You are on page 1of 188


To print, select PDF page nos. in parentheses

Beyond the Language Classroom: A Study of Communicative Abilities in Adult Immigrants (9-29) Following Intensive Instruction 185 Alison d'Anglejan, Gisle Painchaud, and Claude Renaud Pronunciation Revisited 207 (32-49) Martha C. Pennington and Jack C. Richards English for Specific Purposes: Content, Language, 227 ( 5 1 - 6 9 ) and Communication in a Pharmacy Course Model Janet G. Graham and Robert S. Beardsley Coherence and Academic Writing: (72-89) Some Definitions and Suggestions for Teaching 247 Ann M. Johns The Meaning and Discourse Function of the (91-110) Past Tense in English 267 Elizabeth Riddle Dont Put Your Leg in Your Mouth: Transfer in the Acquisition of Idioms in a Second Language 287 (111-128) Suzanne Irujo Information Gap Tasks: Do They (129-149) 305 Facilitate Second Language Acquisition? Catherine Doughty and Teresa Pica

The Computer Book 327 Mohyeddin Abdulaziz, William Smalzer, and Helen Abdulaziz Reviewed by Jack Kimball Review of the State-of-the-Art of Educational Technologies Implemented in Programs Serving LEP Students Funded by the 331 Department of Education COMSIS Corporation Reviewed by Richard E. LeMon


Recognition of Sentences by Native and Nonnative 335 Speakers of English: Probing the Role of Imagery Patricia Dunkel, Shitala Mishra, and Alfred Stover 338 Dominant Administrative Styles of ESL Administrators Alfred W. Reasor

The Effects of Modifying the Formality Level of ESL Composition Questions 343 Keiko Hirokawa and John Swales

Women Students in Our Institutions: A Response to the U.N. Decade for Women (1976-1985) 347 Linda A. Moody Comments on Bernard A. Mohan and Winnie Au-Yeung Los Academic Writing and Chinese Students: Transfer and Developmental Factors 354 A Reader Reacts Joan Gregg Response to Gregg On Evidence for Cross-Cultural Rhetoric Bernard A. Mohan Information for Contributors 363 Editorial Policy General Information for Authors Publications Received 367 Publications Available from the TESOL Central Office 371 TESOL Membership Application 384


In This Issue
s The articles in this issue of the TESOL Quarterly address a variety of topics, including the effect of intensive language instruction on the communicative ability of adult immigrants; specific issues in the teaching of pronunciation, grammar, and academic writing; the role of first language transfer in the acquisition of idioms in a second language; the contribution of particular types of student-directed activities in shaping classroom discourse; and the potential for a communication course based on principles of English for specific purposes and content-area ESL teaching to meet general and specialized English language learning needs. Together, these articles reflect our fundamental commitment to base changes in classroom instruction and program design on new perspectives on the nature and uses of English, on a fuller understanding of second language development, and on a more precise specification of the language needs which learners face in studying, working, and living in English-speaking communities.
In examining the ability of adult immigrants to communicate in an interview situation as they were completing a 30-week, 900-hour second language program and then 6 months later, Alison dAnglejan, Gisle Painchaud, and Claude Renaud explore a number of important issues in adult second language program development and policy. Although the subjects in this study were learners of French as a second language, the authors findingsincluding the progress of their subjects during the intensive language program, the relationship between the learning which took place in the program and continued development during the posttraining period, and the effect of such factors as high unemployment and limited social contact with native speakers on language skills improvementhave challenging implications for the development of policy guidelines for all adult immigrant language training in urban North American settings and elsewhere.

Arguing that the view of pronunciation embodied in traditional approaches to language teaching trivializes its true nature and that recent changes in perspective in TESL/TEFL methodology have produced uncertainty about [its] role, Martha Pennington and Jack Richards review current perspectives on pronunciation to show that it must be seen not only as part of the system for expressing referential meaning, but also as an important part of the interfactional dynamics of

the communication process, involving a complex interaction of perceptual, articulatory, and interfactional factors. Conceding that in the domain of pronunciation . . . there is not likely to be a one-to-one relationship between teaching and learning, the authors provide five guiding principles for the teaching of pronunciation in the second language classroom and offer an agenda for research which will help determine more clearly the potential and limitations of classroom instruction in this area.

Janet Graham and Robert Beardsley describe an experimental course in communication which they designed for nonnative English-speaking pharmacy students. The course was designed to combine principles of content-area ESL and English for specific purposes, with an analytic syllabus developed from a multidimensional needs analysis. Activities were planned to focus on skillful communication and increased fluency rather than on formal correctness. While evaluation of the course, in which a total of 10 advanced students participated, revealed a number of areas for improvement, the course was viewed overall both by the students and the authors as quite successful. Graham and Beardsley suggest that similar courses could be designed for other health preprofessionals and professionals, as well as for specialists in other occupations, and that the addition of content-area instruction to an ESP course may be a productive response to the demand for relevant and effective English language instruction. Ann Johns explores the problem of teaching ESL students the concept of coherence in academic writing. Her review of the literature points out that this complex concept, which is often discussed in vague terms with students, is both text based (involving the ordering and interlinking of propositions within a text by use of appropriate information structure) and reader based (requiring the writer to consider the expectations and needs of the reader). The author outlines a three-lesson revision unit incorporating a top-down approach to the teaching of coherence: Student writers are led from global considerations, including the reconstruction of the essay prompt and the development of a discourse theme, to increasingly local, text-based and reader-based concerns. The result, Johns suggests, is a learning sequence which may help students produce the kind of coherent writing which is particularly realistic for the academic milieu, but certainly appropriate for any writing task. Elizabeth Riddle argues that a major source of the problem which ESL students have in learning to use the past tense correctly is an inadequate understanding of its actual meaning and discourse functions. The author proposes that the best denotation of the simple past tense may be simply true before speech time in the real world or in the speakers belief world and provides several examples to illustrate the extent to which a speakers point of view and purpose in performing a speech act condition the choice between the present and past tenses. The article concludes with a description of a number of communicative and TESOL QUARTERLY


contextually based exercises which aim at raising students awareness of the past tense as it is actually used in discourse.

Suzanne Irujo reports the results of a study of the acquisition of English idioms by native speakers of Spanish. The study sought to determine the role of transfer in the ability of learners to comprehend and produce English idioms of three kinds: those for which there is an identical equivalent (both in form and meaning) in Spanish; those for which there is a similar, but not identical, Spanish equivalent; and those from which the Spanish equivalent differs considerably. The study found evidence of both positive and negative transferthe latter affecting in particular the ability of her subjects to produce similar idioms. In addition, there were some fairly clear cases of target language strategies. On the basis of her findings, the author offers guidelines for deciding which idioms to teach and suggests activities appropriate for developing students ability to comprehend and produce English idioms. Catherine Doughty and Teresa Pica report the findings of the most recent in a series of studies conducted to determine the effects of task type and participation pattern on language classroom interaction. Noting that efforts to teach second languages within a communicative framework have led to certain methodologically motivated changes in the classroom environment, the authors compared the extent to which teacher-fronted, small-group, and dyadic learning arrangements for the completion of a required information exchange task generated the kind of modified interaction thought to be crucial to successful second language acquisition. In conjunction with findings from their earlier research, the results of this study lead to the conclusion that both group and pair work must be carefully planned by the classroom teacher to include a requirement for a two-way or multi-way exchange of information. Reviews: Jack Kimball reviews Mohyeddin Abdulaziz, William Smalzer, and Helen Abdulazizs The Computer Book, and Richard LeMon reviews the COMSIS Corporations Review of the State-of-theArt of Educational Technologies Implemented in Programs Serving LEP Students Funded by the Department of Education. Brief Reports and Summaries: Patricia Dunkel, Shitala Mishra, and Alfred Stover report the results of a study of the role of imagery in native and nonnative speakers memory for sentences; Alfred Reasor summarizes the results of his study of dominant administrative styles of ESL administrators; and Keiko Hirokawa and John Swales describe the findings of research on the relationship between the formality of task statements and selected features of ESL writing ability. The Forum: Linda Moody discusses the needs of women students in ESL institutions in Women Students in Our Institutions: A Response to 183

Also in this issue:



the U.N. Decade for Women (1976-1985 ), and Joan Greggs reaction to Bernard Mohan and Winnie Au-Yeung Los Academic Writing and Chinese Students: Transfer and Developmental Factors is followed by Mohans response, On Evidence for Cross-Cultural Rhetoric. Stephen J. Gaies



TESOL QUARTERLY, VoL 20, No. 2, June 1986

Beyond the Language Classroom: A Study of Communicative Abilities in Adult Immigrants Following Intensive Instruction
University of Montreal

Douglas Hospital Research Centre, Montreal

The study reported in this article examined the ability of adult immigrants to communicate in an interview situation as they were completing a 30-week, 900-hour second language program and then 6 months later. Foreign Service Institute (FSI)-type interviews were carried out with two cohorts of learners of varying levels of ability. The first cohort was comprised of immigrants from Southeast Asia; the second included Poles and Latin Americans. Multivariate statistical comparisons of the FSI levels at the time of the first and second interviews showed significant gains for both cohorts. However, the two cohorts differed on the subtests which contributed to those gains. Progress by Cohort 1 was mainly attributable to vocabulary gains, while gains in vocabulary, grammar, and comprehension contributed significantly to the progress made by Cohort 2. Descriptive analyses of results for subjects who progressed, regressed, or remained stable over the 6-month period showed a greater tendency to progress among lower level subjects than among more advanced subjects.

The acquisition of a new language is undoubtedly one of the major hurdles which the adult immigrant must face in adapting to a new society. A study carried out in Canada by Mastai (1979) showed that while finding suitable employment ranked as the most critical task confronting the newcomer, success in doing so was largely contingent upon second language skills. The ability of the adult immigrant to adapt to a new environment is thus conditioned by both employment opportunities and opportunities to learn the

language of the workplace. This article explores the communicative skills of two groups of adult immigrants as they completed a period of intensive language training and then 6 months later. In Quebec, the province in which the study was carried out, immigrants are offered particularly generous conditions in which to acquire Frenchthe provinces sole official language and the principal language in the workplace. Under joint funding agreements between the province and the federal government of Canada, special intensive French as a second language courses for immigrants are offered in Centres dorientation et de formation des immigrants (COFIS). Immigrants receive a small stipend while they attend 900-hour courses (6 hours per day for 30 weeks) designed to provide the necessary language skills to enable them to enter the work force. A wide range of instructional approaches and methods are used within the COFI classrooms. Although the European structural approaches are still in widespread use, the most typical classroom tends to be eclectic, combining highly form-focused instruction with more functional approaches such as simulated real-life tasks. For the teaching of structures, teachers rely on commercially available methodsfor example, De Vive Voix (Moget & Neveu, 1975), Dialogue Canada (1974), Le Franais International (Calv, Germain, LeBlanc, & Rondeau, 1973)as there is no prescribed curriculum. More flexibility is shown for the portion of classroom time incorporating content areas generally considered to be functional skills. Emphasis is placed almost exclusively on the spoken language and on the development of language skills per se, with some attention being given to consumer economics, health, community resources, employment-seeking skills, and so on. Since the students entering the COFIS vary greatly in terms of their age, prior schooling, nationality, and occupation, predictably, there is considerable variation in their competence in French upon completion of the COFI program. After 30 weeks of study, some 15% to 20% of the learnersthe figures vary considerably according to the characteristics of the learner groupstill experience considerable difficulty in communicating in French. An in-depth study of the relationship between learner characteristics and achievement in the COFIS (dAnglejan, Renaud, Arseneault, & Lortie, 1981) showed that lower levels of schooling and of nonverbal reasoning ability, greater age, and greater classroom anxiety levels tended to be associated with learning difficulties. These difficulties were exacerbated by the fact that few learners had the degree of social or occupational contact with native 186 TESOL QUARTERLY

speakers which would enhance language learning outside the classroom. Studies carried out in Europe (e.g., Klein& Dittmar, 1979; Meisel, 1980; Perdue, 1982), in the United States (e.g., Schumann, 1976a, 1976b), and in Africa (e.g., Obanya, 1976), to cite only a few, have shown the importance of naturalistic language learning through social interaction, particularly in the case of learners with little formal education (dAnglejan et al., 1981). Learners who cannot draw on advanced literacy skills to further their knowledge of a new language are greatly dependent on social interaction and the availability of enlightened classroom instruction. Our present understanding of the process by which adult immigrants attain adequate language skills to achieve professional mobility is still fragmentary. However, it would appear that both adequate instruction in formal learning situations plus opportunities for interaction with well-disposed native speakers in social or workplace environments are necessary to ensure that immigrant learners will progress beyond a minimal knowledge of the target language toward more advanced levels necessary for securing satisfactory employment. The study described below was designed to explore aspects of the communicative skills of immigrants upon completion of the 900hour COFI program and to reassess these skills 6 months later. More specifically, the study had the following aims: (a) to identify the level of functional competence in spoken French of two cohorts of immigrant learners upon completion of the 30-week COFI program; (b) to determine whether, after a 6-month period following the cessation of instruction, the exit level of French had remained stable, progressed, or regressed; and (c) to compare the communicative ability of the two cohorts of learners with respect to the five performance criteria which constitute the Foreign Service Institute Oral Interview Check List of Performance Factors (Jones, 1979). Our data thus address practical issues such as the impact of the intensive French as a second language (FSL) program, as well as more theoretical issues concerning the stability of selected aspects of classroom-acquired language proficiency and the potential contribution of informal learning environments. METHOD Data Elicitation: The Foreign Service Institute Interview Since our aim was to describe the oral proficiency of our subjects in functional rather than formal terms, we chose as our data-eliciting technique the Foreign Service Institute (FSI) interview (Jones, COMMUNICATIVE ABILITIES OF ADULT IMMIGRANTS 187

1979). The use of this oral interview makes it possible to identify a subjects level of oral proficiency on a scale ranging from 0 to 5. The scale actually captures finer distinctions through the use of midpoints, that is, 0, 1, 1+, 2, 2+, 3, and so on. While the scale provides a global measure of spoken language, the subjective evaluations are based on five weighted constituents of proficiency: accent, grammar, vocabulary, fluency, and comprehension.1 Studies by Bachman and Palmer (1981) and by Clark (1978) have shown that the technique has external and internal validity. While it can be argued (see Olynyk, dAnglejan, & Sankoff, 1985) that the unplanned speech context which characterizes the oral interview may give rise to the most hesitant type of interlanguage speech, we felt that the interview situation is quite representative of the types of demands placed on immigrants when seeking employment or social services. A study by Hinofotis et al. (1982) showed that the subjective ratings elicited by an FSI-type interview did, in fact, reflect underlying development patterns in the use of English morphology as established by statistical counts. Interview Content and Procedures Several ethnic organizations were contacted to establish the appropriateness of the topics to be discussed in the interviews, since subjects were from varied cultural backgrounds. It was agreed that the first interview, to be conducted as subjects entered their final 2 weeks of language training, would focus on the following themes: demographic information (e.g., age, schooling, occupation in the country of origin) and sociolinguistic information (e. g., languages spoken, opportunities for the use of French in the home and in the community, contact with the media). Subjects would then be invited to discuss such topics as their arrival in Montreal, housing, the type of employment they hoped to find, social contacts, and the preparation of regional dishes. Special attention was to be given to any other topic which might be introduced spontaneously by the subject. The second interview, to be held 6 months after the first, would serve to describe the characteristics of the main social and occupational environment in which subjects had spent their time since leaving the language center. In addition, topics such as leisure
1 A description of the scoring procedure based on Jones (1979) can be found in the Appendix.

Since our analyses are based on this version of the test, we will continue to use the term FSI interview, although the term Interagency Language Roundtable has since been adopted.



time, childrens schooling, and the outlook for the future would be introduced. Interviews were conducted in the language centers. The second interview took place after working hours for those who were unavailable during the day. All subjects received a $10 fee to cover transportation costs. Interviews were tape-recorded on a Uher 4000 and averaged about 30 minutes in length, although some were shorter owing to subjects limited ability to communicate. Two university-educated francophone females conducted the interviews. Considerable pretesting was done to establish appropriate procedures. The tape-recorded, randomly ordered interviews were rated by two experienced FSI teachers who underwent prior training in the FSI interview technique. Interrater reliability coefficients (Spearman rho) reached .826 for the first interview and .825 for the second (p < .001); t tests carried out on the mean scores for the two evaluators revealed no significant differences. Subjects When we began our research, the student population in the COFI consisted mainly of immigrants from Southeast Asia, including a large proportion of political refugees. We decided to focus initially on these Southeast Asian learners (Cohort 1). The second group (Cohort 2), tested some 18 months later, was comprised mainly of Polish and Latin American immigrants. All subjects had undergone at least 26 weeks of language training and were volunteers. The principal demographic characteristics of the two cohorts are shown in Table 1 (for a more detailed account of our demographic data, see Painchaud, dAnglejan, & Renaud, 1985).
TABLE 1 Demographic Characteristics of Cohorts 1 and 2



Subjects pretraining placement levels reflected assessments made by COFI personnel on the basis of learners level of education, language background, prior knowledge of French, and so on. However, we were unable to respect these sampling specifications for several reasons. In a number of cases, subjects who had agreed to participate when the proposed research was explained to them by interpreters later declined to be interviewed or simply failed to keep appointments. Moreover, when the second interviews took place, some of the original cohort could not be traced or were unwilling to be interviewed. Thus, in spite of our efforts to explain that the research team was university based, that the results of the interviews were in no way to serve as a verification of the subjects language competence as individuals, and that the usual guarantees of anonymity would be given, the interviews were clearly perceived as threatening by some individuals. Difficulties such as these are fully predictable and in our opinion must be viewed as inherent to any research carried out in similar settings. However, these limitations place obvious constraints on the inferences which can be drawn from our data. Of the 113 subjects who participated in the first interview, 82 were able to be traced 6 months later and agreed to the second interview. Of these, 36 belonged to the first cohort and 46 to the second. 2 As can be seen in Table 3, rapid learners were overrepresented in Cohort 1, while Cohort 2 included a disproportionate number of average learners. The low number of slow learners in both groups can be explained by the fact that in spite of having undergone 850 to 900 hours of FSL instruction, many lacked the second language skills necessary to participate in the interview and may have felt particularly threatened by the research. A Mann-Whitney test on the results of the first interview revealed a significant difference between the global mean scores for the two cohorts (z = 3.23; p < .001). Global evaluations for Cohort 2 were significantly higher than those for Cohort 1. RESULTS AND DISCUSSION The results of the two interviews for the two cohorts are presented in Table 2. At the time of the first interview, subjects belonging to Cohort 1 were rated in almost equal proportions to be at Levels 1 and 2 of the FSI scale. In functional terms, these data suggest that following 30 weeks of language training, 50% of these
2 Although

all 46 subjects in the second cohort were interviewed twice, FSI interview scores are available for 45 subjects only; the recording for 1 subject was inaudible.



subjectsthose whose ratings placed them at Level 2had acquired the minimal knowledge of French necessary for limited functions in a workplace setting. The other 50%those rated Level lhad a knowledge of French barely adequate to fulfill their personal needs. According to FSI criteria, this level is not considered adequate for the workplace. The pattern of results for Cohort 2 is more diversified than that for Cohort 1. At the first interview, 20% of the subjects were rated at Level 1, while the majority (64.4%) were at Level 2. Some 15.6% of
TABLE 2 Distribution of Subjects According to Change in FM Levels Between Interviews 1 and 2



the group of learners had reached Level 3a level considered adequate to participate in most formal and informal conversations on practical, social and professional topics. The aim of the second series of interviews and evaluations was to specify the level of subjects functional knowledge of French following a 6-month interval. Data for Cohort 1, shown in Table 2, reveal that the vast majority of subjects still ranged across Levels 1 and 2, with 1 subject at Level 3. However, it is interesting to note that the percentage of subjects at Level 1 or 1+ declined from 50% to 22.2% whereas the percentage at Level 2 and above increased from 50% to 77.8%. The global mean (the sum of the weighted ratings for the five FSI subscales) also increased significantly: Interview 1, M = 43.00; Interview 2, M = 46.38; t (35) = 3.61; p <.001. A similar pattern of results was found for Cohort 2 with a significant increase in global ratings: M = 51.30 to M = 54.97; t (44) = 6.04; p <.001. Between-group comparisons reveal significantly higher ratings for Cohort 2 than for Cohort 1 with respect to global ratings (z = 3.51; p < .001) and FSI levels (z = 3.18; p < .001). To summarize, results for both cohorts improved significantly over the 6-month period, with a slightly greater increase noted for Cohort 2 than for Cohort 1. These results are not surprising, given initial differences between the groups in terms of their demographic characteristics. We next focused more closely on the nature and direction of the changes which took place over the 6-month period. We divided our subjects into three groups: Category A, those whose functional competence in French increased between the two interviews; Category B, those who showed no change; and Category C, those whose competence actually declined. An examination of the data for Cohort 1 (see Table 3) shows that 42% of the subjects progressed during the 6-month period, whereas nearly 53% showed no change and 5% actually regressed. Of the 9 subjects classed as potentially slow learners, 5 (55.5%) progressed during the postinstruction period. Among the 10 learners classified as average, 3 (30%) registered progress, while 7 of the 17 fast learners (41.2%) progressed. Data for Cohort 2 show similar patterns: 40% of subjects progressed, 58% remained at the same level, and only 1 subject (2%) regressed. A statistical comparison of the two groups with respect to Categories A, B, and C showed no significant difference (z = .03; p, n.s.). As can be seen from the chi-square figures in Table 3, the initial placement groupings established by COFI personnel were 192 TESOL QUARTERLY

not highly predictive of progress made by the two groups during the period following language training. In a further analysis we examined the degree of change during the postinstruction period in relation to subjects ratings at the end of language training. These data, shown in Table 3, suggest that it is relatively easy for subjects to progress from Level 1+ to Level 2; however, the fact that subjects attain Level 2 is not an assurance that their learning has stabilized. Two subjects who had attained Level 2+ by the end of their training were found to have regressed 6 months later. Progress from 2+ to 3 appears difficult to achieve (only 5 subjects made that transition), and the highest levels of
TABLE 3 Distribution of Subjects According to Initial COFI Placement Groupings and Degree of Progress During Posttraining Period

stabilityneither progress nor regressionare to be found at Levels 2 and 2+. Changes over the 6-month period were mainly confined to increases or decreases in half levels, for example, 1+ to 2, with improvement occurring most frequently among students at the lower ability levels. Subjects Occupational Status The second interview provided us with data concerning subjects employment experience during the 6-month postinstruction period. Given the potential importance of opportunities to use French in the workplace in the development of functional language skills, this information was tabulated and is presented in Table 4. The most striking finding is that whereas only 6 members of Cohort 1 had found no employment whatsoever during the 6-month period, the figure for Cohort 2 was 24, a reflection of the severe economic recession which prevailed at that later time. Given the wide range of between-subject and between-group variability in terms of background characteristics and employment experience, no systematic attempt was made to draw inferences
TABLE 4 Distribution of Subjects According to Their Employment Status During the Posttraining Period

Note: Some percentages total to slightly less than 100 due to rounding. Category A = subjects whose competence in French increased between the two interviews; Category B = subjects who showed no change; Category C = subjects whose competence actually declined. 194 TESOL QUARTERLY

about the possible effect of workplace experience on subjects functional competence at the time of the second interview. The vast majority of those subjects who did find jobs were employed in factory or business settings as skilled or unskilled laborers. Many were reticent to discuss the extent to which they used their mother tongue during work hours; most reported only rare opportunities to use French with fellow workers or in social settings. (For a more extensive set of data covering subjects language-related activities during the posttraining period, see Painchaud et al., 1985.) Examination of the FSI Subtests Since the previous analyses indicated that subiects as a whole performed significantly bitter on the second interview than on the first, we decided to look at the changes which had occurred in the five subtests which made up the FSI interview scores: accent, grammar, vocabulary, fluency, and comprehension. There was every reason to believe that there might be considerable overlapping among the five FSI subtests, and we also wanted to identify as unambiguously as possible the locus of the variance in performance between the two evaluations. Therefore, we adopted an exploratory multivariate statistical approach. Table 5 presents the results of the multivariate comparisons of the means for the five tests (Hotellings T2 for paired samples).
TABLE 5 Means and Standard Deviations for the FSI Subtest Scores



For both cohorts, the results showed a significant difference between the sets of ratings for Interviews 1 and 2 (F= 3.01 and 6.78; p <.03 and <.01, respectively). The discriminant function coefficients indicate the contribution of each subtest to the between-test variance. For Cohort 1, the vocabulary subtest accounted for the greatest amount of variance between the two sets of interviews, whereas for Cohort 2 the difference was accounted for by the comprehension, vocabulary, and grammar subtests. The contribution of the other subtests was in each case marginal. We next performed a multivariate analysis of variance (MANOVA 2 x 2) with repeated measures on the variable of interview. The results (see Table 6) showed a significant between-group effect for the subtests from the two interviews: F (5,75) = 9.68, p <.001. The discriminant function coefficients shown in Table 7 indicate that the subtests of vocabulary, comprehension, and accent contributed significantly to the between-group variance. In addition, the MANOVA revealed a significant within-subject main effect for the variable of interview: F (5,75) = 8.94, p <.001. Subjects as a whole were rated significantly higher on the Interview 2 subtests than on the Interview 1 subtests. Finally, the Interview x Cohort interaction was not significant, indicating that the increment in scores between the first and second evaluations was equivalent for the two cohorts. To summarize the results of the multivariate statistical analyses, subjects as a whole were rated significantly higher for Interview 2 than for Interview 1 with respect to the set of five FSI subtests. The vocabulary subtest made the greatest contribution to this growth effect. However, the cohorts performed differently with respect to this increment. For Cohort 1, comprised of Southeast Asian learners, improvement was restricted almost exclusively to the area of vocabulary; Cohort 2 progressed in comprehension, vocabulary, and grammar. Moreover, Cohort 2 performed significantly better than Cohort 1 on the vocabulary, comprehension, and accent subtests. Finally, the groups did not differ significantly in terms of their overall improvement between the two interviews. According to Higgs and Clifford (1981), Level 2 of the FSI scale constitutes for many learners a terminal level of performance beyond which progress is difficult. Once basic communicative needs have been met, incorrect grammatical forms tend to fossilize, even in the case of highly motivated, well-educated individuals. Our data provide some confirmation of this hypothesis: While both cohorts registered some progress during the posttraining period, for the Southeast Asian group (Cohort 1), comprised of relatively less 196 TESOL QUARTERLY

TABLE 6 Multivariate Analysis of Variance on the Results of the FSI Subtests for the Two Interviews

TABLE 7 Discriminant Function Analysis of the FSI Subtest Scores

educated but younger learners, the advance was limited to vocabulary. Cohort 2s improvement was attributable to comprehension, grammar, and vocabulary. However, for both groups, the grammar subtests registered the lowest scores of the five FSI subtests. Since the grammar scores are subject to a relatively heavy weighting (the highest of the five subtests), they play an important role in determining the FSI level. The fact that our subjects had little opportunity to use their language skills in the workplace or in social interactions makes it impossible to assess the potential interaction between this factor and formal training. As Spada (1985) points out, in the absence of finely tuned language-contact assessment instruments and detailed information regarding classroom teaching approaches, it is unlikely that researchers will be able to unravel fully the complex interaction

between the intentional learning resulting from classroom language training and the incidental learning which occurs through contact with native speakers in occupational or social settings. Some Observations From Transcribed Conversations In a detailed qualitative analysis of the transcribed interviews with two Level 1 subjects, we examined the devices used by subjects to express temporal distinctions. It should be noted that the morphology of the French verb system is highly inflected. Level 1 learners frequently omit, fail to mark, or incorrectly mark verbs for tense and aspect. This is particularly true in extended discourse in which subjects attempt to describe or explain personal experiences or events removed in time and space. Subjects rely heavily on calendrical reference and on adverbs such as avant (before), aprs (after), and maintenant (now), often used emphatically in opposition to each other to place events in time. This can be seen in the following example, quite typical of Level 1 subjects, in which little support, or scaffolding (see Slobin, 1981), is provided by the interviewer, who seems unable to follow what the learner is trying to say. The interviewer is inquiring about the availability of Cambodian and Chinese newspapers in Montreal, but the subject shifts the conversation to an entirely new topic involving complex temporal relationships: I: Est-ce quil y a des journaux . . . ? S:

I: S:

Examples such as this are frequent in our transcriptions. They show that while temporal distinctions in some narratives can be successfully established through the use of place names, calendrical references, or adverbs, the interlocutor must bring shared 198 TESOL QUARTERLY

knowledge to the task of understanding. When meaning can be inferred, as in the following example, the interlocutor is able to prop up the narrative by providing ongoing confirmation of meaning in the form of vocabulary, paraphrase, or other speech markers.

In our transcriptions, however, many narratives take place with little scaffolding provided by the interlocutor, an indication that when the conversation is taking place, the meaning is not transparent at the phrase-by-phrase level, even if it can be inferred later from transcripts. The fact that our well-disposed interviewers had difficulty following such narratives is an indication of the difficulties our immigrant subjects might expect to encounter in less benign workplace or social settings. CONCLUSIONS AND COMMENTS The results of this study suggest a number of issues which are discussed in relation to second language program development and policy. The points to be raised concern levels of proficiency, the contribution of formal instruction to postinstructional learning, and policy guidelines. It is both disappointing yet challenging to discover that after 900 hours of formal instruction, the vast majority of subjects have attained proficiency levels which can best be described as minimal. They are not equipped with the language skills necessary to enter into competition with native speakers in the job marketother than in low-status jobs requiring little language. The advent of computerized technology may make such jobs more scarce, and those available may increasingly call for the ability to handle computerized, printed information. In any event, there is evidence that the inability to find appropriate employment on account of inadequate language skills can be a source of considerable dissatisfaction among immigrant workers such as those in our study. The results of our study also give rise to some more positive findings: Whatever learning did take place through extensive instruction served as a foundation for continued development COMMUNICATIVE ABILITIES OF ADULT IMMIGRANTS 199

during the posttraining period. It is significant that the lower level subjectsthose who appeared to derive the least benefit from formal instructionwere subsequently able to make progress through exposure in the course of routine social interactions. This suggests that it might be appropriate to view intensive formal instruction, such as that offered to our subjects, as a useful preparation for later learning in naturalistic settings. Our data show that as a result of the high levels of unemployment which prevailed at the time the study was conducted, few subjects were able to improve their language skills in the workplace; furthermore, they had limited social contact with native speakers. These facts have implications for the development of policy guidelines for language training in urban North American settings. In a recent policy statement, the TESL Canada Association (1982) points out that language training programs for immigrants should be viewed within a framework of continuing education, closely linked with social and occupational development objectives. As a follow-up to preemployment language programs, such as that described in this article, we advocate the continuation of language training in the workplace and the sharing of responsibility for this training by the private sector. The recent scarcity of employment opportunities leads us to suggest the development of even more broadly based structures for the enhancement of language learning by immigrants. These might include vocational training centers, social services, and even a permanent network of native-speaker voluntary associations working in cooperation with language training institutions. The role of these units would be to ensure that immigrants who are outside the job market and who do not have adequate social contacts with native speakers might be provided with other structured opportunities to gain the exposure to English so necessary to sustain their language development. At the same time, a long-term goal should be the encouragement of literacy skills and sensitivity to the role which literacy plays in our society. While the development of an appreciation and understanding of literacy may not have an immediate impact on the employment prospects of the marginally literate immigrant, it may have important, more diffuse implications for second language learning and the acquisition of literacy by children in such families.



Alison dAnglejan is Professor in the Faculty of Education at the University of Montreal. She holds M.A. and Ph.D. degrees in psychology from McGill University. Her published work includes articles on cognitive and social aspects of language acquisition, language attitudes, and bilingualism. Gis1e Painchaud is Associate Professor of Education at the University of Montreal, where she teaches in the Andragogy Department. She is conducting research and has published in several areas, including adult second language acquisition, teacher training, and self-directed learning. Claude Renaud obtained his M.A. in psychology from the University of Quebec at Montreal. He has published several papers on psychological aspects of second language learning among immigrants to Quebec. He is now a Research Associate at the Douglas Hospital Research Centre in Montreal.

d'Anglejan, A., Renaud, C., Arseneault, R. H., & Lortie, A.M. (1981). Difficults dapprentissage de la langue seconde chez limmigrant adulte en situation scolaire [Second language learning difficulties in adult immigrants in a school situation]. Quebec: Centre international de recherche sur le bilinguisme, University of Laval Press. Bachman, L. F., & Palmer, A.S. (1981). The construct validation of the FSI Oral Interview. Language Learning, 31, 67-86. Calv, P., Germain, C., LeBlanc, R., Rondeau, F. (1973). Le Franais international (2nd ed. ). Montreal: Centre Educatif et Culturel Inc. Clark, J.L.D. (Ed.). (1978). Direct testing of speaking proficiency: Theory and application. Princeton, NJ: Educational Testing Service. Dialogue Canada 1, Units 1-15, Units 16-30, Units 31-45. (1974). Ottawa: Commission de la Fonction publique du Canada. Higgs, T. V., & Clifford, R. (1981). The push toward communication. In T.V. Higgs (Ed.), Curriculum, competence, and the foreign language teacher (pp. 57-79). Skokie, IL: National Textbook. Hinofotis, F., Schumann, J., McGroarty, M., Erickson, M., Hudson, T., Kimbell, L., & Scott, M.L. (1982, May). Relating FOPT (FSI) scores to grammatical analysis of the learners speech. Paper presented at the 16th Annual TESOL Convention, Honolulu. Jones, R.L. (1979). The oral interview of the Foreign Service Institute. In B. Spolsky (Ed.), Advances in language testing (pp. 104-115). Arlington, VA: Center for Applied Linguistics. Klein, W., & Dittmar, N. (1979). Developing grammars: The acquisition of German syntax by foreign workers. Berlin: Springer. Mastai, J. (1979). Immigrant adult education: Tasks of adaptation. In Proceedings of the twentieth annual Adult Education Research Conference (pp. 94-99). Ann Arbor, MI: Adult Education Research Conference. COMMUNICATIVE ABILITIES OF ADULT IMMIGRANTS 201

Meisel, J.M. (1980). Linguistic simplification. In S. Felix (Ed.), Second language development: Trends and issues ( p p . 205-231). Tbingen: Gunter Narr Verlag. Moget, M.T., & Neveu, P. (1975). De vive voix. Paris: Didier. Obanya, P. (1976). Second language learning out of school. ITL Review of Applied Linguistics, 31, 15-26. Olynyk, M., d'Anglejan, A., & Sankoff, D. (1985). A qualitative and quantitative analysis of speech markers in the native second language speech of bilingual. Manuscript submitted for publication. Painchaud, G., d'Anglejan, A., & Renaud, C. (1985). Apprentissage du franais par des immigrants adultes au Qubec [The learning of French by adult immigrants to Quebec]. Quebec: Centre international de recherche sur le bilinguisme. Perdue, C. (1982). Second language acquisition by adult immigrants: A field manual. Strasbourg: European Science Foundation. Schumann, J.H. (1976a). Second language acquisition: The pidginization hypothesis. Language Learning, 26, 391-408. Schumann, J.H. (1976b). Social distance as a factor in SLA. Language Learning, 26, 135-143. Slobin, D.I. (1981). Reference to the not-here and not-now. P a p e r presented at the Max Planck Institut, Nijmegen. Spada, N. (1985). Some effects of the interaction between type of contact and instruction on the L2 proficiencv of adult learner. Unpublished manuscript, University of Michigan. TESL Canada. (1982). The provision of ESL training to adults. Six principles toward a national policy. TESL Canada Newsletter, 2, 2-12.

APPENDIX Check List of Performance Factors Foreign Service Institute native accurate uneven even complete TESOL QUARTERLY



Instructions for Use of Check List to Determine S-Ratings 1 2 3 4 5 6 (A)

Procedure: Place in Column (A) the credits to be given for each scale on the Check List. For example: a check mark in position 3 on the Accent scale is given a credit of 2. Add the credits to find the total score. The final S-Rating is to be equated with the total by the following table.

If the total scores by different examiners yield different S-Ratings, an average of the total scores should be used to determine the final ratings. Accent 1. Pronunciation frequently unintelligible. 2. Frequent gross errors and a very heavy accent make understanding difficult, require frequent repetition. 3. Foreign accent requires concentrated listening and mispronunciations lead to occasional misunderstanding and apparent errors in grammar or vocabulary. 4. Marked foreign accent and occasional mispronunciations which do not interfere with understanding. 5. No conspicuous mispronunciations, but would not be taken for a native speaker. 6. Native pronunciation, with no trace of foreign accent.

1. Grammar almost entirely inaccurate except in stock phrases. 2. Constant errors showing control of very few major patterns and

frequently preventing communication.

