Writing to Make Meaning or to Lean the Language?

A Descriptive Study of Multi-ethnic Adults

Learning Japanese-as-a-Foreign-Language

Sufumi So

A thesis submitted in conformi ty wiih the requirements

for the degree of Doctor of Philosophy
Department of Cumculum, Teaching and Learning
University of Toronto

O Copyright by Sufumi So 1997

u*m National Library
of Canada
Bibliothéque nationale
du Canada
Acquisitions and Acquisitions et
Bibliographic Services services bibliographiques
395 Wellington Street 395. nie Wellington
OttawaON K I A ON4 OttawaON KtAON4
Canada Canada
your iiy, votre refertmce

Our iUe Notre reference

The author has granted a non- L'auteur a accordé une licence non
exclusive licence allowing the exclusive permettant à la
National Libraty of Canada to Bibliothèque nationale du Canada de
reproduce, loan, distribute or sell reproduire, prêter, distribuer ou
copies of this thesis in microform, vendre des copies de cette thèse sous
paper or electronic formats. la fome de microfiche/film, de
reproduction sur papier ou sur format
électronique.

The author retains ownership of the L'auteur conserve la propriété du

copyright in this thesis. Neither the droit d'auteur qui protège cette thèse.
thesis nor substantial extracts fiom it Ni la thèse ni des extraits substantiels
may be printed or otherwise de celle-ci ne doivent être imprimés
reproduced without the author' s ou autrement reproduits sans son
permission. autorisation.
 Wn'ting to Make Meaning or to Learn the Language?
A Descriptive Shuly of Mzrlh'ethnic Aduits
Leaniing Japnesessa-Foreign-Lnngwge

Ph-D. Thesis
Department of Cumculurn, Teaching and Learning, University of Toronto, 1997

This study was motivated by the epistemological question of what "writing as a

mode of leaming" means in the context of foreign language instruction. I t analyzes how
Japanese-as-a-foreign-laquage (JFL) learners and their teacher deal t wi th the w n ting
cornponent in a natural instructional setting over a p e n d of two university semesters. The
thesis details how this biliterate situation functioned in terms of individual, psychological
constructs contingent on the learners' levels of JFL proficiency and their perceptions of
JFL writing and leaming.

In seeking to identily comprehensively a set of educationally-relevant factors that

may influence students' leaming through writing in JFL, 1 considered (a) the presage
variables in leamers' backgrounds and (b) the educational contest dong with (c) the
cognitive processes of writing during the learners' production of JFL compositions and (d)
the qudiiies of their J E written products. These data were genented through classroom
observations, interviews, questionnaires, tape-recordings of think-aloud sessions. and essay
assessrnents, involving 13 multi-ethnic students enrolled in an upper-level JFL literacy
course at a major anglophone university in Canada. The data were described, analyzed, and
interpreted through the combined use of quantitative and qualitative methods,
incorporating a unique statistical technique. dical scalirig (Nishisato, 1994).
 Multiple sources of evidence showed the superior-level and the advanced- and
intermediate-level Japanese-proficient speakers dernons trating di fferent patterns of
behavior. In general, the former group of students produced essays of better qudity in
terms of textual and cognitive structural features, and they focused on both semantic and
lexico-grammatical components to produce their written discourse, suggesting
opportunities for learning about content and language concurrently. 1n contrast, the
Japanese essays written by the latter group of students were consuained in their textual and
cognitive structural features, and their primary focus in their production of wri tten
discourse was on the lexico-grammatical system, probably limiting their learning to the
aspect of language use. The study suggests needs for further research particularly on L3
writing as a metalinguistic activity, L2 writing as a personal construct, and L1 -L2tnnsfer.
 When I started to work on this thesis research, 1 had no way of knowing what the

days ahead would hold in store for me. Upon the completion of the thesis, 1 now see the
entire thesis process to be exhilanting. But the experience was oftentimes daunting and
difficult. Many individuals directly or indirectly helped me to go through this challenge.

This thesis would not have k e n possible without the contn bution of Kei and her
students, who shared with me their classroom, their t h e , and their thoughts. 1 am deeply

indebted to them.

1 am also very much indebted to Alister Cumming, my thesis supervisor, whose

expert input was especially important to the development and completion of this work. In
fact, his ideas and works have influenced me enormously ever since 1 came in contact with
him five years ago. 1 have greatly benefited frorn numerous opportunities to work with him
at the Ontario Institute for Studies in Education of the University of Toronto (OISUUT). 1
also wish to thank Shizuhiko Nishisato and Sharon Lapkin, my thesis cornmittee members,

for their constant encouragement and useful comments.

T h e Collowing people assisted me in data analyses o r production o f the thesis:

Christine Guro, Mari Haneda, Mary Landolf, Jiang Li, Angel Lin, Hiromi Masurni-So.
Alfred Sakyi, Ling Shi, Billy So, Junko Tanaka, and Chihiro K. Thomson. Thank you dl.

Gai1 and Gerry Jordan, Hannele and Lloyd Robinson, Mayumi Kondo and her

family, and Lillian S o generously supported m y undertaking in more private ways

throughout the p e n d of my doctoral study at the OISUUT.1 am very grateful to them.
 1 am dso indebted to Martha C. Pennington whose input has k e n a constant source

of encouragement and inspiration for my academic work since 1 came to know her ten
years ago.

1 also acknowledge the financial support of the OISUUT in the forms of Gmduate

Assistantships (1992-96) and "high visa tuition waivers" (1993-95) as well as that of the
Ontario Ministry of Education and Training who awarded the Ontario Graduate
Scholarships (1994%) to me.

Last, but not, of course, least, 1 a m very thankful to my husband, Billy, who
constantly encouraged me to have confidence in my work,and to my son, Duane, who put
up with my hectic schedule and ternperarnental behavior throughout the writing of this
thesis.
 TABLE OF CONTENTS

CHAPTER 1: INTRODUCING AND FRAMINC THE STUDY 1

1.1. OVERVIEW 1
1.2. RATIONALE 2
1.2.1. PEDAGOGICAL CONSIDERATIONS 3
1.2.2. THEORETICAL CONSIDERATIONS 5
1.3. BIGGS' S PRESAGE-PROCESS-PRODUm MODEL OF C LASSROOM LEARNLlYG AND
TEACHING 7
1.4.OI&JECI'AND FOC1 OF THE STUDY 10
1.4.1. RECENT CONCEPTUALIZATIONSOF FOREIGN W G U A G E CURRICULA 11
1.5. ORIENTATION AND ORGANIZATION OF TEiE DISSERTATION 15

C H A P T E R 2: TEACHING CONTEXT 17

2.1. SITE
2.2. TEACHER
2.3. CENERATING RESEARCH MATERIALS
2-3.1. CLASSROOM OBSERVATION
2.3.2. INTERVIEWZNG
2.4. ANSWERING THE WSEARCH QUESTIONS
2.4.1. K EI AND CLASSROUMACTIVITIES LEADING TO WRlTZNG TASK I
2.4.2. K El AND CLASSROOMA CTIVITIES LEADING TO WRITING TASK II
2.4.3. SUiMMARY

CHAPTER 3: STUDENTS' CHARACTERISTICS 36

3.1. SOLICITING STUDErYT PARTICIPANTS

3.2. G ENERATING RESEARCH MATERIALS
3.2.1. QUESTIONNAIRE
3.2.l.I. PARTICIPANTS ' ETHNICITIES AND LI S
3.2.1.2. PARTICIPANTS ' AGES AND MMOR FIELDS
3.2.1 -3.PARTICIPANTS ' EXPERIENCES W T H THEJAPANESE LMGUAGE
3.2.1.4. PARTICIPANTS' MOTIVATIONS FOR LEARNING JAPANESE
3.2.2. ASSESSMENT OF JFL PROFICIENCY AND LI WRITING EXPERTISE
3.2-2.1.W RITING ASSESSMENT I NSTRUiMENT
3.2.2.2. WRITING ASSESSMENT R ATERS
3.2.2.3. PARTICIPANTS ' LANGUAGE-RELATED CHARACTERISTICS
3.2.3. INTERVIE WING
33. ANSWERKNG THE RESEARCH QWSTION
3.3.1. JANE'S AND COLIN'S PERCEPTIONS OF JFL WRITING AND LEAANING
3.3.1 -1. JAhrE
3.3.1.2. COLIN
CHAPTER 4: WRITING PROCESSES 55

4.1. GENERATiNG RESEARCH MATERIALS

4.1.1. COLLECTION OF THINK-ALOUD PROTOCOLS
4.1.1.1. REACTlVZTY
4.1.2. TRANSCRZPTZON OF THINK-ALOUD PROTOCOLS
4.1 -3.CODING OF THINK-ALOUD PROTOCOLS
4.1.4. POST-W RITING QUESTIONNAIRE
42. ANSWERING THE RESEARCB QUESTIONS
4.2.1. OVEUVIEW OF THINK-ALOUD PROTOCOL DATA
4.2.2. OVERVIEW OF POST-WRiTlNG QUESTIONNAIRE DATA

CHAPTER 5: WRITTEN PRODUCTS 71

5.1. GENERATING RESEARCH MATERIALS

5.1.1. ASSESSMENT OF JAfANESE WrUirTEN TEXTS
5.1 -1.I. KEI
5.1.1.2. INDEPENDENT RATERS
5.1 -2. KEI'S AND INDEPENDENT RATERS ' ESSA Y SCORES
5.13 . ASSESSMENT OF STRUCTURAL COMPLEXITY IN JAPANESE WRZiTEN TEXTS
5.2. ANSWERING TEIE RESEARCH QUESTIONS
5.2.1. RELATZONS BETWEEN ESSA Y SCORES AND O THER FACTORS
5.2.2. RELATIONS BETWEEN SOLO LEVELS AND OTHER FACTORS
5.3. GENERAL REMARKS ON THE INITIAL ANALYSES OF THE DATA

CHAPTER 6: DUAL-SCALING ANALYSES 87

6.1. THE DATA REVISlTED

6.2. DUAL SCALING
63. DUAL-SCALING ANALYTIC PROCEDURES
6.4. OUTCOMES OF D UAL-SCALING A NALYSES
6.4.1. ANALYSZS 1
6.4.2. ANALYSZS 2
6.4.3. ANALYSIS 3 ( A )
6.4.4. ANALYSIS 3(B)
6.4.5. ANALYSIS 4 ( A )
6.4.6. ANALYSIS 4(8)

CHAPTER 7: CONCLUSIONS 116

7.1. UNDERSTANDING TBE C ONTEXT O F JFL WRITING 117

7.2. TüE RESEARCH QUESTIONS REVISITED 120
7.2.1. RESEARCH QUESTIONS PERTAINING TO THE TEACHING CONTEXT 120
7.2.2.RESEARCH QUESTION PERTcVNING TO STUDENTS' CHARACTERlSTZCS 121
7.2.3. RESEARCH QUESTIONS PERTAINZNG TO STUDENTS ' WRiTZNG PROCESSES 122
7.2.4. RESEARCH QUESTIONS PERTAINZNG TO STUDEhiTS' WRITTEN PRODUCTS 124
7.3. M AKING C ONNECTIONS BETWEEN THE INTERPRETIVE AND DU AL-SCALING
ANALYSES 125
7.4. REFLECTING ON MY RESEARCH OVERALL 128

REFERENCES 136
 LIST OF TABLES

Tàbfe2-1. Kei 's Respomes to rhe 6-item Questionnaire on Teacher Beliefs and Practices of Teaching of
Writing 25
Table 3-1. Studenrs ' Charackristics 38
Table 3-2. Students ' Language-Related Characteristics 47
Table 3-3. Summas, of Students ' LI Writing,JFL Oral, and JFL Writing Skilk 47
Table 4- 1. Timeline of Wriring Assignmenrs 57
Table 4-2. Transcription Conventions 59
Table 4-3. Categoriesfor Cuding Think-Alod ProrocoL 61
Table 4-4. Surnrnary oJ Think-Abud Prorocol Data 65
Table 4-5. Aspecrr of Attenrional Foci and Rejlectiveness (Think-Aloud Protocol Data) 67
Table 4-6. Post- Writing Questionnaire Responses. JFL Projîciency. and LI WrifitzgExpertise 69
Table 4-7. Post- Writing Questionnaire Responses and Perceptions ofJFL Writing and Learning 70
Table 5-I. Description of SOLO LeveIs 78
Table 5-2. Summary of Japamse Essq Scores and Cornments 82
Table 5-3. SOLO Lmek of Japanese Written T e m 84
.
Table 6- 1. Major Contriburors and T k i r Bipofar Options (Analysis I Solulion I ) %
Table 6-2. Major Concribufors and Their B e l a r Options {Analysis 1, Solution 2) 97
Table 6-3. major Contributors and Titeir Bipolar Options (Analysa 2, Solution I ) IO1
Table 6-4.Major Contributors and Their Bipolar Options (Analysis 2. Solufiori 2) 102
Table 6-5.Major Contributors and T k i r Bipolar Options ( A n a b i s 3(a),Solution 1) IO5
Table 6-6. Major Contributors and Their Bipolar Options (Analysis 3(a).Solution 2 ) IO5
Table 6-7. kîajor Contributors and Their Bipolar Options (Anaîysis 3(b).Solurion 1) 109
Table 6-8-Major Conrributors and Their Bipolar Options (Analysk 3(6). Soluzion 2 ) Il!
Table P I . Background Questionnaire Data (Iterns 1-10) 179
Table P2. Language Skills Assessrnent Data ( I t e m 11-13) 180
Table P3. Interview Data (ltems 14-16) 180
T d l e P4. Posr-Wrih g Quesrionnaire Data (Items 17-22) 181
Table P5. Think-Aloud Protocol Data (Items 23-30) 182
Table P6. Japanese Essay Assessrnent Data ( I t e m 31-44) 183
Table QI. Major Con&ributorsand Their Bipolar Oprions (Anafysis 4(a).Solution I ) 184
Table Q2. Major Contriburors and Their Bipolar Options (Atzalysis4(a),Solution 2) 185
Table Q3. Major Conzribrtmrs and Their Bipolar Oprions (Analysis 4(b),Solurion 1) 186
Table Q4. Major Concributors alid Their Bipolar Oprioru (AnaZysis4{b),Solurion 2) 187

LIST OF FIGURES

Figure 1-1. The 3P Model as a CIQssroornSystern Relared to CVriring 9

Figure 6-1. Distribution of 13 Participants in A~raZysisI 95
Figure 6-2. Diwibution of 13 Participunr~in Analysis 2 100
Figure 6-3. Distribution of 13 Participants in Analysis 3( a ) 104
Figure 6-4. Distribution of 8 Participa)ifsin AnalysLr 3(b) 108
F i g m 6-5. Distribution of 13 Participants in Analysis 4(a) 113
Figure 6-6.Distribution of 8 Participants in Artalysis 4(b) 115
Figure 7-1. The Presenr JFL Writing Contexi as a Biliterate Instance Situated on Honzberger 's Nine
Contitiua 119
Figue 7-2. Conriectiotzs Nt the JFL CVriting Perforrrrarice Data 126
 LIST OF APPENDIXES

APPENDIX A: LETTERS O F CONSENT

APPENDIX B: CLASSROOM OBSERVATION SCHEME

APPENDIX C: QUESTIONNAIRE O N T E A C H E R BELEFS AND PRACTICES OF

TEACWING OF WRITING

APPENDIX D: LETTER TO SOLICIT P A R T I C P A N T S

APPENDIX E: QUESTIONNAIRE O N STUDENT BACKGROUND, JAPANESE

LANGUAGE STUDY, AND WRITING IN L1 AND JAPANESE

APPENDEX F: SAMPLE JAPANESE ESSAYS

APPENDIX G: ESSAY RATING SCALE

APPENDIX H: INSTRUCTIONS F O R ESSAY RATING PROCEDURES

APPENDIX 1: STUDENT INTERVIEW GUIDE

APPENDIX J: FULL SUMMARY OF STUDENT INTERVIEW RESPONSES

APPENDIX K: INSTRUCTIONS F O R THINK-ALOUD PROCEDURES

APPENDIX L: POST-WRITING QUESTIONNAIRE

APPENDIX M: SAMPLE PROTOCOLS W I T H CODED CATEGORIES

APPENDIX N: FULL SUMMARY OF THINK-ALOUD PROTOCOL DATA

APPENDIX O: SUMMARY O F PRODUCT AND PROCESS DATA

APPENDIX P: FORTY-FOUR ITEMS OF STUDENT DATA

APPENDIX Q: RESULTS OF DUAL SCALING (ANALYSIS 4(A) AND ANALYSIS 4(B)) 184
 Chapter 1: Introducing and Framing the Study

1.1. Overview

Why do I ask my students to write compositions? What are they getting from this
task? These plain questions, which I have repeatedly asked myself as a teacher of Japanese

as a foreign laquage1 (JFL), are the impetus for the present research. I t is commonly held,
with the current pedagogical emphasis on processes of writing, that writing provides a n
effective opportunity for learning in al1 subject areas. What does this statement mean in the
context of foreign language pedagogy? What and how well do students learn through
wnting compositions in a foreign language? These questions form the core of my inquis.
They were addressed within the recent tradition of research on situated learning (e-g.,
Minick, Stone, & Forman, 1993; Newman, Griffin, & Cole, 1989). Following this trend,
the analyses in my study involve individuals working in a natuni, goal-directed, activity

setting (cf. Wertsch, 1981,1985) nther than focusing on the decontestualized performance
of an isolated individual in a laboratory-Iike setting.
The study adopts some tenets and practices of grounded theory (B. Glaser &
Strauss, 1%7; Strauss, 1987; Strauss & Corbin, 1990); it focuses initialiy on interpreting
certain elernents of experience then tries to develop theory from an examination of those
elernents and their interrelationships. The ultimate aim of my study is to contribute to the

1 In this thesis different tcms are usai to d e r COa non-native language. The tenu foreign langwge is
defined as a language learned and used outside its spcech community and the rem second language as the
language learned and used wittlln its speech commuaity. The tcrm LZ includes the concepts of both foreion
and second languages. Sec Ashworth (1985. pp. 16-20). Iüein (1986. pp. 19-20).and Stem (L983a. pp. 75-
17) for discussion of these rems.
construction of an integrated theory that enables the researcher to understand the nature
and meaning of an experience for a particular group of people in a particular setting (B.
Glaser & Strauss, 1%7). My orientation to analysis, however, is practical. not entrenching
itself in rigid prescriptions on what oirgltt to be done. It favors a science that emphasizes
the ernergent nature of mind in activity and that acknowledges a central role for
interpretation in its explanatory frarnework (Cole, 1996).
The objects of my study are JFL university students' experiences of learning
through writing JFL compositions in a formal instructional setting. This setting, the
classroom, is viewed as a system that is "an interconnected set of elements in equilibrium"
(von Bertalanffy, cited in Biggs, 1991, p. 36) or a "functional system" (Anokhin, cited in
Newman et al., 1989, p. 69, pp. 71-73). The classroom consists of various elements
pertaining to leamer, teacher, task, and leaming processes, as well as leaming outcornes,
which interact with each other. The integrated system of the classroom is well captured and
represented in Biggs' presage-process-product (or 3P) model (Biggs, 199 1; Biggs &
Moore, 1993; see section 1.3 for a description of the model). which has provided the
present study with a conceptual frarnework upon which to build an organized, disciplined,
and systematic investigation.

The theme of my study is the leaming-and-writing relation in the contest of formal

instruction of JFL. The need for such a study is obvious from both pedagogical and
theoretical viewpoints.
 The past fifteen years or so have witnessed remarkable developments in foreign
language education in North Amenca (Lange. 1992; Met & Galloway, 1992) as well as in
other parts of the world (Dichon & Cumming, 19%). One notable phenornenon in North
Amenca was the mainstrearning of the Japanese language in education generally (Jorden &
Lambert, 1991 ) 3 Facing the sudden growth of Japanese language programs, the need for
research that would prornote and assist the leaming and teaching of Japanese became
urgent. A new group of foreign language specialists and researchers emerged focusing on
vanous aspects of JF'L education. The number of JFL researchers is still on the rise of an
upward mood among the foreign language profession in general; their professional
activi ties are becorning prominent. Many efforts to tnnslate research fi ndings in to practice
are being made. However, most of this research has thus far concentrated on the
components of reading, speaking, and grammar learning while leaving the writing
component under-researched. For example, Kanagy's ( 1995) bibiiography of research on
the acquisition of Japanese as a second or foreign language (JL3) contains 127 entries,
including published book sections and journal articles, theses, and conference papers.
Among them, only 2 concen writing in contrast with 34 on reading, 22 on speaking, and
23 o n gramrnar leaming. The paucity of research on JL2 wnting is also evident in a
bi bliographical search 1 did of the ERIC Database, MLA In ternational Bi bliography
(Febniary 1981-1997), PsychLIT, and Dissertation Abstracts (January 1982-December
1996). The search resulted in the total (cumulative) number of 195 references conceming

JL2. Among them, only 9 focus directly on JL2 wnting.

Why has writing failed to attract the attention of JFL researchers? It is clear from
Jorden and Lambert's (1991) survey that composition writing is not a popular activity for

2 Jorden and Lambert's ( 199 1) report on Japanese lan,ouge education was iimited to situations in the
United States; no equivdent account is available documenting developments of Japanese ianguage education
of recent years in Canada, mg research site. However. it rnay well be presumed that much of the situation in
Canada mirrors what bas k e n observed in ihe United States.
American students enrolled in Japanese language study. Rather, they are annious to
develop speaking and reading cornpetence. Such students would natural ly request more
instruction that emphasizes speaking and reading. And teachers and curricula have
probably responded to such students' expressed needs. This view of classroom needs could
have set me a research agenda focused on speaking and reading.
Recently, however, sorne foreign language teaching professionds have begun to
value wnting for its having a potentially wide range of pedagogical uses. Koda (1993)
even sees composition writing as "a major component of foreign language (FL) instruction,
whatever the pedagogical orientationn (p. 332). Writing is now considered, at Ieast by
sorne, as an effective means to develop integrative language skills. It is also said to
enhance cognitive and metalinguistic awareness by affording writers opportunities to think
in and refiect on the target language (Cumming, 1990; Swain & Lapkin, 1995; cf. Olson.
199 1). This renewed conceptualization of writing within the foreign language teaching
profession is probably tied to the current emphasis on writing, the teaching of writing, and
the development of writing abili ty in education in general (Valdes. Haro. & Echevaniarta,
1992). For example, the extent of empirical research on the teaching and learning of
English-as-a-second-language (ESL) writing in recent years might have k e n expected to
yield a direct impact on foreign language instruction. Nonetheless. this has scarcely
happened; composition writing deserves more attention from foreign language
professionds.
Presently, many things are unknocvn about foreign language wnting. An obvious
area for investigation to explore promisi ng pedagogical implications is the widel y-held
assumption that writing is a way to leam. The qualities of learning that might occur
through writing in a foreign language need to be better understood so as to allow educators
to make inforrned decisions about what to do with this tool of writing in a foreign language
context. T o draw relevant pedagogical implications, however, such research must also
account for the situated nature of learning, tahng into consideration the instructional and
contextuai variables that corne to play in leaming.
The question I have taken on--how lemming through writing can be identified and
studied in a forma1 instructional context--is a practical problem for foreign language
educational research. This is because wnting activity, though often an integral part of
foreign language instruction, typically lacks a clear definition of pedagogical purpose.
Wri ting instruction is not so well-defined in foreign language programs as in its ESL
counterparts. ESL wnting courses are designed to prepare an ever-increasing number of
international students in English-speaking countries to meet entry levels in academic
programs as well as their continuing writing requirements. Much research effort, for
insiance, has been directed toward anal yzing how well students exiti ng from ESL courses
may be able to write in English and to cornpete with their peers of English as a rnother

tongue. Such pragmatic reasons of academic and social integraiion for writing instruction
and research are seldom present in foreign language situations where contacts with the
target language and people and opportunities for using the language are by definition
limited. Nonetheless, many people believe that writing has a role to play in foreign
language instruction. My research tries to subject this folk belief to close investigation.

The object of my study, as outlined above, is a practical problem. I t is also a

theoretical problem, as indicated by those seeking to build a comprehensive theory of L?

wnting that accounts for the phenornena of writing in both second atid foreign languages

(Cumming & Riazi, 1997; Gass & Magnan, 1993; Gnbe & Kaplan, 19%; Hirose &
Sasaki, 1994; Kroll, 1990; Sasaki & Hirose, 19%; Silva, 1990, 1993; VaIdes et al., 1992),
and similarly for those concerned with the construction of an integrated theory of writing
as a social and cognitive process (e-g., Bridwell & Beach, 1984: Fiower. 1990a).
 Recently there have been a number of laudable attempts to account systematically
for students' L2 writing behaviors, cognitive processes during composing, and the ways

these behaviors and cognitive processes interact with written products and their

environmental and persona1 contexts (Cumming & Riazi. 1997: Grabe & Kaplan, 1996;
Hirose & Sasaki, 1994: Sasaki & Hirose, 1996). These studies have emerged after an
earlier accumulation of a substantial number of empirical--albeit exploratory, small. and
separate--studies on L2 writing over the past 15 years or so. It should be noted, however,
that virtually al1 of the previous work has conspicuously concentrated on English in
situations where the language is learned as a second language (as noted by Grabe &

Kaplan, 1996: Silva. 1993; Uzawa & Cumming, 1989; Valdes et al., 1992). and mainly
focused on such particular populations as school-age children. international students
preparing to study at English-speaking universities, and immigrant adults (as noted by
Cumming, 1994). These tendencies may be presenting only a partial picture of L2 writing;

it needs to be arnended by more research on foreign language writing.

Foreign language writing research is still in its infancy. Our immediate task at this

stage of research development is to identify the facets of foreign language writing, doing
so in a systematic and principled fashion toward the construction of a modular theory (cf.
Ellis, 1994. chap. 15). Furthemore, to provide a "rounded picture" (Flower, 1WOb, p. v) of
L2 writing, the facets to be discovered should be seen as dynamics of cognitive and social
phenomena. In particular, the blended socio-cognitive perspective currently popular among
English-as-a-mother-tongue writing researchers, and gaining some attention from ESL
writing scholars (Prior, 199 1; Riazi, 1993, has not been taken up by foreign language
professionals in writing research. In short, the current scope of L2 writing research needs
to be expanded by looking at not only ESL writing but also the writing of other ianguages
being leamed as a foreign language under various conditions, and by taking an inclusive
and unified view of human cognition and its environments.
 2.3.Biggs's Presage-Process-ProductMo&l of Classroorn Learning and Teaching

I used Biggs's presage-process-product (3P) model to frame rny inquiry. The 3P

model represents an integnted system comprising three phases with four components that
interact with each other to account for the total context of classroom learning and teaching.
The three phases are presage, process, andproducf (following Dunkin and Biddle's earlier
formulation of a similar rnodel, 19743). The presage phase has two components, one
pertaining to students (i.e., leaming-related student characteristics such as abilities,
expectations and motivations for leaming, and conceptions of learning) and the other
relating to the instructional context at the level of individual teachers (e.g.. teachers'
personal characte~sticsand their conceptions of leaming and teaching) and at the
institutional level (e.g., curric~lumcontent and assessment). Student and instructional
presage factors intemct among thernselves and feed into leaming processes, a component
of the process phase. For instance, depending on the degree of their genuine interest in a
given academic task, or depending upon how they see the demands made by the
instructional context, students set up certain goals of their own in dealing ivi th the task and
accordingly constme their particular ways of approaching the task. Furthemore, different
approaches are lihly to bnng about quantitatively and qualitatively different learning
outcomes. Outcornes are then prone to affect the presage conditions. That is, students'
leaming outcornes may cause a change in teachen' conceptions of learning or their
approaches to instruction. Similarly, students' perceptions of their own outcomes may
affect their levels of confidence in their abilities or may alter their expectations of
achievement. These directional links between the model's components appear as one-way,
chain-reacting phenornena. But interactions between al1 components are possible.

3 Stem (19Sa. chap. 22) used Dunkin and Biddle's mode1 to discuss L2 teaching in educational
tenns. See also Chaudron (lm)for an adoption of Dunkin and Bidde's model to L2 classroorn research.
indicating the systemic, rather than additive, nature of the whole?
Two main features of this mode1 fit neatly with the perspectives taken in my
research. First, learners are the focus. The instructional context provides a superstmcture
set by the teacher and the institution, affecting learners' states at the presage, process, and
product phases. However, leaming does not occur without leamers' actual engagement in a
learning task. And how they approach the task determines the quality of learning
outcornes. This logic gives good reason for my research to focus on learners' activities,
rather than teaching activities, as the study is concemed with individual students' leaming.
Second, the model takes the entire context of classroom learning and teaching into account
and represents its complexity in a succinct fashion. My study investigated the mechanisrns
of learning through composition writing within the system of a university JFL classroom, a
situation which is uniquely organized arnong as well as within individual mernbers of the
class and one that differs in certain ways (descnbed above) from the situations for ESL
learning addressed in most other research on L2 writing. The 3P mode1 of the classroom
gives an appropriate structure to f m e my inquiry.
Figure 1- 1 displays the model schematically and shows how 1 have modified some
of its original terminology, adapted slightly to suit the object of my study. I t should be
noted that systemic pnnciples represented in the model opente at not only the classroom
level but also higher levels Iike school, school board, and community (Biggs & Moore,
1993, chap. 16). I t is important to know how each level of the educational structures forms
a larger coherent system at the nest level in order to increase Our understandings of
teaching and learning. My present inquiry, however, limits its scope to the classroom
context focusing on writing tasks.

4 Biggs and ~Moore(1993. p. 449) provided a concise explanation on the systemic versus additive
view of phenornena: "A system is a working whole made up of a set of component parts, which interact with
each other to forrn an equilibrium."Thus, whatcvcr happens to one part will affect the rernaining parts. An
additive model. on the other hand, is "a multistructurd deficit modei" where addition. subtraction, or
modification of a part does not change the rest of the whole. A relevant discussion can dso be found in
Salomon ( 199 1 ) which contrasts the epistemologies of analytic and systemic approaches.
 Presage Process

Ethnicity
Orallwritten L I
L I writing skill
LOF!
TL proficiency
Perception of TL writing
Area of difficulty in TL writing
Area of learning through writing
etc

Multiple trait assessrnent

Conceptual complexity

Teacher
Classroom

F i p w I - I . Tlie 31' iiiodel as a classrooiii systeni related [O wriiiiig. Adnpted froiii f i e Pr-ocess o/ Lecimittg (p.
45 1 ), by J. B. Biggs and P. J. Moore, 1993, Sydney: Preritice Hall.
 1.4. Object ami Foci of tl2e Shdy

My research began and ends with my interest in the epistemological question of

what "writing as a mode of learning" (cf. Emig, 1977) means in the context of foreign
langage instruction. To facilitate this i nquiry, 1 took up the hypothetical proposition that
foreign language writing is more than jusi wri ting a foreign language, inspired by the titie
of Janet Swaffar's article, "Language leaming is more ihan leaming language" (Swaffar.
199 1). 1 structured my esploration of this proposition around four sets of research
questions related to the four main components of Biggs's 3P mode[:
Teaching context: How are the writing tasks situated in the JFL course? How does
the instructor perceive the role of writing, the teaching of writing, and the teaching of
EL?
Stiiderrrs ' cliarncferisfics:How d o students' characteristics relate to their perceptions
of JFL wri ting and leming?
Writing processes ( Meta-leaming processes): What linguisuc and cognitive be haviors
are invoked by JFL writing activities? How are the. related to students'
characteristics and instructional contest variables?
Writteii prodiicts (Outcornes of leaming): What testual qualities of JFL wnting do
students manifest in terms of language use, content, and organization? How 1s the
complesi ty of cognitive structure manifested in their w n tten texts? How are these
elements related to students' characteristics, instructional contest variables, and in-
process behaviors?

At the heart of the proposition "Foreign language wnting is more than just wnting a

foreign language" is a broader curricular issue. I t concerns the question of what role
foreign language instruction plays in education aiming at the development of students'
minds or intellects. Considering the historical impact that literacy has had on the
development of people's thought and cognition (e.g., Olson, 1994). focusing on the writing
component of language instruction brings this issue to the fore. To make this perspective
clear, 1 review in the following section the emerging voices of foreign language
professionals who have been concemed about articulating a mtionale for foreign language
instruction in cumcula at schools, colleges, and universities.

1A.1. Recent ConceptualiZatrZatroons

of Foreign Lunguage Cruric~~la

As foreign language educaiion has become "a secure part" of school and university
cumcula due to recent institutionalized enforcements of foreign language requirements in
the United States (Lange, 1992, p. 528; see also Davis, 1997) and elsecvhere (Dickson &
Cumming, 1996), an urgent need has appeared to carefully consider the relationship
between foreign language curricula and the goals of educational programs overall.
Historically, language teaching has been influenced primarily by theones and pnnciples of
linguistics, psycholi nguistics, sociolinguistics, and second language acquisition research
(Nunan, 1988; Richards & Nunan, 1991). Only recently have foreign language
professionals begun to see themselves as educators as well as applied linguists. This
renewed conception of foreign language teaching as an educational matter invites us to
consider what needs to be learned beyond the linguistic aspects of a target language.
Littlejohn and Windeatt ( 1989) called such leaming "additional, non-language, leaming"
(p. 158) as emphasized by, for instance, Lange (1992). To Lange, however, such leaming
is a priority rather than "additional" component. He has strongly advocated that foreign
language cumcula move toward a view that prizes "its contribution to human development
rather than the developinent of language features" (p. 538).
To understand Lange's proposal, one needs to consider esisting foreign language
cumcular orientations, particularly their sense of what the cumculum is for. Lange (1990)
insightfully anal yzed foreign language curricular orientations using Schubert's ( 1986)
threefold classification of curriculum inquiry: the scientific-technical, also known as
empirical-analytical; the practicai or hermeneutic; and the cntical o r emancipatory. Recent
trends in foreign language curriculum have linked the empirical-analytical orientation and
the interpretive o r hermeneutic orientation. The former is represented by audiolingual and
grammar-translation teachi ng while the latter emphasizes the developmen t of language
proficiency a s a means of expressing individual meaning and discovenng meaning in the
world of the individual. Lange suggests that one such example is Stem's multidirnensional
curriculum. consisting of linguistic, cultural, communicative, and general language
education syllabi (Stem, 1983a. 1983b, 1992). Although curricula of this combined

orientation (or what Lange calls the cotiriec~ive)focus on uses of language for

communication. the understanding of the self and of the siirroiciiding other as well as any

use of l a n g u a p to act upon the human condition are fundamentally byproducts of leaming
the linguistic components (phonology, morphology, syntas, iesicon) and basic
communicative functions of language (Lange, 1990).For ihis reason Lange suggested that
cumcular efforts should go beyond the analytic and scientific to help leamers find personal
meaning in their learning processes (through a hermeneutic orientation) or to allocv them to
contemplate and act upon the society in which they live (through a n emancipatory
orientation). This suggestion aims toward the development of the enlightened self o r the
betterrnent of the human condition.
Such orientations and purposes are difficult to find in the cumcular wntings on
foreign language education (lange, 1990). There are, however, some exceptions, such as

Moskowitz (1978) and Stevick ( 1990) who have presented instructional techniques for the
cause of hurnanism based on a hermeneutic philosophy of language learning and teaching,
or Crawford-Lange and Lange (1987) and Kubota (19%) who have applied the work of
Freire (1973) to reorient language learning from the scientific-analytic acquisition of facts
and information toward emancipatory reflection and action on such facts and information.
 These voiccs speak for language leaming as a total "educative experience" (Dewey,
193811 x 3 ) . They overlap wi th those advocating content-centered ianguage instruction5 as

a means to counter traditions of language instruction through "bland and contentless"

textbooks (Littlejohn & Windeatt, 1989,p. 159).The intent of con tent-based instruction is
twofold (Crandall & Tucker, 1990). One is to help leamers attain language proficiency
beyond the level of basic communicative skills, that is, expanded foci of instruction for the
development of "basic interpersonal communicative skills (BICS)" a n d
"cognitivekademic language proficiency (CALP)" (Cummins, 1980; Cummins & Swain,
1986, chap. 8). The other intent is to meet the specific needs of language students who
corne to language classrooms for a variety of reasons. Although integnted language and
content instniction is often associated with second language leaming where special
language skills are required to cany out cognitively demanding academic tasks, i t is dso
gaining acceptance among Foreign language profcssionals who recognize the increasing
importance of students developing foreign language proficiency that goes beyond the BICS
level (Crandall, 1993; Crandall & Tucker, 1990; Leaver & Stryker, 1989; Snow, Met, &
Genesee, 1989; Sudermann & Cisar, 1992).
This curent movement of foreign language curricular reform is supported by a
general understanding of L2 acquisition that suggests a n L2 is Icamed most effectively in
contests that are personally meaningful and socially purposeful, that is, when rneaning
rather than form is a focus of instruction. But there is a danger in this line of thinking, as
Sudermann and Cisar ( 1997) have cautioned in their cri tical appraisal of foreigli kmgimge
across ~ h eciirricldio~rprognms, narnely , there is a danger that corn ples form-meaning
relations may be reduced into dualistic or simplistic pedagogical formulas. The reduction

5 See Crandall (1993) and Crandail and Tuckcr (L99û)for a comprehensive overview of content-
centered language instruction. The former includes an annotated bibliography on the subject. Pmctical
suggestions and guideIines for the implcmentation of content-centered instruction cm be found in blohan
(1986) especiaily for ESL professionais, Brinton, Snow. and Weschc (1989) for L2 specidists at
postsecondary acadernic institutions, and Cantoni-Harvey (1987) for elemcotary and secoodary educators
deaiiag with minonty students. Sec also Adair-Hauk's (1996) strategies for the practical application of a
whole Iringuage approach in second and foreign language classrooms at the secondq and universi ty levels.
has appeared, for esarnple. in the long-standing debate over the value of phonics or whole
language in reading instniction or more recen tly the product-process debate in wri ting
instruction. When two seemingly oppositional views are put forth. there appear voices to
admonish extremism, calling for a more balanced, integnted stance as seen in the
interactive model of reading instruction (Carrell, Devine, & Eskey, 1988). the " balanced
and flexible literacy diet" (Willows, 19%), and the "middle ground" approach to writing
instruction advocated by, for esarnple, Arndt (1981), Connor (1987). or Raimes (1985).
For the development of balanced foreign language curricula that place importance on form
and meaning and language learning atid non-Ianguage leaming objectives, 1 (So, 1994a)
have suggested that integntion of language and non-language goals might be realized
successfully through Miller's (Miller, 1988, 1993; Miller, Cassie, & Drake, 1990) holistic
curriculum model consisting of three overlapping cornponents of transmission
(corresponding to the scientific-technical orientation to curriculum), transaction (the
hermeneutic) and transformation (the emancipatory).
Ideologies in foreign language education are changing progressively and
enpanding. The purpose, content, and rationale of the foreign language curriculum are in
transition accordingiy--from loci on the mechanical learning of [inguistic forms isolated
from contest, to emphases on the rneaningful, communicative use of language, and to aims
of broadening leamers' visions and perspectives. In view of these recent trends in foreign
language curricular orientations, 1 wanted my research to ask, how is the writing
component treated in relation to the overail curriculum of JFL? What would be the
experience of leamers writing in JFL? My simple and situated inqui.--what is really going
on with JEX wnting in the classroom?--has never, to my knowledge, been taken up as a
research subject in thc field of JFL education.6 Taking up this hnd of "reflective inquiry"

6 As rnentioned =lier in this chapter, JFL writing has received Little research treatment in the pzist.
Somc exceptions include writing process reswch such as Kato (1995). Ootrike (19%). Pennington and So
( 1993).Sbibata ( 19%). and Uzawa and Cumining (1989). However, thesc investigators' primary conccrn wris
not eduwtionai or leaming processes through writing but rarher the process of writing itself in specific tasks.
has involved drawing upon al1 sorts of howledge and hunches, without which "empirical

research c m hardly move" (Bereiter & Scardamalia. 1987. pp. 35-36).

1.S. Orientalion and Orga~uurtionof the DissertWon

Through the experience of writing up this dissertation 1 am reassured that writing ir

thinking; writing is a way to discover what 1 am thinhng and to discover gaps in my
thinhng. Writing offers a private way to capture and give concrete form to ideas that are
oftentimes elusive. This dissertation is where that private process of mine, that was at once
emergent, evolving, and iterative, has been made public. The written account of this
research, Le., my dissertation, focuses on my coming to understand the phenornena of JFL
writing and learning as 1 collected data. dispiayed information. and interpreted and

expressed the meanings of this information. It is a report of an exploratory, descriptive

case snidy in which 1 proceeded inductive1y. 1 consider that the best way to reconstruct this
inductive process is first to present the data collected in the field and then to show how I
analyzed it to seek data-based understandings of JFL students' and teacher's expenences. I
followed B. G. Glaser's (1978) suggestion for a much later handling of relevant literature

so that I would not be unduly influenced by others' ideas and foci regarding the problem
under ~ t u d y Thus,
.~ as I present the research, 1 will also show how 1 consulted and
reviewed published literature to relate rny esperiential understandings to them and their

relevant ideas.
After this brief introductory chapter, I will in chapters 3, 3, 4, and 5 display
different sets of data collected wiihin the frame of Biggs's 3P mode1 and address the four

7 This does not mean that 1 did not read before collecting data. 1 did read wideiy, as B. G . Glaser
(1978) suggested, in substantive areas including L2 acquisition and eduwtion. writing, and research methods
and perspectives. This activity tumed out higbIy beneficial as it enabled me to "skip and dip, thereby gaining
greater coverage"(B. G . Glaser, 1978. p. 32).
sets of guiding questions formulated within the sarne fnme (see section 1.4). Various sets
of research materials were genented through classroom observations, interviews,
questionnaires, tape recordings of think-aloud sessions, and wrï tten documents in order to
capture the complesity or educational phenomena as they actually occurred in relation to
JFL wri ting tasks. These materials are orgaanized according to the four cornponents of the
3P model and described in separate chapters: the reachirzg content data in chapter 2, the
strrderits' clznrczcteris~icsdata in chapter 3, the ivrifing processes data in chapter 4, and the
writteri prodircts data i n chapter 5. In each chapter 1 surnmarize the steps that 1 have
followed and the pnctical concems that 1 dealt with in the process of converting the
materials into an analyzable and interpretable form and denving scientific evidence in the
present study. I use a narrative fom to report my research when 1 wish to highlight "the
storied quali ty of esperience" (Clandinin & Connell y, 199 1. p. 259). Furthermore, I
address each corresponding set of research questions by describing, analyzing, and
interpreting the data through the combined use of quantitative and qualitative rneth~ds.~
In chapter 6 1 present the results of statistical analysis of the combîned data
generated Srom the students--the stiuients' cluvncteris~ics, writi~igprocesses, and writteti
prod~ccfsdata--to describe how these components are interrelated. 1 used a unique
statistical technique called &al scnlirig to go beyond a descriptive, interpretive account
based on my impressions and intuitions. This chapter is followed by chapter 7, which
concludes the dissertation by providing a coherent and valid account of the phenomena
represented in the data as a whole. In this final chapter I also retum to the original question
of what "writing as a mode of learning" means in the contest of foreign language
instruction and in relation to the proposition that foreign language writing is more than just
writing a foreign language.

