ROD

REFERENCE,

BERTOLET

FICTION,

AND FICTIONS

Among those who think there are truths to be told 'about fictional objects' two views prevail. One is the so-called Meinongian view that, e.g., 'Pegasus has wings' is true for the reason that other sentences of subject-predicate form are true: because there is a horse, Pegasus, which is winged. T h a t this horse does not exist is beside the point. T h e other view is that the sentence 'Pegasus has wings' is really about a myth, and that whatever truth there is to be told is a truth about that myth. T h e second view allows us to avoid nonexistent objects, but is otherwise not particularly plausible; the sentence 'Reagan has a ranch' makes no covert appeal to California legend, and it is rather unclear why 'Pegasus has wings' should be any different. Both views, I think, should be rejected, but I wish to defend something like the second view in this paper. I shall begin by setting out the problems posed by empty names and other vacuous noun phrases. T h e proliferation of recent rehearsals may make one more discussion seem tiresome, but rehearsals are sometimes flawed, to the detriment of the final production. My reason for what may strike some as plodding over familiar ground is that I think the problem is often misstated, and not without prejudice to the range of plausible solutions. After setting out how I think the problem is - and is not properly posed, I shall present, clarify, and defend my own suggestion. Before getting to work, however, I may as well place my metaphysical cards on the table. Meinongian views, as Parsons has shown, need be neither confused nor contradictory, even if they are not wholly freefrom problems. 1 My objection to them is the usual one - that there are no objects which do not exist (or, as T e r r y Penner likes to put it, what doesn't exist isn't anything at all). At least, we would be well advised to search carefully for a view which does without Meinong's objects, embracing them only if all else fails. This will not much impress Meinongians, but it will give the rest of us reason to press on, and at the very least we need a more fully articulated version of the competing views in order to intelligently decide which deserves our support. Synthese 60 (1984) 413-437. 0039-7857/84/0603-0413 $02.50 © 1984 by D. Reidel Publishing Company

414

ROD BERTOLET

Parsons has made the case for the Meinongian; this paper contains the beginnings of the case for the opposition. While I have tried to keep it to a minimum, a certain amount of overlap with various presentations of the second view noted above has proven unavoidable. What I hope is distinctive about the present discussion is that it provides a general framework for the claims made which is plausible and which provides some reason for taking the view seriously beyond the fact that it obviates appeal to fictional objects. .

As a moderately competent speaker of English who knows a bit of mythology, I claim that speakers generally say something true when they utter (1)-(3), and that they generally say something false when they utter (4)-(6): (1) (2) (3) (4) (5) (6)

Pegasus is a winged horse. Pegasus was captured by Bellerophon. Santa Claus has a white beard. Pegasus is a winged cow. Pegasus lives at the bottom of the sea. Santa Claus brings presents every Fourth of July.

When I say that generally something true is said, I mean that something true is said in the conversational contexts in which such sentences are ordinarily uttered, recognizing that there are other contexts, such as telling stories, in which such sentences might be uttered without anything true being said. The more usual cases will be considered here. So if, for example, someone who can't find the relevant material in Bulfinch drops in and asks who captured Pegasus, I give the correct answer if I utter (2) - moreover, what I say by uttering (2) is true. It seems extremely difficult to deny this, although it certainly calls for explanation. But before turning to any theory for help, let us focus on some rather different examples. (1)-(6) all, as the phrase goes, involve fictional entities. For some, this constitutes a reason to shelve them with the provisional assumption that they work differently, somehow, from ordinary discourse, and need not delay us in our attempts to understand the more mundane uses of language. Perhaps there are some uses of such sentences that do work differently from ordinary discourse (again, storytelling), but this is not

REFERENCE,

FICTION,

AND FICTIONS

415

true of the sort of case I have mentioned. This, at any rate, is what I shall argue later. In the meantime, notice that there are sentences much like (1)-(6) in that (a) some of their constituents seem to be nondenoting singular terms, (b) we can say something true by uttering them, but (c) which are not comfortably describable as 'fictional' or 'mythological'. Suppose that on the basis of current seismic conditions an expert predicts that an earthquake will strike in three hours, and that the expert broadcasts this to the public. T h e r e is no earthquake at the time of the prediction, it is just that conditions make the prediction quite reasonable (at least, there is no earthquake at the place where it is predicted there will be one). 2 Now it seems quite clear to me that we as residents of the town who are aware of the prediction can say something true by uttering any of (7)-(9): (7)

T h e e a r t h q u a k e /~will i s supposed toJ/ be here in three hours.

