Principle Investigator: Johanne Paradis, Department of Linguistics, University of Alberta

Co-Investigator: Martha Crago, School of Communication Sciences and Disorders, McGill University

Collaborators: Lydia White, Department of Linguistics, McGill University; Mabel Rice, Child Language Doctoral Program, University of Kansas

Specific Language Impairment (SLI) is a relatively frequent developmental language disorder that can have non-trivial consequences for the lives of affected children, particularly in the domain of educational outcomes. Nevertheless, researchers are still grappling with fundamental questions regarding the nature of this disorder. For example, there is (1) limited knowledge of the characteristics that definitively mark impaired language development, (2) a lack of agreement on what the underlying mechanisms of this disorder are, and (3) almost no systematic understanding of the limitations specific language impairment imposes on the ability to learn more than one language.

Why is there no fully agreed upon set of characteristics that definitively marks a child’s language as impaired? One reason is because many of the problem areas in the speech of children with specific language impairment are also found in other populations of language learners. Thus, the linguistic properties that are exclusively the outcome of a disorder need to be clearly distinguished from the normal developmental process of language learning. While there has been a great deal of research aimed at differentiating the incompletely-learned language of children with SLI, and younger normally-developing children, much less attention has been paid to differentiating the language of children with SLI from another group of normally-developing children, those who are learning a second language. A major thrust of this research programme is to comprehensively compare and contrast impaired versus second language toward the goal of isolating the unique characteristics of impaired language.

Part of being able to distinguish between normal language learning and impaired language learning lies in the ability to pinpoint the underlying mechanism of impairment. Some researchers have proposed that specific language impairment is the result of a general processing deficit that makes it difficult for affected children to take in linguistic input, and to access the linguistic knowledge they have. Other researchers argue that children with SLI have specific deficits at the level of linguistic representation, and thus, cannot develop all aspects of linguistic knowledge that unaffected children can. Comparing second language learning children and children with SLI regarding whether they display deficits in representation or processing is another dimension of this research programme. This dimension will not only contribute to the goal of defining SLI, but will also inform a fundamental issue about the nature of impaired language and what sets it apart from other incompletely-learned language like that of second language learners.

Can children with SLI learn a second language? Even though there is virtually no research directly addressing this question, it is commonly assumed among educators and speech-language pathologists that acquiring two languages would be too challenging for children with an impaired language faculty. A third strand of this programme of research involves examining second language learning by children with SLI, in order to provide ground-breaking evidence to address this question. Furthermore, documenting the language of second language learners with SLI would contribute substantially toward the goal of defining SLI development by providing a contrast between normal and impaired second language learning.

Our research programme will yield a set of comparative learner group profiles for (1) monolingual children with specific language impairment, (2) children with language impairment who are learning a second language, and (3) normally developing children who are learning a second language. We will develop these profiles from a series of interconnected studies looking at longitudinal expressive language abilities as well as receptive language abilities in terms of representational knowledge and processing of linguistic knowledge. These studies are focused on grammatical morphology because this aspect of language has been found to be a hallmark vulnerable area for both children with SLI and second language learners, and thus, is a pivotal aspect upon which to differentiate these two groups. Comparing these profiles will enable us to contribute substantially to defining the characteristics of SLI and of second language learning, to understanding the mechanism underlying disordered language, and to exploring the limits of the impaired language faculty for bilingualism.

Finally, the outcomes of this proposed research programme will make original and important contributions not only theoretically, but also to the lives of children with specific language impairment. These learner profiles can be used to inform the differential diagnosis of impaired versus normal second language learners, and to inform the educational debate on the suitability of bilingual and immersion programmes for children with SLI. As such, these outcomes would be relevant to many multilingual communities across the globe.