3. Frequent errors showing some major patterns uncontrolled and

causing occasional irritation and misunderstanding. 4. Occasional errors showing imperfect control of some patterns but no weakness that causes misunderstanding. 5. Few errors, with no patterns of failure. 6. No more than two errors during the interview. Vocabulary
1. Vocabulary inadequate for even the simplest conversation. 2. Vocabulary limited to basic personal and survival areas (time, food, transportation, family, etc.). 3. Choice of words sometimes inaccurate: limitations of vocabulary prevent discussion of some common professional and social topics. 4. Professional vocabulary adequate to discuss special interests; general vocabulary permits discussion of any non-technical subject with some circumlocutions. 5. Professional vocabulary broad and precise; general vocabulary adequate to cope with complex practical problems and varied social situations. 6. Vocabulary apparently as adequate and extensive as that of an educated native speaker.

Speech is so halting and fragmentary that conversation is virtually impossible. Speech is very slow and uneven except for short or routine sentences. Speech is frequently hesitant and jerky; sentences may be left uncompleted. Speech is occasionally hesitant, with some unevenness caused by rephrasing and groping for words. Speech is effortless and smooth, but perceptibly non-native in speed and evenness. Speech on all professional and general topics as effortless and smooth as a native speakers. TESOL QUARTERLY

1. Understands too little for the simplest type of conversation. 2. Understands only slow, very simple speech on common social and touristic topics; requires constant repetition and rephrasing. 3. Understands careful, somewhat simplified speech directed to him, with considerable repetition and rephrasing. 4. Understands quite well normal educated speech directed to him, but requires occasional repetition or rephrasing. 5. Understands everything in normal educated conversation except for very colloquial or low-frequency items, or exceptionally rapid or slurred speech. 6. Understands everything in both formal and colloquial speech, to be expected of an educated native speaker.



TESOL QUARTERLY, Vol. 20, No. 2, June 1986

Pronunciation Revisited
University of Hawaii at Manoa

In this reexamination of the status of pronunciation in language teaching, the traditional phonemic-based view of pronunciation is contrasted with a broader, discourse-based view comprising segmental, voice-setting, and prosodic features. A description of the nature and interaction of these three aspects of pronunciation serves to raise issues which are then reviewed in a survey of research on the acquisition of pronunciation. Central issues are the influence of the first language, the acquisition processes operative in L2 phonology, psychosocial and individual factors, and the role of instruction. A broader focus on pronunciation in the context of discourse is suggested as the emphasis of both second language acquisition research and second language teaching. From this perspective the effects of voice setting, stress and intonation, as well as coarticulatory phenomena, assume greater importance for teaching. Pronunciation should be taught as part of the means for creating both referential and interfactional meaning, and not merely as an aspect of the oral production of words and sentences.

While pronunciation has in the past occupied a central position in theories of oral language proficiency, the view of pronunciation embodied in traditional approaches to language teaching trivializes its true nature. In older methods such as audiolingualism, pronunciation has been largely identified with accurate production of isolated sounds or words, and this view is reflected in more contemporary methods such as the Silent Way. The Contrastive Analysis Hypothesis (Lado, 1957) regards pronunciation as central to second language proficiency, but it likewise largely restricts the domain of pronunciation to the segmental level. The goals of language teaching have changed under the impact of communicative views of language and interactive theories of language learning. Pronunciation, traditionally viewed as a component of linguistic rather than communicative competence or as an aspect of accuracy rather than of conversational fluency, has come to be regarded as of limited importance in a communicatively oriented curriculum. Comprehension-based approaches to teaching 207

such as the Total Physical Response and the Natural Approach deemphasize the need for accurate production in the early stages of second language learning. In addition, the value of instruction in pronunciation has been called into question by the limited success reported for the direct teaching of this aspect of proficiency. Recent changes in perspective in TESL/TEFL methodology have produced uncertainty about the role of pronunciation. In view of this uncertainty, there is a need to assess and clarify the current status of the teaching of pronunciation. This article addresses that need, first by presenting an overview of the nature of pronunciation and its role in spoken language interaction and then by considering issues in the learning of pronunciation and drawing implications for language teaching. THE COMPONENTS OF PRONUNCIATION A number of dimensions of speech are included within the description of pronunciation, or phonology. For most language teachers, pronunciation is largely identified with the articulation of individual sounds and, to a lesser extent, with the stress and intonation patterns of the target language. This reflects the traditional view that pronunciation is primarily associated with the expression of referential meaning and that individual sounds, or phonological segments, are the building blocks for higher level meanings. From the perspective of contemporary research in discourse analysis (Brazil, Coulthard, & Johns, 1980), however, pronunciation is seen not only as part of the system for expressing referential meaning, but also as an important part of the interfactional dynamics of the communication process. According to this view, it is artificial to divorce pronunciation from communication and from other aspects of language use, for sounds are a fundamental part of the process by which we communicate and comprehend lexical, grammatical, and sociolinguistic meaning. Pronunciation involves a complex interaction of perceptual, articulatory, and interfactional factors. In this article, that complex of factors is described in terms of three types of features: segmental features, voice-setting features, and prosodic features. Segmental Features Segmental features are minimal units of sound defined in phonetic terms. Traditionally, the fundamental components of 208 TESOL QUARTERLY

pronunciation are phonemes, and acquisition of the target language phonological system is viewed as mastery of the phonemic distinctions embodied in its phonological inventory and of the phonetic variants of phonemes which occur in particular environments within syllables and words. Linguistic theory has shifted toward viewing sound segments in terms of distinctive features (the underlying minimal components comprising speech segments). There has been a parallel shift in speech perception toward dynamic top-down approaches to language processing (those which work from global to local meaning) rather than the static bottom-up model of perception (those which work from local to global meaning) seen in earlier models of speech processing (Dirven & Oakeshott-Taylor, 1984). Nevertheless, language teaching has continued to adhere to the traditional emphasis on phonemes as the principal units of pronunciation. While phonemic and other types of features (e.g., aspiration) that function at the level of individual segments provide a valuable basis for detailed analysis of languages, this kind of micro-perspective on phonology needs to be complemented by a macro-focus on voice-setting and prosodic features. Voice-Setting Features Whereas segmental features refer to specific phonetic characteristics of individual sound segments, voice-setting features refer to general articulatory characteristics of stretches of speech. The tendency of speakers of a particular language to adopt certain habitual positions of articulation in connected speech, resulting in a characteristic voice quality, can be described in terms of voicesetting features. Such features comprise what are sometimes referred to as voice quality, voice quality settings (Esling & Wong, 1983), phonetic settings (Laver, 1980), or certain paralinguistic features (Brown, 1977). Laver (1980) gives an example of such a setting as a quasi-permanent tendency to keep the lips in a rounded position throughout speech. Another would be a habitual tendency to keep the body of the tongue slightly retracted into the pharynx while speaking. Another would be the persistent choice of a characteristically whispery mode of phonation. Settings give a background, auditory colouring running through sequences of shorter-term segmental articulations. (p. 2) Voice quality setting is the phenomenon which accounts for our impressions of, for example, certain male Japanese and Arabic PRONUNCIATION REVISITED 209

speakers as speaking their language (or English) with a hoarse- or husky-sounding voice, or of female speakers from some cultures as speaking with a high-pitched, pinched quality to their voices. This phenomenon has also led some nonnative speakers of English to observe that Americans appear to overuse their lips when speaking. In learning to speak a language, mastery of a characteristic array of voice-setting features appears to contribute substantially to a nativelike accent and possibly to overall intelligibility as well. Prosodic Features and Related Coarticulatory Phenomena The third dimension of pronunciation is stress and intonation, the so-called prosodic, or suprasegmental, domain, together with the related coarticulatory phenomena of the blending and overlapping of sounds in fluent speech. Prosodic features involve the relative levels of stress and pitch within syllables, words, phrases, and longer stretches of speech. Coarticulation causes elisions, contractions, and assimilations of neighboring sounds in the stream of speech under the influence of stress and intonation (Ladefoged, 1982, pp. 52-56, 98). Stress refers to the degree of effort involved in the production of individual syllables or combinations of syllables making up a word or longer utterance. For longer utterances a combination of strong and weak syllables comprises a rhythmic pattern. English, like any language, is spoken with a distinct rhythmic pattern. The ability to produce English with an English-like pattern of stress and rhythm involves stress timing (the placement of stress only on selected syllables), which in turn requires speakers to take shortcuts in how they pronounce words. Natural-sounding pronunciation in conversational English is achieved through blends and omissions of sounds to accommodate its stress-timed rhythmic pattern (Clark & Clark, 1977). Brown (1977) documents the patterns of blends and omissions in conversational speech which can result in change of consonant or vowel quality, loss of consonant or vowel, or even loss of entire syllables. Examples of these phenomena are given below.



Stress and intonation interact with other phonological features and with choices made about the meaning or information conveyed in an utterance. According to Brown and Yule (1983a), stress and intonation mark the elements which the speaker [does or] does not require the hearer to pay attention to (p. 164). Syllables or words which are articulated precisely are those high in information content, while those which are weakened, shortened, or dropped are predictable and can be guessed from context (Dirven & Oakeshott-Taylor, 1984). In every language, characteristic intonation contours carry both referential and affective meaning (Ladefoged, 1982). In their referential function, intonation contours provide an interpretation for a sentence by indicating which part of the information is viewed as new versus known, salient versus less salient, or topic versus comment. Intonation and stress are highly context-dependent, so that the patterns of stress and pitch that characterize isolated words or phrases are typically modified when these words or phrases occur in the context of longer utterances. For example, pitch level tends to be reduced in later parts of a discourse as predictability of information increases. Thus, intonation is an essential component of the prosodic continuity that makes connected stretches of speechas opposed to individually spoken words or syllablescoherent and interpretable by the listener. To interfere with stress, timing, fundamental frequency [and other aspects of prosodic continuity in discourse] usually has more drastic consequences for comprehension than removing the cues of a particular [phonological] segment (Dirven & Oakeshott-Taylor, 1984, p. 333). Certain intonational features distinguish statements from questions or indicate interest, doubt, certainty, and other aspects of the speakers attitude toward the topic or the person spoken to. Brazil et al. (1980) emphasize that intonation cannot be adequately described except in relation to the interaction between speaker, and hearer: We see the description of intonation as one aspect of the description of interaction and argue that intonation choices carry information about the structure of the interaction, the relationship between and the discourse function of individual utterances, the interfactional given-ness and newness of information and the state of convergence and divergence of the participants. (p. 11) The point of view expressed by Brazil et al. can be extended to PRONUNCIATION REVISITED 211

other features of pronunciation as well. Phonological features at each of the levels described above carry a variety of interfactional meanings. For example, a vowel or consonant may be pronounced in a novel or unusual way to achieve a certain effect on the hearer, as for example when an American lisps an s or pronounces the u in tune, due, and soon in the British fashion as [yu]. Similarly, voicesetting features such as lip stretching, creaky voice, or low pitch, when used in an interaction by an American female, may indicate her actual or attempted dominance or may serve to mark her higher social status relative to the person spoken to. The view of pronunciation described above emphasizes that pronunciation in a second language involves far more than the correct articulation of individual sounds. Pronunciation is not simply a surface performance phenomenon but is rather a dynamic component of conversational fluency. When contrasts such as those between accuracy and fluency are made, it is misleading to depict pronunciation as belonging to the domain of the former rather than the latter. The acquisition of the phonology of the second or foreign language involves learning how to produce a wide range of complex and subtle distinctions which relate sound to meaning at several different levels. Articulatory, interfactional, and cognitive processes are all equally involved.

THE LEARNING OF PRONUNCIATION The preceding description of the constituents of pronunciation provides a basis for considering how these constituents interact and shape the processes of phonological development in a second or foreign language. While phonology has not occupied as central a position as syntax in second language acquisition research, some important characteristics of the phonological learning process have been isolated. These include the extent to which the second language phonological system is influenced by the phonological system of the first language, the role of universal acquisition processes in the development of L2 phonology, psychosocial and individual factors, and the context of language learning and use.

The Influence of the First Language Language transfer has always been recognized as basic to any theory of second language phonological development (Lado, 1957). The notion of interlanguage acknowledges the role of language 212 TESOL QUARTERLY

transfer (Selinker, 1972), and current views of the nature of interlanguage consider the learners phonological representations as constituting a system intermediate between the native language and the target language (Flege 1980, 1981). Other researchers argue that the phenomenon of transfer extends beyond the level of individual phonemes to include syllable structure (Hecht & Mulford, 1982; Johansson, 1973; Macken & Ferguson, 1981; Tarone, 1980) as well as prosodic and voice-setting features (Esling & Wong, 1983). Faerch, Haastrup, and Phillipson (1984) report the transfer of the following mother tongue intonation patterns in the speech of Danish learners of English: 1. A tendency to pitch the unstressed syllables higher than the preceding stressed ones (the normal pattern in Copenhagen Danish), creating a weaving or lilting impression. 2. Instead of full [intonational] glides (falls, rises, fall-rises), flattening them out (as is the case in Danish) and consequently making them less clearly marked. (p. 125) Gumperz (1982), in his studies of cross-cultural interactions, demonstrates that transfer of voice-setting and prosodic features of the first language can lead to serious intercultural misunderstanding in the target culture.

Acquisition Processes in L2 Phonology Second language acquisition (SLA) research has confirmed that many other processes interact with language transfer in shaping the L2 (second language) phonological system. Some of these acquisition processes are similar to those found in first language phonological development and may be interpreted as a reactivation of first language development strategies. For example, children acquire voiceless consonants before voiced consonants (Macken & Ferguson, 1981), and the same order of acquisition has been observed in second language phonological development, even when the learners native language possesses voiced final consonants (Hecht & Mulford, 1982). Mulford and Hecht (1980) stress the interrelation of transfer and universal developmental processes in determining the particular range of persistent sound substitutions that occur in the acquisition of second language phonology. The course of acquisition, both in terms of rate and order, has been a focus of SLA studies, though most of these studies have PRONUNCIATION REVISITED 213

addressed L2 syntactic rather than phonological development. In first language learning, however, there has been a considerable amount of research on the rate and order of L1 phonological development (Macken & Ferguson, 1981), and L2 phonology needs to be examined from a similar perspective. Another phenomenon cited as a developmental process in first and second language acquisition is simplification (e.g., of syllable structure, as in Tarone, 1978, 1980; however, see Hodne, 1985, and Sato, 1984). Large-scale simplification of the target language by nonnative speakers has been seen as a kind of pidginization (Ferguson, 1971; Schumann, 1978). For second language learning the pidginization model predicts that a learners interlanguage forms will either fossilize at some distance from the target or go through a process of de-creolization to approximate the target phonology over time (Andersen, 1983; Schumann, 1975, 1978). The development of L2 phonology can be viewed as a dynamic process involving cognitive, psychomotor, linguistic, and interactive factors. Markedness theory has been invoked to account for the fact that certain phonological features are more difficult for second language learners to acquire than other features (Eckman, 1977). It has also been suggested that part of the problem of pronunciation is psychomotor. The development of phonological representations, or schemata, for the target language (i.e., second language phonological competence), through opportunities to hear and to speak the language, appears to be a long-term cognitive process which may not develop at the same rate as the corresponding motor skills required for articulation (i. e., second language phonological performance). As a consequence, perception and production may not develop in parallel (Leather, 1983; Neufeld, 1977; Neufeld & Schneiderman, 1980; Sheldon& Strange, 1982). Other features of second language acquisition are similar to processes found in the context of language change and variation. As in the case of first language phonological change, in second language learning a new item or rule is not acquired categorically: Learners do not immediately begin to use a new phonological rule or feature in all its contexts or in all its appropriate phonetic variants. Rather, learners acquire variants of target language features and gradually refine the range of contexts in which the variants are used. At the same time, they add new features to their repertoire, which at an earlier time they may have avoided altogether (Celce-Murcia, 1977). L. J. Dickerson (1975) shows that Japanese learners of English, in 214 TESOL QUARTERLY

acquiring the phoneme /z/, produce a higher percentage of target variants for the phonemes in initial than in medial or final positions. As learning proceeds, they gradually produce more target variants in medial and final positions. W. B. Dickerson (1976), drawing on sociolinguistic variation theory (e.g., Cedergren & Sankoff, 1974; Labov, 1972), describes the process of acquiring new environments for target language rules as a wave mechanism in which rules are learned in specific contexts and then spread throughout the learners interlanguage. Through this wave mechanism, phonological features learned in the context of specific lexical items, phrases, or grammatical constructions may be applied by analogy to additional items, phrases, and constructions. Lacking from the SLA studies, however, has been a thorough examination of interfactional effects in connected spoken discourse, including the universal and language-specific mechanisms for highlighting and de-emphasizing information, and how these affect the acquisition and spread of new phonological features in interlanguage development. Psychosocial and Individual Factors Pronunciation is a central component of face-to-face interaction and is consequently part of the process by which speakers present an image of themselves to others. The concepts of face-work (Goffman, 1972) and language ego (Guiora, Beit-Hallahmi, Brannon, Dull, & Scovel, 1972) may help to explain the fact that phonological features are among the most salient linguistic dimensions used by speakers to create a sense of personal identity. Certain first language phonological features may be consciously retained as markers of ethnic or group identity (Giles, Bourhis, & Taylor, 1977), and so caution should be exercised in regarding [phonological] intrusions simply as instances of interlingual interferences, particularly in the cases of second and third generations of immigrants, as they may often be adopted by them deliberately as ethnic speech markers to establish a distinctive linguistic identity. (Giles, 1979, p. 260) Some learners, wishing to integrate actively into the target culture and to be identified with its speakers, may be motivated to try to attain a native accent in the foreign language. Others, in contrast, may not have a strong integrative motivation toward the target culture and so may consciously or unconsciously seek to maintain a distinctive accent. PRONUNCIATION REVISITED 215

The fact that pronunciation is intimately associated with a persons identity may also explain why considerable individual variation is found in rates and ultimate levels of achievement in phonology (Leather, 1983; Macken & Ferguson, 1981). Personality variables such as introversion, extroversion, or sociability have been suggested to explain differences among individuals in phonological attainment (Busch, 1982). Individual language aptitude (e.g., the ability to mimic sounds) has also been cited as a contributing factor. Leather (1983) reports findings which support the view that it is individual perceptual ability which remains in principle the limiting factor in developing second-language pronunciation (p. 206). The age factor remains an unresolved issue in language acquisition research. Some studies have shown a biological advantage for younger learners (Scovel, 1969; Seliger, Krashen, & Ladefoged, 1975), while others have shown no such advantage (Olson & Samuels, 1973; Snow & Hoefnagel-Hhle, 1977). Lowenthal and Bull (1984) suggest that these contradictory findings reflect varying psychosocial conditions of the testing situations under which data were gathered in the studies. Research findings suggest not a critical period of language development but, rather, that the way in which language is processed can change throughout the course of development (Menyuk, 1978, p. 154). There is evidence that retention of the native accent by adult second language learners may be the result of the use of a speechprocessing strategybased, perhaps, on motor speech perception (Liberman, Cooper, Shankweiler, & Studdert-Kennedy, 1967) established in childhood (Menyuk, 1978; Menyuk & Anderson, 1969). However, adults do not necessarily lose the linguistic capabilities which were present at earlier stages of development (Leather, 1983), and dormant capabilities could possibly be elicited in adults through germane experience (Menyuk, 1978, pp. 156-157).

The Context of Learning and Use The context and conditions for learning and using the language may also affect levels of attainment in phonology. The degree and type of exposure to the second language in classroom and naturalistic settings may in part determine eventual outcomes in phonology. Traditionally, explicit instruction in phonology (e.g., via minimal pair drills) was thought to influence the students ability to articulate new sounds and to improve the learners capacity for self216 TESOL QUARTERLY

monitoring (Acton, 1984; Morely, 1979). Currently, acquisition- or communication-based methodologies do not assign a central role to direct instruction in pronunciation, nor do many bilingual education models, which set the goal as intelligibility rather than native-like phonology. It is assumed in these models that target-like pronunciation will eventually result from interaction with native speakers in naturalistic settings and cannot be achieved through formal instruction. A number of research studies have investigated the effects of instruction on the learning of pronunciation, but the results are inconclusive. While Suter (1976) and Madden (1983) find no positive effect for formal training on achievement in pronunciation, two studies report positive effects for phonetic training of adults. Murakawa (1981) shows that a 12-week program of phonetic training can produce significant changes in the articulation of individual phonemes by adult learners of English. Similar results are reported by Pennington (1984) after six instructional sessions incorporating training in both articulation and listening discrimination. Positive effects on production or perception are also reported for training in prosodic features by Gilbert (1980), Neufeld and Schneiderman (1980), de Bot (1983), and de Bot and Mailfert (1982). Differences in results in the reported studies appear to be due to the great variation in their experimental design, particularly in the type of training which was provided. Phonological performance in the target language is affected by the communicative demands of the situation or task in which the learner is engaged. Some aspects of L2 phonological learning can be viewed from the perspective of information processing, which describes the learning of any complex task or form of behavior as the integration of a number of subskills. Initially, the use of these subskills requires conscious attention, but as learning proceeds, they become routinized and are performed without conscious attention, through what is referred to as automatic processing (McLaughlin, Rossman, & McLeod, 1983). Under certain circumstancesfor example, in a public presentationperformance conditions may inhibit access to automatic processing. In such cases the learner may have to resort to the domain of conscious processing and so plan and monitor speech more closely. Thus, a learners phonological performance may differ in the controlled and automatic modes of processing. In particular, performance may suffer when it must be consciously maintained under stressful conditions. PRONUNCIATION REVISITED 217

IMPLICATIONS FOR TEACHING The model of pronunciation as a context-dependent and dynamic interaction of segmental, voice-setting, and prosodic features has a number of implications for language teaching. This view leads to the articulation of goals different from those set forth for traditional approaches to the teaching of pronunciation. Accuracy at the segmental level is no longer the fundamental aim of teaching, since it is now known that accurate production of segmental features does not in itself characterize native-like pronunciation, nor is it the primary basis for intelligible speech: In view of the fact that segmental information in the acoustic signal may well be of limited scope and reliability, it is of the greatest importance that the learners attention is directed to non-segmental information. (Dirven & Oakeshott-Taylor, 1984, p. 333) As the emphasis moves away from a narrow focus on segments to a broader focus on stretches of speech, the effects of voice setting, stress and intonation, as well as coarticulatory phenomena such as shortenings, weakening, and assimilations, assume greater importance for teaching. This top-down perspective on pronunciation highlights the overarching role of context in determining phonological choices at all three levelssegmental, voice-setting, and prosodic features. Teaching isolated forms of sounds and words fails to address the fact that in communication, many aspects of pronunciation are determined by the positioning of elements within long stretches of speech, according to the information structure and the interfactional context of the discourse as determined by speaker and hearer. The research we have reviewed on the learning of pronunciation also supports a different focus in teaching. Intervention by the teacher may not be able to alter the learners path of development in mastering second language phonology. Learning is a gradual process involving successive approximations to the target language system over time and a progression from controlled to automatic processing. In addition, the learners performance and competence may not develop in synchrony; that is, lower level motor skills may develop at a different rate from their higher level corresponding mental representations. Immediate results from pronunciation training may not be achieved if the learner has not reached an appropriate stage in phonological development and so lacks the developmental 218 TESOL QUARTERLY

prerequisites for what is being taught. Such training may, however, assist in the development of new articulatory habits and contribute to the reorganization of higher level systems, or schemata, eventually resulting in a change in performance. For the same reason, immediate improvements in pronunciation resulting from direct training may take time to become a part of spontaneous language use. In the domain of pronunciation, then, there is not likely to be a one-to-one relationship between teaching and learning. These conclusions support the following general recommendations regarding pronunciation and its place in second language teaching: 1. The teaching of pronunciation must focus on longer term goals; short-term objectives must be developed with reference to longterm goals. 2. The goal of any explicit training in pronunciation should be to bring learners gradually from controlled, cognitively based performance to automatic, skill-based performance. 3. Teaching should aim toward gradually reducing the amount of native language influence on segmental, voice-setting, and prosodic features but should not necessarily seek to eradicate totally the influence of the native language on the speakers pronunciation in the second language. 4. Pronunciation ought to be taught as an integral part of oral language use, as part of the means for creating both referential and interfactional meaning, not merely as an aspect of the oral production of words and sentences. 5. Pronunciation forms a natural link to other aspects of language use, such as listening, vocabulary, and grammar; ways of highlighting this interdependence in teaching need to be explored. DIRECTIONS FOR RESEARCH AND TEACHING Clearly, there is a need for basic research into the nature, learning, and teaching of pronunciation in a second or foreign language. From the time of audiolingualism, a direct relation between listening and pronunciation has been tacitly assumed, though the small amount of research on this basic issue is inconclusive (Leather, 1983). Whether there is a natural developmental sequence for the acquisition of second language phonology, as has been posited for the development of grammar, remains to be explored. PRONUNCIATION REVISITED 219

Information of this kind would be relevant in determining the content and sequence of instruction in the spoken language. Detailed information on the nature and functions of voice quality setting and on the use of phonological features in interaction is also required as a basis for planning materials and instructional strategies. As is true for other areas of second language acquisition and teaching, research in sociolinguistics and spoken discourse analysisof English and of other languagesis yielding important results for phonology and the teaching of pronunciation. This line of research, in helping to clarify the nature of human communication and the linguistic means for achieving specific effects in interaction, is providing a basis on which to develop materials and techniques representing authentic phonological productions in real communication. In addition, work in sociolinguistics and social psychology on the psychosocial processes involved in acquiring new phonological systems should also be consulted. We can design more realistic and effective approaches to teaching spoken language if we understand the variety of social and psychological factors which play a role in the acquisition of a new phonological system. Finally, instrumental analysis (including computer-aided analysis) of phonetic data from English and other languages is providing more specific and detailed phonetic descriptions, thus making it possible to represent and compare with increasing precision the phonological features of languages. This type of data, in combination with the interfactional data provided through discourse analysis, provides essential baseline information needed to make decisions about the content and nature of instruction in second language pronunciation. While a variety of suggestions have been made concerning the teaching of pronunciation (Acton, 1984; Brown & Yule, 1983b; Haycraft, 1971; MacCarthy, 1979; Parish, 1977; Stevick, 1978), too little is known about specific instructional practices to assess their overall contribution to L2 phonological development. There is some evidence (as reviewed above) that training can produce positive effects on pronunciation in a classroom setting in a relatively short period of time. However, it is not clear whether such effects persist over time and carry over to other situations. Controlled studies of what might be achieved through pronunciation training in the context of information structure and interaction (e.g., as advocated in Brown & Yule, 1983b, and Gilbert, 1984) or in the context of real-life psychological and social concerns (e.g., as 220 TESOL QUARTERLY

advocated in Acton, 1984) have not yet been undertaken. Such studies would help us to determine whether these intuitively appealing teaching programs are actually effective in improving the pronunciation of second language learners, especially in adulthood. In order to progress in the teaching of pronunciation, we must have data which fully support claims made for the effectiveness of specific instructional programs or procedures. In particular, the kinds of information needed are (a) clear specifications of the precise aspects of pronunciation being taught, (b) precise descriptions of the instructional procedures used, and (c) valid measures of the effects, positive or negative, of the procedures used. Stricter attention to the question of research design is necessary before the results of particular instructional programs and teaching procedures can be evaluated. With more complete information of this kind, it will be easier to determine not only the relative value of teaching pronunciation as opposed to other aspects of language proficiency but also the amount of attention which should be paid to pronunciation within the context of a whole language teaching program.

Martha C. Pennington, Assistant Professor in ESL, is conducting research in second language phonological acquisition and teaches courses in phonology, ESL methods, and program administration. Jack C. Richards, Professor of ESL, teaches courses in language curriculum development, methodology, and materials design at the University of Hawaii, where he also teaches in an ESL program. His research interests include listening comprehension and conversational analysis. His most recent books are Approaches and Methods in Language Teaching (with Ted Rodgers), published in 1986 by Cambridge University Press, and the Longman Dictionary of Applied Linguistics (with J. Platt and H. Weber), published in 1985.

Acton, W. (1984). Changing fossilized pronunciation. TESOL Quarterly, 18, 71-85. Anderson, R. (1983). Pidginization and creolization as language acquisition. Rowley, MA: Newbury House.



de Bot, K. (1983). Visual feedback of English intonation, I: Effectiveness and induced practice behavior. Language and Speech, 26, 331-350. de Bet, K., & Mailfert, K. (1982). The teaching of intonation: Fundamental research and classroom applications. TESOL Quarterly, 16, 71-77. Brazil, D., Coulthard, M., & Johns, C. (1980). Discourse intonation and language teaching. London: Longman. Brown, G. (1977). Listening to spoken English. London: Longman. Brown, G., & Yule, G. (1983a). Discourse analysis. Cambridge: Cambridge University Press. Brown, G., & Yule, G. (1983b). Teaching the spoken language. Cambridge: Cambridge University Press. Busch, D. (1982). Introversion-extraversion and the EFL proficiency of Japanese students. Language Learning, 32, 109-132. Cedergren, H. J., & Sankoff, D. (1974). Variable rules: Performance as a statistical reflection of competence. Language, 50, 333-355. Celce-Murcia, M. (1977). Phonological factors in vocabulary acquisition: A case study of a two-year-old English-French bilingual. Working Papers in Bilingualism, 13, 27-41. Clark, H. H., & Clark, E. V. (1977). Psychology and Language. New York: Harcourt Brace Jovanovich. Dickerson, L. J. (1975). The learners interlanguage as a set of variable rules. TESOL Quarterly, 9, 401-408. Dickerson, W. B. (1976). The psycholinguistic unity of language learning and language change. Language Learning, 26, 215-232. Dirven, R., & Oakeshott-Taylor, J. (1984). Listening comprehension (Part 1). State of the art article. Language Teaching, 17, 326-343. Eckman, F. R. (1977). Markedness and the contrastive analysis hypothesis. Language Learning, 27, 315-333. Esling, J. H., & Wong, R. F. (1983). Voice quality settings and the teaching of pronunciation. TESOL Quarterly, 17, 89-95. Faerch, C. K., Haastrup, K., & Phillipson, R. (1984). Learner language and language learning. Clevedon, England: Multilingual Matters. Ferguson, C. A. (1971). Towards a characterization of English foreigner talk. Anthropological Linguistics, 17, 1-14. Flege, J. E. (1980). Phonetic approximation in second language acquisition. Language Learning, 30, 117-134. Flege, J. E. (1981). The phonological basis of foreign accent: A hypothesis. TESOL Quarterly, 15, 443-455. Gilbert, J. B. (1980). Prosodic development: Some pilot studies. In R. C. Scarcella & S. D. Krashen (Eds.), Research in second language acquisition (pp. 110-117). Rowley, MA: Newbury House. Gilbert, J. B. (1984). Clear speech. New York: Cambridge University Press. Giles, H. (1979). Ethnicity markers in speech. In K. R. Scherer & H. Giles (Eds.), Social markers in speech (pp. 251-290). Cambridge: Cambridge University Press. 222 TESOL QUARTERLY

Giles, H., Bourhis, R. Y., & Taylor, D. M. (1977). Towards a theory of language in ethnic group relations. In H. Giles (Ed.), Language, ethnicity and intergroup relations (pp. 307-348). London: Academic Press. Goffman, E. (1972). Relations in public. New York: Harper& Row. Guiora, A., Beit-Hallahmi,H., Brannon, R., Dull, C., & Scovel, T. (1972). The effects of experimentally induced changes in ego states on pronunciation ability in a second language: An exploratory study. Comprehensive Psychiatry, 13, 421-428. Gumperz, J. J. (1982). Discourse strategies. Cambridge: Cambridge University Press. Haycraft, B. (1971). The teaching of pronunciation: A classroom guide. London: Longman. Hecht, B. F., & Mulford, R. (1982). The acquisition of a second language phonology: Interaction of transfer and developmental factors. Applied Psycholinguistics, 3, 313-328. Hodne, B. (1985), Yet another look at interlanguage phonology: The modification of English syllable structure by native speakers of Polish. Language Learning, 35, 405-422. Johansson, F. A. (1973). Immigrant Swedish phonology: A study in multiple contact analysis. Lund, Sweden: CWK Gleerup. Labov, W. (1972). Sociolinguistic patterns. Philadelphia: University of Pennsylvania Press. Ladefoged, P. (1982). A course in phonetics (2nd ed.). New York: Harcourt Brace Jovanovich. Lado, R. (1957). Linguistics across cultures. Ann Arbor: University of Michigan Press. Laver, J. (1980). The phonetic description of voice quality. Cambridge: Cambridge University Press. Leather, J. (1983). Second-language pronunciation learning and teaching. State of the art article. Language Teaching, 16, 198-219. Liberman, A., Cooper, F., Shankweiler, D., & Studdert-Kennedy, M. (1967). Perception of the speech code. Psychology Review, 74, 431-461. Lowenthal, K., & Bull, D. (1984). Imitation of foreign sounds: What is the effect of age? Language and Speech, 27, 95-98. MacCarthy, P. (1979). The teaching of pronunciation. C a m b r i d g e : Cambridge University Press. Macken, M. A., & Ferguson, C. A. (1981). Phonological universals in language acquisition. Annals of the New York Academy of Sciences, 379, 110-129. Madden E. (1983). The effect of training on pronunciation. ORTESOL Journal, 4, 69-80. McLaughlin, B., Rossman, T., & McLeod, B. (1983). Second language learning: An information processing perspective. Language Learning, 33, 135-159. Menyuk, P. (1978). Language and maturation. Cambridge, MA: The M.I.T. Press. PRONUNCIATION REVISITED 223

Menyuk, P., & Anderson, S. (1969). Childrens identification and reproduction of /w/, /r/ and /1/. Journal of Speech and Hearing Research, 12, 39-52. Morely, J. (1979). Improving spoken English, Ann Arbor: University of Michigan Press. Mulford, R., & Hecht, B. F. (1980). Learning to speak without an accent: Acquisition of a second-language phonology. Papers and reports on child language development, 18, 16-74. Murakawa, H. (1981). Teaching English pronunciation to Japanese adults. Unpublished doctoral dissertation, University of Texas, Austin. Neufeld, G. (1977). Language learning ability in adults: A study on the acquisition of prosodic and articulatory features. Working Papers in Bilingualism, 12, 45-60. Neufeld, G., & Schneiderman, E. (1980). Prosodic and articulatory features in adult language learning. In R. C. Scarcella & S. D. Krashen (Eds.), Research in second language acquisition (pp. 105-109). Rowley, MA: Newbury House. Olson, L. L., & Samuels, S. J. (1973). The relationship between age and accuracy of pronunciation. Journal of Educational Research, 66, 263268. Parish, C. (1977). A practical philosophy of pronunciation. TESOL Quarterly, 11, 311-317. Pennington, M. (1984, March). Can pronunciation be improved by instruction in pronunciation? Paper presented at the 18th Annual TESOL Convention, Houston. Sato, C. J. (1984). Phonological processes in second language acquisition: Another look at interlanguage syllable structure. Language Learning, 34, 43-57, Schumann, J. H. (1975), Affective factors and the problem of age in second language acquisition. Language Learning, 25, 209-235. Schumann, J. H. (1978). The pidginization process: A model for second language acquisition. Rowley, MA: Newbury House. Scovel, T. (1969). Foreign accents, language acquisition and cerebral dominance. Language Learning, 19, 245-253. Seliger, H. W., Krashen, S., & Ladefoged, P. (1975). Maturational constraints in the acquisition of second language accent. Language Sciences, 36, 20-22. Selinker, L. (1972). Interlanguage. International Review of Applied Linguistics, 10, 209-232. Sheldon, A., & Strange W. (1982). The acquisition of /r/ and /1/ by Japanese learners of English: Evidence that speech production can precede speech perception. Applied Psycholinguistics, 3, 243-261. Snow, C. E., & Hoefnagel-Hhle, M. (1977). Age differences in the pronunciation of foreign sounds. Language and Speech, 20, 357-365. Stevick, E. W. (1978). Toward a practical philosophy of pronunciation: Another view. TESOL Quarterly, 12, 145-150. 224 TESOL QUARTERLY

Suter, R. W. (1976). Predictors of pronunciation accuracy in secondlanguage learning. Language Learning, 26, 233-254. Tarone, E. (1978), The phonology of interlanguage. In J. C. Richards (Ed.), Understanding second and foreign language learning (pp. 15-33). Rowley, MA: Newbury House. Tarone, E. (1980). Some influences on the syllable structure of interlanguage phonology. International Review of Applied Linguistics, 18, 139-152.



