Linguistic profiles of pragmatically disordered children


Material Information

Linguistic profiles of pragmatically disordered children three case studies
Physical Description:
vii, 244 leaves : ; 28 cm.
Coleman, Thalia Jean
Publication Date:


Subjects / Keywords:
Pragmatics   ( lcsh )
Communicative disorders in children   ( lcsh )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph. D.)--University of Florida, 1988.
Includes bibliographical references.
Statement of Responsibility:
by Thalia Jean Coleman.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001125409
notis - AFM2512
oclc - 20117442
System ID:

Full Text








This research is dedicated to the memory of my mother,

Mrs. Lily Mae Coleman, who died on August 13, 1986. She

made noble sacrifices throughout my entire life in the hope

that her children would enjoy a better standard of life than

she had. From 1954 when she sat by the light of a kerosene

lamp washing one of my two dresses so I could look nice in

school, until 1986 when she hid the seriousness of her

illness from me for fear that I would abandon this effort, I

remember her determined efforts to ensure my happiness and

that of her other children.

It was Mother's dream to live long enough to see me

reach this milestone. One of my most cherished memories is

the smile upon her face when she learned that I had decided

to embark upon this endeavor, and one of my most painful

recollections is the sadness of her countenance when she

realized that she would not be able to see me finish it.

I am grateful for every floor Mother scrubbed, her toil

in the hot sun, the neglect of her needs so I could have the

things I needed, each humble job she accepted in order to

buy food for her children, and for the times, when all else

failed, she begged for food and clothing for her children.

Mother through her integrity, strength, and perseverance

left an unforgettable legacy to me. The thought of her

sacrifices has encouraged me to persevere through many

dismal periods. For that I shall be eternally grateful.


I gratefully acknowledge the contributions of my

mentor, Dr. Linda J. Lombardino, to this research effort.

Dr. Lombardino provided valuable technical advice and moral

support during my matriculation at the University of

Florida. I also appreciate the support given me by the

other members of my supervisory committee: Dr. Thomas

Abbott, Dr. Vivian Correa, and Dr. Alice Dyson. Dr. Abbott

valiantly waded through administrative technicalities while

aiding me in my effort to finish this project and ensuring

my happiness in my future employment setting. Dr. Dyson and

Dr. Correa gave unwavering support in all of my endeavors.

Each of them was a great source of inspiration for me.

Special thanks are extended to the following very dear

friends for their sacrificial love and support throughout

this project: Dr. D.R. Williams, Mrs. Jamie B. Stappell,

Miss Laverne Q. Crawford, Mrs. Idella R. Dash, Mrs. Jean

Avery, Larry and Janice Palmer, and Sherry and Herbert

Dupree. I appreciate the support of Dr. Harold Powell and

my colleagues at South Carolina State College. I am also

grateful for the support of my family; the townspeople of

Lake View, South Carolina; and my friends at Williams Temple

Church of God in Christ, Gainesville, Florida and at Faith

Tabernacle Deliverance Temple, Orangeburg, South Carolina.



ACKNOWLEDGEMENTS ........................................ iii

ABSTRACT....... ..........



Review of the literature........................... 14

2 METHODOLOGY.........................................62

Subjects........... ...................................... 62
Assessment Model...................................65
General Procedures ................................... 67
Data Collection and Analysis .......................78
Reliability......................................... 3

3 RESULTS ..............................................88

Case Study One ...................................... U
Case Study Two ..................................... 119
Case Study Three................................... 145
Chapter Summary...................................171

4 DISCUSSION.........................................175

Methods of Investigation ..........................175
Subject Characteristics...........................179
Patterns Observed Across Experimental Subjects....194
Conversational Profiles............................198
Are Pragmatic Deficits Dissociable from Other
Dimensions of Language ........................206
Pragmatic Deficits: Autism Versus Non-autistic
Syndrome............................................ 21
Relationship of Findings to Previous Research..... 214
Implications for Future Research..................218




WORKSHEET:SAMPLING TECHNIQUES.......................224


PARENT CONSENT FORM ............ ................... 227

BARRIER TASK ANALYSIS....... ......................228

REFERENCES...... ......... ...............................230

BIOGRAPHICAL SKETCH......................................242

Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the degree of Doctor of Philosophy



Thalia Jean Coleman

December 1988

Committee Chairperson: Dr. Linda J. Lombardino
Major Department: Speech

Research in the area of discourse development has

focused primarily on descriptions of early language

behaviors in speakers from exemplary backgrounds. Few

attempts have been made to systematically investigate the

conversational behavior of children who are language

impaired. Most research efforts in the pragmatics domain of

language disorders have focused on examining selected

aspects of conversational performance. Little emphasis has

been placed on evaluating the multiple dimensions of

pragmatics in language-impaired children. Conclusions drawn

based on studies of isolated conversational skills may be

different if these same skills were examined as part of the

child's overall linguistic performance.


The purpose of this study was to conduct detailed

analyses of the linguistic behavior of children who had been

identified as poor conversationalists, in an attempt to

describe and isolate the specific pragmatic variables that

are impaired and to develop profiles of a small number of

pragmatically disordered children. In contrast to previous

research that had focused on selected aspects of pragmatics,

this study examined all domains of language. The results of

this investigation indicate that there are multiple patterns

of pragmatic deficits and that language-impaired children

can have a pragmatic deficit in specific areas but not in

others. Findings also revealed that weaknesses in other

areas of language may contribute to the perceived severity

of the pragmatic disorder.




General Statement

Human communication in its broadest meaning is the act

or acts that produce some kind of response between two or

more persons. Because this is true, then those persons must

clearly understand each other, or the response may be far

from what is expected (Bangs, 1968). Perhaps that is the

reason effective communication has been an area of concern

for mankind since the beginning of recorded history. Much

of that concern has focused on oral communication. An

Egyptian essay, written on parchment around 3000 B.C. and

addressed to the son of the pharaoh, contained some

elementary advice on effective speaking. During the fifth

century, B.C., several interesting theories concerning

effective public speaking were developed by Greek

philosophers (DeVito, 1964). This fascination with oral

communication skills has persisted into the twentieth


One aspect of oral communication that has received

considerable attention within the past three decades is

language development in children. Several disciplines have

combined efforts that have yielded an impressive volume of

information concerning how language is acquired. Most

recently, much of this research has focused on the

acquisition of conversational skills. Although conversation

is a universal human activity performed routinely in the

course of everyday interactions with our fellow beings, the

nature of conversational speech and the means by which

children learn to participate in conversations are still

poorly understood. We have some ideas about what

constitutes a well-formed and well-performed conversation

and can recognize at least gross deviations from these

(McTear, 1984). Recent investigations have yielded results

that identify the specific skills involved in normal

conversations of children and of adults. Additional studies

have focused on how these skills are acquired (Bedrosian &

Willis, 1987; Chapman, 1981; Coggins & Carpenter, 1981;

Dale, 1980; Dunn & Kendrick, 1982; Fey, 1986; Fivush &

Slackman, 1981; Gallagher & Prutting, 1983; Lucas, 1980;

Nelson & Seidman, 1984; Terrell, 1985). There are many

questions yet to be answered, however, regarding the nature

and development of discourse skills in children.

Citation of Related Studies

Many theories have emerged concerning the acquisition

of language. One school of thought that has been widely

accepted is the interactionist viewpoint that was presented

by Bloom and Lahey (1978). In his discussion of this

theory, Fey (1986) reported that language must be understood

as a complex system involving three separate but highly

interdependent knowledge bases: content, form, and use.

Content consists of two kinds of knowledge: referential and

relational. Referential knowledge involves the child's

understanding of the particular objects, events, and

relations that underlie the meanings of the words we use.

Relational knowledge is the child's understanding of the

roles that objects can play in the context of actions or

states of affairs as well as the ways in which events can

relate to one another. According to Fey (1986), much

interest has developed concerning how this knowledge is

acquired and the impact that this knowledge has on a child's

ability to participate in on-going discourse. Children must

learn how knowledge (content) and form (structure) interact

to serve different functions in conversation.

Nelson and Seidman (1984) proposed that three

components of conversational context would contribute to

maintenance of dialogue: presence of objects, shared topic,

and shared script knowledge. Nelson and Seidman discussed

scripts as mental or cognitive representations of events

that are familiar to an individual and the temporally

related acts that typically characterize such events.

According to Nelson and Seidman, conversation takes place

around a shared topic. Conversational episodes are defined

in terms of topic shifts, which begin and end each episode.

Thus shared topics are necessary in order for dialogue to be

sustained. They further stated that, in order to account

for the variable potential of topics to sustain play-related

discourse, we need to invoke the notion of the shared

script. Additionally, they pointed out that when a script

becomes the topic of play, it provides the structure within

which a coherent sequence of activity and talk may proceed.

Fey (1986) suggested that children can be encouraged to

converse about things not present in the immediate context

through the use of scripts. He further speculated that such

activities enhance not only the child's command of language

as a tool for communication of nonpresent events but also

the child's use of language as a tool of thought.

Other research in the area of discourse skill

development has yielded results that indicate that children

who learn language normally acquire a variety of skills

necessary for effective communication prior to entering

school. These children are sensitive toward several

parameters of nonverbal behaviors by the time they are three

years old. Prinz (1982) reported that certain socio-

cognitive aspects of development, such as symbolic play,

role-taking abilities, social dominance and persuasion

techniques, and presuppositional and referential

communication skills, all contribute to pragmatic

advancement at this early stage. Ervin-Tripp (1977) found

that these children are capable of using both direct and

indirect request strategies. Sachs and Devin (1973)

reported that by the age of four, children are able to vary

politeness depending on whether they are speaking to adults,

peers, younger children, or baby dolls. Topicalization and

turn-taking are two major aspects of discourse that continue

to evolve during this period.

Fey and Leonard (1983) identified three patterns of

conversational participation in children with specific

language impairment. Children whose behavior is consistent

with the first pattern are generally unresponsive in all

types of social interactions relative to same-age normal

language children. Children who fit the second pattern are

generally more responsive to their social partners than are

the children in pattern one. They are less likely to

initiate conversational exchanges than are their normal

language peers. These children can be described as

responsive but nonassertive. Pattern three describes the

behavior of specific language impaired children who assume

the roles of both speaker-initiator and listener-respondent

in a manner that is very similar to normal language children

their own age. They do not, however, assume these roles in

the same manner that is consistent with that of their peers.

They rely more on nonverbal means of encoding. They also

seem to be limited in specific linguistic forms required to

produce certain speech acts, such as the assertion of rules.

Fey and Leonard (1983) reported that the most general

conclusion, however, is that the population of specific

language impaired children is no more homogenous with

respect to pragmatic skills than it is with respect to

semantic, syntactic, or phonological abilities.

Snyder (1976, 1978) found that language-impaired

children appeared deficient in several aspects of

pragmatics. They had difficulty generating linguistic

performatives, encoding the most informative contextual

element, and disengaging themselves from the immediate

context. Gallagher and Darnton (1978) found that language

impaired children revised their utterances in response to

listener misunderstanding. The revision strategies they

used, however, were different from those found in normal

children and were not systematically related to their levels

of structural knowledge. Gale, Liebergotti, and Griffin

(1981) found that language impaired children's request for

clarification patterns were qualitatively different from

those of normal language-learning children of the same

language stage. The language impaired children used more

nonverbal requests and exhibited a restricted number of

linguistic forms. Conti-Ramsden and Friel-Patti (1983)

found differences in the discourse skills of language

impaired children. These children participated in dialogue

less often and produced more inadequate and ambiguous

responses. Wiig and Semel (1984) reported that learning

disabled children exhibit problems in the area of social

perception. They fail to perceive, interpret, and

appropriately use facial expression, gestures, and other

body language.

Preschool children also have consistently shown certain

error patterns in conversational interactions. Preschoolers

have difficulty in managing interruptions and also do not

have well-developed devices for acknowledging and

terminating topics (Ervin-Tripp, 1979). Terrell (1985)

suggested that they may experience problems gaining a turn

in on-going discourse as a result of these difficulties.

According to Terrell, children at this level find it

difficult to monitor the conversation and determine the

appropriate point to "jump in." They cannot monitor the

speech of others and simultaneously remen.ber their own

contributions. Because of this deficiency, preschoolers

find it more difficult to participate in multi-party

interactions with adults and older children than in dyadic

interactions (Ervin-Tripp, 1979). According to Terrell

(1985), during the early stages of linguistic development

children lack the specific devices for verbally terminating

and initiating topics; therefore, their contributions are

often viewed as irrelevant. As a result, they are sometimes

ignored by older listeners (Ervin-Tripp, 1979). Topic

introductions may also be unsuccessful (Terrell, 1985).

