ED 373 305 AUTHOR TITLE Report No. 14.DOCUMENT RESUME ED 373 305 CS 011 804 AUTHOR Booth, James R.;...

DOCUMENT RESUME

ED 373 305 CS 011 804

AUTHOR Booth, James R.; Hall, William S.TITLE Relationship of Reading Comprehension to the

Cognitive Internal State Lexicon. Reading ResearchReport No. 14.

INSTITUTION National Reading Research Center, Athens, GA.;National Reading Research Center, College Park,MD.

SPONS AGENCY Office of Educational Research and Improvement (ED),Washington, DC.

PUB DATE 94CONTRACT 117A20007NOTE 44p.PUB TYPE Reports Research/Technical (143)

EDRS PRICE MF01/PCO2 Plus Postage.DESCRIPTORS Higher Education; Intermediate Grades; Reading

Achievement; *Reading Comprehension; ReadingResearch; Secondary Education; VocabularyDevelopment; *Word Recognition

IDENTIFIERS District of Columbia; University of Maryland

ABSTRACT

A study compared students' cognitive word knowledgeof the cognates of "think" and "know" within a theoretical frameworkfocused on hierarchical levels of meaning. Subjects were 31 fifth, 32seventh, and 21 tenth graders attending single-gender private schoolsin the Washington, D.C. metropolitan area, and 70 collegeundergraduate students in an introductory psychology course at theUniversity of Maryland. Cognitive words form a category within theinternal state lexicon and may be central to accessing, monitoring,and transforming internal states, all of which seem to be processescritical to reading comprehension. Cognitive word knowledge waspositively correlated with achievement scores. The correlations withcognitive word knowledge were higher for verbal (vocabulary andreadint comprehension) than quantitative achievement scores, andcognitive word knowledge increased with age. However, the order ofacquisition of cognitive words depended on a complex interactionbetween the frequency of the cognitive word in established wordfrequency counts, the level of meaning as determined by theconceptual difficulty hierarchy, and whether the cognitive word was acognate of "think" or "know." (Contains 61 references, 6 tables, and3 figures of data. The cognitive word task is attached.)(Author/RS)

Reproductions supplied by EDRS are the best that can be madefrom the original document.

***********************************************************************

Relationship of Reading Comprehensionto the Cognitive Internal State Lexicon

James R. BoothWilliam S. Hall

University of Maryland

U S DEPARTMENT OF EDUCATIONou.er, Lth.cat. na Resca rn anc$

EDUCATIONAL RESOURCES INFORMATIONCENTER (ERICI

ti(Ths document ' ',eon reproduced asreceived from I :ton o. organizationoriginating it

Minor changes have been made toimprove reproduction quality

Points of view or opinions stated in thisdocument do not nocessanty representoffival OERI position or policy

NRRC NationalReading ResearchCenter

READING RESEARCH REPORT NO. 14

Spring 1994

PEST COPY AVAILABLE

NRRCNational Reading Research Center


James R. BoothWilliam S. Hall

University of Maryland College Park

READING RESEARCH REPORT NO. 14Spring 1994

The work reported herein is a National Reading Research Project of the University of Georgiaand University of Maryland. It was supported under the Educational Research andDevelopment Centers Program (PR/AWARD NO. 117A20007) as administered by the Officeof Educational Research and Improvement, U.S. Department of Education. The findings andopinions expressed here do not necessarily reflect the position or policies of the NationalReading Research Center, the Office of Educational Research and Improvement, or the U.S.

Department of Education.

r)0

NRRC NationalReading ResearchCenter

Executive CommitteeDonna E. Alvermann, Co-DirectorUniversity of Georgia

John T. Guthrie, Co-DirectorUniversity of Maryland College Park

James F. Baumann, Associate DirectorUniversity of Georgia

Patricia S. Koskinen, Associate DirectorUniversity of Maryland College Park

Linda C. DeGroffUniversity of Georgia

John F. O'FlahavanUniversity of Maryland College Park

James V. HoffmanUniversity of Texas at Austin

Cynthia R. HyndUniversity of Georgia

Robert SerpellUniversity of Maryland Baltimore County

Publications Editors

Research Reports and PerspectivesDavid Reinking, Receiving EditorUniversity of Georgia

Linda Baker, Tracking EditorUniversity of Maryland Baltimore County

Linda C. DeGroff, Tracking EditorUniversity of Georgia

Mariam Jean DreherUniversity of Maryland

Instructional ResourcesLee Galda. University of Georgia

Research HighlightsWilliam G. HollidayUniversity of Maryland College Park

Policy BriefsJames V. HoffmanUniversity of Texas at Austin

VideosShawn M. Glynn, University of Georgia

NRRC StaffBarbara F. Howard, Office ManagerCanine R. Bush, Senior SecretaryUniversity of Georgia

Barbara A. Neitzey, Administrative AssistantValerie Tyra, AccountantUniversity of Maryland College Park

National Advisory BoardPhyllis W. AldrichSaratoga Warren Board of Cooperative EducationalServices, Saratoga Springs, New York

Arthur N. ApplebeeState University of New York, Albany

Ronald S. BrandtAssociation for Supervision and CurriculumDevelopment

Marsha T. DeLainDelaware Department of Public InstructionCarl A. GrantUniversity of Wisconsin-Madison

Walter KintschUniversity of Colorado at Boulder

Robert L. LinnUniversity of Colorado at Boulder

Luis C. MollUniversity of Arizona

Carol M. SantaSchool District No. 5Kalispell, Montana

Anne P. SweetOffice of Educational Research and Improvement,U.S. Department of Education

Louise Cherry WilkinsonRutgers University

Production EditorSusan L. YarboroughUniversity of Georgia

Dissemination CoordinatorJordana E. RichUniversity of Georgia

Text FormatterAnn VanstoneUniversity of Georgia

NRRC - University of Georgia318 AderholdUniversity of GeorgiaAthens, Georgia 30602-7125(706) 542-3674 Fax: (706) 542-3678INTERNET: [email protected]

NRRC - University of Maryland College Park2102 J. M. Patterson BuildingUniversity of MarylandCollege Park, Maryland 20742(301) 405-8035 Fax: (301) 314-9625INTERNET: [email protected]

About the National Reading Research Center

The National Reading Research Center (NRRC) isfunded by the Office of Educational Research andImprovement of the U.S. Department of Education toconduct research on reading and reading instruction.The NRRC is operated by a consortium of the Universi-ty of Georgia and the University of Maryland CollegePark in collaboration with researchers at several institu-tions nationwide.

The NRRC's mission is to discover and documentthose conditions in homes, schools, and communitiesthat encourage children to become skilled, enthusiastic,lifelong readers. NRRC researchers are committed toadvancing the development of instructional programssensitive to the cognitive, sociocultural, and motiva-tional factors that affect children's success in reading.NRRC researchers from a variety of disciplines conductstudies with teachers and students from widely diversecultural and socioeconomic backgrounds in prekinder-garten through grade 12 classrooms. Research projectsdeal with the influence of family and family-schoolinteractions on the development of literacy; the interac-tion of sociocultural factors and motivation to read; theimpact of literature-based reading programs on readingachievement; the effects of reading strategies instruction

comprehension and critical thinking in literature,science, and history; the influence of innovative groupparticipation structures on motivation and learning; thepotential of computer technology to enhance literacy;and the development of methods and standards foralternative literacy assessments.

The NRRC is further committed to the participationof teachers as full partners in its research. A betterunderstanding of how teachers view the development ofliteracy, how they use knowledge from research, andhow they approach change in the classroom is crucial toimproving instruction. To further this understanding,the NRRC conducts school-based research in whichteachers explore their own philosophical and pedagogi-cal orientations and trace their professional growth.

Dissemination is an important feature of NRRC activi-ties. Information on NRRC research appears in severalformats. Research Reports communicate the results oforiginal research or synthesize the findings of severallines of inquiry. They are written primarily for re-searchers studying various areas of reading and readinginstruction. The Perspective Series presents a widerange of publications, from calls for research and

commentary on research and practice to first-personaccounts of experiences in schools. instructionalResources include curriculum materials, instructionalguides. and materials for professional growth, designedprimarily for teachers.

For more information about the NRRC's researchprojects and other activities, or to have your nameadded to the mailing list, please contact:

Donna E. Alvermann, Co-DirectorNational Reading Research Center318 Aderhold HallUniversity of GeorgiaAthens, GA 30602-7125(706) 542-3674

John T. Guthrie, Co-DirectorNational Reading Research Center2102 J. M. Patterson BuildingUniversity of MarylandCollege Park, MD 20742(301) 405-8035

r

NRRC Editorial Review Board

Patricia AdkinsUniversity of Georgia

Peter AfflerbachUniversity of Maryland College Park

JoBeth AllenUniversity of Georgia

Patty AndersUniversity of Arizona

Tom AndersonUniversity of Illinois at Urbana-Champaign

Harriette ArringtonUniversity of Kentucky

Irene BlumPine Springs Elementary SchoolFalls Church. Virginia

John BorkowskiNotre Dame University

Cy nthia Bow enBaltimore County Public SchoolsTowson, Maniand

Martha CarrUniversity of Georgia

Suzanne ClewellMontgomery County Public SchoolsRockville. Marlwid

Joan ColeyWestern Maryland College

Michelle CommeyrasUniversity of Georgia

Linda CooperShaker Heights City SchoolsShaker Heights, Ohio

Karen CostelloConnecticut Department of EducationHartford, Connecticut

Karin DahlOhio State University

Lynne Diaz-RicoCalifornia State University-San

Bernardino

Pamela DunstonClemson University

Jim FloodSan Diego State University

Dana FoxUniversity of Arizona

Linda GambrellUniversity of Maryland Colic. ! Park

Valerie GarfieldChattahoochee Elementary SchoolCumming, Georgia

Sherrie Gibney-Shermanithens-Clarke County Schools

Athens, Georgia

Rachel GrantUniversity of Maryland College Park

Barbara GuzzettiArizona State University

Jane HaughCenter for Developing Learning

Potentials:Iver Spring. Maryland

Beth Ann HerrmannUniversity of South Carolina

Kathleen HeubachUniversity of Georgia

Susan HillUniversity of Maryland College Park

Sally Hudson-RossUniversity of Georgia

Cynthia HyndUniversity of Georgia

Robert JimenezUniversity of Oregon

Karen JohnsonPennsylvania State University

James KingUniversity of South Florida

Sandra KhnbrellWest Hall Middle SchoolOakwood, Georgia

Kate KirbyGwinnett County Public SchoolsLawrenceville. Georgia

Sophie KowzunPrince George's County SchoolsLandover, Maryland

Linda LabboUniversity of Georgia

Rosary LalikVirginia Polytechnic Institute

Michael LawUniversity of Georgia

Sarah McCartheyUniversity of Tex-as at Austin

Veda McClainUniversity of Georgia

Lisa McFalLsUniversity of Georgia

Mike McKennaGeorgia Southern University

Donna MealeyLouisiana State University

Barbara MichaloveFowler Drive Elementary SchoolAthens. Georgia

Akintunde MorakinyoUniversity of Maryland College Park

Lesley MorrowRutgers University

Bruce MurrayUniversity of Georgia

Susan NeumanTemple University

Caroline NoyesUniversity of Georgia

John O'FlahavanUniversity of Maryland College Park

Penny OldfatherUniversity of Georgia

Joan PagnuccoUniversity of Georgia

Barbara PalmerMount Saint Mary's CoPege

Mike PickleGeorgia Southern University

Jessie PollackMaryland Department of EducationBaltimore, Maryland

Sally PorterBlair High SchoolSilver Spring, Maryland

Michael PressleyState University of New York

at Albany

John ReadenceUniversity of Nevada-Las Vegas

Tom ReevesUniversity of Georgia

Lenore RinglerNew York University

Mary RoeUniversity of Delaware

Nadeen T. RuizCalifornia State University-Sacramento

Rebecca SammonsUniversity of Maryland College Park

Paula SchwanenflugelUniversity of Georgia

Robert SerpellUniversity of Maryland Baltimore

County

Betty ShockleyFowler Drive Elementary SchoolAthens. Georgia

Susan SonnensrheinUniversity of Maryland Baltimore

County

Steve StahlUniversity of Georgia

Anne SweetOffice of Educational Research

and Improvement

Liqing TaoUniversity of Georgia

Ruby ThompsonClark Atlanta University

Louise TomlinsonUniversity of Georgia

Sandy TumarkinStrawberry Knolls Elementary SchoolGaithersburg, Maryland

Sheila ValenciaUniversity of Washington

Bruce VanSledrightUniversity of Maryland College Park

Chris WaltonNorthern Territory UniversityAustralia

Janet WatkinsUniversity of Georgia

Louise WaynantPrince George's County SchoolsUpper Marlboro, Maryland

Priscilla WaynantRolling Terrace Elementary SchoolTakoma Park, Maryland

Dera WeaverAthens-Clarke County SchoolsAthens, Georgia

Jane WestUniversity of Georgia

Steve WhiteUniversity of Georgia

Allen WigfieldUniversity of Maryland College Park

Shelley WongUniversity of Maryland College Park

About the Authors

James R. Booth is a graduate student in Devel-opmental Psychology at the University of Mary-land College Park.

