A Puzzle about Desire (and Intention etc.)

A Puzzle about Desire (and Intention etc.)

In “A Puzzle about Belief” Kripke introduces his puzzle about belief as a puzzle about belief—specifically, the behavior of names in belief contexts. I will contend that it is a not a puzzle about belief specifically and not about names specifically; that is just one version of the underlying puzzle. Kripke’s protagonist Pierre first learns about London (“Londres”) from a book, later visiting that city and learning English. He forms contradictory beliefs about the attractiveness of London from these two sources. But he could have formed these contradictory beliefs from the same kind of source: he could have read two books, one in English and one in French; or he could have paid two visits to different parts of the city. The puzzle has nothing essentially to do with testimony-based belief and observation-based belief (not that Kripke says it does). For simplicity, let’s assume the beliefs are formed from reading two books each ascribing different properties to London. Suppose Pierre forms the desire to visit London (“Londres”) by thus reading about it, but that he also forms the desire not to visit London by reading another book in English (the first book talks mainly about Kensington, the second about Hackney). He desires to visit London and he desires not to visit London. The same kind of disquotation principle applies to desire as to belief (“I hate London”, “J’aime Londres”). Accordingly, Pierre has contradictory desires. The same goes for intention, obviously enough: Pierre may intend to visit London and intend not to visit London, depending on the information he acquires. The names “London” and “Londres” feature in his vocabulary and they can generate the same result as they do for belief. It isn’t the concept of belief that gives rise to the puzzle; it’s the way names interact with propositional attitudes in general (but see below).

Does the puzzle arise only in the case of names? Apparently not: we can generate the same kind of puzzle using demonstratives or pronouns. Pierre may express himself by saying “That city is attractive” (in French) and “That city is not attractive” (in English), referring to London both times, while not knowing this. Indeed, we can get the same result even if he is monolingual. We can copy his linguistic preference by reporting him as believing that that city is attractive and also that that city is not attractive, unknowingly pointing at the same city twice (same for “it”). So, names are not essential to the puzzle either. Nor need reference to a particular object be part of the story: Pierre could have contradictory beliefs about a natural kind (e.g., water) or even about a physical magnitude (e.g., a mile). All he needs is two words in different languages (or the same language) associated with different bodies of information. So, it is not strictly accurate to say, as Kripke does, that the puzzle concerns “the behavior of names in belief contexts”: it is more general than that both with respect to belief and names—better to say, propositional attitudes and reference more generally. Nothing specific to belief or names is raised by the underlying puzzle.

Can the puzzle be generalized even further? Is language even necessary? I think not: Pierre could wander around a district of London one day and think that the city of which it is a part (“this city”) is attractive, while the next day forming the opposite belief while wandering around a different part—not realizing he is in the same city. Similarly for desire—he desires to stay in the first city but not the second, as he would put it. He need not express his beliefs in a public language, simply forming them without speaking. He need not even be able to speak, having never learned a language. A speechless animal could likewise form contradictory beliefs, as long as they are formed from different bodies of information. It isn’t language as such that is generating the puzzle; rather, it is propositional attitudes considered in themselves—desires, intentions, hopes, regrets, etc. We might even say it is a puzzle about concepts. No disquotation principles are needed to get it going, let alone proper names.

What about perception—can it generate the puzzle? I don’t see why not, though we might need to exercise more ingenuity to find a convincing example. Take someone looking at a tomato and believing it is red. Unknown to him, there is a mirror in his visual field reflecting that tomato, but cleverly disguised to give an impression of greenness. He accordingly believes that tomato not to be red—even though it really is. If he gives the tomato two names, under the impression that he is seeing two objects, he will commit himself to a pair of perceptual beliefs that are contradictory without realizing it. He has contradictory visual impressions: it seems to him that what is in fact a single tomato is red and not red—as we would put it, but not he. Or we could have an example in which an object seems square visually but seems oval tactually: the subject perceives it as square and at the same time as oval—his perceptions contradict each other (though he fails to see that). Or suppose an animal espies a potential predator and has the impression of a scary animal over there, but also sees its reflection in water and has the impression of a harmless animal (it seems to be on the point of drowning). The same animal seems to be both dangerous and not dangerous, and this seeming may not be a case of belief proper. So, concepts in the full sense are not even required to construct a case like Kripke’s, if we exclude perception from the conceptual domain. The puzzle really concerns intentionality in general—any kind of mental representation. It isn’t about beliefs in particular, and certainly not about names in particular. It’s about the representational mind, and clearly derives from the possibility of two perspectives on the same thing—two appearances of the same reality (in conjunction with other auxiliary factors). Kripke’s paper might well have been called “A Puzzle about the Mind”.[1]

[1] I don’t know if Kripke would reject the position here put forward, because he never denies that the puzzle generalizes in these ways. But he doesn’t explicitly accept it either; the possibility is simply left open. However, there is a strong impression that he takes the puzzle to be more restricted. I would be amazed if he had thought of these extensions but simply decided to leave them out. Nothing in the argument would be lost by generalizing it, as far as I can see.