(8)

It is feared that the earthquake will cause millions of dollars of damage and numerous deaths. If we leave now we can be in Nevada when the earthquake comes.

(9)

And we can say something true by uttering any of these even though ex hypothesis there is, as of now, no p h e n o m e n o n of which we can say, 'the earthquake' refers to that. Further, if the day passes without incident, (10)

I'm sure glad the earthquake never materialized

is not out of line. Further cases involving predicted or anticipated events could be detailed, but this one will serve to make the point. While someone might suggest that Santa Claus and Pegasus 'exist in fiction', I see little plausibility to the" claim that an earthquake 'exists in prediction'. W e can begin to see what sense it might make to say Santa is a fictional entity (that there's a fictional work 'about Santa'), but it is difficult to follow out the parallel with a predicted entity. If someone successfully predicts an earthquake, then a genuine earthquake occurs. And if the prediction fails, then no earthquake occurs at the time and place it was predicted: the prediction turns out to be wrong, empty. So it seems to me that the postulation of predicted, anticipated, etc., entities as another species of intentional entity is not intuitively satisfactory, and

416

ROD

BERTOLET

that we must look elsewhere for the explanation of how we can say something true by uttering (7)-(9). 3 I thin'x the explanation does lie elsewhere, and that it is the same as the one which accounts for how we can say something true by uttering (1)-(3) (and something false by uttering (4)-(6)). But it is not time for that story yet. In its most general form, the problem is to explain how we can say something true by uttering sentences of the form (i)

A is 04

when there is nothing to which A 5 applies, or sentences of the form (ii)

T h e 0 is lfl

when nothing satisfies the description 'the 0'. For if A names nothing, or nothing satisfies 'the 0', what could there be to say something true about? T h e r e is, finally, the question of how we can say something true by uttering sentences of the form (iii)

a does not exist 6

since the sentences which (iii) schematizes appear to predicate of whatever it is that a denotes, that it does not exist. If there is nothing to which a refers, nothing it denotes, then there would seem to be little to be truly asserted of it, including its failure to exist. Thus it may seem that something true can be said by uttering sentences of the form (iii) only if a does denote something - but if a does denote something it is hard to see what truth there is to be asserted. So the attempt to assert something true by uttering such a sentence has often seemed selfdefeating, as early as the Sophist (237b-239b). . Now notice what I have not claimed. It has not been asserted that (1)-(3) or (7)-(9) are themselves true, or that (4)-(6) are themselves false. T h e claim is rather that speakers who utter these sentences (in certain circumstances) say something true or false. T h e claim I have foresworn is, moreover, one I think we should not make, at least not on the basis of the intuitions that are always brought to bear here - and it is here that I wish to part company with most of those who have written about empty names. It is natural enough to m o v e from the claim that a speaker says

REFERENCE,

FICTION,

AND FICTIONS

417

something true on an occasion, to the claim that the sentence he utters on that occasion is true. But not all that is natural is good and right, and this move seems to me both avoidable and preemptive. This can be seen by attending to yet another sort of example. Explaining the psychological theses that underlie Hume's account of belief to my class, I utter (11)

A belief is a particularly vivid or lively impression.

