Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Thinking outloud: the problem-solving language of preschoolers with and without language impairment Sturn, Paula A. 1993

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

Item Metadata


831-ubc_1993_spring_sturn_paula.pdf [ 4.29MB ]
JSON: 831-1.0086169.json
JSON-LD: 831-1.0086169-ld.json
RDF/XML (Pretty): 831-1.0086169-rdf.xml
RDF/JSON: 831-1.0086169-rdf.json
Turtle: 831-1.0086169-turtle.txt
N-Triples: 831-1.0086169-rdf-ntriples.txt
Original Record: 831-1.0086169-source.json
Full Text

Full Text

to the required standardTHINKING OUTLOUD: THE PROBLEM-SOLVING LANGUAGE OF PRESCHOOLERSWITH AND WITHOUT LANGUAGE IMPAIRMENTbyPAULA ARLENE STURNB.A., Simon Fraser University, 1985A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFMASTER OF SCIENCEinTHE FACULTY OF GRADUATE STUDIES(School of Audiology and Speech Sciences)We accept this thesis as conformingTHE UNIVERSITY OF BRITISH COLUMBIAApril 1993© Paula Arlene Sturn, 1993In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.(Signature) SC,k0o1Deraftffleitt of AVAA,0i0y^.1::e.ecPA. The University of British ColumbiaVancouver, CanadaDate^Ci; 04_ 2_5(DE-6 (2/88)ABSTRACTEighteen preschoolers were asked to build bridges out of diverse construction material, twochildren working in parallel. Their utterances were coded for task relevance, function(regulative vs. affective) and addressee (private vs. social). The language of six children withspecific language impairment was compared to that of six age-matched, and six language-matched peers. Correlations between dependency, impulsivity and efficiency and language usewere also investigated. 39% of the speech was narrowly task relevant and regulative. Similarproportions of private and social speech occurred but problem-solving language was more likelyto be private (61 %). Group differences were found only in amount of speech. These findingsindicate first that the correlation between function and addressee, while consonant withtraditional accounts, is far from perfect: researchers interested in the intellectual functions oflanguage should study both social and private speech. Second, language impaired children usespeech for problem-solving to an age appropriate degree, though formulation problems may stillreduce the efficacy of such use. A positive correlation was found between impulsivity and taskirrelevant speech, while dependency was negatively correlated with the total number ofutterances, the number of broadly task relevant utterances, and the number of social utterances.Correlations between efficiency and language use revealed a split between children with normallangauge and children with specific language impairment. Efficient children with normallanguage spoke more and devoted more of their speech to narrowly task relevant utterances.Efficient children with specific language impairment spoke less but also devoted a largerproportion of their speech to narrowly task relevant speech. These findings indicate that theintellectual function of language is used productively by efficient children regardless of languagestatus but efficient children with specific langauge impairment rely on it less.iiTABLE OF CONTENTSABSTRACTLIST OF TABLES^ viiLIST OF FIGURES viiiACKNOWLEDGEMENTS^ ixCHAPTER 1: LITERATURE REVIEW^ 1The Emergence of Semiotic Function 1Connections Between Thought and Language^ 2Language Communicates Thought 3Inner Speech: Language Represents Thought^ 5Egocentric Speech^ 6Piaget and Egocentric Speech: Failure to Adopt^ 7Vygotsky and Egocentric Speech: Speech for Self 11Internal and External Speech^ 12The Function of Egocentric Speech 15Current Developments: Private Speech^ 17The Potential Cognitive Role of Social Speech^ 18Contemporary Studies of Private Speech 21The Social Roots of Private Speech^ 22Developmental Course^ 26Cognitive Self-Guiding Function^ 33iiiPersonality Characteristics and Private Speech^ 37The Effects of Specific Language Impairment (SLI) on Cognition^39The Current Study^ 40CHAPTER 2: METHODOverview^ 42Subjects 42Procedures^ 48Assessment Procedures^ 48Standardized Tests 49Experimental Task^ 49Procedures 50Experimenter Conduct^ 52Transcription and Coding of Language Data 53Coding Scheme^ 53Function 54Addressee^ 55Cognitive Complexity^ 56Inter-Rater Reliability 56CHAPTER 3: RESULTS^ 58Speech and Problem Solving^ 59ivThe Roles of Social and Private Speech in Problem Solving^60The Effects of SLI on Regulatory Language^ 63Interactive vs. Solitary Problem Solvers 66Personality Characteristics and Language Use^ 67CHAPTER 4: DISCUSSION^ 75Intellectual Functions of Language^ 75Broadly Task Relevant (BTR) Speech^ 76Narrowly Task Relevant (NTR) Speech 77General Incidence of Private Speech^ 82Effects of SLI on Language Use 83Personal Style^ 85Solitary or Interactive^ 85Personality Characteristics 87Impulsivity^ 87Dependency 88Efficiency^ 89Clinical Implications 92Theoretical Implications^ 93Future Implications 95Conclusion^ 96vBIBLIOGRAPHY^ 98APPENDIX A: DESCRIPTION OF LANGUAGE ASSESSMENT MATERIALA^102APPENDIX B: LETTER AND CONSENT FORM SENT TO PARENTS^105APPENDIX C: PERSONALITY CHARACTERISTIC MEASURES 108viLIST OF TABLESTABLE 1^Subject Selection Data^ 46TABLE 2^Subject Selection Data Comparison^ 47TABEL 3^Coding Categories for Problem Solving Utterances^ 57TABLE 4^The Functions of SPeech in Problem Solving Tasks 60TABLE 5^Proportion of Total Utterances that were Task Relevant^62and Either Private or Social SpeechTABLE 6^Proportion of Task Relevant Utterances that were 63Private and Social in NatureTABLE 7^Narrowly Task Relevant, Regulating Speech^ 64TABLE 8^Personality Characteristics^ 68TABLE 9^Correlations Between Personality Characteristics and^70Language UseviiLIST OF FIGURESFigure 1^Diagram of River Valley^ 50viiiACKNOWLEDGEMENTSI would like to extend my heartfelt thanks to Judith Johnston, without whom this study wouldnever have happened. I would also like to thank: the children and their families for theirparticipation in this study; Carolyn Johnson and Ingrid Jeffrey for their advise; and the staff andfaculty at the U.B.C. School of Audiology and Speech Sciences for their support andencouragment.ix1CHAPTER ONELITERATUREB  REVIEWThe Emergence of the Semiotic Function During most of the first two years of life infants are unable to use symbolic representations,which leaves them anchored to their immediate environment. They begin interacting with theirsurroundings through reflexive sensori-motor movements and through this interaction learn toattribute meaning to actions and objects (Ginsburg & Opper, 1979; Piaget & Inhelder, 1969).At this point, however, thought and action cannot be separated, leaving the child tied to thepresent. During this period infants make many cognitive advances, including the ability toimitate something previously seen and to anticipate the future position of an object before itmoves, as well as the understanding of object permanence (Cromer, 1991). Nonetheless, it is onlyas children develop the capacity to use a symbol to represent something not immediately present,i.e. the semiotic function, that they are able to expand their perspective beyond the immediateenvironment.By approximately age two, certain cognitive skills emerge which indicate that the child isbeginning to employ the semiotic function. At this point children develop the ability to usemental symbols, usually words, and thought and action become separated. This reflective abilityallows children to think about objects that are not immediately present and engage in symbolicplay routines such as "house", where they use role play and toys to symbolize real life. Thisreflective ability also allows children to consider past experiences and future possibilities (Piaget& Inhelder, 1969). The child is able to remember that "yesterday I didn't come up 'cause I was2too sick" as well contemplate plans for tomorrow, e.g. "are you going to bring this tomorrow sowe can play again'?" Since this and other such skills require the use of mental symbols, theadvent of the semiotic function is a significant step in the child's development. The child, freedfrom the immediate environment, is now able to begin exploring the world of possibilities.Connections Between Thought and Language Language, being the most common, overt form of semiotic function in humans, appears aboutthe same time as other manifestations of the semiotic function. Although language is only oneexpression of the semiotic function, it seems to play a major role in the way children learn tostructure their thoughts. "It contains a notation for an entire system of cognitive instruments...foruse in the service of thought" (Piaget & Inhelder, 1969), or, as Moffett argues, it is an abstractionof the structure of the world into a symbolic system that more closely resembles the structure ofthe mind (as cited in Blank & Franklin, 1980). Researchers argue that language allows the childto: 1) represent a number of factors simultaneously (Cromer, 1991; Piaget & Inhelder, 1969); 2)make spatial, temporal and causal discriminations (Blank, 1974; Cromer, 1991; Opper &Ginsburg, 1979); and 3) increase the speed of representation over other nonverbal representationalsystems (Blank, 1974; Cromer, 1991).Language is also commonly used to refer to absent objects and actions, states of affairs,emotions, and abstract concepts. It allows communication between individuals of not only whatis seen, but also about what is unseeable, e.g. propositional connections and thematic schemes(Ginsburg & Inhelder, 1979; Heath, 1989). In short, language is both a mode for the symbolic3representation of thought and a means for communicating that thought to others. These two rolesof language, and their developmental connections, are discussed below in more depth.Language Communicates ThoughtLanguage is the human's tool of choice for social interaction. It is both learned and used in asocial setting. Children actively seek out new information in their environment. Throughcommunication with peers and adults they can learn about, and from, the experiences of others.Cultural knowledge, for example, is passed down from adult to child, from generation togeneration, largely via language (Ginsburg & Opper, 1979). Perhaps more interestingly, socialinteractions contribute to the development of language as a tool for expressing thought. "Thisis of the very first importance for the understanding of child thought...intelligence, just becauseit undergoes a gradual process of socialization, is enabled through the bond established bylanguage between thoughts and words" (Piaget, 1955, p.64). By means of social interaction,children learn to express their own thoughts in a way that is comprehensible to others. Althoughthought and language are not innately connected, through social interactions with others, childrenlearn to use language to generate and express complex thought sequences.With development, thought and speech unite into verbal acts of thought that reflect reality in away quite different from sensation. Researchers differ somewhat in their account of this change.Piaget believed that children began life in an individualized "autistic" state (as cited in Vygotsky,1934/1968, p.16). Though interacting with others, the child must progress through an egocentricperiod before developing socialized, increasingly mature thought (Duncan, 1991). The child4advances through many intermediary stages between autism and mature thought but the mostimportant one is egocentric thought (Piaget, 1955). At the egocentric stage, children attempt toadapt their thoughts to reality but have not yet learned to communicate them as such.Accordingly, their early language not only reflects this egocentric style of thought, but alsoinfluences and fosters it . "The mere fact, then, of telling one's thought, of telling it to others,or of keeping silence and telling it only to oneself must be of enormous importance to thefundamental structure and functioning of thought in general and child logic in particular" (Piaget,1955, p.64).Piaget stressed peer interaction rather than adult-child interaction as the crucial factor in thedevelopment and socialization of children's thought and language (Forman & Kraker, 1985;Piaget & Inhelder, 1968). Children often play a passive role in adult-child interactions, simplycomplying with demands placed on them by the adult. In peer interactions, however, cognitivedevelopment is fostered because a mismatch is often set up between the thoughts of the childand those expressed by peers. Children are more likely to contest the ideas of their peers andtry to reconcile them to their own ideas. When there is a disagreement between peers the newinformation is frequently assimilated into the child's thought, resulting in adjustments being madeto the child's view of the world (Forman & Kraker, 1985, Opper & Ginsburg, 1979; Piaget &Inhelder, 1968).In opposition to Piaget, Vygotsky saw speech as being socialized from the beginning. Heconsidered gestures and cries, as well as speech and language proper, to be socialized, adaptive5communication. (Vygotsky, 1934/1968). Rather than developing from an individualized to asocialized state, Vygotsky believed that children developed from a socialized state to anincreasingly individualized state of being. In early interactions, adults often use language todirect and regulate children's actions. Adults guide children through difficult tasks, enablingthem to solve problems they would not have been able to work out on their own. As the child'scompetence at the task improves, the adult gradually shifts the control of problem-solving to thechild. As children mature they internalize the adult's regulation and the emphasis shifts fromother-regulation to self-regulation of thought and action. "Since the main course of the child'sdevelopment is one of gradual individualization, this tendency is reflected in the function andstructure of his speech" (Vygotsky, 1934/1968, p.133). Children begin to use language toregulate their own thoughts and behaviour in the same way that the adult's language onceregulated them (Vygotsky, 1934/1968). As children develop, social interactions with othersfacilitate their ability to function as independent self-regulated problem-solvers for whomlanguage functions as a private-cognitive as well as a social-communicative tool.Inner Speech: Language Represents ThoughtWhichever of these approaches one takes, it is clear that well before adulthood, languagefunctions as more than just a medium for social communication. One special form of language,"inner speech", plays a private rather than a social function. As it is usually conceived, innerspeech is a form of language that is used strictly for thinking for oneself (Vygotsky, 1934/1962;Feldman, 1977). Vygotsky's account of this construct is the most fully elaborated. He arguesthat speech turns inward as its function changes from social to personal adaptation. Inner speech6is an abbreviated, internal form of language in which only the predicate is expressed. It is aninstrument for symbolically representing ideas in order to delineate and clarify them more clearly(Feldman; 1977). A single word or nonconventonal sign may become so saturated withmeaning that it takes many words to explain it in external speech (Vygotsky, 1934/1962). Innerspeech is used for thought and, as such, Vygotsky (1934/1968) believed it to be the intermediarybetween language and thought. "Inner speech is to a large extent thinking in pure meanings. Itis a dynamic, shifting, unstable thing, fluttering between word and thought, the two more or lessstable, more or less firmly delineated components of verbal thought" (Vygotsky, 1934/1962,p.149). If, as Vygotsky hypothesized, inner speech is the connection between thought andlanguage, then determining the specific characteristics and functions of inner speech could revealthe connection between the two. Unfortunately, by its very nature, inner speech is inaccessibleto regular methods of research, making it a difficult notion to study.Egocentric Speech One avenue that is open to investigation, however, is the phenomenon of "egocentric speech".Vygotsky (1962/1934) postulated that egocentric speech is an intermediate step between external,socialized language and internalized, inner speech, and that it occurs in the developmental stagebefore inner speech emerges. It is expressed like external speech but fulfils many of the samefunctions as inner speech. "When circumstances force (the child) to stop and think, he is likelyto think aloud. Egocentric speech, splintered off from general social speech, in time leads toinner speech" (Vygotsky, 1934/1968, p.19). For Vygotsky, egocentric speech is speech on itsway inward, on the whole already incomprehensible to others and meant for self, but still7manifesting itself externally. The structural and functional idiosyncrasies of verbal thought areshaped as it progresses from external speech to egocentric speech to inner speech.Both Vygotsky and Piaget theorized about the implications of egocentric speech in thedevelopment of language and cognition in young children but drew different conclusions aboutits role. According to Vygotsky (1934/1962), egocentric speech assists in cognitive self-guidance.It is used to plan solutions to problems and to direct thoughts and actions toward consciouslydirected goals. Piaget, on the other hand, hypothesized that egocentric speech plays little functionin children's thinking or activity. It primarily occurs in conjunction with them because youngchildren do not differentiate between actions and words (Ginsburg & Opper, 1979). Childrenspeak their thoughts out loud because they are "incapable of keeping to [themselves] the thoughtswhich enter his mind" (Piaget, 1955, p.59). Egocentric speech, then, provides a window into thechild's thought. The following sections will consider the ideas of each of these theorists in moredetail.Piaget and Egocentric Speech: Failure to AdaptPiaget, who introduced the term egocentric speech (as cited in Goudena, 1987), described it asaudible speech that does not appear to be directed or adapted to a listener's point of view. Hewas not interested in egocentric speech per se, but considered it an avenue for studying children'scognition (Goudena, 1987). For Piaget, egocentric speech plays no important role in the8psychological or cognitive development of children. Their speech is abbreviated and missinginformation the listener needs to understand what the child is trying to say because children arecentred around themselves and fail to take the listener's point of view into consideration whenspeaking. Their speech, then, reflects the egocentric nature of their thought. "Since languageand logic are obviously interdependent" (Piaget, 1955, p.28), egocentric speech provides insightinto the child's thoughts.While Piaget believed that the unfolding of the semiotic function contributes to cognitivedevelopment, he felt it was the physical realizations rather than linguistic realizations that are ofprimary importance. At this age inner speech is not yet possible, so children use action to fulfilsimilar functions. "Rather than simply recalling an event, the child has need of a more directsymbolism which enables him to relive an event" (Piaget & Inhelder, 1969). The childsymbolizes an event through acting it out. When egocentric speech occurs "the child is simplythinking out his action aloud" (Piaget, 1955, p.36). Before approximately age seven, children havenot learned to differentiate between word and action so the word is considered part of the objector action. Thought and word are linked so tightly that they also occur simultaneously. The childsymbolizes thought through action, and egocentric speech occurs simply as an attribute of theaction.In studying the spontaneous speech produced by two six-and-a-half-year-old subjects, Piaget(1955) discovered that egocentric speech accounted for nearly half of the speech produced byboth children. This high proportion of egocentric speech in their spontaneous language led Piaget9to regard egocentric speech as a useful instrument for the study of child logic. He believed itprovided great insight into the child's comprehension of the world, and studied the phenomenonclosely enough to devise a developmental scheme (Opper & Ginsburg, 1979). Under Piaget'sscheme, egocentric speech is broken down into three categories: collective monologue, individualmonologue and repetition.The collective monologue occurs when two or more children are together and one child isspeaking but is ignored by the other(s). Piaget hypothesized that when such collectivemonologues occur the speaker is attempting to communicate with those present. The child mayeven believe that communication is taking place, but the egocentric nature of the speech makesit impossible for others to understand (Ginsburg & Opper, 1979). The speech is often lackingin definite reference and is centred about the child's activities, real or imagined. The child doesnot consider what information is required for the listeners to understand. There is no attempt atdialogue, or conversation development with the other children. In fact, it is sometimes the casethat two or more children will carry on collective monologues simultaneously with no regard tothe interest or understanding of the others (Piaget & Inhelder, 1969). In such cases the languageis no longer communicative or social in nature. It merely occurs in the presence of others.The individual monologue is similar to the collective monologue except that there is no onepresent to ignore the child: the child is alone and yet speaks out loud. Two hypotheses wereproposed by Piaget to explain the occurrence of this type of speech (Ginsburg & Opper, 1979).As a first hypothesis, Piaget suggested that in some cases the child's speech is a type of word10fantasy. The child uses language as a kind of wish fulfilment, to accomplish what action cannot."If the child talks even when he is alone as an accompaniment to his action, he can reverse theprocess and use words to bring about what the action itself is powerless to do" (Piaget, 1955,p.36). Unable to achieve a goal through actions, the child attempts to attain the goal with wordsinstead. During one of the experimental sessions in this study, ME, after numerous failedattempts to insert a pipecleaner into a straw, told the pipecleaner "hey, come on, come on in".The other hypothesis suggested by Piaget is that words and actions have not become separatedand the child sees the two as an undivided whole. The word is a part of the action or thing andso accompanies the child's actions. "In such cases speech does not communicate the thoughtsof the speaker; it serves to accompany, to reinforce, or to supplement his action" (Piaget, 1955,p.39). The arbitrary relationship between language and action is not yet clear in the child's mind,so the child treats the word as a component of the action. The individual monologue, then, issimply part