AB STR A CT
This is a cross-linguistic w ithin-subject designed exploratory study intended to investigate the relationships betw een com prehension strategies in first (L I) and second (L2) language reading. T h e research questions are as follow s: 1. W hat specific co m prehension strategies w ere u sed by these C hinese readers reading ex p ository texts in C h inese (1.1) and English (L2)? 2. T o w h at extent did these C hinese readers use sim ilar o r differen t
com prehension strategies w hen reading in C hinese (L I) and in E nglish (L 2)? E ig h t C hinese graduate students at the U niversity o f V ictoria participated in this study. T hink-aloud technique and com prehension strategy check-lists w ere used jo in tly to collect d a ta on the participants’ co m p reh en sio n processes when they read expository texts in C hinese ( L I ) and English (L2). T h e p a rticip a n ts’ reading and think-aloud perform ances w ere audio-taped. The think-aloud tapes w ere transcribed afterw ards to obtain real-tim e d a ta o n the
participants’ co m p reh en sio n processes. T h e com prehension strategy ch eck-list w a s u se d as a backup data source and an instrum ent to triangulate the d ata o btained from the think-aloud protocols. T he resu lts o f data triangulation indicated that the tw o data collection instrum ents were valid and reliable fo r the collection o f data o f p a rticip a n ts’ com prehension processes. Seven findings w ere o b tain e d fro m the study. T he findings su g g est that p articip an ts used the sam e types o f strategies to process reading i r the tw o languages. T h e script o f the ianguage did n ot c h an g e the p attern o f participants’ co m prehension p rocesses. H o w ev er, the script o f the lan guage d id ch an g e the frequency o f occurrences o f som e type o f
strategies. T h at is, so m e participants d id use som e strategies m o re freq uently in L I than in L2 o r v ice versa. T h e findings o f this study supported and w ere su p p o rted by C u m m in s’s (1 9 8 4 ,1 9 9 1 ) C o m m o n U n d erly in g P roficiency H y p o th esis a n d G o o d m a n ’s (1 9 7 0 ,1 9 7 1 ,
iii
tended to use the sam e types o f strategies w hen reading in L I and L 2. B ased on the findings, the researcher concludes that there is a strong relationship betw een
com prehension strategies in L I and L 2 reading. T he evidence suggests that there is not a single set o f strategies that is m ore effective than others fo r all learners. W hat kinds of com prehension strategies that a reader used depend on the text clues available to the reader and on p rio r know ledge that a reader brought to the reading task. O n the basis o f the results o f the study, the researcher contends that the m ost effective strategy training m ay consist o f encouraging readers to becom e m ore aw are o f their existing com prehension strategies and aw are o f their strategy use. L im itations o f the study w ere addressed. First, the num ber o f research participants w as relatively sm all in term s o f generalization. It is not possible to m ake generalizations from a group o f eight participants. A further lim itation is that the participants in this study w ere reading aloud, w hich is certainly not their norm al m ode o f reading. It is possible that the think-aloud perform ance itself m ay have affected the nature o f data collected. T h e results shed light o n the existing body o f theoretical know ledge about com prehension processes through w hich researchers and L 2 instructors m ay look anew at ou r stu d en ts’ perform ance and at the instructional m ethods and the techniques w e have adopted.
D r. T.D . Johnson, S u p ervisor (D epartm e n t o f C om m unication and Social F oundations)
D r. J. K a r k e r e D e ^ S ^ f it a T M em jb erjP ep artm en t o f C om m unication and Social Foundation^)
D r. J. A nderebn, D epartm ental M em b er (D epartm ent o f Psychological F oundations in E d u c a tio ti)/
Dr. D ^B rySnt, O utside M e m b e rfD ep a rtp ie n t o f A sian and Pacific Studies)
D r. L. G underson, E x ternal E x a m in e r (Faculty o f E ducation, U niversity o f B ritish C olum bia)
T A B LE OF C O N T E N T S C hapter P age Title Page A bstract 1! Table o f C ontents iv Lists o f Tables v Lists o f G raphs vi
A cknow ledgm ents vii
D edication viii
C hapter One: In tro d u ctio n 1
1.1. P u rp o se o f Study 1
1.2. Statem ent o f th e Problem and N eed for the Study 3
1.3. Theoretical F ram ew o rk and R esearch A ssum ptions 4
1.3.1. Existing L2 R eading Theories 4
1.3 .2. R esearch A ssum ptions 4
1.4. R esearch Q uestions 5
1.5. Significance o f th e Study 6
1.6. Definition o f T erm s 7
1.7. Lim itations o f th e Study 9
C hapter T w o: L ite ratu re R eview 12
2.1. T heories U nderlying th e R elationship betw een L I L iteracy Skills
and L2 L iteracy D evelopm ent 12
2.1.1. T he T ran sfer H ypothesis o f R eading 12
2.1.2. T he C om m on U nderlying Proficiency T heory 16
2.1.3. T he R eading U niversal H ypothesis 19
2.1.4. Sum m ary 23
2.2. R esearch on C om prehension Strategies and Strategy U se in L 2 R eading 23
2.2.3. R esearch on C om prehension Strategies 23
2.2.2. R esearch on S trategy U se 28
2.2.3. Sum m ary 31
2.3. T h e R elationship b etw e en C om prehension Strategies in L I and L 2 R eading 31
2.3.1. M ain A rg u m en ts in L2 R eading R esearch 31
2.3.2. H o w d o L I R eading and L2 R eading R elate? 33
2.3.3. Sum m ary 38
2.4. Sum m ary and T heoretical Q uestions 39
C hapter Three: M eth o d o lo g y R eview 42
3 .1. D a ta C ollection A pp ro ach es O ften U sed in L2 R eading R esearch 42 3 .2. V erbal R ep o rtin g as a D a ta C ollection A pproach 44 3.2.1. T hink-aloud as an Introspective D ata C ollection M e th o d 45 3 .2.2. Q uestionnaire as a R etrospective D ata C ollection A p p ro ach 51
V
3.3. Summary 54
C hapter F our R esearch D esign 57
4.1. Pilot Studies 57
4.2. R esearch P articipants 59
4.3. C riteria for Selecting R eading M aterials 64
4.4. D ata C ollection Instrum ents 70
4.4.1. Instrum ents 70
4.4.2. T riangulation o f D ata 74
4.5. Training in th e T hink-aloud Perform ance 75
4.6. D ata C ollection P roced u res 76
4.7. D ata Analysis 77
4.7.1. T ranscribing 77
4.7.2. C oding 78
4.7.3. Frequency o f O ccurrences 80
4.8. In ter-rater Reliability and In tra-rater Reliability 81
4.8.1. In te r-ra te r Reliability 81
4.8.2. In tra -ra te r Reliability 84
4.9. The Role o f th e R esearcher 85
4.10. Definition o f T erm s 86
C hapter Five: R esults 88
5.1. Findings to R esearch Q uestion 1: W hat C om prehension S trategies did
T hese Chinese R ead ers U se W hen They R ead E xpository T exts in Chinese ( L 1)
and English (L 2)? 89
5.1.1. Finding 1: S trategies Identified in the P articipants’ C om prehension
P ro cesses W hile R eading Chinese and English E xpository T exts 89 5.1.2. Finding 2 : T he Frequency o f Strategy U se in LI and
L2 R eading T asks 113
5.2. Findings to R esearch Q uestion 2: T o w hat extent did T hese Chinese R eaders U se Sim ilar or D ifferent S trategies W hen R eading Expository
T ext in L I and L 2? 117
5.2.1. Finding 3: T otal S trategies and Frequency o f Strategies Used
by th e G roup in L I and L2 R eading Perform ances 1 17 5.2.2. Finding 4: S trategies U sed in both L I and
L2 R eading P erform ances 118
