This study analyzed differences between native and nonnative English speakers in how they deliver dispreferred responses during a debate. The study found that native speakers emphasized agreement more at the beginning and were more indirect at the end when disagreeing. Nonnative speakers' responses were more direct throughout. As interactions progressed, both groups became more direct, though nonnatives remained more so. These findings suggest it is important for English language teaching to include conversation strategies to help nonnatives interact successfully with native speakers.
Conversation Analysis: Directness in NNS's Dispreferred Responses
1. 1
Directness of Nonnative Dispreferred Responses in a Debate:
The Need for Implementing Conversation Strategies in EFL
Instruction
By
Roberto Criollo
Columbia University, NYC-Benemérita Universidad Autónoma de
Puebla
Proceedings of the XIV FEULE. Toluca, EDOMEX.
Abstract
Nonnative linguistic production has been found to be qualitatively and
quantitatively different from the native speaker rules at the phonological,
grammatical, lexical, and pragmatic levels. This paper explores the more
challenging and crucial differences that arise at the level of actual
communicative interaction through a conversation analysis in which NNS
were found to disagree more bluntly. The most salient differences of native
and nonnative dispreferred responses are presented and discussed, along
with their implications for ELT.
2. 2
CHAPTER ONE: INTRODUCTION
Dispreferred responses have been studied before by numerous researchers,
many of whom have been particularly concerned with face-threatening acts. Beebe and
Takahashi (1989a and 1989b), Takahashi and Beebe (1987) and Beebe, Takahashi and
Uliss-Weltz (1990) concentrated on refusals, which belong to the broader category of
dispreferred responses (Levinson 1983). While these studies have been sociolinguistic in
nature, the interactive aspect of this phenomenon in conversation remains unexplored.
The purpose of this paper is, then, to analyze and describe pragmatic strategies in
the context of talk-in-interaction. We will be focusing on dispreferred responses
(Levinson 1983, Pomerantz 1984) and how they are performed by native and nonnative
speakers with high proficiency in the language to find out whether and to what extent the
strategies used by both groups are the same or different during the course of a debate. In
so doing, we expect to increase our understanding of conversational structures, and the
development of discourse and strategic competence (Canale 1980) in nonnative speakers,
which are essential components of communicative competence (Hymes 1977, in Brumfit
and Johnson 1979; Coulthard 1985) that have generally been neglected in the language
classroom.
This study sets out to answer the following research questions:
1) What are some characteristics of dispreferred responses produced by native
and highly proficient nonnative speakers of English during the course of a
debate?
2) Do these characteristics change throughout interaction?
3. 3
3) Are there any differences in the conversation strategies used by native and
nonnative speakers?
4) If so, what are the most salient differences between native and nonnative
dispreferred responses?
5) Based on the findings of this research. What conclusions can be drawn about
conversation strategies and what are the implications for the EFL classroom?
4. 4
CHAPTER TWO: LITERATURE REVIEW
Levinson (1983) and Pomerantz (1984) characterize preferred and dispreferred
responses, pointing out that in conversation there is a natural tendency to preferred
responses, which are delivered quickly and are short. On the other hand, dispreferred
responses are performed with some delay (between and within turns), variously hedged,
and indirect.
Pomerantz (1984) points out that dispreferred responses occur for several reasons.
First, when the next speaker does not understand what the current speaker is talking
about, or when the current speaker uses terms unclear or unknown to the recipient.
Second, when the speaker says something on the belief that the next speaker already
knows of. Third, a recipient’s hesitation can be explained on the grounds of disagreement
or lack of support with what is being said, and ‘The recipient may directly address the
prior talk, for example, confirm, elaborate on, challenge, query or disconfirm the
assertion. On the other hand, the interlocutor may look blank or questioning, or may make
hesitating noises such as Uhs, Ums, and Wells.
Levinson (1983)summarizes the correlations of content and format in adjacency
pair seconds as shown below in Table 1.
