
we will focus on how speakers use discourse markers (DMs) to help them
to achieve discourse competence. As we do so, we will attempt to show how
learners use different aspects language to achieve the CEFR (2001: 124/125)
‘Can do’ statements related to discourse competence at different levels, but
we also note that some of these statements are so vague that they are difficult
to measure. As in previous chapters, some comparison will be made to our
native-speaker test data where we feel this illuminates the analysis. Overall,
we wish to show that successful speech, even at lower levels, tends to pay
attention to language use beyond simply constructing sentences and towards
linking ideas within and across turns.
ability to link ideas together to form cohesive and coherent texts of differ-
ent types. In a written form, this may be the ability to write, as an example,
a postcard which is understood by its recipient, while in a spoken form it
could be the ability to request a money transfer from your bank on the
telephone. The CEFR (Council of Europe 2001: 123) defines discourse com-
petence as ‘the ability of a user/learner to arrange sentences in sequence
so as to produce coherent stretches of language’. While both definitions
seem reasonable (if somewhat vague), it is of course inadequate to describe
spoken language in terms of ‘sentences’ and cohesion and coherence need
a more detailed description because in conversation this is achieved both
within a speaker’s turn and across the turns of different participants. For
these reasons, and for the purposes of this chapter, we will use the more
detailed definition of discourse competence given by Celce-Murcia, Dornyei
and Thurell (1995), alongside the broad definition given at the start of this
chapter. Celce-Murcia, Dornyei and Thurell suggest that this competence
can be defined as follows: ‘Discourse competence concerns the selection,
sequencing, and arrangement of words, structures, sentences and utterances
to achieve a unified spoken or written text’ (Celce-Murcia, Dornyei and
Thurell 1995: 15). They go on to suggest that this competence can be further
defined and measured by examining the following elements:
the conversation, to join his or her ideas together across and within turns and
to close down or open new topics and turns. These skills will of course be
achieved in a number of ways, at what Celce-Murcia, Dornyei and Thurell
(1995) suggest is a macro level and a micro level. We are taking the macro level
to mean the larger more holistic view of spoken discourse as often exemplified
in the CEFR ‘can do’ statements and the micro level to mean the ‘smaller’ lan-
guage features which help speakers to realise coherent and cohesive discourse.
The language features we will explore are items set out in section 4.4 and
include items such as it for anaphoric reference. We also explore spoken dis-
course markers, and it is therefore necessary to define these in the next section.
1 DMs are lexical items or phrases (Redeker 1990, Carter and McCarthy
2006), such as ‘right’, ‘I mean’, ‘you know’, ‘I think’.
2 DMs are optional – the absence of a DM does not affect the semantics
or grammar of an utterance. However, the absence will make compre-
hension at least more difficult (Aijmer 2002, Eslami and Eslami-Rasekh
2007).
3 DMs are multifunctional – the same DM can have a variety of functions,
each dependent on context. Fung and Carter (2007) give the example of
‘so’, which can, for instance, both summarise and launch a topic.
112 Discourse competence
4 DMs are not drawn from one grammatical class and are not a closed
grammatical class. Aijmer (2002), Carter and McCarthy (2006) and
Fung and Carter (2007), give examples of DMs drawn from a wide
variety of grammatical classes, such as prepositional phrases (‘by the
way’), response tokens (‘right’) and interjections (‘oh’).
5 DMs have a procedural but not propositional meaning. A DM may
possess a propositional meaning when used as part of another class. An
example of this is the temporal use of ‘now’. The meaning of a DM can
be defined from the broader context in which it operates.
6 DMs function at a referential, interpersonal, structural and cognitive
level (Aijmer 2002; Fung and Carter 2007). They act as signposts for
speakers and listeners as they orientate themselves to the ongoing dis-
course (Schiffrin 1987; Aijmer 2002) by, for instance, signalling that lis-
teners need to time to think or that they wish to show they are listening.
7 DMs are often (but not always) sentence or turn initial (Aijmer 2002;
Fung and Carter (2007).This position occurs often as it fulfils a number
of common functions, such as launching topics (Fung and Carter 2007).
8 DMs ‘should be prosodically independent and be largely separate from
the utterances they introduce’ (Fung and Carter 2007: 413). This will
generally be indicated by the DM occupying a separate tone unit and
(often) being followed by a pause.
that did tended to be more acculturated. Gilquin (2008) found that French
leaners of English used fewer DMs as lexical markers of hesitation (such as
‘like’) compared to native speakers of English. Tsai and Chu (2015) looked
at the use of DMs by Chinese L1 speaking teachers and learners of Chinese
as an L2, studying both inside and outside China. They found that the fre-
quency of use of DMs was linked to greater levels of fluency by individual
speakers. While such studies are illuminating, we wish in this chapter to
explore the use of DMs by successful learners to show how they contrib-
ute to discourse competence in differing ways at each level. We make some
comparison to our native-speaker data but this is not only to illustrate that
learners use fewer DMs but rather to explore which ones they do use, how
they use them and why they do so.
B1 Can adapt his/her expression Can link a series of Can reasonably fluently Can initiate, maintain and close simple
to deal with less routine, even shorter, discrete simple relate a straightforward face-to-face conversation on topics that are
difficult, situations. elements into a connected, narrative or description familiar or of personal interest.
Can exploit a wide range of linear sequence of points. as a linear sequence of
simple language flexibly to points.
express much of what he/she
wants.
B2 Can adjust what he/she says and Can use a limited number Can develop a clear Can initiate discourse, take his/her turn
the means of expressing it to of cohesive devices to link description or narrative, when appropriate and end conversation
the situation and the recipient his/her utterances into expanding and supporting when he/she needs to, though he/she may
and adopt a level of formality clear, coherent discourse, his/her main points with not always do this elegantly.
appropriate to the circumstances. though there may be relevant supporting detail Can use stock phrases (e.g. ‘That’s a
Can adjust to the changes of some ‘jumpiness’ in a long and examples. difficult question to answer’) to gain time
direction, style and emphasis contribution. and keep the turn whilst formulating what
normally found in conversation. to say.
Can vary formulation of what he/ Can intervene in a discussion on a familiar
she wants to say. topic, using a suitable phrase to get the
floor.
C1 As of B2 Can produce clear, Can give elaborate Can select a suitable phrase from a readily
smoothly flowing, well- descriptions and available range of discourse functions to
structured speech, narratives, integrating preface his/her remarks appropriately in
showing controlled use of sub-themes, developing order to get the floor, or to gain time and
organisational patterns, particular points and keep the floor whilst thinking.
connectors and cohesive rounding off with an Can intervene appropriately in discussion,
devices. Can use a variety appropriate conclusion exploiting appropriate language to do so.
of linking words efficiently Can initiate, maintain and end discourse
to mark clearly the appropriately with effective turn-taking.
relationships between ideas.
116 Discourse competence
analysis is necessarily selective but we felt that the language chosen could
reflect key aspects of discourse competence. For these reasons, we chose to
look at the following language areas in order to explore discourse competence:
The corpus was searched for the items above and was then analysed manu-
ally to ensure that the items searched for were being used with the functions
mentioned in 1–4 above and only items which functioned in these ways were
counted. To give one example, it used as a form of cataphoric reference in
figure 4.1 sample ‘A’ would be discounted and it used as a form of anaphoric
reference in figure 4.1 sample ‘B’ would be accepted. Both the examples
below are taken from B2 level.
When analysing DMs in the second part of the chapter, a combination of
quantitative and qualitative analysis was also undertaken. We first simply
looked at the frequency of the following DMs: Er, yeah, I think, Ok, You
know, Well, I mean, Right, Anyway.
These items were chosen because they have been found to be high fre-
quency (O’Keeffe, McCarthy and Carter 2007) in spoken corpora in general
and, as noted in chapter 2, this includes our corpus to a certain degree. They
also fulfil different functions according to the analysis described by Fung
and Carter (2007). These functions can be defined as follows:
Sample A
<$6> some advertisements should er limited it’s because should er from the parents side they
should protect their young children.
Sample B
<$2F> Er maybe winter because I like snow and er I like er play snow with my friends. I think
it’s really interesting
We are aware that not everyone will define er as a discourse marker, but
we wish to suggest that it meets many of the criteria for a DM and this is
something others have also suggested (see for example, Hoey 2004). It is,
for example, mainly turn initial and prosodically independent and carries a
procedural rather than a propositional meaning.
In order to analyse this data, we explored the USTC data at different levels
and qualitatively assessed each instance to check it was being used as a DM
as some of the items listed can have a different propositional meaning. Ok,
for instance, can also be used as an adjective. Only those uses where the
items functioned as DMs were counted. Following this initial analysis, fre-
quency levels were compared using log-likelihood scores. Oakes (1998) and
Jones and Waller (2015) describe this as a measure which allows us to check
frequency of occurrence in different corpora for statistical significance. This
does so by checking the relative frequency of an item in one corpus when
compared to another while taking into account the size of the corpus and
then showing if the difference in frequency of occurrence is significant or
not. Once frequency levels of the DMs were established, the items were ana-
lysed to examine how they were typically ‘chunked’ by speakers, via search-
ing for NGrams of between two and six words. Six was chosen as a ‘cut off’
point because this is generally recognised as the largest number of words in
any meaningful lexical chunk (O’Keeffe, McCarthy and Carter 2007), and
we felt that looking at some larger chunks may illuminate this analysis, even
though, as noted in chapter two, the number of chunks in evidence usu-
ally decreases greatly after four words. To conclude the analysis, the DMs
were examined qualitatively to analyse how different functions vary across
learner levels and to show how they help learners to achieve discourse com-
petence and the various CEFR ‘can do’ statements.