William S. Hall is a Professor of DevelopmentalPsychology and Chair of the Department of Psy-chology at the University of Maryland CollegePark. Please address all correspondence to Wil-liam S. Hall, Department of Psychology, Universi-ty of Maryland, College Park, MD 20742.

National Reading Research CenterUniversities of Georgia and MarylandReading Research Report No. 14Spring 1994


James R. BoothUniversity of Maryland College Park

William S. HallUniversity of Maryland College Park

Abstract. The authors compared fifth-, sev-enth-, and tenth-graders, and college under-graduates' cognitive word knowledge of thecognates of think and know within a theoret-ical framework focused on hierarchicallevels of meaning. Cognitive words form acategory within the internal state lexicon andmay be central to accessing, monitoring, andtransforming our internal states, all of whichseem to be processes critical to readingcomprehension. Cognitive word knowledgewas positively correlated with achievementscores. The correlations with cognitive wordknowledge were higher for Verbal (voca-bulary and reading comprehension) thanQuantitative achievement scores, and cogni-tive word knowledge increased with age.However, the order of acquisition of cogni-tive words depended on a complex interac-tion between the frequency of the cognitiveword in established word frequency counts,the level of meaning as determined by theconceptual difficulty hierarchy, and whetherthe cognitive word was a cognate of think orknow.

Cognitive words such as think and know forma category within the internal state lexicon

(Hall & Nagy, 1986) and may be central toaccessing, monitoring, and transforming ourinternal states (Scholnick & Hall, 1991). Manycognitive words are polysemous and can be de-fined along a hierarchy from simple perceptionto complex planning (Frank & Hall, 1991;Hall, Scholnick, & Hughes, 1987). Accordingto this hierarchy, the higher the level of mean-ing the more conceptually demanding themore internal processing is required. We arguethat cognitive words may be centrally involvedin the development of skilled reading compre-hension.

Cognitive words can provide a medium thatmakes it possible to engage in metacognitiveacts relevant to the reading process, such asgenerating a goal for reading, communicatingthe intended meaning of a text, and evaluatingone's level of understanding. Similarly, cogni-tive words can equip the reader with a vehicleby which to evaluate different comprehensionstrategies critically or to reflect on the logicalorganization and interdependence of the com-

2 James R. Booth & William S. Hall

ponents of a text. Our elaborated cognitiveword lexicon allows us to make fine-graineddistinctions between cognitive states. Cognitivewords "convey shades of meaning which addsuccinctness and precision to the lexicon" andsupply us with "a greater capacity for descrip-tion and definition" (Corson, 1985, p. 61).While skilled reading comprehension requiresthe use of all of the aforementioned skills,direct empirical evidence for the claim thatcognitive ward knowledge is central to thedevelopment of reading comprehension is

sparse (see Olson & Torrance, 1986, 1987).The present study sought to remedy this

situation by providing an empirical test of thehypothesized relationship between cognitiveword knowledge and reading comprehension.In pursuit of this aim, a task was designed tomeasure knowledge of the cognates of thinkand know while simultaneously varying thedimensions of established frequency in theEnglish language (Carroll, Richman, & Dav-ies, 1971) and the level of meaning difficulty(Frank & Hall, 1991). The degree of chil-dren's knowledge of the 'gnitive words wasexpected to correlate highly with readingachievement scores.

Certain cognitive words, such as think andknow, appear very early in a child's lexicon.Children begin to use cognitive words at aboutthree years of age (Bretherton & Beeghly,1982), but their use remains infrequent andlimited primarily to pragmatic functions. Forexample, Shatz, Wellman, and Silber (1983)found that the earliest use of cognitive words isfor pragmatic or conversational purposes , suchas in directing the action. By the end of theirthird year, children begin to use cognitivewords in a way that suggests semantic under-

standing (Shatz et al., 1983), but they do notunderstand the distinctions between manycognitive words, such as remember, know, andguess, until approximately four years of age(Johnson & Wellman, 1980). At around fiveyears of age, children can differentiate betweenknow and think (Johnson & Maratsos, 1977;Moore, Bryant, & Furrow, 1989), know andguess (Miscione, Marvin, O'Brien, & Green-berg, 1978; Moore et al. 1989), and rememberand forget (Johnson, 1981).

Fine-grained distinctions between othercognitive words are not learned until later.Seven-year-olds' judgment of the truth of thecomplement of the cognitive words pretend,know, and think are determined primarily bythe plausibility of the complement and onlysecondarily by the factivity of the verb (Olson& Torrance, 1986); the think and guess dis-tinction is not attained until children are eightyears old (Moore et al., 1989). Furthermore,children do not understand believe, which canbe both factive and nonfactive, until after ageseven (Abbeduto & Rosenberg, 1985), andeven high school and college students haveincomplete knowledge of more complex cog.tive words such as predict, interpret, infer,conclude, and assume (Astington & Olson,1990).

All cognitive words are not acquired simul-taneously, in part because certain cognitivewords differ semantically in specific but subtlerespects. For example, know, remember,forget, and guess refer to the accessibility ofknowledge (Hall et al., 1987); pretend, guess,and know involve presuppositions of disbelief,uncertainty, and belief (Macnamara, Baker, &Olson, 1976); see and know refer to internalversus external experience (Wellman & Estes,

NATIONAL READING RESEARCH CENTER, READING RESEARCH REPORT NO. 14

10

Reading Comprehension and Cognitive Words 3

1987). The development of the internal statelexicon is a gradual, incremental process thatdepends on the particular cognitive words to belearned. An example of this dependence is thatmany cognitive words are polysemous, that is,they have more than one meaning. This poly-semy imposes a constraint on children's acqui-sition of cognitive words.

Hierarchy of Meaning of Cognitive Words

Frank and Hall (1991) proposed that certaincognitive words have a hierarchy of level ofmeaning characterized by increasing abstract-ness and conceptual difficulty (see also Hall etal., 1987). They proposed that the cognitiveinternal state lexicon adheres to a structure thatinvolves the following levels: (1) registering anexperience perceptually; (2) determining thefamiliarity of an experience and embedding itin a factual network: (3) understanding inter-connections among concepts; (4) making one'spresuppositions about the experience explicit;(5) commenting on how internal processing isbeing done; and (6) assessing future intention,which implies an understanding and integrationof past events. They referred to these levels asperception, memory, understanding, evalua-tion, metacognition, and planning, respec-tively. It was proposed that perception is theleast complex and requires a limited amount ofinternal processing, while planning is the mostcomplex and demands the greatest amount ofinternal processing. The middle levels ofmeaning follow a hierarchy of increasingconceptual difficulty between these endpoints.Frank and Hall (1991) hypothesized that thehigher the level of meaning for a word, the lessit is likely to be used in discourse. Indeed, they

found that as the level of meaning increased,the frequency decreased for both adult andchild verbal frequency of the cognitive wordknow. Although, the levels of perception,memory, and understanding were statisticallydifferentiated, the levels of evaluation, meta-cognition, and planning were not. Hughes(1985) obtained similar results in a compre-hension task assessing cognitive word knowl-edge in three-, six-, and nine-year-old children.

Frank and Hall (1991) then tried to incor-porate the cognitive word think into their levelof meaning hierarchy. They found that evlua-non was used most often and that the other fivelevels of meaning were statistically undifferen-tiated for both children and adults. On the basisof these findings, they proposed that cognitivewords are organized hierarchically, but thatdifferent cognitive words may have a differentorganization of levels that depends on bothconceptual difficulty and prototypicality. Ac-cording to their hypothesis, prototypical mean-ings will be acquired first, but cognitive wordswhose prototypical meanings are of a lowerlevel will be semantically mastered earlier thanwords whose prototypical meanings are of ahigher level. This claim is supported by find-ings from an analysis of young children whoused know before think more frequently innatural discourse (Hall, Nagy, & Linn, 1984).Know may have a lower level prototypicalmeaning than think.

Certain cognitive words have several prag-matic as well as semantic functions, and thismay encourage children to develop metalin-guistic knowledge. A semantic use of a cogni-tive word occurs when it contributes directly tothe intended meaning of an utterance, such as"Sally knows the answer." In contrast, a prag-



matic use contributes indirectly, if at all, to themeaning of an utterance; it might be a hedge,a conversational device, an indirect request, oran attention-getting device such as "You know,I need to go to the store" (Hall & Nagy, 1986).Cognitive words appear to have more meaningsand functions than words that name objects,events, or situations. Because the contexts inwhich people use cognitive words vary, expo-sure to cognitive words may be particularlyadvantageous in the development of metalin-guistic abilities. For example, when a childrealizes that a word is only a symbol for itsreferent, that context determines the poly-semous nature of words, and that language canbe an object of thought, his or her linguisticability is advanced considerably. Also, childrenare likely to generalize this knowledge ofmultiple meanings to other lexical domains andcompare different levels of meaning within andbetween lexical domains. Knowing more aboutthe different meanings of one cognitive wordand the relationships between the meanings ofdifferent cognitive words may make childrenmore aware of distinctions they make and howthey make them. This may lead to mastery oftheir knowledge system. All of these processesseem to be critical for high-level text under-standing.

Developmental Course of CognitiveWord Acquisition

Clark's (1983) theory of semantic developmentprovides valuable insights about the develop-mental course of cognitive words. She assertsthat the order of meaning acquisition is afunction of the interaction between both therelative complexity of the meaning of a word

and nonlinguistic strategies with the linguisticenvironment during meaning acquisition. Therelative complexity of the meaning is a functionof the number of meanings, the degree ofoverlap between meanings, and the number ofpossible applications of an individual meaning.Polysemous words such as think and know maybe difficult for children to acquire because theymust develop hypotheses appropriate to eachpossible application of a word's meaning, aswell as contrast all the meanings of this wordwith all the meanings of other words.

According to Clark (1983), meaning acqui-sition also depends on nonlinguistic strategies.Nonlinguistic strategies refer to a particularconceptual organization possessed by a childthat coincides with and therefore makes theacquisition of certain word meanings easier(see Clark, 1980). For example, Clark (1973claims that children acquire locative (in, on,and under) terms in a regular developmentalsequence. Consequently, children may acquirecertain cognitive words comparatively latebecause of their own conceptual organization.Objects, situations, and relations are directlyobservable; therefore, acquisition of wordsreferring to these concepts is probably influ-enced directly by nonlinguistic strategies. Incontrast, most cognitive words are verbs, andverbs have a more abstract and indistinctmental representation than nouns (see Hutten-locher & Lui, 1979; Anglin, 1986). Childrenmay be more able to use nonlinguistic strate-gies in learning simple and concrete cognitivewords, such as perceive and remember, but notin learning complex and abstract cognitivewords, such as reflect and evaluate. At themore complex levels, cognitive words refer tovery abstract, inaccessible, and subtle mental


1 2


states; therefore, it is unlikely that cognitivewords are mapped onto pre-existing conceptualcategories. Instead, language (cognitive words)probably influences the development of chil-dren's concepts of complex mental states invery important ways (Scholnick & Hall, 1991).

Since cognitive words are relatively com-plex and may not benefit from nonlinguisticstrategies, a child's acquisition of cognitivewords may be highly dependent on linguisticinput. Similarly, Scholnick and Hall (1991)conclude that "conscious awareness of mentalstates and the refinement of that awareness ismade possible by socialization into the folkpsychology of a culture through language" (p.444). Unfortunately, children are rarely ex-posed to cognitive words either spoken (Hall etal., 1984; Smith & Meux, 1970) or written byparents or teachers (Carroll et al., 1971;Thorndike & Lorge, 1944).