A central concern in the field of research on specific language impairment (SLI) in children is how to isolate the unique properties of impaired language development from those of language development in normally developing populations, and in so doing, specify the defining characteristics of SLI. While much research has focused on comparing monolingual, normally-developing children with impaired children, very little research has compared children with SLI to other language learner populations such as children learning a second language (L2). Consequently, a 1999 report from an expert study group at the American National Institutes of Health specified that differentiation between L2 learning and SLI should be a priority area for future research on defining the characteristics of SLI. This programme of research will build on and extend our prior comparative research on the grammatical abilities of L2 children and children with SLI with the goal of differentiating between these populations in order to delineate the unique properties of impaired language.

The first major objective of this programme of research is to develop a set of learner profiles for three groups of children: (1) monolingual children with SLI, (2) children with SLI who are learning a second language, and (3) normally developing children who are learning a second language. These comparative profiles will make an original empirical contribution toward defining both the properties of specific language impairment and of second language learning. This contribution will have considerable clinical and educational relevance in a multilingual context, responding to vital concerns in Canada and the world at large. For example, the results of the proposed research will provide a basis for developing explicit criteria for the differential diagnosis of normal L2 children from children with SLI, and impaired L2 children from normal L2 children. The second major objective of our research programme is to contribute to explanatory theories of SLI by using these profiles to address the following issues: Is there resolution of specific language impairment over time? What is the nature of the deficit underlying SLI? and What is the capacity of the mind for second language learning under impairment?

Specific Language Impairment is a developmental language disorder affecting approximately 5-7% of the general population, and the consequences of this disorder typically persist throughout the lifespan, and often negatively affect educational outcomes (Leonard, 1998). Children with SLI exhibit lexical, grammatical and pragmatic deficits when compared with normally developing age-mates, but are within the normal range on other developmental measures such as IQ and social-emotional abilities, and have no neurological damage or significant hearing loss. Thus, their impairment is ‘specific’ to language. Consequently, SLI is partially defined by exclusionary criteria (i.e., by what it is not; for example, language impairment not due to hearing impairment). However, there is not yet a fully agreed upon group of characteristics of SLI that could comprise a well-defined set of inclusionary criteria in the expressive and receptive language abilities of these children. Therefore, an expert study group convened by the American National Institutes of Health has declared that a central concern for research into SLI should be the continued investigation of the characteristics of the language of children with SLI that will differentiate this clinical population from various normal developing populations (Tager-Flusberg & Cooper, 1999).

During the last decade, a number of studies in both English and other languages has pinpointed grammatical morphology as a vulnerable area in the expressive language of children with SLI (reviewed in Leonard, 1998; Crago & Paradis, in press-a). This aspect of language is thus a promising area for further exploration in the search for the defining characteristics of the language of children with SLI. In particular, the problematic development of verb-related morphology is considered a possible hallmark of SLI in English-speaking children. (Leonard, 1998; Rice & Wexler, 1996; Tager-Flusberg & Cooper, 1999). English-speaking children with SLI have greater difficulties supplying obligatory tense and agreement inflections and auxiliary verbs than they do supplying obligatory morphemes in the nominal domain (Bedore & Leonard, 1998; Clahsen et al, 1997; Leonard et al, 1997; Oetting & Rice, 1993; Rice & Wexler, 1996). Errors with morphology in the language of children with SLI are more frequently those of omission (no form is supplied) than commission (incorrect form is supplied), and error patterns tend to be ones of intermittent use in obligatory context, rather than absolute absence of use in obligatory context.

However, a complicating factor for the development of a set of clear characteristics of SLI based on incompletely mastered verbal morphology is the fact that there are other populations of language learners who share the same incomplete learning. Younger, normally-developing English-speaking children demonstrate similar patterns of the optional use of certain verbal morphemes to children with SLI (Rice & Wexler, 1996; Rice, Wexler & Hershberger, 1998). In addition, another group of normally-developing children, those who are in the process of acquiring a second language, make errors with morphology (Dulay & Burt, 1973, 1974; Grondin & White, 1996; Paradis, Le Corre & Genesee, 1998; Prévost & White, 2000). In order to assess the value of incompletely mastered grammatical morphology in general, and finite verb morphology in particular, as defining characteristics of SLI, it is necessary to conduct comprehensive comparisons with both normally-developing younger children and child L2 learners. Substantial research has already been conducted comparing children with SLI to monolingual, younger children, but there has been very little research comparing children with SLI to L2 children.

