To Structure, or not to Structure?

Some accounts of mental content represent the objects of belief as structured, using entities that formally resemble the sentences used to express and report attitudes in natural language; others adopt a relatively unstructured approach, typically using sets or functions. Currently popular variants of the latter include classical and neoclassical propositionalism, which represent belief contents as sets of possible worlds and sets of centered possible worlds, respectively; and property self-ascriptionism, which employs sets of possible individuals. I argue against their contemporary proponents that all three views are ineluctably plagued by generation gaps: they either overgenerate beliefs, undergenerate them, or both. Belief is a relational affair. To be a bit more specific: To have a belief is to stand in an appropriate relation to some sort of semantic object. One major task for a theory of belief, then, is to give an account of these objects. Some philosophers recommend that we treat them as ‘unstructured’, typically using sets or functions. The idea here is not that unstructured doxastic objects lack structure altogether (whatever that might mean exactly) but rather that they lack anything like the constituent structure characteristic of the syntactic objects analyzed by linguists. Other theorists favor a ‘structured’ approach, using entities that formally resemble the sentences used to represent beliefs in natural language. To structure, or not to structure, the objects of belief? That is the question. To address it, we will examine three varieties of unstructured semantics: classical propositionalism, property self-ascriptionism, and neoclassical propositionalism. Though these views don’t exhaust the options for the unstructured camp, the failure of all three would constitute fair cause for alarm. My aim is to present new considerations which suggest that, despite persistent claims to the contrary, none of these accounts work: they either overgenerate beliefs, undergenerate them, or both. Interestingly, each generation gap shows up at a metaphysical extreme. Overgeneration affects an agent’s beliefs about interworldly states of affairs, maximally objective or impersonal facts like the number of primes or the size of the continuum; undergeneration involves beliefs about intraworldly matters, subjective or personal facts like the time of day. Updated versions of the original unstructured scheme finesse the second problem by incorporating local entities into doxastic objects: substituting individuals, or worlds Synthese 139: 55–80, 2004. © 2004 Kluwer Academic Publishers. Printed in the Netherlands.