Table 4.2 Language items used to realise discourse competence across B1 – C1 levels
this changes so that learners refer back to ideas within the speakers’ own
turns but also across previous turns in the conversation. The extracts in
figure 4.2 show some examples of it used as a cohesive device in the ways
described.
In contrast to this, there is a notable increase as the levels increase of
both this and that but this seems to peak at B2 level, where that is used
almost four times as much as at B1 level but with a similar frequency at
C1 level. This seems to show that learners at B1 level will be familiar with
Discourse competence 119
B1
<$0> Thank you. <$1M>. What kind of music do you like to listen to?
<$1M> Er music? Er <$=> s= </$=> er sometimes I like erm er listen some pop music
because er I’m very interested in it
C1
<$2M> I personally think the same about the location the hotel is really important.
<$1F> Yeah.
<$2M> I said if it’s near to the attraction point then that’s good but if it’s far away then you
know it will be costly to travel up.
<$1F> Yeah spend more.
B1
. . . when I coming to this school erm because he she she’s my friend in China so I I just feel
happshe she’s my friend in China so I I just feel happy and er full of <$G1> about this school so
B2
<$1M> + mm and er if er what if we will have party er if tomorrow is er Sunday we can have
the party outside and er if mm they are mm Sunday I think erm most er football match and er
other sports will go on but if have a heav= heavy rain that will be stopped.
<$0> Okay.
<$1M> Or like F E.
<$0> Alright and er what do you think?
<$2F> Er erm I think er in my country I can believe this because in my country maybe it\x92s er
the weather al= always the same but in UK er I you can\x92t believe these because th= they
always change the weather sometimes rains maybe after sunshine <$=> so </$=>.
C1
<$32M> And and if if with the tourists from from other country erm ask ask us ask us some
questions such as the way how the way or some place then we have to we have to speak English
+
<$0> Mhm.
<$32M> + and not Japanese.
<$0> And what do you think <$33M>?
these items but less confident in using them to link ideas across turns, in
a similar way as we saw with it. As previously shown with the example
of it at B1 level, learners tend to use this and that within their own turn
while at B2 and C1, the items are used across a learner’s own turns and
those of their partner. The extracts in figure 4.3 demonstrate this usage at
each level.
120 Discourse competence
These uses show the increased control speakers have over cohesive devices,
as suggested by the CEFR ‘can do statements’. At C1, for example, it is sug-
gested that learners ‘Can use a variety of linking words efficiently to mark
clearly the relationships between ideas’ (Council of Europe 2001: 124/125).
However, it is clear that what is more important is not that the range of
items used to achieve cohesion increases but rather that the frequency and
flexibility of use does, as the study reviewed by Iwashita and Vasquez (2015)
in section 4.3 also found.
The data also shows that there is very little use of spatial deixis at any of
the levels and also in the native-speaker data. This may be because deixis is
normally used to discuss things within immediate time and space and tends to
occur in types of speech such as ‘language in action’ (Carter and McCarthy
2006) where speakers talk about something as they are undertaking a task.
The types of topics discussed in the speaking tests do not develop this type
of talk and as a result, deixis is relatively uncommon.
The items used to realise thematic development had a similar frequency
across the levels, although there is a slight increase in the items used to realise
this from B2 level onwards. Students at all levels tend to develop themes in a
simple additive fashion, much in the same way as native speakers do (Willis
2003), and there is little evidence of the use of spoken forms such as ‘more-
over’ or ‘in addition’ being used in the learner data. A check on the native-
speaker data from our corpus also reveals similar frequency patterns as B2 and
C1 levels, and table 4.3 shows that the differences were not significant. This
suggests that learners from B1-C1 use these items to achieve the sort of the-
matic development as described in the B2 level statement ‘Can develop a clear
description or narrative, expanding and supporting his/her main points with
relevant supporting detail’ (Council of Europe 2001: 124/125). Figure 4.4
gives an example of this usage at B2 level and an example from the NS data.
Table 4.3 Log-likelihood scores for key linguistic items used to realise discourse competence
B2
<$17F> Yes <$E <$0> laughs /$E> yeah it is changing too fast now <$=> the er </$=> for
example the iphone er it takes long time <$E <$0> laughs /$E> yeah <$E laughs /$E> it takes
a long time to be produced now there iphone 1 2 3 and 5 maybe it just changes +
NS
<6F> Erm. I guess one that’s sort of <$=> didn’t doesn’t </$=> doesn’t really go anywhere or I
don’t feel like I’m like utilising my skills so <$=> not to put </$=> not to put like the job down
but for example if if I was to work at say McDonald’s or Weatherspoons for the rest of my life +
B2
<$4M> Erm okay I think that erm if I’m going on a holiday the first thing that I will see is the
location of the hotel. I prefer the hotel to be in a down town so that I can go like for shopping or
go <$=> to </$=> to run some errands as easily as I could. What do you think?
What is also striking is that learners from B2 level onwards are better
able to initiate conversation as we can see in the increased use of what do
you think? This suggests that as learners progress, they have to focus less
on how to manage their own turn and can think more about the interactive
nature of discourse. The fact that the native-speaker data has no examples
of this suggests that there is less need to explicitly invite a conversational
partner into the discourse because they will be aware of a ‘gap’ and can find
it without the need for a signal of this nature. The sample in figure 4.5 shows
this usage at C1 level.
The use of so yeah to round off a turn and to signal the learners are ready
to close it follows a different pattern to what do you think? with an increase
in frequency of use at B1 and C1 levels but as table 4.3 shows, there is no
significant difference in this the frequency of usage. A typical use of so yeah
is shown in figure 4.6 at B1 level.
The findings described so far are confirmed by the log-likelihood scores
in table 4.3 (p. 120), which compare the scores for the most frequent items
above in order to show if increased usage was significant when comparing
122 Discourse competence
<$35M> + er in the time so yeah you know so IELTS is written in English er <$O69> so
</$O69> you know we are just er er not we are not not native speakers +
<$34F> <$O69> yeah </$O69> yeah.
<$35M> + so yeah <$O70> yeah </$O70> yeah it is very difficult to learn +
the learner data across levels and with the native speaker data. These results
show that calculations were undertaken to compare B1-B2, B1C1, B2-C1
and the learner data with the native speaker data. A plus symbol (+) means
the item was used significantly more in the first corpus compared to the
second and a minus symbol (-) means it is used significantly more in the
second corpus than the first. Stars indicate the level of significance, if this
has been found (please refer to table 4.3).
In summary this analysis shows the following:
1 It, this and that were generally used for anaphoric referencing with more
frequency and confidence and flexibility as levels increased, to link ideas
within turns at B1 and then across turns from B2 levels onwards. This
difference in frequency of use was significant particularly in the case of
it, with the frequency of this being more significant at B2 levels. Both
it and that were used significantly less by all learners than the native
speakers, while this was used significantly more.
2 Learners from B2 level onwards were better able to initiate conversation
via the use what do you think? to invite a contribution from their part-
ner. From B2 level onwards, learners used this with greater frequency
than native speakers and this increased frequency was significant. Com-
parisons also show that B2 and C1 levels used this with significantly
greater frequency than B1 level learners (see section 3.5.4). This suggests
that, as mentioned, after B1 level, learners begin to focus less on how to
manage their own turn and can think more about the interactive native
of discourse and that there is a need for learners to be more explicit
when inviting a conversation partner’s turn.
3 Thematic development was realised at all levels (and by native speak-
ers) by use of simple additive expressions such as and also or I also
throughout the levels. The use of for example to explicitly signal that
a turn is going to be extended is significantly more frequent at B2 level
when compared with B1 and C1 levels but not when compared with the
native-speaker data.
4 So yeah is used to round off turns more frequently by both B1 and C1
levels and when compared to native speakers but none of these differ-
ences in frequency are significant.
Discourse competence 123
Having looked at these results, the next section will attempt a similar style of
analysis: examining discourse competence at the micro level, via an analysis
of DMs while exploring how this relates to CEFR competences at the macro
level of discourse.
to compare B1-B2, B1-C1, B2-C1 and B1-NS, B2-Ns and C1-NS. A plus
symbol (+) means the DM was used significantly more in the first corpus
compared to the second and a minus symbol (-) means the opposite. Stars
indicate the level of significance, if this has been found.
Based on the raw frequencies and this data, we can make several further
observations.
Firstly, as noted here and in chapter 2, the raw frequency with which Er
is used decreases as level improves but is still in use at C1 level. As we have
discussed in chapter 1, this is likely to be because it allows for a speaker to
control their own turn and hold the floor. It also buys time and at lower
levels, there is a greater need to do this as you search for language. This is
supported by the log-likelihood scores, which show that this use or Er and
also Ok have a significantly higher frequency at all levels when compared
to the native-speaker data and that they also occur with significantly more
frequency at B1 and B2 when compared with the C1 levels. We would sug-
gest that this is also because Ok also has a key function of buying time and
this is most needed most where learners have less ability to quickly recall
language. The significantly increased use of you know at B1 level when com-
pared to B2 and C1 levels suggests that it is primarily used as a marker to
manage the discourse and to hold the floor, something an examination of the
data demonstrated clearly. There was little evidence of this DM being used
to mark shared knowledge, which Fung and Carter (2007) suggest is a key
function it often fulfils and instead it was part of the greater need of B1 level
learners to manage their turn by buying time. This reinforces the findings of
Fung and Carter (2007), who suggested that learners made less use of the
interpersonal functions of DMs such as you know.
4.6.1 And er
At the most obvious level, this chunk is used by speakers to aid the cohe-
sion and coherence by allowing a speaker to continue a turn by adding new
information to old. It is another way in which successful speakers achieve
thematic development, as for example also does, as discussed in section 4.5.