For example, Astington (1991) found thatjunior high-school science texts rarely containcognitive words and speech. act verbs such as

Epistemic verbs such as define, explain,hypothesize, infer, and interpret are absent, andbelieve occurs only once. This deficiency isparticularly disturbing since many text-basedacademic skills require the cognitive monitor-ing that cognitive words label (Hall et al.,1987). "What schooling appears to provide iscompetence in talking about text, about ques-tions, about answers, in a word, competencewith a metalanguage" (Olson & Astington,1990, p. 563).

RFsearch suggests that the cognitive wordlexicon may develop late in children becausethe necessary linguistic input is deficient. Therole of linguistic input in the development ofcognitive word knowledge is uncertain; only

two studies have compared children's andadults' use of cognitive words. The few studiesthat have been conducted have found a signifi-cant relationship. For example, in a study of13- to 28-month olds, Beeghly, Bretherton, andMervis (1986) found that the numbers andkinds of internal state rances by the motherwere positively correlated with the child'sspoken frequency of different internal statewords, of internal state words referring to selfand other, and of decontextualized internalstate words. Similarly, Hall et al. (1987) founda significant correlation between child use (4years 6 months to 5 years) and parental use inboth levels of meaning and the diversity ofcognitive words used. These studies suggestthat a child's development of cognitive wordknowledge is highly dependent on adult fre-quency of verbal use.

Several studies also suggest that exposureto text may facilitate cognitive word learning.TI different contexts in which a cognitivewoad appears in text may refine its definitionsand functions (Olson & Astington, 1986;Robinson, Goelman, & Olson, 1983; Olson &Hildyard, 1981). Similarly, a child probablymasters the different meanings of know andthink through reading, as the difference be-tween two cognitive words may be highlightedwhen they appear in the same sentence (Robin-son, 1980). Taken together, this research sug-gests that cognitive word knowledge emergesfrom verbal and written exposure to thesewords, but cognitive word knowledge alsoappears to be an essential prerequisite for high-level text understanding.

Despite the research suggesting a strongrelationship between cognitive words and high-level text understanding, only three studies


13


have investigated that relationship empirically(Olson & Torrance, 1986, 1987; Astington &Olson, 1990). In one experiment, Olson andTorrance (1987) found that justifying answersby referring to the text was significantly corre-lated with cognitive word knowledge for third-graders but not for first-graders. In anotherexperiment, Olson and Torrance (1987) foundcognitive word knowledge in third-graders tobe significantly correlated with a listening com-prehension score that measured inferencesdrawn from the story. In addition, Olson andTorrance (1986), in a longitudinal study ofchildren from 5 years 6 months to 7 years 6months, found a total combined score on fourcognitive word tasks to be correlated signifi-cantly with vocabulary ability, conversationability, and reading ability. These three lines ofevidence suggest that reading comprehensionand cognitive word knowledge are stronglyrelated.

Our thinking is in accord with this evidencebecause an understanding of the semantic andpragmatic uses of cognitive words enhances areader's knowledge base a key element intext understanding. Furthermore, knowledge ofcognitive words seems to he centrally involvedin reading comprehension. For example, dif-ferentiating between what a character thinksand knows is essential for the interpretation ofa text. Cognitive words can provide a mediumthrough which a character's mental states canbe interpreted, as well as through which thepast, present, and future goals and motives ofthat character can be analyzed. Cognitivewords allow the reader to designate and reflecton what is true or false, real or unreal, and

ambiguous or unambiguous in text. Thus,cognitive words have special salience in under-standing and evaluating written language.

In summary, child) cn have difficulty inmastering the cognitive word lexicon for atleast three reasons. First, since the cognitiveword lexicon is very intricate, subtle contrastsand comparisons must be learned in order forcognitive words to be used appropriately (Hall& Nagy, 1986; Frank & Hall, 1991). Second,since cognitive words are abstract and elusiveand deal with the inner workings of the mind,they may not benefit easily from nonlinguisticstrategies (Clark, 1983). Third, even thoughexposure to cognitive words appears to beessential for their efficient acquisition (Beeghlyet al., 1986; Olson & Astington, 1986), manychildren are exposed to cognitive words inwritten or oral form infrequently (Astington &Olson, 1990; Carroll et al., 1971). In addition,since cognitive word knowledge appears to bestrongly related to high-level text understand-ing due to the pertinent content knowledge itprovides, this deficient linguistic environmentis especially troubling.

Specific Aims of Research

First, we wanted to test the hypothesis thatcognitive word knowledge is highly related tohigh-level text understanding as measured byreading comprehension achievement scores.We predicted that cognitive word knowledgewould (a) correlate positively with all achieve-ment scores; (b) correlate more highly withVerbal than with Quantitative achievementscores; (c) correlate more highly with Voca-



bulary than with Reading Comprehensionachievement scores; and (d) that there wouldbe a lower correlation between high-frequencycognitive words with achievement scores thanwould be found for low-frequency cognitivewords.

Second. we wanted to investigate the hier-archical taxonomy of cognitive words proposedby Frank and Hall (1991) in order to study itsrelationship to reading comprehension. In thisconnection we made several predictions: (a)cognitive word knowledge would increase withage; (h) the acquisition of cognates of thinkwould he earlier than the acquisition of cog-nates of know; and (c) high levels of meaningwould he acquired after low levels of meaning.

METHODSubjects

Subjects represented elementary, middle-, andhigh-school, and college levels. The gradeschool subjects attended different single-genderprivate schools in the Washington, DC metro-politan area. There were 31 fifth-grade stu-dents, (M age = 11.3; SD = 0.5); 32 seventh-grade students, (M age = 12.7; SD = 0.4);and 21 tenth-grade students, (M age = 15.6;SD = 0.5). The mean ages of the males andfemales within grades were not significantlydifferent; their data are combined for presenta-tion. The 70 undergraduate students, (M age =19.9; SD = 1.6) attended the University ofMaryland and participated in the study to fulfillan Introductory Psychology requirement. All

except two of the grade-school students and allof the undergraduates completed the study.

Materials

Standardized Achievement Scores. The

latest Educational Records Bureau (ERB)independent school norm percentiles wereobtained from the school records of the grade-school subjects. The ERB subscores includedVerbal (combined Vocabulary and ReadingComprehension) and Quantitative scores. Allscores were obtained by a method ensuringanonymity and confidentiality. The undergrad-uate students supplied their subscores on theScholastic Aptitude Test (SAT) and their gradepoint average (GPA) along with documentationsuch as an academic transcript and an SATscore stub. The SAT subscores included Verbal(combined Vocabulary and Reading Compre-hension) and Quantitative scores.

Cognitive Word Task. The cognitive wordtask was modeled after Astington and Olson(1990) and consisted of 24 short stories (four toseven sentences). Each cognitive word passagewas syntactically and semantically simple inorder to ensure the assessment of cognitiveword knowledge primarily, not reading com-prehension. Twelve stories contained the cog-nitive word think, and twelve contained thecognitive word know. Of those, two storiescharacterized each of the six levels of meaningfor think and know (Frank & Hall, 1991). Ofthose two stories, there was one story with alow-frequency and one with a high-frequencyreplacement cognitive word (see Appendix A).Each replacement cognitive word was con-tained within one of four multiple-choicesentences following the story. The subject was

asked to read the story and to then choose the



Table 1. High- and Low-Frequency Replacement Cognitive Words for Each Level of Meaning for Thinkand Know

Level

THINK

High-Frequency Low-Frequency

PerceptionMemoryUnderstandingEvaluationMetacognitionPlanning

noticeforgetrealizedoubt

considerintend

concentrateremind

apprehendinfer

contemplatepredict

Level

KNOW

High-Frequency Low-Frequency

PerceptionMemoryUnderstandingEvaluationMetacognitionPlanning

observerecognize

understandconcludereflectexpect

perceiverecall

comprehendhypothesize

analyzeanticipate

sentence containing the replacement cognitiveword that accurately represented the level ofmeaning of think or know in the story.

Know and think were chosen as the poly-semous words in this study for several reasons.First, several studies have focused on theacquisition of know and think (Macnamara etal., 1976; Johnson & Maratsos, 1977; Johnson& Wellman, 1980; Olson & Torrance, 1986;Astington & Olson, 1990). Second, know andthink have been shown to be the most frequent-ly used cognitive words in a child's lexicon

(Shatz et al., 1983; Hall et al., 1984; Beeghlyet al., 1986). Third, the levels of meaning ofknow and think have already been defined andstudied (Hall et al., 1987; Hughes, 1985;Frank & Hall, 1991). Fourth, both know andthink lend themselves to multiple meanings thatcan be inferred from their context of use.

The replacement cognitive words for knowand think that were included in the list ofchoices were selected from previous studies ofcognitive words (Hall & Nagy, 1986; Asting-ton & Olso:., 1990) as well as by reference to



Roget's Thesaurus. Cognitive words weredivided into the six levels of meaning accord-ing to three independent raters' judgments oftheir typical use. The three raters were gradu-ate students familiar with the Frank and Hall(1991) level of meaning hierarchy. If the threeraters did not agree on the level assigned to aparticular cognitive word, that word was dis-carded from the list. Frequencies of theseprospective cognitive words were collectedfrom the Hall et al. (1984) spoken languagecorpus, the Thorndike and Lorge (1944) writ-ten word corpus, and the Carroll et al. (1971)school textbook corpus.

The low- and high-frequency replacementcognitive words were randomly chosen from acognitive word list created in the followingmanner: The most and least frequent words ineach level of meaning were considered outliersand were eliminated. Then, each level wasdivided in half by the median frequency cogni-tive word (based on Carroll et al., 1971) andthis median frequency was eliminated. High-frequency cognitive words were defined opera-tionally as similar if they had a frequencydifference of less than 1,000 words per 5million words. Low-frequency cognitive wordswere defined operationally as similar if theyhad a frequency difference of less than 150words per 5 million words. Two groups, onelow- and the other high-frequency, separatedby the median, were formed.

The distractor items were also chosen ran-domly from the cognitive word list. For eachquestion, distractor frequency matched correctanswer frequency according to the aboveoperational definitions of frequency similarity.For the four middle levels of meaning, one

distractor item was chosen randomly from thelevel of meaning immediately below, and twodistractor items were chosen randomly fromthe level of meaning immediately above. Forthe lowest level of meaning, two distractoritems were chosen randomly from the level ofmeaning immediately above, and one distractoritem was chosen randomly from two levels ofmeaning above. For the highest level of mean-ing, two distractor items were chosen randomlyfrom the level of meaning immediately below,and one distractor item was randomly chosenfrom two levels of meaning below. All pro-spective distractor cognitive words were evalu-ated to determine whether they fit syntacticallywithin the distractor sentence; if they did not,they were discarded. Table 1 contains a list ofthe levels of meaning of think and know and theappropriate replacement cognitive words.

The order of the passages in the cognitiveword task began with low-frequency and therewere four alternating blocks (six passages each)of high- and low-frequency; within each blockthe level of meaning increased from perceptionto planning. This organization was chosenover randomization because this structureprovided simplicity. By having the cognitiveword task begin with high-frequency and lowlevel of meaning, the reader was not discour-aged initially by unsuccessful performance. Inaddition, having low- or high-frequency cogni-tive word passages clustered together allowedfor symmetry in the distractor items. Thedistractor items were similar among same-frequency cognitive word passages; therefore,the subject was not unnecessarily confused bythe multitude of distractor items. After all, thepurpose of the cognitive word task was to


1 "'


Table 2. Reliability of the Cognitive Word Task Subscales

Reliability

Subscale AlphaInter-Item

Correlation

Perception .292 .093Memory .331 .110Understanding -.303 -.062Evaluation .409 .148Metacognition .440 .164Planning -.032 -.008Low-Level .535 .126High-Level .543 .129Think .605 .145Know .555 .111High-Frequency .437 .079Low-Frequency .652 .158Cognitive Word Total' .722 .120

Note. aUnstandardized alpha was used because we added raw scores to form a composite score beforestandardization; beach score was subtracted from the mean of its age group before alpha calculationbecause mean differences between ages can inflate this coefficient; 'alpha reliability was calculatedwith the level of meaning subscales.

assess cognitive word knowledge, not readingcomprehension or test-taking strategies.

