relations among verbal and nonverbal cognitive skills in normal language and specifically...

J. COMMUN. DISORD. 25 (1992). 205-219

RELATIONS AMONG VERBAL AND NONVERBAL COGNITIVE SKILLS IN NORMAL LANGUAGE AND SPECIFICALLY LANGUAGE- IMPAIRED CHILDREN

MARIA ADELAIDA RESTREPO, LINDA SWISHER, ELENA PLANTE, and REBECCA VANCE i%e University of Arizanu. Tucson. Ari;oncr

This study tested the hypothesis that specifically language-impaired (SLI) children have a qualitatively different cognitive system from that of normal language (NL) children. Twenty NL and 20 SLI children between the ages of 4: 2 (years: months) and 5 : 11 were presented with experimental language-learning measures, experimental nonverbal measures, and verbal and nonverbal norm- referenced tests. A confirmatory analysis of the covariance matrix structures of the two subject groups indicated that relations among cognitive skills differed between NL and SLI children. In addition, a planned comparison indicated that the relation between nonverbal rule-induction and novel bound-morpheme learning differed significantly between groups. The findings indicate that a “qualitative-differences” model of specific language impairment better accounts for the co-occurrence of poor verbal and poor nonverbal cognitive skills in SLI children than a “low-normal” model.

INTRODUCTION

Leonard (1987, 1991) has proposed that the language skills of specifically language-impaired (SLI) children fall at the low end of the normal curve and may not differ from those of normal language (NL) children

in any other way. He utilized this “low-normal” model of specific language impairment to explain both the poor verbal skills (e.g., poor use of bound morphemes and components of the verb system) and the poor

Address correspondence to M. A. Restrepo, The Scottish Rite/University of Arizona, Center for Childhood Language Disorders, The University of Arizona, 33 E. Ochoa St., Tucson, AZ 85701.

01993 by Elsevier Science Publishing Co., Inc. 205 655 Avenue of the Americas, New York, NY 10010 002 l-9924/93/$6.00

206 M. A. RESTREPO et al.

nonverbal cognitive skills (e.g., mental imagery and rule induction) identified in this population.

The low-normal model of specific language impairment specifies no biological cause for the behavioral constellation of poor skills other than normal genetic variation (Leonard, 1987, 1991). Contrary to this assumption, Plante, Swisher, Vance, and Rapcsak (1991) demonstrated that SLI children have atypical perisylvian asymmetries that presum- ably develop in utero. In view of these findings, Plante (1990) proposed that poor verbal and nonverbal cognitive skills co-occur in SLI children as a result of alterations in neuroanatomical development that affect both their left and right hemispheres. We further propose that a qualitative-differences model of specific language impairment better accounts for the co-occurrence of atypical biological signs, poor verbal skills, and poor nonverbal cognitive skills in SLI children than does a low- normal model. The qualitative-differences model characterizes these children as a clinical population; that is, one with unique biological and behavioral characteristics.

Most investigators of SLI children have analyzed variables for quan- titative differences between SLI and NL children. This approach to data analysis has led to findings that can be described adequately by a low-normal model. Another type of analysis, one that examines the relations among skills, would open up the possibility of describing the results from the perspective of a qualitative-differences model. Finding that the relations among skills differ between NL and SLI children, independent of individual skill levels, would indicate that the behavioral differences are qualitative in nature.

Although statistically significant correlations between verbal and nonverbal skills have been described in both SLI and NL children, these correlations have not been tested for significant differences between the two groups. For instance, Kamhi, Catts, Koenig. and Lewis (1984) found that performance on a nonverbal rule-induction task correlated significantly with performance on a receptive vocabulary test in a group of NL and in a group of SLI children. In addition, several studies have established a relationship between language and mental imagery skills in NL and SLI children. For instance, Whitmire and Stone ( 1991) found that both NL and SLI children’s performance on each of three visual-imagery tasks correlated significantly with performance on one or more language measures. Similarly, Kamhi et al. (1984) found that both NL and SLI children’s haptic-recognition scores (also a mental imagery skill) correlated significantly with receptive- language scores.

