UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The generation of academic discourse by ESL learners through computer-based peer tutoring; a case study Rice, Curtis 1995

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_1995-060535.pdf [ 23.3MB ]
Metadata
JSON: 831-1.0054903.json
JSON-LD: 831-1.0054903-ld.json
RDF/XML (Pretty): 831-1.0054903-rdf.xml
RDF/JSON: 831-1.0054903-rdf.json
Turtle: 831-1.0054903-turtle.txt
N-Triples: 831-1.0054903-rdf-ntriples.txt
Original Record: 831-1.0054903-source.json
Full Text
831-1.0054903-fulltext.txt
Citation
831-1.0054903.ris

Full Text

THE GENERATION OF ACADEMIC DISCOURSEBY ESL LEARNERSTHROUGH COMPUTER-BASED PEER TUTORINGA CASE STUDYbyCURTIS RICEB.A., The University of British Columbia, 1986M.A., Simon Fraser University, 1989A THESIS SUBMITTED IN PARTIAL FULFILMENT OFTHE REQUIREMENTS FOR THE DEGREE OFDOCTOR OF PHILOSOPHYinTHE FACULTY OF GRADUATE STUDIESCentre for the Study of Curriculum and InstructionWe accept this thesis as conforminghe r quired standardTHE UNIVERSITY OF BRITISH COLUMBIAMay 1995© Curtis Rice, 1995In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.Department of GZ) c: fThe University of British ColumbiaVancouver, CanadaDate ‘.-- ‘(Signature)DE-6 (2/88)11ABSTRACTDoes peer tutoring using computer-based hypermedia resources help ESL studentsgenerate academic discourse in L2? This relates issues of SL learning and the computer ineducation. I observed 10 upper elementary ESL students research the topic “Earth and theSolar System”, build a HyperCard stack (“Our World”) to record their results, and peer tutorthe stack to younger classmates. Some peer tutored the stack again to ex-ESL Grade 5students. I recorded both peer tutoring sessions (PT(1) and PT(2)) and analyzed discoursetranscripts by quantitative variables, pedagogical objectives, cognitive functions and tutoringstyle; linked our discourse analysis to Krashen’s (1985) Input Hypothesis, Cunimins (1991)concepts of conversational and academic language proficiency, and Halliday’s (1985) modelof language socialization distinguishing interpersonal, ideational and textual components; andfollowed Staab (1986), in dividing the ideational component into “Informing” and“Reasoning”.In PT( 1) students spent more time talking (56%) in computer-based peer tutoring thanin any other activity (10-16%). Informing was high (65%) but Reasoning was low (22%).Tutors used predominantly the traditional I-R-E knowledge-transmission teaching model,speaking 2½ times as much as tutees. In PT(2) tutors were given as aids 1) a Tutorial stackwith knowledge-structure-based computer graphics to represent each topic and 2) training inmoving from I-R-E to more equal dialogue exchanges with tutees. I compared a selected pairin PT(2) with a selected pair in PT(1). In PT(2). tutors produced 27% and tutees 19% morelanguage; Reasoning increased from 22% to 39% overall and to 46% in the Tutor Explanation111tutoring mode; and I-R-E discourse dropped from 62% to 13%. These changes marked amove from traditional knowledge-transmission towards a knowledge-construction paradigm.I conclude that 1) peer tutoring holds great promise for development of academicdiscourse in the L2; 2) but without training, tutors are likely to fall back on I-R-E teacherdominated discourse with a low proportion of Reasoning; 3) interactively using the computerfacilitates a shift from traditional knowledge-transmission to cooperative knowledgeconstruction learning; 4) the students’ use of elementary multimedia technology provides awindow to a future shift from print to electronic technology and towards a knowledgeconstruction paradigm.ivTABLE OF CONTENTSAbstract iiList of Tables xiList of Figures xiiAcknowledgements xivINTRODUCTIONThe research questionSLA perspectivesSignificance of academic discourseCumminsHallidayPeer TutoringResearch methodologyDiscourse AnalysisWork Study concepts and techniquesOVERVIEW“Our World” unit: work flowOutline Flow Process1: Construction of Our World stack & first peer tutoring, PT(1)2: Second peer tutoring sessions, PT(2)3: Comparative analysis of selected pairs in PT( 1) and PT(2)Unresolved issues in comparing PT(1) with PT(2) .Procedure for “cognitive” and “tutoring mode” discourse23232425252526273232354352Work Study techniques 52Discourse Analysis techniquesDetailed Flow Process123355669analysis111111141618181819Results of comparative analysisCHAPTER 1: DESCRIPTION OF “OUR WORLD” UNiTIntroductionBuilding of Our World HyperCard stackPeer tutoring of Our World stackHistory.SourcesTeachers’ collaborationThe Our World unitCooperative teaching .The “Our World” unitObjectivesComponentsThe studentsMethodology55VCHAPTER 2A: SELECTIVE REVffiW OF LiTERATURE (1)INTRODUCTIONSecond language acquisition 59Academic discourse 60Peer tutoring 61The computer in education 61PERSPECTIVES ON SECOND LANGUAGE ACQUISITIONKrashen’s (1985) Input Hypothesis 63Krashen’s five hypotheses 63Claimed evidence for the Input Hypothesis 64Criticisms of the Input Hypothesis 66Cummins (1991) and academic language proficiency 67Conversational and academic language proficiency 68The interdependence hypothesis 69The threshold hypothesis 70Operationalizing the concept of academic discourse 70Hallidaytsfunctional model of language 71Schieffelin and Ochs and language socialization 74ACADEMIc DISCOURSEHistory of research neglect of cognitive observational studies 78Neglect of the Smith, Meux studies 78The behavioural research stream 79The linguistic research stream 80Research studies citing Smith, Meux 81Summary 91Concepts for approaching the cognitive aspects of teaching 92The importance of the cognitive aspects of teaching 92A rationality model of teaching 95Hirst and epistemological structures 97Smith, Meux and the Logic of Teaching 99Smith, Meux, Coombs and the Strategies of Teaching 102Mohan’s Knowledge Framework 105Knowledge Framework compared with Smith, Meux structures . .. 109Knowledge Framework compared with other cognitive structures . . 116Summary and conclusions 120The methodology of discourse analysis 124Definitions 124Sinclair and Coulthard’s linguistic discourse analysis 124Systematic classroom observation 125Cognitive and pragmatic discourse analysis 126The typical I-R-E structure of lessons 126Discourse language functions 128viCHAPTER 2B: SELECTIVE REVIEW OF LITERATURE (2)PEER TUTORINGHistoryComponents for successful peer tutoringEvaluation of peer tutoringBenefitsEffectiveness researchPeer tutoring and second language acquisitionPeer tutoring issuesPoliticalPsychologicalAdministrativeTASK-BASED LANGUAGE TEACHINGConstructing a task syllabusVacuum between syllabus and methodSteps to construct a task syllabusComparison with the “Natural Approach”Task defined in the context of syllabusThe procedural syllabusThe process syllabusThe task-based language teaching syllabusOther definitions of taskSummaryPsycholinguistic and sociocultural approaches to taskA psycholinguistic approachA sociocultural approachTask related to concepts of practice and theoryTheory and practice dichotomizedTheory and practice reintegratedTask and the language of practice and theoryCummins’ continua of communicationMohan’s Knowledge FrameworkSociocultural tasks related to action and theoryCoMPuTER-AssIsTED LANGUAGE LEARNING (CALL)Changing directions in CAl and CALL researchCAl and CALL researchMove towards experimental and ethnographicComputer as facilitator of social (group) interactionComputers and organization of learning .Machine as humanizerSocial factors in CALLComputer as dependent variable130131132132133140141141142142143143144145146146147148148149149150152154154155157157159159160162161162163167168169studiesWhat sort of language does computer use foster7 171viiComputer-based group work and cooperative learningStudents as computer expertsSocial learning from writing with computersComputer as reorganizer of cognitive processesPea: Cognitive reorganization vs amplification .Hawkins & Sheingold: LOGO and databaseHooper: Integration of language and contentMETHODOLOGYTask designCommunicative task designWork Study (Industrial Engineering)Industrial engineering and metaphors of organizationWork study origin and featuresWork study methodologyImplementationCHAPTER SUMMARY173174175176177178180182182185186195197202205208208209212213214214215216218220220220222224224225226228228230233234235237CHAPTER 2C: SELECTIVE REVIEW OF LITERATURE (3)THE COMPUTER IN EDUCATIONTeaching style and classroom technologyKnowledge transmissionKnowledge constructionAssisted performanceTwo false assumptionsTechnologies of communicationWritingPrintingThe revolution of electronic informationWhy do schools resist technological change?Computers and new patterns of thinkingPrint vs computer mediated thinkingThe computer as reorganizer of mental functioning .Computers as facilitators of collaborative learning .Inappropriate uses of technologyUsing new technology like the oldComputer assisted instruction (CAD)Using new technology to preserve old curriculaAppropriate uses of technologyThe CSILE projectLearning networksMultimedia (hypermedia)The InternetCognitive effects of hypermediaThe futureviiiCHAPTER 3: STUDENT ACTIVITY SAMPLING RESULTSScheme of presentationLevels of aggregationSets of activitiesTotal students for all periodsBy student work modeBy student communication modeTotal students for each periodComputer LabLibraryHomeroomEach group for all periodsBy student work modeBy student communication modeEach student by group for all periodsBy work modeStudents’ evaluation of group workStudents’ evaluation of working environmentsCHAPTER 4: PEER TUTORING (1) RESULTSPEDAGoGIcAL ANALYSISTasks and objectivesThe discourse taskThe pedagogical objectivesThe studentsTranscript conventionsTutor/tutee pairsQualitative analysis of discourseAnalysis by pedagogical objectivesAnalysis by tutors’ teaching style .Qualitative summaryQuANTITATIvE ANALYSISMethodologyLength of turn by # utterancesLength of utterances by # wordsWords per minuteMean length of utterance (MLU)241241242243243246249249252254257257260262262268270276276277277277278278278291298300301302304306Quantitative summary 307CHAPTER 5: PEER TUTORiNG (2) RESULTSINTRODUCTIONReview of peer tutoring (1)The tutors’ teaching styleEvaluationAn alternative approachNew tutoring sessionsTutors’ trainingThe studentsTranscript conventions and interpretationPEDAGoGIcAL ANALYSISQualitative analysis of discourse by pedagogical objectiveObjective 1: Understanding the visual structuresObjectives 2 & 3: Production of language to matchObjective 4: Courteous treatment of tuteesLength of turn by # utterancesLength of utterance by # wordsWords per minuteMean length of utterance (MLU)Quantitative summaryCHAPTER 6: COMPARISON OF PT(1) ANDIntroductionDifferent settingsPositive aspectsPurpose of comparisonOverall quantitative comparisonPT(2) RESULTSProblems of quantitative comparabilityAnalytical strategyQuantitative comparison of selected tutor/tutee pairsOverall qualitative comparisonPersistence of I-R-EAnalytical strategySummaryFocus on academic discourse309309311314314318322323324324structures . . . . 333341348362362362363364368370370373373375375377ixQualitative summaryObjective 5: Moving away from the “Questioning Mode” 344QuAN’rrrATIvE ANALYSISBackground 351Methodology 351Tutor/tutee comparability 351Hypothesized results of PT( 1) and PT(2) comparison 352352356357359360xCHAPTER 7: COGNITIVE ANALYSIS OF PT(1) AND PT(2)Introduction 379Cummins’ model of communicative activity 379Mohan’s Knowledge Framework model 380Methodology 382Tutor/tutee selection for comparison 382Macro analysis by Venture 383Micro analysis by language function 383Tutoring modes 387Selected features 388Cognitive analysis results 388Analysis by language function 389Analysis by language function and tutoring mode 390A special look at “Hands-On” 394Analysis of tutor/tutee % total words by tutoring mode 396Analysis by selected features and tutoring mode 399Examples of reasoning discourse 403Summary of results 409CHAPTER 8: DISCUSSION AND CONCLUSIONSConclusion 412The research question answered 412Summary 412Peer Tutoring 1 412Peer Tutoring 2 417Discussion 422The issues 422Strategy for lessening discourse domination by tutors 423Traditional teaching paradigm grounded in technology 426Role of new technologies in changing teaching paradigm 428“Our World” unit a window to the future 431BIBLIOGRAPHY 434APPENDIX 1 “Our World” Sample Q&A Cards 451APPENDIX 2 “Our World” Tutor’s Checklist 457APPENDIX 3 “Our World” Classification Work Sheet 458APPENDIX 4 “Our World” Sample Topic Cards 459APPENDIX 5 “Our World” Sample Tutorial Cards 470xiLIST OF TABLESOver. 1 Language Functions and Peer Tutoring Modes 191.1 Our World Topics 241.2 Student data and profiles 442.1 Smith, Meux Logical Operations 1002.2 Smith, Meux Strategies of Teaching: Ventures, Objectives and Moves . . . 1043.1 Activity Sampling: Accuracy limits for Groups by Work Mode 2573.2 Activity Sampling: Accuracy limits for Groups by Communication Mode . 2603.3 Activity Sampling: Accuracy limits for Students by Work Mode 2634.1 Transcript conventions 2776.1 Data for comparative analysis of total PT( 1) and PT(2) tutor/tutee pairs . . 3646.2 Data for comparative analysis of selected PT( 1) and PT(2) tutor/tutee pairs 3706.3 PT(1) and PT(2) Pedagogical Objectives 3777.1 Ventures related to Knowledge Structures 3837.2 Halliday/Staab Language Functions 3847.3 Analysis of a Venture 386xiiLIST OF FIGURESIntro. 1 Three views of language learning 2Over. 1 “Our World” Outline Flow-Process Chart 11Over.2 “Our World” Flow-Process Chart with Inputs/Outputs 121.1 Model of Cooperative Teaching 301.2 Schematic of Our World HyperCard Plan 382.1 The Knowledge Framework (Mohan) 1062.2 Main types cognitive structure in ethnographic semantic analysis (Mohan) . 1082.3 Mapping of Logical Operations onto Knowledge Structures (KSs) 1092.4 Mapping of Ventures onto KSs 1122.5 Mapping of Meyer’s (1985) text structures onto KSs (Mohan) 1172.6 Langer and Applebee text structure schematic (Mohan) 1182.7 Mapping of Langer and Appletree content nodes onto KSs (Mohan) 1182.8 Mapping of Martin and Rothery’s genres onto KSs (Mohan) 1192.9 Mapping of graphic conventions onto KSs (Mohan) 1202.10 Contextual support & cognitive involvement in communication (Cummins) 1572.11 The Knowledge Framework (Mohan) 1592.12 Knowledge Transmission 2082.13 Knowledge Construction 211Our World Unit, Activity Sampling:3.1 Total Student time by Learning Environment and Work Mode 2433.2 Computer Lab: Breakout of Student Task Time 2453.3 Library: Breakout of Student Task Time 2453.4 Student Talk by Learning Environment 2473.4a Correlation of Student Talk with Group/Peer and Share Time 2483.5 Computer Lab: Student Time by Lesson Period and Work Mode 2503.6 Library: Student Time by Lesson Period and Work Mode 2533.7 Homeroom: Student Time by Lesson Period and Work Mode 2553.8 Computer Lab: Student Time by Group and Work Mode 2583.9 Library: Student Time by Group and Work Mode 2593.10 . 2613.11 . 2613.12 . 2643.13 2643.14 2684.1 3014.2 3034.3 3054.4 3065.1 3535.2 3565.3 3585.4 3596.1 3666.2 3717.1 . . . . 3797.2 . . . . 3817.3 . . .. 3897.4 . . .. 3917.5 .. . . 3957.6 . . . . 3977.7 . . . . 3987.8 . . .. 399xiiiLibrary: Student Time by Group and Communication ModeComputer Lab: Student Time by Group and Communication ModeComputer Lab: Individual Student Time by Group and Work ModeLibrary: Individual Student Time by Group and Work ModeSolo and Group Seating ArrangementsPeer Tutoring (1) Linguistic Analysis:Length of Turns by # UtterancesLength of Utterances by # WordsWords per MinuteMean Length of Utterances (MLU)Peer Tutoring (2) Linguistic AnalysisLength of Turn by # UtterancesLength of Utterances by # WordsWords per MinuteMean Length of UtterancePeer Tutoring (1) & (2): Comparative Linguistic AnalysisSelected Measurements: Totals All PairsSelected Measurements: Gladys and her TuteesPeer Tutoring (1) & (2): Comparative Cognitive AnalysisCummins’ Model of Communicative ActivityMohan’s Knowledge FrameworkTutor/Tutee Analysis by Language FunctionTutor/Tutee Analysis by Language Function and Tutoring Mode . . .PT(1) Analysis of Hands-On Tutoring Mode by Language FunctionTutor/Tutee Analysis by Tutoring ModeTutor/Tutee Analysis by Language Function and Tutoring Mode. .Tutor/Tutee Analysis by Selected Features and Tutoring ModexivACKNOWLEDGEMENTSThere is not space enough to list all who have helped me over my past 5-year apprenticeship in educational research . A few, however, I must name for special help along the way:Dr. Deborah Court who encouraged me not to “drop out” after my Master’s programand positively insisted that I see Dr. Leroi Daniels to whom she introduced me.Dr. Leroi Daniels who as Director of EDCI accepted me into the program, gave megood counsel and smoothed my path through the academic labyrinths of UBC.Dr. Walt Werner who wonderfully concentrated the minds of all us in the doctoralseminar and patiently led us from planning to research the universe of education to settlingfor something less with a sharp, narrow focus.My committee of three: Dr. Jerry Coombs (Chairman) who gave me a needed philosophical perspective and contributed his personal recollections of working on the Smith, Meuxteam, whose work was central to my thesis; Dr Rita Watson who was available to me at alltimes, including by e-mail during her sabbatical; and Dr. Bernard Mohan, my thesis advisor.The three teachers whom I call Karen, Jim and Harry to preserve their anonymity whowere extraordinarily generous in opening their classrooms to me over a 7-month period and ingiving of their time in many interviews and discussions. My thanks also go to their livelyclass of students, also identified by pseudonyms, who were welcoming and cooperative.I wish to thank the Social Sciences and Humanities Research Council of Canada(SSHRCC) for support from the grant #484890018 to Dr. B. Mohan entitled “Computers,language proficiency and academic achievement of immigrant ESL students”.I must return to one name, Bernie Mohan. I felt Bernie as much my collaborator asadvisor. He gave freely and generously of his time and skillfully acted as the catalyst toprecipitate the ideas which became central to my thesis. Above all, I appreciated Bernie’sthoughtfulness for me and all his students and the soundness of his counsel.1INTRODUCTIONThe research questionIs peer tutoring, using a computer-based hypermedia resource, an effective means ofhelping ESL students generate academic discourse? We explore this question through a casestudy of a science unit called “Our World”, a study of Earth in the context of the Solar Systemand of Earth’s physical features, conducted by the more experienced half (10) of an ungradedelementary ESL class. The students, divided into four groups, researched the subject matter(content) in the school library, organized content in the school computer lab with MacintoshHyperCard 1 multimedia authoring software, and peer-tutored the content in the resultingHyperCard stack to the less experienced half of the class. The homeroom teacher, the librarianand the teaching computer coordinator together planned and taught the unit:.Our research question links a number of research areas: second language acquisition(SLA), academic discourse, peer tutoring, task-based learning, and technology in education,particularly the use of computers and multimedia. In Chapter 2 we selectively review theresearch literature for these areas as relevant to our case study. Although our study looksprimarily at only 10 students, we believe that the way they work and the tools they use offer awindow through which we can perceive how the new electronic technologies, as they overtakeHyperCard is a multimedia (or hypermedia) authoring program for the Macintosh computer.Unlike a word processing program, which is restricted to text and some graphics, the HyperCardprogram integrates text, graphics, sound, animation and embedded computer programs in aproduct which on screen looks like an index card. A number of cards comprise a stack. Unliketext, which is read serially, each card contains “buttons” which, when clicked, branch directly toother cards in the stack. A stack is therefore multiply cross referenced giving users a choice innavigation:- serially, randomly (browsing), or to a predetermined plan.2the long-standing technology of print, may facilitate a change in the classroom teachingparadigm from knowledge transmission to knowledge construction.SLA perspectivesThree views on second language acquisition are relevant to this study (Figure Intro. 1):1. Krashen’s (1983) Input Hypothesis: The learner acquires the second language (L2)code through exposure to comprehensible input, i.e. L2 input at no more than one stage above thelearner’s current stage of grammar acquisition. As the learner progresses through the L2 learningstages, so the complexity range of comprehensible input increases.Figure Intro.1: Three views of language learning2. Cummin’s (1984) distinction between conversational and academic languageproficiency: On a continuum from “cognitively demanding” to “cognitively undemanding”,3academic discourse is in varying degrees “demanding” and conversational discourse is in varyingdegrees “undemanding”. The second language learner’s success in mastering “content” and soprogressing satisfactorily through the educational system depends upon acquiring the moredifficult academic language skills rather than the more quickly acquired and less difficultconversational language skills3. Halliday’s (1975, 1985) language-socialization model of language learning: Language,perceived as a sociosemantic system for making meanings, is both a medium of learning and aninstrument of socialization within specific contexts of situation and culture. Thus secondlanguage learning is seen in terms of acquiring the L2 code not in isolation but as an integral partof learning content and the L2 culture. Halliday’s model provides a system for a holistic analysisof discourse that provides insights into ways discourse contains ideational, interpersonal andtextual components (“textual” relating particularly in this case study to use of the computer). Italso links with the Cummin’s issue of the difference between academic and conversationaldiscourse.Significance of Academic DiscourseWe use “academic discourse” to mean “the language of teaching and learning” as distinctfrom the language of everyday social intercourse. Cummins and Halliday both make thisdistinction.CumminsCummins (1991) distinguishes between conversational and academic language proficiency. He has discussed this distinction previously (1984:136-7) in terms of “cognitive academiclanguage proficiency” (CALP) and “basic interpersonal communicative skills” (BICS):4(BICS is) defined in terms of the manifestation of language proficiency ineveryday communicative contexts whereas CALP (is) conceptualized interms of the manipulation of language in decontextualized academicsituations.He draws out the significance of this interpersonal/academic distinction:The major points embodied in the BICS/CALP distinction are that someheretofore neglected aspects of language proficiency are considerablymore relevant for students’ cognitive and academic progress than are thesurface manifestations of proficiency frequently focused on by educators.In other words, coping well with social situations is not an indicator that students (andparticularly L2 students) will be prepared to cope with the academic discourse of classroom andlecture hail as their education proceeds.Mohan (1989:100) points out the negative impact of low CALP on students’ performancein the “content” subjects in their education:While the exact definition of these (BICS/CALP) is controversial, there isevidence that while immigrant students may acquire basic interpersonalcommunication fluency in 1-2 years in the host country, they may requireas much as 4-8 years for academic language proficiency, during whichtime subject area achievement may be considerably depressed.In Cummins’ terms, the academic future of these ESL students -- completing high school,readiness for university-- depends on their cognitive academic language proficiency, i.e. theircompetence to understand and produce academic discourse in the classroom and its writtencounterpart, expository writing. We will adopt the term “academic discourse” as a convenientlabel for an important issue while noting that Cummins’ use of the term does not escapecompletely from the criticisms raised about his previous term “cognitive academic language5proficiency” (CALP). Our prime example of academic discourse will be the language of teachingand learning as it occurs in peer tutoring.HallidayHalliday (1985: xiii) distinguishes between the two main kinds of meaning in alllanguages: the “ideational” and the “interpersonal”:These components, called “metafunctions” in the terminology of thepresent theory, are the manifestations in the linguistic system of the twovery general purposes which underlie all uses of language: (I) tounderstand the environment (ideational), and (ii) to act on the others in it(interpersonal).Wherever it takes place, ideational component embraces “the whole of the transitivitysystem in language -- the interpretation and expression in language of the different types ofprocess of the external world, including material, mental and abstract processes of every kind...