Many different analyses of the concept of de re belief have been proposed in recent years. Most of these analyses may be called ‘reductionist’ since they attempt to “reduce” de re belief to de dicta belief or to analyze de re belief in terms of de dicta belief. Some reductionist analyses are extremely liberal in their attribution of de re beliefs — they imply that people have de re beliefs in a variety of situations in which more restrictive analyses have no such implication. In this paper I will show that the most liberal of the reductionist theories, those Roderick Chisholm calls “latitudinarian theories”, are unacceptable.

Latitudinarian analyses have been proposed by many philosophers, including Ernest Sosa, Mark Pastin, and also, perhaps, W. V. Quine and Wilfrid Sellars.

In his article, “Stroll on Russell's ‘Proof’” (this Journal VI, (1976), Robert Fahrnkopf takes issue with four comments I made about Russell's theory of descriptions in a paper, “Russell's ‘Proof,’” which appeared in the June 1975 issue of this Journal. Though I disagree with Fahrnkopf on the points in question, there would be no point in washing our private conceptual linen in a public place were it not for the ingenious and highly original suggestions he makes in his defense of Russell. I think some additional discussion may be useful in carrying the current debate about the theory a step further. In what follows, I will therefore deal with each of the four points he speaks to.

If on a general knowledge test, you were given the question ‘“Sherlock Holmes lived on Baker Street,’ true or false?” you would quite correctly answer “true.” On the other hand, if approached near Baker Street Underground Station by a naive tourist with the question “Is it true that Sherlock Holmes lived on Baker Street?'’ you might quite correctly answer “No, it is not true. Sherlock Holmes was a fictional character.” Cases of this sort seem to some people to create a problem, while to others they seem merely to pose a puzzle, the solution of which is obvious enough in principle though perhaps somewhat tricky in detail. I belong to the second group, and in this paper I undertake to give the main lines of a solution to the puzzle along with some, though not all, of the trickier details.

Let me begin with some statements which I will try to defend in due course: there is a human activity which we may call “tale-telling” which consists of putting forward certain sentences in such a way that they are neither asserted nor denied.

Bill Anglin has an ingenious argument in support of the classical utilitarian view that there is an obligation to create extra people if the people thus created will, on balance, be happy, and creating them will not reduce the happiness of others by a comparable amount. Ingenious as it is, I believe the argument is fallacious.

Anglin's argument rests on a case in which a woman has a choice between having a child whose expected level of happiness is zero or undergoing a minor operation, involving only a tiny amount of unhappiness for herself, and then having a child certain to be very happy. Anglin assumes that the woman ought to have the operation, and I do not think we should quarrel with this assumption.

The remainder of the argument is swiftly stated. First Anglin claims that for a utilitarian it must be morally indifferent whether one brings into existence an extra person who would experience neither happiness nor unhappiness at any time in his life, or does not bring anyone into existence.

Suppose we live in a very advanced age of neurosurgical skills. We have a patient, Jones, with intractable pain. We have a second patient, Smith, with a long history of unintentional self-injury due to congenital absence of pain centres. So we decide to do double service by transplanting Jones's pain centres into Smith's brain.

The distinctive feature of Professor Purtill's interesting, though somewhat promissory, paper, is its willingness to have the tail of pragmatics wag the dog of semantics. I myself find the pre-emption unfortunate, though I should hasten to add that Professor Purtill and I share something of a common view about the problems that should be solved by a decent account of fictionality; and some of our own solutions happen in fact to coincide. We part company, however, in respect of the following two theses.

Thesis (1). Professor Purtill holds that the sentences literally constitutive of a piece of fiction are neither true nor false, that they do not make assertions, that (therefore) they do not make assertions about what they would appear to be about. Such sentences in fact tell tales, and in such nonassertive uses, they are spared the burdens of all but the limits of semantic significance. That is, they are not true and they are not false.