I have, it seems, said something true by uttering (11). But this is not much of a reason to say that (11), qua type or token, in this or any other context, is itself true. Indeed, as I go on to disparage Hume's views a bit later in the hour, I might point to the very inscription of (11) I wrote on the board earlier and say this is false, continuing with the usual objections. I have not, initially, said that a belief is a vivid or lively impression, and I do not later lapse into self-contradiction when I attempt to motivate the rejection of the Humean doctrine on beliefs. Otherwise put, I never asserted (11), I never said that what it says is true. But if we were to say that (11) is true, on the grounds that what I said is true and (11) is the sentence I asserted, we would need an explanation of this truth, and would be led down one of two paths: postulating some extraordinary objects, Humean beliefs, which are vivid or lively impressions, or making some semantic readjustments, either in our claims about the semantic structure of (this token of) (11), or in our semantics for (this token of)7 (11). Neither path is overly attractive, but I do not think this fork exhausts the possible routes. We might, and I think we ought, say instead that while the speaker says something true by uttering (11), (11) is not itself true. This tack would allow us to avoid both dubious nonactual beliefs and semantic complexities. The explanation of the speaker's saying something true will, I think, be provided by speech-act theory rather than semantics, which will be discussed later. 8 The point here is that our intuitions do not clearly or obviously support more than I have claimed, and in particular, they do not seem to support the claim that (11) is true. This is a further claim motivated by a philosophical theory rather than our intuitions about the example, where the theory needlessly blinds us to another way of approaching the problem. The same can be said for other cases having nothing to do with the special problems concerning fictional objects. This discussion casts the problem for the view I wish to defend in a

418

ROD

BERTOLET

somewhat different, and more favorable, light. What is initially not terribly promising is giving an account of how 'Pegasus is a winged horse' is itself about a myth, and thereby true. But it is perhaps less hard to see how a speaker might say something about a myth by uttering 'Pegasus is a winged horse', and thereby say something true, on the model of saying something true by uttering (11) without it being true, or said, that a belief is a vivid or lively impression. This may seem less than fascinating because it is all too familiar. That, however, is precisely why it is crucially important that we remind ourselves of it. This phenomenon is quite familiar to anyone who teaches philosophy, or for that matter explains the views of others, b u t the moral to be drawn from this fact is that there is nothing at all odd or unusual about the claim that someone says something rather different from what is properly said by using the sentence he utters (from what that sentence means). 9 What I shall attempt to develop, in the next three sections, is an account of how speakers say something true by uttering (1)-(3), and something false by uttering (4)-(6), which does not force us to traffic in intentional objects, and which introduces no semantic complexities into our views about such sentences. It will also be suggested that this same apparatus allows us to account for the possibility of saying something true by uttering (7)-(10), and the possibility of saying something true by uttering (11) while denying the Humean account of belief. .

My attempt to support my claims will utilize the notion of a speaker's presupposition, which I shall now characterize. If a speaker presupposes (that) p, then that speaker (S) takes it to be understood that p. (Or, if you like, S takes p for granted.) For S to presuppose that p will be for S to believe p, and also to believe that his audience believes that p. Thus what S presupposes will be distinct both from what S says, and from what S implicates in Grice's sense) ° To illustrate the work I intend this notion to do, let us return to the classroom exposition of Hume. When I utter (11)

A belief is a particularly vivid or lively impression

I presuppose that Hume's philosophical and psychological views are under discussion. (Perhaps I presuppose this because I said it five

REFERENCE,

FICTION, AND FICTIONS

419

minutes earlier, but perhaps not; I may simply presume that it is clear from the context, a reasonable enough presumption in a philosophy class.) This presupposition is what enables me to expect it to be clear that I am talking about Hume, or Hume's philosophical system, and making an assertion about him or it. What I am actually doing is referring to Hume (or his philosophical system) and saying that according to him (or it) a belief is a vivid or lively impression. I shall analogously be claiming that when, in ordinary conversational circumstances, I utter (1)

Pegasus is a winged horse

I refer to a myth, asserting that it says that Pegasus is a winged horse. What I say is that the myth says that Pegasus is a winged horse; this is facilitated by my presupposition that mythology is under discussion. What I want to point out is that the philosopher who objects to this may be on no firmer ground than the student who criticizes me for contradicting myself in both asserting and denying that beliefs are vivid or lively impressions. Did I not, he might ask, say or assert that (11) is the case, the proof being that I wrote (11) on the board? No. I wrote it on the board, but I never asserted it, I never said what it says: he has simply misconstrued the nature of my speech act, and his objection is simply irrelevant.