Table 1: Levinson’s Categorization of First and Second Parts
FIRST PARTS:
Request Offer/Invite Assessment Question Blame
SECOND PARTS:
Preferred: Acceptance Acceptance Agreement Expected Denial
answer
Dispreferred: Refusal Refusal Disagreement Unexpected Admis-
answer or sion
non-answer
5. 5
Schegloff, Jefferson and Sacks, for their part,(1977) provide a classification of
speakers’ preferred strategies for repair in conversation. In their paper, they conclude that
self-initiated, self-repair is the preferred strategy and other-initiated, other repair is the
least preferred one. In other words, when a breakdown in conversation is produced,
speakers prefer to correct themselves before they may be corrected by somebody else.
The models above will be useful to describe cases of dispreferred responses as
they appear in the transcribed talk of native and nonnative speakers to analyze how both
groups perform dispreferred responses when disagreeing to opinions, assessments, and
answering questions.
6. 6
CHAPTER THREE: METHODOLOGY
3.1 Participants
In order to compare and contrast the different elements and particularities of talk-
in-interaction for native and nonnative speakers, two groups of four people were selected
to participate in this study. All eight subjects were graduate students at Teachers College,
Columbia University. In the first group, all were native speakers of English and in the
second group, all members were highly proficient nonnative speakers from different
linguistic backgrounds: one Korean, one Chinese, one Albanian, and one Spanish.
The gender and sex for both groups was mixed and they were randomly chosen, to
avoid gender- or age-related particularities. However, both groups were formed with
graduate students at Teachers College, Columbia University, because they had to be able
to control the same topic and being in equal position to the different members of the
group.
3.3 Instruments
The data was collected from a recorded interaction elicited with a task sheet. This
task asked group participants to select the most appropriate kind of curriculum to be
chosen for a particular group of students (see simulation task in Appendix 1). Participants
were supposed to support a particular kind of curriculum and had to persuade the others
that their proposal was the most appropriate, using their convincing strategies. This kind
of task was ideal to elicit dispreferred responses, since every person had to support one
argument and try to win.
7. 7
3.3 Procedures
First of all, both groups were given the task. Task sheets were handed separately
to each participant in the groups 20 minutes before the discussion, so that they could
study it and prepare individually for the conversation with the other people in their group.
Once they finished studying their task, both groups were taken to a different room
were they would talk for 20 minutes, and try to come to an agreement as to which kind of
curriculum they would choose. Participants were audio and videotaped in order to
achieve a good level of accuracy when transcribing, and for subsequent study or reference
in case the research wants to be extended to a broader domain (i.e. including body
language).
3.3 Data Analysis Procedures
The conversations were transcribed, first from the audiotapes and the difficult
parts (where the sound was not good) were contrasted and complemented with the
videotapes. A group of four people crosschecked the transcripts for a higher level of
accuracy. Four transcription drafts were written before getting to the final version. The
symbols used in the transcription are those proposed and used by Atkinson and Heritage
(1984).
8. 8
CHAPTER FOUR: FINDINGS
First of all, the data was examined to find instances of dispreferred responses
(DR), as defined by Levinson (1983) and Pomerantz (1984). Once the different adjacency
pairs including dispreferred responses were identified and isolated, the particular
characteristics of each sequence were examined.
4.1 DR at the Beginning of the Debate
A really interesting pattern was found in native-speaker dispreferred responses. In
this regard, participants exhibited almost all of the features described by Pomerantz
(1984) and Levinson (1983). They were delivered with some delay, a preface (generally a
token of agreement), a marker of dispreferred (but), and an account of the dispreferred.
Emphasis on agreement at the beginning of the turn containing the dispreferred was
prominent. Consider, for example, Fragment 1a, below.
Fragment 1a, NATIVE SPEAKERS: Opinion VS disagreement (p.2)
23 E: Uh I thought about-um- all the options. I
24 thought about grammar uh based curriculum
25 the TOEFL base curriculum and the TOEFL
26 opinion curriculum. Um and I think that - uh - both
27 of those kinds of curriculum are too narrow
28 for the students. Who are your students
29 first of all. Um they’ve been especially
30 prepared twenty of them uh to take parts in
31 (experimental) program. Chances are
01 they’re language teachers (although
02 necessarily >they have to pass the test<)
03 E: Um these people have learned other
04 languages in the past. So probably,
05 apparently they learned English as well
06 as others. And to- uh give’em the TOEFL-
07 account based curriculum, I think it’s giving them
08 a too narrow curriculum ‘cause they really
09 want to be prepared for experiences at TC.
