Data Collection For Speech Act Production example essay topic

1. Introduction Interlanguage pragmatics (ILP) is new branch of pragmatics and second language acquisition (SLA). Initiated by Kasper (1981), it has received more attention and become quite fruitful through eighties, then was established as an independent discipline with the publication of! ^0 Interlanguage Pragmatics!

+/- (Kasper & Blum-Kulka, 1993). It was first defined by Kasper & Dahl (1991: 216) as! ^0 nonnative speakers! comprehension and production of speech acts, and how that L 2-related knowledge is acquired! +/-. Later Kasper provided another definition: !

^0 ILP can be defined as the study of nonnative speakers! comprehension, production, and acquisition of linguistic action in L 2! +/- (Kasper, 1996: 85). These two definitions differ in that the latter one widen its research scope from speech acts to linguistic action. Nevertheless, both of them consider speech acts as the core of study in ILP.

As the first definition suggests, speech acts can be approached from their comprehension, production and acquisition, but this paper is confined to the production of speech acts. Research methodology is crucial to studies in every discipline, and it also holds true with ILP. Liu (1997) calls for studies on research methodology in ILP but there has been no paper concerned published in China so far. Kasper & Dahl (1991) provided a good review of the data collecting methods in ILP research, but it is a little old and needs a new paper adding more information. This paper focuses on the methods of data collection for speech act production in ILP study. In the following sections, I will first discuss on some questions to consider in collecting data.

Then I survey some important data collection techniques one by one. Finally I show how multiple techniques are used in data collection with illustrations by some typical studies. 2. Questions to consider during data collection 2.1 Initial questions There are at least two questions needing consideration before data collection starts. The first is whether to gather data longitudinally (that is, over a period of time from the same subjects) or cross-sectionally (that is, at a single point in time). Usually longitudinal studies focus on only a small numbers of subjects, while cross-sectional studies pool data from large numbers of subjects.

There are advantages and disadvantages to both types of study in research on variation in interlanguage (Tarone, 1988: 115-117). A focus upon longitudinal data will provide valuable detail about the performance of a few individuals, but not reliability in the sense that results cannot confidently be generalized to other individuals. A focus upon cross-sectional data will provide reliable information about general patterns among large bodies learners, but may obscure important patterns of variation over time and across individuals, and patterns such as the influence of one part of the IL system upon another. The second question is how to keep intervening variables constant in order to collect data of certain speech act.

The production of speech act is influenced by many sociopragmtic factors. Among them the most important comprise relative power (P), distance (D) and ranking (R). These three independent and culturally sensitive variables that subsume all others and that play a principled role in speech act realization (Brown & Levinson, 1987; Fraser, 1990). Before data collection begins, we should keep those variables constant in selecting participants. Hudson, Detmer & Brown (1991: 13) provides a summary of these variables quoted as below. Relative Power (P) of the speaker with respect to the hearer; in effect, the degree to which the speaker can impose will on the hearer due to higher rank within an organization, performance of job / duty, control over money in the situation, or professional status.

Social Distance (D) between the speaker ad the hearer; in effect, the degree of familiarity and solidarity they share. Absolute Ranking (R) of imposition in the culture, in terms of the expenditure of goods and / or services by the hearer, or the obligation of the speaker to perform the act. As our concern is the production of speech acts by non-native speakers, they are affected by additional variables, such as proficiency level of the target language and first language transfer. The participants in any study should give attention to these variables as well. If they are the focus of a study, we should try to keep all the other variables constant; if they are not, we should keep them constant. 2.2 Secondary questions The choice of data collecting methods is closely related to our research purposes and what types of data are expected.

This relation is clearly demonstrated by a series of questions below. It is necessary for any researcher to answer all or some of these questions when he is to collect data on speech act production. (1) Do you need performance data or data on cognitive process? Earlier research is usually aimed to provide a detailed description of how certain speech act is expressed, so performance data are collected accordingly. Later researchers go further to explore how participants plan and monitor their utterances while performing a speech act. To achieve this goal, the data of cognitive process are needed.