.

I shall discuss negative existentials, sentences of the form (iii), first. For ease of exposition, I shall begin with a simplistic version and introduce the needed complications along the way. Briefly, my claim is that when S utters (12), (13) or (14), (12) (13) (14)

Pegasus does not exist Dragons do not exist Santa Claus does not exist

S presupposes that

(15)

[Pegasus ] There is a ~myth/in which ]dragons ~ figure(s) t story J /Santa Claus)

420

ROD BERTOLET

and S says or asserts that (16)

[Pegasus ] There is no individual which the ~dragon ~ myth is lSanta ClausJ about.

In asserting (16), moreover, S is simply saying that there is no one or thing of whom or which we can say the story is a story - unlike, say, certain World War II propaganda films concerning Hitler which, no matter how inaccurate, were about Hitler. Such a film might portray Hitler as a kindly soul trying only to bring peace and brotherhood to the world; no matter how far off the mark the portrayal is, it is still a portrayal of Hitler and the film is about Hitler. The Pegasus myth, so far as I know, is not about anything, even in this way (better, it is not about any horse). S is saying, in other words, that Pegasus et. al. are not really part of the world; differently still, that they are not real. H Besides providing a framework which makes my claims about saying something true by uttering sentences of the form (iii) predictable and explicable, I think my suggestion is rather plausible on intuitive grounds. If I utter (13), what I mean to convey is I think pretty clearly that certain things which were posited in mythology are in fact not part of the world's furniture, which is just to say that the myth is a myth and not a story, true or distorting, about anything. And in saying so I do not say that there are certain creatures, called dragons, which do not in fact (do not really) exist; rather, I am presupposing that someone else has claimed or suggested or represented it as being the case that there were such things, and saying that this suggestion, etc., is not true. a2 So I think that my view accords nicely with what actually goes on when we utter sentences such as those under discussion. I have not yet said, however, what is packed into (15) and (16), and should explain how these are to be understood. I begin with a somewhat naive interpretation of

(15)

[Pegasus ] a / m y t h / i n which ]dragons ~ figure(s). There is tstoryJ /Santa Claus)

This might be thought to undermine the whole approach, since it may seem (15) will be true only if there are dragons, etc., so I shall say why it

REFERENCE,

FICTION,

AND FICTIONS

421

does not do so. (15) is an abbreviation for something like this: there is a myth or story, some of whose constituent sentences are such that the word 'dragons' ('Pegasus', etc.) occurs in them. And that is (almost) all - if there is a myth containing the sentence 'St. George slew seventyeight dragons before breakfast on Tuesday', then dragons figure in the myth, as does St. George. All this means, however, is that the word 'dragons' occurs therein. Well, not quite; as I said, this version is naive. It is not true that the occurrence of 'Pegasus' in a myth or story always suffices for Pegasus to figure in that myth, or in the looser formulation, for Pegasus to be a figure in that myth. Certain kinds of occurrences of 'Pegasus' in a myth or story would not be sufficient for Pegasus to figure in that myth or story. For example, the occurrence of 'Pegasus' in 'Horace dreamt that a horse named Pegasus flew into his room and ate his granola' would not, by itself at least, make Pegasus figure in the story in which that sentence occurs. Nor would dragons figure in a story if the only (sort of) occurrence of 'dragons' in that story were 'All was of course mistaken in thinking Rover was a dragon, for there were no dragons in Alstacia'. (Less fancifully, Ivan and Alyosha figure in The Brothers Karamazov but the Grand Inquisitor does not.) That a certain noun phrase ('NP') occurs in these ways does not suffice to make NP figure in the story. Why not? Notice that in the examples, 'Pegasus' and 'dragons' are governed by the verbs 'dreamed' and 'thinking', verbs which express psychological attitudes. This suggests, first, that we may be able to state the requirement regarding the way in which 'Pegasus' and company must occur in purely linguistic terms, and, secondly, that we can use our ordinary discourse as the model for the criterion. For it does not follow from the sentence 'John believes that there are dragons', that there are dragons. Nor does it follow, just from the sentence 'Jane imagined that a nuclear reactor exploded in California', that there are nuclear reactors. And, pacd the Eleatic Stranger and others, 'There are no dragons' does not entail that there are dragons. Other relevant verbs are verbs of saying; if Stephen claimed or said that the aliens have landed, there might or might not be aliens who landed. Neither the assertion of such sentences, nor their truth, requires the existence of the relevant entities, nor any commitment to their existence. This latter, the matter of existential commitment, is where I propose to place the weight of the criterion for the 'right way' for 'NP" to occur in a story in order that NP figure in that story. We might adopt

