what matters in second language

Upload: andrew-salanik

Post on 07-Apr-2018

216 views

Category:

Documents


0 download

TRANSCRIPT

  • 8/6/2019 What Matters in Second Language

    1/21

    TESOL QUARTERLY, Vol. 18, No. 2, June 1984

    What Really Matters in Second

    Language Learning for Academic

    A c h i e v e m e n t ?MURIEL SAVILLE-TROIKE

    University of Illinois at Urbana-Champaign

    This re t rospec t ive ana lysis seeks to expla in why a group of

    chi ldren who had been matched for Engl ish prof ic iency and

    socioeconomic status (SES) when they started a school year, and

    who were subsequently taught and tested through the medium of

    English, differed in their school achievement at the end of that

    year. Factors considered include relative productive competence

    in English morphology, syntax, and vocabulary; verbosity; patterns

    of social interaction; first language performance; and personality

    factors. Extensive intra-group variability is reported, but several

    generalizations are drawn which have relevance for ESL curricu-

    lum organization and instructional practice: vocabulary knowledge

    is the single most important area of second language (L2) compe-

    tence when learning content through that language is the depen-

    dent variable; grammatical accuracy is of l i t t le importance to

    students immediate academic needs; communicative competence

    in social interaction does not guarantee communicative compe-

    tence in academic situations; and the use of the first language (Ll)

    enhances conceptual development, even when it is tested through

    the medium of the L2.

    Most of the research exploring factors presumed to contribute tosecond language learning has treated language proficiency as the

    major dependent variable, that is, as the principal outcome to be

    tested. For those of us who teach students who must learn to acquire

    knowledge through the medium of English, however, there is a

    different outcome to be considered. Academic achievement in

    reading and in content areasnot just the learning of Englishhas a

    clear priority in our curricula.

    This study is a retrospective analysis which has been conducted in

    an effort to explain why a group of children who began their schoolyear with an equal lack of proficiency in English were at very

    199

  • 8/6/2019 What Matters in Second Language

    2/21

    different levels of achievement in English-medium instruction by

    the end of that year. The findings are necessarily correlational in

    nature, and while no cause and effect relationships can be claimed,

    i t is hoped that such an analysis wil l contr ibute to a greater

    understanding of the factors which influence the success or failure

    of limited English-speaking students in schools in the United States.

    BACKGROUND

    In the fall of 1981, Erica McClure, Mary Fritz, and I began a study

    of a group of middle-class non-English-speaking elementary school

    students with the expectat ion that our f inal report would be a

    description of successful learning. Many of us have heard, at least

    anecdotal ly, about chi ldren from advantaged backgrounds who

    enter schools in the United States with little or no English and yet

    who manage to cope quite well both socially and academically. Oursubjects were all just such advantaged children, with well-edu-

    cated and supportive parents, generally high expectations on the

    part of their teachers, and positive prior learning experiences. We

    decided to document their natural ly occurr ing success story

    during the 1981-1982 school year and hoped that what we observed

    might be used as a model toward which we could adapt instructional

    settings for less advantaged students. Instead, our final result was

    the documentation of a full range of academic achievement, from

    expected success to unanticipated failure. In addition, even some ofthe cases of success turned out to be surprising.

    A number of possible reasons for such differential achievement

    have been suggested by previous research. I wish to explore several

    of these variables in trying to answer the question I have posed in

    the title of this article: What really makes a difference in second

    language learning for academic achievement?

    INDEPENDENT VARIABLES

    The first variable to be explored is the correlation of various

    aspects of the subjects oral English language proficiency with their

    academic achievement. The general use of language tests to deter-

    mine whether students will be able to learn in English-only pro-

    grams or whether they will require bilingual instruction is based on

    the assumption that the proficiency:achievement correlation must

    be high. However, this assumption has been brought into questionby evidence such as that provided by Ulibarri, Spencer, and Rivas

    (1981), who found that language proficiency tests are not veryuseful for predicting elementary school childrens achievement in

    200 TESOL QUARTERLY

  • 8/6/2019 What Matters in Second Language

    3/21

    mathematics and reading. More information is needed on how

    English language performance and school achievement interact. In

    addition to the correlation of rankings on achievement measures

    and language proficiency tests, the analyses in this study include

    consideration of differential correlations with grammatical com-

    plexity, morphological accuracy, and vocabulary production.

    The relation of both the subjects school achievement and oral

    language proficiency to the interaction structures they participated

    in or experienced is also examined, including the nature and quan-

    tity of their communication with other children and with adults in

    various contexts as well as the extent to which they used English as

    opposed to their native language. Recent research with children has

    generally supported the conclusion that the mere quantity of contact

    with English-speaking children has little effect on second language

    learning but that the degree of active participation in communica-

    tion with English-speaking peers is a significant factor (e.g., Johnson

    1983 and Strong 1983). This conclusion relates to other findings that

    have suggested that the style of interaction is significant in language

    learning: children who are willing to become actively involved in

    conversation-play are more successful in learning a second language

    than those who are either less willing to interact socially or who are

    more concerned with the accuracy of their speech than with its

    social function and content (e.g. , Wong-Fillmore 1976 and Nei-

    moianu 1980).Another variable to be considered is home language influence,

    including transfer from both native language structures and patterns

    of use and from the subjects out-of-school language and educational

    experiences. While the notion of interference as it was conceived

    by Fries (1945) and other applied linguists of his day has proved to

    be inadequate for explaining many phenomena in second language

    learning, recent research has demonstrated that transfer from first to

    second language does indeed occur and, in fact, goes well beyond

    the surface-level structural features noted by Fries. Transfer effectsmay also be found at the metalinguistics level of knowledge about