of the child's actions and not communicative.The third component of Piaget's classification scheme, repetition, makes up a very smallproportion of egocentric speech. Repetition is simply that: the child echoes something just heard.According to Piaget, when children mimic they often are not aware that they are merely repeatingsomething another person has just said (Ginsburg & Opper, 1979). Imitation occurs when thechild confuses the activities of others with the activities of self (Piaget, 1955). Repetition is notcommunicative in nature. Children repeat both their own words and those of others to practise11verbal skills just as they repeated behaviour patterns to practise sensori-motor skills during theearlier sensori-motor period.As children mature, their thought becomes socialized and they learn to take others' points of viewinto account (Ginsburg & Opper, 1979). Just as infants progress from an "autistic" sensori-motorstage to an egocentric period, so toddlers advance through egocentricity and on to maturing,socialized thought. The child speaks "first and foremost to himself, and that speech, before itcan be used to socialize thought, serves to accompany and reinforce individual activity" (Piaget,1955, p.59). As children's thoughts become more socialized, egocentric thought, and thereforeegocentric speech, decreases. Both gradually diminish until they finally disappear.Vygotsky and Egocentric Speech: Speech for SelfThe Piagetian view emphasized the communicative inadequacy of egocentric speech and paidlittle attention to its function. Egocentric speech was seen to provide a window to the child'sthought, but it was not viewed as important to the development of mature logic and reasoning.Vygotsky (1934/1968) differed dramatically from Piaget on each of these points. He believedthat the development and internalization of egocentric speech was key to the maturation ofintellect and thought. For Vygotsky, the egocentric speech of young children is thought whichhas not yet been internalized. More than that, however, it is a tool for cognitive self-guidance,aiding children in planning, orienting, organizing and structuring their behaviour as well asguiding and regulating their thoughts (Vygotsky, 1934/1968; Berk, 1986; Goudena, 1987; Harris,1990). Egocentric speech "does not merely accompany the child's activity: it serves mental12orientation, conscious understanding; it helps in overcoming difficulties; it is speech for oneself,intimately and usefully connected with the child's thinking" (Vygotsky 1962/1934, p.133).Internal and External Speech: Vygotsky (1934/1968) considered all language, both social andprivate, to be made up of two fundamental elements: the internal semantic representation and theexternal phonetic one. The relationship of meaning (semantic) to "word" (phonetic) is a dynamicone. Word meanings evolve and change as the language and culture evolve. "The relationshipof thought to word is not a thing but a process, a continual movement back and forth fromthought to word and from word to thought. In that process the relation of thought to wordundergoes changes which themselves may be regarded as development in the functional sense"(Vygotsky, 1934/1968, p. 125). The external and internal aspects of speech also change anddevelop over time, but each does so in its own unique way. As the maturation of these twocomponents progresses, the relationship between thought and word changes. At each stage wordmeaning has a distinct, complex relationship with both speech and thought.Though forming a complex whole, the external and internal aspects of language have their ownindependent development. As a child learns to speak, the external phonetic component of speech(external speech) progresses from single word to complete sentence - from part to whole(Vygotsky, 1934/1968). Children typically progress from single word to two word sentences.Gradually, as their language skills develop, children advance from simple to more complexsentences. Finally, as grammar and morphology are fully developed, speech becomes a coherentstream of sentence linked to sentence, expressing complex thoughts.13In contrast, the internal semantic component of language (internal speech) begins as a complexwhole that only later gets broken down into its simpler semantic units (Vygotsky 1934/1968).Children's first utterances are often single word sentences where one word expresses a completethought. Only as the semantic component of the language develops do children begin to masterthe meaning of words and break down complex thoughts into smaller differentiated segments.Word meanings "change as the child develops; they change also with the various ways in whichthought functions" (Vygotsky, 1934/1968,p. 124). Children learn to use words as representationsof simpler, discrete semantic units which can then be combined to express intricate thoughts.Ergo, the external and internal aspects of language develop in reverse directions: external speechdevelops from single word to multiple sentences, while the internal component of languageevolves from complex whole to discrete part.Because these two aspects of language develop in opposite directions it is important todistinguish between them. This does not mean, however, that the two are unrelated. Thedevelopment of each one plays a significant role in the development of the other (Vygotsky,1934/1968). The structure of language does not duplicate the structure of thought. Languagedoes not merely communicate thought, it shapes thought as well. Thought undergoes significantchanges as it turns into speech. "Thought is not merely expressed in words; it comes intoexistence through them" (Vygotsky, 1934/1968, p.125). Children's thoughts begin as vaguecomposites that can only find expression in single words. As their thoughts becomedifferentiated children are able to construct complex strings of words to express those thoughts.14Conversely, as children progress from single to multiple word sentences their amorphous thoughtsget broken down into discrete units that can then be combined in a variety of ways.It is still necessary to look past the external and internal components of language, however, tounderstand the relationship between thought and language. This relationship goes beyond thephonetic and semantic aspects of the word to the plane of inner speech, where thought andlanguage eventually unite. "The relationship of thought and word cannot be understood in allits complexity without a clear understanding of the psychological nature of inner speech"(Vygotsky, 1934/1968, p.130). Once language and thought have developed some degree ofsophistication, the child begins to use language for thinking. At first, private speech and socialspeech are indistinguishable, but as language develops further and the relationship betweenthought and language progresses, distinctly personal uses of language can be seen. Suchegocentric speech is a transitional phase in the evolution of social speech to inner speech, andas such, exhibits characteristics of both. Although it is still vocalized, audible speech, it has thefunction and structure of inner speech, making it a key to the connection between language andthought.In the early stages of language, egocentric speech and social speech are undifferentiated becauselangauge is "parasocial" in nature (Vygotsky, 1934/1968). When speaking, children distinguishbetween themselves as speakers and others as listeners, but they do not distinguish betweenthemselves as listener and other as listener (Kohlberg et al, 1968). Because they do not15distinguish between themselves and others as listener, neither do they distinguish between speechfor self and speech for other.As the child matures, however, the structural and functional idiosyncrasies of egocentric speechevolve. It becomes less audible, less socially communicative and more abridged in nature(Vygotsky, 1934/1968). As the specific function of egocentric speech develops, its syntacticpeculiarities become more pronounced and it becomes less and less similar to external speech.These characteristics are most prominent by approximately age seven. By this point, egocentricspeech has developed a structure and function that is totally different from social speech, andchildren realize that it plays a private, not communicative role. As they become capable ofabstracting the sound from the language, vocalizations decrease and become moreincomprehensible. The child is able to think the words instead of saying them, and inner speechbegins (Vygotsky 1934/1968).The Function of Egocentric Speech:  If Vygotsky is correct in hypothesizing that children useegocentric speech in much the same way as adults use inner speech, then egocentric speech playsa central role in the development of children's cognitive processes. It is speech for self,intimately connected with the child's thoughts. Children will use speech for self to focus anddirect their thoughts. Egocentric speech, then, does not merely accompany children's actions,it aids in mental orientation and conscious understanding, increases attentional focus, and guidesand directs motor activities (Vygotsky, 1934/1968; Bivens & Berk, 1990; Frauenglass & Diaz,1985). Egocentric speech makes the relationship between action and goal explicit and therefore16clearer (Pellegrini & DeStefano, 1979). It also serves as an affective release, as an avenue toexpress both positive and negative emotion about the state of affairs (Vygotsky, 1934/1968;Goodman, 1981; Berk & Garvin, 1984).Vygotsky (1934/1968) particularly emphasized the self-regulatory, planning function of speechfor self. He believed that egocentric speech "represents the child's attempts to use language asan instrument of thought...a tool to plan, guide and monitor problem-solving activity"(Frauenglass & Diaz, 1985, p.35'7). Its function is to communicate with the self: to plan, guideand direct actions. Unlike adults, children vocalize these thoughts because they cannot directtheir actions in a covert manner (Kohlberg et al., 1968). Using egocentric speech helps them todevelop strategies and action-plans for problem-solving. Verbalizing thoughts about possiblesteps to take assists children in mapping out feasible alternatives before and during problem-solving tasks.The semantic aspect of speech for self assists in orienting and organizing behaviour, accordingto Vygotsky (1934/1968). Children use egocentric speech to consciously understand a situationand overcome difficulties manifested by that situation (Harris, 1990). Speech for self increaseschildren's focus on a problem and allows them to explicitly outline its exact nature. It helpschildren direct their attention more closely to the problem at hand, focus on it, and identify theprecise obstacles that need to be overcome. By aiding in mental orientation, it permits childrento delineate the problem-area, eliminate factors that are not necessary to the problem-solving taskand concentrate on the factors most pertinent to the problem.17Speech for self also serves an expressive function, including motivational utterances, commentsabout task performance, and other positive and negative affective verbalizations (Vygotsky,1934/1968; Meichenbaum & Goodman, 1979; Furrow, 1984). Children often make commentsto themselves, both positive and negative, about their performance and ability to do a task(Harris, 1986; Goodman, 1981). Expletives are expressed when failure or success occurs.Throughout a task children remark on their performance and compare it to that of other children.These types of utterances all communicate emotion, and egocentric speech serves as the meansof expressing it.Given the rich theorizing and intrigue on the part of these two pioneers of egocentric, or private,speech, it is not surprising that there is recurrent interest in this phenomena. The perspectivesof Vygotsky and Piaget differ so dramatically on this point, however, that the current literatureis often difficult to interpret.Current Developments: Private Speech Since Vygotsky first hypothesized speech for self as an instrument of human cognition, theconcept has evolved from a functionally defined construct with incidental behaviourialaccompaniments to a behaviourially defined construct with less clear function. In the currentliterature, speech for self has been transformed into the notion of private speech, a term coinedby Flavell (as cited in Goudena, 1987). Contemporary researchers share Vygotsky's view thatprivate speech is speech that guides thought, but his purely functional definition has given wayto a more restrictive definition that can be operationalized. As well as having content that refers18principally to the speaker's own activities and consciousness about those activities, private speechmust now be speech that is clearly not social. In behaviourial terms, this usually reduces to thecriterion that there be no evidence of an addressee. Vygotsky made no distinction betweenspeech for self and social speech: an utterance could be spoken in the presence of a listener, withclear intent to communicate, and still be speech for self. It "may come very close in form toexternal speech or even become exactly like it" (Vygotsky, 1934/1968, p. 47). WhereasVygotksy's speech for self was defined by the cognitive function it fulfils, private speech isimportantly defined by what it is not. The effect of this modern change has been, of course, adramatic reduction in the range of speech and language behaviours that are viewed as servingthought. A careful look at the literature suggests that valuable information has been lost.The Potential Cognitive Role of Social SpeechAlthough a large proportion of speech generated during private speech studies is social in nature,its occurrence is normally considered an unwanted concomitant of the experimental design andis discouraged by the experimenter (Duncan, 1991; Berk, 1986; Goodman, 1981; Kohlberg et al,1968). Those studies that investigate children's speech in social settings with adults discuss theinfluence that adult speech has on children's cognition and problem-solving abilities (Behrendet al., 1989; Goudena; 1987), but offer no reflection on the influence the children's own socialspeech has on task performance. Yet, the form of language children use to interact with othersfundamentally affects the type of information they, in turn, receive (Blank & Franklin, 1980).Since children produce social speech during problem-solving tasks and other activities, even when19it is discouraged, it seems feasible that such speech performs functions beyond that ofcommunication.Very little information is available on the function that social speech performs in children'sactivities. Goodman (1981) examined how children's verbal and motoric behaviour are related.Although social speech was not directly examined, this is the only study on private speech thatreports the proportion of social speech occurring during the task. Thirty-eight four-year oldchildren were video-taped during a jigsaw-puzzle solving task. The children's verbalizationswere transcribed and categorized into social and private speech. Social speech was categorizedas task-relevant or task-irrelevant, while private speech was classified as fulfilling one of sixfunctions. In addition, the children's actions were also unitized and categorized on a five pointscheme: 1) immediately successful placement; 2) success following spatial reorientation; 3)success following trail and error; 4) failure due to placement error; and 5) failure, piece returnedto table. The co-occurrence of private speech and puzzle-solving behaviours was then examined.The results pertaining to private speech will be discussed in detail at a later point in the chapter.What is of interest here is that task-relevant social speech made up 24% of the children'sutterances during the task. This equals the proportion of utterances that were private and devotedto verbalizing plans or thoughts. Another point of interest is that very little of the social speechwas task-irrelevant. Only 3% of the verbalizations fell into this category.There is only one study to date which presents an overall view of the function of children'sspeech. Furrow (1984) compared the uses of social and private speech in two-year-olds,20employing the same categorization scheme for all utterances. Twelve children were audio- andvideo-taped in a free play session with an adult. Two hundred spontaneous utterances were: 1)classified as social (eye-contact), other social, and private; and 2) assigned one of twelvecommunicative functions. Because Furrow's classification system accounts for both social andprivate utterances, utterances are categorized in a manner that is quite different from theclassification schema used for private speech. The twelve categories used in this schema include:1) Instrumental: an utterance refers to the child's wants or is "whined"; 2) Regulatory: anutterance refers to an event that might be immediately carried out; the child is not the statedagent and the action is not performed by the child; 3) Self-regulatory: an utterance refers to anevent that might be immediately carried out; the child is the stated agent and the action isperformed by the child; 4) Attentional: an utterance refers to a sensory event that is ongoing ormight immediately be carried out; 5) Interactional: utterance content is a conventional greeting;6) Expressive: Utterance content is an evaluative opinion, an expression of an internal state orstock phrase that expresses feeling; 7) Referential: an utterance refers to a present object or apresent event that does not involve the child; 8) Describing own activity: an utterance refers toan ongoing or just completed event in which the child was involved; 9) Question: utteranceintonation contour resembles adult rising question intonation and/or an utterance is syntacticallya question; 10) Imaginary: an utterance is sung, is word play, or represents a transformation ofreal objects or events, whether present or not; 11) Informative: an utterance refers to a nonpresentobject or event; and 12) Incomprehensible: an utterance is inaudible or incomprehensible. Resultsindicated that regulatory, attentional and informative functions were exclusive to social speech.While private speech tended to be self-regulatory, expressive and descriptive of own activity, no21functions were exclusively private in nature. All functions performed by private speech werealso performed by social speech.Much still needs to be learned about the role of social speech in problem solving, but these twostudies establish that while social speech appears to have unique functions, it performs all of thefunctions of private speech as well. If this is true, then social speech will contribute to thecognitive development of a child in a manner that is parallel to private speech.Contemporary Studies of Private SpeechDespite the fact that studies of private speech have examined only a subset of the functionallyinteresting data, they have largely supported Vygotsky's view that language can guide thought.Private speech has been most commonly defined as speech that is either not adapted to a listeneror not clearly and definitely addressed to another (Berk & Garvin, 1984). It is speech that is notovertly marked as intended for a listener: (1) it does not occur in combination with eye contactand/or body orientation toward the listener; (2) there is no explicit verbal reference to the listener;(3) the remark is not repeated until a response is received; (4) a response is not demanded; (5)no physical proximity/contact is initiated; (5) volume is reduced relative to social utterances(based on Feigenbaum, 1989). Using some version of this definition, a number of studiesperformed over the past fifteen years support the three main hypotheses put forth by Vygotsky(1934/1968): 1) private speech develops out of social speech; 2) private speech follows adevelopmental pattern that culminates in true inner speech; and 3) children use private speechto guide and direct behaviour. Although each study investigates a different facet of Vygotsky's22theory on private speech, together they provide clear support for these three hypotheses. Sinceeach of the studies is designed to investigate slightly different aspects of private speech, each alsouses a slightly different categorization scheme for private speech. The schemes are similarenough, however, to allow a general discussion on the role of private speech in children'sproblem-solving and the developmental ascendency of private speech, with its roots in socialspeech and its maturity in true inner speech.The Social Roots of Private Speech: A study by Berk & Garvin (1984) investigated thedevelopment of private speech in low income Appalachian children. Subjects in this studyconsisted of thirty-six low income Appalachian children, ranging in age from four to ten yearsold. Two observers dictated detailed narratives of the children's behaviour into tape recorders.The children were observed for two continuous one-hour periods, one by each of the observers.Each child was observed in each of the following natural school settings: the classroom, theplayground, the halls and the lunchroom. All speech was coded as private or social. Twoenvironmental conditions were also examined in this study: 1) cognitive demands of tasks; and2) adult presence. The data pertaining to private speech and the effects of these conditions willbe discussed at a later point in the chapter. The overall frequencies of social and private speechproduced by each child were positively correlated. The more verbally interactive children alsoproduced more private speech. An additional finding suggests that this was not due to meretalkativeness. The earlier forms of private speech, e.g. word play and repetition, were morestrongly correlated to social interaction than were the later forms, e.g. describing own23activity/self-guidance and inaudible mutterings. This supports the notion that social interactionassists in the development of early forms of private speech.Similar findings were obtained in the first of four pioneering studies by Kohlberg et al. (1968).This study was designed to investigate the incidence of private speech, its relationship to thechild's cognitive level, and the correlation between social interaction and incidence of privatespeech. The third purpose will be considered here, while the first two will be discussed below.Twenty-eight children between four and seven participated in the first study. They were dividedin to four equal groups: bright 4's, average 4's, bright 6's and average 6's. The mean Stanford-Binet IQ of the bright subjects was 129 and of the average subjects, 104. Each child wasfollowed by an observer for two morning school or preschool sessions, and detailed recordingswere made of all speech and behaviour for a period of two hours. Approximately equal portionsof each narrative were devoted to outdoor play, indoor free play and adult-structured