In C.M. Tung, B.C. Huang, & C.M. Liao (Eds.), Papers from the third conference on English language teaching and learning in the Republic of China (pp. 365-380).  Taipei: Crane Publishing Co., Ltd., 1986.

Introduction

Recent statistics show that the largest group of foreign students in United States universities comes from Taiwan.  The majority of these are graduate students.  In the United States, most graduate students work as teaching assistants or research assistants during at least part of their graduate study to gain professional experience and to help fund their education.  As a result, teaching assistants (TAs) who are nonnative speakers of English teach courses in American universities.

Hinofotis and Bailey (1980) cite a number of reports and communications which mention the limited communication skills of nonnative speaking TAs.  Bailey's (1982) dissertation terms this situation the "foreign TA problem," and she provides a thorough review of the literature on nonnative TAs' limited teaching abilities.  Obviously, the solution to the problem in not in prohibiting nonnative TAs from teaching, but in testing the oral communication skills of prospective TAs and providing training for those people deficient in said skills.  Consequently, a number of universities, for example, the University of Minnesota (Mestenhauser et al., 1980; Land & Perry, no date) and the University of California at Berkeley (McKay, 1977-78) have put together detailed outlines of their courses to help nonnative TAs improve their communication skills in the classroom.

Pronunciation, grammar, and fluency present the most obvious problems for nonnative TAs; Hinofotis and Bailey (1980) also reported that a group of TESL TA raters judged language proficiency to be the most important factor affecting nonnative TAs teaching ability.  The TA may also have to adjust to cultural differences in the classroom.  These differences may range from the move overt standards of student and teacher behavior, such as the title used to address a teacher, to subtle paralinguistic and nonverbal behavior, such as intonations, gestures, eye gaze patterns, and posture.

Within the past 20 years there has been an increasing interest in the functions of language (e.g., Halliday, 1970).  This shift in interest to the purely structural aspect of language (e.g., grammar) to the functional aspects (e.g., how to make a polite request, how to ask a question) has also been carried over to the area of second language teaching.  By this shift in emphasis, the importance of acquiring communicative competence as well as linguistic competence is recognized; that is, the ability to use language in the right way, at the right time, in the appropriate style, etc., to achieve a communicative purpose (Hymes, 1974) as well as the ability to use the correct grammatical forms.  Moreover, linguistic competence does not imply communicative competence.

In the United States, one of the functions of a teacher is to answer student questions.  But what happens when the teacher is a nonnative speakers of the language in which he teaches?  Do nonnative speaking teachers use the same strategies as native speaking teachers in answering student questions?  One particular feature of teacher answers, repetition, is dealt with in this paper.

The present study investigates TA speech within the structure of student-initiated question sequences in classes taught by American and Taiwanese teaching assistants.²  A question is defined as an utterance that functions as a request for an answer, response, or explanation.  A question sequence is an exchange which contains a question, e.g., question-answer-acknowledgement.   This particular type of interaction sequence was chosen because it is the least predictable from the point of view of the TA.  Krashen (1978, 1980) and Ochs (1979) have argued that second language learners show different levels of competence when they can plan what they will say as compared to utterances produced when they cannot plan or practice the grammatical forms in advance.  Ochs (1979) gives a broad definition of planned discourse as "discourse that has been thought and organized (designed) prior to its expression" (p. 55), whereas unplanned discourse is broadly defined as "discourse that lacks forethought and organizations preparation" (p. 55).

Hinds (1976) proposed a scale of spontaneity for various types of discourse, with "spontaneity being characterized as a lack of editing and preparation" (p. 55).  Conversations, on one end of the scale, are generally completely spontaneous, whereas prepared speeches, on the other hand, are devoid of spontaneity.  Most types of discourse may and do contain varying degrees of spontaneity.  This variation is especially evident in teaching, where some parts of the lesson may be prepared word for word and others may be completely spontaneous.

By virtue of his role as a teacher, the TA can plan in advance and control his lecture, the examples he gives, and the questions he asks his students.  However, the TA cannot know in advance exactly what questions his students will ask nor what form these questions will take, so he cannot plan the exact grammatical form of the answers.  Yet student questions are very frequent in American university classroom interaction.  In this situation, the TA must use all of his linguistic and communicative competence to understand the questions and to answer them.  Unplanned discourse, therefore, provides a better medium for the expression of communicative competence.