    how language works as a system (Zehler 1981), as well as in the

    organization and sequencing of discourse (see Kaplan 1983) and in

    styles of social interaction (Neimoianu 1980). Additionally, when

    academic achievement in the second language is being considered,

    the possibility of interdependence (Cummins 1979) between first

    and second language development must be taken into account.

    Finally, I will briefly explore some of the personality and attitudi-

    nal factors which may contribute to the subjects differential aca-

    demic achievement. Strong (1983) has presented a comprehensive

    review of the traits which have been related to success in second

    language learning by chi ldren, and he concludes f rom his own

    SECOND LANGUAGE LEARNING FOR ACADEMIC ACHIEVEMENT 201

  • 8/6/2019 What Matters in Second Language

    4/21

    research that chi ldren who are character ized as talkat ive and

    verbally responsive in their native language are more efficient in

    learning English. Research has thus far yielded conflicting results

    concerning the importance of childrens motivation to become part

    of the target language group to their success in second language

    learning (see Wong-Fillmore 1976 and Strong 1983).

    SUBJECTS

    The population for our research consisted of nineteen children,

    nat ive speakers of seven languagesJapanese, Korean, Hebrew,

    Arabic, Spanish, Icelandic, and Polish. They were enrolled in the

    second th rough sixth grad es and ranged in age from 6 years/ 11

    months to 12 years at the beginning of the project. Criteria for

    subject selection were: 1) very little or no prior exposure to English,

    2) well-educated families (at least one parent possessing a graduatedegree), and 3) initial literacy in their native language. All of the

    subjects lived in the same university-owned housing complex and

    thus shared many of their out-of-school activities as well as their

    instructional experiences. Most of the families were to be residents

    in the United States only for a single year, and none planned to

    remain on a permanent basis.

    In the public elementary school to which the children were bused,

    41% of the students were native speakers of a language other than

    English, and all native English speakers were required to study aforeign language. The nineteen children in our sample were divided

    into two groups, according to their age and maturity level, for daily

    English as a second language (ESL) instruction, and all but the

    Icelandic and Polish speakers also received pull-out instruction in

    their native languages another 30 minutes a day. For the remainder

    of the school day, these children were assigned to regular English-

    medium classrooms with native English-speaking students. Twelve

    of the subjects were in classrooms where one or two other children

    were fluent in their native language; in no case were more than threespeakers of the same nat ive language assigned to one English-

    medium class. All but one of the nineteen subjects had opportunities

    during each school day to interact with native language peers, with

    the greatest amount of L1 interaction possible for speakers of

    Japanese and Korean.

    ACHIEVEMENT

    At the end of the year all students in the school were administeredth e Comprehensive Test of Basic Skills (CTBS) in English, and it is

    202 TESOL QUARTERLY

  • 8/6/2019 What Matters in Second Language

    5/21

    that measure that serves as the dependent variable for this retrospec-

    tive analysis. The CTBS is a standardized battery which includes

    subtests for reading, language, social studies, science, and mathe-

    matics. The matrix in Table 1 shows that the correlations of relative

    student performance on the first four subtests are all significant,

    suggesting that they tap a common factor in ability to comprehend

    and perform academically in English (see Oller 1980).1

    To illustrate the wide range of achievement under consideration,

    the most successful student (a third-grade Japanese girl) scored atthe 65th percentile in reading according to national norms after only

    one year of English, and the least successful student (also a Japanese

    girl, in the fourth grade) scored only at the first percentile. This most

    successful achiever also scored at the 99th percentile in the English

    language subtest, which includes spelling and mechanics, the 97th

    percentile in science, and the 85th percentile in social studies.

    Quantitative math scores were frequently at variance with scores

    on these language-dependent subtests and seemed to reflect the

    differential coverage and sequencing of math concepts which thestudents had already encountered in their home country curricula.

    Twelve of the nineteen subjects scored above the 90th percentile in

    math computation (including all of the students from Korea and

    Japan), and half of these scored at the 99th percentile according to

    United States norms. That there was no relation to English ability is

    shown by the fact that these high achievers in math included subjects

    who scored among the lowest i n r e ad i n g an d o t h e r l an g u ag e -dependent content areas. Their high math scores also show that the

    1 All correlations reported in this study are based on the rank ordering of subjects according toeach variable.

    SECOND LANGUAGE LEARNING FOR ACADEMIC ACHIEVEMENT 203

  • 8/6/2019 What Matters in Second Language

    6/21

    low reading achievement was not due to any general cognit ive

    deficit, although we did not attempt to test for intelligence.

    Scores on the reading subtest of the CTBS serve as the primary

    basis for the rank ordering of achievement in this study, with other

    scores (except math) taken into account when the reading scores

    were identical for two or more individuals.