This function can be seen in particular at B1 and B2 levels and is used in a
Table 4.6 Two to six word chunks with er, yeah, I think, Ok, you know, well and I mean
simple additive fashion, within and across turns. When speakers progress to
C1 level, there is also evidence of this but in addition, at this level, speakers
also use the chunk to hold the floor within longer, more extended stretches
of talk and across turns, as the examples in figure 4.7 show.
4.6.2 Yeah/yeah/yeah I
In addition to use of so yeah to close conversations at B1 and C1 levels, as
noted in section 4.5, this data shows how the DM changes function across
levels. At B1, it seems to be used most often to agree, pause or buy time,
(yeah yeah) and at B2 to agree and to add a viewpoint (yeah I) and buy
time. These functions continue at C1 but yeah I also begins to function
as a more interactive marker to agree and also continue a topic, even if it
was begun by a different speaker. This usage begins at B2 level (whereas at
B1 it tends to be used to manage within a speaker’s own turn), but there
are more examples of this at C1. The samples in figure 4.8 show these
functions.
B1
<$29F> Oh okay so well actually er I need describe a game er so the game must be the the same
like the sport or any games?
<$0> Any game. Any game.
<$29F> Ah okay so erm I I think er erm in <$G1> games I just enjoy play the computer game
because I don’t like sport. Yeah so er you know when I play the indoor the online games I can
stay in my rooms don’t need go out and er and sometimes I can lie in the bed play this game with
my friend.
B2
<$16F> Yeah erm I I like you I prefer outdoor outdoor sports <$O38> just like </$O38> er I
like play badminton and er maybe sometimes I think just lie +
<$15F> <$O38> Yeah </$O38> Yeah.
<$16F> + on the grass is very comfortable.
<$15F> Yes I think erm you do some sports and some sports is on just like football and erm er
basketball and er you can also erm play play football er or basketball inside or outside.
C1 level
<$0> Okay. Thank you. Is there a particular place that you would like to visit?
<$20F> Er I went to Bristol. I like that city +
<$0> Mhm.
<$20F> + and er Lancashire er and er Liverpool er Manchester. I hope I will go to other cities.
B1
Agreeing
<$1M> If the language is a problem maybe we can talk talk with each other.
<$2M> Yeah. I’m very agree with your opinion and er in addition I would like to say
Buying time
<$20M> Okay. Er <$E> coughs </$E> er when I was young I used to play er football everyday
but er now it’s because we= I have a lot of things to do it I er I always busy and <$G5> and no
more time I er I I didn’t have any time to play football but I like to watch it and er er yeah
football yeah I like to watch it.
B2
<$2F> Okay. Yeah erm but er different er like er our break is very warm but is really hot so
people er always er always re= can’t erm go out and er to en= to enjoy they job I think.
<$1M> Yeah.
<$2F> And er +
<$1M> Mm yeah I+
<$1F> The hotel amenities before I stay there has to be good like they has to be very good
before I stay in a hotel because I can’t just stay in normal hotel that just have run down stuff +
<$2M> Yeah.
<$1F> + it has to be really good.
<$2M> Yeah I do the same and most importantly first is the hotel room right?
<$1F> Mhm.
<$2M> Yeah the place where you would spend the night over yeah how important do you think
it is?
<$1F> Yeah.
<$2M> Yeah it’s very important yeah.
B1
<$0> Okay. Would you begin please?
<$4F> Er I will begin first <$E> coughs </$E> Er yes I think er individual personality will
affect the choice of work job because sometimes if the person is confidence maybe they will do
some own <$G1> not follow some other people’s doing some job applications.
B2
<$7M> + so what do you think?
<$8M> Yeah I definitely agree with you because I think er maybe sometimes it’s depend
depends on different average er different <$G2> of application maybe maybe some developed
countries er the average is very high so people maybe have a quite good of er knowledge so they
can just communicate very easily <$=> and er mm I er </$=> besides I think mm +
C1
<$0> <$29M> sorry. When is the best age to learn to drive?
<$29M> To learn to drive I think it must be sixteen or seventeen but to start driving itshould be
eighteen because if people just learn to drive in one month or in one session they won’t learn
everything so <$=> they start </$=> two years ago they start having classes some safety some
safety classes about the classes about the cars <$=> can </$=> what they are capable of what
they can do what they can’t do they must tell them everything before they can use the car.
4.6.4 Ok er/ ok I
The function of this DM is used to buy time, to show understanding, and to
launch a turn or topic. The first two functions are most common at B1 and
B2 and the final function becomes increasingly used at C1 level. Figure 4.10
shows these uses.
4.6.5 Well I
As noted previously, the use of well increases as levels increase and its use
is significantly more frequent at B2 and C1 levels when compared with B1
level. If we compare B1 to C1, this difference is highly significant. At B2
level, the most common use is to signal some form of hesitation at the start
of the turn. At C1 level, this use is also in evidence but there is also some
evidence that it is used to mark a slightly unexpected answer, which is often
listed as a common function of this DM (see for example Carter and McCar-
thy 2006) The examples in figure 4.11 show these uses at C1 level.
Overall, what these samples show is that DMs vary to some degree in fre-
quency across levels and this is also shown in the log-likelihood scores, but
more importantly they vary in terms of how they function across levels. As
with the cohesive devices explored in the earlier part of this chapter, there is
a clear tendency for successful speakers to use DMs to manage the discourse
within their own turn at B1 level. While these functions continue at higher
130 Discourse competence
B2
Showing understanding
<$0> Okay would you begin please?
<$8M> Ok er <$7M> let’s talk about something about er the companies’ team working. So what
do you think about it if we just refer to these two words?
<$7M> The training er +
C1
Launching a turn
<$4M> how important is tourism to your country?
<$3M> Ok can I go first? Ok erm tourism in my country is not really important why because
erm my country has a lot of problems like erm there are a lot of things that er need to be atte=
attended to before tourism so basically what they are focussing on is not tourism at all they are
trying to focus on agriculture and the production the industries and the rest so like tourism is
like neglected in my country
Marking hesitation
<$22F> Okay what do you think?
<$21M> Er well these points important points in tourism but er I think well I agree with this
paper and the first point sites of historical interest especially erm in my country there is many
many sites historical site
Unexpected response
<$0> Film or a book?
<$27M> Well I don’t read books.
levels, DMs tend to become used more across turns from B2 and in particu-
lar at C1 levels. In terms of the CER descriptors, the B1 descriptor suggests
that regarding flexibility, learners ‘Can exploit a wide range of simple lan-
guage flexibly to express much of what he/she wants’ while at B2 and C1
levels learners ‘Can adjust to the changes of direction, style and emphasis
normally found in conversation’ (Council of Europe 2001: 124/125). This
Discourse competence 131
data suggests that learners use DMs more flexibly as levels increase to adjust
to the ongoing needs of the discourse and of their conversational partner.
It would seem then that successful discourse competence at lower lev-
els is reflected in the ability to manage and extend your own turn, and as
learners progress this becomes increasingly part of what McCarthy (2009)
has termed ‘confluence’ in relation to fluency, that is co-constructing the
discourse across turns. It is also interesting that in some cases the learners
used more DMs than was evident in the NS data, although, as noted, these
differences in frequency were not always significant. We can suggest that this
is likely to be because successful learners have a greater need to explicitly
signal as a way to manage the discourse in spoken test conversations such
as the ones which make up this data. This is likely to be because the topics
are not of their choosing and more time is thus taken processing, managing
and responding than it was by native speakers.
4.7 Conclusion
The findings in this chapter lead us to several conclusions related to the
language learners use at different levels to achieve discourse competence and
have clear implications for teaching and learning.
The first thing we can note is that speakers generally use similar language
to produce discourse that is coherent, cohesive, developed and managed but
that frequency of occurrences does differ across levels in significant ways as
do the specific functions for which such language is used. This was demon-
strated through the analysis in the first part of the chapter. We can see that
in terms of cohesion, for example, that when learners refer back to some-
thing said previously they used it more frequently as the levels increase but
for this and that the frequency of use reached a peak at B2. This increase in
use of it was significant in terms of log-likelihood scores when we compare
B2 and C1 levels with B1 level. However, the usage of this was significantly
more frequent at B2 level in comparison to B1 and C1, while that was sig-
nificantly more frequent at both B2 and C1 levels when compared with B1.
Such differences reflect the fact that successful discourse competence at B1
level generally reflected the ability of learners to manage their own turn, and
as learners progress this becomes increasingly part of co-constructing the
discourse across turns and developing what a speaking partner says.
Secondly, the chapter shows that certain DMs are an important feature of
discourse competence at all levels. Some items were used with significantly
higher frequency by learners at all levels when compared with the NS data,
namely er and ok. In the case of I mean, the usage was significantly higher
in the NS data. The frequency of items also varied across the levels and these
variations can be significant alongside differences in how they function to
give two examples, you know was used with significantly higher frequency
at B1 level, when we compare this to B2 and C1. At B1, it is used to mark
132 Discourse competence
hesitation or buy time rather than to mark shared knowledge. Well was used
with significantly higher frequency at C1 levels when compared to B1 and
B2 and the functions also increase. At B2 level, well is primarily used to mark
hesitation and to buy time while at C1 is also used to mark an unexpected
response. As a whole, the data also supports the findings of Gilquin (2008),
that hesitation was more commonly marked with er before any other DM,
even though usage decreases as levels increase. The low frequency of I mean
in the USTC data suggest that learners have less focus on explicitly signalling
a reformulation in their output and may be more focused on the message
they are producing. DMs also formed chunks such as so yeah which fulfil
important functions across the levels. In the case of so yeah, it was used most
frequently at B1 and C1 levels to close turns.