activities.All speech was coded as private or social. Each child was also rated by a teacher on a 7-pointscale of popularity, indicating general social participation. A positive correlation was foundbetween the amounts of private and social speech produced. A positive correlation was alsofound between ratings of popularity and the amount of private speech generated. These findingssuggest that private speech and social speech both develop out of a social orientation and reflectthe "parasocial" nature of private speech.Goudena (1987) hypothesized that children would be more likely to produce private speechfollowing interaction with a collaborative adult than a non-collaborative adult. Twenty-two 4-24year-old children participated in the study. Each child was observed under two conditions, thecollaborative (C) and the noncollaborative (NC) conditions. The order of presentation of the twoconditions was randomly counterbalanced. In condition C one of two female experimenterspresented two puzzle-type tasks to the child. The experimenter completed the first puzzle andthen asked the child to also complete the task. If the child displayed difficulty, the experimenter,without verbalizing during the task, put the child's pieces together correctly and then asked thechild to try again. The second task was done in a similar manner except that the experimenteronly arranged the first item of the puzzle before the child started to work. In the NC conditionone of two experimenters also presented two puzzle-type tasks. The experimenter did notcomplete the puzzle in the presence of the child but presented an already completed figure forthe child to copy. If the child failed at the task the experimenter simply asked the child to tryagain and did not offer any assistance. When the presentation of the two puzzles was finished(constituting the C or NC conditions), the experimenter handed the first of two new puzzles tothe child and asked the child to try to complete it, indicating that she was going to be occupiedwith something else. The experimenter then sat down in a corner of the room behind the child'sback, started to read and was minimally responsive to social comments made by the child. Thechildren produced more private speech overall in condition C than in condition NC. Whenchildren have received direct regulation and that regulation is then withdrawn, they were morelikely to produce private speech than when they have not received any regulation in the firstplace. This supports the view put forth by Vygotsky (1934/1968) that children internalize theadult's regulation and the emphasis shifts from other-regulation to self-regulation of thought and25action. Children begin to use language to regulate their own thoughts and behaviour in the sameway that the adult's language once regulated them.Goudena also reports, however, that when only the instruction periods are considered, childrenproduced approximately three times as much private speech during the NC instruction sessionas during the C instruction session. When the children received direct regulation from an adultwhile learning a task, they produced far less private speech than in situations where they receivedno adult regulation. This finding seems to oppose the Vygotskian position, but is probably anartifact of the definitional decisions discussed earlier. Children who were interacting with theadults already might well choose to do their cognitive work socially, rather that privately.These studies support the hypothesis put forth by Vygotsky (1934/1968) that private speech hasits origins in social speech. The overall frequencies of social and private speech are correlated(Berk & Garvin, 1984; Kohlberg et al. 1968), and earlier forms of private speech are more highlycorrelated to social speech than later forms of private speech (Berk & Garvin, 1984). Sociallyoriented children are more likely to produce private speech (Kohlberg et al., 1968), and, whenchildren received direct regulation but were later required to work on their own, they were morelikely to produce private speech than when they have not received any regulation in the firstplace (Goudena, 1987).The Goudena findings also support the view put forth by Vygotsky (1934/1968) that childreninternalize the adult's regulation and the emphasis shifts from other-regulation to self-regulation26of thought and action. Children begin to use language to regulate their own thoughts andbehaviour in the same way that the adult's language once regulated them. This regulatory themewill be discussed in a later section. The major point here is that egocentric speech develops outof the social paradigm as the child learns to transfer behaviours learned in the social realm to thatof the inner realm.Developmental Course: Vygotsky (1934/1962) proposed that private speech followed a distinctdevelopmental path. He proposed that private speech first monitors action and then precedes it.Early forms of private speech play a descriptive role, occurring in conjunction with the child'sactions, assisting the child in monitoring behaviour. As private speech develops, it begins toprecede action and play a guiding, planning role in the child's activities. Over time it becomesless audible, and finally ceases as it becomes fully internalized, inner speech. Kohlberg et al.(1968) was particularly interested in understanding the relationships between cognitivedevelopment and private speech. The classification system they proposed consists of adevelopmental sequence of six categories: 1) word play and repetition; 2) remarks addressed tononhuman objects; 3) describing own activity; 4) questions answered by self; 5) self-guidingcomments; 6) inaudible or indecipherable muttering. Kohlberg et al. (1968) hypothesized thatprivate speech reaches the end of its developmental sequence as it becomes muttering. After thispoint it becomes internalized, increasingly mature verbal thought. Most subsequent studies inprivate speech have either used this categorization scheme or established their schemes upon it(Duncan, 1991; Berk, 1986; Berk & Garvin, 1984).27As mentioned above, the first study of Kohlberg et al. (1968) was designed to investigate theincidence of private speech and its relationship to the child's cognitive level, as well as thecorrelation between social interaction and incidence of private speech. In this study only the firstfive categories of the classification system were used because the children were followed at adistance, making it impossible to record inaudible mutterings. The results showed a largedecrease in the proportion of private speech from age four to age six. In addition, the bright andaverage 4's generated similar proportions of private speech while the bright 6's produced lowerproportions of private speech than the normal 6's. These findings are consistent with a cognitive-development explanation of private speech. Younger children use more private speech than olderchildren and cognitive maturity quickens the developmental process.The second study by Kohlberg et al. (1968) attempted to confirm the correlations between privatespeech, chronological age and intellectual maturity found in the first study. A subsample wasalso followed up one year after first testing to investigate longitudinal trends. One hundredtwelve subjects between the ages of four and ten participated in this study. Twenty-six subjectsparticipated in the follow-up session. As in the first study, intelligence groups were determinedby Stanford-Binet performance, with bright subjects having an average IQ of 131, and normalsubjects having an average IQ of 106. Each child participated in two sessions, one with a maleexperimenter and one with a female experimenter. The children participated in parallel play withthe experimenter, making sticker designs. The experimenter made a check in one of fourcategories for each sentence-like verbal remark the child made. The categories consisted of:28egocentric speech; statement of information; question; and request for help/approval. Egocentricspeech was further categorized using the six point coding system noted above.The results from the second study differed slightly from those of the first study. In the secondstudy the bright children produced more private speech than the normal children at the early age,whereas in the first study both young groups produced similar amounts of private speech. Thisis likely due to the inclusion of inaudible mutterings under egocentric speech. If private speechdoes follow a developmental pattern and inaudible muttering is the most mature form of privatespeech, then the inclusion of category six in the classification scheme would be expected to alterthe results in just such a manner. The bright younger children would be expected to producemore advanced levels of private speech than the normal younger children, and the inclusion ofsuch speech in the data collection would increase the amount of private speech produced by thebrighter children. Long term results indicated a curvilinear trend, with private speech decliningaround age six to seven. The age decline in the bright group occurred at an earlier chronologicalage, but similar mental age, presenting even clearer support for a cognitive-developmentinterpretation of private speech.The third study by Kohlberg at al. (1968) examined the feasibility of the proposed developmentalclassification system listed above. This category system was applied to the data gathered duringthe follow up portion of study 2. Inaudible mutterings, the presumed most mature category ofprivate speech, increased regularly with age. Mirror image trends between self-guiding speech(Category 5) and inaudible mutterings (Category 6) suggest that mutterings come to take the29place of self-guiding speech. The infrequency of self-answered questions (Category 4) at any agecasts doubt on the developmental significance of this category. Describing own activity(Category 3) declined steadily with age. Repetition and talking to inanimate objects (Categories1 and 2 respectively) had very low occurrences, suggesting that children in this age range hadprogressed beyond these immature levels of private speech. The results of this study indicate thatthe classification system put forth by the researchers (with the possible exception of Category 4)captures the developmental pattern of private speech.Similar confirmation of Vygotsky's developmental account can be found in these more recentstudies by other research teams. Berk (1986) investigated the relationship between the privatespeech of elementary school children and their behaviourial accompaniment to task, attention andtask performance. Seventy-five children in the first and third grades were observed during dailymath seatwork periods. Their private speech, motor accompaniment to task and level of attentionwere recorded by four observers. Each child was observed for four periods (from the beginningto the end of one assignment), once by each observer. Private speech and behaviourialaccompaniment were only recorded when: 1) the children were at least moderately attentive tothe task; and 2) children were working independently, i.e. were not seeking help from the teacheror being offered assistance by the teacher. The classification system for the private speech wasreduced to three categories: 1) self-stimulating, task-irrelevant (word-play/repetition; task-irrelevant affect; comments to nonhuman others); 2) task-relevant, externalized (describing ownactivity/self-guiding comments; self-answered questions; reading aloud; task-relevant affect); and3) task-relevant, external manifestations of inner speech (inaudible muttering involving clear30mouthing of the words; lip and tongue movement only (with no clear mouthing of words).Behaviour was categorized as: 1) tension-reducing behaviour, e.g. fidgeting; 2) task-facilitatingbehaviour; and 3) no movement. Attention level was also categorized on a three point scale:1) focused; 2) moderately focused; and 3) diverted.The main findings of Berk (1986) were: task irrelevant private speech was positively correlatedwith fidgeting and negatively correlated with no movement; externalized, task relevant privatespeech was positively correlated with task-facilitating motor behaviour and negatively correlatedwith no movement; and task relevant externally manifested inner speech was positively correlatedwith no movement and negatively correlated with tension-reducing, task irrelevant behaviour.These correlations indicate that private speech is related to functionally similar forms ofbehaviour. The most immature level of private speech is positively correlated with task irrelevantbehaviours and negatively correlated with no movement, implying that at this level the speechis occurring with, but not yet imposing any regulation on, the behaviour. The intermediary levelof private speech occurs in conjunction with task relevant behaviour but is negatively correlatedto no movement, suggesting that private speech at this level is regulating behaviour. At the mostmature level of private speech, externalized inner speech, no movement is positively correlatedwhile task irrelevant movement is negatively correlated. By this level private speech has becomeinternalized and is regulating thought but is no longer associated with behaviour. The child isable to abstract the word from the action. This pattern of private speech and behaviourcorrelations also support the developmental pattern proposed by Vygotsky (1934/1968).31Berk and Garvin (1990) performed a longitudinal study based on the work of Berk (1986) above.Thirty-three children (grades one, two and three) were observed over a three year period. Thechildren were observed only when they worked on assignments requiring individualized practicein arithmetic concepts and skills. Each child was observed for four entire seatwork periods (frombeginning to completion of an assignment or until the math seatwork session for the whole classended, whichever occurred first). Observations were randomly assigned to observers with eachchild's four periods being assigned to a different observer. The same classification system forprivate speech, behaviour and attentional level as described in Berk (1986) was used in thisstudy.Although the overall incidence of private speech did not change with age, it did undergolongitudinal age-related changes. Task irrelevant (Level 1) private speech, and task relevantexternalized (Level 2) private speech decreased linearly with age, while task relevant internalized(Level 3) private speech increased linearly. The maturation process is more evident whensubtypes within the three levels of private speech are considered. Decreases in the amount ofthe Level 1 subtype of "wordplay/repetition" appeared responsible for linear decrease in Level1 speech over time. The other subtypes of "expression" and "comments to nonhuman objects"were low in frequency and showed no developmental changes. All Level 2 subtypes, except"self-answered questions" decreased significantly across all ages. The most frequent Level 2subtype "describing own activity/self-guiding comments" decrease sharply between first andsecond grade at the same time Level 3 subtype "inaudible muttering" showed a rapid increase,nearly doubling in frequency. "Inaudible mutterings" then dropped off between second and third32grade as Level 3 subtype "lip and tongue movement" increased dramatically. As the childmatures task relevant internalized private speech not only becomes the most dominant but alsothe most consistent type of private speech.Finally, Behrend et al. (1989) evaluated the effects of age, task difficulty and parent presence.Twenty-four children at each of three age levels (2, 3 1/2, and 5 years) participated in the study(seventy-two children in all). An easy, medium and difficult puzzle was presented to each childin each of two conditions: mother present, mother absent. Different puzzles were used for eachcondition. The difficulty level of the puzzles was matched to the child's age, to ensure a roughlyequal range of difficulty for each age group. The effects of task difficulty and parent presencewill be considered later in the chapter. The number of self-directed utterances increased with ageand older children were more likely to produce self-directing utterances than younger children,supporting a developmental view of private speech.Each of these studies supports some aspect of the cognitive development interpretation of privatespeech. Private speech shows a curvilinear development with very young children (2- to 3-year-olds) producing less private speech than young children (4- to 5-year-olds) who generate moreprivate speech older children (six years plus) (Kohlberg et al., 1968; Behrend et al., 1989). Moreintellectually mature children use more mature forms of private speech earlier than averagechildren and the curvilinear trend with private speech declining around age six or seven isassociated more strongly with mental age than chronological age (Kohlberg et al., 1968). Privatespeech also appears to be related to functionally similar forms of behaviour. More mature forms33of private speech are positively correlated with more mature forms of behaviour and negativelycorrelated with more immature forms of behaviour (Berk, 1986). The inverse relationshipbetween self-guiding utterances and inaudible mutterings implies that external manifestations ofinner speech replace task-relevant self-guiding externalized forms of private speech withincreasing age. Private speech, therefore, becomes increasingly internalized and task relevantwith age. Intercorrelations among the private speech levels and categories both within and acrossage groups provides confirmation that private speech goes underground to become inner speech(Bivens & Berk, 1990; Kohlberg et al., 1968). Private speech appears to follow a distinctdevelopmental pattern culminating in internal verbal thought. This pattern appears to be similarto the one suggested by Kohlberg et al. (1968), building on the observations and theories ofVygotsky.Cognitive Self-guiding Function: As private speech develops it plays an increasingly importantrole in orienting and guiding children's behaviour. Children use private speech to plan theiractions, monitor their behaviour and delineate the problem space. The fourth study by Kohlberget al. (1968) explored the self-guiding nature of private speech as well as the influence thatculture and sex have on its incidence. Thirty-four 4- to 5-year-olds participated in this study.Each child was given four tasks in the following order, from easiest to most difficult: 1) beadstringing; 2) easy jigsaw puzzle; 3) building a tower with fifteen 1-inch cubes; and 4) hard jigsawpuzzle. Speech was coded for private and social and all private speech was coded using theKohlberg categorization scheme mentioned above. No cultural or sex differences were evidentin the findings. The amount of cognitive activity required for solving a task, i.e. task difficulty,34had a significant influence on the amount of private speech generated by the task. Children weremore likely to use private speech to talk themselves through difficult tasks than easy ones.In the Goodman (1981) study described above, faster rates of action and shorter completion timeson the puzzles were associated with a higher rate of verbalizations. The children who were mostsuccessful at the task also produced the most private speech. The children produced speech whenthey were successful in placing puzzle pieces, but they were more likely to verbalize when theyfailed in an attempt to place a puzzle piece. This suggests that the children were usingverbalizations in an attempt to overcome difficulty. These results indicate that private speechdoes more than simply accompany children's actions. Children use it as a strategy to overcomeobstacles.Duncan (1991) examined the use of private speech by preschool children over two sessions.Thirty-two preschoolers were video-taped during two sessions, while arranging sets of picturesto make brief stories. During the first session, half the children were given collaborativeassistance by the experimenter and half were not. During the second session all children workedon the tasks alone. During the first session they worked on easy and difficult story-sequencingtasks. During the second session, they worked on familiar and novel items. The familiar taskswere tasks that had been presented during the first session whereas the novel items had not.Sessions were approximately one week apart. The children produced more private speech whenworking on the difficult tasks than on the easy ones but produced similar amounts on familiarand novel tasks.35As in the Goodman (1981) study, the children were more likely to verbalize when theyexperienced difficulty, suggesting that private speech was used to perform a self-guiding function.Behrend et al. (1989) also found that private speech occurred most frequently when childrenworked on puzzles that were appropriate for, or slightly above, their general level of cognitiveability. In each of these three studies children actively used language to guide themselvesthrough difficult tasks.The findings of Berk & Garvin (1984), described above, also support the hypothesis that theprimary function of the private speech is one of self-guidance. Focused attention increased asmore internalized levels of private speech were exhibited, indicating that attention comesincreasingly under the control of private speech. The most frequently occurring categories ofprivate speech utterances were describing one's own activity/self-guidance and inaudiblemuttering. The children in this study were more likely to use private speech for guiding andorienting themselves to the task than for any other purpose.Recall that Berk (1986), described above, found that: 1) task irrelevant private speech waspositively correlated with fidgeting and negatively correlated to no movement, and 2) negativelycorrelated to no movement; and 3) task relevant externally manifested inner speech was positivelycorrelated with no movement and negatively correlated with task-facilitating behaviour. Thesecorrelations indicate that private speech is related to functionally similar forms of behaviour. Self-guiding speech occurs in conjunction with appropriate behaviour. For children who displayedlarge amounts of off-task behaviour, greater use of externalized task-relevant private speech was36related to lower levels of inattentiveness. This indicates that distractible children use privatespeech in an attempt to control and focus attention. This is in keeping with the interpretation thatchildren use private speech to focus and guide behaviour.Finally, Berk and Garvin (1990) described above, found similar correlations between privatespeech and behaviour to Berk (1986). As Level 1 (task irrelevant) speech declined, so didtension reducing, task irrelevant behaviour (e.g. scratching, chewing). As Level 2 (task