PROCEDURE

The cognitive word task was administered togroups of fifth-, seventh-, and tenth-gradestudents in their regularly scheduled classesand to the undergraduate students in a groupsetting at the University of Maryland. Theexperimenter read the instructions aloudthese were printed on the first page of the taskbooklet (see Appendix A) and answered allquestions the students asked. The task took a

maximum of 25 minutes to complete. Allsubjects finished within the allotted time.

Because the cognitive word task was devel-oped for this project, its reliability had to bedetermined. To do that, alpha and item-totalcorrelations were computed for total and sub-scores. (A reliability of a = .60 or greater isrecommended for basic research with broad-band instruments [Nunnally, 1978].) Theoverall reliability of the cognitive word taskwas high, a = .72. Furthermore, five out ofthe six reliability coefficients for the main ef-fects were greater than a = .53. Because of theexploratory nature of this research, the reli-



ability lower limit was set at a = .15, whichallowed us to include most of the subscales inthe analysis. Unfortunately, the levels of un-derstanding (a = -.30) and planning (a= -.03)were not sufficiently reliable to include infurther analyses (see Table 2 for reliabilitycoefficients of the main effect variables). Reli-ability coefficients were not computed withineach grade due to the small number of subjects.However, Cochran's C-test for the homogene-ity of variance for cognitive word total wasinsignificant across age classes, suggestingequal error of measurement within each gradelevel (for extensive information regarding theestablishment of reliability and validity of thecognitive word task, please contact the au-thors).

RESULTS

We will discuss the results in terms of the tworesearch aims: (1) to investigate the relation-ship between cognitive word knowledge andtext understanding as revealed through readingachievement scores, and (2) to apply a modelof the development of the internal state lexiconto cognitive word knowledge as measured by amultiple-choice task.

Cognitive Word Knowledge and ReadingComprehension Scores

In order to confirm the strong relationship wehypothesized between cognitive word knowl-edge and reading comprehension, a correlationmatrix was computed using cognitive wordtotal, high-frequency cognitive words, andlow-frequency cognitive words with achieve-ment scores (see Table 3). There were signifi-

cant correlations of the cognitive word totalwith achievement scores suggesting that thecognitive word task was validly measuring boththe construct of verbal ability and the academicachievement. Correlations between cognitiveword total and achievement scores were testedfor significant differences using Steiger's t-testfor dependent r's (1980). For all grade-school-ers, the correlation of cognitive word total washigher (but not significant) for Verbal (r =.33, p < .01) than for Quantitative (r = .22,p < .05) and higher (1 = 1.56, p < .1) for Vo-cabulary (r = .49, p < .001) than for ReadingComprehension (r = .35, p < .001). For un-dergraduates, the correlation of cognitive wordtotal was higher (t = 1.74, p < .05) for Verbal(r = .55, p < .001) than for Quantitative (r =.36, p < .01), and about the same for Vocabu-lary (r = .45, p < .01) and Reading Compre-hension (r = .47, p < .01). Convergent/discriminant construct validity was revealed bythe fact that cognitive word total correlatedhigher with Verbal than with Quantitativeachievement scores for the undergraduates andhigher with Vocabulary than with ReadingComprehension achievement scores for thegrade-school subjects. Furthermore, the corre-lations with cognitive word total were higherfor low than for high-frequency cognitivewords (see Table 3 for these correlations).Note that the correlations were typically lowerfor the seventh-graders. This relative lack ofassociation may be attributed to the less-than-ideal testing conditions for the seventh-grademales, as there were many environmentaldisturbances during their testing session.

In summary, our predictions that cognitiveword total would correlate significantly withachievement scores, and that cognitive word

NATIONAL. READING RESEARCH CENTER, READING RESEARCH REPORT NO. 14

12 James R. Booth dc William S. Hall

Table 3. Correlation of Total, High-Frequency, and Low-Frequency Cognitive Words with StandardizedAchievement Scores by Grade

Cognitive Word Total

Subscale

Grades/Level Means

5 7 10 College GradesAll

Groups

Verbal .47" .15 .34 .55" .33" .42-Quantitative .13 .16 .22' .28-Vocabulary .74.- .17 .55" .45" .49-Reading Comprehension .54- .21 .25 .35' .39-GPA .12

High-Frequency

Verbal .31' .26 .16 .33- .28*-Quantitative .25 .11 -.08 .19 .15'Vocabulary .61- .08 .44' .28' .39- .35-Reading Comprehension .58- .12 -.05 .34- .27" .29"GPA .12

Low-Frequency

Verbal .48- .15 .36 .51- 32" .41-Quantitative .31' .11 .30 .35- .23' .29-Vocabulary .62- .18 .44' .40- .40- .40-Reading Comprehension .26 .41. .38- .31" .34-GPA .08

Note. .p< .05, "p < .01, "'p < .001. Grades Means and All Groups Means adjusted for mean differences betweenageclasses. Since cognitive word total was significantly different between grades, pooling the different age classes maylower its correlation coefficient with the achievement scores variables: therefore, the cognitive word mean of each ageclass was subtracted from the individual scores within it.

total would correlate more highly with Verbalthan Quantitative achievement scores, andmore highly with Vocabulary than ReadingComprehension achievement scores weresupported. Moreover, our prediction that low-frequency cognitive words would correlatehigher than high-frequency cognitive wordswith achievement scores was supported.

Taxonomy of Cognitive Words Proposed byFrank and Hall (1991)

Since the academic ability of each age levelwas statistically similar, subject selection biasescould not account for any developmental differ-ences found in the current research. One-wayANOVAs with Verbal, Quantitative, Vocahu-


2V


Table 4. Means and Standard Deviations by Grade for Student Standardized Achievement Scores

Grades/Level

Subscale 5 7 10 College

Verbal 49.68a 55.07 53.80 530.2(29.5)h (24.9) (30.7) (80.9)

Quantitative 52.74 61.72 52.00 600.1(30.7) (24.0) (31.3) (90.4)

Vocabulary 49.68 60.24 51.95 540.7

(30.0) (19.9) (31.1) (80.9)

Reading 53.10 57.80 51.00 540.2

Comprehension (31.6) (21.6) (25.8) (90.9)

Grade Point 2.66

Average (.76)

Note. 'mean; hstandard deviation. Grade-schoolers' scores are ERB independent school norm percentiles;undergraduate scores are SAT-scaled scores. GPA is on a 4-point scale.

lary, and Reading Comprehension achievementscores revealed no age differences in achieve-ment scores. The grade-schoolers' ERB scoresand the undergraduate SAT scores were con-verted to z-scores to allow statistical compari-scn. Table 4 displays the means and standarddeviations by grade for the independent schoolnorm percentiles for the grade-school students.and the SAT-scaled scores for the undergradu-ate students.

Furthermore, the reading difficulty of thecognitive word passages did not confound anyof the observed differences in cognitive wordknowledge. Four 2 (Frequency) x 2 (Think/Know) x 2 (Level of Meaning) ANOVAs withthe Readability Index, the Fog Index, theaverage number of syllables per word, or theaverage number of words per sentence asdependent variables revealed no significant

differences. (The Right Writer computer pro-gram was used to compute these dependentvariables.) In addition, the reading difficulty ofcognitive word passages could not account forthe mean percent-correct difference observedbetween cognitive word passages. Correlationsbetween the mean percent-correct for thecognitive word passages and the above readingdifficulty indices for those passages wereinsignificant within each age level and acrossall age levels. Therefore, all age, frequency,think or know, and level of meaning differencesmust be accounted for by cognitive wordknowledge, and not the reading difficulty ofthe passages. (However, the reading compre-hension abilities necessary for reading thesepassages may account for some of the varianceexplained in achievement scores by the cogni-tive word task).


21


As mentioned previously, the cognitiveword passages were intentionally very simplein order to assess only cognitive word know-ledge. The readability indices of 16 of thecognitive word passages were less than thefifth-grade level the youngest age group inour study. These low readability indices sug-gest that reading comprehension abilities, asopposed to cognitive word knowledge, were aminor factor in the students' performance onthese passages. Furthermore, the readabilityindices for all of the cognitive passages werebelow the eighth-grade level, suggesting thatreading comprehension abilities were even lessof a factor in the performance of the tenth-graders and undergraduates.

Regarding cognitive word acquisition, wepredicted that proficiency in cognitive wordswould increase with age, low-frequency cogni-tive words would be acquired after high-fre-quency cognitive words, high level of meaningwould be acquired after low level of meaning,and cognates of know would be acquired aftercognates of think. In order to confirm ourprediction these age, frequency, think orknow, and level of meaning differences in

cognitive word acquisition, a 4 (Grade) x 2(Frequency) x 2 (Think/Know) x 2 (Level ofMeaning) ANOVA was computed. For thisANOVA, the perception and memory levelswere combined to form the low level of mean-ing category and the evaluation and meta-cognition levels were combined to form thehigh level of meaning category in order toattain at least four cognitive word passages percomparison group necessary for sufficientreliability. The means and standard deviationsby grade for cognitive word task subscales arepresented in Table 5. We subtracted 0.33 for

each wrong answer to correct for guessing;negative scores were counted as zero (seeAstington & Olson, 1990). In addition, all ofthe subscales scores were reduced to percent-correct to enable the comparison of subscalesthat had different numbers of passages withinthem.

All main effects were significant. Percent-correct for low level of meaning (M = .66)was significantly greater (F(1,156) = 34.42,p < .001) than percent-correct for high level ofmeaning (M = .52). Percent-correct for high-frequency cognitive words (M = .64) wassignificantly greater (F(1,156) = 20.94, p<.001) than low-frequency cognitive words (M= .50). Percent-correct for think (M = .65)was significantly greater (F(1,156) = 26.11.p< .001) than know (M = .53). Post hoc one-way ANOVAs with Scheffe range comparison(p < .05) between grades revealed the follow-ing: (a) fifth- and seventh-graders scoredsignificantly lower than tenth-graders andundergraduates on high level of meaning.know, low frequency, and cognitive word total;(b) fifth-, seventh-, and tenth-graders scoredsignificantly lower than undergraduates on lowlevel of meaning and think: (c) fifth-gradersscored significantly lower than tenth-graders onlow level of meaning, think, and high-frequen-cy; and (d) fifth- and seventh-graders scoredsignificantly lower than undergraduates onhigh-frequency. Our expectations that develop-mental differences would exist in cognitiveword acquisition, that low-frequency would heacquired after high-frequency, that high levelof meaning would he acquired after low levelof meaning, and that cognates of think wouldhe acquired after cognates of know were sup-ported.



Table 5. Means and Standard Deviations by Grade of Percent-Correct for Cognitive Word Task Subsea les

Grades/Level

Subscale 5 7 10 College

Perception .42' .53 .62 .79

(.32)' (.32) (.32) (.22)

Memory .45 .55 .70 .84

(.29) (.25) (.28) (.23)

Understanding .44 .36 .48 .30

(.25) (.29) (.20) (.23)

Evaluation .26 .42 .63 .75

(.25) (.27) (.34) (.26)

Metacognition .29 .36 .65 .64

(.31) (.33) (.25) (.29)

Planning .74 .75 .85 .78

(.22) (.25) (.25) (.25)

Low-Level .44 .54 .66 .82

(.25) (.22) (.21) (.17)

High-Level .25 .39 .63 .69

(.21) (.21) (.27) (.22)

Think .44 .52 .68 .84

(.23) (.21) (.27) (.16)

Know .31 .40 .64 .67

(.18) (.24) (.21) (.21)

High-Frequency .46 .58 .69 .79

(.23) (.19) (.23) (.15)

Low-Frequency .29 .35 .63 .72

(.18) (.22) (.26) (.23)

Cog. Word Total .37 .47 .66 .75

(.17) (.18) (.21) (.15)

Note. 'mean; 'standard deviation



Percent Correct

0.8

0.6

0.4

0.2

Fifth Seventh Tenth

Grade

College

IN Low Level Think

FIN Iigh Level Think.

Low Level Know

Li I high Level Know

Figure 1. Three-way interaction of Think or Know, Level of Meaning, and Grade. Note: Low or HighLevel (level of meaning), Think (cognate of think), Know (cognate of know).

