Propositional attitudes

DOI: 10.4324/9780415249126-V028-1
Version: v1,  Published online: 1998
Retrieved March 07, 2021, from

3. Propositional attitude contents

The contents of propositional attitudes are propositions. That much is simply a matter of definition. But what are propositions? There are at least the following contenders: (1) collections of circumstances of evaluation, such as collections of possible worlds; (2) syntactic entities, such as interpreted sentences in natural language; (3) set-theoretic structures, such as Fregean thoughts (see Frege §§3–4). Only some of the strengths and weaknesses of these different approaches can be examined here (see Propositions, sentences and statements).

(1) One important virtue of the theory that propositions are collections of possible worlds – that is, collections of complete ways that things could be (see Semantics, possible worlds) – is that it directly captures the intuition that propositions involve a sorting among ways that things could be. For example, in believing that snow is white, one believes that the actual complete way that things are is among those complete ways that things could be in which snow is white; in desiring that snow falls at Christmas, one desires that the actual complete way that things are is among those in which snow falls at Christmas; and so on.

Perhaps the most important drawback of this theory is that its most plausible formulations seem to give the wrong results for attitudes with logically equivalent contents. For example, it entails that the belief that there are husbands is identical to the belief that there are wives. Because ‘There are wives’ is logically equivalent to ‘There are husbands’, the set of worlds where there are wives is one and the same as the set of worlds where there are husbands. Likewise, it makes the desire to prove that two and two are four identical to the desire to prove that arithmetic is not decidable, and the intention to draw an equiangular triangle identical to the intention to draw an equilateral triangle. As the objection is commonly put, the possible worlds theory delivers objects of the attitudes that are insufficiently fine grained – there are more distinct attitude contents than it allows.

In addition, there is a complication which needs to be addressed. Consider, for example, my belief that I am cold, or your belief that you are spilling sugar on the floor of the supermarket. These attitudes de se (about oneself), as they are called, need to be thought of as involving a sorting among ways that things might be centred on the subject – me, here and now, as it could be, rather than of complete ways things might be (see Demonstratives and indexicals; Content, indexical).

(2) One important virtue of the theory that propositions are syntactic entities – that is entities with sentence-like structure – is that it provides objects to discriminate between attitude contents which, as we have just seen, the possible worlds theory is unable to discriminate. For example, the sentence ‘There are husbands’ is distinct from the sentence ‘There are wives’; and the theory can exploit this fact to distinguish the belief that there are husbands from the belief that there are wives. They are attitudes to different though equivalent sentences.

However, if the sentences are thought of as being in a natural language, this advantage is purchased at a price; it provides distinctions where it is plausible that there are none. Consider, for example, our practice of translating and interpreting those who speak different languages. Clearly, ‘Snow is white’ and ‘La neige est blanche’ are distinct sentences – but surely we are not thereby obliged to say that it is impossible for a monolingual French speaker to believe that snow is white? Consider, too, the belief that p, and the belief that p and p. There is at least some temptation to think that these are not distinct beliefs.

One response is to argue that the sentences in question do not belong to natural languages but rather to a language of thought – that is, to a neural system of information or information storage which is supposed to have syntactic structure, and that this is something we all share (and share with animals like dogs that lack public language but do seem to have propositional attitudes) (see Language of thought). One difficulty with this suggestion is that it seems wrong that a mere analysis of propositional attitudes will commit us to the existence of a language of thought, a substantive view about our neural natures. It seems to be an empirical question, for example, whether mental representation is language-like or map-like (see Belief §3), but the language of thought hypothesis commits us in advance to the former alternative.

(3) One important virtue of the theory that propositions are set-theoretic structures – that is entities whose basic structure is described by set theory (see Set theory) – is that it promises to cope with the difficulties mentioned above for the syntactic and possible worlds accounts. However, before we can see why this is so, we need to give some explanation of the theory.

Consider, for example, the sentence ‘London is pretty’. A natural thought is that the proposition expressed by this sentence is somehow made up of the city of London (or some way of thinking about it) and the property of being pretty (or some way of thinking about the property of being pretty). More generally, the theory holds that propositions are composed from constituents which are somehow put together in an orderly (set-theoretic) fashion. The constituents in question include: individual objects (and perhaps modes of presentation thereof), which correspond to proper names and other singular terms; properties, which correspond to predicates; functions, which correspond to sentential operators (such as ‘and’ and ‘possibly’), and so on.

This view has the resources to discriminate between the belief that there are husbands and the belief that there are wives, one of the examples that gave the possible worlds approach trouble. Roughly, the first has in its propositional object women and the property of being married to men, whereas the second has men and the property of being married to women. The view also has the resources to identify the propositions expressed by the sentences ‘Snow is white’ and ‘La neige est blanche’ and so allow that speakers of different languages can share beliefs, one of the examples that gave the sentential theory trouble. Although the words differ, the objects and properties they stand for are the same and hence the propositions thought of as constructed from them are the same. However, the theory in its typical form has to distinguish between the belief that p, and the belief that p and p – the propositional object of the second is special in involving the operation of conjunction – and this is not clearly desirable.

Another problem with set-theoretic approaches is that they involve various kinds of ‘disreputable entities’ – intensional properties, modes of presentation (‘Fregean senses’), perhaps even set-theoretic functions – which many philosophers would like to exclude from their theories of the attitudes (see Intensional entities). They worry about how these entities can be tied to propositional attitudes, especially if these attitudes are thought of functionally. What causal role can intensional properties or sets, for instance, play? True, this is part of a hard question that arises for any account of the attitudes. We have been discussing competing views about what the objects of the attitudes are. But there is also the question as to how a given propositional attitude state gets attached to the proposition it is attached to; the question of what makes it true, for example, that a certain state is the belief that snow is white rather than that grass is green. We will set this important question aside except to remark that the set theoretic view of the nature of the attitudes is often thought to be especially ill placed to tackle it.

Plainly, there is much more to be said about each of the options discussed here (and about other options as well). The important point is that there are many dimensions to propositional content, and that it is far from easy to provide a theory which scores well on every dimension. A good theory of propositional content should have all of the following features:

  1. it should be naturalistically respectable – that is compatible with the outlook of natural science;

  2. it should capture the normative dimensions of propositional content;

  3. it should allow a fine-grained discrimination of the contents of propositional attitudes (logical equivalence is not sufficient for identity);

  4. it should not tie the possession of propositional attitudes too tightly to particular languages (for example, it should allow that the same attitudes can be had by creatures which do not share a language – and, indeed, that attitudes can be possessed by creatures which do not speak any language at all);

  5. it should make it clear how states which possess that kind of propositional content can play a role in the production of (rational) action.

Citing this article:
Oppy, Graham. Propositional attitude contents. Propositional attitudes, 1998, doi:10.4324/9780415249126-V028-1. Routledge Encyclopedia of Philosophy, Taylor and Francis,
Copyright © 1998-2021 Routledge.

Related Searches


Related Articles