These children do, however, have the skill to readjust when

their topic introductions fail, and to make relevant and

contingent responses to the topics on the floor (Prelak,

Messick, Schwartz, and Terrell, 1981).

Shortcomings of Previous Investigations

The typical methodology used in data collection for

conversational dialogues includes the observation of

naturally occurring behaviors. According to Friel-Patti and

Conti-Ramsden (1982), the goal of observational research is

to gather samples of communicative behaviors that are

representative of the subjects' actual response repertoire.

They suggest that the adequacy of these samples depends in

part on a series of decisions made by the investigator.

These decisions involve the means of observing and

recording, data collection procedures, physical facilities,

the time of day, and the people involved other than the

subjects themselves. Furthermore, according to Friel-Patti

and Conti-Ramsden, once the data have been collected, they

must be analyzed in a manner that seeks to maximize

objectivity and minimize examiner bias or subjectivity.

They stated that of particular importance in the study of

children's conversational discourse are the decisions made

concerning the children's behavior and contextual notations.

Furthermore, according to the writers, in studying atypical

populations such as language impaired children, one is

confronted with the same methodological problems as those

with normal children. Problems arise given the large range

of performance of language-impaired children with changes in

the conditions under which their behavior is observed.

Friel-Patti and Conti-Rarasden warned that caution must be

exercised when interpreting the results of language-impaired

children's dyads in specific situations as they may not be

representative of the full repertoire of the dyad. The

results obtained in only one condition may seriously

underestimate the communicative abilities of these dyads.

Friel-Patti and Conti-Ramsden (1982) reported that

children's styles of interaction from situation to situation

were different. Although the style of interaction for

children with normal language was stable, the style of

interaction of language-impaired children was not. The

researchers felt that the effect on language-impaired

children's performance would be similar when other aspects

of data collection are manipulated. Friel-Patti and Conti-

Ramsden (1982) concluded that there clearly are problems

involved in the data collection, transcription, and analyses

of language-impaired children's discourse skills and that

their cumulative effect raises questions specific to the

population of atypical language learners. They suggested


that further research with language-impaired children should

aim to tighten the methodological decisions.

According to Fey (1986), there is considerable

variation in the rate, sequence, and pattern of language

development in normal children. Normative data are usually

gathered on very small samples of children, and these

children generally come from exemplary backgrounds. They

are observed over long periods of time in a variety of

contexts. Fey pointed out that these same procedures are

generally not used when studying language-impaired children.

In fact, very little developmental data is available about

the conversational skills of language-impaired children.

The data that have been gathered often do not reflect more

than one context of conversational participation. The

results obtained from the studies of normal language

children are used as a data base to indicate what is normal

for the general population even though they are generally

gathered on a very small sample of children and on limited

variables at any one time. These data are commonly accepted

as regular targets or goals in intervention. Until more

extensive research is done on a wider segment of the

population and on overall conversational skills, a question

remains as to whether there are differences between various

groups of children, or if there are specific patterns that

may be considered deficits that need to be targeted for


Several researchers have suggested formats for

assessing certain aspects of language use (Chapman, 1981;

Coggins & Carpenter, 1981; Prutting, 1982; Simon, 1984;

Gallagher & Prutting, 1983; Keenan & Schieffelin, 1976;

Bedrosian, 1985). According to Fey (1986), even when

pragmatic factors are considered, the children's skills are

tested outside typical contexts, and consequently the

children's true linguistic abilities are distorted.

Comparisons of a child's performance with those of other

children are not likely to be valid and reliable when the

validity and reliability of the instruments and/or

procedures used to measure specific skills are not well

established. Because most of the research in the area of

pragmatics has suffered from the same methodological

problems that have plagued language research in general,

there is a lack of an appropriate or complete data base from

which to draw information when making assumptions about the

conversational adequacy of children from diverse


Friel-Patti and Conti-Ramsden (1982) reported that

research with language-impaired children has typically

lagged behind that of normally developing children and has

often been fragmentary. They speculated that this has

resulted from the fact that investigations with atypical

language learners tend to look at skills in particular tasks

at a particular moment in time rather than looking at

patterns of behavior over time. Further, they stated that

there has been little developmental emphasis on research

with language-impaired children beyond the one-word stage.

The approach to the study of discourse abilities of

preschool language-impaired children has generally been to

select a group of language-impaired children and compare

them with chronological age-mates and/or MLU matches.

There has been no systematic attempt to discover

developmental patterns in the language-impaired population.

In addition, Friel-Patti and Conti-Ramsden pointed out that

research with this population has been limited in scope

(only a few aspects of discourse have been examined) and

limited by the methodological problems previously discussed.

Purpose of the Study

The development of discourse skills is currently a

popular area of study in language. Bedrosian (1985) and

Craig (1983) reported that there has been a strong

theoretical shift from syntactic ana semantic models of

child language acquisition to communicative or discourse

models of language acquisition during the last decade.

According to Bedrosian (1985), this theoretical shift has

also been evident in the area of language disorders.

Information on how normally developing children develop

conversational skills is just emerging. It is almost

nonexistent for language-impaired children (VanKleeck &

Frankel, 1981).

The majority of the research in the area of discourse

development has focused on descriptions of early language

behaviors of speakers from exemplary backgrounds. Few

attempts have been made to systematically investigate the

conversational behaviors of children from diverse

backgrounds or of children who are language-impaired. Very

small samples of children have been used to examine selected

aspects of conversational performance. It is possible that

assumptions made based on studies of isolated conversational

skills may be quite different if those skills were examined

as part of a child's general conversational abilities. This

study examined children's conversational skills in different

contexts in response to three different elicitation

strategies: "Script-Based Play," "Elicited-Schemes Play,"

and "Barrier-Task Play."

According to Wanska, Bedrosian, and Pohlman (1986),

over a period of years investigators have documented the

effects of materials or physical context variables on the

syntactic performance of children. They reported, however,

that only a few studies have examined the effect of these

variables on pragmatic performance. Few, if any, studies

have performed detailed analyses on single subjects in order

to describe and isolate tne specific pragmatics varlaoles

that are impaired.

The purpose of this study was to do in-depth analyses

of the conversational behavior of young children who have

been identified as poor conversationalists in an attempt to

describe and isolate the specific pragmatic variables that

are impaired. The researcher developed indepth profiles of

children who had pragmatic problems in order to quantify and

qualify the nature of pragmatic disorders aiid to examine

pragmatics in relation to

other dimensions of language including phonology, semantics,

and syntax.

Review of the Literature

The vast amount of literature available on the topic of

language development in children is beyond the scope of this

study. The impact that the study of pragmatics has had on

language research will be discussed under the general

heading of Language Use in Children. Under this heading two

areas will be addressed: the development of pragmatic skills

in children and the effect of materials and physical context

variables on conversational performance. The topic of

conversational skills in the language-impaired population

will be reviewed under the heading of Discourse Development

in Language-Impaired Children. Some concerns about research

and assessment methodology will be discussed under the

heading of Research and Assessment Issues. A rationale for

conducting an extensive analysis of conversational skills in

children who have been identified as having pragmatic

disorders will be presented under the heading of Statement

of Need.

Language Use in Children

Most recently the shift in the study of language

development and disorders in children has been toward the

study of the pragmatic aspects of language. Prutting (1982)

stated that the contemporary term, pragmatics, has its

origin in an early Greek word that meant action. Greek

philosophers such as Plato and Aristotle were interested in

the relationship between the speaker and the audience.

Prutting concluded that rhetoric, therefore, may be thought

of as the earliest form of contemporary pragmatics. Bates

(1976) is generally given credit for introducing the term

pragmatics into the field of speech-language pathology.

Researchers now realize that in addition to learning the

semantic, syntactic, and phonologic rules of language, a

child must also master the rules that underlie how language

is used for the purpose of communication (Hymes, 1971; Roth

& Spekman, 1984).

A child acquires language because of its usefulness.

Even before children learn to talk, it becomes apparent to

them that communication can be a valuable tool. Pragmatics

is concerned with how and why the speaker uses language,

what the purposes are, and what the speaker hopes to

accomplish through different utterances (Nation & Aram,

1984). The definition of pragmatics that is used most often

is based on the semiotic theory (Morris, 1938). According

to Morris, pragmatics presupposes both syntax and semantics.

During the evolution of transformational grammar, Morris

made many contributions to the study of language use. The

transformational grammar era dominated the study of

linguistics for many years. It was only in the decade of

the 1970s that there was renewed general interest in the

study of pragmatics. Two researchers made significant

contributions during the 1960s that helped to usher in this

period of renewed interest in the use of language. Austin

(1962) suggested that the speech act is the basic unit of

pragmatics. He explained the speech act as a linguistic

communication between a speaker and a listener. Austin

described the speech act process as consisting of three

phases: locutions, illocutions, and perlocutions. The most

advanced of these phases is the locutionary act, which

consists of the utterance of words plus the proposition or

content of the message. Searle (1969) also discussed the

importance of speech acts in communication. He said that

the speech act is composed of the doing aspect (utterance

act), the meaning aspect propositionall act), and the

function (illocutionary act).

Johnson, Johnston, & Weinrich, (1984) reported that a

review of literature in speech-language pathology identified

a number of pragmatic functions that must be correctly

coordinated if language is to serve as a useful tool.

Language must be appropriate to the context and the

listener, be polite enough, provide sufficient information,

be relevant to the situation, and have sufficient truth

value (Grice, 1976). In addition, Johnson and her

colleagues (1984) stated that language must incorporate the

expected rules for discourse -- establishing topic,

maintaining conversation, using appropriate register, and

signaling pragmatic information by using specific syntactic

forms. According to Bloom and Lahey (1978), if a child can

produce generally appropriate and meaningful discourse, his

or her language will function as a useful tool. They

explained that the child will be able to use language to do

such things as ask and answer questions, give directions,

comment on problems, suggest actions, give orders,

speculate, joke, argue, and eventually to use the higher

literary and argumentation forms of the adult world.

According to Roth and Spekman (1984), a large number of

clinical reports have indicated that language-impaired

children with disabling conditions demonstrate najor

communication deficits that transcend their problems with

form and content or that exist even in the presence of

normal linguistic skills. The writers speculated that

problems exhibited by these children illustrate that

linguistic knowledge alone does not guarantee appropriate

language use. Ochs (1979) stated that expanding the study

of child language to include pragmatic dimensions enables us

to see the variety of communicative skills the young child

uses in everyday social situations. This, she suggested, in

turn provides a richer and more accurate account of the

child's linguistic and social knowledge. Damico and Oller

(1980) reported that teachers were able to identify language

disordered children more readily by using pragmatic criteria

than they could by using morphological/syntactic criteria.

They further reported that results from many studies show

that even persons who are untrained in linguistics make less

reliable judgments about discrete points of surface form

than they do about more global aspects of communicative

effectiveness. Investigations in the area of pragmatics

have focused primarily on three aspects of communication

skills: communicative intentions, presupposition, and social

organization of discourse. Several variables represent the

essential components of discourse and may serve to guide a

clinician's observations in this area. These include topic

maintenance, on-topic exchanges, topic termination, and

conversational repairs (Roth & Spekman, 1984).

Communicative Intentions

Bates (1975) examined the onset of intentional

communication before speech begins and traced it to the very

first uses of gestures as forms of intentional

communication. Much of Bates' early work was based on that

of philosopher J. Austin (1962). Bates followed the

development of language through Austin's three stages. Two

precursors to the first functional uses of speech were

defined: protodeclaratives and protoimperatives. Halliday

(1975) talked about language as "doing" and language as

"learning." He proposed eight communicative functions of

language: instrumental, regulatory, interactional, personal,

heuristic, imaginative, representational, and ritual. He

also suggested that there are three distinct phases of

pragmatic development. Dore (1975) described children's

one-word utterances as primitive speech acts. He described

how intentions emerge and contribute to the child's

acquisition of words. Dore also discussed the idea of

language universals and suggested that all language can be

analyzed in terms of these fundamental notions. He

developed an elaborate scheme for coding what he called

conversational acts in the speech of young children.

Roth and Spekman (1984) suggested that at the level of

communicative intentions, it is important to analyze both

the range of intentions understood and expressed as well as

the verbal and/or nonverbal forms in which they are coded.

They claimed that the analysis of intention may be

approached in several ways. First, one can examine whether

a child uses gestural, paralinguistic, and/or linguistic

means. Second, the linguistic structures used to convey an

intention can be analyzed. According to Roth and Spekman, a

classification system of semantic relations can be employed

for a young language learner. For a more syntactically


advanced child, messages can be coded for sentence types. A

third approach involves an examination of the degree of

explicitness with which an intention is expressed.