    METHODS

    All subjects were videotaped weekly in their ESL classes through-

    out the year, observed weekly in regular classroom and playground

    settings, and videotaped in these contexts approximately once a

    month. By design, ESL sessions included an unsupervised period

    before the teachers entered the room, which allowed us to videotape

    the childrens unstructured intra-group communication. Addition-

    ally, all subjects were tracked through a complete school day, fromthe moment they stepped off the bus in the morning until they

    boarded it again 5 hours later . A schedule of activit ies and

    transitions was recorded, and at 30 second intervals the observer

    tallied the nature of verbal interaction in which the subject was

    involved (if any), including which language was used, whether with

    an adult, a child, or a text, and whether the interaction took the form

    of speaking, listening, reading, or writing.

    At the end of the year pairs of children who did not share a

    common native language were videotaped as they were interviewedby an adult with whom they had frequently interacted. An attempt

    was made to match children within each pair as closely as possible

    for aggressiveness and talkativeness in order to prevent one child

    from dominating the situation. Each interview lasted 30-45 minutes

    and included quest ions about the chi ldrens past , habi tual , and

    future activities to provide grammatical as well as content informa-

    tion. Each subject was also asked to recount one of his or her

    favorite folktales in English (e.g., The Three Pigs, Goldilocks and

    the Three Bears, or Cinderella). Three English language tests werealso administered during the latter part of the year: the Northwest

    Syntax Screening Test (NSST), the Functional Language Survey

    (FLS), and the Bilingual Syntax Measure (BSM).

    Data on home language and personality factors were elicited in

    informal interviews with parents and teachers. These factors were

    explored in more depth for a subsample of the population by Koda

    (1982), a native Japanese speaker. After the school year ended, sheinterviewed four of the Japanese students and their parents at some

    length in their first language and observed the. families interacting intheir homes.

    204 TESOL QUARTERLY

  • 8/6/2019 What Matters in Second Language

    7/21

    ENGLISH LANGUAGE PROFICIENCY

    Student rankings on the three language tests that were adminis-

    tered did not predict achievement on the reading subtest of the

    CTBS, corroborating the findings of Ulibarri, Spencer, and Rivas

    (1981). As can be seen in Table 2, the subjects performance across

    language tests was consistent, but there was a very low correlation

    between performance on any one of those tests and the CTBS

    reading subtest administered in English.

    One clear pattern that did emerge was that speakers of Indo-

    European languages performed better on these language tests than

    did speakers of unrelated languages (e.g., r = .799 between the NSSTand the language relatedness factor). On the other hand, no specific

    first language appeared to have any particular relationship to rank

    order of achievement on the reading subtest of the CTBS.

    Other indices of end-of-year oral English language proficiency for

    this study were rank orderings of subjects performance based on

    analyses of the videotaped interviews. Quantitative measures in-

    cluded calculations of verbosity (ratio of total vocalizations per turn

    to number of words, with hesitation markers and repetitions re-

    moved), mean length of T-units (MLT) calculated in morphemes,grammatical accuracy rate (calculated as the inverse of the error

    ra te2), and the total number of different vocabulary items used by

    each child. The correlations of the rank ordering of subjects on each

    2 The grammatical error rate was calculated by McClure and Blomeyer (1983) for 16 of thesesubjects, using segments of the interviews which included the folktales and anecdotes.Three of the subjects were excluded from their study because their very short utterancesprovided too little data for this type of grammatical analysis. The error rate is the sum ofdeterminer, syntactic, and morphological errors divided by the number of words: morpho-logical and determiner error rates are omissions or substitutions per obligatory context; the

    syntactic error rate is syntactic errors per clause plus phrases. The correlation wassignificant between ranking according to this rate and the Northwest Syntax Screening Test(r= .726; p < .05).

    SECOND LANGUAGE LEARNING FOR ACADEMIC ACHIEVEMENT 205

  • 8/6/2019 What Matters in Second Language

    8/21

    of these factors with their ranking on the CTBS reading subtest and

    the NSST are shown in Table 3.

    The range in quanti ty of product ion was very wide, f rom a

    minimum of 57 to a maximum of 182 turns at talk for each subject,

    vocalizations from just over 1 (1.198) to almost 16 (15.947) per turn

    (mean = 8.416; sd = 3.714), and the number of different vocabulary

    items from less than 50 (48) to more than 250 (262) (mean= 187.316;

    sd = 60.844). Verbosity had a positive but non-significant correlationboth with language test scores (r= .353) and with reading achievement

    (r = .404). The correlation of MLT with reading achievement was

    even lower (r = .114).

    Our first finding was that accuracy in English morphology and

    syntax in spoken language appears to make l i t t le di f ference in

    academic achievement, which probably accounts for the low predic-

    tive power of the language tests; the Japanese girl who scored at the

    65th percentile in reading at the end of the year still systematically

    omitted plurals, articles, and tense markers in her spoken English(al though she had scored at the 99th percent i le on the CTBS

    language subtest), as did four out of five others among the top six

    academic achievers. On the other hand, the Polish, Icelandic, and

    Spanish speakers ranking 13th, 14th, and 18th, respectively, out of 19

    in reading achievement scores had a far better mastery of English

    morphology, perhaps because of positive transfer of inflections

    from their nat ive Indo-European languages. The correlat ion be-

    tween students rank order of reading achievement and grammatical

    accuracy was only .025.Based on our interview data, the only factor in oral English

    206 TESOL QUARTERLY

  • 8/6/2019 What Matters in Second Language

    9/21

    language production which was found to be significantly correlated

    wi th reading achievement was the number of different vocabulary

    items each child used (r = .633; DF = 17, p < .05). Among language

    factors, there was a significant positive correlation of MLT with

    verbosity and with grammatical accuracy.