No predictions were made regarding cogni-tive word subscale interactions. There werethree significant two-way interactions (but nosignificant two-way interactions includingGrade), and three significant three-way inter-actions, all of which were the significant two-way interactions with the addition of the Gradevariable. Post hoc t-tests (p < .05) were com-puted in order to compare the different cells inthe two-way interactions, while post hoc one-way ANOVAs with Scheffe range comparison(p<.05) were computed between and withingrades for each of the three-way interactions.

The significant Think/Know x Level ofMeaning interaction (F(3,156) = 7.51, p < .01)revealed that for think, low level of meaning(M = .76) was significantly greater in percent-correct than high level of meaning (M = .62:p< .001), while for know, high (M = .56) andlow (M = .51) level of meaning did not signi-ficantly differ in percent-correct (p> .05). Inother words, think or know did not significantlyaffect percent-correct for high level of meaning(p> .15), but for low level of meaning, knowsignificantly lowered percent-correct as com-pared to think (p< .001).

NATIONAL READING RESEARCH CENTER. READING RESEARCH REPORT NO. 14

24


Percent Correct

(1.8

Fifth Seventh Tenth

Grade

College

NI I Iigh Freq. Think

El Low Freq. Think

Iigh Freq. Know

L..] Low Freq. Know

Figure 2. Three-way interaction of Think or Know, Frequency, and Grade. Note: High or Low Freq.(frequency in written language), Think (cognate of think), Know (cognate of know).

For the Think/Know x Level of Meaning xGrade interaction (F(23,156) = 10.03, p<.001; see Figure 1), between grades analysesrevealed: fifth- and seventh-graders scoredsignificantly lower than undergraduates on lowlevel of meaning think and know and high levelof meaning think: and fifth- and seventh-grad-ers scored significantly lower than tenth-grad-ers and undergraduates on high level of mean-ing know (p< .05). In sum, there was a devel-opmental trend of increasing cognitive wordknowledge with a period of accelerated cogni-tive word acquisition after seventh-grade and

before the undergraduate level. Within gradeanalyses indicated: (a) for fifth-graders, lowlevel of meaning think was answered correctlysignificantly more often than the other three;(h) for seventh-graders, low level of meaningthink was answered correctly significantlymore than high level of meaning think: (c) fortenth-graders, no significant differences werefound; and (d) for undergraduates, low levelof meaning know was answered incorrectly sig-nificantly more than the other three (p < .05).In sum, low ievel of meaning think appeared tohe easier for the fifth- and seventh-graders, and



Percent Correct

Fifth Seventh Tenth

Grade

College

ai High Freq.

PM Low Freq. !Am Level

/High Freq. High Level

[ I Low Freq. high Level

Figure 3. Three-way interaction of Frequency, Level of Meaning, and Grade. Note: High or LowFrequency (frequency in written language), Low or High Level (level of meaning).

high level of meaning know appeared to beharder for undergraduates, while all combina-tions appeared to be equal in difficulty fortenth-graders (see Figure 1).

The Think/Know x Frequency interaction(F(3,156) = 13.74, p< .001) revealed that athigh frequencies, think (M = .74) was signifi-cantly greater in percent-correct than know (M= .56: p < .001), while at low frequenciesthink (M = .56) and know (M = .52) did notsignificantly differ in percent-correct (p> .60).In other words, different frequencies did notsignificantly affect percent-correct for know(p> .90), but for think, low-frequency signifi-

cantly lowered percent-correct as compared tohigh-frequency (p < .001).

For the Think/Knoiv x Frequency x Gradeinteraction (F(23,156) = 7.71, p < .001, seeFigure 2), between grades analyses revealed:fifth- and seventh-graders scored significantlylower than the tenth-graders and undergradu-ates on low-frequency think and know, andhigh-frequency know; and fifth- and seventh-graders scored significantly lower than under-graduates on high frequency think (p < .05). Insum, there was a developmental trend of in-creasing cognitive word knowledge with aperiod of accelerated cognitive word acquisi-


26


tion after seventh and before tenth grade.Within grade analyses indicated: (a) for fifth-and seventh-graders, high-frequency think wasanswered correctly significantly more than theothers; (b) for tenth gravers, no significantdifferences were found; and (c) for undergrad-uates, high- and low-frequency think wereanswered correctly significantly more thanhigh- and low-frequency know (p < .05). Insum, high-frequency think appeared to beeasier for fifth-graders, seventh-graders, andundergraduates, but not for tenth-graders.

The Frequency x Level of Meaning inter-action (F(3,156) = 37.29, p< .001) revealedthat at high frequencies, high (M = .64) andlow (M = .64) levels of meaning did notsignificantly differ in percent-correct (p> .80),while at low-frequency, high level of meaning(M = .40) was significantly lower in percent-correct than low level of meaning (M = .68;p< .001). In other words, whether the wordwas high- or low-frequency did not significant-ly affect percent-correct for low level of mean-ing (p> .25), but for high level of meaning,low-frequency cognitive words significantlylowered percent-correct as compared to high-frequency cognitive words (p < .001).

For the Frequency x Level of Meaning xGrade interaction (F(23,156) = 3.60, p< .05;see Figure 3), between grades analyses re-vealed a complex pattern of grade differences.There was a developmental trend of increasingcognitive word knowledge with a period ofaccelerated cognitive word acquisition afterseventh and before tenth grade. Within gradeanalyses indicated that for fifth-graders, sev-enth-graders, arid undergraduates, low level ofmeaning low-frequency cognitive words were

answered incorrectly significantly more thanthe others, while for tenth-graders, there wereno significant differences (p < .05). In sum, allgrades, except tenth, appeared to have diffi-culty with high level of meaning low-frequencycognitive words.

We predicted that cognitive word knowl-edge would significantly increase with age andthat all levels of meaning would be signifi-cantly different with the lower levels of mean-ing being acquired earlier than the higher levelsof meaning. Since the limited number of pas-sages in the cognitive word task prevented theanalysis of all four levels of meaning withfrequency and think/know, a 4 (Grade) x 4(Level of Meaning) ANOVA was computed;the data were collapsed across frequency andaccording to whether the word was think orknow. This analysis revealed significant maineffects for grade, F(3, 156) = 71.71, p< .0C. 1,and level of meaning, F(3, 156) = 12.20,p< .001. There was no significant interaction.

Post hoc one-way ANOVAs with Schefferange comparison (p < .05) were computed.Between level of meaning analyses indicatedthat perception (M = .63) was significantlygreater in percent-correct than metacognition(M = .49), and memory (M = .68) was signifi-cantly greater in percent-correct than evalua-tion (M = .55) or metacognition (M = .49).Between grade analyses indicated that fifth- (M= .37) and seventh-graders (M = .45) scoredsignificantly lower in percent-correct thantenth-graders (M = .66) and undergraduates(M = .75). Then, post hoc one-way ANOVAswith Scheffe range comparison (p < .05) werecalculated between grades for each level ofmeaning. For perception, memory, and evalua-



tion fifth- and seventh-graders scored signifi-cantly lower than undergraduates, and formemory and evaluation, fifth-graders scoredsignificantly lower than tenth-graders. Con-trary to expectations, seventh-graders scoredsignificantly lower for metacognition than fifth-graders and undergraduates. Our two predic-tions that cognitive word acquisition wouldsignificantly increase with age and that alllevels of meaning would significantly differ,with the lower levels of meaning being ac-quired earlier than the higher levels of mean-ing, were partially supported.

SUMMARY OF RESULTS

Cognitive word knowledge increased with age;low level of meaning cognitive words were ac-quired before high level of meaning cognitivewords; high-frequency cognitive words wereacquired earlier than low-frequency cognitivewords, and cognates of think were acquired be-fore cognates of know. However, the order ofacquisition of cognitive words depended on acomplex interaction between frequency of thecognitive word in established word frequencycounts, the level of meaning as determined bythe Frank and Hall (1991) conceptual difficultyhierarchy, and whether the cognitive word wasa cognate of think or know. More specifically,high-frequency think and low level of meaningthink were acquired earlier, and high level ofmeaning low-frequency cognitive words wereacquired later. Furthermore, cognitive wordswere positively correlated with achievementscores, and the correlations were higher forVerbal than Quantitative and higher for Vocab-ulary than Reading Comprehension.

DISCUSSION

Several issues ensue from the findings reportedhere concerning the relationship betweencognitive word knowledge and reading com-prehension. The first has to do with the validityof the cognitive word assessment. The secondturns on the nature of the development of thechild's understanding of cognitive words. Thethird issue has to do with the relationshipbetween reading comprehension and cognitiveword knowledge. We discuss these issues inturn.

Experimental Design Issues

The design of our study was important to theexpansion of our knowledge of the relationshipbetween cognitive words and reading compre-hension for several reasons. First, most re-search on cognitive word comprehension hasemployed children under eight years of age(except Astington & Olson, 1990); therefore,we know little about cognitive word knowledgein older children and adults. Second, despitethe voluminous research on internal statewords, only two studies have assessed knowl-edge of more than three cognitive words at onetime (Astington & Olson, 1990; Olson &Torrance, 1986). It is essential to assess a widevariety of cognitive words to uncover theimpact of the mental state lexicon on cognitivedevelopment. Third, only three studies haverelated cognitive words to reading comprehen-sion empirically (Olson & Torrance, 1986,1987; Astington & Olson, 1990). Fourth, weused a questionnaire to assess cognitive words,so we were not dependent on frequency of ex-


2cS


pression as the measure of conceptual diffi-culty. Frequency of expression is susceptible tomisinterpretations (see Frank & Hall, 1991).Our comprehension task gives a more accuratemeasure of a child's cognitive word knowledgebecause young children can often understand aword before they can produce it (Clark &Hecht, 1982; Benedict, 1978; Goldin-Meadow,Seligman, & Gelman, 1976) and observationalstudies of production may unfairly attributemore or less proficiency to children than de-served (Clark, 1983).

Finally, the use of abstract words to devel-op theories of word meaning (e.g., Frank &Hall, 1991; Hall et al., 1987) is unique, asmost experimental investigations of wordmeaning have employed concrete nouns to testpsychological theories of word meaning (for anexception see Miller & Johnson-Laird, 1976).Yet, many theories do not generalize easily toother types of words, such as abstract nouns,verbs, adjectives, and function words (Gam-ham, 1985). "The meaning of different wordsmay be of different types: a single analysis ofconcepts may well not suffice" (Carey, 1982,p. 360). Therefore, the semantically complicat-ed lexicon of cognitive words may itself re-quire an equally intricate theory of acquisition(Scholnick, 1987).

Cognitive Word Acquisition: ConceptualDifficulty and Prototypicality

As reviewed earlier, Frank and Hall (1991)developed a level of meaning hierarchy ofincreasing abstractness and conceptual difficul-ty, ranging from perception through memory,understanding, evaluation, and metacognition

to planning. This hierarchy adequately des-cribed both adult and child verbal frequency ofthe cognitive word know (Hall et al., 1984); asthe level of meaning increased, the frequencydecreased. Furthermore, the levels of percep-tion, memory, and understanding were statisti-cally differentiated, but the levels of evalua-tion, metacognition, and planning were not.Frank and Hall (1991) then tried to incorporatethe cognitive word think into their cognitiveword hierarchy. They found that evaluationwas used most often and that the other fivelevels of meaning were statistically undifferen-tiated for both children and adults.

The present study provided some supportfor the Frank and Hall (1991) level of meaninghierarchy, as all of the level of meaning differ-ences were in the hypothesized direction. Morespecifically, perception was significantly great-er in percent-correct than metacognition, andmemory was significantly greater in percent-correct than evaluation and metacognition.Furthermore, when perception and memorywere combined into a low level of meaningcategory and evaluation and metacognitionwere combined into a high level of meaningcategory to allow for sufficient reliability, wefound that low level of meaning was answeredcorrectly significantly more than high level ofmeaning. A more reliable measure of cognitivewords would undoubtedly differentiate statisti-cally between all levels.