In our own prior research on SLI, we made a first attempt at comparing the expressive language of French-speaking children with SLI and French L2 children in order to investigate the similarities that we suspected might exist based on extant child SLI and L2 research. We found striking similarities between our two groups of seven-year-old subjects in their lexical verb diversity, use of clitic pronouns, omissions of obligatory arguments, but above all, in the use of nominal and verbal morphology (Crago & Paradis, 1999; in press-b; Paradis & Crago, 1998, 2000a, 2000b, in press). The extent of similarities we found between the L2 children and the children with SLI poses a challenge for the particular proposal that finite verb morphology could be a definitive clinical marker distinguishing between normal and impaired populations beyond the primary acquisition years (Rice & Wexler, 1996). Because children learning a second language display many surface similarities in their language to those of children with SLI, teasing apart the differences between SLI and L2 development is crucial to isolating the unique properties of impaired language development. Furthermore, the comparison of L2 learners and children with SLI is particularly informative because same-age groups of both populations can be matched not only on language ability but also on cognitive and mental maturity. This means that SLI-L2 learner comparisons can help uncover deviant-and-incomplete from normal-but-incomplete language learning, without mental maturity discrepancies as a confounding variable. However, our own prior research comparing children with SLI to L2 learners was limited in scope, since we only examined expressive language use, and at single point in time. Since it is reasonable to assume that some differences exist between SLI and L2 development, further in-depth comparisons between these two learner groups aimed at uncovering differences will bring us closer to isolating the characteristics unique to impaired language learning.

Defining the characteristics of these learner populations also helps to address certain fundamental yet unsolved explanatory theoretical issues concerning specific language impairment and second language learning, namely whether these children’s incompletely-learned grammars resolve over time, and whether SLI and L2 learning differ in their resolution rates and patterns, as well as whether their grammatical difficulties are due to deficits in access to linguistic representation, or in the representation itself, and whether children with SLI have the capacity to learn two languages.

Rice, Wexler & Hershberger (1998)’s longitudinal study of nominal and verbal morphology use documents the patterns and rates of acquisition for children with SLI. In comparison to Rice’s findings, it is likely that L2 children differ from children with SLI in the speed with which they reach ceiling accuracy for the use of these morphemes (rate), or in the shape of their developmental curve (patterns) over time, or both. In other words, children with SLI and L2 learners may share an intermediate stage of grammar at one point in development, but not over time. Well-planned longitudinal comparisons between these two populations are needed since the lack of resolution and the extended period of grammatical difficulties evidenced in children with SLI may well be one of the most significant characteristics of this population. These comparisons will thus serve to delineate impaired grammars from other intermediate grammars. The data needed to conduct such comparisons cannot be obtained from the extant child L2 literature (cited above) because of methodological differences with Rice et al (1998) in data collection procedures and differences in the target morphemes examined.

Another unsolved theoretical dimension of SLI is the nature of the deficit underlying this disorder. Broadly construed, two positions have emerged on the nature of receptive language deficits in SLI. Some argue that children with SLI have specific representational deficits in the underlying knowledge of abstract grammatical properties like tense and agreement (e.g. Rice, Wexler & Redmond, 1999; Rice & Wexler, 1996). In contrast, others propose that limitations in the processing of linguistic input, access to linguistic knowledge stored in memory, or both, is at the root of language impairment (e.g. Ellis-Weismer et al, 1999; Windsor et al, 2001). Thus, researching receptive language skills in the domains of both representation and processing in these two populations is needed to provide multiple levels for detecting L2-SLI differences, as well as to inform the fundamental question concerning the underlying deficit in SLI. The need to resolve this theoretical issue of processing versus representational deficit has also been described by the NIH expert study group on SLI as a priority area for research. Comparisons of receptive abilities with morphology in children with SLI and L2 learners can lead to a particularly interesting contribution because there is no a priori reason to suppose that L2 learners would show information-processing limitations, and furthermore, because grammatical morphemes would be an important linguistic target for investigating the role of processing limitations, since both children with SLI and L2 children make intermittent omission errors with them. Studies of receptive language processing in children with SLI typically have been focused on word access and recall (Leonard, 1998), and studies of children with SLI’s receptive abilities with morphology have focused mainly on off-line representational knowledge (e.g. Rice et al, 1999). To the best of our knowledge, receptive language studies on processing and representation of L2 morphology have not been conducted with children.