Cognitive process is seldom studied independently, instead it is studied after the description of speech act realization in one paper, see Cohen & Olshtain (1993) as an example. (2) If data on cognitive process, choose the technique of verbal report. (3) If performance data, do you need linguistic data or non-verbal data? (4) If non-verbal data, video taping is favorite. (5) If linguistic data, do you need elicited data or naturalistic data? (6) If naturalistic data, ethnographic observation is recommended.

(7) If elicited data, you can use discourse completion texts (DCT) or role play. The above seven steps can only serve as a general guiding clue during the course of determining the techniques appropriate to your study. More importantly, researchers should understand the features of these techniques in order to use them where they are desirable. In the next section, I will give a survey of the five data collecting techniques mentioned in the above seven steps.

3 Data collection techniques 3.1 Discourse completion texts (DCT) The format was first developed by Levenson & Blum (1978) to study lexical simplification, and first adapted to investigate speech act realization by Blum-Kulka (1982). There are three similar definitions of DCT. One is given by Kasper & Dahl (1991: 221): Discourse completion tasks are written questionnaires including a number of brief situational descriptions, followed by a short dialogue with an empty slot for the speech act under study. Subjects are asked to fill in a response that they think fits into the given context. A second definition is given by (Houck & Gass, 1996): Discourse completion tests are defined as written questionnaires consisting of a brief description of a situation followed by dialogue with a blank line where the subject is to put in what's / he believes to be an appropriate response (p. 46). Gass & Houck (1999) offered the third definition.

A discourse completion test is a pencil and paper test that requires subjects to write what they believe they would say in a particular context. In most tasks a situation is described, blank lines are drawn for the response (see example 1), and (sometimes) a follow-up response is provided (see example 2) (p 26). In addition, DCT can sometimes used in an oral version. To be specific, learners listen to a tape in which a situation is described, then they respond orally to the situation.

(1) Sample item 1 (Olshtain & Wein bach, 1993: 121) Situation 9 It is not the first time that loud rock music is heard from your neighbor!'s apartment quite late at night. You pick up the phone and say: (2) Sample item 2 (Hudson, Detmer, & Brown, 1991: 14) Directions: Read each of the situations on the following pages. After each situation write what you would say in the situation in a normal conversation. Example: Situation: You live in a large apartment building. You are leaving to go to work. On your way out, you meet your next door neighbor, whom you haven! t seen for a long time.

You: Neighbor: No, you haven! t. I have been out of town for a week. There are advantages and disadvantages to this type of data collection (Gass & Houck, 1999: 26). The advantages are clear. Large amounts of data can be collected in a relatively short amount of time and from participants with a wide range of individual characteristics. Furthermore, because of the consistency of the situation, responses can be compared along a number of dimensions (e. g., age, gender, ethnicity).

On the other hand, disadvantages are noteworthy. There is the question of the extent to which the data collected actually reflect the sociolinguistic constraints that operate on the speech act in question in natural circumstances. This is argued by Wolfson, Marmor, & Jones (1989: 182), who point out that! ^0 short de contextualized written segments! +/- may not be comparable to what takes place in actual interaction. Another problem with DCT is that certain kinds of information are not collected this way, such as the prosodic and nonverbal features of oral interaction. Furthermore, participants usually have more time to respond when doing so in writing than when doing so rally.

Written and oral versions (sometimes referred to as closed role plays) of the same discourse completion test have produced comparable results. Rintell & Mitchell (1989) used written and oral versions of the same discourse completion test (eliciting apologies and requests), which were given to low advanced learners of English and to native English speakers. They found that the! ^0 language elicited! - is very similar whether collected in written or oral form! +/- (p. 270).

They argue that the discourse completion test is actually a role play. That is, both the written and spoken forms provide data that resemble spoken language rather than written language... 3.2 Acceptability rating Acceptability rating involves the obtaining of respondents! judgments as to how appropriate certain responses are for a give situation (Olshtain & Blum-Kulka, 1985). In this technique, a series of possible responses are presented and the respondent has to select the most appropriate of them for the given situation.