The present study also provided support forthe developmental differences in level of mean-ing acquisition found among children (Hughes,1985), and between children and adults (Frank& Hall, 1991; Hall et al., 1987). In general,cognitive word knowledge increased with age,


29


and developmental differences were observedin all cognitive word subscales. For example,younger children scored significantly lower inpercent-correct than older children in low levelof meaning, high level of meaning, think,know, high-frequency, low-frequency, andcognitive word total. More specifically, forperception, memory, and evaluation, fifth- andseventh-graders scored significantly lower thanundergraduates, and for memory and evalua-tion, fifth-graders scored significantly lowerthan tenth-graders. For nietacognition, seventh-graders scored significantly lower than under-graduates. Contrary to expectations, for meta-cognition, seventh-graders scored significantlylower than fifth-graders, probably because thefifth-grade girls were enrolled in a study skillsdevelopment class, thereby enhancing theirknowledge of metacognitive words.

Frank and Hall (1991) proposed that cogni-tive words are hierarchically organized, butthat different cognitive words may have adifferent organization of levels that depends onboth conceptual difficulty and prototypicality.They further hypothesized that prototypicalmeanings will be acquired first, but cognitivewords whose prototypical meanings are lowerlevel will he semantically mastered earlier thanwords whose prototypical meanings are higherlevel. This hypothesis was supported by thefact that children in the Hall et al. (1984) studyacquired know before think. Know's prototyp-ical level of meaning may be at the perceptionlevel while think's prototypical level of mean-ing is at the evaluation level.

In contrast, our results indicate that thinkwas answered correctly significantly more

often than know. But a linguistic analysis ofthink and know may provide the answer. Verysimply, it is easier to know what anotherperson thinks rather than what another personknows, given veridical self-report of that per-son. According to Webster's UnabridgedDictionary of the English Language (1989),know means "to perceive with certainty; tounderstand clearly; to be sure of or well-in-formed about; as we know the facts . . . to

have a firm mental grasp of . . to have clearand certain perception," while think means "toform or have in the mind . . . to bring theintellectual facilities into play; to use the mindfor arriving at conclusions, making decisions,drawing inferences etc.; to perform any mentaloperation; to reason." From reflection on thepreceding definitions, it is clear that to knowmeans to have a definite, veridical understand-ing of something, whereas to think merelymeans to reflect about something, probablywith an indefinite, uncertain understanding. Inshort, knowing what another person thinks ismuch easier than knowing what another personknows. Our discrepancy with Frank and Hall(1991) may also be accounted for, in part, bythe fact that our study measured comprehensionwhile theirs measured word production. How-ever, Hughes (1985) found the correlationbetween production and comprehension ofcognitive words to be rather high (r = .66).

The two-way Think/Know x Level ofMeaning and three-way Think/Know x Levelof Meaning x Grade (see Figure 1) interac-tions also illuminate the Frank and Hall (1991)conceptual difficulty and prototypicality hy-pothesis. Closer analysis of Figure 1 reveals

NATIONAL READING RESEARCH CENTER. READING RESEARCH REPORT NO. 14


that the two-way interaction appears to holdonly for fifth- and seventh-graders; that is, lowlevel of meaning think is easier than high levelof meaning think, and high and low level ofmeaning know. In contrast, there appear to beno significant differences in tenth-graders, andonly the two main effects exist in undergradu-ates. Again, Frank and Hall (1991) suggest thatthink has a prototypical meaning at a higherlevel of meaning than know and that is why itis learned later. Our only significant differ-ences were that children learned the low levelof meaning think earlier than high level ofmeaning think, low level of meaning know, andhigh level of meaning know. According to theFrank and Hall (1991) hypothesis, this indi-cates, at least for the fifth- and seventh-grad-ers, that the prototypical level of meaning ofthink may be at the lower and not the higherlevel of meaning, and that know may not havea prototypical level of meaning. Further re-search must address this question.

The two-way Think/Know x Frequency andthree-way Think/Know x Frequency x Gradeinteraction (see Figure 2) revealed high-fre-quency think to be easier than low-frequencythink, low-frequency know, and high-frequencyknow for fifth-graders and seventh-graders.There appeared to be no significant differencesin tenth grade, and only the main effects offrequency and level of meaning appeared toexist in undergraduates. The interactions sug-gest that frequency of exposure to cognitivewords, especially to cognates of think, is veryimportant in cogniti' word acquisition, sug-gesting that linguistic input is essential forenhancing lexical development in this domain.This finding supports previous research that re-

ports a high correlation between adult and childuse of cognitive words (Hall et al., 1987;

Beeghly et al., 1986).The Frequency x Level of Meaning inter-

action and the Frequency x Level of Meaningx Grade interaction (see Figure 3) revealedthat all grades except tenth appeared to havedifficulty with high level of meaning low-frequency as compared to low level of meaninglow-frequency cognitive words, and high andlow level of meaning high-frequency cognitivewords. In other words, the probability ofgetting a certain level of meaning correctdepended on the frequency of the word. If thecognitive word was low-frequency, low levelof meaning was answered correctly more thanhigh level of meaning, but if the cognitiveword was high frequency, children were equal-ly likely to answer it correctly regardless of itslevel of meaning. In sum, frequency appearedto be very important in cognitive word acquisi-tion (high-frequency even overcame the levelof meaning effect) suggesting that linguisticinput is essential for efficient cognitive wordacquisition.

The important question then became wheth-er there was a Think/Know x Level of Mean-ing x Frequency interaction. The 4 (Grade) x2 (Frequency) x 2 (Think/Know) x 2 (Levelof Meaning) ANOVA revealed a nonsignificantinteraction. Even if this interaction were signif-icant, it would be meaningless; two cognitiveword passages per cell does not allow forsufficient reliability. Nevertheless, from thepreviously described three two-way interactionsit appears that the high-frequency and low levelof meaning think are relatively easy at least forfifth- and seventh - graders, while low-frequency

NATIONAL READING RESEARCH CENTER, READING RESEARCH REPOR'F NO. 14

31


high level of meaning appears to be difficultfor all subjects. On the basis of these findings,we predict that future studies may find thatfrequency and level of meaning will have littleeffect on percent-correct for know, while onlyhigh-frequency combined with low level ofmeaning will increase percent-correct for think.

Frequency of exposure to cognitive wordscould account for the above interactions. Morespecifically, children may be exposed to lowlevel of meaning think more often than lowlevel of meaning know, but they may receivemore equal exposure to high level of meaningthink and high level of meaning know. Thiswould account for the Level of Meaning xThink/Know interaction (see Figure I ). In addi-tion, children may be exposed to high-frequen-cy think more often than high-frequency know,but they may receive more equal exposure tolow-frequency think and know. This would ac-count for the Frequency x Think/ Know inter-action (see Figure 2). These predictions will betested empirically in future studies.

In summary, all three-way interactions re-vealed a developmental trend of increasingcognitive word knowledge, and a period ofaccelerated cognitive word acquisition afterseventh and before tenth grade. YoUngerchildren have difficulty in mastering the cogni-tive word lexicon for many reasons. First,children's exposure to written arv' spokencognitive words is seriously deficient (Asting-ton, 1991; Hall et al., 1984; Carroll et al.,1971; Smith & Meux, 1970; Thorndike &Lorge, 1944), yet adult/child correlations ofcognitive word use are very high (Hall et al.,1987; Beeghly et al., 1986). Second, the cogni-tive word lexicon is very intricate and many

words within it are polysemous, thus the ap-propriate use of cognitive words requires arecognition of the subtle contrasts and com-parisons between them (Hall et al., 1987).

Third, cognitive words are abstract and elusivebecause they deal with the inner workings ofthe mind; therefore, acquisition cannot benefiteasily from nonlinguistic strategies such asmapping language onto pre-existing conceptualcategories (Clark, 1983). Given that a periodof accelerated cognitive word acquisition existsand that linguistic input may be essential forthis rapid period of growth to occur, children'soral environments and reading materials shouldbe enhanced by higher frequencies of cognitivewords if their intellectual development is to beaccelerated.

In support of this hypothesis, Corson (1985)observed that most cognitive words are usuallyencountered late and learned later in life.

Access to them and use of them are furtherrestricted by their introduction in literature ortextbooks, not through dialogue or discourse.Corson (1985) calculated the frequency ofcognitive words and speech act verbs in 43types of text and found their frequency to varysubstantially, for example, from 40% in philos-ophy of education to 0% in five- to six-yearreading age children's fiction (see Table 6).Thus, the period of accelerated developmentbetween the seventh grade (M = 12.7) andtenth grade (M = 15.6) children in our studymay be a product of increased exposure tocognitive words in high school texts and otherreading material. This increased exposure tocognitive words may determine the conver-gence of cognitive word knowledge in tenthgrade (see three-way interactions, Figures 1, 2,



Table 6. LiteratureGraeco-Latin Content

Ranked According to

Literature % Content G-L

Philosophy of Education 40Linguistic Philosophy 40History of Religion 37

Theology 36Sociology 38Science Education 33

Psychology 32Physics 31

Music 29Mathematics 28-29English Literature 16-24Newspaper 5-24Children's Fiction:

12 years R.A. 10

9-11 years R.A. 7

10-11 years R.A 4

9-12 years R.A. 4

7-8 years R.A. 3

7-11 years R.A. 2

5-6 years R.A. 0

Note. Based on random passages of 100 words(Corson, 1985). R.A. = Reading Age.

& 3). The lack of convergence in the undergraduates may be due to the diverse academicability levels at the University of Maryland ascompared to the homogeneous population ofstudents in the private high schools.

Cognitive Words and ReadingComprehension

Children's ability to assign meanings to wordsis a developmental process involving both cog-nitive and linguistic skills in interaction withcontextual constraints. The current educationalliterature on reading shows conclusively that

the development of word knowledge in chil-dren is essential for high-level text understand-ing (e.g., Stanovich & Cunningham, 1992;Stahl, Hare, Sinatra, & Gregory, 1991; El-dredge, Quinn, & Butterfield, 1990; Stahl,Jacobson, Davis, & Davis, 1989; Dixon, Le-Fevre, & Twilley, 1988). Furthermore, severalsuccessful ,direct vocabulary instruction pro-grams have shown beneficial effects on chil-dren's high-level comprehension of text (e.g..McDaniel & Pressley, 1989; Reutzel & Hol-lingsworth, 1988). Unfortunately, the role ofthe domain of "states" the temporary orpermanent properties of objects and situations

in high-level text understanding has receivedlittle attention, compared to the domain ofobjects, situations. or events (see Clark, 1983).

Only three investigations of the relationshipof cognitive word knowledge to reading com-prehension have been conducted. Our studysupports the importance of cognitive wordknowledge for skilled reading comprehension.For all grade-schoolers and undergraduates,adjusted for mean differences, the correlationof cognitive word total was higher for Verbal(r = .42, p< .001) than for Quantitative (r =.28, p < .001) and higher for Vocabulary (r =.47, p < .001) than for Reading Comprehension(r = .39, p < .001). To a greater extent thanknowledge of other vocabulary words, knowl-edge of cognitive words seems to be centrallyinvolved in reading comprehension. For exam-ple, differentiating between the subtle differ-ences in mental and motivational states of theprotagonist or his or her adversary is essentialfor the interpretation of a literary text. Cogni-tive words provide the reader a vehicle bywhich to analyze the past, present, and future


3


actions and goals of a character with precisioneven when presented with ambiguous informa-tion. Similarly, Corson (1985) describes cogni-tive words as allowing people "to order thoughtwhere such ordering of thought might notoccur without the words themselves" (p. 61)and to increase their "capacity to deal withexpectations and hypotheticals" (p. 48).

Future research should consist of studies inwhich children's cognitive word knowledge isenhanced, and subsequent observation is takenof his or her high-level text understanding andmetalinguistic acquisition in a variety of learn-ing environments. The rationale of this asser-tion is supported in part by Paul and O'-Rourke's (1988) suggestion that teachers mustbe more aware of these multimeaning words,and perhaps include direct vocabulary instruc-tion in their programs (see Paul & O'Rourke,1988). We make several predictions regardingthe role of cognitive word knowledge in devel-opment of high-level text understanding andmetalinguistic knowledge. We suggest thatchildren with high cognitive word knowledge,as compared to those with low cognitive wordknowledge, may be more likely to extract,process, and recall in an organized way themetalinguistic information they encounter intexts (see McDaniel & Einstein, 1989; Alex-ander & Judy, 1988; Recht & Leslie, 1988).Because this metalinguistic information is moreeffectively gleaned from reading, children maylearn reading strategies faster and earlier.Furthermore, children may be more likely tomaintain and generalize these reading strategiesbecause cognitive words allow children todescribe under what conditions metalinguistic

activities are successful, why they are success-ful, and how they may be used in solvingdifferent problems. These issues await furtherresearch.

