Annotated Bibliography for 5 articels

Teachers of English to Speakers of Other Languages, Inc. (TESOLyf The Influence of Teachers and Peers on Second Language Acquisition in Bilingual Preschool Programs Author(syf 5 D \ & K H V W H U I L H O G . D W K H U L Q H + D \ H V / D W L P H U . D W K O H H Q % D U U R Z V & K H V W H U I L H O G D Q d Regino Chávez Source: TESOL Quarterly, Vol. 17, No. 3 (Sep., 1983yf S S 9 Published by: Teachers of English to Speakers of Other Languages, Inc. (TESOLyf Stable URL: http://www.jstor.org/stable/3586255 Accessed: 06-02-2017 01:31 UTC JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected]. Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://about.jstor.org/terms Teachers of English to Speakers of Other Languages, Inc. (TESOLyf is collaborating with JSTOR to digitize, preserve and extend access to TESOL Quarterly This content downloaded from 137.148.11.31 on Mon, 06 Feb 2017 01:31:20 UTC All use subject to http://about.jstor.org/terms TESOL QUARTERLY, Vol. 17, No. 3, September 1983 The Influence of Teachers and Peers on Second Language Acquisition in Bilingual Preschool Programs RAY CHESTERFIELD ]ubrez and Associates KATHLEEN BARROWS CHESTERFIELD Jubrez and Associates KATHERINE HAYES-LATIMER University of California, Los Angeles REGINO CHAVEZ Juiirez and Associates The interactional patterns of eleven Spanish-speaking preschool children enrolled in two different bilingual programs were observed over the course of one year. The programs were characterized by a majority of Spanish-speaking children in the classrooms at one site and of English-speaking children in the classrooms at the second pre- school. Consistent with the linguistic composition of the classes, teaching staffs were generally found to use a predominance of Spanish or English respectively in their classrooms. For all classrooms, language choices of individual children were examined. In classrooms where English-speaking children predominated, greater language proficiency was found to be more consistently related to the use of English with peers than with teachers. In classrooms where Spanish- speaking students were prevalent, on the other hand, interactions with the teachers in English were more consistently related to greater English language proficiency. Implications for the teaching of young children in a dual language environment are discussed. The appropriate use of two languages by the teacher is generally considered to be a key factor in the implementation of bilingual education programs. This is especially true when the programs are designed for young children who are still in the process of acquiring both a first and a second language. It is often assumed that teachers in bilingual programs provide most of the input received by individual learners in their second language and that, generally, teachers' language use in the classroom will be relatively balanced. Such assumptions, however, ignore both the interactional choices made by the learners themselves and the linguistic composition of the students making up a given classroom. This study examines the experiences of Spanish-speaking children INFLUENCE OF TEACHERS AND PEERS ON SLA 401 This content downloaded from 137.148.11.31 on Mon, 06 Feb 2017 01:31:20 UTC All use subject to http://about.jstor.org/terms learning English in two preschools with markedly different ratios of Spanish- and English-speaking students. It is argued that the linguistic composition of the class may affect the overall language environment of the classroom as well as the relative importance of the interactions with teachers or with peers in the acquisition of English. The few studies undertaken on the behavior of teachers in bilingual classrooms have generally focused on aspects of the teachers' language use. Studies such as those of Hughes and Harrison (1971yf D Q G W K e United States Commission on Civil Rights (1973yf K D Y H G H D O W U H V S H F - tively with how much pupil talk in Spanish and English different teachers permitted and the disparities in teachers' use of positive responses in dealing with children of different ethnic backgrounds. Others (Townsend 1974, Zamora 1974yf K D Y H F R P S D U H G W K H Y H U E D O D Q d nonverbal interaction patterns of bilingual preschool teachers to those of assistant teachers. Similarly, Legarreta (1977, 1979yf H [ D P L Q H G W K e differential use of Spanish and English by teachers implementing different approaches to bilingual education. Although all of these studies suggest a relationship between the type of program and children's language proficiency, none of these studies has explored the extent to which children's interactions with peers (as compared to those with teachersyf D U H U H O D W H G W R L Q F U H D V H G O D Q J X D J H S U R I L F L H Q F \ . Researchers using observational techniques have begun to study social interaction in bilingual classrooms. Shultz (1975yf 3 K L O O L S V \f, and Genishi (1976yf K D Y H H [ D P L Q H G W K H S K H Q R P H Q R Q R I O D Q J X D J H X V H L n teacher-child interactions. These studies have found that bilingual teachers conducted formal instruction about equally in both Spanish and English; however, as a result of their use of English in non- instructional contexts, this language predominated in the classroom. Moll (1981yf I R X Q G W K D W W K H Y H U E D O E H K D Y L R U R I D V L Q J O H J U R X S R f bilingual children who participated in reading lessons in Spanish and English in different classrooms was highly conditioned by their teachers' presuppositions about the children's competence. These presuppositions affected the teachers' subsequent organization of the lessons, despite the fact that the children were in the high ability group in both classrooms. While these studies have provided valuable insights into classroom interaction in structured lessons with specific children, they tell us little about the general language use of teachers in classrooms with different student populations or of the effect of interactions with teachers, as compared with peers, on the language acquisition of individual students. The effects of naturally occurring social interaction on children's acquisition of a second language have been dealt with at least peripherally in a few studies. Bruck and Shultz (1977yf I R X Q G W K D W W K e rules for peer-peer language use by Spanish-speaking children