relevant,externalized) speech diminished, task facilitating movements (e.g. pointing, gesturing) alsodeclined. As children used more Level 3 (internalized) private speech, focused attention alsoincreased. Private speech appears to facilitate, or at least occur in conjunction with, increasingself-control over behaviour and attentional focus.The above studies provide clear evidence of the self-guiding nature of private speech. Childrenconsistently produced more private speech on difficult tasks than easy ones (Duncan, 1991),particularly if the task was at or slightly above their general cognitive level (Behrend et al.,1989). They were also more likely to verbalize when experiencing difficulty on a task(Goodman, 1981). Private speech was not only correlated to the ability to focus attention (Berk& Garvin, 1984; Berk, 1986; Berk & Bivens, 1990), but was related to functionally similar formsof behaviour (Berk, 1986; Bivens & Berk, 1990). Children used similar types of behaviour andprivate speech simultaneously. As the cognitive level of the children's private speech increasedso did their ability to maintain focus on the task (Berk, 1986; Bivens & Berk, 1990). Thesefindings support the view put forth by Vygotsky (1934/1968) that children use private speech to37guide and regulate behaviour and to focus attention on the task at hand. Children use privatespeech to guide their actions and thoughts as they work through problems and difficultconditions.In summary then, current research provides confirmation of the three main hypotheses postulatedby Vygotsky (1934/1968) concerning private speech. Private speech is rooted in socialinteraction and is used as a tool for guiding behaviour and focusing attention. As childrenmature, the other-directed regulation provided by social interaction with adults is replaced withthe self-regulation of private speech. As children become more proficient at using private speechfor self-regulation, it becomes progressively more task-relevant and internalized until finallyevolving into fully internalized verbal thought.Personality Characteristics and Private Speech Thus far I have looked at the literature on private speech as it relates to the Vygotskian view ofdevelopment. There is one further aspect of modern literature on private speech, however, thatis not developmental. Since there is a high degree of interpersonal variation in the amount ofprivate speech produced by children (Behrend et al., 1989; Frauenglass & Diaz, 1985; Kohlberget al., 1968; Fuson, 1979), it is possible that this variation is correlated to certain personalitycharacteristics. Language use is very individual. Some people are highly verbal while others arenaturally taciturn. Personality may influence the manner in which individuals incorporatelanguage into daily routines and problem-solving activities. Almost no research has been doneon the influence that intrapersonal characteristics have on the production of private speech. What38research does exist, discusses three characteristics that are believed to be correlated with the useof private speech: dependency; impulsivity; and efficiency. Dickie (1973) found that impulsivechildren appear to use lower levels of private speech (as cited in Fuson, 1979). Impulsivechildren were more likely to use task-irrelevant and outward directed private speech thanreflective children. Klieman (1974) found that independent children were more likely use privatespeech than dependent children (as cited in Meichenbaum & Goodman, 1979). Goodman (1981)found that efficient children produced a higher rate of private speech. Goudena (1987) performedmeasures of impulsivity and dependence for each of his subjects because he hypothesized thatthese factors were related to private speech production, however, he failed to report on thepresence or absence of any correlations between these two characteristics and the speechproduced by the children. While the evidence is sparse supporting of the relationship betweenpersonality characteristics, e.g. dependency, impulsivity and efficiency, and private speech, thesestudies provide some confirmation that private speech in general, and certain types of privatespeech in particular, are more likely to be produced by children displaying certain personalitycharacteristics than others.In summary, then, private speech is one visible example of the way in which language servesthought. The empirical data and theory provide insight into the development of private speechand therefore the development of cognition. As such, private speech provides a key tounderstanding the connection between thought and language and the influence that each has onthe other. As the precursor to inner speech it also provides information on children's cognitivedevelopment, because advances in cognitive ability will be reflected in the private speech39produced. There are two gaps in the current literature, however. Little information exists on therole of social speech in cognitive development and the influence of personality on language use.These two areas are important in understanding the larger picture, i.e. the association betweenlanguage and thought. An awareness of the role these two factors play in the relationshipbetween language and thought is necessary for a full comprehension of the intellectual uses oflanguage.The Effects of Specific Language Impairment (SLI) on Cognition I come now to the final motivation for this study - the language impaired child. Although it wasonce believed that specific language impaired children had normal cognitive functions, researchover the last decade has indicated that these children display widespread cognitive delays(Johnston, 1988, 1991; Thal, 1991). They are more likely to display chronic learning problemsthan normal children and exhibit cognitive dysfunctions over a wide range of tasks (Johnston,1988, 1991). There are three possible explanations for the cognitive delay exhibited by SLIchildren: a) the cognitive impairment is a result of a cognitive processing problem that underliesthe language impairment; (b) it is a direct consequence of the language impairment; or (c) bothare factors. The work on private speech suggests that explanations (b) or (c) are most likely.Since private speech appears to play a role in cognitive development, a language impairment willpresumably affect a child's ability to effectively use language as a cognitive self-guiding strategy.While there is no published work on the use of self-regulating speech by young children withspecific language impairment, it seems possible that these children have not learned to use theregulatory, planning functions of language. Interesting confirmation of this fact comes from40research by Harris (1990; 1986), which indicates that the explicit teaching of private speech mayhelp learning disabled children learn cognitive self-guidance.The Current Study The purpose of this study, then, is three-fold. First, it was designed to investigate the intellectualrole of social as well as private speech. By categorizing both elicited social speech and privatespeech using a cognitively based taxonomy it may be possible to determine whether both socialspeech and private speech play similar roles in cognitive development, with some children usingone over the other as their preferred learning strategy. Second, I was interested in therelationship between personality characteristics and language use. The literature suggests thatchildren who display certain personality characteristics appear to be more likely to use privatespeech. Three such characteristics are independence, efficiency and impulsivity. If a relationshipcan be revealed between personality characteristics and the tendency to use private speech oversocial speech as a learning strategy, then it may be possible to identify children who wouldbenefit most from explicit teaching of self-guiding strategies. Finally, I felt it was important toinvestigate the effect of SLI on cognitive uses of language. To this end the study also includeda preliminary investigation of the use of self-regulatory speech by children with a specificlanguage impairment.In conclusion, the five questions asked in this study are: 1) What are the problem solvingfunctions of speech? (2) Do private speech and social speech play similar roles in problem-solving? (3) Do some children show a preference for private speech or social speech as their41main problem-solving strategy? (4) What correlation exists between language use and personalitycharacteristics? and (5) Do children with a specific language impairment differ significantly fromnormal children in their use of self-guiding speech?42CHAPTER TWOMETHODOverviewThis study was designed to investigate questions concerning the intellectual uses of language bynormal and specific language impaired preschoolers, in particular: 1) what role does speech playin problem-solving; (2) do private and social speech play similar roles; (3) do children with aspecific language impairment use self-regulatory language in a way that is fundamentallydifferent from normal children; (4) do some children show a preference for private speech orsocial speech as their main problem-solving strategy and if so; (5) what correlations existbetween this preference and the child's level of independence, efficiency, and cognitive style.Preschool children were audio- and video-taped while participating in a fantasy-play sessioninvolving a building task (the bridge task). The audio- and video-tapes of each session weretranscribed and the children's utterances were coded for task relevance (irrelevant, broadlyrelevant, narrowly relevant). Those utterances pertaining to the task were further coded forfunction (regulative, affective, word play), addressee (private, social) and level of cognitivecomplexity (Blank, 1974).SubjectsEighteen preschool children participated in the study: six 3-year-olds (YNL) and six 4-year-olds(ONL) with normal language development and six 4-year-olds (SLI) with delayed languagedevelopment. All spoke English as their first language and scored within +1 Standard Deviations43on the Leiter International Performance Scale (LIPS), indicating age-appropriate cognitivedevelopment. The language impaired children were recruited from a pool of children receivingservices from the Speech-Language Pathologists associated with the Central Fraser Valley HealthUnit. The children with age appropriate language abilities were recruited through a daycare inNew Westminster and a preschool in Vancouver. All the language impaired children and twoof the 4-year-olds with age-appropriate language were at home with their mothers. Theremainder of the children were in daily child care.Prior to inclusion in the study, the children underwent a screening assessment of cognitive andlanguage skills. Children were selected to participate on the basis of:1. Cognitive Ability: Children performing within ±1.00 Standard Deviations of the mean on theLeiter International Performance Scale  (LIPS) (Leiter, 1969) were considered as displayingage-appropriate cognitive development. The LIPS is a non-verbal test requiring judgementsof perceptual and conceptual similarity. Only children with age-appropriate cognitivedevelopment were considered for the study.2. Language Ability: Children were grouped according to their expressive and receptivelanguage skills. The language portion of the screening assessment was primarily based on theDevelopmental Sentence Scoring (DSS) (Lee, 1974), performed on a spontaneous sampleaudio-taped during a 15 minute free play session. Three secondary language measures werealso used to corroborate the DSS findings. These included: (1) the Expressive One WordPicture Vocabulary Test-Revised (EOWPVT-R) (Gardner, 1979); (2) the Peabody Picture Vocabulary Test (PPVT) (Dunn & Dunn, 1981); and (3) the Test of Auditory Comprehension 44of Language-Revised, subtests 2 and 3 (TACL-R) (Carrow-Woolfolk, 1985). See AppendixA for a brief description of these measures.Children performing more than 1.5 standard deviations below the mean on the DSS and atleast two of the other three language measures were considered as displaying a significantlanguage delay. Children within + 1.0 standard deviations of the mean on the DSS and atleast two of the other three language measures were considered as having age-appropriatelanguage development. The 3-year-old age-appropriate children were matched as a group tothe 4-year-old delayed children on the basis of the DSS score obtained during the screeningassessment. There was no statistically significant difference between the mean of the DSSscores of the 4-year-old delayed children and that of the 3-year-old age-appropriate children.Children with normal language development were solicited through a daycare centre andpreschool in the Greater Vancouver Area. Fifty letters and consent forms were sent to theparents of children attending the daycare and preschool, asking if they would consent toparticipate along with their child in the study. See Appendix B for a copy of the letter andconsent form sent to parents. Fifteen responses were received and these fifteen children werescreened. One child was eliminated from the study because her score on the LIPS was more than1.00 standard deviation above the mean. Two more children were dropped from the studybecause they stopped attending the daycare they were canvassed through. The other twelvechildren met the screening requirements for children with normal language development and wereincluded in the study.45The selection process for the SLI children began with the director of Speech Services at theCFVHU and the experimenter reading the files of each child receiving speech-language servicesthrough the CFVHU (approximately thirty-five children). Fifteen children were eliminated fromthe potential subject list because they exhibited global or cognitive delays. Of the remainingtwenty children, the twelve children exhibiting the most significant language delays were chosenas potential subjects for the study. A letter and consent form were sent from the CFVHU to theparents of each child, asking if they would consent to participate with their child in the study.See Appendix B for a copy of the letter and consent form. Five parents, one with two SLIchildren, responded to the letter. These six children met the screening requirements for SLIchildren and were included in the study.A summary of the screening results can be found in Table 1.46TABLE 1: SUBJECT SELECTION DATAGROUP SUB # CA^LIPS^DSS DSS%ileGRP 1SLI 1 61 110 6.7 14SLI 2 58 110 6.3 10SLI 3 56 101 4.4 07SLI 4 57 115 4.2 07SLI 5 57 115 4.9 08SLI 6 60 110 5.7 08MEAN 58.2 110 5.4 9.0GRP 2ONL 1 57 110 8.1 48ONL 2 59 116 8.8 55ONL 3 51 116 7.1 37ONL 4 58 110 7.4 29ONL 5 51 101 8.3 60ONL 6 54 94 7.8 45MEAN 55.0 108 7.9 45.7GRP 3YNL 1 42 91 6.2 30YNL 2 38 108 4.6 25YNL 3 39 105 5.6 40YNL 4 44 107 6.4 27YNL 5 42 105 7.4 80YNL 6 45 100 6.4 27MEAN 41.7 103 6.1 38.2A one-way ANOVA, using the post-hoc Tukey Studentized Range Method test option, wasperformed on each of the selection variables to see if there were significant differences amongthe groups. The results are given below in Table 2.47TABLE 2: SUBJECT SELECTION DATA COMPARISONCHRONOLOGICAL AGE (CA) SLI ONL YNLSLI^**ONL **LIPSSLIONLDSS SLIONL SLI ONL YNLSLI ONL YNL****DSS PERCENTILESLI ONL YNLSLI^**^**ONL** denotes significance at p = 0.01* denotes significance at p = 0.1The mean chronological ages of the SLI and ONL children are not significantly different;however, the mean chronological age of the YNL children is significantly lower than that of theSLI and the ONL children. There are no significant differences between the mean LIPS scoresof the three groups. The mean DSS score of the SLI group is significantly lower than the mean48DSS score of the ONL group but is not significantly different from that of the YNL group. Themean DSS score of the ONL group is also significantly higher than the mean DSS score of theYNL group. The mean DSS percentile of the SLI group is significantly lower than the meanDSS percentile of both the YNL and the ONL groups. The mean DSS percentiles of the YNLand ONL groups are not significantly different.ProceduresEach subject participated in three to four sessions, including two or three assessment sessions andone experimental session. All sessions were conducted in surroundings familiar to the child(child's home, daycare/preschool centre) by the same female experimenter. During theassessment sessions the child and experimenter were alone in a quiet room. The experimentalsessions were conducted with two children who were familiar with each other and, on fiveoccasions, a mother of one child was also present. In each experimental session at least onechild of the dyad displayed normal language development.Assessment ProceduresThe assessment was conducted over two or three 45 minute sessions. When two sessions wererequired, the LIPS and the PPVT were administered during the first session and subtests two andthree of the TACL-R, the EOWPVT and a 15 minute free play period completed the secondsession. If three sessions were required to complete the assessment, the LIPS was administered49during the first session, the PPVT and subtests two and three of the TACL-R were administeredin the second session, and the EOWPVT and a free play period completed the third session. Thefree play period was audio-taped, the child's first 50 utterances were transcribed and the DSS wasperformed on those utterances.Standardized TestsIn addition to the cognitive and language tests described earlier, each child was given the KansasReflective-Impulsivity Scale for Preschoolers (KRISP) (Wright et al., 1978). This test measuresthe cognitive style (impulsive/reflective) and efficiency of the child. If two sessions wererequired to complete the assessment the KRISP was administered during the first session. Ifthree sessions were required, the KRISP was administered during the second session. SeeAppendix C for a brief description of this test.The Beller Questionnaire (Beller, 1957) was also given to a parent of each child. The childrenwere not matched on these variables but both measures were taken into account during the dataanalysis. See Appendix B for an example of the questionnaire.Experimental TaskMaterials: At the beginning of the experimental session a 70" X 32" piece of cardboard with acoloured diagram of a river valley was placed on the floor in the middle of a large open area.The diagram consisted of a winding river running the length of the cardboard with trees and twohouses drawn kitty-corner to each other.50FIG,A pile of building supplies was located on the floor beside each house. Each pile consisted offifty blocks of assorted sizes, shapes, and colours, fifteen flexible straws, ten - 12-inchpipecleaners, five pieces of coloured string approximately 12 inches in length, a 3-inch diameterball of playdough, and twenty popsicle sticks. Also included in each pile were three pictures ofbridges made by the experimenter out of the above listed supplies.A video camera and tripod were used to video-tape each session. Two Samson remote lapelmicrophones, one for each child, and a two-channel Marantz tape recorder, Model No. PMD430,were used to audio-tape the sessions. The lapel microphones were placed in custom-made vestswhich the children wore during the task. The audio and visual equipment was set upapproximately 10 - 15 feet from the river valley diagram.Procedures: The children played in pairs. The diagram of the river valley was placed on the floorand two sets of building supplies were piled on the floor on opposite sides of the diagram. Each51child was asked to sit down by a pile of building supplies. The experimenter then placed a 6-inch clay monster in the middle of the river and told the children the following story:You (pointing to one child) live on this side of the river (pointing to the house the child issitting beside) and you (pointing to the other child) live on this side of the river (pointing tothe house the child is sitting beside). A monster lives in the river (picking up the monsterand putting him back down again). He likes to eat little children. You need to build a bridgeto get across the river so that you can play with each other. Here are some picture of bridgesthat other children have built (picking up the pictures in each of the piles). You have straws,and string, and blocks, and playdough, and popsicle sticks, and pipe-cleaners to build thebridge (picking up an example of each from each child's pile as they are listed). You buildthe biggest, best bridge that you can to get over the river so the monster doesn't eat you.Use as many of these (indicating the supplies) as you can. I'm going to sit over here(indicating a location near the audio-visual recording equipment, approximately 10 - 15 feetaway) while you build. Both of you build whatever you can to get across the river.The experimenter then withdrew to the indicated location where she could observe the childrenand the audio-visual equipment but be apart from the task. If the children became involved ina task that was broadly related to the bridge building task, i.e. building a boat or a trap for themonster, they were permitted to continue uninterrupted. If the children were off-task for longerthan two minutes, completed part of the task and were unsure of how to proceed, or announcedthat the task was complete before 30 minutes had expired, the experimenter would prompt them52to continue with the assigned task. Sessions lasted 30 minutes or until the children lost interest(minimum = 15 minutes).The task was called to a halt after 30 minutes or when the children indicated that they no longerwished to continue playing. This was signified by direct verbal comments such as "I don't wantto play any more" or an unwillingness to stay on task regardless of prompting by theexperimenter. The minimum time period for the experimental session was 15 minutes.Experimenter Conduct: When the children went off task, were unsure how to proceed with thetask, or indicated that they had completed the task, the experimenter prompted them on withcomments such as "what else can you build?", "can you build a bigger bridge that the monstercan go under?", or ""could you build a trap for the monster?". If the children questioned theexperimenter about the supplies or asked for clarification/repetition of the task she wouldrepeat/clarify the information as requested. Other comments and questions about the task wereresponded to with non-specific comments such as "you put it wherever you want", "that's prettygood", "see what you can make", "looks good", "wow", "mm-hmm", etc. If the childrenpersisted in requesting a more precise answer they were told "this is your job to work on. I'mnot going to help you. I have my own work to do over here." Questions and comments thatwere irrelevant to the task were answered as briefly as possible and, if necessary, the child'sattention was directed back to the task at