Cognitive Science Program

Knowledge of binding in normal and SLI children (1996)
Steven Franks, Phil J. Connell
Journal of Child Language, 23 (2), 431-64

Abstract

The properties of reflexives vary across adult languages with respect to (i) the domain in which a reflexive may be bound and (ii) the syntactic positions an appropriate antecedent may occupy. These two issues have been approached in GB theory in various ways, each with specific implications for acquisition. In this paper we examine these implications by testing normal and Specific Language Impaired (SLI) children for evidence of the binding domain and orientation properties of their grammars. The investigation reveals that, contrary to most previous claims, normal children acquiring English pass through a long-distance binding stage. SLI children, however, do not display this pattern, tending instead to behave like very young normal children in requiring the nearest available NP to be the antecedent. We argue that this constitutes an early binding stage not previously identified. Finally, we interpret our findings in terms of a conception of acquisition dubbed the 'competing grammars' model, according to which competing incompatible grammars may coexist in the mind of the learner.

The Conceptual Basis for Morpheme Learning Problems in Children With Specific Language Impairment (1994)
Phil J. Connell, C. Addison Stone
Journal of speech and hearing research, 37 (2), 389-98

Abstract

Two groups of children were exposed to instances of a nonlinguistic conceptual rule under controlled instructional conditions to determine whether the problems children with specific language impairment (SLI) have learning and accessing language rules extend beyond the language domain into the general cognitive domain. The performance of 20 children with SLI, aged 5:0 to 6:11 (years:months), was compared to that of 20 normally developing children matched for age and nonverbal ability. These children were taught under two instructional conditions that differed only in whether the child was asked to imitate the solution to a conceptual problem after each demonstration (imitation) or merely to observe it (modeling). Contrary to previous findings regarding linguistic rule-learning using auditory or visual symbol systems and similar instructional conditions, no difference was found between the extent of overall learning displayed by the normally developing children and those with SLI. Also, the performance of the children with SLI was not uniquely better under the imitation condition than under modeling, as had been the case with the learning of a novel morpheme in an auditory linguistic task. These results are interpreted as confirming the earlier assumption that the generally lower overall learning rate of the children with SLI on both the auditory and visual tasks reflects a specific linguistic rule-learning difficulty, rather than a general deficiency in rule induction.

Induction of a Visual Symbolic Rule in Children With Specific Language Impairment (1993)
C. Addison Stone, Phil J. Connell
Journal of speech and hearing research, 36 (3), 599-608

Abstract

In a design modeled after that of Connell and Stone (1992), two alternative instructional conditions (Modeling and Imitation) were used to teach a novel morpheme to a group of 21 children with Specific Language Impairment (SLI) (ages 5:0 to 6:11) and to 25 age- and ability-matched peers and 21 language-matched younger peers. In contrast to the use of the auditory modality in the Connell and Stone study, in the present study, the morphemes were embodied in a visual symbol system. The objects and meanings conveyed by the symbols were identical to those used in the earlier study. The present results indicated that both the children with SLI and their age-matched peers showed a benefit to their morpheme production from the requirement to imitate the use of the morpheme during learning trials. These results were interpreted as indicating that the group-specific benefit from imitation instruction on the part of the children with SLI found by Connell and Stone with their auditory task is unique to morpheme learning in the auditory modality. As in the earlier study, the children with SLI also tended to learn less than their age-matched peers in all conditions, indicating a possible general deficiency with symbolic rule induction.

Morpheme Learning of Children With Specific Language Impairment Under Controlled Instructional Conditions (1992)
Phil J. Connell, C. Addison Stone
Journal of speech and hearing research, 35 (4), 844-52

Abstract

Three groups of children were exposed to instances of a novel morpheme under controlled experimental conditions. The performance of 32 children with specific language impairment (SLI), aged 5:0 to 7:0 years (years:months), was compared to that of 24 normally developing children matched for age and nonverbal ability and 20 younger normally developing children matched for language development and nonverbal ability. The children were taught under two instructional conditions that differed only in whether the child was asked to imitate the new language form after each instance (imitation) or just to observe its use (modeling). Consistent with past research (Connell, 1987b), the children with SLI performed significantly better under the imitation condition than under modeling, but the age-matched controls showed no difference in response to instruction. The performance of the language-matched controls was similar to that of the age-matched controls, suggesting that the instruction-specific effect for the children with SLI is not merely a function of general language immaturity. Although the superiority of the imitation condition for the children with SLI was evident for test trials requiring production of the new morpheme (as in past research), no such effect was evident for comprehension trials. This differing effect of output demands suggests that the SLI-specific response to instruction is not a matter of different mastery of the new rule but rather is specific to the need to access the newly induced rule on production trials. The accessing of phonological representations as a possible explanation for the effect is discussed.

The Use of an Invented Language Rule in the Differentiation of Normal and Language-Impaired Spanish-Speaking Children (1991)
Celeste A. Roseberry, Phil J. Connell
Journal of speech and hearing research, 34 (3), 596-603

Abstract

The purpose of the present study was to determine whether the results of a language-teaching procedure could be used to identify specific language-impaired children in a group of bilingual children with limited English proficiency (LEP). An invented morpheme was taught to two groups of LEP children who had been previously identified as normal and specific language-impaired. The language-impaired group learned the morpheme at a slower rate than the normal children, thus allowing the two groups to be differentiated. The approach promises to circumvent many of the obstacles that impede current practices for identifying language impairment in the LEP population.