Sei sulla pagina 1di 15

Teachers of English to Speakers of Other Languages, Inc.

(TESOL)

Effects on Reading Comprehension of Building Background Knowledge Author(s): Patricia Johnson Source: TESOL Quarterly, Vol. 16, No. 4 (Dec., 1982), pp. 503-516 Published by: Teachers of English to Speakers of Other Languages, Inc. (TESOL) Stable URL: http://www.jstor.org/stable/3586468 . Accessed: 25/01/2014 23:44
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Teachers of English to Speakers of Other Languages, Inc. (TESOL) is collaborating with JSTOR to digitize, preserve and extend access to TESOL Quarterly.

http://www.jstor.org

This content downloaded from 148.206.53.9 on Sat, 25 Jan 2014 23:44:10 PM All use subject to JSTOR Terms and Conditions

TESOL QUARTERLY Vol. 16. No. 4 December1982

Building Knowledge Bac.kground

on Reading Comprehension of Effects


PatriciaJohnson
Thisexperimental theeffects on reading study investigated comprehensionofbuilding were Thefollowing background knowledge. specific points addressed. effects on ESL cultural havedifferent Wouldprior experience students' ofinformation linked to a familiar reading comprehension aspect ofan American toan unknown custom incomparison toinformation linked custom? ofpreWouldtime ofpresentation of themeanings aspectofthis selectedunfamiliar wordsin thesamepassagehave different vocabulary effects onreading twoadvancedESL students on comprehension? Seventy theuniversity levelreada passageon thetopicofHalloween. Thepassage contained unfamiliar andfamiliar information recent basedonthesubjects' of thiscustom. of preselected the meanings experience Subjectsstudied unfamiliar inthetext. words before found them and/or vocabulary reading Statistical oftherecallofthepassageand ofthesentence analysis recognitiontask indicates that cultural for comreaders prior experience prepared of the familiar aboutHalloweenin thepassage. information prehension to meanings ofthetarget wordsbyanyof However, exposure vocabulary thetreatments seemsnotto have a significant on reading effect comprehension. Characteristics of a text can have a large impact on readers' ability to it. The cultural comprehend backgroundof thetopic and thelevel ofvocabof a passage influence ularydifficulty readingcomprehension. Goodman (1971) has described reading as a cyclical process by which a reader reconstructs a textby sampling,predicting, and confirming. testing, Goodman's argument is thatthe good reader takes advantage of theredunin language whichenables a reconstruction of thewhole text dancy inherent a of the material has been extracted. Once such althoughonly part graphic an internal reconstruction of thetextualmessage has takenplace, the reader testsits accuracy againstpreviousinformation. Previousinformation can be the information extractedfroma text as well as the reader's background thatthereconknowledge on the topic of a text.When thereader confirms struction is in agreementwithpreviousknowledge,thenthecyclicalprocess of samplingbegins again. If thereis some inconsistency or inaccuracyin the reconstruction of a textbecause of a conflictwiththe reader'sbackground knowledge,thereader may reread thetextor may not believe theinforma-

of Wisconsin, Ms.. Johnsonis AssistantProfessorin Applied Linguisticsat the University Green Bay. Her researchinterests includesecond language acquisition, readingcomprehension, and teachereducation.

503

This content downloaded from 148.206.53.9 on Sat, 25 Jan 2014 23:44:10 PM All use subject to JSTOR Terms and Conditions

504

TESOL Quarterly

can be tion in the text until later when a more plausible reconstruction confirmed. Fluent and proficientreaders of a passage in theirnative language will recover quickly fromwrong guesses fortextualmeaning. ESL readers, on the otherhand, may not recoverin such a successfulmannerand inaccurate information extractedfromthe textmay lead to inaccuratelaterpredictions formeaning. ESL readers would also probably have to sample much more from the text in order to derive the same amount of meaning as fluent readers.They may also depend moreon backgroundknowledgeof thetopic and reconstruction thanon linguistic analysisof the textforcomprehension the of of a passage because of their language. (For a incompleteknowledge discussionof ESL readingbased on the Goodman model, see Buck, 1975.) in processThe relationof backgroundknowledge to textcomprehension The theorists. schema and been studied information has ing recalling by of a that the earliestworkwas by Bartlett who organization (1932) proposed reader's past experiencesdirectly influencesthe comprehensionand retena passage tionof materialsin a passage. In otherwords,readersunderstand text an internal the their to schema, organizationof by analyzing according and Woodworth Kant (1938) have used the (1963) past personalexperiences. Lehnert (1977) Charniak and frame; term,schema; (1975) Minsky(1975), and Schank and Abelson (1975), script;Becker (1973), Bobrow and Norman and Ortony(1972), schemata. (1975), and Rumelhart The view of schema theoryasserts that activatingor building readers' existingknowledge prior to reading would improve and/or alter reading comprehensionand recall. Thus, the provisionof vicarious or real experiences would fill in or expand the readers' existingculturallydetermined backgroundknowledge of a topic and would prepare themto comprehend and retainmaterialon thattopic in thereadingpassage thatfollowed. show Two studies on the effectof providingprior culturalinformation resultsforESL students' reading comprehension.Yousef (1968) contrasting Middle Eastern stufound thateven afteran intensive culturalorientation, dents were not able to interpret textsfromthe Americanculturebecause of negative attitudes.He concluded thatthe students'resistanceto American culturereduced theirmotivation to learn.Gatbontonand Tucker and efforts (1971) suggest that inappropriatevalues, attitudes,and judgmentsled to of Americanliterature. However, Filipino ESL students'misunderstanding when they provided relevantculturalinformation, theirstudents'reading comprehensionimproved. Both of these studieswere conducted in foreign settingswhere there was littlecontact with American culture.Could real experiences withinthe Americancultureprovide background information of a and have a positive effecton thereadingcomprehensionand retention context? passage set in the Americancultural The emphasis on the development of vocabulary knowledge in reading textbooksmay encourage word-by-wordreading and consequently may