Overall, this chapter shows that discourse competence at B1 level is devel-
oped largely within a speaker’s own turn where the learner focuses on creat-
ing coherence and cohesion mainly within their own contribution. At B2 and
C1 levels, this starts to expand across turns and learners can co-construct
turns and increasingly develop what a speaking partner says by linking their
contribution to what has been said.
References
Adolphs, S. and Carter, R. 2013. Spoken corpus linguistics: From monomodal to
multimodal. London: Routledge.
Aijmer, K. 2002. English discourse particles: Evidence from a corpus. Amsterdam:
John Benjamins.
Byrne, S. 2015. Examining successful language use at C1 level: A learner corpus
study into the vocabulary and abilities demonstrated by successful speaking exam
candidates. Journal of Linguistics and Language Teaching, 6(1), 67–88.
Canale, M. and Swain, M. 1980. Theoretical bases of communicative approaches to
second language teaching and testing. Applied Linguistics, 1(1), 1–47.
Carter, R. and McCarthy, M. 2006. Cambridge grammar of English: A comprehen-
sive guide: Spoken and written English grammar and usage. Cambridge: Cam-
bridge University Press.
Celce-Murcia, M., Dornyei, Z. and Thurell, S. 1995. Communicative competence:
A pedagogically motivated model with content specifications. Issues in Applied
Linguistics, 6(2), 5–35.
Chen, Y-H. and Baker, P. 2016. Investigating criterial discourse features across sec-
ond language development: Lexical bundles in rated learner essays, CEFR B1, B2
and C1. Applied Linguistics, 37(6), 849–880.
Cook, G. 1989. Discourse. 6th ed. Oxford: Oxford University Press.
Council of Europe 2001. Common European framework of reference for languages:
Learning, teaching, assessment. Cambridge: Cambridge University Press.
Fung, L. and Carter, R. 2007. Discourse markers and spoken English: Native and
learner use in pedagogic settings. Applied Linguistics, 28(3), 410–439.
Gilquin, G. 2008. Hesitation markers among EFL learners: pragmatic deficiency or
difference? In J. Romero-Trillo ed. Pragmatics and corpus linguistics: A mutualistic
entente. Berlin: Mouton de Guyter, 119–143.
Discourse competence 133
Halliday, M.A.K. and Hasan, R. 1976. Cohesion in English. 2nd ed. London:
Longman.
Hellermann, J. and Vergun, A. 2007. Language which is not taught: The discourse
marker use of beginning adults learners of English. Journal of Pragmatics, 39(1),
157–179.
Hoey, M. 2004. Spoken discourse: Discourse markers er, erm and OK.MED Maga-
zine [Online] Available from: <www.macmillandictionaries.com/MED-Magazine/
March2004/17-language-awareness-discourse-uk.htm> [Accessed 12 February
2017].
Iwashita, N. and Vasquez, C. 2015. An examination of discourse competence at dif-
ferent proficiency levels in IELTS speaking part 2. IELTS research reports, 5, 1–44.
Jones, C. 2011. Spoken discourse markers and English language teaching: practices
and pedagogies. PhD. Thesis. Nottingham: University of Nottingham.
Jones, C. and Waller, D. 2015. Corpus linguistics for grammar: A guide for research.
London: Routledge.
Jucker, A.H. and Ziv Y. 1998. Discourse markers: Descriptions and theory. Amster-
dam: John Benjamins.
McCarthy, M. 1991. Discourse analysis for language teachers. 2nd ed. Cambridge:
Cambridge University Press.
McCarthy, M. 2009. Rethinking spoken fluency. Estudios de linguistica inglesa apli-
cada, 9, 11–29.
McCarthy, M. and Carter, R. 1993. Language as discourse: Perspectives for language
teaching. New York: Longman.
Oakes, M.P. 1998. Statistics for corpus linguistics. Edinburgh: Edinburgh University
Press.
O’Keeffe, A., McCarthy, M. and Carter, R. 2007. From corpus to classroom: Lan-
guage use and language teaching. Cambridge: Cambridge University Press.
Reder, G. 1990. Ideational and pragmatic markers of discourse structure. Journal of
Pragmatics, 14, 367 –381.
Schiffrin, D. 1987. Discourse markers. Cambridge: Cambridge University Press.
Thornbury, S. 2005. Beyond the sentence: Introducing discourse analysis (methodol-
ogy). 3rd ed. Oxford: Macmillan Education.
Tsai, P-S and Chu, W-H. 2015. The use of discourse markers among Mandarin Chi-
nese teachers, and Chinese as a second language and Chinese as a foreign language
learners. Applied Linguistics. Advanced online access [Online]. Available from:
<https://academic.oup.com> [13th March 2017].
Waller, D. 2015. Investigation into the features of written discourse at levels B2 and
C1 of the CEFR. Unpublished PhD dissertation, University of Bedfordshire.
Willis, D. 2003. Rules, patterns and words: Grammar and lexis in English language
teaching. Cambridge: Cambridge University Press.
Chapter 5
Pragmatic competence
5.1 Introduction
This chapter will focus on the speech acts of requests and apologies with
the aim of establishing the prerequisite components for their successful pro-
duction. As outlined in chapter 1, the data used to analyse the pragmatic
features of successful spoken request and apology language were based upon
a 39,437-word corpus of international learner responses. As Geluykens and
Kraft (2008) highlight, in the absence of suitable reference corpora which
often do not lend themselves well to cross-cultural analyses, or are too gen-
eral to be useful for pragmatics research in particular, researchers may be
left with little option but to devise their own. It is for these reasons that the
Speech Act Corpus of English (SPACE) originally took shape, though it is
limited to the speech acts of request and apology in its current form, at the
time of writing this book.
The majority of the NNS and NS data for the SPACE corpus were gath-
ered over a series of intervention studies (Halenko 2016; Halenko and Jones
forthcoming) and captured via the use of innovative and interactive virtual
role plays, designed to elicit large amounts of data in a controlled envi-
ronment, in an attempt to simulate authentic NS-NNS interactions more
closely (see chapter 1 for a more detailed description of this instrument).
As discussed in chapter 1, simulated spoken data cannot entirely replicate
naturally-occurring discourse but is a more efficient means of capturing
large samples of language with a specific focus, in controlled environments.
As a reminder, the test featured interlocutors from an academic context who
international students were likely to encounter on campus at a British uni-
versity, playing high and low social distance roles, within scenarios typical
of staff-student exchanges such as high imposition requests and apologies
(see figure 5.1). The SPACE corpus comprised those request and apology
responses considered successful from B2 level learners, which were rated as
being ‘appropriate’ or ‘very appropriate’ for the scenarios presented.
Without revisiting communicative models from chapter 1 in too much
depth, section 5.2 first outlines features which underpin the notion of prag-
matic competence, its importance as a component to successful language
Pragmatic competence 135
You:
Figure 5.1 An example of a scenario from the CAPT to elicit spoken data
production, and how the CEFR embeds the notion of the sociolinguistic
competence in its descriptors. Given the central focus of requests and apol-
ogies in the SPACE corpus, these speech acts are also defined, providing
frameworks for analysing the data more closely, in addition to an acknowl-
edgement of the substantial role formulaic language plays in production
of pragmatic language, in particular. Following a review of previous stud-
ies and how the data were analysed in sections 5.3 and 5.4, the chapter
turns to examining the linguistic aspects of successful request (section 5.5)
and apology (section 5.6) language and cross-referencing the findings to the
CEFR descriptors. Section 5.7 concludes the chapter by identifying the prag-
matic strategies and organisational patterns of moves considered requisite
for structuring successful requests and apologies. As a point of comparison
of alternate ways success can be achieved, native-speaker data within the
SPACE corpus will also be examined, as in previous chapters. The aim of this
chapter is to be able to highlight components of successful B2 level request
and apology language to encourage prioritising these aspects in teaching and
learning programmes.
listener obliges and complies with the request. The success of utilising IFIDs
appropriately, however, is less commonly achieved by NNS (Barron 2003),
the reasons for which have been one of the motivating drives for second
language pragmatic investigations. Building on the descriptions in chapter 1,
Crystal (1997: 301) provides the most widely referenced definition of sec-
ond language pragmatics:
Pragmatics is the study of language from the point of view of the users,
especially of the choices they make, the constraints they encounter in using
language in social interaction, and the effects their use of language has on
other participants in the act of communication.
These two areas of choice and constraint are conveniently differentiated
by Leech (1983) and Thomas (1983) as pragmalinguistic and socioprag-
matic components of pragmatics. Pragmalinguistics refers to the knowledge
of linguistic resources available and the choices made to convey messages.
Sociopragmatics, on the other hand, is primarily concerned with the effect
constraints such as social distance and status will have when realising a
communicative act.
As noted in chapter 1, pragmatic competence as referred to in the CEFR,
aligns to Canale’s discourse competence, representing ‘the functional use of
linguistic resources . . . the mastery of discourse, cohesion and coherence’
(CEFR 2001: 13). With the descriptions of pragmatic competence outlined
in 5.1 in mind, which place an emphasis on the influence of social and cul-
tural conventions on language use, it seems a much closer fit to the CEFR’s
notions of sociolinguistic competence: ‘sociolinguistic competence is con-
cerned with the knowledge and skills required to deal with the social dimen-
sion of language use . . . [including appropriate use of] linguistic markers
of social relations, politeness conventions and differences in register at the
different levels’ (CEFR 2001: 118). The CEFR descriptors are captured in
table 5.1 and suggest B2 learners (whose responses comprise the SPACE cor-
pus) should have a sociopragmatic awareness of politeness and behaviour
relative to the context and interlocutor, and be able to make appropriate
(pragma)linguistic choices based on this.