General nonverbal intelligence scores also have been found to corre- late with language scores in SLI children. For instance, Aram, Ekelman

VERBAL AND NONVERBAL RELATIONS 207

and Nation (1984) found that, out of several language, intelligence, and demographic measures, the Leiter Internutional Performance Scale (LIPS) (Leiter, 1969) was the best predictor of language and academic performance in language-impaired children (most of whom were SLI) 10 years later. Similarly, Schery (1985) found that performance 1Q was a significant predictor of language performance and language gain in SLI children.

The present study tests the theory that SLI children have a qualitatively different cognitive system from that of NL children. To test this theory, this study first will determine whether the relations among selected verbal and nonverbal cognitive skills are significantly different between NL and SLI children. In addition, in the presence of significant findings, this study will investigate whether qualitative differences between the two groups extend beyond the language system by testing three verbal-nonverbal relations between subject groups.

METHOD

Subjects

Subjects were 40 English-speaking children between the ages of 4:0 and 5 : 11 (years: months). To obtain representative samples, consecutive referrals who met the selection criteria for classification as SLI or NL were selected for study. Twenty of the subjects were SLI (4 girls, 16 boys) with a mean age of 4: 10 and an age range of 4: 2 to 5 : I 1; 20 were NL (4 girls, 16 boys) with a mean age of 5: I, and an age range of 4:4 to 5 : 11. A t-test for age differences was not significant. Children were not matched for language age or mental age’. All subjects met the following criteria: (a) passed a hearing screening at 25dB HL for 500 Hz, 1000 Hz, 2000 Hz, and 4000 Hz (ANSI, 1969), (b) scored above - I .67 SD (IQ > 75) on the Leiter International Performance Scale (LIPS) (Leiter, 1969), and (c) were between the ages of 4:0 and 5: 1 I.

SLI subjects met the following additional criteria: (a) they were enrolled in a. special needs program in a local school system that provided them with speech and language services, (b) they had been identified by a certified speech-language pathologist as SLI, (c) they were judged by a certified speech-language pathologist to be able to produce the sounds in the novel vocabulary items that would later be presented to

’ Matching would not have resulted in a representative sample of both groups of children. In addition, matching according to one composite score has limited utility with multidimensional constructs such as “language” and “intelligence” (cf. Johnston and Kamhi, 1984).


them. No referrals were excluded with these criteria. NL subjects met the following additional criteria: (a) they were enrolled in a local preschool or daycare program, and (b) they had obtained a passing score for their age level on the Fluharty Preschool Speech and Language Screening Test (Fluharty, 1978). One child referred as NL was excluded from study with this measure.

The school system’s criteria to qualify children as SLI were as fol- lows: (a) failed only the language portion, but not the cognitive portion, of the Batelle Screening (Newborg, Stock, Wneck, Guidubaldi, and Svinicki, 1984), (b) failed the Fluharty Preschool Speech and Language Screening Test (Fluharty, 1978), (c) scored - 1.5 SD on at least one comprehensive language test, most commonly the Sequence Inventory of Communicative Development-Revised (Hedrik, Prather, and Tobin, 1984), and (d) teachers reported no concerns regarding intelligence.

Measures and Procedures

The measures and procedures utilized in this study were adapted from a study by Nakamura, Plante, and Swisher (1990). Table 1 lists the order in which each measure was presented to each subject.

Experimental Language-Learning Measures. TRAINING STORY. The purpose of the training story was to familiarize

subjects with the experimental procedures. The presenter read the training story while he or she manipulated two animal figures-a pig and a bear. At preset intervals throughout the story, subjects were

Table 1. Order in Which Each Measure Was Presented to the Subjects

Day I Day 2

Hearing

Fluharty”

Articulation

Screeningh

Training

Story

Vocabulary

story

Vocabulary

game

Day 3

Vocabulary game

Vocabulary

posttest

Day 4

Bound-morpheme

game

Bound-morpheme

story 2

Bound-morpheme

story I Bound-morpheme

Posttest

Generalization

Day S

LIPS

Snail

Trail

Rule

Induction

ITPA-GC

PPVT-R

” Task was presented to normal-language children only.

h Task was presented to specifically language-impaired children only.