- - -- -

8 My methodologicai orientation in ths research is manifest in the combincd use of an interpretive

approach and a statistical technique; this orientation reflects a growing recognition of the dialectic
rclationship betwçen meaning and number and of the partncrship between qualitative and quantitarive
research methods. Dey (1993).for instance, supported such an eclectic vicw of data analysis, stressing "the
interdependence and mutual enhancernent of apparcntly opposing approaches"(p. 266).
 Chapter 2: Teaching Context

2.1. Site

It was a small seminar room of approximately 300 square feet with no windows,
sandwiched between another small seminar room and a large room wi th a seating capaci ty
of 100. There was only one door to get in and out of this room. The r m m was located in a
large modem building of a major anglophone university in Canada. Students mshing to
their respective classes o r hanging around the hallivay were a diverse m i s of ethnici ties,
though Asian students were conspicuous, 1 thought. It was the day of my first visit to Kei's
Japanese class. Kei (a pseudonym, pronounced K a y ) introduced me btiefly as a Ph.D.
student collecting data for her dissertation to 15 students sitting around a large old wooden
table in the room. The table took up s o much space of this small room that the students had
to be squeezed into the remaining narrow open space around the table. AI1 in the room
were Asian escept for two Caucasian students.
The class was a Japanese reading and writing course meeting once a week for two
hours in the moming throughout the academic year, from September to April. The course
is offered each year to students who have completed a third-year general Japanese
language course o r who possess an equivalent level of Japanese proficiency. There is,
however, no upper iimit of proficiency set for admission into the course. It is one of the
two highest levels of courses in the Japanese language offered a t this university. Most of
the students were in the fourth and final year of their undergraduate studies; most had

taken an elementary Japanese language course with Kei. Although it was only the second
meeting of the class, there was not the stiffness typically found in the initial meetings of
any course. 1 wrote i n my field notes that day, "There was a sense of warmness and at-
homeness in the class" (Field notes, September 29, 1994).
This class was selected for my research site for the following reasons: (a) The
course focused on literacy skills in J E , including a component of composition wnting; (b)
Kei, the course instructor, was willing to collaborate with me in canying out this research;
(c) the course participants were educated adult students, a fact which would avoid

confusing foreign language writing with such issues as maturation and basic litency; and
(d) the location was convenient for me to make frequent visits.

2.2. Teacher

1 first met Kei in the summer of 1993, a year prior to my data collection. We were

in the sarne class for a course in cumculurn studies at a graduate school of education. At
that time Kei was cornpleting her M.Ed. degree in L2 education as a part-time student
while teaching full-tirne at the above-rnentioned university. In the cumculurn studies class
we were the only native Japanese. There were a nurnber of other factors that might have
put us close together. We were both female, around the same age, with several years of
overseas experience as adults; we both aspired to be specialists of JFL education at p s t -
secondary academic institutions outside Japan. Our friendship continued and developed
after the summer course. We would talk for hours discussing goals, objectives, content.
and methodology of JFL programs at institutions of higher education and shanng one
another's teachi ng experiences.
One episode highlights our common concems about JFL education. In May 1994
Kei related to me some feedback she had received about her teaching [rom one of her
students at the end of a Japanese language course she had been teaching. Her teaching was
highly evaluated by this student saying, "Your course was most interesting. 1 was always
looking fonvard to corning to your class." This student then added, "But the course which
made me tliitzk the most was Mr. X's." Mr. X was Kei's colleague teaching a history course
in the same department. Both Kei and Mr. X were apparently very conscientious educators
and deeply concemed about their students' development for future responsibilities and for
success in li fe. They were particularl y reflective about their teaching practices. Having

recounted the story to me, Kei continued, "1 wonder if it is possible to go deep into the
rninds of students in a foreign language classroom" (from my notes and recollection; my
translation of what was spoken originally in Japanese). In a series of subsequent

conversations ivith her, 1 discovered what she had meant by this. That is, can we, as
y to memorize grammar rules and
foreign language instructors, encourage students not 01-11
vocabulary or practice using the language for daily communication, but also to ihitik--the

kind of thinhng to make sense of one's experience and the world around oneselî?
When 1 approached Kei for her collaboration in my research work in May 1994.

she agreed immediately. She was in fact quite excited about having me in her class every

week. There was. by then, mutual trust between Kei and me at both professional and
personal levels. She appreciated my weekly visits to her classes in the coming academic
year, especially because she felt she needed someone who could listen to and understand
her pedagogical concems in situ. Such professional support was, according to Kei, scarce
in her department. During the past three years of her teaching JFL at the universi ty, she
worked in isolation; she wished for meaningful discussion of pedagogical matters with
special reference to university-level JFL. My 9-year experiences of teaching in similar
contexts could benefit her as much as her participation in my study would benefit me. I

gave her a copy of the proposal for the present research (So, L994c) in June 1994, she
discussed i t with me for clarification i n August, then signed a consent letter prepared
according to the OISE gi<ideliriesfor theses atzd orals (Ontario lnstitute for Studies in

Education, 1993)(A ppendis A ) in September before the commencement of data collection.

The chairperson of her department also signed a leiter of consent (Appendix A ) approving
my research activities in Kei's class.

1 visited every class throughout the year with a few esceptions; 8 out of 10

meetings were obsewed in the fîrst semester and 10 out of 12 in the second semester. 1 sat
through each class from the beginning to the end. My persistence in full and frequent
observations was necessary in order to experience and understand "the syllabus in action"
(Breen, 1984, p. 50).which is the enactrnent of a predesigned syllabusg through continual
interaction between teacher and learners during the actual processes of teaching and
learning. As Breen stated, "Although. as teachers, we may follow a predesigned syllabus,
every teacher inevitably interprets and reconstructs that syilabus so that i t becomes
possible io implement it in his or her classroorn" (p. 50) to meet the needs and abilities o l
individual learners and of the socio-cultural context. The fluid and dynamic nature of
syllabus construction and reconstruction processes was evident in Kei's classroom. The
original written schedule of teaching for the first semester was revised twice and the
schedule for the second semester received minor changes frorn time to time i n
considention of student and time factors. For the purpose of my research i wanted to know
exactly in what cumcular çontext wri ting activi ties took place. To do so. i t was important
for me to be on site at ail times to observe the tlow of classroorn activities.

9 In this thesis the term syiiabus is used in a more restricted sense than the t e m curricuiurn. Syllabus
refers essentialiy to the specifrcation and sequencing of course content. Curriculum, on the orher hand. refers
to the whole of teaching, including those elements of syllabus dong with considentions of goals,
rnethodology, and evduation of a course. There has k e n a great deal of debate over the definition of the tenn
curriculum among educationai researchers in general (see Jackson. 1992, for a detailed summary of
definitional issues of curriculum). For definitions of the terms curriculum and syllabus used in the field of L2
teaching, see Johnson ( 1989).Nunan ( 1988, chap. 2). Rodgcrs (1989).and Stem (1983a. chap. 19. 1992).
 One corner near the door was the spot 1 chose to sit during my observations. There
was no more than one and a half feet between me and the students sitting nearby. My
primary role was that of a non-participant observer, though there were a few occasions in
which 1 was asked to participate as a resource person supplying Japanese words and
expressions to students working in srna11 groups. The class was conducted entirely in
Japanese. 1 carried a clipboard to hold sheets of an observation scheme to record in real
time what happened and what the teacher and students said in class. The scheme
(Appendis B) was used no&for the purpose of quantitative analysis but to systematize my
observational records. 1 developed then revised it during my first few visits to the classes.
It employed 3-minute intervals and for each 3-minute interval it recorded the substance of

activity (e.g., test reading, discussion. grarnmar exercise), format of activity (e.g.. whole
class, pair work, small group), area(s) of instructional focus (e.g., read-aloud of a text,
content comprehension, lexis, grarnmar rules), utterances made by teacher and students,
and students' behaviorsY Ample space was provided for each item to be written down in
detail. I used both English and Japanese to record classroom events and discourse on
observation sheets. With the permission of Kei and the students, parts of some sessions
particularly relevant to wnting were audiotaped, lest some details should be missed in my
notes. 1 gathered al1 handouts, teaching materials, quizzes, and mid-term examinations to
supplement the information [rom my observations.
Based on the on-site notes, 1 converted the observations into 18 field notes in d l ,
each produced in English either on the day of observation or shonly aftenvards. Each note
was 500 to 1ûûû words in length, consisting of two parts; one described what happened in

the class, and the other recorded my reflections and comments. These notes necessady
reflected what 1 wanted to see in the classes. 1 concur wi th Fanselow's proposition ( 1990):

Io This approach rcsembles ihat taken in many other. more routinized scbemes for classroom
observation (e.g., Allen. FrohIich, & Spa&. 19% Riazi, Lessard-Clouston, & Cumming, 19%). See dso
Allwngbt (1988) for an account of the historical devclopment of systçmatic observation in lmguage
classrooms.
"What we see is not what takes place but what we value as important to see: observing is
selecting" (p. 186). Going through the 18 field notes at the time of writing this thesis, I
realize that what 1 selected to observe were the content and sequence of teac her-directed
learning activities. That is to Say, 1 tended not to focus on individual students or their
actions irrelevant to the teacher-directed activities. Writing my reflections and comments
facilitated my thinking about the research; it helped me to know what needed to be
clarified about Kei's teaching. Thus, this part of the field-note construction served the
function of generating questions to be asked during the subsequent interviews wi th Kei.
But at the same tirne 1 was aware that in this process I was reconstructing what happened
in the classes from my own perspective.

To understand Kei as a JFL teacher, 1 interviewed her frequently throughout the

period of data collection. 1 wanted to know her general conceptions of learning and
teaching of JFL on the one hand and on the other her thinking about the course content,
objectives, and teaching and assessrnent methods, as well as the students' progress in
leaming during the course. These issues constituted a general plan for the topics of my
unstructured interviews with Kei. However, no special meetings were arnnged for the
purpose of interviews per se; the interview agenda, so to speak, emerged during the course
of our usual, ongoing informal and intimate conversations as had begun before my entenng
the classes. Neither of us felt uneasy about these conversations because litile really
changed from the talks w e had had before in terms of topics or style. We talked in
Japanese for 30 minutes to 3 hours once or twice a week on the phone or over lunch, as we
would always do. Since 1 was watching her classes during the period, i t was only natural to
touch u p n our shared experiences in the class while talking. Wi thout deliberate efforts Our
conversations developed d o n g my genenl plan for the research.
 Benjamin (1974) spoke of two types of interviews, one for the interviewer wanting
help from the interviewee and the other for the interviewer wanting to help the interviewee.
These demarcations were not always clear in my conversations with Kei. Kei would
express her concems about what happened in the class that day or what she was planning
to do next week. In response 1 would bring in my experience and expertise to help her
resolve her problems. A t the same tirne I learned, through her explanations of ihese
concems, hocv Kei was conceiving of JFL teaching in general and thinking about the
course of my research objectives in particular.
A tape recorder was not used in these discussions; nor were on-site field notes

taken. Circumstances did not permit the use of such tools. Besides, the- could have spoiled
the rapport and trust between us that contnbuted to the generative process of our candid
conversations on these topics. Nonetheless, 1 did scribble quick notes about our talks
immediately after we parted, and later on the same days I produced retrospective notes of
Our conversations in English on my cornputer. These notes were 300 to 1OOO words in
length and recorded when, where, how long, and how the conversation took place.
sumrnarized what we talked about, and included Kei's words transcribed from memory
when they were deemed important. 1 genented a total of 32 of these notes betiveen
October 19, 1994 and March 19, 1995. The- were written from the perspective of me as a
researcher wishing to uncover the teacher's points of view by accounting for what was
observed in the classes.
In addition ro the materials collected during my observations of the course. 1 also
wished to be informed of Kei's retrospective thinking about the writing component of the
course a year later. 1 used a survey questionnaire as a guide for probing this. It comprised 6
items (Appendix C) taken from a 22-i tem questionnaire developed by Pennington
(Pennington. Costa. So. Shing, Hirose, & Niedzielski, 1997) to survey ESL teachers'
practices in writing instruction. I was aware that the survey instrument designed

11 Pennington et d ' s ( 1997) instrument highlights contrasts between process and product approaches
expressly for writing classroorns per se wouid not be exactly suitable to the situation where
Kei had to incorporate both reading and writing components into one course. As 1 had
espectcd, Kei found it difficult to answer the questions. Nonetheless, her responses were
useful to veri fy what I thought were her conceptions of writing. language, and education
more generally.

2.4. A l i s wering the Research Questions

How are the ivrititrg ~nskssituated in the JFL course? Hoiv does tfw
Nrstrrictor perceive the role o/writittg, the reachirzg of ivriling, and the
teachitzg 01JFL?

1 address these questions by telling a story about Kei and her classroom. The story

tells how Kei's patterns of interaction with this particular teaching context evolved over
the period of my research, shaped by her students and her moment-to-moment thinking. 1
summanze what happened in her classroom in relation to writing activities. focus on key
episodes. and delineate central elements, so as to construct an illuminating nanative.
As mentioned earlier, there are two highest-level courses in the Japanese language
at the university where Kei taches. One focuses on speahng and listening and the other
on reading and writing. The latter is the course Kei was assigned to teach in the year of my
research data collection. Kei was given freedom in the development of a syllabus for this
course. There were no institutional guidelines that she had to follow as to what, how, or for
what ends to teach. Nor was she constrained by any particular cumcular ideologies. A
pensa1 of m y classroom observation and teacher interview notes presents the following
pp - - - - - - -

to the teaching of writing, solicitiag the respondent's general view on the teaching of writing as well as his or
her specific ciassmm practiccs. .Most of the 22 items on the questionnaire are to be answered on two 5-point
scales: ideal and actud. AIso space is provided in each item for open-ended comments.
synopsis of Kei and her teaching.
Progressivism may be the best word to c h a n c t e n z e Kei's orientation to this
cumculum. This cumcular ideology was tacit, if not explicit, in what 1 heard from Kei and
what 1 saw in her classes. I t provided a direction to the role and purpose of her cumculum
constnicted around a series of themes, focusing on the expressive and creative functions of
written language. The cumculum developed in the context; it received modifications and
changes as Kei felt the need for pedagogical adaptability or encountered unplanned
teachable moments in the course of her teaching. Moreover, students were consulted about
the specifics of the syllabus such as the thernes to be studied and the g n d i n g scheme. In
short, her curriculum was, o r at least intended to be, leamer-centered, process-onented, and
meaning-focused. Kei's cumcular orientations were manifest in her responses to the 6-item

questionnaire (Table 2-1, Appendis C ) asking her to rate retrospectively on the ideal and

acti~al5-point scales certain aspects of her teaching of J E writing in the year that 1

observed.

Table 2-1
Kei 's Respomes 10 the 6-hem Qrrestiomiaire on Teacher Beliefs alzd Practices oJ Teacking
of Writirig
' Item# IDEAL ACTUAL
1 4 2.5

Note. The higher the rating, the more proccss- and leamer-centered the teacbing orientation.

In this curricular context, how did Kei perceive and approach the writing
component of the course? AIthough it was the third time for her to teach this Japanese
reading-and-writing course, Kei was particularly unsure of and uneasy about the teaching
of the wri ting component. The what, how, and why of JFL w riting instruction were Kei's
long-standing. unresolved pedagogical questions. Partly because of these unresolved
questions and also because of time constraints, class tirne was spent alrnost exclusively on
reading activities. Initially Kei was not cornfortable about the skewed weight placed on
reading in the enactment of her cumculum as she felt she was not teaching to the ritle of
the course. Toward the end of the course, however, she came to terms with the reality in
her own rvay. She said, "In the fïrst term 1 was overly concerned about not spending
enough time on the wnting component. But 1 have redizcd that i t is almost impossible to
teach both s h l k satisfactonly in a 7-hour-per-weekcourse. If airning at two is unrealistic, 1
should instead locus on one of the two and work torvard the students' substantial gain in i t"
(Interview notes, March 1, 1995).
Writing esercises rarely took place in class; they were instead given as homework
assignmenis. Nonetheless, connections were made between the reading activities in class
and the wnting assignments to be done at home. 1 now tum to cumcular events that put
these two components together. 1 describe a series of classroorn activities leading to each
of the two major wnting tasks (Task 1 and Task II), in which the students participating in
my study produced think-aloud protocols, as well as the espenences and images12 Kei
brought to her teaching practice.

2.4.1. Kei and Classroow Activities kading to Writing Task I

Both Task 1 and Task II were given to conclude two respective cycles of theme
study through reading matenals. The first theme was Japanese popular culture. Students
and teacher together decided to read about rnnrrgn-bzrimrr (a phenomenal boom in comic-
book stories) in Japan. By teaching to the interests of her students, Kei was hoping to

12 Clandiriin (1986) provided a unique account of the link betwcen experience and image as is relevant
to teachers' "personal practical howledge" (Connelly & Clandinin, 1988).
create intellectually stimulating and engaging events that wouid enable the expansion of
the students' minds. Kei selected reading passages from authentic materials such as
newspaper articles as well as contrived JFL texts. Her approach to the instruction of
reading and her rationde for i t may be best understood by reading the following excerpt
from rny interview notes in which 1 constmcted my interpretation of Kei's current vision
and action,

Kei wants her students to thitzk. Rather than just asking them factual
questions about reading texts or transmitting information on particular
topics. Kei wishes to go one step further. She wants her students to
reconstmct the text they are reading and make personally relevant meanings
of the text--Le., one step beyond the level of understanding what the text
says. Most common questions that foreign language teachers ask to students
in class tend to be very simple and purposeless. Kei would not be satisfied
with such instruction; she would rather seek for spnce during the class time
for students to reflect upon themselves, Iearning materials, and relations
between themselves and the matenals. To get this point across to me, Kei
referred to the current lessons undenvay on the topic of rrratiga-biciinu. In
these lessons she would ask display and referential questions about the
content of reading passages to make sure that the students have understood
the content. Furthermore, she would try to exploit the students' cognitive
capacities to discuss relevant issues such as influences of reading comic
stories upon the human mind. Kei's purpose for this latter activity is to help
her students espand their mental lives. (Interview notes. October 10. 1994.)

Kei found her students overly concemed with word- and sentence-level problerns,
far from the stage of treating JFL reading as "real reading" (Le., reading for meaning) or of
engaging in an active process of constructing testual and personal meaning of written test.
Based on her belief that no student should ever be encouraged to limit him- or herself to
mere decoding skills, Kei decided to devote a substantial amount of time and attention to
such top-down concems as reading for global meaning until her students would develop
skills to read Japanese tests for meaning and make educated guesses at meaning without
definitively knowing each and every word. I n response to some students' persistent
concerns about the accurate identification of lexical and grammatical forms, she would
Say, "Now you are reading in quanti ty and for meaning, so do not try to read word by
word. Do not stop at each and every gramrnar point that you are not sure about. Instead,
use context to guess the gist of the text" (Filed notes, October 27, 1994). Through this
meaning-focused instruction students were encouraged to regard the JET written text as
something that has content--nther than something that merely consists of linguistic mies--
as they would normally do with their native language reading.
Did Kei succeed in realizing her instructional goals? Yes. to some extent. During
the four class sessions on the topic the students were willing to go bepnd what they were
reading in class. Some students voluntarily brought a comic magazine, audiotape. or
videotape that contnbuted to animated discussions in the classes. They were also happy to
share their personal esperiences with comic books. 1 rvrote my reflection on the day of rny
observation of the second session on the topic as follows:

I found the entire class were enjoying the topic. Especially during the
discussion of Japanese comics, they were so animated and engaged. And
they did this so naturally that I did not realize that they were using a foreign
language to espress themselves. Kei seemed to have felt the same way as I
did. She said later that she was arnazcd at the students' receptiveness to her
teaching style and their involvement in reading and talking about their
collectively chosen topic. (Field notes. Oclober 20, 1994.)

The active student discussions observed that day were not a one-time incident. On my visi t
to the fourth session on the topic 1 made a similar observation:

When the discussion started, the classroom climate changed suddenly. That
was most striking to me. That was far from a common image of a "bland
and contentless" foreign language classroom. The use of the Japanese
language did not secm to be hindenng the students from speakng up. (Field
notes, November 3, 1994.)

During the fourth and final session on the topic of Japan's nranga-burtmii. the students were
given a list of questions in writing about the merits and drawbacks of reading comics. They
were asked to discuss the questions in a small group of ihree or four and formulate their
collective responses to share with the rest of the class. My fie d notes on this session
record:

During the 30-minu te discussion the students and teacher were drawn into
 ivlint was k i n g said. Even the norrnally quiet students like Jane and Charles
s p o h up to offer their opinions. The whole class was, it seemed to me,
competing to say their say irz Jupurzese . (Field notes, Novem ber 3, 1994.)

To wrap up this discussion, Kei wrote down in point form (and in Japanese) o n a
blackboard ail the positive and negative points of reading comics raised by the students.
Most of the students diligent1y copied what Kei wrote on the board in their notebooks. A
few months later Kei retrospectively descnbed this discussion as a pre-writing activity
utilizing brainstorming and listing techniques (Interview notes, February 23, 1995). In the
last five minutes of this session Kei explained dl in Japanese what was expected of the

first major writing assignment (Task 1). Foilowing is an abbreviated English translation of
her instructions. They were given orally; no wntten instructions about the task were
provided.

Write the merits and drawbacks of rnartga-buurna.. So you might Say,

"There are such and such good points but there are such and such bad points
as well." Then write your own opinions. You don't have to write al1 of these
(referring to the good and bad points written on the blackboard). You can
focus on a few of these or other points. It is also possible to state your
opinions a t the beginning of your essay then you talk about good and bad
points in order to support your viewpoints. What 1 redly want to see in your
essays is how you look a t this manga-brironit in Japan, that is, your own
opinions about the phenomenon. Also when you discuss the merits and
drawbacks. give esamples and use them to support your stances. (An
abbreviated English translation of Kei's instructions for Task 1.)

Kei repeated the above information, in slightly different ways, a couple of times while
giving the instructions for this assignment, probably wishing to be assured that her points
got across to the students. Kei also required the students to make a n outline in either
Japanese, English, or both, to organize their ideas M o r e starting to wri te an essay. The
students were asked to turn in their outlines as well as their final products. There were no
questions about this assignment [rom the students.
I n hindsight Kei thought this task was too controlled to allow originality.
According to her, most of the students' essays contained no more than what ivas talked
about in class (Interview notes, January 21, 1995). The students perceived variously the
provision of specific information on content, rhetoncal patterns, and language for Task 1.
as revealed in their written comments given in the Task 1 pst-writing questionnaire. 13 For
instance, Judith cvho "ivas not so much motivated to wnte on rriatiga [comics]" felt that the
pre-writing discussion in class had made the writing task easier. But there were Cathy and
Mary who had initially thought that the task could be easier because of the pre-wnting
discussion but in fact found the task to be still difficult. Both attributed the difficulty to
their lack of linguistic cornpetence in Japanese. Colin had something interesting to say:
It is easy to :nite a composition when al1 the necessary information
including vocabulary and sentence structures [as Kei did] are laid out on the
table. In a sense, the difficulty for a foreign language leamer is that the
writing process c m be quite technical. After all, to achieve greatness in
writing, unconsciousness plays a major role. (Colin's wntten comments in
the Task I pst-writing questionnaire.)

3.4.2. Kei alid Clarsroom Activities Leading tu Wriang Task II

The session right after the completion of the first theme study was the beginning of
the second theme study, living in Japan. This study lasted 5 weeks with 3 sessions before
and 2 sessions after the winter break. It proceeded differently [rom the first theme study.
First of all, instead of focusing on one particular topic of the chosen theme. reading
materials for this theme study covered different topics such as young people's employment
opportunities, male-female differences, and societal structures in Japan. They were al!
contrived to suit the leamers' Japanese proficiency levels; a list of vocabulary needed for
each reading was given. Kei's instruction included lexical and structural exercises.
comprehension check (display) questions, and extended (referential) questions for
discussion. 1 noted a shift of instructional foci. Statements Iike the one below, which Kei
made in the second session of the new theme study, were never heard during the course or
the first theme study.
- -

13 See chapter 3 for detailed information about individual students and chapter 4 for an explanaûon of
 Do you want me to go over some grarnmar as well? (Some students nodded
their heads in agreement.) Some of you seem to be needing some grarnmar
exercises. So let us work on this (refening to structural and lexical exercise
worksheeis accompanying the reading text under study) today. Also while
reading, if you have any questions about grammar, ask me. 1 will explain to
you. (An English translation of Kei's Japanese utterances tnnscribed in
Field notes, November 17, 1994.)

In the interview we had two days after this class session, Kei made the following rernark as
if she were speahng to herself:

What 1 [i-e., Keij really want is to have the students express their opinions
and share their own experiences in relation to what we are reading in class. 1
want to do something more than explaining grammar and vocabulary,
something substantial and meaningful for their lives. But at least one third
of the students in the class are still struggling with structural and lexical
problems. How can they possibly engage in meaning-construction processes
without a sufficient degree of grammatical and lexical control over the
Japanese language? (Interview notes, November 19, 1994.)

Around this time Kei was having individual conferences with the students to discuss their
Japanese language studies during her office hours and noted that "they have different

problems and these problems are mostly micro-level issues" (Interview notes, November
17, 1994).Apparenlly she saw the firne had corne, the time to pay more attention to micro-
level linguistic problems. On the other hand, meaning-focused activities. mostly in the
form of discussion, continued concurrently.
1 saw the curricular even& happening in these sessions a s a pedagogical application

of the interactive view of language processing in reading (e.g.. Adams & Collins, 1979;

Carrel1 et al., 1988; Lesgold & Perfetti, 1978) that aims at the development of top-down
and bottom-up language processing ski[ls, although Kei did not articulate her pedagogical
stance in such terms. I n making instructional decisions, Kei constantly experienced an
intemal stmggle to be coherent and accountable. For instance, in an interview after the
class that ernphasized the study of small units of text such as words and phrases, Kei
confided to me her concems about effects of language-focused esercises that tended to bc

rhis p s t - w n ting questionnaire.

mechanical: "1 wonder if those exercises were really useful to the students. It seems to me
that they were intellectually not stimulating at dl; nor were they facilitating the snidents'
leaming of language" (Interview notes, November 25. 1994). Her pedagogical concems
were, on the one hand. how to stnke balance between attention to language and attention to
content (described by Mohan, 1986, and van Lier, 1996, as one of the recumng problems
language teachers face) and on the other hand, how to provide meaningful language-
focused instruction.

During the winter break Kei was in Japan participating in a 5-week JFL teacher
training course. As it turned out, this experience brought significant changes to her
teaching in the subsequent classes after the break. There are two things about her
experience in Japan that are important and relevant here. Among the courses offered
during the training session, Kei found the course on the teaching and learning of E L
reading most usefuI. The course served as a refresher of what she had learned--such as an
interactive approach to reading and a schema theory-in a course on reading taken at her
graduate school in Canada in the previous year. She was amazed at the advanced Ievel of
her understanding of the issues lectured on and discussed in the classes this time. She tried
to explain to me, "My better understanding of theoretical issues this time was. I think,

because the course was offered in Japanese" (Interview notes, January 5, 1995). She did
not dismiss another possible reason, her previous exposure to the same or similar
information. But she apparently preferred to emphasize the former reason. More than once
in the same interview, she said to me, "It was so much easier to do things in Japanese. I
could read so much and so quickly yet 1 could understand better [in Japanese]" (Interview
notes, January 5, 1995). She then added. "I've never realized that doing things in a foreign
language can be so difficult" (Interview notes, January 5, 1995). This realization of hers
made her more sympathetic toward her JFL students grappling with language problems.
Consequently, she began to see micro-level language issues more senously than before.
Another important point about Kei's experience in Japan was her learning of the
J F L text readability measurement research based on Klare's ( 1974) work. Using the
formula given by this JFL readability measure, Kei assessed the readability of the written
texts she had used in the first hall of the Japanese reading-and-wnting course and figured
that they were far beyond the proficiency levels of most students in her class. She began to
think "the texts were probably linguistically too difficult for the students to activate fully
the cognitive faculty that processes language for connection and comprehension"
(Interview notes, January 5, 1995). My interview notes record:

Kei paused to laugh a t herself, saying, "Ignorance is bliss." She then

reflected on her teaching in the previous term: "While having the students
read authentic o r quasi-authentic texts containing words, expressions and
synîactic structures that were beyond the readability level, 1 would tell
them, 'Don't mind about the words and expressions you don? understand.
Don't be too concerned about them. Read it through quickly, and see if you
can grasp the gist of it."' Kei now thinks that such an approach could have
been of little use to the students and that it is a manifes~tionof the teacher's
lack of responsibility. (Interview notes, January 5. 1995.)

Classes after Kei's trip to Japan differed from the preceding ones. They were more
carefuliy pre-organized and systematic, following the sequence of content schema-setting
activities, linguistic schema-setting activities, and reading comprehension esercises. The

wnting component remained as a subordinate o r extended activity of reading exercises.

But there was language-focused instruction, though limited, reiated to wnting assignrnents.
In one class session where the second wnting assignment was announced, Kei drew the
students' attention to the "power of expression" (Field notes, January 16, 1995) and spent
30 minutes of the class time introducing and practicing a great variety of expressions to be

used in expressing one's opinions and making an assertion that are subtly different from
one to another depending on the wri ter's intention and the degree of his o r her confidence
in the statement he or she is making.14 This particular focus of instruction l5 was, as Kei

14 [a)
Forty-seven expressions of this kind were presented to the students. They included ornou think],
omowareru [it secms to (me)], dervmakarooka [isn't that ...?], and nakerebariaranai [ m u t beldo].
15 The usefdness of such instruction was meotioned expressly by Jme in her written comments in the
Task II pst-writing questionnaire and during the interview. It is unknown, however, how the other students
thought of such language-focused instruction.
had exphined to me in our interview two days earîier (Interview notes. January 34, 1995).
prompted by her discovery of the flat and du11 tone of the students' writing in the previous
assignment. Kei thought their compositions lacked the variety of such expressions.
The second writing assignment (Task II) was to write an essay based o n the
readings about living in Japan. Kei t a i h d about the assignrnent for 2 minutes, a11 in
Japanese. As in the case of Task 1, no written instructions were given. No question was
raised by the students about this assignment. The following is an English translation of
Kei's instructions about the task:

W e have been reading on various topics about living in Japan f o r the past
several weeks. For this writing assignment, I want you to pick up one from
those topics you've found most interesting and present your opinion on it.
When you write this composition, you may, for instance, Say something like
"1 have read such and such a thing in class but I think such and such. 1 agree
with these points but not with those points. And my opinion i s such-and-
such" o r something like that. 1 would like you to follow this kind of
presentation pattern. And &O, you don't have to limit yourself t o the course
readings. You may refer to o r quote from the readings you may have done
on your own. (An English translation of Kei's instxuctions for Task II.)

A s it turned out, nor al1 the instructions were heeded by the students. It was
"disappointing" for Kei to find that most of the Task II essays were merely descriptive or
expository without the statement of one's opinions. She regretted that not al1 of her
instructions about this writing task were acted upon by the students. Nonetheless, she was
pleased to see more originali ty in the Ta& LI essays and thought that "they are better
wri tten than the Task 1 essays" (Interview notes, February 23, 1995).

Kei's approach to JFL writing was undoubtedly what Reid (1993) called "writing-
based" n t h e r than "language-based" (p. 29). Kei's teaching emphasized JFL wnting as a
cognitive ski11 rather than a language or grammar skill. This d w s not mean. however, that
Kei dismissed the importance of "micro-monitoring" ski11 in discourse production; rather,
her intended focus was to allow interaction between "macro- and micro-moni toringn (Ellis,
1994, p. 132) skills in the process of creating compositions.
For both the Task 1 and Task II assignments Kei required the use of a specific
organizational technique for presenti ng wri tten discourse: the development of paragraphs
by presenting pros and cons of a selected issue and stating one's opinions about them.
However, this organizational pattern was inconspicuous or simply non-esistent in the tests
written for Task II, whereas it was manifest in evet-y student's writing for Task 1. As a
consequence, the two writing assignments stand in contrat in tenns of their modes of
discourse; Task I was argumentative and Task II fundamentally descriptive or informative.
There are other differences as well. The Task 1 assignment was, in a way, more
controlled than the Task II. No choice of topics was given in the former task, whereas in
the latter the students were free to wnte on a topic of their choosing under the broad
common theme of living in Japan. A discussion preceded the assignment of the former as a

pre-writing activity utilizing brainstorming and listing techniques. No sirnilar kind of

discussion took place for Task II; there was, however, a provision of a list of predicative
expressions prior to the Task II writing assignment. From the students' perspective, to

paraphrase some of their remarks about the two tasks from individual interviews (see
section 3.9.3), "not on1y the necessary information" (i-e., content) "but also the structure"
(Le., rhetorical organization) "was provided" for Task 1, but Task II was "a write-whatever-
and-however-you-want kind."
 Chapter 3: Students' Characteristics

It was on my second visit to the class that I was given the first 10 minutes to talk

about my research and to distribute a letter asking students to volunteer for the research.
Volunteering was considered desirable as participants would have to be willing to offeer
extra time and energy to complete such tasks for the research as producing extensive

verbal reports while writing compositions and attending interviews. The letter was written
in English. stating the nature of the research tasks to be camed out by participants
(Appendis D). I t also indicated that the research findings 1 would obtain were to be used
for doctoral research and that participants would be given a stipend of $100 as tvell as
individualized suggestions for irnproving their Japanese wri ting in retum for their
participation in the study. Thirteen students signed a form of consent to participate in the

study by the end of that class session. l 6 This unespectedly large tumout (86.7%)was. as I
found out later. mainly due to my offer to pay US$60 for each person to take a
standardized onl proficiency intenriew test, one of the required research tas ks. Apparent1y
they were anxious to know how proficient they were in Japanese through a standardized
means of assessment.

16 One student decidd to participate Iater but withdrew after a trial think-aloud session.

36
 A questionnaire and interviews were the primary methods 1 used to gather profile

information about the student participants. Al1 necessary contacts with the participati ng
students were made while 1 was in the classroom, normally during the time waiting for the
teacher to a m v e and begin a lesson. No request was made of Kei to spare any of her class
time for this purpose. Additional contacts were made by phone later to rnake sure that the
participants received my instructions correctly and to provide them with an opportunity to
ask me questions.

A week after they signed the consent form for their participation in my study. 1
gave the students a Cpage questionnaire (Appendix E), asking them to complete i t at home
and retum it to me in the following class session. The questionnaire was wntten in English
and based on the background questionnaires used by Cumming (1988, pp. 322-224) for the
study of ESL writing and by Oxford (1990, p. 282) as a supplement to the Stntegy
Inventory for Language Leaming instrument. The questionnaire sought information on (a)
general and linguistic background, (b) experience with the Japanese language, (c)
motivation for learning Japanese, (d) self-evaluation of Japanese overall and wri ting
proficiency and of first language (Ll)17 wnting proficiency, and (e) writing habits in
Japanese and LI. Table 3-1 presents a surnmary of the participants' responses to the
selected items of (1) ethnicity, (2) oral LI, (3) written L1, (4) self-rating of L l writing
proficiency, (5) degree of importance of learning Japanese, ( 6 )length of residence in Japan

- - - - -

17 In this thesis, the term Jîrst language or LI refers to the language a person feels most cornfortable
using. It may or may not be a person's rnother tongue or the language aquired first. A persoa may have
different LIS for oral and written communication. Sm Richards, Platt, and Platt (1992. p. l m ) and Stem
(1983a. pp. 9-13).
in the past 5 years, (7) self-rating of overall Japanese proficiency in cornparison to JFL
pers in class, and (8) in companson to native speakers of Japanese, and (9) self-nting of
Japanese wnting proficiency in cornparison to JFL peers in class, and (10) in cornparison
io native speakers of Japanese.

Table 3-1
Strtderr ts ' Characteristics

Partici-
- Wri tten Self-
7
Self-
pant L1 rating of rating of
Japanese Japanese

Jasmine English
-writing writing
cf. cf ass cf. native

[air

Jane English Poor

Judith
JOy
FPnlpl
Ja ese En ish

,
Engiish
English
poor
fair

Jack EngIish fair

Cathy

Colin

Charles Chinese

Mary

Mike
-4-
Chinese s~andyin

English
- -

fair

Martin English poor

l
Eliot English exallent ~~poriïnt L year g d poor poor

Eddie goai
"m i year fuir POO' fair

3.2.1.1. Participants' Ethnicities and Lls. Each of the 13 students who stayed

throughout the research penod was identified by a pseudonym beginning with the first
letter of the ethnic group to which he or she belonged. Five of them (Jasmine, Jane,
Judith,l8 Joy, and Jack) were Japanese by ethnicity; ail escept Jasmine were born to native
Japanese parents and schooled in Canada. Jasmine, whose parents were also native
lapanese, was born and raised in Brazil until s h e moved to Canada with her family at
Grade 4. Six students (Cathy, Colin, Charles, Mary, Mike, and Martin) were Chinese by
ethnicity. Cathy, Colin, and Charles spoke Cantonese as a mother tongue and were
educated in Hong Kong until they rnigmted to Canada with their families 5 o r 7 years ago.
Mary, Mike, and Martin spoke Mandarin. Mary and Mike came to Canada as immigrants
when they were in Gndes 7 and 4 respectively, w hereas Martin was bom to a Mandarin-
speaking family in Canada and schooled in English here. [t may be important to note,
because of some shared elements of the Chinese and the Japanese writing system (see
Taylor & Taylor. 1995), that the 3 Cantonese-speahng students and Mary possessed a
relatively high levef of Chinese litemcy, that Mike's Chinese litency was at an elementary-
school level, and that Martin could speak but neither read nor write Chinese. Eddie and
Eliot were white anglophone Canadians of Anglo-saxon origin. Since 1 know al1 these
languages myself, 1 felt capable of handling these participants in the research.
3.2.1.2. Participants' Ages and Major Fields. The participants' ages ranged [rom

18 to 25. Al1 the students were in the final year of their undergraduate studies escept for

the 9- sophomores, Jasmine and loy. Five students (Joy, Jane, Jack, Martin, and Eliot) did
not choose Japanese-related subjects as a major field of their studies; their majors were,
respectively, immunology, psychology, commerce, electncal engineering, and physical
pography. The other students majored in Japanese Studies alone o r in combination with
other subjects such as economics and linguistics. Colin was then p u n u i n g a second
bachelor's degree in Japanese Studies; he had previously obtained a bachelor's degree in
science [rom the same university. Most of the smdents were multilingual, knowing
languages other than Japanese and English.

l8 Judith had a Japanese mother and a Chinese father and comrnunicatcd with the mother in Japanese
but in English with the father. Her Chinese roou were thin as she spoke no Chinese. For this reason. 1
wtegorized her wi t ' i n the Japanese ethnic group in this shdy .
 3.2.1.3. Participants' Experierices with the Japanese Lunguage. The students'
esperiences with the Japanese language varied. Jasmine. Jane, Judith, and Joy had, as it
were, lived with the Japanese language throughout their lives, whereas Jack stopped using
Japanese altogether on entering kindergarten; he then decided in adulthood to re-leam the
language by taking Japanese language courses at the university. Ail the other students took
up the study of the Japanese language through formal instruction after entenng the
univenity, though some students had tned to learn some Japanese by themselves before.
Eddie spent a year at a university in Japan as an eschange student in his junior academic
year in 1993-94. Martin and Eliot spent the same year in Japan but as Company trainees
rather than as students. This arrangement was part of the university's special Japanese

language prognm for professional purposes in which Martin and Eliot received a year of
intensive formal instruction in the language prior to their departures to Japan. by attended
for two years an international high school in Japan, where al1 subjects were taught in
English, before she enrolled in the university in Canada. Judith was enrolled in a JFL
teacher-training course in Japan for a year in 1992. Al1 the other students had neither been
to Japan nor visited there.
3.2.1.4. Participants' Motivatiotis for Learning Japanese. Despite their
differences in age, major field. ethnicity, and experience with the japanese language, the
students were al1 motivated to pursue their study of Japanese. The?; indicated high levels of
cornmitment, interest, and motivation at various occasions. According to their responses to
Item II-(@ of the background survey questionnaire, they were al1 interested in the Japanese
language and culture and needed or wished to use the language for their future careers. The
students of Japanese ethnicity, except Judith, mentioned that their Japanese language study
was also motivated by their desire to maintain ties to their relatives in Japan and to
preserve the sense of their own heritage. Furthermore, the latest news of the students 1
received in spring 1996 proved that their motivation at that time was substantial and
continuing. lasmine, Jane, and Judith had since been training to be JFL teachers. When his
plan to live in Japan upon graduation failed, Jack went on to a law school at another
university in Canada espressly to becorne a specialist in legal matters concerning Japan-
Canada relations; he chose to live in a dormitory where the majority of residents were
Japanese students [rom a university in Japan. Cathy had a full-tirne job but was worlung as
a volunteer clencai assistant for a Japanese heritage-language program every Saturday just
because she wanted to stay in touch with the Japanese language and people. Colin was now
preparing to write a thesis on Japanese history for a master's degree which he had been
pursuing at the same university alter graduation. He was also hoping to be awarded a
scholarship to conduct research in Japan. Upon graduation Mike retumed to his native
Taiwan [O help his father's business that had a number of contacts with companies in
Japan. Martin was completing a master's degree in engineering a[ an Arnerican university
and had k e n promised ernployment by a large Company in Japan. Eddie had recentl y been
admitted to a master's program at the university of his undergraduate study and was
planning to do research on Japanese religion. Eliot was pursuing a master's degree in Japan
on a scholarship. It was toward these goals that the students probably wanted to improve
and refine their Japanese during the year 1 conducted my research. I n fact most of them
were concurrently taking the other highest level course of Japanese speaking and listenine
taught by another instructor, another apparent indication of their high motivation for
Japanese language study.