This content downloaded from 148.206.53.9 on Sat, 25 Jan 2014 23:44:10 PM All use subject to JSTOR Terms and Conditions

ofBuilding Effects Knowledge Background

505

preventthe ESL reader fromthe development of the skill of processing for meaning. Smith (1971) syntaxand contextin samplingand confirming to readthat or readingis detrimental letter-by-letter word-by-word argues the before be of word one The forgotten may meaning ing comprehension. can nextword is built or understoodand thus,no meaningful relationships be establishedbetween words. In researchon perceptual processes in reading,Kolers (1973) foundthat readingis not at all a precise perceptualprocess of sequentialword recogis contributed nition.More information by the reader thanby the printon what theyread because theyare able thepage. That is, readersunderstand to an appropriate to take the stimulusbeyond its graphic representation in memories. their of stored group conceptsalready Nevertheless, vocabulary developmentis consideredessentialto reading have regularlybeen found for native comprehension.Strong correlations readers between knowledgeof word:meaningsand abilityto comlanguage and thesewords and between word difficulty prehendpassages containing and see Anderson a review of these studies, (For readingpassage difficulty. Freebody 1979.) of word knowledge on nativelanguage studiesof theeffect Experimental mixed resultsin comparisonwith the shown have reading comprehension correlational studies. For example, Wittrock, Marks,and Doctorow (1974) foundthatpassages in which 15%of the words had been changed to lowerled to about 25%decrease in performanceon subsefrequencysynonyms quent comprehension questionsfornativelanguage readers.In a laterstudy words improved that on low-frequency found directinstruction (1975), they of readers on comprehension questions. performance On the otherhand, thereare studies which have failed to establishthe directrelationship between vocabulary difficulty and reading comprehension. Tuinman and Brady (1974) found that an instructional program on in in readers' native performance vocabularyresulted an increase language on a vocabularytestby an average of about 20%.However, themeans forthe reading comprehensionmeasures were almost identical for passages with and withoutpreteaching Pany, and Schreck (1978) revocabulary.Jenkins, transported similarresults.They found thatwhereas vocabulary training due to no effect there was ferred to singlesentencescontaining words, target of increased word knowledge on broader measures of the comprehension nativelanguage readers. In research on reading comprehensionin a foreignlanguage, Johnson seemed to have no of vocabularyand syntax (1981) foundthatsimplification effect on thecomprehension of a passage of thesame cultural backgroundas was betterfor a the ESL readers. However, theirreading comprehension passage of foreignculturaloriginwhen the vocabulary and syntaxwere simplified.If the meanings of words crucial to comprehensionhad been before reading the passage of foreignculturalorigin,would the pretaught

This content downloaded from 148.206.53.9 on Sat, 25 Jan 2014 23:44:10 PM All use subject to JSTOR Terms and Conditions