“ (Halliday, 1973:39). The “content” subjects in the school curriculum are primarily concernedwith the world and its processes (including the social world). We may, therefore, expect theideational component of discourse to be of special interest in the classroom where those subjectsare taught.Halliday thus offers an approach to discourse analysis which can be used to illuminate theprocesses of peer teaching.Peer TutoringPeer tutoring has been widely claimed to be effective in first language content learningand developing cognitive skills (Goodlad 1979; Gerber & Kauffman 1981; Sharpley & Sharpley1981; Topping 1988; Johnson & Johnson 1983; Bloom, 1984; Goodlad & Hirst 1990; Johnson,6Johnson & Holubek 1991). Researchers like Flanigan (1991:141) have also claimed its value inpromoting simultaneously second language discourse and content understanding. Flanigan notes:that recent studies of interactive learning in the SL classroom haveemphasized teacher-student and student-student discourse as a means ofbreaking the tradition of teacher-fronted one-way instruction.Taking Cummin’s perspective, it is therefore natural to ask whether peer tutoring can be valuablefor developing academic language proficiency.In our case study, student discourse resulted from cooperative work in the library andcomputer lab and from peer tutoring at the end of the Our World unit. We measured the amountof student talk as a percentage of total time in these activities and in teacher fronted whole classinstruction to determine the effectiveness of peer tutoring in generating student talk comparedwith the other classroom environments. We then used discourse analysis to focus on the type andquality of language generated from the peer tutoring activity.Research methodologyWe used two major research methodologies:1. discourse analysis for the recording, analysis and evaluation of peer tutoring discourse;2. work study for the recording, analysis and evaluation of student tasks and activities.Discourse AnalysisWe analyze the students’ peer-tutoring discourse under three separate headings:1. Quantitative. A quantitative analysis of volume of talk by the variables: word count, wordsper minute, mean length of utterance, turns by number of utterances, utterances by number ofwords. For this analysis we draw on the Systematic Analysis of Language Transcripts (SALT)7system (Miller & Chapman, 1984-199 1) from the University of Wisconsin-Madison LanguageAnalysis Laboratory.2. Cognitive/Tutoring Mode. A quantitative and qualitative analysis by:a. Halliday/Staab discourse functions:i. Interpersonal broken down by Social Needs and Controlling;ii. Ideational broken down by Informing and Reasoning;b. Tutoring Modes: Unstructured, Hands-On, Set Questions, Review, Tutor Explanation,Tutee Emulation, and Joint Construction.For cognitive analysis we draw on Staab (1986) who, following Halliday (1973),distinguishes between cognitive (ideational) functions and sociocultural (interpersonal)functions with further breakdowns shown above; and we draw on Mohan’s (1986) Knowledge Framework for cognitive analysis of computer graphics support, recognizing thatknowledge structures apply across modes of communication to graphics as well as text.For tutoring mode analysis we use our own categories, derived from inspection of thediscourse transcripts. We provide detailed definitions of cognitive discourse functions andtutoring modes when describing peer tutoring results in Chapters 4, 5 and 6.3. Pedagogic. A qualitative analysis by pedagogical objective for evidence of:a. students achieving the teachers’ objectives for the unit or lesson;b. interpersonal dynamics between tutors and tutees.For this analysis we draw on general concepts of classroom discourse (Mehan 1979; Cazden1988), and on Mehan’s identification of the Initiation-Response-Evaluation (I-R-E) style of8classroom discourse, which Cazden (1988:53) calls the “default” pattern: “what happensunless deliberate action is taken to achieve some alternative”.Functional model of languageIn our “pedagogical” and “cognitive” analyses, we follow Halliday’s (1985) functionalmodel of language which requires an examination of the discourse situation in terms of field(content or subject matter) manifested in the ideational function of discourse, tenor (participants’ role relationships) manifested in the interpersonal function and mode (communicationchannels), represented by the textual function which shapes the other discourse structures. Modein the peer tutoring situation includes oral discourse supported by computer graphics, recordedvoice, and on-screen text (i.e. “multimedia”), as well as printed worksheets and, in a hands-oncontext, practical demonstration of and experimentation with physical operations. The ideationaland interpersonal components are not present in the text as discrete blocks but rather asinterwoven strands bound together by the textual elements and expressed through the students’words, actions and behaviour.Taken as a whole, our discourse analysis covers a range of discourse aspects: ournarrowly defined “quantitative” analysis; our more broadly defined “cognitive” analysis; and ourqualitative “pedagogical” analysis which evaluates students’ words and actions to provide notonly an analysis by pedagogical objectives but also a qualitative framework for interpreting theother discourse aspects.9Work Study concepts and techniquesOur case study is an example of task based instruction. The students had tasks to performthat were not simply classroom exercises: they had to produce a final end product in the form ofa HyperCard stack that would be a resource for teaching other students what they had learned;they had to organize and produce hard copy of their stack in order to publish a printed booklet forthe school’s Visitors’ Day; and they had to perform necessary intermediate tasks of planning,research, note taking, organization of information, and computer/HyperCard operations in orderto deliver their end products. Finally, they had the task of tutoring their less experiencedclassmates who knew nothing about the Our World stack. Their tasks were, in principle, the sameas “real-world” tasks for producing teaching materials for an institution or business organization.Because of this strong natural task orientation we elected to use Work Studymethodology to record and analyze student activities rather than classroom observation systemssuch as the Flanders System of Interaction Analysis, which are more oriented to whole classteaching. Simon and Boyer (1975:92) characterize Flander’s as “the most widely known andused classroom observation system” in their anthology of 99 classroom observation instrumentsto which we refer in our review of literature in chapter 2A.Work Study, a British term, falls under the discipline of industrial engineering in theUnited States. It represents a rich research tradition developed within business and industry forthe study of real-world tasks in context in order to improve products and services, and increaseproductivity. It has much in coimnon with educational case study research. In particular, we drawon the following techniques:10Activity Sampling. We use systematic activity sampling (SyAS) based on anapproximately two minute interval (Haines 1958), derived from simple random sampling (SRS)(Tippett 1935), to answer the question “What is going on in the classroom/computer lab/library”?The answer is expressed as percentages of time spent on student activities, e.g. at one level:group work, solo work, “teacher fronted” whole-class work; and at another level: speaking,reading, writing.Recording. We use a variety of data gathering techniques including: the quantitativetechnique of activity sampling (already described); techniques to record the nature and sequenceof task work such as operation flow process charting; and the qualitative techniques ofparticipant and non-participant observation and interviewing;.Critical Examination. As well as general qualitative analysis, we use informally the workstudy technique, Critical Examination, to identify key operations in order to develop task designenhancements for the future. We do not report on suggested enhancements in this dissertation.Having described our purpose, basic research perspectives, and methodologies employed,we turn to an Overview of the project to briefly describe outcomes.11OVERVIEW“Our World” unit: work flowOutline Flow ProcessFigure Over 1 shows the three broadactivities in the Our World unit: library researchin which the students work in groups of two orthree in the library; HyperCard operations inwhich they work in groups of two or three in thecomputer lab, and peer tutoring in which they r’work in pairs: all pairs together in the computerlab or one pair at a time in the computercoordinator’s office. We analyze the discoursefrom the peer tutoring operation (activity) with a focus on academic discourse.Detailed Flow ProcessFlow process analysis is typically expressed in terms of inputs and outputs. Adopting thisas a convention, Figure Over.2 expands the Outline Flow-Process Chart in Figure Over. 1 to showinputs and outputs in greater detail and also to show knowledge structure linkages to operations.Each operation (activity) produces simultaneously two outputs: subject matter content andstudent discourse. We examine in particular the student discourse from the peer tutoring activityfrom which we develop “pedagogic” and “quantitative” analyses for all tutor/tutee pairs, and“cognitive/tutor mode” analysis for one representative pair.12The flow process shown is the one finally put into place as described later in this Sectionunder the heading “Second peer tutoring sessions”. It includes the use of the “Tutorial” stackcontaining knowledge-structure-based computer graphics as a tutoring aid.Figure Over.2: Flow Process Chart with Inputs/OutputsDescription of processResearch. The students working in groups draw on books, software and other material inthe library and produce as outputs 1) their written research notes and drawings and 2) oraldiscourse.13HyperCard operations. In the second dual operation, HyperCard and Graphics, they drawon their research notes and drawings and, working in groups on the computer, produce as outputs1) a HyperCard stack integrating text, graphics and sound and 2) oral discourse.Peer tutoring. In the third and final operation, Peer Tutoring, each tutor/tutee pair,working on the computer, draws on the Our World HyperCard stack and on the front endTutorial stack and produces as outputs 1) their negotiated and internalized subject matterunderstanding and 2) their negotiated and externalized oral discourse.Knowledge structures. Knowledge structures (KSs) (Mohan, 1986) are shown as a linkingdevice. They link with operations as process-- describing, classifying etc. -- and with outputs asproduct-- description, classification etc. In the research operation in the library, the students areencouraged to recognize the KSs implicit in their library source material, particular the KSs ofdescription, classification and sequence, and to use KSs as an organizing device as they preparetheir notes and charts. The KSs as product in the source material accordingly tend to becomeembedded as product in the students’ notes and charts. Then in the computer lab the studentsdraw on their library notes and charts and again use KSs as an organizing device as they build the“Our World” HyperCard stack. The KSs as product in the students’ notes and drawingsaccordingly tend to become embedded as product in the “Our World” stack. Finally, the studentsdraw on 1) the “Our World” stack, where KSs are implicitly embedded, and 2) on the “Tutorial”stack where, through key visuals, KSs for each topic are made explicit, and negotiate anunderstanding of the subject matter. This negotiated content is the final and intangible endproduct of the peer tutoring operation, present as understanding in the students’ minds.. Because14the discourses are recorded, the KSs as product can be identified in the tangible end product ofthe transcripts.The case study falls into three parts:Part 1 which includes: a) the construction of the Our World HyperCard stack; b) the peertutoring sessions, PT(1), at the end of the unit and of the school year; and c) thequantitative and qualitative discourse analyses of the peer tutoring sessions.Part 2 which includes a) new peer tutoring sessions, PT(2), at the beginning of the newschool year with new tutees and the addition of a “Tutorialt’stack; and b) the quantitativeand qualitative discourse analyses of the peer tutoring sessions.Part 3 which includes a cognitive discourse analysis by Halliday/Staab languagefunctions and by peer tutoring modes, applied to two selected and comparable tutor/tuteepairs: one from the first peer tutoring sessions, PT(1), and the other from the second peertutoring sessions, PT(2).Part 1: Construction of Our World stack and first Peer Tutoring sessions (PT1)We conducted a 3½ month ecological study of the more experienced half of the class (tenstudents) who worked in four groups to produce the “Our World” stack and, in tutor/tutee pairs topeer-tutor its content to the less experienced half of the class. The teachers had decided on peertutoring as an experiment at one of their planning meetings after teaching the unit had started.The objective was twofold: 1) to see if peer tutoring would be an effective means for the tutors toconsolidate their understanding of the unit’s content; and 2) to let the tutors share the results oftheir work with the other half of the class. Results were as follows:151. Peer tutoring proved effective in producing student talk (Krashen, 1985; Long & Porter,1985) in quantity: activity sampling showed 56% of total time was student talk in peer-tutoring compared with 10-16% in other activities in homeroom, library and computer lab.2. A quantitative analysis of peer tutoring discourses showed that the tutors benefited more thanthe tutees. They produced relative to the tutees:• two-and-a half times as many words per minute (a measure of total student talk);• more than three times as many multi-utterance turns;• more than three times as many utterances of three words or more;• one-and a-half times the mean length of utterance (MLU).3. Peer tutoring appeared effective in achieving the teacher& pedagogic objectives for the OurWorld unit. A qualitative “pedagogical analysis” of the discourse transcripts offered evidencethat:a. The tutees gained an appreciation for the content of the Our World stack, shown by theirsatisfactory completion of a written evaluation booklet after the peer tutoring sessions.b. The tutees learned HyperCard operations and showed by the end of the tutorial sessionsthat they could navigate the Our World stack with minimal assistance from the tutors.c. The tutors consolidated their understanding of that part of the content they had worked onand became familiar with the content they had not worked on. This was shown by theirquestions and discussions with each other and their teachers, as recorded in the discoursetranscripts.4. Additional qualitative analysis showed that:16a. The teaching style most commonly used by the tutors was their version of the traditional“Initiation-Response-Evaluation” (I-R-E) style (Cazden, 1988) to which they had beensocialized (Halliday, 1975) in their previous classroom experience. They initiated (I)questions (thereby maintaining control) to which they expected to receive memorizedresponses (R) which, in turn, they evaluated (E) as right or wrong.b. The tutors’ authoritarian control of the discourse in their question-driven approachintroduced some unproductive interpersonal tensions between tutors and tutees.c. A notable exception to the use of the I-R-E teaching style in Part 1 was discourseassociated with hands-on instruction in the form of learning and applying the HyperCardcommands with mouse and keyboard. This all took place in Part 1. Although the tutorsspoke more than the tutees in these sequences as they did in the I-R-E sequences, thedialogue between tutor and tutee was more spontaneous and varied and the tuteesexperienced the satisfaction of “getting it right” by direct feedback from the computerscreen.Part 1 ended with the end of the school year in June, 1992.Part 2: Second Peer Tutoring sessions, PT(2)Although Part 1 realized many benefits, the teachers were not satisfied with the quality ofdiscourse produced by the tutors’ question-driven approach. To help future tutors move awayfrom this “default” teaching style (Cazden, 1988), the computer coordinator, in preparation forthe new school year, developed a “Tutorial” stack as a front end to the “Our World” stack. TheTutorial stack was designed to provide students with contextual support in the form of computergraphics (key visuals) to represent the knowledge structures (classification, description, sequence17etc.) inherent in each topic. The tutors were shown how to use the key visuals and practised bytutoring each other. Over a one month period at the beginning of the new school year I observedand recorded new tutoring sessions on four of the original tutors with new tutees. Results were asfollows:1. The same high level of student talk as achieved in Part 1 was maintained in Part 2.2. The tutors moved away from one-way, question-driven “transssion teaching towardsexpository discourse based on the key visuals.3. The tutors encouraged the tutees to explain in their own words the topics which the tutors hadexplained and provided assistance, as required, with explanatory dialogue.4. Control was more equally shared between tutors and tutees.5. The tutors were more empathetic to the tutees’ needs and feelings.6. However, despite these positive qualitative changes:a. The tutors tended to fall back on the I-R-E teaching mode when the tutees had more thanminor difficulty in emulating the tutors’ explanations.b. The tutees showed only modest quantitative improvements relative to the tutors in Part 2over Part 1: up 7.5% on words per minute (the tutors still spoke more than twice as muchas the tutees overall); up 7% on mean length of utterance (MLU); and less than 1% up ontotal # utterances. The highest rate of increase was 17% on 5+word utterances.18Part 3: Comparative analysis of selected pairs in PT(1) and PT(2)Unresolved issues in comparing PT( 1 with PT(2)At this point we recognized three problems:1. The tutor/tutee pairings in Part 2 were not directly comparable to those in Part 1 because thetutees’ profiles in Part 2 differed significantly from Part 1.2. Our pedagogic and SALT quantitative analyses of discourse had not addressed the questionof measuring academic discourse quality. Therefore we did not have objective evidence forour judgement, based on observing the tutoring sessions and reading the dis-coursetranscripts, that the academic discourse quality was better in Part 2 than in Part 1.3. Also, a careful reading of the discourse transcripts indicated that the tutors varied theirtutoring strategies in the course of any one session. Therefore our quantitative analyses ofeach whole discourse could conceal potentially important differences among the differenttutoring strategies used, a consideration that would also apply to the cognitive analyses thatwe were about to develop.Procedure for “cognitive” and “tutoring mode” discourse analysisWe therefore took the following steps to prepare for a comparative cognitive analysis ofPart 1 and Part 2 in total and by tutoring mode:1. We based our “cognitive” analysis on Halliday/Staab language functions and our “tutoringmode” analysis on our own determination of peer tutoring modes, as shown in the followingTable:19Table Over.1: Language Functions and Peer Tutoring ModesLanguage Functions Peer Tutoring ModesHalliday Staab Part 1 Part 2Interpersonal Social Needs Unstructured Tutor ExplanationControlling Hands-On Tutee EmulationSet Questions Joint ConstructionIdeational Informing Review I-R-EReasoning2. We identified two tutor/tutee pairs, one from Part 1 and one from Part 2, that couldreasonably be compared because the tutor was the same in both pairs and the two differenttutees had similar profiles.3. We carried out a cognitive discourse analysis in total and by tutoring mode on the twoselected tutor/tutee pairs.4. We also included in this analysis a count of a) exophoric references because they areparticularly to be expected in computer based discourse, the computer screen being the pointof reference; and b) I-R-E utterances because we wished to see if the tutors succeeded inmoving away from this teaching style in PT(2).Results of comparative analysisCognitive Analysis by Tutoring Mode (see charts in Chapter 7)1. Peer Tutoring proved equally effective in generating Informing/Reasoning discourse in PT(1)(87%) as in PT(2) (86%).202. However, the split of Ideational between Informing and Reasoning changed significantlybetween PT(1) and PT(2): Reasoning increased from 22% in PT(l) to 39% in PT(2) whileInforming correspondingly decreased from 65% in PT(l) to 47% in PT(2).a. A key factor in the Reasoning gain in PT(2) appeared to be the ability of both tutors andtutees to raise the quality of their academic discourse by replicating in it the knowledgestructures embedded in the computer graphics.3. In PT( 1), the peer tutoring mode least productive of Reasoning discourse is “Unstructured”(9%), in which tutor and tutee browsed through the stack without a pre-set agenda. Thiscompares with 27% for the Set Questions and Review modes, both of which are highlystructured.4. In Part 2, the peer tutoring mode least productive of Reasoning was Tutee Emulation 34%).This was higher than the modes most productive of Reasoning in Part 1, Set Questions andReview (both 27%).5. Reasoning in Part 2 further increased from 34% in the Tutee Explanation mode to 38% inJoint Construction and 46% in Tutor Explanation.Hands-OnThe “hands-on” work in this unit was learning and applying the HyperCard commandswith mouse and keyboard. All of this occurred in Part 1. Therefore, as we were not concernedwith comparing Hands-On with Part 2 but rather with determining its functional profile, wecould use data from all the tutor/tutee pairs in Part 1. The results were: Informing, 53%,Reasoning, 20%, Controlling, 23% and Social Needs, 4%.211. Controlling (23%), a function shared between the tutor and tutee, was higher for Hands-Onthan for any other tutoring mode in either Part 1 or Part 2. Joint Construction in Part 2 wasclosest at 16%. Both cases were in the nature of co-operative