LIPS = Leiter International Performance Scale, ITPA-GC = Illinois Test of PsycholinguistIc

Ability-Grammatic Closure. PPVT-R = Peabody Picture Vocabulary Test-Revised.


asked to name each animal and each action that the animal performed (i.e., running, rolling).

VOCABULARY LEARNING MEASURES AND POSTTEST. Clay figures, a story, and a board game were used to teach four novel (consonant- vowel-consonant) vocabulary items (e.g., “ped,” “gack”) to which novel bound morphemes would later be attached. The vocabulary items included two nouns, one animate and one inanimate, and two verbs representing actions performed by the animate noun on the inanimate noun. The story contained 10 presentations of each of the four vocabulary items, and 20 questions to elicit each vocabulary item five times. The questions (e.g., “What is he?” “What does he like to do?“) were asked at increasing intervals following the presentation of each vocabulary item. The presenter provided feedback concerning the accuracy of the subjects’ responses. Positive feedback included a “yes” response and the vocabulary item (e.g., “Yes, he’s a dap.“); negative feedback included the correct vocabulary item (e.g., “He likes to gack.“). The “no” utilized by Nakamura et al. (1990) was omitted because pilot SLI subjects were reluctant to respond following this type of negative feedback (see Appendix A for a sample of a vocabulary story).

The vocabulary board game also contained 40 presentations of vocabulary items (10 of each) and 20 questions to elicit each item five times. Feedback was provided in the same manner as in the vocabulary story.

A vocabulary posttest was presented immediately after the vocabulary learning measures to reflect vocabulary learning. During the posttest, each subject had the opportunity to respond to eight questions designed to elicit each of the four vocabulary items twice (e.g., “What is this?” (ped)). In order to proceed to the bound-morpheme learning measures, each subject had to reach a criterion of six correct responses out of eight. If the subject did not reach this criterion, a short board game was played. During this game, each vocabulary item was presented and elicited two times. Following this game, the posttest was readministered, and the presenter proceeded to the bound-morpheme learning measure.

BOUND-MORPHEME LEARNING MEASURES AND POSTTEST. Two stories and a board game were used to teach two novel morphemes in meaningful contexts. The bound morphemes were the suffixes /A/ and iul attached to the novel noun and verb vocabulary items. The noun- bound morpheme represented the larger version of each object, and the verb bound morpheme represented increased magnitude of each action. These concepts were selected because neither is coded with bound morphemes attached to nouns or verbs in English. Because vowels are

M. A. RESTREPO et al.

longer in duration than consonants. IA/ and iul were selected as the novel bound morphemes to control for a saliency effect that may con- tribute to bound-morpheme difficulties in SLI children (Rom and Leon- ard. 1990).

The bound-morpheme stories contained 40 presentations, ten of each vocabulary item, six in the affixed form (e.g.. “pedu.” “gackA”). and four in the unaffixed form (e.g.. “ped, ” “gack”). Twenty responses were elicited from each subject, three in the affixed form and two in the unaffixed form of each vocabulary item. No feedback concerning accuracy of response was provided during these stories (see Appendix A for a sample of a bound-morpheme story).

The board game contained 20 sentences that presented each vocabulary item ten times, six in the affixed and four in the unaffixed form. Each subject had the opportunity to answer a total of 20 questions, five times for each vocabulary item; three in the affixed and two in the unaffixed form. No feedback was provided during the game.

A bound-morpheme posttest followed the presentation of the bound- morpheme stories and game to measure learning of the affixed forms and generalization of the morphemes to untrained stems. The posttest involved four trained and four untrained vocabulary items to which the subject was to attach the bound morphemes (e.g., “He is a gack. What is he‘?” (gacku); “He likes to ped. What else does he like to do’?” (pedA)). Sixteen responses were obtained, two for each of the four trained vocabulary items and two for each of the four untrained vocabulary items (see Appendix B for the list of vocabulary items and morphemes presented).