5.2.3. Finding 5: T h e F requency o f S trategy C ategories 1 19 5.3. R esults o f Individual R e a d e rs’ C om prehension P ro cesses 121 5.3.1. Finding 6: E ach Individual R e a d e r’s S trategy R ep erto ire 1 2 1 5 .3 .2. Finding 7: Frequency o f S trategy U se o f E ach Individual R ead er 124 5.3.3. D escription o f Individual R e a d e rs’ C om prehension P ro cesses 125
5 .4 . Sum m ary o f Findings 144
C hapter Six: D iscussion
6.1. Findings to R esearch Q uestion 1: W hat C om prehension S trategies Did
these C hinese R eaders Use W hen They Read E xpository T ex ts in C hinese ( L I )
and English (L 2 )? 146
6.1.1. Finding 1: T w enty-four C om prehension S trategies
O bserved in the Study 146
6.1.2. Finding 2: T he Frequency o f Strategy U sed in LI
and L 2 R eading Tasks 150
6.2. Findings to R esearch Q uestion 2: To W hat E xtent D id T hese C hinese R eaders U se Sim ilar o r D ifferent Strategies W hen R eading E xpository
T ext in L I and L2? 152
6.2.1. Finding 3: T otal Strategies and Frequencies o f S trateg y U sed by
th e G ro u p in L I and L2 R eading P erform ances 153 6.2.2. Finding 4: Strategies U sed in B oth L I and L 2 R eading P erform ances 153 6.2.3. Finding 5: T h e Frequency o f Strategy C ategories 154 6.2.4. Finding 6: E ach Individual R ead er’s S trategy R e p e rto ire 155 6.2.5. Finding 7: Frequency o f S trategy U se by E ach Individual R ead er 156 6.3. R eading C om prehension H ypotheses and the Findings o f T his S tudy 158
6.3.1. T h e R eading T ransfer H ypothesis 158
6.3.2. T h e C om m on U nderlying Proficiency H ypothesis 159
6.3.3. T he R eading Universal H ypothesis 161
6.4. Sum m ary 162
C hapter Seven: C onclusions and Im plications 166
7.1. G eneral C onclusions 166
7.2. Im plications 169
7.2.1. Im plications fo r L2 R eading Instruction 169
’ 7.2.2. Im plications fo r L2 R eading R esearch 171
7.3. L im itations 172
References 175
vii
List o f Tables
Table Page
Table 1 Table o f C riteria and R ational for Selection o f Participants 60 Table 2 T able o f B ackground Inform ation on the Participants 63
Table 3 T able o f C riteria for R eading M aterials 67
Table 4 T able o f D ata Triangulation 74
Table 5 T able o f In ter-rater Reliability 33
Table 6 T able o f In tra-rater Reliability 84
Table 7 Table o f F requently-used Strategies in L I and L2 reading tasks 114
Table 8 T able o f Infrequently-used Strategies 116
Table 9 T able o f G roup Total Strategies/Frequencies: M ean, SD and
P assage L anguage 118
Table 10 T able o f S trategies U sed in B o th R eading Perform ances 119 Table 11 T able o f G ro u p F requency and P ercentage o f Strategy C ategories 120 Table 12 T able o f E ach Individual R ead er’s C om prehension Strategies
in L I and L2 122
Table 13 T able o f Individual Frequency o f Strategy use in LI and L2 Reading 124 Table 14 T able o f P i ’s Frequently-used Strategies and Frequency o f
S trategy U se/P ercentage 130
Table 15 T able o f P 3 ’s F requently-used S trategies and Frequency o f
S trategy U se/P ercentage 134
Table 16 Table o f P 4 ’s Frequently-used S trategies and Frequency o f
S trategy U se/P ercentage 138
Table 17 T able o f P 8 ’s Frequently-used strategies and Frequency o f
G raph G raph 1 G raph 2 G raph 3 List o f G raphs Page T otal S trategies in Chinese and English R eading 122 T otal S trategies used in B oth L I and L2 R eading T asks 123 Frequency o f O ccurrences o f Strategy Use in Chinese and
ACKNOWLEDGMENTS
A special acknow ledgm ent is reserved for my supervisor, Dr. Johnson. H e w as always available fo r consultation, guidance, creative suggestions and support, w hich w ere very im portant to me.
A ppreciation will g o to Dr. A nderson, Dr. Bryant, Dr. Evans, and Dr. H arker not only fo r their constructive suggestions, but also for the superlative training they have provided rne w ith ov er th e last th ree years. I am fortunate to have w orked w ith such understanding and helpful people.
Sincere th an k s will be extended to the University o f V ictoria for offering me the fellowship th ro u g h o u t the tim e I pursued my doctoral degree.
I am very grateful and in debted to sixteen Chinese graduate students w ho helped me and donated th eir tim e to participate in this study. W ithout their cooperation and support, I could not have finished my dissertation successfully.
I will express my thanks to my family and friends. They w ere alw ays there to nourish me, never letting m e consider for a m om ent that I w ould not successfully com plete this m anuscript.
Thanks will go to m y husband, for being reliable source o f encouragem ent and m otivation. Special thanks will also g o to my son for his understanding, patience and support.
This dissertation is dedicated to my parents. I will be eternally grateful for their love. It is through their love th at I have developed the strength to continue to w o rk for im provem ent in education.
C h ap ter One: Introduction
1.1. P u rpose of th e Study
T he p u rp o se o f this research is to study the relatic :os betw een reading com prehension strategies in first language and second language reading. M ost first language (L I) read ers achieve reading proficiency from many thousands o f hours o f practice over a lifetim e o f reading. It is am azing that m any second language (L 2) readers can achieve a sim ilar level o f reading proficiency in a second language from few er hours o f form al reading practice. D o L I literacy experience and skills facilitate reading
com prehension in a second language? W ithin the literature on L2 reading, there is relatively little research on th e relationship betw een L I and L2 reading, and less o.i the relationships b etw een L I literacy and L 2 literacy skills within educated adult L2 readers.
R esearchers in second language reading have recently, as B lock (1992) claimed, "ceased debating w h e th e r reading is a boU om -up, language process or a top-dow n,
knowledge' based process" (B lock, P. 319). A num ber o f significant factors influencing L2 lead in g , (e.g., effects o f orth o g rap h ic stru ctu re o f the first language, rhetorical structure, cultural factors) have been investigated, and som e insights have been obtained. Previous research has enabled us to u nderstand m ore about the im pact o f cultural factors, content and rhetorical stru c tu re on th e com prehension o f L 2 w ritten texts (Carrel, 1984a; 1984b; Johnson, 1981). R esearch o n schem a theory has helped us to understand why L2 readers som etim es fail to com prehend an L2 passage (Johnson, 198 ; Lee, 1986). H ow ever, this
do not know w hat L2 readers do w hen they realize they have failed to com prehend w ritten m aterial or w hether L2 readers use the com prehension strategies obtained in their LI reading experience to solve their L2 reading problem s. N o r do w e know if they use different strategies in L 1 and L2 to solve their com prehension problem s, o r w h ether L 1 literacy affects L2 readers' com prehension strategies.
I f second o r foi eign language instructors are to deal efficiently w ith readers o f various linguistic backgrounds, they need to know m ore about the relationship b etw een L I literacy and L2 reading. M o re research is th erefore needed on the relationship betw een reading com prehension strategies in L I and L2 o f particular individuals.
