Page 312 - Encyclopedia of Philosophy of Language

Language and Logic
possible world (at least on many theories; otherwise it is not a world, but corresponds to a unique world), and is a member of every true proposition. If John really is Tom's father, then the actual world is one in which John is Tom's father, so it is a member of the proposition that John is Tom's father. Sets of worlds bear relations to each other of inclusion, exclusion, and so on, and in possible-worlds' theory these become logical relations. For instance, if proposition P is a subset of proposition Q, then P entails Q. The set of worlds in which grass is green and snow is white is a subset of the set of worlds in which snow is white. Possible-worlds propositions are in this way neatly suited to tasks of set-theoretically modeling logical relations.
Construed in this way, propositions have constitu- ents, but these are worlds, as opposed to individuals and properties or concepts. So their structure is nothing like the structure of statements that express them. For instance, conjunctions usually have fewer constituents than their conjuncts, and all necessarily true propositions have exactly the same constituents (and are identical). So while possible-worlds prop- ositions may do semantic duty as truth conditions of statements, they are not useful as semantic 'structures' reflecting the way in which the truth conditions of a statement depend on its parts.
While possible-world semantics seems not to make the mistake of distinguishing between identical prop- ositions, it does seem to make the opposite mistake, of conflating distinct propositions. According to possible-worlds' theory, all necessarily equivalent propositions are identical, since they are true in all the same worlds. Thus, the conjunctive proposi- tion that snow is white and two equals two, is on this view identical to the proposition that snow is white. In addition, all true mathematical claims express the same proposition (the necessary one con- taining all possible worlds).
6. Prepositional Attitudes
Part of the intuitive motivation for using propositions as central theoretical tools comes from their serving not only as the things one claims, but also as the things one believes, hopes, doubts, and so on. The difficulty in explaining semantically linguistic practices of talk- ing about these prepositional attitudes has led to an embarrassment for theorists who have long since adopted propositions for theoretical use.
The naive semantic view of belief reports, for exam- ple, is just this: a use of 'A believes that />,' expresses the proposition that A bears the relation of belief to the proposition expressed by p. This view has been held in very different forms by Frege, Russell, Carnap, Barwise and Perry, Salmon, and Soames. Yet, the view in all its forms has never been able satisfactorily to handle the simple, though stubborn, puzzles about
substitution, including Frege's example of the agent who believes that Hesperus is seen in the morning, yet does not believe that Phosphorus is seen in the morn- ing (where Hesperus and Phosphorus are really both the planet Venus). The problem seems to result from a tension between on the one hand the theoretically useful public nature of propositions, and on the other the fact that belief reports seem to make very fine, idiosyncratic distinctions among beliefs.
Frege and his followers opt to solve the problem by holding that just which proposition is expressed by a sentence is a very idiosyncratic, speaker-relative matter, which cannot be determined simply by the individuals and properties referred to by the words and phrases in the sentence uttered. Thus it is no surprise that someone can believe that Hesperus is seen in the morning, but not that Phosphorus is seen in the morning (because the sentences 'Hesperus is seen in the morning' and 'Phosphorus is seen in the morning' can express very different propositions). One trouble with this approach is in connecting the prop- osition that the speaker of such a belief report expresses by uttering, say, 'Hesperus is seen in the morning,' with a proposition that the agent might really believe. If propositions are idiosyncratic for given speakers and believers, then it is difficult to ensure that what an agent believes is ever what we say she believes.
Another way out is to explain away the intuition that an agent really can believe something about Hesp- erus without believing the same thing about Phos- phorus. The prevalent view of this sort (see Barwise and Perry 1983; Salmon 1986; Soames 1987) involves holding that it is strictly speaking true, but very mis- leading, to say that the agent believes that Phosphorus is seen in the morning. It is wrong, though not necess- arily false, to say things that are misleading, so one would not be comfortable in saying that the agent believes that Phosphorus is seen in the morning— though in fact one is wrong in one's naive intuitions about the truth of that claim, according to these theorists.
Responses to these problems have emphasized the context sensitivity of prepositional attitude state- ments, and have abandoned the naive view that such a statement simply claims that an agent is related to a proposition (see Asher 1986; Crimmins and Perry
1989; Richard 1990). This allows these semanticists to retain the picture of propositions as public entities that people state, believe, hope, and so on, while still allowing for the idiosyncrasy and speaker relativity of attitude reports. The idiosyncrasy is traced not to idiosyncratic prepositional contents of beliefs, hopes, and statements, but rather to idiosyncratic con- texrually determined conditions that reporters claim to hold of agents' beliefs, hopes, and statements, beyond merely having a certain content. Thus, to claim that the agent believes that Hesperus is seen in
290