TESOL QUARTERLY, Vol. 20, No. 2, June 1986

English for Specific Purposes: Content, Language, and Communication in a Pharmacy Course Model
University of Maryland, Baltimore County

University of Maryland, Baltimore

After presenting an overview of content-area ESL and English for specific purposes (ESP), this article describes an experimental course in communication offered to nonnative English-speaking pharmacy students at the University of Maryland and reports the results of an evaluation of the course. A combination of contentarea ESL and ESP, the course, which met weekly for one semester, was team-taught by a pharmacist specializing in communication for pharmacists and an ESL specialist. Speech functions deemed necessary for effective oral communication by pharmacists in their professional settings were used as an organizing principle for the syllabus, which also provided for instruction in relevant linguistic structures, for instruction in communication principles and techniques, and for much active student participation. Although better evaluation measures are needed, student evaluations and comparison of the results on preand posttests of the Speaking Proficiency English Assessment Kit (Educational Testing Service, 1982a) suggest that the course was effective.

The experimental communication course for pharmacy students described in this article grew out of the concern of a number of administrators and faculty members at the University of Maryland School of Pharmacy that some of their nonnative English-speaking graduates were professionally handicapped by a lack of communicative ability in English. While standards for admission to the School of Pharmacy and the required level of English, as measured by standardized objective tests and course grades, are high, it had been observed that a number of graduates had difficulty finding 227

good positions due to their insufficient speaking proficiency. An ESL specialist at a sister campus of the university was contacted, and a course, Communication for Nonnative English-Speaking Pharmacy Students, was conceived. This course was envisioned as a combination of a content-area ESL course and an English-for-specific-purposes (ESP) course. Before the course itself is described, a brief overview of these two types of instruction is presented. Content-area ESL uses content (e.g., mathematics, art, science) to provide the context for language instruction (Mohan, 1986). Ideally two instructional objectivescontent-area knowledge or skill and increased language skillare accomplished in one class (Chamot, 1983). Content-area ESL is called by various names, including subject-related ESL (Allen & Howard, 1981), ESL through subject matter teaching (Krashen, 1982), and content-based ESL (Chamot, 1984). Particularly at the elementary and secondary levels, this approach combining ESL and content instruction has been the subject of considerable attention in recent years (Allen & Howard, 1981; Cantieni & Tremblay, 1979; Chamot, 1983, 1984; Early, Thew, & Wakefield, 1985; Mohan, 1986; Rodriguez, 1981; Tucker & dAnglejan, 1975). Teaching content-area ESL does not mean submersing ESL students in classes with native English speakers. Rather, special classes are provided for ESL learners which take into account their level of English proficiency and their communication needs (Mohan, 1986). The classes are taught by teachers who make linguistic and cultural adjustments in order to help their students understand (Krashen, 1982, p. 167). For successful language learning in content-area classes, it is necessary that students comprehend the material and the teachers messages (either through prior knowledge or strong contextual clues) and that teachers can understand the students messages well enough to provide feedback (Mohan, 1986). Language learning (or acquisition, following the Krashen model) in content-area ESL courses is stimulated by the rich context the subject matter provides, by the inherent interest and relevance of the content, and by the fact that the learners focus on messages and not on language form (Krashen, 1982; Mohan, 1986). In this approach language ceases to be taught in isolation (Mohan, 1986, p. 18). While much remains to be learned about teaching contentarea ESL, the research into its practicability and effectiveness is encouraging (Chamot, 1983). English for specific purposes (also known as English for special 228 TESOL QUARTERLY

purposes) is that area of English language teaching which focuses on preparing learners for chosen communicative environments (Mohan, 1986, p. 15). It differs from general English in that it is based on a close analysis of the learners communicative needs for a specific occupation or activity, as well as a detailed analysis of the language of that occupation or activity (Strevens, 1980). Unlike in general English courses, in an ESP course, English is taught for a clearly utilitarian purpose of which there is no doubt (Mackay, quoted in Robinson, 1980, p. 6), and it is taught not as an end in itself but as an essential means to a clearly definable goal (Mackay, 1978, p. 28). In other words, the language in an ESP course is not the subject matter but is being learned as part of the process of acquiring some quite different body of knowledge or set of skills (Robinson, 1980). The element that gives ESP its identity as a distinctive area of language teaching activity is learners purpose, a purpose that is not restricted to linguistic competence alone but that does involve the mastery of language skills in which language forms an integral part (Phillips, 1981, p. 92). While often addressing the communicative needs of professionals such as physicians or engineers, ESP also addresses the needs of nonprofessional workers (Crandall, 1984) and of special groups such as students (English for academic purposes). The following definition of an ESP course encompasses the elements widely considered essential to ESP, plus indicating some of the possible variations: An ESP course is purposeful and is aimed at the successful performance of occupational or educational roles. It is based on a rigorous analysis of students needs and should be tailor-made. Any ESP course may differ from another in its selection of skills, topics, situations and functions and also language. It is likely to be of limited duration. Students are more often adults but not necessarily so, and may be at any level of competence in the language: beginner, post-beginner, intermediate, etc. Students may take part in their ESP course before embarking on their occupational or educational role, or they may combine their study of English with performance of their role, or they may already be competent in their occupation or discipline but may desire to perform their role in English as well as in their first language. (Robinson, 1980, pp. 13-14) Content-area ESL and ESP share several of the same guiding principles: for example, the importance of context; the importance of attending primarily to meaning, not language form; and consideration for the needs of the learner (Mohan, 1986). They A PHARMACY COURSE MODEL OF ESP 229

employ many of the same methods based on communication and meaning. However, in some ways the two types of instruction are quite different, the most important differences stemming from their respective objectives. Content-area ESL aims to teach content and to improve overall English proficiency, in addition to teaching particular language skills required for understanding the content. Language learning is achieved primarily through careful presentation of understandable material (Chamot, 1984; Krashen, 1982). Even when the content material is carefully organized to support language learning, as recommended by Mohan (1986), the organization considers both content and language. ESP, on the other hand, has the narrower objective of preparing learners to function in very specific environments. Thus, ESP courses are structured principally to promote efficient and effective acquisition of particular language and communicative skills.1 The experimental pharmacy course described in this article combined various features of content-area ESL and ESP. Like content-area ESL courses, the experimental course taught English partly as a by-product of teaching some other content area, in this case communication principles and techniques for pharmacists. Offered as a three-credit elective for nonnative speakers, the experimental course was to be taken after completion of the communication course required of all pharmacy students, native and nonnative English speakers alike. It offered further instruction in communication principles and techniques, as well as skill development in spoken English. As an ESP course, the experimental communication course had as its overriding objective more effective oral communication by the learners in situations likely to be encountered by pharmacists in their professional lives. Procedures for analyzing the students needs and for designing the syllabus were influenced by the ESP literature as well as by recent writing on functional/notional syllabuses and communicative competence. DESIGNING THE SYLLABUS Because the needs of the learners are paramount in an ESP course, one must analyze what the learners need to be able to do before attempting to design the syllabus (Mackay, 1978; Munby, 1978; Richterich & Chancerel, 1980; Robinson, 1980; Schmidt,
1 See

Krashen (1982, pp. 169-170) for a discussion of the difference between ESL through subject-matter teaching and ESP.



1981). A needs analysis considers the needs expressed by the learners themselves, by the teaching establishment, by the userinstitution, or by all three (Richterich & Chancerel, 1980). To determine what skills would be needed for effective oral, communication by pharmacists with their patients and with other professionals, we relied on (a) the knowledge of the instructor of the required communication course, a practicing pharmacist who was to be one of the instructors of the course for ESL students, and (b) information gleaned from materials (videotapes and books) developed by pharmaceutical companies for the improvement of the communication of pharmacists. We also considered the extensive and valuable research done by Candlin, Bruton, Leather, and Woods (1981) on the communication of physicians in consultation with their patients. In addition, prospective students were interviewed before the syllabus design began, and, as is recommended by many writers on ESP (Robinson, 1980), a questionnaire was administered to the students on the first day of class. In the design of the course, we chose an analytic approach, as described by Yalden (1983), as opposed to a synthetic approach. That is, our starting point was not a set of linguistic items to be learned but rather linguistic and extralinguistic behaviors needed to achieve the goal of communicative competence. We wanted our syllabus to be based on meaning rather than on formal structure and to emphasize language as an interpersonal behavior rather than as a personal one. In analytic approaches, learners rely on their ability to analyze, since there is no attempt to carefully control the language in the classroom. Rather, components of language are not seen as building blocks which have to be progressively accumulated. Much greater variety of linguistic structure is permitted from the beginning and the learners task is to approximate his own linguistic behavior more and more closely to the global language. (Wilkins, 1976, p. 2) In designing the syllabus, we attempted to take into account three components of language recognized by Wilkins (1976): the semantic component (what to communicate), the functional component (why we communicate), and the formal component (how w e communicate). The following components of a communicative syllabus (Yalden, 1983) were considered: (a) the purposes for which the learners are studying the language, (b) the setting in which they will be using it, (c) the role the learners will assume, as well as the roles of their interlocutors, (d) the communicative events or situations in which they will participate, (e) the language functions A PHARMACY COURSE MODEL OF ESP 231

involved in these events, (f) the notions involved, (g) the discourse and rhetorical skills involved, (h) the varieties and levels of language the learners will need, (i) the grammatical content needed, and (j) the lexical content needed. Like others (Krashen, 1981; Yalden, 1983), we believe the most effective learning takes place if the emphasis in the class is on communicative competence rather than formal accuracy. Particularly with advanced learners, the emphasis should be, in Brumfits (1982) words, on fluency work rather than on accuracy work (p. 76). Therefore, we planned classroom activities that would focus on skillful communication and increased fluency rather than on formal correctness. However, we realize, as Brumfit does, that communicative competence is aided by greater control over grammatical form and a larger repertoire of lexical items from which to choose. For this reason we decided to include some review of relevant structures and to teach idioms which might be used in the pharmacy or hospital. THE STUDENTS The class consisted of 10 students, 9 from the School of Pharmacy and 1 from the School of Nursing. The latter was allowed to participate because no other suitable English classes were available. The pharmacy students had been identified by faculty as likely to benefit from an ESL course. All were individually informed about the course, and all elected to take it. The nursing student was referred by an administrator at the School of Nursing. All students were of Asian descent, and most had been in the United States for several years. Before the course began, the students were administered the Educational Testing Services (1982a) Speaking Proficiency English Assessment Kit (SPEAK) tests, the institutional version of the Educational Testing Services (1982b) Test of Spoken English. The SPEAK test audiocassettes were sent to the Educational Testing Service for professional rating, and most of the students scored on the high end of the scale on overall comprehensibility (M = 228; see Table 1) in the range described by the Educational Testing Service (1982b) as generally comprehensible with some errors in pronunciation, grammar, choice of vocabulary items, or with pauses or occasional rephrasing (p. 11). Four of the students tested achieved overall comprehensibility scores of over 250, in the range described by the Educational Testing Service (1982b) as completely comprehensible in normal speech, with occasional 232 TESOL QUARTERLY

grammatical or pronunciation errors in very colloquial phrases (p. 12). Before sending the test cassettes to the Educational Testing Service for rating, the ESL specialist listened to them to evaluate the need of each prospective student for the class and also to see where improvement was needed. Only 1 student was considered too advanced for the class. (However, this student chose to participate anyway and, as it turned out, showed significant improvement.) As a whole, the greatest need seemed to be for increased fluency, a need which was corroborated by the official Educational Testing Service ratings. The prospective students were also given the listening section of Harris and Palmers (1970) Comprehensive English Language Test (CELT). The mean score of the pharmacy students was 94.5%, indicating a very high level of listening proficiency. (The nursing student, who had been in the country a very short time, scored much lower than all the pharmacy students.) On the first day of class, the students were asked to fill out a questionnaire asking them what they thought they needed most to work on. As a group, the students, corroborating the results of the SPEAK test, felt they most needed to increase their fluency. The area that was ranked as the second greatest area of need was writing, and the third was vocabulary. (To satisfy the desire of many of the students for writing instruction, two required writing assignments were given on pharmacy-related topics, as well as a number of optional writing assignments.) DESCRIPTION OF THE COURSE The course was structured around a number of speech functions considered necessary for effective oral communication by pharmacists in their professional settings (see Appendix A). In selecting the speech functions, we relied heavily on the work done by Candlin et al. (1981) on the language of physicians in clinical practice. During the course, each function was first illustrated, either by using professionally videotaped dialogues prepared by pharmaceutical companies or by live demonstration by the instructors. Following the illustration, practice in recognizing the function was given. Next, lists of expressions for performing the function were discussed, and additional expressions were elicited from the students existing knowledge (see Appendix B for an example of a list of expressions for a function). At the same time, extralinguistic A PHARMACY COURSE MODEL OF ESP 233

behaviors suitable for performing the function were considered. When necessary, analysis and practice of the grammatical forms were provided. The appropriateness of various behaviors, both linguistic and extralinguistic, was always considered. Finally, roleplaying activities incorporating the function under study were performed, followed by analysis of the role-playing (see Appendix C for role-play situations for practicing the directing function). Role-playing was fundamental to the course because our pharmacy students, like other ESP students (Robinson, 1980), were studying primarily in order to perform roles. The role plays were designed to familiarize the students with the kinds of situations they might face as practicing professionals. In addition, role-playing was employed because it is believed to be an effective technique for developing communicative competence (Krashen & Terrell, 1983; Mockridge-Fong, 1979; Rivers & Temperley, 1978). By creating dramatic situations in the classroom, teachers can, in Newmarks words, expand the classroom indefinitely and provide imaginatively natural contexts for the language being used (cited in Mockridge-Fong, 1979, p. 96). We found that role-playing also contributed in ways mentioned by Donahue and Parsons (1982): It served as a vehicle for learning the appropriateness of certain linguistic and extralinguistic behaviors in various situations; it provided sensitizing situations for developing an understanding of American culture and of roles within American culture; and it gave the students an opportunity to compare their reactions to the problems and solutions presented in the role plays to those of their classmates. In addition, the role plays elicited from the pharmacist-instructor much information about pharmacy and pharmacy practice which was of considerable value to the students. While role-playing continued throughout the course (see Appendix D for some additional role-playing situations), procedures were changed in response to student suggestions. In the beginning, students were given situations to role-play in front of the rest of the class, and then their performance was reviewed, as tactfully as possible. About midway through the course, students complained that watching other students was not useful; from then on, the instructors would role-play before the students did and then solicit criticisms from the class. This achieved two results: (a) A generally positive model was presented, and (b) because the instructors criticized their own performance and solicited criticisms from the students, the latter realized that the aim of the activity was not perfection but practice and an increase in awareness. 234 TESOL QUARTERLY

Toward the end of the course, role plays were videotaped and then played back and reviewed. Students seemed to feel that a better result could be achieved by videotaping in private and reviewing in public. Although several students complained that they were too self-conscious to do well at role-playing in front of the class (and at the beginning there was a considerable amount of clowning), in their final evaluations they rated role-playing quite highly (see Table 2). In hopes of increasing the students repertoires of appropriate expressions, which in turn would lead to increased fluency, structured work on idioms was included in the course. The work on idioms was also intended to improve their comprehension of colloquial speech, a need expressed by a number of students in the class. Only idioms which the instructors thought might be used in professional settings were studied, and students were asked to incorporate the idioms in dialogues which they themselves wrote (see Appendix E). Colloquial terms for parts of the body (e.g., belly for abdomen) and for bodily functions were also introduced. It was emphasized that vocabulary development was a continuing process and that the students themselves were responsible for making efforts to increase their own vocabularies both during the course and after. Work on pronunciation, particularly the suprasegmental elements, was incorporated into the course. Most of the instruction on pronunciation took place in the language laboratory using the texts Clear Speech (Gilbert, 1984) and Stress and Intonation (English Language Services, 1967). Because it is crucial that pharmacists make their important points understood, both the laboratory work and the pronunciation work in class emphasized ways English speakers signify the relative importance of ideas they are expressing. (Instruction in other ways of focusing attention and indicating significance was also an important part of the class. ) The language laboratory was also used for listening comprehension activities. While students listened to articles about communication taken from pharmaceutical journals, they followed the written text, filled in missing words, and took dictation read from the articles (see Appendix F). Discussions of the content of the articles sometimes followed. Not only did these listening exercises give practice in listening comprehension and information about communication problems and solutions in pharmaceutical practice, they also focused attention on grammatical structures (particularly articles) that led to brief but, as it turned out, valued grammar lessons. The language laboratory was also used for practicing the A PHARMACY COURSE MODEL OF ESP 235

names of the 200 most common drugs, using a tape prepared by the pharmacist-instructor at the request of the students. Instruction on principles and techniques of effective communication was woven into virtually all the class work that was done: listening comprehension exercises, discussions of the appropriateness of various linguistic and extralinguistic behaviors, pronunciation work, and especially the critiques of the role-playing. EVALUATION AND DISCUSSION At the completion of the course, a second version of the SPEAK tests was administered. The results of a paired t test showed statistically significant improvement in all categories (see Table 1). The positive results of these tests should not be overrated, since there was no control group with which to compare the gains made. In addition, these 20-minute proficiency tests were not designed to measure the specific kinds of communicative competence that were the overriding objective of the course. Nevertheless, the results are gratifying, especially because of the advanced proficiency level of the students. It was also gratifying to see that the mean fluency



score rose the most (14.2%) of all the scores, since increasing fluency was one of our primary objectives. The students were also asked to evaluate the course at its completion (see Table 2). On a scale of 1 to 6 (6 being the highest), the mean rating for the course overall was 4.8. The students least liked the work done in the language laboratory. They most liked the model role plays done by the pharmacist-instructor (M = 5.3), and in second place, surprisingly, came instruction on articles (M = 5.1). Student role-playing exercises received a mean rating of 4.9. A number of comments written in response to the question, What was the best thing about the class? applauded the informal atmosphere, presumably because it led to a great deal of student participation.
TABLE 2 Evaluation by Students

On the basis of both student evaluations and instructor observations, several changes will be made in the course. First, better language laboratory activities will be developed, and laboratory assignments will be based on individual needs. For some very advanced students, reading and writing activities can occasionally be substituted for oral-aural activities. However, in spite of the feeling on the part of a number of the students that the laboratory work was too easy, the instructors felt that the increase in the pronunciation and overall comprehensibility scores on the SPEAK tests was very likely attributable, in large part, to the laboratory work. Therefore, laboratory sessions on pronunciation will continue to be included. A PHARMACY COURSE MODEL OF ESP 237

In the future, more modeling of successful communicative acts will be provided. This can done by live modeling on the part of the instructors, by videotapes produced by the instructors, and by professionally prepared videotapes. Also, when student role plays are to be videotaped so that students can see themselves, the role plays will be either taped out of class or viewed out of class, in order not to waste class time. The greatest need for improvement is in the area of evaluation. In addition to tests measuring proficiency, such as the SPEAK and CELT, tests are needed that will determine the learners ability to perform the kinds of communicative tasks being taught (Robinson, 1980; van Ek, 1973; Yalden, 1983). Ideally, the tests would involve oral interaction, although such tests can be difficult to devise properly and to assess, as Carroll (1980) points out. APPLICABILITY This course was designed specifically for pharmacists in training, but very similar courses could be designed for other health preprofessionals and professionals. We believe that communication courses designed for more than one health profession, though not ideal, could be useful if courses targeted for specific professions were not possible. The nursing student who participated in the course described here seemed to benefit as much as the pharmacy students. Because this course was team-taught, it was relatively expensive. Could a similar course be taught by an ESL instructor alone? While it is ideal to have as one of the instructors a practicing health professional experienced in teaching communication, such professionals are rare. It is probably not absolutely necessary provided that such a professional work closely with the ESL specialist in designing the course and in preparing materials and assignments, that the health professional act as a consultant during the course, and that the health professional participate in at least some of the classes. In addition, the ESL specialist would need to become acquainted with the literature on communication in the health professions. The more the ESL specialist becomes conversant in communication principles and techniques, and the greater the availability of professionally prepared materials relating to communication in the health fields, the less the ESL specialist needs to rely on a health professional. However, it must be emphasized that when one is working with preprofessionals, the participation of a health 238 TESOL QUARTERLY

professional is extremely valuable. Preprofessionals often are not sure of the correct professional information or advice to be given to patients in, for example, the role plays. The course described in this article taught communication principles and extralinguistic communication skills as well as ESL, an approach we feel has much to recommend it. However, given the continuing concern about the communicative competence of foreign-born health professionals, even courses more narrowly focused on the linguistic aspects of communication in English are needed. In attempting to convince health professionals and administrators of the value of such courses, ESL professionals must first convince them that these courses are specifically designed to meet the communicative needs of health professionals in their professional settings. Such traditional, linguistically oriented ESP courses appear to meet the needs of particular groups more effectively than general English courses. The addition of content-area instruction to an ESP course, however, may prove to be more effective yet. Content-area information and skill development enhance the learning of specialized communication, we believe, by enriching the context in which the language is learned, by stimulating interest in the course, and by increasing the courses perceived relevance. Content and language instruction can mutually reinforce one another, producing a sum greater than its parts. In addition to meeting the needs of health professionals, courses which combine attributes of content-area ESL and ESP could serve well, we believe, to provide training for specialists in other occupations. The demand for varieties of ESP, as opposed to general English courses, appears to be growing both in the United States (Crandall, 1984; Huckin & Olsen, 1984) and throughout the rest of the world (Strevens, 1980); thus, imaginative responses to the demand for relevant and effective English language instruction are called for.

Janet G. Graham is Language Coordinator at the Learning Resources Center of the University of Maryland, Baltimore County, where she teaches ESL and administers ESL programs. She is a Ph.D. candidate in English education at the University of Maryland and recently served as president of the Baltimore-area TESOL.



Robert S. Beardsley, Associate Professor at the School of Pharmacy, University of Maryland at Baltimore, received his Ph.D. in pharmacy administration from the University of Minnesota. He is co-editor of the journal Patient Counseling in Community Pharmacy and of the textbook Communication in Pharmacy Practice. His current research efforts include patient education/communication skills of pharmacists.

REFERENCES Allen, J. P. B., & Howard, J. (1981). Subject-related ESL: An experiment in communicative language teaching. Canadian Modern Language Review, 37, 535-550. Brumfit, C. (1982). Methodological solutions to the problems of communicative teaching. In M. Hines & W. Rutherford (Eds.), O n TESOL 81 (pp. 71-77). Washington, DC: TESOL. Candlin, C., Bruton, C., Leather, J., Woods, E. (1981). Designing modular materials for communicative language learning: An example: Doctorpatient communication skills. In L. Selinker, E. Tarone, & V. Hanzeli (Eds.), English for academic and technical purposes: Studies in honor of Louis Trimble (pp. 105-133). Rowley, MA: Newbury House. Cantieni, G., & Tremblay, R. (1979). Use of concrete mathematical situations in learning a second language: A dual learning concept. In H.T. Trueba & C. Barnett-Mizrahi (Eds.), Bilingual multicultural education and the professionals: From theory to practice (pp. 246-255). Rowley, MA: Newbury House. Carroll, B, J. (1980). Testing communicative performance. Oxford: Pergamon Press. Chamot, A.U. (1983). Toward a functional ESL curriculum in the elementary school. TESOL Quarterly, 17, 459-471. Chamot, A.U. (1984). A transfer curriculum for teaching content-based ESL in the elementary school. In J. Handscombe, R.A. Orem, & B.P. Taylor (Eds.), On TESOL 83 (pp. 125-133). Washington, DC: TESOL. Crandall, J. A. (1984). Adult ESL: The other ESP. The ESP Journal, 3, 9196. Donahue, M., & Parsons, A.H. (1982). The use of roleplay to overcome cultural fatigue. TESOL Quarterly, 16, 359-365. Early, M., Thew, C., & Wakefield, P. (1985). ESL instruction via the regular curriculum: A framework and resource book. Victoria, B. C.: Ministry of Education. Educational Testing Service, Test of English as a Foreign Language Program. (1982a). Speaking proficiency English assessment kit (SPEAK). Princeton, NJ: Author. Educational Testing Service. (1982b). Test of spoken English: Manual for score users. Princeton, NJ: Author. English Language Services. (1967). Drills and exercises in English pronunciation: Stress and intonation, Part 1. New York: Collier Macmillan. 240 TESOL QUARTERLY

Gilbert, J. (1984). Clear speech. Cambridge: Cambridge University Press. Harris, D. P., & Palmer, L.A. (1970). A comprehensive English language test for speakers of English as a second language (CELT). New York: McGraw-Hill. Huckin, T. N., & Olsen, L. A. (1984). The need for professionally oriented ESL instruction in the United States. TESOL Quarterly, 18, 273-294. Krashen, S.D. (1981). Second language acquisition and second language learning. Oxford: Pergamon Press. Krashen, S.D. (1982). Principles and practice in second language acquisition, Oxford: Pergamon Press. Krashen, S. D., & Terrell, T.D. (1983). The natural approach. Hayward, CA: The Alemany Press. Mackay, R. (1978). Identifying the nature of the learners needs. In R. Mackay & A. Mountford (Eds.), English for specific purposes (pp. 2137). London: Longman. Mockridge-Fong, S. (1979). Teaching the speaking skill. In M. CelceMurcia & L. McIntosh (Eds.), Teaching English as a second language (pp. 83-89). Rowley, MA: Newbury House. Mohan, B.A. (1986). Language and content. Reading, MA: AddisonWesley. Munby, J. (1978). Communication syllabus design. C a m b r i d g e : Cambridge University Press. Phillips, M.K. (1981). Toward a theory of LSP methodology. In R. Mackay & J.D. Palmer (Eds.), Languages for specific purposes: Program design and evaluation (pp. 92-105). Rowley, MA: Newbury House. Richterich, R., & Chancerel, J.L. (1980). Identifying the needs of adults learning a foreign language. Oxford: Pergamon Press. Rivers, W., & Temperley, M.S. (1978). A practical guide to the teaching of English. New York: Oxford University Press. Robinson, P. (1980). ESP (English for specific purposes). O x f o r d : Pergamon Press. Rodriguez, I.Z. (1981). An inquiry approach to science/language teaching and the development of classification and oral communication skills of Mexican-American bilingual children in the third grade. Unpublished doctoral dissertation, The University of Texas at Austin. Schmidt, M.F. (1981). Needs assessment in English for specific purposes: The case study. In L. Selinker, E. Tarone, & V. Hanzeli (Eds.), English for academic and technical purposes: Studies in honor of Louis Trimble (pp. 199-210). Rowley, MA: Newbury House. Strevens, P. (1980). Teaching English as an international language: From practice to principle. Oxford: Pergamon Press. Tucker, G. R., & dAnglejan, A. (1975). New directions in second language teaching. In R.C. Troike & N. Modiano (Eds.), Proceedings of the First Inter-American Conference on Bilingual Education (pp. 63-72). Arlington, VA: Center for Applied Linguistics.



van Ek, J. (1973). The threshold level in a unit/credit system. In J.M.L. Trim (Ed.), Systems development in adult language learning (pp. 89146). Strasbourg: Council of Europe. Wilkins, D. (1976). Notional syllabuses. London: Oxford University Press. Yalden, J. (1963). The communicative syllabus: Evolution, design, and implementation. Oxford: Pergamon Press.

APPENDIX A Working Taxonomy of Speech Functions for Pharmacists 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. Greeting Asking for informationclosed-ended questions Asking for informationopen-ended questions Reflecting Paraphrasing Reassuring Acknowledging receipt of information Encouraging further communication Requesting feedback Directing Making sure Extending (to test a deduction) Socializing Taking leave APPENDIX B Expressions for the Directing Function
Take these three times a day, with meals. The first thing you have to do is . . . After youve done that you . . . The next thing you do is . . . Be sure to . . . Make sure to . . . Dont forget to . . . Be careful not to . . . Make sure you remember . . . The most important thing is to . . . Its really important to . . . TESOL QUARTERLY


Its extremely important that you . . . I would like you to . . . I feel its important to . . . Some people find that . . . Research has found that . . . In my experience, Ive found that. . .

APPENDIX C Role Plays for the Directing Function Tell a patient how to do the following: 1. Administer an eye drop, ear drop, nose drop, enema, rectal suppository, or nitroglycerin patch 2. Take blood pressure or pulse 3. Put in a contact lens 4. Adjust crutches, canes, or walkers 5. Care for ostomies 6. Take temperature of pediatric patient 7. Measure for support stocking, braces, or supports

APPENDIX D Additional Role Plays The following role-playing situations focus on asking reluctant people open-ended questions. Players receive their instructions on separate slips of paper and do not know what others are told. Situation 1 Patient: Two days ago you accidentally dumped 10 of your digoxin tablets down the sink, and now you need to get a refill for some more. If the pharmacist asks you what happened, dont tell him or her, since you are really embarrassed about the accident. Pharmacist: A patient returns for a refill of digoxin. After checking the patient profile, you notice that the person is coming back 10 days early. Try to determine why the medication is needed so soon. Is the patient taking it correctly? Situation 2 Patient: You come back for a refill for your diuretic medication. You have not been taking it correctly, missing some days, since you dont think it does any good. You are embarrassed to tell the pharmacist the real reason, and you become defensive when asked about it. A PHARMACY COURSE MODEL OF ESP 243

Pharmacist: A patient returns for a refill for a diuretic medication. After checking the patient profile, you notice that the patient should have refilled the prescription 10 days ago. Ask the individual what happened. Situation 3 Staff pharmacist: You have been arriving late to work (15 minutes) for the past week, since you need to take your son to nursery school. Your wife/ husband usually drops him off, but he/she is severely ill and you dont want anyone to know about itit is a private, family matter, plus you stay late each day to make up the time. Hospital pharmacy director: You notice that this staff pharmacist has been coming late every day (15 minutes) for the past week, and you want to know why. You wont tolerate tardiness. APPENDIX E Example of a Student-Written Dialogue Incorporating Idioms The idioms which were being studied are underlined. Ph = Pharmacist; Pt = patient. Ph: Im just about to call your doctor for your medication. It will only take a second. Pt: I appreciate you doing this for me. Im so absent-minded. I keep forgetting to have a checkup and then I run out of my pills. Ph: What about your medication for high blood pressure? Are you taking it regularly? Pt: Well, I try to remember to take it everyday, but with all these different medicines I need to take, sometimes its very confusing. Ph: I understand, Mrs. Jones. Would it be easier if I made you a drug schedule so you can check off the one you took? Thats really a good idea. I will definitely take advantage of it. APPENDIX F I.istening Exercise Directions: Listen to the following article and fill in the missing words. Paragraph 5 is missing entirely: Listen to the tape and take dictation from it. When you have finished listening and taking dictation, check your work against the correct version of the article, and prepare yourself to discuss it. Patient Compliance Improves When Practitioners Emote* A review of the social scientific literature on physician-patient communication, published in the Journal of the American Medical * From American Pharmacy (1985), NS 25 (2), p. 18. 244 TESOL QUARTERLY



TESOL QUARTERLY, Vol. 20, No. 2, June 1986

Coherence and Academic Writing: Some Definitions and Suggestions for Teaching
San Diego State University

Coherence in written text is a complex concept, involving a multitude of reader- and text-based features. Perhaps because of this, we writing instructors and the textbooks we use often discuss coherence in a vague or incomplete manner. This article reviews current coherence literature, defines coherence in broad terms, then presents a three-lesson revision unit based on modern coherence principles. In this unit, ESL students deconstruct the assignment prompt and prepare their own first drafts of an essay response. Then they examine a fellow students first draft from the top down, evaluating the thesis in relationship to the prompt and to the assertions within the essay and analyzing the information structure intended to guide readers through the text. Conclusions are drawn about the success of this group revision technique and the necessity for providing sequential exercises to improve coherence.

A recent survey of college instructors teaching lower-division general education classes (Johns, 1985), conducted to determine the tasks which they assign and concerns which they have about ESL student writing, found that the task types were predictable: Most required the integration of information from sources (e. g., lecture, assigned readings, or library sources) into written assignments. In terms of concerns, a number of those responding commented that students academic writing is often incoherent, a feature which appears to cover a large number of perceived weaknesses. Although all of us may believe we have a sense of what the term incoherent means, we often find ourselves discussing incoherence and coherencein vague terms with students. Sometimes we do not get past comments such as those cited in Jacobs (1982):
A piece of writing is coherent when it elicits the response: I follow you.