hand. If the child's attention wandered from the taskor the child completed a task and was unsure of how to proceed, the experimenter would prompt53the child to " use as many of the supplies as you can", "see what else you can build using thosethings" or " can you build a bigger bridge/trap/boat/etc.?"Transcription and Coding of Language DataThe audio- and video-tapes from each experimental session were transcribed and the children'sutterances were evaluated for their connection to the problem-solving task according to ataxonomy of Problem Solving Utterances, derived from the work of Blank and Franklin (1980).With two exceptions, single word utterances were not coded. Single wh-question words werecoded as questions and "okay" was coded if it was not being used as a synonym for "yes". Allother utterances were coded for task-relevancy (task irrelevant, broadly relevant, or narrowlyrelevant, unknown). Utterances that were broadly or narrowly task-relevant were further codedfor function (affective, regulative, word play), addressee (private, social) and cognitivecomplexity. Further details of these coding schemes are given below.Coding SchemeRelevance: Utterances were coded for one of three levels of relevancy: narrowly task relevant,broadly task relevant and task irrelevant. In a few instances it was impossible to determine themeaning of the child's utterance and these were coded as relevancy unknown. Those utterancesthat were totally unrelated to the bridge task were coded as task-irrelevant. This includedcomments about family and friends - "my Mummy's at school"; toys at home - "my baby'sall examples are taken from the transcripts54pretend and her name's Sissy"; or comments about events other than those pertinent to the taskat hand - "we don't want bugs to live in your hair, yuck".Utterances that were directly relevant to solving the bridge building problem were coded asnarrowly relevant. This included questions and comments about what to build - "hey, I think I'mgonna build a trap for it"; how to build it - "hey, how you build that?"; why it was being built -"then the monster can't come out"; and other problem-solving types of utterances - "and this ishow I'll stop it".All other utterances were coded as broadly relevant. These included fantasy play utterances -"he's karate and then the karate man get that monster out and then he kick and they kick";drawing attention to the structure being built - "see what I made?"; general comments about thebuilding supplies - "how do you get this mixed up playdough?" (referring to the multi-colouredplaydough); and other utterances generally related to the bridge task - "I'm bigger than themonster is 'cause he's little". Only utterances coded as narrowly or broadly task relevant weresubsequently coded for function, addressee, and cognitive complexity.Function: All task relevant utterances were subsequently coded for the general sort of functionthey fulfilled: regulating, affective, or word play. Utterances that primarily served to expressemotions were coded as affective. These include positive or negative evaluations - "mine almostdone" and "can't do that one"; fantasy affect - "that monster's gonna be really scary"; emotionalexclamations - "oopseys"; or exclamations such as "ahhh" when a structure crumbles. Word play,55nonsense words, singing, and single words and nonsense words accompanying actions were codedas word play. This includes utterances like "log wog hog, I got log". Action accompanimentincluded such things as bouncing the monster down the river and saying "hop, hop, hop".Singing included a range of from "do do do do do" to singing the Batman theme song whenattacking the monster.Utterances that were coded as regulatory in nature included statements about action - "I'm gonnaput everything away"; statements about states of affairs - "can't reach over here" or "I don't haveany more bricks"; commands (including requests for attention, clarification, action) - "look", "tellus the story" (about the monster); "don't put it there"; questions - "what these are?"; and backchannel responses - "let's see". In a very few cases, it was not possible to determine whatfunction the utterance served and these were coded function unknown.Addressee: All task relevant utterances were coded for whom they were addressed to: self(private) or other (social). Typically, when speech is social, i.e. is addressed to another, eyecontact is at least briefly established and/or the speaker's body is oriented toward the listener.When eye contact is difficult or inconvenient, speakers tend to compensate and mark the speechas intended for a listener by: (1) making explicit verbal reference to the listener; 2) repeatingtheir remark until a response is received; 3) demanding a response; 4) initiating physicalproximity/contact; or 5) increasing volume (Feigenbaum, 1989). In this study, task relevantspeech was be judged to be social if it met any one of these criteria. Speech was judged to beprivate if it: 1) was subvocal; 2) was reduced in volume; or 3) failed to meet the criteria for56social speech. On occasion it was difficult to determine if an utterance was social or private innature. These utterances were coded as addressee unknown.Cognitive Complexity: The fourth level of coding performed on the transcripts evaluated theutterances for the cognitive complexity of the material being expressed, based on Blank (1974).Cognitive complexity was sometimes judged over a span of utterances when the utterances wereconceptually linked, e.g. "I'm gonna build the water" "so he can't get over". Level 1 utterancesmatched or analyzed the here-and-now experience. These included utterances that notedattributes - "long blocks"; locations - "right over there"; current actions - "me put that on"; ordeficiencies in the present situation - "he (the block) can't fit".Level 2 utterances were those which restructured the here-and-now, including fantasy creations -"I'm making a blast off castle"; similes - "like a high jump slope" , generalizations - "nomonsters can go through there"; simple statements about possible states of affairs - "don't putthat there or it will break" and plans of action - "now I'm gonna make a tunnel". Utteranceswhich expressed reasoning about the restructured reality were classified as Level 3 utterances.These included conditional relationships - "I need one of those to stop the truck goin' fallin'back" and explanations - " we put this" "so the monster can't get out". Those utterances thatwere incomprehensible were coded as level unknown.57Inter-Rater ReliabilityA second rater, trained on the coding criterion by the experimenter, independently coded 18%of the total utterances, with approximately equal numbers of utterances being sampled from eachchild. Reliability measures were calculated for each level of coding as well as for transcriptionof the utterances. A 98% agreement was reached on the transcription of the utterances. Onclassification of Relevancy, a 93% inter-rater agreement was reached. Ninety-nine percentagreement was reached on Function, and 89% on Addressee. Inter-rater reliability was 95% forlevel of Cognitive Complexity.Table 3 summarizes all coding categories and criteria.58TABLE 3: CODING CATEGORIES FOR PROBLEM SOLVING UTTERANCES Level 1 Matching/analyzing here and now experience.Includes: noting attributes, locations, current actions; noting deficiencies in present situation.Level 2 Restructuring the here and now.Includes: fantasy creations, similes generalizations, simple statements about possible statesof affairs and plans of action.Level 3 Reasoning about the restructured reality.Includes: conditional relationships, explanationsRegulating: occurs at all 3 levels of cognitive complexity)includes utterances that are primarily regulating in nature include: statements about action;statements about states of affairs; commands(including requests for attention, clarification,action); questions; back channel responses, etc.Affective: occurs at all three levels of cognitive complexity: fantasy affect is level 2 complexity;explanation of affect is level 3 complexityincludes utterances that primarily serve to express emotion including: positive and negativeevaluations; emotional exclamations and fantasy affectWord Play: occurs only at level 1 of cognitive complexityincludes word play; nonsense words; singing; single words and nonsense words accompanyingactions59CHAPTER THREERESULTSThis study was designed to investigate the intellectual uses of language by normal and specificlanguage impaired preschoolers. In particular I was interested in the role that speech plays inproblem-solving. To this end I asked:1) Do young children use speech to assist in problem-solving?2) Do private speech and social speech play similar roles?3) Do children with a specific language impairment use self-regulatory languagein a way that is fundamentally different from normal children?4) Do some children show a preference for private speech or social speech as theirmain problem-solving strategy?, and5) What correlations exist between the child's level of dependence, cognitive styleor efficiency and language use in problem-solving tasks?Two data bases were set up to investigate the research questions. The first data base includedall utterances generated by each child during the bridge task, to a maximum duration of 30minutes. Because of the degree of individual variation in the transcript length (15 to 30 minutes),a second data base included only on the utterances generated during the first fifteen minutes ofthe bridge task. Since there was little difference in the patterns of results produced by the totalutterance and 15 minute data bases, all results presented in this chapter are based on the totalutterance data base.60Speech and Problem Solving The first analysis examined the functions that speech performs during problem-solving tasks.As a first step, an utterance coded as "task-irrelevant" was not coded further for function. Alltask-relevant utterances were coded for the function they fulfilled e.g. "regulating", "affective"or "word play". If an utterance fulfilled more than one function it was coded for each functionas well as coded for fulfilling a "multiple function". Also, some utterances were impossible tocode for function, e.g. confused or subvocal utterances, and these were coded as "unclearfunction".Table 4 lists the group means for the total utterances and the proportion of utterances in each ofthe above categories. Although all the children produced utterances that were irrelevant to thetask, the majority of their speech was about the task at hand (83.8% - 91.5%). A relativelysmall proportion of utterances produced were irrelevant to the task (8.5% - 16.3%), with theyounger children generating a larger proportion of off topic speech than the two older groups.Only a small proportion of utterances fulfilled the functions of "affective" (8.6% - 9.5%) or"word play" (2.0% - 2.3%). The majority of utterances produced by all three groups wereregulatory in nature (76.6% - 84.2%).It is clear that speech plays a large regulatory role during problem-solving tasks. When presentedwith a problem to overcome, children will use language to discuss and consider solutions for thatproblem. The language may play a variety of functions but it is overwhelmingly centred aroundthe task. At this level of analysis, only normal developmental differences were evident. The61younger children were more likely to go off task and produce less regulatory language than theolder children. Differences between the SLI and NL children will be considered in more detaillater in this chapter.TABLE 4: THE FUNCTIONS OF SPEECH IN PROBLEM-SOLVING TASKS SLI ONL YNL MEANTOTAL UTTERANCES 155.3 186.3 136.7 159.4% TASK IRRELEVANT UllERANCES 8.5 9.4 16.3 11.4% REGULATING UT1ERANCES 84.2 81.2 76.6 80.7% AFFECTIVE UTTERANCES 8.6 9.4 9.5 9.2% WORD PLAY UTTERANCES 2.4 4.6 3.5 3.5% UNCLEAR FUNCTION UTTERANCES 2.3 2.0 2.2 2.2% MULTIPLE FUNCTION 6.0 5.8 8.0 6.6The Roles of Social and Private Speech in Problem Solving The second set of analyses investigated whether private speech and social speech play similarroles in problem-solving activities. I was interested to see if children preferred to use privatespeech for guiding their thinking on the bridge problem or whether they were just as likely touse social speech. To this end the proportion of all task relevant private speech and social speech62used by each child was calculated as a function of the total number of utterances. The proportionof narrowly task relevant (NTR) private and social speech was similarly calculated. Table 5 liststhe group means for these variables, as well as the mean for all three groups combined. Totalutterances was used as the denominator in these variables to facilitate the eventual use ofANOVA techniques. As can be seen, the proportions of social and private speech areapproximately equal when all task relevant language is considered. When only narrowly taskrelevant language is considered, more private speech is produced than social speech. It is ofinterest, though, that there is as much social problem-solving language as there is. Even thoughchildren are more likely to use private speech when focusing narrowly on the problem at hand,they still produce considerable amounts of social speech.TABLE 5: PROPORTION OF TOTAL UTTERANCES THAT WERE TASKRELEVANT AND EITHER PRIVATE OR SOCIAL SPEECH ALL TASK RELEVANTUTTERANCESNARROWLY TASK RELEVANTUTTERANCESSOCIAL PRIVATE SOCIAL PRIVATESLI 44.7 45.5 14.8 29.0ONL 48.3 41.1 20.1 26.0YNL 44.8 35.2 12.2 19.4MEAN 45.9 40.6 15.7 24.863The reliability of the observed differences was tested with two - two-way repeated measuresANOVAS, Group (3) X Addressee (2), with percent of all utterances falling within a givenAddressee category as the dependent variable. I will report here only those effects that concernthe Addressee factor. The main effect for Addressee was not statistically significant when all taskrelevant language was considered, F = 0.97; df = 2, 1; p = 0.34. The second ANOVA consideredonly those utterances that were narrowly task relevent and regulating, again looking fordifferences in Addressee. The dependent variable was percent of all utterances falling within agiven Addressee category. In this analysis, the main effect for Addressee did prove to bestatistically significant, F = 6.81; df = 2, 1; p < .02. There were no significant Group XAddressee interactions in either ANOVA.The nature of the Addressee effect can be more clearly seen if the proportion of task relevantutterances that were either private or social is calculated, i.e. further restrict the denominator ofthe proportions. Table 6 lists group means for these variables. Children used the sameproportion of private and social task relevant speech throughout the session (53% vs. 47%), butwhen speaking specifically about the problem-solving aspects of building the bridge they usedmore private than social speech (61% vs 39%). Although neither ANOVA yielded significantGroup X Addressee interactions, the SLI children consistently produced equal or greater amountsof private speech than social speech, while the normal children produced more private speechthan social speech only when narrowly task relevant language was considered.64TABLE 6: PROPORTION OF TASK RELEVANT UTTERANCES THAT ARE PRIVATE AND SOCIAL IN NATURE ALL TASK RELEVANT LANGUAGE ^SOCIAL SP^PRIVATE SPSLI^50.0 50.0ONL^54.0^46.0YNL^56M 44MMEANS^53.0^47.0NARROWLY TASK RELEVANT LANGUAGE ^SOCIAL SP^PRIVATE SPSLI^34.0 66.0ONL^44.0^56.0YNL^39M 61.0MEANS^39.0^61.0The Effects of SLI on Regulatory Language The earlier analyses examined the function of language in problem-solving tasks across groupsof children. In this third set of analyses I was interested in whether the SLI children usedregulatory language in a manner that was different from normal children. Since I wasparticularly interested in language produced by children to guide and direct their thinking abouta problem, I focused the analyses on those utterances that were specifically about the problem-65solving aspects of building the bridge and were regulatory in nature, as opposed to affective orword play. Table 7 indicates the number and proportion of such utterances that were producedby each group.TABLE 7: NARROWLY TASK RELEVANT, REGULATING SPEECH SLI ONL YNLNO. OF REGULATING 33.3 43.4 19.3UTTERANCES% OF REGULATING 0.44 0.47 0.33UTTERANCESThe ONL children produced more narrowly task relevant, regulating utterances than the SLIchildren, who in turn produced more than the YNL children. The YNL children tired morequickly. The SLI children kept working but were less talkative. There are no differences in theproportion of language produced by the three groups, however, when only narrowly relevant,regulating speech is considered.A two-way repeated measures ANOVA, Group (3) X Addressee (2), with absolute number ofnarrowly task relevant utterances as the dependent variable was performed to test the reliabilityof the observed group differences. The main effect for Group was statistically significant, F =7.11; df = 2,1; p > 0.01. The ONL children produced more regulating utterances than the SLI66children, who produced more regulating utterances than the YNL children. When a similarANOVA was performed, using the proportion of narrowly task relevant, regulating utterancesas the dependent variable, the main effect for Group no longer proved to be statisticallysignificant, F = 2.17, df = 2,1; p = 0.15. Although the three groups produced different amountsof language during the task, they devoted the same proportion of their speech to the problem-solving aspects of the task.Interactive vs. Solitary Problem Solvers As a group the children displayed a strong trend for using private speech when problem-solving.I was interested in seeing whether this trend also held at the individual level or whether somechildren showed an inclination for social speech over private speech as their main problem-solving strategy. To answer this question, only narrowly task relevant utterances wereconsidered. Children who produced at least 30% more private NTR utterances than social NTRutterances during the task were considered solitary problem-solvers. Those who produced at least30% more social NTR utterances than private NTR utterances during the task were consideredinteractive problem-solvers. Those who produced less than a 30% difference between social andprivate utterances were considered versatile problem-solvers. Out of the eighteen subjects, threechildren (one SLI, one ONL, and one YNL) showed a marked preference for social speech whenproblem-solving. Ten children (four SLI, four YNL and two ONL) showed a strong preferencefor private speech and 5 children (one SLI, one YNL, and three ONL) showed no strongpreference between the two types of speech when problem-solving. Although there was anoverall preference among the children to use private speech over social speech as the mainlanguage strategy when problem solving (ten children vs. three children) this was not universal.67Some children appeared to prefer to problem solve in an interactive setting while others seemedto use social and private speech to the same degree when attempting to construct solutions toproblems.Personality Characteristics and Language Use Since there was a distinct difference among the children in the style of language used duringproblem-solving (private vs. social), the next logical step was to ask if there were anycorrelations between the personality characteristics and the tendency to use private or socialspeech as the main problem-solving strategy. During the assessment sessions data was collectedon three personality characteristics of the children: cognitive style (impulsivity), efficiency andlevel of dependency. The raw latency and error scores collected on the KMSP were transformedinto z-scores, based on the distribution of scores across all three groups, and the efficiency andimpulsivity scores were derived from these. Highly positive impulsivity scores indicate that thechild has an impulsive cognitive style while highly negative impulsivity scores indicate areflective cognitive style. High positive efficiency scores indicate low efficiency while highlynegative efficiency scores indicate high efficiency. Dependency scores were calculated from theparent responses on the Beller Questionnaire. The higher the dependency score the moredependent the child was judged to be. A summary of the results can be found in Table 8.TABLE 8: PERSONALITY CHARACTERISTICS GROUP SUB # IMPULSIVITY EFFICIENCY DEPENDENCYGRP 1SLI 1 -1.54 -0.52 46SLI 2 0.75 -1.05 41SLI 3 1.13 -0.11 50SLI 4 -0.53 -1.29 56SLI 5 -1.02 0.28 36SLI 6 1.07 0.39 46MEAN -0.02 -0.38 45.83GRP 2ONL 1 0.27 -0.13 46ONL 2 -0.85 -2.09 51ONL 3 -0.89 -0.73 47ONL 4 -2.32 -0.18 46ONL 5 0.69 -0.55 47ONL 6 -0.61 -1.01 46MEAN -0.62 -0.78 47.17GRP 3YNL 1 1.01 0.89 50YNL 2 -0.97 2.87 59YNL 3 3.00 -0.22 53YNL 4 -1.78 2.36 43YNL 5 3.31 1.23 39YNL 6 -0.67 -0.51 51MEAN 0.65 1.10 49.176869The personality characteristic scores of the seven children showing the most distinct differencesin the production of social and private narrowly task relevant speech (three showing a strongsocial preference and four showing the strongest private preference) were compared. There wasno evidence of a pattern between any of the characteristics and style of language preferred. Thepersonality characteristics of the four children displaying the least difference between the amountof social and private narrowly task relevant speech were also compared and again no pattern wasfound. It appears the preference for social or private speech in problem solving is unrelated tothe child's level of dependency, efficiency, or cognitive style.The final analysis looked at whether correlations exist between the personality characteristics ofefficiency, cognitive style and level of dependency and other aspects of language use. Unlessotherwise noted, the correlations were calculated across all children, regardless of groupmembership and are based on absolute numbers of utterances, not proportions. Analyses wereperformed to investigate correlations between the three personality characteristics and the use ofprivate speech, social speech, total utterances, task irrelevant utterances, narrowly task relevant(NTR) utterances, broadly task relevant (BTR) utterances, and affective utterances. Both theproportion and the absolute amount of each of these aspects were compared to the threepersonality characteristics.70TABLE 9: CORRELATIONS BETWEEN PERSONALITY CHARACTERISTICS ANDLANGUAGE USE EFFICIENCY^IMPULSIVITY^DEPENDENCYPRIVATE SPEECHSOCIAL SPEECHTOTAL U1 I E^RANCES -.53 **-.61 **TASK IRRELEVANTUTTERANCESNTR UTTERANCES^.47 **.58 **BTR UTTERANCESAFFECTIVE UTTERANCES-.71 **** indicates significance at p < 0.05There was a positive correlation between the level of impulsivity displayed by the child and theamount of task irrelevant speech produced. Impulsive children produced more task