E.~-pc~rir?lt~nt~~I-Nonr~P~hrrl MPLI.FUI~P.S. The Snail-Trail measure (adapted from Savich. 1984) was selected to assess each subject’s spa- tial rotation skills, which appear to be depressed in SLI children (Sav- ich, 1984; Kamhi, 1981). The Snail-Trail measure required subjects to look at three-dimensional figures (a circle. an oval, an hourglass, a figure-eight “trail”), and to predict the future position and orientation of a snail as it moved along the figures or trails. Each subject demonstrated the appropriate position and orientation of the snail by placing a second snail on the trail, rather than by drawing the position on paper as in Savich’s version. We adapted the task because pilot work indicated that NL preschool children could not perform Savich’s procedure without extensive training. The Rule-Induction measure (adapted from Kamhi et al.. 1984) was selected to assess subjects’ ability to induce rules nonverbally. This task required each subject to determine which of two containers held a ball. The container lids had a figure or color designed to serve as a clue to the ball’s location (e.g.. the ball was always in the container with the butterfly on the lid). The presenter


said: “There is a ball in one of these containers. I want you to find it every time. There is a clue on the lid to help you. Are you ready?” If a subject found the ball six consecutive times, we concluded that he or she had induced a rule regarding the location of the ball. The task was discontinued after 40 trials if a subject had not succeeded. Fewer trials to criterion reflected better performance; that is, fewer trials to induce the rule regarding the location of the ball. We adapted the task from its published form (Kamhi et al., 1984) because pilot work indicated that children within the age range to be studied could not perform the origi- nal task in 40 trials.

Verhul und Nonverbal Norm-Rejkrenced Tests. Three norm-referenced tests of verbal and nonverbal skills were administered to each subject. The verbal measures included the Grammatic Closure subtest of the Illinois Test of Psycholinguistic Abilities (ITPA-GC) (Kirk and Kirk, 1968), selected to reflect each subject’s current expressive bound- morpheme level; and the Peabody Picture Vocabulary Test-Revised (PPVT-R) (Dunn and Dunn, 1981), selected to reflect each subject’s current receptive vocabulary level. The LIPS was selected to reflect each subject’s general nonverbal intelligence level.

PresentersiTest Administrcztors. Five students unaware of the purpose of the experiment were trained to administer the language-learning task. Six other students unaware of the results of the language-learning task administered the test battery. A certified speech-language pathologist supervised students at all times.

Task Presentation. The hearing screening, the speech and language screenings, and the training story were administered on Day 1. On Day 2, the vocabulary story and vocabulary game were presented. On Day 3, the vocabulary game was played again, the vocabulary posttest was administered, and the first bound-morpheme story was presented. On Day 4, the bound-morpheme game was played, the second bound-morpheme story was presented, and the posttest/generalization probe was administered. The test battery was administered on the fifth and final day (see Table 1 for a summary and order in which measures were presented to the children).

The supervisor present during all sessions scored subjects’ responses live. In addition, all sessions were audiotaped for double scoring.

CountPrhulancing. The novel-bound morphemes were counterbalanced across novel nouns and verbs to decrease phonological effects; 20 subjects (10 NL and 10 SLI) received /u/ as the noun suffix and /A/ as the verb suffix: the other 20 subjects (10 NL and 10 SLI) received


the reverse. To decrease word-saliency effects, vocabulary items were counterbalanced across story presentations and generalization probes. The animate figures were counterbalanced across story presentations and posttests to decrease figure-saliency effects. To minimize order effects, the order of presentation of the nonverbal experimental measures and the norm-referenced tests was randomized across

subjects.

Reliability. Procedural reliability was evaluated independently by four observers for two of five randomly selected presenters. The story and game presentations for four subjects (20%) in each group were selected at random, videotaped, and checked by a trained observer for accuracy of item presentation on a sentence-by-sentence basis. The results indicated that items were presented with an average of 98.1% accuracy, with a range of 95.00/o-100%. In addition, trained observers checked all scored responses to experimental language-learning measures, experimental nonverbal measures, and verbal and nonverbal norm-referenced tests to establish point-to-point reliability. Point-to- point reliability for the scoring of the vocabulary measures and posttest ranged from 85.6%-lOO%, with an average of 97.0%. Reliability for scoring the bound-morpheme tasks and posttest/generalization probe ranged from 91. I%-100% with an average of 97.8%. The experimental nonverbal measures and the norm-referenced tests were scored with an average of 95.2% accuracy, with a range of 87.9%-98.5%.