422

ROD BERTOLET

a certain variant of the dictum that we are committed to exactly those objects which are the values of the variables of quantification, less those which are governed by some psychological attitude verbs and certain others. 13 It must be a variant of this dictum, because truth and existential commitment are not obviously relevant concepts to apply to works of fiction. Authors of fiction are not concerned with speaking the truth, in any direct and ordinary way, and are not to be chastised for not believing that the things they appear to write about (dragons, for instance) exist. 14 T h e r e is nonetheless a connection, which arises from the sort of relation Wittgenstein observed between meaning and truth namely, that to know what a sentence means is to know under what conditions it would be t r u e . 15 An essayist sits down and composes a story of some sort by writing out various sentences. What is he doing? One thing he is not doing, in any direct way, is offering us truths about the world. So how exactly is truth relevant here? What the essayist produces are quite unavoidably sentences, and these sentences have a certain meaning. Because they mean what they do, they represent the world as being a certain way, even though the author of them does not. So there is a connection, albeit an oblique one, between fiction and truth. Because we know what the sentences mean, we know what w o u ld be the case if they were true, how the world would have to be in order for them to be true. Bound up in this knowledge is knowing what things would have to exist for the sentences to be true. Those are precisely the things which figure in the story. Let us look, briefly, at an example. Stately, plump Buck Mulligan came from the stairhead, bearing a bowl of lather on which a mirror and a razor lay crossed. A yellow dressing-gown, ungirdled, was sustained gently behind him by the mild morning air. 16

What would have to be the case in order for this to be true? Well, there would have to be a person named Buck Mulligan, a stairhead, a bowl of lather, a mirror . . . . ; Mulligan would have to be plump, bearing the b o w l , . . . Had the book continued, as we may be grateful it does not, Remembering that he'd dreamt that an elf seized the razor and slit his throat, Mulligan mounted the gunrest and cast the instrument to the ground

it would not need to be the case that an elf had slit Mulligan's throat, or done anything else. I would say that Mulligan figures in the story, while