Wells (1974) categorized speech acts in relation to the

parameters that control the intents and forms used. In his

system, speech acts are categorized within each of the

following functions of communication: ritualizing,

informing, controlling, feeling, and imagining. He pointed

out the critical importance of the infant's physical and

verbal environments. Wells suggested that the meanings that

the child constructs prelinguistically are heavily dependent

upon the direction of his focus of attention by caregivers.

Babies engage in various forms of eye contact with

their mothers (Argyle & Ingram, 1972 cited in Prutting,

1979). These acts seem to signal intent, readiness, and

turn-taking. Coggins and Carpenter (1978) defined a set of

eight pragmatic categories identifiable in preverbal

children. Beginning at about 9 months of age, children have

been found to use a variety of intentions that they express

through gestures and early vocalizations. Several

researchers have identified a set of these early

communicative behaviors (Bates, Camioni, & Volterra, 1975;

Escalona, 1973; Halliday, 1975). The first gestures of

infants that are consistent with the gestures of the

community include showing, giving, and pointing. Subjeccive

sensorimotor pointing leads to communicative pointing

(children use objects to obtain attention from adults)

(Bloom & Lahey, 1978). Infants gradually progress to

communicating with intention to communicate in the second

half of the first year. Although still unable to use the

conventions of language, infants come to know that their

behavior can influence the behavior of others, and they

behave with the intention of achieving that influence (Bloom

& Lahey, 1978). Coggins et al. (1987) reported that by 12

months most subjects were able to establish joint reference

and direct their parents' attention to interesting sights or

sounds. DeVilliers and DeVilliers (1978) found that by the

time a child is 12 months old, he can engage another

person's attention through gestures and eye contact to

indicate what he needs. According to Chapman (1981),

studies typically fail to find examples of intentional

communicative vocalization until about 9 to 12 months, or

sensorimotor stages 1 and 2. Beginning at about 9 months of

age children use a variety of intentions that they express

through gestures and early vocalizations. Some of the

preverbal communicative intentions that have been identified

are attention seeking, requesting, greeting, transferring,

protesting/rejecting, responding/acknowledging, and

informing (Roth & Spekman, 1984). According to Prinz

(1982), the specific categories of communicative

interactions that occur between 6 and 12 months are:

participation in routine games, requesting and rejecting


objects, and commenting on events in the environment. Prinz

further stated that participation in routine games,

requesting objects or assistance, rejecting proffered

objects or activities, and indicating an object to the

mother or commenting on its appearance are common to 9 to

12-month olds.

Dore (1974) identified several early communicative

behaviors: labeling, answering, requesting, calling,

greeting, protesting, repeating, and practicing. Tough

(1977) was primarily concerned with the roles that language

plays in problem-solving and thinking. Tough devised a

system of categorization of communicative functions and

intents that views language from a cognitive-semantic

perspective. Within this system, four major functions of

language are identified: directive, interpretative,

protective, and relational.

According to Roth and Spekman (1984), the clinical

assessment of communicative intent involves gaining some

idea of 1) the types of intentions comprehended and

expressed, 2) the linguistic forms of intention comprehended

and expressed, 3) the nonverbal and paralinguistic means for

communicating intent, and 4) the social conventions that

govern interpretation and selection of particular linguistic

and nonverbal forms of intention.


Assessment in the area of presupposition focuses on the

ability of children to take the perspective of their

conversational partners. This skill is commonly referred to

as role taking. The role taking skills necessary for

communicative success typically must be inferred from the

linguistic, paralinguistic, and extralinguistic

modifications that children make when communicating with

different partners, in different situations, and for

different purposes (Roth & Spekman, 1984). Greenfield

(1979) reported that children at the single-word level are

not necessarily aware of the listener's perspective, of what

might be "old" or "new" information for the listener. She

further stated that the power of a process of information

extraction common to child and adult is that it can make

verbal communication between a child and an adult possible

long before the child has developed any such awareness of

the listener's point of view.

According to Bloom and Lahey (1978), the learning of

the social conventions for the uses of language begins in

the third year but continues through adulthood. Early

development includes use of presupposition, listener-role,

conversational discourse between speakers, and forms of

reference, deixis, and ellipsis. Bloom and Lahey believe

that children between 3 and 4 years of age can modify form

and content of their speech according to what they


presume that their listener knows. Among the indicators of

the child's ability to take account of the listener's

orientation are: existence of relative clauses, presence of

identifying adjectives, use of orienting terms and

transition indicators, and adjustment of language complexity

(Lund & Dunchan, 1983). Four-year-old children have been

found to modify their speech as a function of the age of the

listener. These children produce fewer, shorter, and less

complex sentences when speaking to 2-year-olds as compared

to adults (Shatz & Gelman, 1973, cited in Prutting, 1979).

Presupposition and inference can account for

distinguishing between old (or given) and new information,

retrieving information deleted from elliptical utterances,

determining the referents of pronouns, and creating

information by combining elements from more than one source.

Conversational inferences occur when an individual assesses

another person's intentions and responds appropriately.

Inferring involves the ability to accurately evaluate the

social, cognitive, and communicative dimension of a

situation and then appropriately assign propositional and

illocutionary meaning to the utterances. One source of

conversational inferences is prosodic and paralinguistic

features (Rees, 1982).

Rowan, Leonard, Chapman, and Weiss (1983) did a study

in which the presuppositional and performative abilities of

language-disordered and normal children were compared,

controlling for the children's ability to use the lexical

items required in the experimental tasks. Results revealed

that both the language-disordered and the normal language

children showed a tendency to encode changing rather than

unchanging situational elements. The two groups of children

also demonstrated similar levels of imperative and

declarative performative intent. DeVilliers and DeVilliers

(1978) cited Volterra (1973) who reported that children

between 1.5 and 2 years are aware of the rules for simple

negations. This ability, they asserted, is one of the best

proofs of presuppositions. A speaker must be able to adjust

the form and content of an utterance according to the

situation and the listeners. The formality of the

situation; the relative social status of speaker and

listener; the age, sex, and culture of the communicative

participants; and the purpose of the communication are

factors that must be considered by the speaker in order to

be effective. Bloom and Lahey (1978) cited Smith's (1933)

study in which children were observed to ask significantly

more questions when they were interacting with adults than

when they were interacting with their chronological age

peers. By the age of 3, children can make reference to

absent persons and events (DeVilliers & DeVilliers, 1978).

Bloom, Lightbrown, and Hood (1975) reported that when MLU

was about 2.5, children knew that it was possible to use

both nouns and pronouns in reaction to verbs.


Roth and Spekman (1984) pointed out that, because there

are currently no formalized coding systems available that

address different aspects of role-taking, the clinician will

want to be sensitive to the informativeness of a child's

messages as well as to variables related to the social

context. They suggested that to evaluate this aspect of a

child's message, the following questions can serve as

guidelines: What does a child choose to talk about in a

given situation? Does the child encode what is novel or

merely comment on what is already given or known? Does the

child encode new information gesturally and/or

linguistically? Are messages informative, vague, or

ambiguous? Are different referents clearly established for

the listener? Does the child talk differently about objects

or events present in the environment than about those not


In addition to examining message information,

presuppositional abilities may be reflected in linguistic

devices, including dietics, indirect/direct reference forms,

and other forms of cohesion. Because the use of deixis and

indirect/direct reference is always bound to the particular

contexts in which it occurs, definitive developmental

milestones cannot be given. In fact, even adults frequently

use these terms incorrectly. Therefore, the purpose of

assessing these abilities is to obtain descriptive

information about dimensions of role-taking skills (Roth &

Spekman, 1984). Presuppositional abilities can also be

assessed by asking children to describe a movie, retell a

story, or describe how a toy works. These formats provide

units of extended discourse that can be analyzed for

introduction of referents and instances of deixis and

cohesion. Finally, a variety of role-playing activities can

be developed to examine role-taking abilities (Roth &

Spekman, 1984).


Topicalization involves the selection of a topic by the

speaker and the specification of that topic in sufficient

detail so that the listener will be able to follow. The

amount and type of topic specification required will be a

function of the amount of shared information that the

speaker and the listener can assume at a given moment in

conversation. The degree of topic specification required

may also influence the selection of the topic. In general,

however, topic selection emerges out of the flow of previous

discourse (Bates & McWhinney, 1979). What determines topic

selection when there is no given element? Bates and

McWhinney (1979) suggested several factors. First of all,

when all the elements in a proposition are equally novel,

speakers are particularly likely to choose as the starting

point the element with the greatest similarity to the

speaker. Secondly, topic selection is motivated by some

combination of at least three intentions; giveness,

perspective, and salience. Further, when the topic is old

information, definite articles and demonstratives are

devices typically used in topic specification. When,

instead, the topic is new information, indefinite articles

and quantifiers are more likely to be used. According to

Bates and McWhinney (1983), hierarchical structuring and

course of topicalization in discourse may include the

following: initial greeting, openers, logical development of

the topic, resolution of the topic, and closing remarks or

termination of the topic.

Topic coherence involves the ability to relate

utterances through the use of meaningful and structural

ties. Utterances can be tied to the previous speaker or to

the current speaker's own utterances. Topic continuation

involves the ability to maintain a topic, to elaborate on a

topic, or to introduce new topics with appropriate topic

shifts. Halliday and Hasan (1976) suggested that

utterances can be tied together through the use of

structural links they termed cohesive devices. Keenan and

Klein (1975) described multiple strategies used by a set of

twins during the early preschool years to sustain

interaction and topical focus in child-child conversations.

These strategies included simple repetition of the previous

speaker's utterance, affirmation or denial of the previous

utterance, repetition of the previous utterance with a

change in a single linguistic or paralinguistic feature,

extension of the previous utterance by predicating something

new, and including ties such as anaphoric pronouns,

conjunctions, and ellipsis. Conversational structure

consists of openings (greetings and exchange of amenities),

topic discourse (exchange of information about the topic,

each taking turns in an orderly way), and closings (Lund &

Dunchan, 1983). Sinclair and Coulthard (1975) discussed

three categories of utterances: initiations, responses, and

feedback. Initiations are those utterances in which the

child predicts or expects a response. The response is an

utterance in which there is a fulfillment of the

expectations of a preceding utterance. Feedback is a

response type in which neither of the above occurs but where

there is a reaction to the preceding utterance.

Brinton and Fujiki (1984) investigated the manipulation

of discourse topic in spontaneous conversation in three age

groups. They found that the number of topics introduced and

reintroduced in the sampling period decreased with age,

whereas the proportion of topics maintained increased with

age. Topic shading also increased with age. A topic was

considered to be shaded when the topic focus is not strictly

maintained, and while some aspect of the current

propositional content was derived from the preceding

utterance. Topic shading may occur in an utterance that

either adds or requests information.

According to Prutting (1979), one characteristic of a

child around 2.5 years old is that there is rapid topic

change throughout the discourse. Bloom and Lahey (1978)

suggested that one of the earliest techniques for monitoring

discourse exchanges begins when children seek to establish

that someone is listening before they begin to talk. They

reported that the use of the vocative to gain the attention

of a listener before continuing with an utterance has been

observed often among 2-year-olds. Prutting (1979) cited a

study by Bloom, Rocissano, and Hood (1976) in which they

found that somewhere around 3.5 years the child acquires an

ability to maintain a topic over adjacent successive

utterances. Rees (1982) cited studies by Keenan (1974) and

Garrad (1979) in her conclusion that children in their third

and fourth years use devices like anaphoric pronouns and

ellipsis to relate their conversational contributions to

prior linguistic materials. Somewhere between the age of

two and three years the child learns to produce utterances

that are relevant to preceding adult utterances, not only on

the basis of intimate context, but on the basis of the

actual linguistic content of the adult's utterance. The

child is able to repeat part of the adult's utterance to

respond, answer questions, use deixis, and apPropriately

substitute pronominal storms for previously introduced

nominal forms (McLean, Snyder-McLean, Sack, & Decker, 1981).


The examination of context requires attention to

characteristics of the listener, the communication channels

available, and the environment in which an interaction takes

place. A wide variety of audience variables warrant

attention including age, status, level of familiarity,

cognitive level, linguistic level, and shared past

experiences. To examine a child's appreciation of the

physical context, researchers might ask: Does the child make

modifications that reflect an awareness of the channels

available for communication? An additional question might

be: Does the child recognize that rules governing behavior

may change in different social environments such as home,

playground, and classroom? (Roth & Spekman, 1984). Camaioni

(1979) argued that a thorough study of conversational

competence must consider, not only dialogue exchange as

such, but also the physical, social, and cultural context in

which the conversation takes place. Ochs (1979) stated that

context includes minimally the language user's beliefs and

assumptions about temporal, spatial, and social settings;

prior, ongoing, and future actions (verbal, nonverbal); and

the state of knowledge and attentiveness of those

participating in the social interaction at hand. According

to Scollon (1979), the earliest stage of language has

traditionally been taken to be when an uninitiated speaker

of the adult language can clearly understand sentences

spoken by the child without reference to the context.