ventures with both tutor andtutee proposing, counter-proposing and discussing how to proceed. Therefore a high score onControlling is to be expected.2. The score on Reasoning (20%) is perhaps surprisingly high as there is a tendency to think ofhands-on instruction as essentially the tutor telling (informing) the tutee physically what todo. Reasoning becomes necessary when the tutor explains why an operation is needed orwhy a wrong move has produced a certain result.Exophoric references1. Exophoric references were highest in the Hands-On mode (16%) because of the frequency ofpointing to screen or keyboard both in giving directions and asking questions.2. They were present but lower (3-5%) in all other modes in both Parts 1 and 2 except Review,which was zero.3. They were zero in Review because Review was carried out like a classroom exercise with noreference to the computer screen, an exophoric area of reference for the other modes.I-R-E1. I-R-E predominated in Part 1. A breakdown by tutoring mode showed that it was found in theSet Questions (76%) and Review (81%) modes but did not appear at all in the Hands-Onmode. This may be because I-R-E is a technique for transmitting textbook knowledge (the“facts”) and is irrelevant to teaching practical (“how to do”) operations.222. I-R-E appeared residually in Part 2. It did not appear in Tutor Explanation, was negligible inJoint Construction, but accounted for 34% of the Tutee Emulation discourse. This wasbecause the tutor would resort to the I-R-E default when the tutee got beyond a certain pointof difficulty in his or her explanations.3. I-R-E is associated with teacher direction. A qualitative examination of the discoursetranscripts showed that the tutor controlled the I-R-E sequences, starting from the first “I”question. This changed in Part 2. With the tutor in Tutor Explanation mode, the tutee wouldbreak in with a question or a challenge so that control was shared. In the Tutee Emulationmode, the tutee always started in control. Only beyond a certain point of difficulty would thetutee lose control -- and not regain it when the tutor reflexively reverted to I-R-E.4. Only when control is shared does the tutee begin to benefit in L2 usage to the same extent asthe tutor. Therefore the tutee will gain substantially less than the tutor in I-R-E sequences.In the next Chapter, we describe in some detail the history, participants and progress ofthe Our World unit to provide an adequate context for understanding the discourse analyses andresults that are described in later chapters.23CHAPTER 1: DESCRIPTION OF OUR WORLD UNITIntroductionThe Unit divides into two parts:1. ESL students’ library research and computer lab operations to produce aHyperCard stack called “Our World” on the physical features of Earth and its place inthe Solar System.2. Peer tutoring of the stack by more experienced to less experienced students.The peer tutoring part divides into two:a. Peer tutoring of the Our World stack at the end of the unit by the moreexperienced half of the class who had built it, to the less experienced half whohad worked on a different computer lab unit.b. Peer tutoring of the Our World stack three months later by some of thestudents who had built it to some younger ex-ESL students from other classes.On this occasion, the tutors had the advantage of a new Tutorial front endwhich had been provided for them to use with the Our World stack.Building of Our World HvperCard stackThe site was a large metropolitan elementary school with a high ESL enrolment. Thehome room teacher, “Karen Black”, divided her ungraded ESL class of twenty studentsbetween the ages of 10 and 13 into two on the basis of L2 experience. The more experiencedhalf of the class completed a science unit designed to introduce basic concepts about the solarsystem and planet Earth. The unit took the form of research in the library under the librarian,“Jim White”, and building a Macintosh HyperCard stack in the computer lab under the24computer lab coordinator, “Harry Green”. The ten students were divided into four groups.Each group chose one “overall topic” and at least two “physical features” from Table 1.1.Table 1.1: Our World TopicsOverall Topics j Physical Features of Earth ISolar System S Oceans Mountains IDayandNight I JI*’ IAtmosphere 2IIhOur Earth Vo1cØFThe students produced at least three HyperCards on each topic and prepared anIntroduction and Table of Contents. The Our World stack was published in two formats, bookand electronic, as exhibits for the school’s Visiting Day. Three additional topics, Lakes,Jungles and Soil were added to the stack after Visiting Day. Appendices 1 and 4 showsamples of the cards produced.Peer tutoring of Our World HyperCard stack1. Peer tutoring to students from same classAt the end of the unit, the half of the class who built the stack peer tutored itscontents, with an emphasis on the Overall Topics, to the other half of the class who had beenworking on a different unit.2. Peer tutoring to students from other classesBased on the results of the first peer tutoring sessions, Harry Green (the computercoordinator) built a supplementary Tutorial stack as an aid to future tutoring. Early in the nextschool year, Harry arranged for four of the original tutors, now mainstreamed from ESL to25Grade 6 or 7 classes, to use the new Tutorial stack with the Our World stack in specialtutoring sessions to younger Grade 5 ex-ESL students.HistorySourcesThe sources for this history are my observation records, tape recorded interviews, anddocumentation provided by the three teachers connected with the Our World unit: KarenBlack, the homeroom teacher; Jim White, the librarian; and Harry Green, the non-enrolling,teaching computer coordinator and English language support teacher. Documentation includeslesson plans, minutes of meetings, and samples of student products from the Our World unitand from previous units using a combination of homeroom, library and computer labresources. I attended the teachers’ joint planning meetings (the first was held on February 14)and regularly observed Our World periods in both library and computer lab.Teachers’ collaborationThe three teachers collaborated in producing the Our World unit in the school year199 1-92. They had previously worked cooperatively in the 1990-9 1 school year. Karen haddiscussed with Jim the idea of producing a unit whose output could be presented at thecommunity Arts and Letters Festival in the Spring of 1991. They settled on the Weather,which Karen was planning to teach for science in her ungraded ESL class, with the idea ofthe students producing and publishing a book. They asked Harry to teach the class HyperCardin the Macintosh computer lab so that HyperCard could be the producing medium. Thestudents researched their topic in the library. Jim, the Librarian, taught them how to takenotes and supervised their production of sentences, paragraphs and illustrative diagrams and26pictures. The students took their written materials and illustrations to the computer lab whereHarry taught them Macintosh HyperCard basics and how to enter their materials into aHyperCard stack (a computer file based on a card index metaphor with an emphasis on thecoordination of text, graphics and sound). The end product of the Weather unit was a bookletLet’s Find Out About the Weather, reproduced from printouts of the Weather HyperCardstack.The Our World UnitThe Weather unit became the model for the Our World unit in Karen’s new class forthe 199 1-92 school year. The ten more experienced students who worked on Our World hadbeen beginners the previous year and had not, except for one, worked on the Weather unit.They became familiar with HyperCard in their first project of the new year through a unit onCanada which Karen taught them in collaboration with Harry. Karen supervised theirproduction of content and Harry taught them HyperCard basics. Karen wanted to move on toOur World because she found that even her ten more experienced students (ages 11-13) knewvery little about the earth. They knew that it was round but could not say what caused dayand night, what were the effects of the sun and moon on the earth, and what were the broadphysical features of the earth. On a scale of one to ten, Karen rated even the highest of thisgroup at little beyond two in this regard. Therefore, in October, 1991, she spoke to Jim aboutteaching them research skills in the library with Our World as the focus and in early January,1992, asked Harry to do a full HyperCard unit with them as he had with last year’s class onthe Weather unit. All the instruction on Our World was done by Jim in the library and Harryin the computer lab. This left Karen free to work independently with her ten less experienced27students (seven, ages 9-11 and three, age 13). She worked with this group in the computer labat the same time as Harry worked with the Our World group and would occasionally help outwith Our World students’ questions as they worked at their computers.Cooperative teachingKaren, Jim and Harry worked together in two ways, joint planning and joint teaching,building upon their previous successful collaboration in the previous year. We will look athow they did this but first describe their teaching background.The teachers backgroundKaren. Karen had been ten years in teaching, all ESL, in several schools. This washer third year in the present school, which was the only school in her experience to offeraccess to collaborative team support. She identified several advantages to this approach:1. The ESL students rotate among different teachers and therefore get variedconversational experiences.2. They get used to the demands of English as the language of discourse where teachersare not ESL trained and do not slow down and simplify their speech as an ESLteacher does.3. The students benefit from the expertise of other teachers.4. When the class is ungraded, the ESL teacher can spend time more intensively with thebeginners, who need more individual attention, while the more experienced studentswork in the library, computer lab or elsewhere with other teachers.Karen enjoys working on a resource team and takes every opportunity to draw on the school’slibrary and computer lab services, represented by Jim and Harry.28j111. Jim had been teaching for just on 30 years in elementary schools, just over 20years as a teacher-librarian, and the last 10 years in the present school. What he particularlylikes about the teacher-librarian’s job is the opportunity it gives to know and work with thewhole school from K through 7. He also finds working with other teachers stimulating. Henoted that cooperative teaching in his experience had come to the fore only in the last tenyears. Before that the theory of the library being the “heart of the school” and actual practicewere very different matters. However, there has now been a shift and other teachers arebecoming accustomed to using the services of a librarian-teacher, at least in his presentschool. He collaborates with all the teachers in the school, although more intensively withsome than others. He works particularly closely with Harry who provides a similar teachersupport service in the computer lab.Harry. Harry had been teaching just over 12 years. He majored in English literatureand then did his Master’s in English literature and ESL methods. He joined the present schoolten years ago at the same time as Jim. Harry was interested in the potential uses of computersin education right from the advent of the personal computer. In his first school he used thesingle PET computer extensively and learned programming. In the present school he used theearly generation Commodore 64s. He was particularly interested in LOGO programming,simulation applications like Rocky’s Boots and graphics applications like Koala Pad, all ofwhich made it possible for students to play an active role in using the computer rather thanthe passive role dictated by traditional drill-and-practice computer-assisted instruction (CM).In his present school, Harry pushed for the installation of a Macintosh computer lab becausehe could see the potential of HyperCard (bundled with the “Mac”) to allow students not only29to use the computer interactively to construct and store knowledge but also to communicatethat knowledge in creative ways through the “knowledge navigation” of a computer fileintegrating text, graphics, animation and voice. Students, who grow up as passive consumersof TV, can learn how to use a TV-like medium actively as an authoring system to produce amulti-media electronic text with a greater potential to communicate than simply printed texton a page. Harry finds that the use of HyperCard both to author and communicate anelectronic text strongly motivates students to apply themselves and brings out their pride ofauthorship. He also feels that a dynamic HyperCard stack is more inviting to a viewingstudent, particularly in ESL, than paper products. Moreover, the students get a practicalintroduction in an elementary way to the multi-media concept which is now a dominatingtrend in the use of PCS in the workplace for producing training materials, catalogues, in-house communications and business presentations.Jim and Harry trained together on the Knowledge Framework (Mohan, 1986) in agovernment “Funds for Excellence” program and have found the representation of knowledgestructures in key visuals a positive factor in presenting content material. They haveinternalized Knowledge Framework concepts in the way they teach because they have foundthem effective both in lesson planning and in helping students understand content.Figure 1.1 presents a visual model of the cooperative teaching employed.30Figure 1.1: Model of Cooperative Teaching%%>: 4F!p- CLASSROOM ThACHER: .: >g::: :f:.::. -s •‘$\ A %%%. .....v ... . s \\‘ .; s•/‘PlannmgA /i ...W4PI . -.(IiIMr— UNIT ITeachmg jf : \‘. . . . . V.. • ... .4 i S I—\‘:%dr / IiI 44 V ‘VI COMPUTER-SUPPORT TEACIfr4 LIBRARIAN - TEACHER/Joint and md d*.i%on unitJoint PlanningKaren produced an outline of the unit as a basis for discussion first with Jim in thelibrary and later with Harry in the computer lab. On January 29, Harry gave Karen a formalplanning guide for setting out the overall objectives and requirements of the computer labcomponent of the unit. Karen completed this and on February 7 met with Harry to discussand expand it into greater detail. Harry designed a prototype HyperCard template and somesuggested language activities based on notes prepared by the students in their library sessions(obtained from Jim). On February 12 and 13, Harry spent time in the library with Jim and thestudents to see how much the students had completed and what they still had to do and to gettheir ideas on suitable questions and answers to include on the cards for the different topics.31On February 14, Harry, Jim and Karen met formally for the first time in Harry’soffice. I also was invited to this meeting and attended as an observer. It was agreed: to dividethe students’ topics into two groups, four overall topics and 9-12 topics on physical featuresof the earth (see Table 1.1 above); to have the students develop questions and answers onlyfor the overall topics because of time constraints; and to have them develop brief narrationsfor the more numerous topics on the physical features of earth. Karen then presented a newwritten proposal for peer tutoring sessions at the end of the unit so that the Our Worldstudents could share what they had done with the other half of the class and, in so doing,consolidate what they had learned. The first computer lab session was scheduled for February20. I was invited to attend this and also the two library sessions in that week in order todevelop a proposal for a case study.On February 26, I presented to Harry, Jim and Karen my proposal for making aformal case study of the Our World unit. They willingly agreed to let me attend all sessionsin the library and computer lab, conduct activity sampling, make audio and video recordingsof the students at work, and individually interview both teachers and students at convenienttimes. Karen also agreed to let me observe a series of lessons in the students’ homeroom (noton Our World as this was taught only in the library and computer lab) to get a comparisonbetween the homeroom and the library and computer lab environments.On May 22, Harry, Jim and Karen met formally for the last time. I attended as anobserver. Jim reported on the production of the printed book from the completed HyperCardstack. It was agreed that the students would continue to produce extra topic cards for Lakes,Jungles and Soil. Therefore the stack in its final form included more material than the book.32Plans for the peer tutoring sessions were discussed. Harry pointed out that this wouldintroduce the students to a new use of HyperCard. They had to this point used HyperCard toproduce and store knowledge. Now they would use it to navigate and present knowledge.Karen pointed out that another valuable result of the unit was the students’ acquisition oflibrary and research skills.Joint teachingIn the library. Jim was responsible for work in the library. Once Harry was involvedin the unit, he attended most of the library sessions, in which the students worked in groups.He and Jim circulated among the groups answering questions, making suggestions andgenerally facilitating the students’ work. The two teachers also took turns in conducting briefwhole class sessions to review what had been accomplished and to plan succeeding activities.In the computer lab. Harry was responsible for work in the computer lab. He workedwith the Our World students in one half of the lab while Karen worked with the rest of theclass on another unit in the other half. The Our World students worked both in their groupsand individually. When the Our World work became intensive, Karen would leave her half ofthe class (they had highly structured activities on which they could often work withoutextensive help) and assist the Our World groups or individuals who were asking for help.The Our World UnitObjectivesKaren presented a list of objectives at the joint planning meeting on February 14. Thefollowing were agreed as feasible within the time constraints of the school year, including ascontent knowledge the topics Jim had brainstormed with the students in the library:33Content knowledge about our world. Basic geographical and scientific facts:a. Earth as a globe and its position in the solar system.b. Physical characteristics of earth:i. land: the continents.ii. water: the oceans.iii. atmosphere: the various layers.iv. tides: caused by the moon.v. light and heat: from the sun as giver of life.c. The earth spinning on its axis, causing day and night.d. The earth moving around the sun, causing the four seasons.e. The earth’s surface: rivers, lakes, oceans, rocks, soil, caves, mountains,volcanoes, islands, forests, jungles, deserts.2. Language and related skills.a. Notetaking, sentence composition, oral presentation.b. Understanding the process of library research.c. Compiling and publishing a book.3 Computer skills. Using a Macintosh computer and HyperCard software to organize,write, illustrate and display content.A fourth and important set of objectives was added by Karen at the joint meeting withHarry and Jim on February 14:4. Sharing and learning through peer tutoring.a. Promotion of sharing and cooperative learning experiences.b. The tutors’ development of teaching strategies.34c. The tutees’ acquisition of vocabulary and knowledge about Our World.d. The tutors’ consolidation of their knowledge about Our World.Karen developed a Key Concepts Checklist (see Appendix 2) and attached it to the listof objectives together with an outline of the following steps for the tutorial sessions:Check that the tutors:a. have acquired sufficient background knowledge about Our World;b. know how to run and demonstrate the use of HyperCard stacks;c. know how to use the checklist for key concepts;d. know how to make and use evaluation booklets;(Karen will later provide evaluation questions to enter into a bookletwhich the tutors and tutees will make and complete in the homeroom).e. understand that they can use a mutual first language with their tutees2. Have each tutor, working with one tutee in the tutoring sessions:a. tour the HyperCard stacks with the tutee in order to look and enjoy;b. teach the tutee the use of HyperCard features and buttons;c. read out vocabulary and paragraphs;d. read out vocabulary and paragraphs a second time, along with the tutee;e. use the Key Concepts Checklist to record completion of the concepts taught.3. Have each tutor, after the tutoring sessions:a. make evaluation booklets jointly with his or her tutee;b. help the tutee to record the answers;c. illustrate the booklets jointly with the tutees.35ComponentsThe Our World unit therefore has three components: the first two, the library periodsand the computer lab periods, were based on the previous year’s Weather unit; the third, peertutoring, was new.The library periodsThe Our World students (the more experienced half of the class) spent one single andone double period in the library each week for the duration of the unit. The computer labperiods did not start until late February. From mid-October to mid-February, Jim worked withthe students in the library on selecting topics for study, introducing them to basic conceptsand teaching them research and notetaking skills.Whole class work. At first, Jim worked with the students as a whole. He found, asKaren had said, that they did not understand the causation of basic phenomena. He tookpractical steps to remedy this deficiency, e.g. by demonstrating the causation of day and nightwith the aid of the library globe and a flashlight to model the sun. He also introduced them toconcepts with the commercial HyperCard program Earthquest, which they viewed on thelibrary computer. He taught them how to write notes, using one phrase for each main idea. Hethen discussed their notes with them, put on the board the results of discussions and helpedthem individually to write out full sentences and paragraphs in their notebooks. He went on toshow them the “web” system of notetaking, diagrammatically grouping sub theme keywordnodes around a central theme node and joining the nodes with lines of relationship. This hadthe advantage of saving them from immediately writing out sentences. Whole class work alsoincluded brief introductions to and reviews of each session.36Group work. As soon as Jim had covered these basics, he asked the students to dividethemselves into four groups, two or three to a group. One student, Hetty, wanted to work onher own but Jim insisted she join a group. She did not work out with the first pair of girls shejoined. She moved to the other pair of girls where she remained for the rest of the unit. Asobservation was later to show, this was an early indication of Hetty’s problems workingcooperatively. Once the groups were formed, Jim divided the research topics amongst themon the basis of group choice with negotiation when necessary. He showed the students how tofind books on different topics by using the library’s computerized book catalogue, indexed byauthor, title and subject. The students were enthusiastic about the catalogue and were proud toknow how to use it. It was a good example of “empowerment”: they quickly learned how towrite down the call numbers and then find the corresponding books on the shelves. Thebooks they retrieved were left on a cart for their use for the duration of the unit. Jim alsotaught them how to use a book, taking advantage of the table of contents and index and“skimming” before reading in depth. As the students worked in their groups, Jim or Jim andHarry circulated among them, helping the students both in their groups and individually.Individual work. The students jointly researched their topics and agreed on subtopicsthat they would