10 opinion Um uhhh that’s why I think the content
11 based curriculum - that is- based on what
12 they will experience here. That’s what I
13 think is more necessary for them.
14 Agreemt A: I agree with you on many points, but I
9. 9
15 think - we need to go a little more toward
16 delay the grammar. Uh the TOEFL is, is, a lot of
17 grammar on that. =it’s almost scientific. =
18 account you have to know A plus B plus C. hhh
19 Although there will be content, - of course
20 you’ll get a lot of, in a time in America
21 living here, they need a lot of (doing)
22 disagremrnt function type of things, but I think it’ll
23 be much more useful to teach grammar
24 because that will help greatly in getting
25 over the TOEFL.
The nonnative group, in contrast, did not emphasize agreement, but only gave a
slight sign of acknowledgement, and even though they still made use of the features of
dispreferred responses, their responses were rather direct and immediate. Fragment 1b
below illustrates this.
Fragment 1b, NONNATIVE SPEAKERS: Opinion/stance VS
disagreement
07 opinion R: In my opinion, grammar is like the co:re -
08 of - language and the kind of skills that
09 they’re gonna need. (1.3) uh - I dunno,
10 stance what do you think? [ Maybe ]
11 token J: [Yeah, but] for six
12 disagreem months grammar based instruction - would be
13 boring I guess.
4.1 DR in the Middle Part of the Debate
Similar to the DRs at the beginning of the interaction, native speakers kept
emphasizing agreement. However, Fragment 2b below shows an even more complex
structure. After an opinion has been stated by A, M elaborates on his dispreferred. First,
he delays the dispreferred (‘well’) and makes use of a very strong token of agreement (‘I
agree with both of you one hundred percent’) followed by a marker of dispreferred
10. 10
(‘but..’, 2:30-3:01), then he offers an account, which is also very elaborate (3:03-18).
Finally, he gives the dispreferred (‘they don’t need it’,3:03-18).
Fragment 2a, NATIVE SPEAKERS: Opinion VS disagreement
25 opinion A: I feel that, uh grammar based course - will
26 help greatly in - both uh getting over the
27 TOEFL and toward life (2.0) at TC and -
28 life in New York.
29 Al: ((cough))
30 token M: Well - I agree with both of you one hundred
31 percent but that’s because you’re
01 supporting my argument.
02 All: ((laugh))
03 M: And this is why.
04 All: ((laugh))
05 delay M: Because [uh (1.0)
06 Al: [((clearing throat))
07 account M: You both were speaking about preparing
08 them for their experiences at TC, but that
09 makes me question. =We have to get them
10 into TC first, and they can’t get into TC
11 token unless they pass that test. So I agree we
12 do need a curriculum that address - the uh
13 addresses the, their very wide range of
14 interests and abilities, but -
15 unfortunately if they can’t pass that test,
16 then uh then their ability to ( )
17 acknowledge the grammar will never be put
18 disagreement to the test. =They don’t need it.
In contrast, nonnative speakers’ DRs became more direct. In fragment 2b, R’s
opinion (‘I thought it was very interesting’, 16:9) and ensuing stance (‘right?’, 16:12) are
followed by L’s immediate repair (‘I wouldn’t say exactly...’, 16:15). (other repair is the
least preferred type of repair). R’s dispreferred is immediate and categorical (‘you said
situations’, 16:17).
11. 11
Fragment 2b, NONNATIVE SPEAKERS: Opinion/stance VS other repair,
followed by disagreement
07 R: So what I was saying he was uh - with
08 Leonard something that Leonard said and I
09 opinion thought it was very interesting [is that
10 J: [Mmhm,
11 R: =(0.6) by providing them with situations,
12 stance right?=
13 J: =Mmhm,
14 R: To, [to
15 repair L: [I wouldn’t say exactly with situations
16 in the real sense
17 disagr. R: You said situations=
4.1 DR at the End of the Debate
Finally, Fragment 3a was taken from the last part of the interaction in the native
speaker group, where the conversation was coming to an end. Even when dispreferreds
needed to be more direct, NS were reluctant to disagree bluntly, as shown below.