506

TESOL Quarterly

subjects have understood the unadapted version as well as the simplified version?If culturalbackgroundknowledge had been expanded, would the culturaloriginas well as the subjectshave understoodthepassage of foreign with their own culturalbackground? passage dealing Such questionsled to thestudyreportedin thisarticle.The specificpoints addressed were the following: Would prior experience in the American culture affect ESL students' comprehensionof a passage on the topic of an American custom, Halloween? Would this prior experience have different effectson the reading of information linked to a familiar comprehension aspect of Halloween in to that linked to an unknown of Halloween? comparison aspect Would the exposure to meanings of difficult vocabulary words in the affect of ESL readers? Would passage comprehension priorstudyof vocabin on the the different effects have and/or ulary glossingvocabulary passage of a comprehension readingpassage? A studywas designed to obtainmore data on theextent of theseinfluences on thereadingcomprehension of ESL students.The effects of priorexperience in the Americancultureand exposureto difficult vocabulary words on the readingcomprehension of the passage were measured by an analysisof the subjects' written recall, theirrecognitionof sentences containingtrue and their on a vocabulary doze test. information, performance Methodology The studysample consistedof 72 students fromthe advanced level readclasses at the Center for as a Second ing English Language at Southern Illinois University.The studentsrepresented23 nationalities,the largest groups were 16 studentsfromMalaysia and 13 fromTunisia. The average lengthof thesubjects'stayin theUSA at thetimeof theresearchtestwas 16W weeks. A comparison of the standarized scores for these studentson the withthenormsprovided withthistest MichiganTest of English Proficiency indicatesthattheirEnglish proficiency was below the level usuallyrecommended forfulltime academic work. The topic of thereadingpassage was thecelebrationof Halloween. All but one subject indicated participation in the city-widecelebrationtwo weeks priorto theresearchtest.The readingpassage had two sections.One section contained information assumed to be familiarto the subjects: a description of thecostumes,masks,and thecustomof trick-or-treat of Halloween today (firstparagraph), a descriptionof the witch as a central symbol (second paragraph), and a concluding generalizedcomparisonof Halloween today with the past (fifthparagraph). The other section contained information believed to be unfamiliar to even native readers: the celebrationof Halloween in the historicalpast (thirdparagraph) and the different degrees of witchcraft at thattime (fourth paragraph). The familiar and unfamiliar with sectionsof thepassage were constructed

This content downloaded from 148.206.53.9 on Sat, 25 Jan 2014 23:44:10 PM All use subject to JSTOR Terms and Conditions

Effects ofBuilding Knowledge Background

507

and West's (1953) countof semantic words,usinga thesaurus low-frequency with these lowfrequenciesof words. Moreover, the subjects' familiarity a vocabulary test.This vocabulary frequencywords was checked through testfollowed proceduressuggestedby Andersonand Freebody (1979) and yielded a reliable estimateof word knowledge,correctedforresponseson words knownby more thebasis of partialknowledge. Those low-frequency than25%of thesubjectswere no longerregardedas target vocabularywords for the research test. However, a balance was still maintained since the familiarsection of the passage had a total of 172 words with 22% lowsectionhad a totalof 191 words with frequencywords while theunfamiliar 16% words. low-frequency In the linguistic analysisof the passage, therewere two types of propositions.Lexical propositionswere clauses, phrases, or single words which were expressedseparable idea units(Kintsch1974). Relationalpropositions linksbetween propositions and expressedideas of attribution, time,place, or explanation, (Grimes 1972, manner,cause, more specific information the same numberof types The contained adapted by Meyer 1975). passage of propositions in of the familiar and unfamiliar sections. (total 122) The sentence recognitiontask was designed to measure the subjects' of the targetvocabulary words and of the familiarand ununderstanding in the passage. All sentenceswere paraphrasesof pasfamiliar information sentences the targetvocabularywords. There were twelve sage containing sentences:an equal numberof trueand falsesentencesfromthefamiliar and unfamiliar sectionsof thepassage. A cloze teston theoriginaltextmeasured theretention of the target vocabularywords. Membersof each readingclass (6) were randomlyassigned to one of the fourfollowing conditions: testing 1: the Group reading passage without a vocabulary list to study before or to referto while reading. reading 2: the definitions of thetarget words beforereadingbut not Group studying able to refer to this list while being reading. words glossed in thepassage. Group 3: readingthe passage withthetarget 4: the words beforereadingwiththe deGroup studying targetvocabulary finitions of the target in words glossed the passage. Aftersubjects read the passage, they were asked to recall the storyin written formwithoutreferenceto the textand to recognize sentencescontrue information fromthe passage. The students were testedin their taining classroomsduringtheirreadingclass. They worked at theirown rates,and consequentlytherewas some variationin completiontime,(30-35minutes). Two weeks afterthe experimental treatments, subjectswere givena vocabularycloze teston the passage. The dependentvariablesin thelinguistic analysisof data were thenumber and types of propositionsand relationsbetween propositionscontainedin thesubjects' written recalls when compared withthe propositionalanalysis

This content downloaded from 148.206.53.9 on Sat, 25 Jan 2014 23:44:10 PM All use subject to JSTOR Terms and Conditions