develop individually. Each individual was responsible for drafting paragraphsand illustrations for his or her own subtopics. This led to additional individual research,reading and notetaking although the students were free to discuss and consult with each other.Each subtopic would eventually become the subject of a card in the HyperCard stack in thecomputer lab.37Sharing time. A special feature of work in the library was “sharing time”. As soon asa group was ready with a topic, Jim gathered all the students into a circle, sitting on the floor,and had the students in the group share in reading out what they had got. Jim then invitedquestions and discussion. These sessions provided good opportunities for informal dialogue.HyperCard orientation. In mid-February, at about the time when the groups starteddeveloping their note “webs”, Harry came into the library sessions and started talking to thestudents about the HyperCard stacks they would shortly be producing. He used the Weatherunit as an example to show them what was required. They were able to relate well to what hesaid because of their previous experience doing the Canada stack (see History section above).The Computer Lab periodsKaren’s students started their computer lab sessions on February 20 and continued withone double period per week until the end of the school year. The Our World students, calledthe Experts, worked in one half of the lab under Harry. The other half of the class, called theBeginners, worked under Karen in the other half of the lab on two programs: Pictionary, inwhich they matched written words to pictures, and Typing Tutor, in which they followed agraduated series of keyboard tasks and timed exercises. They clearly enjoyed these programsand were quietly engrossed in them for most of each double period. For the first few weeks,the “experts” helped the “beginners” set up their computers with the required programs.The Our World HyperCard plan. Before the computer lab sessions started, Harrydesigned, in consultation with Karen, two HyperCard templates: one for Question and AnswerCards, the other for Picture and Paragraph Cards (see Appendices 1 and 4). These became thetwo basic formats which the students used to enter and edit their text, graphics and pictures38which they drafted in the library. The design of these templates typically required three drafts,sometimes more, with discussions and reviews before the final product was agreed. Harryplanned a HyperCard stack as shown schematically in Figure 1.2.Figure 1.2: Schematic of Our World HyperCard PlanThe stack is headed by a Menu which points in two directions: to the Question andAnswer cards and to the Picture and Paragraph (Topic) cards. Each student group will beresponsible for producing at least three cards on the themes and topics negotiated with them:one Question and Answer theme and two or three Picture and Paragraph topics per group.Harry will consolidate the productions from the four groups into a single stack of39approximately 12 Question and Answer Cards (four themes) and between 24 and 36 Pictureand Paragraph Cards (8-12 topics), making up 36-48 cards in total (The finally produced stackcontained 40 cards plus five for introduction and menus). All cards will be linked bothsequentially with each other and with the menu. Thus viewers can go through it sequentiallylike a slide show or at will to and from the menu (knowledge navigation). Each group willagree on the general layout of its cards and each student will be responsible for the finalpreparation of individual cards.Harry pointed out to me later that the students made suggestions of their own formodifications of the stack design as they came to understand the dynamics of the stackthrough using it. He helped them evaluate these suggestions and implement those that werepracticable. Many of the students’ ideas came from using commercial HyperCard applicationsin the library. As in the library, the students worked in the computer lab as a whole class, ingroups and individually.Whole class work. As in the library, whole class sessions included brief introductionsto goals and periodic reviews. The introductions included instruction or review of features ofHyperCard to be used in a session. Unique to the computer lab were whole class “lockstep”sessions in which Harry would call out the sequence of HyperCard instructions to be used fora given objective while the students implemented those steps on their individual computers.Harry would check that their screens corresponded to his as he also implemented the steps.These sequences therefore involved the students in practical action with immediate feedbackas they watched their screens for the effect of their instructions. Alternatively, if Harrythought the students familiar enough with a particular sequence of HyperCard instructions, he40would demonstrate requirements for the next phase of work to the whole class around onecomputer. So prepared, the students would move on to work in their groups or individually.Group work: Question and Answer cards. The students produced the Question andAnswer cards (Solar System, Day and Night, Atmosphere, Our Earth) in group mode.Examples of these cards are shown in Appendix 1. Each student had his or her owncomputer. To work as a group the two or three students in a group gathered around onecomputer. Their task was to design a visual to illustrate their theme and then execute it withcomputer graphic tools (Macintosh Super Paint or the HyperCard painting tools). Harry hadtaught them basic computer graphics earlier in the year, giving them experience of combininggraphics and text in the production of Halloween cards and various signs. He now introducedthem to two new techniques. One was interactive naming-and-voice buttons: clicking themouse on a part of the graphic would produce the written label for that part and clicking onthe label would produce its spoken rendition (Harry had prepared digital recordings of thelabels with the students on the library computer, using their own voices, and had copied thesevoice files into the students’ computers in the lab). The other technique was animationbuttons: clicking a “fast forward” button would trigger a sequence of cards in rapidsuccession, creating the illusion of movement. E.g. the students working on Day and Nightcreated several cards of the earth as a globe, each one with the land masses drawn a littlefurther forward; then when the viewer clicked the “fast forward” button, the earth would seemto turn a half revolution on its axis, simulating the movement from day to night. The groupsthen agreed on the final wording for their questions and answers (consulting their librarynotes) and entered the text in the question and answer boxes on the cards. When the cards41were complete, Harry copied them onto the lab network so that they became available to allstudents on their individual computers. Harry circulated among the groups helping them whenthey needed technical help, discussing with them design ideas and assisting them with designimplementation.Individual work: Picture and Paragraph (Topic) cards. The students producedindividually the topic cards on the physical features of earth (Oceans, Islands, Rivers etc.),using the notes and drawings they had prepared in the library. Examples of these cards areshown in Appendix 4. Harry digitized the original drawings by scanning them onto thestudents’ disks. The students then copied the scanned images onto their computer hard disks,retrieved and sized them into the left hand half of the screen, and edited them with theHyperCard painting tools, varying thicknesses of lines, creating different shading patterns andadding and deleting as they saw fit. They then entered and edited the text for the pictures inthe text box on the right hand side of the screen. When complete, Harry also copied thesecards onto the lab network so that they became available to all students on their individualcomputers. Harry circulated among the individual students, helping them as required.“Experts” help “Beginners”. At the beginning of each period the Our World students(the “Experts”) helped the other half of the class (the “Beginners”) start their computers andretrieve SuperPaint or Typing Tutor ready for their session with Karen. After about sixsessions, the Beginners were ready and demanding to set up their own machines. At this pointHarry thanked the Experts for their help and said that the Beginners would now manage ontheir own. The distinction between Experts and Beginners disappeared except in HyperCardapplications to which the Beginners still had not been exposed.42The Peer Tutoring sessions1. Peer tutoring to students from same class (PT(1)).The half of the class who built the stack peer tutored its contents, with an emphasis onthe Overall Topics (Question and Answer cards), to the other half of the class who had beenworking on a different unit. Tutors were matched with tutees one-on-one and assigned to onecomputer per pair in the computer lab. If both had a first language in common, they werepermitted to use it. I observed, recorded and analysed the tutoring sessions of four pairs whoconversed in English. The teachers were impressed with the quantity of dialogue generated bythe sessions but noted from the discourse transcripts that the tutors spoke more than the tuteesand that the dialogue showed a simple question and answer pattern led by the tutors, usingtheir concepts Checklist (see Appendix 2), rather than more equal exchanges between thepairs. After discussion, Harry built for future use a HyperCard Tutoring stack as a front endto the Our world stack with key visuals to help tutors explain underlying concepts and tofacilitate discussion between tutor and tutee. Appendix 5 shows samples of cards produced forthe Tutorial stack.2. Peer tutoring to students from other classes (PT(2).The first peer tutoring sessions took place at the end of the school year. Harrytherefore planned to use the new Tutorial stack with the Our World stack early in the nextschool year when four of the original tutees would remain in the school, mainstreamed fromESL to Grade 6 or 7 classes. Arrangements were made for them to give special tutoringsessions to six younger Grade 5 ex-ESL students who were selected by their teachers as43particularly in need of or able to benefit from extra tuition, plus two students from Karen’snew class who were also tutees in the first sessions but then using Li.Harry first arranged three orientation sessions to show the tutors how to use theTutorial stack. He then arranged for two tutoring sessions: the first in the computer lab withfour tutor/tutee pairs simultaneously; the second in his office with four tutor/tutee pairs one-at-a-time. Each tutor had a different tutee in each session. All sessions were conducted in L2.I observed all four pairs in the computer lab and recorded three of them (one recorder failed);and observed and recorded all four pairs in the computer coordinator’s office.The StudentsBoth the student data and student profiles in table 1.2 which follows were supplied bythe students’ teachers. All the tutors and the “same-class” tutees came from Karen’s 199 1/92class. I have only included the tutees whose sessions were conducted in English, which werealso the sessions I observed and recorded. The tutors’ profiles correspond well with myindependent observations made in the early weeks of the Our World unit before I formallyinterviewed Karen; and agree also with independent observations made by Jim in the libraryand Harry in the computer lab. The “other-class” tutees, who were tutored three months laterin the following school year, came from four classes, including Karen’s new class for thatyear. The information in Table 1.2 is relevant to the qualitative observations, presented later,of both the Our World unit as a whole (Chapter 3) and the two series of tutoring sessions(Chapters 4 and 5).44Table 1.2: Student data and profiles Page 1 of 8STUDENT DATA I STUDENT PROFILETutorsAlbert Promoted from the School primary ESL program as soon asmale, age 1 1½, from China, he was 9. His physical size made him stand out. Hisin Canada 3½ yrs, elementary teacher felt he had a speech problem because hein three-member Group #1; spoke (and speaks) in broken English with unclear pronuncFirst sessions: iation. A speech therapist found him within acceptabletutored same-class tutee in range for ESL students. He goes to Chinese school MondayLI; to Friday. Karen says his Cantonese is excellent andSecond sessions: switching between English and Chinese is difficult. Hematched with other-class thinks in Chinese and tends to superimpose Chinesetutee Manny, age 1 1¾, from structure on his English. He knew his printed alphabet whenVietnam, in Canada 2 yrs, a he came to Karen but she taught him cursive script. He hassame-class tutee with Bobby greater aptitude for math and science than language. Hein first sessions, using Li. comes from an intact family.Bobby Has younger sister Flo (also a tutor) in same class. Hadmale, age 13½, from knowledge of English, numbers and alphabet when arrivingVietnam, from a refugee camp with Flo and family. He is outgoingin Canada 2 yrs, and athletic (well thought of by his PE teacher) and wonin three-member Group #1; “all star” award in School badminton team. He isFirst sessions: developing good work habits, now completing histutored same-class tutee homework which he did not do last year. He is maturingManny in Li; socially and developing responsibility. Father is VietnameseSecond sessions: and mother Cantonese. Both he and Flo speak bothnot in second sessions languages fluently and Cantonese at home. He learned tobecause transferred to read and write in Vietnam but had no drawing skills whichanother school. he has now developed well.45Table 1.2: Student data and profiles Page 2 of 8STUDENT DATA STUDENT PROFILETutorsChancy Father was a teacher of English in China. He says thatmale, age 134, from China, Charley had behavioural problems in China where all hisin Canada 1½ yrs, teachers spoke of his lack of concentration. Karen says hisin three-member Group #1; problems persist in Canada. He tends to interfere and seemsFirst sessions: to invite negative attention. Earlier in the year he set off thematched with same-class School fire alarm to “see what would happen”. Hetutee Kenny, age 13¾ from sometimes does not get on well with the other students. HisVietnam, in Canada 6 ms; grandmother told Karen that in China the child is kingSecond sessions: because of the one child per family rule and the one childnot in second sessions becomes “spoiled”. Karen thinks this applies to Charley.because transferred to However, he likes his Canadian school and its learninganother school environment.Denis Has older sister Jeannie in same class. He had four years ofmale, age 1 1½, from Hong English, one period per day, in the Hong Kong system. HeKong, in Canada 18 ms, arrived with good handwriting, knowledge of alphabet andin two-member Group #2; numbers and able to read and write in English within theFirst sessions: confines of a limited vocabulary. He did not havetutored same-class tutee conversational English as the Hong Kong system deals withLenny, age 10¾, from reading and writing only. He is sociable and friendly and aPoland, in Canada 6 ms; good badminton player. He got a ribbon for his team,Second sessions: making first place in this year’s badminton tournament. Hetutored other-class tutees is particularly adept at working with computers.Sam, 10 and Vic, 9¾, bothCanadian-born Chinese.46Table 1.2: Student data and profiles Page 3 of 8STUDENT DATA STUDENT PROFILEHas older brother Bobby in same class and has similarbackground to him. She is chatty in her native Vietnamesebut not communicative in English in which she has troubleexpressing herself, even in structured situations. She,Jeannie and Ivy are Karen’s “silent ones” that she feels need“spoonfeeding”. She is passive in class. Karen thinks Her“passivity” is partly cultural as oriental girls are notencouraged to express themselves actively, and partlyI I IHas younger brother Ollie in same class. Came to Canadafrom a refugee camp with his parents and Ollie. Spent sixmonths in Calgary before coming to B.C. but this is his firstschool in Canada. His has a good attitude, wants to learnand is settling down well.because she lacks basic study skills.47Table 1.2: Student data and profiles Page 4 of 8STUDENT DATA I STUDENT PROFILETutorsGladys Shows herself academically md med anti grasps conceptsfemale, age 12½. from quickly. She too is from a one-child family in China. HerChina, in Canada 2½ yrs. mother is a schoolteacher. Her father does work inin three-member Group #3; connection with computers. She did not do well in herFirst sessions previous school. Karen thinks that her previous teacher, nottutored same-class tutee. a Chinese speaker, did not recognize the problem ofNancy, age 10½, from learning English against the totally different backgroundVietnam, in Canada 6ms: structure of Chinese. She said that Gladys was slow. KarenSecond sessions: finds the reverse to be true. Gladys spent four months oftutored other-class tutees the previous year in a another school because of her familyPatty, 10¼. from China, in moving away for this period. She has settled down and isnada 4 yrs and Ura, progressing well.10¼, Canadian-born Chinese.Hetty Also from a one-child family in China. Her father dotes onfemale, age I 1½. from her but puts some pressure on her to do well. She is closeChina. in Canada 2½ yrs. to him and talks to him about what happens at school. Sheiii three-member Group #3; learns quickly, has a good attitude to learning and has goodFirst sessions: insight into “how to learn”. She had problems adjusting totutored same-class tutee rules and regulations in the school. In her first year sheOllie, age 8¾. from Vietnam, refused to make friends. She has an underlying defiancein Canada 2 yrs; which she masks with apparent compliance. She is verySecond sessions: determined to do things her own way.tutored other-class tuteesRoxy. 10’ . Canadian-bornChinese and Win, 10, fromVietnam, in Canada 2yrs.48Table 1.2: Student data and profiles Page 5 of 8STUDENT PROFILEBorn in the school neighbourhood, she is the only one inKaren’s class born in Canada. After only a few months inkindergarten, her family returned to Taiwan for five years.Her first schooling was in Taiwan where she was immersedin Mandarin (conunon to mainland China, Taiwan, andSingapore while Cantonese is spoken in Hong Kong). Sheis a bright, diligent student but is another of Karen’s three“quiet ones” and needs drawing out to speak in class. Sheand fellow “quiet one” Jeannie are best friends and make upHas younger brother Denis in same class and a similareducational background to Denis. Like Ivy, her team mate,she is a diligent student but, also like Ivy, is one of Karen’sthree “quiet ones” and needs drawing out to speak in class.Seems timid and shy. Karen does not think this is a naturalcharacteristic but rather a product of the classroomenvironment in Hong Kong where teachers maintain adistance from students.STUDENT DATAthe two-member Group #4.inItutored s.49Table 1.2: Student data and profiles Page 6 of 8STUDENT PROFILEOlliemale. age 9¾ from Vietnam,in Canada 2 yrs;tutored by Hefty female.age 11½ from China,in Canada 2½ yrs.Has older brother Eric in same class. Unlike Eric, who hadbasic primary education in Vietnam, Ollie had no educationat all when arriving at the age of 5 in a Hong Kong refugeecamp. He is illiterate in Vietnamese and is thereforelearning literacy in English. He is slow recognizing wordsand at the moment is better at math than language.KcSTUDENT DATA I IHas cousin Nancy in same class and a sister in Grade 2. Hisfather was a lawyer in Vietnam. He learns quickly and is aconscientious student. He has a speech impediment due tomalformed teeth. Karen thinks that this reinforces atraditional Asian teacher-submissiveness and makes himappear less able than he is.Lives with his mother and two brothers: one younger andthe other older. His mother is divorced. His language isgood orally but his writing and math is below average. Hismind easily wanders in class.Both her parents are professionals: her father a pharmacist,her mother a teacher. She was well educated in Vietnamand has a comfortable economic background. She had noEnglish on arrival in Canada and had to start with thealphabet. She is highly motivated arid academically orientedwith superior performances in written language and math.fViage I /z fromin Canada 2½ yrs.50Table 1.2: Student data and profiles Page 7 of 8STUDENT DATA STUDENT PROFILEfemale, age lOiA, Canadianborn Chinese;tutored by Hetty, female,age 12 from China,inCanada:3.3Li is Mandann She was recommended to the EnglishLanguage Centre (ELC) and the Learning Assistance Centre(LAC) She is low in confidence, has cognition problems(“doesn’t have it”), is embarrassed and poor in talking, hasdifficulty expressing herself, and is weak in wnting Rated“average minus”Li is Mandarin. She was recommended to the LearningAssistance Centre (LAC). She is still “fairly ESL”. She hasan outgoing personality, likes to participate, asks questions,is bright, confident and takes risks. If she had sufficientEnglish she would be a top student.L 1 is Cantonese. She was recommended to EnglishLearning Centre (ELC) where she had a major breakthroughin mastering English. She is shy, not overly confident butwill talk and take risks. Rated an average student.Was in Karen’s class last year so she knew Hetty. She leftVietnam without schooling at 4½, lived in refugee camps inHong Kong for 3 years, changed camps 3 times, andlearned some Vietnamese and math from a Vietnamesevolunteer. She came to Canada not “school ready”, did notintegrate into primary Grade 3 and so came to her presentESL class at age 9. She learns quickly and willingly, hasgood math skills, is lively and eager to please, listens welland likes to help in order to show what she can do.female,Other-Class Tutees (second sessions)tutoredage 13frotn’4.indian-.leRoxyWinfemale, fromVietnam,in Canada Ttutored by I51Table 1.2: Student data and profiles Page 8 of 8STUDENT DATA I STUDENT PROFILE)Li is Cantonese. He was recommended to the EnglishLanguage Centre (ELC) and also last year to the LearningAssistance Centre (LAC) because he appeared “learningdisabled” although with good potential. Not well motivated,poor listener, and may have processing and comprehensionWas in Karen’s class last year so he knew Albert. He isChinese from Vietnam, speaks Cantonese and a littleVietnamese, and had some schooling in Vietnam. Hearrived in Canada with his family from a refugee camp inthe Philippines.0age [1yrs. problems.Still “very ESL”, lacks confidence, has low language skills,and may have a learning disability: he has difficultyprocessing and “glazes up”. Very much below Grade 5norm.52MethodologyThe purpose here is to state in outline what methodologies were used and when theywere used for gathering and analyzing data throughout the case study. Their origin andrationale are discussed in greater detail in Chapters 2A & 2B, Selective Review of Literature,under the headings “The role of discourse analysis” and “Work Study”. They fall under twomain headings: work study methodology for recording and analyzing task activities anddescribing the larger context in which those activities are set; and discourse analysismethodology for recording and analyzing student discourse.Work study techniauesActivity SamplingActivity (or work) sampling was used in the library and the computer lab to determinewhat percentage of their time the students were spending on various observable activities.The particular type of activity sampling used was Fixed Interval (or Systematic Random)Sampling. During the first ten weeks of the study, I observed each student’s activityapproximately every two minutes over six single and five double periods in the library