in a 402 TESOL QUARTERLY This content downloaded from 137.148.11.31 on Mon, 06 Feb 2017 01:31:20 UTC All use subject to http://about.jstor.org/terms bilingual classroom differed from those used in a regular classroom. Although the findings of this study indicated that the use of Spanish in both types of classrooms by the two children making up the sample decreased over the year and provided some evidence on the language directed by the teacher to the individual children, the limited time periods in which data were collected did not allow investigation of the effects of such interactions. Fillmore (1976yf L Q H [ D P L Q L Q J W K H F R Q G L - tions which facilitated or inhibited the learning of English among five young Spanish-speaking children, found the social aspects of language learning to be important since social skills enabled the learners to participate in situations where the new language was used. Although the author discussed the role of social abilities in the individual differences exhibited among her sample in learning a second language, the experimental nature of the research, which paired English- and Spanish-speaking subjects, did not lend itself to a discussion of the effects of peer interactions in which speech partners are freely chosen, as in the naturally occurring activities of a bilingual classroom. The same author (Fillmore 1980yf K D V H [ D P L Q H G W K H U H O D W L Y H X V H R f English by teachers and students in a monolingual English and a bilingual classroom in which Chinese students predominated. She found that English was generally used in both classrooms and that her subject child in each classroom was acquiring English. Although the author suggested that a balance of English and non-English-speaking (NESyf F K L O G U H Q L Q W K H F O D V V U R R P Z D V Q H F H V V D U \ I R U W K H Q R Q Q D W L Y e speakers of English to learn to speak the second language, the predominance of Chinese speakers in both classrooms under observa- tion did not allow her to examine the influence of different ratios of students on individual children's performance. In a subsequent study, Fillmore (1982yf D U J X H V W K D W L Q V L W X D W L R Q s where teachers organize classroom activities to allow NES children to interact with their peers, English-speaking children, where available, can serve as a resource in their classmates' second language acquisition. Although she suggests that the linguistic composition of the class may influence the role of teachers and peers in providing input in English, she does not empirically examine the interactions of individual learn- ers. Our own findings (Barrows Chesterfield et al. 1982yf R I F R Q V L V W H Q W O y high correlations between the use of English with peers and increased English language proficiency among Spanish-speaking preschoolers lend support to Fillmore's contention about the value of peer interac- tion in learning a second language. As our study was limited to classrooms in which English-speaking children predominated, how- ever, it told us nothing of the effect of the linguistic composition of the class on the importance of interacting with peers. The above studies show that social interaction is an important factor INFLUENCE OF TEACHERS AND PEERS ON SLA 403 This content downloaded from 137.148.11.31 on Mon, 06 Feb 2017 01:31:20 UTC All use subject to http://about.jstor.org/terms in the learning of a second language. None, however, provides a systematic investigation of the nature and effect of individual contact in different classroom settings. In the present study, we investigate the linguistic composition of bilingual preschool classes as it influences the language use of teachers and the verbal interactions of individual students. In addition, the relationship between the conversational partners of a sample of children in classes of different linguistic composition and the increased second language proficiency of those students is explored. METHODOLOGY The classroom social interaction of eleven Spanish-preferring' chil- dren (eight males and three femalesyf U H I H U U H G W R W K U R X J K R X W W K H W H [ W D s target children, all between the ages of 42 and 50 months, was investigated. Six of the children were enrolled in a preschool in Corpus Christi, Texas; the remaining five attended a preschool in Milwaukee, Wisconsin. At the time of the study, both of the schools served lower income families and each was implementing a similar bilingual curricu- lum. The principal data collection strategy was that of participant obser- vation, in which researchers trained in the techniques of naturalistic inquiry in the classroom provided written fieldnotes on the verbal and nonverbal behavior of individual children and audiotaped samples of speech throughout a school year. Focused observations were made of the children at each site three times during the school year-Novem- ber-December,2 February-March, and April-May. Children were ob- served on randomly selected days for specific amounts of time until the total amount of observation time approximated that of a normal preschool day. Data collection combined the strategies of time and event sampling; specific contexts (mealtime, independent play, transition, large or small groupyf Z H U H U D Q G R P O \ V D P S O H G D Q G H D F K F K L O G Z D V R E V H U Y H d for amounts of time proportional to the percentage of time devoted to 1A child's preferred language is defined as that which the child uses most often with teachers and peers in a variety of classroom situations. This was determined for all children in the classroom through a rating form used over a two-week period at the beginning of the school year prior to the focused observations of the sample children. The concept of language preference, rather than the traditional language dominance classifications of Spanish/English monolingual, Spanish/English dominant, and balanced bilingual, is used here as young children in a bilingual setting are often in the process of acquiring two languages and may call upon either depending on the situation, the context, and the actors involved. The term preference is not intended to suggest a conscious language choice on the part of the children. 2 Owing to a series of fieldtrips and vacations related to the holiday season, the total observation time was somewhat shorter during this observation period than during subsequent ones. Thus, relative frequencies are used to determine language use patterns. 