508

TESOL Quarterly

of the reading passage. Sentences in the writtenrecalls which showed a correct recall of the passage were counted. In addition, those sentences which contained relations between propositionswere given a separate count.A proposition was countedas recalled ifitwas verbatimor expressed thegistof thepropositionin thepassage. Lexical and relationalpropositions as well as compatible and incompatibleinferences were totaledseparately. in textual made recallincluded summary statements, Compatible inferences and elaborationsof textualinformation based on background implications, cultural were distortions of thepassage knowledge.Incompatibleinferences in recallbecause of a misunderstanding of textual or because of implications a lack of backgroundcultural biased misinterpreknowledge or a culturally tationof the topic of the passage. (Examples of these different types of in recall will be discussedlater.) propositions The dependent variable in the sentence recognitiontask was the total number of confirmations of true statements and rejectionsof false statements.The dependentvariablesin thecloze testwere an exact-wordscoring of the insertion of the targetvocabulary in the passage and a semantically or acceptable scoringof the recall of both the targetwords and definitions of thesewords givenin theoriginal synonyms vocabularylist. Results Priorexperiencein the Americancultureseemed to affectESL students' comprehensionof a passage on the topic of an Americancustom,Halloween. This priorexperiencehad different effectson the reading comprehension of information linked to a familiaraspect of Halloween in comparison to that linked to an unknown aspect of Halloween. Four of the eleven dependent variables in Table 1 indicate greaterrecall of familiar information from the passage. Subjects wrote more sentenceswhich exthe relations in the passage and rebetween propositions pressed correctly called more total propositions, includingboth lexical and relationalpropositions. There are no significantdifferencesin means in the recall of unfamiliar information fromthepassage. Data for eight of the dependent variables in Table 2 indicate a signifirecall of thefamiliar in thepassage thanof information cantlybetterwritten theunfamiliar information. Subjectsin all fourgroupswrotemoresentences fromthe familiarsection of the passage. containingaccurate information Their use of relationsbetween propositions in written recall reflected more accurately the organizationand semantic cohesion of the passage. They recalled more lexical and relationalpropositions and made more inferences which were compatible with the textof the familiarsection.Althoughinmade in recall of spectionof means suggestsmore incompatibleinferences theunfamiliar sectionof thepassage, all differences are non-significant. When the means of the dependentvariable of recognition of information

This content downloaded from 148.206.53.9 on Sat, 25 Jan 2014 23:44:10 PM All use subject to JSTOR Terms and Conditions

Effects ofBuilding Knowledge Background


TABLE 1
Analysisof Variance Results For Effectsof Familiar (and Unfamiliar) Information on Recall of Text withinGroups"
Group 1 X Sentencesin Recall Total Sentences 6.22 (3.50) WithRelationalPropositions 5.50 (2.72) Recalled Propositions Total Propositions RelationalPropositions Lexical Propositions Compatible Inferences Total Propositions RelationalPropositions Lexical Propositions IncompatibleInferences Total Propositions RelationalPropositions Lexical Propositions 30.33 (15.00) 13.00 ( 6.17) 17.33 ( 8.83) 5.06 ( .22) 2.44 (..11) 2.61 ( .11) .72 (4.72) .39 (2.11) .33 (2.61) Group 2 X 6.39 (3.50) 6.22 (3.22) 36.28 (18.89) 15.95 ( 8.28) 20.33 (10.61) 8.00 ( .94) 3.89 ( .50) 4.11 ( .44) 1.89 (2.72) .94 (1.39) .94 (1.33) Group 3 X 5.61 (1.78) 4.94 (1.67) 25.33 (10.61) 10.78 ( 6.33) 14.56 ( 4.28) 7.06 ( .61) 3.44 ( .33) 3.61 ( .28) 1.78 (2.17) .94 (1.11) .83 (1.06) Group 4 X 4.83 (2.28) 4.22 (2.11) 22.06 (11.06) 9.33 ( 4.67) 12.72 ( 6.39) 7.61 (1.06) 3.67 ( .56) 3.94 ( .50) 1.00 (1.67) .50 ( .83) .50 ( .83)

509

F-value 2.07 (2.57) 2.75*(1.67) 3.71'(1.44) 3.72*( .83) 3.62"(1.87) .86 ( .94) .85 (1.00) .86 ( .85) .91 (1.11) .88 ( .96) .94 (1.21)

Note: Means and F-values in parenthesesindicate the analysisof variance resultsforeffects of unfamiliar information on recall of textwithin groups. aN=18 foreach group;overall N=72. *p <.05.