RESULTS

A confirmatory analysis of the covariance matrix structures of the two subject groups was utilized to test the hypothesis that SLI children have a qualitatively different cognitive system from that of NL children. This analysis tests a mode1 that accounts for the relations among variables. Data were analyzed using LISREL7 (Joreskog and S&born. 1989).

First, we obtained a correlation matrix for each group for the following seven variables: performance on the bound-morpheme learning task, vocabulary posttest, ITPA-GC, PPVT-R, Rule Induction, Snail Trail, and LIPS. Correlations, means, and SDS for all the variables for NL children are presented in Table 2. Significant correlations were present (p < 0.05) between performance on the novel bound-morpheme variable and each of the following variables: vocabulary posttest, ITPA- CC, and Rule Induction. A significant correlation was also obtained between performance on the ITPA-GC and PPVT-R.

Correlations, means, and SDS for all the variables for SLI children are presented in Table 3. A significant correlation (p < 0.05) was ob-


Table 2. Means, SDS, and Pearson Correlation Coefficients between All Pairs of Variables for Normal-Language Children

Variable

1. BMT

2. ITPA-GC

3. LIPS

4. PPVT-R

5. RI

6. ST

7. VP

x

SD

Variable

1. BMT 2. ITPA 3. LIPS 4. PPVT

1.00 0.48” 0.35 0.27

1.00 -0.03 0.76”

1 .oo - 0.05

I .oo

48.45 17.15 119.95 67.45 19.10 25.95 6.50

5.46 5.03 17.54 15.53 5.53 4.61 1.62

5. RI 6. ST 7. VP

~ 0.58“

-0.36

-0.16

~ 0.28

I .oo

0.24

-0.01

0.17

~ 0.05

-0.21

1.00

0.59”

0.27

0.03

~ 0.09

-0.10

0.39

I .OO

” p < 0.05. BMT = Bound-Morpheme Task. ITPA-GC = Illinois Test of Psycholinguistic Ability-Grammatic Clo-

sure. LIPS = Leiter International Performance Scale, PPVT-R = Peabody Picture Vocabulary Test-Rev- ised. RI = Rule-Induction Task. ST = Snail Trail Task. VP = Vocabulary Posttest.

tained between performance on the novel bound-morpheme task and vocabulary posttest. Other significant correlations among variables were obtained between performance on the ITPA-GC and PPVT-R, and performance on Rule Induction and PPVT-R.

We then converted the correlation matrices into covariance matrices

Table 3. Means, SDS, and Pearson Correlation Coefficients between All Pairs of Variables for Specifically Language-Impaired Children

Variable

Variable 1. BMT 2. ITPA 3. LIPS 4. PPVT 5. RI 6. ST 7. VP

I. BMT

2. ITPA-CC

3. LIPS

4. PPVT-R

5. RI

6. ST

7. VP

x

SD

1 .oo 0.38 -0.19 0.35 -0.12 -- 0.06

1.00 0.03 0.60” -0.22 -0.10 -

1.00 0.44 - 0.43 ~0.24 ~

I .oo -0.51” PO.15

I .oo 0.01

1.00 -

26.75 6.90 99.50 43.80 26.80 22.45

13.09 5.07 13.80 10.09 15.25 5.77

0.75”

0.09

0.25

0.13

0.09

0.08

1.00

4.40

2.50

” p i 0.05.

BMT = Bound-Morpheme Task. ITPA-GC = Illinois Test of Psycholinguistic Ability-Grammatic Clo- sure. LIPS = Leirer International Performance Scale, PPVT-R = Peabody Picture Vocabulary Test-Rev- ised. RI = Rule Induction Task, ST = Snail-Trail Task, VP = Vocabulary Posttest.