404 TESOL QUARTERLY This content downloaded from 137.148.11.31 on Mon, 06 Feb 2017 01:31:20 UTC All use subject to http://about.jstor.org/terms a particular activity in the school day. For example, if large group activities took place for 30 minutes and small group activities for 50, the researcher would observe a child in those activities for time samples of three and five minutes, respectively, until the daily total was reached. The researcher noted the time at which an observation began and then proceeded to describe the behaviors of the designated child and his or her verbal interactions with others. Note was made of any transitions in activities occurring during the observation period and of the time of such transitions. After each day's observations, the observer rewrote the rough notes taken in the classroom and coded them. Given the focus of the study on the relationship of language proficiency to the target children's productive abilities and language use, only utterances of the target children were coded. All utterances, regardless of whether or not a target child initiated an interaction, were coded for type of interaction (child with peer(syf F K L O G Z L W K W H D F K H U F K L O G Z L W K K L P K H U V H O I D Q d ambiguous-to no one in particular, or with a group of mixed language preferenceyf O D Q J X D J H X V H G E \ W K H W D U J H W F K L O G 6 S D Q L V K ( Q J O L V K , language mixingyf F R Q W H [ W P H D O W L P H L Q G H S H Q G H Q W S O D \ W U D Q V L W L R Q , large or small groupyf D Q G W K H O D Q J X D J H S U H I H U H Q F H R I W K H Q R Q W D U J H W " child or addressee interacting with the target child. An utterance was defined as a statement containing a complete message or thought, the boundary of which was indicated by speakers' pauses. Single words uttered in response to questions, including yes and no, were regarded as utterances since the context of the discourse provided the complete message. In the case of language switching by the target children, the language in which the majority of words in the utterance appeared was considered to be the chosen language.3 In addition, all of the utterances of the teaching staff, made either to the target child or to other children in the room, were coded for language use (Spanish, English, or language mixingyf . Consistency in data collection and coding was maintained through a number of quality control procedures. Fieldworkers underwent a briefing session prior to each phase of fieldwork. These sessions provided a review of field procedures and notetaking and included coding exercises to insure consistency among fieldworkers. Research- ers were also provided with a field manual which complemented the briefing sessions by presenting guidelines for carrying out the field research and operational definitions of the phenomena to be studied. Finally, the supervisor of fieldworkers conducted parallel observations with the researchers during the fall and spring observation periods. These observations were coded individually for verbal and nonverbal 3 In 100yb R I W K H F D V H V W K L V P D W F K H G W K H O D Q J X D J H S U H I H U H Q F H R I W K H W D U J H W F K L O G . INFLUENCE OF TEACHERS AND PEERS ON SLA 405 This content downloaded from 137.148.11.31 on Mon, 06 Feb 2017 01:31:20 UTC All use subject to http://about.jstor.org/terms behavior and an acceptable level of agreement (above .85yf Z D V I R X Q d at both sites. Using the codes, the language use of teachers in classrooms with linguistically different compositions of students was characterized for each classroom, and for each classroom type, the children's interaction patterns with teachers and peers were compared with the children's growth in English language proficiency. To investigate the influence of class composition on teacher language use, the relative frequency with which the classroom staff used English and Spanish was computed for each of the classrooms in which the sample children were located. In order to determine differing degrees of interaction with adults and peers, the relative frequency of each child's verbal interactions in Spanish and English was calculated for each of the three time periods. Language proficiency was assessed through use of the children's average mean length of utterance (MLUyf L Q ( Q J O L V K 5 D Q G R P V D P S O H s of the children's language production in each classroom context during the first and third data collection periods were used to calculate the MLU.5 Increase in language proficiency was determined by calculating the change in average MLU from the first to third observation periods. Rank order correlations were computed to examine the relationship between peer and adult interaction and the relative increase in English language proficiency of the children over the year. THE RESEARCH SETTINGS The sample children at Corpus Christi were enrolled in three bilingual classrooms. Each was staffed by a teacher and an aide, all but one of whom were bilingual. The Milwaukee sample was found in two bilingual classrooms, each of which had a bilingual teacher and two bilingual aides. Classrooms at the two sites were generally similar in that at each the daily schedule included two meal periods, small group language sessions, large group activities, independent play/work activi- ties, and outside play. At both sites, the structured activities of large and small groups were designed to encourage the use of both Spanish and English. Activities at both sites generally took place in one of a number of learning centers or areas set up around the rooms. The 4 Criteria for calculating MLU were those of Brown (1973yf 0 / 8 Z D V F K R V H Q D V W K H P H D V X U H R f language proficiency as it permits an estimate of the syntactic diversity of individuals, such as the children in this study, who are at the earliest stages of language learning. 