Methodology

The subjects in this study are male American and male Taiwanese teaching assistants in the mathematics department of a large United States university.  Taiwanese were chosen because they were the second largest group of international students in the United States in 1982 (TESOL Newsletter, February 1983).  Consequently, a study comparing Taiwanese and American TAs has wide application.

Subject selection was limited to the mathematics department in order to limit the influence of subject matter on teaching style.  Most TAs in math teach recitation sessions: they meet with a group of approximately 20 - 30 students once a week to go over homework, review, answer questions, and in general supplement the large group lectures the students receive from another instructor.  Because answering student questions is one of the major TA activities, it is an appropriate vehicle for obtaining unplanned speech.

This study has a total of four subjects.  Two of the subjects were native speakers of Chinese from Taiwan; two were native speakers of American English.  The students described are undergraduates enrolled in the classes the TAs were teaching.  Each Taiwanese subjects was taped twice for a total of four hours of videotaped data.  Each American subject was videotaped once for a total of two hours of videotaped data.  All videotaping was dine in black and white using Scotch U-Matic 3/4 inch videocassettes which have a total length of 60 minutes recording time.

The data collection yielded six 60-minute videotapes and accompanying audiotapes.  The investigator first transcribed the verbal portion of the data from the audiotapes, then used the videotapes as a check on the verbal transcription.  The transcription was checked again with each audiotape.  AS an added check on the accuracy of the transcription, a graduate student with transcribing experience, unaware of the specific purposes of the study, also checked the transcriptions with the audiotapes.  These transcripts were then used as the actual data for the analysis.

The transcription system used in this study is taken and adapted from those used by researchers in the sociology of language.  Standard English orthography and punctuation are used.  In cases where the subject's pronunciation is unusual, a spelling is approximated to reflect it (e.g., numerate for numerator).  A partial word is represented as closely as possible (e.g., k- can you).  The pause fillers are represented by a and hm.  The basic transcription for the verbal portion of the data is taken from a system developed primarily by Jefferson (reprinted in Schenkein, 1978, pp. xi-xvi).  Speakers are not identified by name; TA speech is denoted by T and student speech is denoted by S.  Ss are not differentiated except when they speak simultaneously or directly after one another with no intervening T speech; in this case the second student speaker is denoted by S2.³

Due to the small number of subjects, no statistical analyses were done on the data.  Percentages of occurrences were determined for each subject and for each group as indicative of possible tendencies in the behaviors of the subjects.  This study represents a beginning step in the analysis of adult classroom interaction comparing native and nonnative TAs.  Therefore, at this point, it is much more useful to describe qualitatively the various behaviors exhibited rather than to do a strictly quantitative analysis when it is not known what the determining factors are that influence the interactants' behaviors.

Results

The following hypothesis was tested in this study: Taiwanese TA answers will exhibit more repetition than American TA answers.  This hypothesis was supported.  For Taiwanese, the percentage of answers with repetition was 78.0% (32/41); for Americans, the percentage of answers with repetition was 32.8% (21/64).

In classifying TA answers, first a distinction must be made between the answers to student requests that the TA do a specific problem and other questions.  In answer to the former, the TA begins to do the problem.  This is the only appropriate kind of answer under normal circumstances.  But to the latter question type, there are more possible ways to answer the question, and it is in this area that differences between Taiwanese TA and American TA behavior may be found.

TA answers were coded as either elaborated or non-elaborated.  An elaborated answer essentially contains an explanation that goes beyond the mere repetition of lexical items in the question.  The following is an example of an elaborated answer.

A non-elaborated answer contains repetition of lexical items in the question, as in the following example.

For TW1 (Taiwanese subject 1), seven of his answers contain partial repetitions, as in TW1,2,7 and TW1,2,4.

In these examples T repeats the last part of the question before he goes on with his answer.  Even when T does not repeat the content of the question, he self-repeats, as in TW1,2,6.

In only two cases TW1 begins his answer without other repetition or self-repetition, as in TW1,2,2.

Even in this example, T first repeats the zero for cri- critical value? as a request for confirmation before he actually begins an elaborated, non-repeated answer.

TW2 also uses repetition, but it is other repetition, as in TW2,2,7-9.

This is an important strategy on TW2's part.  By other-repeating he first of all keeps the conversation going; T lets S know that he still has not understood the question, prompting S to elaborate until T does understand.  T finally does begin an elaborated answer when he has understood the question by saying Yeah, that's a good question.  Let's see...  Even this last statement is a set phrase that keeps him talking while he formulates his answer. 