This study w as designed to explore the relationship betw een com prehension strategies used in L I and L2 reading, to find o u t w hether Cum m ins' (1984, 1991)
C om m on Underlying Proficiency m odel can be applied to educated adult L2 readers, and to determ ine w h ether reading com prehension is a cognitive universal process. In o rd er to know m ore about academ ic L2 readers' com prehension p rocesses and to b e tte r understand the C om m on U nderlying Proficiency H ypothesis (Cum m ins, 1984, 1991), and th e
Universal H ypothesis (G oodm an, 1970, 1971, 1973) the goal " ' the intended study is tw o-fold:
(1) to provide descriptive inform ation on the reading com prehension strategies used by academ ic L2 readers "/ho are Chinese native speakers and w ho are functioning
(2 ) to ascertain to w hat extent these readers' com prehension strategies are similar o r different when they read in their native language and in their second language.
1.2. S tatem ent o f th e P roblem and N eed for the Study
In today's w orld, m ore and m ore people need to be literate not only in their native language, but also in a second language. H ow ever. little inform ation on reading
com prehension pro cesses in a second language has been obtained. W e are still uncertain if proficient readers o f English as a second language (E S L ) use similar com prehension strategies w hen reading in th eir native language and in the second language. M oreover, w e have little know ledge ab o u t th e actual com prehension processes o f ESL readers w ho do obtain acceptable scores on proficiency tests (e.g., T O E FL ) and function quite well in an academ ic environm ent. C onsidering o u r lack o f know ledge ab o u t this type o f second language reader, investigation is needed to determ ine how these already-literate reader, read in a second language, w h at th e relationship is betw een their L I reading
com prehension strategies and their L 2 reading perform ance, and w hether these reading com prehension strategies a re universal and interdependent across languages, o r w hether there a re generic p ro cessing strategies th at accom m odate both first and second languages (B ernhardt, 1991). A nalyses o f the relationship o f th e com prehension strategies betw een L I and L 2 reading m ay lend insight into the C om m on U nderlying Proficiency H ypothesis (Cum m ins, 1984, 1991), and the U niversal H ypothesis (G oodm an, 1970, 1971, 1973).
1.3. Theoretical Fram ew ork and R esearch A ssum ptions
This study w as based primarily on the follow ing existing theories and research assum ptions w hich are related to L2 reading:
1.3.1. E xisting L2 R eading T heories
1. Psychological perspective. Esling and D ow ning (1986) suggested the possibility o f literacy transfer across languages and, based on this assum ption, put forw ard their transfer hypothesis. A ccording to this hypothesis, the skill o f literacy, Si'.ce the skill o f oracy, is learned only once in an individual's lifetime, th o u g h he o r she may transfer those skills to o th er languages.
2. C ognitive perspective. T here is a cognitive/academ ic proficiency th at is com m on fo r all languages (Cum m ins, 1984). A lthough th e surface orthographic structures o f languages m ay be different, an underlying proficiency is presupposed w hich is com m on across languages.
3. Psychoiinguistic perspective. T he prim ary goal o f reading is com prehension. T he com prehension process is th o u g h t to be universal since no m atter in w hich language, the ultim ate p u rpose o f reading is the same. G oodm an (1971) suggests th at the reading process cannot vary to any extent from one language to an o th er since the key question is how m uch background know ledge the reader brings to th e specific reading task.
1.3.2. R esearch A ssum ptions
This study w as based on the follow ing research assum ptions indicated in the literature, and attem pted to m eet th e needs o f L2 reading research.
1. Second language reading research has not sufficiently em phasized that first language literacy is a significant com ponent in second language reading (B ernhardt 1991),
2. The m ajor problem o f previous studies o f L2 reading is that they provide insufficient inform ation ab o u t the relationship o f reading com prehension strategies betw een L I and L 2 reading in the sam e individuals (A lderson, 1984).
3. Insights into th e relationship betw een L I and L2 reading com prehension processes can only com e from em pirical studies using a cross-linguistic w ithin-subject design. Such studies co m pare the perform ance o f the sam e individuals in various L 1 reading tasks to th eir perform ance in similar L2 reading tasks (Hulstijn, 1991).
4. The m ethodological assum ption is th at utilizing a think-aloud technique during reading perform ance and a com prehension strategy check list im m ediately after the reading task will provide realistic descriptions o f readers' com prehension processes.
1.4. R esearch Q uestions
F o r the p u rp o se o f th e research, and based on the theoretical fram ew ork and research assum ptions, th e research questions are addressed as follows:
1. W hat specific com prehension strategies w ere used by these Chinese readers reading expository tex ts in C hinese ( L I ) and English (L2)?
2. T o w h at extent did th ese C hinese readers use similar o r different com prehension strategies w hen reading in C hinese ( L I ) and in English (L2)?
1.5. Significance o f the Study
A crucial question concerning the n ature o f the effect o f LI reading
com prehension strategies on L2 reading rem ains unansw ered. This question appears to be o f great im portance in tw o aspects. First, w e stand to gain in o ur theoretical understanding o f the notions o f L2 proficiency in general and L2 reading proficiency in particular if they could be linked w ith existing L 2 reading theories such as the C om m on Underlying
Proficiency H ypothesis (C um m ins, 1984; 1991), and the R eading U niversals (G oodm an, 1970; 1971; 1973). Second, such research m ay hold im portant consequences fo r L2 instruction. This kind o f research m ay enable us to understand L2 reading com prehension processes better, and provide L 2 researchers and L2 instructors w ith insights into L2 readers' reading com prehension processes. C onsequently, such research m ay also change o u r view o f L2 reading instruction.
Even though L2 reading research has begun to attend to com prehension processes (B lock, 1986; 1992), inform ation on th ese com prehension processes is still limited. In addition, m ost previous research ers draw assum ptions about readers' com prehension strategies by exam ining readers' com prehension scores, th a t is, how m any c o rre c t answ ers the students provide for a cloze te s t o r a m ultiple choice test (Carrell, 1989, C arson et al.,
1990). Strictly speaking, since th ese kinds o f instrum ents test the end-product o f
com prehension rather than th e pro cesses o f task perform ance, w e still do not know w hat L2 readers are actually doing w hen they p ro cess the tex t to w ard com prehe on. It is essential th at w e learn m ore a b o u t th e strategies w hich readers use to m ake sense o f expository texts in bo th their L I and their L2. This study attem pts to address th at need.
This research is an ex ploratory study intended to exam ine the relationship betw een LI and L2 reading com prehension strategies w ith academ ic L2 readers. Theorizing about reading com prehension strategies in both L 1 and L2 reading activities requires an
experim ental design which directs research attention to ongoing process. Ericsson and Simon (1984) suggest that p ro cess-tracing m ethods provide a conscientious analysis and syn hesis o f the strategies readers use. This study portrays readers' verbalizations o f com prehension processes by using th e think-aloud technique as a d ata collection
procedure, and a reading com prehension strategy check list to collect additional d a ta so as to com pensate fo r som e w eaknesses o f the think-aloud protocols (w hich will be discussed in C hapter 3) and to triangulate the d a ta obtam ed through the think-aloud protocols. By collecting the co n current d a ta o f the sam e individual readers reading in C hinese ( L I ) and English (L2), the research er obtained m ore insights into how these L2 readers processed the tex ts in the tw o languages concerned, w hat specific strategies they used, and to w hat extent these strategies w e re sim ilar o r different in L I and L2 reading. Thereby, this study helps us understand better th e C om m on Underlying Proficiency H ypothesis and the U niversal H ypothesis.
1.6. D efinition o f T erm s L I a n d L I R e a d e rs
L I refers to the first o r native language o f a person. L I readers are readers w ho read in their native language. In the case o f this study, L I refers to Chinese.