Author Notes. This research was supported bygram #01-4-32735 from the Office of EducationalResearch and Improvement, U.S. Department ofEducation, to William S. Hall and James R. Boothof the National Reading Research Center. We

thank David R. Olson, Linda Baker, and twoanonymous reviewers for very helpful critical

comments on earlier versions of this manuscript.We also thank the administrators at Landon Schooland Holton-Arms School for their organizationalassistance, and their students for volunteering toparticipate in this investigation.

REFERENCES

Abbeduto, A., & Rosenberg, S. (1985). Children'sknowledge of the presuppositions of know andother cognitive verbs. Journal of Child Lan-guage, 12(3), 621-641.

Alexander, P.A., & Judy, J.E. (1988). The interac-tion of domain-specific knowledge and strategicknowledge in academic performance. Review ofEducational Research, 58(4), 375-404.

Anglin, J. M. (1986). Semantic and conceptualknowledge underlying the chi I's words. In S.A. Kuczaj & M. D. Barrett (Eds.), The devel-opment of word meaning: Progress in cognitivedevelopment research (pp. 83-97). New York:Springer-Verlag.

Astington, J. W. (1991). Narrative and the child'stheory of mind. In B. Britton & A. Pellegrini(Eds.), Narrative thought and narrative lan-guage. Hillsdale, NJ: Erlbaum.


3 (1"


Astington, J. W., & Olson, D. R. (1990). Meta-cognitive and metalinguistic language: Learningto talk about thought. Applied Psychology: AnInternational Review, 39(1), 77-87.

Beeghly, M., Bretherton, I., & Mervis, C. B.(1986). Mothers' internal state language totoddlers. British Journal of DevelopmentalPsychology, 4, 247-260.

Benedict, H. (1978). Language comprehension in9-15 month old children. In R. N. Campbell &P. T. Smith (Eds.), Recent advances in the psy-chology of language: Language developmentand mother-child interaction. New York:Plenum.

Bretherton, I., & Beeghly, M. (1982). Talkingabout internal states: The acquisition of anexplicit theory of mind. Developmental Psy-chology, 18, 906-921.

Carey, S. (1982). Semantic development: The stateof the art. In E. Wanner & L. Gleitman (Eds.),Language acquisition: The state of the art (pp.347-389). Cambridge: Cambridge UniversityPress.

Carroll, J. B., Davies, B., Richman, P. (1971). TheAmerican heritage word frequency book. NewYork: American Heritage.

Clark, E. V. (1983). Meanings and concepts. In J.H. Flavell & E. M. Markman (Eds.), Thehandbook of child psychology (pp. 787-840).New York: John Wiley & Sons.

Clark, E. V. (1980). Here's the top: Nonlinguisticstrategies in the acquisition of orientationalterms. Child Development, 51(2), 329-338.

Clark, E. V. (1973). Non-linguistic strategies andthe acquisition of word meanings. Cognition,2(2), 161-182.

Clark, E. V., & Hecht, B. F. (1982). Learning tocoin agent and instrument nouns. Cognition,12(1), 1-24.

Corson, D. (1985). The lexical bar. Oxford:Pergamon Press.

Dixon, P., LeFevre, J., & Twilley, L. C. (1988).Word knowledge and working memory aspredictors of reading skill. Journal of Educa-tional Psychology, 80(4), 465-472.

Eldredge, J. L., Quinn, B., & Butterfield, D. D.(1990). Causal relationships between phonics,reading comprehension, and vocabulary achiev-ement in the second grade. Journal of Educa-tional Research, 83(4), 201-214.

Frank, R. E., & Hall, W. S. (1991). Polysemy andthe acquisition of cognitive internal state lexi-con. Journal of Psycholinguistic Research,20(4), 283-304.

Garnham, A. (1985). Psycholinguistics. New York:Methuen.

Goldin-Meadow, S., Seligman, M. E. P., & Gel-man, R. (1976). Language in the two-year-old.Cognition, 4, 189-202.

Hall, W. S., & Nagy, W. E. (1987). The seman-tic-pragmatic distinction in internal state words:The role of the situation. Discourse Processes,10(2), 169-180.

Hall, W. S., & Nagy, W. E. (1986). Theoreticalissues in the investigation of the words ofuernal report. In I. Gopnik & M. Gopnik

(Eds.), From models to modules: Studies incognitive science from the McGill workshops(pp. 26-65). Norwood, NJ: Ablex.

Hall, W. S., Nagy, W. E., & Linn, R. (1984).Spoken words: Effects of situation and socialgroup on oral word usage and frequency.Hillsdale, NJ: Erlbaum.

Hall, W. S., Scholnick, E. K., & Hughes, A. T.(1987). Contextual constraints on usage ofcognitive words. Journal of PsycholinguisticResearch, 16, 289-310.



Hughes, A. T. (1985). Comprehension and pro-duction of cognitive words. Unpublished mas-ter's thesis, University of Maryland, CollegePark.

Huttenlocher, J., & Lui, F. (1979). The semanticorganization of some simple nouns and verbs.Journal of Verbal Learning and Behavior, 18,141-162.

Johnson, C. N., & Maratsos, M. P. (1977). Earlycomprehension of mental verbs: Think andknow. Child Development, 48, 1743-1747.

Johnson C. N., & Wellman, H. M. (1980). Child-ren's developing understanding of mentalverbs: Remember, know, and guess. ChildDevelopment, 51(4), 1095-1102.

Macnamara, J., Baker, E., & Olson, C. L. (1976).Four-year-olds' understanding of pretend,forget, and know: Evidence for propositionaloperations. Child Development, 47, 62-70.

McDaniel, M. A., & Einstein, G. 0. (1989).Material appropriate processing: A context-ualist approach to reading and studying strate-gies. Educational Psychology Review, 1(2),

113-145.

McDaniel, M. A., & Pressley, M. (1989). Key-word and context instruction of new vocabularymeanings: Effects on text comprehension andmemory. Journal of Educational Psychology,81(2), 204-213.

Miscione, J. L., Marvin, R. S., O'Brien, R. G., &Greenberg, M. T. (1978). A developmentalstudy of preschool children's understanding ofthe words "know" and "guess." Child Devel-opment, 49, 1107-1113.

Miller, G. A., & Johnson-Laird, P. N. (1976).Language and perception. Cambridge, MA:Harvard University Press.

Moore, C., Bryant, D., & Furrow, D. (1989).Mental terms and the development of certainty.Child Development, 60(1), 167-171.

Nunnally, J. C. (1978). Psychometric theory. NewYork: McGraw-Hill.

Olson, D. R., & Astington, J. W. (1990). Talkingabout text: How literacy contributes to thought.Journal of Pragmatics, 14, 557-573.

Olson, D. R., & Astington, J. W. (1986). Chil-dren's acquisition of metalinguistic and meta-cognitive verbs. In W. Demopoulos & A. Mar-ras (Eds.), Language learning and conceptacquisition: Foundational issues (pp. 18-1-

199). Norwood, NJ: Ablex.Olson, D. R., & Hildyard (1981). On the mental

representation of oblique orientation. CanadianJournal of Psychology, 31(1), 3-13.

Olson, D. R., & Torrance, N. G. (1987). Lan-guage, literacy, and mental states. DiscourseProcesses, 10, 157-168.

Olson, D. R., & Torrance, N. G. (1986). Somerelations between children's knowledge ofmetalinguistic and metacognitive verbs andtheir linguistic competencies. In I. Gopnik &M. Gopnik (Eds.), From models to modules:Studies in cognitive science from the McGillworkshops (pp. 66-81). Norwood, NJ: Ablex.

Paul, P. V., & O'Rourke, J. P. (1988). Multimean-ing words and reading comprehension: Implica-tions for special education students. Remedialand Special Education, 9(3), 42-52.

Recht, D. R., & Leslie, L. (1988). Effect of priorknowledge on good and poor readers' memoryof text. Journal of Educational Psychology. 80,16-20.

Reutzel, D. R., & Hollingsworth, P. M. (1988).Highlighting key vocabulary: A generative-reciprocal procedure for teaching selected



inference types. Reading Research Quarterly,23(3), 358-378.

Que Software. (1990). Right Writer, Version 3.1.New York: Macmillan Computer Publishing.

Robinson, E. (1980). Mother-child interaction andthe child's understanding about communication.International Journal of Psycholinguistics, 7(1& 2), 85-101.

Robinson, E., Goelman, G., & Olson, D. R. (19-83). Children's understanding of the relationbetween expressions (what is said) and inten-tions (what is meant). British Journal of Devel-opmental Psychology, 1, 75-86.

Scholnick, E. K. (1987). The language of mind:Statements about mental states. DiscourseProcesses, 10, 181-192.

Scholnick, E. K., & Hall, W. S. (1991). Thelanguage of thinking: Metacognitive and condi-tional words. In J. P. Byrnes & S. A. Gelman(Eds.), Perspectives on language and cogni-tion.. Interrelations in development. New York:Cambridge University Press.

Shatz, M., Wellman, H.M., & Silber, S. (1983).The acquisition of mental verbs: A systematicinvestigation of the child's first reference tomental state. Cognition, 14(3), 301-321.

Smith, 0. B., & Meux, M. 0. (1970). A study ofthe logic of teaching. Urbana, IL: University ofIllinois Press.

Stahl, S. A., Hare, V. C., Sinatra, R., & Gregory.J. F. (1991). Defining the role of prior knowl-edge and vocabulary in reading comprehension:The retiring of number 41. Journal of ReadingBehavior, 23(4), 487-508.

Stahl, S. A., Jacobson, M. G., Davis, C. E., &Davis, R. L. (1989). Prior knowledge anddifficult vocabulary in the comprehension ofunfamiliar text. Reading Research Quarterly,24(1), 27-43.

Stanovich, K. E., & Cunningham, A. E. (1992).Studying the consequences of literacy within aliterate society: The cognitive correlates 01print exposure. Memory and Cognition, 20(1),51-68.

Steiger, J. H. (1980). Tests for comparing elementsof a correlation matrix. Psychological Bulletin,87, 245-251.

Thorndike, E. L., & Lorge, I. (1944). Teacher'swordbook of 30,000 words. Teachers College,Columbia University, NY: Bureau of Publi-cations.

Wellman, H. M., & Estes, D. (1987). Children'searly use of mental verbs and what they mean.Discourse Processes, 10, 141-151.

Wellman, H. M., &Johnson, C. N. (1979). Under-standing mental processes: A developmentalstudy of "remember" and "forget." ChildDevelopment, 50, 79-88.

APPENDIX A

COGNITIVE WORD TASK

Instructions

After the cognitive word test booklet was handedout, the experimenter read the instructions. Theinstructions, which were on the first page of thebooklet, read:

The words think and know are very commonlyused in our language and they have manymeanings. The meaning of think and knowdepends on the context that they are used in.Other words can describe a particular meaningof think and know. Sometimes another wordcan more accurately or narrowly define what ismeant by think or know.


BEST COPY AVAILABLE


For example, the statement ( I ) "Mike knowsthe clock is on the wall in front of him" may bemore specifically stated by (2) "Mike sees theclock on the wall."

The purpose of this test is to see if you knowwhat other words can be used to describe whatis meant by think and know in a certain context.You will be asked to read a short story that tellsof a person who thinks or knows something.After the story, there are four possible sent-ences containing words that may describe whatis meant by think or know. Circle the letter ofthe sentence that contains the replacement wordthat goes along with the context of the story thebest. There are 24 stories. You have exactly 25minutes to complete this Cognitive Word Task.If you finish early, please go over your an-swers. Are there any questions?

Cognitive Word Passages

Perception: "Notice"

I. Jeanne just poured herself a bowl of cereal forbreakfast. She put the cereal box back into thecupboard. Her mom sees her eating cereal atthe kitchen table. Her mom asks, "I am goingto the food store to shop. How much cereal isleft? Do we need more cereal?" Jeanne says, "Iwas thinking of something else, when I waspouring my cereal."

