Modules, modals, maths and the mind

The standard cliché has it that the human mind-brain is the most complex entity in the universe. Understanding such complexity requires the genius of simplification, and noone has been better at this than Jerry Fodor. He has been responsible for two of the most controversial, and fruitful, claims in the psychology of mind and language. In 1975 he argued that you cannot learn a language with greater expressive power than one you already know (hence that there must be a rich innate language of thought); and in 1983 in arguing that the structure of the mind is in part modular, he suggested that the ‘‘central system’’, responsible for puzzle solving and the fixation of belief, was both non-modular and pretty much inscrutable. Though there have been significant changes in his thinking over the last 20 to 30 years, these still seem to reflect his basic beliefs. But recent research indicates that both positions may be wrong. Let’s start with modularity. I wish to impugn only one half of Fodor’s claim about the modularity of mind: I am still convinced of the validity of his characterisation of ‘‘input systems’’ such as vision and audition in terms of their being domain-specific, fast, mandatory, informationally encapsulated, and so on. But the structure of the central system seems to be less opaque than his pessimistic observations suggest. Of the proposals around, many owe their inspiration ultimately to Chomsky’s (1975, 1984) notion of module. For current purposes I shall use the framework that Ianthi Tsimpli and I have been developing over a number of years, exploiting the idea that the central system is compartmentalised into a number of quasimodules (Smith, in press; Tsimpli & Smith, 1998). Quasi-modules are like Fodorian modules, but use a conceptual (not a perceptual) vocabulary, and are not informationally encapsulated: i.e. they can ‘‘talk’’ to each other. Examples are Theory of Mind, Moral Judgement, Music, Common Sense, Social structure, the Number Sense, Personality structure, Folk Physics, Folk Biology, and more. The full set of such faculties is a matter of ongoing debate, but reasonably clear evidence for at least this inventory comes from the existence of ‘‘double dissociations’’, where different subjects may lose one ability while retaining another and vice versa (Smith, 1998). Language is a special case: it seems to be partly modular in Fodor’s sense, and partly quasi-modular (Smith & Tsimpli, 1995). The details are complex, and there is also the need to accommodate the emotions (see Smith, 2002), but this is enough to be getting on with. Looking at the inter-relations among these various entities should cast light on the claims I started with. The evidence comes from the growth of the child’s linguistic abilities and how these feed into and are influenced by the development of other modules of the mind, especially Theory of Mind and the Number sense. ‘‘Theory of Mind’’ is the mental faculty that enables us to adopt the viewpoint of other people, even where this leads to our entertaining representations of the world that differ from our own or from the ‘‘truth’’. I may be convinced that the prime minister is a war criminal, but nevertheless know that you consider him a paragon of every virtue. Such conflicting views do not lead to contradiction, precisely because I can attribute them ‘‘meta-representationally’’ to different people. This ability is not present in two-year-old children, but develops gradually around the ages of three to four, as manifest in the ability to pass so-called false belief tasks such as ‘‘Sally-Anne’’ and ‘‘Smarties’’ (see e.g. Wellman, 1990). For current purposes what is relevant is the role of meta-representation in the child’s developing mastery of modal verbs like must. There is an (apparent) ambiguity between ‘‘deontic’’ (or ‘‘root’’) and ‘‘epistemic’’ uses of modals (plus a possible third ‘‘alethic’’ interpretation (Smith, 1989)), as shown in the three versions of (1) given in (2):