Scollon claims that this stage can be pushed back to an

earlier age when the adult speaker's understanding is

enhanced by interaction with the child and by close

attention to the contextual detail in tape recordings.

Researchers seem to agree that the scope of context is

not easy to assess and define. Ochs (1979) stated that the

researcher may have access to the immediate physical

environment in which communication takes place, and may have

access to the verbal environment in which a verbal act is

couched. However, she asserts that although these

dimensions of context are significant, they do not exhaust

the range of utterance-external variables that affect the

use and interpretation of verbal behavior. Ochs further

stated that in order to assess the import of a language

user's behavior, one must consider the social and

psychological world in which the language user operates at

any given time. She suggested that this world is shaped

both by culture-specific values and expectations, and by

cultural and interactional processes that affect language

users across cultures and languages. According to Ochs

(1979), what we have learned about child language and

context lies outside the scope of any one field. It draws

on insights achieved in several fields, including

psychology, anthropology, sociology, linguistics, and

philosophy. She concluded that accounting for context is an

awesome task, and that researchers are reluctant to attend

to levels of context that the methodology of their field

excludes. A consequence of this, in turn, is that

researchers have been unable to develop a unified theory of


According to Bloom and Lahey (1978), speech acts that

include different message content in different contexts,

such as requesting an action, requesting permission,

warning, inviting, or promising, are a later development;

they usually appear in the later preschool years (ages three

to five years) but continue to develop and change into

adulthood as the individual encounters new and different

contexts. Schwartz, Messick, and Pollock (1983) defined

pragmatics as the study of language use as a function of

context. They stated that contextual influences include

characteristics of the conversational participants as well

as characteristics of the particular communicative

situation. According to Schwartz and his colleagues, a

number of studies have demonstrated that both normal and

speech impaired children vary phonetic aspects of production

in relation to various contextual factors. Ochs (1979)

concluded that context cannot be ignored by language users.

She stated that to be competent, a language user must know

the multitude of norms for adapting language to the

situation at hand. According to Chapman (1981) and Rees

(1982), discourse has to do with the role that utterances

play in their contexts. This involves how utterances

operate to open, maintain, or close conversations; how

utterances fulfill the expectations set up by the previous

speaker utterances; and how utterances serve to clarify, or

change the topic of conversation.

Prutting (1982) stated that context is the interrelated

conditions in which something exists or occurs. She

discussed several dimensions of context. The cognitive and

social context involves the knowledge of the physical world

including the setting, the communicative partner, and the

rules of interaction. The physical context is concerned

with the perceptual properties of people, places, and

objects. The linguistic context involves the prior, co-

occurring, and past verbal behavior used in composing and

interpreting communication. Nonlinguistic context

is concerned with the nonverbal and paralinguistic behavior

in generating and interpreting meaning. The linguistic and

nonlinguistic dimensions are specifically involved in


Conversational Repairs

Conversational repairs involve the ability to

recognize, indicate, and clarify unsuccessful utterances.

The speaker must be able to recognize that a listener has

not understood the intended message, determine the possible

reasons for the breakdown, and clarify the message via

elaboration, repetition, or paraphrase. The listener's

responsibility is to communicate to the speaker that the

message has not been understood. He may use a number of

verbal and nonverbal means such as questions and quizzical


One device used by listeners to alert speakers to that

their messages are unclear is the contingent query (Garvey,

1977). The contingent query is a question addressed to the

speaker which signals his immediately preceding message was

unclear or that additional information is desired or needed

in order for the message to be fully understood. Garvey

identified four kinds of contingent queries used by three-

to-five year-olds: nonspecific request for repetition,

specific request for repetition, specific requests for

confirmation, and specific requests for specification.

Garvey found that children do not begin to use contingent

queries productively until around three years.

James and Sebach (1982) investigated the pragmatic

function of preschool children's spontaneously produced

questions. Questions produced during these observation

periods were categorized by pragmatic function. The major

function of the questions produced by the two and three-

year-olds was clearly information seeking, but the tour- and

five-year-olds' questions were more evenly distributed among

the functional categories. The four-year-olds used a high

percentage of conversational questions in comparison to the

other age groups. The children's question use appeared to

follow the principle of using new forms for old functions

and old forms for new functions. Prutting (1979) cited a

study by Gallagher (1977) who investigated children between

the ages of 21-29 months with an MLU of 1.5 to 2.9. She

found that all children responded to requests for revisions.

She also found that these revision behaviors were systematic

and that the type of change was a function of the child's

linguistic sophistication.

Roth and Spekman (1984) discussed three major

categories of revision strategies: linguistic structure,

linguistic content, and extralinguistic factors. Linguistic

structure includes phonologic, morphologic, lexical, and

syntactic factors. Linguistic content involves such

strategies as repetition, confirmation, specification, and

elaboration. Extralinguistic factors include pitch change,

stress, volume change, and demonstration. Terrell (1985)

cited a study by Beal and Flavell (1983) in which they

showed that children in preschool and kindergarten had

difficulty rejecting listener's inaccurate statements

regarding their comprehension of messages. This was true

even when the children themselves were aware of the quality

of the messages. First graders were able to use their

evaluations of message quality to make decisions about

whether or not listeners actually understood.


According to Roth and Spekman (1984), there is a large

body of evidence that demonstrates that very early mother-

child interactions involve synchronous alterations of turn

and temporally linked behaviors. Turn taking behavior is

also seen in the highly ritualized and repetitive games

played between mothers and their infant children. Roth and

Spekman suggested that the presence or absence of turn-

taking should be documented during assessment. In dialogue,

infants very early discover the essentials of turn-taking,

how to achieve affective synchrony in interaction with

others, the capacity to predict the behavior of others, and

the capacity to intentionally influence another's behavior

(Bretherton, McNew, & Beegley-Smith, 1981). By the age of

14 months, children can initiate vocal turn-taking sequences

with their mothers (Dunn & Kendrick, 1982). Uinio and

Bruner (1978) reported that children aged 8-18 months

produced nearly perfect turn-taking cycles. Bloom,

Rocissano, and Hood (1976) noted that the child during the

one-word stage shows the rules of turn-taking and

systematically applies them in conversational exchanges with

a partner.

The allocation of turns involves a knowledge of the

rules used to initiate and coordinate the speaker's

participation in conversation. Initiation involves the

ability to begin a speaking turn as well as to provide

opportunities for others to participate. It requires that

conversational participants recognize turn requests, such as

questions. Nonverbal cues and/or verbal statements may be

used. Ervin-Tripp (1979) found that children between two

and four years old produce more turn-relevant interruptions

than younger children. Children over 4.5 were less likely

to ignore overlaps and when interrupted they attempted to

remedy the situation more often than younger children. It

appears that turn-taking is an important and natural

occurring phenomenon that may not be fully developed until

the child is well advanced in general communicative



Garvey (1975) studied three-to-five year-olds' ability

to respond to different forms of requests for action. Both

3.5 and 4.5 year-olds responded better than younger children

to indirect requests, but 4.5 year-olds responded better

than younger ones to indirect requests. Ervin-Tripp (1977)

reported that preschool children use indirect forms like

questions to make requests. James (1978) suggested that the

use of indirect forms to achieve politeness is well

developed by the time a child is five years old. Lund and

Dunchan (1983) discussed Ervin-Tripp's (1977) study where a

developmental sequence for directives was presented. Two

year-olds use attention getting words along with gestures

and rising intonation, request words, problem statements,

and verbal routines. Three year-olds used modal forms

(would you + request). Four year-olds used hints with a

statement of the problem. Five year-olds made requests

distinct from the intent ("Pretend this is my car").

Lund and Dunchan (1983) described an unpublished study

by Lawson (1967) of a two-year old child who used different

degrees of directness and politeness when making requests to

peers as opposed to adults. According to Lund and Dunchan,

the different strategies included embedded imperatives,

permission statements, hints or implications, and

manipulations. Children between three and five years old

begin to produce indirectives or hints in their

communicative exchanges (Ervin-Tripp & Kernan-Mitchell,

1977). Utterances such as "My mother always lets me have

ice cream before lunch" are used to request indirectly

rather than to inform (Prutting, 1979). When using

directives, four and five year-olds are more polite when

making requests than when commanding, and they include more

polite markers when addressing adults and age peers than

when addressing younger children (James, 1978). When

addressing younger children, preschoolers demonstrate many

of the characteristics of mothers' speech to children. They

simplify their speech, use shorter and less complex

sentences, use frequent repetition, ask more questions, use

more attention-getting devices, and simplify their phonology

(Sachs & Devin, 1976; Shatz and Gelman, 1973). As they

modify the style of their speech, children are not only

aware of age differences in listeners but also of

differences in developmental status. Bates (1971) suggested

that comprehension of polite forms precedes production of

them. "Please" and "intonation items" were reportedly

recognized by the age of four. Sachs and Devin (1973)

reported that four-year-olds varied politeness depending on

whether talking to adults, peers, younger children, or baby

dolls. In two experiments, Nippold, et al. (1982) examined

children's use and understanding of polite forms and the

relationships these abilities share with an ability to adopt

the perspective of another. Development in the

understanding of polite forms seemed best characterized as a

sensitivity to the function of these forms in an increasing

number of sentence types. Development in the use of polite

forms could be viewed as an increasing tendency to produce

more than one polite form in the same utterance and to use a

wider variety of interrogative types when expressing

politeness. Emergence of the use and understanding of

politeness in speech at age three seemed related to an

emerging ability to adopt the perspective of another.

However, development of politeness beyond this age seemed

related only to a form of perspective taking in which

knowledge of social roles plays a part.

Discourse Skill Development of Language-Impaired Children

Lucas (1980) discussed five major areas of specific

semantic/pragmatic disorders. Topical or referential

identification problems are most significant with respect to

asking and answering questions. Off-target responding

occurs when a child is unable to scan the previous

utterances or the surrounding context for appropriate

referential possibilities. Topic closure difficulties occur

when the child cannot determine the boundaries of his

utterances in terms of ideas or topics. Boundaries are

arbitrary divisions around ideas or semantic bits of

information that a speaker wishes to convey to a listener.

Pauses and other prosodic devices are used to linguistically

mark the division between the boundaries.

Most children acquire the language of their environment

with considerable ease. A significant minority however,

acquires language slowly and with great effort. Children

experiencing difficulties learning language often exhibit

other types of handicaps that may relate or contribute to

their problems with language. There is one group of

language deficient children however, for whom this is not so

clearly the case. These children manifest what has been

termed specific language impairment (Leonard & Schwartz,

1985)). This condition is seen in approximately 1 of every

1000 children. A number of other terms have been used in

the literature to refer to children with specific language

impairment. These include: "delayed language," "infantile

speech," "deviant language," "language disorder," and

"congenital," "developmental," or "childhood" aphasia

(Leonard & Schwartz, 1985). The term language-impaired will

be used to describe this population of children throughout

this research document. Although language-impaired children

constitute a heterogeneous group, they share several key

characteristics. These children show significant deficits

in linguistic functioning although they have adequate

auditory acuity, age-appropriate performance on nonverbal

tests of intelligence, and the absence of gross neurological

disability (Leonard & Schwartz, 1985).

According to Fey and Leonard (1983), children with

language impairment have traditionally been identified on

the basis of deficits in lexical and syntactic development,

and studies attempting to provide a description of these

deficits have focused on such factors as the modality that

is most affected. Fey and Leonard stated that although the

lexical and syntactic abilities of language-impaired

children are still a major concern, investigators recently

have broadened their perspective on language impairment to

include not only considerations of linguistic form, but also

of the ways in which language-impaired children use

language. Several studies were reviewed by Fey and Leonard

(1983) that examined the ability of language-impaired

children to participate in conversation. Conflicting

conclusions can be drawn from the results of the

investigations cited. Fey and Leonard concluded tnat the

population of language-impaired children is no more

homogeneous with respect to pragmatic skills than it is with

respect to any other aspect of language. Fey (1986)

discussed three patterns of conversation identified by Fey

and Leonard (1983) in an earlier study of the language-

impaired population. Children were categorized as being

either assertive, responsive, or neutral. Assertive

children are highly verbal and tend to make functional use

of the limited linguistic skills that they possess. Other

language-impaired children are responsive to their

conversational partners but are not likely to initiate

conversations and/or topics. Children in the third group

display general deficits in their patterns of social-

conversational interaction. They are neither as assertive

nor as responsive as children developing normally, and they

exhibit a reduction in motivation to interact verbally with

their social partners.