L2 refers to a second language, or u language besides the native language, that a person is learning o r dealing with. L2 readers are persons w ho can read in a language oth er than their native language. In this study “L2 readers” refers to readers w ho are not only proficient read ers in their native language, but also have m astered the fundam ental know ledge o f the ta rg e t language, so th at reading in an L2 has becom e part o f their life. In the case o f this study, L2 refers to English.
Academic L2 Readers
In this study, "academ ic L2 reader" refers to g raduate students w ho are
undertaking g rad u ate studies and research in various subject fields at th e U niversity o f Victoria.
English as a Second Language (ESL)
English is tau g h t and learned in a situation w here the language has social status in the com m unity as an official language, as th e m edium o f instruction in som e sector o f public education. In th e E S L situation, learners can not only learn the language in a form al classroom , but also acquire th e language in their daily lives. A s for reading, E S L learners can read m ore extensively outside o f class, e.g., reading advertisem ents, notices,
new spapers, and b o o k s according to their ow n interests.
Reading Comprehension Strategies
R eading com prehension strategies are cognitive m oves selected during reading to solve problem s o r to achieve particular goals. R eading com prehension strategies (i.e., predicting, sum m arizing, etc.) are the m ental operations involved w hen readers
9
purposefully approach a text to m ake sense o f w hat they read. T hese may be either conscious techniques controlled by the reader or unconscious processes applied autom atically (B arnett, 1989; van Dijk & Kintsch, 1983).
Authentic Material
A uthentic m aterial refers to an article o r a passage chosen from a publication w ithout any editing o r change in language or organizational style. In other w ords, the m aterial retains its original form .
Cross-linguistic Within-subject Design
C ross-linguistic w ithin-subject design refers to a study in w hich the same
individuals are exam ined on both their LI and L2 reading perform ances in similar reading tasks. Such a study com pares the perform ance o f the sam e individuals in similar reading tasks in L I and L 2 so as to find com m onalties o r differences betw een individuals' reading com prehension strategies w hen they deal with different languages.
1.7. L im itations o f th e Study
1. T he study is an exploratory investigation into the relationship o f reading com prehension strateg ies betw een L I and L2 reading am ong C hinese g raduate students. B ecause o f the ch aracteristics o f the subjects in the study, the results obtained may not be generalized to L2 readers o f different educational o r linguistic backgrounds.
2. T he stu d y does n ot assess the subjects' L I reading ability since the participants in th e study are assum ed to be proficient LI readers on acco u n t o f their educational background. I f this study is to be replicated w ith o th er language groups, or w ith subjects
o f different educational backgrounds, and if a standard reading test is available, L 1 reading ability should be m ore fully assessed.
3. T he results obtained in this study are specific to th e unique com bination o f text, task, context, and reader background know ledge. T he resu lts do n ot reflect the entire C hinese g raduate student population since the study w as undertaken w ith particular people reading particular te x ts under particular circum stances.
4. F urther research is needed to understand m ore ab o u t th e relationship betw een readers' L I and L2 reading com prehension strategies. F u rth e r research can be conducted w ith subjects o f different linguistic backgrounds, o r w ith larger groups, using reading m aterials o f different discourse types, such as narrative, fiction/nonfiction.
5. U se o f think-aloud techniques is a tim e-consum ing task. C onsidering the
lim itations on the tim e o f th e participants and the researcher, the research w as done w ith a small g ro u p o f participants. T he generalizability o f a lim ited scale study is alw ays open to question.
6. T he instrum ents o f d ata collection, namely, th e think-aloud task itself and the texts, m ay have affected the participants' com prehension strategies.
In the follow ing chapter, literature relevant to th e present study will be review ed. C hapter T hree will present th e m ethodology review ; C h a p te r F o u r will address th e design o f the present study; C h ap ter Five will present the result o f th e study from both g ro u p and individ al perspectives; C hap ter Six will discuss th e findings o f th e study in the theoretical fram e presented in C hapter T w o. T he study concludes w ith C h ap ter Seven, in w hich
findings are sum m arized, conclusions are developed, and im plications for future research and instructional pedagogy are presented.
C hap ter Tw o: L iteratu re Review
T he relationship o f reading com prehension strategies betw een L 1 and L2 is, o f necessity, a com plex and m ulti-dim ensional topic. T o grasp its fundam ental structure, a vast body o f literature m ust be surveyed. In this chapter, I include current research from three points o f view relevant to th e study: 1) existing L2 reading theories w hich underlie the relationship betw een L I and L 2 literacy skills; 2) reading com prehension strategies and strategy use; and 3) the relationship b etw een L I and L2 reading. Som e contrary
view points are analyzed and discussed as well.
2.1. T heories U nderlying th e R elationship b etw een L I L iteracy skills and L2 L iteracy D evelopm ent
In this section, th ree existing L 2 reading theories w hich analyze the relationship betw een L I and L2 literacy a re included: Esling and D ow ning's (1986) T ransfer
H ypothesis, Cum m in's (1983, 1991) Interd ep en d en ce H ypothesis, and G oodm an's (1970; 1971; 1973) universal hypothesis.
2.1.1. T he T ransfer H ypothesis o f R eading
Esling ana owning (1986), w o rk in g from a psychological perspective, suggested the possibility th a t literacy tran sfer could happen across languages. T ransfer here, from the current researcher's understanding, is being used in a general
psychological sense, and n o t in th e lim ited sense that it has in language acquisition studies, in which it refers to the transfer o f specific linguistic structures. Salom on and Perkins (1988; 1987) offer a fram ew ork for explaining the conditions under which transfer will occu r m ore reliably. They distinguish tw o different kinds o f transfer, the "low road" and th e "high road" (Perkins, et al. 199C, p .295). A ccording to the above researchers, low ro ad transfer occurs spontaneously, w hen patterns o f cognitive processing autom atized in o n e co n tex t are triggered in another by a similarity o f stim ulus conditions. As explained by Perkins' exam ple, sitting behind the wheel o f a truck will trigger one's car-driving skills, w hich fortunately suit the truck context fairly well. H igh road tran sfer happens by m indful decontextualization o f principles from one context and application to another. F o r instance, a chess player, getting involved in politics, m ight a b stract the principle "control the center" from chess and apply it to the political arena (Perkins, et al., 1990).
R eading com prehension in an L2 m ight involve both low road and high road transfer. F o r exam ple, an L 2 reader, facing a text in the target language, may find that his o r h e r experience o f decoding, deriving o r constructing m eaning from the context will fit th e L2 reading context, an instance o f low road transfer. As for the high road transfer, L2 readers m ight use their aw areness o f know ledge held in long-term m em ory to facilitate their reading com prehension.
B ased on research o n skill acquisition in general, Esling and Dow ning apply the findings to reading in particular. T hey argue that reading is a skill that can be
developm ent. Any skill, such as reading, is developed th ro u g h the universal pattern w hich consists o f three overlapping stages: cognitive, m astering, and autom aticity. At the first stage, learners try to figure o u t w hat they should attem p t to do in perform ing the skill. A t the second stage, the learners w ork to perfect th e perform ance o f the skill. Then they practice until they obtain autom aticity. "O nce autom aticity has been
achieved, a skill like reading does not atrophy" (Esling & D ow ning, 1986, p. 56). B ased on this assum ption, Esling and D ow ning have p ro p o sed th e tran sfer hypothesis as follows:
The skill o f literacy like the skill o f oracy is learned only once in an individual's lifetime th ough he o r she m ay transfer those skills to o th er specific languages, (p .60)
T h e transfer hypothesis assum es th at reading, as a skill, can be transferred because reading is also considered to be developed th ro u g h th e th re e typical stages till it reaches its autom aticity. O nce autom aticity is obtained, a rea d e r is expected to apply his reading skills developed in one language to any oth er languages, despite differences in the w ritten system s o f th e languages concerned. H ow ever, research on L2
acquisition in general and L 2 reading in particular has n ot yet provided sufficient evidence to su pport E sling and D ow ning's transfer hypothesis. W hat is notable in research on transfer is that th e great m ajority o f studies fo cuses on such m anifestations as the transfer o f specific linguistic structures th at lead to erro rs (i.e., negative
transfer) and on transfer in language production (R ingbom , 1992). In o th e r w ords, this kind o f research focuses on cross-linguistic similarity or difference rath e r than on language com prehension processes, tha t is, th e involvem ent o f higher-level cognitive
processes such as using previous know ledge in long-term m em ory to facilitate language com prehension.