6 As is generally the case in studies of naturally occurring classroom discourse, our audiotaped samples were limited to structured small group activities. Consistent with various research studies (Cole et al. 1978, Wells 1979yf Z K L F K K D Y H V K R Z Q W K D W W K H 0 / 8 R I \ R X Q J F K L O G U H Q D Q d the functions served by their speech vary with the context in which the discourse takes place, most of the taped samples were short answers to specific questions. Thus, given the focus of the study on a variety of classroom contexts and the high inter-observer agreement found during the parallel observations at both sites, data recorded by hand were considered the most appropriate means of estimating MLU. 406 TESOL QUARTERLY This content downloaded from 137.148.11.31 on Mon, 06 Feb 2017 01:31:20 UTC All use subject to http://about.jstor.org/terms major difference in the sites was that small group activities at the Texas site emphasized the separation of Spanish and English whereas those in Milwaukee encouraged concurrent use of both languages. The linguistic composition of the classes at each site differed considerably. The three classrooms in Texas (designated A, B, and Cyf had populations of four, four, and five Spanish-preferring children and ten, eight, and seven English-preferring children respectively. Class- room I at the Milwaukee site had ten Spanish-preferring children and seven English-preferring children, while Classroom II contained nine Spanish- and seven English-preferring children. RESULTS As illustrated in Table 1, the linguistic composition of the class greatly influenced the overall language use by the classroom teaching units. English was the predominant language in all of the preschool classrooms at Corpus Christi. This language was used in at least 53 percent of a teaching team's utterances during each observation period and reached a high of 78 percent of all utterances during the third observation period in Classroom C. The relatively greater use of English in Classroom C at the second and third observations was a result of a monolingual English-speaking teacher replacing the original bilingual teacher at midyear. In Milwaukee, Spanish usage ranged from 45 percent to 61 percent of the utterances in Classroom I, which had the highest ratio of Spanish-preferring to English-preferring children. In Classroom II, on the other hand, the teaching team used more English than Spanish at each observation period. The extent of their English language usage, however, remained substantially below that at Corpus Christi. The relatively higher percentages of language mixing found at the Milwaukee site were a reflection of the emphasis given to concurrent translation and language switching in the small group language sessions at that site. It might be argued that the greater use of English at the Corpus Christi site was indicative of the greater bilingual abilities of the Spanish-preferring children at that site rather than simply the composi- tion of students in each classroom. However, the relatively greater use of Spanish in both Milwaukee classrooms at the third observation period, despite the fact that it was found that the children had increased their second language abilities, would seem to negate this argument. An examination of the change over time in the language proficiency of the sample children, shown in Table 2, indicates that at the end of the year four of the five target children at Milwaukee were relatively more proficient in English than the two most proficient children in the Corpus Christi sample had been at the beginning of the INFLUENCE OF TEACHERS AND PEERS ON SLA 407 This content downloaded from 137.148.11.31 on Mon, 06 Feb 2017 01:31:20 UTC All use subject to http://about.jstor.org/terms Cj H t~l ? q c. TABLE 1 Relative Frequency of English, Spanish, and Language Mixing' by Classroom Teaching Staffs Over Time CORPUS CHRISTI Time 1 Timune 2 Time 3 Language Language Language English Spanish Mixing English Spanish Mixing English Spanish Mixing Classroom A 58yb \b 3yb \b 46yb \b 60yb \b 2yb Classroom B 63 37 0 63 30 7 56 40 4 Classroom C 62 33 5 72 25 3 78 20 2 MILWAUKEE Time 1 Time 2 Time 3 Language Language Language English Spanish Mixing English Spanish Mixing English Spanish Mixing Classroom I 41yb \b 6yb \b 45yb \b 33yb \b 6yb Classroom II 43 35 22 52 40 8 48 47 5 'Language mixing indicates switching of languages within a single utterance. This content downloaded from 137.148.11.31 on Mon, 06 Feb 2017 01:31:20 UTC All use subject to http://about.jstor.org/terms z tTI z 0 C z TABLE 2 Change Over Time in English Language Proficiency' of Sample Children CORPUS CHRISTI MILWAUKEE Period I Period III Change from Period I Period III Change from Period I-II Period I-III Ricardo 5.1 +5.1 Javier - 4.2 +4.2 Carolina 2.6 6.2 +3.6 Ramona - 4.0 +4.0 Luis 2.5 4.5 +2.0 Roberto - 3.6 +3.6 Dolores 3.3 4.7 +1.4 Mario - 3.6 +3.6 Jesus 3.3 4.3 +1.0 Mateo - 2.7 +2.7 Jaime 2.6 3.4 +0.8 Indicated by the mean length of utterance (MLUyf . 2 - indicates that the child failed to speak the minimal number of utterances set for calculating MLU. This content downloaded from 137.148.11.31 on Mon, 06 Feb 2017 01:31:20 UTC All use subject to http://about.jstor.org/terms year. Despite such relative proficiency, however, in both classrooms in which the Milwaukee children were found, the relative use of Spanish by teachers increased. In general, the children at both sites exhibited similar patterns of development. Five of the six children at Corpus Christi, as a result of having started the year with some proficiency in English, exhibited greater proficiency in their second language at the end of the year than did the most proficient of the Milwaukee children. The target children in Milwaukee, as might be expected from their almost total lack of English at the beginning of the year, generally showed greater gains in their mean length of utterance than their Texas counterparts. In each group there was considerable variation in the degree of change. At Corpus Christi, two of the children, Ricardo and Carolina, exhibited relatively large increases of at least 3.6 in English MLU, whereas at Milwaukee four of the children showed such a pattern. The gains of Mateo at Milwaukee (2.7yf D Q G / X L V D Q G ' R O R U H V D W & R U S X V & K U L V W L 0 and 1.4 respectivelyyf Z H U H O H V V G U D P D W L F Z K H U H D V - H V X V \f and Jaime (0.8yf P D G H W K H V P D O O H V W J D L Q V $ Q D Q D O \ V L V R I W K H L Q F O D V V O D Q J X D J H X V e patterns of these children serves to explain, at least in part, the variable increases of the sample children. Table 3 presents the relative frequency of language use for each sample child over the preschool year. For most Corpus Christi children, there is an obvious change in the balance of Spanish and English