I see what you mean. It is incoherent when it elicits the response: I see what youre saying here, but what has it got to do with the topic at hand or with what you just told me above? (p. 1)

These remarks, though true to the recent discussion of coherence as a phenomenon involving the interaction of reader with text (Carrell, 1982; Rumelhart, 1977) and as primarily a function of topic development (Grabe, 1984), are not of much help to our students, who need more specific definitions and sequential, task-dependent exercises to produce prose judged to be coherent by experienced graders. To refine my understanding of coherence and methods for teaching coherent writing, I reviewed the literature on coherence and prepared a series of questions to guide group editing. These questions have proved much more useful to my students and me than my previous general comments and fragmented activities. The purpose of this article is to share what I have discoveredhow coherence is variously defined in the recent literature and how I assist my ESL students in revising their papers to improve coherence. TWO DEFINITIONS OF COHERENCE: AS TEXT BASED AND READER BASED Text-Based Coherence Coherence is defined by some as a feature internal to text. In traditional handbooks (see, e.g., Hodges & Whitten, 1972), this feature is divided into two constructs: cohesion (i.e., the linking of sentences) and unity (i.e., sticking to the point). Often, these constructs are introduced separately, as if, in fact, they could be separated in written text (see, e.g., Bander, 1978; Martin, 1974). The appearance of Halliday and Hasans Cohesion in English (1976) has had a major impact on the understanding and teaching of coherence features. These linguists speak of coherent text as having two characteristics somewhat different from those in the traditional definition: cohesion (i.e., ties between sentences) and register (i.e., coherence with a context): A text is a passage of discourse which is coherent in these two regards: it is coherent with respect to the situation, and therefore consistent in register; and it is coherent with respect to itself, and therefore cohesive. (p. 23) Though Halliday is concerned with register appropriateness in other writings (see, e.g., Halliday, 1978), Cohesion in English 248 TESOL QUARTERLY

focuses almost exclusively on cohesion as a text feature. This work has created more controversy and interest (see, e.g., Carrell, 1982; Markels, 1983) and spurred more ESL writing research (see, e.g., Connor, 1984; Johns, 1980a; Scarcella, 1985; Witte & Faigley, 1981) than have Hallidays register features. The category types which appear in Cohesion in English (reference, substitution, ellipsis, conjunction, and lexical cohesion) have become common subjects for discussion in well-respected ESL teacher reference textbooks (e.g., Hughey, Wormuth, Hartfiel, & Jacobs, 1983, p. 129; Raimes, 1983, pp. 53-55). However, some of the points made in Cohesion in English have been misinterpreted or misused in the classroom. The literature on cohesion has warned against clustering cohesive items in semantic groups, such as teaching all additives (and, in addition, furthermore, etc. ) together (see, e.g., Johns, 1980b; Kantor, 1985); yet some textbooks continue to list these related items in groups, disregarding register or semantic variation. Cohesive items have also been taught prescriptively, in isolated exercises, without consideration for constructed texts. Witte and Faigley (1981) warn teachers against using these practices: Coherence conditionsconditions governed by the writers purpose, the audiences knowledge and expectations, and the information to be conveyedmilitate against prescriptive approaches to the teaching of writing. Indeed, [our] exploration of what cohesion analysis can and cannot measure in student writing points to the necessity of placing writing exercises in the context of complete written text. (p. 20) Whereas Halliday and Hasan speak of coherent text as being cohesive (i.e., having appropriate ties among sentences), other modern text analysts have concentrated upon the sticking to the point feature of coherence. More important, they have discussed the relationship of the points, or propositions, to each other. Selection of cohesive items and other features of the information structure are subsumed in their analyses; for them, meaning, realized in propositional relationships, drives the text. Some of the most interesting work on sticking to the point comes from the Prague School (e.g., Lautamatti, 1986), whose members have investigated how sentence topics combine to lead the reader through text and to an understanding of the discourse theme or topic. Witte (1983) and Connor and Farmer (1985) have applied the topic depth and maintenance models of the Prague School to the writing of native-speaker and ESL students. They have found, among other things, that passing essays have fewer topics and more T-units per topic than do failing essays, thereby demonstrating that COHERENCE AND ACADEMIC WRITING 249

topic support is one of the most important features of coherent essays. Grabe (1985), in a useful review of text linguistics literature including the work of the Prague School, speaks of coherence as generally defined by text analysts as a theoretical construct in text structure [referring] to the underlying relations that hold between assertions (or propositions) and how they contribute to the overall discourse theme (p. 110). Grabe cites various well-known textanalytical models, which share three interacting features essential to coherence: (a) a discourse theme (or thesis); (b) a set of relevant assertions relating logically among themselves by means of subordination, coordination, and superordination (see, e.g., Nold & Davis, 1980); and, (c) an information structure imposed on the text to guide the reader in understanding the theme or intent of the writer. (This last includes cohesion and a number of other features; see Vande Kopple, 1985.) Reader-Based Coherence So far, coherence has been defined principally as a feature of text, either in terms of the linking of sentences (cohesion) or as the relationships among propositions in the text (sticking to the point). However, others claim, on the basis of schema-theoretical models, that a text cannot be considered separately from the reader and that coherence requires successful interaction between the reader and the discourse to be processed (Carrell, 1982; Rumelhart, 1977). According to this view, the degree to which a reader grasps the intended meaning and underlying structure from text (and therefore finds it coherent) depends, to a large extent, upon whether the reader-selected schemata (or expectations) are consistent with the text (see, e.g., P. Johnson, 1982; Miller & Kintsch, 1980). These expectations are founded in the readers prior knowledge, both of the content to be introduced and the form it takes (Carrell, 1983). As the reader processes the text, these expectations are modified to establish consistency with text structure or content, for reading is a process of continuous interpretation. Rumelhart (1977) and others have noted that text processing takes place on a number of levels, from the bottom up (the processing of letters, words, and phrases), as well as from the top down (from the readers prior knowledge and expectations). However this is done, the important point is that reading is considered an interactive and interpretive process. Therefore, the writer must continuously keep 250 TESOL QUARTERLY

the intended audience in mind (see Johns, in press). Armbruster and Anderson (1984) speak of discourse which meets reader expectations and provides guidance through the text as reader considerate. . This review of current literature provides a number or principles to guide instructors in teaching the concept of coherence: 1. Coherence is text based and consists of the ordering and interlinking of propositions within a text by use of appropriate information structure (including cohesion). 2. At the same time, coherence is reader based; the audience and the assignment must be consistently considered as the discourse is produced and revised. 3. Instructors have an obligation to teach coherence comprehensively, that is, to take into account these two approaches (text based and reader based), at a minimum. TEACHING COHERENCE Many students at every level are unfamiliar with the conventions of English writing which, if well integrated, result in coherent prose. Numerous ESL textbooks present sentence-level grammar in a discourse context (see, e.g., J. A. Johnson, 1983; Sheehan, 1986) and teach students to write generalizations (topic sentence and theses) and to provide supporting examples and details (see, e.g., Huizenga, Snellings, & Francis, 1982; Rice & Burns, 1986). However, none that I am aware of examines or teaches the multitude of coherence features discussed in recent literature. Therefore, published textbooks, though they may supplement and augment the following suggestions, do not provide sufficient introduction to the depth and variety of coherence features necessary for proficient writing. Many of these recent textbooks also make use of the process approach, which is based on a theory of writing development that has revolutionized teaching in the past 10 years (see, e.g., Sommers, 1980; Spack, 1984; Zamel, 1983). However, we may be doing our students a disservice by strictly adhering to all tenets of this approach, for it must be examined in light of the tasks which students are required to perform. The classical process approach requires two conditions for student writing: (a) time to plan (Spack, 1984), draft, and revise (Sommers, 1980) and (b) student-generated meaning and form. COHERENCE AND ACADEMIC WRITING 251

Horowitz (1985) points out that time for revision is not available for students writing essay responses in timed academic (i.e., other than English) examinations. For students writing an essay for an English class, some time may be available, as it is when students prepare out-of-class assignments for their academic courses. However, the second condition for the process approach, that of student-generated meaning and form, is contradictory to the authentic requirements of most academic classrooms. Horowitz (1985), Swales (1982), and Johns (1985) have all found that an academic assignment (or prompt) will generally designate the content, form, aims, and strategies required for response to the prompt. To prepare our students for authentic tasks, then, we must generate representative prompts requiring expository writing. And, when teaching and evaluating the writing which results from these assignments, we must insist upon student adherence to the requirements of the prompts in order to insure reader-based coherence. THE REVISION UNIT My advanced students have studied grammar and have some familiarity with the essay model (see Martin, 1974), with topic sentence and thesis development. Yet they seem to be unable to transfer these essay-writing conventions to their own prose, especially in response to a prompt, and most continue to revise their work at the sentence level only. Because of these problems, I begin teaching coherence to students by moving from the top down, that is, from more global to more local considerations. Research shows that good revisers who develop coherent text begin at the top, rather than at the bottom, that they correct spelling, grammar, and mechanics last (see Kroll, 1983; Sommers, 1980). In successive task-dependent activities, the class is asked to consider coherence systematically in terms of prompt requirements, thesis development, the relationships among assertions and to the thesis, and the adequacy of the information structure. Only in the final stages do students edit for sentence-level errors. The unit consists of a minimum of three lessons, each of which has multiple goals, drawn from both task- and reader-based coherence literature. Understanding the prompt and developing a discourse theme (or thesis) in response to it are the goals of the first 252 TESOL QUARTERLY

lesson, which draws principally from reader-based considerations. This lesson results in the production of first drafts of an essay. Goals of the second lesson, which is principally text based, are to analyze a thesis statement and the relationships between propositions in an essay. The first draft of one students essay is examined during this lesson. After the second lesson, all students revise their essays, using the same techniques that they applied to the sample student essay. The third lesson again focuses on the work of a single student, this time on the students second draft. This lesson concentrates on reader-based considerations in the information structure. After the lesson, the students again take their own papers home and revise them, asking the questions posed in the class. Finally, they edit for sentence-level errors before handing the papers in to me. Before the in-class instruction began for the class discussed here, students were asked to read a short essay by Mead and Metraux (1984) entitled The Gift of Autonomy. We discussed main ideas and vocabulary together, with students making marginal notes, until I was satisfied that they understood the reading and could respond to the content of the essay. They were then given the following prompt, requiring the integration of information from the Mead and Metraux essay into their writing: According to Mead and Metraux, parents show their hopes for their children through gift-giving. Using examples from this article and from your own life, discuss how parents show their hopes through the gifts they give. Lesson 1: Deconstructing the Prompt and Preparing a Thesis Most writing for academic classes is in response to a specific assignment or prompt (Horowitz, 1985; Johns, 1985). One of the most importantand perhaps the most difficulttasks for the academic writer is to understand what the prompt writer wants. For this reason, my class began by deconstructing the prompt (Carlson, 1985) in order to analyze better the directions and limitations of their assigned task. Academic prompts often have a number of instructions and key task-related terms, such as list and describe (Swales, 1982), and they often indicate the form, content, and strategies of the assignment (Horowitz, 1986). Therefore, it is wise to give students a number of prompt types so that they can develop strategies for successfully deconstructing a variety of task instructions (Flower, 1985). Because they had had little experience with deconstructing COHERENCE AND ACADEMIC WRITING 253

prompts, I assisted the students in answering the following questions: 1. What is the function of the first sentence in the moment? What is the prompt writer asking you to do, if anything? The students decided that the first sentence was just a statement of the thesis of the Mead and Metraux essay, creating a context for the instructions which were to follow. 2. What does the second sentence tell you about your writing task? What does it tell you about the required aims or strategies for writing? The students decided that the writer of the prompt told them to discuss, using examples from their own lives and from the reading. They decided that discuss was a general word which did not tell them much about structuring their argument. However, they had specific instructions about aims and strategies: They were to support their argument by using examples from this article and their own lives. 3. What does the prompt tell you about the focus of the content? The students decided that hopes was the central term and that it must appear or be implied in their thesis sentences and be the central topic of the essays they were to produce. Once the prompt had been deconstructed and students understood the directions of the writer of the prompt regarding aims, strategies, and content, we went on to develop the discourse theme, generally called the thesis in writing textbooks. In the prose of experienced writers, the thesis can be either explicit or implicit (see Lautamatti, 1986; Witte, 1983). However, requiring an explicit thesis is useful for inexperienced writers because it provides guidance as they organize and redraft their essays. To make the theses their own, not just a repetition of the thesis from the article, the students were asked to do some divergent thinking through the use of invention strategies (Daubney-Davis, 1982; Spack, 1984), such as clustering and listing, all based upon the central idea of parents showing hopes through gift giving. One possibility for approaching thesis building was to think of the gifts which parents might give (both mentioned in the article and occurring in their own lives), then decide what hopes these gifts represent. Another, of course, was to think of a possible thesis, then come up with examples appropriate to it. Still another was to think about the hopes of parents, then to discover gifts which express 254 TESOL QUARTERLY

these hopes. Though I provided examples of theses and made suggestions for how students might go about planning their own, thesis development became an individual, divergent-thinking activity, for students solve problems of thesis formation in a number of different ways (Flower, 1985; Williams, 1985). When the students had developed some tentative theses, by whatever method, they were asked to test the theses against the requirements of the prompt, then prepare essays based upon these theses for the next class. Lesson 2: Examining a Thesis and the Relationships Among Assertions in an Essay On the second day, one student essay was examined by the entire class, acting as readers and text analysts. The essay examined, by a student whom I will call Yoko, was chosen because it was considered typical of the essays which had been written for our classes, in terms of response to prompts and topic development. The students were assigned to permanent groups of four which assembled each time an essay was examined in this manner. These groups were organized according to diverse linguistic backgrounds and proficiency levels, so that each group would, by necessity, speak English and would have at least one highly proficient member to lead discussions. The sample essay (see Appendix A) and the questions for analysis, given below, were distributed and discussed by the groups. 1. Is the thesis in the paper appropriate for the prompt provided? The students had decided that the central proposition of the prompt was how parents showed their hopes through gift giving. Therefore, the thesis was considered appropriate, since Yoko planned to show that the gifts are different from each other. (The students assumed that these differences are based on hopes.) 2. What does the thesis pre-reveal to the reader? Does it reveal the writers argument and the organization which the argument will follow? Perhaps because this prompt does not require a particular organizational framework, Yoko did not indicate how her essay would be organized. She did pre-reveal content and argument, however, when she said that the gifts are different from each other. Thus, the students decided that the thesis was present, complete, and adequate to the prompt and that it revealed the COHERENCE AND ACADEMIC WRITING 255

content of the student paper but not the form in which the content would be organized. Once they had considered Yokos thesis, the students turned to the development of topics within her essay. Coherence, the students were reminded, depends not only on the introduction of a clear thesis, but on topic hierarchies and relationships, all of which relate back to the thesis itself (see, e.g., Lautamatti, 1986; Witte, 1983). This discussion led to the next question: 3. What are the relationships among the assertions? In their groups, students were asked to determine the gist of Yokos paper and the relationship among assertions, by writing a single sentence summary of each paragraph in the essay. Here is an example of a summary from one group of students: Thesis: And according to its hope the gifts are different from each other. Para 2: Different gifts for different people have different meanings. (implied) Para 3: The most difficult thing is what kind of thing or how to give rather than what to give. (stated) Para 4: An example of hopes for giving is my fathers book buying. (implied) The students were then asked questions which explored the topic relationships and the relationship breakdowns in the text: Which paragraphs were most difficult to summarize? Why? Did you have difficulty understanding the relationships among topics in the paragraphs? How could the author make these paragraphs easier to summarize, for example, by showing the relationship with the thesis or by providing topic sentences? The students made a number of comments about the summarizing process which demonstrated their sensitivity to the relationships among the topics within the paragraphs and to the thesis. They mentioned, for example, that in Paragraph 2 the author might demonstrate a closer relationship with the thesis if she were to provide a topic sentence which repeated the key word different and showed explicitly how these gifts were examples of different hopes. They suggested that this paragraph begin with a topic sentence such as, Different gifts may mean different things. They had more trouble with Paragraph 3 because they could not understand how the difficult thing had anything to do with differences among gifts, which was the key phrase in the thesis. They suggested that this paragraph, too, should be devoted to differences among gifts. They asked the writer to show how the 256 TESOL QUARTERLY

examples in this paragraph were somehow distinct from those found in Paragraph 2. Further problems arose when they considered Paragraph 4. The example in this paragraph, unlike those which had been taken directly from the Mead and Metraux essay, was personal. But was it an example of differences? Some students suggested that a possible topic sentence for both Paragraphs 3 and 4 might be, Even when the same kind of gift is given, it may have various meanings. They also suggested the addition of a conclusion to restate the thesis and comment upon the points made in the internal paragraphs. After discussing topic relationships and topic breakdowns, the students turned to a reader-based consideration. 4. How do you think the ideas presented in Yokos essay would affect the reader, a native-speaker teacher? Noting that a number of details from Mead and Metraux were included in Yokos paper, the students concluded that the ESL teacher, who already was familiar with the assigned essay, would be much more interested in details from Yokos own life. They suggested lengthening the personal experience section or integrating it with information from Mead and Metraux. Each paragraph should be longer, they pointed out, with more details about each example. These comments are consistent with Wittes (1983) findings that passing essays have fewer topics and more topic support. After these two lessons were completed, all members of the class took their first drafts home and revised them, using as a guide the questions which had been posed for Yokos essay in class. Lesson 3: Examining the Information Structure In Lesson 3, Yokos paper was again employed for class discussion, this time in second draft form (see Appendix B). In this lesson, students were to discuss the information structure, consisting of cohesion and other features which lead the reader through the text (i.e., meta-discourse). Most information structure is reader based; however, some features, such as the cohesive ties discussed in Halliday and Hasan (1976), are considered to be text based. To develop an understanding of these text-based features, students were asked to take a look at the cohesive ties in Yokos essay. They were asked the following questions: COHERENCE AND ACADEMIC WRITING 257

1. Did the author link sentences through use of vocabulary? Are there related words which appear throughout the paragraph? How are these words related (by synonymy, as superordinates/ subordinates, etc.)? Is the linking of vocabulary successful, or are there words which do not fit? 2. What reference items are used? Does the writer use this, the, or it to provide a tie with earlier sentences? Are the reference items appropriately used? Do they lead you through the text? 3. What types of conjunctions are there? A group of students was assigned to each set of questions. One group was assigned to go through Yokos essay to find related words, to suggest other words which might be used as synonyms, and to see that word relationships were carried through the text. A second group of students went through the paper to identify the ties between the pronouns of reference and the words to which they referred. They were warned that indefinite reference is a problem for many students (see Johns, 1980b). The third group of students went through the paper to find and evaluate connectives introducing independent and dependent clauses. Once students had examined and commented on cohesive items, they turned to features which are reader based, that is, the metadiscourse items discussing information in the text. Vande Kopple (1985), who has an excellent taxonomy of meta-discourse items, quotes Williamss definition of meta-discourse as writing about writing, whatever does not refer to the subject matter being addressed (p. 84). The Vande Kopple taxonomy includes narrators (e.g., according to Mead and Metraux), validity markers (e.g., may), topicalizers (e.g., for example), and reminders about what is discussed earlier and later in the text, After being given a list of meta-discourse items and examples, the students were asked to evaluate the essay for these features. They were given the following questions: 4a. What meta-discourse items appear in this essay? Are they effectively used? The students found topicalizers (as for hopes), validity markers (might and can), and an allocation marker (for example). Most decided that the meta-discourse items were effectively but sparingly used. b. What other items might be necessary, considering the prompt and the information in the essay? 258 TESOL QUARTERLY

The students decided that the major problem with this draft of the essay was that there was little indication of which examples or quotations came directly from the Mead and Metraux essay and which were Yokos own examples. They recommended narrators (e.g., according to Mead and Metraux) when examples from the essay appeared. After the students had completed exercises covering this unit, in which reader-based and text-based questions were asked about the prompt, the thesis, the relationships among the topics, and the information structure, they revised their own papers in the same manner, edited them, and turned them in for a grade. CONCLUSION At first, students have some difficulty with this unit, since they have had little experience with reconstructing prompts and evaluating their own writing. Therefore, I carefully take them through the entire process, explaining the questions and suggesting answers. As the semester progresses, the students are exposed to additional student essays and become more adept at the groupediting process. Clearly, this is not a unit which can be done once and forgotten; it must be repeated with a variety of prompts and a number of student essays. It then becomes the organizational structure for a writing course with the goal of improving coherence in essays. If consistently employed, this approach using revision to improve coherence is successful for several reasons. 1. The approach considers coherence to be both reader based and text based. Traditionally, coherence (generally cohesion) was thought to be text based; revisions were based upon incongruities and errors in the written text (Hodges & Whitten, 1972). Recently, the emphasis has been placed on reader-based coherence (Carrell, 1982). Yet an approach to teaching need not be one or the other; it should, in fact, include both reader- and text-based considerations. 2. The exercises integrate a number of important features of coherence. For example, whereas many writing textbooks still consider cohesion, as defined by Halliday and Hasan (1976), as central to text-based coherence, meta-discourse is introduced here as also necessary for leading the reader through the text. 3. One of the emphases in the process approach to writing has been COHERENCE AND ACADEMIC WRITING 259

on asking students to be concerned with meaning before imposing form upon their written work. Zamel (1984) notes: As students continue to develop their ideas in writing, considerations of organization and logical development come into play. The question, then, is not of choosing to attend to organization or not, but of when and how to do so. (p. 154) In the approach to academic writing discussed in this article, form, content, aims, and strategies are often integral to the prompt. Therefore, students must be able to reconstruct the prompt before planning their writing. This restricts their creativity, of course, but results in a product which is more acceptable to the grader who wrote the prompt and is therefore more realistic for the academic milieu. In fact, since we usually write for an audience, this approach may be appropriate for any writing task. 4. This unit provides sequenced questions and activities which go from the top down. The students must answer the first set of questions before they can go on to the next lesson. They cannot be poor revisers (i.e., edit on the sentence level only) if they answer the questions and revise their essays as suggested. 5. And last but not least, when using this approach, the teacher does not see the students papers until they have redrafted them and edited them at least twice. Rather than relying on teacher correction, the students devote more time to monitoring their own work and to providing an audience for their peers. The teacher sees a more finished product, which cuts down considerably on grading time. Defining and teaching coherence are difficult tasks. However, my students and I have found that if we break the tasks into manageable, task-dependent parts, we can achieve more success in the writing classroom.

The author would like to thank Ulla Connor, Bill Grabe, Dan Horowitz, and two anonymous TESOL Quarterly reviewers for their useful comments on earlier versions of this article.



Ann M. Johns is Associate Professor of Academic Skills and Linguistics at San Diego State University. She is the co-editor of the ESP Journal and has published in CATESOL Occasional Papers, Language Learning, The Journal of Basic Writing, and elsewhere.

Armbruster, B. B., & Anderson, T.H. (1984). Producing considerate expository text: Or easy reading is damned hard writing (Reading Education Rep. No. 46). Champaign: University of Illinois, Center for the Study of Reading. Bander, R.G. (1978). American English rhetoric. New York: Holt, Rhinehart and Winston. Carlson, P.A. (1985, November). Procedural knowledge required in inclass essays. Handout discussed at the Seminar on Cognitive Frameworks and Higher Order Reasoning, University of Chicago. Carrell, P.L. (1982). Cohesion is not coherence. TESOL Quarterly, 16, 479488. Carrell, P.L. (1983). Some issues in studying the role of schemata, or background knowledge, in second language comprehension. Reading in a Foreign Language, 1, 81-92. Conner, U. (1984). A study of cohesion and coherence in English as a second language students writing. Papers in Linguistics: International Journal of Human Communication, 17, 302-316. Conner, U., & Farmer, M. (1985, March-April). The teaching of topical structure analysis as a revision strategy: An exploratory study. Paper presented at the AERA Conference, Chicago. Daubey-Davis, A.E. (1982). Using invention heuristics to teach writing. Part II: A look at methods. TECFORS Newsletter, 5 (3), 1-3. Flower, L. (1985, November). Problem-solving strategies for reading to write. Paper presented at the Seminar on Cognitive Frameworks and Higher Order Reasoning, University of Chicago. Grabe, W.P. (1984). Towards defining expository prose within a theory of text construction. Unpublished doctoral dissertation, University of Southern California. Grabe, W.P. (1985). Written discourse analysis. In R.B. Kaplan (Ed.), Annual review of applied linguistics (Vol. 5, pp. 101-123). Rowley, MA: Newbury House. Halliday, M.A.K. (1978). Language as a social semiotic. Baltimore: University Park Press. Halliday, M. A. K., & Hasan, R. (1976). Cohesion in English. New York: Longman. Hodges, J. C., & Whitten, M.E. (1972). Harbrace College Handbook. New York: Harcourt Brace Jovanovich. COHERENCE AND ACADEMIC WRITING 261

Horowitz, D. (1985). Process not product: Less than meets the eye. Unpublished manuscript, Western Illinois University, International Programs, Macomb. Horowitz, D. (1986, March). What professors actually require: A study of academic writing tasks. Paper presented at the 20th Annual TESOL Convention, Anaheim, CA. Hughey, J. B., Wormuth, D. R., Hartfiel, V. F., & Jacobs, H.L. (1983). Teaching ESL composition: Principles and techniques. Rowley, MA: Newbury House. Huizenga, J., Snellings, C. M., & Francis, G.B. (1982). Basic composition for ESL: An expository workbook. Glenview, IL: Scott, Foresman. Jacobs, S. (1982). Composition and coherence: The writing of eleven medical students. Washington, DC: Center for Applied Linguistics. Johns, A.M. (1980a). Cohesion in written business discourse: Some contrasts. The ESP Journal, 1, 35-44. Johns, A.M. (1980b). Preventing global discourse errors: Approaches to ESOL writing. CATESOL Occasional Papers, 6, 65-70. Johns, A.M. (1985). Genre and evaluation in general education classes. Unpublished manuscript, San Diego State University, Academic Skills Center. Johns, A.M. (in press). The ESL student and the revision process: Some insights from schema theory. Journal of Basic Writing. Johnson, J.A. (1983). Writing strategies for ESL students. New York: MacMillan. Johnson, P. (1982). Effects of reading comprehension on building background knowledge. TESOL Quarterly, 16, 503-516. Kantor, R.N. (1985, April). Topic management: Problems of sentence boundaries, transition and reference. Paper presented at the 19th Annual TESOL Convention, New York City. Kroll, B. (1983, November). Bad essays in good English and good essays in bad English: Implications for the teaching of writing. Paper presented at the Fifth Los Angeles Second Language Research Forum, Los Angeles. Lautamatti, L. (1986). Observations on the development of the topic in simplified discourse. In V. Conner & R.B. Kaplan (Eds. ), Writing across languages: Analyzing L2 texts (pp. 92-125). Reading, MA: AddisonWesley. Markels, R.B. (1983). Cohesion paradigms in paragraphs. College English, 45, 450-474. Martin, L.J. (1974). The five-hundred word theme. Englewoods Cliffs, NJ: Prentice-Hall. Mead, M., & Metraux, R. (1984). The gift of autonomy. In G. Levin (Ed.), Prose models (6th ed.) (pp. 304-309). New York: Harcourt Brace Jovanovich. Miller, J. R., & Kintsch, W. (1980). Readability and recall of short prose passages: A theoretical analysis. ]ournal of Experimental Psychology: Human Learning and Memory, 6, 335-353. 262 TESOL QUARTERLY

Nold, E., & Davis, B. (1980). The discourse matrix. College Composition and Communication, 31, 141-152. Raimes, A. (1983). Techniques in teaching writing. New York: Oxford University Press. Rice, M. K., & Burns, J.U. (1986). Thinking/writing: An introduction to the writing process for students of English as a second language. Englewood Cliffs, NJ: Prentice-Hall. Rumelhart, D. (1977). Toward an interactive model of reading. In S. Dornic (Ed.), Attention and performance (Vol. 6, pp. 33-58). Hillsdale, NJ: Lawrence Erlbaum. Scarcella, R. (1985, May). Cohesion and coherence in writing development of native and non-native English speakers. Paper presented at the Eighth Annual Linguistics Colloquium, San Diego State University. Sheehan, T. (1986). Comp One! An introductory composition workbook for students of ESL. Englewood Cliffs, NJ: Prentice-Hall. Sommers, N. (1980). Revision strategies of student writers and experienced adult writers. College Composition and Communication, 31, 378-388. Spack, R. (1984). Invention strategies and the ESL college composition student. TESOL Quarterly, 18, 649-670. Swales, J. (1982). Examining examination papers. English Language Research Journal, 3, 9-25. (Department of English Language and Literature, University of Birmingham, England) Vande Kopple, W .J. (1985). Some exploratory discourse on metadiscourse. College Composition and Communication, 36, 82-93. Williams, J. (1985, November). Writing and knowing: A pragmatic interpretation of development and critical thinking. Paper presented at the Seminar on Cognitive Frameworks and Higher Order Reasoning, University of Chicago. Witte, S.P. (1983). Topical structure and writing quality: Some possible text-based explanations of readers judgments of student writing. Visible Language, 17, 177-205. Witte, S. P., & Faigley, L. (1981), Coherence, cohesion and writing quality. College Composition and Communication, 32, 189-204. Zamel, V. (1983). The composing processes of advanced ESL students: Six case studies. TESOL Quarterly, 17, 165-187. Zamel, V. (1984). The author responds . . . (in the Forum). T E S O L Quarterly, 18, 154-157.

APPENDIX A The First Draft of Yoko (Lesson 2) Every gifts from parents to their children carries with it love, hope and expectation for children. As for the hopes, there are various kinds of hopes: to be autonomous, to become like studying, to grow up favorably, to be an COHERENCE AND ACADEMIC WRITING 263

honest child and so on. And according to its hopes the gifts are different from each other. For example, to give a girl a diary with key might mean that the parents hope her to grow her sense of identity and independence. And giving a boy a desk may mean that they hope him to foster his sense of personal privacy. To give money to go to college is to give opportunity to be come educated. More difficult thing as to gift-giving is this: what kind of thing or how to give rather than what to give. As for to give a doll to a girl, there are many kinds of dolls and according to how extend of autonomy parents expect to the child the choice will be different. And as for to giving money, there are also some choices. For example, only to give money saying nothing or to give it with telling him to buy a thing, etc. When I was a child, my father used to take me to a used book store. He recommended me several books and bought me some books I chose. I learned by doing so pleasure of reading and to use everything with care not only books. And I believe he wanted me to learn such things. I think its a good idea to go and choose something with children. It shows that you love them and worry about them. It is easy way to show children hopes of the parents.

APPENDIX B The Second Draft of Yoko (Lesson 3) Every gifts from parents to their children carries with it love, hope and expectation for children. As for hopes, there are various kinds: to be autonomous, to become like studying, to grow up favorably, to be an honest child and so on. And according to its hope the gifts are different from each other. Different gifts may mean different things in terms of the parents wishes. For example, to give a girl a diary with a key might mean that the parents hope she will grow in a sense of identity and independence. And giving a boy a desk may mean that they hope him to foster his sense of personal privacy. To give money to go to college is to give opportunity to become educated. I was given money to go to college; my parents hope that I will become very smart and wise and be able to think for myself. The same kind of gift may have various meanings for the parents and the children. There are many kinds of dolls you can give, for example. Choosing a doll for a little girl, do I buy her a perishable costume doll for which she will make dresses out of the materials I also give her? The costume doll can perhaps be dressed and undressed, but that is all. A sturdy doll with a ready-made wardrobe places choice in the childs own hands. She herself can dress and undress it, bathe it safely and decide whether the little girl will wear pink or blue, plaid or plain. Giving money



can also have many meanings. Parents can give money and say this is yours, or they can give money for specific things to buy or to be saved. My father gave me books for many reasons. When I was a child, my father used to take me to a used book store. He recommended me several books and bought me some books I chose. I learned by doing so pleasure of reading, use of reading for study and how to take care of books. He hoped that I would love books, take care of them and use them for my study. So parents show hopes for their children by giving them different kinds of gifts or the same types of gifts but with different meanings. When I become a parent, I will probably do the same thing, because I will have many hopes for my children, just like my parents.



TESOL QUARTERLY, Vol. 20, No. 2, June 1986

The Meaning and Discourse Function of the Past Tense in English

Ball State University

The past tense is often used inconsistently, even by very advanced ESL learners. In this article, it is suggested that a major cause of the problem is a failure by teachers and students to recognize the extent to which a speakers point of view and purpose in performing a speech act condition the choice between the present and past tenses in actual discourse, Following an analysis of the meaning and discourse function of the past tense, suggestions are offered for teaching and practicing this tense in context.

Even very advanced students often use the past tense inconsistently, despite the fact that this tense appears to have a simple and readily explainable meaning. This is true not only of speakers of languages without past tenses per se, such as Chinese or Indonesian, but also of Korean and Japanese speakers, for example, whose languages do have past tenses. A key source of the difficulty may be that the past tense is generally taught as having a completive sense, while a more general meaning and discourse conditions on its use go unrecognized. 1 Specifically, it is argued here that the past tense simply means true before speech time and that completion is not part of its denotative meaning, although it is an implication often associated with the past tense in many contexts. Thus, the past tense can be used to describe situations which may still exist, objectively
1 Although

this article discusses only the simple past and present tenses, the generalizations apply to the corresponding progressive tenses as well. For this reason, I use the labels past and present rather than simple past, etc. Technically, however, the progressive tenses are actually combinations of a simple tense plus a progressive aspect form. The counterfactual use of the past tense, as inIf I were you, is not discussed in this article. Although some linguists have proposed unitary accounts of the past tense which include this use, I feel that these treatments are too abstract to be suitable for pedagogical purposes and that this use is best covered separately. Also excluded from the discussion is the alternation between the past and the historical present. See Wolfson (1979) and Schiffrin (1981) for analyses of the latter.


speaking, at the time of speech if the purpose of the speaker2 is to present the information in terms of its past psychological relevance. As shown below, this approach makes it clear that past tenses said to result from tense-harmony (also called sequence of tenses and backshifting) copying rules really occur for the same semantic and pragmatic reasons as do main clause past tenses. First, common assumptions about the use of the past tense are briefly identified, followed by a demonstration that the speakers point of view and communicative purpose play a crucial role in the selection of the past tense in main clauses. It is then argued that the same factors condition the choice of tense in indirect speech clauses as well and that therefore tense harmony per se does not exist. Next, it is suggested that the past tense should be analyzed as meaning true in the past and that the notion of completion is a contextually based implication. The final sections of the article show how this analysis explains the occurrence of a particular type of error in past tense use by advanced ESL learners. Teaching suggestions are also offered. COMMON ASSUMPTIONS The completive sense of the past tense, as illustrated in Example 1, is the one most commonly recognized by teachers and most consistently used by advanced students.3 la. I had a VW, but I sold it. b. That was my pudding she ate. In Example la, the speaker no longer owns the car, and the act of selling is over. In Example lb, the eating has been brought to completion, and the pudding no longer exists. These are examples of what may be thought of as the prototypical use of the past tense. ESL teachers and textbooks, however, generally overlook the fact that the past tense lacks a sense of completion in other contexts. 2. [Leaving a movie theater] That was a great movie. Although the movie continues to be great, it is described in the past. A second common but mistaken assumption about the past tense is that completed acts are described only in that tense. This
convenience, the terms speaker and speaking are used to cover writer and writing as well. Speech time also refers to time of writing. 3 Examples without a source specified were constructed by the author. Those labeled spontaneous speech were naturally occurring utterances recorded by the author. The remaining examples were taken from books and television as indicated. Italics in all examples have been added by the author for emphasis.
2 For



assumption is particularly reinforced when the past and present perfect tenses are compared and contrasted. As Example 3 shows, however, a verb may occur in the present tense, even though the actual action denoted by the verb in its general sense took place in the past. 3. Werner (1948) writes about primitive languages in the following terms. (Johnson-Laird & Wason, 1977, p. 439) Finally, it is generally thought that the present tense in an indirect speech clause embedded below a past tense main verb is converted to the past tense. The operative rule is that the tense of the main verb is copied. This particular use of the past tense is thus considered not to reflect the true meaning of the tense. An example of this so-called tense harmony or sequence of tenses is given below. 4a. Jane said, I am tired. b. Jane said she was tired. Was in Example 4b is in the same tense as the reporting verb said, which indicates that the act of saying and the state of fatigue were simultaneous. THE PAST TENSE IN DIRECT DISCOURSE In addition to the objective temporal relationship between the time of a situation (this is intended to include events and states of affairs) and speech time, the speakers point of view and purpose in communicating play a crucial role in the choice between the past and present tenses.4 A situation whose time frame extends from the past to the present may be described in the past rather than the present or present perfect tenses if the purpose is to present information or ask a question from a past point of view. The specific reasons why a speaker would have a past point of view fall into two major categories. 1. Past association: The fact or nature of a persons association with a particular situation in the past is more relevant to the purpose
4 Reichenbach

(1947) points out the importance of point of view in the determination of tenses. He defines tenses in terms of the relations between the point of event (E), point of reference (R), and point of speech (S). For example, in English, all three points are simultaneous for the present tense, while R and E are simultaneous with each other and precede S for the simple past. This system is not used here, however, because it wrongly implies that all simple past tense verbs describe completed events (i. e., not in existence at speech time). Lakoff (1970) was perhaps the first to draw attention to the role of speaker viewpoint in the use of the noncompletive past. See Riddle (1978) for a theoretical discussion of her analysis.



in speaking than the objective current existence of that situation. (The person in question may be the speaker or another person being described.) In this case, the past tense functions as an indicator of subjective attitude.5 2. Background information: Although the information to be presented is about a situation subjectively viewable as existing in either the present or the past (as above), this is considered to be background to other information whose present existence is to be emphasized. 6 Here the past tense functions as a discourseorganizing device which backgrounds information. Past Association This category comes into play when a personal past experience of a situation is presented as the motivation for a past action, as in Example 5. 5. [The speaker, gesturing with an old potholder in her hand at a flea market, is explaining why she just bought it.] It didnt have any stains on it. (spontaneous speech) Even though the potholder described here has no stains on it at the moment of speech, the speaker uses the past tense to show that its lack of stains was the motivation for its purchase, an act which has already taken place. Another aspect of past experience contributing to a past point of view is the time a piece of knowledge was acquired. In Example 6, a diver employed by the police to locate a submerged car describes the body he finds in terms of the moment of discovery. 6. Diver: [Just sticking his head out of the water] Theres a car down here all right. Brand new convertible. Tragg: Anybody in it? Diver: Fellow who was driving. He looked like a minister. (TV dialogue, Perry Mason) It is important to note here that the mans appearance as a minister was referred to for the purpose of identification, not as a comment on his personal attributes. For this reason, the fact that he is dead is not the critical factor in the choice of the past tense in this particular exchange. The present tense is also possible in Example 6 but would
5 Ard (1984) independently proposes some similar conditions on the use of the past tense in scientific writing but does not relate them to the use of the past tense in general or to the distinction between foregrounding and backgrounding. 6 This use of the past tense is mentioned by Aristar and Dry (1982) and Wallace (1982), among others.



indicate a different point of view; the connection between the mans appearance and the moment of discovery would not be emphasized. One might think of the use of the present tense here as less egocentric than the use of the past. Likewise, the past tense may be used in questions to focus on the addressees experience of visiting a place rather than to request an objective description of the person or thing in question. 7. How was Poland? The present tense is appropriate in such a question about the addressees experience only in a letter or phone call while he or she is still in the place referred to. The same emphasis on past experience can be seen in Example 8, in which the speaker wants to describe an interesting sight experienced before the moment of speech. 8. [Said a few moments after a man passing in a car has driven out of sight] He did have an earring in his ear. (spontaneous speech) The present tense, on the other hand, would emphasize the present characteristics of the man himself. Since the man is not in view at the time of the speech act and since the speaker has no current relationship with the man, the past tense is used. Use of the present tense in such contexts tends to imply the existence of a current relationship between the speaker (or hearer) and the person described. 9. Anne: Jane just bought a Volvo. John: Maureen has one. Anne: John, youve got to quit talking about Maureen as if you were still going together. You broke up three months ago. The use of the present tense in this example wrongly implies a continuing relationship with Maureen on the part of the speaker. The speaker uses the present tense, however, because he is still emotionally attached to his former girlfriend and is reluctant to break the psychological tie. Another type of context in which information objectively known to be true in the present may be expressed in the past tense is when the speakers purpose is to attribute a belief to a deceased person rather than to present the belief as the speakers independent assertion. (The asterisk in the example below indicates that revolves is not acceptable here.) revolved 10. According to Copernicus, the earth {*revolves} around the sun, and nothing could persuade him otherwise. THE PAST TENSE IN ENGLISH 271

If a name can refer either to a dead person or to an existing work by that person, in conversation the past tense more strongly suggests the person and the present tense, that persons work.