M. A. RESTREPO et al.

to account for the underlying SD of each variable in the relations. Next, we specified a seven-factor model to account for the relations among the seven variables with no error variance for the NL children. (This procedure produced an error-free model of the relations among variables in NL children.) Finally, we attempted to fit the SLI data to the NL model using LISREL7 (Jiireskog and Siirbom, 19X9) in a confirmatory analysis of the covariance matrix.

The covariance matrix for the SLI children did not fit the model of relations derived from NL children. The x2 was 73.55 (28. N = 40. 1, < 0.001). This finding indicates that the two covariance matrices are significantly different from each other. and thus the relations among

skills between the two groups are significantly different. After the confirmatory analysis was run. three planned comparisons

of correlations were tested for a significant difference between the two groups. (These analyses determine whether specific correlations con- tributed to the statistically significant difference between covariance matrices. In addition. they indicate whether qualitative differences in relations among skills extend beyond the language system because each of the three correlations involves a nonverbal cognitive skill.) We utilized Fisher z,’ transformations to test three correlations for group differences: performance on the bound-morpheme learning measure with performance on Rule Induction, Snail Trail, and LIPS. A summary of the Fisher z,‘ transformations and the z tests is presented in Table 4.

Only the correlation between performance on the bound-morpheme learning measure and Rule Induction was significantly different between the NL and SLI groups (p > 0.01; :’ = 3.48). This correlation difference between the two groups identifies a specific relation between

two skills that is different for the SLI and NL children.

Table 4. Correlations. Fisher :’ ‘fran~fc~rmation~. and ; Value4 in Norma- Language and Specifically L~anguage-Impait-ed Children

Variables

Correlations ;’ transfwmation

NL SLI N L SLI ; value

BMTiLlPS 0.35 -0. I9 0.365 0. 19’ I .h3

BMTIRI - 0.58 -0.12 0.678 0. 121 2.48”

BMTISI 0.24 0.06 0.245 0.060 0.83


DISCUSSION

The results of the present study support the hypothesis that the cognitive systems of NL and SLI children are qualitatively different. As predicted by the qualitative-differences model of specific language impairment, a confirmatory analysis of the covariance matrix structures indicated that NL and SLI children are characterized by different relations among verbal and nonverbal cognitive skills.

Further analysis indicated that the correlation between scores on a nonverbal Rule-Induction task and a bound-morpheme learning task significantly differed between the two groups. This result indicates that qualitative differences between NL and SLI children extend beyond the language system and include at least one component of the nonverbal cognitive system. This finding adds to the large body of literature that indicates NL and SLI children differ quantitatively on a wide variety of nonverbal cognitive tasks (e.g., Johnston and Ramstad, 1983; Kamhi et al., 1984; Nelson, Kamhi, and Ape], 1987; Nippold, Erskine, and Freed, 1988; Savich, 1984; Weismer, 1991).

That qualitative differences between the cognitive systems of NL and SLI children involve language skills is also in line with the large body of evidence identifying qualitative differences between the language skills of NL and SLI children. For example, Johnston and Kamhi (1984) found that the types of semantic and syntactic forms in SLI children’s sentence production significantly differed from those of NL children matched by mean length of utterance. In addition, van der Lely and Harris (1990) found that the type of response errors to a comprehension task involving reversible sentences differed between NL and SLI children. Thus, SLI children’s language skills may appear low normal (e.g., Curtiss, Katz and Tallal, 1992) because they have been analyzed quantitatively, other types of analyses reveal qualitative differences.

The findings support the view that SLI children form a unique population (Aram, 1991; Tomblin, 1991). This implies that theories and inter- vention procedures developed for NL children cannot be assumed to be maximally beneficial to SLI children. Instead, assessment and inter- vention programs should be derived from the SLI population itself and tested for effectiveness with this population. Toward a similar purpose, Connell (1987) and Connell and Stone (1992) found that SLI children responded differently from NL children to two different language teaching programs. In 1987 this finding led Connell to conclude that treatment should be adapted to SLI children’s “unique learning styles” (p. 105).

