Programa de Optimización Prosocial
3.3 Cartas descriptivas
3.3.3 Sesión 3 El laboratorio personal
Joanisse (2004) and Tomblin and Pandich (1999) argue that children with G-SLI are at the lowest end of a normal distribution of grammatical abilities. Therefore, contrary to van der Lely, they claim that G-SLI does not exist as a phenotypically and genotypically separate subgroup. This claim appears to be based, at least in part, on misinterpretations of van der Lely's data and the conclusions she and her colleagues draw from them. For example, Joanisse stresses that a rule-deficit account (e.g. Pinker, 1991) would not predict van der Lely and Ullman’s (2001) finding that irregular past tense forms are impaired to the same level as regular past tense forms, and he says that a delay in irregulars is also found in typically developing children. However, what van der Lely and Ullman’s data show (and I
replicate this finding in Chapter 8) is that typically developing children show an advantage for regular verbs over irregulars, whereas G-SLI children show no such regularity advantage. Therefore van der Lely and Ullman are justified in interpreting their data as showing that regulars are affected more than irregulars. Tomblin and Pandich (1999) find no children who have grammatical deficits but no vocabulary deficits in their studies, but in fact van der Lely has never claimed that G-SLI children show a profile of poor grammar but intact vocabulary. Instead she contends that lexical learning requires, among other skills, the ability to use syntactic cues, and that difficulties in lexical learning could be secondary to the syntactic deficit (van der Lely, 1999).
Regardless of the validity of G-SLI as a subgroup, the point remains that van der Lely and colleagues have strict criteria for their choice of participants, and this detailed characterization remains essential. The issue of whether G-SLI children are at the tail-end of a normal distribution of grammar abilities, or a subgroup with grammatical behaviour different to the norm, is one that does not impact on the value of the linguistic findings presented in this thesis. The work in this thesis provides a detailed phonological phenotype of G-SLI, particularly as regards the impact of phonology on morphology. Establishing precise phenotype/genotype relations is an essential pre-requisite for studies seeking to identify the genetic basis of SLI (e.g. Lai, Fisher, Hurst, Vargha-Khadem & Monaco, 2001; SLI consortium, 2002). Ultimately, it may only be through understanding the genetic and neurological underpinnings of SLI that we can settle the issue of whether G-SLI children are qualitatively different from children with normally-developing language and from other SLI children. Comparing the genetics of different subgroups may provide an explanation for the heterogeneity in SLI, as it is possible that the observed phenotypic variability is a result of genetic variation (van der Lely, 1999; van der Lely & Stollwerck,
1996).
A detailed investigation of language is also valuable because it may be that when language structures break down, they break down in similar ways whatever the deficit, be it SLI, language delay, Down’s Syndrome, Williams Syndrome etc. The linguistic findings from the G-SLI population, and the tools developed for probing language structures, may then help advance our knowledge of language deficits in other developmental disorders. Relevant to this issue is the existence of a conflict between clinical and theoretical aims when selecting SLI children for research. Clinically-orientated research can use inclusive criteria in participant selection because impaired language may require the same therapy whether or not children have co-occurring deficits such as low non-verbal IQ, Asperger’s Syndrome, dyspraxia, or Down’s Syndrome. For theoretically-oriented research, however, the inclusion of children with co-occurring deficits makes it harder to distinguish causal
factors. For example, if children with both low non-verbal IQ and SLI are studied, we cannot rule out the possibility that the group’s poor performance on a language task is the result of low IQ.
In addition to the clinical/theoretical divide is a further conflict of aims: those of the cognitive sciences versus those of linguistics. Whereas developmental cognitive science is interested in what population as a whole does, current linguistics is inspired by the Chomskyan notion of an individual grammar, whereby studying what goes on inside the head of any particular speaker is of value in exploring the boundaries of grammatical knowledge. Therefore a tension exists when researching language disorders between using pure groups such as G-SLI and using more inclusive groups: the more inclusive the group, the larger the number of children that can be studied, but such research tends to miss out on the linguistic details. At the opposite extreme, concentrating on the detailed linguistic behaviour of one or two children runs the risk that this behaviour is rare and not representative of the wider population. The work in the chapters that follow tries to maintain a balance between investigating the characteristics of the G-SLI group as a whole, while also commenting on the individual linguistic behaviour of individuals, and how this can inform theories of linguistics and cognitive science. I do not pretend to have got the balance right, but I believe that the framework I set out has benefits for both levels of analysis.