andover eight double periods in the computer lab. Although Our World was not taught in thehomeroom, I observed four double and three single periods in the home room on units similarin content demand to Our World in order to compare the home room with the library andcomputer lab environments. I recorded observations under two sets of broad headings:Students’ work modes.1. Teacher-led activity- teacher-fronted, whole-class work.2. Solo - students working on their own.533. Solo with teacher - students working individually with a teacher.4. Group work - students working in their groups of two or three.5. Group work with teacher - students working in their groups with a teacher.6. Working with peers - students working out of their groups with peers.7. Pause - time out of the room, waiting for a teacher or not on assigned task.8. Other - time spent at library shelves, the board, or the computer lab printer.Students’ communication modes.1. Speaking - time spent in conversation.2. Reading - time spent reading notes or printed material (not computer screen).3. Writing - time spent writing on paper (not computer screen).4. Other - time spent on everything else, e.g. listening to teacher and “pause” time.I also used activity sampling to record communication mode activities in the peer tutoringsessions, using video recordings of four tutor/tutee pairs.InterviewsTeachers. Early in the study, I formally and separately interviewed, with theirconsent, the three teachers, Karen Black, the homeroom teacher, Jim White, the librarian, andHarry Green, the computer lab coordinator, to get their individual perspectives on the OurWorld unit, a description of their roles and an account of their previous working relationships.I also informally interacted with all three teachers throughout the unit and spoke with Karenseveral more times to get her profiles of the students and their previous experience in herclass.54Students. Karen invited me to talk as much as I liked with the students in thehomeroom during any of their free weekly periods. I therefore spoke with them on severaloccasions throughout the study, individually and in small groups, to get their perspectives onthe unit, their experiences in working together in groups and in peer tutoring sessions, andtheir reactions to the three environments in which they were working: the library, thecomputer lab and their homeroom.ObservationI observed all sessions of the Our World unit except one period in the computer laband two in the library. During the periods that I activity-sampled (the first nine weeks of theunit) I jotted down brief qualitative notes to help in the later interpretation of the quantitativesampling data. During the remaining eight weeks of the unit (including the first of the peertutoring sessions) I made extensive qualitative notes as I circulated among the student groups.During the peer tutoring sessions held in the computer coordinator’s office (two pair-periodsfrom the first sessions and all eight pair-periods from the second sessions) I was able tocontinuously observe the tutor’s and tutee’s interactions with each other and the computerbecause only one pair was working in the office at a time. In addition, I had a complete videorecording of two periods of peer tutoring to supplement my personal observations.Participant observationThe students quickly accepted me as just another teacher doing special things in theirunit like recording, writing notes and occasionally asking questions. Both Jim in the libraryand Harry in the computer lab encouraged me to help out with students’ questions duringgroup work. So I observed the groups as I interacted with them from time-to-time in the role55of “TA”. I also took the opportunity to occasionally sit with each group and watch it work formore extended periods of time and at the same time answer questions as they arose.Discourse analysis techniquesAudio and video tane recordingThe Our World unit. I recorded all the Our World library and computer lab sessionsthat I attended with four audio tape recorders, one for each of the four groups. Theserecordings were useful as an audio record of the progress of the unit and the content of thelesson periods. However, although they clearly recorded the teachers’ discourse, they caughtvery little of the students’ discourse because the students spoke on-the-whole softly againstconsiderable background noise. I also recorded the sessions (not on Our World) that Iobserved in the home room with one audio recorder.The peer tutoring sessions.It was imperative to get clear recordings of student discourse in the peer tutoringsessions so that accurate transcripts could be made for discourse analysis. I tried a number ofalternative configurations of equipment and eventually found the answer in small, highlysensitive Shure unidirectional microphones attached to headsets (as used by recordingvocalists) so that the microphone came to within about one inch of the wearer’s mouth. Wefitted the tutors and tutees with these headsets and connected the microphones either to aregular desk-model tape recorder through a Shure mixer or directly to a stereo tape recorder.The results were good. Relatively little background noise was picked up and it was possibleto decipher even whispers and below breath utterances.56The first peer tutoring sessions, PT( 1), extended over two periods. Two tutor/tuteepairs were audio-recorded in the first of these periods in the computer lab. Two pairs wereaudio- recorded in the second period in the computer lab. One of these pairs was additionallyvideo-recorded. A third pair was audio- and video-recorded in a special second period in thecomputer coordinator’s office (the tutor had been absent from the second regular period). Inaddition two pairs were audio- and video-recorded in a supplementary session in the computercoordinator’s office.The second peer tutoring sessions, PT(2), extended over three periods. Threetutor/tutee pairs were recorded in the first of these periods, held in the computer lab. Thesecond and third sessions were held in the computer coordinator’s office. Four pairs followedeach other on one day and four pairs (same tutors, different tutees) followed each other onanother day in the succeeding week.Ouantitative analysis of discourse transcriptsAll peer tutoring pair-periods recorded were transcribed, using the Systematic Analysisof Language Transcripts (SALT) software (J. F. Miller and R. S. Chapman, 1984-1991) fromthe Language Analysis Laboratory of the University of Wisconsin-Madison. This software hasthe advantage, once discourse has been key-entered using the required conventions, ofanalyzing discourse over a great number of variables of which the following were used:1. Total utterances per speaker, broken down inter alia by complete, incomplete,interrupted and questioning utterances, and utterances containing mazes and overlaps.2. Speaker turns by number of utterances.573. Timing of total utterances and total words per minute for each speaker (excludingwords in mazes).4. Number of utterances per speaker by utterance length in words.5. Standard word lists per speaker broken down by grammatical categories of directquestions, negatives, conjunctions, auxiliary verbs and pronouns.6. Word frequency lists per speaker.The SALT data thus provides the basis for a quantitative analysis of the discourse transcripts.Pedagogical analysis of discourse transcriptsThe discourse transcripts were also analyzed qualitatively to assess the students’success in realising the pedagogical objectives of the tutorial sessions. These objectiveschanged from PT(1) to PT(2) and are stated in detail in the context of Chapters 4 and 5.Cognitive analysis of discourse transcriptsA cognitive analysis of the discourse transcripts was required to determine the amountand quality of the ideational as distinct from the interpersonal component generated. The roleof discourse analysis is discussed in the selective review of literature in Chapter 2A. Themethod of its application for cognitive analysis is discussed in detail in Chapter 7 dealingwith the cognitive analysis of the transcripts from PT( 1) and PT(2). In brief, a method forboth quantitative and qualitative cognitive analysis was developed for this study, drawing on:1 Mazes are defined as false starts, repetitions, and reformulations and are codedwithin parentheses (like this). Words found in mazes are excluded fromcalculations of Number of Different Word Roots, Total Number of Words andMean Length of Utterance (MLU).581. The concept of Ventures, i.e. units of instruction categorized by their “overarchingobjectives”, any one or number of which make up a “strategy” of teaching (Smith,Meux et al., 1967, “Strategies of Teaching).2. Knowledge Structures (KS5), as developed in Mohan’s (1986) Knowledge Framework,a high level cognitive analytical framework of six KSs that may be applied to bothverbal and non verbal (e.g. graphical) presentations of knowledge.3. Staab’s (1986) language functions, based on Halliday’s (1970) dichotomy ofInterpersonal and Ideational language functions. Staab collapses Tough’s (1979, 1973)seven functions of language into five: three interpersonal and two ideational. The twoideational functions are Informing and Reasoning of which Reasoning is key indetermining academic discourse quality.4. Mehan’s (1979) concept of Initiation-Response-Evaluation (I-R-E), the “default” modelof teachers’ classroom discourse (Cazden, 1988).5. Our own determination of different tutoring modes or styles.We now turn to a selective review of the literature relevant to the activities observedin this case study and to the methodologies used for their investigation.59CHAPTER 2A: SELECTIVE REVIEW OF LITERATURE (1)INTRODUCTIONOur Research Question is: “Is peer tutoring, using a computer-based hypermediaresource, an effective means of helping ESL students generate academic discourse?” Thistouches on four main themes which are inseparably interlinked in our enquiry:1. Second language acquisition;2. Academic discourse;3. Peer tutoring4. The computer in education.Because the peer tutoring in our case study comprised a series of tutor tasks builtaround the computer, we review task-based language teaching and computer-assisted languagelearning in conjunction with peer tutoring.Each theme contains a number of historical, conceptual and methodological issuescontained in the literature. Our scheme of review is therefore as follows:1. Second language acquisitionConceptually, we review the following perspectives on second language acquisitionwhich bear upon our study: Krashen’s (1986) input hypothesis; Cummins’ (1991) distinctionbetween academic and conversational proficiency; Halliday’s (1975) functional model oflanguage, which provides a conceptual framework for the practical analysis of discourse in asociocultural setting from a linguistic position; and Schieffelin and Ochs’ (1986) presentationof language socialization from an anthropological position.602. Academic discourseAcademic discourse is a convenient label for the “language of teaching and learning”.It contains, in Halliday’s terminology, both ideational and interpersonal components withemphasis on the ideational. We review historical, conceptual and methodological aspects ofthis theme:HistoricalA review of observational studies of classroom discourse and interaction in whichthere is a striking neglect of cognitive as distinct from linguistic and behaviouralanalysis.Conceutual:a) A rationality model of teaching vs a behavioural, linguistic or other type of model;b) Two schemes for the cognitive analysis of discourse:I) From the Smith, Meux Logic and Strategies of Teaching studies:Logical Operations, Ventures and Moves;ii) From Mohan’s Language and Content studies: the Knowledge Framework.MethodologicalDiscourse analysis:a) its development for linguistic and interpersonal analysis;b) its identification of the Initiation-Response-Evaluation (I-R-E) teaching model;c) its development for cognitive analysis using the above conceptual schemes.613. Peer tutoringPeer tutoring was a task assigned to the students with the object of generatinglanguage use while absorbing content. Specifically it was a computer-based task. We thereforereview:Peer tutoringThe concept of peer tutoring, its historical development and assessments of itseffectiveness as a method of teaching.Task-based language teachinga) The concept of task-based teaching;b) Concepts of psycholinguistic and sociocultural tasks;c) Computer-assisted language learning (CALL);Methodologicala) Task design;b) Work Study, a discipline for task recording, measurement and analysis.4. The comnuter in educationThis theme is broader based than the methodological review of CALL. It looks at theissue of technology in the classroom, including the technology of print. Students of mediaclaim that print was the origin of the traditional knowledge-transmission model of teachingwhile the computer and other electronic technologies have a natural affinity with aknowledge-construction model of learning.The unit we studied was an example both of computer use and of knowledgeconstruction learning. We therefore regard it as a window through which we can discern the62problems and potential of the rapidly maturing electronic technologies for both language andcontent learning in the classroom. We therefore review:a) The links between technology in the classroom and teaching style;b) The technologies of communication (media): Writing, Printing, Computers;c) Underlying factors in the school system inimical to technological change;c) Computers and new patterns of thinking;d) Appropriate and inappropriate uses of technology;e) The future.Although our case study is a second language learning unit, many of the concepts andmethods we review are equally applicable to content learning by Li students.PERSPECTIVES ON SECOND LANGUAGE ACQUISITIONIn this Division we review the perspectives on second language acquisition relevant toour study. We begin with Krashen’s (1985) Input Hypothesis and notion of comprehensibleinput; move to Cummins (1984, 1991) who provides the distinction between conversationaland academic discourse; then to Halliday (1975, 1985) whose functional theory of languageprovides a firm basis for identifying and analysing the ideational and interpersonalcomponents of discourse; and conclude with Ochs/Schieffelin (1986) and Ochs (1988) whoprovide for analysis in depth beyond the contours of the language to the ways in whichlanguage users (our students) are socialized into the culture (including that of the classroom),culture both defining and being defined by language use.63Krashen’s (1985) Input HypothesisKrashen’s five hypothesesThe Input Hypothesis is the central one of five hypotheses that restate Krashen’s(1976) comprehensive theory of second language acquisition. The theory, known as Krashen’sMonitor Theory (MT), was “one of the best known and most influential theories of SLA inthe 1970s and early 1980s . . . MT began life not as a theory of SL acquisition, but as amodel of SL performance” (Larsen-Freeman & Long 1991). Krashen’s five hypotheses can besummarized as follows (Krashen 1985:1-4):1. The Acquisition-Learning Hypothesis distinguishes acquisition, a subconscious processas used by children in acquiring their first language (which can be likened to Chomsky’s“Language Acquisition Device” or LAD), from learning, a conscious process used byadults in traditional language learning, which is based on the L2 grammar system. Thesubconscious process is by far the most powerful.2. The Natural Order Hypothesis states that rules of language are acquired in a generallypredictable staged order independent of the order in which they are taught in class.3. The Monitor Hypothesis states that conscious learning has only a limited function inadult SL performance and that is as a monitor or editor of language produced fromacquired subconscious knowledge.4. The Input Hypothesis states that language (whether Li or L2) is acquired only by theprocess of receiving comprehensible input. “We progress along the natural order(hypothesis 2) by understanding input that contains structures at our next ‘stage’ --structures that are a bit beyond our current level of competence” (4). Input at stages64beyond this represents uncomprehended noise. The received input is processed throughthe Chomskyan LAD, the “mental organ” for language, which generates rules forperceived new constructions and works equally for adults as for children. The InputHypothesis has two corollaries:1. Acquisition, the subconscious process, not learning causes speaking. Speaking does notcause acquisition.2. Given sufficient comprehensible input, the LAD will automatically generate therequired grammar without need of teaching.5. The Affective Filter Hypothesis: Negative affective factors, particularly lack ofmotivation, lack of self-confidence, and anxiety prevent various amounts ofcomprehensible input from reaching the LAD. This works in adults from puberty onwards,not in children, and is the reason language acquisition declines sharply after childhood ispassed.Language acquisition for adults, very simply, depends on sufficient comprehensible inputgetting through the Affective Filter. For the argument that output is important, see Swain,1985)Claimed evidence for the Input HypothesisKrashen & Terre!! (1983) claim extensive evidence for the Input Hypothesis:1. Caretaker speech, teacher talk and foreigner talk (Krashen, 1981) make concessions tothe interlocutor’s lack of language experience by negotiating meaning through variousmeans (e.g. simplification, expansion of utterances, repetition) to assist the decoding process. This ensures a steady flow of comprehensible input which is not more than (I + 1)65structures beyond the acquirer’s level of competence. Interlanguage talk (speech ofpeers) is also important for the same reason.2. The “silent period” in children’s L2 acquisition (Krashen 1983, Newmark 1966,Rodriguez 1981, Fourcin, 1975) can be explained by the Input Hypothesis. Largeamounts of comprehensible input are assimilated during this period so that speech, when itbegins, is at a level comparable to that reached by children who have not remained silent.3. When there are age differences, e.g. between adults and children or between older andyounger children, those who are older acquire L2 at a faster rate (Krashen, Long &Scarcella 1982). This can be explained by the greater amount of comprehensible input towhich they have been exposed because of their greater life-experience. Younger acquirersovertake adults in the long run because of their lower affective filters.4. Research findings that greater L2 acquisition sometimes results from instruction within aninformal environment compared with a formal one (Krashen 1982) can be explained bythe larger amounts of comprehensible input often provided in informal environments.5. Comprehensible input may be the vital difference in the results of different teachingmethods. The low efficacy of grammar-based and drill-based methods can be explainedby paucity of comprehensible input. Success of some unconventional methods like NaturalApproach (Krashen & Terrell, 1983) and Suggestopedia (Dhority, 1984) may be explainedby plentiful comprehensible input. Findings that delaying oral production has no negativeeffects (Gary 1975, Postovsky 1974) may also be explained by the Input Hypothesis -- theLAD silently processes the comprehensible input with or without speech.666. The success of Canadian immersion programs (Swain & Lapkin, 1982) can beattributed to an abundance of comprehensible input. In particular, this experience showsthat the subject-matter classroom can be a superb language classroom with natural, notartificially contrived, input. Similar benefits have been achieved in “sheltered” contentclasses taught to university students in L2. (Edwards, Wesche, Krashen et al. 1985).7. Successful bilingual programs provide content instruction in Li together withcomprehensible input in L2 (Cummins, 1981, 1983). The Li instruction increases learners’“cognitive academic language proficiency” (Cummins, 1979) which, in turn, makes the L2input more comprehensible, i.e. increases comprehensible input.8. The Reading Hypothesis states that a relationship exists between writing competence andreading for interest and pleasure (Krashen, 1984; Smith, 1983). This may be explained byextensive reading placing added comprehensible input at the service of writing.Criticisms of the Input HypothesisKrashen’s theory attracted considerable criticism in the SLA literature and perhaps“more than its fair share (because) it was one of the first ‘theories’ developed specifically toexplain SLA, and so was for a time a, and for many the major position to be subjected toempirical test” (Larsen-Freeman & Long 1991:245). Legitimate criticisms included, inter alia,challenges to the learning/acquisition, conscious/subconscious distinctions and the use ofsubjective, anecdotal evidence to support them (McLaughlin 1978); and Krashen’sidiosyncratic use of the term hypothesis: the Natural Order Hypothesis is a generalizationbased on empirical studies; the Acquisition-Learning, Monitor and Input hypotheses areclaims about SLA processes; and the Affective Filter Hypothesis is a metaphor. None of these67is a hypothesis in the usually accepted sense of the term, i.e. being empirically testable(Larsen-Freeman & Long 1991:291). Gregg (1988) presents a rigorous critique of what heconsiders vague linguistic notions in Krashen’s theory and some of the defences of it andconcludes: “It may be the case that second language acquisition is simply too hard, toocomplex a field to be susceptible of coverage by a unified theory” (Gregg 1988:95). Larsen-Freeman & Long (1991:249) give a balanced summing up:In conclusion, the Monitor Model and its successor, MT, served SLA researchers wellby offering an early attempt to make sense of a wide array of disparate researchfmdings. In addition, Krashen’s ideas themselves stimulated a good deal of data-basedresearch, and forced some fresh thinking in language teaching circles. While some ofthe original claims no longer excite much interest among researchers and/or have beensuperseded by other developments, they served a valuable purpose by identifying someof the relevant issues and, where apparently wrong, by obliging critics to seek out andsubstantiate alternatives.Cummins (1991) and academic language proficiencyCummins’ research has focused on first and second language (Li and L2) academicdevelopment in bilingual contexts. Over a number of years he has developed three theoreticalconstructs: 1) a distinction between conversational and academic language proficiency; 2) theinterdependence hypothesis which posits a relationship between Ll and L2 proficiency; and 3)the threshold hypothesis which states that bilingual children must first attain threshold levelsof linguistic competence to enjoy the cognitive benefits of bilingualism (Cummins 1991).68Conversational and academic language proficiencyConversational and academic language proficiency were originally termed ‘basicinterpersonal communicative skills” (BICS) and “cognitive academic language proficiency”(CALP) (Cummins 1979). The problems Cummins addressed in making this distinction were:1. The unwarranted conclusion that poor performance on an L2 verbal IQ test indicatesdeficient cognitive abilities: testers have assumed that the language proficiency needed forface-to-face communication is no different from that needed to perform academic tasks.2. The premature transfer of minority language students to L2 only classrooms on the basisof their “fluency” in superficial interpersonal communication: the result is that suchstudents fall behind academically and then get “tagged” as cognitively or culturally“deficient” (Cummins 1984:131).Cummins points out that while data suggests that “many minority students can developa relatively high degree of English