T other-repeats twice in this exchange.  These repetitions illustrate another one of their functions: T's repetition lets S know if T has understood what S said and gives S a chance to confirm or deny if the repetition was correct.

TW2 does exhibit non-repeated, elaborated answers, as in TW2,2,6.

Even in this example, T uses a pause filler to gain time to encode his answer.

A1 (American subject 1) exhibits six examples of other repetition.  In A1,16 he repeats part of the problem as he writes it on the board so S can confirm or deny it before he goes on with his question.

In two other cases T repeats part of the question but incorporates it into the syntax of his fluent answer.

In A1,19 T self-repeats by finishing a sentence when S had become nonfluent in his question.

In another very interesting example, T actually paraphrased S's question without repeating it as a concession to S that he has understood his question.

Most of A1's answers (76.2%), however, are not repeated and elaborated, as in A1,2.

Here, even though T does not repeat, he does gain time for encoding by using the initial pause filler and a second one before the word exponents.

A2 shows one example of other-repetition initially as a request for confirmation.

In A2,3, T finishes a student's utterance by beginning his answer with part of what S said.

A2 also incorporates part of S's question into the syntax of his answer, as in A2,10.

On the other hand, 55.2% of A2's answers are elaborated and non-repeated, as in A2,11.

Conclusions

The results of this study indicated that Taiwanese TA answers showed a greater use of self-repetition and other-repetition than American TA answers.  AS has been reported in previous literature (Huang & Hatch, 1978; Scarcella & Higa, 1981; Wagner-Gough, 1978), the use of repetition, especially other-repetition, may be an important strategy for the nonnative speaker.  By repeating part of the question, he gains time to encode his answer; repeating is easier than saying something new.  It also keeps the conversation going.  Furthermore, by repeating, T can ask S if that is what he said or meant; that is, T may repeat if he is not sure what S said, giving S a chance to confirm, clarify, or deny the correctness of the repetition.

It was also noted that when Taiwanese TAs other-repeated, they repeated a word or a phrase of the student question, then they began their answers as new sentences.  When American TAs repeated part of the student question, they incorporated it into the syntax of their answers in most cases.  This occurrence supports the explanation that nonnative speakers use other-repetition to encode and to make sure they understand what the other person said; native speakers may also use that strategy (e.g., A2,1), but they use it less frequently, reflecting their greater English language ability in comprehension and production.

Obviously, not all second language learners show the same level of ability in that second language.  One of the two Taiwanese TAs in this study, TW1, had a lower level of English language competence compared to TW2.  This difference was evident not only from scores on the Michigan Test and teaching evaluation,⁴ but also from subjective impressions.  For example, TW1 often self-repeated, giving his speech a stammering quality: while explaining, he inserted the word okay after every few words.  He frequently seemed to be grasping for words and he used a simplified vocabulary.  He often had difficulty comprehending student speech.  Although TW2, compared to native speakers, spoke a little more slowly and sometimes took more time to comprehend student speech, he was approximately native-like competence much more closely than TW1.

In addition to the above subjective impressions, the two Taiwanese subjects showed differences in some of the areas investigated in this study.  Although both TW1 and TW2 used other-repetition, TW1 was the only subject who used self-repetition through both classes, not only when he was answering questions.  He repeated words or parts of words to correct mispronunciations and sometimes repeated whole phrases and even sentences, reflecting his limited vocabulary and lack of grammatical control by his inability to paraphrase.  TW2 showed no examples of self-repetition in answering student questions, which most likely reflects his higher level of English language competence.

This study is, of course, only a beginning, and any conclusions drawn from it can only be tentative hypotheses which must be tested more thoroughly.  The number of subjects was quite small, and individual factors may have affected their teaching performance.  For example, A1 and A2 had been teaching for six and eight years, respectively.  TW1 had been teaching or two years in the United States and had taught for two years in Taiwan; he had been in the United States for three years.  TW2 had been teaching in the United States for two years, had never taught in Taiwan, and had been in the United States for five years.  A1 had not taught that level recitation course previously; A2 had taught the lecture course twice before.  TW1 had not taught that level recitation course previously; TW2 had taught his recitation course twice previously.

Thus it is not the case that repetition is necessarily a mark of someone with lower language ability.  On the contrary, the native speakers used repetition for confirmation (Is that what the S really asked?).  Repeating a part or the whole student question or paraphrasing is also a good teaching strategy.  It may even be a most necessary strategy if the room is large or if the student asked his question in a quiet voice.  Therefore we can say it is not the use of repetition that marks the nonnative speaker, but the kind and manner of the use of repetition.