TABLE 2 Dependent T-Test Results For Effectsof Familiarvs. Unfamiliar Information on Recall of Text withinGroupsa
1 4 2 3 Group Group Group Group Mean diff. T-value Meandiff. T-value Meandiff. T-value Meandiff. T-value inRecall Sentences TotalSentences With Relational Propositions Recalled Propositions TotalPropositions Relational Propositions Lexical Propositions Compatible Inferences TotalPropositions Relational Propositions Lexical Propositions Incompatible Inferences TotalPropositions Relational Propositions Lexical Propositions Op < .01. 2.72 2.78 15.33 6.83 8.50 4.83 2.33 2.50 -4.00 -1.72 -2.28 4.55* 5.53* 4.90* 4.64' 4.99* 3.56* 3.61" 3.52* -1.94 -2.04 -1.84 2.89 3.00 17.39 7.67 9.72 7.06 3.39 3.67 -.83 -.44 -.39 4.58* 4.68* 6.01 6.08* 5.79* 6.61* 6.38* 6.76* -.58 -.61 -.54 3.83 3.28 14.72 4.44 10.28 6.44 3.11 3.33 -.39 -.17 -.22 7.88* 6.68 4.53* 2.21* 3.87* 3.45* 3.33* 3.55* -.49 -.38 -.61 2.56 2.11 11.00 4.67 6.33 6.56 3.11 3.44 -.67 -.33 -.33 3.94* 3.24* 3.67* 3.68* 3.61P 4.08' 3.97* 4.15* -.71 -.71 -.71

aN= 18 foreach group;overall N = 72.

This content downloaded from 148.206.53.9 on Sat, 25 Jan 2014 23:44:10 PM All use subject to JSTOR Terms and Conditions

510

TESOL Quarterly

to those found with the written recall of the passage. Subjects of all four sectionof fromthefamiliar groupsmore accuratelyrecognizedinformation thepassage in comparisonto the unfamiliar section,althoughthe difference between means is significant foronlyGroups 1, 2, and 4.
TABLE 3 Dependent T-Test Results For Effectsof Familiarvs. Unfamiliar Information on SentenceRecognition within Groups" Familiar Information X Group 1 Group 2 Group 3 Group 4 5.11 5.22 5.06 5.33 Unfamiliar Information X 3.73 4.39 4.61 4.28 T-value 4.93* 2.73* 1.64 2.89*

aresimilar results from the inTable3,itislearned that areexamined passage

"N=18 foreach group;overall N=72. *p < .01

The exposure to meaningsof difficult vocabulary words in the passage seemed not to affectthe comprehension of ESL readers.Post-hoccomparisons (Scheff6test) indicate thatsignificant effectsforthe dependent variables in Table 2 can be explained by a difference between the means of 2 recall of the of Group 4. Group 2 and those Group subjects' passage who studied the of the words beforereading the subjects, meanings target recalled more the textthanGroup 4 from passage, accuratelypropositions who studied the words before subjects, vocabulary reading and had the definitionsof those words glossed in the passage, F = 3.70, p < .01. A comparisonof types of propositionsshows that Group 2 subjects recalled more lexical propositions or idea units, F = 3.61, p < .01, and morerelations or linksbetween propositions, F = 3.72, p < .01, thanGroup 4 subjects.The written recallsof Group 2 subjectsalso containedmoresentencesexpressing relationsbetween propositionsand thus reflectedbetterthe semanticcohesionof the passage thanthoseof Group 4 subjects,F = 2.75, p < .05. A univariateF-testindicatesno significant differences in means between for the of true in information groups recognition paraphrases of passage sentencescontainingthe targetvocabulary words. UnivariateF-testsalso indicate no significant in means between groups foreitherthe differences exact-wordscoringor the semantically acceptable scoringof the cloze test on the target words of the vocabulary passage.

This content downloaded from 148.206.53.9 on Sat, 25 Jan 2014 23:44:10 PM All use subject to JSTOR Terms and Conditions

Effects ofBuilding Background Knowledge Discussion.