Usually the responses are scaled on continuum according to some dimension (Cohen, 1996: 26). For example, (3) Sample item (Hudson, Detmer, & Brown, 1991: 15) Directions: Please read the following descriptions of situations in which you might find yourself. Pretend that you are in the situation and imagine what you would do and say. Read the choices and indicate which you think is the most correct response. Example: You are in a department store buying an item and you accidentally knock over but do not break something on the counter. a. Oh, I! m terribly sorry. b.

Do you think you can ever forgive me. c. Oh, I! m terribly sorry. Can you find the way to give me forgiveness for my clumsiness. d. Sorry. 3.3 Role play Role plays may be open or closed.

In closed role-plays, participants usually are usually given an oral version of the discourse completion test. In open role-plays, an entire dialogue is observed and recorded. In an investigation into the sociocultural competence of L 2 speakers, Cohen & Olshtain (1981), used a procedure in which subjects were presented with role cards such as the following: (4) Sample item (quoted from Nunan, 1996: 369) Directions: you will be asked to read eight brief situations calling for an apology. In each case, the person who you owe the apology to will speak first. I will role play this person. Respond as much as possible as you would in actual situation.

Your response will be tape-recorded. Example: You! re at a meeting and you say something that one of the participants interprets as a person insult to him. He: I feel that your last remark was directed at me and I take offense. You: . Role-plays have the advantage of providing data in an oral mode rather than a written mode. In a closed role-play, participants are given a situation and are asked how they would respond.

In Walters! study (1980) children were asked to make a request to a puppet. These puppets varied in age, sex, and race. As! ^0 closed! +/- data do not allow a free range of answers, they suffer from the possibility of non-symmetry with naturally occurring data. Open role plays have the advantage that they allow examination of speech act behavior in its full discourse context (Kasper & Dahl, 1991: 228). Role plays have the advantage over authentic conversation that they are replicable and allow for the comparative study of Nss and L 1 and L 2 NS controls.

A disadvantage that open role plays share with authentic conversational data is that they need transcribing. Coding open role play data is more difficult than coding data from more tightly controlled tasks, since illocutionary force and the precise function of conversational markers often cannot be unambiguously determined and these facts make inerrater reliability harder to achieve (Kasper & Dahl, 1991: 229). Comparatively speaking, open role-plays are better in collecting naturally occurring speech events. They have the advantage of allowing the researcher to set up situations in which the occurrence of a particular speech act can be recorded or video-taped, so long interaction sequences of comparable data are available for close analyses. 3.4 Ethnographic observation Ethnographic observation involves the collecting of naturally occurring data. This method has proven effective in collecting data on certain speech acts, such as compliments (for example, Wolfson, 1989).

Yet for other speech acts, such as apologizing, it may be extremely time-consuming and not very productive (see Cohen, 1996 for details). Wolfson (1981) and Wolfson, Marmor & Jones (1989) have argued that ethnographic observation is the most reliable means of learning about the social and linguistic constraints on a particular speech act. This methodology allows for observation of naturally occurring speech events with precise recording about the social setting, location, and the participants, thereby providing information about the linguistic and social constraints on the use of a given speech act. However, it also has limitations. Not only can contextual variables not be controlled, but also the occurrence of a particular speech act cannot be predicted.

3.4 Video taping Video is seldom used within the context of speech act research, this may partly be due to the fact that there has been an emphasis on either discourse completion tests, where clearly video data are superfluous, or spontaneously produced speech, where video data cannot be planned. Videotaped data also have some weaknesses (Gass & Houck, 1999: 50): 1) Video equipment is cumbersome to use in data collection. At least with adults, it can be used only in planned and perhaps contrived instances, as with role plays; 2) Gathering data with video is more intrusive than with audio-recorder only; 3) some subjects who do not mind being audio-taped do mind being videotaped; 4) Showing videotaped data may result in viewers being distracted by the physical appearance of the participants rather than focusing on the communication event that they are engaged in; 5) Videotaped data are unusually difficult to transcribe. 3.5 Verbal report When researchers want to discover the cognitive processes that learners go through to produce speech acts, verbal report is usually employed. Considering the intrusive nature of verbal report techniques, it would be unreasonable to ask speakers to provide such data while they are engaged in communication.

