Version: v1, Published online: 1998
Retrieved March 07, 2021, from https://www.rep.routledge.com/articles/thematic/propositional-attitudes/v-1
5. Defining propositional attitudes
The characterization of propositional attitudes in §2 was very rough, and leaves room for the thought that functionalism involves circularity of definition. We in effect identified the behaviour beliefs cause as behaviour that tends to realize desires, and this bit of inter-defining looks rather like circularity. However, the characterization can be improved in a way which makes it clear that this is not so – and, as an added bonus, also makes clear just what would be involved in the denial of the existence of propositional attitudes. The secret is to take the general account of the definition of theoretical terms provided by David Lewis and apply it to the propositional attitudes.
Roughly, the idea is to collect together a class of sentences involving the verbs of propositional attitude and treat this class as a simultaneous definition of the propositional attitudes. The same idea can be applied to other webs of theoretical terms. A very simple case is the definition of a wife as someone with a husband, and a husband as someone with a wife. We remove the air of circularity when we note that we could spell matters out as follows: x is a wife if there are two people, one male and one female, they are married to each other, and x is the female one. Consider, for a more complex example, the artworld and its denizens. Artists are producers of art works. Audiences are viewers or consumers of art works. Critics are critical assessors of art works. Galleries are places for the exhibition of art works. And art works are objects produced by artists for consumption by audiences, assessment by critics and display in galleries. Taken together, these sentences – or some more elaborate version thereof – can be taken to provide a simultaneous definition of the key terms: artist, audience, critic, gallery, artwork. And there is no (vicious) circularity involved because the other terms which feature in the various sentences – ‘production’, ‘consumption’, ‘display’, ‘object’, ‘view’, and so on – are (assumed to be) independently understood. These terms in effect specify the key interconnections, and the definition of artist, work of art, and so on that is delivered amounts to saying that an artist is anyone who is appropriately interconnected.
The sentences which are collected together are either prima facie analytic truths – that is, conditionals and biconditionals which encode the inferential practices constitutive of possession of the concepts under analysis – or else claims involving the concepts to be analysed which are regarded as prima facie common knowledge by those, or most of those, who possess the concepts. Under the former account, there is no guarantee that the sentences will be immediately available to those who possess the concepts; however, it might be hoped that reasonable and reflective people can be brought to assent to them when the sentences are drawn to their attention. Under the latter account, there is no requirement that the sentences encode inferential practices constitutive of possession of the concepts under analysis: for all kinds of synthetic and a posteriori claims may be selected. However, the following constraint is to be observed: it should be commonly held to be a condition on attribution of possession of the concepts which are up for analysis to a subject that they do not reject, or fail to accept, too many of the chosen sentences.
On either account, one might doubt that there are sentences of the kind required by the analysis. So, for example, it might be said that it is notoriously difficult to provide exceptionless generalizations about the connections between beliefs, desires, intentions, actions, and so on. However, it is important to note that there is nothing in the account which requires that the platitudes in question must be couched as exceptionless generalisations. Indeed, a plausible thought is that most of the platitudes will instead make claims about what is normally the case – that is about what happens when all other things are equal, or when conditions are normal. So, for example, on either account, there are sentences like these: a system of beliefs and desires tends to cause behaviour that serves the subject’s desires according to that subject’s beliefs; beliefs typically change under the impact of sensory evidence; desires aim to have the world fit them; beliefs aim to fit the world, and so on.
The final point to note is that our definition tells us that the propositional attitudes are whatever it is that satisfies the chosen set of sentences and so stand in the specified relations. In the simplest case, there will be a unique natural collection of candidates for the propositional attitudes. In that case, the sentences provide an explicit definition of the propositional attitudes. But things may not be so simple: perhaps there is nothing which satisfies the sentences exactly; and perhaps the only thing which satisfies the sentences is extremely unnatural, that is, gerrymandered. In these cases, we shall look for the best near-satisfier, where the criteria which guide our search are (1) considerations of naturalness; and (2) satisfaction of as many as possible of the more central platitudes. Put another way: application of a procedure which aims at achieving reflective equilibrium may lead to the rejection of some of the sentences, and perhaps to the elevation of further sentences to membership of the favoured class. Moreover, the procedure which aims at achieving reflective equilibrium is guided by judgments about (1) the naturalness of the resulting concept; and (2) the centrality or revisability of the sentences which are incorporated or rejected.
Of course, there is nothing in the account of this procedure which guarantees that there will be a point of reflective equilibrium: and, in the case in which there is no point of reflective equilibrium, we will need to give an error theory of the propositional attitudes – that is we will hold either that there were no coherent concepts characterized by our pre-reflective usage, or else that there simply is nothing to which our concepts apply.
Oppy, Graham. Defining propositional attitudes. Propositional attitudes, 1998, doi:10.4324/9780415249126-V028-1. Routledge Encyclopedia of Philosophy, Taylor and Francis, https://www.rep.routledge.com/articles/thematic/propositional-attitudes/v-1/sections/defining-propositional-attitudes.
Copyright © 1998-2021 Routledge.