511

The variable of topic familiarity or culturalcontextof a passage provides information about theinterpretation of personalknowlof a passage in terms seem to that real experiences indicate edge. Resultsof thestatistical analysis withinthe culturalcontextprovided backgroundinformation formore effectivereadingcomprehension of a passage. Subjects were not onlyable to recall moreexplicitand implicit but their in information individualsentences written recall reflectedmore accuratelythe cohesive ties and relationsbetween idea unitsor propositions in thefamiliar sectionof thepassage. In order to illustrate more clearlythe interaction of backgroundknowlthereaders'written from edge and cohesionin recall,uncorrected examples recallsof the textfollowan excerptfromthereadingpassage. is funforeverybody. Halloween incostumes, dress and Peopleputon masks, turn intomonsters. to have funand maybe go outintothestreets Theythen scaresomeone. from (excerpted 1) Paragraph 1. They wear masks [and change completelyhow theyreallylook.] 2. [It depends on theindividualwhich typesof] masks [theywant] to wear. The portionof the recall in brackets (#1)is a textualinference fromthe second sentence in the passage. Cultural inferences(#2) fromthe textual proposition, (people put on masks),are based on thesubject'sexperiencein theculture.Both inferences are compatiblewiththemeaningof thetext. The followingexamples containculturalinferences: an example inferred fromthefirst sentence(#3)and an example inferred sentence fromthethird in the from the Both (#4) inferences, excerpt passage. compatible withthe of the are on based the meaning passage, subjects' personalexperiencesof Halloween and illustrate theirattemptsto reconstruct a plausible recall of thepassage. 3. People enjoy thisday [because theyprepare forthe celebrationand use the colors orangeand black.] 4. [Everyyear] people go [downtown]to celebrateHalloween. [You can see people who come fromothertowns]to enjoy thisbig day. Expressed relationalpropositions(#3)are cause (because ... black) in relationto enjoy, additive (... and . . .) and more specific information to colors. Relationalpropositions in thefirst (orange and black) in relation sentence(#4)expresstime(everyyear),place (downtown),and cause (to celebrate Halloween) in relation to go. Relational propositionsin the second sentence (#4)expressmore specificinformation (who ... towns) in relationto people and cause (to ... day) in relationto come. All ofthese relational propositionsare culturalinferencesfrom the subjects' backof a coherent, groundknowledge of Halloween and aid in reconstruction meaningful passage. Halliday and Hasan (1976) detail varioustypesof cohesive tiesevident in texts:reference, and lexical cohesion. substitution/ellipsis, conjunction,

This content downloaded from 148.206.53.9 on Sat, 25 Jan 2014 23:44:10 PM All use subject to JSTOR Terms and Conditions

512

TESOL Quarterly

The connectionof such itemsacross sentencesdefinesthe semanticconor cohesivenessof a text.The cohesive items exemplifiedin the tinuity recalls reflectthe relationship between sentencesin the subjects' written familiar sectionof thepassage. 5. Halloween is [a veryspecial day]/when[old and young]put on costumes/ and [theygathertogether in a chosen place, forexample, a streetin the city.] The first two portions/Halloween ... costumes/contain explicitinformation and textualinferences(in brackets)recalled fromthe first and second sentencesin theexcerpt.The third ... city/, is a cultural inference portion,/and fromthe thirdsentencein the excerpt.Cohesive ties are identifiedas reference by theuse of theyand ellipsisby youngand old, both in reference to is identified to day and by street people. Substitution by when in reference in referenceto place. An example of semantic connectionis and, a conjunctiveitem. 6. People [walk in groups]/anddress in different kindsof clothes [like animals or othercountry or whatever]/and[theytalkwitheach otherand do to have fun. everything] The first and thirdportions(#6)containexplicitpropositions and cultural inferences fromthe thirdsentencein the excerpt.The second portioncontainsrecall of explicitpropositions and textualand culturalinferences from the second sentence.Conjunctivecohesionis identified by and between the first and second portionsand between the second and thirdportions.Reference cohesion is identifiedby they in the thirdportionin referenceto is identifiedby whatever and do people in the firstportion.Substitution in reference to a cultural everything experienceof thesubject. The preceding examples show thattextualcohesion can be understood when thereader appropriately thetopic of a passage accordingto identifies an existing schema. In otherwords,recognizingthata textis about thetopic of Halloween in the first sentence of the reading passage seems to make in the possible the processingof cohesive elementsor semanticcontinuity text.However, recognition of the topic may resultin a loss of textualcohesionin recall when thereis a lack of backgroundinformation on thetopic. Subjects may recognize the topic, Halloween, but the descriptionof the historical to them.Therefore, their aspect of the customis new information sentences in recall may contain relationsexpressed between propositions and cohesive tieswhichdo notreflect thepassage accurately.The following is from the section of the unfamiliar information excerpt passage containing forthereaders,followedby uncorrected from the readers' written examples recall of the text. The assembled witches vowedto obeytheir witch who was god,themaster as an animal. children to thegod and thanked disguised Theypledgedtheir himfor foodandfor life. Thereligious was followed and ceremony byfeasting The witches dressed from dancing. up likeanimals. (excerpted 3) paragraph

This content downloaded from 148.206.53.9 on Sat, 25 Jan 2014 23:44:10 PM All use subject to JSTOR Terms and Conditions