5.2.1 Requests
Derived from Searle’s (1976) classification of ‘directives’, requests are seen
as illocutionary acts in which a speaker conveys to a hearer his/her wish
for the hearer to perform an act which is of cost to them but has benefit
to the speaker. This can be a request for verbal goods, such as information,
or non-verbal goods, such as an object or service (Trosborg 1995). It is
characterised as a pre-event act given the expectation that the act will take
place in the immediate or near future time. As the request imposes on the
hearer, it is also, by nature, a face-threatening act (FTA). Within Brown and
Levinson’s (1987) theory of politeness, a request specifically threatens the
138 Pragmatic competence
Sociolinguistic appropriateness
Direct strategies
• direct
e.g., Give me a lift
Non-conventionally
indirect strategies • least direct
e.g., I’m late for the train
or ‘hints’, are employed when the speaker does not wish to overtly state the
desired action but instead prefers to make a statement or ask a question. It
requires the hearer’s interpretation of the speaker’s intent, e.g. I’m late for
the train ought to signal to the hearer that he/she might offer the speaker a
lift to the train station.
The head act is able to function independently but is typically embedded
within a range of mitigating supportive moves which serve to soften the
request. These comprise internal and external modifiers. Internal modifiers
are those which form part of the head act itself and include softeners which
reduce the force, e.g. Could you possibly . . ., items used to fill in the gaps
of the utterance, e.g., Could you, erm, possibly . . . or alerters which serve
to gain the interlocutor’s attention, e.g. Excuse me . . . or the token please.
In contrast, external modifiers surround the head act, serving to further
absorb the impact of the impending imposition. These include Preparators,
employed to set up the request, e.g. Mr Waters, I’ve got a question about my
assignment, and Grounders, devices used to provide a reason or explanation
for the request, e.g. Could I have an extension? I’ve had computer problems.
Observations about the context and social environment need to be made
before deciding on the appropriate construction of the request itself.
5.2.2 Apologies
As with requests, apologies are considered face-threatening acts (FTAs).
To repair the damage of FTAs, interlocutors may engage in a number
of facework strategies to ‘redress’ the incident which include apologies.
Olshtain and Cohen (1983: 20) maintain that when an action, utterance
140 Pragmatic competence
(or lack thereof) causes offence on the part of the recipient(s), the culpable
person(s) needs to apologise to re-establish social harmony. In other words,
‘[an] apology is an instance of socially-sanctioned hearer-supportive behav-
iour’ (Edmondson 1981: 273) and defined as a post-act event (Blum-Kulka
and Olshtain 1984).
The conditions to the apology being fulfilled, however, are dependent on
the culpable person acknowledging or recognising the offence has occurred,
which may be determined by sociocultural norms just as linguistic norms will
determine whether the utterance actually qualifies as an apology (Olshtain
and Cohen 1983: 20). In Brown and Levinson’s (1987) terms, the act of
apologising is face-saving for the hearer and face-threatening for the speaker.
Leech (1983) qualifies this by maintaining there is some kind of benefit
for the hearer at a cost for the speaker through the act of apology, unlike
requests which are costly in the reverse.
Researchers have posited a number of general and more detailed classifi-
cations for the semantic formulae contained in acts of apology. Most build
on the influential work of Goffman (1971) who describes apologising as
‘remedial work’ accomplished by accounts (excuses/explanations), requests
(begging sufferance) and apologies. Goffman classifies apologies as either
‘ritual’, motivated by social habits, or ‘substantive’- the wish to repair any
damage or harm caused by the initial act.
The limited categorisations proposed by Goffman have since been modi-
fied and expanded by a number of scholars based on cumulative research
conducted in the 1980s (Olshtain and Cohen 1983; Owen 1983; Blum-
Kulka and Olshtain 1984; Trosborg 1987). As a result, these studies have
developed and described a range of strategies to be undertaken for appro-
priate apology behaviour. Observational and interventional investigations
often cite the seminal work of Blum-Kulka, House and Kasper (1989) Cross-
Cultural Speech Act Realisation Project (CCSARP) for a basic conceptual
framework of the semantic formulae involved in apologising, though this
is largely a reformulation of those proposed in the earlier studies (Blum-
Kulka and Olshtain 1984; Olshtain and Cohen 1983; Owen 1983; Trosborg
1987). It consists of a set of five formulae which individually may be con-
sidered sufficient to placate the hearer, although a combination, signifying
a more intense apology, is also commonplace (see figure 5.3). The apology
may also be accompanied by a strategy to signal intensification I’m very
sorry to amplify the speaker’s regret (Dalmau and Gotor 2007: 1). It is use-
ful to view them on a continuum as Trosborg (1987) suggests. In this case,
the cumulative total of formulae a) to e) in figure 5.3 increase in indirectness
and potential for placating the recipient.
Explicit expressions of apology a) are generally realised through some kind
of performative verb such as ‘apologise’ or ‘forgive’. An explanation b) provides
a reason for the violation or damage which has occurred and often provides
supportive evidence to a). An admission of responsibility for the offence is
realised through strategy c) which Nureddeen (2008: 290) suggests is the
Pragmatic competence 141
Figure 5.3 Formulaic strategies for the apology speech act (based on Trosborg 1987)
‘most explicit, most direct and strongest apology strategy’. An offer to repair
or pay for the damage caused is provided through d), whilst promising not to
repeat the offence in the future is acknowledged in strategy e).
Strategies a) and b) are said to be the basis of any remedial work, whilst c)
to e) are situation-dependent (Blum-Kulka, House and Kasper 1989) in the
event further mitigation is required. In contrast, Bergman and Kasper’s
(1993) review of a range of empirical apology studies have since concluded
the essential components of an apology contained strategies a) and c);
explicit expressions of apology and statements of responsibility (Holmes
1990; House 1988; Kasper 1989; Trosborg 1987) and it is the severity of the
infraction which dictates the redressive strategy preferences (e.g. Brown and
Levinson 1987; Holmes 1990; Maeshiba et al., 1996). A single apology in
the form of an IFID may be adequate for being slightly late to meet a friend
(ritual apology), but a more elaborate formulation incorporating multiple
strategies may be required if the offence is much more serious such as break-
ing a person’s treasured possession (substantive apology). Contextual fac-
tors such as power and social distance between interlocutors also influence
apology performance (Maeshiba et al., 1996).
As many of the ways to realise both requests and apologies comprise the
use of formulaic chunks, it is worth visiting this area further in section 5.2.3
with specific reference to pragmatic language.
that formulaic language plays a central role in effective and efficient commu-
nication (Pawley and Syder 1983; Schmitt 2004; Sinclair 1991; Wray 2008),
as described in chapter 2. For L2 learners in particular, formulaic language
can save mental capacity which can be used more effectively to internalise
syntactic rules (Wang 2011), relieves pressure on memory which may benefit
L2 acquisition (Weinert 1995) and is said to improve fluency (Fillmore 1979).
For these reasons, language learning from a formulae-based approach can be
an effective learning strategy. Interlocutors also easily recognise formulaic
language and, in fact, have an expectation that conventionalised sequences
are used in order to expedite effective communication in many formal and
informal situations. Kecskes (2014) notes that much formulaic language
which occurs in social situations is culture-specific, so an understanding of
the social norms of a speech community is dependent on successful use and
interpretation. Much conventional formulaic language can be found in real-
ising speech acts such as requests and apologies (Schmitt and Carter 2004),
but as Wray (2012: 236) notes, ‘instructed L2 learners have an impoverished
stock of formulaic expressions’. Bardovi-Harlig (2009) and Kecskes (2000)
suggest lack of familiarity with expressions and sociopragmatic knowledge,
overuse of familiar expressions, level of proficiency and inadequate L2 expo-
sure as factors for this lack of resource.
Schauer and Adolphs (2006) and Jones and Halenko (2014). Collectively,
these studies demonstrate how corpora may be used successfully for work-
place training and language teaching purposes.
Cheng’s (2004) study analysed a six-hour sub-corpus of service encounter
exchanges at a hotel reception desk from the 500,000 Hong Kong Cor-
pus of Spoken English (HKCSE). Her quantitative and qualitative analysis
included identifying a need for pragmatics training of frontline hotel staff
who failed to meet the sociopragmatic and pragmalinguistic expectations of
their role and the hotel’s high customer care ethos. Specifically, the discourse
highlighted staff members’ preoccupation with settling the bill during the
speech event of ‘checking out’, at the expense of expected moves which show
sufficient concern and interest in the guests’ overall satisfaction. This was
exacerbated by the infrequent use of honorifics, hedging and the marker
‘please’ whilst guiding guests through the check-out process.
Schauer and Adolphs (2006) focussed on formulaic sequences in expres-
sions of gratitude elicited from written production tasks compared to the
multi-million word Cambridge and Nottingham Corpus of Discourse in
English (CANCODE). Findings revealed the former successfully provided a
wide range of interactional formulaic sequences and contemporary expres-
sions such as that’s wicked which could be used as classroom resources for
learners to gain insights into current language practices. In addition, the
CANCODE data was able to provide examples of how false starts and hesi-
tations keep conversations moving, and revealed expressions of gratitude
often comprised several phrases rather than a single expression e.g. ‘Great
stuff. That’s brilliant. Thank you very much. That’s much appreciated’. The
authors advised combining both data sets for maximum teaching benefits.
Jones and Halenko’s (2014) exploratory study also aimed to encourage
the use of corpora in classroom teaching by showing how the freely available
corpus tool Lextutor (2013) could be used to examine learner language, in
this case Chinese EAP learners’ successful request language, as a founda-
tion for teaching materials. The authors’ analyses concluded that conven-
tional indirectness comprising high frequency modals can, could and would,
and formulaic expressions to encourage listenership (e.g. sorry to bother
you, excuse me) and convey politeness/indirectness (e.g. would you mind,
could you help me) were key features of the 3,919 word spoken sample.