Yet once the interaction is over, participants may not be able to retrospect fully as to the strategy selection that they carried out a few minutes ago. There are some studies using verbal report in speech act production. Only two studies are reviewed here, one used with discourse completion tests and the other with role-play. Motti!'s study (1987) involved 10 intermediate EFL university students in Brazil. After filling out a discourse completion task calling for apologies in English, the students were asked to retrospect individually in Portuguese regarding a series of variables.

These variables include the students! depth of analysis of the situation before response and the extent to which they thought in L 2 or L 1 while preparing and writing their responses. Motti found that participants thought slightly more in English than in Portuguese in the planning and execution of their utterances and were preoccupied with correctness. They also reported paying more attention to another!'s status than to age. The first study combining verbal report lb? lb? is that of Cohen & Olshtain (1993). The study sought to describe ways in which nonnative speakers plan and execute speech act utterances.

The subjects, 15 advanced English foreign-language learners, were given six speech act situations in which they were to role-play along with a native speaker. The interactions were videotaped, and after each set of two situations of the same type, the tape was played back and the respondents were asked both fixed and probing questions regarding the factors contributing to the production of their response to that situation. The probing interviews conducted after each set of two speech act situations were intended to employ retrospective self-observation in order to obtain verbal report data about the cognitive processes that went into the production of speech act realizations... 4 Multiple data collection techniques Some studies are devoted to a comparison of methodology. In addition to Kasper & Dahl!'s (1991) survey on methodological issues, there are a number of other studies that present results based on a comparison of methodologies.

Some of the most interesting studies involve comparisons of DCT with naturally occurring data. Beebe & Cummings (1996) studied refusals using two types of data for their analysis, real telephone requests and a discourse completion test based on these requests. Data were collected only from native speakers of English. What they found was that in the oral data there was more elaboration of the refusal; in the written data, the layout on the page allowed for only a minimum amount of data to be produced. The elaborate nature of spontaneously produced refusals has been further pointed out by Hartford & Bardovi-Harlig (1992). They compared semantic formulas resulting from naturally occurring data with those elicited by a discourse completion test.

Data were collected in an institutional setting in which the interact ants and situations are relatively invariant. The interactions were audiotape d, and participants were aware of being observed and recorded. Their analysis focused on rejections of advice in academic advising sessions. The eight discourse completion tests developed for comparison were based on data from these sessions.

They included situations which were frequent in the advising sessions with non-native students, but rarely occurred with native speaker students. Hartford & Bardovi-Harlig discovered a number of differences between data from the discourse completion tests and the real advising sessions. They found that the most frequent strategies used in responses to the DCT differed from the most frequent strategies in the naturally occurring data: 1) a number of common semantic formulas in real advising sessions did not occur in responses to DCT and 2) semantic formulas that occurred in the DCT data did not occur in the advising sessions. Hartford & Bardovi-Harlig concluded that DCT result in limited data. Clearly, written responses, especially those that are sandwiched between an opening statement and a follow-up statement (as in many DCT), do not allow a speaker to exhibit the full range of response types that are found in naturalistic data. This does not necessarily suggest an overall superiority for the collection of naturalistic data since the elicited data allow for the testing of hypotheses, which is not possible when data are sporadic and uncontrolled.

Other research on methodology has focused on multi-method comparison. Yamashita (1996) investigated testing instruments for the assessment of pragmatic knowledge English native speakers learning Japanese. Six tests were used, including 1) a self assessment, 2) an oral production test, 3) an open discourse completion test, 4) an role play 5) a role play self-assessment, and 6) a multiple choice discourse completion test. In the self-assessment, learners were presented with a situation (written format), as in a discourse completion test, and were asked to think about what they say in Japanese in such a situation. In the oral production test (an oral discourse completion test), learners listened to a tape in which a situation was described, after which they responded orally to the situation. The open discourse completion test used a written format, both in the presentation of the situation and in the response requested.