3.2.2.Assessrnent of JF L Proficiency arid L1 Writittg Expertise

Relatively objective methods were used to assess the participants' Japanese

proficiency as well as their LI writing expertise. The Japanese Speaking Test (JST)
(Center for Applied Linguistics, 1992) was administered in January 1995 to measure the
participants' oral Japanese proficiency, and its score reports were retumed in March 1995.
The JST is a simulated oral proficiency interview test derived from the Speaking
Proficiency Guidelines of the American Council on the Teaching of Foreign Languages
(ACTR). It is suitable for evaluatîon of the level of oral proficiency in Japanese attained
by English-speaking leamers o l Japanese in pst-secondary institutions. Euaminees' taped
responses are rated accordîng to the ACTFL Guidelînes at the Center for Applied
Linguistics by a cenified rater.
Since no standardized test of proficiency in JFL writing was currently avvailable, the
two compositions (sec chapter 2 for a detailed description of these tasks) that the students

had wntten for homecvork assignments in the course (and that are the foci of my
investigation) were used as sarnple essays for the assessrnent of the participants' Japanese

writing proficiency. The length required in both tasks was about 2 pages of regular-six
writing paper or gerikooyooslii (Japanese and Chinese manuscript paper conwining 400 or
500 small squares into each or which one chancter fi ts). The average num ber o l characters
per essay was around 800 to 1000. When 1 obtained copies of the submitted essays, which
were al1 handwri tten, 1 typed them before giving them to nters for fear that the writers'
idiosyncratic handwriting and the untidy appeannce of some of their tests would affect the

nters' judgments of the quaii ty of the essays 19 (see Appendis F for typewritten sample
essays). In making these transpositions to typed test, 1 did not correct or edit the essays,
however.

For the L1 essays, al1 the "J."~O"M" (escept Mary), and "E" students wrote

compositions in English of MO to 600 words in length under the ritle. "What the English
language means to me." The "C" students and Mary wrote compositions in Chinese of 700

to 1 0 0 characters in length under the title, "What the Chinese language means to me."
Cathy, Colin, Charles, and Mary, who considered Chinese to be their wntten LL and wrote
Chinese essays, also produced essays in English under the same title " What the English

19 Some people may consider the quality of handwriting as one feature not to bç negfected in assessing
an essay (Vaughan, 199 1). For the purpose of the present research, however, 1 decided that readers should not
confuse the appearancc of a papa with the quali ty of what is said and how i t is said in the papcr (CarIson.
1991, p. 308).
20 Joy failed to produce an L1 essay.
language means to me" 4 months dter their production of Chinese essays, to see if there
were any differences in the quali ty of their writing in the two languages. Al1 these essays
were wri tten expressly for my research and produced at home on participants' own time.
Most of the English essays were wordprocessed and al1 the Chinese essays were
handwritten. The handwritten English essays were typed without correction or editing, for
the sake of consistency, before submitting them to nters. The Chinese handwriting was
neat and the appearance of the essays was tidy. Therefore, the Chinese essays rvere given
to n t e n as they were. The topics yielded both narrative and descriptive modes of rvnting.
Incidentally, these essays provided me with useful information about the writers' linguistic
experiences and helped me understand the individuai students better.
3.2.2.1. Writing Assessnlent Itislshument. There is no single essay scoring scheme

applicable to al1 Ianguages or dl contexts (Purves, 1992). But there seem to be certain
common elemen ts essential to corn p s i tions wri tten in any contest. Jacobs, Zinkgraf,
Wormuth, Hartfiel. and Hughey (1981, p. 34) pointed out three essential elements focused
o n by composition teachers, testers, and students since the time of Arktotle: what 20 sav
(content), hoiv IO orgnniie it (organization), and Izow to say it efectively (use of the units
of language). Because of these salient features of any piece of writing, 1 iavored multiple-
trait, nther than primary-trait or holistic, scoring (see Hamp-Lyons, 19%b, for an account
of different scot-ing procedures) for the assessrnent of the English, Chinese, and Japanese
essays. Besides, "the multiple trait procedure possesses psychometric properties that
enhance the reliabiliiy of single number scores built from its components" (Harnp-Lyons,
199 1b, p. 151).Deciding how to obtain comparable scores between essays of different

languages in different contexts, 1 was faced with two choices. One choice was to ask raters
to make judgments on their own intuitive criteria and to rate on a 4- or 5-point scale the
quality of each of the three main components--content. organization. and language use.
The other option was to use an established instrument with scoring criteria capable of
scoring an essay on the above components, pending the question of "tnnsferability"
(Hamp-Lyons & Henning, 1991, p. 338) across languages o r tasks. 1 chose the latter option
and considered Ham p-Lyons's 9 - p i n t s c d e as a candidate (Harnp-Lyons, 199 1a).

Hamp-Lyons's scale, known as the New Profile Scale, was developed to be used in
scoring the wnting component of the British Council's English Language Testing Service
(ELTS) test (1980- 1989) taken by pst-secondary nonnative speakers of English wishing
to attend college o r uni versity in Britain (Hamp-Lyons, 199 1a; Hamp-Lyons, 199 1b;
Hamp-Lyons & Henning, 1991). Naturall y the task type and scoring procedures of this
writing test reflect what is vaiued in writing in the western academic contest. One obvious
esample is the inclusion of the argiirnentatioti traits in the scoring instrument. Considering
apparent differences (in languages involved, purposes, tasks, writing conditions, etc.)
between the ELTS and rny research context, the application of Hamp-Lyons's scale (the
version given in Hamp-Lyons, 1991a. pp. 149-151) to the latter may seem problematic. On
my examination of the instrument, however, 1 found the criteria or band descriptors fairly
general and appropriate for the assessment of the essays 1 collected. 1t certainly remains a

big question whether Harnp-Lyons's 9-point scale is a valid measure in assessment

contests different from that for which the instrument was designed. 1 used the instrument
in its original English r o m , nonetheless, with a slight modification. The modification was
to esclude the category of argrrineti~a~iorr .
which did not fit the modes of writing used in
my study (except for the Japanese essays written for Task 1). Although the recommended
practice of multiple trait assessment is to treat each trait separately (Harnp-Lyons, 1995). 1
corn bined trait scores and used the total score as an indication of each participant's wri ting
proficiency in each of the two languages (Japanese and Ll), following the method of Jacob
et al.' s ( 1981) b'composition profile." The maximum total score was thus 36. buil t from iis
component scores in cotntnrtriicative qrcality, organiration, litrguis~icaccicracy , and
linguistic npproprincy, which are weighted a maximum of 9 points each, equally?'

21 According to Hamp-Lyons (1991b). "the safest way to combine scores is to weight each facet
equally" (p. 249; cf. Jacobs et al., 1981).
(Appendix G). 1 used trait scores to examine the qualities of Japanese essays (see chapter

3.2.2.2. Writing Assessrneni Raiers. In all, seven volunteer raters were involved: 2

English, 3 Chinese." and 2 Japanese (see section 5.1.1.1for more information about the
Japanese raters). They were al1 native speakers of the respective languages. The Chinese
and Japanese raters were al[ balanced bilinguals of Chinese or Japanese and English, which
made it possible to present the scoring instrument in its original English form. Al1 the
raters were experienced language teachers familiar with recent research on writing. One
Chinese rater had taught for 5 yars and al1 the others had teaching esperience of 10 to 18
years. One held a Ph.D. in language education; two Engiish raters held master's degrees in
Teaching English as a Second Language; the rest were complcting a Ph.D. in second
language acquisition at the time of the nting. They were told that scores would be used
only for the purpose of my research and that the wnters would have no access to the
scores. They were also informed that the essays were written by native or non-native
speakers, but no other information on the wrî ters was disclosed for ethical and halo eflect
considerations. In fact the raters had never met the participating students. Due to some
diffïculties in making arrangements, no special training was provided to the raters except
the written instructions on the nting procedures (Appendix H); nor were the conditions
under which the rating was performed strictly controlled. None of the raten had used
Harnp-Lyons's scale before. Despite al1 these unfavorable conditions, intemter reliability
correlations turned out relatively high, .76 for the Japanese essay ratings, .95 for the
English, and .90, -92, and .98 for the Chinese. These correlation coefficients were
considered sufficient to assure the ntings for this study were reliable.

22 Three raters, instead of 2 as are commonly utilized in many writing assessments, scored the Chinese
essays. A third rater. a Hong Kong Chinese. was added to the fmt two volunteers. who werc bolh cducated in
the People's Republic of China (PRC). because al1 the Chinese essays in the present study were wrirten in the
original form of Chinese charaeters used in Hong Kong and Taiwan nther than the simplified version which
is the standard written Chinese in the PRC.Among these C h i n e s e - s p e - gplaces, however, there should be
no differcnt standards applied ta rtietorical and grammatical requirements in a piece of writing.
 3.2.2.3. Participants' Ltzriguage-Related Cltaracteristics. The students'
characteris tics on language-related factors are summarized in Table 3-2 (an expanded
summary of the students' characteristics appearing in Table 3- 1). This summary presents
the individual's Japanese and English or Chinese essay scores given by the two or three
nters as averages. '3

Some interesting features appear in Table 3-2. Al1 the ethnic Japanese students
except Jack were nted sirperior or ndvmrced-higlr on the JST and recei ved relative1y high

scores on their Japanese essays (28 to 34) as well as on their L1 cssays (30 to 35). The 3
non-ethnic Japanese students with recent residential experience in Japan (Martin, Eliot, and
Eddie) were nted odvauced-liigli or advnnced on the JST but their scores on the Japanese
essays (23 or 23) were not as high as the scores of the high JFL-proficient ethnic Japanese
students. although their scores on the L1 essays (31 to 34) were comparable. The other

students fell in the infermdia~eband of the JST rating scale; their Japanese essay scores
m g e d between 15 and 26 while their LI essay scores were between 23 to 35.A condensed
surnmary of these trends is given in Table 3-3.

23 1 was toId during interviews that Mary, Martin, and Eliot had their Japanesc friends r a d and correct
their Japanese compositions for Task II before handing h e m in. Thus, the scores given to their Task II essays
were excluded from this calculation. Ai1 other essays were submitted without any input from others, as 1
verified during the interviews. Two sets of Japanese essays (Le., Tasks I and II) were averaged. for al1
participants except h l q , Xhîin, and Eliot, to generate scores indicating their Japanese writing proficiency.
Cathy, Colin, Charles. and Xlary received two sets of scores for their L1 writing, one for the Cbinese
and the other for the English essay. Thcir Chinese çssay scores wçre substantidly highcr by 4 to 8 points so 1
u s a i that measure as an indicatian of rheir L 1 w ri ting expertise.
 Table 3-2

M c i -
P t overdl
writing
witing
WY.36
cf. clas; native
I cf. ctass: oative

Jasmine fair 30 O pood; fair superior fair. poor 34

Jane fai r 34 O g d ; fair

Judith 1 fair 35 1 year good; fair

Joy English 1 EngIish superior 1 fair. poor 1 18

Jack English 1 English I excellent 35 2 months gaxi; poor

Cathy

Colin
-L
Cantonese

1
Cantonese
I
Chinese

Chinese
1
fair 30 O Pmr. Pmr

4-
+
Charles Cantonese Chinese

Mary Mandarin Chinese fair 26 1 O fair, poor

1
Mike Mandarin EngIish fair 1 2 3 1 O 1 fair.poor

~Martin excellent 32 1 year good; poor

advanced
1
Eiiot

Eddie
-4-
English English excellent 32 1y m
1
gd;poor

advanced 1
g d ; poor

[air. poor 1 33

Table 3-3
Surnrnary of Striderrts' L1 Wriiirzg,JFL Oral. and JFL Writirzg Skiiis

1
Participants LI essay scores E L orai proficiency leveis JFL essay scores
superior
"Ji's (except Jack) 30 - 35 advanced-hi& 28 - 34
advanced-high
"E"s& iMartin 32 -34 advancd 22 -23
Jack, "C"s& %l"s
23 -35 intermediate-high, mid, low 15- 26
(except Martin) h

Another curious trend in the data is that the high JE-proficient ethnic Japanese
students evaluated their cornpetence in writing in Japanese to be only fair or poor w hen
they were asked to assess it as compared to other students in their class, but they
considered thei r overall Japanese proficiency to be good in corn panson wi th thei r peers.
And their Japanese compositions, in fact, received high scores. Eliot, Eddie, and Martin
rated their overall and written Japanese likewise as good orJair relative to other students in
the class, which does not fit well the gap we find between their JFL oral proficiency levels
and J E essay scores assessed by JFL specialists. The intermediate JFL-proficient students
(escept for Jack) also rated their o v e d l and written Japanese similarly but at the lower end
of the scale, i.e., fair or poor. 1t is difficult to make sense of this trend. But one thing that
might be said is that the orally proficient ethnic Japanese learners were overly concerned
about their less skilled writing ability and underestimated their actual cornpetence in
writing in Japanese. In other words, writing was probably a specific concem for them,
whereas, for other students, it was not necessarily a focal point of their Japanese language
studies but just one of many aspects to be improved.
For the participants' self-ntings of L1 writing skill, Jack. Colin, Martin. and Eliot
rated theirs as excellent. If we suppose that the rating of excelle~ztrefiects a high degree of
confidence in written language, such confidence indicated by Jack and Colin does not
seem to have been camed over to their Japanese writing esperiences: they assessed their
Japanese writing skills as only fair or poor. This may be because of their insufficient
ability to use the target language (though Jack indicated his confidence by rating his
overall Japanese proficiency as good). On the other hand, Martin's and Eliot's ratings of
their Japanese writing skills as good rnight be related to their confidence in writing,
Japanese, or both. The other students nted their L1 ivriting skills as good or fair and might
have considered themselves "average" writers who could write (but not particularly well)
or "inexperienced" wri ters who had little pnctice in wri ting. Self-reporting and self-
assessrnent data pose questions of reliability and validity, and they necessarily have
limitations. Nonetheless, such data can assess people's psychological s tates about the
matters quened. In my study these self-rating data were consulted to examine participants'
attitudes toward Japanese and L1 writing.

My study focused o n the two major writing tasks the students worked on for the
course. How did the students perceive the tasks? What were their expectations o r targets in
these tasks? T o understand the students' viewpoints, 1 met wi th the participants
individually or in pairs (according to their preferences)ZJ for 20 minutes to 1 hour between
Febniary 28 and March 16, 1995, after their completion of the second task in rnid
February. The interviews were conducted in English and structured around four main
questions probing for information on the students' perceptions of JFL writing and leaming.
The questions were:
1. What do you think you learned from writing the two compositions for the course?

2. What expectations d o you have for your own compositions in Japanese?

3. What difficulties d o you have in writing Japanese compositions? Or what are the
frustrations you feel while wnting in Japanese? And how do you deal with them?
4. What functions does composition writing play in your study of Japanese?
Before they attended the interviews, I told the students briefly about the questions and
encounged them to give them some thought. My interview guide (Appendix 1) contained
not only these questions but also synoptic notes of what 1 was to say at the beginning of the
interview, follow-up questions to each main question, and questions for additional
information. Dunng the interviews 1 was primady a listener seeking information from the
interviewees with occasional interruptions for clarification and follow-up. AI1 the
interviews were audiotaped with the interviewees' permission and later transcnbed
verbatim.

24 There were 7 individuals and 3 pairs (Jasmine and Martin. Judith and Colin, and Joy and hllary).
 The prepared questions for the interview were deliberately made broad and
redundant, probing for the interviewees' prime concems from different angles. A perusal
of the interview transcripts led me to identify five themes commonly talked about: (a)
conceptions of wtiting in Japanese, (b) expectations about one's own Japanese writing, (c)
problems in writing in Japanese, (d) experiences of leaming through wtiting in Japanese,
and (e) conceptions and experiences of wri ting in LI.

3.3. Ans wering the Research Question

Hoiv do stzide~lts'characteristics relate to their perceptiom of JFL

What stood out in the interview data (the full summary of the data appears in
A ppendix J) was the leamers' preoccupation wi th language production. Their prirnary
concem was how to put their thoughts into appropriate language. Apparently writing in
Japanese in this context, for them, was synonymous with ski11 in usage and structure of
language. It appeared that the students, regardless of their levels of proficiency in
Japanese, viewed JFL writing from the standpoint of language rather than ideational
content. Some s tudents were more explicit on this point than others during their interviews
with me. For instance, Martin did not "think the content is really importantwx; for Mike.
writing a Japanese composition was "a chance to practice using Japanese, focusing on the
right word choice and stuff'; and Eliot viewed writing Japanese compositions as "one of
the ways of getting better Japanese." Grammatical correctness was emphasized by Cathy,
Charles, Mary, Martin, and Eliot. Judith, Joy, and Charles felt that their lack of vocabulary

25 This and other quotes in section 3.3 wcre talrcn from the transcripts of respective individuals'
rernarks during their interviews with me, unlcss indicated othcnvise. The words are minimaiilly edited for
readability.
was hindering their writing processes. Such specific problems as the writing of kanjiz6 and
sentence-final expressions were raised by lasmine, Jane, Judith, Joy, and Eddie. Further,
most of these students associated their experiences of leaming through JFL writing with
the learning of some aspects of the target language. Apparently, writing compositions in

Japanese forced these IFL leamen to think about syntactic and lexical properties of the
Japanese language (cf. Swain & Lapkin, 1995). and this process might have served
functions conducive to L2 Iearning, as suggested by Swain (1995).
On the other hand, no one really mentioned the problem of coming up with ideas
for essays during the interviews (although Judith, Joy, Cathy, and Eliot spoke of their
concems about putting ideas together or the organization of ideas). Even Judith, who
struggled in vain to write something "personally meaningful" and "interesting" in order to
enjoy the writing processes, had enough things to write and did not talk about content as
her problem in writing in Japanese; instead, she saw kanji, vocabulary, and expressions as
hurdles. Joy "had to do a lot of research for the second [writing] task." Similarly, Charles
"had to get sorne information for the content of the essays." But it appeared that neither of
them saw this as a problem; rather, they attributed their content-related leaming
experiences to such work, and identified vocabulary, grammar, or both as their problems in
writing Japanese essays.
It seems that these adult lemers writing in a foreign language experience just the
opposite of what Bereiter and Scardamalia (1987) discovered about children's problems in
writing essays in their native language, ix., "problems of finding content, not of finding

26 Japanese texts are written in a mixture of different types of scripts. They use mostly Chinese
characters called kanji for content words and one form of a syllabary called hiragana for grammatical
morphemes. Because of the onhographic and semantic complexity, Japanese people expend an enormous
amount of time and effort on learning kanji. It is possible to write a Japanese text al1 in hiragana without
using kanji as in most texts intended for little children. But such texts, as Jane correctly pointed out dunng
her interview, "look very childish." See Taylor and Taylor (1995) for a detailed description of the Japanese
writing system.
27 Exceptions were Cathy and Charles. Cathy did not find anything substantial to have the sense of
learning in. Charles mentioned content learning, not languagc learning, as a result of his writing Japanese
essays; for language Iearning, he, rather, talked about "read[ing] a lot" as "the major point of studying
Japanese well." Jasmine. Jane, Joy, and Eddie indicated content leming (cf. Gere, 1985) as well as language
learning that might have resulted from writing Japanese compositions.
language to express it" (p. 62).*8Participants in my study, nther, had problems of finding
language to express themselves, not of finding content to wnte. In a way their attitude

toward J F L wnting overlaps with the picture of wnting behavior of unskilled mother-
tongue college students that has emerged [rom a nurnber of writing process studies (e.g.,
Perl, 1979; Pianko, 1979; Shaughnessy, 1977). depicting such writers as persistent1 y
concerned about sentence-level correctness.
For the students in my study, JFL writing was primarily a language production
activiiy. However, some students did imply that they also paid some attention to the aspect
of wnting concemed with "meaning making" (Flower, 1993, pp. 17-23) that pictures
writers engaged in the process of representing their thoughts, feelings, and images in text.
Jasmine, for instance, "tried to wnte [a Japanese essay] as a short essay like putting the
points down and esplaining why"; Judith emphasized the importance of finding topics ihat
are "persona11 y meani ngful" and "interesting"; Joy, Cathy, and Colin approached the
Japanese writing task in such a way as they would d o in creating text in their Lls,

following "the prewriting [process] that is basically the same" (Joy), focusing on the points
to be made in a piece of writing (Cathy), or "searching Tor an image, the sou1 of an essay"

(Colin).
It is difficult to find any clear patterns about the relations between students'
characteristics and their perceptions of JFL writing and learning. However, it seems that
highly proficient, native-like leamers are more able to expend their cognitive energy on

content-related matters. This inference makes sense theoretically, if we apply concepts of

automaticity (e-g., LaBerge & Samuels. 1974; McLaughiin. 1987, chap. 6; Shiffrin &
Schneider, 1977). According to automatici ty theory, people w ho process linguis tic

elements skillfully and automatically in a certain task leave attention free to be used for

28 See also Durst's (1989) and Langer's (1986a. 19û6b) studies indicating that elernentary and
secondary school mother-tonguc students of English paid proportionally less attention to such issues as
mechanics, syntax, and lexical choices whilc they composed; instead. these students focused more on
fonnulating ideas and relhement of meaning.
higher levels of information processing such as deriving and making meaning. There are,
however, a few cases in my study that do not fit exactly with this picture. I examine two
such cases--Jme and Colin--below.

3.3.1. Jane's and Colin's Perceptiorrs of JFL Writing and Learnirrg

3.3.1.1. Jarre. it appeared that Jane, a native-like speaker of Japanese nted as

sicperior on the JST. considered JFL wri ting primaily as a language production activi ty. In
her interview she pointed to the "big gap" between her speahng and wri ting abili ties in

Japanese, saying, "1 didn't h o w [my Japanese writing skill] was this bad." Because of her
strong sense of her hpanese heritage, she felt that her Japanese writing should be as good
as the writing of native Japanese people of her age, using "more decontive, mature, and
sophisticated forms" of words and espressions and more ka~rji.Writing hpanese essays
provided her with opportunities to be aware of the need of refining her Japanese and to
work toward that laquage-focused goal. Asked about her sense of leaming through JFL

writing, Jane did mention some learning of the content she wrote for the essays as well a s
language use. But throughout her interview, she emphasized her concems about language
issues in conjunction with her cultural heritage. Apparently she did not give much attention

to meaning making during her production of Japanese test; instead, she focused on
language issues.
3.3.1.2. C o h . Colin had a unique view of writing. Writing was an essential means

for him to make his thinking clear. He compared the process of composing a piece of
wnting to that of "playing a jigsaw puzzle" as both seek for a coherent whole. This
process, according to Colin, is valid in any kind of writing including the JFL writing

assignrnents done in the course. Conceming specific problems of writing in Japanese, he

spoke of the importance of "getting into the logic of the Japanese mind and seeing things
from the Japanese point of view" by "getting exposed to and absorbing the way [Japanese
people] deal with a particular iopic and present it in writing." This is, as Colin said, "like

fiilling into the sea of Japanese culture." He continued, "There you c m pick up words and
expressions in the way [Japanese people] use natunlly." Bei ng an i ntennediate-level
leamer of Japanese, Colin saw language use as "a major barrier" in his production of
Japanese essays. But he took this fact calmly and confidently, saying, Wight now 1 am not
[at the stage where one can use language like a native] but this situation cannot iast for
cver ... When 1 read, 1 try to assimilate the [Japanese] language, give myself to it, and be

friendly to it instead of stmggling with it." Such remarks by Colin about Japanese language
learning are suggestive of his unique approach to language and language learning.
T o Colin, an experienced (but not professional) writer in his L1, writing in any
language is a culture-bound, meaning-making activity. He wanted to write something in
Japanese that "could move the Japanese reader." Creating a piece of writing that "is
coherent and well thought-out," as Kei spoke of Colin's Japanese writing (Teacher
interview notes, November 3-4, 1994), was probably an integral part of' Colin's conscious
and subconscious knowledge. What came into his mind as new knowledge or learning as a
result of writing JFL compositions were some aspects of Japanese language use. The JFL
wri ting assignments provided Colin with opporiunities to " think how to match the content
with linguistic expressions" of Japanese.
 Chapter 4: Writing Processes

4.1. Cenerathg Researck Mater*

4.1.1. Coüection of Thirik-AlotuiProtocois

1 employed the think-aloud procedure (Ericsson & Simon. 1993). which has k e n

used profitably in psychological studies of, m o n g others, reading (Pressley & Afflerbach.
1995) and wri ting (Bereiter & Scardmalia, 1987; Smagorinsky, 1989, 1994b), to trace the
thought processes of writers at work.29 In accordance with the emphasis of my study on the
social and psychological contexts of writing, the collection of think-aloud protocols was
placed in as authentic composing situations as possible. Rather than asking students
individually to appear in an isolated place to conduct a protocol session at an appointed

time in the presence of the researcher, protocols were produced at the participants' own
homes (cf. Smagorinsky's in-progress work reported in Smagorinsky, 1994~;see also
Smagorinsky, 1997).
On the day the writing task to be used for my research was given as an assignment
in ciass, the participating students received from me a packet containing 3 120-minute
blank audiotapes,'O a sheet of general instructions, a sheet of think-aloud instructions
(Appendix K), and a sheet of pst-ivnting questions (see section 4.1.4; Appendix L). They

29 Smagorinsky (1989, 1994b) provided a review of recent claims and cnticisms of protocol analysis
wi th particuiar referace to writing research.
30 The participaois were told to supply lheir own blank tapes in wse they needed more thm 2 tapes to
record the entirc process of writing. They were dso told that the same number of extra tapes as they provided
would be issued Iater. As it turned out, most snidents were able to complete their think-doud sessions within
2 hours, thus leaving 1 tape provided by me unused. A few needed io use the second tape but only part of it.
used their own tape recorders to produce think-aloud protocols. Al1 instructions and
questions were wri tten in English. The think-aloud instructions were adapted from
Ericsson and Simon ( 1993, pp. 376-377) and Flower ( 19W, pp. 73-73) and stressed two
points: (a) to work on the writing task as one normally would do, and (b) to Say out loud
and clearly everything that cornes into one's mind while writing. No tirne limitations were
irnposed; nor was a restriction placed on the use of dictionarîes or reference books. The
students were free to use any languages to report their thoughts while composing. In case
they needed to take a break, they were told to record the time and date of signing off and
signing back in io the tape-recording. They were also requested to fil1 in, right after
completing the writing task (i.e., having produced a final version to be submitted to the
instructor), the pst-wn ti ng questionnaire.
Al1 the participants wrote essays wi th a pen or pencil nther than using word-
processing although a few students mentioned, w hile thinking aloud or during interviews,
their desire to produce essays with word-processing. Audiotapes of the think-aloud
protocols, one copy of an essay produced during a think-aloud session, one copy of the
final version to be submitted to the instructor, and a filled-out pst-writing questionnaire
were retumed to me on or shortly after the day the assignment was due.
There were two main writing assignrnents during the course which were the objects
of my study, as rnentioned earlier. Prior to these tasks, a pilot trial of the think-aloud
procedure was carried out i n exactly the way explained above for the main tasb, using a
summary writing taslc assigned by Kei as homework in October 1994. The participants
were informed that it was a practice session and told of the importance of this trial. The
timeline of the assignrnents from which protocol data were generated is shown in Table 4-
1.
After the trial session, the participants were asked about their experiences in
thinking aloud. None of them had had the esperience of systematically thinking aloud
before. Seven students reported that they felt at ease with the task while the remaining 6
found i t difficul t and were unsure whether they were doing as expected. Nonetheless. al1
the verbal protocols from the trial session turned out more or less satisfactory. A11 the
participants said they liked the idea that 1 was not present during their think-aloud sessions.

Table 4-1
Tirnelirte of Writitzg Assigmneizts

Date of Assigmnent Given Due Date of A s s i p e u t I

Pilot !ziai October 20, 1994 October 27. 1994

r
Task 1 November 8, 1994 November 17,1954

1 Task U 1 February 2,1995 1 February 14, 1995 1

4.1.1.1. Reactiviîy. Some participants' accoun ts of thei r espenences in thi nking

aloud, obtained dunng the interviews, showed traces suggestive of reactivity to the think-
aloud method which might have resulted in these people's altering their cognitive
processes from what they might have used to c a n y out these tasks under normal (i.e.. non-
protocol) conditions (see Smagorinsky , 1989; Stratman & Hamp-Lyons. 1994). For
example, Eddie said:
[Thinking-aloud] took a lot more thought than 1 thought it would. For
instance, 1 didn't really ever think of the process of looking up words or
creating a sentence when 1 just wrote without talking. It was interesting,
though, to see my own thought processes at work. Actually the method was,
1 thought, very useful because it helped me clanfy what it was exactly that 1
was doing in my mind. On the other hand, this, 1 think, might affect the way
in which 1 write, and therefore these essays may be different from those 1
have written without speaking. (Eddie's interview, M u c h 3, 1995.)

Jasmine's comment was:

I found it hard to Say everything 1 think because 1 might be thinking
different things at different times. While writing down one phrase, 1 might
be thinking of a whole bunch of other things at the same time. So 1 can't
really put all the thoughts into words. Besides thinking goes so fast that I
can hardly get hold of it. (Jasmine's interview. March 9, 1995.)

Cathy also found it difficult to verbalize her thoughts while worhng on an essay and said,
"To tell the tmth, 1 had to think of what to Say into the tape first" (Cathy's interview,
March 7, 1995). Eliot gave an outright disapproval of the method:
1 don't like to think d o u d because speahng and thinking are two different
things. 1t's really painful to write and speak at the same time. When 1 hear
my own voice, 1 cannot think clearly any more. (Eliot's interview, March
16, 1995.)

Jack "wasn't sure exactly what was expected at first" but gradudly he "felt comfortable
with [the think-aloud method]" (Jack's interview, March 8, 1995) and invented what he
called "the stream-of-consciousness approach" inspired by the thi nk-aloud me thod, using i t
in the second of the two composition tasks. Using the stream-of-consciousness approach,
Jack first thought aloud into the tape in Japanese to compose, rewound the tape, and
transcribed what was recorded onto paper. Jane was the only participant who esperienced
no difficulty a t al1 in thinking aloud from the stan: "1 don't think it affected the way 1
wrote. 1 don't think it disturbed the process of m y writing. And 1 don't think it made me
nervous o r anything" (Jane's interview, March 16, 1995).
These data suggested that reactive effects did exist in the think-aloud protocols
collected and were rather idiosyncratic, as the prevailing assumption holds (cf. Stratman &

Hamp-Lyons, 1994). Despite the possible esistence of such effects, think-aloud protocol
andysis "offers a unique glimpse into the workings of the human rnind, and has a distinct
persuasiveness due to the storytelling chamcter of the data" (Smagorinsky, 1994a, p. siii). 1
regarded it as one of the essential means to validate my exploratory and descriptive study
of situated writing in JFL.

4.1.2. Trariscnptiot~of Think-Alouâ Protocoh

Once the audiotapes of think-aloud protocols were returned to me, I began

transcribing them into cornputer files in preparation for coding. 3 1 They involved four
languages: Japanese, English, Cantonese, and Mandarin. 1. a balanced bilingual of
Japanese and English with some knowledge of spoken Cantonese and Mandarin,
transcribed al1 the Japanese and English verbalizations exactly as they were spoken. For
the Cantonese and Mandarin data, a well-educated native speaker of Cantonese who speaks
and wntes fluent English and Mandarin and has some knowledge of Japanese translated
these segments into English as we listened to the tapes together. In my transcriptions
Japanese utterances appear in romanization, English verbalizations are typed in italics, and
English translations of Cantonese and Mandarin utterances are underlined. Other
conventions used in transcription (described in Table 4 2 ) are pnrnanly lrorn Curnming
( 1989,pp. 9 1-92).

Table 4-2
Tratwriptimi Conventiom

Mark Signif~cance
A period indicating the closure of an utterance with falling intonation, usually followed
by a pause.
? A question mark indicating a rising (questioning) intonation.

A comma indicating a shorter pause or abrupt shift in the flow of an utterance.

n n
Double quotation marks indicating that words, phrases, or sentences inside the marks
are treated as test to be wri tten down.
... Three dots indicating an unfinished utterance.

[..] Two dots inside square brackets iodicating inaudible sounds.

{ } Curiy brackets used for original Cantonese or Mandarin words in romanization where it
is more meaningful to do so.
( ) Parentheses used for the transcriber's cornments.

31 The process of transcription was faciliiated by consulting students' written notes and earlier drafts.
which I collecteci whenever ihey were made available to me.
4.1.3. Coding of Think-Abud Protocois

Following the practice used by Cumming ( 1989), 1 analyzed only verbal reports of
the participants' thinking about their writing. This excluded remarks directed at the
researcher. the reading aloud of text already written, and the sounding out of test to
emerge. Exceptions were words and phrases of text uttered in conjunction with comments
about wri ting. These selected protocols ivere segmented into co~nrnitiiicationuriits (c-units)
and numbered chronologically. Following Langer ( 1986a. l986b), I defined the c-uni t3' as
a separately identifiable remark about a thought or behavior. A c-unit "may have sevenl
sentence nodes as a consequence of having several sentences, several clauses or k i n g a
run-on or compound sentence" (Freed, 1978, p. 43) to express an idea. C-units are not
always grammatical sentences due to false starts and pauses common in speaking. C-unit
segmentation was applied across English (original English utterances and English
translations of Cantonese and Mandarin verbalizations) and Japanese p r o t ~ c o l s . ~ ~
As Smagorinsky (1994) aptly pointed out, a coding system applied to protocols
reflects the researcher's assumptions and agenda. 1 had my own assumption and agenda in
deriving a system to code my protocol data. With my prima. concern about writing-and-
l e m i n g relations, 1 wished to highlight the wnters' mental operations while engaging in
actual production of writlen test My assumption was that such thought processes would bc
suggestive of opportunities for learning. 1 consulted a number of protocol studies of
wnting and decided to adopt parts of Cumming's (1989) and Langer's (1986a, 1986b)

32 The c-unit has k e n used as a measure of the syntactic complexity of native English-speakng
cfiildren's (e-g.,Loban, 1963; Smith, 1978) and non-native EngIish speakers' (e-g.. Brock, 1986; Freed,
1978) oral utterances; it has aiso k e n used to distinguish the stretches of verbalimtions from one another in
overall think-aloud protocols of native English-spea)iing children and adolescents during theu production of
written text (Durst. 1989; Langer. 1986a, 1986b; mars hall. 1987; Newell, 1994). Operational defiaitions of
the c-unit may differ sligfitly from one study to another. This linguistic unit differs from T-unit (Hunt. 1965,
1970)or idea unit ( C M ,1980).
33 The validity of the c-unit as a cross-language measure has yet CO be demonstrated. Howeuer.
Harrington's (1986) successful application of T-unit--a linguistic unit comparable to the c-unit--anaiysis.
which has thus far been most widely used with EngIish and other European (such as French and German)
languages, to spoken Japanesr: tends some support to my application of c-units to Japanese and English
discourses in the same study.
schemes. These satisfied my interest in the what and hoiv of menial opentions while
composing. Thus, 1 first coded each c-uni t for aspects of the ivriter's atterztiorz (gist,
discourse organiwtion, language use-lesis, language use-syntas, language use-spelling, or
writing procedures; Cumming 1989, adapted from Scardamaiia & Pans, 1985), and then
for the riature o f ~ h ewrirer'r rnomloring behizviors (more or less refiective, corresponding
to what Langer called awnreness and use, following Baker & Brown, cited in Durst,
1989v p. 345). Categones are shown in Table 4-3 with esamples. English translations of
these sample transcripts are provided in cases where they tvere Japanese utterances.

Portions of representative protocols in a continuous text with coded catepries are shown
in Appendix M.

Table 4-3
Categories for Codirzg Th Nik-Aloud Protuc~Is

1. Aspects of the writer's attention

gist = remarks that focus on the substantive content of the emerging text.
How shouid 1 look at this ~robiemof ijime [bdlying] 2 (Colin, Task II)
ee kono sakubun wa mangabunka no pi ten warui ten soshite soreni tsuite no jibun
no iken O kaku koto dakedo [un in this composition 1 am supposed to writc about good and bad
points of uimic culture and my opinion on it.] (Jasmine, Task I)

discourse organization = remarks that focus o n the organization of the test at the
levels beyond the sentence, including concems about cohesive devices that make a
link between two or more sentences.
unto "konoyoona" unto "kono" unto dokokan rieiv paragraph ni shite ii no ka zenzen
wakannai na [um "as such" um "such"um I have no idea where I should k g i n a new pan,gaph.]
(Jane, Task II)
Okay [.. J spïir a parngrapk liere arid acid a feiv settletzces. (Martin, Task II)

- - - - --

34 Durst (1989) adapted and çxpanded one component of Langer's Analysis of Mcaning Construction
scheme (1986a, 1986b), focusing on two aspccts of the writer's monitoring behaviors--awarcness and use of
the writer's thinking processes.
01) language use-lexis = remarks that indicate word- and phrase-Ievel concems.
(Im" tokoro ga" [place] basho @lace] that's better. isn't it? " basho" (Cathy, Task 11)
Everythirig ehe I say everythirzg efse U ~ everythitig
L uh banji [everything] (consulting
dictionary) (Eddie, Task II)

(1s) language use-syntax = remarks that indicate concems about syntactic and
morphosyntactic rules (e.g., postpositions, sentence endings, conjugations) as well as
the formation of an entire sentence.
Yeah biit in this case sfloidd I rue the rrrn ukemikei [passive voice]? (Mary. Task i)
Okay 1 use ni ri] " zensekai ni [in rhe whole world] " (Cathy, Task 1)
(lm) language use-spelling35 = rernarks that indicate concems about orthognphic
conventions.

Hmm hoiv do you write omoshi roi [interesting]? (Martin, Task 1)

gaman gaman kanji de gaman tte kakoo [ i'm going to wri te 'gamao' [patience] in kanji.]
(Jane, Task II)

(p) writing procedures = remarks that locus on one's procedures in completing a

writing iask. This includes the ivri ter's consul ting dictionaries and course handouts.
Okay yoii shoiild nlwnys firiisli the sectiori iike tiiis ivith the best point yori fiave or
certaitily riot the weakest so (pause) (Jack, Tasli 1)
So well 1 read the earlier part. (Colin, Task II)

2. Nature of the writer's monitorin~

- behaviors

Note. The first of the two example statements provided under each of the above categones
is coded "more reflective" and the second esample " Iess reflective."

(m) more reflective = remarks that indicate esplicit on-line monitoring, Le., remarks in
which the w n ter says "1 think" or " 1 know", expresses an opinion, makes a judgment,
or indicates hislher uncenainty while writing.

(1) l e s reflective = rernarks that describe the writer's on-line behavior o r report what
happened and what has been decided in the writer's mind, not s o esplicitly reflecting
on hislher thinking processes o r the task.

35 Al1 the statements under this category concemed the spelling of kanji or katahria (phonetic scripts
used to represent foreign--mostiy Enplish and other European--1oan words in Japanese). except one siatement
that was about the expression of a number: "Um ah veah 1 shouldn't write two in Arabic nurnbers" (Cathy.
Task 1). For descriptions of the Japanese wri ting system. see Taylor and Taylor ( 1995).
 While examining the protocols to establish the above category system, 1 decided to

include another dimension, i.e., languages the wri ter used in making metacogni tive

remarlcs (Japanese, English, Cantonese, Mandarin, or any combinations of these). 1 wished

to see which language or languages these foreign language writers might use in the process
of generating the target language discourse. Intuitively 1 thought such inquiry could shed
some light on the role of the mother tongue in the process of thinking for target language
verbal production (cf. Ringbom, 1987). Analyses of this and a second dimension could

document the Izow of concurrent thinking about writing while a fmt dimension the whaf.

Once the coding system was established, I coded each segment for the three
dimensions of the system. Because c-units are. as mentioned earlier, segmentable
staternents for a thought or behavior distinct from another, one category was sufficient to
code separate segments in most cases; double or larger configurations of coding were very
few; for a second dimension, of course, either one of the two mutually excIusive categories
had to be used to code a segment. Also because some segments contained only minimal

information, the coding decisions of those segments had to be made by consulting

preceding and following segments as well as wri tten tests (see Ericsson & Simon, 1993,
chap. 6, for the use of contest in coding protocols).
I checked inter- and intra-coder agreements to make sure the reliability of my
coding judgments was sufficiently high. About 10 segments (i.e.. c-units) selected
randomly from each of 16 different protocols--amounting to 153 units. about 10% of the
total codings--were coded by Kei who volunteered to be a second coder. This took place
one year after she had taught the course that 1 had observed for the present research. Her
continued interest in rny work and her cornpetence in Japanese and English served the
purpose well. On the assigned day she was first given an explanation of the coding system
then practiced on 10 separately selected coding units before the actual coding. Only the
first and second dimensions were coded for reliability checking as the third dimension (i.e.,
the language used for metacognitive remarks) was stnightfonvard and devoid of concem
about bias. Agreement was 83.7% for the first dimension and 81.7% for the second

dimension. Intra-coder agreement, checked on 10%of the total segments 1 weeks after the
initial codings, was 97.5% [or both dimensions.

To supplement the think-aloud protocol data, the participants were asked to fil1 out
a post-cvn ting questionnaire (Appendix L) right after completing each task. The
questionnaire asked about (a) the level of satisfaction the writer felt with his or her own
essay just produced, (b) the level of difficu1t the writer esperienced in genenting ideas,
and (c) the level of difficulty the writer esperienced in expressing ideas in Japanese. It also
provided space for comments o n the writing task just completed. The participants'

comments on the questionnaire were informative, esplaining the strategies they used in
writing essays, elaborating on specific points of difficulty, or reasoning about laborious
and less labonous parts of the writing process. I utilized these pieces of individual,

anecdotal information in my attempt to undentand better the overail system of learning

and teaching about writing activities in this JFL classroom.