Craig and Gallagher (1986) investigated the potential

relationship between interactive play and the frequency of

related responding to comments within the dyadic

interactions among a five-year old boy and normal language

users. They found that the language-impaired boy's

conversational behavior was different from that of his same-

aged or language-similar peers. The data as a whole seemed

to indicate that unlike normal language children whose

frequencies of related responses were essentially stable,

the language-impaired boy was more frequently able to

produce a related response after the referent had been

explicitly verbally identified by both he and his partner.

Fey and Leonard (1984) investigated partner age as a

variable in the conversational performance of language-

impaired and normal-language children. They found that the

language-impaired subjects were as assertive as their same-

age peers on all accounts, and that they demonstrated signs

of being even more active, in general, than were the younger

language-matched subjects. The researchers concluded that a

general characterization of language impaired subjects as

conversationally unassertive is, at best, premature and, at

worst, incorrect. The language-impaired subjects also

demonstrated the ability to modify their speech styles when

speaking to different-aged partners in the same manner and

to the same extent as the same-aged subjects in all areas

except internal-state questions and utterance length and

complexity. The investigators felt that their findings

corroborated the results of Meline's (1978) and Skarakis and

Greenfield's (1982) studies which showed that the typical

higher cognitive, social, and language comprehension skills

of language-impaired children may give them advantages on

certain pragmatic tasks over normal language children who

are roughly equivalent in expressive language ability.

Tomasello, Farrar, and Dines (1984) cited a study by

Gallagher and Darton (1978) in which they found that

language-impaired children have the ability to adjust their

speech in response to listener feedback. The subjects

revised their utterances with a high frequency, although the

frequency with which they used different types of revisions

was different from that of normal children in the same

stages. They used mixed revision strategies. The

strategies were not related to their levels of structural


Van Kleeck and Frankel (1981) discussed a study in

which Keenan (1975) concluded that their are two strategies

operating during children's conversations. The focus

operation is where one or more lexical items in an

antecedent utterance is focused on and repeated in a

subsequent utterance. These kinds of utterances require a

minimal level of competence. The other strategy,

substitution operations, involves repeating part of the

antecedent utterance but replacing a lexical item. In

substituting, the child begins to go beyond reproducing what

he hears or sees. Like the adult, the child at this point

begins to draw on his background knowledge to respond

relevantly to some prior utterance. In a study of language-

impaired children's use of these two devices through which

utterances are related to ongoing discourse, Van Kleeck and

Frankel (1981) reported that children who were at lower

linguistic levels used more focus operations while the

linguistically more mature child used more substitutions.

Contrary to expectations, however, the language-impaired

children were able to appropriately use substitution

operations even at the lowest language level observed. In

addition, a developmental trend was suggested because the

use of substitutions was greatest for the child with the

highest language level. Consequently, no noteworthy

qualitative differences were apparent between normally

developing and language-impaired children in the discourse

devices analyzed.

Friel-Patti and Conti-Ramsden (1982) cited Snyder

(1976, 1978) who developed a task to elicit declarative

performatives, imperative performatives, and encoding of the

most informative element in the context. Snyder found that

language-impaired children were deficient in their ability

to generate linguistic performatives, encode the most

informative contextual element, and disengage themselves

from the immediate context. Although both the language-

impaired and the language-normal groups were similar in

their nonverbal use of communicative behaviors expressing

declarative and imperative functions, the language-impaired

children were far less likely to express these intentions


Most of the emphasis on discourse research with

language-impaired children has not extended beyond the one-

word stage. Friel-Patti and Conti-Ramsden (1982) stated

that researchers who study the discourse abilities of

preschool language-impaired children have generally selected

a group of language-impaired children and compared them with

their chronological age-mates and/or MLU matches. The

writers pointed out that there have been no systematic

attempts to discover developmental patterns in the language-

impaired population. Additionally, they asserted that

research with the language-impaired population is limited in

scope and limited by various methodological problems.

Conti-Ramsden ana Friel-Patti (1983) found quantitative

differences in the amount of dialogue participation of

language-impaired children and language-normal children of

the same language stage. Specifically, the language-

impaired children initiated dialogue and introduced topics

less often than did the language-normal children. The

researchers reported that qualitative differences in the

type of responses used by language-impaired children in

dialogues with their mothers were found. Language-impaired

children had more inadequate and ambiguous responses to

mother's initiations than did the language-normal children

of the same language stage. Brinton and Fujiki (1982) found

similar results in child-to-child interactions. Friel-Patti

and Conti-Ramsden (1982) cited a study by Geller and Wollner

(1976) in which they compared the speech act repertoires of

three language-impaired children with those of age-matched

subjects reported by Dore (1977). The language-impaired

subjects were notedly deficient in the range of speech acts

they used.

Donahue, Pearl, and Bryan (1980) and Donahue (1984)

examined aspects of a model for responsive listening. In

both studies, the child's task was to listen to a clue

provided by an experimenter and then select one of four

pictures of the same object, which differed in varied

features. Some messages were completely informative,

providing enough details for the child to identify just one

picture. On the other trials, the experimenter created a

communicative breakdown by giving general clues that applied

to two pictures or to all four pictures. Donahue and

colleagues (1980) found that children in grades one through

eight had no difficulty identifying the correct picture

after hearing the informative clues. However, both younger

children and language-impaired children were less likely

than their comparison groups to request clarification of the

inadequate messages. As a result, they made significantly

fewer correct picture choices. Gale, Liebergotti, and

Griffin (1981) found that language-impaired children used

more nonverbal requests and a restricted number of

linguistic forms when making requests for clarifications

than did language-normal children at the same language

stage. Courtright and Courtright (1982) compared language-

impaired and language-normal children in terms of their

ability to interpret emotional meaning from the vocal cues

of an adult speaker. This ability is an important factor in

a child's pragmatic language functioning. The findings

suggested that language-impaired children were less

sensitive than language-normal children to vocal cues.

However, these results did not address the issue of whether

language-impaired children are actually simply less

sensitive to such vocal cues or whether they provided

different response patterns to these emotions.

Donahue, Pearl, and Bryan (1980) presented results that

indicate that the productive language deficits may be

sufficiently significant to interfere with even the informal

and elliptical conversations characteristic of communication

among peers and family members. Graybeal (1981) compared

the ability of language-impaired children and their

chronological age-mates to recall stories read to them. She

found that the language-impaired group recalled

significantly fewer ideas than their normal controls. The

language-impaired did not differ from normals in the

ordering of the story information, the amount of plausible

information added to the story, or the number of errors in

recall. Lucas (1980) believes that the basis of pragmatic

problems is a semantic disability. She suggested that

children who have limited language repertoires may not

understand that communication is useful. Wiig and Semel

(1984) outlined some of the difficulties that learning

disabled children exhibit in the area of social perception.

They fail to perceive, interpret, and appropriately use

facial expressions, gestures, and body language.

Methodological Issues in Studying Language-Impaired Children

According to Kirchner and Skarakis-Doyle (1983),

although the understanding of language disorders has

broadened considerably over the past two decades, the

results of research in this area are fragmentary and

contradictory. They agree with Johnston (1982) that the

only consistent conclusion from research in this area is

that learning-delayed children may learn to speak slowly and

late, but little else about their language has proven


Gallagher (1983) stated that spontaneous language

sampling is the centerpiece of child language assessment.

According to Gallagher, it has been recommended as one of

the main sources of information about children's language

for over five decades. Language analysis procedures have

changed over time. The procedures for obtaining a language

sample, however, have remained basically unchanged since

they were first described by McCarthy (1930). Gallagher

points out that one of the concerns in gathering language

samples has been the issue of collecting a representative


sample. She believes that this issue is complicated by the

fact that language use varies with context. Any language

sample would be the product of a combination of contextual

variables and the child's knowledge of language structure.

The major response to this dilemma has been to obtain

language samples in more than one context. Miller (1981)

suggested that three 15-minute language samples be obtained:

one with the mother in free play, one with the clinician in

free play, and one with the clinician directing the child

with questions and commands. In addition, he suggested that

a sample could be collected with a playmate or a sibling.

McLean and Snyder-McLean (1978) recommended that language

samples be obtained in a "warm" atmosphere. This, tney

suggested, might be the home, the classroom, or a playroom

with someone familiar to the child present. They further

suggested that the assessment should consist of three

observation sessions scattered over several days. They

stated that the sessions do not need to differ in terms of

the communicative partner, materials, or setting. Muma

(1978) also suggested that three language samples should be

collected. Bloom and Lahey (1978) recommended that the

child's language be sampled in more than one setting if that

is possible.

According to Gallagher (1983), directly comparable data

regarding physical context variables are not available for

language-impaired populations. Based upon their clinical

experience assessing language-impaired children, however,

Bloom and Lahey (1978) suggested that a greater number of

utterances and a greater variety of utterances can be

obtained by structuring the clinician-child interaction

around a concrete activity using items such as science

experiments, art projects, and construction toys. Stalnaker

and Creaghead (1982) conducted a study in which they

examined three experimental conditions to determine if there

were significant differences in the quality and quantity of

language produced in these situations when compared to each

other. There were three test conditions: retelling a story

with toys; playing with the toys; and playing with the toys

with questions from the examiner. The latter condition

produced a larger number of total utterances. The

researchers suggested that some kind of adult contribution

such as asking questions or telling a story may give the

children more to talk about and keep then from getting so

involved with the toys that they do not talk. The results

of the study did not demonstrate that one method of

obtaining a language sample was superior, but gave support

to the notion that methods need to be varied among children

and provided suggestions for choosing a strategy based on

the child's manner of responding.

Simon (1985) suggested a format for a functional-

pragmatic evaluation procedure and presented several

considerations that should be taken into account when

developing an assessment strategy for language-impaired

children. The procedure should be comprehensive and

descriptive and yet meet time constraints within a school

setting; therefore, it must have structure. There should be

a series of evaluation tasks that stimulate communicative

contexts and probe metalinguistic awareness. It should

include informal evaluation procedures which allow the

clinician to apply recent research immediately. The

informal procedures must have standard and consistent

instructions and materials. In the past two decades the

speech-language pathologist has become increasingly

sophisticated in the assessment of several areas of language

functioning, including analyzing the child's syntax and

evaluating semantics. The ability to select lexical items

and to structure them appropriately is important, but the

ability to make good sentences can be useless unless the

child understands how and when the sentences can be used

(Johnson et al., 1984). Thus, a complete language

evaluation must include tasks that will examine these

aspects of oral communication.

Gallagher and Prutting (1983) discussed a protocol that

is organized around four areas: the utterance act, the

propositional act, the illocutionary act, and the

perlocutionary act. Dale (1980) developed measures of

pragmatic development for children in the second year of

life. He concluded that the range of pragmatic functions in

a child that age is measurable and contributes information

not provided by a measure of syntactic development alone.

Stalnaker and Creaghead (1982) found that preschool children

produced the longest MLUs when they retold a story using

objects to act it out, compared to two other toy

manipulation conditions. The literature as a whole suggests

that every aspect of physical context probably can, but will

not necessarily, affect a particular child's use of

language. However, it is clear that the number of

contextual variables that have the potential of influencing

children's language use is too great to suggest that all

contexts should be sampled during each language assessment.

Sampling across a few contextual environments, however,

meets the practical requirement of efficiency (Gallagher,

1983). For several years, investigators have documented the

effects of physical context variables on the syntactic

performance of children. Only a few studies however, have

examined the effects of these variables on pragmatic

performance, and few, if any, have included an analysis of

topic categories or subject matter. In view of the current

emphasis on the use of play in language assessment and

intervention in the area of pragmatics, documentation of the

effect of specific play situations on the types of topics

discussed by children is warranted (Wanska, Bedrosian, &

Pohlman, 1986).


The term discourse topic appears frequently in current

research on language but investigators have not agreed on a

precise definition (Brinton & Fujiki, 1984). Bedrosian

(1985) described topic as the proposition or set of

propositions or subject matter about which the speaker is

either providing or requesting new information. Once a

topic has been initiated by a participant, continuous

discourse constitutes those subsequent turns that are linked

in some manner to the topic that was initially introduced.

Chapman (1981) suggested that the types of continuous

discourse turns used by a child may offer important

developmental information regarding his or her level of

communicative competence. Bedrosian (1985) asserted that

topic offers an all encompassing framework for viewing

communication skills and reported that it has been suggested

that researchers code the communicative intent of each topic

initiation. She explained how this procedure may help in

identifying those clients who are having difficulty in

initiating and maintaining topics. Brinton and Fujiki

(1984) offered a description of Keenan and Schieffelin's

(1976) model for the establishment of topic, topic

continuation, and topic shading. They suggested that

language analysis should focus on the number of topics

introduced, the number of topics reintroduced, the number

and proportion of topics maintained, the length of topic

maintenance, and the number of topics shaded.