    INTERACTION PATTERNS

    Child-Child Interaction

    Analyses of the weekly videotapes indicate that in situations with

    other limited English speakers, the childrens primary reasons for

    communicating were either interfactional or performative (based on

    a system of coding adapted from Dore 1977), and they had little

    need to request or give information of any kind. The most common

    communicative goals were to engage another child in play, to tease,

    to protest , to call attention to things, and to claim objects or

    territory. As our subjects amply demonstrated, children really do

    not need language to achieve these goals.3The form of most of these

    acts was entirely non-verbal (i. e., gestures and physical contact

    accompanied by non-speech sounds). Observations and recordings

    on the playground yielded the same finding: children can and do

    participate in a great deal of social activity, even when fairly

    complicated rules are involved, with little or no language. Other

    English-speaking children taught games to non-English speakers by

    demonstrating what to do and by correcting their mistakes with a

    simp le No, or with physical intervention.

    There was almost no increase in the amount of verbalization used in

    these events throughout the year, even as the children developed their

    proficiency in English. One significant change in the verbalizations

    that did occur was from the use of the native language to protest and

    claim (even in addressing children who did not understand the lan-

    guage) to the use of such English routines as dont do and thats mine.

    Early greetings between children from different language back-

    grounds who had had no pr ior exper ience wi th Engl ish were

    accomplished by such means as shaking raised fists at one another or

    exchanging non-linguistic vocalizations. For instance:

    Child #1 (initiating exchange): Ummm.

    Child #2 (respond s): Ummm .

    The following invitation to playfrom a fourth-grade Korean

    boy (K) to a Japanese classmate (J)involved some of the first use

    of English that was recorded:

    3 Analysis of this interaction for the first half of the school year is also reported in Saville-Troike, McClure, and Fritz (in press).

    SECOND LANGUAGE LEARNING FOR ACADEMIC ACHIEVEMENT 207

  • 8/6/2019 What Matters in Second Language

    10/21

    K: Excuse me. Me. Me-me.

    (He points to h is own chest.)OK?

    J: (Nods head and s t ands up . )Me? OK.

    (They go off to play together.)

    The earliest tactics used by our subjects to communicate with oneanother in more complex interaction events can be illustrated by an

    episode that occurred in the third week of videotaping in the fall.

    This sequence occurred as three boysspeakers of Japanese (J) ,

    Korean (K), and Spanish (S)worked together to assemble a puzzle

    map of the world. Conversational rounds consisted mainly of one

    child naming a referent and the other children taking turns repeating

    the term in round-robin fashion.

    S:

    J:

    America.(He finds a puzzle piece of the United States and puts it inplace.)(He picks up Ss hand and moves it out of the way.)

    Japani, Japani.(He points to the place in the puzzle frame where the Japanpiece w ill go.)

    Japani.Japani.Japani.Japani.Ah. Japani.

    (He find s the Japan p iece to the p uzzle.)Korea.

    (He also claims to have found Korea.)

    this point, add ing only No and using d ifferent stress and intonationalcontours, the children d isagree and question one another,

    K: No Korea.

    J: Korea.

    K: No.J: Korea.S: Korea?K: No.S: Si? (Yes?)

    The Korean boy then offers an explanation, which the others acknowl-edge by repeating it after him:

    K: Yes, yes, yes. No Korea. Pink Korea.(Stress on pink. Pinkrefers to the color of the Korean map piece;

    that is, the child is saying that this piece cann ot be Korea becau seKorea is pink.)

    208 TESOL QUARTERLY

  • 8/6/2019 What Matters in Second Language

    11/21

    J: Pink.

    S: Pink.K: Korea.

    (He finds the p iece and pu ts it in p lace.)

    At this point the ESL teacher enters the scene, telling the class to put

    away th eir gam es.

    As is evident, all of the participants in this complex eventcommunicated quite effectively for their purposes. Yet very little

    English was actually required for them to do so. They successfully

    cooperated in the construct ion of the event and negotiated the

    accomplishment of their communicative aims with a total vocabu-

    lary of six words.

    Even joking among students was accomplished with minimal

    English. One day during the unsupervised interval prior to the ESL

    teachers arrival for class, a second grader took the teachers usual

    place at the head of the table, tapped the chalkboard, and addressedthe other students in a loud voice with, What day? At the same

    time he made a gesture indicating he had on glasses, which she

    wore. His only other verbalization involved repetit ions of Be

    quiet! to several children, but his pantomime included pretending

    to write on the board and sitting at the head of the table to check the

    roll in the teachers book. He then deliberately fell out of her chair

    and was applauded enthusiastically.