Esling and D ow ning's hypothesis suggested a transfer beyond formal or linguistic sim ilarity and dissimilarity. W hat they suggested is that once a person acquires reading skill in any language, he o r she can apply the skill to any other language. This hypothesis implies th at reading skill includes "know ing how" (Esling & D ow ning, 1989, p. 57). In other w ords, the focus o f the transfer hypothesis o f reading is on cognition. T h e "know ing how" is also called "procedural know ledge" (O'M alley & C ham ot, 1990), o r operational know ledge (B ernhardt, 1991). T he so-called procedural know ledge includes know ing w hen-w hat, that is, w hen you need to do w hat to help achieve th e goal. This kind o f know ledge can only be developed and acquired gradually th ro u g h plenty o f practice. O nce the procedural know ledge is obtained, it can possibly be applied to other similar situations. In accord v/ith Esling and D ow ning, B ern h ard t also suggests that once a person has learned how to read, no m atter in w hich language, he o r she does not need to learn to play this gam e again.
N onetheless, L 2 reading research has not yet provided any persuasive evidence for the tran sfer hypothesis. Liu et al. (1992) com pare sentence-processing strategies by observing the listening com prehension processes o f bilingual speakers o f English and Chinese. T he findings o f th eir study indicate th at transfer o f sentence-processing strategies occurs am ong th e subjects. H ow ever, the study focuses on linguistic-specific structure, namely, w o rd -o rd e r in th e tw o ianguages concerned. It indicates language transfer rath er th an cognitive strategy transfer. T herefore, w e still do r o t know if or
how this transfer - ppens, o r under w hat circum stances th e transfer happens. Still, research studies in L2 reading indicate that readers perform reading task s similarly in LI and L2 (C larke, 1979, 1980; B enedetto, 1984, am ong others). C an the transfer hvpothesis c f reading explain the phenom ena th at L2 readers perform ed similarly in LI and : ,2 :& v; a rsibie to test w hether a strategy used by a read er w as developed in L I
!irg experience o r w as acquired in L2 reading practice? T he tran sfer hypothesis o f reading m ight be adeq u ate in theory, but not testable in research. W e need further research to explore th e n atu re o f transfer if it happens in L 2 reading. Cum m ins' (1983, 1991) w o rk w ith bilingual children provides a consistent point o f view concerning the possibility o f transfer across languages, and addresses a theoretical position on the issue o f transfer o f literacy skills.
2.1.2. C om m on U nderlying Proficiency T heory (T he Interdependence H ypothesis) Cum m ins (1981, 1983) m akes a strong case for th e tran sfer o f literacy skills across languages. H e states th a t an underlying cognitive/academ ic proficiency exists which is com m on to all w ritten languages. This th eory o f a com m on underlying proficiency is also called the inferdependence hypothesis, and suggests, from a
cognitive point o f view , that there is a cognitive/academ ic proficiency that is com m on for all languages. C um m ins (1983, 1984) explains that th ere is only so m uch space o r capacity available in o u r brains for language o r literacy. H e co m pares th e space o f language o r literacy to a balloon. I f w e divide the space b etw een tw o languages, blowing into the L I balloon will succeed in inflating L I b ut not L2. I f that is th e case,
literacy in neither language will develop properly. For the purpose o f bilingual education, C um m ins argues that w e can b e tte r inflate the L2 balloon by blowing into the L I balloon because the space fo r literacy developm ent is not, and should not be separated. L2 learners can thus benefit from proficiency either in L 1 o r in L2, o r from both, since literacy skills can be seen as com m on o r interdependent across languages. This com m on underlying proficiency is believed to facilitate the transfer o f
cognitive/academ ic skills from o n e language to another. As such, a theoretical
foundation for th e interdependence hypothesis is given. R esearch in bilingual education has provided evidence fo r Cum m ins' C om m on Underlying Proficiency m odel from different aspects, such as studies relating age on arrival and im m igrant students' L2 acquisition (Cum m ins, 1980, 1981); results o f bilingual students' L2 acquisition (Cum m ins, 1984); and studies o f th e relationship betw een L I and L2 proficiency (Cum m ins & M ulcahy, 1978).
T he C om m on U nderlying Proficiency theory expresses the point that experience w ith either language can pro m o te developm ent o f the proficiency underlying perform ance in bo th languages. It also indicates tha* L2 learners, particularly edu cated adult L 2 readers, have tw o channels by w hich to inflate their capacity fo r literacy developm ent. C um m ins (1981) further contends that surface aspects, such as oral fluency, develop separately in LI and L2, but that an underlying cognitive com petence required fo r th e developm ent o f literacy skills is com m on to LI and L2. This suggestion im plies th a t th e level o f L2 reading proficiency is largely predicted by L I literacy com petence. It should b e pointed o u t th a t Cum m ins' theory is
contrary o Esling and D ow ning's in its view o f oracy developm ent. H ow ever, oracy skills are not the focus o f the present study, and will not be discussed further.
W ith respect to Cum m ins' theory, the com m on underlying cognitive/academ ic proficiency has different L I and L2 surface m anifertations. A lthough the surface orthographic structures o f languages are different, an underlying proficiency is
p resupposed which is com m on across languages. In reading, the surface features o f L I and L2 are the objects o f low er-level cognitive processing, having becom e relatively autom atized or less cognitively dem anding, w hereas th e underlying proficiency is one that involves higher-level cognitive processes in reading com prehension.
This m odel, from a cognitive point o f view , lends su p p o rt to th e possibility o f transfer o f literacy skills across languages, since literacy-related skills involve cognitive processes. Cum m ins believes that the com m on underlying proficiency w ould predict that reading instruction in one language not only leads to literacy skills in that
language, but also to a deep er conceptual and linguistic proficiency w hich is Wrongly related to literacy and oth er general academ ic skills in th e o th er language. Cum m ins (1991) presum es th at the unidirectional literacy skill tran sfer occu rs from L I to L2. P revious research (e.g., L am bert & T ucker, 1972; am ong o thers) on bilingual education o f children provides evidence for this interdependence hypothesis.
Providing a transfer occurs in reading across languages, transfer should also occu r in adult L2 readers' reading processes and should n o t be only unidirectional as C um m ins (1991) proposed. T he com m on underlying proficiency m odel suggests th at e ducated adult L2 readers w h o are already literate in their L I m ay have tw o channels
available to them as they develop literacy skills in their L2. They can draw on their literacy skills and know ledge o f literacy practices from their L I, and they can also draw on input from th e second language. T herefore, research on educated adult L2 readers m ight indicate a tw o -w ay transfer, th at is, from L I to L2 and vice versa, since adult educated L 2 readers have som ething to transfer both ways. Liu et al.'s study on listening com prehension processing m entioned earlier provides evidence o f "backw ard transfer" (Liu, et al., 1992, p. 472), th at is, transfer o f listening com prehension skills from L 2 to L I . H o w ever, th ere has been no research do n e to test the Com m on U nderlying Proficiency H ypothesis w ith educated adult L2 readers.