Finally, there is another theoretical issue central to both the fields of impaired language development and second language learning. This concerns the capacity of the language faculty in children with SLI to encompass the acquisition of two languages. The scarce research on this topic does not directly address this issue. Paradis, Crago, Genesee & Rice (2000; under review) examined use of morphology in the expressive language of children with SLI who were simultaneous bilinguals from birth (not L2 learners), and examined only one point of time in their development. Bruck (1982) studied language impaired children learning an L2 through immersion education over time, but she mainly reports scores from standardized tests of language and other cognitive abilities, and does not report details of the children’s spontaneous expressive language. A rigorous study of the ability of children with SLI to learn a second language is needed to inform our understanding of the human mind and its language capacity, as well as to help resolve the often heated and singularly under-informed educational and clinical controversy over whether children with SLI should be expected to learn a second language

Our programme of research has the overall goal of gathering comprehensive data on the expressive, receptive, and on-line processing grammatical abilities of three groups of children: (1) normally developing children learning a second language, (2) children with SLI learning a second language and (3) monolingual children with SLI. The series of research questions and studies described below are intended to uncover the differences in developmental language between these groups, and in so doing, to identify some essential properties of impaired versus normal grammars as well as first language versus second language grammars. They are also designed to inform the various explanatory theoretical issues described above. In particular, this programme of research is intended to address the following three sets of questions with a study designed expressly for each set of questions:

Grammatical morphology in expressive language over time: [1] Do L2 children acquire mastery of grammatical morphology faster than children with SLI? [2] Do L2 children show different growth curve patterns in their development of grammatical morphology than children with SLI? Study 1 is aimed at collecting longitudinal, expressive L2 data directly comparable to those of Rice et al, in order to probe more extensively the differences between these learner groups with respect to the acquisition of grammatical morphology in expressive language over time. Results of this comparison will also bear on whether lack of resolution is indeed a hallmark of SLI.

Receptive knowledge and processing abilities with grammatical morphology: [3] Do children learning a second language and children with SLI show deficits in their receptive knowledge of grammatical morphology? [4] If they do show deficits in the receptive domain, can these be attributed to processing grammatical knowledge, the representation of that knowledge, or both? [5] Are the patterns the same for L2 children and children with SLI? Study 2 consists of a comparison of off-line and on-line tasks to investigate whether processing limitations are implicated in receptive language deficits in children with SLI and L2 children. These results will complement the expressive language data in that they may reveal important and additional differences between the two groups of children. Furthermore, results of Study 2 will inform the debate on the nature of the mechanism underlying SLI.

Grammatical morphological development in SLI children who are learning an L2:. [6] What are the characteristics of morphological acquisition in normally developing second language learners compared to children with SLI who are learning a second language? [7] How does specific language impairment affect the pattern and rate of acquiring the grammatical morphology of a second language? Studies 1 and 2 are designed to uncover differences between normally-developing L2 and impaired first language (L1) development in order to further our understanding of the nature of SLI, and to provide a basis for differentially diagnosing normal versus impaired incomplete language learning. An additional and essential piece in this puzzle is the ability to distinguish L2 learners who are impaired from the normal L2 population. The purpose of Study 3 is to document the development of grammatical morphology over time in children with SLI from the outset of their L2 learning, and in so doing, not only contribute essential information regarding the characteristics of impaired second language development, but also explore the capacity of children with SLI to become second language learners.