used in the classroom. During the first observation period, the children's preference for Spanish was evident from the fact that, with the exception of one child," at least 60yb R I W K H L U Y H U E D O L Q W H U D F W L R Q V L n the classroom were in that language. By midyear, three of the six children were using a majority of English in the classroom. At year's end all but one child, Jaime, were speaking a predominance of English in the classroom. Consistent with the general language environment of the Milwaukee classrooms, children at that site generally spoke more Spanish at the beginning of the year than did their Corpus Christi counterparts. With the exception of one child, Javier, who made the greatest gains in English proficiency at that site, all of the children continued to use a majority of Spanish at the midyear observation. By the end of the year, however, three of the five children were using a greater percentage of English than Spanish in the classroom. This occurred despite the fact that two of these children, Javier and Ramona, were in Classroom I, where teachers were using primarily Spanish. 6 Dolores, the one child who did not follow this pattern, was characterized as a shy child who had to be drawn out at the beginning of the year. The relatively low percentage of Spanish usage in comparison to her Spanish-preferring peers was probably due to her few utterances-a total of seven-during the first observation period. 410 TESOL QUARTERLY This content downloaded from 137.148.11.31 on Mon, 06 Feb 2017 01:31:20 UTC All use subject to http://about.jstor.org/terms When the verbal interactions of the children are examined by conversational partner, it becomes clear that in spite of the open classroom environment designed to encourage child-child interactions at both sites, the majority of most sample children's speech was with teachers. In 75yb R I W K H L Q G L Y L G X D O R E V H U Y D W L R Q V F K L O G U H Q L Q W H U D F W H d more frequently with the teacher than with peers. The patterns of these interactions, however, varied by site and over time. At Corpus Christi, four of the six children showed a trend away from using relatively large percentages of both languages with teachers at the first observation period to a predominance of English at the third set of observations. The two exceptions to this pattern were the children who showed the smallest gains in English proficiency over the course of the year. At Milwaukee, with the exception of Javier, Spanish predominated in the interactions of all children with the teachers throughout the year. There was, however, a relative increase in the amount of English spoken with the teacher at each point in time for four of the five children, and the children exhibiting the greatest increases in English proficiency were generally those who interacted more frequently with the teacher in that language. Early in the year, teachers were the speech partners for most interaction in English at this site. An examination of the sample children's interaction with peers reveals that the children at both sites generally increased the relative frequency of their interactions in English with peers. With the excep- tions of Jaime, who spoke totally in Spanish even with English- preferring peers throughout the year, and Luis, who interacted primari- ly with Spanish-preferring peers throughout the year, all other children at both sites established a pattern of talking principally in English with peers. The increasing English proficiency of most children in the preschool classrooms was reflected by the fact that two of the children at Corpus Christi and all of those at Milwaukee addressed some of their Spanish-preferring classmates in English late in the year. While those children who exhibited the greatest increased English proficiency at Corpus Christi were also those who interacted most frequently in English with their classmates, this was not always the case at Milwau- kee. For example, Javier, the child who ranked first in increased English proficiency, ranked only fourth among the sample children in the relative frequency of his interactions in English with peers. To determine more precisely the relationship between the frequency of verbal interaction with teachers and peers and growth in English language proficiency of children in classrooms having different linguis- tic compositions, the children at each site were ranked by change in English MLU from the first to the third observation periods. Spearman rank order correlations of this measure of language proficiency and INFLUENCE OF TEACHERS AND PEERS ON SLA 411 This content downloaded from 137.148.11.31 on Mon, 06 Feb 2017 01:31:20 UTC All use subject to http://about.jstor.org/terms rP dr C3 Fl cn O P C 3+ 15 CI trl 35 P cC TABLE 3 Relative Frequency of English and Spanish Use in the Classroom by Sample Children CORPUS CHRISTI Observation Interaction RICARDO CAROLINA LUIS DOLORES JESUS JAIME Period Type Spanish English Spanish English Spanish English Spanish English Spanish English Spanish English PEERS' Spanish Preferring 20yb \b 41yb \b 17yb \b 0yb J \b 0yb \b 0yb English Preferring 0 0 0 6 4 4 0 14 0 0 57 0 I TEACHERS 20 40 35 18 26 22 43 43 61 17 36 7 SELF 0 0 0 0 0 0 0 0 0 0 0 0 AMBIGUOUS' 20 0 0 0 26 0 0 0 0 6 0 0 TOTALS. 60 40 76 24 73 26 43 57 78 23 93 7 PEERS Spanish Preferring 0yb \b 6yb \b 12yb \b 0yb \b 7yb \bg 13yb \bg English Preferring 0 42 0 13 0 4 23 8 4 7 0 0 II TEACHERS 23 27 23 35 28 32 27 19 48 26 60 3 SELF 0 0 0 0 0 0 0 0 4 0 0 0 AMBIGUOUS 0 8 16 6 0 12 4 19 0 0 20 3 TOTALS 23 77 45 54 40 60 54 46 63 37 93 6 PEERS Spanish Preferring 0yb \b 7yb \b 325 5yb \b 0yb \b 0yb \b 0yb g English Preferring 0 53 0 29 0 11 11 19 0 12 8 0 III TEACHERS 7 27 12 33 5 42 23 37 23 12 64 8 SELF 0 0 0 0 0 0 0 0 6 0 0 2 AMBIGUOUS 0 13 7 7 0 5 4 7 0 35 4 0 TOTALS 7 93 26 74 37 63 38 63 41 59 89 10 Indicates speech addressed either to a single child or to a homogeneous peer group of the indicated speech preference. 2 Refers to the child talking to him/herself. 3 Refers to speech directed to no one in particular or to a mixed group of Spanish- and English-preferring children. STotals may not equal 100 due to rounding. This content downloaded from 137.148.11.31 on Mon, 06 Feb 2017 01:31:20 UTC All use subject to http://about.jstor.org/terms C-. 