communicative skills within about two years of exposureto English-speaking peers, television and schooling”, analysis of results from language testsadministered in the Toronto Board of Education shows “that it took immigrant students whoarrived in Canada at age 6-7 or later, between 5 and 7 years, on the average, to approachgrade norms in English verbal-academic skills” (Cummins 1984:133). For empirical evidence,see Collier (1987). This general conclusion is supported by comparison with successfulbilingual programs for minority students in the United States (San Diego City Schools 1982)and French immersion programs in Canada (Swain & Lapkin 1982). Similar results werereported by Skutnabb-Kangas & Toukomaa (1976) who, in their study of Finnish immigrantstudents in Sweden, drew a distinction between “surface fluency” and “conceptual-linguistic69knowledge”. Cummins formalized this distinction in terms of “basic interpersonalcommunicative skills” (BICS) and “cognitive academic language proficiency” (CALP). BICSwas “defined in terms of the manifestation of language proficiency in everydaycommunicative contexts whereas CALP was conceptualized in terms of the manipulation oflanguage in decontextualized academic situations” (Cummins 1984:137). Cummins (1991)now refers to BICS and CALP as simply “conversational language proficiency” and“academic language proficiency” in bilingual contexts.The interdependence hypothesisThe relationship between Li and L2 proficiency is expressed in the “interdependencehypothesis” (Cummins 1978, 1991) which states formally:To the extent that instruction in Lx is effective in promoting efficiency in Lx, transferof this proficiency to Ly will occur provided there is adequate exposure to Ly (eitherin school or environment) and adequate motivation to learn Ly.In other words, “the development of literacy-related skills in L2 is partly a function of priordevelopment of literacy-related skills in Li”. What is transferred from Li to L2 is primarilycommon underlying conceptual knowledge rather than specific linguistic elements. Thishypothesis is advanced to explain the research data that consistently indicate that instructionin a minority language has no adverse effect on development in the majority language, afactor of obvious concern to the parents and teachers of children entering a bilingual program.Cummins (1991:77-78) notes the relevance of the interdependence hypothesis to Krashen’snotion of “comprehensible input”:Li literacy and conceptual knowledge constitute central attributes of the individual thathelp to make academic input in the L2 comprehensible. If a second language learner70already understands concept x in her Li then L2 input containing that concept will beconsiderably more comprehensible than if she does not understand concept x in her Li.The threshold hypothesisThe threshold hypothesis was developed to account for the findings of some empiricalstudies of cognitive deficits resulting from bilingualism. Cummins 1991:76) states:there may be threshold levels of linguistic competence that bilingual children mustattain both in order to avoid cognitive disadvantages and allow the potentiallybeneficial aspects of becoming bilingual to influence their cognitive functioning.Cummins claims a “valuable heuristic impact” for this hypothesis but agrees that itcannot be empirically validated because of the difficulty in specifying and operationalizing thenature of language proficiency.Operationalizing the concept of academic discourseIn the Our World case study we needed, in Cummins’ terms, to differentiate between“conversational language” and “academic language” in order to assess the extent to which thestudents were successful in generating academic discourse. Cunimins does not offer aprocedure for operationalizing this distinction but claims “solid linguistic evidence fordistinguishing between conversational and academic aspects of language proficiency” and inparticular refers to Biber’s (1986) “psychometric analysis of an extremely large corpus ofspoken and written material” along three dimensions:1. high personal involvement and real-time constraints vs conditions permitting considerableediting and high explicitness of lexical content but little personal interaction.2. a detached formal style vs a concrete colloquial one based on analysis of linguisticfeatures such as nominalizations, prepositions, and passives;713. a primary narrative emphasis, marked by considerable reference to a removed situation vsnon-narrative emphases marked by little reference to a removed situation but a highoccurrence of present tense forms.While these linguistic criteria may in general distinguish between conversational andacademic discourse, they are not satisfactory as a basis for specific analyses. For instance, acolloquial style may indeed characterize conversational discourse. This does not, however,preclude the use of a colloquial style in an academic dialogue, e.g. an excited discussion,replete with exophoric references to data presented on a computer screen. Conversely, someconversational discourses may be clothed in great fonnality at high levels of abstraction, e.g.a disciplinary interview. We felt that we needed direct criteria that we could apply to eachutterance to determine whether it was “conversational” or “academic”. We therefore turned toHalliday’s (1973, 75, 85) functional model of language which allows for a multi-dimensionalanalysis of discourse to include not only linguistic but also sociosemantic dimensions.Halliday’s functional model of languageHalliday’s model of language embraces not only meanings expressed throughgrammatical patterns but also meanings which are themselves “the expression, or realization,of options in behaviour” realized in socially significant contexts. Halliday (1973:5) has, in hismodel building, taken into account Bernstein’s theories of social order and social change,“theories in which language plays an essential part”:Language is the primary means for the transmission of culture from one generation tothe next; and Bernstein’s work has shown that there are certain types of social context,especially forms of interaction between parent and child, which are critical for thissocialization process. By taking a functional viewpoint we can gain some idea of how72it is that ordinary language, in its everyday uses, can so effectively transmit to thechild the deepest patterns of culture.Halliday further points out that a functional model of language was advanced byMalinowski from an anthropological viewpoint with the perception that the functional originsof the language system could be seen most clearly in the language of young children. At thechild level, language is what language does. The child uses it to realize his/her intentions.Halliday identifies eight child-talk functions which the child uses one at a time according tothe end in view: The instrumental model, the “I want” function, is a way of manipulating andcontrolling the child’s environment; the regulatory model, the “do as I tell you” function is away of regulating the behaviour of others; the interactional model expresses “me and you”,the self interacting with others; the personal model expresses the self as speaker and aware ofindividuality; the heuristic model with the “tell me why” function uses language to investigatereality, to learn; the imaginative model with the “let’s pretend” function, creates privateworlds; finally, developing later, the representational model with the “I’ve got something totell you” function communicates and expresses propositions. Halliday (1973:36) links themicro functions of child talk to the macro functions of adult language:What happens in the course of maturation is a process that we might from one point ofview call ‘functional reduction’, where the original functional range of the child’slanguage- a set of fairly discrete functional components each with its own meaningpotential- is gradually replaced by a more highly coded and more abstract, but alsosimpler, functional system. . These ‘macro-functions’ are the highly abstract linguisticreflection of the multiplicity of social uses of language.Halliday identifies the macro functions as three in number: the ideational, theinterpersonal and the textual:731. The ideational is concerned with the content of language; its function is to express “ourexperience, both of the external world and of the inner world of our own consciousness --together with what is perhaps a separate sub-component expressing certain basic logicalrelationships” (66). It represents the meaning potential of the language system in itsvastness and complexity: “for example, the whole of the transitivity system in language --the interpretation and expression in language of the different types of process of theexternal world, including material, mental and abstract processes of every kind” (39).2. The interpersonal refers to the non-ideational elements in the adult language system: it is“language as the mediator of role, including all that may be understood by the expressionof our own personalities and personal feelings on the one hand, and forms of interactionand social interplay with other participants in the communication situation on the otherhand” (66). Grammatically, the interpersonal function is represented by mood andmodality: “the selection of the speaker of a particular role in the speech situation, and hisdetermination of the choice of roles for the addressee (mood), and the expressions of hisjudgements and predictions (modality)” (41).3. The third component, the textual, “has an enabling function, that of creating text, which islanguage in operation as distinct from strings of words or isolated sentences and clauses.It is this component that enables speakers to organize what they are saying in such a waythat it makes sense in the context and fulfils its function as message” (66), thusdistinguishing “a living message from a mere entry in a grammar or a dictionary” (42).The “living message” is contained in discourse which is embodied in text andgenerated within the context of specific sociocultural situations. Halliday offers a conceptual74framework for the practical analysis of discourse that systematically relates the “context ofsituation” to the “functions of language”. The functions of the semantic system of languagewe have briefly examined above: (a) the ideational, subdivided into logical and experiential,(b) interpersonal, and © textual. The context of situation is expressed in terms of field ofdiscourse, tenor of discourse and mode of discourse. The two, context of situation andfunctions of language, are then brought together into:a correlation between the categories of the situation and those of the semantic system,such that, in general terms, the field is reflected in the experiential meanings of thetext, the tenor in the interpersonal meanings, and the mode in the textual meanings.We could express this the other way round by using a complementary metaphor andsaying that experiential meanings are activated by features of the field, interpersonalmeanings by features of the tenor, and textual meanings by features of the mode”(Halliday 1985:29).Halliday’s system firmly sets language with recognizable functions in a socioculturalcontext. This marks a departure from the psycho-linguistic movement of the 1960s, whichwas concerned primarily with the mechanism of language rather than its meaning andfunction. Above all, language is a system for making meanings which becomes the “primarymode of transmission of the culture”. The semantic system of language is therefore parallelwith and pointing to the social system, itself a system of meaning. This gives point toHalliday’s notion of language development as a semiotic process (Halliday 1975:iv).Schieffelin & Ochs and language socializationSchieffelin & Ochs (1986) describe the process of language socialization (LS) from ananthropological position in a way that is entirely consonant with Halliday’s functionalapproach to language. They note the same intimate linking between the semantic and social75systems as Halliday, which they express as two interlocking processes: “socialization throughthe use of language and socialization to use language”. Importantly, they add that languagesocialization “is not limited to the role of language in integrating children into society, but isopen to investigating language socialization throughout the human lifespan across a range ofsocial experiences and contexts” (163). This makes the LS approach to classroom discourseanalysis particularly relevant.They note that the notion of LS “draws on sociological, anthropological, andpsychological approaches to the study of social and linguistic competence within a socialgroup”. Like Halliday, they acknowledge the work of Bernstein (1975) and his associates, e.g.in showing the link between school children’s styles of communication and local concepts ofsocial identity and social roles. They also note the relevance of Hymes’ (1967, 1972) notionof communicative competence, e.g. in Fischer’s (1970) study of linguistic socialization inJapanese and American family communication in which varying degrees of communicativecompetence maintain and progressively change the subject’s social position as member ofsociety (164). In summation, they state: “the contribution of linguistic anthropology todevelopmental psychology is to make explicit that Anglo white middle-class verbal interactions with infants and young children are also culturally organized and to outline the natureof that cultural organization” (165).A key theme in LS literature is that language socialization is an interactive processmeaning that “the child or novice is not a passive recipient of sociocultural knowledge butrather an active contributor to the meaning and outcome of interactions . . (an idea)compatible with the Piagetian (1952) concept of the child as an active constructor of his or76her own development” (165). This perspective draws on the early symbolic interactionistapproach taken by Mead (1934) and on the phenomenological notions “that reality, includingconcepts of self and social roles, is constructed through social interaction” (e.g. Berger &Luckmann 1967; Schutz 1967; Ricoeuer 1974); and “that members’ perceptions and conceptions of entities are grounded in their subjective experiences and that members bring somewhatdifferent realities to interpersonal encounters” (Mehan & Wood 1975). Interactivity is also adominant research theme within Vygotsky’s (1962, 1978) framework, e.g. Bruner (1966), Cole& Bruner (1971), Leontyev (1981), Cazden (1981), with the notion that “novices are able tocarry out particular tasks through ‘guided interactions’ (and) develop skills in a ‘zone ofproximal development’ as they move from guided or collaborative to independent action”(166).Ochs (1988) emphasises the role of activity in the whole process of languagesocialization and points out that activity is both a behavioural unit associated with particularmotivations and goals (Leontyev 1981) “and a process, in the sense of praxis (Marx 1959;Vygotsky 1962) “drawing on “the psychological approaches of Piaget (1952) and Vygotsky(1962, 1978) and the sociological approaches of Bourdieu (1977) and Giddens (1979, 1984)• . What is common to all these approaches is the view that knowledge and praxis createeach other” (14-15). Ochs represents this in a simple model:Linguistic knowledge < > Activity < > Sociocultural knowledgeThe relevance of the sociohistorical approach to this model is “that sociocultural andlinguistic knowledge structure activity, and activity creates (in the case of the novice/acquirer) and recreates (in the case of the member/competent language user) knowledge in77both of these domains” (16). This model is directly relevant to task-based content andlanguage learning in the classroom for both Li and L2 learners, and more particularly if thetasks are “real” rather than pedagogical activities. The sociocultural knowledge imparted isboth that of the host culture in general and of the classroom sub-culture in particular.ACADEMIC DISCOURSEIn this Division we review the history of cognitively oriented observational studies ofthe classroom and consider the reasons for their neglect after the ground-breaking work ofSmith, Meux et al. in the 1960s. Conceptually, we examine reasons for the importance ofcognitive observational studies, particularly in L2 learning; review the two Smith, Meuxstudies, Logic of Teaching and Strategies of Teaching, and their concepts of LogicalOperations, Ventures and Moves; review Mohan’s Knowledge Framework; note the linksbetween the conceptual units of the Knowledge Framework (knowledge structures) and a) theconceptual units of the Smith, Meux studies (logical operations. ventures and moves), and b)formal schemata in reading and writing research, which, together with the KnowledgeFramework, marked a return in the mid-1980s to an emphasis on the cognitive aspects ofclassroom teaching. Methodologically, we look at the origins of discourse analysis; how ithas concentrated on the interpersonal and procedural rather than the cognitive aspects theclassroom discourse; and how it identified the I-R-E (Initiation, Response, Evaluation) modelof classroom teaching. We end with a summarization and some conclusions.78History of research neglect of cognitive observational studiesNeglect of the Smith. Meux studiesTwo classroom observation studies by Smith, Meux, Coombs et a!: A study in thelogic of teaching (1962) and A study in the strategies of teaching (1967) are amongst theearliest and, perhaps, the most extensive in the literature. We will refer to them from thispoint on as Logic and Strategies. Although frequently cited, they have received little detailedattention. For example, Miller (1976) refers to the hundreds of observation schedulesdeveloped in the 1960s and 1970s and, although he includes the Smith, Meux studies in hisreferences, he omits them from his list of significant examples:Withall’s Social-Emotional Climate Index (1949)Medley & Mitzel’s Observational Schedule and Record - OSCAR (1958)Flanders’ Interaction Category System (1960 and 1970)Bellack and colleagues’ Pedagogical Moves (1966)Taba’s Teaching Strategies and Cognitive Functioning (1965)This neglect may be because the Smith, Meux studies focused exclusively on thecognitive aspects of teaching:Teacher behavior is many-sided. For this reason, we have been forced to select someaspect of it for study rather than attempt to analyse the behavior in its entirety. Wehave, therefore, concentrated upon the cognitive aspects of teaching behavior and haveignored those behaviors that have to do with the affective domain and with suchelements as reinforcement and discipline (Smith, Meux, Coombs et al. 1967:3).Subsequent research neglected the cognitive content of teaching in favour of two otherstreams: the behavioural aspects of teaching from the mid-1960s on; and additionally, thelinguistic analysis of classroom discourse starting in the mid 70s.79The Behavioral research streamNuthall, who was on the Smith, Meux, Coombs Strategies of teaching team, andAlton-Lee (1990:549) suggest that classroom-interaction researchers were influenced to movetowards a study of teacher behaviour in the classroom by the challenge mounted in the 1960sby the application of behavioral psychology to education:Teaching machines, programmed instruction, and behavioral objectives were the productsof that application. The behaviorists claimed that, compared with their own scientificallybased achievements, classroom interaction research was a case of the blind leading theblind. . .Teaching needed reforming, not studying.Accordingly, classroom-interaction researchers felt the need to show that suchvariables as teacher questions, lesson structuring, and patterns of teacher-student interactionhad a causal relationship to student learning. Process-product research became the order ofthe day with the assumptions that 1) teacher behaviour (the process) causes student learning,and 2) the amount of student learning (the product) measures teacher effectiveness. Theresearcher’s task was to observe various teacher behaviors and correlate them to student scoreson achievement tests. As in behavioral studies, the focus was on stimulus and response -- theteacher’s stimulus and the student’s response -- with no interest in the “black box” ofcognition in between. The findings of Smith, Meux et al. were irrelevant to this research.Tom (1984) notes that this line of research eventually ran its course. The initial claimsof researchers that the process-product approach had isolated the key teacher variables thatcause student learning (Rosenshine and Furst 1971) were sharply reduced in the face ofcritical challenges (Rosenshine 1979). Critics particularly called into question the top-downapproach of this research with its emphasis on the teacher rather than student and its relianceon aggregated data and average scores, which concealed individual student differences. By the80mid-1980s, despite a number of experimental studies which attempted to hold constant theteacher variable (Nuthall and Church 1973; Gage 1985), researchers were unable to reach aconsensus about the value of this research paradigm. Some (Tom 1984, 1985; Shulman 1986)expressed doubt that it had any positive value at all.Nuthall and Alton-Lee (1990) speak of these doubts and differences as a crisis ofconfidence in research on teaching. They suggest a new direction for research that, inresponse to earlier criticisms, will focus on learning rather than teaching and on individualstudents rather than aggregated data. Research in this new direction may indeed yield valuableresults. The focus, however, is still on the relation of behaviours to learning under variouscontextual conditions and not on the cognitive processes employed.The Linguistic research streamSinclair and Coulthard (1975), who pioneered the use of classroom discourse forlinguistic analysis, were not educationists but linguists. Their interest was “not to throw lighton strategies of teaching and learning, but to provide linguistics with an impetus and apreliminary model for the systematic study of discourse” (Edwards & Westgate 1987:139).They chose the classroom as a setting because classroom discourse has a clear, overtstructure, topics of discourse are well defined and one person (the teacher) has acknowledgedresponsibility for directing the discourse (Sinclair & Coulthard 1975:6). Their work stimulateda stream of research, taking as its focus the role of language in structuring classroominteractions. Edwards and Westgate (1987:5) review this line of research into classroom talkand summarize its focus as “a concern with the communicative consequences of transmittingknowledge, and a concern for the often limited and limiting quality of language experience81which schools offer children”. They criticize earlier classroom interaction studies, takingFlanders (1970: 135) as representative, for omitting an analysis of language:Each focuses upon behaviour involving or giving rise to language rather than uponlanguage itself, on control of the topic or setting rather than on the structure of thediscourse. There is no linguistic analysis of the behaviour at all.They make only the briefest of references to classroom interaction studies with acognitive focus. They dismiss the Smith, Meux studies in one short sentence (146): “Earlystudies drew more on logic than linguistics (for example, Smith and Meux, 1970)”. Bellack(1966) also receives a one-line mention with credit for being one of the first to use verbatimtranscripts of classroom talk (overlooking that Smith Meux (1962) had used verbatimtranscripts some years earlier). The cognitive aspects of teaching were considered no morerelevant in the linguistic than in the behaviorist research stream.Research studies citing Smith. MeuxThe two studies -- Logic and Strategies -- address a deficiency