Although the study of isolated variables in a small

number of subjects raises several methodological concerns,

the use of small numbers of subjects with careful attention

to research design may in some cases yield very significant

data. McReynolds and Thompson (1986) pointed out that our

profession has gained valuable information from studies of

single subjects. They stated that descriptive studies of

the speech and language behavior of one subject abound in

our literature. As an example, they cited the work of Paul

Broca (1869) which involved study of only one subject, the

results of which continue to influence current theory and

research regarding the brain and its relationship to


Another example of small N research that has had

widespread impact on what is currently known about language

is Halliday's (1975) study of his son Nigel. Bloom and

Lahey (1978) reported data based on Bloom's observations of

her daughter Allison. Ninio and Bruner (1978) used

Greenfield and Smith's (1976) data on one child to report on

the longitudinal development of indicative and volitional

performatives. Several other investigators of child

language have used sample sizes of ten children or less to

conduct studies that have contributed significantly to our

understanding of child language development (Dore, 1977;

Bloom, Rocissano, & Hood, 1976). McReynolds and Thompson

(1986) concluded that these small N studies have had a

strong impact on our understanding of speech and language

development and disorders. Among the practical advantages

of using small N studies which they discussed is that this

type of design may sometimes be more economical in terms of

time, and that researchers are not usually faced with the

sometimes impossible task of locating large numbers of

allegedly homogeneous subjects with a particular

communication disorder, randomly selecting a sample for

study, and assigning them to the required groups.

There are three basic types of small N studies: diary

studies, single-subject experimental studies, and case

studies. Diary studies are usually longitudinal in nature.

Researchers study some aspects) of speech and language

development for several months or years. Single-subject

experimental studies generally explore the efticacy of

selected treatment approaches. They involve small numbers

of subjects who receive several variations of treatment in

an attempt to determine advantages and disadvantages of

using the experimental treatments for each subject in the

study. Case studies are observational research projects

that involve detailed analysis of behavior that occurs

during a specified period of time. While there are unique

advantages for each of these types of studies, McReynolds

and Thompson (1986) point out that the case study approach

has prevailed throughout the history of our profession. The

1983 issue of Topics in Language Disorders was devoted

exclusively to articles about case studies of phonological

disorders. Writers of the various articles in the volume

emphasized the individuality of children as language

learners and stated that as researchers acquire information

from careful and extensive examination of one child's

methods of acquisition, a data base will eventually emerge

which may have universal application in the study of

language development and disorders.

Statement of Need

Prutting and Kirchner (1987) suggested that future

research should address the performance of well-defined

clinical groups which are matched on diagnostic profiles to

extract patterns of clusters or dimensions on which the

subjects perform well or poorly. They believe that this

kind of research would allow us to better understand the

nature and impact of a pragmatic deficit in a population of

disordered subjects based on pattern analysis from

relatively homogeneous groups. They further suggested that

an indepth descriptive account of linguistic and cognitive

performance should yield data that could be used that will

emerge as strengths and weaknesses at the pragmatic level.

To date, no research investigation has presented data

on the multiple dimensions of pragmatics in language-

impaired children. Previous studies have examined only

isolated aspects of pragmatics (i.e., conversational

repairs, presupposition, communicative functions).

Therefore, researchers in the area of child language do not

have a clear picture of the linguistic strengths and

weaknesses in this group of children. Furthermore, it is

not known if pragmatic deficits occur in isolation, in

conjunction with other deficits or in variable patterns.

Further, no previous research has presented

comprehensive analyses of the multiple dimensions of

language in language-impaired children who have been

diagnosed as having a primary deficit in the domain of

pragmatics. Lastly, no study has used a language processing

model as a basis for determining the dimensions of language

to be included in a comprehensive study of this nature.

There is a need to develop comprehensive communicative

profiles of language-impaired children with primary

pragmatic deficits (Prutting and Kirchner, 1987).

According to Ochs (1979), expanding the study of child

language to include pragmatic dimensions of communication

enables us to see the variety of communicative skills the

young child uses in everyday social situations. She feels

that this, in turn, provides a richer and more accurate

account of the child's linguistic and social knowledge.

There is no standardized method of coiiecting language

samples, and no definitive evidence regarding the most

effective methods tor determining the child's maximum

potential (Stalnaker & Creaghead, 1982).

Over several years, investigators have documented the

effects of physical context variables on the syntactic

performance of children. However, few studies have examined

the effects of these variables on pragmatic performance, and

few, if any have involved an analysis of topic categories or

subject matter. Although little research regarding the

effects of physical context variables on the discourse

performance of normal and language-impaired children has

been conducted, the consideration of these variables in

language assessment and intervention is warranted

(Bedrosian, 1985).

This study examined the conversational skills of

language-impaired children who had been diagnosed as having

primarily pragmatic deficits. Extensive analyses of the

language skills of the children in this study were performed

on semantic, syntactic, phonologic, and pragmatic dimensions

of language. A carefully selected battery of standardized

language test measures was administered to each child.

Additionally, several informal analyses of various aspects

of communicative function were conducted on language samples

of each child. A descriptive case study approach was

selected as the most appropriate research design for this

study. The study examined the children's language

performance on standardized tests and in the context of

language sampling procedures. The purpose of this study was

to attempt to describe and isolate the specific pragmatic

variables that are impaired. This study sought to develop a

profile of pragmatically disordered children that should aid

speech-language clinicians in the assessment of language-

impaired children and in developing appropriate targets for


The experimental questions were

1. What types of communication profiles do language-

impaired children with a primary deficit in the

area of pragmatics exhibit when their language is

evaluated using a specific theoretical model?

a. Are there any patterns of deficits observed

among the language-impaired i.e., do some show

strictly pragmatic deficits while others show

deficits across domains?

b. Are there specific deficit patterns within the

domain of pragmatics?

2. Does an autistic child who is pragmatically

disordered present a linguistic profile that is

different from the profiles observed in non-

autistic children?


The purpose of this study was to conduct detailed

analyses of the conversational performance of young children

who had been identified as poor conversationalists, in an

attempt to describe and isolate the specific pragmatic

variables that were impaired. The researcher developed

indepth profiles of children who had pragmatic disorders, in

order to quantify and qualify the nature of pragmatic

disorders and to examine pragmatics in relation to other

dimensions of language such as phonology, semantics, and



The subjects were six black male children attending

public schools in Gainesville, Florida. There were three

case studies, each of which was comprised of a language-

normal and a language-impaired child. The language-impaired

children were matched with the language-normal children on

the basis of nonverbal intelligence, chronological age,

socioeconomic status, sex, and race. The children ranged in

age from 7:5 to 11:7 years old. The subjects had been

diagnosed by an American Speech-Language and Hearing

Association (ASHA) certified speech-language pathologist as

language-impaired with a primary deficit in the domain of


A.D., an 11-year-old high functioning autistic black

male, was in the fifth grade in a regular public school

program. He was served by resource teachers twice each day.

In 1986 A.D. was administered the Kaufman ABC Test of

Intelligence. His performance resulted in a nonverbal IQ

score of 100. His initial diagnosis of autism was made when

he was two years old and further confirmed when he was

tested at the Children's Mental Health Unit, Shands

Hospital, University of Florida. A.D. received speech and

language therapy for two years at the University of Florida

Speech and Hearing Clinic. He was also enrolled in an

intervention program at the Children's Mental Health Unit,

Shands Hospital, for three years. He is no longer enrolled

in either program. R.A., an eleven-year-old black male,

served as A.D.'s language-normal control. He was also in

the fifth grade in a regular public school program. He was

administered the Otis-Lennon Test of Intelligence in 1986.

His performance resulted in an IQ score of 102. R.A. had

never been the recipient of special services. Both children

were being reared by college-educated parents and were

involved in a variety of community and church-related


B.N., a seven-year-old language-impaired black male,

was enrolled in kindergarten in a regular public school

program. He was receiving speech and language therapy twice

per week at the school he was attending. B.N. was

administered the Weschler Intelligence Scale for Children-

Revised in 1988. His performance resulted in a nonverbal IQ

score of 92. D.J., a seven-year-old black male, served as

B.N.'s language-normal control. He was in the second grade

and had not been the recipient of any special services.

D.J. was administered the Test of Nonverbal Intelligence in

1988. His performance resulted in a nonverbal IQ score of

97. Both children were from low socioeconomic backgrounds.

T.M., a seven-year-old language-impaired black male,

was in the first grade in a regular public school program.

He was receiving speech and language therapy twice per week

at the school he was attending. T.M. was administered the

Weschler Intelligence Scale for Children-Revised in 1988.

His performance resulted in a nonverbal IQ score of 99.

G.N., a seven-year-old black male, served as T.M.'s

language-normal control. He had not been the recipient of

any special services. He was administered the Columbia

Mental Maturity Scale in 1988. His performance resulted in

a nonverbal IQ score of 100. Both children were from low

socioeconomic backgrounds.

All subjects met the following selection criteria:

1. All of the children exhibited nonverbal intelligence

within a normal range as measured by a standardized

intelligence test;

2. All children exhibited normal structure and function of

the speech mechanism on the basis of results of

administration of the Dworkin-Culatta Oral Mechanism

Examination (D-COME) (Dworkin & Culatta, 1980);

3. All children passed a hearing screening test at the 25dB


4. All children had normal or corrected-to-normal vision;

5. All children came from homes where English was the only

language spoken;

6. All language-impaired subjects were diagnosed as

language-impaired on the basis of standardized language

testing which was conducted less than 12 months prior to

the onset of the study; and

7. All language-impaired children had been diagnosed as

presenting with pragmatic deficits by an ASHA-certified

speech-language pathologist.

Assessment Model

The areas for assessment were directed by the Child

Language Processing Model (CLPM), which was developed by

Aram and Nation (1977). The tests that were used based on

the model are presented in Figure 2-1. The CLPM is a


language processing model comprised of three segments which

suggest the processes by which children receive, comprehend,

repeat, formulate, and produce messages. The model is based

on the assumption that disordered language behavior is a

reflection of disrupted language processes within its three

segments. Language disorders are considered to be a

reflection of disrupted underlying processes. The model is

comprised of three interrelated processing segments. Each

segment is made up of a series of processing stages. The

first segment is the Speech to Language Processing Segment.

It involves prelinguistic processing of the language message

before it becomes meaningful. There are three processing

stages within this segment, sensation, perception, and

repetition. Sensation is concerned with receiving the

intensity and frequency of sounds that are produced.

Auditory acuity is the area that was assessed in this stage

of the model. The speech perception stage involves the

transformation of acoustic information into the speech code.

Repetition involves the ability to reproduce verbal

information. The areas assessed at this stage include

auditory discrimination, auditory memory, auditory

sequencing, and repetition.

The second segment of the model is the Language to

Thought to Language processing segment. It is concerned

with the interpretation of the speech code as processed by

the first segment. There are four processing stages within

this segment of the model: speech and language repetition;

language comprehension; language integration; and language

formulation. Language comprehension is the processing stage

in which the child understands the words and grammar that

the child hears but does not necessarily fully interpret the

complexities of meaning that are embedded in the message.

The integration stage links comprehension and formulation in

the CLPM. The process of integration permits the listener

to understand the pragmatics of the language heard and to

relate what is heard to past experiences, to information

from other sensory modalities, and to other aspects of

cognition. The final stage in this segment, language

formulation, is concerned with the selection and retrieval

of words and the organization of these words into a

logically acceptable and syntactically appropriate form

(Aram and Nation, 1977). At this stage, the speaker's

intentions and messages are translated into language

structure. The Language to Speech processing segment is the

final component of the model. It is primarily a

postlinguistic segment that has two stages, speech

programming and speech production.

General Procedures

There were two basic data collection procedures, norm-

referenced testing and language sampling. These procedures

will be discussed separately in detail below. Each child


was seen for four test sessions within a three-week period.

All norm-based tests and procedures were administered during

the first three sessions. Each of the three sessions lasted

approximately 60-90 minutes.