When the sentence or the context unambiguously refers to a dead person, however, only the past tense is possible. 12. Einstein thought that . . . Background Information When it is logically possible to use either the past or the present tense, according to the conditions discussed above, the past may be used as a backgrounding device and the present as a foregrounding device. For example, when a claim from an existing work is cited as background information to a study, perhaps to offer a historical perspective, the past tense is preferred. 13. Bartlett (1932), the earliest of the theorists discussed here and the first psychologist to use the term schema, in effect said it all: . . . the past operates as an organized mass . . . (Tannen, 1979, p. 139) When an existing work is discussed in terms of the current validity of its claims and their bearing on the main point to be developed, the present is usually chosen, thereby foregrounding the information. 14. Bartlett contends that an individual has an overmastering tendency simply to get a general impression of the whole. (Tannen, 1979, p. 139) Example 15 illustrates the difference between backgrounded and foregrounded information within a single paragraph. 15. In November 1859, Charles Darwins The Origin of Species, one of the greatest and most controversial works in the literature of science, was published in London. The central idea in this book is the principle of natural selection, In the sixth edition, which appeared in 1872 and which Darwin regarded as the definitive one . . . Darwin wrote: This principle of preservation or the survival of the fittest, I have called Natural Selection. (Eigen & Winkler, 1981, pp. 53-54) In this passage, was is in the past because it refers to a completed event and because the sentence provides background information. The verb in the next sentence, however, is in the present because 272 TESOL QUARTERLY

Darwins book still exists and because the principle of natural selection is the main topic in this section of Eigen and Winklers book. The first condition must be fulfilled in order for the present tense to be a possible choice. Given this choice, the present may then be chosen for the purpose of foregrounding. Appeared in the third sentence also refers to a completed event, and regarded is in the past tense because it refers to the personal view of a deceased individual. Wrote occurs in the past because it represents information which was true before speech time and which is background matter with respect to the main topic in the section of the book from which Example 15 is taken. That is, the quote from Darwin and the fact that it was written in a certain edition at a certain time constitute historical information and provide support for Eigen and Winklers more primary point expressed in the present tense in the second sentence. For the past tense to be selected as a backgrounding device, the noncompletive predicate must represent information that was true before speech time. Descriptions of experiments or empirical studies which focus on their completion in the past, as in Example 16, also tend to constitute background information, and the past tense is therefore preferred in such contexts. 16. An experiment was carried out to determine whether . . . But when reference is made to something that continues to exist in the present and has immediate current relevance, such as to the results of an experiment or to data being reported, then the information tends to belong to the foreground, and the present tense is usually preferred. 17. The results of this experiment show that . . . Finally, the present tense is usually used in an abstract of a paper because it represents foregrounded information as an independent summary of the paper as it exists in its entirety in the present. 18. This paper demonstrates that sound changes do not always affect the most frequent words first. (Phillips, 1984, p. 320) When reference is made in a later section to something said in an earlier part of a paper, on the other hand, the past tense is usually preferred, as in Example 19. 19. As pointed out earlier, literary and ethnographic methods in folklore research complement each other. (Ben-Amos, 1981, p. xxxvi) The past tense is preferred in later references because the writer is describing the paper in terms of the process of writing and of THE PAST TENSE IN ENGLISH 273

making the claims rather than in terms of an entity existing independently and in the present for the reader. (Summaries and conclusions are often in the present perfect for reasons which go beyond the scope of this paper.) When the point of an entire section or a whole work is to present historical information, however, then within that context the past occurs in its completive use or only as a marker of past point of view rather than as a backgrounding device. Up to this point, the discussion has been restricted to the use of the past tense with main clause verbs. Let us now turn to instances of the past tense traditionally thought to be cases of tense harmony. THE PAST TENSE IN INDIRECT DISCOURSE Tense Harmony The conversion from direct to indirect speech is often associated with a change in the tense of the reported speech verb. For example, the past tense of a main clause reporting verb is believed to trigger a change in the indirect speech clause: A tense reflecting real time reference changes to a tense maintaining the relevant temporal relationship with respect to the reporting verb (see Example 4b). Many ESL teachers and textbook writers believe that the change from present to past is automatic and exceptionless. Others, such as Aronson (1984) and Azar (1985), recognize that the present tense can be used in indirect speech clauses embedded below past tense main verbs but mistakenly attribute this to informal style. Frank (1972) goes further in recognizing that the present tense may be retained in a that clause object expressing a generalization (He said the train always arrives late.) (p. 62). In their ESL grammar for teachers, Celce-Murcia and LarsenFreeman (1983) present a more complete description. In addition to noting the point about general truths, they say that immediate repetitions and descriptions still true or considered still true by the speaker occur in the present tense. Citing Chang (1981), they also remark that the present tense may be retained in indirect speech clauses containing a present or future adverbial, such as in Jeremiah said he hopes to begin his new exercise program by early next week (p. 462). The best generalization, however, is that the present tense occurs in a reported speech clause embedded under a past tense verb if the situation described there is of current relevance to the speaker. The past tense in such subordinate clauses is motivated by the same 274 TESOL QUARTERLY

types of factors contributing to a past point of view as in the main clause examples discussed above.7 Current Relevance Versus Past Association Many types of situations involve current relevance. For instance, the present tense in Example 20 implies speaker belief in the truth of the information in the subordinate clause. 20. Columbus recognized that the earth is round. In Example 21, the present tense serves to emphasize the current firmness of the speakers plans. 21. I am sorry, she said, but I told Mr. Martin yesterday that we are not thinking of selling. (Wentworth, 1965, p. 23) In Example 22, the present tense implies that the situation described in the subordinate clause is unresolved. 22. The paper claimed today that the mayor is involved in that big tax fraud cover-up theyve been investigating. The past tense, on the other hand, indicates a past point of view. As the next set of examples shows, the noncompletive past tense in indirect speech clauses occurs under the same conditions as those determining this use of the past tense in direct speech. (The same type of analysis can also apply to the alternation between other pairs of tenses in indirect speech clauses.) Consider Example 23, in which the noncompletive past in the indirect speech clause reflects a past point of view on the part of the speaker, motivated by the same factors shown to be relevant for direct speech. 23. [Spoken about 15 minutes after it was discovered that there was no garlic] I couldnt make garlic butter because we didnt have any garlic. (spontaneous speech) The lack of garlic is given as a reason for a past act and is therefore described in the past, even though there is still no garlic at the moment of speech. In Example 24, although the speaker still thinks that Miss Marple has something to say, had is in the past tense because this conclusion was drawn on the basis of an observation of her demeanor before the moment of speech.
7 See Costa (1972), McGilvray (1974a, 1974b), and especially Riddle (1978) for detailed

theoretical discussions of the alternation between the present and past tenses in such contexts.



24. Have you got something you want to tell me, Miss Marple? Now why should you think that? You looked as though you had, said Davy. (Christie, 1968, p. 135) Another common context for framing a description in terms of the speakers past perceptions is when prior ignorance about a currently existing situation is indicated. 25. B. Im jolly lucky if I do Elstead in four hours. ...... c. I thought it was four hours regularly. (spontaneous speech recorded in Svartvik & Quirk, 1980, p. 293) The present tense rarely occurs in such contexts, not because of a need for tense harmony per se, but because its discourse function would conflict with the speakers communicative purpose. One of the functions of the present tense in an indirect speech clause is to indicate speaker belief in the truth of the information presented in that clause. The purpose of Speaker c, however, is to emphasize past lack of awareness of certain information. In Example 26, an unchanged geographical location is described in the past because the speakers presence in the past in a place where he believed justice always prevailed was the motivation for his behavior at that time. 26. [An innocent boy accused of murder is being questioned in a TV interview about why he had said something honest but selfincriminating to the police.] I thought that this was the United States and that nothing could happen to me. (spontaneous speech, 60 Minutes) The fact that the speakers only contact with a certain car was a brief glimpse is the motivation for the use of was in Example 27. 27. A: Did Jack say what the make of the hit-and-run car was? B: Yeah. He said it was a Ford Mustang. In Example 28, Janes ex-husband is described in the past because she and/or the speaker has no current relationship with him. Whether or not the man in still a gambler is left unspecified. 28. Jane said that her ex-husband was a pathological gambler and thats why she divorced him. If the speaker wishes to describe some aspect of a current relationship, however, the present tense is more appropriate, as shown in Example 29. 276 TESOL QUARTERLY

29. Jane said that her ex-husband is a pathological gambler and she really worries when he has the kids. When a subordinate verb presents background information with respect to a present tense main verb, it occurs in the noncompletive past tense. The same information could be described in the present tense in another context. Compare Examples 30a and 30b. 30a. Chomsky (1965) ascribes a filtering function to transformations. b. In his history of transformational grammar in America, Newmeyer writes that Chomsky (1965) ascribed a filtering function to transformations. Conversely, a subordinate verb occurring below a noncompletive past tense main verb can be in the present tense if it expresses information believed currently valid and essential to the main point (thus belonging to the foreground) and if the main verb expresses background information. 31. I would like now to illustrate furthur the need for filters by discussing briefly a number of cases from English grammar. . . . It was observed many years ago by Fillmore (1965) that indirect object NP based on for-prepositional phrases in general behave differently under passivization than indirect object NP based on toprepositional phrases. (Postal, 1972, p. 141) Thus, the same reasons for the use of the noncompletive past in direct speech clauses, where there is no model for tense copying, account for its use in indirect speech clauses. This conclusion means that there is no need for a rule of tense harmony. The past or present tense is chosen in indirect speech on the basis of its general meaning and discourse function. THE MEANING OF THE PAST TENSE Since the past tense commonly occurs in both the completive and noncompletive senses, the best denotation of the simple past tense may be simply true before speech time in the real world or in the speakers belief world, with the completive sense being determined by context and the meaning of the verb. Thus verbs denoting actions, such as read or swim, for example, usually carry a completive sense because of the nature of the activity, but verbs denoting states, such as be or have, will quite often be associated with a noncompletive sense. The final arbiter is the context, however. THE PAST TENSE IN ENGLISH 277

The fact that the past and present tenses may be joined by and supports this analysis. If the past tense included in its literal denotation no longer true, then the conjunction of the past tense and the present tense, which denotes true at speech time, would result in a contradictory sentence. Instead, the inclusion of the present tense in the second conjunct merely cancels the potential implication that something is no longer the case or that a person no longer exists.8 As Example 32 shows, a sentence with conjoined past and present tenses referring to the same proposition is not contradictory. 32. Wife: Husband: Do you think he could actually have fallen in love with Diana? Theyve only known each other 2 days. Why not? I flipped for you in just 24 hours. Of course you were the most beautiful girl in the worldand still are. (TV dialogue, Love Boat)

Although the completive sense is prototypical of the past tense and commonly occurs with many verbs, it does not occur in all contexts, and that sense may be canceled. When describing a situation which could exist at speech time, the speaker may use the past tense to focus on his or her experience or perception of that situation in the past, and there is no completive connotation. Conversely, one may choose to describe past acts or opinions in the present tense if they constitute information to be foregrounded as currently relevant. IMPLICATIONS FOR ESL/EFL A major source of the problem advanced students have in using the past tense consistently may be that they do not adequately understand its actual meaning and discourse functions as outlined in this article. Consider the dialogue in Example 33. 33. Susan: Mei-Li: Susan: Mei-Li: Did you do anything interesting during the break? Yeah, I went to the Grand Canyon with some of my friends. We drove and camped out on the way. It sounds like fun. It is.

In this dialogue, Mei-Li uses the past tense for completed actions but mistakenly switches to the present tense after the native speaker Susan uses the present. The problem is one of point of view. Susan
8 For a theoretical discussion of the nature of contextual implications and the cancelability

criterion, see Grice (1975) and Sadock (1978).



uses the present as a reflection of her evaluation of the trip at the moment she speaks. Mei-Li, however, should have continued to use the past tense to reflect the fact that her experience took place in the past. She incorrectly switches to a present point of view and uses a tense with a habitual sense in that context rather than continuing to describe the trip in terms of her own experience. The paragraph in Example 34 was written by a fairly advanced ESL student from Vietnam.9 34. While I was in the gym, I saw someone looks very much like somebody I know from my hometown. I was afraided to come and ask her if she is from my hometown or not. Anyway, I guessed she felt the same way, because not too long she came and asked about myself. Then I found that she is from my hometown. I could not be sure because she changed a little. However, we talked for awhile, then she had to leave. In this paragraph, whenever a verb denotes an act, state, or situation completed in the past, the past tense (indicated by italics), is correctly used (although the past perfect would be more standard for changed). However, when states which can be considered to exist in the present as well as in the past are described, the verbs (indicated by italics and underlining) incorrectly appear in the present tense. Adoption of the wrong point of view is also the source of the error in Example 35, taken from a very advanced Indonesian students paper. 35. [Opening line of an article review] The writer was of the opinion that the first thought of many language educators . . . The student mistakenly describes the writers opinion in terms of the time that it was expressed in writing, that is, the past. He failed to realize that an opinion held in the past may still be described in the present if expressed in an existing work with current relevance for the reader and if the information belongs to the foreground. The analysis presented in this article is not intended to account for all instances of past tense errors in interlanguage. For example, Godfrey (1980) proposes other discourse factors which can affect tense use, including distraction from maintenance of past tense continuity by extralinguistic details at episode boundaries as well as the intrusion of forms not related to the main topic continuity. In another discourse-based study, Kumpf (1984) examined the interlanguage of an individual speaker who had not had formal instruction in English. She concluded that tenses in general and the

am grateful to Jan Edwards for supplying me with this writing sample.



past tense in particular were generally omitted in references to completed actions in the foreground but were virtually always used for states and inconsistently used for noncompleted actions in the background. My conclusions appear to be inconsistent with hers in that I have found that the past tense is used in some interlanguage for completed actions but not for states and actions with both present and past time reference. However, Kumpf does not claim that other speakers have the same interlanguage system as that of her subject. More important, our analyses agree in finding the aspectual contrast between completed and noncompleted actions relevant to the analysis of interlanguage, both analyses providing some support for the claim that tense as a category is universally secondary to aspect. In addition, both analyses point to a relationship between backgrounding and tense selection, although in different ways and in different language systems. Finally, I agree with Wolfram (1985) that surface-level constraints such as the form of the past tense marking, phonological environment, and verb frequency play the primary role in determining the incidence of past tense use by many ESL learners. TEACHING SUGGESTIONS Various teaching activities can be used for both discovery and practice purposes. Most important, the past and present tenses must be discussed in context, using excerpts from novels, stories, newspaper articles, academic prose, transcribed natural conversation, television dialogue, and so on. Brief notices about recent research in magazines such as Psychology Today are good sources of examples of academic prose which can easily be presented to students in their entirety and which are of common interest. The newspaper, especially the school paper, is also an excellent source of examples, one which helps students to see that the distinctions being taught are part of everyday usage and are not grammatical oddities. Also, the subject matter is often of personal interest to them. Rather than presenting ready-made explanations to the students, one approach the instructor can take is to provide them with examples first and then ask them to formulate their own hypotheses. This encourages active analysis by the students of real language input and prepares them for a more complete explanation by the teacher. 280 TESOL QUARTERLY

The instructor can also offer explanations in terms of individual students experiences. For example, an instructor might say: 36. Mohammed, remember how when you first came to the U.S. you didnt want to eat hamburgers? In telling us about that experience, you could say, When I first came to the U.S. I didnt know that hamburgers were made from beef, not ham. Exercises can be prepared from excerpts collected by the instructor, with the present and past tense verbs from the original given only in the bare infinitive form. The students then change to the appropriate tense, based on the preceding and following discourse. Example 37 is drawn from the opening paragraph of an Agatha Christie novel (1934). 37. Mr. Satterthwaite sat on the terrace of Crows Nest and watched his host, Sir Charles Cartwright, climbing up the path from the sea. Crows Nest (be) a modern bungalow of the better type. It (have) no half-timbering, no gables, no (be) a excrescences dear to many a builders heart. It plain, white, solid building, deceptive as to size, since it (look). It (be) a good deal bigger than it (owe) its name to its position, high up, overlooking the harbor of Loomouth. Indeed, from one corner of the terrace, (be) a sheer protected by a strong balustrade, there (be) a drop to the sea below. By road, Crows Nest (run) inland and then mile from the town. The road (zigzag) high up above the sea. (p. 5) In this exercise, all blanks require the past tense. Although there is no tense contrast, it is a useful exercise in that students generally expect the present tense in such contexts. In addition, the exercise stresses the relationship between tense choice and maintenance of point of view. Other exercises can be devised in which the past and present tenses contrast. When students do not choose the tense of the original, it should be made clear whether their choice is impossible, and on what grounds, or whether it is a possible alternative description but changes the point of view or adds or subtracts connotations. Students themselves can also be asked to collect examples in context from speech or writing and to explain why the past or present tense was used in each case, perhaps coding their answers to generalizations presented by the instructor on a handout. This helps to build monitoring and analytical skills which enable learning beyond the classroom walls. Students can even be asked to prepare exercises based on these examples for each other. This helps them to

solidify their understanding of the tenses and can heighten their interest. Mini-dialogues with blanks for the relevant tenses can be composed by the instructor, with space provided for the students to give brief reasons for their choice of tense. Examples 38 and 39, which require the student to use a form of be in each blank, are typical of such exercises. in 38. a. Mary: Did you see Passage to India when it town? it any good? b. Bob: No, c. Mary: Yes, it fantastic. Its too bad you didnt get to see it. 39. Anne to Kathy: You really should go to see Passage to India before it leaves town. It a great movie. For oral or written practice of the noncompletive past, students can be asked to describe situations which would naturally call for this sense. For instance, they might describe a former teacher presumed still to be living but not encountered since the fifth grade, reactions to the scenery at a particular place visited 5 years ago, feelings during a moment of danger and how the surroundings looked at that moment, a former home not seen in some time, and so on. Examples 40 (by an Arabic speaker) and 41 (by an Indonesian speaker) are descriptions written by my students. 40. When I was a child, my father and my mother decided to visit my uncle in Egypt. . . . We started our trip in the morning. . . . We reached my uncles house in the evening on the same day. . . . On the next day my cousin and I went to the pyramids. They were very nice. By the way, I had never seen the pyramids before that day. 41 When I was a child. I lived in a small village, in the area of South of Sumatra; it was called Campang Tiga. In this village I went to elementary school when I was seven years old. . . . While I studied at this school, I had a teacher who was called Mr. Mohammed. He looked horrible and was very dogmatic in his opinions. He had a loud voice. The noncompletive and completive pasts can also be illustrated and/or practiced in the context of a role play in which, for example, one student takes the part of a police officer questioning another representing a mugging victim. The latter describes the events and the muggers appearance from a past point of view. A third student then issues a police bulletin, describing the clothing and other temporary characteristics of the suspect in the past tense and 282 TESOL QUARTERLY

permanent characteristics such as height in the present. The students can perform this spontaneously or prepare the dialogue beforehand. In the former case, the instructor can give the students feedback on their use of the tenses via hand signals. A sample role play, with minor variations in parentheses, is given in Example 42. 42. Mugger: Give me your purse (wallet). Victim: [Hands over purse. After mugger leaves, runs to phone and calls police.] Police: Muncie Police Station. Can I help you? Victim: Yes, I just got mugged. He (She) took my purse. Police: OK, Maam (Sir). Where and when did this take place? Victim: A few minutes ago, on the comer of Vine and 8th. Police: Did you see what the mugger looked like? Victim: Yes. He (She) was tall, had brown hair, and was wearing jeans and a red shirt. He (She) had a gun. Police: OK, Maam (Sir), well get right on it. ....... Police Dispatcher: [Making announcement to police cars via radio] The suspect is male (female), tall, has brown hair, and was last seen wearing jeans and a red shirt and carrying a gun. Obviously the situation (and its language) are not realistic in all details, but it dramatically illustrates the use of the past tense in addition to offering an opportunity for practice once the generalizations have been presented. This situation may be contrasted with another scenario in which the victim can still see the mugger down the block but must describe the mugger to someone who is helping him or her so that the suspect can be picked out from other people on the street. In this case, the victims initial description of the mugger will be in the present tense. Another possible role-play situation consists of two people riding in a bus or car and commenting on what they see, as in Example 43. 43. A: Did you see that house we just passed? B: No. What about it? A: It had a green and orange roof. It is also helpful to devise as many real situations in the classroom as possible to illustrate and practice the noncompletive past. For example, the instructor can arrange for someone unknown to the THE PAST TENSE IN ENGLISH 283

students to come into the class to give him or her a message. After that person leaves, the instructor elicits from the class details about the persons appearance, clothing, and so on in the noncompletive past by saying, Did you notice that woman who was just in here? Describe her to me. The instructor can ask questions such as What was she wearing? What color was her hair? and so on, if necessary. The contrasting use of the present tense can be illustrated and/or practiced by having someone the students know come into the class and by later asking them, for example, What is Mary wearing today? The instructor can also start a conversation with students about their studies, native customs, or past experiences and at various points animatedly exclaim, for example, I didnt know you and Keiko were in the same chemistry lab! or I didnt realize that the Vietnamese ate curries! Communicative and contextually based exercises such as these are of great help in raising students awareness of the past tense as it is actually used in discourse and may contribute to greater consistency in their use of the past tense.

This article, the writing of which was partially supported by funds from the Office of the Provost, Ball State University, is partly based on my 1978 Ph.D. dissertation, Sequence of Tenses in English, University of Illinois, Urbana-Champaign. Earlier versions were read at the INTESOL Conference, Bloomington, November 1984, and at the TESOL Convention, New York, April 1985. I am grateful to the reviewers and to Christopher Ely, Herbert Stahlke, and especially Paul Neubauer for their helpful comments and discussion.

Elizabeth Riddle is an Assistant Professor of English at Ball State University, Muncie, Indiana, where she teaches linguistics, ESL, and TESL methods. She has also taught in Poland as a Fulbright lecturer. Her primary areas of research are pragmatic and functional grammar.



Ard, J. (1984, March). The semantics of tense and aspect in written scientific discourse. Paper presented at the 18th Annual TESOL Convention, Houston. Aristar, A., & Dry, H. (1982). The origin of backgrounding tenses in English. In K. Tuite, R. Schneider, & R. Chametzky (Eds.), Papers from the eighteenth regional meetingChicago Linguistic Society (pp. 1-13). Chicago: Chicago Linguistic Society. Aronson, T. (1984). English grammar digest. Englewood Cliffs, NJ: Prentice-Hall. Azar, B. S. (1985). Fundamentals of English grammar. Englewood Cliffs, NJ: Prentice-Hall. Ben-Amos, D. (1981). Folklore genres. Austin: University of Texas Press. Celce-Murcia, M., & Larsen-Freeman, D. (1983). The grammar bookAn ESL/EFL teachers course. Rowley, MA: Newbury House. Chang, Y.-W. (1981). A contextual analysis of tense shifts in reported speech. Unpublished English 215 paper, University of California, Los Angeles. Christie, A. (1934). Murder in three acts. New York: Popular Library. Christie, A. (1968), At Bertrams Hotel. Glasgow: Fontana/Collins. Costa, R. (1972). Sequence of tenses in that-clauses. In P. Peranteau, J. N. Levi, & G. C. Phares (Eds.), Papers from the eighth regional meeting Chicago Linguistic Society (pp. 41-51). Chicago: Chicago Linguistic Society. Eigen, M., & Winkler, R. (1981). Laws of the game: How the principles of nature govern chance (R. & R. Kimber, Trans.). New York: Alfred A. Knopf. Frank, M. (1972). Modern English (exercises for non-native speakers), Part II. Englewood Cliffs, NJ: Prentice-Hall. Godfrey, D.L. (1980). A discourse analysis of tense in adult ESL monologues. In D. Larsen-Freeman (Ed.), Discourse analysis in second language research (pp. 92-100). Rowley, MA: Newbury House. Grice, H.P. (1975). Logic and conversation. In P. Cole & J.L. Morgan (Eds.), Syntax and semantics 3: Speech acts (pp. 41-58). New York: Academic Press. Johnson-Laird, P. N., & Wason, P.C. (Eds.). (1977). Thinking: Readings in cognitive science. Cambridge: Cambridge University Press. Kumpf, L. (1984). Temporal systems and universality in interlanguage: A case study. In F.R. Eckman, L.H. Bell, &D. Nelson (Eds.), Universals of second language acquisition (pp. 132-143). Rowley, MA: Newbury House. Lakoff, R. (1970). Tense and its relation to participants. Language, 46, 838849. McGilvray, J. (1974a). A proposal for the semantics of tenses in English. Unpublished manuscript, McGill University, Montreal.



McGilvray, J. (1974b). Tenses and beliefs. Unpublished manuscript, McGill University, Montreal. Phillips, B. (1984), Word frequency and the actuation of sound change. Language, 60, 320-342. Postal, P.M. (1972). The best theory. In S. Peters (Ed.), Goals of linguistic theory (pp. 131-170). Englewood Cliffs, NJ: Prentice-Hall. Reichenbach, H. (1947). Elements of symbolic logic. New York: Macmillan. Riddle, E. (1978). Sequence of tenses in English. Unpublished doctoral dissertation, University of Illinois, Urbana-Champaign. Sadock, J.M. (1978). On testing for conversational implicature. In P. Cole (Ed.), Syntax and semantics 9: Pragmatics (pp. 281-297). New York: Academic Press, Schiffrin, D. (1981). Tense variation in narrative. Language, 57, 45-62. Svartvik, J., & Quirk, R. (Eds.). (1980). A corpus of English conversation. Lund, Sweden: CWK Gleerup. Tannen, D. (1979). Whats in a frame? In R.O. Freedle (Ed.), N e w directions in discourse processing (pp. 137-181). Norwood, NJ: Ablex. Wallace, S. (1982). Figure and ground: The interrelationships of linguistic categories. In P.J. Hopper (Ed.), Tense-aspect: Between semantics and pragmatics (pp. 201-223). Amsterdam: John Benjamins. Wentworth, P. (1965). The gazebo. Harmondsworth, England: Penguin Books. Wolfram, W. (1985). Variability in tense marking: A case for the obvious. Language Learning, 35, 229-253. Wolfson, N. (1979). The conversational historical present alteration. Language, 55, 168-182.



TESOL QUARTERLY, Vol. 20, No. 2, June 1986

Dont Put Your Leg in Your Mouth: Transfer in the Acquisition of Idioms in a Second Language
Brown University and Boston University

The study reported in this article investigated whether second language learners use knowledge of their first language to comprehend and produce idioms in the second language. Subjects were 12 Venezuelan advanced learners of English. Comprehension of 45 English idioms15 identical in form and meaning to their Spanish equivalents, 15 similar to their Spanish equivalents, and 15 different from the corresponding Spanish idiomswas tested with a multiple-choice test and a definition test. Production of the same 45 idioms was tested with a discourse-completion test and a translation test. Results showed identical idioms were the easiest to comprehend and produce. Similar idioms were comprehended almost as well but showed interference from Spanish. Different idioms were the most difficult to comprehend and produce but showed less interference than similar idioms. Subjects used both inter- and intralingual strategies to produce idioms they did not know. Within each type, the idioms that were comprehended and produced most correctly were those which were frequently used and transparent and which had simple vocabulary and structure.

Second language learners encounter such difficulty using English idioms that they often prefer to avoid them altogether. This difficulty may result from confusing part of an idiom they have heard but not mastered in English, as in the case of to go out on a stick instead of to go out on a limb. It may also be the result of transferring part of an idiom in their first language to an English idiom, as in the case of to spread the voice instead of to spread the news (from the Spanish correr la voz, to run the voice). When the first and second language have identical idioms, the use of transfer can result in a correct idiom, such as to take the bull by the horns (the Spanish idiom agarrar al toro por los cuernos is identical in form and meaning). The investigation reported in this article was undertaken to 287

determine whether learners would use knowledge of their first language to help them understand and produce idioms in a second language. Specifically, the study examined whether first language idioms that are very similar to their equivalents in the second language would cause more interference than idioms that are different. The study was also intended to provide information about the strategies learners use when they have to produce idioms they do not know and the characteristics of those idioms which are the easiest to learn. An idiom is a conventionalized expression whose meaning cannot be determined from the meaning of its parts. For example, the idiomatic meaning of I was pulling your leg cannot be derived from the meanings of pull and leg. Idioms differ from other figurative expressions, such as similes and metaphors, in that they have conventionalized meanings. Native speakers of English know immediately that I was pulling your leg means I was teasing you, whereas they have to deduce their own meaning from a metaphor such as I was greasing your mind. The distinction between idiom and metaphor is not always precise because many idioms are dead or frozen metaphors figurative expressions which have acquired conventionalized meanings. For instance, Hes nothing but skin and bones could be understood as an idiom through knowledge of its conventionalized meaning, Hes very thin. If that conventionalized meaning were unknown, however, it could be interpreted metaphorically with the same meaning. Idioms can also be distinguished from other kinds of conventionalized language. Certain fixed, literal expressions are commonly used in particular situations but are not idiomatic. For example, just between you and me indicates confidentiality, I beg your pardon is a formulaic apology, and see you later is a common farewell. These expressions are not idioms because their meanings can be determined from the meanings of the words which comprise them. Yorio (1980) calls such expressions routine formulas and provides a comprehensive classification of their properties. Idiom and routine formula are not mutually exclusive categories, however; some routine formulas, such as take five or lets call it a day, are also idiomatic. TRANSFER AND CONTRASTIVE ANALYSIS The concept of transfer is based on the idea that previous learning affects subsequent learning. In language learning, this means that the forms and patterns of the native language are imposed on the second language (Gass, 1979). When these are identical in the two 288 TESOL QUARTERLY

languages and the learner uses the first language in producing the second, positive transfer occurs. The result is a correct second language form or pattern. When they are different, using those of the native language to produce the equivalent form or pattern in the second language causes negative transfer. The errors that result are called interference errors. During the 1950s and early 1960s, interlingual transfer was assumed to be the most important factor in learning another language (Politzer, 1965). However, the paradigm shift that occurred in linguistics and psychology in the 1960s created a change of focus. The emergence of generative grammar and cognitive psychology created the new discipline of psycholinguistics. Language acquisition was no longer seen as a process of forming correct habits through repetition and reinforcement, but as the result of an innate language acquisition device which operates through a process of hypothesis testing (Chomsky, 1959). Because transfer had been associated with the habit formation theory of language acquisition, the shift to generative grammar brought with it much less emphasis on interference and more emphasis on developmental processes, learning strategies, and the structure of the target language as sources of error (Richards, 1974). Second language acquisition came to be seen as a creative construction process rather than the transfer of habits from the first language to the second (Dulay & Burt, 1975). In spite of the claims that interference is not really a very important factor in second language acquisition, investigators have continued to find evidence showing substantial influence of the first language on the second (James, 1980; Sheen, 1980; Steinbach, 1981). Recent investigations have focused on the question of what is transferred, what the domains of language transfer are, and whether transfer can be predicted (Gass & Selinker, 1983). Transfer and contrastive analysis are linked in the literature because a comparison of two languages can help to show how an item from one language can be transferred to the other. The strong version of the contrastive analysis hypothesis has claimed to be able to predict areas of difficulty by comparing the native language of the learner with the target language (Lado, 1957). Similar patterns would be easy to learn because they could be successfully transferred from the first language. Different patterns would cause interference and therefore be difficult to learn. Stockwell, Bowen, and Martins (1965) hierarchy of difficulty proposed that the more different two items were, the more difficult they would be. TRANSFER AND THE ACQUISITION OF IDIOMS 289

Contrastive analysis has been criticized on theoretical grounds (Sajavaara, 1976; Whitman, 1970) and on the basis of empirical investigation (Brire, 1968; Buteau, 1970; Tran-Thi-Chau, 1975). Attempts have been made to modify the contrastive analysis hypothesis to make it more viable. Wardhaugh (1970) has proposed a weak version that explains errors after the fact rather than predicting them before they are made. Based on the results of empirical research, Oller and Ziahosseiny (1970) have suggested a moderate version, which proposes that more difficulty occurs when the differences between languages are slight. Very little work has been done on the role of transfer in the acquisition of idioms. Two studies done in The Netherlands tested whether structures such as idioms, proverbs, and slang, which were called language-specific, are considered by learners to be nontransferable. Both Jordens (1977) and Kellerman (1977) asked second language learners to judge the grammaticality of correct and incorrect sentences containing idioms. Some of the idioms had first language equivalents; some did not. In both studies, learners tended to judge those idioms which had first language equivalents as ungrammatical, indicating a reluctance to transfer language-specific items. These studies, however, dealt only with grammaticality judgments, not with the actual comprehension and production of idioms; nor did they differentiate idioms according to their degree of similarity to first language equivalents. The study reported in this article was designed to assess the differential effects of transfer on the comprehension and production of idioms. THE STUDY Hypotheses The following hypotheses were formulated: 1. Identical idioms would show evidence of positive transfer; they would be easiest to comprehend and to produce correctly. 2. Similar idioms would show evidence of negative transfer; while comprehension might be almost as high as for identical idioms, production of these idioms would reflect interference from the first language. 3. For different idioms, there would be no evidence of either positive or negative transfer; subjects would comprehend and produce fewer different idioms than idioms of the other two types. 290 TESOL QUARTERLY

Subjects A total of 12 advanced learners of English from Venezuela served as subjects. Subjects of the same nationality were chosen to ensure that they all used the same variety of Spanish and would be familiar with the Spanish equivalents of the idioms chosen for the study. From a list of all Venezuelan students at a major university, potential subjects were initially contacted on a random basis. However, the final group was self-selected in the sense that they were interested enough in the study to be willing to give 2 hours of their time in return for a small reimbursement. All subjects were regularly enrolled undergraduate students at the university. All had scored at least 500 on the TOEFL; the mean TOEFL score was 570. Average length of residence in the United States was 2.75 years, and average age was 21.8 years. Materials and Procedures The idioms chosen for the study were selected on the basis of two versions of a questionnaire, one in English and one in Spanish. The questionnaire consisted of three parts, each containing 50 idioms of one type (identical, similar, or different). Twenty-three Spanish speakers and 30 English speakers completed the questionnaire in their native language. They were asked to define each of the idioms and rate, on a scale of 1 to 5, the frequency of use of each. Based on these results, 15 idioms of each type (see Appendix) were chosen; all had been defined unambiguously by all of the respondents, had equivalent figurative meanings in both languages, and had received a median of at least 3 on the frequency-of-use scale. Tests were written to assess recognition, comprehension, recall, and production of these idioms. The recognition test was a multiplechoice test, with the choices including the correct paraphrase of the idiom, a sentence related to the correct paraphrase, a sentence related to the literal interpretation, and an unrelated sentence. An item from the multiple-choice test is presented below: Im fed up with him. a. Ive been seeing him too much. b. Im very tired of him. c. Im full from eating too much. d. Im crazy about him. The comprehension test asked the subjects to write a definition of the idiom in either English or Spanish. The recall test was a discoursecompletion task consisting of a paragraph containing the idiom TRANSFER AND THE ACQUISITION OF IDIOMS 291

with one word missing; subjects had to supply the missing word. The following item is an example: Tims parents were tired of hearing loud rock music all the time. Turn that music down, his mother yelled. Im up with your loud music! The production test was a translation task, although the subjects were not told that they were to translate the idiom. Each item consisted of a paragraph in Spanish containing the idiom and an English translation of the paragraph with the idiom omitted. Subjects were asked to supply the English idiom which they would use in that situation. Examples of all three types of idioms were given in the instructions, so subjects would realize that a literal translation was not always possible. A sample test item is given below: Los dos hermanos siempre se peleaban. Por fin su madre no poda ms, The two brothers were always fighting. Finally their mother couldnt take any more, and she shouted, Enough! with these fights! Subjects were tested individually or in small groups; all subjects took the tests in the following order: discourse-completion, translation, definition, multiple-choice. Scoring and Analysis The items on the two comprehension tests were scored as correct or incorrect, and the items on the production tests were scored as correct, incorrect, or incorrect with interference. It was often difficult to determine when interference had occurred. In this study, interference was defined as the incorrect use of a translation of a content word from a Spanish idiom. However, in many cases, especially with similar idioms, an incorrect word in the English idiom could be either a translation from the Spanish idiom or an overgeneralization or overextension of a word in the English idiom. For example, does put your leg in your mouth result from interference from the similar equivalent Spanish idiom meter la pata (to put in the leg), or is it an overextension of the English word foot? In unclear cases, the error was not considered to be interference. To check the reliability of the scoring, four graduate students were asked to score one of the translation tests which contained a large number of errors; interrater reliability among the researcher and the four independent raters was .87. 292 TESOL QUARTERLY

One-way analyses of variance with repeated measures were performed to test for differences among the three types of idioms on total number of idioms correct. Where there were significant differences, Tukeys Honestly Significant Difference test was used to ascertain which type of idiom had significantly higher or lower scores than the others. Analysis of variance could not be used to test for differences in interference scores because the mean scores for identical idioms were zero. Therefore, paired, two-tailed t tests were used to test for differences between similar and different idioms on number of interference errors. In addition, responses given on the translation test were analyzed to determine what strategies the subjects were using to produce unknown idioms. The idioms that were comprehended and produced most correctly were also examined to see if they had any characteristics in common. RESULTS Analyses of Variance One-way analyses of variance with repeated measures indicated clearly that subjects were performing differently with the three types of idioms (see Table 1). Planned multiple comparisons were done to show exactly where the differences were (see Table 2). On each of the two tests of comprehension, subjects did equally well
TABLE 1 Means, Standard Deviations, and F Scores for Idiom Tests Task Idiom type Multiplechoice Definition Translation Discoursecompletion



with identical and similar idioms but had significantly more difficulty with different idioms. On the translation test, similar and different idioms were equally difficult; on the discourse-completion test, performance differed for all three types of idioms. It appears that when the task was to recognize the meaning of an idiom, subjects were able to generalize from the meaning in their first
TABLE 2 Differences Among Means of Types of Idioms (Tukey's Honestly Significant Difference)

language to the meaning in the second language if the form was identical or similar; slight differences in form did not affect this process. When the task was to produce the entire idiom, slight differences in form made similar idioms just as difficult to produce as different idioms; whether the differences were great or very small did not matter. When the task was to supply part of an incomplete idiom, the more different an idiom was, the more difficult it was. Only the two tests of productive ability were scored for interference. There were no interference errors on identical idioms, but paired t tests showed that there were differences between similar and different idioms (see Table 3). However, transfer was more in evidence with similar idioms than with different ones, probably because subjects recognized the similarity and assumed that they could transfer. The hypotheses of the study were supported. Subjects comprehended identical idioms as well as similar idioms, and both were comprehended better than different idioms. They correctly 294 TESOL QUARTERLY

produced more identical idioms than similar or different idioms (Hypothesis 1). Similar idioms were comprehended as well as identical idioms, but production of similar idioms showed interference from the first language (Hypothesis 2). Fewer different idioms were comprehended and produced correctly (although the difference between similar and different idioms was not significant
TABLE 3 Means, Standard Deviations, and t Scores for Interference Errors

on the translation test), and there was little evidence of interference for different idioms (Hypothesis 3). Strategies While the results of this study show that subjects did use their native language to comprehend and produce idioms in the second language, they also used target language-related strategies. It is impossible to make any definitive statement about the relative influence of first and second language strategies, however, because of the difficulty of assigning responses to a specific category. For example, is I am filled up a confusion of filled and fed (I am fed up was the expected English idiom), or is it the result of interference from the different but equivalent Spanish idiom estoy hasta la coronilla (I am up to the crown of my head, with being up to the crown of your head meaning that you are filled up) ?1
1 An

anonymous TESOL Quarterly reviewer has pointed out that there are at least two other possible explanations for this error: the phonological similarity of filled and fed, or semantic reanalysis of I am filld up from I am full. Either of these explanations would be considered a target language strategy.