T he reaso n th at w e lack research on tw o -w ay transfer am ong educated adult L2 readers m ight be th at a transfer phenom enon is difficult to test am ong these readers, o r th a t th is phenom enon is neglected or overlapped by o th er factors such as the reading universals proposed by G oodm an (1970; 1971; 1973).
2.1.3. T he R eading U niversal H ypothesis
Cum m ins' C om m on U nderlying Proficiency T heory assum es that the same basic literacy skills underlie th e processing o f m eaning in L I and in L2. In accordance w ith Cum m ins, b u t from a psycholinguistic point o f view , G oodm an, (1970, 1971,
1973), p u t fo rw ard th e R eading Universal H ypothesis. This hypothesis suggests that the prim ary goal o f reading is com prehension. C om prehension can only be obtained by m aking use o f b o th sensory and nonsensory so urces (R um elhart, 1977), or seen and unseen inform ation (B ernhardt, 1991) to reco n stru ct th e m eaning o f a given passage.
D uring reading, readers predict the m eaning o f a text by relating it to their background know ledge; they use o th er strategies available along w ith th e incom ing sensory data. Then th e readers m ay refine, confirm , o r even reject the initial prediction.
In accord w ith G oodm an's suggestion. M iller (19 8 8 ) analyzes the reading com prehension p ro cess and contends that reading com prehension includes visual decoding processes, cognitive processes th at integrate th e inform ation in a tex t w ith general know ledge. O f these, only decoding is specific to a certain language. T he o th er tw o processes m ay be universal to any language. H ence, reading is th o u g h t to be universal in any language since no m atter in w hich language, th e ultim ate p u rpose o f reading is the sam e. G oodm an (19 7 3 ) suggests th at th e read in g p ro ce ss cannot v ary to any extent from o n e language to another, since the key q u estion is ho w m uch
background know ledge (e.g ., linguistic know ledge, co n te n t know ledge, strategy know ledge) th e rea d e r brings to the specific reading task. H e g o es fu rth er and postulates the reading universal hypothesis. T he reading universal hypothesis argues that the reading p ro cess will be m uch th e sam e fo r all languages, w ith m inor variations to accom m odate th e specific characteristics o f the w riting system s and th e gram m atical structures o f each.
R esearch o n orthographic differences in L I and L 2 reading (K oda, 1989; 1990, am ong o th ers) indicates som e difference in w o rd recognition p rocesses in different languages. H o w ever, w ord recognition is ju s t one o f th e prim ary step s o f reading com prehension. R ecognizing w ords is a low er-level and m echanical p rocess com pared w ith th e o th er p rocesses such as relating to previous relevant know ledge stored in the
long-term m em ory. In being consistent w ith G oodm an's hypothesis, R osenblatt (1985) strongly suggests th a t w e need n o t pause here to deal w ith th e prim ary level o f
recognition o f th e printed signs. A s so o n as w e turn to the m atter o f their lexical or sem antic interpretation as sym bols, w e find ourselves involved in consideration o f "what th e reader brings to th e tex t — a fund o f past linguistic, literacy, and life experiences".
B ased on th e universal hypothesis, G oodm an claim s th at learning to read a second language should b e easier fo r som eone already literate in an o th er language, and good L I readers are m ore likely to b e g o o d L2 readers, regardless o f how similar or dissimilar the o rth o g rap h y is, since th e essential characteristics o f th e reading com prehension p ro ce ss a re universal. Som e research on L2 reading (C larke, 1979; 1980; B enedetto, 1984), lends su p p o rt, to som e extent, to G oodm an's universal theory and his claim s’
G oodm an calls fo r research ers "to te s t and challenge the hypotheses in term s o f languages and orth o g rap h ies o th er th an English" (G oodm an, 1970, p. 103 ).
U nfortunately, no research has been do n e in L2 reading to test this hypothesis. H ow ever, previous studies w ith ad u lt second language readers (H auptm an, 1979; B enedetto, 1984; B lock, 1986; 1992), no m atter w hat th e initial intentions o f the studies w ere, have revealed som e sim ilarities in reading com prehension processes across th e languages concerned. W hat is th e underlying cause o f th e similarities? C ould w e interpret th e sim ilarities as being caused by reading universals? O r is th ere a universal pattern th a t proficient rea d e rs follow w hen th ey read in any language?
as L ee and M usum eci (1 9 8 8 ) and B lock (1986, 1992), present som e empirical
evidence suggesting that th e reading com prehension process is a "stable phenom enon". B lock (1986) investigated the reading com prehension strategies o f L 2 readers w ho w ere native speakers o f C hinese and Spanish com pared w ith th o se o f L I readers w hose native language w as English. H e r findings indicate th at the subjects w ith different L 1 backgrounds perform ed similarly w hen reading in English (L2), and th at they also perform ed similarly to the native speakers o f English. B ased on this result, she claim s that strategy use in reading is a stable phenom enon w hich is not tied to specific language features.
B lock's findings im ply that reading com prehension is a universal process. Previous research also sug g ests th at reading is n o t m erely a linguistic process, beyond the prim ary stage o f readm g, say, th e stage o f aecoding or w ord recognition, b u t also a cognitive process. R eading com prehension in any language requires cognitive and m etacognitive skills. It is th e se cognitive and m etaco g ritiv e skills th a t enable a read er to m ake use o f his b ack ground know ledge, to sam ple and predict th e printed
inform ation, to evaluate fo rm er predictions, to . aitor understanding, and to repair compreh<*"sicii w hen necessary. It is assum ed I'lat these cognitive and m etacognitive strategies are universal. O n ce these kinds o f skiiis have been developed, no m atter in w hich language, readers should be able to apply ,.iem to reading in any language providing they have attained a sufficient m astery o f th e language in question.
2.1.4. Sum m ary
T he foregoing theories from th re e different theoretical standpoints, namely, the cognitive, psychological, and psycholinguistic perspectives, claim that reading is a skill transferable, interdependent, and universal across languages. T hese three existing theories share som e com m onalties, and their co re features overlap. A ccording to the present researcher's understanding, all these th eories em phasize cognitive
com m onalties in reading com prehension processes across languages. All these theories indicate th a t th ere is a relationship b etw een L I literacy and L2 literacy developm ent. W hat w e need is em pirical research to explore th e nature o f the relationship betw een L I and L 2 reading. It should be p o inted o u t th a t previous studies (Block, 1986; 1992) com pare L I readers' com prehension pro cesses w ith those o f L2 readers. To b etter understand the relationship b etw een L I and L 2 reading com prehension processes, w e need inform ation ab o u t ho w th e sam e readers c o p e w ith reading tasks in their LI and L2. W e should p ro b e ho w these rea d e rs g a th e r and use inform ation available to them to co n stru ct their understandings. E vidence can only be obtained from empirical experim ents on com prehension p ro cesses in L I and L2 w ith the sam e individual readers.