    Such abi l i ty to communicate with peers did not necessar i ly

    generalize in a way which proved helpful in other school situations.In particular, students who demonstrated early skill at combining

    English and miming to communicate successfully with other limited

    English-speaking children and their ESL teachers did not experience

    comparable success with teachers in their regular classrooms. This

    strategy, however, proved to be very successful for communicating

    with their English-speaking peers and was used by all of those of the

    research population who from the beginning of the year interacted

    most with native English-speaking children at school. But important-

    ly, this group included none who scored highest on the standardized

    achievement tests given in English at the end of the year.

    This was a surpr ising f inding, since our or iginal assumption

    (based primarily on Wong-Fillmore 1976) had been that children

    who interacted socially the most extensively with other English-

    speaking children would learn English faster and thus do better in

    English-medium instruction; this turned out not to be the case. In

    fact, at least for the first half of the year, three of our five highest-achieving subjects used their native languages with peers to the

    virtual exclusion of English, while the other two top achievers rarelyspoke at all in any language to other children during the ESL or

    SECOND LANGUAGE LEARNING FOR ACADEMIC ACHIEVEMENT 209

  • 8/6/2019 What Matters in Second Language

    12/21

  • 8/6/2019 What Matters in Second Language

    13/21

    probably unders tood only the word shape, he appropr ia te ly re-

    sponded, Triangle. The teacher, when asked shortly afterward to

    evaluate the students English, reported that he was quite fluent.

    Each ESL session we videotaped also included less structured

    segments, where interact ion was character ized pr imari ly by the

    need for a real exchange of information, such as the teachers asking

    the whereabouts of absent chi ldren before the st ructured lesson

    began, or the explanation and implementation of an art project or

    game fol lowing the st ructured lesson. I t was only in this semi-

    directed type of event throughout the year that the students

    communicative production regularly included the functional cate-

    gories of requests, descriptions, and statements, and it was only in

    this type of event that they went beyond memorized patterns as

    they struggled to obtain needed information or to express themselves

    to the teacher and other children. The results were often ungram-

    matical, but i t is interesting to note that grammar was never

    corrected when real communication was at stake. Successful com-

    munication, in fact, often included a combination of English, non-

    verbal miming, and the speakers native language.

    Analysis of Interaction D ata

    An analysis of the language forms presented in the structured ESL

    lessons and those language forms actually used by the students for

    genuine communicative purposes in less structured events shows

    very li t t le carry-over from one context to the other, except for

    vocabulary. The is of this is a pencil, for instance, was generally not

    used for weeks or months in other contexts and, as in the interview

    data (along with ar t icles and past tense or plural inf lect ions) ,

    continued to remain absent in the speech of some who achieved

    considerable fluency in both communicative and academic uses of

    English.

    Observation and recording during the rest of the childrens five-

    and-a-half-hour school day showed that the real world communi-

    cative demands of the regular classroom most closely resembled

    those found in the less structured events in the ESL classroom, so it is

    not surprising that students carried over strategies and behaviors

    from one to the other.

    There was, of course, intra-group variation. A few of the students

    sought interaction only with the teacher during these information

    exchanges, and a few interacted almost entirely with other children.

    When another child in the group spoke the same native language,

    some chose to obtain information and express themselves primarily

    through that language rather than English, especially during the first

    half of the year. The children we studied who preferred to request

    SECOND LANGUAGE LEARNING FOR ACADEMIC ACHIEVEMENT 211

  • 8/6/2019 What Matters in Second Language

    14/21

    information and express themselves in their native language in the

    ESL class did so throughout the school day, often seeking explana-

    tions of directions that had been given in English from bilingual

    adults or classmates. Not infrequently, these limited English speakers

    joined forces to figure out problems they could not understand. Our

    recordings show that a very high percentage of native language

    communicat ive acts were direct ly concerned with assigned aca-demic tasks. Some students used their native language even when

    there were no other speakers present, verbalizing to themselves as

    they worked. As noted ear l ier , this nat ive language preference

    group included three of the five children who scored highest in

    English reading and content areas (e. g., science), as tested through

    English at the end of the year (CTBS).

    Two other kinds of participant structures were frequently ob-

    served and recorded in regular classrooms (but never in ESL, partly

    because of the smaller class size), and it was in these that we sawmost of the limited English-speaking children entirely unable to

    cope: staring out windows, doodling, poking their neighborsone

    even crawling under his desk in retreat. These are the teacher-to-

    whole-class participant structure, where the group was talked (or

    read) to as a whole, and its opposite, the fully independent partici-

    pant structure, where students were given written or oral instructions

    in English and expected to proceed without additional interaction

    with adults or peers.

    The amount of time allocated to these two types of participantstructure varied greatly from class to class, but they were found to

    some extent in all. In the classrooms we observed through the year,

    these kinds of participant structures occupied from ten to over fifty

    percent of the instructional day. The percentage generally increased

    with grade level, and there is every reason to assume that we would

    have found students engaged inor disengaged fromeven more

    of these kinds of instructional situations if our research had continued

    into junior high and high school.