0 C,, C1I C,, 0 zl TABLE 3 (continuedyf MILWAUKEE Observation Interaction JAVIER ROBERTO RAMONA MARIO MATEO Period Type Spanish English Spanish English Spanish English Spanish English Spanish English PEERS' SpanishPreferring 33yb \b 33yb \b 40yb \b 63yb \b 0yb \b English Preferring 0 0 0 7 20 0 0 0 9 0 I TEACHERS 56 11 60 0 40 0 37 0 82 9 SELF~ 0 0 0 0 0 0 0 0 0 0 AMBIGUOUS' 0 0 0 0 0 0 0 0 0 0 TOTALS' 89 11 93 7 100 0 100 0 91 9 PEERS Spanish Preferring 7yb \bY 29yb \b 0yb , \b 18yb \b 3yb \b English Preferring 0 29 0 21 0 0 0 4 0 7 II TEACHERS 28 36 21 15 90 10 52 19 71 12 SELF 0 0 0 0 0 0 0 0 7 0 AMBIGUOUS 0 0 7 0 0 0 0 8 0 0 TOTALS 35 65 57 43 90 10 70 31 81 19 PEERS Spanish Preferring O0 6yb \b 16yb \b 3yb \b 7yb \b 10yb English Preferring 0 21 3 22 0 38 3 25 0 11 II TEACHERS 12 52 31 16 27 20 33 5 51 13 SELF 0 9 0 3 0 0 2 5 3 0 AMBIGUOUS 0 0 0 3 0 0 5 0 5 2 TOTALS 12 88 40 60 37 61 58 42 65 36 Indicates speech addressed either to a single child or to a homogeneous peer group of the indicated speech preference. Refers to the child talking to him/herself. 3 Refers to speech directed to no one in particular or to a mixed group of Spanish- and English-preferring children. 4 Totals may not equal 100 due to rounding. This content downloaded from 137.148.11.31 on Mon, 06 Feb 2017 01:31:20 UTC All use subject to http://about.jstor.org/terms measures of different types of classroom interaction indicative of the children's language choice patterns were then computed. Children at each site were ranked on four measures of classroom interaction with both peers and teachers. All measures represent aspects of the child's English language use with different speech partners. Two of the measures dealt with the child's general use of English with peers and teachers during the year. These measures were 1yf S H U F H Q W D J H R I ( Q J O L V K X V H G H L W K H U Z L W K S H H U V R U W H D F K H U V \f out of the child's total verbal interaction for the year, and 2yf S H U F H Q W D J H R f English used with peers (out of all verbal interaction with peersyf D Q d percentage of English used with teachers (out of all verbal interaction with teachersyf G X U L Q J W K H \ H D U 7 K H U H P D L Q L Q J W Z R P H D V X U H V I R F X V H d on change over time in each child's use of English. These were 1yf change in the total percentage of English used by each child (either with teachers or peersyf R X W R I D O O L Q W H U D F W L R Q V F R P S D U L Q J W K H I L U V W W R W K e third observation period, and 2yf F K D Q J H L Q W K H S H U F H Q W D J H R I ( Q J O L V h used by each child with teachers and change in the percentage used with peers out of all verbal interactions with each respective speech partner during the first as compared to the third observation period. The results in Table 4 suggest that the relationship of verbal interaction with a particular speech partner to increased English proficiency is dependent to some degree on the linguistic composition of the class. At Corpus Christi, all measures of English use with peers correlated highly with increased proficiency in that language. Botli in terms of all verbal interactions and those verbal interactions solely with peers, significant relationships were found between the percentage of English used throughout the year and increased language proficiency (rho = .94 and .83 respectivelyyf 6 L P L O D U O \ L Q F U H D V H V L Q W K H U H O D W L Y H X V e of English with peers in the classroom in general correlated significant- ly (rho = .83yf Z L W K L Q F U H D V H G ( Q J O L V K O D Q J X D J H S U R I L F L H Q F \ , Q F U H D V H V L n the use of English solely with peers over the course of the year was also related to greater language proficiency, although not significantly so (rho = .70yf . When the Corpus Christi children's verbal interactions with teachers were examined, less consistent patterns were found. There appears to be little relationship between greater English language proficiency and either the amount of English used with the teacher as part of the child's total classroom discourse (rho = .49yf R U F K D Q J H I U R P W K H E H J L Q Q L Q J W o the end of the year in the percentage of all verbal interaction ac- counted for by interactions with the teacher in English (rho =-.14yf . This may be a result of the number of English language resources available to the children among their classmates. Given the limited access to the teacher, who had to divide her time among all of the children in the classroom, the children learning English as a second 414 TESOL QUARTERLY This content downloaded from 137.148.11.31 on Mon, 06 Feb 2017 01:31:20 UTC All use subject to http://about.jstor.org/terms TABLE 4 Rank Order Correlations of Classroom Language Interaction Patterns with Increased English Language Proficiency VERBAL INTERACTION IN ENGLISH WITH PEERS CORPUS CHRISTI MILWAUKEE Use of English with peers as a percentage of the child's total verbal interaction for the year .94' .98* Use of English with peers as a percentage of all of the child's verbal interaction with peers for the year .83* .33 Change from Observation Period I to III in the percen- tage of the child's total classroom verbal interaction that is represented by interaction with peers in English .830 .37 Change from Observation Period I to II in the percentage of the child's interaction in English with peers only .70 .86 VERBAL INTERACTION IN ENGLISH WITH TEACHERS Use of English with teachers as a percentage of the child's total verbal interaction for the year .49 .73 Use of English with teachers as a percentage of all of the child's verbal interaction with teachers for the year .90? .86 Change from Observation Period I to III in the percentage of the child's total classroom verbal interaction that is represented by interaction with teachers in English -0.14 .98? Change from Observation Period I to III in the percentage of the child's interaction in English with teachers only .77 .98? "p .05 language were able to take advantage of their English-speaking classmates as their English proficiency increased. When the children interacted with the teacher, however, the use of English in such interactions was significantly related to increased English language proficiency (rho = .90yf 6 L P L O D U O \ L Q F U H D V H V R Y H U W L P H L Q W K H X V H R f English in the children's discourse with only the teacher were highly related to greater proficiency in English (rho =.77yf . In the Milwaukee