4.2. Atzsweririg the Research Questiuris

Whnt IBrgitistic ntid cogrritive belraviors are invoked by JFL ivrititig

activities? How are tliey related ro stridenfs' characteristics and

itisirtictiorial cotztext variables?
 4.2.1. Oventiew of Tliink-Aloud Protocol Dala

Due to poor conditions of 5 participants' recordings on one or both tasks, only the
two sets of complete protocols obtained from 8 participants (Jasmine. Jane, Jack. Cathy,
Colin. Mary, Martin, Eddie) were analyzed. i tallied al1 c-units coded under respective
categories for each protocol then converted these to percentages. The tai lied figures are
treated as cumulative percentages for the protocols containing uni& with multiple codings
(Le.. more than one category coded). An overview of the resul ts is given in Table 4-4. In

this summary four fundamental categories are highlighted that represent four mutually
exclusive aspects of writing people may focus their attention o n while composing: gist,
discourse organization, language use, and writing procedures. Thus, the three sub-

categories of language use--lesis, syntas, and spelling-are confiated.

T a l e 4-4
Slimmaq of Thirik-AloridProtocol Data

Participant discourse writing "reflective" cornments in

gist (%) organization (%) language use procedures {%) cornments (%) Ll (%>
L
Task1:TaskII Task1:TaskII TaskI:TaskII TaskI:TaskII Task1:TaskII Task1:TaskII
Jasmine 41 : 23 8 : 1 49 : 54 6 : 2 6 . W : 3 9 O : O
Jane 5 l : E 3 : 8 30 : 61 14 : 10 72 : 72 1 : 6
Jack 33 : 31 3 : 7 23 : 17 43 : 45 -1L.l : 31 98 : 100
Cathy 39 : 52 0 : 7 41 : 2.8 19 : 12 SI : 56 98 : 100
Colin 76 : 66 1 : 5 4 : 17 10 . 14 67 : 35 n : 88
Mary 35 : 40 2 : 3 58:43 6 : 10 78 : 55 9 3 : 90
Martin 33 : 32 6 : 4 34:34 27 : 31 72 : 54 99 : 98
Fddie 2 - 74 O6 6 65 - 63 12 : 8 79 : 59 74 : 62

In this summary 1 took note of the following trends:

A large proportion of attention focused on gist, language use. or both in contrast wi th

attention paid to discourse organization or wri ti ng procedures. Two cases (Jack and
Martin) went against this general pattern, hoivever. The largest proportion of Jack's on-

line cornments focused on the aspect of rvriting procedures, followed by gist and
language use. Martin's attention more or less evenly divided between gist, language
 use. and writing procedures. Another unique case was Colin, who focused most of his
attention on gist and paid much less attention to the other three aspects. On-line verbal
comments about organization were very few across the cases (cf-. Cumming, 1989;
Sasaki & Hirose, 1996).

More than a half of the on-line comments in al1 but a few cases (Jasmine, Task II; Jack,
Tasks 1 and II; Colin, Task II) were reflective in nature, suggesting, on the assumption
that learning is a manifestation of thought (e.g., Smith, 1990), a great number of
opportunities for these students' potential learning while engaging in JFL composition
writing. In generai, Task 1 induced more comments o f a refiective nature than Task II.

Al! the students, escept Jasmine and Jane, relied heavily on their Lls in verbalizing
their thoughts; Jasmine and Jane, who spoke native-like Japanese, produced their
protocols primanl y in Japane~e.3~

To make better sense of trends in the protocol data, I created three broader
categor-ies of ideotiorznl thirzkitlg (combining the aspects o f gist and discourse
organization), farrgimge-relnred thirrkiizg (cornbi ning the three sub-categories of language
use--lelis, syntas. and spelling), and procediirnf thi~ikiiig(w hich is the aspect of writing
procedures renamed for consistency with the other categories' names). These new
categories of attentional foci were examined according to the nature of the writerts
monitoring behaviors (more or less reflective). The iallied figures of these combined
categories appear in Table 4-5 (see Appendis N Cor the full tallies of the six original
categories of attentional foci combined with the two categories for monitoring behaviors).

36 Thcse trends accord with the percentagcs of L1 (Japanese) use in the production of L2 (English)
written discourse reported by highcr and iower proficient L2 learners in Kobayashi and Rinnert's (1994. p.
239) study. Note that Jasmine considered Japanesc as her oral L1 but English as her writtcn L1.whereas Jane
chose Engiish for her orai and written L1.
 Table 4-5
Aspects of Attentional Foci and Re/ectiverzess (nzirzk-Moud Protucoi Data)

Participant Task ideational thinlring language-related

p r d d
(w
more ref*: Iess ref
thinking (%)
thinking (%)
more ref : less ref
more ref : less ref
Trisk 1
,.*.-.--.....-P.
40.7 : 8.2 42.5 : 2.9 1.7 : 4.1
.*.--..-................*.*.*..-.*..*..-.
..*.**..----..-.......-.......-......*-*....*--.-........-..-...**,
....-..m.-
Task II 9.1 : 15.6 26.0 : 28.6 3.9 : 22.1
Jarme ,........-.....-... 44.8 : 11.3 253 : 4.5
Task 1 .............................................................................................. 2 3 : 11.5 ri..rr.........-....-............

Tasli II 2 1 . 6 : 7.5 46.7 : 13.3 3 3 : 7.5

Jack Task 1 15.8
..-.-.....-.........*.-..--...........*..-....*........
-4
: 19.9 15.7
.......... : 6.9
.........-....*......1I-..*............-..4.....-*.....-....-...-...-~.
13.7 : 30.1
Task II 10.9 : 27.1 9 3 : 7.8 10.1 : 34.9
cath~ Task 1 ...........17.7
I.._.-p
_.. ................:........21.9
.... 31.2 : 8.4 ............................................
...................................... 4.2 : 16.7
Task II 29.6 : 29.6 24.6 : 3.9 1.7 : 10.6
Colin ...Task 65.4 25.0
.............1.... . .......*.-......-...-.-......*.*.....*.
: ........-
S
.........S
2.9
.-........
: 1.5 2.2 : 7.3
................-.......---....*~-..~..--~
Task II 26.6 : 45.6 6.8 : 10.9 2.0 : 12.2
Mary ...-......-.-.-p.
Task 1 ......................................... 49.0 : 9.8
27.7 : 8.9 ...................................................................................... 1.0 : 5.0
Task II 20.5 : 22.3 33.8 : 13.8 0.6 : 9.0
Martin Task 1
..................... 313 : 7.8
*.-.artni artni
283 : 6.0 133 : 13.3
.................................................................................................
Task II 20.4 : 14.9 24.0 : 9.3 11.1 : 20.4
Eddie Task I 19.1
................................... : 4.1
....-.-..*....*.......... 573 : 8.1 2 . 3 : 9.2
........................................................................................
Task II 15.6 : 14.2 -42.5 : 17.9 3.0 : 6.7

* ref = reflective

This set of composite data shows:

Generally, reflective thinking occuried while the wnter thought about ideas, language.
or both. For Jasmine, Jane, and Colin in Task 1, renective thinking about ideas
occupied a large proportion of their writing processes.

Language was distinctly an object of reflective thinking for most of the writers. Even
the highl y proficient leamers of Japanese, Jasmine and Jane. focused a large proportion
of their attention on language use. But their linguistic concems were at a somewhat
higher levei.37 Jack and Colin rarely verbaiized their mental processes of Japanese
language producti~n?~

37 Jasmine and Jane wrote about their linguistic concerns in the pst-writing questionnaires; they both
indicated their attempts to use a variety of predicative expressions to avoid the monotonous tone of written
text and to use "harder words" instead of simple and basic words. Their thînk-aloud protocols also attested to
this.
38 The very few verbalizations about languiigc issues these intermediate-proficient students produced
made me wonder why. Colin's comments written in the Task II post-writing questionnaire provided some
dues. He wrote. "Although it took only a short period of tirne to finish writing the assignment, important
 Among the three sub-categories of language use, lexis was the aspect to which mos t of
the writers devoted most of their reflective thinking39 (see Appendix N). Syntax was

less of a concern to them generally (but Mary was apparently much concemed about
syntax4). They cared little about mechanical matters in general. Jasmine was the on!y
one who showed excessive concems over spelling or the wnting of karzji.

Individual writers showed similar patterns for both tasks in terms of the proportion of
attention paid to respective aspects of writing and the proportion of more to less
reflective thinking under each category. Exceptions were Jasmine, whose intensive
ideational and language-related thinking of a refiective nature in Task 1 reduced
dramatically in Taslc II, and Jane, whose ideational and language-related thinking of a
reflective nature in Task 1 and Task 11 were just the opposite.

Table 4 6 displays the resuits of the individuai participants' responses to the post-
w n ting questionnaires, together wi th their JFL oral proficiency levels and their L 1 and
Japanese essay scores. This matris table left me with the impression that more students
found Task II (which the students interpreted as "write whatever you want") easier than
Task 1 (in which the students were aware that the statement of one's opinions was
required) both in terrns of idea generation and language use. They produced more self-
-

ideas were generated during a long pcriod of time. Extensive preparation before the writing process was
required." A similar retrospection was given by Jack during bis interview with me. Mthough they wcre
referring to the processes of idea generation at a pre-writing stage (which were not tape-recorded) and the
idea generation might have been done in their Lls. they could have been thinking, consciously or
subconsciously. about Japanese expressions lhat would fit the generated ideas during the "iong period of
extensive preparation." Because of such preparation, Ianguagc issues may not have corne to the fore of their
consciousness for verbalizations during the actual taping.
39 Comor-Linton and Haichour's (1997), Cumming's (1990). and Shibata's (19%) stüdics of L2
writing processes showed vocabulary howledge to be a sie4cant factor in writing.
JO This observation is in line with what Mary wrote in her pst-writing questionnaires: "1 found this
assignment quite difficult ... 1 used many [Japanese lanmwge]dictionaria ruid , o r m a r handbooks" (Task 1);
and "There were some difficuit things about writing this composition. I think they are grammar and
vocabuiaxy" Cask II).
satisfying essays in Task II (cf. Kei's better impressions on the Task II essays in
cornparison to the Task 1 essays; see section 2-42).Neither JFL proficiency nor L1 writing
skill correlated with the level of satisfaction the students had with their own essays or the
level of difficulty they experienced in idea generation and language use.

Table 4-6

-
Post-Writing Qtiestiotrnaire Respomes, JFL Proficiency. and LI iVrititzg Expertise

Partici-
- TASKII
Difficul ty
pant
tion with in

30 1 satisfied
not so
difficult satisfied di fficult
-
,
language
use
not so
difficult

difficul t difficult difficult

advanced- not so n d so not so
Judith stisfied difficul[ di fficult
not so
-JOY superior
intermedi 1
28
4
missing satisfied
not SQ
difficult
flot so
Jack ate-high 1 17 difficuit satisfied dirficuit difficult
intermedi not so
-
Cathy ate-low
intemedi
16
I not so
difficult
not ~o
Colin ate-mid 20 difficuit
intermedi l notso not so
ate-mid 24 di fficult difficult ciifficuit
intermedi not so
ate-rnid 26 difficult difficuk satisfied di fficult
intermedi not so
noLw
satislied difficult
ate-mid 15
not so
Martin advanced 93 difficult

-Eliot
advanced-
high 1 22
not so not so
difficult

Eddie 34 1 satisfied
- not so
di fficult

Table 4-7 shows the results of the individual participants' responses to the p s t -
writing questionnaire, together with their perceptions of JFL writing and learning-the
information elici ted in the students' interviews. I noted the following general trends in this
sumrnary:
 The students who mentioned some leaming of content as a result of JFL wri ting also
said they experienced difficulty genenting ideas in both or one of the tasks. On the
other hand, the students who mentioned only the learning of language through JFL
writing round idea generation either easy or not so difficult.

Al1 the students (except Eliot) perceived language use as a problem in their J E
writing; language processing was either difficult or not so difficult. but never easy for

Table 4-7
Post- Writing Questionnaire Responses arzd Perceptions of JFL Writing arzd karning
-
*
Partici- satisfaction

Task 1
-
Task II
very sarisfied
satisfied
1 difficult
difficult
not so difficult
not so difficult
rneaning-
making
problem area in area of leamin

language
thm JFLwritin

language &
content
Jase
Task 1 satisfied difficult difficult language language language &
Task II satisfied not so difficult difficult exercise content
Judith
Task 1 satisfied difficult diffïcult meani ng- language Ianguage
Task II not s o satisfied not so difficul t not so di fficutt making
JOY languagt: &
Task 1 çatisfied not so difficult difficult meaning- language
Task II not so satisfied not so difficult not so difficult content
making
Jack
Task 1 not so satisfied not so difficult difficult language language language
Task 11 satisfied not so difficult not so difficult , exercise
Cathy
Task l not so satisfied difficult di fficuit meaning- language & "nothing in
Task iI satisfied not so difficult not so difficult making orgmization particular"
Colin
Task 1 not so satisfied not so difficult difficult m-ng- language language
Task II not s o satjsfied not s o difficult not so difficult makinp
Charles
Task 1 not so satisfied difficult difficult Ianguage language content
Task (1 satisfied not so difficult not so difficult exercise
MW
Task 1 not so satisfied not so difficult difficult language language language
Task II satisfied not so difficult difficult esercise
Mike
Task I not so satisfied difficult language language language
crisy
Task II satisfied not so difficult not so difficult axercise
Martin
Task i satisfïed W not so difficult language Imgurtge language
Task II satisfied not so difficult not so difficult exercise
Eliot I

Task 1 satisfied not so difficult language organization la"guage

easY
Task II satisfied not so difficult not so difficult exercise
Eddie
language &
-
Task 1
Task II
satisfied
satisfied
not so difficult
difficult
difficult
not so difficult
language
aercise
language
çontent
 Chapter 5: Written Products

According to a systemic view of classroom leaming and teaching, outcornes are

determined by leamers' activities in the processes of carrying out given tasks which
interact with student and teaching presage factors (Biggs, 1991). 1 collected copies of the
final written products submitted to Kei, the course instmctor. to try to find links between
process and product factors suggestive of potential learning. However, 1 soon found this
task far too comples and fuzzy to be carried out thoroughly in a single study. 1 thus
focused on two types of descriptive analyses. First, I looked at scores and cornrnents given
to individual essays by Kei and two independent JFL teacher iaters (who were the
Japanese raters described in section 3.3.3.3). Second. 1 evaluated the quali ties of thought
manifested in the written products. applying the SOLO (Structure of the Observed
Learning Outcorne) tasonomy (Biggs & Collis, 1982).

5.1.1. Assessînenl of Japanese Wriiten Texts

Kei and the two raters used multiple trait scoring procedures that evaluate
linguistic, rhetoncal, and content features of a piece of writing. What do these scores mean
in my research contest? Certainly they do not provide surnmative assessments of the
students' leaming as a result of writing, but they do provide sorne indications of the
students' current abilities to compose essays in Japanese. Loohng at the component scores
can help to identify, albeit only approximately, where the studenis' strengths and
weaknesses lie, to relate them to the aspects of writing the writers paid attention to during
the process of producing an essay, and to suggest potential opportunities for ittcidentol
learning (in contrast to interttional learning that is active1y desired and controlled by
leaners; see Bereiter & Scardamalia, 1989) attributable to writing experiences that
required effortfui processing of ideas and language.
5.1.1.1. Kei. In search of an instrument to assess Japanese essays wri tten by her

students, Kei wanted one that (a) ivould be available for immediate use (i.e., with no or
Ii ttle modification required), (b) would be congruent wi th her instructional goals and her
beliefs about how Japanese writing skills should develop, and (c) would provide clear and
concise feedback to focus on the leaming efforts of students. She decided on an adaptation
of the ESL Composition Profile developed by Jacobs et d.( 1981), which 1 had translated
into Japanese for my master's thesis on JFL writing (So, 1989).41 Kei considered this
scoring instrument to satisfy her requirements and used it cvithout modification. She noted
that the weights assigned to the content (30%)and organization (70%) components are
relatively large," but she observed "that suits the purpose of my coursen (Interview notes,
February 10, 1995). She liked particularly the fact that the band descriptors and al1 other
information in the scoring report were given in Japanese, providing the students exposure
to the Japanese language in a relevant contest.
A s this description implies, Kei approached the assessrnent of essays [rom the
standpoint of a person involved in both teaching and assessing. She mentioned in one of
Our interview sessions that essay assessrnent was primanly meant to be a n encouragement
for her students. i.e., to help her students gain confidence by highlighting the good traits of
their writing (Interview notes, February 10, 1995). On the same occasion she also
expressed the difficulty she expenenced with choosing the "right" descriptors and

41 Hirose and Sasaki (1994, p. 228) also translated Jacobs ct d.'s profile into Japanese (O assess
Japanese compositions written by native speakers of Japanesc participating in their study on Engiish-as-a-
foreign-lmguagewriting.
42 The other components consist of vocabulary. laquage use. and rneclranzcs, which are weighted
20%. 25%. and 5% respectively.
assigning scores to respective essays because of her familiarity wi th the essay wnters, a
kind of halo effect. It may be said that Kei's scores on her students' compositions make
sense pedagogically, but they do not necessarily carry psychometric values. To understand
the mechanism of the classroom system in relation to writing tasks, which is an
ovenrching purpose of my research, it is important to look at Kei's scores and comments.
on the one hand; on the other hand, 1 balanced this perspective by also weighing the scores
obtained from the two independent JFL teacher raters who had had no contact with the
essay writers or their instructor (Kei). The latter scores are thus free from halo effects and
thus rnay be considered "more objective."
5.1.1.2. Independent Rates. The two independent raters were carefully selected to

satisfy the following criteria: (a) being a female native-speaker of Japanese, (b) k i n g
experienced in teaching J E to university students in an English-speaking country. and (c)
king expenenced in assessing JFL writing for i ts communicative effectiveness. In other

words, I searched for i ndependent raters w ho shared Kei's identity, sociocultural context,
and values about language teaching." 1 wished to obtain from these nters essay scores that
were free from halo effects yet "intemally valid." The two volunteer raters, fernale native
speakers of Japanese, were very similar to Kei in terms of educational and professional
experiences as well as their current teaching contests and orientations. I had cvorked wi th
both of thern in tertiary-level JFL education at an institution where 1 was previously
employed in Singapore and knew them cvell. A t the time of data collection for this study
they were teaching JFL at a university in Australia. Al1 correspondences with them were
done through e-mail and regular mail. Since their scores were also to be used as measures
of the participants' Japanese writing ski11 (see section 3 - 7 2 ) . it was considered more
meaningful to apply the same rating scale used for L1 writing assessment, i.e., Hamp-

43 How different readers in different contexts (e-g., teachers of Japanese as an LI. non-language
teachers, non-native speaker teachers of JFL) might read and respond to the same essays is an important
question in order to create a universal definition of "good JFL writing." A growing body of such research in
ESL contents ( e g , Brown, 1991; Connor-Linton, 1995) attests to the usefulness of such inquiry. but it goes
beyond the scope of my current research.
Lyons's 9-point scale (see section 3.3.2.1). I mailed the raters a packet containing the
essays of Tasks 1 and II,U an instruction sheet, the rating scale, and instructions for n t i n g
procedures (Appendixes G and H), allowing them a month to complete the assessments.
Although they worked in the same institution, the two raters assessed the essays
independently without any discussion of the assessment task between them. Both raters
were also assiduous in providing comments in a space given on the score report form.
These comments were mostly idiosyncratic responses and reactions to unusual or
intangible qualities of the writing that they had noticed but thought were not encornpassed
by the traits and their descriptors of the assessment instrument (Harnp-Lyons, 19% b).

5.1.2. Kei's und I~ulependentRaiers' Essay Scores

To summarize Kei's and the two independent raters' scores on the studenis'
Japanese compositions. 1 conflated Harnp-Lyons's scale's original categories of linguistic
accuracy and linguistic appropriacy into one category called larlgrîage use. and called the
original category of communicative quali ty simply confenl. Likewise, I confiated Jacobs et
al .'s scale's original categories of vocabulary, language use, and mechanics into one
category cailed Iungzîage me. These changes enabled me to focus on the three essential
aspects of a composition (i.e., content, organiuiion. and language use) and helped me in
making meaning of Kei's and the two raten' scores. 1 avenged the composite scores or
linguistic accuracy and linguistic appropriacy on Hamp-Lyons's scale for the single
category of language use. 1 adjusted Kei's scores to make them numerïcally comparable to
the scores on Hamp-Lyons's scale by multiplying the content score by 9130. the

44 To eliminatc the possibility that the rater wouid cany over her judgments on the quality of one essay
to that of another essay written by the same author, different sets of names were used for the Task 1 and Task
II essays COdisguise matched pairs. Since ail compositions wcre typed beforc submitting them to the raters,
the nters had few clues to identiry which cssays of Task 1 and Task II wcre written by the same author.
organization score by 9/70, and the language use score by 9/50.45
The table given in
Appendis O displays the modified scores for the three aspects of the students'
compositions that Kei assigned and the two other nters also assigned for the two taslrs, as
well as from the thi nk-aloud protocol data the percentages of ûtten tion (where these figures
are available) each student paid to these aspects of wri ting during the processes of wri ting
the compositions. The table also shows Kei's and the nters' comments, originally wntten
in Japanese but translated into English"
My impressions of these ratings ivere as follorvs:

Generally, Kei's scores were higher than the independent nters'. This phenornenon was
particularly acute with the students of less-than-near-native Japanese proficiency (Le.,
Jack, Cathy, Colin, Charles, Mary, M i h . Martin, Eliot, and Eddie). For esample, Jack
received 23.4 points [rom Kei for his Task II essay which was nted 10.5 and 11.5 by
the raters. Kei awarded Cathy 23.2 for her Task I essay that scored 14 and 1 1 according
to the nters' judgments. Kei's judgment of 75.8 for Colin's Task I essay received 13.5
and 7 from the nters. Eliot's Task I essay received 34.4 liom Kei, compared to 17 and

12 from the raters. On the other hand, discrepancies were very small for the near-native

proficient students, Jasmine. Jane, Judith, and Joy. There ivere 1'5 cases in al1 where
Kei rated the compositions lower than the independent raters: 6 cases with the less-
than-near-native group, amounting to 17%. and 6 with the near-native group,
amoun ting to 38%" A mong the 6 cases wi th the former (i .es,less-than-near-native)

- -

45 1 derived these manipulation formulas from the fact that the maximum points are 30 for content, 20
for organization, and 50 for language use (the sum of 20 [vocabulary),25 [language use], and 5 [mechanics])
on Jacob et aL's s d e , and 9 for each of these three components on Hamp-Lyons's scale. I do not believe that
these mathematical manipulations alone enable valid cornparison between scores on Jacobs et d.'s scale and
Hamp-Lyons's scale. 1 manipulated the scores, however. for the sake of convenience, so these adjusted
scores should be taken as appro.ximations when compared to the scores based on Hamp-Lyons's scale. As a
result of these manipulations. the total masirnum score is 27 points, buitt from its component scores in
newly-created categories of content, organization. and language use, which are weighted a maumum of 9
points each equally.
46 Long comments (most of which ivere Kei's) arc abridged for the sake of brcvity.
47 6 + (2 cornparisons Wei vs. Rater 1, Kei vs. Rater 21 x 2 t a s h x 9 students-Jack, Cathy. Coiin.
Charles, Mary,Mike. Martin, Eliot. and Eddie) x 100 = 16.66.
6 + (2 comparisons [-ei vs. Rater 1, Kei vs. Rater 23 x 2 tasks x 4 students-Jasmine. Jane, Judith,
 group, 3 cases were where the essays had previously been corrected by native speakers,
mostly in respect to the use of language.

There was a consistent pattern for the same reader's ntings of each essay's three

components. That is, the same reader tended to rate the three aspects (content,
organization, and language use) of a piece of writing similarly. Only 19 (out of 2349
ases showed a difference of 2 or more points between the three cornponent scores; the
rest showed a difference of less than ? points?

The same readers tended to rate each writer's two essays similarly. Note that the two
nters were unaware which essays were written by the same author and that Kei nted
the two sets of writings at an interval of 3 months and she did not refer to the scores of
the first essays while scoring the second ones. Only 7 (out of 3 9 M ) cases went against
this general pattern, showing a difference of 5 or more points be tween the scores of the
essays for Task 1 and Task II written by the same author. Four (Rater 2's ntings of
Mary's, Martin's, and Eliot's essays; Rater 1's ratings of Martin's essays) of these 7
cases were where the writers had sought native speakers' advice on one of the two
tash. 51

The readers commented about the writing generally or o n each aspect of the writing
(content, organization, or language use). Genenl comments, posi t h e and negative,

were made for 38% of the total number of essays; comrnents about content for 35% of
the total; comments about organization for 13%;and comments about language use for

and Joy) x 100 = 37.5.

3 cornparisons (content vs. orgimization. content vs. language use, organization vs. Ianguage use) x
2 task s 3 readers x 13 students = 234.
49 If we look at this phenornenon by cornparhg the uear-native and the less-than-na-native group, we
find only 2 out of the 19 more-than-2-point-differencecases in the near-native group and 12 in the las-than-
near-native group. In fact the differences observed in al1 these 19 cases arc just 2 or 3 points. except for 2
cases. The 2 exceptionai cases occurred in Eliot's Task 1 essay. which was nted 2 points on content. 2 points
on organization, and 8 points on language use by Rater 2. Rater 2 cornmented on the cssay in question as "1
see no trace of planning in tbis essay. 1 cannot tell what the wnter wants to say."
50 1 cornparison (Task 1 score vs. Task II score) x 3 readers x 13 students = 39.
51 The 3 other cases happcned to be Rater 2's assessrnents of Jack's, Colin's, and Mike's two essays.
which scorcd 19 c a s k 1) vs. 11.5 (Task II), 7 vs. 25.5, and 7.5 vs. 13.5 respectively.
 46s.52 General statements such as "well written" and " r h e author] must be a near

native" were the most cornmon type of comments given to the students of the near-
native group (Jasmine, Jane. Judith, and Joy) (6789). and al1 of these were positive.
On the other hand. the students of the less-than-near-native group received comments
on language use most frequently (57%%)and such comments included both positive
and negative rernarks, such as: "good at connecting one sentence to another." "The
spoken and written foms of language are mised," "The writer seems to possess
sufficient vocabulary but laclis grammatical cornpetence." and "There are grammatical
mistakes but they do not bother me too much."

5.1.3. Assess~nentof SCrr~cfiualComp&xLty irr Japanese WnLten Texts

1 applied the SOLO taxonomy (Biggs & Collis, 1982; see also Biggs & Moore,

1993) to assess the cognitive structural cornplexity of the students' written texts. 1 wanted

to see what might distinguish mature [rom immature or deep from shallow thinking
manifested in the written products.55. The SOLO tasonomy kvas developed based on the
assumption thai the quality of students' thoughts would closely relate to the quality of their
learning. It is a classification system that describes the quality of learning outcornes
according to five levels: presrrricrriral, ariislrr<ctiual,rnrcltisiritcrriral. rela~io~rai,
and
exfended nbstraci. It is a conceptual fnmework built upon the empirical study of students'

52 Differenc types of comments could be made by a reader reading one piece of writing. For instance,
Kei's comments on Colin's Task I essay reads. "Very well organized (organization). lnsightful (content).
Enjoyable read (general). Rich vocabulary, though some mistakes in their usage (language use)." Such
multiple configurations of comment types were not very common. though. The great majority of comments
had single or double configurations. The percentages were derived from dividing the frequency of each
comment type by the total number of essays on which Kei. Rater 1. and Rater 2 commented (13 students x 2
tasks x 3 readers = 78).
53 16 general comments (given to the near-native group) i (3 readcrs x 2 tasks x 4 near-native
students) x 100 = 66.66.
54 3 1 comments on languagc use (givcn to the less-than-near-native group) + (3 readers x 2 mks x 9
les-thiui-na-native studcnts) n 100 = 57.j0.
55 In relation to the comection between the thinking process and the written produci, Smith (1 994, p.
34) contended, 'We cannot observe ourselves thinking, but we can observe the products of thought. And one
responses in sevenl subject areas i ncluding history, mathematics, Englis h. and modem
Ianguages. In applying this scheme to my written product data, 1 looked to the illustration
of SOLO analysis of English essays written by Austraiian high school students in Grades 7
through 12. Biggs and Collis defined the rive SOLO levels (applied to the analysis of
English essays) in ternis of characteristics such as the defining and refining of meaning
through word choice, appropriate syntax, and so on. 1 list the main features of each level in
Table 5- 1.

Table 5-1
Descrip~iorrof SOLO Leve fs

N o consistent component. Incoherent. The words appear as they stnke

the wnter. giving a series of unrelated and fleeting impressions. Little
considention for the reader.
Unistriictriral. Linear and spread d o n g a single line with a beginning, middle, and
end. Highly concrete, particularistic, and sirnplistic.
Mziltistrzictriral. Appropriate use of description but conventional. The ovenll effect is
flat and unimaginative.

Rela t i o mf . A novel, cdculated effect achieved. Sufficient information is present

for readability (considention for the reader). Limited to the particular
contest of the wri tef s choosing.
Recognition of different layers of rneaning that estend beyond the
chosen contest. Innovative use of rhetoric, permitting the
metaphoncal and layered access to meaning.

Note. Adapted from Evaluating the Quality of learning: The SOLO Taronomy (Siructure of rhe Observed
Learning Ourcorne) (pp. 109- 1 11). by J. B. Biggs and K. F. Collis. 1982, New York: Academic Press.

The purpose of this assessrnent was to measure the depth of ideational thinking
during the production of JFL written discourse. so I tried as much as possible to focus on
the "deep structure [italics in the original] of language" (Smith, 1994, p. 49), Le.. its

of the most powerfd tools for doing so is writirrg [italics in the original]."
meaning. 1 tried as much as possible to read the students' compositions as a semantic
representation, ignoring the language use problems typical of JFL leamers. This attempt,
however, created a paradox because meaning can be conveyed only through the "suî$zce

structure [italics in the original] of language" (Smith, 1994, p. 49) or its physical,

observable properties. Deficiency in the language (Japanese in this case) skill wa.5 sure to
affect the meaning conveyed, and thus i t s readabiIity. The reverse is true, too. The accurate
and fluent use of language might have boosted the readabili ty of an essay, making its

content look more sophisticated or comples than it actually was. Although I was well
aware of this difficulty, 1 still camed out an assessment, on an exploratory bais, of the
qualities of ideational thinking represented and organized in the students' two
compositions. 1 hoped the SOLO analysis, attempting to evaluate the quality of thinking
behind the observable product, would be able to provide a different perspective on the
students' compositions than Jacobs et al.3 or HampLyons's essay nting scales had,
offering more profound insights into the leaniing of their JFL authors.
Once again Kei vohnteered to collaborate with me as a second rater in this
assessment task. 1 met with Kei on an appointed day, which was already one and a hall
years after the completion of my data collection. We first went over an esample of SOLO
analysis on the high school English writing given by Biggs and Coilis (1983, pp. 108-131).
We spent about 40 minutes discussing the features of the five SOLO levels in light of the
sample essays given by Biggs and Collis and typical JFL students' writings we had often
corne across in our teaching. We found that the level descriptors designed for the anal ysis
of English expressive writings would be applicable to the essays 1 had gathered in my
study. Cognizant of my intention for the SOLO analysis. Kei agreed to focus on meaning,
ignoring grammatical or lexical mis takes inasmuch as they would not bnng the resul ting
text to the point of unintelligibility. To facilitate this process, we decided that we should
not spend more than 2 minutes reading a piece of writing to detemine its SOLO level. To
minimize halo effects, neither the authorsf real names nor pseudonyms were used to
idcntify the essays; al1 the essays were simply numbered, for the purpose o l identification,
1 to 36 with the 13 nndomly ordered essays of Task II bearing the numbers 1 to 13 and the

13 nndomly ordered essays of Task 1 bearing the numbers 14 to 36. We also anticipated
that there might be essays showing features belonging to more than one level. In such
cases, we agreed we should indicate al1 possible levels.
Once boih of us felt cornfortable about the nting criteria and procedures, we began
the assessments from Essay 1 to Essay 16 in order, at the sarne time, but independently.
Kei completed the assessments in 35 minutes and 1 in 45 minutes. There were no more

than two levels assigned to one essay. We called the essays belonging to two levels
"transitional," following Biggs and Collis's system. Perfect (8 cases) and partial56 ( 10
cases) matches were put together to calculate an intemter agreement of 69%. Considering

the absence of formal training and Our limited time for discussion of the criteria. this level
of agreement suggests our judgrnents were fairly reliable. For the 8 cases of complete

mismatches, Kei and I re-read the essays in question together and discussed their
appropriate levels. This session took place on the same day, dter a break of ?. and a haff
hours during which we did not touch on the subject; the session lasted 35 minutes. The re-
nting process went smoothly and we came to agreed levels easily, requiring little debate.

Most of the time ei ther one of us changed the chosen level to the level selected by the other
immediately after re-reading the essays in question, typically accornpanied by such
remarb as "How did 1 corne to rate this so high (or low)?" or "1 want to change rny nting."

Two cases required more deiailed discussion and re-assessrnent and we seitled on a mid-
level between the levels we had chosen on ginally.

--

56 Partial matches happened where both or either one of us assigned two levels to one essay. In these
cases ody the level w e both chose was taken to Iabcl a given essay.
 5.2. Answering the Researclt Questions

Wiat textrial qrtalities of JFL writirig do strtderru manijèst irr tenns of

langmge rise, corrtent, and orgnnicatiorr? How is rhe cornplexify of
cognitive structure rnatrifated irr their writte~rtexts? How are filese
efernerrts relrted to studerrts' chartrcteristics, iristrrtctiorral corrtext
variables. arrd irz-process belrnviors?

52.1. Relorions between Essay Scores and Other Factors

Table 5-2 summarizes the trends in the essay scores and cornmentary data,
demarcated by a dichotomy between the near-native and less-than-near-native groups. In
al1 these comparisons, the near-native students' performances (as judged by Kei and the
two raters) were superior and stable relative to the performances of the Iess-than-near-
native leamers. This finding may be explained by the presence or absence of the "foreign
language effect," which "refers to a temporary decline in the thinking abili ty of people who
are using a foreign language in cvhich they are less proficient than in their own native
language" (Takano & Noda, 1995, p. 658). Takano and Noda explained the cognitive
mechanisms of the foreign language effect as follows: Because working memory is
limited, one cannot easily concentnte on the more abstnct levels of meaning or semantic
integrity until the processing of the lower, subordinate levels (e-g.,orthographics, word
choice, wi thin-sentence gnmmar) of discourse production has become automated (cf.
automatic processing vs. controlled processing; see McLaughlin, 1987). This is to say, lack
of ski11 in linguistic processing reduces one's capacity for ideational thinking. Does this
mean that linguistically less proficient students cannot utilize writing as a means to gain
access to, explore, and develop their ideas? 1s i t futile to espect that linguistic a n d
ideational gains will be achieved by these students engaging in a writing task?
 Table 5-2
Srcrninary of Japanese E Î s n y Scores arzd Cornmeizis
- - -

cornparison nw-native group

(Jasmine. Jane, Judith. and Joy- -) 1 las-than-near-native group
(Jack. Cathy. Colin. Charles, Mary. bWte.
hl&in. Eiiot. and ~ d d i e ) -
Kei vs. Raters Kei's and the raters' scores were similar. Kei's and the raters' scores differcd. Kei's
Kei's scores were in gened slighrly higher 1 scores were substantiaiiy higher.
- --
but not always.
3 trait scores Scores of the 3 traits of each piece of writing Scores of the 3 traits of each piece of writing
(content vs. were high and similar COeach other. were Iower than the near-native group but
organization vs. similar to each other generally. There were
language use ) uses that showed a difference of 2 or more
points between the 3 trait scores.
Task 1 vs. Task Scores of the Task 1 and Task II essays of Scores of the Task 1 and Task II essays of
iI the sarne writer were high and simil& to the same writer were lower than the near-
each other. native group but similar to each other
generaiiy. There were cases chat showed a
&feren& of 5 or more points between the
same wnter's 2 task scores.
comment typçs Positive general comments were mos t Commenis on language use. both positive
(general vs. cornon. and negative. were most common.
content vs.
organization vs.
lanpuage use)

The low nti ngs on content and organization for the less-than-near-native learners'
essays appear to suggest a presence of foreign language effects. However, if we look at the
percentages of their attentional foci during the process of writing (see Appendis O), we sec
there was in fact little quantitative difference between the near-native and less-than-near-
native groups.9 The highly proficient Jasmine and Jane paid as much attention to content
or gist (somewhere between 20 and 50%)atid language use (somewhere between ?O and
60%) as did the less proficient Cathy, Mary, Martin, and Eddie. The unique case was Colin

whose on-line verbal comments disproportionately concentnted on content (76%in Task 1

and 66 % in Task LI). Judging frorn the level of his Japanese proficiency, 1 would presume
that linguistic knowledge should be an element Colin lacked or was constantly in doubt

57 A further study may bc carried out to examine the students' think-aioud protocols to see if therr:
were differences in the qualities of knowledge or W g they brought to bear.
about; it thus should have become the focus of his attention during his production of
Japanese written discourse. On the contnry, Colin nrely thought out loud about his
language use (4%in Task 1 and 17% in Task II) while composing (see botnote 38).
What is the meaning of Colin's intensive engagement in content genemtion and
refinement during the production of his Japanese compositions? How might i t be related to
his learning esperiences? Next 1 looked at the results of the SOLO analysis, hoping that
they would shed some light on these questions.

5.2.2. Relah'ons between SOLO LeveLs atui Other Factors

The results of the SOLO analysis are given in Table 5-3. Here again the near-native
versus less-than-near-native distinction seems to best describe the results. The essays of the
near-native Jasmine, Jane, Judith, and Joy were judged rniiltistriictiiral, rniiltistriictwal
trarzsitional, and relnrionnl whereas the essays of the less-than-near-native students ranged
from prestri<ciiirczl to ii~iisiriicinralrrairtitioiial. The hi gher-order t hinki ng58 of the near-
native group might have been enabled by their Iàcility with the language (i.e., their
automaticity in language processing); on the other hand, such thinking rnay have been
hindered by the "foreign language effect" (Takano & Noda. 1995) in the case of the less-
than-near-native group. But how about Colin, who emerged with multistructural
transitional and relationai discourse structures? The SOLO analysis evidenced the quality
of his thinhng despite his lack of Japanese proficiency. Apparently there are more things
than foreign language effect to enplain foreign language writing perlormances.

58 This tenn requires a definition. 1 used the term broadly here as refemiag to the kind of thinZring thrit
leads to an intercomected and hierarchical stnrcturing of one's howledge and a better organization of this
hierarchy of howledge around explanatory or causal relationshps (Carnine & Kameenui, 1992; R. Glaser,
1985). Smith's (lm,pp. 23-26) phi~osophicalargument against the notion of higher-ordcr thinking is
signifiant and should be borne in mind.
 Table 5-3
SOLO Levels cf Japntrese Witten Texts

) hhcipant 1 Task 1 1 Task II 1

S

Jasmine relational relationai

Jane mdtistniccural mu1tistructural transi tional
Judith mu1tistntctural transitionai multistnrciural
1 Joy 1 mu1tiscnictural 1 mu1tistructural vansitional I
Jack prestructural presmctural
Cathy unistructural unistnicnual
Colin rnd tistructurd transitional relationril
Charles unistnictural unistructural
Im
~Mke
I unistructural transitiond
unistructural
I unistructurai uansitiod
unistnicniral
I
iLiartin unistrllctural unismcniral .
Eliot unistmctural unismcniral
1 Eddie 1 unistructurai transit i o d 1 Unistnicturai transitional I

I t was tme in most cases that the students' linguistic cornpetence in Japanese was a
main factor to distinguish the high from the low qualities of their Japanese essays. The
essays written by the students of native-like proficiency consistently scored much higher
than the less-than-native-like students' essays in al1 three components (content,
organization, and language use). Their tests were also judged to present more complex
cognitive structures. Other factors such as iask and cognitive performance during the
production of the essays appeared to have little association rvith the qualities of the written
products (see the table in Appendis O and Table 5-3).
Al thoug h there were some qualitative differences between the near-native and the
less-than-near-native students' Japanese essays, al1 the tests (including the near-native
students' essays) were, to a greater or less degree, written in the "concrete-symbolic" mode
(Biggs, 1987, p. 117) where language was used to dexribe the esperienced world or retell
class notes or information obtained from other sources. AIthough some traces of
reflectiveness were observed in the essays wi th mul tistructural transitional and relational
stnictures, most actually displayed the kind of "knowledge-telling" in writing (Bereiter &
Scardamalia, 1987) that is said to resul t from a sentence-by-sentence routine during which
the writer's concerns are simply what to say next, in suitable language, rather than to
attempt to transfonn one's knowledge in the processes of composing or what Bereiter and
Scardamdia d l e d "knowledge transforming" in writing.