irrelevantutterances than reflective children, r = .58; p > 0.02. No other significant correlations were foundbetween impulsivity and language used.A negative correlation was found between dependency and quantity of speech produced. The lessdependent children produced more total utterances and more broadly task relevant utterances,71r = -.61; p < .01; r = -.71; p > 0.001. Less dependent children also produced more social speechthan dependent children, r = -.53; p > 0.03. Children displaying more independence are morelikely to talk in general than dependent children. In particular they are more likely thandependent children to interact socially using language and talk more extensively about the taskat hand.A positive correlation existed between efficiency and the proportion of narrowly task relevantspeech produced. The more efficient children devoted a larger proportion of their speech tonarrowly task relevant remarks, r = .47; p = 0.05. No other correlations were found betweenefficiency and language use when all the children were considered as a homogenous group.To further explore possible correlations between personality and language use, it was decided tosplit the children into two groups: those who displayed a specific language impairment (SLI) andthose who did not (NL). I felt that the delayed language development of the SLI children mightbe masking developmental trends between language use and the three personality characteristics.The SLI children would be more similar to the ONL children in nonverbal maturity but moresimilar to the YNL children in language ability. They would be less impulsive, less dependentand more efficient but display less mature forms of language.All the above mentioned analyses were also performed on the two subgroups and similar resultswere obtained for both groups. Other analyses performed on the efficiency measure, however,indicated significant correlations and significant group differences between SLI and NL children.SLI children exhibited a negative correlation between efficiency and the total number of72utterances produced, r = -.78, p = 0.07, while NL children displayed a positive, though non-significant, correlation between the two, r = .35; p = 0.27. The same converse relationship wasobserved between efficiency, narrowly task relevant speech and private speech. The moreefficient SLI children produced less narrowly task relevant speech, r= -.74; p = 0.09, while themore efficient NL children produced more narrowly task relevant speech, r = .61; p = 0.04. SLIchildren scoring high in efficiency produced less private speech, r = -.61; p = 0.20 (non-significant), while NL children scoring high in efficiency produced more private speech,r = .60; p = 0.04. In all three instances the SLI children exhibited a negative correlation betweenefficiency and language while the NL children displayed a positive correlation. More efficientSLI children produced less total speech, less narrowly task relevant speech and less privatespeech than less efficient SLI children. Conversely, the more efficient NL children producedmore language overall, more narrowly task relevant speech and more private speech than the lessefficient NL children.When absolute values of speech are considered, SLI children displayed negative correlationsbetween efficiency and language while NL children displayed positive correlations between thetwo. As noted above, however, when proportions of language are considered, all the childrendisplayed a positive correlation between efficiency and the proportion of speech devoted tonarrowly task relevant comments. The SLI children talk less than the NL children. The moreefficient SLI children were likely to produce even less speech. When they did talk, however, itwas more likely to be narrowly task relevant. Although they spoke less they devoted a largerproportion of their speech to narrowly task relevant comments than the less efficient SLI children.73In contrast the more efficient NL children were more likely to produce more speech as well asdevote more of their speech to narrowly task relevant remarks.In summary, the major findings of this study were as follows:1) The vast majority of utterances generated by all the children during the bridge taskwere task-relevant and regulating.2) The same amounts of social and private speech were used in broadly task relevantremarks but private speech dominated in narrowly task relevant remarks.3) The SLI children produced less speech than the NL children. The older children(both SLI and NL) talked more than the younger children and the youngerchildren were off task more frequently. Although there were group differences inthe absolute amount of speech produced, all three groups devoted the sameproportion of remarks to narrowly task relevant speech.4) Some children displayed a preference for social speech as their main problem-solving strategy, some preferred private speech, while some used both equally.5A) There were no correlations between the preference in (4) above and the threepersonality characteristics, but there were correlations with other aspects oflangauge use.5B) A positive correlation existed between impulsivity and task irrelevant speech.5C)^A negative correlation was found between dependency and two aspects oflanguage: the total utterances produced and the amount of broadly task relevantspeech produced.74A positive correlation was found between efficiency and the proportion ofnarrowly task relevant speech produced by all the children. The data on SLIchildren displayed negative correlations between efficiency and three aspects oflanguage use: the total utterances produced; the number of narrowly task relevantutterances; and the amount of private speech produced. Conversely, the data onthe NL children displayed positive correlations between efficiency and these threeaspects of language.75CHAPTER FOURDISCUSSIONThis study was motivated by two major issues: the possible intellectual functions of social andprivate speech in children's language; and the possible influence specific language impairmentexerts on a child's ability to take advantage of these intellectual functions. A secondary purposeof the study was to investigate whether there were any correlations between personal style andthe manner in which the children used language. Since this is the first study in this area, someof the findings, while supportive of general research, the results are unique in the questions theyraise concerning children's use of language.Intellectual Functions of Language A major goal of this study was to explore the intellectual functions of language, be it social orprivate. In the present study the amount of task relevant speech increased with age from 83%for the younger children to 91% for the older (both NL and SLI) children. The majority of speechproduced by all the children during the bridge task was directly concerned in some manner withthe task itself, but the younger children produced more task irrelevant speech than the olderchildren. This is in accordance with results produced by Kohlberg et al. (1968) and Behrend etal. (1989) indicating that very young children produce less task relevant speech than olderchildren. The younger children may also have been off task more frequently because they wereless interested in the task. It was more difficult to get the younger children involved in the taskthan the older children, and they also appeared to lose interest in the task more quickly than theolder children.76Also consistent with previous findings (Berk, 1986; Berk & Garvin, 1984; Goodman, 1981), alarge proportion of the task relevant speech (77 - 84%) was regulatory in nature. Again, theyounger children produced less regulatory speech than the older children. The remaining speechwas purely affective or word play. The children used language throughout the bridge task. Theresults suggest that they used it primarily to talk about the task, to direct themselves and othersin possible task solutions and to focus attention on the problem at hand. These results will bediscussed more fully throughout the following pages.Broad Task Relevant (BTR) SpeechBroadly task relevant (BTR) speech is speech that was generally relevant to the bridge task butwas not specifically problem-solving in nature. This speech included fantasy play utterances,general comments about the building supplies or other utterances generally related to the task.BTR speech was equally likely to be private or social in nature (53% vs 47%). When thechildren were discussing general parameters of the task or were building on the fantasy aspectsof the bridge task, they were as likely to use social speech as private speech. Although there wasa great deal of variation from dyad to dyad in the story-line behind the fantasy play, all thechildren's solutions were influenced by the story-line developed in their particular session. Takefor example the following monologue of private BTR speech by JP:JP^a diving boat.JP^you know he ate that.JP^went into the water then.77Later in the session JP built a boat to get across the river. Social BTR speech appears tosimilarly influence the children's approach. The following is a short conversation between CWand NC, about the monster, the wind and their castles.NC^look it the wind blew everything.NC look at the wind.CW you not gonna break down my castle.CW because my castle's a very very very strong one.CW something a hurricane can't break down.NC^no mine is the strongest.NC^see he [the monster] likes my castle 'cause he thinks is it high.Shortly after this conversation CW began fortifying her castle so the wind/monster could notbreak it down. BTR speech appears to supply an overall context from which to draw a solution.It is possible that the BTR speech provides a method of "brain-storming" or "free association"until the child develops a working theory to take to the problem-solving task. Each explanationis equally plausible given the results of this study. Regardless of the role that BTR speech playsin problem-solving, at this level of task relevance the language of thought is equally private andsocial in nature. As the language became more narrowly task relevant, however, it also becamemore inwardly directed.Narrow Task Relevant (NTR) SpeechBy definition narrowly task relevant (NTR) speech focused on the problem-solving aspects of thetask. These utterances were directly relevant to solving the bridge problem. NTR utterancesincluded questions and comments about what to build, how to build it and why it was beingbuilt. NTR speech was used as an attentional aide to assist the children in focusing theirthoughts on the task. The children were also mapping out general plans for and identifying78specific steps needed to complete the project. The children were more likely to use privatespeech than social speech to focus their thoughts and direct their actions when working onproblem-solving aspects of a task. A high proportion of NTR speech was private in nature (61%vs 39% social). The following example illustrate the focusing function of language. MM istalking to himself while instigating an extensive search for a particular block. He appears to beusing language to keep his thoughts focused on what he is searching for.MM red.MM where where red.MM where where where where red.This focusing function of language was also exhibited in other ways. The children were notconstantly focused on the bridge-problem throughout the task. They would occasionally go offtask or indulge in fantasy play. When returning to the specific problem-solving aspects of thetask, they often repeated empty phrases such as "let's see" or "okay" once or twice as if to focusthemselves on the problem. These utterances were frequently subvocal and private.The planning and problem-solving functions of language were also frequently performed byprivate speech. The children often talked to themselves when they encountered difficulties toovercome. Following are two monologues of private speech produced during the bridge task.TK^now I can't fit.TK^which one fits?TK nope.TK^why can't it?TK why can't?TK where's the XX for it?TK^where did it go?and79ME me do it too.ME then let me do that too.ME no girl.ME I got one.ME I need more playdough for it.ME move it there.ME no, come around.ME then hammer.ME there.ME okay, okay.The children talked to themselves more than they talked to others when they were specificallyfocused on solving the bridge task. These results add to those in the current literature on thecognitive self-guiding use of private speech. While previous research indicates that much ofprivate speech is self-guiding in nature (Berk, 1986; Berk & Garvin, 1984.; Berk & Bivens, 1990;Goodman, 1981), this research is the first to indicate that the majority of direct problem-solvingspeech is private in nature.A fair proportion of NTR speech (39%), however, was also social. Focusing on private speechwhen investigating the cognitive self-guiding functions of language ignores a full third of thelanguage used for this purpose. Furrow (1984) reported that no function of language was totallyfulfilled by private speech alone. These results indicate that children's social speech may alsoplay a significant role in regulating actions and thoughts about a task. Such social speech maypotentially be self-regulating or other-regulating. If it is self-regulating then it may very wellplay an identical role to private speech. During the following monologue AM is attempting toget a plug of playdough out of a straw. She stops working and makes eye contact with theexperimenter as she says each remark then returns to her work.80AM I gonna break it pieces then it will come out.AM I (break) it up.AM XX it come out?This is similar to NTR private speech. Although it is directed to an external listener it appearsto play a self-regulatory role. This NTR speech seems to be directed at self, in spite of its socialform.Social NTR speech may also other-regulating. In these types of interactions one child directlyregulates the other child's actions. In the following exchange AG is helping SG rebuild astructure that has just collapsed. SG is in the process of adding a large block.AG^no (the structure falls again).AG see.AG see what I told you.AG^wouldn't work (begins to rebuild the structure).SG^mm-hmm (picks up another large block).AG nope, no tall one.AG nope nope.SG okay.AG's comments are directly focused on SG's actions. Through his comments he successfullyprevents SG from repeating an error. His language is specifically intended to direct the actionsof the other child he is interacting with.The children also provided ideas for each other just by working on the bridge task at the sametime. For example, in the session with EW and ST, EW decided to make a trap for the monster.When the monster entered her trap, a block fell and hit him on the head. Within moments STalso began to build a trap with a falling block. This type of imitation occurred in a number of8 1sessions. The children were able to watch each other and duplicate the strategies and solutionsadopted by the other child in the dyad.In other exchanges the children spoke to each other about their relative successes and providedsuggestions and opinions to each other. In these types of interactions one child did not directlyregulate the other's actions. The child might identify a problem but leave the solution up to theother. This is similar to scaffolding, where an adult, or peer, provides structure for a taskkeeping it within the child's ability level but allowing the child to complete the task alone.Following is an excerpt of a conversation between SK and TK about the bridge that TK has built.TK^now that's high enough.TK see?SK ya but how you supposed to get over now?TK what?SK^you gotta get over there and you can't climb (TK gestures jumping up to the bridge).SK^no, I mean if you had a person, how you going to get up and onto the other side?TK^easy (lifts the monster up and over the bridge).SK^'cause you can't do that the way you're building.TK jump.SK^jumping won't - you can't jump.TK just watch.TK^I'll just drive up there on these blocks.TK went on to build steps up to the bridge so that a truck could drive up and over the bridge.By pointing out flaws and making suggestions to each other on how to improve the structure, thechildren used language to scaffold the task for each other. Children who were scaffolded in thisway had the advantage of extra input into the bridge problem without direct regulation curtailingtheir actions.82The children not only received general problem-solving information about the bridge task fromeach other, but also from the experimenter. The experimenter did not directly regulate thechildren's behaviour but on occasion refocussed their attention on the task or reminded them ofthe parameters of the bridge task as in the following conversation with AG and SG.EXP how about trying to build a bridge that the monster can go through?AG he can break down that one but I'll make a taller one for he can't break down.EXP mm-hmm.SG^I'm gonna make a bridge that he can go under it.EXP that would be a good idea.The children then went on to integrate the new information into their solutions. This type ofinteractive situation offers alternative solutions that might not have occurred to the children ontheir own. They do not have to follow through on the offered solutions, but at least had theopportunity to consider the potential of each one and then adopt or discard the solutions as theychoose.Language appears to play an important role in the way children learn to structure their thoughts.Children use language not only to communicate their thoughts to others, but to communicatethought to themselves. Both social and private speech may be used as the language of thought.Although direct problem-solving types of language tend to be expressed through private speechthe correspondence is not perfect. Social speech also plays a significant role in the child'sunderstanding and solving of a problem-solving task.83General Incidence of Private SpeechOne of the reasons this study was successful in looking at the intellectual functions of languageis because we were able to elicit a high percentage of private speech during the bridge task. Thechildren in this study generated a large number of utterances, a large proportion of which wereprivate in nature (27% - 63%), and private speech was used by all subjects. Many previousstudies report a low incidence of private speech, with some children producing no private speechduring the experimental session. Beaudichon (1973) reported as few as 0.60 utterances perminute. In the first three studies by Kohlberg et al. (1968), only 50% of the children producedprivate speech. The difference between the present study and these earlier studies is likely theresult of differences in the experimental setting. In the present study the experimenterencouraged interaction between the children. She also interacted with the children herself. Berk& Garvin (1984) found that adult (teacher) presence hindered the production of private speech.Other studies, however, have reported if the adult does not regulate the child's behaviour, theincidence of private speech is either not affected or actually increases. Behrend et al. (1989)found that parent presence increased the frequency of both social and private speech producedby the children, while Goudena (1987) noted that the children produced more private speechwhen they were in the company of a collaborating adult. It seems that environments thatencourage interaction promote the overall use of both social and private speech. While thepresent experimenter did not encourage interaction herself and the children, neither did shediscourage it. Also, the children were encouraged to actively interact with each other. Althoughno previous studies have investigated children's private speech in interactive group situations, itis likely that such an environment would also elevate overall language use. The presence of84peer/adult and the interactive setting are probably responsible for the high levels of private speechin this study.The Effects of SLI on Language Use. The question here is how SLI effects a child's ability to make use of language. The first aspectof the data of interest concerns the intellectual functions of language. Although the SLI childrenproduced less speech overall than the ONL children, there was no significant difference in theproportion of speech devoted to the task by the two groups. Language appears to play agenerally regulatory role, regardless of language ability. The similarity in the proportion ofspeech devoted to NTR utterances and private speech by both SLI and NL children indicates thatspeech for self functions in a similar manner in SLI and NL children. Although SLI childrentalked less than NL children, children of similar cognitive levels (i.e. same chronological andmental ages) appear to use the same proportion of regulatory language. Language status seemsto have little effect on the use of this function of language. Such results indicate that SLI childrenattempt to use the cognitive functions of language to the same degree as NL children. It doesnot follow, however, that language functions as productively in SLI children as in NL children.Whether they are as effective in their use of language as NL children is still unknown.A second aspect of SLI langauge concerns the social nature of language. The NL childrenproduced slightly more social speech than private speech during BTR talk, and significantlymore private than social speech in NTR utterances. The SLI children, however, consistentlyproduced more private speech than social speech across situations. They also consistently85produced a higher proportion of private speech than the NL children. In other words the SLIchildren talked more to themselves than they talked to others.Two possible explanations exist that may explain this apparent difference in the way that NL andSLI utilize language. It is possible that the SLI children produce more private speech becausethey understand themselves better than others do. SLI children find it harder to communicatewith others and make their thoughts understood, so they do not talk as frequently. One of thecharacteristics of private speech is that it is somewhat idiosyncratic. SLI children may not beat the same fundamental disadvantage when they are talking to themselves as when they aretalking to others, so private speech is not reduced. The second possible explanation is that SLIchildren are less effective in social interaction than NL children. The current literature on thesocial relations among SLI children indicates that they are less likely to interact socially and areless successful than NL children when they do interact (Bryan, 1986). Such children are oftenless socially accepted and do rouse negative expectations in others through inappropriate anddeficient use of social language.Personal Style Personal style appears to exert