of Process-Productresearch-- that it examines teaching strategies but does not look at the logical acts ofteaching. Smith, Meux introduced the notion of teaching as a manipulation of content forachieving stated objectives, using rational dialogue. Above all, they provided a richdescriptive natural history as a basis for future experimental research. No other studies of thisdepth followed on the cognitive aspects of teaching in the classroom; nor has there been anymajor follow-up of the issues raised. However, a number of studies have cited Logic and/orStrategies, sometimes in simply a historical sense and sometimes indicating being influencedby them.82Early linking of teacher behaviour and student achievementWright & Nutgall (1970) -- Nutgall was one of the researchers on the Strategies study-- sought to identify significant relationships between student performance on an achievementtest and the sorts of teacher behaviours identified in Bailiwick, et al (1967), Smith, Meux, etal (1962), Tab., Levine and Elsey (1964), and Nutgall and Lawrence (1965). Despite theciting of Smith, Meux, the results were expressed in tenns of teachers’ operational, notcognitive, strategies: the teachers producing greater pupil subject knowledge were those whoused direct closed questions, provided informative summaries after each episode of discussion,involved more pupils by redirecting questions, frequently thanked pupil responses, andprovided revision at the end of lessons.This study is an example of process-product research which sought links betweenteacher behaviour and student outcomes. It came under criticism for the inability ofcorrelational studies to hold variables constant (e.g. characteristics of teacher and students,physical environment etc.) in natural settings as distinct from controlled laboratory conditions(Tom 1984; McKeachie 1984). Nutgall himself in Nutgall and Lee (1990) recognizes anddocuments these criticisms in a review of teaching and learning research over the precedingthirty years.Part of “a proliferation of instruments”A number of researchers have referred to the proliferation of observation systems inthe late 1960s and the 1970s (Rosenshine & Furst 1973; Dunkin & Biddle 1974; Miller 1976;Galton, 1979). Most of these systems have an affective, work process or behavioural but notcognitive orientation. Some of the more significant observation systems were brought together83by Simon and Boyer (Eds., 1967, 1970, 1975) in an anthology called Mirrors for Behaviour.Over half of the 79 systems in the 1970 edition of Mirrors were derived from Flanders’Interaction Analysis System (Flanders, 1964) which has an affective and behaviouralorientation and has proved to be the most popular of all observation systems (Galton 1979).Mirrors classifies the systems it lists under four headings: Affective Domain,Cognitive Domain, Work Process (Control), and Behavior. The Affective Domain refers to theaffective climate of the classroom as indicated by teachers’ reaction along a continuum fromrejecting to accepting students’ ideas (cognitive output), feelings (emotional output), attemptsto manage classroom procedure, and non-verbal behaviours. Few systems deal only with thecognitive domain and they tend to be more complex. Those listed include both Logic andStrategies. However, as Nutgall and Church (1972) note, the “cognitive domain” in Mirrorscan be subdivided to distinguish between instruments with “special theoretical concerns” andthose directed towards “linguistic analysis”. Logic and Strategies are concerned with thetheoretical concerns (the cognitive in the sense of logical operations applied to lessoncontent), not with linguistic analysis.Miller (1976) observes that most observational systems concentrate on the recordingand coding of teachers’ rather than students’ talk and behaviour. He advocates a return to arelatively unstructured yet careful observation of students’ activities. He recognizes, however,the need for some system of classification of teacher and student behaviours to facilitate theformulation of working hypotheses that, in turn, may lead to further observations underexperimental conditions. The Logic and Strategies studies had made precisely that point. Healso advocates the use of audio and video recording to supplement the human observer and84contribute towards as complete a record as possible of both teacher and student talk. Logicand Strategies had also pioneered the use of recording methodology. He cites both Logic andStrategies but not in connection with the points we have just noted but for their use ofepisodes of classroom speech as units of reporting.Dunkin (1976) speaks of “the arduous task of reviewing a large body of systematicobservational research” which he had completed in Dunkin and Biddle (1974). More than athird of the 200 studies reviewed use a classification of classroom behaviour based onFlanders’ (1964) Interaction Analysis. The many other approaches are each represented byonly a handful of studies, which makes it difficult to assess their contributions to acumulative understanding of the effects of teaching. Dunkin speaks of “a tremendousdevelopment in the realization and representation of the patterns and complexities ofclassroom behaviour,” and acknowledges advances in the sophistication of research designand data gathering techniques. On the other hand, he deplores the conceptual confusionresulting from the lack of uniform terminology and the use of the same terminology forvariables that are conceptually different. He compares the cognitive systems used by Smithand Meux (1962) (Logic) and Bailiwick Ct al. (1966), and analyses the meanings of the terms“defining”, “explaining”, “evaluating” and “opining”. He concludes that only “opining” hasapproximately the same meaning in both studies and that “it would, therefore, be invalid toconclude that either study had provided evidence additional to the other’s concerning theremaining three terms” (179). It would seem more appropriate to say that Bailiwick’s studyhad not provided evidence additional to Smith and Meux’s as Logic was the first study of thetwo by several years. Dunkin deplores the lack of compatibility in the large number of85observation systems used. Galton (1979) echoes his frustration because lack of consistencyand compatibility has made this large body of work of much less accumulative value thanmight otherwise have been the case.A task analysis analogMoore (1981) makes an interesting connection between task analysis and the analysisof language interactions in schemes like Logic. He led a three-year research project at theSchool for External Learning (SEL), an experience-based high school program in a large city.He and his team were seeking some form of organizing principle for the analysis of cognitivedata in the performance of tasks. They rejected various psychological schemes generated fromlaboratory studies because they were not generalizable to natural settings. In reviewing theliterature on teaching and learning, they came upon the body of work on the structure oflanguage interaction in classrooms, including Smith, Meux et a!. 1962 (Logic), Bailiwick et a!.1966, Tab. 1966, Sinclair & Coulthard 1975, and Mehan 1979. Moore (291) comments:“Several of the studies suggest categories for classifying ‘moves’ made respectively byteachers and students in the stream of classroom discourse”.Of the studies cited, only Smith, Meux deals exclusively with cognitive discoursewhile Bailiwick and Tab. deal with cognitive along with affective and procedural discourse.However, Moore went in a different direction, of interest to our case study because of ourtutors’ emulation of the traditional I-R-E classroom discourse pattern. He and his team wereintrigued with the frequency with which the I-R-E sequence (Mehan, 1979) -- initiation bythe teacher, response by the student and evaluation by the teacher -- appears in theclassroom discourse transcripts they were examining and were puzzled by their inability to86find the same sequence in the discourse transcripts from their resource sites. Moorecomments: “We found . . . this form does not appear often in natural speech -- there issomething peculiarly ‘teacherly’ about it.” What they did discover, however, was a structuralanalog of the I-R-E sequence in a sequence, not on the level of speech alone, that they calleda “task episode”:The task is established for the student. The task is accomplished (or not) by the student.The student’s performance on the task is monitored or processed. . . . Task, therefore, tookthe logical place of “cognition” in our original problem, as it had the advantage of beingan observable phenomenon (Moore 1981:291-2).Moore proceeds to develop a “rational pedagogy of experience” in which thedimensions of tasks and social means together form pedagogical strategies. He offers this as away of better understanding the process of education in non-classroom settings. We willsuggest in reviewing task-based language teaching later in this Section that Moore’s notion isalso applicable as a pedagogical structuring device in classrooms and that, often, the closerclassroom tasks are to the workplace setting, the more meaningful and hence more effectivethey are.Relevance in math and science educationBishop (1982), whose orientation is towards constructive altemativism (Kelly, 1955),looks to educational research not for “results” which should be “applied” but for helpfulconstructs that will sensitize student teachers towards problems and help them generateexperiences that will broaden horizons both for them and their students. He distinguishesbetween the classroom of teaching-methods research, which tends to be an extension of thelearning laboratory with a focus on individualized learning, and the real life mathematics87classroom which is, above all, a social arena. He cites Smith, Meux (1970) (Logic) as anexample of researchers who have chosen the classroom as their research site and haverecognized the centrality of discourse for understanding what is going on:Language has traditionally been of interest and researchers have studied aspects fromthe level of questions being used (Friedman, 1976) to the whole logic of classroomdiscourse (Smith and Meux, 1970), an area which could have tremendous value formathematics education but which has not yet been exploited systematically (125).Bishop advocates moving much of the work in teacher training in mathematics fromthe general educational theory context to the mathematics-education context because of theunique character of much of the discourse in the mathematics classroom. Hirst’s theory offorms of knowledge and the findings in Strategies of predominating patterns of moves indifferent subject areas lends support to this view.Gordon (1984) modified the methodology of Logic for a special purpose. He wantedto identify the image of science as part of the “hidden curriculum”: that which is transmittedunintentionally by teachers and picked up unconsciously by students over a very long period.Direct evidence to support this thesis could be established only over a period of many years.Gordon therefore decided to gather indirect evidence in the form of highly redundantmessages transmitted in the hidden curriculum and to take redundancy as prima facieevidence that these messages had been transmitted successfully. He used the same sort ofcategory system as developed by the Logic study but modified to suit his special needs:The category system was used to analyze all teacher verbal behaviour excluding classmanagement and discipline issues. For most of the categories the unit of analysis wasthe lesson’s logical episodes, although the notion of an episode used resulted in far88shorter sections of discourse being defined as episodes than in Smith’s (1960 [sic])pioneering work in this field (Gordon 1984:394).Later linking of teacher behaviour and student achievementWe started this review of research studies citing Smith, Meux with an early process-product study (1970) by Wright and Nutgall linking teacher behaviours and studentachievement in science lessons. Needels (1988), nearly twenty years later, recognizes themany criticisms of the traditional process-product approach, e.g. Macmillan & Garrison(1984), Erickson (1986), but believes that many of these criticisms can be addressed so thatvalid relationships between teacher actions and student outcomes can be established bybuilding safeguards against inappropriate generalizations into a study design. Of particularinterest to us is her attention to an additional issue, the quality of instruction. In particular,She finds research on logic in education lacking in its assessment of quality of discourse.For the most part, research on logic in education has been conducted on the followingthree areas: (a) the kind of instructional treatment that best increases students’performance on a post-test of formal logic, (b) the development of logical reasoning inthe child. . . and (c) the kinds of logical operations used during classroom instruction(e.g. Bailiwick et al. 1966; Smith & Meux 1962 (Logic); Wright & Nutgall 1970).Only the last of these kinds of research was concerned with the kinds of logical movesused by teachers in a particular subject area. . . These studies, however, failed toassess the quality of discourse. This failure is a major weakness because to understandthe logic of the discourse, the quality of the segment must be assessed (Needels, 1988,p. 505).In other words, if teachers waffles or uses poor syntax when attempting to clarify a statementor give an explanation, they are more likely to confuse than to enlighten the student.89Needels therefore uses communicative logic in her study as distinct from formallogic, following Grice (1975) and his Cooperative Principle with its four categories:Quantity, Quality, Relevant, and Manner. She established six variables: 1) confused syntax; 2)omitting necessary definitions; 3) omitting causal factors; 4) irrelevant use of words; 5)Incorrect use of words; 6) incorrect causal relation. Based on this analysis, she found apositive relationship between the quality of teacher discourse and student achievement,especially for lesson content at a high level of cognitive complexity.This adaptation of the methodology of Logic (she does not refer to Strategies) toestablish communicative logic categories is constructive and addresses the issue of quality ofteacher discourse which neither Logic nor Strategies does. However, Needel’s notion ofquality seems applicable more to one way teacher discourse in an I-R-E mode than tointeractive discourse from which the participants construct meaning.Link with natural language interfaces for expert systemsKellerman et al. (1989) found the concept of “venture” in Strategies valuable inproviding conceptual support for the notion of “scenes” as a higher level unit than “scripts” asused in natural language programming for expert systems and intelligent tutoring systems.They use Schank’s (1982) concept, Memory Organization Packet (MOP), as an addition to theconcepts “scripts” and “scenes”. “MOPs are knowledge structures that organize scenes so thatsome higher level goal(s) can be accomplished” (30) and replace scripts as an organizingprinciple of memory. Scripts are redefined as short and containing only those actions thatdetail general actions specified by the scene: e.g. in an introduction scene, one script willinvolve the specific actions for shaking hands, another the actions for revealing the90relationship with the conversation partner (“we’ve met before” or “I’m your student”) etc. Thescene, therefore, is the independent memory unit of a MOP and represents a grouping ofscripts for cross-situational and situation-specific actions. Whatever becomes a scene musthave a discourse topic and an associated instrumental goal. Scripts will be brought into playto realise that goal. Kellermann et a!. comment on the similarity between their concept of“scene” and the concept of “venture” in Strategies:Smith, Meux, Coombs, Nutgall, and Precians (1967) have identified what they term a“venture” as a segment of discourse consisting of a set of utterances dealing with asingle topic and having a single overarching content objective. The venture is believedto be the equivalent of the scene in the informal initial conversation MOP. Greetingsserve the goal of recognizing each other, introductions serve the content objective ofexchanging identifying information, and each topic serves as an instrumental goal oflocating commonalities, extracting information, and the like. (33).If “ventures” are the equivalent of “scenes”, it would seem reasonable to suppose that“moves” are the equivalent of “scripts” or, at least, closely resemble them. Moves, like scripts,specify a number of limited, detailed alternatives at the disposal of the conversation partnersin the venture. Kellermann et al. find that Strategies does what discourse analysis in generaldoes not do -- provides a hierarchical as well as a linear structure to the discourse: “Thus, theseeming profusion of discourse units and their resultant implications for discourse structures ismore a function of a linear model of discourse, without as consistent a concern for itscorresponding hierarchical nature” (28). Just as the MOP is made up of scenes which containscripts so the Strategy is made up of ventures which contain moves.91SummaryThe Smith, Meux methodologies in Logic and Strategies were noted in reviews of the“proliferation of instruments” for classroom observation in the 1960s and early 1970s. Theywere also included in successive editions of Simon & Boyer’s Mirrors for Behaviouranthology (1967-75). These, however, catalogued methodologies as distinct from using them.There was no full scale follow-up on the cognitive aspects of classroom discourse to explorethe questions that Smith, Meux raised: in particular, the effect that explicit understanding oflogical operations in classroom discourse may have on the quality of students’ thinking andacquisition of knowledge, and of teachers’ performance and their command of subject matter.The Smith, Meux studies did, however, establish units of analysis and logicalcategories that proved useful to some subsequent research: Wright & Nutgall (1970) used the“logical operations” of the Logic study, together with other units, to study links betweenteacher behaviour and student achievement. This was, however, behaviourally rather thancognitively oriented. Moore (1981) found the concept of “move”, used by Smith, Meux andothers as an organizing principle for analysing classroom discourse, a helpful analog fororganizing task work in experienced-based instruction in non-classroom settings. Gordon(1984) modified the Smith, Meux unit of “logical episode”, although more to analyze scienceteachers’ verbal behaviour in communicating “hidden” science curriculum than in analyzingcognitive exchanges with students. Needels (1988) used an imaginative adaptation of “logicaloperations” in devising a system of communicative logic to measure the quality of teachers’discourse. This was a valuable complementary study to Smith, Meux but was not a directfollow-up. Kellerman et al. (1989) found the concept of “venture” in Strategies valuable in92providing conceptual support for the notion of “scenes” as a higher level unit than “scripts” asused in natural language programming. This work, although highly cognitive, lies in adifferent field than the cognitive aspects of classroom discourse.Therefore we may say that although the constructs devised in the Smith, Meux studiesprovided conceptual support and/or leads to various researchers, there was no effectivefollow-up on the line of cognitive classroom research they pioneered.Concepts for approaching the cognitive aspects of teachingThe importance of the cognitive aspects of teachingA major weakness in the process-product research paradigm and in the behaviourismfrom which it sprang is an implicit denial of the autonomy of students (and, for that matter,of teachers too). As Smith, Meux (1962:5) point out: Thorndike, in his studies of animalintelligence, reduces thinking and learning to a stimulus-response model. From this it isconcluded that classroom teaching is simply a matter of the teacher supplying correct stimulito students and reinforcing correct responses. All that matters is observable behaviour.Because mental processes are not observable, laws of logic and modes and procedures ofcognition are irrelevant. Dewey, at the same time, was developing a totally different nonbehaviourist notion that learning resulted from reflective thought and logical enquiry elicited(i.e. motivated) by the need to resolve perplexing situations. “Between the perplexing situationand its resolution, Dewey put reflective thought. Thorndike on the other hand, filled the gapbetween stimulus and response with neuron connections” (6). For Dewey, the rules of logicare the rules of successful inquiry and therefore normative for the resolution of perplexingsituations. The psychologists, however, ignored these normative aspects of Dewey’s theory of93learning and, emphasizing his psychological elements, developed a psychologized version ofhis theory of logic which drew upon motivation rather than cognition theory. The way theyhad it, students do not think their way but behave their way out of troubling situations by trialand error, motivated to act by discomfort and guided and reinforced by the teacher. From thiswas born the notion of learning by the problem (discovery) method: “In this version ofinquiry and teaching, there is no distinction between valid and invalid thinking, for suchdistinction cannot be made within this kind of psychological analysis of problem solving” (7).Smith, Meux seek to redress the balance. They hold that learning results from theexercise of autonomous thought, not from automatic behavioural responses. They believe thatlearners will benefit in terms of increased knowledge and understanding if they can be madeaware of these cognitive processes operating in various situations. Because their study isdescriptive and not experimental, they express these potential benefits in the form ofquestions to be addressed by subsequent experimental research:• . . is the student better able to monitor his own thinking as well as that of others ifhe has knowledge of the logical structure and rules governing the performance of these(logical) operations? Does an explicit understanding of these operations increase thestudent’s knowledge of the subject he is studying? Does the more rigorous performanceof these operations require that the teacher have command of his subject matter inways different from that which he ordinarily possesses? Would the ability of a teacherto explore subject matter logically free him from overdependence upon the textbook?Should answers to these questions be affirmative, would not knowledge about theselogical operations and how to perform them constitute a new content for courses inpedagogy? (Smith, Meux, 1962:10-il)Mohan (1986, 1989) also is concerned with increasing the learner’s knowledge andunderstanding through awareness of cognitive processes, in particular those underlying94academic discourse and expository texts. He is concerned particularly (but not exclusively)with students learning in a second language (L2). He points out the significance of Cummins’(1991) distinction between conversational and academic language proficiency, particularly inthe length of time indicated for mastery: 1-2 years for conversational, 4-8 years for academicdiscourse. Immigrant students are “mainstreamed” after two years (or less) of second languagestudy into the native speakers’ content classes on the strength of coping satisfactorily withinterpersonal communication only to find themselves unable to understand the far morecognitively demanding academic discourse and expository texts with which they areconfronted. As a result, they do poorly