5.3.Generd Remmks on the I I L WA~nalyses of the &ikz

When 1 was done with the coIlection of data (which are displayed in this and the
preceding three chapters) in April 1995, 1 found myself ovenvhelmed by the amount and
vanety of data 1 had collected. They seemed rich yet difficult to understand for their
significance. They appeared connected in a complex way, if only by virtue of their having
corne from the same context, which was what 1 wished to understand better; yet they were
still disconnected. This sensation is probably one that frustrates most grounded, natudistic
researchers; wi thout a pre-ordained research design, researchers feel out of control and
" unscientific." Inductive analyses in naturally esisting settings wi th naturally functioning
human participants ma- not be theoretically tidy, but their compleaity is potentially as
exhilarating as the real life they attempt to study (cf. Newman et al., 1989).
In this and the preceding three chapters 1 have laid out, discretel y and as completely
as I c m , al1 the data I collected from Kei, her classroom, and her students. Putting al1 the
pieces of collected information into a wntten text was itself a process of reduction and
interpretation, as were al1 the preliminary analyses, codings, and ntings 1 did. This process
was also a self-conscious one, shaping where my focus on the JFL students' expenences of

wntingto-leam was heading. A t that time some imagery of the phenornena 1 experienced
as a researcher began to form in my mind. I saw images of a JFL teacher concemed about
the current state of language instruction skewed toward mechanistic, shallow-level leaming

of language foms without encounging ideational thinking and mindfulness. I also saw her
highly motivated JFL adult students of the three different ethnic groups and varied levels
of Japanese proficiency and writing skill exhibiting similar and differing characteristics in
the processes and products of their Japanese writing.
My foregoing impressions and intuitions about the data have shaped these images
into various f o m s of questions about the JFL wnters' cognitive operations of linguistic
and ideational processing. From the pedagogical point of view, the principal issue is
whether thinking and mindfulness can be encounged in the production of foreign language
written discourse. This problem is big and complex. For instance. although the genenl
belief is that ideational thinking skills are complex and language slcills more rudimenîary,
just the reverse may be the case in activities of foreign language discourse production.
Furthemore, the hierarchy of basic skills to cornplex processes--for instance. from wri ting
a simple single sentence to the production of meaningful written discourse--is so ingrained
in language activity that it may be difficult to conceive of the interdependence of language
processing skills and the skills of reasoning and thinhng.
1 had prepared al1 the data in prelirninary forrn for analyses. and consolidated the

image of the prima- objects of my inquiry as human agents in a purposefut educational

context, and the image of human agents as information processors who would process
different arnounts and types of information according to their own information-processing
capacities, their "personal needs and interests ... inclinations and impulses" (Vygotsky,
1986, p. 10). and the nature of tasks they were engaged in. I then decided to subjeci my

impressionistic interpretations of the phenornena to more rigorous and systematic

procedures of anal ysis. 1 used a unique statistical technique called dzial scali~zg(Nishisato,
1994), a technique to analyze categorical data that are not strictly quantitative but

nonmetric and thus qualitative, to re-assess the data. 1 present the procedures and results of
dual-scaling analyses in the next chapter.
 Chapter 6: Dual-Scaling Analyses

6.1. Tite Data Revisited

In chapters 3, 4, and 5, 1 used matrices to compare information across cases (i.e.,

participants). Creating these matrices involved the reduction of values and variables and

the elimination of detail. 1 saw this process as a virtue, for it helped to clarify the foci of
my analyses. In the same chapters 1 used the matrices to explore relationships between
different elements by visual inspection. A further process of abstraction was, however,

necessary (and proved invaluable) for dual-scaling analysis. This process involved the re-
examination and re-organization of the existing categories and as the necessi ty arose, the
splitting, slicing, combining, creation, and elimination of categories as well as the

translation of re-established categories into variables together with subsequent re-coding.

The key principle in this process was to assign as unambiguous meanings as possible to
data in order to enhance the elegance and power of the mathematical manipulations behind
the dual-scaling technique. Efforts were made to be as explicit as possible about the criteria
used for allocating values to the cells in six case-by-i tern rnatns tables (Tables Pl -P6 in

Appendix P), which provide an overview of al1 the main features of each case. Tables Pl

to P6 contain 44 different kmds of information for each student. I devised this large set of
data so as to make it resemble 44 items of multiple-choice data. Table Pl containing Items

1 to 10 shows a summary of the data collected through the initial questionnaire surveying

the participants' backgrounds and linguistic experiences. Table P2 summarizes the results
of the assessrnent of three language skills (Items 11-13),The data summarized in Table P3

corne frorn the interviews conceming the participants' conceptions of JFL writing and
leaming (Items 14-16). Table P4 gives a summary of the participants' responses to the
post-writing questionnaires (Items 17-22). Table P5 is a summary of the think-aloud
protocol data focusing on reflective mental activities (Items 73-30). Table P6 summarizes
the component scores of Japancse essays assessed by Kei and the two independent raters as
well as the SOLO levels of the essays (Items 3 1-44).
Note that numbers were used, for simpiicity, to represent the values (3 to 4 options
per item) chosen by individual participants or identified by the researcher; however, they
do not carry any numerid values. Also, numbers are shown in bidface where 1 was able
to collect complete sets of data, including the think-aloud protocols. The entry 'O'indicates
missing information. Nishisato (1994, pp. 300-306) described different procedures to
handle missing pieces of information in dual-scaling analyses. The D U A U Statistical
Software Series package (see Appendix of Nishisato, 1994; Appendix II of Nishisato &
Nishisato, 1994) that 1 used for data analyses adopts the method that analyzes only
observed responses, ignoring missing ones. Although this means an inevitable influence on
the outcome of scaling, this method, according to Nishisato (1994), is probably the best
approach to handle missing bits of information so long as they occur infrequently (which
was the case in my research). In my data, Joy's L1 essay score was the only missing piece

of information that affected the outcome of scaling (see Table P? in Appendix P). Cases
where think-aloud protocols were unavailable were excl uded from analyses involving
these data.
In Tables PI to P6 the same option categories are not repeated so as to avoid
redundancy whenever they appear more than three times i n the same table. The
demarcation of the essays' total and component scores into the poorJair, good, and
excelleilt categories (Items 11, 13, 3 1-42) was made following Jacob et al.'s ( 198 1)
assignment of these terms to certain numencal ranges. Percentage ranges used as values to
assign to each case's think-aloud protocol data are arbitrary; they were demarcated in the
way that made sense to me after my careful esamination of the data.
 My understanding of dual scaiing cornes from the writings and lectures by, and
persona1 communications wi th, Professor Shizuhi k o Nishisato w ho proposed the name
duni scalitzg for the technique and developed it into a practical t m l for data analysis. 1 owe

the following description of d u d scaling to his work (Nishisato, 1994, 19%; Nishisato &
Nishisato, 1984, 1994).
Duai scaiing is, a s mentioned earlier, a method of analyzing categoncal data whose
origins can be traced back to the 18th century (Nishisato & Nishisato, 1984, Appendix II).
It is capable of handling a variety of categorical data such as contingency, frequency,
mu! ti ple-choice, paired com parison, rank-order, successive-category , sorti ng, and
multidimensional categorical data. As the amount of research in duai scaling has increased
markedly over the past few decades, the scope of its applications has broadened as well
(Nishisato & Nishisato, 1984, 1994).
T h e method enables the measurement of the reIative "distances" between cases
(Le., participants), and between items (or options in the case of mu1tiple-choice data), by
assigning optimal weights both for items (or options) and for participants. In othcr words,
dual scaling is an analytical procedure that attaches weights to items and scores to
participants in a way that masirnizes the squared-correlation ratio (i.e., a ratio of the
between-column, o r between-row, surn of squares to the total sum of squares). The duality
implied by its name refers to this symmetry of the analysis. Maximizing the ratio of the
between-item sum of squares to the total sum of squares results in optimal weights for the
items. from which optimal scores for participants can be directly derived. Likewise.
optimally assigned participant scores imply that weights for the items are optimal as well
and can be expressed as a function of participants' scores.
 The crus of dual scaling lies in its simultaneous maximization of the between-row
and the between-column surn of squares. This simultaneous, optimal quantification of the
rows and columns of a data rnatnx is one of the most remarkable katures of the dual-
scaling procedure. In addition, since no particular distribution for responses is assumed in
dual scaling, no assumptions have to be made regarding the distribution of the data. This is
a strength in terms of its applicability as well as a weakness in its inferential potential.
Indeed dual scaling is sample-dependent so the results of its analysis cannot be genedized
to other populations. What dual scaling does is to provide "a simpler, often ciearer,
description" of a given data set (Nishisato, 1994, p. 17). To use Eisner's ( 199 1, p. 95)
distinction between description and interpretation, dual scaling gives an account of, but
does not necessarily account for, the given data. My research, like many other studies in
educational research, values a full description of a particular group of people and the
patterns of their behaviors more than genenlizabili ty. 1n this circumstance dual scaling
offers a particularly effective technique for obtaining rich descriptive analyses. It is also an
excellent tool Tor esamining the interrelationships among items and participants and
provides a wealth of information that would probably be lost or overlooked in other types
of analysis.

6.3.Dua.i-Scaiing A nalytic Procedures

The D U A U Statistical Software Series package offers various computer

applications of dual scaling. I used one of its five prognms, called DUAUMC (Version
2.36, July 1992), for multiple-choice data, to perform a series of analyses.sg I n total I

59 There arc two types of analysis that c m be carried out with dual scding. One procedure is called
standard andysis or option weighting that focuses on the entire data set. The other is called forced
ciassification, the procedure developed by Professor Shizuhiko Nishisato. Forced classification allows the
investigator to focus his or her anaiysis on a specific aspect of the data; he or she can specify a pariicular
canied out 6 analyses in a cumulative fashion. These analyses addressed the three sets of
research questions 1 posed related to the threc cornponents of Biggs's 3P model focusing o n
students' characteristics, iheir writing processes, and their written products.60 1 did not
analyze data on the tacher, Kei, in this way, as they were not in a form that was amenable
to such analyses. The first anal ysis (Analysis 1) subjected the information col lected
through the initial background survey questionnaire (Items 1- 10, Table Pl) to dual scaling
to provide an overail picture of the particular group of people participating in the study.
Analysis 2 espanded the scope of Analysis 1 to include the assessrnent of the participants'

language skills (Items 1 1- 13, Table P7) and their conceptions of JFL writing and leaming
(Items 14- 16, Table P3) to answer the research question pertaining to students'
chmcteristics (see section L .4). My next level of analysis included the wri ting process data

to address the third set of research questions. Responses to the pst-wnting questionnaire
(Items 17-12, Table P4) were added to the preceding 16 items to make Analysis 3(a) for dl
of the 13 participants. Moreover, Anal ysis 3(b) included the thi nk-aloud protocol data
(Items 23-30, Table P5) as well as the pst-writing questionnaire data for 8 participants (as
the rernaining 5 participants failed to produce analyuble think-aioud protocols). The
fourth and final set of rcsearch questions concerns al1 aspects OC the 3P model as a

classroom system. To address these questions, 1 used al1 the items escept Items 23 to 30
for the 13 participants in Analysis *a), and Analysis 4(b) included al1 the items for the 8
participants who produced usable thinhaloud protocols.

The DUAL3MC prognm extracted three solutions (or orthogonal cornponents) for
each analysis by derault; however, I present only the first two of these solutions in
discussing the outcornes. Dual scaling yields many solutions. According to Nishisato

(1994; see also Nishisato & Nishisato, 1984, 1994), there is no clear-cut answer to the

- - - - -- --

variable (or a set of variables) of his or her interest and examine the relationships in the data with respect to
this predetermined focal point. I used standard dual scaiing in this thesis research as an aid in summarizing
cornplex data in a simple, clear. and useful m m e r .
60 I also adcirascd these questions scparately in chapters 3.4, and 5, where 1 described the prelimhary
d y s e s of the relevant data and their results. See section 1.4, where the research questions are inaroduced.
question of the number of solutions to obtain. He recommended interpretability as a
prac tical cri tenon for deciding how many solutions to extract. Typicall y the first solution
tends to reveal a general division while the subsequent solutions capture more
idiosyncratic traits attributed to a small number of participants. In the present research I
considered the first two "best" or "optimal" solutions sufficient for explaining g e n e d
patterns of the information in the data sets.

Dual scaling provides a number of statistics. The following statistics are

particularly useful in interpreting scaling outcornes of multiple-choice data (Nishisato,
1994; Nishisato & Nishisato, 1984, 1994); I ivill refer to them in presenting the

interpretation of the computer outputs:

Perceritage horrrogerreity is an index of how good the solution is as compared with the
perfect case (i.e., al1 the inter-item correlations are 1 and a single item contains al1 of
the information in the data). This statistic equals 100 limes the squared correlation
ratio.

Item stntistics--S%), the sum of squares of weighted responses of each item, is a vital
statistic in test construction and indicates the relative contribution of Item j to the test
(Le., the set of al1 the items subjected to the analysis), which 1s proportional to the

square or item-total correlation, Eut).&i,, the positive square root of R2üt) or item-
total correlation. indicates the estent to which Item j is correlated with the test; the
higher the value of Rtit), the greater the relevance of Item j to the test. Thus, this
statistic can be used to select a subset of homogeneous items by discarding those items
) retaining those items with large values of the statistic.
with small values of R Q ~and

Participatits' projec~edscores represent a set of weights which reflect the relative

importance of the solution and are in contrast to the normed option weights
representing a set of weights that have a fixed or common unit. More technically
spealring, normed weights alone do not contain sufficient information to reproduce the

input data, but projected weights do (Nishisato, 19%).

 6.4. Outcornes of Dml-ScaCSng Analyses

I took the following steps to interpret the results of the dual-scaling analyses. First,
I looked at percentage homogeneities of Solutions 1 and 2 to see if the given data set
contained a dominant dimension or component. Second, to grasp the nature of the
solutions, 1 plotted each participant on a two-dimensional gnph using his or her projected
scores for Solutions 1 and 3 as the coordinates for the horizontal and vertical &.es.61 I then
partitioned participants into groups and identified those options shared by participants in
each group. Each group was further partitioned into subgroups by looking at items whose
options were not shared by al1 the participants in the proup.6? My esamination of these
patterns was important because these "patterns contain the full information of the [given]
data" (Nishisato, 1994, p. 160).
To facilitate this process, 1 identified the items with hîgh values of R üti rvhich were
major contributors to the analysisa; these items were thus capable of par-tiiioning the space
relatively neatly and chancterizing the clusters by their respective options. That is to Say, 1
looked at only a subset of strongly relevant items, nther than the entire set. In this contest
the importance of the percentage hornogenei ty, the use of w hich often mises questions,

61 Nishisato (1994) devoted one entire chapter to discussing problems associated with graphical
representations of duaI-scaling outcornes. He stated, "Granting that graphicat dispIay is almost indispensable
for interpretation of the quantification results, one should nevertheles be aware of some potential pitfaiis and
logical difficulties associated with graphical display" (p. 26 1). Despi te many uaresol ved problems wi th
grapbiul display in generai, Nishisato found a theoreticaily sound and p r a c t i d y useful graphical method to
present the results of dual s d i n g of multiple-choice data. The essence of his method is plotting the projected
scores of participants in two-dimensionai space for a pair of solutions and labeling them by their response
patterns. 1 used this method; however, the large number of i t e m included in my ruialyses caused difficulty in
labeling participants by their response patterns, which are too long to be displayed on a single graph.
Therefore, responsç patterns do not appear in my graphs. Rather. the participants' names are used to denote
the points in the plot. The concurrent opecation of classifying participants into clusters and distributing item
options into separate regions was, nonetheless. canied out in interpreting clusters of participants in tenns of
their common response/behrivior patterns by consulting Tables PI to P6 in Appendix P.
62 This operation can be repeatcd to yield "a tree structure, an outcome of hierarchid clustering"
(Nishisato. 1994, p. 289).
63 The numbcr of contributing items becomes smallcr as we move from the first to the subsequent
solutions, pointing to the decrease in the number of participants conuibuting to the latter solutions (Nishisato.
1994).
should diminish since the percentage homogenei ty applies to the entire set of items, not to

the selected subset of items (S. Nishisato, personal communication, June 30, 1997).
What the appropriate significance Ievel of Rut) is is a question yet to be explored
(Nishisato, 1994), however. Nishisato (1994) noted the nature of this question differs from
the traditional test for significant correlations, and the "relation of item-total correlation to
the smooth transition oï response patterns is a topic that has never been explored" (p. 160).
In the present study 1 considered items with Rg.0 of .50 o r higher to be major contributors,
following the practice demonstrated by Nishisato and Nishisato ( 1984, chap. 3) in their
dual-scaling analyses of multiple-choice data.

T h e percentage homogeneity was 47.3 1% for Solution L and 19.61% for Solution 3

in this analysis. suggesting that this instrument (Le., the questionnaire from which the data
were derived) contained medium and lower degrees OS homogenei ty for respective
solutions. Similarly. these medium and relaiiwly low values of the percentage

homogeneities imply that this group of participants (or respondents) were rather
~ ~ is, the instrument did not contain n dominant dimension o r
h e t e r o g e n e ~ u s .That
component; rather, the anal ysis captured individual di fferences in answering di fferen t

questions. Figure 6- 1 shows each participant o n a two-dimensional graph, from ivhich

three distinct clusters can be identified to partition the participants.

a This kind of argument applics both to the naturc of item distributions and to the nature of participant
distributions because of the symmetry of dud swling (Nishisato & Nishisato, 1984). This unique
characteristic of dual scaling, inentioned in section 6.2 in this thesis, is evident in ihe way Nishisato and
Nishisato (1994) explained the m&ng of ihc technique's name: "'symrnetric quantification' of a two-way
table of cxtcgorical daia" (p. 115).
 T o interprei the clusters in ternis of their cornmon response patterns. I considered
the items ivith RGC)of -50o r higher for the two solutions (Tables 6- 1 & 6-2) and examined
response pattems of these items by l o o h n g at Table Pl (Appendix P). Tables 6- 1 and 6-2
also include the values of SS(j) as well as options wi th outstanding bipolar weights65 (i-e.,
options of a particular item that stand o n the dichotomous ends of a scale), which indicate
distinct traits of people in a certain group. These two ways of interpreting the information
about the participants' response patterns complemented each other and helped m e g n s p
the nature of the solutions better.

Table 6-1
Major Coritributors ami Their Bipolnr Optiotis (Aoiysis 1. Solrctioii 1 )
i tern description
oral LI Chinex

self-rating of overall Japnese (cf-pers) P r

ethnici t y Chinese
witten LI Chi nese

importance of Japanese study very important

length of residence in Japan less than 6 months

sef f-rating of Jripmcse writing (cf.pecrs) PoQr

Table 6- 1 contrasts the oral and written Lls of Anglo-saxon students against those
of the ethnic Chinese students. The other contnstive factors were self-ratings of o v e n i l
and written Japanese compared to peers (good vs. poor), importance of Japanese study
(important vs. very important), and length of residence in Japan (more vs. less than 6
months). Table 6-2 similarl y contrasts the ethnic Japanese students and the Anglo-saxon
students.

65 In case the samc or very close weights were derivcd for more than one option of a particular item.
aii of these options are listed; for instance, in SoIution 1 of h d y s i s 2 (see Table 6-3) the projected wcights
were minus 0.92 for the option of superior and minus 0.80 for the option of advancrd of the item of JST
rating, and the projected wcight was minus 0.77 both for the option of Japanese and for the option of h g l o -
Saxon in the item of ethmcity.
 Table 6 - 2
Major Contributors and Their Bipolar Options (Andysis 1. Soktio>r2)
item description - - -

ethnicity Japan-
self-rating of overd1 Japanese (cfmtive) fair
self-nting of L1 writing iair
on1 L1 Japanese
self-rating of Japanesc writing (cf.peers) poor~fair
length of rcsidence in Japan Lcss than 6 months

Among the items listed in Tables 6-1 and 6-2, so-called "very good" items for the
purpose of classification--that is, items that are highly correlated within the questionnaire
(Le.. items with higher Rüi))--in this analysis were: ethnicity (Item l ) , oral L1 (Item 2).
wntten L 1 (Item 3), self-ntings of Japanese proficiency overall in comparison to peers
(Item 7), and self-ratings of Japanese proficiency overall in comparison to native speakers
(Item 8). The other items in these tables did contribute to uncovering some response
pattems hidden in the data, but they are less useful than Items 1, 7. 3, 7. and 8 in
classifying the participants into distinct clusters by their options. In the following
descriptions of the clusters, 1 compared the participants' options on the selected items to
identify the response pattems underlying the solutions.

One cluster in Figure 6- 1 consists of ethnic Japanese students-Jasmine, Jane,

Judith, Joy, and Jack--who used English as their oral and written L1 (but note that Jasmine
preferred to use Japanese for daily conversation and English for writing) and considered
their overall Japanese as good in comparison to other students in the class (i.e., had
relative1y high confidence in their Japanese). W hen their ovenll Japanese ski11 was
compared to that of native speakers, Jasmine, Jane, and Judith nted i t as fair while Joy and

Jack were less confident. nting it as pour. Judith and Joy had recently stayed in Japan for
more than 6 months but the others had not.
 Another cl uster consists of ethnic Chinese students-Cathy, Colin, Charles, Mary,
and Mike--who spolie either Cantonese or Mandarin Chinese as their L1. Although they
had been studying in an English-speaking environment in Canada for some time, they still
felt more cornfortable writing in Chinese than any other Ianguages, escept for Mike who
had been in Canada since Grade 4, the earliest aniving immigrant among the five, and who
chose English as his written L1. Al1 these students rated their ovenll Japanese as poor in
companson to native speakers. Cathy, Colin, and Charles also felt that their overall
Japanese was poor in cornparison to their peers in the class; but Mary and Mike nted theirs
asJiair, rating higher than the 3 Cantonese-speaking students. Supplementary information 1
obtained during the interviews might esplain Mary's and Mike's higher ratings on this

item. Both Mary and Mike had relatives in Japan and had visited Japan in their young
childhood, whereas the Cantonese students neither had such contacts nor had visited Japan
at d l . None had a recent experience of king in Japan for an estended p e n d of tirne.
The third ciuster consists of native English speakers. Eliot and Eddie were white
native-born Canadians while Martin was a Chinese-Canadian bom and raised in Canada-
These 3 students had spent one academic year in Japan to study or work prior to their
enrollment in Kei's course. Although they ail le1t their Japanese was poor compared to
native speakers, Martin and Eliot rated their overall Japanese, compared to their peen, as
good, and Eddie rated his as fair. Their relatively high confidence in their Japanese

indicated by these ratings rnight have something to do with their recent experience of
l 1tem 6 (length of resîdence in Japan in the past 5 years)
living in Japan. In fact the R ~ t of
in this analysis was reasonably high (-72 for Solution 1 and .51 for Solution 2 ) and could
be useful in distinguishing this particular group of people from the other groups.

Some other interesting trends are worth noting here. Discrepancies between the
self-ntings of overall Japanese proficiency and written Japanese were most notable for
ethnic Japanese students, while the other students' ratings on these two items were
consistent. This finding points to a gap that ethnic Japanese students perceived between
their overall proficiency and writing sliills in Japanese. It may be further inferred from this
that these edinic Japanese students had set higher goals for their Japanese wnting, wishing
to be able to wnte as well as they could s@, for instance. Another interesting point is the
self-rating of L1 writing ability (which was in fact not a very good item, with Rct) of -31
for Solution 1 and -68for Solution 2. contributing little to the clear demarcation of the
participants). That is, those w ho rated their LI wri ting ability as fair were al1 female escept
Mike, while those who considered themselves as either good or e.rcelletrr writers in their
Lls were male students. Although gender was not included in this statistical analysis, it
may have influenced these people's self-ntings of abilities and attitudes.

For Analysis 2, the percentage homogeneities of Solutions L and 2 were 43.43%

and 30.87% respectively. Figure 6-2 shows a picture similar to the one that appeared in
Analysis 1. where three clusters were distinguished from one another by ethnicity, oral LI.
written L1,and self-ratings of ovenll proficiency in Japanese. One main difference from
Analysis 1, though, is that Jack has moved away from the cluster of ethnic Japanese
participants and stands alone. To interpret this analysis, 1 loohd at the items with Rüt) of
-50 or higher for the two solutions (Tables 6-3 & 6-4) and esamined the participants'

responselbehavior patterns to these items by consulting Tables Pl, P?. and P3 (Appendis

P)-
 Table 6-3
Major Co~tributorsatld Tileir Bipolar Options (Attaiysis 2, Solution 1 )
item description <------ ---- ---------->

oral L1 JapanesdEnglish Chinese

JST rating superiorladvanced intemediate
self-rating of overall Japanese (cf-peers) goai POO'
written LI English Chi nese

ethnicity Japanese/Angle Saxon Chinese

perceived area of leaming ihni E L w-ritiag language & content "nothingMIcontent

length of residence in Japan more than 6 months less than 6 months
LI essay score excellent
importance of Japanese study important very important

self-rating oï Japanese writing (cf.peers) goodlfair P r

self-rating of ovcrall Japanese (cfmtive) fair P r
perceiveci area ol problem in E L writi ng organization language & organization

Table 6-3 highlights the contrat between the ethnic Japanese and Anglo-saxon
students and the ethnic Chinese studen ts; Table 6-4 contrasts the Anglo-Sason students
with the ethnic Japanese students. And these contrasts are clearly indicated by the three
distinct clusters shown in Figure 6-3. These clusters are main1y characterized by the three
distinct leveis of JST ratings (whose Rut) was -91 for Solution 1 and .78 for Solution I ) ,
which happened to correspond roughly to the three distinct ethnicities (with R G ~of) -83 for

Solution 1 and .78 for Solution 2). All ethnic Japanese students were rated sriperior on the

JST except Judith, who was advarrced,66 and Jack, who was i~itermdiate.Al1 Chinese

66 Judith was rated advmced-plus, the level characterizcd by "remarkable fluency and ease of speech"
according to the ACTFL Japanese Roficiency Guidelines (American Council on the Teaching of Foreign
Languages. 1987). The advanced-plus speaker may experience a breakdowu under the demands of superior
leveI complex tasks. Eliot was also an advanced-plus speaker while Eddie and Martin wcre nted advanced.
Ai though Judith and Eliot received the sanie rating on the IST, my observation was that Judilh's ability to use
the Japanese language was much better and more native-Iike than Eliot. This difference 1 noticed between
these two students may be due to the fact that Japanese was one of Judith's two home lanemges while that
was not the case for Eliot who had had a much more limited amount of formal and informal Japanese
instruction compared to Judith. This issue certainly speaks to the complex nature of the concept of
proficicncy and the mechanism of tcsting, but it is not relevant in the present thesis and thus not discussed
students were rated intermediate. Eddie, EIiot, and Martin were advarrced.

Table 6-4
Major Contributors arid Tlieir Bipoiur Optiom (kialysir 2. Solutioiz 2)
item description <----------

J S T rating advanced superior

ethnicity .AngbSa.on Japuiese

Japanese essay score g d exellent

self-rating of LI writing excellenUgood fair

self-nting of Japanese writing (cf-pers) g d poorl fair
self-rating of ovcrall Sapanese (cf.native) P='r fair

perceived area of problem in JFL,witing organization Ianguage & orgrinization

oral L1 English Japanesc

perception of JFL writing language exercise meaning-making
length of residence in lapan more than 6 rnonths iess than 6 months

perceived area of learning thni IFL wriring languagdcontent "nothing"/language& cont

Other items listed in Tables 6-3 and 6-4 further charicterize these three groups
generally. For example, the superior Japanese-proficiency group (including Judith) scored

in the excellerit range on their Japanese essays, consîdered JFL writing as a meaning-
making activity, and had a felt-sense of having learned about language and content through
engaging in JFL writing tasks. The Japanese essay scores of the advanced Japanese-

proficient group fell in the good range, and they regarded JFL writing primarily as an
activi ty for language exercises. This conception of JFL writing as a mere language esercise
was also a characteristic of the Chinese-background participants w h o were j udged
intemediate on the JST and produced Japanese essays in the /air or good range. Boih the

intermediate and advanced Japanese-proficiency groups perceived their main Ieaming

through JFL writing to be language use, if anything. Across the groups the rnajority scored
excellent on their L1 essays (an indication that little difference may have existed among
the participants' writing skills in their Lls) and found language use as a major stumbling
block to writing in JFL regardless of their proficiency levels in Japanese (see Tables P2 &
P3 in Appendix P).

The percentage homogeneities of Solutions I and 2 in this analysis were 37.60%

and 37.74%respective1y, indicating a slight decrease in the accountabili ty of each solution

compared to the previous analyses. This suggests that this particular set of data contains a
cornplex hidden structure. The distribution of the participants plotted in Figure 6-3

resembles that in Figure 6-2, showing three distinct clusters, with Jack isolated from any of
these clusters.
Looking at Tables 6-5 and 6-6 as well as Tables Pl to P4 in Appendix P helps to
identify each cluster's common charac teristics in particular relation to the items added to
this analysis, i .e., the retrospective information on wri ting processes gathered through the

pst-writing questionnaires (1tems 17-72).Tables 6-5 and 6-6 highlight respectively the

con tnsts between the Chinese students of intermediate proficiency in Japanese and the
Anglo-saxon and ethnic Japanese students of advanced or superior Japanese-proficiency.
and between the Anglo-saxon and the ethnic Japanese students. The three clusters shown
in Figure 6-3 reflect these contrasts.
 Table 6-5
Major Contribulors arrd Their Bipolur Optiom (kzalysis 3(a).Soluliorr 1 )
item description <-- ----------- -- - -

orai L1 Chinese JapaneseGq$ish

level of satisfaction with Task 1 essay not so satisfied very satisfied/satisfied

JST nting intemediate supenor/advanced

sel f-rating of overail Japanese (cf-pers) PQor good

witten LI Chinese English

ethnicity Chi nese AngbSason/Japanese

pcrceived a r a of leaming thm JFL witing "nothing"lcontent language &content

length of residence in Japan l e s than 6 months more than 6 months

Ll essay score go=' excellent

self-rating of Japanese witing (cf.peers) P r g d

Japancse essay score fair e~cellent

importance of Japanese study very important important

difficdty in Ianguage use in Tasli 1 difficult not so difficult

perceived a r a of problem in JFL writing language & organization orgmizrition

Table 6-6
Major Corztribzitors orid Their Bipolar Op~iorls(A~zniysis3(a). Solrition 2 )
item description

difficulty level in idea generation in Tasic f difficult

JST rating superior

Japanese essay score excellent

ethnicity Japanese

self-ratingof Japanese writing (cf-peers) poorl fair

self-rating of overall Japanese (cf-native) fair

self-nting of LI writing fair

oral L1 Japanese

satisfaction level in Task 1 essay very satisfied

perceived area of lerirning thm E L writing language & contl"nothing"

perceived area of problem in JFL writing language & orpnization

perception of JFL wnting meaningrnaking

 The students' responses to Items ?O to 22 of the post-writing questionnaire
conceming the level of satisfaction with and difficulty in idea genention and language use
in the Task II essay did not help much in demarcating each cluster's uniqueness; the
participants' generaily uniform responses to these items yielded relatively low values of
Rüt) (.4?. .63, and .O3 for 1tems ?O, 2 1, and 72 respectively in Solution 1; .14, -13, and 2 6
in Solution 2). By consulting the relevant tables in Appendis P, we can tell the majonty of
the students, showing little individual difference among them, were reasonably satisfied
with their Task 11 essays and did not esperience much difficulty in idea generation or
language use while writing these essays.
There were some distinct tendencies. however. in the three groups' response
patterns concerning Items 17 to 19, which concemed their Ievel of satisfaction wiih and
difficulty in idea generation and language use in the Task 1 essay. Students in the supenor
Japanese-proficiency group (including Judith) were al1 satisfied with their essays, and most
of thern esperienced difficulties in both idea genention and language use (but for Jasmine
and Joy, either one of these aspects posed difficulty). The advanced Japanese-proficiency
group were also satisfied with their essays; but neither idea generation nor language use
was a problem for thcm (encept Eddie who had difficulties in language use). The
intemediate Japanese-proficiency group's responses to Items 17 and 19 were unanimous;
they were not so sarkfied wi th their essays, and they expenenced difficutties in language
use. Their responses to Item 18 about the level of difficulty in idea genention were a m i s
and rior so diJiccrilf.
of dific~ilf
1 further compared, though speculatively, the levels of difficulty the students felt
about idea generdtion in Task 1 wi th the levels of confidence they had in writing in their
Lls, which I associated with one of the background questionnaire items-i-e.. self-ntings
of LI writing (Item 4)--contributing mainly to distinguishing the advanced-proficient
Anglo-Saxon group from the superior-proficientethnic Japanese group in Solution 3 in this
analysis ( R ~ t =
l .69);1 did this by examining the participants' option patterns of these items
shown in A ppendix P. 1t appeared generally true in the case of Task 1 that the students who
raied their L 1 wri ting competence as fair evperienced difficulties in idea genention
(Jasmine, Jane, Judith, and Cathy), and the students who considered their L1 wnting skills
as good or excellerit (Jack, Colin, Martin, Eliot, and Eddie) did not find the generation of
ideas so difficult.
I also compared the levels of difficulty the students felt about language use in Task
1 with the students' self-ntings of Japanese writing skills in comparison to their peers

(Item 9 with Rüt) of .58 for Solution 1 and of .71 for Solution 3). The students who nted
their Japanese writing skills as pwr or fair in comparison to their p e r s mentioned (on the
Task 1 post-writing questionnaire) difficulties they had in language use (Jane, Judith. Joy,
Jack. Cathy. Colin, Charles, Mary, Mike, and Eddie). Jasmine rated her Japanese wn ting
ski11 only as fair yet she responded on the pst-writing questionnaire that the use of
language was not so difficult for her. The students who mted their Japanese writing skills
as good (Martin and Eliot) did not find the aspect of language use so difficult, according to
their responses to the pst-wri ting questionnaire.

For Analysis 3(b), the percentage homogeneities of Solutions 1 and 3 were 37.69%
and 30.66% respectively. The clustering patterns shown in Figure 6-4 are basically what
appeared in Figure 6-3. Jack is again isolated [rom any of the three clusters. Tables 6-7 and
6-8 and Tables Pl to P5 in Appendis P help us esamine the participants' chancteristics in
tenns of the items highly correlated within the analysis.
 Table 6-7
Mnjor Coatribiriors ami Their Bipolar Optiotrr (Arialvsis 3(b),Solution I )
item description ---- ---------------- >

JST rating intemediate superior

level of satisfaction with Trisk 1 essay not so satisfied very satisfi ed

on1 L1 Chinese Japanese

perceiveci area of learning thm JFL writing "nothing" Ianguage & content

Japlnese essay score fair excellent

wri tten L1 Chinese English

self-rating of overall Japiincse t cf.native) POO' lai r

n t i o of thinking aIoud in L1 in Task 1 more ihrin KI% less than 2040

n t i o of thinking aloud in Ll in Task II more than J0% ICSS than ?@%

ethnicity Chinese Japanese

sel f-rating of overal l Japanese (cf.peers) P r g - J

reflectivc ihinking about ianguage in Task U 3&-!@?0 over 43%

refiective thinking about l q g e inTmk 1 3-0 20-3û%

difficulty in idea generation in Task II not so difficult difficult

difficulty in language use in Task 1 difficu!t not so difficult

self-rating of Japanese writi ng (cf.peers) POO' fair

difficulty in idea generation in Task 1 not so difficult di fficult

The contrasting characteristics of the item options listed in Tables 6-7 and 6-8 bring
us certain images of the three respective groups demarcated clearly in Figure 6-4. The
intermediate Japanese-proficient people were mostly Chinese and were not very satisfied
with their Task 1 essays, in which they esperienced dilficulties using language but not
necessarily in generating ideas. Idea genention was not a problem for them in Task II,
either. Thirty to 10% of their verbalizations of concurrent thoughts in both Tasks 1 and II
were about language use. In contrast, the supenor Japanese-proficient people were very
satisfied with their essays for Task 1. They round it difficult to generate ideas, but they did
not find it so difficult to use language for Task 1. And they verbalized less their thinking
about language and more about content for Task 1. But their verbalizations of thoughts
about language increased for Task II. As in their experiences in Task 1, they had difficulty
generating ideas for Task II. The advanced Japanese-proficient people viewed JFL writing
pri mari1y as a language esercise (in contrast to the superior-proficient students '
conceptualization of JFL writing as meaning-making), expressed no difficulty in
genenting ideas for Task 1, and stated a reasonable degree of satisfaction with their Task 1
essays.
Dual scaling captured these characteristics. From this we c m d n w distinct images
of the three groups.6' But there existed an a m y of individual differences that do not fit

neatly into the three clusten shown in Figure 6-4. For instance, if we look at the items
conceming thinking processes for the production of a piece of writing (Table P5 in
Appendis P68), the corn plesi ty of the data structure involving 8 participants becomes
apparent. Sorne manifested consistent performances across Tasks I and II (Jack. Mary,
Martin, and Eddie). but others did not. Eddie and Mary paid more attention to language use
than gist in both tasks. Cathy was like them in Task 1, verbalizing her concems about
language use more oiten than about pist. But in Task II her renective thinking about
language use decreased. In both tasks about 30 to 30% of Martin's thinking was of a
reflective kind and about language use; another ?O to 30% was of a reflective kind and
about gist. Jack's protocol data showed little reflectiveness in his thinhng. Jasmine, Jane.
and Colin thought reflectively about gist in Task 1, but they did much less in Task II.
Jasmine was concemed about language use in Task 1 but not so much in Task II. Jane
manifested the opposite phenornenon. Colin did not verbalize rnuch of his thinkng about
language use in either task. The aspect of writing procedures was least attended to
reflectively by al1 the participants. In sum, it is hard to find any consistency wi thin this set
of think-aloud protocol data, nor to see consistency in their association wi th other variables

-- - -- - --

67 This implies that thesc images are not necessarily the exact descriptions of al1 ihe individuais
cornprising respective groups, but rather prevalent tendencies.
68 The idonnation contained in Table P5 focuses on reflective remarks of concurrent thinking while
writing and does not inciude less reflective remadcs. for it is reflective thin)iing that is believed to accornpany
"meaningful learning" (Smith. IW, p. 127).
included in this anal ysis.

Table 6-8
Major Corilribirtors attd TTiir Bipoiar Optiom (Amlysis 3(b).Soiu~ion2)
item description <-------- ----------- --------A--------
>

JST rating superior advanced

importance of Jripanese study very important important
length of residence in Japm Iess than 6 rnonths more than 6 months
difficult). level in i d e generation in T& I difftcult easy
self-rating of LI writing fair g"=)d
reflcctive thinking about gist in Task 1 -la-% 3M

self-rating of Japanese witing (cf-peers) P r iPd

satisfaction level in Task 1 essay very saiisfied satisfied
o n l L1 Japanese English
Japanese ssay score excellent g d
ethicr ty Japuiese XngbSa~on

perception of JFLwriting rneaning making language enercisc

self-rating of overdl Japanese (cf-native) fair P r
ntio of thinking riioud in Li in Task I -?@?%O -KI-%

ratio of thinking doud in L1 in Task II -20% 404

For Analysis 4(a),the percentage hornogeneities of Solutions 1 and 3 were 38.52%

and 3.47% respectively. Figure 6-5 shows the participants clustered into three groups,

rnainly characterized by three levels of Japanese proficiency (Le., 1tem 12 was a major
contnbutor with Rüt) of -89Tor Solution L and -76 for Solution 2). Jack is now plotted with
the other intermediate proficiency students. To identi fy the chancteris tics of product data
for these respective clusters, I examined the matns presented in Table P6 in Appendix P as
well as the lists of al1 the contributing items (i-e., items with Rüt) of .5û or higher) for
Solutions 1 and 2 (which are given in Tables QI and Q3 in Appendix Q because of their
length).
T h e lists of items with high values of Rtit) in this analysis (Tables Q 1 & Q?)
include a number of items concerning the qualities of written products. This rneans these
items contributed to the segmentation of the space of Solutions 1 and 2 into three distinct
clusters shown in Figure 6-5. This further rneans these clusters can be identified by distinct
characteristics of written products. In fact, regulanty in the transition of option pattems can
be easily identified in the results of the assessments of the participants' Japanese essays
(Table P6 in Appendix P). Generally speahng, the supenor Japanese-proficient group nas
associated with excellent qualities of Japanese essays on three components (content.
organiution, and language use) across the trisks, the advanced-proficient group with good
qualities, and the intermediate Japanese-proficient group with fair o r good qualities.69
In terms of SOLO Ievels, distinctions were clear between the superior and the
intermediate/advanced Japanese-proficient groups. The writings (across the two tasks) of
the superior proficient students showed more complex structure, whereas the wntings of
the less proficient students appeared simplistic. Of special note, though, is Colin who
clustered into the intermediate Japanese-proficient group. Although the qualities of other
testual features of Colin's Japanese essays were seen to be about the same as those of other
intermediate students, his essays were judged to be as cornples as those of the superior
Japanese-proficient students in terms of cognitive structure mani fested in the tests.

69 In seelring to identify some patterns in che results of the Japancse essay assessrnenu, the Task II
essay scores of b l q , blartin, and Eliot were not taken into consideration. for those essays had previously
been correcteci based on native speakers' input before submission, as rnentioned in chapter 5.
\ :. -2,.
. lu-
.- - E.
 For Analysis 4(b), the percentage homogeneities of Solutions 1 and 2 were 37.95%
and 31.01% respectively. The distribution of the participants plotted in Figure 6-6 was
essentially the same as that in Figure 6-5. An esamination of major contributing items with

R ~ t of
l -5or higher for Analysis 4(b) ( s h o w in Tables 43 & 44 in Appendix Q) confims
the images emerging from the previous analyses. That is, this data set contains three
distinct clusters divided by the three proficiency levels and the three ethnicities. The
superior Japanese-proficient group consisting of ethnic Japanese students were inclined to
approach JFL writing as a meaning-making activity and to feel they had lemed about not
only language use but also content through their JFL writing. During the processes of
writing, idea genention or language use was never easy for them. The protocol data
indicated that they did devote a fair amount of cognitive energy to thinking about
language, gist, or both. Another characteristic of this group was the high quali ty of the

testual features and cognitive structure manifested in their Japanese essays.

The advanced and intermediate Japanese-proficient groups comprised pnmanly
English- and Chinese-speaking individuals, respectively. The two groups showed distinct

characteristics in their backgrounds and less distinct but unique features in their views of
and behaviors in JFL writing. The advanced Japanese-proficiency group was apparently
more confident in Japanese and approached JFL writing with less struggle than the

intermediate Japanese-proficiency group, who understood well and admitted openly their
lack of cornpetence in the Japanese language. Despite such differences. the qualities OC the
two groups' Japanese essays were almost i ndistinguishable. This fact brought the two
groups into close prosimity in Figures 6-5 and 6-6 compared to the previous Figures for
the analyses without essay assessrnent data. AI[ of the component scores of both groups
were in the fair or good range, and the cognitive structure manirested in their writing was
less comples than that of the supenor Japanese-proficiency group. Also. they shared the
viecv of JFL writing as a mere language esercise and of having leamed, if anything, about

the use of language, but not s o much about content. while they wrotc in Japanese.

a ) .
 Chapter 7: Conclusions

This chapter concludes the dissertation. In this final chapter t retum one more time
to the entire data set to corne to a better understanding of the nature and meaning of the

students' experiences of J R wri iing and learning in this one JFL classroom. At the same
time, 1 attempt to situate the study in a broader context of scholarly work on literacy,
biliteracy, and bilingual praessing. First, 1 describe the situational and individual factors
of rny study in terms of the "continua of biliteracy" (Homberger, 1989),70which provide a
system to define biliterate situations comprehensively, enabling meaningfui cornparisons
among different instances of them. Once my research context is defined in biliterate terms,
1 revisit the four sets of guiding questions pertaining to the main components of Biggs's 3P

mode1 (see section 1.4)--the teaching context, students' chancteristics, wnting processes,
and written products. I summarize and discuss my answers to the questions based on the

results of the analyses presented in the previous chapters. Finally 1 reflect on my research
overall and address my ovenrching question: What does wnting as a mode of leaming
mean in the contest of foreign language instruction and in relation to the proposition that
foreign language writing is more than just wnting a foreign language? Suggestions for
further research and implications for classroom instruction and curriculum design are also
provided.