    To give a general idea of the quantity of interaction across thesample, the tallies made at 30 second intervals throughout the day

    showed that, almost half of the time (45.48%), our subjects were

    involved in no verbal interaction at all, with a range of approxi-

    mately 30-60%. When they did interact, they interacted more with

    adults in English (18.03%) than with other students (10.76%), and even

    more with written texts (19.78%). Perhaps predictably, students who

    interacted more with their peers spoke English more frequently thandid others, while those who interacted more with adults spent a

    greater percentage of time listening.

    212 TESOL QUARTERLY

  • 8/6/2019 What Matters in Second Language

    15/21

    As shown in Table 4, there was a very low correlation between

    school achievement and t ime spent using English. The overall

    amount of English verbal interaction has only a small positivecorrelation with achievement, and the amount of interact ion with

    other students a slightly negative one. Correlations of achievement

    with the amount of interaction with adults and with written texts

    during the hours sampled are also non-significant. (However, overall

    amount of English verbal interaction does correlate positively with

    language proficiency rankings on the Northwest Syntax Screening

    Test [r= .692] and with MLT in the interviews [r= .498].)

    NATIVE LANGUAGE INFLUENCE

    As for what does seem to make a difference, the childrens first

    language appeared to influence both second language learning and

    academic achievement in several important ways. Most obvious was

    the transfer of native language forms into English. As mentioned

    earlier, speakers of Indo-European languages were more accurate in

    their production of English morphology than were speakers of non-

    Indo-European languages. There was transfer in syntax as well,

    including the following examples from end-of-year interviews withJapanese students, which illustrate native-language word order used

    in English:

    They had Mr. Smith teach they my class three people.

    If I go my mother-with shopping and my mother didnt shopped my

    like-things, I dont like.

    In terms of transfer of communicative style, while in some cases

    children who were quite talkative in their native language did not

    choose to speak much in English, all quiet children were quiet inboth languages.

    SECOND LANGUAGE LEARNING FOR ACADEMIC ACHIEVEMENT 213

  • 8/6/2019 What Matters in Second Language

    16/21

    Another significant factor was first language proficiency. We did

    not use formal tests of the students speaking and reading abilities in

    their native languages, but in almost all cases bilingual instructors

    judgments of students relat ive competence in nat ive language

    studies coincided with the same students relative achievement in

    English. A reasonable hypothesis is that reading achievement in

    English is more dependent on native-language reading ability thanon proficiency in oral English, but this clearly requires further

    invest igat ion. The Japanese gir l who scored highest in reading

    English, for instance, was reported to read several years above

    grade level in Japanese, and her Japanese vocabulary and grammar

    were considered exceptional for a child her age.

    There appears to be some specific transfer of reading skills

    involved, such as the strategies used for inferring the meanings of

    unfamiliar words. This speculation is based on different types of

    errors made in multiple-choice vocabulary tests, where Japanesechildren seemed to be more likely to choose an alternative which

    had the same sequence of two or more letters in any word position

    (a strategy which may be similar to conducting a visual search for a

    common radical in a Chinese character) . Speakers of Western

    European languages, on the other hand, appeared more likely to

    look only for a similar stem (a search for cognates), and much less

    likely to choose an alternative which had the same sequence of

    letters at the end of a syllable or in suffix position.

    One exception to the bilingual achievement pattern was a Japa-nese girl who read Japanese well, but not English. She did not talk in

    either language, although she was very friendly and interacted non-

    verbally with both Japanese- and English-speaking children. At the

    end of the year her American teacher said she had never even heard

    her voice and interpreted her silence and passivity in class to mean

    she was a slow learner. As a result, little was expected of this student,

    and little was achieved. The other exception was an Israeli boy who

    had difficulty reading Hebrew but for some unexplained reason did

    very well in learning to read English.

    PERSONALITY AND ATTITUDES

    The substudy conducted by Koda (1982) focused on four of the

    Japanese students. The four students were all girls: one was the

    highest achiever, one above average, one below average, and one

    the lowest in the total sample. All of their fathers are professors in

    Japan, and all of their mothers are housewives who spoke little or no

    English upon arrival in the United States.After extensive interviews and observations, Koda concluded that

    214 TESOL QUARTERLY

  • 8/6/2019 What Matters in Second Language

    17/21

    the two better achievers in the subsample shared active and competi-

    tive coping styles at school and a confident attitude at the beginning

    of the year that they would succeed, while the two poorer achievers

    were passive, compliant, and dependent as they faced the challenges

    of learning in a new language. The high achievers also both initiated

    independent learning act ivi t ies. Although the highest achiever

    avoided speaking English at school for the first three months of the

    year, she took an early interest in writing the language and, on her

    own initiative, spent over an hour each day after school translating

    sentences f rom Japanese to English with the aid of a bi l ingual

    dictionary. She noticed the different syntax and enlisted help from

    her mother with word order . Both low achievers did al l of the

    assigned work diligently but never attempted anything extra. The

    lowest was especially fearful of making a mistake and very cautious

    about giving the impression that she did not understand, whether in

    Japanese or in English Koda attributes this in part to the girls very

    tradi t ional social izat ion to obedience and feminini ty and to her

    fathers high expectations for her and his attitude that failure would

    be most shameful.

