Content: wide and narrow

5. Theories of wide content

In the philosophy of psychology, one of the main motivations for varieties of externalism has been a desire to provide a naturalistic theory of content, one that explains content in terms amenable to science (see Reduction, problems of; Mind, philosophy of). Externalist intuitions that arise in relation to examples like those involving Beth and Anne and water and XYZ in §2 above naturally lead to the idea that causal relations between subject and environment play a key role in determining content. It appears that it is because Anne enters into causal interactions with Beth that the contents of the beliefs she expresses with the name “Beth” are about Beth and no other thing. Equally, it appears that it is because Oscar₁ has acquired his concept through causal interactions with water that his concept applies to water and not to other things. And causal relations are amenable to scientific explanation.

The basic idea of causal theories of content is that a type of mental representation R extends over (is true of) Fs if, and only if, a certain kind of causal relation holds between instances of F and occurrences of R. The idea has no obvious application to logical contents, such as the negation or conditional, and causal theorists do not typically think that it applies to them. But they do typically take it to apply to all empirical concepts. We also have representations that do not apply to anything in the real world, such as unicorn and Vulcan. Since these things do not exist, they can have no causal impact on our brains. Fodor (1990, 101) suggested that the contents of non-singular representations like unicorn could be accounted for by counterfactual causal relations between properties and representations. The property of being a unicorn is real, though uninstantiated, and instances of it would cause occurrences of representations. Segal (2009) argued that such contents would be narrow, not wide.

Even within the range of representations that have a real-world extension, it is obviously not true that R extends over Fs if, and only if, all and only instances of F cause an occurrence of R. Among the exceptions are these: not every planet causes people to think planet; someone might mistake a star for a planet, in which case a star would cause an occurrence of planet. On seeing a picture of a planet, someone might think planet. When someone sees a planet and thinks planet, a proximal cause of the occurrence of planet is a pattern of cell firings in the retina; an occurrence of planet might be caused by an electric shock caused by an electrode.

Proponents of causal theories have tried to deal with these problems in various ways. Some appeal to normal conditions, holding that R extends over Fs only if instances of F would cause an occurrence of R under normal conditions (Stampe 1977). Others appeal to conditions that obtain during a period of learning, the period when R came to acquire its extension: under those conditions, all and only Fs caused occurrences of R, and that is why R comes to have the extension F (Dretske 1988). A more popular type of proposal appeals to evolutionary functions. The idea is that representational systems in the mind/brain have evolved to perform specific functions that benefit the organism. The belief system, for example, has evolved so that its states are causally sensitive to events in the environment in systematic ways. The system then, in turn, causes the organism to act in ways that promote survival, given those events. So, the theory goes, R extends over Fs if, and only if, a causal connection of a certain kind between Fs and R is derived from the brain’s evolutionary function (e.g., Millikan 1989; Papineau 1993).

Fodor (1987) addressed the problem of error (mistaking a star for a planet) with an “asymmetric dependence” condition. He noted that the only reason a star might cause an occurrence of planet is because planets cause occurrences of planet. The converse is not true: planets would still cause occurrences of planet, even if stars did not. The causal relation between stars and planets is in this way asymmetrically dependent on that between planets and planets. Fodor constructed a causal theory of content around this insight.