2.2. R esearch on C om prehension S trateg ies and Strategy U se in L2 Reading 2.2.1. C om prehension S trategies
L anguage com prehension is view ed in cognitive th eory as an active, constructive p ro cess that applies equally to listening o r to reading. A ccording to
perceptual analysis, parsing, and utilization o f the m eaning uncovered in oral or w ritten text. Similar to O 'M alley and C ham ot's notion, oth er researchers, such as Ulijn (1976) and Schank (1972; 1975) contend th at com prehension includes parsing and inferring. Parsing, w hich is largely controlled by conceptual inform ation, represents the m eaning o f each individual sentence o f th e text. A process o f inference, on the o th er hand, helps to co n stru ct the m eaning o f a tex t since a tex t usually leaves many concepts unexpressed, so th at the read er has to infer them from his conceptual know ledge. A read er has som e alternatives to g en erate conceptual know ledge, or background know ledge to form hypotheses and expectations about th e text. T hese alternatives can be called reading com prehension strategies. C ognitive theory holds that the com prehension process is a p rocess in w hich the read er uses various kinds o f know ledge (e.g., linguistics, content, strategy, etc.) selectively and interactively to reconstruct m eaning o f a w ritten text. A t each o f th e stages, com plex processing and strategic analysis ta k e place to assist readers in detecting o r inferring m eanings and in relating th e inform ation to their existing know ledge.
T h e notion o f com prehension strategy w as first introduced in 1970 by B ever in the context o f sentence processing. Since then, several o th er researchers (K intsch & van Dijk, 1983, am ong others) have em ployed th e notion, and extended the notion from the sentence level to the discourse level. T hese researchers suggest that understanding a w ritten tex t involves n ot only th e processing and interpretation o f printed data, but also the activation and use o f internal, cognitive inform ation. T hus
the read er has th re e kinds o f data, namely, inform ation from the text, inform ation from the situation o r context, and inform ation from the cognitive presuppositions (van Dijk & K intsch, 1983).
R esearchers also claim th a t com prehension strategies are p art o f our
know ledge (van D ijk and K intsch, 1983; B lock, 1986, 1992; B arnett, 1988), Reading com prehension is considered to be a com plex activity involving various strategies that readers use to c o n stru c t m eaning o f a w ritten text. A m ong them , som e strategies may be called linguistic strategies, since they are language-specific. Som e are considered to be lo w er level asp ects o f reading, such as feature analysis o f th e print and w ord
identification, w ith th e parsing o f phrases and sentences and th e assem bly o f idea units. Also som e are considered to be th e higher level aspects o f reading com prehension, including draw ing up o n p rio r know ledge and integrating th e inform ation in the text w ith p rio r k now ledge (Perfetti, 1985; Stanovich, 1991). T h e high-level aspects o f reading com prehension strategies are m ore cognitive and involve th e use o f w orld know ledge, personal experience, p rior related inform ation, and oth er cognitive inform ation, such as opinions, beliefs, attitudes, interests, plans, and goals. W hat strategies readers use, how ever d e m o n stra te'th e know ledge they have about understanding a text, th e w ays in w hich they p rocess the text, and certain kinds o f schem ata th at they activate to facilitate their com prehension.
So far th e effective use o f reader schem a is the only com prehension strategy to have occasioned extensive L2 research (B arnett, 1988). C assanave (19 8 8 ) contends that th e are a o f schem a has been m ore appealing because th e test designs are less
com plicated. As well, the concept o f schem a is m ore co n crete and easier to gain insights into than oth er cognitive processes. R esearch on schem a th eo ry provides inform ation ab o u t why L2 readers fail to understand a given type o f text, and how to provide L2 readers w ith ap p ro p riate inform ation they lack before reading.
N onetheless, no inform ation indicates w hat readers actually do w ith schem ata o r how readers overcom e their com prehension obstacles to reach their com prehension o f a given tex i. R esearch on schem a th eo ry provides evidence th a t w orld know ledge o r linguistic know ledge will b e used to aid in interpreting th e m eaning o f a text. R esearch studies indicate th a t L2 readers fail to understand a w ritten tex t in th e ta rg e t language either b ecause th ey lack ap p ro p riate schem ata ( o f content, cultural, or rhetorical), o r because they fail to activate them. R esearch also show s bo th th at activating
ap p ro p riate schem ata o r providing necessary background helps L 2 readers
com prehend m ore, and th a t fam iliarity and unfam iliarity o f th e co n te n t area o f texts will affect L2 readers' u se o f com prehension strategies (C arrell, 1984, 1989;
B ernhardt, 1986; Johnson, 1982; Lee, 1986 am ong others).
T o b e tte r understand the com prehension process, w e should not only know w hy L 2 readers fail to understand a w ritten text, but also k n o w h o w they process a w ritten tex t, ho w they understand it, and w hat strategies th ey use to solve their com prehension problem s and to fix their com prehension failures. In this sense, research on strateg y use is, at l e a s t , as im portant as learning w hy L 2 readers fail to understand a w ritten text. H o w ever, researchers have different notions o f strategies. Som e research ers believe th a t strategies are tactics th at read ers use on p u rp o se and
consciously (Paris et al., 1983). T hese researchers try to distinguish skills from strategies. They define skills as actions th at readers take unconsciously, w hereas strategies are conscious m oves th a t readers take in order to solve com prehension problem s. A nother notion holds th a t "com prehension strategies are generally unconscious. Strategic behaviour is neither necessarily unconscious nor necessarily autom atic" (van D ijk & K intsch, 1983, p. 31). F o r the p u rpose o f this study, the current researcher considers all readers' alternative moves, operations o r procedures to achieve the goal o f com prehension, bo th conscious and unconscious, to be strategies.
These researchers also suggest th at one o f the fundam ental properties o f strategies is th at they are not independent. Thus, readers can use strategies jointly and selectively to reach their com prehension goals. H ow ever, w hat strategies are to be used, ho w to p rocess th e text, and w h at kinds o f know ledge structures are needed are decisions that can only be m ade by a certain read er in a certain context, and for a certain reading purpose, van D ijk and K intsch (1983) pro p o se that:
T he strategies applied are like effective w orking hypotheses about th e c o rre c t stru ctu re and m eaning o f a text fragm ent, and these m ay be disco n tin u ed by further processing. Also, strategic analysis depends n ot only on textual characteristics, b u t also o n characteristics o f th e language user, such as his o r her g oals o r w orld know ledge. This m ay mean that a reader o f a text will try to re c o n stru c t n ot only the intended m eaning o f the te x t — as signaled by th e w riter in various w ays in the tex t o r co n tex t — but also a m eaning th at is m ost relevant to his o r her ow n interest and goals, (p. 11)
A nother im portant characteristic o f strategies is that they are difficult to study because they happen in readers' brains. Som e kinds o f problem solving and decision
m aking can be applied very quickly and in a highly autom atized way in language understanding. Furtherm ore, only certain kinds o f strategies are open to empirical assessm ent via protocol analysis (van Dijk & K intsch, 1983).
2.2.2. Strategy Use
M o st research on co m p rehension strategies used in L 2 reading has focused on specific com parisons and c o n tra sts betw een o f L I and L2 readers, particularly on com paring native readers' reading perform ances w ith those o f L2 readers w hen they process a given language (B ernhardt, 1987; B lock, 1986; 1992). T hese studies use those native readers' reading com prehension processes as a yardstick by w hich to m easure L2 readers' perform ance. In essence, w hat they have found is that L2 readers read m uch m ore slow ly than native readers (C ziko, 1979, O iler & Tullius, 1973); L2 readers tend to be m ore attentive to visual details in te x t than native readers (H atch et al, 1974; Cziko, 1980); L2 read ers spend m ore tim e for each eye fixation than LI readers (O iler & Tullius, 1973; B ernhardt, 1983); L 2 readers' oral m istakes in m iscue analysis are less sem antically related to the original w ords; they have difficulty profiting from sem antic and disco u rse constraints within a tex t (C hihara et al. 1977; Cziko, 1978; Clarke, 1.979); and L2 read ers have difficulty in generating ap p ro p riate background know ledge to facilitate com prehension (Carrell, 1983; 1984, Lee, 1986; Johnson, 1981). As w e have seen, L2 readers lag behind L I readers in term s o f the efficiency w ith w hich low er level p ro cessing is carried out. In o th e r w ords, th e great difference betw een native readers and L2 readers is a local difference (B row n, 1985).