There were some fairly clear cases of target language strategies, however, such as the substitution of words with similar meanings in kill two birds with one rock, swallow it hook, cord, and sinker, and hit the nail on the tip, or the collocation of antonyms in come low or high water. In other cases, the confusion came from a different English idiom: put something fast on her, where to pull a fast one is confused with to put something over on her; kicked the towel, where to throw in the towel is confused with to kick the bucket. Other target language-related strategies included providing an incomplete idiom (cost an arm for cost an arm and a leg); using an English idiom different from the expected one, either an acceptable equivalent (Ive had it for Im fed up) or an unacceptable nonequivalent (play all my cards does not mean the same as put my cards on the table); using a figurative expression which is not a known idiom (a nail in the backyard for a needle in a haystack); or using a routine formula which is not idiomatic ( whats wrong with her? for whats bugging/eating her?). An interesting first language strategy was used by several subjects who thought of another Spanish idiom equivalent to the one given and translated that one. For example, it was by the clouds (a direct translation of estaba por las nubes, which means that something cost a lot) was given for cost un ojo de la cara (it cost an eye of the face), when the expected equivalent was it cost an arm and a leg. Use of the first language in producing idioms in English varied by individual; some showed virtually no interference for any type of idiom, while others had high rates of interference for both similar and different idioms. These differences may be related to whether the subjects kept their two language systems separate or not. Levenston (1979) discusses the possibility that some learners may consciously try to keep their two languages separate and thus reject second language forms which are too close to those of the first language. It would be interesting to investigate whether the use of first and second language strategies in producing idioms varies according to the proficiency level of the student. Previous research has shown that this is true for syntax (Taylor, 1975), for lexical simplification strategies (Jarujumpol, 1984), and for lexical compensatory strategies (Poulisse, Bongaerts, & Kellerman, 1984). In all cases, the less proficient learners used more first language strategies, while the more proficient learners used more second language strategies. Paribakhts (1985) study of communication strategies found two strategies related to idioms that were used exclusively by one group or the other. In a concept-identification task, only the less proficient learners used idiomatic transfer (reference to some semantic or 296 TESOL QUARTERLY

syntactic feature of a first language idiom), while only the more proficient group made use of target language idioms and proverbs to refer to a specific context where the target item was used. In the study reported here, there was no correlation between proficiency (as measured by TOEFL score) and the amount of interference evident in each subjects production of idioms. However, the language proficiency of the subjects was relatively homogeneous (TOEFL scores ranged from 520 to 620), and the study had not been designed to examine differences among them. Further research comparing intermediate and advanced learners would be of interest to see if advanced learners use more second language strategies. Best Known Idioms The 45 idioms used in this study were rank ordered according to the number and percentage of total correct responses. Within each type, the best known idioms (point of view, lend a hand, sleep on it) were frequently used, useful in a university setting, short and relatively simple, and fairly transparent in their meaning. The least known idioms (look for a needle in a haystack; swallow it hook, line, and sinker; take the cake) were less frequently used, more colloquial, and often contained difficult vocabulary. An additional factor, semantic similarity, may help account for why some idioms are easier than others. Degree of similarity is probably not a question of discrete categories, and perceived similarity may vary from one learner to another. The fact that two of the different idioms may have semantic rather than formal similarities to their Spanish counterparts could account for their being the easiest idioms in this group. To sleep on it is part of the same semantic domain as consultarlo con la almohada (to consult it with the pillow), and to pull his leg can be associated with tomarle el pelo (to take to him the hair). In both cases the literal meanings involve doing something to some part of the body. DISCUSSION On the two comprehension tests, it appears that subjects were able to generalize from the meaning of the idiom in Spanish to its meaning in English, even when the form was slightly different. On the two production tests, they were able to produce correctly many more identical idioms than idioms of the other two types. Both of these results indicate that positive transfer was being used. Negative transfer (interference) was also evident on the two production tests, TRANSFER AND THE ACQUISITION OF IDIOMS 297

more so for similar idioms than for totally different idioms. When differences are slight, the tendency may be to generalize and ignore those differences. When differences are so great that two forms have nothing in common, there would be no reason to try to use one form to produce the other, so little transfer would occur. The results of this study support the notion that advanced learners of a second language whose first language is related to the second can use their knowledge of idioms in their first language to comprehend and produce idioms in the second. This result differs from those of Jordens (1977) and Kellerman (1977), who found that learners considered idioms to be nontransferable, but the tasks subjects were asked to do are not comparable across the three studies. There is also support for the notion that structures which are very similar in the first and second languages will produce more interference than structures which are different. This supports the moderate contrastive analysis hypothesis of Oller and Ziahosseiny (1970). This study has theoretical implications for the investigation of transfer in the acquisition of a second language. The results suggest that similarities between languages encourage interference and that idioms are not always considered nontransferable. It must be remembered, however, that the results apply only to the specific subjects and tasks of this study. Further research is needed with subjects from other language and cultural backgrounds. In addition, it is possible that subjects would avoid using idioms if they had a choice of using an English idiom or not. Finally, the findings of this study can be applied to the teaching of idioms in ESL and foreign language classes. If students are using their knowledge of idioms in their first language to comprehend and produce second language idioms, teachers should take advantage of this. In bilingual and foreign language settings, overt comparisons can show students which idioms can be transferred from their first language and which are likely to cause interference. In ESL settings, where students come from various language backgrounds and the teacher does not know all of the students native languages, students can be encouraged to make such comparisons themselves. The data on best and least known idioms also provide a basis for deciding which idioms to teach. Infrequent, highly colloquial idioms with difficult vocabulary should be avoided. Students will obviously have difficulty producing them correctly; in addition, these colorful idioms, even when correctly produced, often sound 298 TESOL QUARTERLY

strange and unnatural when spoken by nonnative speakers of English. Activities for teaching comprehension of idioms should provide students with skills in guessing meaning from context and in dealing with figurative speech. For example, shown a paragraph from which an idiom has been deleted, students can supply a word or phrase which fits the context. When an appropriate word or phrase has been supplied, students are shown how they have actually guessed the meaning of the deleted idiom. They can then be guided to discover whether there is an equivalent idiom in their first language and, if so, whether the two idioms are identical, similar, or different. Activities which compare literal and figurative meanings of idioms help students to realize the absurdity of the literal meanings and provide a link from the literal words to the nonliteral meaning. Examples of activities of this type include matching pictures showing literal and idiomatic meanings of an idiom, drawing or acting out literal meanings, playing idiom charades, and making up stories or dialogues in which the literal use of an idiom creates a misunderstanding or a humorous situation. Activities of this type would be particularly useful with idioms which have no first language equivalent or a totally different one. When idioms have identical or similar first language equivalents, the native language already provides the link between the literal words and the nonliteral meaning. Activities which encourage production of idioms can be based on lists of idioms collected by the students or supplied by the teacher. These lists should include idioms which are similar in the first and second languages and are therefore likely to cause interference. Students can tell add-on stories containing idioms; retell a story they heard that contained idioms; write and present short plays, puppet shows, stories, or dialogues with idioms in them; and role-play situations that lend themselves to production of idioms. All of these activities will be easier if the idioms used are thematically related, and having students decide which idioms are related will help them learn their meanings. (See Irujo, in press, for complete descriptions of these activities, as well as criteria for deciding which idioms to teach.) The issue of teaching for recognition versus teaching for production is particularly important in the teaching of idioms, since interference will be most obvious in production. It has been suggested that at beginning and intermediate levels, idioms be taught for recognition only (Yorio, 1980). However, the use of idioms is so common in English that it can be difficult to speak or TRANSFER AND THE ACQUISITION OF IDIOMS 299

write without them (Seidl & McMordie, 1978). Even beginning students can successfully learn to produce some idioms if they are carefully chosen on the basis of frequency, need, transparency, and syntactic and semantic simplicity. Students should be taught how to utilize positive transfer and avoid interference, and they must be given enough opportunity to practice using idioms in contextualized situations. By doing this, we can help students overcome their idiomphobia and learn to produce English idioms correctly, both in and outside of class.

This article is based on a portion of the authors doctoral dissertation (Irujo, 1984b) and on a paper (Irujo, 1984a) presented at the Ninth Annual Boston University Conference on Language Development. Thanks are due to Bruce Fraser for guidance and comments and to an anonymous TESOL Quarterly reviewer for helpful suggestions.

Suzanne Irujo is Director of Project BELT (Bilingual Education Leadership Training) at Brown University and Coordinator of Student Teaching for Bilingual Education, TESOL, and Modern Foreign Languages at Boston University. Her research interests include transfer in second language acquisition and strategies of lexical acquisition.

Brire, E.J. (1968). A psycholinguistic study of phonological interference. The Hague: Mouton. Buteau, M.F. (1970). Students errors and the learning of French as a second language: A pilot study. International Review of Applied Linguistics, 8, 133-145. Chomsky, N. (1959). Review of Verbal behavior by B.F. Skinner. Language, 35, 26-58. Dulay, H. C., & Burt, M.K. (1975). Creative construction in second language learning and teaching. In H.C. Dulay & M.K. Burt (Eds.), New directions in second language learning, teaching and bilingual education (pp. 21-32). Washington, DC: TESOL. Gass, S. (1979). Language transfer and universal grammatical relations. Language Learning, 29, 327-344. Gass, S., & Selinker, L. (1983). Language transfer in language learning. Rowley, MA: Newbury House. 300 TESOL QUARTERLY

Irujo, S. (1984a, October). Dont put your leg in your mouth: Transfer in the acquisition of idioms in a second language. Paper presented at the Ninth Annual Boston University Conference on Language Development, Boston. Irujo, S. (1984b). The effect of transfer on the acquisition of idioms in a second language. Unpublished doctoral dissertation, Boston University. Irujo, S. (in press). A piece of cake: Some thoughts on learning and teaching idioms. ELT Journal. James, C. (1980). Contrastive analysis. London: Longman. Jarujumpol, W. (1984, March). Lexical simplification strategies employed by high-proficiency and low-proficiency non-native speakers. P a p e r presented at the 18th Annual TESOL Convention, Houston. Jordens, P. (1977). Rules, grammatical intuitions and strategies in foreign language learning. lnterlanguage Studies Bulletin, 2 (2), 5-77. Kellerman, E. (1977). Towards a characterization of the strategy of transfer in second language learning. Interlanguage Studies Bulletin, 2 (1), 58-145. Lado, R. (1957). Linguistics across cultures. Ann Arbor: The University of Michigan Press. Levenston, E.A. (1979). Second language lexical acquisition: Issues and problems. Interlanguage Studies Bulletin, 4 (2), 147-160. Oller, J.W., Jr., & Ziahosseiny, S.M. (1970). The contrastive analysis hypothesis and spelling errors. Language Learning, 20, 183-189. Paribakht, T. (1985). Strategic competence and language proficiency. Applied Linguistics, 6, 132-146. Politzer, R.L. (1965). Teaching French: An introduction to applied linguistics (2nd. ed.). Waltham, MA: Blaisdell Publishing. Poulisse, N., Bongaerts, T., & Kellerman, E. (1984, August). The use of compensatory strategies by Dutch learners of English. Paper presented at the Seventh International Congress of Applied Linguistics, Brussels. Richards, J.C. (1974). Error analysis: Perspectives on second language acquisition. London: Longman. Sajavaara, K. (1976). Contrastive linguistics past and present: A communicative approach. Jyvskyl Contrastive Studies (No. 4, pp. 930). Jyvskyl, Finland: Jyvskyl University, Department of English. Seidl, J., & McMordie, W. (1978). English idioms and how to use them. Oxford: Oxford University Press. Sheen, R. (1980). The importance of negative transfer in the speech of nearbilinguals. International Review of Applied Linguistics, 18, 105-119. Steinbach, H.R. (1981). On the classification of errors in translation papers with some consideration of interference phenomena. In J. Fisiak (Ed.), Papers and Studies in Contrastive Linguistics (No. 13, pp. 249-259). Poznan, Poland: Adam Mickiewicz University Press. Stockwell, R.P., Bowen, J. D., & Martin, J.W. (1965). The grammatical structures of English and Spanish. Chicago: University of Chicago Press. Taylor, B. (1975). The use of overgeneralization and transfer learning strategies by elementary and intermediate students of ESL. Language Learning, 25, 73-107. 301


Tran-Thi-Chau. (1975). Error analysis, contrastive analysis and students perception: A study of difficulty in second language learning. International Review of Applied Linguistics, 13, 119-143. Wardhaugh, R. (1970). The contrastive analysis hypothesis. TESOL Quarterly, 4, 123-130. Whitman, R.L. (1970), Contrastive analysis: Problems and procedures. Language Learning, 20, 191-197. Yorio, C.A. (1980). Conventionalized language forms and the development of communicative competence. TESOL Quarterly, 14, 433-442.

APPENDIX List of Idioms Identical Idioms point of view punto de vista (point of view) to have on the tip of my tongue tener en la punta de la lengua (to have on the tip/point of the tongue) to look for a needle in a haystack buscar una aguja en un pajar (to look for a needle in a hayloft) to play with fire jugar con fuego (to play with fire) to break the ice romper el hielo (to break the ice) to have your hands tied tener las manes atadas (to have the hands tied) to stick your nose into everything meter las narices en todo (to stick the noses in everything) to be all ears ser todo odos (to be all ears) a vicious circle un crculo vicioso (a circle vicious) to run a risk correr riesgo (to run risk) to be in charge estar a cargo (to be at charge) the black sheep of the family la oveja negra de la familia (the sheep black of the family) to put your cards on the table poner las cartas sobre la mesa (to put the cards on the table) to open his eyes abrirle los ojos (to open to him the eyes) to wash my hands of it lavarme las manos de ello (to wash myself the hands of it) 302 TESOL QUARTERLY

Similar Idioms to kill two birds with one stone matar dos pjaros de un tiro (to kill two birds from one shot) to catch him red-handed cogerle con las manos en la masa (to catch him with the hands in the dough) cant make heads or tails of it no tiene ni pies ni cabeza (it doesnt have feet or head) to lend a hand echar una mano (to give a hand) to cost an arm and a leg costar un ojo de la cara (to cost an eye of the face) to put your foot in your mouth meter la pata (to put in your leg) to burn the midnight oil quemarse las pestaas (to burn oneself the eyelashes) to have a screw loose faltarle un tornillo (to be missing to him a screw) to swallow it hook, line, and sinker tragar el anzuelo (to swallow the hook) to take a load off his mind quitarle un peso de encima (to take to him a weight from on top) to hit the nail on the head dar en el clavo (to give/hit in the nail) come hell or high water contra viento y marea (against wind and tide) to hold your tongue morderse la lengua (to bite oneself the tongue) to have your back to the wall encontrarse entre la espada y la pared (to find oneself between the sword and the wall) the coast is clear no hay moros en la costa (there are no Moors on the coast) Different Idioms to pull his leg tomarle el pelo (to take to him the hair) to keep someone posted tener a alguien al corriente (to have someone at the current) to have a free hand tener carta blanca (to have a letter white) to kick the bucket estirar la pata (to stretch the leg) to take the rap pagar el pato (to pay the duck) whats eating him? qu mosca le ha picado? (what fly to him has bitten?) TRANSFER AND THE ACQUISITION OF IDIOMS 303

to make a killing hacer su agosto (to make his August) to sleep on it consultarlo con la almohada (to consult it with the pillow) to go along with the crowd dejarse llevar por la corriente (to let oneself to carry by the current) to put two and two together atar cabos (to tie ends) to stick to your guns no dar el brazo a torcer (not to give the arm to bend) to put something over on dar gato por liebre (to give cat for hare) to be fed up with estar hasta la coronilla de (to be up to the crown of the head from) to throw a fit poner el grito en el cielo (to put the shout in the heaven) to take the cake ser el colmo (to be the summit)



TESOL QUARTERLY, Vol. 20, No. 2, June 1986

Information Gap Tasks: Do They Facilitate Second Language Acquisition?

University of Pennsylvania

This article reports the findings of the latest of a series of studies conducted to determine the effects of task type and participation pattern on language classroom interaction. The results of this study are compared to those of an earlier investigation (Pica & Doughty, 1985a) in regard to optional and required information exchange tasks across teacher-directed, small-group, and dyad interactional patterns. The evidence suggests that a task with a requirement for information exchange is crucial to the generation of conversational modification of classroom interaction. This finding is significant in light of current theory, which argues that conversational modification occurring during interaction is instrumental in second language acquisition. Furthermore, the finding that group and dyad interaction patterns produced more modification than did the teacher-fronted situation suggests that participation pattern as well as task type have an effect on the conversational modification of interaction.

Efforts to teach second languages within a communicative framework have led to certain methodologically motivated organizational changes in the classroom environment. To aim at specific needs of students as well as to captivate their interest, current ESL classrooms often feature a diverse assortment of instructional materials, learning activities, and student-teacher or student-student interactional patterns. In addition to using lessons in which they fully control classroom interaction, many teachers have regularly begun to employ small-group and pair work as a means of increasing their students target language practice time. Classroom assignments now feature not only activities involving the introduction and practice of usage rules, but also tasks which encourage 305

the use of the target language in problem-solving and decisionmaking situations. In sum, the kinds of activities students are engaged in and the interlocutors with whom they interact have changed with recent years. In light of these organizational changes in the ESL classroom, a series of empirical studies was conducted to examine the possible effects on classroom second language acquisition of learning tasks and interactional patterns currently in use (Pica & Doughty, 1985a, 1985b, 1985c). PREVIOUS RESEARCH: THE EFFECT OF PARTICIPATION PATTERN An initial study (Pica & Doughty, 1985a) compared conversational interaction in teacher-to-student and student-to-student interactional patterns during decision-making exercises of the kind well known in ESL materials. Of particular interest was the identification of differences in (a) grammaticality of input, (b) the amount of speech produced, and (c) the amount of modified interaction which occurred during these conversations. Modified interaction is defined here as that interaction which is altered in some way (either linguistically or conversationally) to facilitate comprehension of the intended message meaning. In the teacherfronted activity, individual classes, together with teachers, who directed the interaction, had to arrive at a solution to a problem. As a class, they were given information about five families living in the 21st century, and then they had to choose which one was most eligible to adopt a child. In the group situation, 4 students working together had to choose among six potential recipients for a heart transplant. Thus, both teacher-fronted and group tasks involved arriving at a decision based on a description of a situation. Although it had been hypothesized that there would be more conversational modification (operationalized as confirmation and comprehension checks and clarification requests, as defined by Long, 1980, and repetitions) by students in groups than with their teachers, these predictions were not borne out. In fact, the teacherfronted situation engendered more conversational adjustments than did the group format. These counterintuitive results could not be considered to have great significance, however, because very little conversational modification was observed in either situation. In view of the importance attached to conversational modification in making input comprehensible and thereby promoting second language acquisition (Long, 1981), it appeared that neither 306 TESOL QUARTERLY

participant-pattern format was especially conducive to the acquisition of a second language in the classroom environment. However, two potentially confounding factorsone having to do with the task, the other related to classroom patternmay have influenced the results. These two factors, discussed below, led to the design of a second experiment. The first concern was with the task employed in the investigation. Davies (1982) and Long (1980, 1981, 1983a, 1983b) have stressed the importance of using activities with a built-in, two-way information gap. Information gap refers to the existence of a lack of information among participants working on a common problem, but the term does not define the nature of the gap. Two-way information gap tasks are here defined (following Long, 1980) as those tasks which require the exchange of information among all participants, each of whom possesses some piece of information not known to, but needed by, all other participants to solve the problem. In this article, such tasks are referred to as required information exchange tasks to emphasize the obligatory nature of the gap and to avoid confusion, as the exchanges that occur are actually multidirectional rather than two-way. Long (1981) claims that such activities promote optimal conditions for students to adjust their input to each others levels of comprehension (i.e., modify the interaction) and thereby facilitate their second language acquisition. One-way information gap tasks are usually defined as tasks which do not require an exchange of information; they are referred to here as optional exchange tasks. In optional exchange tasks, participants decide whether or not to contribute to the solution of the problem. Often, as discussed below, confident and proficient speakers carry the conversation, and weaker students tend to opt out of the task altogether. The decision-making activities used in the first study, while communicative in emphasis, were nevertheless not required information exchange tasks. Each participants contribution to the decision, primarily in the form of arguments and opinions, may have been useful in helping other participants arrive at a group solution but was not necessarily required for making the final decision. In other words, completion of the task did not oblige participants to pool information known only to individuals as would be required by a multi-way information gap task. As a result, the teacher and a few class members monopolized the conversational interaction in the teacher-fronted lesson, and the more fluent students did likewise within their individual groups. Thus, there were no constraints on all students to participate or to INFORMATION GAP TASKS AND SLA 307

adhere to any one topic. Many students tended to go along with the majority opinion of both their class and group when it came time to articulate the final decision; this occurred in spite of the fact that they had given prior indications of disagreement with their classmates. In some cases, if the students were not able to reach a unanimous decision, they would simply shift to a different aspect of the problem, thereby abandoning the topic at hand altogether. Typically, in the face of group or class conflict of opinion, the less linguistically proficient students opted to avoid participation, and the less skillful debaters tended to capitulate rather than to make sure that their opinion was taken into account. The more expressive participants, including the teacher of course, dominated the interaction and supplied most of the input. The input generated by the proficient students and the teacher apparently was either beyond the processing capacity of weaker students, and hence incomprehensible to them, or simply was at their current processing level, and therefore did not necessitate interfactional modification. In the second instance, when students did not go beyond their existing level of understanding, they may have been influenced by the lack of motivation to reach a truly unanimous decision. Thus, in both teacher-fronted and group interaction during decision-making tasks, students may either have failed to have any idea of message content or may have understood messages so well that they did not need to ask for or provide adjustments in target language use. The second possible explanation for the counterintuitive outcome of the initial study was that group work, for many of the reasons outlined above, may not have been the optimal format for activating modified interaction among the students. As happened in the teacher-fronted situation, the more fluent student(s) among the 4 in each group studied tended to dominate the decision-making activity, often providing input so far above the comprehension level of the other students that it was not challenged. At other times, the language produced by individual group members was easily understood so that little modification was needed; hence few adjustments were requested or produced in the group interaction. We suspected that a combination of factors contributed to the null findings of this teacher-fronted versus group-work comparison. Potentially the most important factor was that the tasks employed did not require an exchange of information and thus resulted in a small number of confirmation and comprehension checks and clarification requests, all of which are believed to be vital to second language acquisition. For that reason, the number of conversational modifications which occurred in either participation-pattern format 308 TESOL QUARTERLY

was extremely low. Second, we had predicted that during interaction with a full class, the teacher would control the interaction in such a way that little modification would be required. Surprisingly, however, the students working in small groups also tended to structure the discourse so as to limit the need for adjustments. Thus, what we had thought were different participation patterns were more similar than we realized. PRESENT RESEARCH: THE EFFECTS OF TASK AND PARTICIPATION PATTERN Purpose and Hypotheses A second study was conducted to examine these two factors. The major differences between this study and the earlier research are that (a) tasks were employed which had a requirement f o r information exchange and (b) in addition to comparing teacherfronted versus group work on these tasks, a third interfactional patternthe student dyadwas introduced into the experimental design. Our first hypothesis was that activities which required an information exchange for their completion would generate substantially more modified interaction than those in which such exchange was optional. Thus, there would be more comprehension and confirmation checks, more clarification requests, and more repetitions in the former than in the latter activity. Furthermore, we predicted that the number of interlocutors and the presence or absence of the teacher would influence the amount of modified interaction in the activity. We believed that the teacher, more experienced in making sense out of interlanguage productions, would be less likely to seek clarification or confirmation of student utterances. The more proficient students would be more confident that their target language could be understood and therefore would be less likely to check the comprehension of their interlocutors. Less linguistically proficient students might feel reluctant or embarrassed to indicate their lack of comprehension in front of their teacher or a large number of classmates. Thus, we anticipated that the presence of the teacher and the dynamics of a large group of interlocutors should reduce the amount of modified interaction. In the group situation, on the other hand, we felt that participants, sitting in closer, face-to-face view than in the teacher-fronted situation, might notice confusion on the part of fellow interactants and would therefore be inclined to check their comprehension. In INFORMATION GAP TASKS AND SLA 309

addition, we believed that the face-threatening nature of the task would diminish as the number of interactants decreased. Thus, opportunities for modification would be even more pronounced in the dyad situation, in which participants interacted only with each other. This reasoning led to our second hypothesis: Although interaction is generated by all required information exchange tasks, more modified interaction would occur in the dyad situation than in the group situation, which would in turn provide more opportunity for modification than the teacher-fronted situation. Subjects The subjects in both the earlier and present studies were adult students and teachers from six intermediate ESL classes (three classes in each of the two studies). Classes were selected according to proficiency level: Pilot testing revealed that the task was challenging, yet not too difficult, for intermediate-level students. Those students who participated in group and dyadic activities were chosen at random by the classroom teachers. The students came from a variety of L1 backgrounds; the teachers were native speakers of English, all of whom had had several years of teaching experience. Data Collection To insure the validity of comparisons, data were collected for the present study through the same procedures used in the earlier study. Since the two sets of data were collected from different sets of subjects, the classrooms selected to participate in the present study were carefully matched to those in the earlier study on the variables of proficiency level, age, size (in both studies, class size ranged from 11 to 15), and teacher experience. Each activity was audiotaped, and as in the previous study, the researchers were not present during taping so that data could be collected as unobtrusively as possible. Materials and Procedures The required information exchange task developed for this study was carried out in each of three interactional patterns: teacher fronted, small group, and dyad. For the teacher-directed activity, each participant, including the teacher, was given a felt-board 310 TESOL QUARTERLY

garden and a number of various loose felt flowers which were to be planted (see Figure 1). At the beginning of the task, each board contained a tree, which was glued down in the center and served as a point of reference, and a display of a small number of flowers which had already been planted (i.e., glued down). No two boards contained the same display of already-planted flowers. The object of the task was to plant the garden according to a master plot, which was not shown to participants until after they had completed the task. Individual boards displayed a different portion of the master plot to each participant, who was to instruct other participants on which flowers to plant and where to put them. Together, the participants possessed all the information to complete the task. (All the felt-board gardens superimposed on each other would comprise the master plot.) Individually, however, participants possessed only a few pieces of the garden puzzle. All work had to be carried out by each participant behind the board, which was held in a semi-vertical position. The students and teacher were required to keep their own gardens and unplanted flowers out of sight of the other participants and were not allowed to hold up the unplanted flowers so that they could be seen by others. After completing the task, the students and teacher together compared their own gardens to the master plot. Each individual was required to contribute because no other participant possessed the same information regarding the location of certain flowers on each felt-board garden. Furthermore, all participants had to understand each others information about flower locations in order to accomplish the task successfully. Thus, we predicted that more modified interaction would be generated. For the small-group task, the teacher was asked to choose, at random, a group of 4 students. This time the task involved arranging a new set of flowers of different shapes and colors into another configuration. For the dyad situation, the teacher chose 2 students from the group of 4, again at random, and a third distinctive arrangement of flowers had to be planted. Ten-minute samples from each activity were later transcribed and analyzed to compare several features of interaction generated by the multi-directional, required exchange tasks in the three interfactional situations.1 These activities were conducted in the following order: teacher fronted, small group, dyad.
1 Each sample was coded independently by both researchers. Interrater reliability scores of

.88 for repetitions and .93 or higher for all other features of interactional modification (see Analyses) were obtained.



To insure that differences among the three participation patterns were not due to a practice-on-task effect, two precautions were taken. First, the teacher conducted a demonstration lesson with the class, during which directions for planting the garden were given, the various materials to be employed in the tasks were introduced and described, and frequent checks of students comprehension were made. Second, the teacher-fronted lesson, although always conducted before the group or pair work, was carried out in two parts. After 15 minutes of activity, the teacher stopped the task and conducted a question/answer period and class discussion. Then the task was completed. The 10-minute sample used for research purposes was taken from the last third of this phase, when the activity had been taken up again. In all cases, the activity had been in progress for at least 20 minutes before the 10-minute sample for transcription was selected. We believed that by that time, students would be familiar with all the materials and with the procedures involved in exchanging information about them. Thus, any modification which arose would be due only to the need to exchange information (equal in all three tasks) and not to a need to clarify the procedures of the tasks (likely to be unequal across the tasks, as the first time through would be more difficult than the third). Analyses The features of modified interaction used in the analysis of the data collected for the present study are the same as those used in the earlier research. They include clarification requests, confirmation checks, and comprehension checks. Clarification requests occur when one interlocutor does not entirely comprehend the meaning and asks for clarification, as in the following example: A: She is on welfare. B: What do you mean by welfare?

In making confirmation checks, the listener believes he or she has understood but would like to make sure: A: Mexican food have a lot of ulcers. B: Mexicans have a lot of ulcers? Because of the food?

In making comprehension checks, the speaker wants to be certain that the listener has understood: A: Do you know what I mean? INFORMATION GAP TASKS AND SLA 313

Several other features of modification are subsumed by the general label repetition. The categories of repairing, preventive, and reacting repetitions (Doughty & Pica, 1984) were developed to distinguish between classroom-related moves and the modification of interaction which has been claimed to be necessary for second language acquisition (Long, 1981). In the analysis of the data for earlier research (Pica & Doughty, 1983), it was observed that many classroom repetitions are used for such purposes as (a) initiating topics during structuring moves, (b) insuring adherence to a topic or completion of a task when students attention wanders, or (c) offering feedback to students regarding appropriateness of student responses. These classroom-related moves, called structuring and feedback repetitions (described more fully in Pica & Doughty, 1985b), were eliminated from analysis in the present study. The only repetitions considered were those which occurred during actual or perceived communication breakdowns or when both interlocutors took an active role in establishing or developing topics. Such repetitions were examined both in the case of repeating ones own utterance (self-repetition) and restating anothers utterance (other-repetition). RESULTS AND DISCUSSION The aims of this discussion are (a) to compare the amount of modified interaction generated in teacher-fronted and group interactional patterns when the nature of the task was manipulated, specifically, optional versus required information exchange tasks; (b) to compare the amount of modified interaction generated when the task was held constant and the participation pattern was manipulated, in this case, teacher-fronted versus small-group versus dyad participation patterns in required information exchange tasks; (c) to examine the role of repetition; and (d) to present ancillary findings on the total amount of interaction produced during a task. The Effects of Task and Participation Pattern on the Modification of Interaction A requirement for information exchange generated more modification of interaction than did a task with no such requirement. A two-way analysis of variance (ANOVA) revealed that the main effect for task was statistically significant, thus confirming the first hypothesis of this study (see Tables 1 and 2). The ANOVA also 314 TESOL QUARTERLY

showed that while the main effect for participation pattern was nonsignificant, there was a significant interaction of the two variables of task and participation pattern.
TABLE 1 The Effects of Task Type and Participation Pattern on Total Interactional Modification

TABLE 2 Two-Way ANOVA: Task Type x Participation Pattern

The Effect of Participation Pattern on the Modification of Interaction The results of a one-way ANOVA revealed a statistically significant main effect for participation pattern (see Tables 3 and 4). Modification of interaction was higher in the group than in the INFORMATION GAP TASKS AND SLA 315

teacher-fronted participation pattern. However, there was virtually no difference between the group and dyad interaction patterns in the amount of modification, as can be seen in the very similar modification scores. Thus, the second hypothesis of this study was confirmed insofar as group participation pattern resulted in more modification than did the teacher-fronted pattern; however, our prediction that the dyad would facilitate even more modification was not borne out. In comparing the results of the analyses for both hypotheses, we observed that when both task and participation pattern are independent variables (i.e., manipulable by the teacher), task type has the overwhelming influence on the amount of modification. However, participation pattern is not unimportant: This is suggested by the interaction obtained between the two variables and is confirmed when task is removed as a variable and participation pattern then produces a significant main effect.
TABLE 3 The Effect of Participation Pattern on Total Interfactional Modification


Regarding the experimental design, although the group and dyad activities always occurred after the teacher-fronted task, there was clearly no practice-on-task effect. Practice on task would have 316 TESOL QUARTERLY

resulted in fewer modifications as participants became more familiar with the task and the task-related materials and thus became less likely to need clarifications on how to complete the activity. Instead, more interfactional modification occurred during the group and dyad activities than during class interaction with the teacher. One factor which may have helped produce these results is the interactional experience that comes from repeating a task. As Pica and Long (in press) have argued, native speakers (NSs) become more skilled in modifying interaction with nonnative speakers (NNSs) as they accumulate experience in NS-NNS conversation. It is possible that NNSs make similar gains through experience in interacting with other NNSs; however, there is as yet no research which equates NS gains in experience with NNS gains in experience. Assuming that experience does affect NNS-NNS interaction, the NNS students who completed all three tasks would have been better at modifying interaction on the third task than on the first. Clearly, NNS gains in linguistic and conversational modification skills through repeated experience in NNS-NNS interaction is an area of research which demands fuller investigation. The Role of Repetition In further analysis of the data, additional effects of the manipulation of task type and participation patterns were examined (see Tables 5, 6, and 7). These analyses, though somewhat more speculative, suggest implications for future research. As discussed elsewhere (Pica & Doughty, 1985b), repetitions, while functioning in an important role as modifications of interaction, are puzzling at best to analyze. Much repetition occurs without affecting the interaction at all (e. g., the case of a teacher who repeats an utterance several times, even though students understood the first time). We found it useful to eliminate repetitions entirely from the analyses to insure that the results would be robust (i.e., not influenced by coding or interpretive factors). Three features form a crucial subset of interfactional modifications: clarification requests, confirmation checks, and comprehension checks. Table 6 presents the results of a two-way ANOVA of the difference in the amount of these modifications across task and participation pattern. These results are consistent with those which included repetition (see Tables 1-4) and thus eliminate any INFORMATION GAP TASKS AND SLA 317

TABLE 5 The Effects of Task Type and Participation Pattern on a Subset of Interfactional Modification

TABLE 6 Two-Way ANOVA: Task Type x Participation Pattern

apprehension about the definition and role of repetition in interfactional modification.2 The smaller number of confirmation and comprehension checks and of clarification requests which occurred during both teacher2 In

work currently under way (Pica, Doughty, & Young, 1985), we have attempted to clarify the definition of repetition, and by using videotaping techniques, we are now able to determine when and how repetition affects interaction. Repetition, we have found, may in fact be the most critical interfactional modification; thus, it is important to continue to develop sophisticated data-collection instruments which can accurately record this variable.



fronted task types suggests that students may have been reluctant to indicate a lack of understanding in front of their teacher and an entire class of students. Thus, they may have attempted to behave as though they understood, even when they did not. It is therefore possible that both teacher-fronted task types did not generate enough modification to make classroom input comprehensible to individual students. In the case of the gardenplanting task, one of the participating teachers noted informally that in the teacher-fronted format, individual students boards did not often correspond to the instructions given. In striking contrast, the participants in group and dyad interaction did manage to replicate the master plot quite closely.3 Total Amount of Interaction Another area of interest is the total amount of interaction produced during a task. Total amount of interaction is defined as the sum of all T-units and fragments (Hunt, 1970). As shown in Table 7,4 when teacher-fronted and group participation patterns were compared on both optional and required information exchange tasks, we found that there was more total interaction produced in the teacher-fronted pattern than in the group in both types of task. We also found that for both participation patterns, the total amount of speech increased when the exchange of information was required. However, the increase in the group was almost 10 times that in the teacher-fronted situation45.6% and 4.6% respectively. Thus, on the participation pattern variable, more total interaction was generated whenever the teacher was present, and on the task variable, more interaction was generated during the compulsory information exchange task. The teacher-fronted interaction on a required information exchange task generated the most total interaction, while the group interaction on the optional exchange task generated the least.
3 Based

on these informal findings, another series of studies is now being conducted to determine whether modification makes input sufficiently comprehensible for the successful execution of such tasks. In these studies, students are being videotaped to determine if their comprehension is sufficient for following directions about the placement of items on a board game (Pica, Doughty, & Young, 1985, 1986). 4 The data in Table 7 are presented somewhat more informally than those presented thus far because two of the data samples for the groups working on decision-making tasks (collected in the earlier study) were 5 instead of 10 minutes in duration. For purposes of comparison with all other samples, the number of modified and unmodified utterances for these two samples was doubled, and percentages were calculated on the basis of these adjusted numbers. Only these two scores were adjusted; the others were used in their original form. However, since these are extrapolated numbers, no formal statistical procedures were performed.