Implications for TA Selection and Training

This study was conducted in order to provide some answers and recommendations for the foreign TA problem.  As was stated previously, with the greater number of international students in United States universities, the answer does not lie in preventing international students from teaching, but in establishing criteria and guidelines for evaluating the classroom communicative competence of foreign students who would be TAs.  The initial steps have already been taken by many universities, with each university choosing or developing its own test(s) and cut-off points.

Clearly, English language competence should be the most important criterion in evaluating nonnative TAs.  Without a good command of the language, no teacher can be effective.  It is not the purpose of this study to say exactly what level should be attained, but it must be high enough to handle classroom interaction adequately.

TA evaluation was in its third year at the university at which this study was conducted.  The recommendations made by the Center for English as a Second Language for each TA evaluated were still not binding on departments in their TA selection.  Consequently, one of the subjects in this study, TW1, should not have been teaching according to his evaluation.  We are fortunate, however, to have had him as a subject, along with another Taiwanese TA with a positive recommendation and the two American TAs, for the comparison has proven to be extremely enlightening.

Results of this study suggest that prospective TAs be given practice in how to respond to student questions.  First, the TA must be able to understand the question; if he does not, he must find a way to understand it.  Both TA and student have to negotiate the meaning.  The TA should not do in the classroom what he might do in an out-of-class situation--guess the meaning and hope his response is correct, putting the responsibility for clarification on his interlocutor.  This behavior is inconsistent with the role of a teacher, who is expected to lead classroom discussion and activity.  Many students may be unwilling or unable to persist in their quest for information if the teacher has misunderstood and does not answer their question.  The teacher is in the role of knower: if a student question is unclear, the TA must take the lead in attempting to clarify the meaning.

Simple repetition may be effective for small bits of information (e.g., page seventy-five), in seeking confirmation or in thinking aloud before beginning an answer (e.g., times x to the third, let's see...), but it is of little value when more complex questions are involved.  Here the teacher needs to know just what it is he does not understand (Miyake & Norman, 1979).  He can repeat or paraphrase what he thinks he understands for confirmation, so the student is free to clarify or continue.  Or the TA can ask the student a further question.  This study did not look at paraphrase, but practice with paraphrase can only help the TA as he later searches for the meaning of student questions.

Another portion of this study not presented here reports that students asked Taiwanese TAs fewer questions (41 questions in 4 hours of data) than they asked American TAs (64 questions of 2 hours of data).  Barring the unlikely possibility that the students in the Taiwanese TA classes were somehow different, the results suggest that the students were unwilling or unable to communicate effectively with their nonnative As (the tapes were made between the fourth and eighth weeks of a ten-week term).  Perhaps the TA was unwilling to answer questions because of his previous expectations of the role of teacher as lecturer.  Or, more likely, the TA may have attempted to answer a question he misunderstood, and any negotiation of meaning quickly became tense and frustrating for both TA and students.  In this case, the students may have just given up.  This seemed true for TW1's classes.  TW2, on the other hand, had much more interaction with is students; he also seemed more self-confident and less afraid of them.

Presumably, these prospective TAs are competent in their subject areas (this is not for the English teacher to determine), but students may mistakenly interpret difficulty with English as incompetence.  If on the first day of class the TA tells his students something about himself and his educational and work experience, without unduly praising himself, and encourages questions about himself and his country, he may be initiating a pattern of more open communication for the rest of the semester.  Furthermore, if the TA honestly admits to the students that, because he is not a native speaker, he may sometimes make mistakes in English, and, if from the beginning, he asks students to help if his English gets muddled, most students will respond favorably.  In short, if the TA can admit to himself and to his students: "I know math better than you do, but you know English better than I do, so let's help each other," he is well on his way to becoming an effective communicator.

Although other variables (e.g., personality, age) certainly have an effect on teacher and student behaviors, the simple most important factor affecting nonnative TA classroom behavior is English language competence.  Certainly it is useful to give TAs practice in the use of audiovisual aids, effective use of the board, and even practice in the various verbal acts, e.g., giving a homework assignment, introducing and summarizing material.  These techniques will augments any teacher's effectiveness.  But without an adequate level of English language competence, these techniques will be ineffective.  Techniques are only tools.  It is still the teacher who must lecture, explain, and answer questions, and these activities require the use of the verbal channel.