a great deal of influence over how language is used. While theprivate speech was more commonly used for NTR utterances, there was also individual variationin the manner in which the children used social and private speech. Correlations were alsoevident between language use and the personality characteristics under investigation, i.e.impulsivity, efficiency, and dependency.86Solitary or InteractiveAlthough most children preferred to use private speech when problem-solving, some embraceda consensus-style approach. Three of the eighteen children (inter-active problem-solvers) reliedheavily on social speech for most of their problem-solving. The majority of their NTRutterances were social rather than private. Ten children preferred private speech (solitaryproblem-solvers) as their main language strategy for problem-solving, while five children showedno preference for either (versatile problem-solvers). These children used social and privatespeech equally throughout their problem-solving (NTR) utterances. There were no correlationsbetween any of the personality characteristics under consideration and this individual style. Also,the children from the three groups were randomly distributed throughout the three styles.To date no studies are available which investigate this aspect of language use, but severalpossible explanations exist: the developmental course of private speech and personal style mayexplain this pattern of language use. The curvilinear development of private speech, as suggestedby Vygotsky (1934/1962), provides one possible answer to the observed individual variation. Ifthe children were at different stages of internalizing regulation then this pattern of results wouldbe expected. Although children pass through similar stages around similar times, languagedevelopment follows an idiosyncratic time line and there is much individual variation. Theinteractive problem-solvers (3 children) may have not yet learned how to step other-regulationtoward self-regulation. They remain dependent on others to regulate their actions, and this isreflected in their language. The versatile problem-solvers (5 children) are balanced betweenother-regulation and self-regulation, relying on both equally. Vygotsky might argue that their87self-regulatory function is beginning to develop. At these ages, however, most (10 children) havebegun to take full advantage of the self-guidance and self-mediation functions of speech for self.Plausibility notwithstanding, the lack of other developmental trends in this study make aVygotskian explanation unlikely. Children who preferred private speech were not older, lessimpulsive, more efficient or more independent. It thus appears more likely that this individualvariation is simply due to personal style. While the majority of children seem to prefer usingprivate speech when problem-solving, this does not hold true for all. Some children appear touse social speech to a lesser or greater extent for the same purpose.Personality CharacteristicsIn spite of the tremendous individual variation in the way language is used in problem solvingthere is very little in the current literature about the effects of personality on language use. Thereis no information on correlations between personality and the use of social vs. private speech andonly preliminary results exist on other characteristics. What information there is, however,indicates that impulsivity, efficiency and dependency all exert some influence on aspects oflanguage use.Impulsivity: Impulsivity appears to affect the child's ability to focus on the problem space.Impulsive children are fast but inaccurate workers (Salkind & Wright, 1977). Although theywork quickly, they make a lot of errors, sacrificing accuracy for speed. Dickie (1973) found thatimpulsive children were more likely to use "self-stimulating", i.e. task-irrelevant, private speech.88Impulsivity was also positively correlated with the number of task irrelevant utterances in thisstudy. The more impulsive children were more likely to produce task irrelevant speech than theless impulsive children. These results may reflect the combination of two developmental trends.Children become less impulsive as they become older and older children tend to use less taskirrelevant speech than younger children. This correlation between impulsivity and task irrelevantspeech may just indicate the independent co-occurrence of these two aspects of the maturationprocess.Alternatively, there may be a true connection between impulsivity and language use. The resultsindicate that impulsivity affects the degree to which the child talks about the task. Impulsivitymay not affect language so much as it affects thought. If impulsive children go off task morefrequently, their language may reflect this thought pattern. Interestingly, if language is alsocapable of regulating thought, then therapy designed to develop the focusing and orientationfunctions of speech for self may assist impulsive children in staying on task. Berk (1986) foundthat among children who were more off-task, the greater use of private speech was related tolower levels of inattentiveness. Although no similar correlation was found in this study this maybe due to the measure used for impulsivity. The impulsivity measure was a Z-score, rankingeach child's level of impulsivity in relation to the other children in the study. A limited rangeof impulsivity in this population may account for the observed lack of correlation. The childrenin this study did appear to use private speech to focus attention. It is possible that impulsivechildren would benefit from the explicit teaching of this function.89Dependency: The less dependent children were not only more likely to talk than the moredependent children but they were also more likely to engage in social language use. Thestrongest correlation, however, was between the use of BTR speech and independence. Lessdependent children were more likely than dependent children to engage in BTR speech. If BTRspeech acts as a springboard for NTR speech then less dependent children may be better atdeveloping this springboard than more dependent children. These children may engage in moresocial speech because they are more able to generate independent ideas to add to the overall topicdevelopment. They may be less reliant on others to provide scaffolding for a task. WhileKlieman (1974) did not measure the use of social speech, she found that independent childrenwere more likely to use private speech than dependent children. The lack of similar findings inthis study may be an artifact of the small sample size. It may also be due to the relatively smallrange of dependency in the children. The dependent children were not that different from theindependent children in their scores on the Beller (1957) Scale. This may account for the lackof confirmation of this correlation between private speech and independence.Efficiency: Efficiency was the only personality characteristic where significant group differenceswere exhibited between NL and SLI children. Efficient children work quickly and accurately(Salkind & Wright, 1977). These children balance speed and accuracy without sacrificing onefor the other. Recall that analyses were run on both proportions and absolute numbers of eachaspect of language under consideration. One aspect of language use generated the samecorrelation with SLI and NL children. Highly efficient children devoted a larger proportion oftheir speech to NTR remarks than less efficient children, regardless of language status.90In other aspects, however, the efficient NL children differed from the efficient SLI children.The SLI children display mirror image correlations to those of NL children when absolute valuesof language were considered. SLI children displayed a significant negative correlation betweenefficiency and total utterances produced, a non-significant negative correlation between efficiencyand the absolute number of NTR remarks, and a non-significant negative correlation betweenefficiency and the amount of private speech produced. The NL children, on the other hand,exhibited a non-significant but positive correlation between efficiency and total utterancesproduced, a significant positive correlation between efficiency and the number of NTR remarks,and a significant positive relationship between efficiency and the amount of private speechproduced. Even though the correlations were not always significant the trend is apparent. Themore efficient SLI children generated fewer total utterances, less narrowly task relevant speechand less private speech than the less efficient SLI children. With the NL children the trend wasreversed. Efficiency was consistently correlated to greater language use in these three areas.Goodman (1981), who defined efficiency as a faster rate of motoric acts and shorter time to solvethe puzzle, also reported a correlation between efficiency and private speech with NL children.The more efficient (NL) children in her study also produced more private speech.This difference in language use between NL and SLI children is a new finding in the context ofboth private speech and language use by SLI children. There appear to be two pieces to thecorrelation between efficiency and language use in SLI children. Efficient SLI children talk less,but when they do talk it is narrowly task relevant. When they use speech, it tends to be specificto the problem-solving task.91The instrument used to measure efficiency in this study, the KRISP, is a visual perceptual task,so children scoring high in efficiency on this task were also strong visual processors. If efficientchildren make the best use of the strategies and tools available to them in the problem-solvingenvironment, then they will also minimize the use of ineffective strategies. Children who arehighly efficient and who also have solid language skills take full advantage of those skills anduse them to their fullest advantage. Language normally is an effective tool in problem-solving,so efficient NL children use it a lot. The current results suggest that efficient children who areweak in language skills will minimize the use of weak skills and develop other tools availableto them, such as visual-perceptual skills. It would also hold that what language they do usewould be highly specific to the task. This explanation is supported by the above reported patternof correlations between efficiency and language use.Stronger support for this particular view could, in principle, also come from the experimentaltask. For example, the more efficient children would be those who were quickest at solving thebridge-task. Unfortunately, it was too difficult to apply a measure of efficiency to the bridgetask. The children were able to choose their own approaches and solutions to the problem.Although the task was set up as a bridge-building task, the children were permitted to devise anysolution that allowed them to get across the river without being eaten by the monster. Some ofthe solutions were very simple, others were very elaborate. Some children created five or morestructures during the experimental session, while others worked on the same structure for mostof the session. This variation in solutions made any direct measure of task efficiency impossibleto apply across all the children.92Further research is required to investigate this apparent link between efficiency and languagestatus, but it may be that successful compensation of SLI over time leads to less language use.This has significant implications for early language intervention with SLI children. Workingsilently may be an indication of an effective adaptation to SLI. Early intervention may decreasethe child's ability to work efficiently but increase versatility. It is important to recognize thegains of either situation and structure intervention accordingly.Clinical Implications There are three points to keep in mind when considering intervention for the SLI child: 1) theintellectual functions of social and private language; (2) SLI children are less effective in socialinteractions; and (3) efficient adaptation to SLI may lead to reduced language use. If socialspeech does provide scaffolding for the task, then the SLI children are not receiving the fullbenefit of the scaffolded setting. If private speech rises out of the social setting, then SLIchildren are at a further disadvantage. Since the children are less likely to interact socially, theamount of other-directed regulation they receive may be reduced. They will receive lessguidance, and therefore, less transfer to self-guidance may occur. Their language impairmentmay also interfere with their ability to make full use of other-regulation when it is provided. Forthis reason, clinicians should consider devoting a proportion of their time to the explicit teachingof self-guiding strategies. They might also use a scaffolding approach to therapy sessions so theSLI children receive increased social input from which to develop self-guiding strategies. Suchsessions may also teach the children better social interaction skills. Children's language directlyaffects the way in which peers and adults interact with them (Blank & Franklin, 1980). Parents93and teachers find SLI children more difficult to talk to (Bryan, 1986). Improving SLI children'ssocial communication skills may ultimately improve the scaffolding they receive from theirenvironment. When SLI children are engaged in these types of activities, however, the clinicianshould not expect high performance on the play task. The children will be applying languageto the task but utilizing a less efficient problem-solving strategy.The effective use of private speech should also be a consideration of language intervention. SLIchildren appear to use private speech similarly to N1_, children. Efficient SLI children use lessspeech when problem-solving but when they use it, it tends to be specific to the problem-solvingsituation. If self-guiding speech arises out of the social context then the above interventionshould also improve the effectiveness of private speech by SLI children. It's effectiveness shouldnot be measured by increased use however. SLI children who have successfully internalized thescaffolding provided during intervention will use a higher proportion of those strategies in theirprivate speech but their overall use of language will not increase. These children will benefitfrom utilizing improved language strategies in private speech but will still function mostefficiently when working silently.Theoretical Implications Vygotsky (1934/1962) proposed that speech for self arises out of social speech. In this study,interaction between the children and the children and experimenter seemed to provide new ideasand information for the children to integrate with their own thoughts. The children appeared tointeract with each other to develop a general framework from which to launch their particular94solutions to the problem. Speech for self may fulfil a cognitive role that is not limited by whomthe speech is addressed to. Through the regulation provided by others, the child may learn self-regulation. Speech for self may begin as social and then turn inward. Goudena (1987), forexample, found that children who had received direct instruction (regulation) on a task producedmore private speech after receiving instruction than children who had received no regulation.In this study, the children were able to interact at will with others. The information garnered insocial interaction may have provided the children with a springboard. Input is received fromothers as well as from self until a possible solution is devised. BTR remarks were equally socialand private. Such speech may have assisted in orienting the child to a specific approach andprovided an outline from which to formulate a solution to the problem. Once the child hasdevised a working theory, private speech becomes the dominant form of speech for self as thechild works out the finer details to the solution. These results indicate that Vygotsky (1934/1962)was correct when he hypothesized that speech for self arises from the social context. Childrenuse the information provided through social speech when using language to regulate theirthoughts and actions.Vygotsky (1934/1962) defined speech for self by the function it played rather than by addressee.The current literature has focused solely on the behaviourially defined construct of private speech.The results of this study indicate that such a narrow definition excludes much valuableinformation about the intellectual uses of language. The language of thought does not appear tobe restricted by addressee. While the intellectual function of private speech has been recognized,the role of social speech in cognitive development has been virtually ignored. The language of95thought transverses the artificial boundaries of social and private speech. It is time to return toa more Vygotskian definition for the language of thought. Language should be defined by thefunction it performs rather than by merely the behaviour that accompanies it.Future Implications The results of this study suggest four main areas for future research.First, researchers interested in the intellectual uses of language should not just study privatespeech. It is important for future research to begin taking into account the intellectual functionsof social speech as well as private speech.Second, the role of fantasy and collaborative narrative in problem-solving remains essentiallyunknown. If social language and fantasy play provide a foundation for children to build uponthen it may very well play a significant role in the children's cognitive development. The effectsof scaffolding and social interaction on children's development are related to this issue. It is timeto drop back from the behaviourially defined construct of private speech and begin investigatingthe notion of speech for self as originally proposed by Vygotksy (1934/1962). The intellectualfunctions of langauge should not be limited by a behaviourial approach to the problem space.Both social and private speech appear to contribute significantly to the cognitive developmentof children so both should be considered when investigating the roles of language in cognitivedevelopment.96Third, the relationship between efficiency and language use in children with normal language andchildren with SLI should be more closely investigated. The results of this study suggests thatSLI has a definite effect on the way children utilize the intellectual uses of language. It isimportant to investigate if these trends hold for the larger population.Finally, it is important to examine if children with SLI use language as effectively as childrenwith normal langauge development. SLI children generally appear to use the cognitive functionsof language in a manner similar to NL children during problem-solving tasks. The questionremains to be asked: do they use it as effectively? The implications of reduced ability and useof social language in the cognitive development of SLI children must also be considered.Uncovering the differences and similarities in the way that SLI and NL children use both socialand private speech to regulate and guide their thoughts will broaden our understanding of howlanguage and thought interact.Conclusion This study was motivated by an interest in the intellectual functions of language. In particular,I was interested in the way children use language as a problem-solving tool and how SLI affectsa child's ability to employ the cognitive uses of language. Of secondary importance was therelationship between personality and language use. Since individual style appears to effectlanguage use and previous studies indicated correlations between dependency, efficiency andimpulsivity and language use, I was interested in duplicating of those findings.97This study has uncovered the cognitive uses of social as well as private speech. Both Piaget andVygotsky stressed the importance of interaction in language and cognitive development. WhilePiaget believed that peer interaction was the crucial factor in the development of children'sthought and language (Piaget & Inhelder, 1968), Vygotsky (1934/1968) hypothesized thatchild/adult interactions exerted the greatest influence on cognition. The results of this studyindicate that both may be correct. Social interaction with both peers and adults appears to fostergreater overall use of language as an intellectual tool.The children with SLI appeared to use language in a manner similar to children with normallanguage development. The SLI children used language to an age appropriate degree, butformulation problems may have decreased its effectiveness as a cognitive tool. The correlationsbetween efficiency and language use are the first results to indicate the children with SLI utilizethe intellectual functions of langauge in a manner different than children with normal languagedevelopment. Given these findings, it is important that more research be done in this area. Therelationship between efficiency and the effective use of language as a problem-solving strategyby both children with SLI and children with normal language development is still unclear. Moreinvestigation is needed in this area before any firm conclusions can be drawn.98BIBLIOGRAPHYBeaudichon, J. (1973). Nature and instrumental function of private speech in problem-solvingsituations. Merrill-Palmer Quarterly, 19, 117-135.Behrend, D.A., Rosengren, K., & Perlmutter, M. (1989). A new look at children's privatespeech: The effects of age, task-difficulty, and parent presence. International Journal ofBehaviourial Development, 12(3), 305-520.Beller, E.K. (1957). Dependency and autonomous achievement striving related to orality andanality in early childhood. Child Development, 28, 287-315.Berk, L.E. (1986a). Private speech: Learning out loud. Psychology Today, 20(5), 34-42.Berk, L.E. (1986b). Relationship of elementary school children's private speech to behaviourialaccompaniment to task, attention, and task performance. Developmental Psychology,22(5), 671-680.Berk, L.E. & Garvin, R.A. (1984). Development of private speech among low-incomeAppalachian children. Developmental Psychology, 20(2), 271-286.Berkowitz, M.W. (Ed.) (1985). Peer Conflict and Psychological Growth. New Directions forChild Development, No. 29. San Francisco: Jossey-Bass Inc.Bivens, J.A. & Berk, L.E. (1990). A longitudinal study of the development of elementaryschool children's private speech. Merrill-Palmer Quarterly, 36(4), 443-463.Blank, M. (1974). Cognitive functions of language in preschool years. DevelopmentalPsychology, 10, 229-245.Blank, M.E. & Franklin, E. (1980). Dialogue with Preschoolers: A cognitively-based systemof assessment. Applied Psycholinguistics, 1, 127-150.Bronckart, J.P. (1973). The regulating role of speech. Human Development, 16, 417-439.Chapman, M. (1991). The epistemic triangle: Operative and communicative components ofcognitive competence. In M. Chandler and M. Chapman (Eds.) Criteria for Competence (pp.209-228). Hillsdale, NJ: LEACromer, R.F. (1991). Language and Thought in Normal and Handicapped Children. OxfordUK: Basil Blackwell Inc.99Deutsch, R., & Stein, A. (1972). The effects of personal responsibility and task interruptionon the private speech of preschoolers. Human Development, 15, 310-324.Dickie, J.R. (1973). Private speech: The effect of presence of others, task, and intrapersonalvariable. Dissertation Abstracts International, 34, 1292B.Duncan, R.M. (April 1991) An examination of Vygotksy's theory of children's private speech.Paper presented at Biennial Meeting of the Society for Research in Child Development,Seattle Washington. April, 1991.Feigenbaum, P. (1990). The development of the planning function and the discourse structureof "private" speech. Dissertation Abstracts International. Feb Vol 50(8-B), 3724-3725.Feldman, C.F. (1977). Two functions of language. Harvard Educational Review, 47(3), 282-293.Forman, E.A., & Kraker, M.J. (1985). The social origins of logic: The contributions of Piagetand Vygotsky. In Berkowitz, M.W. (Ed.) Peer Conflict and Psychological Growth.  