Symposium
Chair: Gina Conti-Ramsden
Discussant: Paul Fletcher
In the field of child language, research with atypical populations hasrecently taken more of a centre stage. Theorists and practitionersinterested in child language development are finding that atypicalpopulations such as children with language difficulties (children withspecific language impairment (SLI) and late talkers) provide them withan interesting natural experiment to contrast different models forlanguage learning.This Symposium aims to foster the dialogue between academicsinterested in child language and those interested in child languagedisorders. Specifically we are interested in focusing on describinglanguage development and understanding predictors of later languagedevelopment. The first paper investigates the possible role of gestures in thelanguage development of late-talking toddlers and normally developingtoddlers. This presentation will discuss methodological issuessurrounding the sampling methods used to obtain information aboutgesture use. It will also address the question of whether gesturescan be used as a positive prognostic sign of later languagedevelopment. The second paper presents results from the first phaseof a detailed longitudinal study of six children with SLI. Thechildren were all at the single-word stage at the beginning of thestudy and were followed weekly via one hour video-recording sessionsat home. The paper addresses the particular role of verbs in childrenwith SLI' early multi-word speech. The third paper also investigatesthe use of verbs in children with SLI. This work is experimental innature and involves children with SLI learning novel verbs and nounsover a period of 5 weeks. This paper aims to discuss the relationshipbetween lexical development and morphological development in childrenwith SLI . This area is of particular interest given that childrenwith SLI are characterised par excellence as having particularproblems with verb morphology. The fourth paper focuses on latetalkers and attempts to identify early predictors for later languageand reading outcomes. This paper puts the work with infant childreninto a longitudinal perspective and addresses the important questionof the possible long-term sequelae of late-talking.They symposium will close with a discussion where the discussant willsummarise and draw conclusions from the above presentations. Thediscussant will also provide a cross-linguistic context for the abovediscussion by drawing from data involving children with SLI learningCantonese. Cantonese is particularly interesting and relevant giventhe very different role verbs play in Cantonese as compared toEnglish.
Details of individual items:
paper
Various studies have reported different relations between gestures andlanguage in young children. In typically developing children, somestudies report a relation between expressive vocabulary and gestureproduction (Acredolo & Goodwyn, 1985; Bates et al., 1988) while othersargue that the important relation is between gesture production andlanguage comprehension (Petito, 198x, 19xx; Bates et al., 1989). Adult use of gestures, on the other hand, appears to supplement thecontent of the verbal message. That is, they seem to function asperipheral or supportive communicative devices. Studies oflate-talking toddlers have indicated relations between use of gesturesand language comprehension that may be predictive of continuedlanguage delay or significant improvement in language skills (Thal,Tobias, & Morrison, 1991, Thal & Tobias, 1992). Research has alsosuggested that helping both typically-developing and delayed toddlersto learn gestures along with words may assist them with the languagelearning process, and does not impede it (Acredolo & Goodwyn, 1990;Brown, 1997). In the seminar proposed here, I will review the methodsused to measure language and gesture in these studies, and thedifferent outcomes reported, in an attempt to clarify the relationbetween use of gestures and both current and later languagedevelopment. I will then review the data relating to prediction ofoutcome and enhancing language learning, including some new studiesfrom the Early Identification of Risk for Language Impairment (EIRLI)Project in San Diego. Specifically, I will review different outcomeswhen gestures are sampled through imitation, confrontation production,spontaneous communication, and parent report. I will also spend sometime differentiating the kinds of gestures that are usually sampledthrough each method. Next, I will review my own studies that showthat gesture use is a positive prognostic sign. Finally, I willdescribe some new work in which we demonstrate prediction of languageat 28-months of age from samples obtained at 16-months. In that studythe number of different words and total number of gestures used forcommunicative purposes in a spontaneous sample, combined with thenumber of words that parents reported their children understood at 16months, predicted 100% of the children in the normal range and 71% ofthe children who fell below the 15th percentile for words produced onthe MacArthur Communicative Development Inventory at 28-months of age. I will also present data showing that imitation and confrontationproduction tasks do not differentiate children with different levelsof language comprehension at 16-months of age while parent report ofgesture production does. In conclusion, I will provide a summary inwhich I suggest a model for obtaining clinically useful informationabout gesture use for language prediction purposes and I will proposesome contexts in which direct teaching of gestures along with languagemay be appropriate.