Fragment 3a: NATIVE SPEAKERS: Stance VS distance
11 stance M: So I think we all agree? TOEFL right?
12 Al: OK. TOEFL
13 stance M: Alex?
14 delay A: (0.1)
15 A: I’m not sure.=
16 All: =((laugh))
17 distance A: I don’t I don’t
18 M: I - I
19 All: ((laugh))
20 disagr. A: I still maintain - that’s too narrow a
21 focus.
Nonnative speakers’ dispreferred responses at this stage of the interaction were
categorical and immediate. Something interesting in their interaction is that a flat
dispreferred triggers another flat dispreferred which did not seem to be the case with
native speakers in this data.
12. 12
Fragment 3b, NONNATIVE SPEAKERS: Opinion VS disagreement
followed by disagreement
19 opinion K: =These people should be continuously motivated
20 (0.4) to study.
21 token&disJ: Okay. But (0.5) I don’t think six months is
22 a (0.2) long term.
23 flat dis.K: It is.
24 flat dis. J: No - I don’t think so. But how can the
con…
The above analysis and data suggest that conversation in both groups displays a
gradually increasing tendency toward directness as the talk-in-interaction keeps
progressing. However, nonnative dispreferred responses were found to be more direct
and immediate.
13. 13
CHAPTER FIVE: CONCLUSIONS
In the previous pages the characteristics of dispreferred responses in the
conversational data from two groups of native and nonnative speakers were described and
discussed in order to identify patterns of differences or any other phenomena in the way
they uttered dispreferred responses.
It was found that native speaker talk displayed a few features distinctive from
nonnative speaker talk, where dispreferred responses are concerned. At the beginning of
the interaction native speakers emphasized agreement, whereas nonnative speakers only
acknowledged it; at the end, native speakers less indirect in their dispreferred responses,
but did not reply with flat dispreferreds, whereas nonnative speakers responded to a direct
dispreferred with another direct dispreferred, as a closing third.
These findings can have many implications for teaching discourse skills in the
language classroom. They suggest that, at least in this data, native speakers seem to use
slightly different strategies in giving dispreferred responses from those of nonnative
speakers. This makes it necessary to include conversation skills into the language-
teaching syllabus, since the ultimate goal of language learners is to successfully interact
with native speakers of the target language. On the other hand, the analysis also shows
that dispreferred responses are not always delivered in the same way, but vary throughout
interaction. A language teacher should keep these things in mind in order to make better
decisions while teaching communicative skills.
14. 14
REFERENCES
Beebe, L. and Takahashi, T. (1989a). ‘”Do you Have a Bag?” Social Status and Pattern
Variation in Second Language Acquisition’. In Gass et al. (eds.) 1989a.
Beebe, L. and Takahashi, T. (1989b). Sociolinguistic variation in face-threatening
speech acts, in Eisenstein (ed.) 1989.
Beebe. L., T. Takahashi, and R. Uliss-Weltz. (1990).Pragmatic transfer in ESL
refusals, in Scarcella et al (eds.) 1990.
Brumfit, C.J: and Johnson, K. (1979). The communicative approach to language
teaching. New York: Oxford University Press.
Canale, M. (1980). From communicative competence to communicative language
pedagogy. In J.C. Richards and R. Schmidt (eds.) Language in communication.
Harlow, Essex: Longman.
Canale, M. & Swain, M. (1980). Theoretical bases of communicative approaches to
second language teaching and testing. Applied Linguistics 1, 1-47.
Coulthard, M. (1985) (2nd
edition). An introduction to discourse analysis. 7: 147-59.
London: Longman.
Levinson, S. (1983). Pragmatics. Cambridge: Cambridge University Press.
Pomerantz, A. (1984). Pursuing a Response. In Atkinson J.M. and J. Heritage (eds.),
Structures of Social Action: Studies in Conversation Analysis (pp. 152-63).
Cambridge: Cambridge University Press.
Richards, J.C., Platt, J., and Platt, H. (1992). Dictionary of language teaching and
applied linguistics. Harlow, Essex: Longman
Schegloff, E., Jefferson, G., and Sacks, H.. 1977. The Preference for Self-correction in
the Organization of Repair in Conversation. Language 53 (2), 361-382.