In conclusion, SLI children appear to be a clinical population characterized by a constellation of atypical neuroanatomical signs, verbal and nonverbal cognitive strengths and weaknesses, and unique relations

216 M. A. KESTKEPO et al.

among verbal and nonverbal skills. The low-normal model of specific language impairment proposed by Leonard (1987, 1991) does not account for the present findings that indicate that the cognitive systems of SLI and NL children are qualitatively different. In addition, it does not account for SLI children’s atypical biological configurations (Plante et al., 1991). Instead, a qualitative-differences model of specific language impairment better accounts for the co-occurrence of atypical neuroanatomical asymmetries, and a cognitive system that is qualitatively different from normal.

This research was funded partially by the Department of Education Grant #H029D90108. We thank L. Coolidge and the Tucson Scottish Rite Charitable Foundation for this unconditional support of research with SLI children.

APPENDIX A

Excerpts from Training Stories

Today, I’m going to tell you a story about a far-away planet. I’ll tell you about the strange creatures that live there and the things they like to do. Listen carefully.

Vocahdury star?

Here is a PED.

It is a toy.

What is this? (ped)

Here is a DAP.

It is his PED.

What is he? tdap) The DAP likes to play on his ped

The DAP likes his toy.

What is this? (ped)

What is he? (dap)

The DAP is having fun.

He can GACK.

Watch him.

What can he do? (gack)

Ir$k~tion stoq

Here is a PED.

It’s a toy.

What is this? (ped)

Here is a PEDU.

It’s a toy too.

What is this? (pedu)

Whose toys are these‘?

He is a DAP.

They are his toys.

What is he? (dap)

He is a DAPU.

Look at him.

What is he? (dapu)

The dapu says “can I play?”


APPENDIX B

List of Vocabulary Items

Vocabulary Gack-lgaekl Nem-/n&m/ Dap-/drep/ Ped-hdi

Sample set Training vocabulary Gack/Gacku-Nouns Nem/Nemu-Nouns Dap/Dapa-Verbs PediPeda-Verb

Daik-/dek/ Bep-lb&p/ Mab-lmrebl Teeg--/tig/

Generalization vocabulary

Daik/Daiku-Nouns Bep/Bepu-Nouns TeegiTeega-Verbs MablMaba-Verbs

REFERENCES

American National Standards Institute. (1969) Specificationfor Audio-

meters (ANSI S3 6-1969). New York: ANSI.

Aram, D.M. (1991) Comments on specific language impairment as a

clinical category. Language, Speech, and Hearing Services in

School, 22:&l-87.

Aram, D.M., Ekelman, B.L., and Nation, J.E. (1984) Preschoolers with

language disorders: 10 years later. Journal of Speech and Hearing

Research, 27~232-244.

Connell, P. (1987) An effect of modeling and imitation teaching proce-

dures on children with and without specific language impairment.

Journal of Speech and Hearing Research, 30: 105- 113.

Connell, P.J., and Stone, C.A. (1992) Morpheme learning of children

with specific language impairment under controlled instructional con-

ditions. Journal of Speech and Hearing Research, 35:844-852.

Curtiss, S., Katz, W., and Tallal, P. (1992) Delay versus deviance in

the language acquisition of language-impaired children. Journal of

Speech and Hearing Research, 35:373-383.

Dunn, L.M., and Dunn, L.M. (1981) Peabody Picture Vocabulary Test-

Revised. Circle Pines, MN: American Guidance Service.


Fluharty, N.B. (1978) Flrrhurty Preschool Speech und Luttgrruge

Screening Test. Allen, TX: DLM Teaching Resources.

Hedrik. D.L., Prather, E.M., and Tobin. A.R. (1984) Seyrrc~nc,Pdftz~,ctt-

tory of‘ Cotnt~~trnic~ution De\~eloptnenr. Seattle, WA: University of

Washington Press.

Johnston, J.R., and Kamhi. A.G. (1984) Syntactic and semantic aspects

of the utterances of language-impaired children: The same can be

less, Merrill-Pmltner Qtturferly, 30:65-86.

Johnston. J., and Ramstad, V. (1983) Cognitive development in preado-

lescent language impaired children. British Jorrtwul of Disorders of Comtnrrnic~arion, 18( 1):49-55.