The norm-referenced tests were administered and scored

by the experimenter and a graduate student clinician under

the direct supervision of an ASHA certified speech-language

pathologist. The graduate clinician was involved in

pre-test training sessions with the experimenter to ensure

consistency and proficiency in administration and scoring of

each of these tests. All tests were administered and scored

according to standardized instructions presented in the

examiners' manuals. The hearing screening procedure was

performed in accordance with ANSI (1969) standards. During

all sessions the children sat at an appropriately sized

chair and table. All testing was conducted in well-lit and

relatively quiet areas of the various schools and test


A language sampling procedure was used to examine

several aspects of communicative performance. The sample

was collected during the fourth test session. While the

language sample was being collected, the subjects in this

study were involved in three activities: "Script-Based

Play," "Elicited Schemes Play," and "Barrier Task Play."


A portable videorecorder and camera were used to record

all data collection sessions. A portable audiometer,

calibrated in January, 1988, was used to conduct the hearing

screening tests.

Norm Referenced Tests

Rationale. The purpose of using norm-referenced tests

was to obtain valid and reliable standardized test scores

reflecting each child's current level of functioning in

several aspects of language including auditory skills,

vocabulary comprehension, syntactic comprehension and

expression, phonological production, and pragmatic

comprehension and expression. Norm referenced measures

allow a direct comparison of one child's behaviors to those

of other children (Newhoff and Leonard, 1983). The specific

tests that were used in this study were selected on the

basis of their available reliability and validity data and

on the basis of their general acceptance and use among

speech-language pathologists. The tests were also selected

to represent the various domains of language (semantics,

syntax, phonology, pragmatics) as discussed earlier in the

assessment model.

Test Battery: Procedures and Analysis

The test battery consisted of the hearing screening

test; the D-COME; The Peabody Picture Vocabulary Test-

Revised (PPVT-R) (Dunn & Dunn, 1981); The Test of Auditory

Comprehension of Language-Revised (TACL-R) (Carrow-Woolfolk,

1985); The Goldman-Fristoe Test of Articulation (Goldman &

Fristoe, 1986); The Kahn-Lewis Phonological Analysis (Kahn &

Lewis, 1986); The Token Test for Children (DiSimoni, 1978);

The Test of Word Finding (German, 1986); The Weschler

Intelligence Scale for Children-Revised (WISC R) (Weschler,

1974) and selected subtests from the Goldman-Fristoe-

Woodcock Auditory Skills Test Battery (Goldman, Fristoe, &

Woodcock, 1974). The processing segments that comprise the

CLPM and the test procedures chosen to correspond to each

segment of the model are presented in Table 2-2. Following

is a description of each of the tests and subtests that were

used in the study.

The PPVT-R is a norm-referenced test designed to

measure receptive vocabulary for Standard American English.

The PPVT-R utilizes a picture selection format. The

examiner says a word and the child selects the appropriate

picture from four black line drawings on a single page. It

was normed on individuals from age two and one-half years to

adulthood. A raw score was obtained following the

procedures specified in the examiner's manual. The raw

score was converted to a language-age equivalent.

The TACL-R measures auditory comprehension of word

classes and relations, grammatical morphemes, and elaborated

sentence constructions. It has normative data for children


aged three through nine with guidelines for use with adults.

Each page consists of three black line drawings from which

the child has to select the picture that best illustrates

the meaning of the word or the appropriate morphological or

syntactic structure used by the examiner. Syntactic

comprehension was analyzed using the results obtained from

administration of the TACL-R. Specifically, the following

variables of auditory comprehension were determined: the

number and types of word classes and relations correctly

understood, the number of grammatical morphemes

comprehended, and the number of elaborated sentence

constructions understood. A raw score was converted into a

language age score.

Phonological development was determined through the use

of The Goldman-Fristoe Test of Articulation. This test

provides a systematic assessment of articulation of

consonant sounds. It provides percentile rank norms which

may be used with individuals who range in age from 2 years

to adulthood. The test utilizes large colorful pictures of

common everyday objects. The child is expected to label

each picture as it is presented by the examiner. The Kahn-

Lewis Phonological Analysis was selected to further describe

the phonological abilities of any child with a disorder of

speech sound production. It allows one to perform a

phonological process analysis on data from the Goldman-

Fristoe and yields a percentile rank score of overall

phonological process usage, a phonetic inventory, and an

age equivalent score.

The Goldman-Fristoe-Woodcock Auditory Skills Test Ba-

ttery consists of four subtests and is designed to measure

several aspects of auditory skills. The test was normed on

individuals aged three years through adulthood. The

Auditory Memory, Auditory Sequencing, and Auditory

Discrimination subtests were administered to the subjects in

this study.

The Token Test for Children was developed to assess an

individual's ability to process and recall verbal directions

of increasing length and complexity. DiSimoni (1978) and

Noll (1970) revised the original version so that it may now

be used with children as young as three years old in

addition to the adult population. The Token Test provides

limited standardization information. The individual being

tested is expected to respond to verbal commands by manipu-

lating tokens of various colors and shapes. The commands

range in length from four to ten words. Morley et al.

(1979) reported that scores on the Token Test correlate

highly with scores on tests of auditory comprehension.

Measures of aspects of expressive and receptive prag-

matics were obtained using results from administration of

the WISC-R. The WISC-R was normed on children ranging in

age from six years to 16 years old. The WISC-R is comprised

of two major scales, verbal and performance. One of the

verbal subtests were administered to the subjects in this

study. The WISC-R provides a point scale. Raw scores were

obtained for each of the subtests. These scores were then

converted into standard scores based on the subject's

chronological age.

The Test of Word Finding (TWF) is based on a diagnostic

model of word-finding assessment drawn from child and adult

literature on word-finding disorders. The TWF is normed on

children aged six and one-half to thirteen years of age. It

has five wordfinding sections and one comprehension section.

Section One is called Picture Naming: Nouns; Section Two is

titled Sentence Completion Naming; Section Three is titled

Description Naming; Section Four is titled Picture Naming:

Verbs; Section Five is titled Picture Naming: Categories;

and Section Six assesses the comprehension of those target

words which the child had difficulty naming on the five

word-finding sections. Raw scores are calculated for each

section. The raw score may be converted to standard scores

and percentile ranks.

Sampling Test Procedures

Rationale. There are many variables that may interact

to influence the nature of the conversational skills

exhibited by a child at any particular time. Therefore, it

is incumbent upon the diagnostician to assess the child's

language behavior in a variety of physical and social

contexts. Fey (1986) stated that in addition to formal

testing, the speech-language pathologist should assess the

child's language in conversational contexts. He further

suggested that a child's performance in all areas of

language content (semantics), form (syntax), and use

pragmaticss) be compared to the clinician's expectations for

a typical child of the same chronological or mental age.

This comparison of the child's total communicative

performance in a variety of informal and formal situations

helps the clinician arrive at a more appropriate decision

relative to the nature of the child's communication skills

in everyday situations.

The sampling procedures were designed to yield a data

base for the analysis of semantic diversity, syntactic

complexity and several aspects of pragmatics (Table 2-3).

Semantic diversity was determined using procedures outlined

by Bennett and Alter (1985). Syntactic complexity was

determined using the Language Sampling Analysis and Training

procedure (LSAT) (Tyack & Gottsleben, 1974). Pragmatic

competence was analyzed along several parameters including:

overall communicative functions, topic manipulation, verbal

turn-taking, presupposition, and conversational repairs.

Several communicative functions were examined including

behavior requests, information statements, information

requests, directives, interpretatives, projectives, and

relationals. The aspects of topic performance that were

examined include topic initiation, topic maintenance, the

proportion of topics maintained, and the number of on-topic

exchanges. Presuppositional skills that were examined

included cohesion, informativeness, and deictics. Several

aspects of communication related to conversational repairs

were examined: contingent queries, repetitions, cause of

breakdown, presence or absence of repair attempts, repair

initiator, repair strategy, and the outcome. The presence

or absence of verbal turn-taking cycles and the nature of

those cycles were also examined.

Procedures. A language sampling procedure was used

to examine several aspects of communicative performance.

The sample was collected during the fourth test session.

The fourth session lasted approximately 60 minutes.

Procedures conducted during this fourth session and

subsequent scoring and analysis of all results was performed

by the experimenter. The play room was equipped with a

common set of age-appropriate toys/props for each session.

The experimenter interacted with each child at least once

prior to the actual videotape session. Only the examiner

and one child were present in the playroom during the first

three test sessions. A third person was called into the

room for the Barrier Task activity during the fourth

session. The same individual, an ASHA-certified speech-

language pathologist, was the third person for the Barrier-

Task activity with each child in the study. Three distinct

phases of interaction occurred during the fourth test

session, "Elicited-Scheme Play" interactions, "Script-Eased

Play" interactions, and "Barrier-Task Play" interactions. A

video camera was positioned in the play room where it would

cause minimal distraction from the on-going activities of

the session. Taping began immediately with the onset of the

session and continued throughout all of the interactions. A

brief period of conversational interaction initiated each

session. This allowed for more natural conversational

output from each child as he became acclimated tc the test

room and to the video equipment.

During the "Elicited Schemes Play" activity, the

examiner sought to evaluate the child's ability to conduct a

conversation about an everyday activity. The activities

selected for this study were Going to School, Art Class

Activity, and Planting a Garden. These particular

activities were selected because all of the children in this

study were attending schools. It was assumed therefore,

that each child would have some basic level of familiarity

with the steps involved in these regular routines. The

examiner started the sessions by engaging each child in

conversation about school in general. She asked the child,

"Do you like going to your school?" The experimenter made

an appropriate comment depending on the child's response.

After the conversation had been initiated, the experimenter

said to the child, "Tell me how you get ready for school in


the morning." When necessary, the examiner provided verbal

prompts such as "and then what happens" in order to

encourage additional speech output from the child. This

activity was considered finished when the experimenter had

presented a list of fifteen standard verbal prompts and had

received some sort of response from the child to each of

them. A complete list of verbal prompts for this activity

can be found in Appendix A.

The "Script-Based Play" activity was designed to

promote conversation around an age-appropriate theme. The

themes Planting a Garden and Art Class Activity were used

with the children for this activity. The experimenter

placed an appropriate set of toys/props on the table by the

child and proceeded through the list of verbal prompts for

each activity. The prompts were presented to each child in

the same order. The activities were allowed to proceed in a

natural manner with the experimenter making comments as ap-

propriate and providing prompts to elicit maximal

conversational output from the child. Each activity lasted

a minimum of fifteen minutes. Standard prompts from the

examiner included "Here are the things we need," "What

should we do first?", "How long will it take for the plant

to grow?", and "Have you ever seen a real garden?" A

complete list of verbal prompts for this activity can be

found in Appendix A. A variety of toys were available

during the play sessions with each child. A list of the

toys can also be found in Appendix A.

The third activity during the fourth test session was

"Barrier Task Play." For this procedure a third person was

included in the interactions. During this procedure, the

examiner performed various manipulations of small objects

and figurines including toy cars and trucks. The child was

instructed to tell the other participant exactly what the

researcher had done so that the third participant could

duplicate the researcher's behavior. For example, if the

researcher put a small ball in the back of the car with a

man in it, the child had to tell the other participant to do

the same. If the other participant made a wrong move such

as putting the small ball in an empty car, the child had to

provide further instructions to the third participant until

the task was completed properly. The researcher used a

prepared script for the Barrier-Task activity. The third

participant also had a copy of the script. Her actions

however, were not based on the written script but were the

direct result of the child's instructions to her. This

barrier task provided an excellent opportunity to assess the

child's presuppositional skills.

Data Collection and Analysis

Videotaped language samples were comprehensively

transcribed in terms of verbal and nonverbal behaviors.

Traditional orthographic symbols were used throughout the

transcript. Symbols from the International Phonetic

Alphabet were utilized as necessary to code unintelligible

or questionable utterances. Contextual notations were

provided throughout the transcript. Semantic, syntactic,

and pragmatic parameters were analyzed. In addition to

type-token ratio (semantics) and structural analysis

(syntax), the following pragmatic parameters were analyzed:

overall communicative functions, topic manipulation, verbal

turn-taking, presupposition, and conversational repairs.

Tough's (1977) taxonomy of communicative functions of

utterances for children was used to analyze each child's

conversational performance (See Table 2-4). Her procedure

identifies four general categories of communicative

functions: directive, interpretative, projective, and

relational. The directive function includes communicative

acts used for the purpose of demonstration, instruction,

forward planning, self-directing, and collaboration. The

interpretative function involves two major areas: reporting

and reasoning. Reporting acts are used when the speaker

wants to show elaboration of detail, association with some

earlier experience or when making comparisons, the recogni-

tion of incongruity, the awareness of sequence, the

recognition of associated action or events, the absence of

conditions, the recognition of a central meaning, and/or

when reflecting on the meaning of experience. Reasoning

acts are used to show the recognition of causal


relationships and recognition of a principle or a condition.

The projective function includes acts used for predicting,

showing empathy, and imagining. Predicting acts are used

for forecasting, anticipating consequences, surveying

possible alternatives, predicting related possibilities, and

recognition of problems with prediction of solutions.