70 See also Cumming (1994) for an adoption of Honiberger's framework as a comprchensive basis to
describe the wide scope of biiiterate issues addresscd in recent studies on U readng or writing.
 7.1. Understanding tlw Conîext of JFL Writing

Based on her extensive review of relevant literatures, Hornberger (1989) proposed

a series of continua that depict key aspects of biliteracy: micro-macro, oral-literate,
monolingual-bilingual , reception-production, oral language-writien language, L 1-L2
transfer, simultaneous-successive exposure, similar-dissimilar language structures, and
convergent-divergent scripts. These nine continua, seen as in timatel y related to each other,
provide a comprehensive framework in which a variety of situational and individual
factors relevant to bili tente performance can be systematically descri bed. 71 According to
Hornberger, any particular instance of biliteracy can be defined at certain points on these
continua. The contest of J K cvriting 1 studied is certainly one such instance, representing
one unique context that shares some common characteristics with other situations, but also
differs [rom others globally, in terms of bilitente practices.
This particular instance of bili teracy--the object of my study-happened at the
micro end of a micro-macro continuum where 1, as an observer of the phenomena, focused
on individual students' performance of JFL writing as well as cumcular events taking
place in that single classroorn. The curricular goal of this classrmm was the development
of reading and wri ting skills in modem Japanese, which p s i tions this contest at the li terate
end of an oral-literate continuum. Kei used Japanese (the students' target language)
exclusively as a medium of instruction, which may well place the classroorn context at the
monolingual end of a mono1ingual-bilingual continuum within this micro setting.
However, I emphasized the multilingualism of individual students i n this particular

71 The importance of such a conceptuai framework for understanding biliterate: 1contexts becomes
apparent if we recognize the cornplesity and diversity of biliteracy around the worid. Cumming (1994)
provided cornmon exampIes that cm be positioned at extreme ends of Honiberger's nine continua. 1 cite
some of his exunples to illustrate complexities inherent in studies of biliteracy: foreign tourists reading maps
at the micro end and the EngIish-language policy in Hong Kong at the macro end of a continuum; use of
Hebrew in Jewish religious pmctice at the literal end and traditions of folk tales at the oral end of a
continuum; and dominant use of English in many parts of North Arnerica at the monolingual end and
customary concurrent use of two or more languages in many parts of Aîrica at the bilingual end of a
continuum.
classroom; moreover, the macro-context, which 1 did not address, was one where Japanese
is not widely used but rather English predominates institutionally and in the society at
large. This JFL ciassroom's focus on the written language defines i t at the written-
language end of an oral-written language continuum. Since Kei concentrated on the
reading component more than the writing component in her teaching, the context was at a
point toward the receptive end of a receptive-productive continuum. In terms of L1-L3
transfer, the students apparently experienced variable situations or different points dong
this continuum, sorne experiencing more transfer across their languages than others, and
Chinese having different affinities to Japanese than English does.
With respect to the media of languages defined by the last three continua (Le.,
simulîaneous-successive exposure, similar-dissimilar language structures, and convergent-
divergent scripts), the contest differed from group to group. The ethnic Japanese or

superior Japanese-proficient people had had early exposure to oral and written Japanese.
whereas the other participants did not The latter group of students started to leam kpanese
well after they had acquired literacy skiils in their Lls. Thus, according to Homberger's

categories, the contest was at the simultaneous end of a continuum of simultaneous-

successive exposure to Japanese for the ethnic Japanese group; for the other ethnic groups,
it was at the successive end. Further, the context was at a point toward the convergent end

of a convergent-divergent scripts continuum for people with Chinese literacy skills because
of the orthographie elements shared by the Chinese and Japanese languages. But it was at
the divergent end for the people who used English as their written LI. On a continuum of
similar-dissimilar language structures, the contest should be placed at a dissimilar end for
al1 the participants since al1 the tanguages involved here (i.e., English. Chinese, and
However. the exact point may be closer
Japanese) belong to di fferent linguistic familie~.~z
to the sirnilar end for Chinese speakers than for English speakers because English is an

72 English is a Germrinic language. Japanese belongs to the i U h c language family while Chinese to
the Sino-TibetanIanguage family (Taylor &Taylor, 1995, p. 5; Fromkin & Raiman, 1983. chap. 9).
Indo-European language according to general typology, whereas both Japanese and
Chinese are different non-Indo-European languages.
In sum, the biliterate context of my study can be situated at a point of intersection
between the micro, the Ii terate, and the multilingual ends of Homberger's continua. The
constituent members of this context are characterized as king at a comrnon point near the
wntten. receptive ends but at different points dong the LI-L2 transfer continuum. In terms
of the media of languages involved, the context presented itself differently for each
subgroup. To al1 the students, the Japanese language was structura11y "dissimilar" to their
Lls. But these people's experiences differed on the continua of simuItaneous-successive
exposure and the convergent-divergent scripts. Figure 7- 1 shows these trends on the nine
continua. 1t should be noted that al though the continua are presented separately here, they
are, as pointed out earlier, inevitably and inextricably interrelated. Similady, any point on a
single continuum is related to al1 other points on the same continuum (Homberger, 1989).

monolingual bilingual

receptive produdive

oral written language

L1 L2 t ransfer

sirnuftaneous <-J C/A -> 1 successive exposure

sirnilar c C-J/A -> 1 dissimilar language structures

convergent C J/A-> ( divergent scripts

Figure 7-1. The present JFL writing contest as a biliterate instance situated on
Homberger's nine continua.
Note. The asterisk (*) indicates a point that applies COal1 the particip,mts; 'J" refers to the point applicable to
ethnic Japanese people of English as their L1; 'C" refers to the point applicable ro ethnic Chinese people
with Chinese literacy skills; and 'A" refers to the point shared by ethnic Anglo-saxon or native English-
speaking people. This representation focuses on common features, rather than idiosyncratic, individual
features. That is to Say, individual ciifferences within the ethnic Chinese group. for example. are not
represented. The Mandarin-Chinese-speakinghiike had Iitde ski11 in Chinese Iiteracy but established literacy
in English and thus should be positioned at a point near the divergent end of the last continuum. Such
individual features are not shown here. No point is indicated on the L 1-L2 transfer continuum as 1 was unable
to identify any common features on tbis continuum.

7.2. The Reseucit Quedons Revisited

7.2.1. Reseurcit Questions Perfaining lo the Teaclcing Context

How ivere fhe writitzg raskr siluafed in the JFL coicrse?

Hoiv did i k iirrriructor perceive the role of writing, the renclring of writing, and the
reachittg of JFL?

van Lier ( 19%) wrote:

In practice, one of the perennial problems a [language] teacher faces is

when to address the formal properties of language, and how to do this. Most
teachers have felt caught up in the periodic pendulum swings from
emphasis on form to emphasis on meaning, and back again (p. 202).

1 observed such tensions in Kei's teaching practices. Her intent was at first to shift

learners' attention from mechanical practice to meaningful practice of language. Her

primary concern was: How might s tudents gain greater cognitive benefi ts in her Japanese
literacy class? Olson (1991), in his attempt to explain intellectual advantages of literacy,
argued that writing offers people opportunitles to simultaneously reflect on the world that
discourse represents and on the language used to represent the world in the discourse. Such
reflective processes, according to Olson, play a key role in thought and developing one's
mind. Kei tried, not necessarily guided by esplicit ideologies about the nature of writing, to
construct a situation where learners might exploit the relations between these two leveis or
awareness to their cognitive advantages. She did this by requiring the statement of one's
opinions on selected topics for both of the two major writing tasks (on which I focused in
my research).

Behind this classroom scenario was Kei 's concern about the ps ychological
dimensions of her J E students attempting to become literate in Japanese. Her specific
question was: What linguistic knowledge would be required to perform li terate tasks in

Japanese? As the classes went dong, Kei became increasingly concemed, but remained
uncertain, about the kinds of linguistic knowledge that might facil itate the processes of
Japanese discourse production. Her lessons contained "planned and improvised elements"

(van Lier, 19%. p. 200) to focirs on larigunge, the term suggested by van Lier (p. 203) to

refer to a variety of ways to reflect on different dimensions of language (including form

and meaning) for various reasons. Kei approached literacy in JFL as "an individual

phenomenon" (McKay, 1993, p. In, attempting to establish a link between li teracy and
cognitive development. What she did not emphasize was the sociocultural dimension of
literacy, Le., social and culturai noms and conventions about the uses or Japanese written
texts.
Note, however, that Kei's teaching context described above constituted only a part
of each learner's unique context shaped by his or her particular needs and purposes.

Naturally, different students devoted their "perceptual, ... emotional, and cognitive
energies" (van Lier, 1996, p. 103)differently to JFL literate tassks in their personal lives,
dthough they were dl recipients of Kei's instruction in the same classroom.

7.2.2. Research Question Pertaining to Shrdents' Cluaracleristics

How did students' characteristics relate to their perceptiorrs of JFL ivrifiiig aizd learriiiig?

Students' personal characteristics and their perceptions of JFL wri ting and leaming
marked them into the following three distinct groups:
 The students who were highly proficient in Japanese were ethnic Japanese and
indicated their confidence in spoken Japanese but not in wntten Japanese. Two of these
4 students had recently stayed in Japan for 1 or 2 years whereas the other 2 had n o such

residential experience. The students regarded English as their L 1 but showed a la& of
confidence in writing in English. They seemed to approach J F L writing as a meaning-
making activity and to feel that they leamed about language and content through
engaging in J E writing tasks.

The students who had an advanced level (but not as high as the proficiency of the
ethnic Japanese group described above) of Japanese proficiency were native-born
anglophone Canadians. They al1 had recently stayed in Japan for 10 or 12 rnonths.
They were relatively confident in their Japanese and ais0 in their L1 writing. These
students tended to regard JFL writing primarily as an activity Tor language esercises
and to feel that their learning through JFL writing involved. if anything, language use.

The students who were intermediate Japanese-proficient were mostly Chinese

immigrant students with oral and literacy skills in Chinese. They showed little
confidence in their Japanese. Some of them indicated their confidence in their LI
writing, but others did not. Like the anglophone. advanced Japanese-proficient
students, this group of students mostly took J R . wn ting as an opportunity to practice
and learn to use the Japanese language without focusing on content.

7.2.3. Research Qicestions Pertaining to Students' Writing Processes

W t m iitigrcistic arrd cogriitive behaviors ivere iiwoked by JFL ivrititig aclivities?

How ivere tley relnted to stiide~iü'cl~nracteristicsniid imtrtictioilal cotife-rtvariables?

Different students performed different degrees of ideational, metalinguistic, and

procedural thinking while writing cssays in Japanese. I t was difficult to Sind, in relation to
the students' chünctenstics or tasks, distinguishable patterns in the students' concurrcnt
verbalizations of their thinking about different aspects of their writing. Nonetheless, there
were some i mportmt, general tendencies including: (a) that reflective thinking occurred
mostly while the writers thought about ideas, language use. or both (cf. L2 learners'
simultaneous thinking about gist and language use during the production of L3 writing
reported by Cumming, 1990, and Swain & Lapkin, 1993, (b) that such reflective thoughts
occurred more than hdf the time during actual writing, and (c) that language tuas distinctly
an object of reflection for many of the leamers writing in J E (cf. Olson's metalinguistic
hypothesis, 1991).
Funher, there were certain pattems in the students' responses to the pos t-wn ting
questionnaire (i.e., the students* retrospective thinking about their writing processes) in
Task 1. These pattems appeared to be best associated with the three groups distinctly
chaiactenzed by JFL proficiency level and other factors. While writing Japanese essays in
Task 1, which required the argumentative mode of wri ting (in contnst to the descriptive
nature of the Task I I assignment), the supenor Japanese-proficient students mostly
encountered difficulties both in idea generation and language use; they were generally
satisfied with their essays. The advanced Japanese-proficient students were also satîsfïed
wi th their essays, but unlike the supenor Japanese-proficiency group. they did not see idea
generation or language use particularly as a problem. The intermediate Japanese-proficient
studenrs were not so satisried with their essays, and they al1 expenenced difficulties in
languagc use. Some of these students also had difficulties in idea generation, but others did
not.
It was also possible to chancterize the students' responses to the post-writing
questionnaire in Task 1 in terms of the level of confidence they had in their L1 wnting or
JFL writing. It appeared that the students who were confident in their L1 wnting did not
see idea generation as a difficult problem (e.g., Jack, Colin, Martin. and Eliot), but those
who were not so confident in their L l writing did; similarly the students who indicated
relatively high confidence in JFL writing did not see [anguage use as a difficult problem
(e.g., Jasmine, Martin. and Eliot), but those who indicated a lack of confidence in JFL
writing did.

7.2.4. Researcli Questions Perfaining fo Students' Wriîten Products

tVha f textir al qualifies of JFL wririrrg nid srrtderirs rruzriifes~in rrrms oJ h g u a g e use.
corrrenk and orgarrization?
How ivns the cornplexiiy of cogriitive sîr~ictiirernatii/ested iri tfreir ivritterr fexîs?
How rvere rlzese efernerrts refated to strtdmts' cfraracteristics, irrstrrîctiorrnl corrtext
variables. and irr -process be haviors?

Testual qualities and the complesity of cognitive structure manifested in JFL

wnting were most closely associated with the students' levels of Japanese proficiency;
their relations to students' other personal characteristics, students' cognitive behaviors
while writing, or instructional contest variables seemed to be relatively thin. For teatual
qualities erpressed in terms of component ntings (which were then converted into the four
levels of poor, fair,good and e.rcellent), major distinctions were between the supenor
Japanese-proficient students who produced Japanese essays of excellent quality, the
advanced-proficient students who wrote Japanese essays of good quality, and the
intermediate Japanese-proficient students whose Japanese essays were assessed as fair or
good on al1 of the three components (content,organization. language use) in both tasks.
I n terms of SOLO levels, distinctions were clear only between the supenor
Japanese- prof icien t s tudents and the intermediate and advanced Japanese-proficien t
students. The former wrote Japanese tests showing more cornplex cognitive structure than
the latter. It appeared that a very high level of Japanese proficiency (as high as that of the
supenor Japanese-proficient students in this study who spoke Japanese like native
speakers) would be needed for J F L Iearners to carry out the sirnultaneous processing of
language and content to produce Japanese compositions w ith comples cognitive structue.
This may be generally true, considenng the possible effects of using a less proficient
language upon one's cognition or concepts of automaticity (as mentioned in earlier
chapten). However, the case of Colin, who was unique in this respect, seemed to suggest
another possible esplanation for effects of language proficiency upon cognition. As
mentioned belore, Colin's Japanese writing showed textual quali lies typical of
intermediate lapanese-proficieni students, but it manifested comple'r cognitive structure
(discussion on Colin follows in section 7.3).

7.3.Making Connections betweert the Znterpretive and DuaI-ScaCi,tg Analyses

The patterns and links that i observed in the data (particularly the data collected
from the students) and those that were established through dual scaling provide evidence
for the connections between categories depicted in Figure 7-3. Figure 7-2 gives a gnphic
summary of the salient features of the data 1 considered and their cor.nections. Solid lines
indicate distinct relations, whereas dotted lines show more tentative links. Multiple sources
of evidence and multiple dual-scaling analyses have shown particular connections between
the students' levels of JFL proficiency, the qualities of their JFL wri tten products, and their
perceptions of JFL writing and leaming.
What has prïmaril y emeged from the dual-scaling analyses are di fferent patterns
demonstrated by the native-like and the less-than-native-like Japanese-prolïcient people.
The native-like Japanese-proficient individuals produced essays of better quality in terms
of testual and cognitive stnictunl features, and they focused on both semantic and lexico-
grammatical components invoived in the production of their written discourse. In contrat,
the Japanese essays written by the less-than-native-like Japanese-proficient students were
clearly constnined in both textual and cognitive structurai features. and their primary focus
in the production of their written discourse was on the lexico-grammatical systcm. These
findings seem to confirm the distinctive stntegic and linguistic features of E writing, in
cornparison to L1 writing, that Silva (1993) pointed out based on a review of previous
studies of E3L writing.
 The relations between in-process thinking patterns and other categories of JFL
writing performance are inconclusive from my analyses, leaving much room for further
investigation. There was nonetheless an indication of some relation, so 1 have indicated
this by dotted lines in Figure 7-2. The category of "LI writing experience" is included in
Figure 7-2 as one varhble that might explain the characteristics of Colin. who digressed
from the i ntermediate Japanese-proficiency group and came closer to the superior
Japanese-proficiency group in terms of the cognitive stnicture mani fested in his wri tten
texts. His paying much attention to gist dunnp his production of the Task I essay and his
view of JFL writing as a meaning-making activity are also characteristics of the superior
Japanese-profïciency group. These interesting observations (i.e., the superior Japanese-

proficiency group's common traits being demonstrated by the intermediate Japanese-

proficient Colin), however, were not captured by the analyses I performed with dual
scaling. I t a p p r s that students' leveis of J R proficiency are not the sole esplanation for
their JFL writing performance (cf., Cumming, 1989).
In search of an alternative esplanatory link for Colin's situation, 1 focused my
attention on the interna1 characteristics of dl the raw data and found that Colin's interview
protocol differed distinctively from those of a11 the other students. His protocol displayed
his own unique and sophisticated view of writing, even a trace of which was hard to find in
the other students' interviews. Writing was Colin's daily routine. He wrote in Chinese, his
wr-itten LI, in quest of the meaning of Me; he would write down whatever thoughts he
might have about life and the world that day in order to be enlightened upon this
philosophical question. In the interview he repeatedly said, " Wri ting is power" and
presented his philosophy of writing at some length. Colin's wantitrg io wnte for perso~zai
reasons was an intrinsic form of motivation and perhaps one of the major contnbutors to
the deveiopment of "the cornpetence to produce a polished piece of writing" (Spaulding,
1993, p. 186). I t is not difficult to think that Colin could also make an esceptionally

interested and engaged writer in a foreign langage. For this reason 1 have added to Figure
7-2 the category of L1 writing experience, which was lost sight of in my processes of data
reduction. as a possible expianation for the idiosyncrasies Colin demonstnted.
One way of viewing the categories represented in Figure 7-3 is as conditions for
leaming (Spolsky, 1989).74 In this view these categories constitute tzecessary, fypical, or
gracied conditions.75 depending on the nature of the relationship between a particular
condition and a particular learning outcorne. However, further research is required to be
precise and clear about the nature of each condition and its relation to other conditions and
to the goals and outcomes of leaming. I t is also important to consider these matters through
an understanding of the configuration of various contextual factors for a particular
biliterate instance under consideration (as 1 did above in section 7.1, using Hornberger's
nine continua of bilitency).

M y ovenll research findings tend to define this particular instance or JFL writing

as "an individual shll" (McKay, 1993, p. 8) or in Street's (1984) terms "an autonomous
mode1 of literacy" (also cited by McKay. p. 8), in contrast to a sociohistorical or
ideological perspective. Among the major issues of this individual, ps ychological
- -

73 The content and amount of information collected on the individual parricipants' writing experienccs
during the interviews were inconsistent and hardly translatabIe into categories to be displayed in matrices.
Two pieces of information concerning the participants' L1 writing included in the matrices were self-ratings
of their Ll writing s k l l and the scores of their LI essays written for this research. The latter did not show
Colin to be any more outstanding than others; al1 students received the highest or the second highest range
scores. The self-rating of his writing skill as excellent seemed to be suggestive of his unique undersmding of
writing. But some other students dso rat& rheir LI writing sklls as e-zellenl. tbough their views of writing
were not comparable to Colin's. Thus, this item was not very useful in distinguishing Colin from others,
eithcr.
74 SpoIsky (1989) applicd in his development of a general theory of second Ianguage Iearning the idea
of a "preferencc modeIWCint proposed by Jackendoff (1983) for linguistic andysis. See also Elfis (1994, pp.
679-6%)) for a brief review of Spolsky's attempt.
75 Spolsky (1989) explained graded conditions as "the more something is tme, the more its
consequence is likely to occur" and typical (or typicality) conditions as llapply[ingJ typically but not
necessarily" (p. 12). In çontrast, necessary conditions arc the conditions without which certain outcomes will
not be produced.
perspective that 1 have highlighted were (a) definitions of literacy as a skill, and (b)
relationships between litency and cognitive development (McKay, L993). I want to
discuss these issues regarding my research findings.

Kei's orientation to J K witing and reading involved a view of literacy as a

language skill and a view of literacy as entailing some content knowledge. Most of her
students, o n the other hand, approached JFL writing tasks with a view that was

predominantly language-skill based. A definition of literacy as metalinguistic activity

(Olson, 1990, 1991) fi ts this content best. Olson ( 1990) described such meralinguistic
activity as "involv[ingJ the increasing awareness of the properties of language independent
of the things in the world chat the language is about" (p. ?O). Frorn this perspective, foreign
ianguage writing is indeed, cannot be more than, just rvnting a foreign language.
But we should remember that the JFL writers partkipating in my study were ail

highly educated and highly litente in their Lls. Consequently, as Olson ( 1990) contended,
these leamers attempting to write in JFL could have benefited tremendously [rom their
abilities to write and think in their Lls while they wrote in JET. This issue reflects

Cummins's (e.g., 1980, 1991) theories of the cross-linguistic interdependence of cogni tive
skills. Reiated to this theoreticai proposition. one central question prevailed throughout rny
research: How does one's proficiency in a foreign language interact with one's cognitive
engagement in idea and test processing during the production of written discourse in that
language (cf. Cumrning, 1989, 1990; Carson & Kuehn, 1994; Edelsky, 1986: Whalen &
Menard, 1995)?Some research has indicated that lack of proficiency in a foreign language

can prevent a writer [rom engaging in the generative and recursive processes of idea
production during the linear process of text production76 (see, for instance, Yau's study,

76 1 have adopted thc terms and concepts of "text production" and "idea production" found in Collins
and Gentnerfs(1980) cognitive description of the cornposing process:
It is important to separate idea production from text prduction. The processes involved in
producing text, whcther they operate on the word level. the sentence level. the paragraph
level. or the text Icvel. must produce a linear sequcnce that satisfies certain grammatid
d e s . In contrast, the result of the process of idea production is a set of ideas with many
intemal connections. only a few of which may fit the iinear mode1 desirable for text. (p. 53)
1987). I t is possible to esplain this phenomenon from the perspective of automaticity
theory: When the language is unfamiliar and little practiced, the task of text production
alone consumes cognitive resources leaving little room for idea production (Ringbom,
1987). Nonetheless, a writer like Colin, a mature writer in his L 1, seemed to suffer very
little. despite a lack of proficiency in Japanese, in engaging himself in idea production

while writing in Japanese. As Cummins's linguistic interdependence mode1 would suggest.

his writing skills were highly practiced and thus independent of his proficiency in the
target language, and could be applied in either language.
Another factor that may account for one's intensive engagement in language or
idea production processes is the writer's pnmary concerns. For instance, Jane was overly
concerned with language production because of her strong desire to be able to use better
Japanese; thus, she viewed JFL writing as an opponunity to practice using the language.
She was the one for whom the elementary processes of the target language had already
approached automaticity, but her attention left free [rom basic text processing was directed
to higher levels of text, rather than idea production processes. Similarly, less proficient
students may have perceived J F L writing as a language exercise and paid most attention to
the production of test that satisfies lexical and morphosyntactic rules because of their
primary concems about language use, leaving little cognitive and perceptual space for idea
production or rneaning mahng. Such cognitive processes and their consequences could
naturally be different (a) if a writer "did not take the content seriously but emphasized the
writing of a grammatically and syntactically correct piece of ivork" (Martin's written
comments in the Task II pst-writing questionnaire), or (b) if a wri ter tried to " think deeper
about the things [she] wanted to writew(Jasmine's wntten comments in the Task II post-
writing questionnaire) (cf. Uzawa & Cumming's, 1989, idea that students can "lower or

raise their standards" for writing in L7).

In view of these issues, what does writing as a mode of learning mean in the
contest of foreign language instruction and in relation to the proposition that foreign
language writing is more than just writing a Coreign language? 1 reflected on my research
contest with this question in mind. The instructor struggled to conceptualize and practice
JFL literacy as a unique tool for promoting leamers' control and capacities over the use of
language and thought genenlly (cf. Gere. 1985; see Ackerman, 1993. for a critical review
of the ivriting-to-learti concept). From leamen' perspectives, these two properties of
language and cognition appeared to exert mutual influence on each other in some cases,
but they seemed to esercise minimal influence on each other in other cases (cf. Williams &
Snipper, 1990). The data in my study have suggested that the former cases would require
native-like proficiency in L2 (cf. Berman, 1994; Cumming, 1989; Pennington & So. 1993;
Sasaki & Hirose. 19%), or a meaning-based conceptudization of writing. In short, as far
as my research findings are concemed, foreign language writing can be more than just
wri ting a foreign language, offering opportuniries to promote language and thinhng skills;
but for them to be so depends on leamers' proficiency in the tarpt language as well as
leamers' views and purposes of wri ting. 1t remains unclear, though, w hat influences one
teacher's activities might have on such inherently personal acts and conceptualizations of
writing that students adopt independently.
My study suggested a number of ideas relevant to foreign language writing that are
worth Iùrther pursuing for refinement. The lollowing issues seem particularly important to
understand better if we are to advance our efforts in developing relatively comprehensive
and educationally meaningful models of L2 writing: (a) L? writing as a metalinguistic
activi ty; (b) LZ w n ting as a personal construct: (c) effects of L2 proficiency, language
distance, and contest upon L l -L2 tnnsfer; and (d) the appropriate or possible balance of
foci in instruction.
1 associated JFL writing in my research context with the concept of litency as

metalinguistic activi ty, based on considerations of the students' think-aloud and interview
protocol data, which featured their concems about language. But 1 did not make any
conceptual or operational distinctions between linguistic proficiency (the abili ty to use a
specific language effectively) and metalinguistic ability (the ability to reflect o n the
language used). Future research that might compare these two h n d s of ability in an L2,
then relate them to the leaming of the language and of content expressed in wtiting using
that language, should not only help to illuminate the relationships between L2 writing and
leaming, but also to clarify the concept of writing as a metalinguistic activity in L7
contests.
Distinct characteristics shown by such participants as Colin, Jane, and Martin led
me to view JFL writing in my study as a personal construct, reminding m e of Bell's (1995)
autobiographical study of learning to write in Chinese. Bell, a native speaker of English,
described her becoming l i t e n t e in Chinese as an emotionally charged process of
transforming her own consciousness. L? students enter their litency classrooms with their
own assumptions, purposes, needs, and preferences in regard to literacy acquisition in a
new language. This is particularly true for adult learners who have had previous
experiences of leaming different foreign languages, or who began learning to read and
wnte in a target language after establishine their litency skills in their mother tongues. In-
depth case srudies that Cocus on students' persona1 espenences and that attempt to derive
m e a n i n g of L3 writing in personal o r esperiential terms will enhance our understanding
about why particular students deal with L3 writing tasks in certain ways (e-B.,
Srnagorinsky, 1997).
Whenever two languages are involved in discourse activity, lnngttnge trarisJer or
cross-lirlguistici@rence (Gass & Selinker, 1993; Kellerman & Shanvood Smith, L986;
Odlin, 1989) becomes an issue. This problem loorned large in my study (and 1 touched on
it above in citing Cummins's mode1 of cross-linguistic interdependence). Although the

transferabili ty of linguis tic and li teracy knowledge between languages has been well
supported empirically (as pointed out by Cumming. 1994, and Bell, 1995; but also note
Bell's cri tical view of this idea), a few key areas need further investigation to renne this

notion. The linguistic interdependence hypothesis and most other ideas about the benefits
of bilingualism assume that education or other societal contexts might provide sufficient
opportunities to enable L2 learners to become fluent users of their L2s (Mayer & Wells,
1996). This is seldom the case in a foreign language context, however. In my study, this
assumption was met only by the few ethnie Japanese students who grew up bilingually.
They ivere fluent speakers of Japanese. Although they were not very confident in their
writing, their Japanese writing was superior in terms of textual features and cognitive
structure manifested in their texts. On the other hand, the so-called "successive exposure"
learners with limited opportunities to use the Japanese language were less fluent in
Japanese and had rnany fewer opportunities to use the language for communication and
purposeful interactions. Although the amount of their Iapanese language input as well as
the level of their Japanese language proficiency differed markedly from individual to
individual, the qualities of these people's Japanese writing were more or less equally
constrained, except for the cognitive structure of Japanese essays wntten by Colin. the
intermediate Japanese-proficient but competent L 1 writer. Based on these observations, we
might ask: What is the threshold level of L2 proficiency required for average writers to
produce L2 essays that manifest cognitive complexity? This is a major issue in need of
future research.
Another aspect of the interdependence hypothesis that could benefit frorn
refinement concerns the effects of "language distance" (Odlin, 1989, p. 32; see also
Corder, 1978) upon bilingual transfer in literate tasks (as rnentioned by Cumming, 1994,
and Cummins, 1991). My research findings have little concrete to say on this because the
variable of language distance was confounded with the variable of L2 proficiency in the
classroom context. However, the context of bilingualism 1 studied differed markedly from
the contexts where most of the research that supports the notion of transferability of
language skills has previously been conducted, involving two Indo-European languages;
for instance, English-speaking children leaming French in immersion programs in Canada
(Swain & Lapkin, 1982). Spanish-speaking children leaming English in bilingual programs
in the Southwestern United States (Edelsky, 1986), and French-speahng adults learning
English in Ontario (Cumming, 1989). Virtually dl of these studies took place in secotzd-
latzgtîage, rather than foreign-laquage, contexts where learning is associated with
"adequate esposure to [the target language] ... and adequate [integrative] motivation to
leam [that language]" (Cummins, 1981, p. 79). This was, only to variable extents, the case
for the learners 1 studied, rvho had limited routine contacts with Japanese and were not
attempting to i n t e p t e themselves in the society of Japan per se. Future, carefully designed
research o n leamers of non-cognate languages in foreign language settings may provide
illuminating insights inro this issue and i ts variable effects o n writing in L2 (cf. Uzawa &
Cumrning, 1989).
The final point 1 want to rnake concems instruction. Kei is now (summer 1997)
preparing for the çame course for the coming academic year as 1 observed for my research
three years ago. She has k e n thinking hard about how to present balanced lessons in such
bipolar relational terms as form and meaning, part and whole, and process and product. 1
would point her (and l i ke-minded teachers) to a number of innovative instructional
approaches to literacy that seemingly have produced impressive results (see Cumming,
1994. p. 9). For esamples, Valerie Anderson's (Anderson, 1995; Anderson & Henne.
1993) Adolescent Literacy Project (ALP) cornes to mind. The core of the ALP lies in an

instructional plan to improve the English literacy of culiunll y and ethnically diverse,
inner-city high school students with literacy skills trvo o r more years below their grade
levels, by helping them to relate reading, writing, and purposeful i n q u i q . I t is based on a
number of learning theones such a s problem-centered leaming, strategic thinking,
intentional leaming , col laborative leami ng, and process wri ting. Successful work of a
similar nature in L2 settings was reviewed by Elley (1994). Studies evaluating the effects
of instruction based o n such innovations are needed; s o is documentation of teachers'
improvisation processes in teaching, in order to understand esactly how theones believed
to enhance leaming actually work to affect students*learningn and cumcular processes.
As Cumming (1994) pointed out, such research "remains remarkably sparse" (p. 9) in

second-language, let alone foreign-language, litency education.

A few years ago, 1 (So, 1994b)discussed major advantages and disadvantags of

case study approaches in educational research. The advantages included (a) the

compieteness and depth of analysis, (b) the hypothesis-generating nature of studies, and (c)
the role of cases that challenge or contradict current beliefs. The disadvantages included
(d) lack of generalizability, (e) lack of objectivity, and (0 a data-dnven, iather than theory-
driven, approach. My thesis research reported here embraced al1 these features, which
together contributed to the way 1 now understand the relations of JFL wnting and learning
as well as the advantages and limitations of case study research.

- - -- --
- - - - -

77 Onc implication from my research concerning the mcasurcment of students' leanting is that
different schemes may assess differcnt aspects of lcruning quafitics (see the multi-trait rating scheme for L2
wri ting and the SOLO a d ysis used io evaluate the same pieces of students' Japmesc cssays).
 References

Ackerman, J. M. ( 1993). The promise of wri ting to leam. WrifferzCo»miunicatio~i.10,

334-370.

Adair-Hauck, B. ( 1996).Prac tical whole language strategies for secondary and university-
level R. students. Foreigtr Langiîage Arzak, 29.33-270.
Adams, M. J., & Collins, A. M. (1979).A schema-theoretic view of reading. In R. O.
Freddle (Ed.),New directions iti discourse processing (pp. 1-22). Norwood, NJ:
Ables.

P., Frohlich, M., & Spada, N. (1984).The communicative orientation of language

teaching: An obsentation scheme. In J. Handscombe, R. A. Orem, & B. P. Taylor
(Eds.), On TESOL '83: The qiiestiorr of conaol (pp. 13 1 - 2 2 ) . Washington, DC:
TESOL.
Allwri ght, D. ( 1988).Observatiori iri the larigliage clmsroom. London: Longman.
American Counci 1 on the Teaching of Foreign Languages. ( 1987). ACTFL Japanese
Proficiency Guideli nes. Foreign Langiiage A>inaLs, 20,589-a.
Anderson, V. ( 1995, Novernber). Teacher deveiopment for improving adolescent litency
through collaborative strategy instruction. Paper presented at the Centre for
Applied Cognitive Science Colloquium, Ontario Institute for Studies in Education
of the University of Toronto.
Anderson, V ., & Henne, R. ( 1993,December). Collaborative, integrated reading and
writing strategy instruction. Paper presented at the annual meeting of the National
Reading Conference, Charlston, South Carolina.
Arndt, V. (1987). Six writers in search of texts: A protocol-based study of Ll and L1
wri ting. ELT Jorir~ral,41,757-267.

Ashworth. M. (1985). Beyond rnetkodology: Second langimge teaching and the

commu~iiiy. Cambridge: Cambridge University Press.
Bell, J. S. (1995).The relationship between L1 and L2 litency: Some complicating factors.
.
TESOL Quarterly 29.68'7-704.
Benjamin, A. ( 1974). The Iielping irirerview (2nd ed.). Boston, MA: Houghton Mifflin
Company.
Berei ter, C.,& Scardarnalia, M. ( 1987). The psycliology of writteir corriposition. Hillsdale,
NJ: h v r e n c e Erlbaum.
Bereiter, C., & Scardamaiia, M. ( 1989). Intentional learning as a goal of instruction. In L.
B. Resnick (Ed.), Kno wing. learning. and ins~uction:Essays in honor of Robert
Glaser (pp. 36 1-39?). Hillsdale, NJ: Lawrence Er1baurn.
Berman, R. ( 1994). Leamers' transfer of writing shlls between l a q u a g e s . TESL Catiada
JoztrrtallRevue TESL dit Canada, 12,39-46-
Biges. J. (1987). Process and outcome in essay writing. In A. H. Miller & G. Sachse-
Akerlind ( U s . ) , Researcli and development in higher education (Vol. 9, pp. 1 1 4
125). Sydney: Higher Education Research and Development Society of A ustralasia.

Biggs, J. B. ( 1991). Enhancing learning in the contest of school. In R. F. Mulcahy, R. H.

Short. & J. Andrews (Eds.), Enhaticitzg learnitzg and thitikitrg ( p p . 35-52). New
York: Praeger.

Biggs, J. B., & Collis, K. F. (1982). Evalriatirig the quality of leortiing: The SOLO
Tuxunoniy (Structure of the Observed Lear~zirtgOutcorne). New York: Academic
Press.
B iggs, J. B., & Moore, P. J. ( 1993). The process o j lenrrzing. (3rd ed.). Sydney: Prentice
Hall.

Breen, M. P. (1984). Process syllabuses for the language classroom. In C. J. Brumfit (Ed.),
General Engfish sylïabirs design: Ciirricslutn and syllabiis desigti for the general
Etzglisil classrootn (pp. 47-60).London: British Counci 1.
Bridwell, L. S., & Beach, R. (1984). Introduction. In R. Beach & L. S. Bridwell (Eds.),
New directiotrs in cornpositiotr resrarch (pp. 1 - 14). New York: The Guilford Press.
Brinton, D. M., Snow. M. A., & Wesche, M. B. ( 1989). Cotzterit-based secotxd la~zgirage
imtricction. New York: New bury House.
Brock, C. A. ( 1986). The effects of referential questions on ESL classroom discourse.
TESOL Qrtarterly. 20,47-58.
Brown, J. D. (1991). Do English and ESL faculties rate writing samples differently?
TESOL Qirarterly. 25,583-603.
Cantoni- Harve y, G. ( 1987). Content-area langiiage imtritctiotr : Approaches and
strntegies. Reading, M A : Addison-Wesley.
Carlson. S. B. (1991). Prograrn evaluation procedures: Reporting the program publicly
wi thin the poli tical context. In L. Hamp-Lyons (Ed.),Assessitig second latigrtage
writirzg iti academic corrtexts (pp. 293-320). Nonvood, NJ: Ablex.
Camine, D., & Kameenui, E. J. (Eds.). (1992). Higlier order thir>kitcg: Designitzg
ciirricz~lrunfor mzittslre~rnedstride~~ls.
A ustin, TX: pro-ed.

Carrell, P. L., Devine, J., & Eskey, D. E. (Eds.). ( 1988). ItiIeractive approaches ta secoiid
latzgziage readitig . Cambridge: Cambridge Uni versi ty Press.
Carson, J. E., & Kuehn, P. A. (1994). Evidence of transfer and loss i n developing second-
language wri ters. 1n A. Cumming (Ed.), Bilingiial perfortnance bi readirlg niid
ivritircg ( p p . 57-38 1). Amsterdam: John Benjamins.
Center for Applied Linguistics. ( 1992). Japarzese Speaking Test. Washington, DC: Center
for Applied Linguistics.
Chafe, W. L. ( 1980). The deployment of consciousness in the production of a nanative. In
W. L. Chafe (M.), The pear stories: Cognitive. cuftiîrd, and linguistic aspects of
rzan-ative proditction (pp. 9-50). Nonvood, NJ: A blex.
Chaudron, C. ( 1988).Second fanguage clussroor~r~:
Research or1 teacliing and learrzing.
Cambridge: Cambridge University Press.
Clandi ni n, D. J. ( 1986). Classroom pracfice: Teaclier irnages in acfion. London: The
Falmer Press.
Clandinin, D.J., & Connelly, F. M. (1991). Nanative and story in practice and research. In
D. Schon (Ed.), The rejlective tzirn: Case sticdies of reflective practice (pp. 358-
28 1). New York: Teachers Colleg Press.
Cole, M. ( 19%). Cirlhîral psychology: A orrce alrd jirtrtre discipline. Cambridge, M A : The
Belknap Press of Harvard University Press.
Collins, A. M., & Gentner, D. (1980). A framework for a cognitive theory of writing. In L.
W. Gregg & E. Steinberg (Eds.), Cognitive processes irr writirrg (pp. 51-72).
Hillsdale, NJ: Lawrence Erlbaum.
Connelly, F. M., & Clandinin, D. J. (1988). Teacliers as curriciifum plantzers: Narratives
of experietice. New York: Teachers College Press.
Connor, U. ( 1987)- Research frontiers in wri ting analysis. TESOL Qiiarferfv,21.677-696.
Connor-Linton, J. ( 1995). Crosscult u r d cornparison of wri ting standards: Amencan ESL
and Japanese EFL. World ErzgIishes, 14,99-115.
Connor-Linton, J., & Haichour, E. H. (1997, March). The relation between linguistics
cornpetence and writing proficiency: ~videncefrom language episodec Paper
presented at the Annual Conference of the Amencan Association of Applied
Linguistics, Orlando, Flonda.
Corder, S. P. (1978).Language distance and the magnitude of the language leaming task.
SZirdies iri Secorid L m g zrage Acqtt isirio~i,2.27-36.

Crandall, J. (1993). Content-centered leaming in the United States. Aanual Revieiv O!

Applied Linguistics, 13, 1 1 1- 126.
Crandall, J., & Tucker, G. R. (1990). Content-based language instruction in second and
foreign languages. In S. Anivan (Ed.), Larigriage tencliing rne~iiodologyfor the
rzineties (pp. 83-%). Singapore: SEAMEO Regionai Language Centre.
Crawford-Lange, L. M.. & Lange, D. L. (1987). Integrating language and culture: How to
do i t. Theory I~rtoPractice, 26,258-266.

Cumming, A. ( 1988). FVritit~gexpertise and second Inriguage proJciency in ESL writing

performarrce. Unpubl ished doctoral dissertation, University of Toronto.
Cumming, A. (1989). Writing expertise and second-language proficiency. Larrgriage
Learrzirrg, 39.81-141.
Cumming, A. ( 1990). Metalinguis tic and ideational thi nki ng in second language
c o r n p i n g . Writterr Cornmtiriication, 7,482-5 11.
Curnming, A. (1994). Introduction: Recent trends in research on bilitency. In A. Cumming
(Ed.), Bilingrial performance iri readirig arid writirig (pp. 1-20). Amsterdam: John
Benjamins.
Cumming, A., & Riazi, A. (1997, in press). Building m d e l s of adult second-language
writing instruction. T o appear in Learriing arid Illstructiorl .
Cummins, J. ( 1980). The cross-lingual dimensions of language proficiency: Implications
for bilingual education and the optimal age issue. TESOL Qicarterfy, 14, 175- LW.
Cummins, J. (1981). The role of prirnary language development in promoting educational
success for language minoriiy students. In Califomia State Department of
Education (Ed.), Schoolirig and Iangtiage rnirtority students: A theoretical
frarnework ( p p . 3-49). Los Angeles: Evaluation, Dissemination and Assessment
Center, Califomia State University.
Cummins, J. ( 1991). Interdependence of first- and second-language proficiency in
bilingual children. In E. B ial ystok (Ed.), Language processing in bilingual children.
Cambridge: Cam bridge University Press.
Cummins, J., & Swain, M. ( 1986).Bilirtgualism irt edricufiott:Aspecls of rlzeory. researclz
m d prncrict.. London: Long man.
Davis, J. N. ( 199'1). Educationai reform and the Babel (babble) of culture: Prospects for the
Statziiards for Foreign Lattgziage LParriiqg. The Modenz Lorigrrage Jotrrnal, 81,
15 1- 163.