ofBuilding Effects Knowledge Background

513

1. The witchwears like theanimals [because] the god is like an animal. 2. They were feasting, dancingand singing[as] theywere flying. relarecall (#1, These examples fromreaders'written 2) containincorrect recalled fromthe text.The tions(in brackets)between explicitpropositions in thefirst clause recalled second example also containsexplicitinformation in the relationto explicitinformation fromthe thirdparagraphin incorrect second clause recalled fromthe second paragraph of the passage. These how thereadersattemptedto make a plausible and consentencesillustrate of materialwhichwas not fully nected interpretation comprehended. thelexical cohesive item,religious In theexcerptfromthetext, ceremony, two sentencesand is identified in the first is a textualinference by the first word in the thirdsentence.The followingexamples (#3, 4, 5) show a misinferences textual in incorrect of thiscohesivetie whichresults interpretation forthereasonsforthewitches'meeting. 3. They promiseGod [to make a good job.] masterand god, 4. There is a story about thewitchwheretheypromisetheir their broken had promises.] [but they 5. The witcheshad a meeting[to discusshow to controlthepeople.] sectionof of theunfamiliar The nexttwo examples (#6, 7) show distortions which referto the familiar the passage. Subjects make culturalinferences sectionof thepassage or to theirown personalexperiencesof Halloween, a timeforfunand not a seriousreligiousceremonyas in thepast. 6. They will pray to God and dress [in different types of masks so as to at theirugly faces and to look the like the people frighten people. They funny characters.] 7. It was important [forthepeople] because [althoughtheydidn'tbelieve in wanted thankin some way fortheirfood and lives. to it,]they between subThe preceding are only a few examples of the interaction their and comprehenjects' culturally experiencedbackgroundknowledge sion of textualcohesion. However, these examples indicate thattheirmislevel is accompanied by a of textual meaningon thelinguistic understanding The level. on the problems withlanguage in conceptual misunderstanding section of the passage seem to be theirwrittenrecalls of the unfamiliar actuallyproblemscaused by a lack of backgroundknowledge on thetopic. Building background knowledge on the historicalaspect of Halloween and written production mighthave improvedtheirreadingcomprehension in English. is a variable most The variable of vocabulary knowledge or difficulty and sentence the word on to the influence reader's comprehension likely on readingcomprehenlevel. However, theeffects of vocabularydifficulty of backgroundknowledge.If the attempt sion are not as clear as theeffects to increase comprehension the explanationof difficult vocabulary through thenGroup 1 subjects words had had an effecton readingcomprehension, worse on the reading comprehension would have performedsignificantly

This content downloaded from 148.206.53.9 on Sat, 25 Jan 2014 23:44:10 PM All use subject to JSTOR Terms and Conditions

514

TESOL Quarterly

tasksin comparisonto theotherthreegroups.Group 1 subjectshad notbeen to study the definitions of the targetwords before given an opportunity the 4 2 and did reading passage (as subjects),norwere thedefinitions group of the words glossed in the passage (as forGroup 3 and 4 subjects). Howon the readingcomprehension tasks ever, the analysisof theirperformance shows quite different results. The sentence recognitiontask, designed to measure the subjects' comprehensionof the targetvocabularywords, and the doze test,designed to measuretheir retention of thesewords,indicateno significant differences in within the four The treatment difcomprehension groups. only significant ferencebetween means within of groupswas explained by theperformance 2 4 and on the of section of the recall the familiar Group subjects passage. These statistical resultsindicate thatstudyingthe definitions of the target words before reading withoutbeing able to referto thislistwhile reading seems to have resultedin better (Group 2 treatment) readingcomprehension and recall of the sectionwhichcontainedfamiliar on thetopic. information the definitions of the words before Studying target reading the passage combined withthesame word definitions in the text(Group 4 treatglossed does not seem to have been as effective for ment) readingcomprehension. In fact,the emphasison vocabularywords forGroup 4 subjectsmight have resultedin word-by-word which have been detrimental to the reading may of the Other such as attistudent generalcomprehension variables, passage. tudes and teacherrelationships, the may also have been a factorinfluencing of the two tests in The were comprehension groups. given the classrooms underteacherdirectionand onlytwo of thefourtreatments were randomly givenin each class. It may take a high percentage of difficult vocabulary items to have a effecton readers' comprehension of a text.The normalredunsignificant wordswithout too dancyin a text mayenable readersto cope withunfamiliar muchdisruption in their Readersseem to be able to construct understanding. a textfrommemorybased on inferences made while reading.Thus, familiwiththetopic of thepassage and generalbackgroundknowledgeof the arity thememay allow the reader to construct highlyplausible meaningsforunfamiliar vocabularywords. treatments of the Reading comprehensionresultingfrom the different four groups in thisresearch study does not support the high correlations between vocabularyknowledgeand readingcomprehension. However, the studieson the effectof vocabularyknowledge on thereading experimental comprehensionof native language readers seem to show the same unclear resultsas the presentresearchwith ESL readers. In conclusion,familiarity with a foreignculturally related topic, knowledge obtained fromreal exis effective forreadingcomprehension of a periencesin the foreign culture, passage on thattopic.