NewDirections for Child Development, No. 29, (pp. 23-40).Frauenglass, M.H., & Diaz, R.M. (1985). Self-regulatory functions of children's privatespeech: A critical analysis of recent challenges to Vygotsky's theory. DevelopmentalPsychology, 21(2), 357-364Frawley, W. & Lantolf, J.P. (1986). Private speech and self-regulation: A commentary ofFrauenglass and Diaz. Developmental Psychology, 22(5), 706-708.Furrow, D (1984). Social and private speech at two years. Child Development, 55(2), 355-362.Fuson, K.C. (1979). The development of self-regulating aspects of speech: A review. In Zivin,G.L. (Ed.) The Development of Self-Regulation Through Private Speech,  (pp. 135-211).Ginsburg, H. & Opper, S. (1979). Piaget's Theory of Intellectual Development. EnglewoodCliffs, NJ: Prentice-Hall Inc.Glucksberg, S. (1988). Language and thought. In Sternberg, R.J. & Smith, E.E. (Eds.) ThePsychology of Human Thought.Goodman, S.H. (1981). The integration of verbal and motor behaviour in preschool children.Child Development, 52, 280-289.Goudena, P.P. (1987). The social nature of private speech of preschoolers during problem-solving. International Journal of Behaviourial Development, 10(2), 187-206.100Guthke, J. (1991). The Learning ability test concept: History and on-going developments.Unpublished mauscript, University of Leipzig, Germany.Harris, K.R. (1986). The effects of cognitive-behaviour modification on private speech and taskperformance during problem solving among learning-disabled and normally achievingchildren. Journal of Abnormal Child Psychology, 14(1), 63-67.Harris, K.R. (1990). Developing self-regulated learners: The role of private speech and self-instructions. Educational Psychologist, 25(1), 35-49.Heath, S.B. (1989). The Learner as Cultural Member. In Rice, M.L. & Schiefelbusch (Eds.)(1989). The Teachability of Language, (pp. 333-350).Johnston, J.R. (1991). Cognitive Abilities of Language Impaired Children. Plenary addresspresented at the International Symposium of Specific Speech and LANguage Disordersin Children, Harrogate, England. May 1991.Johnston, J.R. (1988). Specific language disorders in the child. In Lass, N.L. (Ed.) Handbookof Speech-Language Pathology and Audiology, (pp. 685 -7 1 5).Kohlberg, L., Yaeger, J. & Hjertholm, E. (1968). Private speech: Four studies and a reviewof theories. Child Development, 39, 691-736.Lee, L.L. (1974). Developmental Sentence Analysis. Evaston: Northwestern University Press.McClinton, S.L. (1981). Verbal Problem Solving in Young Children. Journal of EducationPsychology, 73(3), 437-443.Manning, B.H., & White, C.S. (1990). Task-relevant private speech as a function of age andsociability. Psychology in the Schools, 27(4), 365-372.Meacham, J.A. (1979). The role of verbal activity in remembering the goals of action. InZivin, G.L. (Ed.) The Development of Self-Regulation Through Private Speech, (pp 237-261).Meichenbaum, D., & Goodman, S. (1979). Clinical use of private speech and critical questionsabout its study in natural settings. In G. Zivin (Ed.) The Development of Self-Regulation Through Private Speech, (pp.325-360).Olson, D.R. (1970). Language and thought. Psychological Review, 77, 257-272.Pellegrini, A.D. (1981). The social orientation of private speech. Child Study Journal, 11(1),25-31.101Pellegrini, A.D. (1980).^A semantic analysis of preschooler's self-regulating speech.International Journal of Psycholinguistics, 7(4) (20), 59-74.Pellegrini, A.D. & DeStefano, J.S. (1979). The functions of private speech in preschoolchildren. International Journal of Psycholinguistics, 6(3), 27-42.Piaget, J. (1962). Comments of Vygotsky's critical remarks concerning the language andthoughts of the child, and judgements and reasoning in the child. Cambridge, Mass: MIT.Piaget, J. (1955). The Language and Thought of the Child. Cleveland, OH: World PublishingCo.Piaget, J. & Inhelder, B. (1969). The Psychology of the Child. New York: Basic Books Inc.(Original french work published in 1966).Roberts, R.N. (1979). Private speech in academic problem-solving: A naturalistic perspective.In G. Zivin (Ed.), The Development of Self-Regulation Through Private Speech, (pp. 325-360). New York: Wiley.Rice, M.L. & Schiefelbusch (Eds.) (1989). The Teachability of Language. Baltimore MD:Paul H. Brooks Publishing Co.Roberts, R.N., & Tharp, R.G. (1980). A naturalistic study of school children's private speechin an academic problem-solving task. Cognitive Therapy and Research, 4(4), 341-352.Rubin, K.H., Hultsch, D.F., & Peters, D.L. (1971). Non-social speech in four-year oldchildren as a function of birth order and interpersonal situation. Merrill-Palmer Quarterly,17, 41-50.Salkind, N.L. & Schlecter, T. (1982). Comparability of the Kansas Reflection Impulsivity Scalefor Preschoolers and the Matching Familiar Figures Test in kindergarten children. ChildStudy Journal, 12(1), 1-5.Salkind, N.J. & Wright, J.C. (1977). The development of reflection-impulsivity and cognitiveefficiency: An integrated model. Human Development, 20, 377-387.Sternberg, R.J. & Smith, E.E. (Eds.) (1988). The Psychology of Human Thought. Cambridge:Cambridge University Press.Thal, D.J. (1991). Language and cognition in normal and late-talking toddlers. Topics inLanguage Disorders, 11(4), 33-42.Vygotsky, L.S. (1962). Thought and language. Cambridge, MA: MIT Press. (Original workpublished in 1934).102Wertsch, J.V. (1979). From social interaction to higher psychological processes: A clarificationand application of Vygotsky's theory. Human Development, 22, 1-22.Wright, J.C., Gaughan, D.J. & McClanahan, R. (1978). The KRISP: A Normative Evaluation.Unpublished manuscript from The Kansas Centre for Research in Early ChildhoodEducation, Department of Human Development, University of Kansas.Zivin, G.L. (Ed.) (1979). The Development of Self-Regulation Through Private Speech. NewYork: Wiley.103Appendix ADescription of Language Assessment MaterialsFour language measures were completed during the assessment sessions. Following is adescription of those measures.Developmental Sentence Score (DSS): The DSS is a method for evaluating the developmentallevel of a sample of free speech. A detailed evaluation of a child's use of standard Englishgrammatical rules is performed on the first 50 complete sentences of a tape-recorded sample ofthe child's spontaneous speech in conversation with an adult (Lee, 1974). Points, from 1 to 8,are awarded for the occurrence of specified forms within eight categories: 1) indefinite pronounsor noun modifiers; (2) personal pronouns; (3) main verbs; (4) secondary verbs; (5) negatives; (6)conjunctions; (7) interrogative reversals; and (8) wh-questions. Forms that are typically learnedlater within each category earn higher numbers of points.Language samples for the current study were recorded while the child and experimenter wereplaying with the Fisher-Price Farm set. Scores are percentiles for age (in 6 month blocks).Expressive One Word Picture Vocabulary Test (EOWPVT): The purpose of this test is toevaluate the size of children's expressive vocabulary. The child is shown a series of black-and-white line drawings and asked to name each one, using a single word. Drawings primarily elicitlabels for object and event categories. The child's response is written down and the test proceedsuntil 6 consecutive wrong answers/no responses are given. Scores are based standard scores witha mean of 100 and a standard deviation of 15.104Test for Auditory Comprehension of Language - Revised, subtests 2 and 3,  (TOLD-R): Subtest2 is a test of the child's auditory comprehension of the meaning of pronouns and grammaticalmorphemes such as prepositions, noun number and case, verb number and tense, noun-verbagreement and derivational suffixes, tested within the context of a simple sentence. Subtest 3examines the child's auditory comprehension of the meaning of elaborated sentence constructions,including modality (interrogative sentences, negative sentences, active and passive voice,embedded sentences, and partially and completely conjoined sentences. Each subtest consists of40 items. Each stimulus item is composed of a word or sentence and a corresponding plate thathas three black-and-white line drawings. One of the three pictures for each item illustrates themeaning of the morpheme or syntactic structure being tested. The other two pictures illustrateeither two semantic or grammatical contrasts of the stimulus, or one contrast and one decoy. Theexaminer reads the stimulus aloud, and the child is directed to point to the picture that bestrepresents the meaning of the word, phrase or sentence spoken by the examiner. No oralresponse is required on the part of the child (Carrow-Woolfolk, 1985). Scores are percentileranks (in 6 month block for ages 3;0 - 5;11 and year block from 6;0 - 8;11) with conversions forz-scores.Peabody Picture Vocabulary Test  (PPVT): the PPVT is designed to measure a subject's singleword receptive vocabulary. Each stimulus is composed of a single word and a correspondingplate consisting of four line drawings. One of the four pictures for each item illustrates themeaning of the word being tested. The other three pictures are decoys. The examiner reads tostimulus aloud and the child is directed to point to the picture that best represents the meaning105of the word spoken by the examiner. No oral response is required on the part of the child. Thetest is discontinued when there are eight consecutive responses containing six errors. Scores aregiven as standard scores with a mean of 100 and a standard deviation of 15.106Appendix BLetter and Consent Form sent to parentsLetter:Project Title: The Role of Language in Problem-SolvingInvestigators:Arlene Sturn^Speech-Language Pathology, M.Sc. student (U.B.C.) (604) 931-1317Judith Johnston^Professor, Audiology and Speech Sciences (U.B.C.) (604) 822-5591We are asking you to consent to participate with your child in a research project currentlyunderway through the School of Audiology and Speech Sciences at the University ofBritish Columbia and the Speech-Language Pathology Department of the CentralFraser Valley Health Unit.We are investigating the way in which children use language to solve problems. By comparingthe verbal problem-solving strategies of 3- and 4-year old children with normal languagedevelopment and 4-year old children with a language delay we hope to determine if distinctivestyles of language are used indifferent problem-solving situations. This study is specificallydesigned to investigate if children use language in a different manner when they talk tothemselves than when they talk to each other. Also under investigation is Whetherindependent/dependent children and impulsive/reflective children use different language stylesin problem-solving.Assessment: These two sessions will last approximately 20 - 30 minutes each and will be audio-taped. The children will be asked to participate in a series of screening tasks to determine theircognitive style (impulsive or reflective), to verify that their cognitive abilities are within normalrange for their age and their language skills are developmentally normal/delayed. These tasksinclude a picture-naming task, a categorization task, and a free play period.Experimental Task: If your child is appropriate for this study he/she will participate in anexperimental task. This session will also last approximately 20 - 30 minutes and will be audio-and video-taped. During this time children will play a game in groups of two. A large drawingof a river valley will be place on the ground. Each child will sit on opposite sides of the drawingand they will be asked to build a bridge across the river out of a variety of available buildingsupplies.Questionnaires: You will be asked to fill out a questionnaire which rates you child's level ofindependence. This will take approximately 10 minutes to fill out.107Your child's identity will be kept strictly confidential and will not appear in any report ofthis study.Agreements: Any audio- and video-tape records of your child will be kept at U.B.C. and usedonly by the project researchers. Questionnaires and written versions of the taped material willhave all names removed.If you should choose no to participate in this project or wish to withdraw at any time, you areunder no obligation to continue without jeopardizing the services that your child is receiving.If you have any questions or concerns regarding the procedures listed above, please call ArleneSturn at 931-1317 and I will gladly discuss the study with you.Yours sincerely,Arlene Sturn108Consent Form:I consent to the participation of my child and myself in the study: "The Role of Language inProblem-Solving". I am fully aware of the purpose and objectives of this projects and allprocedures have been clearly explained to me. I understand that my identity and that of my childwill be kept completely confidential.I understand that I am under no obligation to participate in this study, that I am free to withdrawat any time, and that doing so will in no way jeopardize the services that my child is receiving.I have retained a signed copy of this consent form for my own records.Please check one:I consent to my child's and my participation in this project.I do not consent to my child's or my participation in this projectSignature of Parent/Guardian^ DatePLEASE SIGN AND RETURN THIS COPY TO PROJECT INVESTIGATORS109Appendix CPersonality Characteristic MeasuresThe following measures were used to evaluate the cognitive style, efficiency and dependency ofthe subjects:Kansas Reflection Impulsivity Scale for Preschoolers: The KRISP was developed as an easierversion of the Matching Familiar Figures Test (MFF) (Kagan, et al., 1964) suitable for youngchildren. It consists of five practise, and ten test, items on each of two forms (A and B). Onlyform A was used for this assessment. Each item presents a standard line drawing of a commonobject, together with four to six similar alternatives, only one of which is an exact copy duplicateof the standard figure. The child is asked to find and point to the one exact copy in the array,and as is done with the MFF, latency to first response and number of errors are recorded.Children are advanced to the next item after a third pointing error on any item. Standardinstructions are used, including standard prompts to be used when the children forget or fail tofollow instructions.Impulsivity and efficiency scores were derived from the raw KRISP scores through a statisticalmethod proposed by Salkind and Wright (1977). The raw latency and error scores weretransformed into z-scores, based on the distribution of scores across all 3 groups. This yieldedtwo continuous scale distributions of all subjects in the sample, from most impulsive to mostreflective, and most to least efficient. Large positive impulsivity scores are indicative ofimpulsivity while large negative impulsivity scores are indicative of reflectivity. Large positiveefficiency scores signify inefficiency and large negative efficiency scores signify high efficiency.110Beller Questionnaire (Beller, 1957): The following questionnaire was given to parents tocomplete. The dependency score was the mean rating across all ten items yielding a continuousscale from most dependent (highest score) to least dependent (lowest score).The following rating scale was used for each question:Very often and^often and^occasionally and little^very rarely andpersistently^persistently persistence^without persistence7^6^5^4^3^2^1DEPENDENCY RATING QUESTIONNAIRE1 HOW OFTEN DOES THE CHILD SEEK HELP?By help is meant any form of assistance from another person. e.g., doing something for thechild like dressing, washing, finding a toy for the child, pushing the child in the swing,protecting it against another child when attacked or something is taken away from it, etc.,giving instructions and guidance, like demonstrating how to build, play, etc., giving what itasks for, e.g., a toy to play with, colour to paint, etc.2 HOW OFTEN DOES THE CHILD SEEK RECOGNITION?By recognition is meant any form of praise and approval. Child comes running to an adultto show what it did, e.g., exclaiming, "I washed my hands"; telling an adult that it carried outa command or request by the adult; e.g., "I put the blocks back," "I drank all the juice," etc.Calling an adult to see what it did, e.g. in the sand box, at painting, in the playroom, etc.Shouts to adult, "Watch me," when on the swing, when on a bicycle, when feeling it isespecially good, doing something praiseworthy, etc.3 HOW OFTEN DOES THE CHILD SEEK PHYSICAL CONTACT WITH AN ADULT?Physical contact: wants to be picked up, holds on to adult's clothing, hugs adult's knee, holdsor reached for adult's hand, puts arm around adult's neck (while adult demonstrates to child,reads to group, on the playground, etc.)4 HOW OFTEN DOES THE CHILD SEEK AllENTION?How often does the child manage to keep other occupied with it? Getting another person tooccupy themselves with the child, e.g., talking to them (answering questions, explaining,watching the child, giving approval, praising the child, scolding, punishing, etc.). Try to111ignore whether the child does it in a pleasant or unpleasant way, whether it is clever andskilful or clumsy or inefficient (a nuisance) in its efforts to draw attention, e.g., talking a lot,asking questions, volunteering answers, making a noise, making faces, being uncooperative,disobeying, excelling others, etc. use as your basic criterion how often it manages to keepothers occupied with it.5 HOW 01- lEN DOES THE CHILD SEEK TO BE NEAR OTHERS?By being near we mean just what it says. The child manages to sit near an adult (or anotherchild), to play where an adult is (or where another child or children are) regardless of whetherit interacts with the other person or not. If the child is active and skilful, this may expressitself in the form of playing with, working with, talking to, offering help, asking for help; onthe other hand, if the child is quiet and timid it may just hang around, watch, stand or sit nearanother individual child, adult (observer) or near a group. This differs from attention becauseit refers just to proximity and does not say anything about the relationship between the childrated to other children or to an adult when they interact.6 HOW OFTEN DOES THE CHILD DERIVE SATISFACTION FROM ITS WORK?This can be judged from the following behaviour: The child finishes its activity, e.g.,painting, building, play, etc., without asking an adult for comment; without makingderogatory comment on the work of other children; or without showing disturbance orirritation by bullying other children, dashing off wildly, destroying its own work, etc., butinstead moving away from a completed activity and getting ready for a new period.7 HOW OFTEN DOES THE CHILD AYIEMPT TO CARRY OUT ROUTINE TASKS BYITSELF?Routine tasks: e.g., dressing,washing,eating, toilet behaviour, etc. The rater is to put specialemphasis on the child's attempts to carry out these routine tasks by itself. The occurrenceof such attempts can be observed directly by seeing the child trying to dress by itself, to dressor undress at toilet or swimming pool, trying to get water running for washing, etc. (whilean adult assisted another child), or the child may be found doing any of these in a clumsyway but doing them as best it can (The rater must be careful not to let the feeling of a self-evident duty to assist the child in all routines when the child needs assistance interfere withan objective appraisal.)8 HOW 01-1EN DOES THE CHILD A 1 1EMPT TO OVERCOME OBSTACLES IN THEENVIRONMENT BY HIMSELF?By obstacle we mean missing a necessary tool or object in play or work, having misplaceda towel, a toy, clothing apparel, etc., desired objects that are placed out of reach, etc. Theextent of the child's striving to overcome such obstacles by itself can be seen when, after itsturning away from an ongoing activity (play or work), it returns and continues after havingovercome the obstacle. This is distinguished from reaction to such obstacles which arecharacterized by the child's interrupting his play or work to join other children or anotherchild, to scream out loud - "I need a hammer," "I need another truck" - to go from child tochild begging, demanding, and finally grabbing the desired object, or simply beginning to112daydream, wandering off aimlessly or crying. How often does the child seek or strive toovercome obstacles in the environment on his own without getting distracted from hisongoing activity?9 HOW OP I EN DOES THE CHILD TAKE INITIATIVE IN CARRYING OUT ITS OWNACTIVITY?When the child comes into the room, playground, etc., it knows what it wants to do andproceeds to do so, e.g., sandbox, bicycle, swing, building a ship, or a tunnel with blocks, etc.This can be distinguished from going out into the playground and looking around forsomeone to join, clinging to an adult, standing or wandering around aimlessly until an adulttakes initiative, asking someone to paly with it, or mostly wanting toys or tools which otherchildren have already begun to use. It does not matter whether another child enters itsactivity occasionally, the main criteria being whether it has it's won ideas and proceeds tocarry them out.10 HOW OFTEN DOES THE CHILD COMPLETE AND ACTIVITY?Once a task is set by an adult or selected by the child, the child carries it out to completion,e.g., construction, play, art, etc. This is to be distinguished from giving up easily, gettingquickly bored, disinterested or distracted. It is also to be distinguished form rigidperseveration, i.e., a child just keeps on doing one thing regardless of whether it is asuccessful or unsuccessful attack on the task. Use as your basic criteria how often the childcarries out an activity to its completion.


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items