Dewey, J. ( lB8/ l%3). Erperierice & edrtcafiort. New York: Macmillan.

Dey, 1. ( 1993). Qualitative da fa analysis: A user-friendfy guide for social scierifists.
London: Routledge.
Dickson, P., & Cumming, A. (Eds.). ( 1996). ProJFles of inrtguage ediicatiorr iri 25
coiortries: Overview of phase 1 of the IEA lariguage educntion srudy. Slough,
England: National Foundation for Educational Research.
Duniun, M. J., & Biddle, B. J. (1974). The study of tencizirrg. New York: Holt, Rinehart
and W i nston.

Durst, R. K. (1989). Monitoring processes in analytic and summary wnting. Writtett

Cornrnrirricatiott,6,340-363.
Edelsky, C. ( 1986). Writirrg irz a bilirlgrial prograrn: Hab i a rirra vez. Nonvood, NJ: Ablex.

Eisner, E. W. ( 199 1). The eriligl~tetzedeye: Qzralilative irtqriiry artd the enharrcernent of
educa~iorrnlpractice . New York: Macmi Ilan.
Elley, W. B. (1994). Acquiring literacy in a second language: The effect of book-based
p r o g m s . 1n A. Cumming (Ed.), Biiiiigrml performarzce in reading atid writiq (pp.
33 1-366). Amsterdam: John Benjamins.

Ellis, R. (1994). The stridy oJsecond la~tggrrageacquisifion Oxford: Oxford University

Press.
Emig, J. ( 1977). Wn ting as a mode of learning. Cofiege Composition and Cormiorrication.
28, 122- 127.

Ericsson, K. A., & Simon, H. A. (1993). Protocol arialysis: Verbal reports as dara (2nd
ed.). Cambridge, MA: The MIT Press.
Fanselow, J. F. (1990). "Let's seen: C~ntrastingconversations about ieaching. In J. C.
Richards & D. Nunan (Eds.), Secorid lariguage teacher edrccariorz (pp. 182- 197).
Cam bridge: Cam bridge University Press.
Rower, L. (1990a). Introduction: Studying cognition in context. In L. Rower, V. Stein, J.
Ackerman, M. J. Kantz, K. McCormick, & W. C. Peck, Readitrg-to-write:
Erploritzg a cogtriiive atrd social process (pp. 3-32) . New York: Oxford University
Press.
Flower, L. (1990b). Preface. In L. Flower, V. Stein, J. Ackerman, M. J. Kantz, K.
McComick, & W. C. Peck, Readitzg-tu-writing: Exploritig a cognitive and social
process (pp. v-vi). New York: Oxford University Press.
Rower, L. (1990~).The role of task representation in reading-to-write. In L. Flower, V.
Stein, J. Ackerman, M. J. Kantz, K. McCormick, & W. C. Peck, Reading to rvriie:
Expkwirzg a cognitive & social process (pp. 3575).New York: Oxford Universi ty
Press.
Flower, L. ( 1993). Probiern-solvirig strategies for writittg (4th ed.). Phi ladel phia: Harcourt
Brace Jovanovich College.

Freed, B. F. ( 1978). Foreigner talk: A study of speech adj~isirnentsrnnde by native

speakers of E q fislr iti coiiversatiotr wifh rion-native speakers. U npu bl i shed
doctoral dissertation, University of Pennsylvania, Philadelphia.
Freire, P. ( 1973). Edldcation for critical cotisciorisness ( M . B. Ramos,Tram.). New York:
The Seabury Press.
Fromkin, V., & Rodman, R. ( 1983). k t introdrictiorz to lariguagr (3rd ed.). New York:
Holt, Rinehart and Winston.
Gass, S. M., & Magnan, S. S. (1993). Second-language production: SLA research in
spealÜ ng and w ri ting. In A. Ornaggio Hadley (Ed.), Researclr irr langroage leurnirzg:
Principles, processes. ami prospecfs (pp. 156- 197). Li ncolnwood, IL: National
Textbook Company.
Gass, S. M.. & Selinker, L. (Eds.). ( 1993). Lorrgi<age tramfer in Infigrrage learnirig .
Amsterdam: John Benjamins.
Gere, A. ( 1985). Roois in sawdust: Writirtg to leariz across the ciirriculum. Urbana, IL:
National Council of Teachers of English.
Glaser, B. G. (1978). Theoretical seitsirivity. Mill Valley, CA: Sociology Press.
Glaser, B., & Strauss, A. L. (1967). The discovery of groititded zlreory: Strazegies for
qiialiïative research. New York: Adeline.
Glaser, R. (19%). Education and thinking: The role of knowledge. American Psycliologist.
39,93- 104.
Grabe, W., & Kaplan, R. B. (1996).Theory nird proctice of writiizg: Ar applied lirrgiiistic
perspective. London: Longman.
Hamp-Lyons, L. ( 1991a). Reconstructing "academic writing proficiency". In L. Hamp-
Lyons (Ed.), Arsessing secoond langiiage ivrititig irt acndernic corztexls (pp. 127-
153).Nonvood, NJ: Ablex.
Hamp-Lyons, L. ( 199L b). Sconng procedures for ESL contests. In L. Hamp-Lyons (Ed.),
AssessBig secorid larzgtiage writirig iri acadernic contexts (p p. 24 1-276). N o w d ,
NJ: Ablex.
Hamp-Lyons, L. ( 19%). Rating nonnative wnting: The trouble wi th holistic scoring.
TESOL Qitarter[y, 29,759-762.
Hamp-Lyons, L., & Henning, G. ( 199 1). Communicative wri ting profiles: An investigation
of the transferabi li ty of a mu1ti ple-trait scoring instrument across ESL wri ting
assessment contests. Langiiage Lear~lirig,41,337-373.
Harrington, M. (1986). The T-unit as a measure of JSL oral proficiency. Descriptive aiid
Applied Lhgaistics, 19,49-56.
Hirose, K., & Sasaki, M. (1994). Explanatory variables for Japanese students' expositoq
w n ting in English: An exploratory study. Jorrnial of Second Lnr~ggnngekvriting, 3,
303-229.
Hornberger, N. H. ( 1989). Continua of bilitency. Review of Edilcntionnl Resenrck, 59,
37 1-296.

Hunt, K. W. ( 1965). Grarninatical structures written al tlcree grade le vels. Urbana, 1L:
National Council of Teachers of English.
Hunt, K. W. ( 1970). Syiztactic mafzcrity NI sclrool-childreii aizd adults ( V 01. 35). Chicago:
University of Chicago Press.
Jackendoff, R. (1983). Setnailtics w d cogtiition. Cambridge, MA: The MIT Press.
Jackson, P. W. (1992). Conceptions of cumculurn and curriculum specialists. In P. W.
Jackson (Ed.), Hmidbook of research ON ciirriciiltirn (pp. 3-40). New York:
Macmillan.
Jacobs, H. L., Zinkgraf, S. A., Wormuth, D. R., Hartfiel, V. F., & Hughey., J. B. (1981).
Testiizg ESL corriposition: A prnctical approacir. Row ley, MA : New bury House.
Johnson, R. K. (1989). A decision-making framework for the coherent language
cumculum. In R. K. Johnson (Ed.), n i e second langilage ci~rricnfurn(pp. 1-23).
Cambridge: Cambridge University Press.
Jorden, E. H., & Lambert, R. D. (1991). Juparzese latigsage imtruction iri the Uriited
States: Resources, practice, atid iri vesirnenr strategy. Washington, DC: National
Foreign Language Center.
Kanagy, R. (1995, July). Bibliognphy of research o n the acquisition of Japanese as a
secondlforeign language. Available HTTP: Hostnarne: uoregon.edu Di rec tory:
http://darkwing.uoregon.edu/-kanagy/jslarhmp.html
Kato, Y. (1995). Differences in cognitive process between slcilled and unskilled writers.
The Joiïrnol of Japaiiese Latigirage Eriiication Methoh, 2,4647.
Kel lerman, E., & Shanvood Smith, M. ( 1986). Crosslinguistic iri/luences irz second
language acquisition. New York: Pergarnon.
Mare, G. R. (1974).Assessing readability. Reading Research Qicarierly, 10,61-103-
Klein, W. ( 1986). Second latiguage acqiisition. Cambridge: Cambridge University Press.
Kobayashi, H., & Rinnert, C. ( 1994). Effects of first language on second language wri t i n g
Translation versus direct corn posi tion. 1n A. Cumming (Ed.), Bilirigiîal
perfurrnatice itz reading and ivritirig (pp. 2Z-255). Amsterdam: John Benjarni ns.
Koda, K. ( 1993). Task-induced vanabili ty in FL composition: Language-specific
perspectives. Foreigrt Lntiguage Aririals, 26,332-346.
Kroll, B. ( 1990). Introduction. I n B. Kroll (Ed.), Secotid Lmzgitage Writit~g:Research
irzsightsfor the classrootn (pp. 1-5). Cam bridge: Cam bridge Uni versi ty Press.
Kubota, R. ( 1996). Cntical pedagogy and cntical literacy in teaching Japanese. Japatzese-
Laiigriage Editcatioti aroiotzd the Globe, 6,354.
LaBerge, D., & Samuels, S. J. (1974). Toward a theory of automatic information
processing in reading. Cogriitive Psydiolugy. 6,393-323.
Lange, D. L. ( 1990). Sketching the cnsis and explonng different perspectives in foreign
language curriculum. In D. W. £3i rckbichler (Ed.), New perspectives atid tieiv
directiorzs iri foreigti latiguage edrocation (pp. 77-109). Li ncol nwood, IL: National
Textbook Company.
Lange, D. L. (1992). Foreign language education. In M. C. A l h n (Ed.), Etzcyclopedia of
edirca~iottafresearcli (6th edi tion ed., pp. 531-530). New York: Macmillan.
Langer, J. A. (1986a). Cliildren reading atid writitig: Strrictiires and strategies. Norwood,
NJ: Ables.
Langer, J. A. (1986b). Reading, writing, and understanding: An analysis of the
construction of meaning. Writteti Comrnioiication, 3 , 319-267.
Leaver, B. L., & Stryker, S. B. (1989). Content-based instruction in foreign language
classroorns. Foreign Lurigringe A~tials,22,269-275.
Lesgold, A. M., & Perfetti, C. A. ( 1978). Interactive processes in reading comprehension.
Discoirrse Processes, 1 ,323-336.
Littlejohn, A., & Windeatt, S. (1989).Beyond language learning: Perspectives on materials
design. In R. K. Johnson (Ed.), rite second language ciirricrilrim ( p p . 155- 175).
Cam bridge: Cam bridge University Press.
Loban, W. D. ( 1%3). The langiiage of elernenfary school cliildren . Urbana, IL: National
Council of Teachers of Engiish.
Marshall, J. D. ( 19â7). The effecis of writing on students' understanding of literary texts.
Research iri rlie Tenchirig of Erzg lish. 21,30-63.
Mayer, C., & Wells, G. (1996). Can the linguistic interdependence theory support a
bilingual-bicul tunl mode1 of literacy education for deaf students? Joiirrial of Deaf
S~iidiesand Deaf Ediicatiori. 1.93- 107.
McKay, S. L. ( 1993). Agrridas for secorid lariguage literacy . Cam bridge: Cam bridge
University Press.
McLaughlin, B. ( 1981). Themies of secmd-lnnguage learrzirig . London: Edward Arnold.
Met, M., & Galloway, V. (1997). Research in foreign language cumculum. In P. W.
Jackson (Ed.), Haridbook of researclr ort cicrricrrlutn: A project of lhe rirnericna
Educatiomzl Research Associafion (pp. 85?-8%). New York: Macmi 1lan.
Miller, J. P. ( 1988). The holistic czirric~ilrrm.Toronto: OISE Press.

Miller, J. P. ( 1993).The holistic teacher. Toronto: OISE Press.

Miller, J. P., Cassie, J. R. B., & Drake, S. M. (1990). Holisric learriirtg: A teaciier's guide
fo iritegrnred strrdies. Toronto: OISE Press.
Minick, N., Stone, C. A., & Forman, E. A. (1993). Integntion of individuai, social, and
institutional processes in accounts of children's learning and development. In E. A.
Forman, N. Minick. & C. A. Stone (Eds.), Corifexfifor leartiitzg: Sociocrilt~crnl
dytiamics il1 cllildrenlr developt?ient (pp. 3-16). New York: Oxford University
Press.
Mohan, B. A. (1986). Lnrrgiiage alid coritent. Reading, M A : Addison-Wesley.
Moskowitz, G. ( 1978). Caring and sltaritig in the foreign larigiiage class: A sozircebook on
hirmariistic ~echniques.Row ley, MA: New bury House.
Newell, G. E. ( 1994). The effects of written between-draft responses on students' writing
and reasoning about literature. Writteri Cornrnririicatiori. 11, 3 1 1-347.
Newman, D., Griffin, P., & Cole, M. ( 1989).The cortsfrrictioti :otre: Working /or cognitive
chaitge iri scliool. New York: Cambridge University Press.
Nishisato, S. ( 1994). Efernents of dual scalirig: Ali introducriori to procrical data arrafysis.
Hilisdale, NJ: h v r e n c e Erlbaum.
Nishisato, S. ( 1996). Gleaning in the field of dual scaling. Psyclmelrika, 61,559-599.
Nishisato, S., & Nishisato, 1. ( 1984). An iintrodiictioti to dual scalirzg . Toronto: MicroStats.

Nishisato, S., & Nishisato, 1. ( 1994).Dial scalitig iti a riiifslzell. Toronto: Microstats.
Nunan, D. (1988). The learrier-ceriired crirricrtliirn. Cambridge: Cambridge University
Press.

Odl i n, T. ( 1989). Lang rmge Trarisfer:Cms-liriguistic iriflrtence iti larzg linge leaniitig .
Cambridge: Cambridge University Press.
Olson, D. R. (1990). Contribution to the Forum: When a learner attempts to become
litente in a second language, what is she or he attempting? TESL Talk, 20, 18-20.

Olson, D. R. (1991). Literacy as metalinguistic activity. In D. R. Olson & N. Torrance

(Eds.), Lilerncy and orcdity ( p p . 25 1-270). Cam bridge: Cambridge Uni versi ty
Press.

Olson, D. R. ( 1994). The world or1 paper: The coricepaal atid cognitive itnplications of
writirig atid readittg . Cambridge: Cam bridge University Press.
Ontario Institute for Studies in Education. (1993). OISE griidelities for theses ntid orais.
Toronto: A uthors.
Ootake, H. (1994). Think-aloud protocol analysis of Japanese language leamers' writing
processes and strategies. Nikorigo Kyoirikri [Journal of Japanese Language
Teaching], 82, 193.

Oxford, R. L. ( 1990). Larigrrage leartzirig rrraregies: Wh~linreverv tenciier slioidd ktiow.

New York: Newbury House.
Pennington, M. C., Costa, V., So, S., Shing, J. L.-W., Hirose, K.,& Niedzielski, K. ( 1997,
in press). The teachi ng of English-as-a-second-languagew n ting in the Asia-Pacific
region: A cross-country comparison. To appear in RELC Jorirtznl.
Pennington, M. C., & So, S. (1993). Comparing writing process and product across two
languages: A study of 6 Singaporean university student writers. Joitrnal of Second
Larigringr IVri~itig, 2,4163.
Perl, S. ( 1979). The composing processes of unshlled college writers. Research in the
Tenchirg of ErigiLÎh , 13.3 17-336.
Pianko, S. (1979). A description of the composing processes of college freshmen wnters.
Research iti the Teaclzitig of EngIisIt, 13, 5-22.
Pressley, M., & Aiflerbach, P. ( 1995). Verbal protocols of readirrg: The m t w e of
cot~striictivelyresporisive readitig. Hillsdale, NJ: Lawrence Erlbaum.
Prior, P. ( 1991). Contestualizing writing and response in a graduate seminar. 1Vritteri
Corrzrnwtica~iorl
, 8, 267-3 10.

Purves, A. (Ed.) (1992). The IEA study of ivrizten cornpositiott II: Edrrcalioti atrd
per/ormmce iti forîrteeri cowtries. Oxford: Pergamon.
Raimes, A. ( 1985). What unskilied ESL students do as they wnte: A classroom study of
composing. TESOL Qrcarterly, 19,329-258.

Reid, J. M. (1993). Tencliiiig ESL writi~ig.Englewood Cliffs, NJ: Prentice Hall Regents.
Riazi, A. ( 1995). Sociali:atiori into acadernic writirig iri n secorid larigrcage: A social-
cogiritive arialvsis of text production and learrririg amorig Irarziarr graduate studetits
ofediccatiori. Ün~ublisheddoctoral dissertation, Uni versi ty of Toronto.
Riazi, A., Lessard-Clouston, M., & Cumrning, A. (1996). Observing ESL writing
instruction: A case study of four teachers. Jormral of Interisive Erigiish Stzrdies. 10,
t 9-30.

Richards, J. C., & Nunan, D. (1997). Second language teaching and leaming. In M. C.
AIkin (Ed.), Eircyclopedia 01edi<cationalresearch (6th ed., pp. 1200- 1208). New
York: Macmillan.

Richards, J. C., Platt, J., & Platt, ii. ( 1992). Lorigtrian dictiorinry of latrgiiage teacliirig and
applied firiguistics (2nd ed.). Halow ,England: Longman.
Ringbom, H. ( 1987). The role of the Jirst larigirnge irz foreigrr laagicage leaniirig.
Clevedon, Avon: Mu1tilingual Matters.

Rodgers, T. S. (1989). Syllabus design, cumculum development and polity determination.

In R. K. Johnson (Ed.),The secorid latigriage citrriculirrn (pp. 74-34). Cmbridge:
Cam bridge University Press.

Salomon, G. ( 1991). Transcending the qualitative-quantitative debate: The anal y tic and
systemic approaches to educational research. Mircational Researclzer. 20. 10- 18.
Sasaki, M., & Hirose, K. (19%). Esplanatory variables for EFL students' espositoq
writing. Lnrrggnage Lear~iirig,46, 137-174.

Scardarndia, M., CQPans, P. (1985).The function of esplicit discourse knowledge in the

development of test representations and composing stntegies. Cogtiitiori ntid
Instritctiori , 2, 1-39.
Schubert, W. H . ( 1986). Czirricrilrirn: Perspeclives, pnradigm, a d possibility . New York:
Macmillan.

Shaughnessy. M. P. (1977). Errors niid expectations: A guide for rlre tencher of basic
writirrg . New York: Osford University Press.
Shibata, M. (19%). The effects of the L1 onhognphic system and L2 language proficiency
on L3 writing processes. In S. T. Hayes & E. Takahashi (Eds.), Proceedirzgs of the
Eiglit Anmal Coti/erence of the Lake Erie T e a d e r s 01Japanese: Buck to the
bmics--bitildirig arttotrraticity (pp. 137-154). The Lake Erie Teachers of Japanese.
Shiffrin, R. M., & Schneider, W. (LW7). controlled and automatic human information
processing: II. Perceptual leaming, automatic attending, and a general theory.
Psychologica l Review , 84, 127-190.
Silva, T. (1990). Second language composition instruction: Developments, issues, and
directions in ESL. In B. Kroll (Ed.), Seco ,id lorigtiage writirig :Research imiglit /or
t k clacsroom (pp. 1 1-23). Cambridge: Cambridge University Press.
Silva. T. ( 1993). Toivard an understanding of the distinct nature of L3 wri ting: The ESL
research and its implications. TESOL Qtinrterly, 27,657-677.
Smagorinsky, P. ( 1989). The reliabili ty and vaiidi ty of protocol analysis. W t e n
Curnrniinicatiort ,6,463-479.
Smagorinsky, P. (19941). Introduction: Potential problems and problematic potentials of
using talk about writing as data about writing process. In P. Smagorinsky (Ed.),
Speokitig about writi~ig : Reflectiotis on research rnethodology (pp. i x-xix).
T h o u s u d Oaks, CA: Sage.
Smagorinsky, P. (Ed.). ( 1994b). Speaking about ivritirig: Rejlectioas on research
rnethodology .Thousand Oaks, CA: Sage.
Smagorinsky, P. ( 1 9 9 4 ~ ) Think-aloud
. protocol analysis: Beyond the black box. In P.
Smagorïnsky (Ed.), Speakitzg abolir writing: Refectiorrr oti research rnethodology
(pp. 3-19). Thousand Oaks, CA: Sage.
Smagorinsky, P. (1997). Personal growth in social context: A high school senior's search
for meaning in and through wri ting. FVritren Cmnrirlicaiiori, 14,63- 105.
Smith, C. T. ( 1978). Evaluating answers to comprehension questions. The Reading
Teaclrer, 3 1,896-900.
Smith, F. ( 1990). ru tlrink. New York: Teachen College Press.
Smith, F. (1994). LVrititrg nttd the writer (2nd ed.). Hillsdale, NJ: Lawrence Erlbaum.
Snow, M. A.. Met, M., & Genesee, F. ( 1989).A conceptual frarneivork Cor the integntion
of language and content in second/foreign language instruction. TESOL Qaarterly ,
23,201-317.

So, S. (1989). Process and prodrrct of six college stildettt writers: A direcf observation
stiidy of LI (Etiglish or Ciiitrese) arid L2 (Japatiese) writitrg . U npublis hed master's
thesis, University of Hawaii at Manoa, Honolulu.
So, S. ( 1994a). A holistic approach to foreign language curriculum. Unpublished
manuscript, Ontario Institute for Studies in Education of the University of Toronto.

So, S. ( 1994b). Case s tudies in second language acquisition research. Unpublished

manuscript, Ontario Institute for Studies in Education of the University of Toronto.
So, S. (1994~).Writing and leaming in the context of J a p n e s e a s a foreign language
education. Thesis proposal, Ontario Institute for Studies in Education of the
University of Toronto.
Spaulding, C. L. (1992). The motivation to read and wnte. In J. W. Invin & M. A. Doyle
(Eds. ) , Rendhg/>vritittg corinectiotzs: Learriitig Potn research (p p. 177-701).
Newark, DE: International Reading Association.
Spolsky, B. (1989). Conditions for second larigrcage learrzirzg: Introductiori tu a gerzeral
iheory. Oxford: Oxford University Press.
Stern, H. H. ( 1983a). Furidarnental concepts of larigrcage teachirzg. Oxford: Oxford
University Press.
Stem, H. H. (1983b). Toward a multidimensional foreign language curriculum. In R. G.
Mead (Ed.), Foreign languuges: Key links in the cllain of learning (pp. 120- 146).
Middlebury, VT: Northeast Conference on the Teaching of Foreign Languages.
Stem, H. H. ( 1992). Issires and options iti lnrrgiiage teachirzg (edited by P. Allen & B.
Harley). Oxford: Oxford University Press.
Stevick, E. W. ( 1990). Hmariisrn Ni lariguage teachhg . Oxford: Oxford Uni versi ty Press.
Stntman, J. F., & Hamp-Lyons, L. (1994). Reactivity in concurrent ihink-aloud protocols:
Issues for researc h. 1n P. Srnagorinsky (Ed. ) , Speakirig nborrt writirzg: Reflecfiotu
on research rnetlwdology (pp. 89- 1 12). Thousand Oaks, CA: Sage.
Strauss, A. L. ( 1987). Qualitative amzlysis for social scientists. New York: Cam bridge
University Press.

Strauss, A. L., & Corbin, J. (1990). Basics of qualitative researclr: Grouaded fheorv
procediires and teclmiqiles. New bury Park, CA: Sage.
Street, B. V . ( 1984).Literacy irz tkeory arzd praciice . Cambridge: Cam bridge Uni versi iy
Press.
Sudermann, D. P., & Cisar, M. A. (1992). Foreign language across the cumculurn: A
critical appraisd. The Modem btzgirage Joiirrml, 76, 295-3023.
Swaffar, J. K. ( 199 1). Language learning is more than learning language: Rethinking
reading and writing tasks in testbooks for beginning language study. In B. Freed
(Ed.), Foreign language acquisitiotr researcli and the classroorn ( p p . 752-279).
Lexington, MA: D.C.Heath and Company.
Swain. M. (1995). Three functions of output in second language leaming. In G. Cook & B.
Seidelhofer (Eds.), Priticiple and prac~iceiri npplied liiiguistics ( p p . 135-144).
Oxford: Oxford University Press.

Swain, M., & Lapkin, S . (1981). Evaliiafitzg bilingiial ediication: A Canadiari. case stirdy.
Clevedon, England: Avon.
Swain, M., & Lapkin, S. ( 1995). Problems in output and the cognitive processes the-
generate: A step towards second language leaming. Applied Lingrcistics, 16, 371-
39 1.
Takano, Y ., & Noda, A. (1995). Interlanguage dissimilarity enhances the decline of
thinking ability during foreign language processing. Larrgiinge Lenrtrirrg, 45,657-
681.
Taylor, I., & Taylor, M. M. ( 1995). Writitig and literucy itz Chinese, Korean and Jupunese.
Amsterdam: John Benjamins.
Uzawa, K., & Curnming, A. (i*89).Wnting strategies in Japanese as a foreign language:
Lowering o r keeping up the standards. The Caaadian Modern Languoge Revîew,
46, 178- 194.

Valdes. G., Haro, P.. & Echevamarza, M. P. (1992). The development of wnting abilities
in a foreign language: Contributions toward a general theory of L2 wnting. The
Modern Lunguage Jotirtzul. 76,333-352.
van Lier, L. (1996). hteractioti itz the latzgiiage curriculiirn: A~varetiess.auionomy. and
aithenticity. London: Longman.
Vaughan, C. ( 199 1). Holistic assessment: What goes on in the rater's mind? In L. Hamp-
Lyons (Ed.), Assessitg srcotzd latrgriage writirig itz acadetnic cotr~exfs(pp. 11 1-
125). Nonvood, NJ: Ables.

Vygotsky, L. S. ( 1986). Tho~cglitatrd latrgiiage (A. Kozulin, Tram.). Cambridge, MA: MIT
Press. (Original work published in Russian in 1934.)
Wertsch, J. V. (Ed.). ( 1981). The corzcept of activity b Soviet psychology . Armonk, NY:
Sharpe.

Wertsch, J. V. (1985). Vygotsky atid the social fortriatiori of tnimi. Cambridge, M A :

Harvard University Press.
Whden, K., & Menard, N. (1995). L1 and L2 writers' strategic and linguistic knowledge:
A mode1 of mu1tiple-Ievel discourse processing. Langiiage Leartzirig, 45,38 1-4 18.

Williams, J., & Snipper, G. ( 1990). Liierncy and bilitrgziulism. New York: Longman.
Willows, D. (1996, November). The balanced and flesible li teracy diet: A teacher
education system for improving litency levels. Paper presented a t the Human
Development and Applied Psychology Research Colloquium, Ontario Institute for
Studies in Education of the University of Toronto.

Yau, M. S.-S. (1981). Two approaches to irtiprovitrg the writitig perforrnnnce of ESL
students. Unpublished doctoral dissertation, University of Toronto.
 Appendix A: Letters of Consent

Givert to Kei

Dear [Kei]:
From our earlier conversation I understand that you are wiIling to support my doctoral thesis
research by allowing me to solicit participants from your class of [course number]. observe some of
the classes during the 1-95 academic year, and interview you monthiy during the year about
your perception of course objectives and instructional techniques and your evaluation of course
progress.
Data obtained from you and your class will be used exclusively for the purpose of this research.
Coofidentiality of the involvement of yourself and your class in the study will be strictly
maintained by using pseudonyrns.
At the completion of the research 1 will provide you with a sumrnq report of the study's findings.
Your participation is voluntary. If you decide to participate at this point, you may withdraw from
the study at any tirne.

Attacheci are a soticitation letter to be distributed to students and my thesis proposal. Please read
thern carefully. If you are still interested in supporting my research project. couid you kindly
complete the attached form and return it to me?
Thank you very much for taking the tirne to read this letter. 1 can be contacted at the telephone
number or e-mai1 address below, if you wish to have more information.
Sincerely,
Sufumi So
Tel. no.: 924-5918
E-mail: [email protected]

Dear Ms. So:

I have read your Ietter and proposal describing the study you are conducting on writing in Japanese
as a foreign language. 1 would like to support and participate in this research.
Namc:
Telephone:
E-mail address:
Mailing address:

Signature: Date:
To whom it may concem:

1 am writing to seek your consent to my research study on writing in Japanese as a foreign

Ianguage that involves one of the courses offered in your department. The research is part
of my Ph.D. thesis at the Ontario Institute for Studies in Education of the University of
Toronto.
Ms. Kei] in your department has given me her permission to solici t volunteer students
from her class of [course number] and to observe some of her classes dunng the 1994-95
academic year for the purpose of my thesis research. Ms. Wei] has reviewed the details of
this research and approved of its design.
Data obtained from the class will be used exclusively for the purpose of this research.
Confidentiality of the involvement of your department's course in the study will be stnctly
maintained.
If you could permit me to work with the instructor and students of [course number] for my
thesis research, could you liindly complete the attached rom and retum it to me?

Thank you very much for taking the time to read this letter. I can be contacted at the
telephone number or e-mail address below, if you wish to have more information.
Sincerely,
Sufumi So
Tel. no.: 924-59 18
E-mail: [email protected]

Dear Ms. So:

1 have read your letter and would like to support your research project.

Name/Title:

Signature: Da te:
Appendix C: Questionnaire on Teacher Beliefs and Practices of Teaching
of Writing

1. What should the purpose of your students' writing activity be?

What was the purpose of your students' writing activity?

writing as writing as a
language pnctice discovery process

IDEAL: I 2 3 4 5

ACTUAL: 1 2 3 4 5
Please elabonte o n your answer:
[More space was given in the actud questionnaire. 1

2. What should the role of the teacher in the writing class be?
What was the role of the teacher in the writing class?
to present information to facilitate each
about written Japanese student's writing process

Please elaborate on your answer:

3. What should the role of the student in the wnting class be'?
What was the role of the student in the wriûng class?
to receive information to take active
frorn teacher responsibility for learning

IDEAL: 1 2

Please elabonte on your answer:

4. How should the lessons be organized?
How were the iessons organized?
student-centered

IDEAL: 5

ACTU AL: 5

Please elaborate on your answer:

5. How shouid topics be detennined?

How were topics detennined?
teachcr-generated student-generated

IDEiiL: 1 5
.L\CTU.L\L: 1 5
Please elabonte on your answer:

6. Should class tirne be devoted to prewriting activities such as brainstorming and discussion
of the topics'?
Was class time devoted to prewriting activities such as brainstorming and discussion of the
topics?
for every
assignrnent

IDEAL: 1

ACTUAL: 1

Please elaborate on your answer:

 Appendix D: Letter to Solicit Participants

Dear friend:

1 am writing to ask if you would be willing to participate in a 6-month-long research study

on writing in Japanese as a foreign language. The research is part of my Ph.D. thesis at the
Ontario Institute for Studies in Education of the University of Toronto. I have obtained the
permission of Ms. Wei] to circulate this letter to ask for your participation in my study.
The main purpose of the study is to esamine the wri ting-and-leaming relationships in the
context of foreign language education. Results of the study should help Japanese and other
foreign language educators understand better what students do, think, and l e m as they
wnte in the target language, and thus wodd contribute to the improvement of foreign
language writing instruction.

If you are interested in volunteering for this study, you will be asked:
1. to talk aloud into the audiotape while writing the course assignments given by
Ms. Wei];
2. to attend an interview w hich will be conducted in English tocvard the end of the
course ;
3. to take a widely-recognized tape-rnediated test of Japanese on1 proficiency,
Japanese Speaking Test (JST), developed by the specialists of the National
Foreign Language Resource Center of the Division of Foreign Language
Education and Testing at the Center for Applied Linguistics in the United States
(The testing fee, US$60 per examinee, will be paid by me. A copy of JST
Esaminee Handbook is available from me or Ms. F e i ] for your information);
4. to complete a questionnaire asking about your background and experience of
writing and l e m i n g Japanese; and
5. to write a short essay in your mother tongue.

In retum for your participation in the entire study you will be paid Can$100. Another
benefit of yours may be my feedback to your writing. 1 will diagnose your writing
performance based on the data 1 will have collected from you and provide you with
feedback and suggestions for the improvement of your Japanese writing.
Data obtained from you will be used exclusively for the purpose of this research. What
you may say or do during the study will not be disclosed to your course instructor. It
will in no way be incorporated into the course grades. Confidentiality of your
involvement in the study will be strictly maintained by using a pseudonym.
Your participation is voluniary. If you decide to participate, you may withdraw from the
study at any time.
If you could participate in the study, could you kindly complete the atiached f o n and
retum it to me. 1 wiil contact al1 people who indicate they are interested and make
necessary arrangements wi th thern.
Thank you very much for taking the tirne to read this letter. 1 can be contacted at the
telephone number or e-mail address below, if you wish to have more information.

Sincerely,

Sufumi So
Tel. no. : 924-59 18
E-mail: [email protected]

Dear Ms. So:

1 have read your letter describing the study you are conducting on wnting in Japanese as a
foreign language. 1 would like to participate in this research.

Narne:

Telephone: (Best time to call: 1

E-mail address:

Mailing address:

Signature: Date:
 Appendix E: Questionnaire on Student Background, Japanese Language
Study, and Writing in L1 and Japanese

1. Identification

a. Narne

b. Age

c. Male Female
d. Major

e. Year undergraduate: first , second , third , fourth

graduate: first , second , third , fourth

f. Country of birth
g. Mother tongue

h. Any other languages you speak andfor wnte in

i. Y our strongest oral language

j. Your strongest written language

II. Japanese language study

a. How long have you been studying Japanese in formal instructional
settings?

b. Describe briefly the history of your Japanese language study (when, where, how long,
what skills--e.g., speaking, writing, reading, etc.).
c. Have you ever lived in Japan before? If yes, when, how long, and for what purpose?
Y es When? From 19
How long? year(s) and month(s)
For what purpose(s)?

d. How do you rate your overail proficiency in Japanese as cornpared to the proficiency of
otlrer shddents in this class?
Excellent ,Good ,Fair ,Pmr
e. How do you rate your overall proficiency in Japanese as compared to the proficiency o l
riutive speakers of Japanese?
Escellent , Good , Fair , Pmr

f. How important is i t Tor you to become proficient in Japanese?

Very important , Important ,Not so important

g. Why do you want to learn Japanese? (Check ail that apply)

interested in the language
interested in the culture
have friends who speak Japanese
required to take a language course to graduate
need it for my future career
need it for travel
other (hst):

h. What has been your favorite esperience in leaming Japanese?

III. Writing in Japanese and LI
japanese wnling
a Do p u wri te in Japanese anything other than course assignments?
Y es (specify):
No -
b. How do you evaluate p u r writing proficiency in Japanese as compared to that of otlter
studertts in this class?
Excellent ,G d , Fair ,Poor

c. How do you evaluate your writing profïciency in Japanese as compared to that of native
speakers of Japanese?
Escellent ,Good ,Fair .
Poor
d. Do you have any specific problems in writing in Japanese?
Y es (list):

a. What types of wri ting do p u perform in English (check al1 that apply)? And how often?
diary for yourself: Always , Sometimes , Not very often
personal letters: Always , Sometimes . Not very often
business letters: Always , Sometimes , Not very often
essays or reports for courses you are taking at the university:
Always , Sometimes , Not very often
essays or reports for newspaper or magazines
Always ,Sometimes , Not very often
books (what types? how frequent?):

other (what? how frequent?):

b. Have you published anything in English?
Y es (specify):
No -
c. How do o u evaiuatc your writing in English?
Excellent .G d .F r , Poor

Please ans wer the satne griestiotis below if yortr strotzge~twritte~ihngruige (Le., the
lnnguage you listed in I-j of this questiutrtraire)is not ErrglisR.

a. What types of writing do you perform in the language listed in I-j or this questionnaire
(check al1 that apply)? And how often?
diary for yourself: Aiways , Sometimes , Not very often
personal letters: Al ways .Sometimes , Not very often
business letters: Always ,Sometimes , Not very often
essays or reports for courses you are iaking at the university:
Always , Sometimes , Not very often
essays or reports for newspaper or magazines
Always .Sometimes , Not very often
books (what types? how frequent?):
-p. - -

oiher (what? horv frequent?):

b. Have you published anything in that language?

Y es ( speci fy):
No -
c. How do you evaluate your writing in that language?
Excellent ,G d , Fair , Poor

TIiurtk yoa very tri~~ch

for atrsweririg tiiese qiieslioris.
- END -
 Appendix F: Sample Japanese Essays

Colirt, Task I
Eddie, Tmk I
 Appendix G: Essay Rating Scale

Adapted from "Reconstnicting '.l\cademic Wnting Proficicncy"' (pp. 149-15 1). by L. Hamp-Lyons, 199 1,
Assessing Secolrd Language Wrifing in Academic Conte-rfs (cdited by L. Hamp-Lyons). Norwood, NJ:
Ablcx.

- - -

communicritive quaÏity otgnnizrition linguistic nccurncy 1 linguistic appropriacy

The wri ting displays an The writing displays a The reader sees no There is an ability to
abiiity to cornmunicrite compfetely logical mors of vocabulary, manipdate the linguistic
in a way which gives the ' organizational structure spelling, punctuation or systems with complete
reader full satisfaction. which enables the ,orammar. appropn acy.
message to be foUowed
effortlessly.
The writiag displays an The writing displays a The reader sees no Therc is an ability to
abiiity to communicate logical organizationai signifiant errors of manipulate the linguistic
without causing the structure which enables vocabulary,speiIing . systems appropriately.
reader any difficulties. 1 the message to be punctuation or ,onmmar.
1 followed easily.
The writing displays an The writing displays The reader is aware of There are minor
ability to communicatc good organizatiod but not troubleci by limitations to the ability
with few difficulties for structure which enables occrisional minor errors
to manipulate the
the reader. the message to be of vocabulary. speiiing,
linguistic systems
followed throughout. appropriately which do
punctuation or ,orammar.
not intmde on the
reader.
The reader is aware of Tbere is Iimited ability
errors or vocabulary. ( to manipulate the
spelling. punctuatich or linguistic systems
,oy;~atnti~,but these appropriately, but this
reader . intmde only intntdes only
ocwsionai~~. ocwsionally .

--t-
The writing displays an The writing is organized The reader is aware of
ability to communicate well enough for the
although there is oftcn
strain for che reader.

The writing displays a

limited ability to
most of the rime.
mors of vocabulary,

frequently.
The writiag lacks a clear The reader Ends the
organizationd structure control of vocabuiary.
There is limited ability
to manipulate the

l
message to be followed spelling. punctuau&i or linguistic systems
prammar which inmde appropriately which
iauudes frequently.
There is inabiiity to
manipulate the l&uistic
communicate which purs and the message is spelling, puncturition systems appropriately.
strain on the reader difficult to follow. and t-ar which causes severe
throughout. inad&ate. strain for the reader.
The writing does not The wnting kas no The reader is primaily There is little or ao
display an abiiity to disceniibte aware of gross sense of linguistic
cornmunicate aIthough organizational structure inadequacies of appropriacy. although
meaning cornes tlirough and a message m o t be vocabulary, spelling . there is evidence of
spasmodicall y. followed. punctuation and 1 sentence smcture.
pmmar.
The reader sees no There is no sense of
evidence of control of linguistic appropriacy.
vocabulary , spelling,
 Appendix H: Instmctions for Essay Rating Procedures

Dear Rater.

Before you read and rate the essays, please examine the rating scale
adapted [rom the Hamp-Lyons scale carefully. You may rate the essays in
any order.

Suggested rating procedures:

1. Read an essay quickly for an overall first impression. Based on the first
i rnpression. identi fy appropriate descriptors in cohfi&?L'ïc.~nvE
Q L A U ~
and ORGXSIZATION, and then record numerical ratings in the provided
rating report.

2. Quickl y re-read the essay. Identify appropriate descriptors for Ui\;GL;[S?+ïc

.-ICCL~ICYand LIXGUSTiC APPROPRIACY and record ratings.
 Appendix 1: Student Interview Guide

0. -Thank the students for their participation.