The role play was an oral video-taped session in which participants acted out their responses to the scenario presented. After learners participated in the role play, they provided a self-assessment of their ability to respond appropriately. Thus, the role play itself provided two measures of assessment (the role play and the self-assessment of the performance of the role play). The final measure used a multiple choice format. Situations just like the ones in the discourse completion tests were presented with three possible responses from which subjects had to select one. The results of this multiple-method comparison suggest that with the exception of the multiple choice instrument, all methods are reliable and valid.

Further, issues such as proficiency and exposure to the target culture are related to some measures, but not others. In particular, a relationship exists between length of exposure to the L 2 and the two oral production tests (oral production and role play) and between proficiency and the three production tests (oral production, role play, and open discourse completion test) A sequential research approach for a study on speech act realization by native and nonnative speakers may include the following data types: 1) observational data of authentic interaction; 2) (metapragmatic) assessment data of socio pragmatic factors; 3) production data, elicited through role plays or discourse completion questionnaires; 4) informants! verbal report on the production task; 5) (metapragmatic) assessment data of the production data by native and nonnative raters. A study which closely approximates this design is Eisenstein & Bodman!'s (1993) on expression of gratitude. An ideal cycle of data collection (see Figure 1) has been proposed by Olshtain & Blum-Kulka (1985) (Cohen & Olshtain, 1994: 148; Cohen, 1996: 21-22).

Investigators would start with the generation of initial hypotheses based on ethnographic data collection of natural speech. Then they would elicit simulated speech such as that in role-plays, to test the initial hypotheses. From there they might go on to a paper-and-pencil task such as discourse completion in order to focus on specific realizations and manipulate the social and situational variables. If they are concerned with the perlocutionary aspect of speech acts, they might want to use acceptability checks in order to validate the range of acceptability within a speech community.

Finally, we might be advised to validate the findings by means of further naturalistic, observational data. Cohen & Olshtain (1994: 149) that verbal report can be used with any of the four data collecting methods listed in Figure 1, as shown by Figure 2. But there have been only studies associating discourse completion and role play with verbal report (see Section 3.5 for details). Ethnographic observation Acceptability role play rating discourse completion Figure 1 The cycle of speech act research methods ethnographic observation acceptability verbal report role play rating discourse completion Figure 2 Verbal report in speech act research 5 Concluding remarks In this paper, I have first discussed on some questions that are needed to consider before or during data collection for speech act production. These discussions will throw light on researchers! determination of data collection methods and specific techniques. After that, I provide a detailed description of five most commonly used techniques in collecting data in research on speech act production.

Moreover, the next section has demonstrated how these techniques are used in one study according to different research purposes and research questions.