Implications for Teaching in Taiwan

What do the results of the above study mean for us, the university teachers in Taiwan?  Our students, especially those in the sciences who choose graduate study in the United States, will find themselves in this situation.  Those graduate students who are asked to teach are typically not given lecture classes, but science labs, problem sessions, and similar activities in which they must communicate orally with students.  This situation could not emphasize more the practical importance of listening and speaking skills for science majors.

Unfortunately, many students come to the university with little experience in speaking skills and a fear and dislike of speaking English.  They may think it is unimportant.  After all, the TOEFL and GRE examinations do not test spoken English.  The rude awakening comes when they reach the United States and find that they are tested quite critically on their speaking skills.  With more and more universities concerned with the so-called "foreign TA problem," spoken English ability can make the difference in whether or not the student receives financial aid in the form of a teaching assistantship.

After freshman English, we seldom see any but the most ambitious students, and freshman English has other goals to meet.  Nevertheless, if these prospective graduate students can be made aware of the need for communicative competence in English at United States universities, perhaps they will include improvement of speaking skills among their educational goals.

Notes

1. This is a revised version of a paper presented at the TESOL Summer Meeting, Georgetown University, July 12, 1985.

2. This research is part of a larger study, Student initiated question sequences in classes taught by American and Taiwanese teaching assistants. Unpublished doctoral dissertation, Pennsylvania State University, 1984.

3. Other transcription notation used in this paper is presented below.  These conventions are presented in Jefferson (reprinted in Schenkein, 1978, pp. xi-xvi).

--Overlapping utterances

When overlapping utterances do not start up simultaneously, the point at which an ongoing utterance is joined by another is marked with a single left-hand bracket, linking an ongoing with an interrupting utterance at the point where the overlap begins:

TOM: I used to smoke a lot

                                   [

BOB:                           He thinks he's real tough.

The point where overlapping utterances stop overlapping is marked with a singe right-hand bracket:

TOM: I used to smoke a lot more than this

                                   [         ]

BOB:                            I see

--Contiguous utterances

When there is no interval between adjacent utterances, the second being latched immediately to the first (without overlapping it), the utterances are linked together with equal signs:

TOM: I used to smoke a lot=

BOB: =He thinks he's real tough

--Characteristics of speech delivery

Double parenthesis are used to indicate details of the conversational scene:

JAN: This is just delicious

          ((telephone rings))

KIM: I'll get it.

--Transcriptionist doubt

When single parentheses are empty, no hearing could be achieved for the string of talk or item in question:

TODD: My (            ) catching

(        ):  In the highest (                 )

where the middle of Todd's utterance, the speaker of the subsequent utterance, and the end of the subsequent utterance could not be recovered.

4. To obtain an approximation of the English language proficiency of the Taiwanese subjects, results of a test given to nonnative speaking TAs were obtained.  Scores on this test included the Michigan Test of Aural Comprehension and a rating of two TA evaluators of the subjects' 10-minute teaching demonstration.  Both subjects were tested in September 1982, approximately seven to eight months before the videotapes used in this study were made.

The first Taiwanese subject (TW1) received a scaled score of 53 on the Michigan Test of Aural Comprehension, a score which is termed "Not proficient enough in English to take any academic work" (Michigan Test of English Language Proficiency, Norms and Interpretation Manual, p. 12).  In the evaluation of his English language skills for the teaching situation, on a 5-point scale--Poor, Fair, Good, Above Average, Excellent--he was judged by two trained TA raters to fall into the following categories:

(1) Comprehension--Good: Understands at slower than normal speed; frequent repetitions necessary.

(2) Pronunciation--Good: Concentrated listening is necessary; errors cause occasional misunderstanding.

(3) Fluency--Fair: Usually hesitant, forced into silence by language problems.

(4) Grammar--Fair: Comprehension difficult; frequent rephrasing; uses basic patterns.

The evaluators recommended that he needed further work in pronunciation, comprehension, fluency, and grammatical control.  In Fall Term 1982 he enrolled in a course for nonnative TAs, and in November 1982 he was evaluated using the TSE (Test of Spoken English, SPEAK).  On this test he scored 130--"generally not comprehensible due to frequent pauses and/or rephrasing, pronunciation errors, limited grasp of vocabulary, and lack of grammatical control" (Guide to Speak, p. 7), and he was determined not to have sufficient control of English to teach.   (Individual academic departments made the final decision at that time who was to teach, despite recommendations from the Center for English as a Second Language).