    Even among the four Japanese girls, however, individual differ-

    ences were more striking than the similarities. While one of the

    higher achievers had a very positive attitude toward English, the

    other did not and consciously avoided playing with American

    children. The lowest achiever, on the other hand, was very positive

    about English and shared only English with her best friend. Further-

    more, the characteristics that were the same for high- and low-

    achieving Japanese girls did not hold true across language groups or

    for Japanese boys. The second and third highest achievers in the

    total sample were very passive and compliant speakers of Icelandic

    and Spanish, respectively, while aggressive and extroverted speakers

    of Spanish, Korean, Japanese, and Arabic were among the lowest in

    achievement scores.

    CONCLUSIONS

    Perhaps the most important point to be made from the analyses

    offered in this article is the extent to which there were individual

    differences among the subjects. Statistics for groups of students too

    often can mask what is actually happening to individuals as they

    succeed or fail in learning English orwhat is more important from

    my perspective herein learning through English. While the diver-sity is particularly apparent in our study of children from such

    different language and cultural backgrounds, there is also great

    diversity among children who share the same mother tongue.

    SECOND LANGUAGE LEARNING FOR ACADEMIC ACHIEVEMENT 215

  • 8/6/2019 What Matters in Second Language

    18/21

    Therefore, to the question about what really made a dif ference

    consistently for all of the children in this study, the only answer must

    be: Very little. There are a few generalizations that can be drawn

    from our findings, however, and these have potential relevance for

    teaching English as a second language:

    1. Vocabulary knowledge in English is the most important aspect of

    oral English proficiency for academic achievement. Vocabulary

    taught in ESL should therefore be related as closely as possible to

    students learning needs in their subject matter classes.

    2. Spoken practice in English may not be necessary for the devel-

    opment o f Eng l i sh p ro f i c i ency and may r e ta rd i t i n some

    instances. Emphasis on interpersonal communication may even

    inhibit academic achievement.

    3. The portions of ESL lessons which focus on structural patterns,

    especially on English morphology, appear to make little contri-but ion toward meeting students immediate academic needs.

    Most beginning students do not use grammatical inflections when

    they are concerned with communicating real information. Mas-

    tery of English grammatical structure is more closely related to

    native language background than to the ability to use English for

    academic purposes.

    4. Most children do not have to be taught to communicate with one

    another; they will do that even without a common language.

    While social interaction between students is certainly to be

    encouraged, we cannot depend on that alone for developing

    English language skills.

    5. Most of the children who achieved best in content areas, as

    measured by tests in English, were those who had the opportu-

    nity to discuss the concepts they were learning in their native

    language with other children or adults. Even in l inguistically

    heterogeneous classrooms such as those in this study, at least

    some degree of bilingual education is proving to be feasible and

    clearly provides the best context for conceptual development

    and for learning English.

    We need to r ecogn ize tha t the re i s a qua l i t a t ive d i f f e rence

    between the communicat ive tact ics and ski l ls that chi ldren f ind

    effective for meeting their social needs and goals and those that are

    necessary for successful academic achievement in the classroom.

    The lowest academic achievers in our sample were among the mostsuccessful at interpersonal communication, especially with other

    children. As Cummins (e.g., 1980, 1981) has argued on theoreticalgrounds, academic success requires competence in using and under-

    standing language in context-reduced situations, where students

    216 TESOL QUARTERLY

  • 8/6/2019 What Matters in Second Language

    19/21

    cannot rely on non-verbal elements of communication. The language

    skill which is most likely to develop this competence is writing, yet it

    is sadly last, not only in the traditional ESL litany of listening-

    speaking-reading-writing but also in the time and attention allotted

    to it in most ESL classes. The few students in our sample who could

    cope with independent instructional activities in regular classrooms

    possessed skills that are generally not taught in ESL classes at theelementary school level but could be, such as how to make good use

    of a bilingual dictionary.

    It is a positive development that we in TESL have broadened our

    focus f rom grammatical competence to communicat ive compe-

    tence. But many who have jumped aboard this newest bandwagon

    have unfortunately misinterpreted communication to apply only

    to social interaction, and such a limited conceptualization still fails to

    fulfill our accountability for students who must learn how to learn

    through the medium of English. We need to develop their academiccompetence as well, and this calls for even more changes in our

    priorities and in our procedures.

    Developing social language skills is a desirable but insufficient

    goal for English teaching and (ironically) may even interfere with. .academic achievement. We must begin to place more emphasis on

    vocabulary learning and less on grammar and pronunciation. Too

    often we in ESL have forgotten that teaching English is not an end in

    itself but only a means to an end; the critical outcome for those of us

    teaching children is how well we equip them to succeed in school.This should become the critical outcome as well for researchers,

    who should no longer content themselves merely with studying how

    well students learn English. If in teaching ESL we fail to teach the

    language needed to succeed in the regular classroom, we have failed

    in our first responsibilitywhich is to our students. If ESL does not

    contribute to this goal, then the time spent on it in elementary and

    secondary schools cannot be justified.

    ACKNOWLEDGMENTS

    This article is a revised version of a Forum Lecture delivered at the 1983 TESOLSum mer Institute in Toronto. I wish to express my ap preciation for comm ents byAlison d Anglejan, Gary A. Cziko, Christina Bratt Pau lston, and tw o anonymousTESOL Quarterly reviewers. Particular thanks are du e to Mary Fritz and to the

    teachers who collaborated in the research, withou t w hose fu ll cooperation thisstudy wou ld not have been p ossible.