paper
This paper will present results from the first phase of a detailedlongitudinal study of six children with specific language impairment(SLI). The children were all between 2;6-3;0 years at the beginningof the study, had normal non-verbal abilities, normal hearing, andnormal social-interactive abilities (non-autistic). All the childrenhad significant problems understanding language (scoring below the 8thpercentile in tests of language comprehension) and also hadsignificant problems using language (using only single words at thebeginning of the study and having less than 50 words in theirvocabulary). The children were seen at home every week and wererecorded in natural parent-child interactions for one hour. Thenaturalistic data was transcribed using the CHAT format of the CHILDESsystem. To our knowledge this is the first detailed longitudinalstudy of children with SLI's early grammatical development.This presentation will focus on how children with SLI constructmulti-word speech. In particular we are interested in the role ofverbs in children with SLI's multi-word speech: a) is there an absenceof verbs in children with SLI's first word combinations (based on thefindings of Conti-Ramsden & Jones, 1997 that children with SLI usemuch more nouns than verbs)? b) Are the verbs used 'model verbs' forparticular construction types (based on the work of Ninio (in press)which suggests that normal children's early multi-woird combinationsinvolve verbs which are excellent exemplars of the construction typethe child is attempting to use)? c) Are they general-all-purpose verbs(GAP verbs, also referred to as light verbs as suggested by the workof Rice & Bode, 1993)? d) Are they specific verbs, tied to particularconstruction types only (based on the work of Tomasello, 1992 whichsuggests that very young children use early verbs in a verb-island wayin specific contexts. Jones & Conti-Ramsden (1997) have suggestedthat children with SLI may undergo and extended verb-island period ofdevelopment where verb use is specifically tied to particular contextsand construction types and no generalisations are evident)?
paper
The present experimental project examines children with specificlanguage impairment (SLI)'s learning and use of verbs. This paperwill focus on children with SLI;s productivity with verbs, word order(i.e. understanding and use of agent and/or patient arguments todescribe transitive, reversible actions), and with verbal morphology(i.e. use of progressive 'ing', and past tense 'ed'). The projectused both a control group (normal language learning children matchedfor language stage and regular verb vocabulary size) and a controlconditions (learning and use of nouns). 20 children with SLI (aged4;5-5;5 years) and 20 normal control children (aged 2;6-3;3 years)participated in the study. The children participated in 10 trainingsessions during a five week period in order to learn four nonce nouns(e.g. toma) and four nonce verbs (e.g. meeking). Each of the fournonce verbs were presented to the children in one of four argumentstructures: no argument (Look! Dacking! This is called Dacking),agent and patient arguments (Minny is meeking Pooh), patient only(e.g. Tamming Daisy), or agent only (tigger is gopping). This sameformat was used to present the children with four nonce nouns. Following the training sessions children were tested on theircomprehension and expression of the novel nouns and verbs.Specifically, the present paper will discuss a)whether children withSLI are slower at learning verbs and require a greater number ofexposures to a given verb form before it can be established in theirvocabulary, b) whether the number and composition of children's verbvocabularies affects verb learning and the ability to generalise anduse verbal morphology (the SLI critical mass hypothesis, Conti-Ramsden& Jones, 1997); and c) whether argument structure affects childrenwith SLI's learning of verbs.Conti-Ramsden, G. & Jones, M. (1997). Verb use in specific languageimpairment. Journal of Speech, Language and Hearing Research, 40,1298-1313.
paper
Early predictors of language and academic outcomes at age 13 in 24 late talkers will be summarized. The late talkers' outcomes will be compared to those of 12 comparison children matched as toddlers on age, gender, SES, and nonverbal ability. At intake (24-to-31 months of age), all of the late talkers had normal nonverbal cognitive ability, a Reynell Receptive Language score within three months of chronological age (CA), and a Reynell Expressive Language score at least 6 months below CA. At age 3, 41% of the late talkers were above the 10th percentile for MLU; by age 4, 71% of late talkers scored in this range. The late talkers scored significantly below the comparison children on both MLU and IPSyn (a measure of emergent morphosyntax) at both age 3 and age 4. Only 15% of late talkers scored below the 10th percentile at age 5 on the PEST, an expressive grammar measure, but the groups were significantly different on this test. The general pattern of results for the age 5-to-9 follow-ups was that late talkers typically performed in the average range on most measures, but that they were significantly worse than comparison children on most language tasks through age 9 and on reading tasks at ages 8 and 9. To analyze the age 13 outcomes, principal components analysis was used to combine the various measures into four factors: reading/writing, vocabulary, grammar, and auditory memory. There were significant group differences on all four factors at age 13, with the effect size between groups ranging from 2/3 and 2 SDs. Despite the fact that the late talkers continued to be inferior to comparison children on most language and reading tasks, they generally scored in the normal range on these measures. For the groups combined, Reynell Expressive z-score at intake had significant correlations with all age 13 outcome factor scores (r.39 for vocabulary, .68 for grammar. .78 for auditory memory, .50 for reading). Expressive One-Word Picture Vocabulary Test (EOWPVT) score at age 3 was also significantly correlated with age 13 factors (r values ranging from .43 to .59). Age 3 MLU and IPSyn, both grammar measures, had significant correlations with auditory memory at 13 and MLU significantly predicted grammar at age 13. The Verbal factor of the Stanford-Binet IV Intelligence Sale administered at age 3-6 was a significant predictor of vocabulary (r.71), grammar (r.66) and reading (r.58) at age 13. Among the age 4 measures, the EOWPVT was a significant predictor of age 13 vocabulary (r .63), and MLU and IPSyn both predicted grammar (r.54 and .50), reading (r.49 and .53), and auditory memory (r.68 and .55) at age 13. In summary, a variety of expressive language measures from age 2-to-4 significantly predicted age 13 outcomes in vocabulary, grammar, reading, and auditory memory in a combined sample of late talkers and matched typically developing peers.