These findings were further confirmed in a data comparison with the British
National Corpus (BNC). In terms of strategy choice, greater social distance
and imposition tended to produce the organisational pattern: apology (for
imposition) + grounder + request or request + grounder e.g. I’m sorry to
bother you, because I have personal reason, could I ask for extension for my
assignment? For interactions with smaller social distance, learners mainly
opted for preparator (e.g. compliment) + request e.g. You are good at com-
puter skills; can you design our group’s presentation?
144 Pragmatic competence
Table 5.2 The twenty most frequent words in the learner and native speaker request data
The data indicate that both learners and native speakers achieve success
through the use of the modal verbs can (NNS) and could (NNS and NS)
which express conventional indirectness in the request head act. The learner
data suggests both modals are used interchangeably though can is the pre-
ferred choice (see figure 5.4). Want also features highly in the ranking but,
on closer inspection, tends to be isolated to transactional exchanges where
services are being provided (library, accommodation office) so increased
directness is less likely to affect the outcome of the request, and the direct
head act is almost always mitigated by supportive external modification
devices such as alerters or internally modified by please. In contrast, NS use
both can and want infrequently. Typical of native-speakers’ requests but not
evident in the learner data are the use of please as a closing and really as an
emphatic marker to boost the urgency of the request, show appreciation,
or as an intensifier to an apology, mitigating the imposition of the request.
<R01S12Ch> I want to find out how to book a study room please. Can you help me? (NNS)
<R01S09Na> Hi. Could you help me book a study room please? (NS)
Rank Four-word chunks (freq.) Three-word chunks (freq.) Two-word chunks (freq.)
Rank Four-word chunks (freq.) Three-word chunks (freq.) Two-word chunks (freq.)
<R05S15Ch> Hi sir. I want to study but the students are too noisy. Could you tell them please?
(NNS)
<R05S08Na> Hi. I just wanted to know if you could please go and talk to those studentsand ask
them to keep the noise down a bit. I’d really appreciate it. (NS)
discussed in section 5.5 has illustrated how this can be achieved. Section 5.6
will now consider this in light of the speech act of apology.
Table 5.5 The twenty most frequent words in the learner and native-speaker apology data
<A07S15Ch> I’m so sorry about the book. Can you tell me what to do now? (NNS)
<A08S08Na> I’m really sorry about the book. I can pay for the damage. (NS)
Figure 5.6 Use of can as a repair strategy in the NNS and NS apology data
150 Pragmatic competence
A feature of the learner data not evident in the NS data is the performative
verb, apologise, though much less frequent than its counterpart sorry (see
table 5.6). The data reveal the instances of apologise are generally limited
to high social distance-high imposition scenarios, hence their acceptability
in these situations. As identified in the request data, NS apologies tend to
incorporate the use of just as a mitigator to the main verb, e.g. just wonder-
ing if, just want to know if just been so busy, just didn’t have enough time
(see table 5.7).
Confirming initial examinations of the raw frequency lists, explicit expres-
sions of regret are common to both the NNS and NS sets but these are
realised in different ways: a combination of (so) sorry about (that) dominate
the NNS requests whilst NS opt for a combination of really (really) sorry
about (that). For both data sets this is generally followed by a description
of the context of the apology, and often concluded with direct offers to
repair the situation using I will or I can (see figure 5.7), in addition to some
requests for help to make amendments which feature in the NNS set only:
Can you help me . . . /What should I do . . .
In comparison to the examples in figure 5.7, there are fewer occurrences
of chunks containing apologise. NNS experiment with several collocations,
in the order of frequency: apologise to; I apologise for; want to/have to apol-
ogise, whereas NS most commonly rely on the sequence want to apologise
for. Figure 5.8 provides an example of several instances of say apologise to
you from the NNS data, where grammatical and syntactic errors are evident,
suggesting that inaccuracies may not always affect the success of a formu-
laic sequence, as noted in other studies (Halenko 2016; Halenko and Jones
2011; Halenko and Jones forthcoming).
<A10S07Jp> Hi I’m sorry I’ve been late to your class every day in this week. I missed the bus
and train every day because I woke up late um I’m so sorry about that. I will make effort to wake
up earlier from next time. (NNS)
<A09S12Na> Hi I’m really, really sorry forbeing late this week I’vebeen ill I’ve had a lot of
assignments erm, I’ll do better next week it’s just it’s just been a really, really tough week. (NS)
<A07S02Sa> Er hello excuse me just I want to say apologise to you about the noise because the
party with my friend er but I promise you that this last time I do it so I apologise to you and the
student.
Rank Four-word chunks (freq.) Three-word chunks (freq.) Two-word chunks (freq.)
1 SO SORRY ABOUT THAT (21) SORRY ABOUT THAT (44) SO SORRY (88)
2 SORRY ABOUT THAT I (13) I HAVE TO (27) I HAVE (74)
3 I HAD A PARTY (12) I WANT TO (26) SORRY ABOUT (62)
4 SPLIT SOME COFFEE ON (12) SO SORRY ABOUT (26) SO I (57)
5 HAD A PARTY AT (8) SO SORRY I (22) AND I (56)
6 HOW ARE YOU DOING (7) I HAD A (16) I WILL (56)
7 I FORGOT TO COLLECT (7) HAD A PARTY (15) SORRY I (52)
8 REALLY SORRY ABOUT THAT (7) HOW ARE YOU (15) ABOUT THAT (50)
9 SAY APOLOGISE TO YOU (7) ABOUT THAT I (14) I CAN (39)
10 THANK YOU SO MUCH (7) APOLOGISE TO YOU (14) I WANT (38)
11 WHAT SHOULD I DO (7) SPILT SOME COFFEE (14) REALLY SORRY (37)
12 A PARTY AT MY (6) EXCUSE ME I (13) I AM (34)
13 CAN YOU HELP ME (6) SOME COFFEE ON (13) TO YOU (34)
14 HI HOW ARE YOU (6) I AM SORRY (12) THANK YOU (33)
15 I HAVE SPILT SOME (6) IN MY FLAT (11) BUT I (31)
16 I HAVE TO APOLOGISE (6) A PARTY AT (10) EXCUSE ME (30)
17 A BOOK FROM YOU (5) COMPLETE MY ESSAY (9) WANT TO (29)
18 A PARTY IN MY (5) I APOLOGISE FOR (9) HAVE TO (28)
18 FLAT WITH MY FRIENDS (5) REALLY SORRY ABOUT (9) FIND IT (27)
20 HAVE SPILT SOME COFFEE (5) SORRY I HAVE (9) ARE YOU (26)
Table 5.7 The twenty most frequent four-, three- and two-word chunks in the native speaker apology data
Rank Four-word chunks (freq.) Three-word chunks (freq.) Two-word chunks (freq.)
1 REALLY SORRY ABOUT THAT (6) REALLY REALLY SORRY (10) REALLY SORRY (39)
2 SPLIT SOME COFFEE ON (6) REALLY SORRY ABOUT (10) AND I (30)
3 TO COLLECT A PACKAGE (6) REALLY SORRY I (10) FIND IT (20)
4 I FORGOT TO COLLECT (5) SPLIT SOME COFFEE (8) THANK YOU (20)
5 I JUST WANT TO (5) I FORGOT TO (7) THAT I (20)
6 REALLY SORRY BUT I (5) I JUST WANT (7) BUT I (19)
7 SORRY FOR BEING LATE (4) COLLECT A PACKAGE (6) I CAN (18)
8 WANT TO APOLOGISE FOR (4) SOME COFFEE ON (6) I WAS (17)
9 A PACKAGE FROM THE (3) SORRY ABOUT THAT (6) SORRY I (17)
10 ACCIDENTALLY SPILT SOME COFFEE (3) TO APOLOGISE FOR (6) HI I (13)
11 COLLECT A PACKAGE FROM (3) TO COLLECT A (6) I JUST (13)
12 FORGOT TO COLLECT A (3) WANT TO APOLOGISE (6) IT I (12)
13 HI I JUST WANT (3) BE ABLE TO (5) REALLY REALLY (12)
14 I BE ABLE TO (3) FIND IT AT (5) SORRY ABOUT (12)
15 I BORROWED A BOOK (3) FIND IT I (5) A PACKAGE (11)
16 I BORROWED FROM YOU (3) FOR BEING LATE (5) THIS WEEK (10)
17 I JUST WANTED TO (3) FORGOT TO COLLECT (5) ABOUT THAT (9)
18 I WAS WONDERING IF (3) JUST WANT TO (5) I FORGOT (9)
18 IT AT THE MOMENT (3) REALLY SORRY BUT (5) I KNOW (9)
20 REALLY REALLY SORRY ABOUT (3) SORRY BUT I (5) I WILL (9)
Pragmatic competence 153
<R06S01Ch> Excuse me. Could you help me find a new place to live? I have some problems
with my accommodation.
<R03S12Jp> I’m so sorry about that. What can I do about this?
