The signs of a savant: Language against the odds (review)

On one level, The signs of a savant (TSS) is a detailed account of how a linguistic savant, Christopher—who is mildly autistic (and unwilling to make eye contact with interlocutors) and apraxic (i.e. has impaired motor coordination and an inability to copy the gestures of others), and has poor spatial awareness—deals with the challenge of learning British Sign Language (BSL). BSL is a language whose mode of expression uses articulators in which Christopher has deficits: head, face, arm, and hand gestures located at different points in space. On another level, it is an ambitious attempt to further develop a modular model of the human mind first sketched in Smith & Tsimpli 1995. The aim of this model is to provide—both for spoken/written and signed languages and for nonimpaired/impaired speakers/signers—an account of the knowledge states that underlie performance. The range of linguistic and nonlinguistic performance domains that the model is intended to cover is impressively broad: the comprehension and production of language, conversation, translation, judgments of well-formedness, IQ tests, ‘theory of mind’ tests, and tests of memory (163). Additionally, in the specific case of signed languages, the model aims to offer some explanation for why certain signs might be perceived as iconic (that is, depicting a characteristic of the referent, as in the case of the BSL sign for ‘milk’ where the hands move as if milking a cow), and for the parallels between facial action and prosody in spoken language. While the evidence base for the model is in principle any study that yields observations on, for instance, comprehension and production, unsurprisingly most of the evidence adduced in TSS comes from Christopher’s performance in the spoken/written languages he knows, from his learning of BSL, and from his performance in various nonlinguistic tasks (e.g. tests of memory, theory of mind, artistic ability).

In addition to his native English, in which his ability ‘indicates a superior IQ in excess of 120 (a level more than sufficient to enter University)’ (1), Christopher can speak, write, and translate more than twenty languages (although with varying degrees of knowledge). Smith and Tsimpli’s (1995) investigations led them to conclude that the source of this talent is an exceptional capacity for learning the lexical and surface morphological properties of new languages. His knowledge of the syntax of second languages (where syntactic properties differ from English), however, is limited: ‘a large proportion of his second language activity consists in introducing items of vocabulary from a wide variety of languages into English syntactic structures’ (24). In contrast to this enhanced linguistic ability, Christopher is impaired in a number of nonlinguistic abilities: a level of drawing ‘that hint[s] at ineducability’ (2), an inability to conserve number (as displayed in the tendency to think that there are more beads in a set spread out on a wire than in a more tightly packed set on another wire, when in fact the reverse is the case), and a failure to adopt another person’s point of view on some theory-of-mind tests (but interestingly, not all), among others. On the basis of these findings, Smith and Tsimpli proposed (part of) a model of the human mind that consisted of Fodorian modules for vision and audition (that is, modules that are domain-specific, mandatory, informationally encapsulated, with representations constructed from a perceptual vocabulary) and non-Fodorian ‘central’ modules for language, theory of mind, and face recognition, as well as a knowledge base and an executive processor whose function is to coordinate information from the different modules.

TSS elaborates and refines this model partly on the basis of new evidence from Christopher’s performance on nonlinguistic tasks, but predominantly on the basis of how he learns BSL. Given the opposing tensions created by Christopher’s enhanced linguistic ability and his impairments, TSS (40–41) predicts that:

1. i. his linguistic talent should override his...