TW2 received a scaled score of 71 on the Michigan Test of Aural Comprehension.  Scores between 65 and 79 are recommended "May take up to 1/2 the normal academic load plus a special intensive course (10 hrs. per week) non--credit, in English as a foreign language (Michigan Test of English Language Proficiency, Norms and Interpretation, p. 12).  In his teaching evaluation, on the same 5-point scale reported for TW1, TW2 was judged by two trained TA raters to fall into the following categories:

(1) Comprehension--Excellent:  Appears to understand everything without difficulty.

(2) Pronunciation--Excellent:  Has few traces of foreign accent.

(3) Fluency--Excellent:  Speech as fluent and effortless as that of a native.

(4) Grammar--Excellent:  Few, if any, noticeable errors of grammar and word order.

The evaluators recommended that he be allowed to teach without any stipulations.  Consequently, TW2 was not required to take any further English tests, so there was no TSE score for him.

References

Bailey, K. M. (1982).  Teaching in a second language: The communicative competence of non-native speaking teaching assistants. Unpublished doctoral dissertation, University of California, Los Angeles.

Guide to SPEAK.  (1982).  Princeton, NJ:  Educational Testing Service.

Halliday, M. A. K.  (1970).  Language structure and function.  In J. Lyons (Ed.), New horizons in linguistics (pp. 140-165).  Baltimore: Penguin Books.

Hinds, J.  (1979).  Aspects of Japanese discourse structure.  Tokyo: Kaitakusha.

Hinofotis, F. B., & Bailey, K. M.  (1981).  American undergraduates' reactions to the communications skills of foreign teaching assistants.  In J. C. Fisher, M. A. Clarke, & J. Schacter (Eds.), On TESOL '80: Building bridges (pp. 120-133).  Washington, DC: TESOL.

Huang, J., & Hatch, E.  (1978).  A Chinese child's acquisition of English.  In E. Hatch (Ed.), Second language acquisition: A book of readings (pp. 118-131).  Rowley, MA: Newbury House.

Hymes, D.  (1974).  Foundations of sociolinguistics: An ethnographic approach. Philadelphia: University of Pennsylvania Press.

Krashen, S. D.  (1978).  The monitor model for second language acquisition.  In R. Gingras (Ed.), Second language acquisition and foreign language teaching (pp. 1-26).  Arlington, VA:  Center for Applied Linguistics.

Krashen, S. D.  (1980).  Lecture notes.  TESOL Summer Institute, University of New Mexico.

Landa, M., & Perry, W.  (no date).  An evaluation of an ESP training course for foreign teaching assistants. University of Minnesota.

Largest number ever of foreign students.  (1983).  TESOL Newsletter, 17(1), 1, 10.

McKay, J.  (1978). Final Report, ESL 350 Teaching Workshop, 1977-1978.  Berkeley, University of California.

Mestenhauser, J., Perry, W., Paige, M., Landa, M., Brutsch, J., Dege, D., Doyle, K., Gillette, S., Hughes, G., Judd, R., Keye, X., Murghy, K., Smith, J., Vandersluis, K., & Wendt, J.  (1980).  University of Minnesota Students' Advisor's Office and Program in English as a Second Language: Report of a special course for foreign teaching assistants to improve their classroom effectiveness.  Minneapolis: University of Minnesota.

Michigan Test of English Language Proficiency, Norms and Interpretation.  (1977).  English Language Institute, University of Michigan.

Miyake, N., & Norman, D. A.  (1979).  To ask a question, one must know enough to know what is not known.  Journal of Verbal Learning and Verbal Behavior, 18, 357-364.

Ochs, E.  (1979).  Planned and unplanned discourse.  In T. Givon (Ed.), Syntax and semantics 12: Discourse and syntax (pp. 51-80).  New York: Academic Press.

Scarcella, R. C., & Higa, C.  (1981).  Input, negotiation, and age differences in second language acquisition.  Language Learning, 31, 409-437.

Schenkein, J.  (1978).  Explanation of transcript notation.  In J. Schenkein (Ed.), Studies in the organization of conversational interaction (pp. xi-xvi).  New York: Academic Press.

Wagner-Gough, J.  (1978).  Comparative studies in second language learning.  In E. Hatch (Ed.), Second language acquisition: A book of readings (pp. 155-171).  Rowley, MA: Newbury House.