Jiireskog. K.G.. and S&born, D. (1989) LISREL7 u Gttide lo lltr Pro-

grutw and Applicutions. Chicago, IL: SPSS Inc.

Kamhi, A. (1981) Non-linguistic and symbolic conceptual abilities in

language-impaired and normally developing children. Journul of’

Speech und Hearing Re.sc)urch, 241446-453.

Kamhi, A., Catts, H.. Koenig, L., and Lewis, B. (1984) Hypothesis

testing and non-linguistic symbolic abilities in language-impaired chil-

dren. Journul of Sptvch und Houring Disordt~rs, 49: 169- 176.

Kirk. S., and Kirk. W. ( 1968) Illinois Test c?f‘Ps~c.holingtristic. Abilities

(Rev. ed.). Urbana, IL: University of Illinois Press.

Leiter, R.G. (1969) The Leiter lntt~mufiontrl P~~t~fi,rntanc~e Swle. Chi-

cago, IL: Stoeliting Company.

Leonard, L. (1987) Is specific language impairment a useful construct‘?

In S. Rosenberg (Ed.), Advances in Applied P.svc,llolingrristics, Vol-

t/me I: Disorders of‘ First Lunguuge Development (pp. l-39). NY:

Cambridge University Press.

Leonard, L. ( 199 1) Specific language impairment as a clinical category.

Lungtruge, Sptjech und Hearing Services in Schools, 22166-68.

Nakamura, M.. Plante, E.. and Swisher, L. (1990) Predictors of novel

inflection learning by preschool-age boys. Jorrrnul of’ Speech und

Ht>uring Reseurch, 331747-754.

Nelson, L.K., Kamhi, A.G., and Apel, K. (1987) Cognitive strengths

and weaknesses in language-impaired children: One more look. Jmr-

nul of Speech uttd Hewing Disorders, 52:36-43.


Newborg, J., Stock, J.R., Wneck, L., Guidubaldi, J., and Svinicki, J.

(1984) Batelie Developmental Inventory Screening Test. Allen, TX:

DLM Teaching Resources.

Nippold, M., Erskine, B., and Freed, D. (1988) Proportional and func-

tional analogical reasoning in normal and language-impaired children.

Journal of Speech and Hearing Disorders, 531433-440.

Plante, E. (1990). Cerrbral Conjigarations among the Parents and Sib-

lings of Language-Disordered Boys. Unpublished doctoral disserta-

tion, University of Arizona, Tucson, AZ.

Plante, E., Swisher, L., Vance, R., and Rapcsak, S. (1991) MRI find-

ings in boys with specific language impairment. Brain and Language,

41:52-66.

Rom, A.. and Leonard, L. (1990) Interpreting deficits in grammatical

morphology in specifically language-impaired children-preliminary

evidence from Hebrew. Clinical Linguistics and Phonetics, 4(2):

93-105.

Savich, P. (1984) Anticipatory imagery ability in normal and language-

disabled children. Journal of Speech and Heuring Research, 27:

494-50 I.

Schery, T.K. (1985) Correlates of language development in language-

disordered children. Journal oj’speech and Hearing Disorders, 50:

73-83.

Tomblin, J.B. (1991) Examining the cause of specific language impair-

ment. Language, Speech, and Hearing Services in Schools, 22:

69-74.

van der Lely, H.K.J., and Harris, M. (1990) Comprehension of reversi-

ble sentences in specifically language-impaired children. Journal qf

Speech and Hearing Disorders, 55:101-l 17.

Weismer, S.E. (1991) Hypothesis-testing abilities of language-impaired

children. Journal oj’Speech mnd Hearing Research, 34: 1329-1338.

Whitmire, B., and Stone, C.A. (1991) Visual imagery skills and lan-

guage abilities of normal and language-learning-disabled children.

Learning Disability Quarterly, 14, Winter, 49-59.

Manuscript received September 18, 1992;

revised November 7, 1992.

relations among verbal and nonverbal cognitive skills in normal language and specifically...

Documents