classrooms, where Spanish-preferring children predominated, a different pattern of relationships is apparent. Only the overall use of English with peers was significantly correlated with increased English proficiency (rho = .98yf 7 K H O D F N R I V L J Q L I L F D Q t relationships on the other measures of peer-peer interaction may be a result of the relatively few English-preferring children in the class- rooms, which made access to peers as a resource more difficult than at the Corpus Christi site. Thus, in general, at the Milwaukee site teachers served as the primary resource for increasing English language pro- ficiency, as can be seen by the relatively high correlations on all of the INFLUENCE OF TEACHERS AND PEERS ON SLA 415 This content downloaded from 137.148.11.31 on Mon, 06 Feb 2017 01:31:20 UTC All use subject to http://about.jstor.org/terms classroom interaction measures. Relatively greater use of English with the teacher over time was especially related to increases in the children's English language proficiency. Both increased use of English with the teacher as part of the child's general classroom discourse and as a percentage of only those interactions involving the teacher were correlated significantly with increased English language proficiency (rho = .98yf . DISCUSSION The findings of this study suggest that the linguistic composition of a bilingual classroom, at least at the preschool level, can influence the teacher's use of Spanish and English in the classroom. Similarly, it would appear that the linguistic composition of the class influences the nature of classroom interaction and the relative importance of the teacher and peers in increased second language proficiency. It was found that in classrooms where English-preferring children predominated, those children who used relatively more English with peers and who increased their English usage over time generally showed the greatest increases in English proficiency, as measured by mean length of utterance (MLUyf , Q W K R V H F O D V V U R R P V Z K H U H W K e majority of students were Spanish-preferring, children who showed the greatest increases in English language proficiency were those who used relatively more English over time with the teacher. These differences may be a result of the greater access all children had to English-preferring peers in the classrooms where such children pre- dominated. Although more consistent relationships were found between chil- dren's interactional patterns with either teachers or peers and increased language proficiency depending on the type of classroom in which they were located, both conversational partners were resources for increasing language proficiency. In each type of bilingual preschool classroom at least one measure of classroom interaction with each speech partner correlated significantly with increased English language proficiency. In addition, in both types of classrooms teachers were the partner for nearly all interactions in English early in the year. IMPLICATIONS Although the results of our research should be regarded as explora- tory, they do suggest a number of implications for the organization of learning experiences for young children in bilingual settings. While it is clear that a bilingual program will have multiple goals, the focus of the present study makes it appropriate to discuss only those implications related to the teaching of English. 416 TESOL QUARTERLY This content downloaded from 137.148.11.31 on Mon, 06 Feb 2017 01:31:20 UTC All use subject to http://about.jstor.org/terms First, we have shown that regardless of the linguistic composition of the class, teachers are generally the only English language resource for limited English-speaking children early in the year. Thus, even in classes with a large number of children proficient in English, teachers should organize activities for limited English speakers in structured teacher-directed small group sessions which will provide them with sufficient vocabulary and grammatical structures to interact with their more proficient classmates in English. Once limited English speakers achieve a minimal level of compe- tence, our results suggest that the linguistic composition of the class should be taken into consideration in determining the most effective means of facilitating these children's acquisition of English. In class- rooms where English-speaking children predominate, interactions with these children were found to be a factor in the acquisition of English by limited speakers of that language. Hence, teachers in such classrooms might promote peer interaction by structuring some learn- ing activities as cooperative tasks to be worked on by small linguistical- ly heterogeneous groups of children. Given the recent evidence on the effectiveness of peer teaching (Cooper et al. 1982yf Z K H U H D E D O D Q F H R f limited English-speaking children and English-speaking children exists, children of different language proficiencies might be paired to work on individual tasks. In classrooms made up largely of limited English speakers, on the other hand, the children's second language develop- ment might be best facilitated through learning activities which emphasize adult-child interactions. These might be small group activi- ties directed by adults who interact individually with each child or adult-directed large group activities in which children are encouraged to answer individually or as a group in English. The findings also suggest that the linguistic composition of the class may predispose teachers to use one language over the other. Three of the five children at Milwaukee went from using predominately Spanish to employing a majority of English in their verbal interactions, and all of the children at that site were relatively proficient in English at the end of the year. The teaching staff in both Milwaukee classrooms, however, actually increased their general use of Spanish and continued to use a majority of Spanish with most sample children. It would seem, therefore, that, at least for young children, ongoing assessment of children's language use in the classroom might assist teachers in designing learning activities that would take advantage of changes in children's language preferences. Similarly, the dramatic changes in the language use patterns of a majority of the children at both sites suggest that classifications of children's language