-Explain the content and purpose of the study.
-M&e the purpose of the interview clear.
-Ask if the' have given some thoughts to the interview questions. (If not. give 5
minutes for them to think.)
-0btain their permission to audiotape the interview.
1. What do you think you learned from writing the two sakuburz [compositions]
for the course?
-from your esperience of wri ting on rtuzriga [comics]
-from your experience of writing on rzihorr tio seikatsrc [living in Japan]
-in terms of content, Ianguage use, and other aspects
2. What expectations do you have for your own sahbwc in Japanese?
-Do you consider what expectations Ms. [Kei] might have o n your
sak~birri? If yes, what do you think are Ms. K e i ] ' s expectations?
-What aspects of writing do you pay conscious attention to while ivnting
Japanese sakzrbrrn ?
3. What dimeulties do you have in writing Japanese sakrrbun? Or what are the
frustrations you feel while writing in Japanese? And how do you deal with
them?
- C m p u compare your experience of writing in Japanese with that of
writing in your first language. What are the common problems you have
across these trvo esperiences? What problems are specific to your Japanese
writing esperience?
4. What functions does sakubsrrz -writing play in your study of Japanese?
-How do you use writing for your study of Japanese?
-Howdoes wri ting help you learn the Japanese language or any thing else?
-How d o you feel after you have wri tten sakubroi in Japanese?
5. Do you like reading and writing in generd? Do you often read and write in
Japanese and your fint language? What types of reading and rvtiting?
-Cm p u tell me your experience of learning to write in your first language?
- C m you tell me your experience of learning to write in Japanese?
6. Do you have anything else o u would like to Say about the writing assignments,
think-aloud task, etc.?
 fipemtionfor~coocema b w l Conccpon & expcricncc d
one's o m Japanese miring n n t i n in
~ Japcsc writinn in L1
It wasn't a rcwarch paper so ih I~ c d r o m l c ~ l u r ~ c r r rIyhave a problcm with scnlmcc- It gave me a good opportwuly to Ihave p r d c m s miting essays i n Incvcr r w t c d wnhng w a y s i n
morc 01 your o p ~ o n l ~ k prttjng
c h e p i n t s &m md final cxprcuions i n miling. thinkabout whaltomilcandIhe . ilis
E n g M or J a p u ~ l eBut Japanese mlrl this yeu. Iwmtcd
explaining why. Idm wankd to mising up formal and i n f w d m y l o phnle Ihings appropnalcly casicr lo mite in Englirh than l n l o improvc my wiling a lot mors
w i l e in a p v r . malure writing expndonr. for wnung, md b l c a m how to Japancw bcaruc it#faskr and I Uun o W things k c a w U k m g
style. Ididnl wan1 u, wrib h a untc h j i . gucu here's no kanji. nnd reading uo nol much d a
m y Idlr Also kanji, Ilookcdup pobicm. Writing hu rlways bcrn
d m a t cvay kanji tomake m e my w&st p i n L
hey wnc ccmcct
IItel UUI since I'm Japancsc. I Iw m t l o k a b l e t o m i t e i n a M y problems u c the [formal] wa: I lcuncd a h Ihc tcpc (but not h g l i r h doan7 have much set I t was my flnttime wnang an
should bc able to nntc Ilks a m a m faslim udng r mom d w i n g w w d s u d kanji.And my from h 2nd task) u well but I style1rhinfr,m1don~hvc~y CSUy d ~ r d c m NWC
k Ilk hl
normal Japanese of my age. bul I sophisticated form d exprowow handwriting sucks lcuncd abwl r big gap bctwccn relistance towar& mitlng i n Iw u w n g really hud
can7 1rs like I'm h n d d stupd and moce kanji. my abiliocs to spak and l o mils. English I'm wcak at cxprcuing Rca&ng w w l d be k t k r for me lc
md Ihat's really slrcdul Ilcuncd my writing ability w u my opnion dwing hud wads. a b m b [linguishc) h n g s
slightly low md 1 didn't know i t h%
w u this bad
It has lo be p n o d l y mcwngful For Ihe 2nd task Ij u t wankd to Kanji is a pmblcm. And Ju, my 1 think it's r good w ~ toy learn Incvalikc to wrib for ochcn Writim might help y w l o bccomc
W I h Ihc first essay Iw o k I n lasl g c i~t ovcr. But the 1st onc Itried vocabuluy m d forms d vocabulary k c a w when you kaw d r f e u dh v i n g r c l s u a speaker.
pI"tours Ihad a lot of fun ro tic rt with uxncching mom cxprcuion uc l l m ~ t r d write about a ccmin Lopic, you m~ne pdge me ttrroughmy 1took a Japanuc c o m p u o n
untlng bccawc 11 wrr rntetertmg pcrrorul l o me so hat it would k d t c n repeat certain wotdr and i f wiuriting. But I've k e n rcccntly c um last ycu.
tome Duf l ddn'l fuul the lop= easier to write, Idm l r i d to theymnotthcwo&ywlmow, very nq&d to ba t d d h a t I I dwayr wanted 10 k able to m d
of h e w o u i g u n e n u In h s imihtc tbe c i d u , indirect lhcy b c c m c familiar to you write v a y cfkienrly. I t r d y and wita karr k c a w hat's
c w x wo lntcrcsong mctoricrl arguritatim d hclpcd me in how If c d about what 1 Wnk I'm marlly lacking
Japanese biramc. mi tin^.
Iwant my csays to k coherent, M y vocabulary i s vcry, very I'm studyrngKicnco so Idon1 I've ncvcr rcllly written essays I n
smight to the paint, md clcariy limited clpccidy f a a tinlc bit have Ihl~many oppabutics b Japuuu, nor h v c 1 c v n k c n
organized so U u t i t i s easy l o mad dilXcull topics And kanji i s vcry write csuyr taught how to m t c
diITicul1 J think rcrdrng 11 vcry Importan1
in m y language just to gun
vocrbuluy.

Srull
Ucs~dcsUic fact hat 11uka much M y upeclation w w l d bc l o Icwldn't h a d a t my h g h l IIhelped mc lo l c u n new wo?& 1 redly don7 have much of a jiXirns ~ h fin1
c hme IO untc a
longer. ~~ollurlg
IS untaual pmgrcu above whal Il d last into Japu~saso clrily, ro Ihad I t and Ihc context d h e words problem m t i n g in En&rh 1fccl l a p u r c ~coolpullon
WnUr~gIn any langugcs can k ~ m c perhaps
. wing aprcujoru wule faa rubotible and wun't mwc w l e u comforublc a h i t Irs Uw poactivc ruturc of wnung
rcdrotl5 uld words Uut Imight have rum II i t was Ihc mosl a p p p a l n wnting I n English a c o r n p t i o n as o p p o ~ dto
picked up from prcviour lcuonr way d cxprtuing Out puoculu mncly reading h a t hclps to
md readings. thou@ And Mrnclimcs 1 consdidarc lane's L n p s u c
encountered Lfilcultics with knowledge] in mind
pudclcr and I h e word ordcr [Dnwing on his pcnollll
cxpcricncc mlh Engltrh mfangl
l c c W o g y c w l d rsdly w s t i n
dcvelopin~writ in^ ~ b l l s
Sou have ro do 11hkc other When Japlnew pcoplc read my Nohng was new [contmt-msc] I My Ctuncrc compoauon IS noc Ibis is really lhc fir$ time hat I
E n g l ~ dessays
~ llunk of wlut you m y . Uxy will a l leu1 get the lcuned shJf Iconknt inlomaPon] ha1 g w d and, wcll. my Engl~sh b d to do such a long
arc gomg wnu.Vunk of at l c u l ~dcawhat Iwas rrylng to say. i n c l u r and then t w k dl thc compostlion is not W l good ~ o m p m i u c qIhad l o wntc $0
thee p m b about Ihc choxn conccpu m d lhcn u-rote an cuay cllhcr many (hings in Japmesc
~oplc.clabrate Ihcm. and rcc d 1 wlm'l ~ d l emphasizing
y
Ihc grammu IS concct grammar oc stnrhuc [rhctoncll
or~ivinitim).
Thcrc IS somcthmg vcry s p c c ~ l l m e ultimale p a l i s to gct info Lhc ~ i b c i n able
g lo we apprupriat The cxpericncc rcwured me of Iread and m t e a lot i n Chinme. 1 Is thc frnt expericncc of
abuf r ) ~ cstyle al wnhng In ogc of !he Japanese mind c x ~ c u i o r and
u words bccauo of thc lmporurvo dacqurnng Ihc m r c to m y x l f to rurd wl what wiling c r v y r in Japuusc
Japanese thc style dceply lcel of lhc langwgc Also 11may my Lfc IS Wnong i s power A s Y o u must not smggJc wlh
rmluddcd tn local culUrc k a 8004 way l o lcun for Enash mdng. Imade mguagc but ruimllate ~tand gve
LtCc any otl~ertun& of unbng. vocabulary c x p c n t i a l growlhrinrc 1 u m c vourwll to il
hc strmng pat IS 16 search for hnc, Thc cffcctivo writing couna If you da't h v c a teacher. you
m Image, thc sod of an ciwy p d c u o r hero inrplrcd ma r la rhwldrcad mm good Japanese
d i n g 3 to sce how lhey u c
Inpap$

(mble cotit irrues)

 Ih n k unling i n Japancx ir I Liy Io wnle rr close as pautblc 1 Mort problcmr ut vocabuluy ind I I c m c d k c a w Ihad to ncr Wnang Chinec, my nauvc 1 quiic
Wnting Lhsucompouliou war
-c i n i m a u o n for the &nient langwgc. IC u i wnlc lt vcry a cNltnpc for ma b c c a w
dihc c w y r fiucnlly Wnling i n Engliih i r the in Ihc p i 3 y c m wc ditin1 wnie
mort difficuILkaw nobây Out much in Japancn
w w l d iindcniuid my Engluh Wriang hclpr y w fo lcun how l a
c o n m m Japancc scnknccr i n
cOnvrmUOllf
'Iho major pain1d radylng
hpancu well is IO rcad a IOI
Il15 Yi 0ppmWUly IOalJLC you Having my J l p u ~ s cfncnâ In T u w m Iwas bughi how to Wnung Japancrc ciuys ir i w l l y
LIU& aloui haw io cxpnu c o m d my gnmniu [ w u my unle Chincsc c o m p u a n r , Itcp a brand ncw Ihing for mc
younclf i n hplnnc Lcanung capencrice] ùy sirp procuiuics, grunmu Rcadtng h c l p w Imgiugc
W n b . yrd vocabulary Icuning. bu1j w t up 10 a cerilin
pan1 Bcyond U t y w h v c 10
pick up from fie clus or itaclicr
Il givcr mc a chancc 10 pnclice ljust want Io bc abla Io upreu Ihave tmuble cxprcsPng uhai I I lcuncd how much I dont know 1& mi like wnung 1 fend I o This ir a W y Ihr fint time I'vc
uring Jrpanesc, focurtng nn Ihe w b i 1 uant to wy, gct ilmore wanl lo wy, ro iha rcrult d my aboui Japancw for c o m p n g an lump arwnd a Idi n my unling, irnllen a c o m p d m d Uur
n g i i word choicc and r d f . undmublc. Japuwc -y iircaily wri of c m y , which ir very Jiffcrcni r m n g s p g n p h Wh one ides lrngrh i n Japanesc
diluird w rùnplificd from vmply w ù n g d o m Yd rort drwitch or jump to
rentniccs anoihcr
IdonsiUunk rhc corilcnl is nally I just wani io gci a good mPrL 1 . 7hO m u ddficuII Urina is U) 1d d think 1 lumcd hlmuch hgîirh w u ncvct a probltm I For ma unUn8 i r no1hl
impoiuni kcaw lhis ~ s n o i a W l ~ c n i l l l it'r y m m acirully put my Ihwghb inlo a b i thC conknt Pmbably. J I ~htnk1'm rully good aluehnical importanL Rciding, rpclling and
rcwuch papcr tud anyhng, ilwould bc m o d y mting, j w l gcltrng grammu lulcningarc m m important
h a way tn u y Mngs ui J a w corrcct ud k i n g able Io upeu m i i n g kaw h m am fcw
And ilis good for kanji. h g r umply and clerrly opponunilicr to gel to wnio i n
lapancie
Wnling i s onc dhwayr d I lcamcd a lot of nrw w o r b jus1 I'vc coma fmm a u i m c e Wnting rrinforccs spcaiung
gculng h i t e r Japuinc fintdrllundtnhnâabîcind ~her&la,rnuchto~y.It'snoca by w o n s cruyr. 1 l c u n d Iiow background so I'm ami w d i o Icnjoy rtudying J a p w , hl'i
secorrd d 111j w t good q d i t y . 1 @cm io wnto an -y but i o i o uro rhc word i n a propcr wnung many p a p m and 1 don'! uhy 1 put SJJ much t m c mio 11
pay puticulu alleniion (o makc a eohcrcnt t u r y wirh contut by h v i n g mncone check Iike m u n g i n Englirh DU 1
gnmmaiical ihuctrrrs ud proper i n v U d u t j w body. and ovcr my unring gucs my m l i n g A I 1 tn f3giirh
word upgc conduum ir good b ~ n# a cxcclfcni
Lcu cmplusu orircscuch and Whai 1 wani torhinl d ir p q x r Iiukes mlonglolookupa w d t lumcd a lot upccially Idwayr &d wcll rnd cnpycd it Iwuh wa h d donc more -y
conrcrii. more on stylc For plinur for mlcnw-final a lunp. l'm &&y ihrnhng vocabiJuy and kmji for vurow wiùng. Ii'r r loi richcr in knnr d
English csriyr. wtnlyou u y ir upsuiml &ad w b i Iwant lo wilc wonit. idd ~ > s e n k n c c - f i d h a üungs io do,compucd IO ruch
itic mosl i m p r i m i h n g Plfnwanbbut I'm sbll aying Io upreuiaru i n composihorill w l i n 8 cxcmxs liko iummuy
figwcout~oncwdprL;uiy. rtylr. IIJra hclpcd Iodclincalc uniin&
what Iwzr üunhng by m u n g il
down

Naft.. Protocols arc mii~imdlyedited io mnkc h m rcadable

 Appendix K: Instructions for Think-Aloud Procedures

In the process of writing, people think and Say many things to themselves that are quickly
forgotten. Y e t these thoughts are interesting and important parts of the wriier's problem-
soIving process.

1 am interested in the thoughts that go through your head as you work on the wri ting task.

Work o n the writing task as you normally would do. This means that you can use
dictionaries and reference books as you normally would do.

But for this writing session, please SAY OUT LOUD AND CLEARLY everv&hirza that
you have in your head while writing. You rnay use any languages you are currently
thinking in. You m q switch between languages according to your trains of thought. Avoid
k i n g silent for an. length of time. Remind yourself constantly to keep talking aloud.

If you wish to take a break. you may do so. But rcmember to record the time and date
before you stop the tape, and record the tirne and date again when you resume worlcing on
the assignment.
 Appendix L: Post-Writing Questionnaire

A bout the writing assignment:

Q. 1 Are you satisfied with this sakubun [composition]?

Very satisfied ( ) Satisfied ( ) Not so satisfied ( ) Not satisfied at al1 ( )

4.2. Was it difficult to generate ideas for this sakr<bwz?

Difficult ( ) Not so difficult ( 1 &Y (

4.3. Was it difficult to express your ideas in Japanese?

Difficul t ( ) Notsodifficult( 1 &Y ( 1

Please comment on the ivnting assignment (e-g.,How did you feel about the nature of this
assignment? What strategies did you use to write this sakrcbioi? What were the relatively
difficult and easy things about writing this sakubsn?). Your comments will be kept
confidential. They will not be disclosed to Ms. F e i ] . So please feel free to make any
comments.
 Appendix M: Sample Protocols with Coded Categories
Note. English translations of Japanese utterances are given in square brackets. .lapanese words that remain in
these translations are wntten in italics.

anoo saikin katsuji banare shi teim kodomo ga fueteim te yuu guai dakara sore wa
totem0 yoi koto aa PVeU, because ifs been said chat children who are ge&g away from p ~ t e d
words arc incrcasing in number. that (refening to what she said earlier) is very g d . um] @d:
reflective: Japanese
" katsuji banare" banare hanarem tte doo yuu j i ["kotrujibanare" w bat is the h r v i for
bcwarr, Imnareru?](consulûng Japanese dictionary) language use (spelling):
refiective: Ja panese
kakijun ga wakannai "katsujibanare" IIdon't h o w w h t is the stroke order. "kam@mare "]
language use (spelling): reflective: Japanese
sorekara watashi nanka ni totte wa manga O -mu koto ni yotte yappari ni hongo O
zuibun oboeta to omou k m a n o nihongo galrushuu tte yuu to oogesani nani keredo.
maa nihongo no vocabidary [And also. to a person Like me. because 1 ihink 1 leamed quite a
.
bit of Japanese by rading comic books, well, if 1 said Japanese language study it might sound too
big a thing. but a n y û y Japancse v o c a b u l ~ gist:
] reflective: Japanese
vocnb te nan d a t t a k e Dvhat is vocab?] (consulting English-Japanese dictionary)
language use (lexis): reflective: Japanese
"nihongo no nihongo O oboeni" de ii ya ["Japanrse.to learn Japanese" would do.]
language use (lexis): reflective: Japanese
soshiie kono "kaigai deno nihon bunka rikai ni yakudatsu" tte yuu n o w a sono
manga nimo yom keredo, tatoeba sazae-san toka dattara m a futsuu n o [..] O oboeru
tte koto wa dekiru kedo, manga no daibubun wa hon mitaini I;oo iwayum naiyoo ga
fi kushon dakara kanarazushirno yakudatsu towa omoenai kan, sore w a botsu. [And
then, çonceming this statemcnt that "(Comic book) promote foreipersf understanding of Japan," it
depends on the type of comics but, for instance, comics like suzae-san caa be a source for learning
orchnary [..], but most comics are, like book, what you d l fiction, so 1 don't t h i d they can r d l y
hclp. so this is n o g d . 1 gist: reflective: Japanese
If it is. then its. it influences. it is because. whv is it so influential? gist: reflective:
Cantonese
Because it itself is itfonnatiotr ful fillment of the Durpose of i@rtnatiori gist:
reflective: CantonesefEnglish
uh tegamni [easily] ht& riten r k n nten [advantage advantage advantagel but because of
-
this tokuchoo [chancteristics) the problern of nichi~msei[ordinariness] n ten [advantage]
-
is ah nten ah nten [advantage um advanrage] is this this this vew easv tegam tegaruni
joohoo o em [easy COobtain information] verv easv to obtain information. gist:
reflective: Cantonese/Japanese
That's riten [advaotage] joohoo [donnationl [..] the problern will be easv t o solve.
gist: non-reflective: CantoneselJapanese
Also this our final ~ r o b l e mof riten [advantagel is inparagrapli six which talks about
-
the tiatiire d tokuchoo nichijoosei [chmctenstics ordinariness] discourse
organizatiodgist: non-reflective: Cantonese/English
What thev reflect han'ei nichijoosei mangabunka no nichijoo mangabunka wa [[O
reflect. ordinariness. routine of cornic culture. mmic culture is] is the same as S ~ U U ~ ~ O
[religion] the same shuukyoo shuukym ni niteiru shuukyoo ni niteini [religion.
resembles religion. resernbks religion] gist: reflective: CantonesdJapanese
So this problem is whv this comic culture, although this comic culture is like
poprilnr culriire popiilar criltiire yori catr cotripure tu popiilar crilture. gist:
reflective: Cantonese/English
The rialrire of iliis poprilar crilture mangabunka [couic cuiture] is is the popitlarity.
gist: reflective: English
It ils iti.iiettce nJfect ifs itzflieticr is so broad ir reaches everv levels of socieîy bi a
ivay tliat a poprilnr religiori coiild do. gist: reflective: English
So tliey sliare the same characteristic ivhich is whicll corild cati be good or cati be
bad. gist: reflective: English
The &ad [Iritig of coiwse is somelitnes yoii cutitiot selecz sometimes yoii cantiot
select yoiir iti/onnatioti. gist: reflective: English
Yoii are forced to be yoii are forced to accept sorne of l e iri$orrna~iomiti/ormttion.
gist: reflective: English
Biit iu n ivav it also sofves the daily probletri. gist: reflective: English
I tliirrk ltve got to give on exatnple. writing procedures: reflective: English
Urn okay \ v h t shorrld I say afer fhis? hm»i let's see rrh Iziirn gist: reflective:
English
[..] yes sasou sasowanakatta sasowanaide sono bu sasowanaide [to invite. did not
invite, without inviting, bat department. without inviting] language use (syntax):
reflective: English
And Rere is a note here. writing procedures: non-reflective: English
rrlr okay "kekkon [mamage]" This is irz brnckets. writing procedures: non-
reflective: English
U m okav wlmf else c m I say? gist: retlective: English
I have one more page to write. writing procedures: non-reflective: English
Ali. karaoke rrm yeah eeto [weJIl let me tlzitrk okav gist: reflective: English
Uh hrirn hurn k t me see vit's coherenr, are fliq? discourse organization:
reflective: English
[../ sutnerlritzg eelse. gist: non-reflective: English
Hrrrrn ler me go [..] biit a k o knraoke ron idea ulr " bure tei kimashita [took (US Qere)] "
[..f sort of ki,rd of divided irh Rirm discourse organization: reflective: English
&rd rrh "sono baa e iku tabini [every time WC went to that bar]" ano "nankyoku kyoku
kyoku [how m;mv songs. pieces of sangs. pieces] " is that a coirriterfor like sorig?
language use (lexis): reflective: English
1..1kyoku [piece] (consulting electronic Japanese-Englis h dictionary) kyoku kyoku
Ipiece piece 1 -ah language use (lexis):reflective: English
"imamade [untii now]" huh "karaokebaa de [in the karaoke bar]" I thitzk if'sgood
"karaokebaa de lutsuu futsuuni [in the h o k e bar. usuaily. usuaiiy]"ah " futsuuni
lutsuuni [usuaiiy us~aiiy]''language use (syntax): reflective: English
ah qiriet qiiief persori qiriet irh qiiiet persotz q-wi (consulting electronic Japanese-
English dictionary) shizuka [quiet]? quiet persori rnaybe shizuka yasashii [quiet kind]?
language use (lexis): reflective: English
uchi ki [inwnrd] shy uchi ki [inward]itmshy is uchi ki [inwardl dcshm? [is it*?]language
use (lexis): reflective: English/Japanese
kirzd of j cari 't everz rernember how to write natsu [summer] language use
(spelling): reflective: English
I'rrr jiist writirig ffirse[..] thirtgs thnr corne fo my hend. writing procedures: non-
reflective: English
Eddie.

050 Utn koiv fo say everydny li/e itrn mainichi seikatsu [everyday Life]? language use
(lexis): reflective: English
Gee. ivliafs tlie ivord. very good ivord? everyciny li/e I car1 language use (lexis):
reflective: English
Huiv ahurit tfzat? uh Iium im alriglit (reading the sentence jusr wn tten) " bangumi O
mi nagara atama ni kangae ga ukanda [it occurred to me while watching the proenml "
language use (syntax): reflective: English
Maybe instead of dentootekni [traciitionai~y1 ~~(rially futsuu [usuaiiy] fudanni fudanni
tabun fudanni no h m ga [ordinarily. ordinarily maybe ordinarily is beiter] language use
(lexis): reflective: English/Japanese
fudan um fudan [ordinarily] here »te go okay fudan (consulting Japanese-English
dictionary) language use (lexis): non-refïective: English
"keredomo" uh "nihonshakaishi O miruton ["however" uh "if one looked at the history of
Japanese society"1 history o j Japanese socieiy ith Jnpntiese sociev jrrst "nihon shakai
O rniruto" [iîone looked at Japanese society] language use (lexis): reflective: English

Ho w corne isn't t h e a word danjoyakuwari [menand women's roies]?language use

(lexis): reflective: English
ZOO mricfi " to n o ijoo ga talcusan takusan ga oo m i " ah 'ooi' no hoo ga ii

["abnomalities are many many ple plentynah plenty is better] ianguage use (kxis):
reflective: Japanese
ion danjo [menand women] istl't there atrother, is tliere n syriotzyrri for yakuwari
[assi,pment of ~OICS]?language use (lexis): reflective: English
Maybe I look rip 'role 'hnri Eriglish orle. writing procedures: reflective: English
s l lurm role role ruot [..J yaku [roiel just yaku [role] rnaybe irr tlie Japatiese otir.
(consulting English-Japanese dictionary) writing procedures: reflective: English
yakuwari yakuwari [assigoment of roies assignment of ro~es]rrh ivfiere ivodd it be?
(consulting Japanese-English dictionary] writing procedures: non-reflective:
English
yaku [mie] tliat's the posi. positiori. language use (lexis): non-reflective: English
yakuwari [ a s s i p c n t of roIes1 iun ilo iio, assirr~ilatea feiv eqitivalrrrt words .
language use (lexis): reflective: English
"danjoyakuwan w a ai mai Dien and women's roies are ambiguous ] " tha1's a riice ivord
"aimai na koto [mmbiguous (hing]" language use (lexis): reflective: English
 Kci: Task 1 Kci: Task II Raiet 1: Task 1 Rakr 1: Tark il Rnicr 2 Tark 1 Rakr 2:Task Ii
Jurnlnr
]011i112?1= 24.2 22 W
conknu9) 8.7 87 9 8 9
organSion(9) 9 8.6 9 8 8
languagc 8 8 9 9 8 8
mq9) Very weU mittcn ia al1 Vcry wcU wriacn. 1s ibauthor reaiiy a T k r e is somcthing to Mat bc a ncar native.
commcn~s aspccu. Iratocil ExceUcnl Isn'l shc imorove in che 3rd
[-noir pffodira a native spcakcr7
l o i ür 'iepoam'
cotriii]
Jmc
lola1ou 252 24 2P
conirny9) 8.1 84 8 7
organizaîion(9) 8.6 86 8 7
language 83 U 8 6
usq9) Wei1 ~ r i l kTbt
l~ Clw evidcnot of UJC Tbc 3rd pangraph ir In ordcr io communlcatt
commenu vaabulary imroduccd in iidbor's e f f o n Tbc auboti suppcrrcd 10 bt LiU;iog ber Id-, tbe writcr
~ ~ ; clms l arcl uxd ~ ; ~point skillfulty &vclopcd,
~ skillfully. about ncgative points. bld ii discussu many h h g s ia-
c*irimi Vcry ftw grammatical uslng îbc expnssions in f r t discws p i i l v c &fi
misulu. Tbe use o f a kugbt in Be clau. poim. Fasy 10 tead
greatci vatiery of
cxpreuions and more h J I
migh impmve ihc asay.
Judlih
toiallal)m 224 22 2 s 2ér
conknU9) 84 7.2 8 8 9
organiLaiion(9) R6 5.1 6 8 9
laquage &I ai 8 7.5 RS
iaq91 Vcry wcll wriuca Ablc b Vcry good wnting. Thc body s e c m sbon in The wrircr mites wcll, but Mint bc a mu native
commenu w a numbtr of abstraci proponion I o the could have given more spcaktr.
',ui: ! , rd~? Somc spokm
wordi. introduction ihoughl to the conlent.
COIU=DI) cxpicsions. A fcw
gramrnaiical rnisiakts, b u
passable.
JOY
iota1 (27) L
u 2 2Q.2 22
7
u
1
conlenU9) 8.4 6.9 8 8
organhlion(9) 9 7.7 7 5 B 5
language 83 7.9 7 7J 7 5.5
urec9i Vcw weii wiitcn in al1 Wcll wriiiea laclurion of Vcry wcll witicn Concise. This essay sccms Good in ihe use of ïbcte is no conncction
:owrcte c w n p l u Thcrc arc some problcms incomplcie. conjunctiom. Easy to icad xtwcco ihe bcginning and
rccommendcd wiih ihc c h i u of and uadcrsidob bc ending, and the wbole
vocabulary, but dbers arc s y is uniocwd
al1 Tinc.

22 UA 1P U
U
8.1 7.8 3 4
8.1 Il 3 3
6.8 5.5 4 3.5
Organization i s oby. 3vidcuce of ihe auihofs From tbt start 1 got an From ihc Tirsi sentence 1 do
Meaningr of somc paru are :ffortful process. An impicssion ihat l i s stu&nt na know w b i l e wnlcr
no1 clear. An a n c m p to use rncmpt io w a viiely of ir bopclcss. WMU 10 Sdy. TOOctyplk.
a variciy of vocabulnry Is vocabulary is laudablt, Necds more cxplanation
laudablc, tbough thty are , b u g b iky are n a dways
not always ustd propcily. ~d pmpcrly.
(table c o ~ i t i r i i t e s )
Cvthy
1o1.11(22) l
u Irf u.2 u
C O I I ~ I I U ~8.) I 6.6 5 5 4
orgYiiraiionl9) 81 68 5 4 4
language 7 61 4 4.5 3
mq9) Wcll o r g m c d essay. An Inurtsiing topic, brd l e T h snidcd mus! k a The wnkr x c m IO posmcsi Bccaw of the many hji
commcnis attmpt Io use a variety of auibds own opinion Chinue spcakrr. S k uiu sujlcicnt voabulary, bu w d in t k -y, 1wi
I":!,rg~?
dimiil
vocabdary is laudablc.
Grammatical mbtakcs u c
duhk Some p m
unintclü gi blc.
to gel away by usiog the
'Chintse wotd + dcaru'
lach grammatical
compctcncc
makc m d n g with effon
But synwctic and olbcr
compicuoiu. But g d exprcuioa enon are disturblng.
~nknc~-sentcnce
conncciiom.
Cdln
l0W11271 a Y KA fi Z
conicni(9) 87 81 5 5 3 9
organizalion(9) 9 81 4 6 2 9
Iangiwge $ 1 6.8 4.5 5 2 7.5
Mrsssge clwly The positive and ncgative The m y could bc a Id 3rammar is alrighi up to Very powcrfd, persuasive.
communicatcd Good points are oot clcarly bcttcr if lbe las1 paragraph .kclaisal lcvei. Bu!
choice of vocabuhy but txplaincd Crypic. wcte funhcr devcloped beyood i b t lcvcl thtre ut
used bppmpnatcly too many mistakcr lhat
somerimcs. ~bscurciht w n k h
co~plexsenences. Somc aiendcd maning

Ch uks
?!l L
a lj U
6 5 6 1 3
63 5 4 3 7
67 5.5 4 5 5
The authoh own opinion Must bc a Chinese speaker. Dotsn'i rhc wnkr have his In spik of many mislaka Easy to rcad through Brd
hing. i fannd help but laugh hi om opinion? Probably in langllage we, the ruay ;becoatenl is vcry simple
ruch a phrase like because of the lack of IIe s y to rcad HLmus1 k uid childish
'shi&lJoo&klm['ctotic' in grammatical compekncc, i good wriirr ln Ll sr well
lapanwe]. Ench paragraph Lbe dacripion of k p a m c
k too shn. pcoplc'r tire bai bccomc
vcry flat and one-slded. - - -

;UI P-U 1P
6.6 7 3
81 8.1 5 5 5
7.6 8.8 5.5 7 7
Wcll written as a wholc, Wcll mitien but h c Iwas troublcd by the Lack it look likc an unfinishcd The Tint half I a h likc an
ibougb ibcrc are Mme authoh own opinlos >f cobcrcncc, ratber t h piece. %y wriuen by a nalive
inua-xnicniial misiakrr. missing. grammatical problcms. ipeaker, But t k iatter hdf
Use of a greakr v h e t y of ieeatobc aJFLIcmch
xPtcncc.codiog mithg

rea
66
u
b
lu
5
W
1
6.8 5 a ! I
5.9 1.5 15 L5 1.5
Suddcn change of lhe k r c b 'I', b u are thcre Wbat iil said in lbe csssy t nec& much ellori io dust bt a Chiaest speaker.
rubjcct maacr. Some parcs 'II*and I was iou mi m i r h the tidr dcrsiaod what tbe wnkr The coaicnt dou not match
unintelligi blc. roubled by tk ' d e d and s ying 10 Say, he lide.
'&' f o m k i n g mixtd in
Grammatical mistalw bc w n e cssay.
conspicuous. Lacking in
 H l
US
6.9
12
6
2.2
8
lx!
6
i%i
9
8.1 72 5 6 7 8
7.6 83 6 8 4.5 85
Very ~ c l wiiuen
l Rut Wcll wriuen bui a Funay iiilc. Swing Ihai bc The cssay lacks a The spokcn and thc wriitn It's a piiy &ai i h cnlling is
sounds a liiilc ioo conclusion miuhg.Uule Lwws about tbc topic wcii conclusion forin of laiiguagc pic m i quik coniplck.
collaquial. evidence of uiilizing ihc is bcry Engiish. Using the inixd. Tbcrc is an Alihough lhe s p o h
cxprcuions iaught in clars. colloqwal expression of ' n . inilucncc of ihc mother laaguage is w d , il is a
de&. tonguc on ihc use of ibc gaod piecc of wiiing.
Japanex hiigugc. Dib the Good linguistic
mewgc is undcrsunikblc. compclcncc ai ihc
discourse levcl.

?.iJ
84
222
6.9
IZ
2
u
9
81 8.1 4 2 8
79 7.9 6 8 65
Vcry W U wriucn. Usc of difkrcnt lonns of The 'de a r r m ' l o m i s 1 sec no uacc of planning iu hlay ix a Chihincsc spcakcr
Exccllcni cboice and use of scatcace-crding cxprcsion used ioo ofrcn and i b i ihis t m y . 1 cannoi tcll DI high JFL proficiency.
vocabulory. U x of more ir cacouaged Little botha me.The wntcr h i+hrtibc wriicr w a m to
complcn xnenccr to bc evidrnce of lnilizing ~ h c dixourse-lcvcl ~aihrrthan say. He musi bc a p r
cncortngcd. cxprcssions (au~hiin clau. grammatical, problenu. wliicr in t1 as WCU.

brl 11 t2
B.4 J 7 B
K6 4 7 B
aI 1 6.5 5
Vcry w l l oiganirtd and Fhc iitlc is somcwhai The conlcnr is suuicwbi iVcll wiiiieu (rom his awn
wcll cxpicucd ushg r mislcading. Ir's n a clcm to sbaiiow. Ba kcarrrc of ibe ~rspcciive.Grammahl
rruicty of vocabulary. :kend ibat thc writci is witcr's high lcvcl of n i s h h do m i boihcr me
Somc intra-scnicaiial od icfcning KIJapancsc liaguislic compciencc, it's oomuch
grammaUc.1mistaka. iociciy in gencral blrt thc casy to ~ c a dhis rssay.
hcaidcal world in Japn.

Nok Figures 111 i~idicsarc for ihr essays r a d and corrccicd by a native s p a k c r of Japüi~csckforc subriiissioii. Asicrisks (.) iiidicaie tliiii the datri are 1\01avciilablc.
 Appendix P: Forty-four Items of Student Data

Table PI
Backgroi~rzdQuestionnaire Data (Zterru 1- IO)

1 Item & Options I

I
1. Ethnicity
( I ) Japanese
(2) Chinese

( 1) Japanese
(2) Chinese

3. Written LI
( 1) Japanese
(2) Chinese
(3) English
4. Self-rating o f L 1 wri ting
(1) ~r
(2) fair
(3)
(4) excellent
5. importance of Japanese study
(1) not so important
(2) important
(3) very important
6. Length of residence in Japan
( 1) less than 6 months
(2)over 6 months
7. Self-nting of o v e d Japancsc
(cf. peers)
8. Self-rating of overail J a p e s e
(cf. natives)
9. SeU-rating o f Japanese writing

Note. Options for Items 7 - 9 are the same as the ones for Item 4.
 Table P2
Lmtguage Skilis Assesment Data (Irerns 1 1 - 13)

1 Item & Options

1 1. L 1 essay score
( 1) P r (0-9)
(2) fair ( 10- 19)
(3) g d (20-29)
(4) excellent (30-36)
12. JST rating
(1) intermediate
(2) advanced
(3) sqerior
13. Japanese essay score
(1) P r (0-9)
(2) fair ( 10- 19)
(3) good (20-29)
(4) excellent (30-36)

Table P3
Interview Data (Items 14-16)

use of lanepage
(2) langupe exercise
15. Perceived area of difficulty in
E L writing
(1 ~ ~ ~ p g e
(2) orgaaization
(3) la&uage & organization
16. Perceived area of Iearning thru
JFL w riting
(2) content
(3) language & content
 Table P4
Posl- Writirig Qirestiomaire Data (Item 1 7-22)

1 Item & Options

17. Level of satisfaction with Task
-
Jure
2
1 essay
(1) not so satisfied
(2) satisfied
(3) very satisfied
18. Level of difficuity in idea
-
3

I
generation for Task 1 essay
(1) =Y
(2) not so difficult
-
. .

(3) difficult
19. Level of difficulty in language 3
use for Task 1 essay
(1) -Y
(2) not so difficult
(3) difficult -
2
20. Level of satisfaction with Task
II essay
(1) not so satisfied
(2) satisfied
(3) very satisfied
2 1. Level of difficulty in idea
-
2
generation for Task II essay
(1) =y
(2) not so difficult
(3) difficult -
3
22. Level of difficulty in langage
use forTask II essay

I
I)
(1) -Y
(2) not so difficult
3 difficult
 Table P5
Tlrink-Aloud Protocol Data (Items 23-30)

1tem & Options

23. Reflective thinking about gist
for Task 1 essay
(1) O - 20%
( 2 ) 20.1 - 30%

1 1
(3) 30.1 - 40%

1 24. Reflective thinking about 1

kmguaipruse for ~ s k - essay
I y
25. Reflective thinking a h u t
writin ocedures for Task 1 essa
26. Thinking about in L 1 while
writing ~ a s 1kessay
27. Reflective thinliing a b u t g i s t
for Task ii essa

29. Refiective thinking about

writing procedures for Task II essay
30. Thinking about in LI while
1 w r i t i n ~ a s II
k essa): I
Note. Options for Items 24 - 30 are the same as the oncs for Item 23.
 Table Pt5
Japariese Ersay Assesstneat Dafa (Items 31-44)

1tem & Options

3 1. Task 1 essay C0NTEh-ï score
by Kei
(1) poor (0 - 2.4)
(2) fair (2.5 - 4.9)
(3) good (5 - 7.4)
(4)excellent (7.5 - 9)
32.Task 1 essay ORGANIZATION
score by E;ei
33. Task 1 essay LINGLIAGE USE
score by Kei
34. Task 1 essay CONTENT score

36. Task 1 essay UNGLTAGE USE

score by Raters
37. Task II essay COhTENT score
by Kei
38. Task II essay ORGANIZATION
score by Kei
39. Task II essay WNGLTAGE USE
score by Kei
JO. Task II essay COhTEhT score
by Rater
1 JI. Task II essay ORGANIZATION

score by Raters
42. Task II essay LANGUAGE USE
score by Raters
43. Task 1 essay SOLO level
(1) prestmctural
(2) unistructurd
1 (3) multistructurril
(4)relationai
44.Task il essay SOLO level
( 1 ) prestnictural
(2) unistnicturai
(3) multistructural
(4)relational
Nore. Options for Items 32 - 42 are the same as the ones for Item 3 1.
Appendix Q: Results of Dual Scaling (Analysis 4(a) and Analysis 4(b))

Table Qi
Major Con~ibi<tors
and Their Bipolar Optiorrr (AtlaZysis 4(a). Solictiotr 1 )

item description <-- ------ ----- >

J ! 3 rating in termediate superior

'orgrinization' score by rater in Task 1 fair excellent
kpanese essay score lair excellent
level of satisfaction with Task 1 essay not so satisfied very satisfied
'language' score by rater in Task I fair excellent

oral L1 Chi nese Japanese

SOLO level in Task 1 prestmctural/unistnictural relationri1

'content' score by rater in Task II goodifair excetlent
'language' score by Kei in Task II g d excellent
SOLO level in Task II prestmctural/unistmctunl mu1tistructural)

self-ratingof overall Japanese (cf.native) Poor fair

'content' score by rater in Task I Tai r escellent
ethnicit y Chinese Japanese

self-rating of overdl Japanese (cfpers) P r good

'language' score by Kei in Task 1 good excellent

'organiwtion' score by rater in Taçk II fair excellent
perceived area of learning thru wrïting "nothing*kontent language & content
'language' score by rater in Task 11 fair excellent
'orgmization' score by Kei in Task 11 twd excellent
written L1 Chinese English
 Table Q2
Major Contribitlors and nwir Bipolor Optiow (Atidysis #(a). Sofufiort2 )

item description
'Iringuage' score by rater in Task 1 fairlexcellent
'orgmization' score by rater in Task II fairiexcellent
lcngth of residence in Japan l e s thm G months

JST rating intennediatdsuperior

self-rating of Japanese writing (cf.peerç) P r
Japanese essay score fairlexcellent
ethnicity ChinesdJapanesc
IeveI of satisfaction with Task 1 essay not solvery satisfied

level of dif. in idea genention in Task 1 difficult

'content' score by Kei in Task 11 excellent
SOLO IeveI in Task 11 prestnicturdlreiationai

SOLO ievel in Task 1 presuucnirdlrelational

perceived area of problem in witing Jpn language & organization
oral LI ChineseiJapanese
importance of Japanese study very important

perception of JFL wriiing rneaning rnaking

self-nting of LI witing rair
 Table Q3
Major Con~ribrmrsand ïkeir Bipolar Optioris (Alialysis 4(b).Solution I )

item description <---------- ----- --------- --->

Japanese essay score fair excellent

'language' score by nter in Task [ fair excellent

'organization' score by nter in Task 1 fair excellent

JST rating intermediate superior

seIf-rating of overall Japanese (cfmtive) P r fair

ratio of thinking aloud in LI in Task 1 UT-% -2Wo

ratio of thinking aloud in L1 in Task II JO-% -20%

ievel of satisfaction with Task 1 essay not so satisfied very satisfied

SOLO level in Task 1 prestmctural/unistnictural relational

perceived a r a of learning thm writing "nothing" language & content

'content scores by nter in Task 1 fair excellent

oral LI Chinese Japanese

SOLO level in Task II presiructuraJ/unistructural multistnrctural

'organization' score by rater in Task II fair excellent

reflective thinking about idca in Task 1 -2000 Je%

ethnicity Chi nese Japancse

'language' score by Kei in Task 11 excellent

'content' score by rater in Task II g d fai r excellent

reflective thinking about lang in Task I 3 W o 20-3Wo

self-rating of overall Japanese (cf.peers) POOC g d

difficuhy in idea genention in Task 1 not so difficult difricuit

wrïtten L1 Chinese) English

'content' score by Kei in Task II good exallent

'language' score by Kei in Task 1 good exceHent

 Table Q4
Major Coritribiitors and ï3eir Bipolar Options (Atialysis 4(b). Sol lition 2 )

item description -------- -- >

'language' score by nter in Task I far

'organization' score by rater in Task [I go=d fair

J!Yï nting advanced intermediate

importance of Japanese study important very important

length of residence in Japan over 6 months less than 6 months

'organization' score by rater in Task 1 g d fair

'content' score by rater in Task 1 g d fair

reflective thinking about in 'lirk 1
'ïaippW 20-30% 3040%

self-rating of overall Japanese (cf.peers1 fair POO'

satisfaction level in Task 1 esmy satisfied not so satisfied

reflective thinking about gist in Task 1 3 w o JOoro-

self-rating of Japanese writing (cf. p e r s 1

difficulty level in idea generation in T&I
perception of JFL writing
-
P d

language exercise
difficult
meaning making

Japûnese a s a y score good lai r

SOLO level in Task II unistructunl relationallprestructunl

SOLO level in Task i unistructunl mu1ti-lpre-stmctunl

'content' score by Kei in Task iI .!Pd excellent

ethnicity AngleSaxon Japanese

self'-rating of L1 witing g='d fair

oral LI English Chinese

reflective thinking about rw~
inTzik iï m o - -2Wo

'language' score by rater in Task II exellent fair

'organization' score by Kei in Task 11 good eircellent
 lMAGE NALUATION
TEST TARGET (QA-3)

APPLIED i IMAGE. lnc

-.-
-
51653 East Main Street
-
-
---
=
-
-
a
I I i
Rochester, NY 14609 USA
Phone: 7161482-0300
F a : 71W88-5989

O 1993. Appiied )mage. Inc.. All Rights Reserved