This content downloaded from 148.206.53.9 on Sat, 25 Jan 2014 23:44:10 PM All use subject to JSTOR Terms and Conditions

Effectsof BuildingBackgroundKnowledge REFERENCES

515

Anderson,Richard C. and Peter Freebody. 1979. Vocabulary knowledge. (ERIC Document ReproductionService No. ED 177 480). Press. FrederickC. 1932.Remembering. London: Cambridge University Bartlett, information. In R. Becker,JosephD. 1973. A model forthe encodingof experiential and language. San Schank and K. Colby (Eds.) Computer models of thought Co. Francisco: W. H. Freeman Publishing Bobrow, Daniel G. and Donald A. Norman. 1975. Some principles of memory and underschemata.In D. G. Bobrow and A. M. Collins (Eds.) Representation Buck, Catherine. 1975. Miscues of non-native speakers of English. In K. Goodman (Ed.) Miscue analysis. Urbana, IL: ERIC Clearinghouseon Reading and ComSkills. prehension in a frame-like Charniak,E. 1975. Organizationand inference systemof common

Press. New York: Academic incognitive Studies science. standing:

An interdisciplinary workshop. Cambridge, MA: Bolt, Baranek and Newman Co. Publishing and the Gatbonton,Elizabeth C. and G. RichardTucker. 1971. Culturalorientation literature. TESOL Quarterly, 5, 2:137-143. studyof foreign in the readingprocess. In P. universals Goodman, KennethS. 1971. Psycholinguistic Pimsleur and T. Quinn (Eds.) The psychologyof second language learning. Press. Cambridge,Eng.: Cambridge University Grimes,JosephE. 1972. The threadof discourse.Ithaca, N.Y.: Cornell University. Halliday, M. A. K. and Ruquaiya Hasan. Cohesion in English. London: Longmans, Green and Co. 1978. andreading Jenkins, J.R.,O. Pano andJ.Schreck. comprehension: Vocabulary Instructional effects.(ERIC Document ReproductionService No. ED 160-999). Patricia. 1981. Effectson reading comprehension of language complexity Johnson, and cultural 15, 2:169-181. backgroundof a text.TESOL Quarterly, Kant, Immanuel. 1963. Critique of pure reason. (IE, 1781; 2E, 1787, Kemp Smith London: MacMillan Publishing Co. translator). Kintsch,Walter. 1974. The representation of meaning in memory.Hillsdale, N.J.: Lawrence Erlbaum Associates. and Kolers,Paul A. 1973.Three stagesof reading.In F. Smith(Ed.) Psycholinguistics Co. and Winston reading.New York: Holt, Rinehart Publishing Lehnert,Wendy. 1977. Human and computationalquestion answering.Cognitive Science 1:47-73.

In Proceedings issues in natural language processing: knowledge. of theoretical

F. 1975.The organization Bonnie Amon memory. Meyer, of proseand itseffects

sterdam:NorthHolland Publishing Co. forrepresenting Minsky,Marvin. 1975. A framework knowledge. In P. H. Winston (Ed.) The psychologyof computervision. New York: McGraw-HillPublishing Co. David E. and AndrewOrtony.1977.The representation ofknowledgein Rumelhart, memory.In R. C. Anderson,R. J. Spiro, and W. E. Montague (Eds.) Schooling

ates. Schank, Roger C. and Robert P. Abelson. 1977. Scripts,plans, goals, and understanding.Hillsdale, N. J.:Lawrence Erlbaum Associates.

and theacquisition N. J.:Lawrence Erlbaum AssociHillsdale, of knowledge.

and learning to read. New York:Holt, Rinehart Co. and Winston Publishing Tuinman, J. Jaap and Mary Ella Brady. 1974. How does vocabulary account for varianceon readingcomprehension tests?A preliminary instructional analysis.In

Frank. 1971.Understanding A psycholinguistic Smith, reading: analysis of reading P. L. Nacke (Ed.) Interaction: Research and practice reading. forcollege-adult

Clemson, S.C.: NationalReading Conference. West, Michael. 1953. A general service list of English words. London: Longmans, Green & Co.

This content downloaded from 148.206.53.9 on Sat, 25 Jan 2014 23:44:10 PM All use subject to JSTOR Terms and Conditions

516

TESOL Quarterly

in M. C., Carolyn Marks,and Marlene Doctorow. 1974. Word frequency Wittrock, Journal readingcomprehension. of Educational Research,67:259-262. M. C. 1975. Reading as a generativeprocess.Journal Wittrock, of Educational Psychology,67:484-489. & Woodworth,RobertS. 1938.Experimental psychology.New York: Holt,Rinehart Winston Co. Publishing Yousef, Fathi S. 1968. Cross-cultural testing:An aspect of the resistancereaction.

18,3-4:227-234. Language Learning,

This content downloaded from 148.206.53.9 on Sat, 25 Jan 2014 23:44:10 PM All use subject to JSTOR Terms and Conditions

Potrebbero piacerti anche