academically. Because, supposedly, they are competentin the language of instruction, their poor academic performance is attributed either to deficientcognitive abilities or poor motivation. Cummins (1984:143) makes the point that L2 studentsbring with them “an underlying cognitive/academic proficiency which is common acrosslanguages”. Working in their own language, they already have a store of conceptual knowledge and experience in the cognitively demanding tasks of analysis, synthesis and evaluation.Mohan’s goal is to enable them to draw on this native proficiency by making them aware ofthe Knowledge Structures (KSs) underlying the academic discourse and expository textspresented to them in L2. This is made easier because KSs can be represented nonverbally:• KSs appear across modes of communication and understanding, being expressedverbally or non-verbally. For instance, classification can be expressed not only throughwriting and reading, and monologue or dialogue, but also across nonverbal modes ofcommunication, such as graphics and database programs (Mohan 1989:104).Heightening cognitive awareness in this way should not only help L2 students come toterms with cognitive discourse and academic texts but also help Li students who have95problems in these areas. Heightened cognitive awareness through teaching text structure hasalso been shown to benefit reading comprehension of expository texts for Li as well as L2students (Carrell 1985; Meyer 1985). Clearly, from the points of view both of teachingmethods and student performance, there appear to be good reasons to repair the neglect ofresearch into the cognitive aspects of teaching.A rationality model of teachingWe have referred to the Smith, Meux studies as early examples of cognitive classroomstudies. We will describe them more fully and show their relationship to Mohan’s (1986)Knowledge Framework. Before doing this, we should first examine the theoretical base forcognitive research in the classroom. This lies in a rationality model of teaching.A rationality model definedCommon to the analyses of teaching and education by Bantock (1965), Peters (1966),Scheffler (1966) and Green (1971) is an emphasis on cognitive transactions conducted in amorally acceptable manner. We may call this a rationality model of teaching. Green allowsexplicitly for those activities that shape behaviour (knowing how) as well as those that shapeknowledge and beliefs (knowing that). However, as in “knowing that” he distinguishesteaching-instructing from indoctrination, so in “knowing how” he distinguishes teachingtraining from conditioning. In both cases the distinction depends upon the cognitive elementof students’ intelligent response. That response must occur to consummate the act of teaching.Some critics, e.g. Cooper (1966) against Scheffler (1960), have charged thatconceptions of teaching that fall under a rationality model fail to address the more affectiveareas of teaching, such as music, the visual and performing arts, and aesthetic appreciation.96Even in these areas, however, cognitive elements are important: students learn performing“know how” through first following procedures and understanding the reasons for them (as weshall see later when we analyze students’ “hands-on” discourse); and they learn to makeaesthetic judgments through first understanding what to look (or listen) for in a work of art.Moreover, a false dichotomy is often drawn between rationality and affect. One does notexclude the other. Many of our beliefs we hold with affective intensity. Some we may evenbe prepared to die for. And belief is without its affective element: the very act of believingis a move away from affective neutrality. We therefore take a rationality model of teaching asa point of departure as expressed in essence by Scheffler (1966:131):Teaching may be characterised as an activity aimed at the achievement of learning, andpractised in such a manner as to respect the student’s intellectual integrity and capacityfor independent judgment.Scheffler puts equal weight on two key aspects of teaching, both focused on thelearner. First is presenting the students with what is to be learned. This is more than thesimple transmission of information from the teacher to the students. It is also the presentationto them of rules of evidence and criteria of judgment for assessing the validity of what theyread and hear. Second is obtaining the students’ acceptance of what is presented to them.This means helping them to develop their capacities to use rules of evidence and applycriteria so that they exercise their independent judgment in accepting (or not accepting) whatis presented to them. The rationality model of teaching therefore implies epistemologicalunderstanding and moral commitment.Scheffler warns against equating the concept of rationality with some innate structureof the human mind. A universal demand for formal consistency is expressed concretely in97uniquely different ways in the different domains of knowledge, e.g. rationality in science isexpressed in different ways than rationality in history. He therefore advances a concept ofrationality that is one of rules and principles embodied in multiple evolving traditions indifferent domains rather than an abstract and general ideal. He concludes (114):Teaching, from this point of view, is not, as the behaviorists would have it, a matter ofthe teacher’s shaping the student’s behavior or of controlling his mind. It is a matter ofpassing on those traditions of principled thought and action which define the rationallife for the teacher as well as the student.Hirst and epistemo1oical structureScheffler does not elaborate on the nature of the traditions of thought and action forthe different educational domains. Hirst (1965) examines this issue from the point of view ofthe epistemological structure of what is taught in the classroom. He revisits the subject againin collaboration with R. S. Peters (Hirst & Peters, 1970); and answers critical comments andelaborates on a number of points in a later paper (Hirst, 1973). The essence of his position isthat within the domain of knowledge there are a number of different forms of knowledge thathave distinct logical characteristics. These, according to his 1973 classification, are:He includes history and the social sciences under “knowledge of minds” because theyare uniquely concerned “with explanations of human behaviour in terms of intentions, wills,hopes, beliefs, etc.” (Hirst 1973:86). They may share with the physical sciences a concern forempirical investigation and experimentation but this is not where their uniqueness lies. Hirstrecognizes that each form of knowledge holds concepts common to others. What he looks for98as a basis for classification is that set of concepts that uniquely predominates in a particulararea. Each form of knowledge, while sharing basic concepts with the other forms, has its ownunique conceptual structure, unique tests for determining truth claims and unique standards ofgood evidence for assessing truth claims.He refutes the idea that “the relations between concepts and propositions in all formsof knowledge must conform to those of mathematical or scientific knowledge”. This, hestates, “is a matter of pure dogmatism” (90). Forms of knowledge are constructs of the humanmind and are constantly changing. The present forms represent today’s objective judgment.Tomorrow’s objectivity may be different. He also denies the existence of an area of commonsense distinct from the particular forms of knowledge: “Common sense knowledge is to mymind simply that collection of elementary knowledge . . . from the different forms, which islargely taken for granted in a given society” (90).Hirst’s analysis of the cognitive content of teaching by forms of knowledge contributestowards understanding what is going on or what should be going on in the classroom. He seesno necessity for structuring the curriculum in terms of his forms of knowledge although hesuggests that from a logical point of view there may be advantages in so doing. He does notclaim to have said the last word on how many forms or sub-forms of knowledge there shouldbe: e.g. we could say that the category “physical sciences” is too broad as it includes physics,chemistry and biology, which all have different intellectual traditions. He also recognizes thatlarge areas of moral, religious and aesthetic teaching lie outside of propositional knowledgeand therefore outside of the forms of knowledge bearing those names (88).99Smith. Meux and the Logic of TeachingHirst discussed the epistemological structure of the knowledge that should be taught inthe classroom but said nothing about the process of teaching it. Smith, Meux (1962) in Logicset out to fill this gap in educational research. They state at the outset: “It is just as necessaryto understand the phenomenon of teaching as a condition of applying ideas and principles toit as it is to understand the ideas and principles themselves” (1). As we saw at the beginningof this Section, researchers had made many observational studies of classrooms and developedvarious observation schedules. However, these had looked primarily at affective elements inteacher-student interactions and at various aspects of classroom management. No study hadmade a thoroughgoing record and analysis of the rational, cognitive content of teaching.Logic was the first. It was also one of the earliest, if not the earliest, major study to makeextensive use of tape recordings of classroom discourse to obtain an exact record of classroom verbal interactions which could be analysed and reanalyzed as many times as required.Logic captured teacher-student discourse in 17 classes from grades 9-12 in English,Math, Science, and History-Social Studies. It was neither an evaluative nor an experimentalinvestigation of teaching but rather “an analytic and descriptive one in the natural historysense” (8). Its purpose was to observe, describe and classify logical behaviour (governed byepistemic rules) in the classroom in terms of logical operations in teacher-student interactions.By “logical operations”, which are the focus of our study, we mean the forms whichverbal behavior takes as the teacher shapes the subject matter in the course ofinstruction. For example, the teacher reduces concepts to linguistic patterns calleddefinitions; he fills in gaps between the student’s experience and some newphenomenon by facts and generalizations related in a verbal pattern referred to asexplanation; he rates objects, events etc., by reference to facts and criteria related in a100pattern called evaluation. If he does not engage in such operations himself, the teachereither requires his students to do so, or more typically, the teacher and his studentsjointly carry on these operations through verbal exchanges (3).These logical operations take place in units of discourse called episodes, which beginwith an expression that triggers a verbal exchange about a topic and end with completingdiscussion of that topic. The logical operations identified from these opening expressions andused in the study are shown in Table 2.1.Table 2.1 Smith, Meux Logical Operations“Describing” was also identified but found too complex with too many sub categoriesto include in the analysis. “Stating” and “reporting” were identified but excluded fromanalysis because they were found to be of little pedagogic significance. “Substituting” wasidentified but excluded from analysis because it occurred in very few episodes. Describing,Designating and Explaining were the three most frequently recurring operations, in that order.Substituting, Reporting and Classifying were the three least frequently occurring operations.The logical operations occur in different patterns in the different subject areas: physiology is more concerned with Defining and Designating; physics with the use of symbols andConditional Inferring; biology and chemistry with Evaluating. Samples were too small for101drawing firm conclusions. Results, however, suggest that differences may exist in the use oflogical operations among teachers and subject areas for examination in future studies.A number of interesting questions are raised, going beyond the purpose of the studybecause they are experimental in import. Among these are:Is the student better able to monitor his own thinking as well as that of others if he hasknowledge of the logical structure and rules governing the performance of theseoperations? Does an explicit understanding of these operations increase the student’sknowledge of the subject he is studying? Does the more rigorous performance ofthese operations require that the teacher have command of his subject matter in waysdifferent from that which he ordinarily possesses? (10)These questions were not followed up by researchers in the following two decades. Theyanticipate the work of Carrell (1985) and Meyer (1985), working with text structures andMohan (1986), working with knowledge structures (KSs). These researchers use logicalanalysis of content in different ways to enable students to increase their comprehension ofexpository and narrative texts and academic discourse. In addition, Christie (1989) has usedgenre analysis and Mohan has used KS analysis to facilitate the writing process.Logic notes at the outset that theories of teaching are based on conceptions of logic orpsychology or both. Although the study sought to establish each logical category based uponthe kind of rule a person would follow in an ideal response, it concludes with the observationthat some categories are more clearly rule-governed than others. Particularly troublesome are“opining” and “describing”. There are no well formulated rules for “opining”:it seems to make more sense to speak of defining well than it does of opining well. Asimilar point can be made . . . with describing or reporting one’s feelings. What rulesdoes one follow in answering questions such as “How do you feel about snakes?” and“Do you have a headache?. . . On the other hand, even in responding to Opining102entries the student is guided by some rules, though they are quite general” (194).The study concludes by suggesting that the categories used could be arranged on alogical-psychological continuum according to the extent to which an ideal response isgovernable by rules. This concession recognizes the inherent untidiness of real-life discourse,which resists neat categorical analysis.The study also deplores the then narrowness of the fields of logic and psychology ofthinking. It suggests ways of broadening them, particularly through a study of logical patternsin ordinary language and in the languages of the different domains; and through aninvestigation of different kinds of logical vs non-logical behaviour. The suggestions wereprescient, preceding as they did the stream of linguistic research that sprang from Austin’s(1962) work on Speech Act Theory, including, notably, Sinclair and Coulthard’s (1975) workon Discourse Analysis; and the stream of research in cognitive psychology that flowed fromthe new interdisciplinary field of Cognitive Science, lying more than a decade into the future.Smith. Meux, Coombs and the Strategies of TeachingSmith, Meux, Coombs et a!. (1967) conducted Strategies as a follow up to Logic.Strategies reanalyzed the recorded data from Logic with a focus on the larger maneuverscontrolling the subject matter of instruction, called Strategies. The term “strategies” wasintroduced in Logic as a pattern of actions concerned with attaining certain outcomes and,hence, directly related to objectives. It was not, however, the focus of analysis as it is in thenew study. Strategies are made up of units of instruction, called Ventures, which arecategorized according to overarching objectives. Ventures are further subdivided into verbalMoves, logical (mostly) and psychological, which constitute the basic elements of the103strategies of instruction. Moves can be viewed statically as units of content (i.e. as products)or dynamically as units of manipulation (i.e. as processes). They reflect the basic discourseschemas that embody the essential features of each venture’s objective.Strategies, like the previous Logic study, is descriptive and focuses on the cognitiveaspects of teaching as distinct from the affective domain and such aspects of teaching asreinforcement and discipline. The eight types of venture identified, with their overarchingobjectives and typical moves are shown in Table 2.2:As with all classification schemes, there are problems with boundaries and overlaps.The Causal and Reason categories are not customarily distinguished in the language and socan be easily confused: “Reason ventures are concerned with considerations a person takesinto account in deciding to take some action. These considerations do not count as causes”(100). The Conceptual and Particular categories strongly resemble each other: “The majordifference between the two is that conceptual ventures discuss the characteristics of a class ofthings whereas particular ventures discuss an individual object, event, person or place” (243).Although some 105 different types of move are identified, with criteria developed for eachone, characteristically only one or two moves predominate within each venture.104Table 2.2: Smith, Meux Strategies of TeachingVentures, Objectives and MovesA rule or several related rules for conventional ways of doing things or foranalytic relations to guide actionsRule formulationRule justificationRule applicationA cause-effect relationship betweenparticular events or classes of eventCause describingEffect describingRelational (14)Reasons for an action, decision, policy or Action elaborationpractice. Purpose elaborationRule centredFactual consideration (10)A set of conditions governing or implied Descriptiveby use of a term and constituting criteria Comparativefor determining membership in the class Instantialof things referred to by the term Usage (15)A body of information which clarifies or Descriptiveamplifies a specific topic or group of Explanatory: related topics Appraisal (9)A rating of an action, object, event, Identificationpolicy, or practice DescriptionRatingCriterialRelationalTangential (20)The meaning or significance of a set of Meaningwords or symbols ExtrapolationFactual elucidationCitationRepresentationEvidential (18)A sequence of actions by which an end Problem centredmay be achieved Performance centredProcedure centred (10)* Macro-level moves are broken down into 105 detailed moves as indicated by the figures inparentheses(9)105Strategies is a deeper analysis of the classroom discourse than Logic. The logicaloperations identified in Logic, governed by epistemic rules, are embedded in the Moveswhich are part of the larger objective-oriented Ventures. The Ventures are governed byinstrumental rules. This analysis foreshadows Green’s (1971) analysis of the logical acts andthe strategic acts of teaching.In a similar way as Hirst identifies unique conceptual structures for his forms ofknowledge, so Strategies identifies certain Ventures which, although not unique, are morecommon in certain subject domains. Conceptual and Causal Ventures occurred mostfrequently in the sciences. Reason Ventures, being more closely associated with humanactions, occurred predominantly in History. Particular Ventures also predominated in History.Interpretive Ventures occurred predominantly in English, mostly in connection withdiscussions about literature. Rule Ventures predominated in English and Mathematics.Evaluative Ventures predominated in English and History. The samples used are not largeenough for drawing firm conclusions but do indicate interesting trends.We turn now to a comparison of the Smith, Meux studies with Mohan’s work.Mohan’s Knowledge FrameworkMohan’s (1986) Knowledge Framework (Figure 2.1) is a research instrument andpedagogic device for the cognitive analysis of texts across the curriculum. Mohan argues thatthe six Knowledge Structures (KSs), Description, Sequence, Choice, Classification, Principles,Evaluation, contained in the Knowledge Framework are logical forms represented withinHalliday’s ideational discourse and, equally, are logical forms represented within theclassroom discourse contained in the Smith, Meux Ventures as defined in their 1967106Strategies study (Mohan, personal communication). A distinction is drawn between the threeKSs of action within specific situations and the three KSs of background knowledge whichrepresent the theoretical principles employed by the KSs of action.Figure 2.1: The Knowledge FrameworkBackground Knowledge (Theoretical)Osifiüon Principles EvaluationSequence ChoiceAction Siluation (Practical)Mohan’s KSs can also be described as high level schemas, a term Smith, Meuxapplied to the concept Moves, which they describe as reflecting the basic schemas underlyingdiscourse. Schema has become a somewhat general and variously used term. The OED datesits usage in English to 1839 from Kant and defines it as “any one of certain forms or rules ofthe ‘productive imagination’ through which the understanding is able to apply its ‘categories’to the manifold of sense perception in the process or realizing knowledge or experience”.From philosophy the term migrated to psychology and was used by Piaget to stand for theinfant’s stored patterns of perceptual-motor coordinations, i.e. schemata of action (“schemata”being Piaget’s preferred plural form). Cognitive psychology uses the term to stand fororganized chunks or packages into which knowledge is schematized to facilitatecomprehension, memory and application (Abelson & Black 1986). Cognitive psychology alsouses knowledge structures as an alternative term to schemas. These structures are flexible,not fixed, constantly changing with inputs of new experience. Mohan (1990:16-17) focuses ona few, broadly defined, high-level realizations:107Because the notion of schema is so general and bland, it is necessary to focus on certainclasses of knowledge structure [see Figure 2.1]. . . There are three pairs of relatedstructures: a description of a particular object or person often involves a classification orset of general concepts; a particular temporal sequence of states, event or actions ofteninvolves general principles (social rules or cause-effect relations) which relate one state toanother; a particular choice or decision often involves general values. These KnowledgeStructures (KSs) are broad and general patterns of the organization of information, at afairly high level of abstraction.Mohan (1987:507) has also been influenced by the anthropological research tradition and byHalliday’s systemic linguistics with the accompanying notions of text and situation:Halliday’s view of language as social semiotic (Halliday 1978, 1985), in the traditionof Malinowski and Firth, has two key terms: text and situation. Every text, spoken orwritten, unfolds in some situation, or context of use.He cites Malinowski’s (1935) description of Trobriand coral gardening, based upon texts oforal discourse. A large part of these texts is concerned not only with the pragmatic languageof action: description, sequence and choice; but also with the more complex language oftheory: classificatory distinctions (taxonomies) among various aspects of land, gardens, cropsand agricultural techniques; social, legal, economic and magical principles; and evaluations.Malinowski (1960) went on to claim that traditional cultural knowledge systems are similar instructure to the knowledge systems of modern science. This leads to the important claim thatthese knowledge structures are cross-cultural -- a significant factor in second languageteaching, which has to provide for students coming from a wide variety of cultural groups.The universality of basic knowledge structures has been well explored in cognitiveanthropology. Figure 2.2 shows Mohan’s mapping onto the Knowledge Framework of the108main types of multi-cultural ethnographic cognitive structure