Cognitive Capacities, Mental Modules, and Neural Regions

Dan lloyd (2011) issues a salutary warning against the assumption of what I shall call neural modularity—the view that there is a one-to-one mapping between cognitive functions and distinct brain regions. He shows how the assumption can distort the interpretation of neuroimaging studies and blind researchers to global structures and activity patterns that may be crucial to many aspects of cognitive function and dysfunction.

In this note, I want to add a further dimension to the discussion by making connections with the notion of mental modularity developed by evolutionary psychologists. What is the relation between mental and neural modularity? Do the arguments for massive mental modularity also support neural modularity? I offer some preliminary remarks on these questions and their bearing on issues in psychopathology.

Modern discussion of mental modularity takes its start from Fodor's The Modularity of Mind (1983). Fodor's conception of a mental module is a strict one, appropriate for peripheral systems, such as vision and language. Modules in his sense are special-purpose computational mechanisms that are to a large extent innate and hardwired, deliver shallow (nonconceptual) outputs, and operate in a way that is typically mandatory, fast, inaccessible, and informationally encapsulated. More recent writers adopt a more relaxed notion of mental modularity, which is applicable to central cognitive processes (e.g., Barrett and Kurzban 2006; Carruthers 2006; Coltheart 1999). Modules in this sense are functionally specialized mechanisms that are dedicated to particular cognitive tasks, such as theory of mind, biological classification, and cheater detection, and that can deliver conceptualized beliefs and desires as outputs. Their key property is domain specificity, and it is an empirical question to what extent a given module possesses the other features on Fodor's list. It is arguable that evolution would have favored the development of such central modules as well as peripheral Fodorian ones. Complex evolved systems are typically composed of specialized subsystems, and it is likely that specialized, domain-specific cognitive mechanisms would be more efficient than general-purpose ones, especially given the wide range of adaptive problems faced by humans and other animals. Building on such arguments, some theorists maintain that the mind is composed wholly, or mostly, of modular subsystems—a view known as the "massive modularity hypothesis" (e.g., Carruthers 2006; Cosmides and Tooby 1992; Pinker 1997; Sperber 1994).

My aim here is not to defend massive modularity, but simply to ask whether mental modularity involves a commitment to neural modularity as well, and thus whether the arguments for the former carry over to the latter. It will be helpful to think in terms of three descriptive enterprises: personal psychology, subpersonal psychology, and neurology (compare Dennett 1981). Personal psychology is an idealized competence theory; it ascribes psychological traits to organisms on the basis of their behavior. Talk of cognitive capacities is pitched at this level. Subpersonal psychology posits internal information-processing mechanisms that support personal-level psychological traits. The massive modularity hypothesis belongs to this level. Neurology, of course, deals with the fine-grained biological structures and processes in which the other levels are realized. Now, neural modularity is a claim about the relation between personal psychology and neurology; it is the claim that each cognitive capacity is realized in a localized brain region. And the question of whether massive mental modularity entails neural modularity breaks down into two subquestions—one concerning the relation between personal and subpersonal psychology, and the other the relation between subpersonal psychology and neurology. The first question is whether each cognitive capacity is subserved by a single proprietary module; the second whether each module is realized in a distinct brain region. If the answers to both questions are positive, then mental modularity is equivalent to neural modularity. I take that, for a given organism, the answers might be positive, but I shall argue that there is no reason to think that they must be, or even that it is particularly likely that they are—and thus that mental modularity does not entail neural modularity.

Consider first the relation between mental modules and...