Bibliography

Bardovi-Harlig, 1999.
Exploring the interlanguage of interlanguage pragmatics: A research agenda for acquisition al pragmatics. Language Learning, 49,677-713. Beebe, L. & Cummings, M.C. 1996.
Natural speech act data versus written questionnaire data. In S. Gass & J. Neu (Ed. ). Speech acts across cultures: Challenges to communication in a second language (pp. 65-88). Berlin & New York: Mouton de Gruyter. Blum-Kulka, S. 1982.
Learning how to say what you mean in a second language; A study of speech act performance of learners of Hebrew as a second language. Applied Linguistics, 3, 29-59. Blum-Kulka, S. 1991.
Interlanguage pragmatics: The case of requests. In R. Phillips on, E. Kellerman, L. Selinker, M. S harwood Smith, & M. Swain (eds. ), Foreign / Second Language Pedagogy Research (pp. 255-72). Clevedon and Philadelphia: Multilingual Matters. Cohen, A., & Olshtain, E. 1981.
Developing a measure of sociocultural competence: The case of apology. Language Learning, 31,113-134. Cohen, A.D. & Olshtain, E. 1993.
The production of speech acts by EFL learners. TESOL Quarterly, 27, 33-56. Cohen, A D. & Olshtain, E. 1994.
Researching the production of second-language speech acts. In Tarone, etc. Research methodology in SLA (143-156). Douglas, D. & Selinker, L. 1994.
Research methodology in context-based second language research. In Tarone, E., Gass, S. & Cohen, A. (Ed). Research methodology in SLA (119-132). Eisenstein, M. & Bodman, J. 1993.
Expressing gratitude in American English. In Kasper, G. & Blum-Kulka, S. 1993 (Eds.
Interlanguage Pragmatics (pp. 64-81), Oxford: OUP. Fae rch, C., & Kasper, G. 1989.
Internal and external modification in interlanguage request realization. In S. Blum Kulka, J. House, & G. Kasper (eds. ). Cross-cultural pragmatics: Requests and apologies (pp. 221-247). Norwood, NJ: Ablex. Gass, S.M. & Houck, N. 1999.
Interlanguage refusals: A cross-cultural study of Japanese-English. Berlin & NY: Mouton de Gruyter. Hartford, B.S., & Bardovi-Harlig, K. 1992.
Experimental and observational data in the study of interlanguage pragmatics. In L.F. Bouton, Y.K. Urbana (Ed. ). Pragmatics and Language Learning Monograph Series, Vol. 3 (pp. 33-52). Urbana, IL: Division of English as an International Language Intensive English Institute, University of Illinois at Urbana-Campaign. Houck, N. & Gass, S. 1996.
Non-native refusals: A methodological perspective. Speech acts across cultures: Challenges to communication in a second language (pp. 45-64). Hudson, T., Detmer, E. & Brown, J.D. 1991.
A framework for testing cross-cultural pragmatics. Technical report 2. Honolulu, Hawai! i: University of Hawai! i, Second Language Teaching & Curriculum Centre. Kasper, G. & Blum-Kulka, S. 1993 (Eds.
Interlanguage Pragmatics, Oxford: OUP. Kasper, G. 1989 a.
Variation in Interlanguage Speech Act Realization. In S. Gass, C. Madden, D. Preston, & L. Selinker (eds): Variation in Second Language Acquisition: Discourse and Pragmatics, Multilingual matters. Cleredn & Philadelphia. Kasper, G. 1989 b.
Interactive procedures in interlanguage discourse. In W. Oleksy (ed. ), Contrastive Pragmatics (189-229). Amsterdam: Benjamins. Kasper, G. & Dahl, M. 1991.
Research Methods in Interlanguage Pragmatics. SLA, 13,215-247. Kasper, G. 1996.
Interlanguage pragmatics. In Karen Ri sager (Ed. ). RO LIG-PAPER 57 (pp. 84-104). Roskilde University Centre. Leather, J. & James, A. 1996.
Second language speech. In W.C. Ritchie & T.K. Bhatia (Ed. ). Handbook of second language acquisition (pp. 69-316). San Diego: Academic Press. Nunan, D. 1996.
Issues in second language acquisition research: Examining substance and procedure. Handbook of second language acquisition (pp. 349-374). Olshtain, E. & Blum-Kulka, S. 1985.
Cross cultural pragmatics and the testing of communicative competence. Language Testing, 2, 16-30. Rintell, E.M. & Mitchell, C.J. 1989.
Studying requests and apologies: an inquiry into method. In S. Blum-Kulka, J. House & G. Kasper (Ed.) Cross-cultural Pragmatics: Requests and Apologies (pp. 248-272). Norwood, NJ: Ablex. Tarone, E. 1988.
Variation in interlanguage. London: Edward Arnold. Walters, J. 1980.
Grammar, meaning and sociocultural appropriateness in second language acquisition. Canadian Journal of Psychology, 34,337-345. Wolfson, N. 1981.
Compliments in cross-cultural perspective. TESOL Quarterly, 15,117-124. Wolfson, N. 1989.
Perspectives: Sociolinguistics and TESOL. Cambridge: Newbury House. Wolfson, N., Marmor, T., & Jones, S. 1989.
Problems in the comparison of speech acts across cultures. In S. Blum-Kulka, J. House, & G. Kasper (Ed. ), Cross-cultural pragmatics: Requests and apologies (pp. 174-196). Norwood, NJ: Ablex. o^Ix^O'E. 1996.