TABLE 7 Total Interaction (T-Units and Fragments)

In itself, this finding is not particularly astonishing. After all, teachers do tend to talk a great deal, speak more quickly, and hesitate less often in comparison with ESL students struggling to learn a new language. Thus, their fluent native speech would add to the total amount of interaction. Indeed, during the decision-making tasks of the first study, this was clearly the case. During these tasks, teachers produced almost half of the total number of utterances in reaching a decision with their classes. In other words, teachers spoke about as much as the total number of students combined. However, in working with their classes on the garden-planting task, the teachers did not contribute as extensively to the interaction. In fact, one of the participating teachers seldom spoke, except when giving directions and when taking a turn to impart information about his flowers. In the 10-minute sample which was analyzed, this teacher contributed only one utterance to the classroom conversationa confirmation check. Thus, the students did more talking on the required information exchange task, whether working with their teachers or in groups of 4. This is probably because the required interaction task places all participants in equal positions, each with the same amount of information, which must be disseminated to other participants. This finding stimulated interest in another question: When the amount of total interaction increased, did the increase occur in the number of utterances characterized by features of modification (here, including repetitions) or in the number of utterances not considered to function to modify interaction? In the teacher-fronted situation, there was an increase of 14% in the area of unmodified interaction, as compared with a decrease of 5% in the utterances 320 TESOL QUARTERLY

which contained features of modification. In the group situation, however, there was a substantial increase in the amount of modification122%and a decrease of 13% in the amount of unmodified interaction (see Table 7). CONCLUSIONS Enthusiasm about group work in the classroom must be tempered by the observation that at times, the teachers absence can limit the amount of modification which takes place when the students interact. This seems most likely to happen in tasks which do not compel the students full-fledged participation. Thus, decisionmaking or optional exchange tasks of the kind used in our earlier study do not trigger modifications among students working independently in groups. This participation pattern facilitates the modification of interaction only if the task requires an exchange of information. Unless a required information exchange task is chosen, students will interact less and will modify their interaction less as well. While a required information exchange task will compel students to talk more in either a teacher-fronted or a group situation, this increase in total production will result in an increase of modified interaction only when students are working in groups. Recent research, reviewed in Long and Porter (1985), has investigated the makeup of small groups. Studies by Porter (1983), Varonis and Gass (1983), and Gass and Varonis (1985) have shown that the presence or absence of native speakers and the group members proficiency levels and L1 backgrounds all influence the amount of modification of interaction in this participation pattern. The most modification was obtained when (a) all members of groups/dyads were nonnative speakers, (b) members of groups had varying proficiency levels, and (c) members of groups had different Lls. These results are encouraging to teachers, as they reflect the makeup of small groups in most second language classrooms. The findings of this recent research, together with the results of the present study, raise an important question: How much of the time do individual students actually engage in modification during a required information exchange? Although the potential for modification among students is present at all times, certain students may not interact because their more limited linguistic proficiency prevents them from processing certain linguistic input. Other students may understand everything that is said during a required exchange of information and therefore may not need to engage in INFORMATION GAP TASKS AND SLA 321

modification of the interaction. Such aspects of the interaction must be investigated further. Another consideration is the effect of the modification of interaction on students who are listening but not participating in a particular exchange. In a typical classroom exchange, these listeners may simply tune out, especially if the interaction is beyond their current processing capacity. However, if all participants need to know each others information, students not directly participating in a modified exchange of information may nevertheless be indirect participants in the ensuing conversational modifications. This would be especially true if they are at the same processing capacity level as at least one of the direct participants. Thus, the indirect effects of the modification of interaction on listeners are another vital area of research. The results of this study have shown that when an exchange of information is guaranteed, a great deal of modification can be generated in a nonnative-speaker group situation. Coupled with the finding from another earlier investigation (Pica & Doughty, 1983) that individual students produce more input and have more input directed toward them in group than in teacher-fronted interaction, it may seem that the exclusive use of group work in the second language classroom is in order. However, such a recommendation would be shortsighted. An important result of the earlier study must be kept in mind: Whether working in a teacher-fronted situation or engaged in group interaction, the students produced a large number of ungrammatical utterances. The teacher, therefore, was the major (if not the only) source of grammatical input in the classrooms. If a primary goal of classroom language instruction is the development of communicative competence, a component of which is linguistic competence (Canale & Swain, 1980), this important finding must not be ignored. (See Doughty, 1985, for a discussion of the effects of exclusive peer work in the classroom and Long & Porter, 1985, for opposing arguments.) Overall, however, on the basis of our combined research, it appears that group workand for that matter, pair work as well is eminently capable of providing students with opportunities to produce the target language and to modify interaction. In keeping with second language acquisition theory, such modified interaction is claimed to make input comprehensible to learners and to lead ultimately to successful classroom second language acquisition (see Long, 1981, 1983a, and Krashen, 1980, 1982, for reviews of this literature). 322 TESOL QUARTERLY

As demonstrated in the above discussion, however, group activities do not automatically result in the modification of interaction among the participants. To be effective, group interaction must be carefully planned by the classroom teacher to include a requirement for a two-way or multi-way exchange of information. Thus, the teachers role is critical not only in providing students with access to grammatical input, but also in setting up the conditions for successful second language acquisition in the classroom.

This is a revised version of a paper presented at the 18th Annual TESOL Convention in Houston, March 1984. The study reported is the second in a series of three. We would like to thank six ESL teachers from the English Program for Foreign Students at the University of Pennsylvania and the Community College of Philadelphia for devoting valuable classroom time to our data collection. We would also like to thank Richard Young and two anonymous Quarterly reviewers for helping us to clarify our statistical approach and for very constructive comments toward revision of the article.

Catherine Doughtys research focuses on classroom second language acquisition and the use of the computer as a research tool. Since 1983, she has been a Research Specialist in the Language Analysis Project at the University of Pennsylvania, where she is currently completing a dissertation in applied linguistics. Teresa Pica has published articles on second language acquisition in Language Learning and Studies in Second Language Acquisition. Since 1978, she has been affiliated with the University of Pennsylvania, currently as an Assistant Professor in the Graduate School of Education and for several years as an Instructor in the English Program for Foreign Students.

Canale, M., & Swain, M. (1980). Theoretical bases of communicative approaches to second language teaching and testing. Applied Linguistics, 1, 1-47. Davies, N. (1982). Training fluency: An essential factor in language acquisition and use. RELC Journal, 13 (1), 1-13. INFORMATION GAP TASKS AND SLA 323

Doughty, C. (1985). Classroom pidginization and creolization. Unpubhshed manuscript. Doughty, C., & Pica, T. (1984, March). Small group work in the ESL classroom: Does it facilitate second language acquisition? P a p e r presented at the 18th Annual TESOL Convention, Houston. Gass, S., & Varonis, E. (1985). Negotiation of meaning in non-native speakernon-native speaker conversation. In S. Gass & C. Madden (Eds.), Input in second language acquisition (pp. 149-161). Rowley, MA: Newbury House. Hunt, K. (1970). Syntactic maturity in school children and adults. Monographs of the Society for Research in Child Development, 53 (1, Serial No. 134). Krashen, S. (1980). The input hypothesis. In J. Alatis (Ed.), Current issues in bilingual education (pp. 144-158). Washington, DC: Georgetown University Press. Krashen, S. (1982). Principles and practice in second language acquisition. Oxford: Pergamon. Long, M. (1980). Input, interaction, and second language acquisition. Unpublished doctoral dissertation, University of California at Los Angeles. Long, M. (1981). Questions in foreigner talk discourse. Language Learning, 31, 135-158. Long, M. (1983a). Native speaker/non-native speaker conversation in the second language classroom. In M. Clarke & J. Handscombe (Eds.), On TESOL 82 (pp. 207-225). Washington, DC: TESOL. Long, M. (1983b). Training the second language teacher as classroom researcher. In J.E. Alatis, H.H. Stern, & P. Strevens (Eds.), Georgetown University Round Table on Languages and Linguistics 1983 (pp. 281297). Washington, DC: Georgetown University Press. Long, M., & Porter, P. (1985). Group work, interlanguage talk, and second language acquisition. TESOL Quarterly, 19, 207-228. Pica, T., & Doughty, C. (1983, October). Native and non-native input in the ESL classroom: An empirical study. Paper presented at the 10th University of Michigan Conference on Applied Linguistics: Input and Second Language Acquisition, Ann Arbor. Pica, T., & Doughty, C. (1985a). Input and interaction in the communicative language classroom: A comparison of teacher-fronted and group activities. In S. Gass & C. Madden (Eds.), Input and second language acquisition (pp. 115-132), Rowley, MA: Newbury House. Pica, T., & Doughty, C. (1985b). The role of group work in classroom second language acquisition. Studies in Second Language Acquisition, 7, 233-248. Pica, T., & Doughty, C. (1985c, April). ESL classrooms: Sizing up the situation. Paper presented at the 19th Annual TESOL Convention, New York.



Pica, T., Doughty, C., & Young, R. (1985, July). Does the modification of interaction lead to acquisition? Paper presented at Summer TESOL, Georgetown University, Washington, DC. Pica, T., Doughty, C., & Young, R. (1986, March). The impact of interaction on input comprehension. Paper presented at the 20th Annual TESOL Convention, Anaheim. Pica, T., & Long, M. (in press). The classroom and linguistic performance of experienced vs. ESL teachers. In R. Day (Ed.), Talking to learn. Rowley, MA: Newbury House. Porter, P. (1983). Variations in the conversations of adults learners of English as a function of proficiency level of the participants. Unpublished doctoral dissertation, Stanford University. Varonis, E., & Gass, S. (1983, March). Target language input from nonnative speakers. Paper presented at the 17th Annual T E S O L Convention, Toronto.



The TESOL Quarterly welcomes evaluative reviews of publications of relevance to TESOL professionals. In addition to textbooks and reference materials, these include computer and video software, testing instruments, and other forms of nonprint materials. Edited by VIVIAN ZAMEL
University of Massachusetts/Boston

The Computer Book Mohyeddin Abdulaziz, William Smalzer, and Helen Abdulaziz. Englewood Cliffs, NJ: Prentice Hall, 1985. Pp. xii+ 203. With the increasing recognition of subject matter as a basis for real language and so-called natural learning, it would seem desirable for ESL teachers to try more natural methodologies in language instruction. And, apart from providing language practice, it would be ideal if these methodologies were in themselves attractive to ESL students. For some teachers and learners, though, the romance of the natural and the attractive is a daily possibility, thanks to computers. As detailed in The Computer Book, computer science and programming can provide a natural, attractive context for learning English. Book-length, theoretical discussions of computer-aided instruction (CAI) in ESL are just beginning. As of now, there is but a small body of literature. Notably, Underwood (1984) has surveyed a few communicative CAI programs and has situated these within the broader frames of generative grammar and the Monitor Hypothesis. Hope, Taylor, and Pusack (1984) have considered a limited range of practical issues like lesson design and software evaluation. Both Underwood and Hope et al. have provided essentially general introductions, with little or no methodological discussion of common issues like word processing and computer programming. For those ESL instructors who are bent on creating their own CAI programs, Higgins and Johns (1984) and Ahmad, Corbett, Rogers, and Sussex (1985) have published teachers guides. These guides are also limited, as they are based on projects designed for British-made computers, with pedagogical emphases placed on developing games, drills, and so on. Gagne and Briggs (1979) and Allessi and

Trollip (1985) have compiled programming manuals that are not focused on ESL but might prove more helpful to language teachers. These manuals, which go far beyond a drills-and-games approach to language learning, discuss methods for setting up language tutorials, for example, detailing such issues as the frequency, placement, and function for various modes of posing questions, for judging learners responses, and for providing appropriate feedback. With so little discussion of CAI in ESL, there are no comprehensive theories as to how computers can be used in natural approaches to teaching. In light of this lack of theory, The Computer Book appears to have instantly advanced the discussion by presenting a practical beginners guide to programming that not only helps establish a content-specific syllabus for language practice but also serves the needs of teachers and students who have never touched a computer. According to its authors, The Computer Book is designed as an introduction to programming and data processing for highintermediate ESL students. The book presents the fundamentals of the BASIC computer language, which students apply in writing their own programs. More generally, the book centers on the development of logical processes for solving problems and supplements this development with exercises in vocabulary, grammar, and reading comprehension. In emphasizing subject matter rather than syntax, The Computer Book offers a compelling context for students to use new lexical items and grammatical forms or, as the authors express it, to get a lot of implicit rather than explicit language practice (p. ix). The Computer Book consists of (a) an introductory chapter, (b) eight lessons (each requiring approximately 6 to 8 hours of class work), and (c) two appendixes containing BASIC commands and a sample sort program. The first few pages of each lesson provide previews of specialized vocabulary and content. The main texts of the eight lessons present intricate procedures for planning, writing, and processing BASIC programs. The sections that follow the main texts are specifically addressed to ESL learners. Entitled Understanding the Text, Review, and Challenge, these sections follow a dual agenda: In terms of implicit language practice, they recap pertinent material by asking learners to complete technical outlines, paraphrase concepts, and define key terms; more explicitly, they test learners command of certain grammatical points. 328 TESOL QUARTERLY

The appropriateness of some of these grammar quizzes is questionable, however, in the larger context of imparting operational data for programming. Similarly, review of technical information by means of cloze formats seems artificial, especially when the parts of speech to be filled in are adverbs, articles, and so on. Nevertheless, one can imagine ESL classes in which such exercises might be applicable; conversely, one can skip these sections if they appear to overemphasize points of grammar that might be better addressed elsewhere. The introductory chapter offers general information regarding CAI and briefly reflects upon social issues that are potentially relevant to the growing use of computers, issues such as unemployment and invasion of privacy. Lesson 1 concentrates on defining descriptive terms for hardware and software. Lesson 2 (which I recommend to all of us who are not programmers) neatly details how to think like a computer, outlining the conventional symbols for constructing flowcharts, for example. The third and fourth lessons describe ways of planning and writing a BASIC program and how to log it in. The last four chapters form an increasingly complex sequence that explains methods for structured programming, table and array processing, composing and maintaining files. (If these terms scare you, you will be relieved by how clearly The Computer Book explicates each topic, page by page.) The unique feature of The Computer Book is that within a computer science framework, it stimulates thinking processes which engage learners in experimenting with and acquiring new language. When attending to the logic of programming concepts, for instance, learners are encouraged in Review sections to explore novel vocabulary to which they have already been introduced in earlier sections. This exercise on paraphrasing is one example: Rewrite the following sentences in your own words. Be sure to convey the same information but not use the italicized words or their derivatives. 1. Division has priority over addition in evaluation. 2. Whenever line 40 is executed, line 50 will be bypassed. (p. 63) Again, when analyzing the meaning of various notations, students can quite naturally practice speaking and review points of grammar as well: Evaluate the following expressions as a computer would. Practice reading them aloud, paying attention to prepositions.



Let A=l, B=2, C=3 1. A+(C B) *2= 2. A**2 + (B * A) C = (p. 65) On the other hand, more explicit grammar reviews, such as exercises in pronoun reference or the use of the passive voice, seem out of place. Relative to learners attention, many of these exercises compete, as it were, with the overall emphasis on logic and mastery of technical data. The Computer Book is a practical extension of the notion that student-directed, interactive technology can assist language learning. Moreover, the focus on content-specific language represents a challenging next step for practitioners who are interested in integrating computers into ESL curricula. Will The Computer Book convince a large number of ESL instructors to adopt programming as a natural approach to teaching? Probably not. But the relatively untechnical material of the first few chapters seems particularly inviting for those instructors who are open to learning a little programming and who see the benefits of teaching programming as one way of thinking logically. In sum, The Computer Book serves both as an excellent primer for novice programmers (be they teachers or students) and as an innovative language textbook for a wide range of ESL learners. While the authors designed The Computer Book for highintermediate precollege learners, many of its problem-solving exercises could be adapted for university-level course work or tutorials. Lastly, since the book concentrates primarily on the development of logical processes, it is conceivable that students could do much of their work without ready access to computers. The Computer Book is thus an ideal vehicle for situations in which there are not enough computers to accommodate all ESL students who want to learn programming. REFERENCES
Ahmad, K., Corbett, G., Rogers, M., & Sussex, R. (1985). Computers, language learning and language teaching. Cambridge: Cambridge University Press. Allessi, S. M., & Trollip, S. R. (1985). Computer-based instruction, Englewood Cliffs, NJ: Prentice-Hall. Gagne, R. M., & Briggs, L.J. (1979). Principles of instructional design. New York: Holt, Rinehart and Winston. Higgins, J., & Johns, T. (1984). Computers in language learning. Reading, MA: Addison-Wesley.



Hope, G. R., Taylor, H. F., & Pusack, J.P. (1984). Using computers in teaching foreign languages. Orlando, FL: Harcourt Brace Jovanovich. Underwood, J.H. (1984). Linguistics, computers, and the language teacher. Rowley, MA: Newbury House. JACK KIMBALL
Harvard University and University of Massachusetts/Boston

Review of the State-of-the-Art gies Implemented in Programs Funded by the Department of Prepared for U.S. Department

of Educational TechnoloServing LEP Students Education: Final Report of Education

COMSIS Corporation. Rosslyn, VA: InterAmerica Research Associates, Inc. (National Clearinghouse for Bilingual Education), 1984. Pp. iv+ 166.

s This highly informative and readable report presents the findings

of a study of the use of new technologies in bilingual programs funded by the U.S. Department of Educations Office of Bilingual Education and Minority Languages Affairs. The funding requests for 604 projects were examined, and 114 projects that used a new technology in instructional methodology were identified. Nine of these projects were then selected for detailed study. Two basic technologiesvideo and computer technologywere represented in the sample, with the majority of the report centering on computer technology. Both computer and video technologies were identified as providing significant changes in the delivery of instruction to limited English proficient (LEP) students. The report asserts that with proper training of teachers and administrators, computers have the potential for permitting students to learn at their own speed in a highly motivating environment. The major impediments to the effective, use of computers were found to be lack of instructionally and technologically sound software and the lack of training for planning computer usage. Another basic problem identified by the report was that objectives tended to be too broad to permit effective planning and evaluation. The report begins with a historical overview of the nine sites visited, which were selected to obtain a distribution of projects by funding year, technology, geographical location, grade level, and native language. In each project, English as a second language was the principal focus of the application of technology, in large part

due to the lack of software in the students native languages. Participation in computer-aided instruction (CAI) varied among the sites from supplementary remediation in laboratory environments to computers in the classroom. Two types of computers were used by the projects, the most prevalent of which was the Apple, manufactured by the Apple Corporation. The other type was Radio Shacks TRS-80. Both systems were perceived to be highly reliable. Factors determining the choice of equipment included familiarity, compatibility with the mainstream curriculum, and availability of software. Computer configurations were of two basic types, stand-alone and networked. The advantages and disadvantages of these two configurations are discussed, including the fact that stand-alone workstations require a lower level of computer technology and that networked stations risk entire breakdown due to the failure of the master workstation. At most sites, the computers were placed in a computer lab, with students having access to them on a pull-out basis. Some projects elected to place the computers in the classroom. Where teachers had direct control of the computers, they tended to become more familiar with their capabilities, and the LEP students used computers in more creative endeavors, such as writing stories complete with computer-produced illustrations. Computers in laboratory situations generally had the effect of reducing the teachers level of involvement or appreciation. Teachers who were best able to utilize computers were those who recognized that effective use required a change in their teaching methods. Apparently, some of the site personnel did not fully understand the proper handling procedures for magnetic storage media. When not in use, computer floppy disks were left on desks or tables, and eating in the computer area was permitted at some of the sites. The report emphasized that computers appeared to motivate students. LEP students in the computer-based projects tended to have better attendance and to show improvement in other class work. Limited access to computers often produced curious results. For example, LEP students who were finally mainstreamed into the general school population often felt as if they were being punished, since they could no longer use computers. Students seemed to prefer using computers to more traditional teaching approaches, although they could not explain exactly why this was so. All of the projects selected and acquired software in a five-step process: definition of requirements, software identification, software screening, staff training, and software evaluation. 332 TESOL QUARTERLY

Underfunding for software acquisition was a common problem. The report repeatedly cited the need for a method of acquiring information on ESL computer software. A major complaint of project personnel was the impact of poor-quality software on student performance. The report lists several quality-related problems, including the inability of some software to find all correct responses or to explain why incorrect responses were wrong. Some 70% to 80% of the software reviewed was considered unsuitable for the sites use, due to the lack of a sound educational approach, technological deficiency, poor quality of programming, or inability to retain a students attention. In projects where teachers played a role in screening software, a greater understanding of the role of computers in education was exhibited. The report found that there often was a lack of positive communication between suppliers of software and the teachers who used their products. For example, one site, having purchased several computer programs from a software publisher, was asked by the publisher to provide reviews of their programs. When the sites review proved to be highly critical, the publisher was never heard from again. The products of traditional publishing houses of educational material which are beginning to enter the educational software market have also been for the most part disappointing to the site personnel. Some of the publishers, according to the report, have apparently adopted a cautious approach to educational software to minimize the impact on their textbook market, although there is evidence that this tendency is changing. Of the nine project sites visited, only two used software which utilized a language other than English, and only one used software designed specifically for ESL use. Computer software was generally geared for drill and practice and for reinforcement of course material. Though graphics may have been imaginative, for the most part, course material was presented using traditional techniques. Overall, the report argues that students do learn more effectively with the proper application of technological tools. On the debit side, however, are the following findings: (a) The application of technology to bilingual education is extremely limited; (b) planning for the use of technology is generally insufficient; (c) methods for evaluating the effectiveness of computer technology are inappropriate; and (d) regardless of the technology used, there is still a need for high quality teachers. Appended to the report are a number of useful items, including a comprehensive 50-page handbook, Debugging CAI: A Handbook REVIEWS 333

for Planning Computer Assisted Instruction (by J.B. Ippolito and R.E. Saunders and edited by R.W. Hoar, Jr.), and a list of available microcomputers. The variety of software used in the projects is listed and evaluated, and a detailed courseware evaluation form is provided, This report should prove valuable to all educators who have an interest in computer-aided instruction. It addresses major sources of optimism and concern about CAI which are often discussed without the benefit of empirical evidence.
Florida State University




The TESOL Quarterly invites readers to submit short reports and updates on their work. These summaries may address any areas of interest to Quarterly readers. Authors addresses are printed with these reports to enable interested readers to contact the authors for more details. Edited by ANN FATHMAN
College of Notre Dame

Recognition of Sentences by Native and Nonnative Speakers of English: Probing the Role of Imagery
The Pennsylvania State University


University of Arizona

The study of imagery and the processes underlying the comprehension and recognition of verbal information forms the foundation stones of cognitive psychology. Research has identified the positive effects of mental imagery on recall (Begg & Paivio, 1969; Paivio, 1971, 1977, 1978; Paivio & Begg, 1971). Paivio and Begg (1971), for example, presented abstract and concrete sentences to subjects and found that on a recognition memory task, changes in the meaning of the concrete, highimagery sentences (e.g., The vicious hound chased a wild animal) were detected more readily than changes in the meaning of the abstract, lowimagery sentences (e. g., The absolute faith aroused an enduring interest). Previously, Begg and Paivio (1969) had found that subjects noticed a semantic change more easily than a change in sentence structure for concrete, high-imagery sentences than they did for abstract, lowimagery sentences. Evidence indicating that visual imagery is an epiphenomenon which accompanies learning but which has nothing functional to do with it has also been presented (Day & Bellezza, 1983; Pylyshun, 1973, 1981); however, the study of mental imagery as a mediator of cognitive tasks continues to be a major concern of cognitive psychologists. Although investigation of the role imagery plays in the processing of information by nonnative speakers has been virtually ignored in memory research, as Rose (1975) puts it, the bilingual is the subject par excellence for investigations of cognitive processes (p. 150). The study reported here yields information concerning the role of mental imagery 335

in the recognition of high- and low-imagery sentences and in the detection of syntactic changes in such sentences by native speakers (NSs) and nonnative speakers (NNSs) of English. The subjects in the study were 22 American and 22 Soviet teachers of English as a second/foreign language. The 40 stimulus sentences, originally constructed by Smith (1981), each consisted of the syntactic arrangement of Article/Adjective/Noun/Verb/Article/Adjective/Noun. Half of the sentences were concrete, and of these, half were affirmative (e.g., The black horse swam the deep lake), and half were negative (e.g., The wooden bowl was not overflowing with ripe apples). The other 20 sentences were abstract, and once again, half were affirmative (e.g., The large quantity lacked any real quality), and half were negative (e.g., His greatest virtue was not his irrepressible confidence). The subjects were instructed to listen over language-lab headphones to the 40 audiotaped sentences and to try to remember as many of the sentences as possible. After presentation of all 40 sentences and a 30second rest pause, the subjects were presented with 40 additional sentences, 20 of which had been presented previously and 20 of which were new sentences in that they had been changed from negative to affirmative or affirmative to negative, as appropriate, compared with the first list. For example, The black horse swam the deep lake was changed to The black home did not swim the deep lake, while His greatest virtue was not his irrepressible confidence was changed to His greatest virtue was his irrepressible confidence. Five sentences from each of the four categories (Concrete, Abstract, Affirmative, Negative) were thus changed to yield 20 new (i.e., changed) sentences to accompany the 20 old (i.e., unchanged) sentences initially presented. Subjects were asked to indicate, by marking yes or no on the response sheet, whether they thought they had heard each sentence during initial presentation of the 40 sentences. Altogether, subjects heard 80 sentences presented. Data from the NS and NNS subjects were analyzed using a 2 x 2 x 2 x 2 (Language Group x Imagery x Verb Structure x Change in Stimulus Presentation) repeated measures analysis of variance. Three variables, imagery (concrete versus abstract sentences), verb structure (affirmative versus negative sentences), and change in stimulus (changed versus unchanged sentences) were repeated on each subject in the language groups (NS versus NNS). The Scheff method of multiple comparisons was used for post-hoc analyses. Examination of the data revealed that NS subjects, in general, recognized more unchanged sentences than did NNS subjects: F (1,42) = 53.55, p <.001. A significant main effect was also found for change in stimulus presentation: F (1,42) = 28.81, p <.001. Unchanged sentences were correctly recognized more often by both NS and NNS subjects than were changed sentences. There was a significant interaction between language group and change in stimulus presentation: F (1,42) = 9.12, p <.01. Scheff test results indicated that NS 336 TESOL QUARTERLY

subjects correctly recognized unchanged sentences more often than changed sentences, and this was also true for NNS subjects. A significant interaction between verb structure and change in stimulus presentation was also found: F (1,42) = 6.04, p <.05. Post-hoc analyses indicated that affirmative-unchanged sentences were recognized more often than were either affirmative-changed, negativechanged, or negative-unchanged sentences. Negative-unchanged sentences were found to be the second most frequently recognized category. This interaction between verb structure and change in stimulus presentation indicates the existence of a three-step hierarchy in the ease of recognition of affirmative-negative transformations. The scale ranges from affirmative-unchanged (easiest to recognize) to negative-unchanged (more difficult to recognize) to both affirmativeand negative-changed sentences (most difficult to recognize). Contrary to expectation, the findings indicated that concrete sentences, either affirmative or negative, were not recognized by either NS or NNS subjects with significantly greater frequency than were abstract sentences. Mental imagery did not appear to play a critical role in the recognition task for the NS and NNS subjects in this study. This studys nonsignificant findings concerning the role of mental imagery in the processing of verbal information suggest examination of several issues pertinent to imagery research. First, the alteration of sentences in this study involved only the syntactic change of an affirmative-negative verb transformation. Making such a single syntactic change may not be an optimal way to examine the function of imagery in the processing of verbal information by NS and NNS subjects. Perhaps other forms of structure or content manipulations (e.g., changes in adjective form or in verb tense) or changes in vocabulary items would yield different results concerning the role of imagery in sentence recognition. Second, using lists of disconnected sentences which have undergone a surface-structure syntactic change may not be an effective method of determining the function of imagery in verbal learning. Other instruments may prove more appropriate for investigating the mediating strategy and effect of visual imagery. Finally, the generalizability of the findings may be restricted by the uniqueness of the sample, which consisted of second/foreign language instructors, who may have been sensitized to the listening and recognition tasks as a result of their language training and professional backgrounds. L2 teachers often engage students in the structural manipulation of decontextualized segments of language during drill and practice sessions. The fact that no main effect was found for the imagery variable (concrete versus abstract sentences) may therefore be a function of the uniqueness of the sample in this study rather than a refutation of the positive effect of imaginal mediators in verbal learning. The use of pictures in the minds eye as causal determinants of verbal BRIEF REPORTS AND SUMMARIES 337

learning and human memory performance should be further investigated using diverse NS and NNS samples. REFERENCES
Begg, I., & Paivio, A. (1969). Concreteness and imagery in sentence meaning. Journal of Verbal Learning and Verbal Behavior, 8, 821-827. Day, J. C., & Bellezza, F.S. (1983). The relation between visual imagery mediators and recall. Memory and Cognition, 11, 251-257. Paivio, A. (1971). Imagery and verbal processes. New York: Holt, Rinehart and Winston. Paivio, A. (1977). Images, propositions, and knowledge. In J.M. Nicholas (Ed.), Images, perception, and knowledge: Papers deriving from and related to the Philosophy of Science Workshop at Ontario, Canada, May 1974 (pp. 47-71). Boston: D. Reidel Publishing. Paivio, A. (1978). A dual coding approach to perception and cognition. In H.L. Pick, Jr., & E. Saltzman (Eds.), Modes of perceiving and processing information (pp. 39-51). Hillsdale, NJ: Lawrence Erlbaum. Paivio, A., & Begg, I. (1971). Imagery and associative overlap in short-term memory. Journal of Experimental Psychology, 89, 40-45. Pylyshun, Z. W. (1973). What the minds eye tells the minds brain: A critique of mental imagery. Psychological Bulletin, 80, 1-24. Pylyshun, Z.W. (1981). The imagery debate: Analogue media versus tacit knowledge. Psychological Review, 88, 16-45. Rose, R.G. (1975). Introspective evaluations of bilingual memory processes. Journal of General Psychology, 93, 149-150, Smith, C.D. (1981). Recognition memory for sentences as a function of concreteness/abstractness and affirmation/negation. British Journal of Psychology, 72, 125-129. Authors Address: c/o Dunkel, Department of Speech Communication, The Pennsylvania State University, 212 Sparks Building, University Park, PA 16802

Dominant Administrative Styles of ESL Administrators

United Nations Institute for Training and Research

The main objectives of this study (Reasor, 1981) were to describe the background and training of ESL administrators and to evaluate their self-perceived, dominant administrative styles. For practical reasons, the study was limited to ESL administrators in U.S. colleges and universities. From the membership of TESOL and the National Association for Foreign Student Affairs, 375 such administrators were identified and surveyed. Data were collected in the late summer of 1980, and responses from 103 ESL administrators (27.5% of the identified population) were analyzed. 338 TESOL QUARTERLY

Two instruments were used to collect data, the first of which was the ESL Administrator Survey, a questionnaire constructed for this study which contained 14 specific questions and 1 open-ended one. The 14 questions focused on the administrators specific educational backgrounds, past teaching experience, past administrative experience, overseas work experience, present supervisory load and responsibilities, size of program administered, status in organization, reporting authority in their organization, tenure status in their organization, L2 testing experience, and self-perception of their orientation to people or tasks in their administrative roles. The open-ended question asked for the respondents ideas about needed training for ESL administrators. These situational variables were selected on the basis of previous research in ESL administration (Escobar & Daugherty, 1975; Gay, 1970; Inman, 1978; Jarvis & Adams, 1979; Norris, 1975; Robinett, 1972 Streiff, 1970; Wardhaugh, 1972; Wilcox, 1980). The second instrument, which measured dominant administrative styles, was selected to reflect a research thrust in educational administration (Hey & Miskel, 1978). This instrument (see Reddin, 1974), the Educational Administrative Style Diagnosis Test (EASDT), has been used and validated in previous research on educational administrators (Arnold, 1977; Bandy, 1976; Landers, 1979; Myers, 1977; Schields, 1976; Welch, 1976). The EASDT is based on the ThreeDimensional Model (3-D) of leadership style analysis (Reddin, 1970) and parallels the structure and format of Reddins (1974) Management Style Diagnosis Test, which has been used to assess the leadership styles of managers in business and industry. The EASDT asks the respondents to project situational elements from their own specific educational setting into those general management concepts referred to in each forced-choice question. There are 56 paired statements on the EASDT, and each of these 112 statements represents both a characteristic of one of eight administrative styles in the 3-D Model and a situational element from educational administration. The test directions state in part, In every case pick the one statement that best describes your behavior if you were faced with the circumstances described. Respondents must choose a statement from every pair, or the test cannot be scored. The following are examples of forced-choice pairs from the test: 1. A He could do a better job at maintaining good relationships with those above him. B He tries to introduce changes very gradually so no one will become upset. 18. A He tends to avoid or to disagree with students, thinking that students often know little of the practical side of things, B He responds to disagreement and conflict by referring to policy and procedures. BRIEF REPORTS AND SUMMARIES 339

The construction of the test is quite complex (Reddin, 1970, 1974); however, no changes were made in the test for this administration (including the sexist language), since the ESL administrators responses were to be compared with those of previous respondents. The three dimensions measured by the EASDT are the administrators task orientation (TO), relationships orientation (RO), and effectiveness (E). From these basic dimensions of leadership analysis, Reddins model, like previous models (Blake & Mouton, 1964; Fiedler, 1967; Halpin, 1956; Likert, 1967), posits that an administrators organizational behavior can be categorized into one of four basic Dominant Administrative Styles (DAS), each characterized by varying degrees of TO and RO. These styles are (a) integrated, which consists of behavior that is both task- and people-oriented; (b) related, which is people-oriented with emphasis placed on interpersonal relations on the job; (c) dedicated, which is task-oriented and often called the authoritarian style; and (d) separated, which is rule- and procedure-oriented and known as the bureaucratic style. The data-collection instruments which were returned were carefully screened and analyzed. All respondents who were not supervising at least one other .ESL teacher were rejected, as were respondents who had not completed the EASDT, leaving a total of 103 respondents. DAS profiles (Reddin, 1974) were determined, and responses from the ESL Administrator Survey were tabulated. Descriptive findings from the ESL Administrator Survey included the following: 61% held a masters degree and 30% a doctorate; 45% had a degree in some form of linguistics or TEFL, while fewer than 1% had a degree outside the humanities. In addition, 71% had 10 years or fewer of teaching experience, 69% had been ESL administrators for fewer than 5 years, 77% reported directly to one person above them in the academic administration, 50% supervised 10 or fewer teachers, and 20% supervised more than 20 teachers. Of the respondents, 73% had worked abroad, 76% had L2 testing experience, and 55% did not have rank or tenure in their college or university. In addition to a DAS profile, each respondents EASDT also provided a score for effectiveness (+) or ineffectiveness () for the DAS. In the 3-D Model, the four basic styles are all neutral; effectiveness or ineffectiveness of each style depends on the situation in which it is used. Thus, the four basic styles are seen as eight stylesfour effective ones and four ineffective ones. Reddin (1970) describes each of the eight styles in detail, noting that no style in this model is considered ideal. Situational variables determine whether ones basic style will be effective or ineffective, and using style flexibility to match ones style with the situational demands of the job is the crux of Reddins theory. The distribution of the eight styles as they occurred among the population of ESL administrators is shown in Table 1. Compared with 537 other educational administrators who had taken the EASDT and who tended to show a more equal distribution among the four basic 340 TESOL QUARTERLY

styles, the ESL administrators clustered heavily (69%) around the separated style. This finding, however, needs to be more thoroughly examined in future research.
TABLE 1 Dominant Administrative Styles of ESL Administrators (N= 103)

Separated-style administrators are characterized in this model as cautious, careful, conservative, and orderly. Such administrators usually prefer paperwork, procedures, and facts to people or tasks. They are accurate, precise, correct, perfectionist, steady, deliberate, patient, calm, modest, and discreet. Reddin (1970) theorized that this style may be engendered by lengthy training programs which individuals in some professions undergo. He also suggested that certain personality types tend to have a separated style when they become administrators in organizations. Whether this is the most appropriate style for ESL administration depends upon the clear identification of the situational variables for the job, which this study only began to explore. Further research is needed to explore why 62% of the administrators in this study perceived themselves to be ineffective in their present styles. Another question raised by this study is whether ESL administrators have received the most appropriate training for the jobs they are now doing. As the data suggest, these administrators are predominantly liberal arts majors with classroom teaching experience; however, the job demands might require knowledge, skills, and abilities in such areas as goal setting, decision making, group dynamics, managerial problem solving, time management, task analysis, human resources development, needs assessment, and budget planning. These and other management skills are not generally included in programs leading to linguistics or TEFL degrees, which 45% of these administrators hold, nor in humanities degrees, which 99% of them hold. Hypotheses concerning the relationship between administrative style and variables described on the BRIEF REPORTS AND SUMMARIES 341

questionnaire were rejected at the .05 level, but further research is recommended. l REFERENCES
Arnold, F. (1977). An investigation of administrative style and job satisfaction of elementary principals. Unpublished doctoral dissertation, The American University. Bandy, L.S. (1976). Relationships of perceived administrative styles of elementary principals and selected situational variables. Unpublished doctoral dissertation, The American University. Blake, R. R., & Mouton, J.S. (1964). The managerial grid. Houston: Gulf Publishing. Escobar, J. S., & Daugherty, J. (1975). Handbook for the ESL/ABE administrator. Arlington Heights: Bilingual Education Service Center. Fiedler, F.E. (1967). A theory of leadership effectiveness. New York: McGrawHill. Gay, B. (1970, June). Thats what ESL programs are made of. NAFSA Newsletter, pp. 7-10. Halpin, A. W. (1956). The leader behavior of school superintendents. Columbus: Ohio State University. Hoy, W. K., & Miskel, C.G. (1978). Educational administration: Theory, research and practice. New York: Random House. Inman, M. (1978). Foreign languages, English as a second/foreign language, and the U.S. multinational corporation. Arlington, VA: Center for Applied Linguistics. Jarvis, G. A., & Adams, S.J. (1979). Evaluating a second language program. Arlington, VA: Center for Applied Linguistics. Landers, T.J. (1979, March). Administrator effectiveness: A comprehensive behavioral model. Paper presented at the annual convention of the Association for Supervision and Curriculum Development, Detroit. Likert, R. (1%37). The human organization: Its management and value. New York: McGraw-Hill. Myers, J.R. (1977). Predicting public education administrator styles through administrator values, organization environment factors, self-actualization levels, and specific demographic information. Unpublished doctoral dissertation, The American University. Norris, W.E. (1975). Guidelines for the certification and preparation of teachers of English to speakers of other languages in the United States. Washington, DC: TESOL. Reasor, A. W. (1981). Administrative styles of English-as-a-second-language administrators. Unpublished doctoral dissertation, The American University. Reddin, W.J. (1970). Managerial effectiveness. New York: McGraw-Hill. Reddin, W.J. (1974). Interpretation manual for the management style diagnosis test. Fredericton, N. B., Canada: Organizational Tests. Robinett, B.W. (1972). The domains of TESOL. TESOL Quarterly, 6, 193-207.

would especially like to thank (Grace S. Burkhart, Robert FOX, Joel Burdin, Edwin Burkhart, and James E. Alatis for their advice and support during this project.