REFERENCE,

FICTION,

AND FICTIONS

423

the elf does not figure in it. This is captured by adopting an amended existential commitment criterion, to the effect that N P figures in the story just in case N P would exist if the story were true. So although novelists and other purveyors of make-believe do not present their works as truths, the existential commitment criterion will, I think, serve to specify what figures in a given story. T h e observation that the sentences which constitute the story represent a certain state of affairs will bear further fruit. We can, by appeal to it, explain why it is that we can say what a story says in more ways than by simply parroting the sentences occurring therein. Since we know what these sentences mean, we know what situation would obtain if they were true. Language is rich enough, however, to represent that situation in more than one Way, with sentences other than those actually occurring in the myth. Thus, whatever sentences fill this same representational role may be uttered by us to say what the myth says 'Santa Claus's cheeks and chin are covered with white hair' will do to inelegantly convey that the myth says that Santa Claus has a white beard. 17 Recognizing the representational role of the sentences of a story also enables us to explain why certain inferences that usually hold seem inapplicable to so-called fictional characters. As is often observed, it does not seem to be the case that we can say something true by uttering 'Santa Claus is a liberal Republican', or by uttering 'Santa Claus is not a liberal Republican'. This suggests that the law of excluded middle fails here. If we claim, as I do, that all we can do with respect to Santa is report what the myth says, we can explain why the law is inapplicable: the story says neither. And in general it is n o t true that persons or books must a s s e r t either a sentence or its negation. I may have to be tired or not tired, but I need say neither - I can refrain from comment, as the Santa Claus myth does on political persuasion. In neither case is there anything to relate on the matter. So far, I have claimed that for, say, Pegasus to figure in a myth or story is simply for 'Pegasus' to occur in that myth or story in a certain way, and I have lately tried to specify what that way is. It is time to turn to what I claim is asserted by an utterance of negative existentials, viz., (16)

[Pegasus ] T h e r e is no individual which the ] d r a g o n ~ myth is about. [.Santa ClausJ

424

ROD

BERTOLET

(16) is perhaps a bit more straightforward: it requires that there be nothing of which the myth says anything true or false. T h e requirement cannot simply be that there is nothing of which the myth is true, for it might be that there is a certain story, such as the one in the propaganda film, which is totally false and yet about someone. That a story is false does not suffice to show that what it purports to be about does not exist; were this so, the U.S. could have simply produced those films and let the troops stay home. It must be the case that there is nothing which the story is a story of. Some cases will be more complicated than others, as e.g., with A g a m e m n o n , or Jesus of Nazareth. Probably, Jesus did have a mother named Mary, who was probably not a virgin when she gave birth, and he probably did address a multitude on a certain mountain, but he probably did not feed all of them with five fish and five loaves of bread. Some of what is said of Jesus is true, and some just false. Similar remarks apply to A g a m e m n o n , though it may be harder still to ascertain which are which. Thus, in each case there is someone whom the story is about - but this is the result we want, since it is (probably) false that A g a m e m n o n or Jesus did not exist, and we can not say something true by uttering 'Jesus didn't exist'. Here, we must make the more subtle point that some of the tales about Jesus and A g a m e m n o n are true and others are false, and to some extent we are unsure which are which. But they are about Jesus and Agamemnon. With many myths and stories, it may be quite difficult to tell whether they are about anything or anyone. But these are difficult questions, and it is not obviously the mark of a good philosophical theory to turn them into easy questions. T h e r e are further complications, one of which is that whether a story is about anything can apparently change over time. It may well be, for example, that the Santa Claus story was once about a certain kindly priest in Asia Minor, but even if it was it is no longer about him - the story has taken over, severing whatever historical roots it may once have had. Similarly, there may well have been someone whose name we render as 'Bellerophon', but no horse, winged or otherwise; perhaps the story was once about Bellerophon but not Pegasus. But even if this is so, it is no longer about that person, or so it seems to me. Still, this suggests that we must be more specific about aboutness, if only because this does seem to vary across time. This suspicion is reinforced by two further complications concerning stories. Ulysses, from which I quoted earlier, seems to be in some clear sense about

REFERENCE,

FICTION,

AND FICTIONS

425

Dublin. Second, there are fictional works which are even more clearly about people who exist or once existed, The Court Martial of George Armstrong Custer and Brecht's "Galileo" being cases in point. The upshot is that 'There is nothing which the myth is about' is too simpleminded. The Pegasus myth may be (or have been) about some things, even if Pegasus is not among them. Notice that it is with regard to 'Pegasus', that name, that the myth is not about anything, although it may be about something with regard to 'Bellerophon'. This suggests that we can once again introduce a linguistic parameter, along these lines: (16a')

There is no individual x such that (a) the Pegasus myth is existentially committed to x by the occurrence in it of the name 'Pegasus', and (b) the Pegasus myth is about x.