Following the formulation of the Identity Theory by Place and Smart a few years ago, there was considerable discussion of subjective mental states and of the importance of first-person reports of mental events. One concern was that the logical status of first-person claims might commit us to belief in irreducible mental phenomena referred to by such propositions. If so the Identity Theory would be false. Kurt Baier went so far as to argue that the incorrigibility of propositions such as “I have a pain” demonstrates that Smart's identification of sensations with brain states and processes is untenable. While there is no point to reintroducing historical discussions it is worth remarking because a more general problem is reflected in this kind of debate. This is the problem of how we expect our ordinary intuitions and ways of talking about something to affect our theoretical view of it or vice versa. In this particular instance it manifests itself as the question of the extent to which a theory of persons is to be dictated to or decided by the common and generally accepted intuitions embedded in our psychological language.

Suppose we are able to transplant Jones's pain centres into Smith's brain. Half way through the operation we test the pain centres by stimulating them electrically in vitro. Would there be pain? Roland Puccetti argues that there would not be. Because (a) pains must have owners and (b) the only available candidate for that role — the excised tissue — is logically unfit to play it. He concludes that the firing of such centres in a normally functioning brain cannot be pain either and that, therefore, materialism is false.

So far as I can see, Puccetti's “refutation” is a series of nonsequiturs from start to finish, and that is what I mainly want to show. It will not follow, of course, that materialism is true. Indeed, I think the view which Puccetti attacks under that name is too naive to be taken seriously by anyone.

Glenn Pearce, labels the identification of the firing of one's pain centres in the brain with feeling pain ‘a naive view,’ the refutation of which cannot much threaten any serious version of materialism. But in fact at least a dozen proponents of contemporary materialism have already hypothesized the identification of feeling pain with activation of a specific neural mechanism, although they picked the wrong mechanism, namely C-fibres. Just to take a recent example, James Cornman and Keith Lehrer, trying to get around the objection to the identity theory that the language of neural processes and the language of feeling pains belong to different linguistic categories, suggest that the materialist could devise a new term to refer to the same entity across categories: ‘fibain.’

They write,

Some philosophers hold that there are nonmoral reasons that can be used to justify being moral and that these are “in a certain way” more fundamental than moral reasons. Presumably these reasons could also be given in some circumstances for not being moral, although this is not clear. Moral reasons, in this view, might be overriding “on the level of everyday life,” but not “at the most fundamental level.” I take this to mean that should there be a conflict between the moral “reasons of everyday life” and those that are “most fundamental” the former must give way, and I take it to mean that the “most fundamental” reasons support the everyday ones and that the latter would collapse without their support. I wish to argue, against this view, that the question “Why should I be moral?” is confused in that it requests a justification where none can be given without radically altering the logic of the discourse that is “supported.”

Gareth Evans's two articles on the syntax and semantics of pronouns unfortunately seem to have been written far more with the desire of showing Peter Geach to have been wrong all along the line than in an endeavour to formulate, clearly and coherently, a comprehensive theory of his own. Some scattered passages of constructive theorizing there are, and in the latter part of this note I shall have something to say about them; but the controversial motive prevails.

It would be wearisome for me, and unrewarding for readers, if I were to work through Evans's arguments against me in detail: his misrepresentations of me are frequent and gross, and exposure of them in detail, together with restatement of the arguments misrepresented, would take up many more pages than the articles themselves. I here document just two of the more flagrant examples; with this warning before their eyes, readers may be disposed not to believe Evans's account of my works without checking for themselves.

Events have played a central role in a number of recent philosophical analyses. In general, there are two different sorts of arguments that might be offered in favour of an event analysis: first, it might be held that the constructions being analyzed contain certain nominals (expressions like “the sinking of the Lusitania,” or “Nixon's resignation“) which intuitively refer to events, and further that any satisfactory analysis must respect these intuitions; second, it might be argued that quite aside from our intuitions, the concept of an event — perhaps as a purely technical device — permits the solution of philosophical problems implicit in alternative non-event analyses (which typically take nominals as non-referring sentence transformations).

I wish to argue that both the above sorts of arguments fail. First, I argue that our intuitions about the reference on nominals are inconsistent, varying from context to context, so that the various sorts of event analyses cannot be talking about a single category of entities.