Schields, R.L. (1976). A study of the validity of the educational administrative style diagnostic test. Unpublished doctoral dissertation, The American University. Streiff, P. (1970). Some criteria for the evaluation of TESOL programs. TESOL Quarterly, 4, 365-370. Wardhaugh, R. (1972). TESOL: Our common cause. TESOL Quarterly, 6, 291303. Welch, D.F. (1976). A comparative study of the administrative styles of public and independent school principals and headmasters. Unpublished doctoral dissertation, The American University. Wilcox, G.K. (1980, March). Suggested guidelines for the selection and/or preparation of ESOL program administrators. Paper presented at the 14th Annual TESOL Convention, San Francisco. Authors Address: Al Khazzan Street Residential Center, King Khalid Building, Apartment 1135, Riyadh, Saudi Arabia

The Effects of Modifying the Formality Level of ESL Composition Questions

The University of Michigan

We know of no investigations of the effect of topic on compositions written by ESL populations under closed-book, essay-examination conditions. In L1 contexts, several studies have been concerned with varying the amount and kind of information in the topic, or stimulus (e.g., Smith et al., 1985). In the most directly relevant of these studies, Brossell and Hoetker Ash (1984) investigated the effects of personal (you) as opposed to neutral phrasing and of a question versus an imperative format. They found no significant results and concluded that their study produced no evidence to support the contention that small changes in the wording of essay examinations of otherwise similar construction affect writers or the holistic ratings given their essays (p. 425). The study reported here examined the extent to which the short and general topics used in the Composition section of the Michigan English Language Assessment Battery (MELAB) (English Language Institute, 1984) generate rhetorically and linguistically unambitious answers that fail to indicate fully the communicative abilities of candidates. Since this investigation was part of a series of studies aimed at developing a new test specifically designed for the evaluation of nonnative-speaking (NNS) graduate and transfer students, it also examined the extent to which more ambitious and more academically appropriate writing could be elicited by simply increasing the level of formality of the topic. An example of simple and academic topic variants is given below.



Simple: Academic:

Would you prefer to be part of a large family or a small one? Family size tends to vary according to a number of factors, such as culture, religion, mortality rate, and level of economic development. What are the advantages and disadvantages of small nuclear families as opposed to larger extended family units? State your personal preference for one of these family types and explain the reasons behind that preference.

The subjects were 32 NNSs enrolling at the University of Michigan in Fall 1985. Each subject wrote two compositions under test conditions. In the first condition, the subjects wrote on academic topics as part of their English proficiency reevaluation (time limit 45 minutes). Two weeks later, the same subjects23 of whom were enrolled in an academic writing course and 9 of whom had been exempted from ESL writing classestook a version of the standard MELAB Composition, responding to simple topics which were different in content area from the topics in the first condition (time limit 30 minutes). Following standard MELAB scoring procedures, all compositions were scored by two experienced raters. Mean MELAB scores were 77.4 for the simple-topic condition and 78.5 for the academic-topic, with scores ranging from 67 to 93. The two sets of performances were analyzed through paired t tests, and statistically significant differences at the .01 level were found for several variables: Simple-topic compositions (a) were longer than the academic topic compositions, as measured by both words written per 30 minutes and sentences written per 30 minutes; (b) contained more subordination (per standardized length); (c) exhibited greater use of the first-person, singular pronoun; and (d) contained more morphological errors. The following variables were found to be significant at the .05 level: Simple-topic compositions exhibited a lower proportion of Graeco-Latin vocabulary (per standardized length), as measured by procedures developed by Corson (1982), and had higher proportions of total errors and of syntactic errors. Nonsignificant variables included sentence length and frequency of logical connectors, second-person pronouns, and the passive voice. Interviews with 5 of the subjects revealed their strong general preference for the more academic and elaborate questions because, as one interviewee put it, it is easier to understand what they want us to write. Interestingly, the academic-topic sessions did not give rise to as many preliminary queries and requests for clarification as is customary in standard MELAB test sessions. A general comparison of performance in the two formality conditions reveals relatively few and relatively small differences ascribable to the



experimental variable. Some of these differences are predictable enough, such as the increased use of first-person pronouns when responding to simple and general topics and the correlation between increased speed of writing and increased error rate. On the other hand, the enhanced utilization of Graeco-Latin lexis in the academic-topic condition is potentially interesting, and the amount of subordination is not easy to reconcile with general expectations. There was evidence of at least two subgroups: one apparently capable of some variation in its rhetorical and stylistic response to the level of formality of the stimulus (most strikingly revealed in the contrasting opening paragraphs) and one apparently unable, unwilling, or unaware of the need to match a response to the level of formality of the stimulus. While it is not clear whether this capacity for variation correlates with writing ability, as measured by MELAB scores, there was certainly no relationship between students self-perception (as manifested in the five interviews) and their actual writing performance. The verdict of this study would seem to be the Scottish one of not proven, especially as the research design did not permit equal writing time in the two conditions. On the other hand, it should be noted that in both conditions, advice to candidates about assessment criteria remained the same: You will be graded on how well you communicate your ideas, not on your ideas themselves. The effect of changing this advice in various ways is currently being investigated, for it would seem easier to communicate well ideas of lesser cognitive complexity.1 REFERENCES
Brossell, A., & Hoetker Ash, B. (1984). An experiment with the wording of essay questions. College Composition and Communication, 35, 423-425. Corson, D.J. (1982). The Graeco-Latin (G-L) instrument: A new measure of semantic complexity in oral and written English. Language and Speech, 25, 1-10. English Language Institute, Testing and Certification Division. (1984). Michigan English language assessment battery. Ann Arbor: University of Michigan. Smith, W. L., Hull, G. A., Land, R. E., Moore, M. T., Ball, C., Dunham, D. E., Hickey, L. S., & Ruzich, C. W. (1985). Some effects of varying the structure of a topic on . college students writing. Written Communication, 2: 73-89. Authors Address: English Language Institute, North University Building 2001, The University of Michigan, Ann Arbor, MI 48109

1 This study was funded by the Testing and Certification Division of the English Language




The TESOL Quarterly invites commentary on current trends or practices in the TESOL profession. It also welcomes responses or rebuttals to any articles or remarks published here in The Forum or elsewhere in the Quarterly.

Women Students in Our Institutions: A Response to the U.N. Decade for Women (1976-1985]
Mills College

A recent issue of the Decade For Women Post Date (International Womens Tribune Center, 1985b) shows a picture of a woman carrying a sign on her head which reads, If its not appropriate for women, its not appropriate. The message carried in these words from the third international womens conference held during the United Nations Decade for Women (1976-1985) rings as true for our field of endeavor as it did for those attending the workshop on appropriate technology at the University of Nairobi, where the stinging words first echoed. If our programs, institutions, and classes are not appropriate for the women we serve, then they are not appropriate. The United Nations World Conference to Review and Appraise the Decade for Women was held in Nairobi from July 15 to July 25, 1985. Over 157 governments and 2,020 official delegates were represented at the World Conference, with an additional 14,000 persons attending the concurrent Forum 85, the nongovernmental organization counterpart to the official World Conference. Including press and rural women from different parts of Kenya, the unofficial estimate of participants in. this international meeting of women rises to 20,000 persons (International Womens Tribune Center, 1985a). The purpose of the World Conference was to review womens gains and losses during the past decade, a decade officially designated by the United Nations as the Decade for Women, and to propose new strategies for the years 1985-2000. One of the major preparatory documents for the conference was Forward Looking 347

Strategies of Implementation for the Advancement of Women and Concrete Measures to Overcome Obstacles to the Achievement of the Goals and Objectives of the United Nations Decade for Women (United Nations Commission on the Status of Women, 1984). This document urges international, national, regional, and local governmental bodies as well as agencies, organizations, and institutions to conduct thorough evaluations of their policies with regard to the treatment of women. With the World Conference in mind, it seems appropriate at this time to review the policies of our ESL institutions and to make those changes necessary for the field of ESL to meet the United Nations goal of equality for women in the area of education (United Nations, 1975). Statistics from the Institute for International Education indicate that there are more than 98,000 foreign women students in the United States ( TESOL Newsletter, 1984). Since 1984, many more immigrant and refugee women have come to participate in our ESL classes. Yet often we do not think of their special needs as we design programs, train teachers, and develop curriculum. The purpose of this discussion is to help administrators and teachers become aware of how to be more sensitive to the needs of women students studying in ESL institutions. To that end, I will discuss how to conduct an in-depth evaluation of ESL institutions to measure our sensitivity to the needs of women students. I will also suggest ways to improve our programs, following guidelines established by the U.N. Commission on the Status of Women (1984) in their Forward Looking Strategies document, The three elements of the evaluation include (a) a workshop to increase awareness of the concerns of women students; (b) a center-wide examination of staff, student services, and curriculum; and (c) opportunities for continued professional development. The task of the workshop, which is intended as a springboard for discussion in the area of womens concerns, is to identify special problems that women students face during culture shock and during the initial stages of cultural adjustment. A secondary purpose is to provide an informal setting in which people can begin to think about the degree to which our ESL programs are appropriate for women. The workshop can last from several hours to a weekend in length, depending on available resources and staff. A good way to begin the discussion is with a review of the definition of culture shock and the various phases in the process of cultural adjustment. Men and women alike share many of the same symptoms of stress that are characteristic of initial culture shock and gradual cultural 348 TESOL QUARTERLY

adjustment. Women, however, may face a unique set of problems, many of which are associated with independence, goal orientation, and assertiveness. Attempting to identify the issues which are unique to women can be interesting and enlightening. Of course, not all women from a particular cultural background will face the same set of adjustment problems in the United States. Personality factors as well as economic, class, and religious factors make each womans adjustment process unique. Nonetheless, certain patterns emerge both within a particular cultural group and across cultures. For example, some women may be faced, during orientation week, with banking and housing arrangements which are baffling to them. In their native countries, a father, brother, or husband may have handled all of the students financial affairs. Similarly, it is customary in some cultures for a woman to live in the parents home until she marries. The freedoms of being in the United States may require a type of independence which is new and stressful for this type of female student. A second type of problem may become apparent as the semester progresses. Some of the women students may seem to lack any clear goal orientation, either for their academic studies or for a future career. Whereas men are under pressure to complete their English programs so that they can become family providers, women may experience stress due to a lack of a clear focus about their future. Not raised with the same expectations as men, women must struggle over their identity as homemakers, students, career women, and mothers. This lack of focus may be especially felt during the extreme stages of culture shock, when women students may need guidance. Assertiveness may also be more problematic for women than for men. While both men and women who come to the United States are often faced with what they perceive as American competitiveness (Shimazu, 1984), the adjustment is often greater for women. Shimazu notes that Japanese male students, coming from a culture which values cooperation and intuition, often report difficulty in the competitive American university setting. Here, students are expected to speak in a direct manner and to present convincing arguments to document their opinions. Even more than Japanese men, however, Japanese women are trained to avoid confrontation. In U.S. coeducational classrooms, these same women are expected to participate, to give opinions, and perhaps even to disagree with a male student or professor. The cognitive dissonance experienced by some Japanese women is highly stressful. THE FORUM 349

While some women may have difficulty being as assertive as is called for in certain situations in the United States, others may be perceived as being overly assertive. For example, one Venezuelan woman from our English Center was mistaken for a prostitute at a local shopping center. She was unnerved by the way in which her assertiveness, friendliness, and manner of dress were perceived in this culture. Since she did not want the incident to be repeated, we talked about appropriate body language and dress for shopping in the U.S. This womans experience, which is offered as an example of a cultural adjustment issue specific to women, should not, of course, be overgeneralized to all Venezuelan women. These are just a few examples of how women may experience culture shock in a way different from men. Most teachers, foreign student advisers, and administrators could offer many other examples of the needs which women bring to an ESL situation. Sharing this information with colleagues is an excellent way to increase awareness and to begin to evaluate the extent to which our programs are appropriate to the needs of women. The second major phase in the evaluation is a center-wide examination of the appropriateness of the program for women in the areas of personnel, student services, and curriculum. Forward Looking Strategies (United Nations Commission on the Status of Women, 1984), which provides clear guidelines for an evaluation of educational programs such as those offered by ESL institutions, summarizes the nature of the examination to be conducted: Curricula of public and private schools should be examined, textbooks and other educational materials reviewed, and educational personnel retrained to eliminate all discriminatory gender stereotyping in education. Educational institutions should be encouraged to expand their curricula to include studies on womens contributions to all aspects of development. (p. 24) The first questions we must ask ourselves, then, are, How sensitive to womens concerns is the staff of our institution? To what extent are we free from discriminatory gender stereotyping, and to what degree is retraining appropriate? If the initial workshop has accomplished its purpose, most staff should be at least open to discussing ways to inform themselves further about the needs of women students and about equality for women in educational settings. Otherwise, as the Commissions report suggests, some type of retraining, perhaps in the form of professional development, may be appropriate. Second, we must ask ourselves, Are student services designed with the needs of women in mind? Are advisers aware of the issues 350 TESOL QUARTERLY

of freedom and independence which may affect womens cultural adjustment? Are college counselors skilled in assisting women to focus on educational goals appropriate for their needs? Are orientation talks geared toward the needs of women? For example, are women informed that they should walk assertively at night, that they should not walk with their heads down, and that they should avoid eye contact with any stranger who might be bothering them? These are personal safety issues which represent, at a very specific level, the types of student services which are appropriate for women entering a new culture. Perhaps the most crucial aspect of student services concerns financial aid. Data collected during the Decade for Women have revealed that women earn approximately 10% of the worlds income and own 1% of the worlds property.1 Because of this unequal distribution of resources, more women than men need financial assistance. Recognizing this inequity, the Commission recommends that efforts be made to ensure that available scholarships and other forms of support from governmental, non-governmental and private sources are expanded and equitably distributed (United Nations Commission on the Status of Women, 1984, p, 24). To whatever degree possible, we should see that scholarship money is provided for women as well as men in our ESL institutions, colleges, and universities. Administrators in particular may be in a position to advocate the need for scholarship moneys for women. For example, at the suggestion of the administrative staff of our English Center, the Board of Trustees agreed to provide a number of tuition scholarships for women. These scholarships were granted as part of the celebration of the United Nations Decade for Women. Grants and donations are an additional means of providing tuition scholarships to refugee women in our professional womens program. Curriculum is the third area of the center-wide examination, and here we must ask, Is our curriculum appropriate for women? The Commissions report also makes clear suggestions in this area: Research activities should be promoted to identify discriminatory practices in education and training and to ensure educational equality. . . . Governments and private institutions are urged to include in the curricula of all schools, colleges and universities courses and seminars on womens history and roles in society. . . . New teaching methods should be encouraged . . . to clearly demonstrate the equality of the sexes.
1 Statistics

provided by the International Wages for Housework Campaign and International Black Women for Wages for Housework, Kings Cross Womens Center, 71 Tonbridge Street, London WC1, England.



Programmed, curricula and standards of education and training should be the same for females and males. Textbooks and other teaching materials should be evaluated and where necessary rewritten to ensure that they reflect positive, dynamic, participatory images of women. (United Nations Commission on the Status of Women, 1984, p. 16) . It is not difficult to apply these guidelines to our ESL courses. For example, are women taught how to write self-recommendations for college applications? I first learned of this need while reading over the college applications of several Japanese women who were writing more negative than positive comments about themselves. Women students in particular may need help with writing recommendations appropriate for admission to U.S. schools. The vocabulary and tone appropriate for self-recommendations can easily be taught in writing classes, and this is something from which both male and female students would benefit. Textbooks and materials also need to be examined. Studies of ESL textbooks reveal that ESL materials, in general, need to be revised to eliminate sexism. Hartman and Judds (1978) study of ESL materials found that they reflected sexist attitudes and values. Porrecas (1984) study showed similar results13 of 15 commonly used ESL textbooks were found to be sexist. Fortunately, some ESL authors have seen the need to produce texts which are free of sex bias. Porreca (1984) found that Byrd and Clemente-Cabetass React/interact (1980) and Azars Understanding and Using English Grarmmar (1981) are far superior to other texts in avoiding sexist usage (p. 719). It is important that textbook writers follow the lead of these authors in producing materials which are free of sex stereotyping, negative images of women, and sexist language. There are many other ways in which a curriculum can be geared toward meeting the needs of women. For example, fertility and contraception information, including vocabulary items regarding body parts, is important to both men and women, but especially to women. This vocabulary, which is essential for a newcomer, can be discussed in class or offered in an optional workshop. When discussing important people in various fields in an oral skills class, instructors should be careful to balance male and female role models. Architect Julia Morgan can be discussed as well as Frank Lloyd Wright. Painter Georgia OKeeffe can be mentioned along with Gauguin or Czanne. Our teaching methods, too, must be examined in light of the Commissions mandate for equity. When we invite open discussion in our oral skills classes, is it structured in a way to encourage womens participation? Or do we allow men to 3.52 TESOL QUARTERLY

dominate our classes and thereby reduce the opportunities for education for women? These are just a few ways to make ESL programs appropriate for women. Many others would no doubt surface in an individual schools center-wide examination. The third and final aspect of an institutional evaluation involves keeping current on new issues. It is simply not enough to make changes in 1986 and then wait until the next United Nations conference in the year 2000 before we make further adjustments to our programs. ESL teachers, administrators, researchers, and staff need to find ways to evaluate our programs continually to make them equally helpful for men and women. Readings from professional journals, workshops, discussions at staff meetings, and outside speakers can all be used to help keep the needs of women students in mind. As a profession, we need to take the time and spend the money necessary to balance our ESL programs so that they provide equity for women. To borrow the message of the placard carried at Nairobi, if our institutions are not appropriate for both the women and men that we purport to serve, they simply are not appropriate at all.

Azar, B. (1981). Understanding and using English grammar. Englewood Cliffs, NJ: Prentice-Hall. Byrd, D.R.H., & Clemente-Cabetas, I. (1980). React/interact: Situations for communication. New York: Regents. Hartman, P. L., & Judd, E.L. (1978). Sexism and TESOL materials. TESOL Quarterly, 12, 383-393. International Womens Tribune Center. (1985a, November). Facts and figures from the Nairobi meetings. Decade for Women Post Date, p. 2. International Womens Tribune Center. (1985b, November). Photograph. Decade for Women Post Date, p. 1. Porreca, K. (1984). Sexism in current ESL textbooks. TESOL Quarterly, 18, 705-724. Shimazu, M. (1984). Japanese students in EFL/ESL classrooms. TESOL Newsbtter, 18 (2), 19. TESOL Newsletter staff. (1984). Foreign student influx into U.S. reaches plateau. TESOL Newsletter, 18 (1), 4-5.



United Nations. (1975). World plan of action for the implementation of the objectives of the International Womens Year (Report of the United Nations World Conference of the International Womens Year, Sales No. E. 76. IV.1). New York: Author. United Nations Commission on the Status of Women. (1984). Forward looking strategies of implementation for the advancement of women and concrete measures to overcome obstacles to the achievement of the goals and objectives of the United Nations Decade for Women (Report of the Secretary General, A/Conf. 116/PC/21). Vienna: Author.

Comments on Bernard A. Mohan and Winnie Au-Yeung Lo's Academic Writing and Chinese Students: Transfer and Developmental Factors
A Reader Reacts. . .
New York City Technical College, CUNY

No one would dispute the assertion by Bernard A. Mohan and Winnie Au-Yeung Lo in their recent TESOL Quarterly article (Vol. 19, No. 3, September 1985) that a variety of factors may be responsible for the rhetorical problems of all nonnative students including native speakers of Chinese. I believe, however, that the authors too readily dismiss traditional conventions of writing in Chinese as a significant source of interference in the expository prose of Chinese ESL college students. Certainly, developmental factors play an important role in the acquisition of expository writing skill; in addition, different methods and emphases of English composition instruction undoubtedly have a significant impact on students mastery of discourse organization in English. But to state rather flatly, as the authors do, that the ability of Chinese students to organize essays in English is not related to a preference for indirectness in the language and culture of Chinese appears at once to reflect too narrow an understanding of the culture-specific style of Chinese writers in English and to be too broad a characterization of Chinese literary and linguistic elements. Without contradicting the findings of Mohan and Los research, I must nevertheless respectfully submit that they are not the whole 354 TESOL QUARTERLY

story and that more integration of Chinese cultural values and rhetorical patterns exists than they are willing to credit. As I do not read Chinese and as my experience in China is limited to a year of teaching literature and advising students writing a research paper in the Peoples Republic of China (P. R.C. ), I hope my remarks will not appear presumptuous. But my extensive experience with Chinese students and the Chinese community in New York, plus my wide reading in the growing field of contrastive rhetoric, does give me some confidence in my views. I agree with Mohan and Lo that Kaplans attribution of contemporary Chinese rhetorical style to the influence of the traditional Eight-Legged Essay is probably mistaken, or at least exaggerated. We need not turn to any such single source for the conventions of expository style favored by the contemporary Chinese academy. Rather, just as we may properly view Western rhetorical patterns as expressions of Western culture, applicable within the context of Western cultural values (Matalene, 1985, p. 789), so too may we profitably view Chinese conventions of written discourse in the context of a very different constellation of values. This is particularly true when a critical posture, whether in regard to literature, history, art, or social behavior, is central to the expository paper. The American critical posture reflects a core of American values: respect for individual autonomy, inventiveness, forthrightness and action, and regard for the individual personality. The American education system emphasizes the individual students search for truth through experimentation, experience, and reading stimulated by personal interest. Americans have a short-range temporal perspective, and authorities of the past are viewed as guides rather than models. No theory is seen as incontrovertible, and all theories, regardless of the high eminence from which they issue, must be tested in the marketplace of ideas. The acceptable expository style in American academic prose is congruent with these values. Prolixity, abstractness, extensive use of figurative expressions, traditional phraseology, and hortatory language are avoided. Clarity of thesis statement, originality of ideas, logical supporting details, neutral academic language, and abundant use of transitional expressions are the hallmarks of the expository paragraph and essay in English. In marked contrast to these values and their resultant expository style are the dominant social patterns and modes of discourse organization in Chinese society (Hsu, 1963). The Chinese value system subordinates the individuals expressive needs to the welfare THE FORUM


of the community, whatever the community happens to be. The traditions of the past are viewed as the basis for contemporary thought and action (even if the models have changed somewhat), and the pronouncements of recognized authorities are not lightly contradicted. The pedagogical system in China is still solidly based on familiarity with and ability to contextualize, explicate, and imitate accepted authorities of the past (see Tsao, 1983, p. 109; Yang, 1983, p. 103). Truth is not an abstract concept inviting analysis by various individuals following their own bents, but a principle rooted in the sociopolitical environment and articulated by approved authorities (Gamberg, 1977). The concrete, idiographic nature of the Chinese language, the necessity to memorize thousands of characters to become literate in Chinese, and the reverence for calligraphy, an art that embraces writing and composing in one inextricable motion, are three additional aspects of Chinese society that must be factored into the Chinese mode of discourse organization (Houghton& Hoey, 1983). In my understanding and experience, these linguistic and cultural elements have indeed shaped a culture-specific expository style, one characterized by what I call, nonpejoratively, imitative, inculcative, and indirect expressive elements. These include the following: 1. An approach to a topic which is descriptive and syncretistic rather than individually innovative and thesis oriented (Scarcella, 1984) 2. Frequent recourse to the pronouncements of authorities, either by continual quotation, extensive paraphrase, and/or unacknowledged reproduction of key thought units 3. A flatly assertive, judgmental tone 4. Inductive, loosely developed topic and subtopic elements, with paragraphs lacking clear topic statements and the relationships between ideas not explicitly signaled (Tsao, 1983, pp. 102-103) 5. Recurrent use of indirect expressive modes such as rhetorical question, metaphor and simile, formulaic phrasing, analogy, and illustrative anecdote (Brooks, 1968; Davis, 1968) It is of utmost importance to recognize these divergencies between Chinese and American expository styles if we are to train our Chinese EFL/ESL students adequately for the advanced levels of expository writing needed for academic and career achievement in the United States and in the international community which uses English as its medium of communication. This is particularly true in 356 TESOL QUARTERLY

regard to the meticulousness with which Americans document their sources, as opposed to the Chinese disregard for such documentation. Our notion of plagiarism is by no means a universal one which can be taken for granted (Gregg & Pacheco, 1981). Instead, it should be thoroughly explored with Chinese student writers on a contrastive basis, rather than from a stance of cultural superiority. Mohan and Los conclusions that discourse organization develops late even in the first language acquisition process and that the mastery of this skill can be influenced by appropriate teaching practices are by no means at odds with my argument. If different cultural patterns do shape different rhetorical patterns, as I believe they do, what could be more logical than providing the nonnative student with explicit instruction in the rhetorical norms of the second language? Given either instrumental or integrative motivation, Chinese students are certainly capable of assimilating new conventions of expository prose, especially if notions of inferiority or superiority of culture-specific rhetorical patterns are avoided. My Chinese students were never persuaded that the American expository style was better than the Chinese. They continued to perceive our emphasis on original thesis development as leading to sometimes bizarre personal judgments. They did not share our disparagement of traditional formulaic languagethe essential furniture of the learned persons writing to them, the abhorred clich to us. They viewed the American near-obsession with clarity and explicitness as inappropriate for the supposedly educated readership of academia. But these students learned to abide by our conventions of expository prose with varying degrees of success, and many, both in my P.R.C. and New York classrooms, were in fact able to strike an interesting balance between the two styles. Obliquely expressive language was used, but sparingly; major subtropical paragraphs were models of deductive clarity and coherence, with inverted and elliptical paragraphs providing variation; rhetorical questions were not wholly eschewed but were used in moderation; and the hortatory tone was subdued and reserved for conclusions. By increasing our own awareness of our students culture-specific writing styles, we lay the groundwork for encouraging them to experiment with new expository patterns in a meaningful and ultimately rewarding way.



Brooks, B.R. (1968). Geometry of the Shr Pin. In T.-T. Chow (Ed.), WenLin: Studies in the Chinese humanities (pp. 121-151). Madison: University of Wisconsin Press. Davis, A.R. (1968). Double ninth festival in Chinese poetry. In T.-T. Chow (Ed.), Wen-Lin: Studies in the Chinese humanities (pp. 45-65). Madison: University of Wisconsin Press. Gamberg, R. (1977). Red and expert: Education in the Peoples Republic of China. New York: Schocken Books. Gregg, J., & Pacheco, B. (1981, November). Research skills development for students in college career programs. Paper presented at the College Learning Skills Across the Curriculum, Grossingers, NY. Houghton, D., & Hoey, M. (1983). Linguistics and written discourse: Contrastive rhetorics. In R.B. Kaplan (Ed.), Annual review of applied lingustics: 1982 (pp. 2-18). Rowley, MA: Newbury House. Hsu, F.L.K. (1963). Clan, caste and club. Princeton, NJ: Van NostrandReinhold. Matalene, C. (1985). Contrastive rhetoric: An American writing teacher in China. College English, 47, 789-808. Scarcella, R. (1984). How writers orient their readers, in expository essays: A comparative study of native and non-native English writers. TESOL Quarterly, 19, 671-688. Tsao, F. (1983). Linguistics and written discourse in particular languages: Contrastive studies: English and Chinese (Mandarin). In R.B. Kaplan (Ed.), Annual review of applied linguistics: 1982 (pp. 99-117). Rowley, MA: Newbury House. Yang, Z. (1983). The mirror and the jigsaw: A major difference between current Chinese and Western critical attitudes. Representations, 4, 101107.

Response to Gregg. . . On Evidence for Cross-Cultural Rhetoric

University of British Columbia

I welcome Joan Greggs comments on the article written by Winnie Au-Yeung Lo and me. The topics of cross-cultural rhetoric and composition in the second language are important ones for several reasons. By exploring discourse variation across cultures and across bilingual competencies, we can not only increase our awareness of student difficulties with composition in the second

language, but also gain insight into composition processes in general and reach a deeper understanding of the nature of academic discourse. Debate on these issues can be most productive, and in this spirit I will state some of our points of agreement and disagreement with Greggs response. In brief, Gregg claims that American academic expository prose is characterized by clear thesis statements, original ideas, logical supporting details, neutral language, and explicit transitions and that Chinese academic prose diverges from these features. She implies that Chinese students writing in English will display these aspects of Chinese prose. In its general approach, her position is similar to the views of Kaplan. These are interesting thoughts, suggesting as they do certain links between academic prose and value systems, and they outline promising directions for future research. However plausible Greggs suggestions are, though, we are given no evidence to support her claims about American prose, Chinese prose, or transfer effects on Chinese students writing in English. Nor does she indicate how she would go about collecting evidence for these claims. This failure to provide evidence is a remarkable omission, and it may account for several ways she has misunderstood our position. For instance, she states that we dismiss traditional Chinese conventions as a source of interference in second language composition. This is simply incorrect. We say specifically that transfer factors should be included in a general model of second language composition. What we insist on, however, is that claims about transfer should be based on well-established cross-cultural differences in rhetoric. They should be based on evidence. But providing the evidence is not as simple as it might appear. Let us see why. A number of second language researchers and composition teachers I have spoken with reason in something like the following standard pattern: We know that there are cross-cultural differences in rhetoric; we know that these differences will cause second language students to make errors in composition which deviate from the norms of English; our students make such errors; therefore, there is negative transfer from the rhetoric of the first language. There are two notable empirical problems with this reasoning. First, cross-cultural differences in rhetoric are not well established but need to be demonstrated. When Kaplan (1966) first suggested cross-cultural differences in paragraph patterns in five different THE FORUM 359

language groups, he was suitably tentative, saying that much more detailed and accurate descriptions are required before any meaningful system can be elaborated (p. 15). In a friendly critique of Kaplan, Hinds (1983) points out that in order to discover foreign language rhetorical patterns, it is necessary to examine compositions in the foreign language; compositions written for an audience which reads that language (p. 186). Accordingly, Hinds analyzes texts written in Japanese for Japanese readers. In our article we compared Kaplans claim that the Chinese paragraph was indirect with two kinds of data: evidence from traditional Chinese writers and statements by specialists in Chinese composition about exposition in Chinese. Research on contrastive rhetoric is not identical with research on composition in the second language. To establish her claims about Chinese rhetoric, Gregg needs to consider evidence from Chinese texts written for Chinese readers. The second problem with this reasoning is that errors in second language composition are not necessarily transfer errors. Consider Greggs characteristics of American prose. Dont American student writers occasionally fail to produce clear thesis statements, original ideas, logical supporting details, neutral language, and explicit transitions? When American students make these errors, they are obviously not the result of transfer from another language. They are more likely to be developmental errors. So when Chinese students writing in English make such errors, we cannot automatically assume that they are due to transfer. The analysis of composition errors is problematic. In some cases in syntax we can identify errors sufficiently well as transfer or developmental to be able to produce strong findings. This is because we can draw on contrastive studies of syntax and on studies of syntactic development. In error analysis in expository discourse we are not yet in such a happy state. The data are typically ambiguous. Hence the importance of using other kinds of evidence in addition. For the group of Chinese students we studied, we were able to show that a comparison of composition practices in Hong Kong and British Columbia indicated that their school experience with English composition was oriented more toward accuracy at the sentence level than toward the development of appropriate discourse organization. Our evidence for this developmental factor was based on teacher and student interviews and surveys, the analysis of composition texts, and facts about the evaluation of composition. We also drew on composition process research. 360 TESOL QUARTERLY

The weakness of the standard pattern of reasoning about the transfer of rhetorical norms, then, is that it proves nothing. It does not prove that contrastive rhetorical differences exist because it does not examine composition in the first language. And it does not prove that transfer errors occur because other explanations of error are available. Assuming what it should try to prove, it argues in a circle. Because it is circular, it merely confirms old dogma and does not lead to new knowledge. This may in part explain why so little progress has been made in this important research area in the 20 years since Kaplan first wrote. Our article did not aim to replace a transfer model of second language composition with a developmental model. To assume so would be completely to misunderstand our general thrust. Our aim was to point a way to more productive work in this area. We suggested that a general model of second language composition should include both developmental and transfer factors, and we emphasized sources of evidence in addition to the analysis of student errors. By broadening our base of theory and research, we can avoid unproductive circularity. We need studies of expository discourse in English and other languages to establish cross-cultural differences. We need studies of the learning and teaching of composition in the first language to establish developmental patterns. Beyond studies of student errors in the writing product, we need studies of the writing process. We need to become more aware of students literacy in their native language and to gather information on their educational experiences with literacy both in the home country and the host country. The nature of academic discourse (or the language of teaching and learning, or of exposition, or of cognitive/academic language proficiencythe labels point to a common issue) is a question of enormous importance throughout education. By using more than one theoretical perspective, more than one language, and more than one form of data, we have the theoretical and empirical variety that leads to a deeper understanding of this vital topic.

Hinds, J. (1983). Contrastive rhetoric: Japanese and English. Text, 3, 183196. Kaplan, R. (1966). Cultural thought patterns in inter-cultural education. Language Learning, 16, 1-20.