Figure 5.9 Examples of NNS strategies and organisation in the SPACE data
154 Pragmatic competence
<R04S12Ch> I really need this book to finish my essay so could you help me to borrow it until
next week? (NNS)
<R04S12Na> Hi. Is it possible for me to borrow this book until next week because I need it for
my homework? (NS)
5.8 Conclusion
In summary, the SPACE corpus reveals that the learner and native-speaker
request and apology realisations share many common features. What is
most noteworthy is that although the learner data suggest the repertoire
of request and apology expressions and strategies are less varied, learners,
in fact, need not have a great pragmalinguistic range in order to be con-
sidered successful. Instead, they can rely on a smaller range of polite core
requests and alerters when formulating requests, and explicit apologies with
intensifiers when apologising, whose forms need not always be grammati-
cally accurate to be successful. Similarly, simply by observing basic request
and apology moves which incorporate appropriate alerters and head acts
(requests) and explicit apologies with attempts at repairing the situation
(apologies), learners are able to successfully show appropriate socioprag-
matic awareness even though disparities in organisational patterns may exist
and choice of strategy may differ from their NS counterparts. Secondly, as
outlined in chapter 1, the frequency of formulaic expressions within the
SPACE corpus lends support to research claims concerning their importance
for effective communication (Pawley and Syder 1983; Schmitt 2004; Wray
2008), and underlines how much formulaic language can be found in speech
acts such as requests and apologies (Aijmer 1996; Wang 2011). That the
learners did not always have native-like command of these formulaic expres-
sions but their responses were still considered successful based on the raters’
scores, raises two interesting points. Firstly, there does appear to be a social
expectation that conventionalised expressions are used, as proposed by Kec-
skes (2000). The influence of formulaic expressions is further underlined in
that simply attempting to employ them, though not always syntactically or
grammatically correct, was still considered positive, as reflected in the high
raters’ scores and the inclusion of the responses in the SPACE corpus. The
second point concerns the perception of pragmatic competence versus lin-
guistic competence for successful communication. As highlighted by Bardovi-
Harlig and Dornyei’s (1998) empirical investigation, ESL tutors in the host
community overwhelmingly favoured pragmatic ability over grammatical
ability, whilst the reverse was true for EFL tutors in the at-home environ-
ment. This perception of the importance of pragmatic awareness seems true
Pragmatic competence 155
for this data and the examples presented above. Thomas’s (1983: 96–97)
seminal paper, seen as a driver of pragmatic studies, provides an emphatic
distinction between the consequences of pragmatic and grammatical errors
and underlines the importance of pragmatic awareness for successful spo-
ken English: ‘while grammatical error may reveal a speaker to be a less than
proficient language user, pragmatic failure reflects badly on him/her as a per-
son’. Overall, this review of what constitutes pragmatic success for requests
and apologies may be a good starting point for learners and practitioners to
enhance sociopragmatic and pragmalinguistic awareness, particularly when
preparing for interactions with members of an academic community.
References
Anthony, L. 2017. AntConc. (Version 3.4.4). [computer software] Tokyo, Japan:
Waseda University. Available from: <www.laurenceanthony.net/> [Accessed 2 March
2017].
Aijmer, K. 1996. Conventional routines in English: Convention and creativity. New
York: Longman.
Austin, J. 1962. How to do things with words. Oxford: Oxford University Press.
Bachman, L.F. 1990. Fundamental considerations in language testing (Oxford
applied linguistics). 5th ed. Oxford: Oxford University Press, USA.
Bachman, L.F. and Palmer, A.S. 1982. The construct validation of some components
of communicative proficiency. TESOL Quarterly, 16(4), 449–465.
Bachman, L.F. and Palmer, A.S. 1996. Language testing in practice: Designing and
developing useful language tests. New York: Oxford University Press.
Bardovi-Harlig, K. 2001. Evaluating the empirical evidence: Grounds for instruction
in pragmatics. In K. R. Rose and G. Kasper, eds. Pragmatics in language teaching.
Cambridge: Cambridge University Press, 13–32.
Bardovi-Harlig, K. 2009. Conventional expressions as a pragmalinguistic resource:
Recognition and production of conventional expressions in L2 pragmatics. Lan-
guage Learning, 59(4), 755–795.
Bardovi-Harlig, K. and Hartford, B.S. 1990. Congruence in native and Nonnative
conversations: Status balance in the academic advising session. Language Learn-
ing, 40(4), 467–501.
Barron, A. 2002. Acquisition in interlanguage pragmatics: Learning how to do things
with words in a study abroad (Pragmatics and beyond new series). Philadelphia,
PA: John Benjamins Publishing Co.
Barron, A. 2003. Acquisition in interlanguage pragmatics: Learning how to do things
with words in a study abroad context. Amsterdam: John Benjamins.
Bergman, M.L. and Kasper, G. 1993. Perception and performance in native and non-
native apology. In: S. Blum-Kulka and G. Kasper, eds. Interlanguage pragmatics.
Oxford: Oxford University Press, 82–108.
Blum Kulka, S. 1982. Learning to say what you mean in a second language: A study
of the speech act performance of learners of Hebrew as a Second language1.
Applied Linguistics, III(1), 29–59.
Blum-Kulka, S. 1983. Interpreting and performing speech acts in a second language:
A cross-cultural study of Hebrew and English. In N. Wolfson and E. Judd, eds.
Sociolinguistics and Language Acquisition. Rowley: Newbury House, 36–55.
156 Pragmatic competence
Conclusion
6.1 Introduction
This chapter will attempt to summarise the main findings of this book and
draw some conclusions and implications for both research and teaching.
In order to do so, section 6.2 gives a summary of findings for each aspect
of communicative competence we have described. We again separate each
competence for ease of analysis whilst acknowledging, as we have through-
out this book, that the different aspects of communicative competence do
overlap and interlock in many ways. We then move on to discussion of the
implications for research in section 6.3 before doing the same of teaching
in 6.4. We then give a final summary of the limitations of the research and
what we hope will be the implications.
of meanings which change across levels. At B1 and B2 levels, the use tends
to be focused on third parties, whereas at C1 level, the learners are also able
to use this to discuss concepts in a general sense.
In addition, in the USTC, er and erm are very frequent at all levels although
hesitation via er does reduce as proficiency grows which reflects changes
documented in CEFR descriptors of fluency. We suggested then that far from
being a mark of an unsuccessful speaker, hesitancy has important functions
as it signals to the listener that the speaker needs more time and it thus can
enable them to hold the floor and extend their own turn (see chapter 4 for
further discussion).
In terms of keywords, it was clear that the verb think is used throughout
in a range of patterns, illustrating how communicative routines for giving
opinions can be realised. Analysis revealed a variety of functions. It enables
learners to successfully sequence utterances, shift focus, express stance and
hedge language. Can also featured as a keyword and is used most often to dis-
cuss general possibility, with the focus shifting as levels increase. At B1, this
is used to talk about things from the speaker’s viewpoint (employing phrases
with I can) and at B2 and C1 levels, it is used to express more hypothetical
or general viewpoints (employing phrases with you can and we can). The use
of these keywords across the levels showed that success at higher proficiency
levels relates to the range of functions an individual word fulfils rather than a
learner simply using a broader range of words. Though increases in frequency
of lexical chunks are evident across the levels, there were no statistically sig-
nificant gains in usage as proficiency grew. Instead, it is clear that learners at
all levels favoured chunks such as I agree with you which can be used with a
range of functions such as expressing agreement or buying time, rather than
employing a range of chunks for one function. A number of chunks enable
learners at different levels to fulfil different aspects of communicative com-
petence. For example, we discussed how so yeah is used by B1 and C1 levels
to signal the closing of a turn as part of discourse competence in chapter 4.
that CSs relate to typical ‘problems’ or gaps in learner language but these do
not dominate successful speech at all; it is the strategies used to maintain,
extend, clarify and preface speech in jointly constructed interaction which
shine through. Arguments that CSs can boost and maximise perceptions of
learners’ overall communicative competence are to some extent supported
in this chapter.
Concerning production strategies, all learners exhibited an ability to
correct or reformulate their speech. Potentially seen as a shortcoming or
disfluency in learner language, the findings instead show that correction is
evidenced more frequently at higher levels both in overall and in learner-
to-learner speech. Whilst corrective focus shows parallels across all levels
with word choice in terms of pronoun and verb selection often receiving
attention, the data also display trends within each level: B1 learners attend
to tense use, B2 learners attend to missing words in their utterances, and C1
learners begin attending to reformulations of entire utterances as opposed
to the individual components within them. We argued therefore that cor-
rection can act as an indicator of proceduralisation as proficiency grows,
with speech monitoring following verbalisation potentially taking the place
of forward planning in speech. Closer inspection of corrective production
strategies thereby uncovered several layers as to the implications correction
has for learners’ successful speech both within and across proficiency levels.
Finally, though prefacing remarks to initiate, maintain and close discus-
sion were found to dominate learner interaction at B1, B2 and C1, it was
their ability to invite others into the conversation and clarify meaning when
necessary that were explored in more detail. Previous findings in chapter 2
suggested that B1 learners overlooked asking questions of partners in place
of communicating their own thoughts and opinions. It was instead found
that it was at B1, rather than at B2, that this interactive CS begins to emerge;
in fact, 21% of learner-to-learner CSs are aimed at conversing with others at
this level. Analysis of USTC data once again demonstrated a change across
B1, B2 and C1 in the nature of the question posed. Though prevalent at
B1, the use of yes/no, one-word questions such as do you agree with me?
and do you think X is good? decrease from B1 to C1; the ability to elicit
longer responses does however rise. We surmise that successful interaction
at higher levels hinges on this ability to seek extended responses from inter-
locutors, rather than short answers which in turn can provide the founda-
tion for jointly constructed discourse. Clarification requests were likewise
found to exhibit changes in both their focus and their expression. Whereas
B1 learners clarify task instructions on the majority of occasions, B2 learn-
ers concentrate on vague examiner questions and C1 learners enquire about
the meaning of vocabulary or questions. This is achieved through the use of
partially and fully formed questions as well as repetition. In our discussion,
we acknowledged that clarification could be viewed in a negative light; ask-
ing questions of fellow learners or native speakers could be taken as a sign
162 Conclusion
6.2.5 Summary
We can summarise the findings overall in table 6.1. This gives simple ‘can do’
statements based on our key findings for each chapter, by levels and in gen-
eral. As the data was not divided by levels when measuring pragmatic com-
petence we have kept these in the ‘B2’ category only. The intention here is
Table 6.1 Summary of successful speakers’ key competences at B1-C1 levels
B1 I can use we to refer to I can correct mix-ups with I can use it to link back to previous
third parties. tenses and expressions that lead ideas in my turn.