A dm ittedly, these findings are in som e sense a truism . It is apparent that w ithout the appropriate functioning o f these local systems, com prehension suffers. From a second language perspective, it is equally true that L2 users rarely reach native-like control ov er these system s. H ence, in second language research, researchers alw ays look to all facets o f the linguistic system when m aking
generalizations ab o u t reading (B ernhardt, 1992). T hat is why w e always explain that L2 readers' com prehension failure is due to their L2 proficiency level. H ow ever, previous research also indicates th at the linguistic system provides a necessary but insufficient understanding o f the second language reading process. W e cannot ignore thai som e L2 readers w ith w ell-developed linguistic know ledge o f the target language fail to com prehend o r that som e L2 readers w ith m eager linguistic know ledge succeed in com prehension. T he explanation for these phenom ena m ust obviously include, but go beyond, the linguistic base (B ernhardt, 1991).
W hat are th e o th er factors w hich affect L2 readers' com prehension process w hen processing an L2 text? L iteracy variables and w orld variables, according to B ernhardt (1991), are the other im portant factors. Literacy variables consist o f operational know ledge (B ernhardt, 1991) or procedural know ledge (O'M alley & C ham ot, 1990), w hich refers to know ing how to approach a text, why a reader approaches it, and w hat to do with a text when it is approached. The relationship betw een L2 and L I reading proficiency clearly operates in this dim ension o f literacy know ledge as well as in th e dim ension o f w orld know ledge, m ore so than in the dim ension o f linguistic know ledge o f the targ et language. In this sense, research into
the relationship betw een L 1 and L2 reading com prehension strategies o f the sam e individuals is crucial. By exam ining the patterns o f com prehension strategies o f the same individuals in both languages, w e may uncover aspects o f reading com prehension processes that are difficult to detect and describe w hen studied in the context o f a single language. Investigation o f th e sam e individuals’ com prehension strategies in reading tw o languages m ay reveal how readers process a te x t in their native language and in a second language, and w hether they process texts similarly o r differently in different languages. Thereby, evidence o f w h ether there is a universal pattern o f com prehension strategies in reading can be indicated.
H ow ever, research on com prehension strategy use in second language reading is still in its infancy. P revious research focusing directly on L2 reading strategy use is slight (B arnett, 1988). M o st suggestions about strategy use have been obtained from studies w hich tre a te d the question as incidental to th at o f ho w reading com prehension related to general language proficiency. T hese studies focus on the questions o f "reading problem o r language problem " (A lderson 1984), "reading ability o r language proficiency" (C arrell 1991), and "a language problem , a reading problem , o r a
know ledge problem " (B ernhardt, 1991). O ther L 2 studies focus on com paring the reading p rocesses betw een L I and L2 w ith native readers versus non-native readers. These studies infer readers' reading processes from their com prehension rate o r from their c o rre c t answ ers to cloze tests or m ultiple-choice com prehension questions or from m iscues (C larke, 1979, Cziko, 1978, Carrell, 1989). O nly a few studies attem pt to ex plore L2 reading strategy use (B lock , 1986, 1992). N o n e o f them focus on the
relationship o f com prehension strategies betw een L 1 and L2 reading o f the same individuals.
2.2.3. Sum m ary
R eading com prehension processes involve linguistic, cognitive and
m etacognitive strategies, and w orld and literacy know ledge as well. D uring reading process, L2 read ers solve their com prehension problem s by using their linguistic know ledge, by consciously o r unconsciously using their cognitive or m etacognitive strategies, and by applying their w orld or literacy know ledge. H ow ever, w hat know ledge so u rces L2 read ers reso rt to and w hat kind o f strategies they use are determ ined by th eir particular reading purposes, by particular text features, and by particular circum stances un d er w hich they are processing the text. T o investigate L2 readers' com prehension strategies, w e should not m erely focus on L2 readers' language proficiency, b ut rath er g o beyond th at to explore their cognitive m oves w hen they read in their L I and L2. By doing this, w e can hope tc explore the nature o f the relationship betw een L I and J.2 reading com prehension strategies.
2.3. T h e R elationship b etw een C om prehension Strategies in LI and L2 Reading 2.3.1. M ain A rgum ents in L2 R eading R esearch
W hen dealing w ith reading in a second language, researchers have been debating such issues as: "psycho o r linguistic?" (C larke 1979), "language com petence o r reading strategies?" (C ziko 1 9 8 0 )," a reading problem o r language problem ?"
language problem , a reading problem , o r a know ledge problem ?" (B ernhardt, 1991). T hese argum ents lead to th e follow ing question: Is L2 reading related m ore to the level o f L2 language proficiency, o r to the level o f L I reading ability?
In general, research o n L2 readers' reading ability and strategy use falls into tw o groups. Som e researchers (C larke, 1980, 1979, 1978; H auptm an, 1979;
M cLaughlin, 1987) com pared the L I and L 2 reading pro cesses o f second o r foreign language learners. They found th a t th o se w ho used successful strategies in L2 reading adopted similar strategies in thair L I reading. H ow ever, th eir limited L2 proficiency did n o t allow th e subjects to read as si' ^ e ssfu lly as they did in th eir L I . T herefore, C larke (1979) p o s’ted his w ell-know n "short-circuii hypothesis", and A lderson (1984) coined th e term "language threshold" fo r second language reading abilities. B o th the short-circuit hypothesis and th e threshold hypothesis arg u e th a t below a language proficiency threshold, the com prehension processes used in students' L I reading are not used as effectively in second language reading. T herefore, A lderson (1984)
believed L2 reading to be m ore a language problem than a reading problem . Language w as seen to play a critical ro le in second language reading abilities. T hese researchers argue th at an L2 proficiency level is a main elem ent w hich affects L2 readers' reading processes.
In contrast, o th er g ro u p s o f researchers (B enedetto, 1984, B lock, 1986, 1992, Carrell, 1991, H udson, 1982) found th a t L2 read ers pro cessed reading m aterials similarly in their L I and L2. A s well, proficient L2 read ers perform ed similarly to
proficient L I readers, and less proficient L2 readers perform ed similarly to less proficient L I readers. It m ust be noticed that these researchers com pare and contrast L i readers' com prehension processes w ith th ose o f L2 readers. R esearchers o f this group m aintain th a t L2 reading depends on the reading ability in a reader's L I literacy level rath er than on the reader's proficiency ievel in the targ et language.
T he fo reg o in g argum ents found in th e literature reveal that relationships exist in L I and L2 reading. B u t th e argum ents do not answ er several questions: H o w does L I reading ability relate to L 2 reading? W hich has m ore im pact on L2 readers'
com prehension processes, L I literacy ?Hility o r L2 proficiency? F o r adult educated L2. readers, does L 1 literacy ability have m ore influence on their L2 reading
com prehension p ro cesses th an L2 proficiency? A nsw ers to these questions can only com e from em pirical studies using cro ss-L ig u istic w ithin-subject designs (Hulstijn,
1991). In o th er w o rd s, such studies should com pare the reading com prehension processes o f th e sam e individuals in L I reading tasks and in similar L2 reading tasks.
2.3.2. H o w do L I R eading and L 2 R eading R elate?
B ern h ard t (1991) p osits th at w e have not sufficiently em phasized that first language literacy is a significant co m ponent in second language reading. H ow do LI reading and L 2 reading relate? B efore w e can begin to answ er such questions, w e need to gain k now ledge o f th e relationships, or interdependence, betw een L2 and L 1