    SECOND LANGUAGE LEARNIN G FOR ACADEMIC ACHIEVEMENT 217

  • 8/6/2019 What Matters in Second Language

    20/21

    THE AUTHOR

    Muriel Saville-Troike, who received her Ph.D. in Linguistics from the University ofTexas, is past p resident of TESOL and the A m erican Association for A pp liedLingu istics and is the author ofFoundations for Teaching English as a Second

    Language and The Ethnography of Communication. She has taught at theUniversity of Texas and at Georgetow n U niversity and is cu rrently Associate

    Professor of Elementary an d Early Childhood Edu cation at th e University ofIllinois at Urbana-Champaign.

    REFERENCES

    Cummins, James. 1979. Linguistic interdependence and the educational

    d e v e l o p m e n t o f b i l i n g u a l c h i l d r e n , Review of Educational Research

    49:222-251.

    Cu m m ins, Jam es. 1980. Entry and exit fallacy in bilingual edu cation.

    NABE Journal 4(3):25-59.Cummins, James. 1981, The role of primary language development in

    promoting educational success for language minority students. In School-

    ing and language minority students: a theoretical framework, 3-49. Los

    Angeles, California: Evaluation, Dissemination, and Assessment Center,

    California State University.

    Dore, John, 1977. Oh them sheriff: a pragmatic analysis of childrens

    responses to quest ions. In Child discourse, S u sa n E r v i n - T r i p p a n d

    Claudia Mitchell-Kernan (Eds.), 139-163. New York: Academic Press.

    Fr ies , Char les C . 1945, Teaching and learning English as a foreign

    language. Ann Arbor, Michigan: University of Michigan Press.Johnson, Donna M. 1983. Natural language learning by design: a classroom

    experiment in social interaction and second language acquisition. TESOL

    Quarterly 17(1):55-68.

    Kaplan , Rober t B . (Ed.) . 1983. Annual review of applied linguistics.

    Rowley, Massachusetts: Newbury House.

    Koda, Keiko. 1982. English language acquisition by four Japanese girls:

    case studies and influencing factors. Manuscript, University of Illinois at

    Urbana-Champaign.

    McClure, Erica, and Charlotte Blomeyer. 1983. Style shifting in the speech

    of children acquiring English as a second language. Manuscript, Univer-sity of Illinois at Urbana-Champaign.

    Ne imoianu , Anca M. 1980 . The boats gonna leave: a study of children

    learning a s e c o n d language from conversations with other children.

    Amsterdam: John Benjamins B.V.

    Oller John W. 1980. A language factor deeper than speech: more data and

    theory for bilingual assessment. In Current issues in bilingual education,

    James E. Alatis (Ed.), 14-30. Washington, D. C.: Georgetown University

    Press.

    Saville-Troike, Muriel, Erica McClure, and Mary Fritz. In press. Communi-

    cative t attics in childrens second language acquisition. In Universals ofsecond language acquisition, Fred R. Eckman, Lawrence H. Bell , and

    Diane D. Nelson (Eds.). Rowley, Massachusetts: Newbury House.

    218 TESOL QUARTERLY

    http://www.ingentaconnect.com/content/external-references?article=0039-8322()17:1L.55[aid=9192741]http://www.ingentaconnect.com/content/external-references?article=0039-8322()17:1L.55[aid=9192741]http://www.ingentaconnect.com/content/external-references?article=0039-8322()17:1L.55[aid=9192741]http://www.ingentaconnect.com/content/external-references?article=0039-8322()17:1L.55[aid=9192741]http://www.ingentaconnect.com/content/external-references?article=0039-8322()17:1L.55[aid=9192741]http://www.ingentaconnect.com/content/external-references?article=0039-8322()17:1L.55[aid=9192741]
  • 8/6/2019 What Matters in Second Language

    21/21

    Strong, Michael. 1983. Social styles and the second language acquisition of

    Spanish-speaking kindergartners. TESOL Quarterly 17(2):241-258.

    Ulibarri, Daniel M., Maria L. Spencer, and Guillermo A. Rivas, 1981,

    Language proficiency and academic achievement: a study of language

    proficiency tests and their relationship to school ratings as predictors of

    academic achievement. NABE journal 5(3):47-80.

    Wong-Fillmore, Lily. 1976. The second time around: cognitive and social

    strategies in second language acquisition. Ph.D. dissertation, Stanford

    University.

    Zehler, Annette M. 1981. The reflection of first language-derived experi-

    ence in second language acquisition. Ph.D. dissertation, University of

    Illinois at Urbana-Champaign.

    SECOND LANGUAGE LEARNING FOR ACADEMIC ACHIEVEMENT 219

    http://www.ingentaconnect.com/content/external-references?article=0039-8322()17:2L.241[aid=9208263]http://www.ingentaconnect.com/content/external-references?article=0039-8322()17:2L.241[aid=9208263]http://www.ingentaconnect.com/content/external-references?article=0039-8322()17:2L.241[aid=9208263]