dominance based on tests administered outside the classroom may not reflect the reality of the classroom. It is likely that INFLUENCE OF TEACHERS AND PEERS ON SLA 417 This content downloaded from 137.148.11.31 on Mon, 06 Feb 2017 01:31:20 UTC All use subject to http://about.jstor.org/terms most of the children in this study would have been judged as Spanish dominant on the basis of their performance on standardized tests, even if these were administered late in the year. Such a designation would not, however, have reflected the language choices made by most of the students within the context of the classroom. This again points out the importance of using methods such as observation to assess children's language use over time and suggests that with young children in a dual language environment a conceptualization akin to our own language preference may be a useful means of classifying children. Finally, it should be noted that the effectiveness with which either classmates or teaching staff were used as a resource for learning a second language depended on the individual child. Our study shows that the students who were least successful in gaining access to the available resources in their classrooms or who were least inclined to use English made the smallest gains in English language proficiency. Thus, teachers should be sensitive to the interactional choices made by individual children which may inhibit their use of English even in situations designed to promote interactions in the second language. An investigation of the socio-psychological variables that condition such choice was, however, beyond the scope of our research. U THE AUTHORS Ray Chesterfield is a Senior Associate with Juirez and Associates in Los Angeles and teaches courses in Educational Anthropology at the University of California, Los Angeles. He has published extensively on informal education in both the classroom and in the home among minority groups of the United States and Latin America. Kathleen Barrows Chesterfield, who holds an M.A. in TESL from the University of California, Los Angeles, is a Research Associate with the Los Angeles-based manage- ment consulting firm of Ju~irez and Associates. Her recent publications include articles in Language Learning (1982yf D Q G W K H + L V S D Q L F - R X U Q D O R I % H K D Y L R U D O 6 F L H Q F H V \f. Katherine Hayes-Latimer is a doctoral student in the Research Methods Department at UCLA. She has conducted extensive anthropological research among the Warao Indians in Venezuela and among minority groups in the United States. Regino ChAvez, a Senior Associate with Juirez and Associates, is a Ph.D. candidate in the Sociology Department at the University of California, Berkeley. His most recent work has centered on early childhood language learning. REFERENCES Barrows Chesterfield, Kathleen, Ray Chesterfield, and Regino ChAvez. 1982. Peer interaction, language proficiency, and language preference in bilingual preschool classrooms. Hispanic Journal of Behavioral Sciences 4(4yf . 418 TESOL QUARTERLY This content downloaded from 137.148.11.31 on Mon, 06 Feb 2017 01:31:20 UTC All use subject to http://about.jstor.org/terms Brown, Roger. 1973. A first language: the early years. Cambridge, Massachu- setts: Harvard University Press. Bruck, Margaret, and Jeffrey Shultz. 1977. An ethnographic analysis of the language use patterns of bilingually schooled children. Working Papers on Bilingualism 13:59-91. Toronto: Ontario Institute for Studies in Education. Cole, Michael, John Dore, William Hall, and Gillian Dowley. 1978. Situation and task in young children's talk. Discourse Processes 1:119-176. Cooper, Catherine, Angela Marquis, and Susan Ayers-Lopez. 1982. Peer learning in the classroom: tracing developmental patterns and consequences of children's spontaneous interactions. In Communicating in the classroom, Louise Cherry Wilkinson (Ed.yf 1 H Z < R U N $ F D G H P L F 3 U H V V . Fillmore, Lily Wong. 1976. The second time around: cognitive and social strategies in second language acquisition. Ph.D. dissertation, Stanford University. Fillmore, Lily Wong. 1980. Learning a second language: Chinese children in the American classroom. In Current issues in bilingual education, James E. Alatis (Ed.yf : D V K L Q J W R Q ' & * H R U J H W R Z Q 8 Q L Y H U V L W \ 3 U H V V . Fillmore, Lily Wong. 1982. Instructional language as linguistic input: second- language learning in the classroom. In Communicating in the classroom, Louise Cherry Wilkinson (Ed.yf 1 H Z < R U N $ F D G H P L F 3 U H V V . Genishi, Celia. 1976. Rules for code-switching in young Spanish-English speakers. Ph.D. dissertation, University of California, Berkeley. Hughes, B. E., and Helene W. Harrison. 1971. Evaluation report of the bilingual education program: Harlandale independent school district; San Marcos independent school district. Harlandale, Texas: Southwest Texas State University. Legarreta, Dorothy. 1977. Language use in bilingual classrooms. TESOL Quarterly 11(1yf . Legarreta, Dorothy. 1979. The effects of program models on language acquisition by Spanish speaking children. TESOL Quarterly 13(4yf . Moll, Luis. 1981. The microethnographic study of bilingual schooling. In Bilingual education technology, Raymond Padilla (Ed.yf < S V L O D Q W L : Eastern Michigan University. Phillips, J. M. 1975. Code-switching in bilingual classrooms. M.A. thesis, California State University, Northridge. Shultz, Jeffrey. 1975. Language use in bilingual classrooms. Manuscript, Harvard Graduate School of Education. Townsend, Darryl R. 1974. A comparison of the classroom interaction patterns of bilingual early childhood teachers. Ph.D. dissertation, University of Texas, Austin. United States Commission on Civil Rights. 1973. Teachers and students: differences in teacher interaction with Mexican American and Anglo students. Report V: Mexican American Study. Washington, D.C.: U.S. Government Printing Office. Wells, Gordon. 1979. Variation in child language. In Language development, Victor Lee (Ed.yf 1 H Z < R U N - R K Q : L O H \ D Q G 6 R Q V . Zamora, Gloria. 1974. A comparison of the nonverbal communication patterns of bilingual early childhood teachers. Ph.D. dissertation, University of Texas, Austin. INFLUENCE OF TEACHERS AND PEERS ON SLA 419 This content downloaded from 137.148.11.31 on Mon, 06 Feb 2017 01:31:20 UTC All use subject to http://about.jstor.org/terms