A. Jeanne did not recognize the cereal wasgetting low.

B. Jeanne did not understand the cereal wasgetting low.

C. Jeanne did not notice the cereal was gettinglow.

D. Jeanne did not realize the cereal was get-ting low.

Perception: "Concentrate"

2. Suzanne is studying in her bedroom for a testtomorrow. The test is going to be very easy.Her brother Tim is playing loud music on hisstereo. Usually Suzanne can study with musicin the background, but she is not interested inwhat she is studying. Suzanne cannot thinkabout her homework.

A.

B.

C.

D.

Suzanne cannot recall the material.Suzanne cannot comprehend the material.Suzanne cannot apprehend the material.Suzanne cannot concentrate on the materi-al.

Perception: "Observe"

3. The other day, Jane was exploring in thewoods. A large bird flew right in front of herface while she was walking. She kept her eyesopen because she wanted to see this large birdup close. When talking to her friend Sally later,she told Sally about the bird. Sally denied thatJane saw a bird. Jane said, "I know a large birdflew in front of my face."

A. Jane observed the large bird.B. Jane recognized the large bird.C. Jane understood it was a large bird.D. Jane realized it was a large bird.

Perception: "Perceive"

4. Jeff just received a new dog and dog whistle forhis birthday. Whenever Jeff blows the whistle,his dog comes to him. The whistle is so high-


3S


pitched that Jeff cannot hear it. Dogs can hearhigh-pitch sounds that people cannot hear.Dogs have better ears than people. The dogknows when the whistle blows.

A. The dog comprehends when the whistleblows.

B. The dog recalls when the whistle blows.C. The dog apprehends when the whistle

blows.D. The dog perceives when the whistle blows.

Memory: "Forget"

5. Sally and Joni just finished watching a movie atthe movie theater. They were talking about themovie on their walk home. Sally asked Joni thename of the hero in the movie. During themovie, Joni thought the hero had a neat name.Now, she cannot think of his name.

A. Joni realizes the name of the hero in themovie.

B. Joni forgets the name of the hero in themovie.

C. Joni notices the name of the hero in themovie.

D. Joni understands the name of the hero inthe movie.

Memory: "Remind"

6. Adam and Tony are riding their bikes aroundthe neighborhood on a late Sunday afternoon.Adam sees a neighbor mowing his lawn. Adamhad told his parents that he was going to mowthe lawn on Sunday afternoon. Adam's parentsare gone for the weekend. Adam thinks he hasto mow the lawn now.

A. Adam perceives he has to mow the lawn.B. Adam comprehends he has to mow the

lawn.C. Adam was reminded he has to mow the

lawn.D. Adam apprehends he has to mow the lawn.

Memory: "Recognize"

7. Last week in class Mr. Smith reviewed thestates in America and their major products.Today, John has a multiple-choice test on thattopic. One of the questions asks: "What is themajor product made by Wisconsin?" Afterreading only the question, John cannot answerit. Lucky for John that there are four choices.After reading the choices, John knows that thechoice "major product is dairy goods" is thecorrect answer.

A. John recognized that their major product isdairy goods.

B. John realized that their major product isdairy goods.

C. John i(nderstood that their major product isdairy goods.

D. John observed that their major product isdairy goods.

Memory: "Recall"

8. Tammy recently made friends with another girlfrom school, Rebecca. Tammy has calledRebecca on the phone a few times. Tanury isat another friend's house. She wants to callRebecca so they can all play. She has nevercalled her before without looking up the num-ber in her address book. Tammy knows thenumber and calls Rebecca.



A. Tammy apprehended the number.B. Tammy perceived the number.C. Tammy comprehended the number.D. Tammy recalled the number.

Understand: "Realize"

9. Claudia is learning how to play a new game.She is trying to teach her friend James how toplay the game. Claudia has already read thedirections, but they are not playing the gameright. Claudia rereads the directions. They arenow playing the game correctly. Claudia thinks

she can play the game right.

A. Claudia concludes this is the way to playthe game.

B. Claudia recognizes this is the way to playthe game.

C. Claudia assumes this is the way to play thegame.

D. Claudia realizes this is tl _ way to play thegame.

Understand: "Apprehend"

10. Bill is putting together a model kit of a car. Hereads the next step in the directions. He gluesthe body together. Next, he is trying to put therear seat in the body, but it does not fit. Herereads the last line in the directions again. Itsays to glue the rear seat in before you glue thebody together. At first, Bill thought the direc-tions said to glue the seat in after you glue thebody together.

A. Bill did not apprehend the directions.B. Bill did not evaluate the directions.C. Bill did not interpret the directions.D. Bill did not recognize the directions.

Understand: "Understand"

11. Danielle has been working on her math home-work for over an hour. She was having troublewith one problem. Finally, she tried to use adifferent formula and it worked. The teachergraded her homework. She got 100%. Now,Danielle knows how to do the problem.

A. Now, Danielle recognizes this is the way todo the problem.

B. Now, Danielle understands this is the wayto do the problem.

C. Now, Danielle assumes this is the way todo the problem.

D. Now, Danielle concludes this is the way todo the problem.

Understand: "Comprehend"

12. Stanley is an expert on the ocean. Stanleyexplains to Geoff why there are tides in theocean. Later, Geoff tells his teacher about thecause of the tides. Geoff says, "The waterconies in and out from shore because the gravi-ty between the earth and moon changes." Histeacher is very impressed. Now, Geoff knows

why there are tides in the ocean.

A. Geoff recalls why there are tides in theocean.

B. Geoff interprets why there are tides in theocean.

C. Geoff infers why there are tides in theocean.

D. Geoff comprehends why there are tides inthe ocean.



Evaluation: "Doubt"

13. Cathy and Erica are visiting the zoo. They aretrying to identify the animals without looking atthe labels on their cages. They come to themonkey cages. Erica has labeled every monkeycorrectly so far. They come to next cage.Cathy labels this monkey a chimpanzee, butErica thinks the monkey is a great ape.

Evaluation: "Conclude"

15. Patrick comes home early from school andfinds that nobody is home. He looks in thegarage and sees that his mother's car is gone.He also sees that his brother Sam's footballequipment is not in the utility room where italways is. He knows that his mother is drivingSam to football practice.

A. Erica doubts the monkey is a chimpanzee. A. He understands that his mother is drivingB. Erica understands the monkey is a chim- Sam to football practice.

panzee. B. He considers that his mother is drivingC. Erica considers tl., inonkey is a chimpan- Sam to football practice.

zee. C. He concludes that his mother is drivingD. Erica imagines the monkey is a chimpan- Sam to football practice.

zee. D. He imagines that his mother is driving Samto football practice.

Evaluation: "Infer"

14. Chris leaves a candy bar on the kitchen counterby snistake. While Chris is outside, his motherfinds the candy bar and puts it in a drawer.When Chris comes home, his candy bar is noton the counter. On his way to his room, he seesa wrapper of the same kind of candy on hisbrother's floor. Chris thinks that his brother atehis candy bar.

A. Chris infers that his brother has eaten hiscandy bar.

B. Chris comprehends that his brother haseaten his candy bar.

C. Chris conceives that his brother has eatenhis candy bar.

D. Chris contemplates that his brother haseaten his candy bar.

Evaluation: "Hypothesize"

16. Kathleen walks outside for recess and meets upwith two classmates in the playground. One isvery big and the other is very skinny. They tellKathleen that they have just balanced them-selves on the seesaw on the other side of theplayground. Kathleen did not see them do it,but she knows they balanced themselves byputting themselves at different distances fromthe center of the seesaw.

A. Kathleen comprehends how they balancedthemselves.

B. Kathleen ponders how they balanced them-selves.

C. Kathleen contemplates how they balancedthemselves.

D. Kathleen hypothesizes how they balancedthemselves.


41


Metacognition: "Consider"

17. Cheryl is studying for a science test tomorrow.Her teacher said that the test is going to coverthe main ideas that she talked about in class.Cheryl is trying to decide whether to spendmost of her time on the parts of the textbookthe teacher covered in class or to study all ofthe details. Cheryl thinks about how to learn thematerial the best.

A. Cheryl concludes how to learn the materialthe best.

B. Cheryl considers how to learn the materialthe best.

C. Cheryl expects to learn the material thebest.

D. Cheryl intends to learn the material thebest.

Metacognition: "Contemplate"

18. Tom and Megan are married and often fightover small things. They just got in a yellingmatch over who has taken out the garbagemore. In the past they have divided the workaround the house evenly. Tom is upset. He isnot sure why they are always yelling at eachother. He thinks it may be because they havedifferent ideas about what it is to be a husbandand a wife.

Metacognition: "Reflect"

19. Georgia sees an animal at the zoo that she hasseen before. Her friend Cynthia asks her whatit is calleci. Georgia calls it a Cheetah. Cynthiasays, "No, it is a Hyena." Georgia knows whyshe called it a Cheetah and not a Hyena. Bothanimals have spots and their names sound thesame.

A. Georgia assumed this is the reason shecalled the animal a Cheetah.

B. Georgia bet this is the reason she called theanimal a Cheetah.

C. Georgia expected this is the reason shecalled the animal a Cheetah.

D. Georgia reflected on why she called theanimal a Cheetah.

Metacognition: "Analyze"

20. Joe and Fred went off by themselves to explorea lake. They came to a group of boys jumpingoff a very high tree into the water. Joe andFred decided to jump. When Fred was about tojump, he became afraid. The many times hejumped off the high dive at the school swim-ming pool popped up in his mind. He jumped.After the jump, Fred knew jumping off thediving board at school helped him to jump offthe tree.

A. Torn apprehends why they fight so often. A. Fred inferred how he jumped off the tree.B. Tom contemplates why they fight so often. B. Fred predicted how he jumped off the tree.C. Tom predicts why they fight so often. C. Fred anticipated how he jumped off theD. Tom anticipates why they fight so often. tree.

D. Fred analyzed how he jumped off the tree.


4'


Planning: "Predict"

21. Dave and Bruce want to go camping in themountains this weekend. They should not go ifit is going to rain. It rained all week. There isnot a cloud in the sky the Saturday morningthey want to leave. Dave listens to the weatherchannel a lot. Dave thinks that it will not rainthis weekend.

A.

B.

Dave imagines that it will notweekend.Dave considers whether itweekend.Dave predictsweekend.Dave infers that it will not rain this week-end.

rain this

will rain this

C. that it will not rain this

D.

Planning: "Intend"

22. Patty has a very busy week next week. She hasa math and science test at school. She has adoctor's and dentist's appointment. She is goingto Heidi's and Linda's birthday parties. Andshe has other things to do too! She studied forher tests during the weekend. She thinks shewill study more after the birthday parties.

A. Patty intends to study after the birthdayparties.

B. Patty conceives of studying after the birth-day parties.

C. Patty concludes to study after the birthdayparties.

D. Patty contemplates studying after the birth-day parties.

Planning: "Expect"

23. A local high school has a very good basketballteam. They are rated number one in theirdivision. They have only lost once during theseason. They lost to the second best team in theleague. The coach tells his team that they areplaying a team tonight that is rated poorly inthe league. The coach knows they are going towin the game tonight.

A. The coach imagines his team will wintonight.

B. The coach expects his team will win to-night.

C. The coach considers his team will wintonight.

D. The coach assumes his team will wintonight.

Planning: "Anticipate"

24. Mary and Alice always go to church togetheron Sunday morning. They always meet in frontof the ice cream store at 10 o'clock and walktogether to the church. One Sunday morningwhen Alice is getting ready to leave, her moth-er asks her, "Have you called Mary to makesure you are meeting in front of the ice creamstore?" Alice says, "No, but I know that Marywill be there."

A. Alice conceives Mary will be there.B. Alice contemplates Mary will be there.C. Alice anticipates Mary will be there.D. Alice infers Mary will be there.


4

NRRC NationalReading ResearchCenter318 Aderhold, University of Georgia, Athens, Georgia 30602-71252102 J. M. Patterson Building. University of Maryland, College Park MD 20742

44

ED 373 305 AUTHOR TITLE Report No. 14.DOCUMENT RESUME ED 373 305 CS 011 804 AUTHOR Booth, James R.;...

Documents

Transcript of ED 373 305 AUTHOR TITLE Report No. 14.DOCUMENT RESUME ED 373 305 CS 011 804 AUTHOR Booth, James R.;...