Empathetic acts are used when the speaker wishes to show

understanding of another person's state of being. Imagining

acts can be seen when speakers represent or rename objects

to be something other than what they are and when making

commentary on the imagined context of a situation. The

relational function involves self-maintaining and

interactional acts. Self-maintaining acts include

utterances used to express one's need, protect self-

interest, provide justifications, make criticisms, or issue

threats. Interactional acts promote various types of social

interplay between speaker and listener. Each of the child's

utterances were coded as fulfilling one of these utterance

acts. The percentage of communicative acts per category

were determined for each child.

Topic manipulation was analyzed based on a procedure

adapted from Keenan and Schieffelin (1976). Two major areas

of topic manipulation were examined: topic

introduction/topic change, topic maintenance (Table 2-5).

Topic introduction/topic changes were identified as those

utterances which were initiated by the speaker which do not

relate to the previous on-going discussion. Topic

introduction/topic change were analyzed in terms of the

number of topics introduced or changed. Topic maintenance

were defined as those utterances that pertain to a

previously introduced topic and that serve to facilitate on-

going discourse by either adding or requesting additional

information. Topic maintenance was analyzed in terms of the

number of contingent queries produced by the speaker, the

number of on-topic exchanges produced by the speaker, and

the average number of utterances per topic. Additionally,

conversational replies were analyzed according to the

criteria specified by Leonard (1986). The presence or

absence of seven types of responses to questions and eight

types of responses to statements and directives were


Turn-taking behaviors were analyzed utilizing

procedures adapted from Prutting and Kirchner (1987), Fey

and Leonard (1984), and Roth and Spekman (1984). Two

aspects of turn-taking were analyzed: proportion of

initiations by the child and proportion of responses by the


Analysis of presuppositional skills involved making a

determination of the presence or absence of deictics,

indirect and direct reference forms, and other forms of

cohesion (substitution, ellipsis, conjunction).

An analysis of the child's responses to communicative

breakdown was based on a procedure suggested by Roth and

Spekman (1984). Five major areas of concern were examined:

cause of the breakdown, the presence or absence of a repair

attempt, repair initiator (listener or speaker), repair

strategy, and outcome.

Semantic diversity was determined using procedures

developed by Bennett and Alter (1985). Specifically the

analysis yielded data that reflected a type-token ratio

(TTR), the number of total words, and the number of

different words in the sample.

Syntactic complexity was determined using the Language

Sampling: Analysis and Training procedure (LSAT) (Tyack &

Gottsleben, 1974). The LSAT separates utterances and their

constituents into forms, constructions, and complex

sentences. Each subject was assigned to a linguistic level

of based on his Word-Morpheme Index. With each child's

assigned level as a reference point, his mastery of forms

and constructions was determined. Further syntactic

analysis was accomplished by using the LSAT to

determine the number/percentage of complex sentences in each

sample. The specific measure that was examined using this

procedure was the proportion of complex sentences within the

child's language sample.


To establish intrajudge reliability the examiner

recorded 25% of each child's data from each of the above

mentioned analyses. Interjudge reliability for the data

from the language sampling procedure was established by

having an ASHA certified speech-language pathologist code

25% of all syntactic, semantic, and pragmatic parameters

from each language-impaired child's language sample.


Language to thought to language processing

Semantics--->---Comprehension Formulation--->Semantics

Speech perception Speech programming
Speech input-->Sensation Speech production
Speech output

Speech to language processing Language to speech
(prelinguistic processing) (postlinguistic

Figure 2-1. Child Language Processing Model (Aram & Nation,

Table 2-1. Assessment Measures Selected to Correspond to Aram and
Nation's Model

Sensation: hearing acuity
Perception: auditory discrim.
auditory memory
auditory sequencing

Comprehension: Semantics

Integration: pragmatics

Formulation: word retrieval

Speech programming: motor speech
Speech production: articulation/

11 Pure-tone test
11 Discrimination Subtest
11 GFW Auditory Memory Subtest
f1 GFW Auditory Memory Subtest
1 TOLD Repetition Subtest

1i Token Test
11 Similarities
11 Test of Word Finding

I1 Goldman-Fristoe/
11 Kahn-Lewis Analysis

Table 2-2. Non-Norm Referenced Procedures

Overall Function

Topic Performancell
Presupposition 11



Conversational 11
Repairs II
Verbal Turntaking 11



1 Speaker Variables 11 Protocols Used
11 11

11 behavior requests
11 information statements
11 information requests
11 directive
11 interpretative
11 projective
11 relational

topic maintenance
topic initiation
number utterances/top


contingent queries
cause of breakdown
presence/absence of
repair attempt
repair initiator
repair strategy
nature of

type-token ratio

structural analysis
complex sentences

lIGuralnick &
11 Paul-Brown (1986)
IlTough (1977)

IlKeenan &
11 Schieffelin (1976)
icil Fey & Leonard
11 (1984)
IIHalliday & Hasan
11 (1976)
1lRoth & Spekman
11 (1984)

1IRoth & Spekman
11 (1984)
IlLeonard (1986)

JlPrutting & Kirchner
11 (1987)
IlFey & Leonard
11 (1984)
J[Roth & Spekman
11 (1984)
IIBennett & Alter
11 (1985)
11 Paul (1981)

Table 2-3. Tough's Communicative Functions (Tough, 1977)

General Categories




Inclusive Acts

forward planning

elaboration of detail
making comparisons
recognition of incongruity
awareness of sequence
recognition of associated
actions or events
absence of conditions
recognition of central
reflecting meaning of
recognition of causal
recognition of principle
or condition

showing empathy


express one's needs
protect self interest
provide justifications
make criticisms
issue threats

promote various types of
social interplay between
speaker and listener


Table 2-4. Types of Conversational Replies

Responses to Questions
1. Affirmative answers to yes-no questions
2. Negative answers to yes-no questions
3. Wh-replacement
4. Alternatives
5. Choice responses
6. Requests for repetition/clarification
7. Imitation

To Statements and Directives

8. Affirmation
9. Compliance
10. Refusal/denial
11. Noncompliance
12. Alternative
13. Expansion
14. Requests for repetition/clarification
15. Imitation


The purpose of this study was to conduct detailed

analyses of the conversational performance of school-aged

children who had been identified as poor conversationalists,

in an attempt to describe and isolate the specific pragmatic

variables that are impaired. The researcher developed

indepth linguistic profiles of children who had pragmatic

disorders, in order to quantify and qualify the nature of

pragmatic disorders and to examine pragmatics in relation to

the other dimensions of language including phonology,

semantics, and syntax. The present investigation was

comprised of three individual case studies. Each case study

involved a language-impaired child and a language-normal

child matched on the basis of nonverbal intelligence,

chronological age, socioeconomic status, and sex. The

children ranged in age from 7 to 11 years.

There were two basic data collection procedures, norm-

referenced testing and language sampling. Formal assessment

procedures were directed by the Child Language Processing

Model (Aram & Nation, 1977). The model is based on the

assumption that disordered language behavior is a reflection

of disrupted language processes within three major

processing segments. Thirteen norm-referenced tests and

subtests were administered to each child.

The sampling procedures were used to yield a data base

for the analysis of semantic diversity, syntactic

complexity, and several aspects of pragmatics. All children

were engaged in four conversational play activities, an

"Elicited-Language" segment called School, two "Script-Based

Play" activities called Art Class and Planting a Garden, and

a "Barrier-Task Play" activity. Videotaped language samples

of the school, art, and garden segments were analyzed to

determine the communicative functions of each of the

children's utterances. In addition, the children's

performance in verbal turn-taking, topic manipulation,

conversational replies, and communicative repairs was

evaluated. The samples were further analyzed to determine

each child's semantic diversity, syntactic complexity, and

phonological performance. The "Barrier-Task Play" segment

was analyzed to evaluate each child's presuppositional

skills. In addition to the pragmatic analysis, the

individual utterances were evaluated to determine the

subject's semantic diversity (type-token ratios) and

syntactic development.

Videotaped language samples were comprehensively

transcribed for verbal and nonverbal behaviors. Traditional

orthographic symbols were used throughout the transcript.

Contextual notations were provided as appropriate.

To establish intrajudge reliability the examiner

recorded 25% of each child's data from each of the above-

mentioned analyses. A coefficient of .97 was achieved for

the intrajudge reliability. Interjudge reliability for the

language sampling data was established by having an ASHA

certified speech-language pathologist recode twenty-five

percent of all syntactic, semantic, and pragmatic parameters

from each language-impaired child's communication sample. A

coefficient of .89 was obtained for the interjudge


Case Study One


The first case study compared the linguistic

performance of two eleven-year-old Black males. A.D. was

the language-impaired subject and R.A. was his language-

normal control. Both of the children were in fifth grade

and attended public schools in Gainesville, Florida.

Norm-Referenced Test Results

Six norm-referenced tests were administered to each

child. These tests measured auditory perceptual skills,

semantic and syntactic comprehension, syntactic repetition,

metalinguistic skills, word-finding skills, and oral motor


skills. The results of the auditory perceptual skills tests

are presented in Table 3-1. All other norm-referenced test

results are presented in Table 3-2.

Auditory Perceptual Skills Battery. The Auditory

Discrimination and Auditory Memory subtests from the

Goldman-Fristoe-Woodcock Auditory Skills Battery were

administered to assess each child's auditory perceptual

abilities: Auditory Discrimination and Auditory Memory.

There was some variability between the performance of

the two children on the auditory perceptual skills subtests.

A.D. obtained a combined raw score of 289 on parts I, II,

and III of the Auditory Discrimination subtest. This score

corresponds to an age equivalency of 7.3 years and placed

him at the fifth percentile. R.A. obtained a combined raw

score of 297 on the Auditory Discrimination subtest. This

score corresponds to an age equivalency of 14.2 and places

him at the sixty-eighth percentile.

The Auditory Memory subtest is comprised of three

segments: recognition, content, and sequencing. Both A.D.

and R.A's performance on the "recognition" section of the

Auditory Memory subtest yielded a raw score of 99 which

corresponds to an age equivalency of 8.2 years and places

them at the fourteenth percentile. Each of the children

obtained lower than expected scores for their ages on the

"content" section of the Auditory Memory subtest. Both

A.D.'s raw score of 6 with its corresponding age equivalency


of 4.10 and R.A.'s raw score of 13 with its age equivalency

of 5.11 places them below the first percentile. The

subjects showed considerable variability in their

performances on the "sequencing" segment. A.D. obtained a

raw score of 47 and an age equivalency of 9.6 which places

him at the thirtieth percentile. R.A. obtained a raw score

of 84 and an age equivalency of 18.11 which placed him above

the ninety-ninth percentile.

In summary, A.D's auditory discrimination scores place

him well below the fiftieth percentile. R.A.'s scores were

considerably above the fiftieth percentile. The subjects

obtained identical scores on the recognition section of the

auditory memory subtest. Both subjects obtained scores on

the content section of the auditory memory subtest that were

well below what is expected for their age. Their scores

placed them below the first percentile. A.D.'s scores on

the auditory sequencing section fell well below the fiftieth

percentile while R.A.'s score was above the ninety-ninth


Semantic and Syntactic Comprehension. Four measures of

semantic and syntactic comprehension and were given to each

child. The subjects demonstrated noteworthy differences in

their lexical comprehension abilities as measured by the

Peabody Picture Vocabulary Test. A.D. obtained a raw score

of 64 which corresponds to a standard score of 43 and a

semantic comprehension age equivalent of 5.6. These results

suggest that A.D. is functioning below two standard

deviations from the mean for his age. R.A. obtained a raw

score of 113. His score corresponds to a standard score of

96 and a semantic comprehension age equivalent of 10.10.

R.A.'s scores suggest that he is functioning within the

normal range for his age.

A.D. and R.A. exhibited significant differences in

their comprehension of syntactic structure as measured by

the Token Test for Children. A.D. received a raw score of

39 which corresponds to a standard age score of 478 and a

grade score of 480. Standard scores between 495 and 505 are

considered to be within the average range. R.A. received a

raw score of 60 on the Token Test. His score corresponds

with a standard age score and grade score of 505. Standard

scores below 494 are considered to be inferior and indicate

a need for additional testing. Scores below 494 are

considered to be inferior and indicates a need for

additional testing in the area of semantics and syntax.

Those children scoring in the range of minus three standard

deviations will have scored more poorly than approximately

97% of the standardizing group of children. A.D.'s scores

place him well below the third standard deviation indicating

a significant deficit in the area of syntactic

comprehension. R.A.'s scores fall within the high range of

normalcy for his age group.