I can use I can to refer to misunderstandings without I can use simple additive
to general possibilities in interlocutor feedback. expressions such as for example,
relation to my own life. I can use simple question forms e.g. and also and I also to develop the
I can use chunks with I do you agree with me? requiring a yes/ theme of my turn.
think to express my own no or one-word answer to invite I can use so yeah to close a turn.
views. others into a discussion. I can use yeah in chunks such as
I can use statements and questions, yeah yeah to agree, pause or buy
though they may not always be time.
accurately formed, to clarify I can use Ok on its own and in
information from native speakers chunks such as Ok er to buy time
and fellow learners. or to show understanding.
I can enquire as to the meaning of
unknown vocabulary or unclear
questions as well as task instructions.
B2 I can use we to refer to I can correct slips and errors I can use it, increasingly this and I can formulate requests with
third parties and general relating to simple word choices that to link back to previous ideas conventional indirectness using modals
topics. and missing words when I become in my turn and a partner’s turn. such as can and could.
I can use I can and you conscious of them or they have led I can use simple additive I can use common 2–4 word chunks such
can to refer to general to misunderstandings. expressions such as and also and I as sorry to bother you to get the listener’s
possibilities in relation I can use open-ended questions also and increasingly for example to attention and could you give me, could you
to my own life and in e.g. how about you? requiring more develop the theme of my turn. help me to form the head act in requests.
relation to others’ lives. detailed responses to invite others I can use what do you think? to I can organise a request using the pattern
I can use chunks with into the discussion and maintain open a turn and invite a partner alerter (e.g. Excuse me) + polite request
agree with to agree with co-constructed conversation. to speak. head act using conventional indirectness.
a speaking, to signal I can use questions, though not I can use yeah on its own and in I can formulate apologies with explicit
some disagreement always accurately formed, to clarify chunks such as yeah I to agree, buy apologies incorporating an intensifier
with I agree with you vague questions that are asked of time and add a viewpoint. such as I’m so sorry.
but partner and to gain me. I can use Ok to buy time and I can use common 2–4 word chunks
thinking time. increasingly in chunks such as Ok such as I’m so sorry about that, I want to
I can use chunks with I Ok to show understanding. apologise (for).
think to give opinions, to I can use well in chunks such as well I can organise an apology using the
seek the views of others, I to mark hesitation. pattern explicit expression of apology +
to hedge my own views offer of repair/explanation.
and to hold the floor.
C1 I can use we to refer to I can backtrack when I encounter a I can use it, this and increasingly that
third parties’ general difficulty in speech relating to word to link back to previous ideas in my
topics and increasingly choice, missing words and tenses. turn and a partner’s turn.
hypothetical topics. I can also monitor utterances for I can use simple additive
I can use I can and you their success in conveying desired expressions such as I also, for
can to refer to general messages and reformulate them example and increasingly and also
possibilities in relation when necessary. to develop the theme of my turn.
to my own life and that I can use open-ended questions I can use what do you think? to
of others we can to e.g. what do you think about X? open a turn and invite a partner
increasingly general and requiring more detailed responses to speak.
hypothetical topics. to invite others into the discussion I can use so yeah to close a turn.
I can use chunks with and maintain co-constructed I can use yeah on its own and in
agree with to agree with conversation. These can also be used chunks such as yeah I to agree and
a speaking partner and to gain more thinking time when add a viewpoint, often continuing
to extend their ideas preparing a response. a topic begun by my speaking
from their turn into my I can use partially and fully formed partner.
own turn. questions, as well as repetition to I can sometimes use Ok to launch
I can use chunks with I clarify the meaning of vocabulary and turns or topics.
think to give opinions, to questions. I can also use repetition I can use well in chunks such as well
seek the views of others, to gain more thinking time when I to mark hesitation and sometimes
to hedge my own views preparing a response. to signal an unexpected response.
and to hold the floor.
In I can use vocabulary I can self-correct my speech when I can use er to buy time, mark
general form the first two errors relating to words choices, in hesitation and hold the floor.
thousand words to talk particular pronouns and verbs, occur. I can use chunks such as I think to
about everyday topics. I can clarify the responses mark hesitation and hold the floor
I can use three- and produced by native speakers and
four-word chunks such fellow learners to aid my own
as a lot of across a range understanding in real-time speech.
of everyday topics
I can use chunks with
think to give my views.
166 Conclusion
not to rewrite the CEFR ‘can do’ statements, as such an undertaking would
involve a much lengthier process and would need to be based on larger, more
comprehensive corpora. Instead, we simply offer these as summaries based
on our data, which we feel could be used as an initial guide to the commu-
nicative competence of successful speakers.
chunks. To take one example, can was a keyword at each level but was
patterned differently from B1–C1 and the range of functions for which
it could be used increased as the level did.
3 The use of chunks in achieving communicative competence was clear
across the chapters. One simple way in which teachers can supplement
courses with instruction on chunks is to use the many useful lists of
these now available based on corpus evidence (for example, Martinez
and Schmitt 2012) as a way of informing their teaching. Another way
teaching could address this would be exploring learner corpora such as
the USTC. The focus here could be upon drawing learners’ attention to
real language learner transcripts of speech and helping them to search
for chunks of language that are used to perform functions such as ini-
tiating, maintaining and ending turns, and seeking clarification. Asking
learners to perform a speaking task (such as a speaking exam practice)
and then examining a recording and transcript of a more successful
speaker (B1 level listening to B2 learners, for example) undertaking the
same task has long been advocated in relation to task-based learning
(for example, Willis 1996). Students can be asked to undertake the task,
then to listen to the recording of more successful speakers, first not-
ing general differences, then exploring the transcripts for differences in
the language used before being guided towards finding and underlining
or highlighting specific chunks with specific functions. Such a process
would help to foster the habit of noticing in learners, something which
has long been suggested is a crucial aspect of second language acquisi-
tion (Schmidt 1990). In addition to such classroom practices, assem-
bling a bank of multifunctional chunks such as the ones highlighted in
this study would also help teachers and learners to focus on using the
same chunks in different ways as their level increases.
4 Related to the use of the USTC corpus for teaching, is also the potential
role it could play as a study aid for learners. The language contained
within the corpus can itself be used to model how some functions can
be realised in learner speech. As a supplementary tool, it can offer a
substitute to resources containing NS speech which may seem unachiev-
able or unrealistic. Giving learners access to the corpus or the texts it
comprises could thus raise awareness of successful speech. For instance,
if the corpus was used to demonstrate discourse competence is realised
in speech at different levels, it could draw learners’ attention to features
which make the process of speaking with fellow language users a more
manageable and perhaps less daunting task.
and difficulty demonstrated that it was the flexibility with which individual
lexis could be used that most exemplified success at different levels, with
particular words and chunks revealing that learner proficiency is in part reli-
ant upon the manner that multifunctionality can be exploited and adapted
in speech. In chapter 1 we described a successful speaker as one who is able
to use linguistic, strategic, pragmatic and discourse competence as appropri-
ate for a particular goal, at a particular language level, as defined by the
CEFR. We hope that this book has contributed in a small way to under-
standing this concept and we hope it is one which is further developed with
the ultimate goal of helping as many learners as possible to be successful
speakers of English.
References
Adolphs, S. and Schmitt, N. 2003. Lexical coverage of spoken discourse. Applied
Linguistics, 24(4), 425–438.
Carter, R., Hughes, R. and McCarthy, M. 2000. Exploring grammar in context:
Upper intermediate and advanced. Cambridge: Cambridge University Press.
Council of Europe 2001. Common European Framework of Reference for Languages:
Language, teaching, assessment. Cambridge: Cambridge University Press.
Folse, K. 2004. Vocabulary myths: Applying second language research to classroom
teaching. Michigan: University of Michigan Press.
Gilquin, G. and De Cock, S. 2013. Errors in disfluencies in spoken corpora. Amster-
dam: John Benjamins Publishing.
Gilquin, G., De Cock, S. and Granger, S. 2010. The louvain international database
of spoken English interlanguage. Handbook and CD-ROM. Louvain-la-Neuve:
Presses universitaires de Louvain.
Haywood, S. 2014. Academic vocabulary. [Online] Available from: <www.nottingham.
ac.uk/alzsh3/acvocab/index.htm> [Accessed on 12 February 2017].
Hunston, S. 2016. Measuring improvements in learner language: Issues of complex-
ity and correctness. In: English Profile, 10th anniversary English Profile seminar.
Cambridge, Cambridge University Press. 5th February 2016.
Hunston, C. 5 February 2016. Measuring improvements in learner language: Issues
of complexity and correctness. Presentation given at 10th anniversary English
Profile seminar. Seminar conducted at the meeting of English Profile, Cambridge,
United Kingdom.
Laufer, B. 1998. The development of passive and active vocabulary in a second lan-
guage: Same or different? Applied Linguistics, 19(2), 255–271.
Lewis, M. 1993. The lexical approach: The state of ELT and a way forward. Hove:
Language Teaching Publications.
Mark, G. and O’Keeffe, A. 2015. Introducing the English grammar profile [Online]
Available from: <www.cambridge.org/elt/blog/2015/11/11/introducing-english-
grammar-profile-1-building-profile-1/> [Accessed on 20 February 2017].
Martinez, R. and Schmitt, N. 2012. A phrasal expressions list. Applied Linguistics,
33(3), 299–320.
McCarthy, M. 1999. What constitutes a basic vocabulary for spoken communica-
tion? Studies in English Language and Literature, 1, 233–249.
172 Conclusion