Similarly, mutatis mutandis, for dragons, Santa Claus and other denizens of fiction. This formulation allows us to capture the claim that Pegasus is a mere concoction without claiming that everything else is too. A better explanation of all this is to be hoped for, but this much is sufficiently clear and free from difficulties to allow us to continue the business at hand. 18 My revised claims, then, are these. When on the sort of occasion in which we are interested S utters (12)

Pegasus does not exist

S presupposes that (15a)

There is a myth in which Pegasus figures

where this is explicated as above, and S says that (16a')

There is no individual x such that (a) the Pegasus myth is committed to x by the occurrence in it of 'Pegasus', and (b) the Pegasus myth is about x.

This, once again, tells us how it is that I can say something true by uttering (12) - by specifying what it is that I say - without any sort of commitment to the ex/subsistence of Pegasus. 19 We may now turn to the other sorts of utterances which I claim need explaining, utterances of sentences of the form (ii) and (iii). Suppose that S utters

426 (1)

ROD

BERTOLET

Pegasus is a winged horse.

It is my contention that S presupposes that (15a)

There is a myth in which Pegasus figures

(again, explicated as above) and that (16a')

There is no individual x such that (a) the Pegasus myth is committed to x by the occurrence in it of 'Pegasus', and (b) the Pegasus myth is about x.

What S says, on my view, is (17)

According to the myth, ~ Pegasus is a winged horse. t The myth says that J

This then tells us how we can say something true by uttering (1) - it is true that the myth says that Pegasus is a winged horse, and this is precisely what I have said to be the case. And clearly, if I am right, by uttering (1) S is not committing himself to the existence or being of Pegasus, for if he takes (15a) and (16a') to be understood he can hardly be taken to believe that there really is some creature named Pegasus when he utters (1). Moreover, this is exactly what goes on when we 'talk about fictional entities'. When we utter a sentence such as 'Pegasus was captured by Bellerophon', we say and are understood by the speech community to convey only that the Greek myth says that Pegasus was captured by a guy named Bellerophon. e° (Assuming, of course, that we are correctly understood. Someone may misunderstand, and mistakenly think that we are saying something else.) Thus I think it is a mistake to distinguish, as Donnellan does, "discourse about fiction" from "discourse about actuality" with the suggestion that the two types of discourse are amenable to different treatment. U n d e r "discourse a b o u t fiction" I m e a n to include those occasions on which it is a presupposition of the discourse that fictional, mythological, legendary (and so forth) persons, places, or things are u n d e r discussion . . . . Discourse about actuality carries the presupposition that the speaker is talking about people, places or things that occur in the history of the world. 2~

The supposition is not that fictional characters are under discussion, but that fiction is under discussion. And fiction surely does exist; it crowds

REFERENCE, FICTION, AND FICTIONS

427

the shelves of public libraries everywhere. There is no philosophically interesting difference between talking about it and talking about Napoleon's defeat at Waterloo, except in the way that we frequently do so. To talk about Napoleon's defeat at Waterloo, we use phrases such as 'Napoleon's defeat at Waterloo', and the connection would appear to be straightforward. But sometimes we talk about the story of santa Claus in a less straightforward way, with the expression 'Santa Claus'. This may seem odd enough at first, but recall Hume's lively impressions. Notice that the view I espouse here meshes nicely with how we adjudicate disputes over whether what we say is true when we utter sentences like (1). If A and B disagree over A's utterance of 'Pegasus is a winged horse' - B responding with 'Pegasus is a seven-headed hydra' - they know just where to look to see who is right. They check mythology books to see whether (17), 'The myth says that Pegasus is a winged horse', is true. They find that (17) is true, and thus that A was right: because it is (17) which represents what A said. There may be further books (or lost Greek manuscripts) to check, but there is no question of them all somehow getting the story about Pegasus wrong. They are the story about Pegasus, and are not to be overruled. 22 So again, I think my view comports well with what we know to be facts regarding such utterances.

.

Furthermore, the account gives us what I think is a natural explication of (7)-(9), the sentences concerning the predicted earthquake. If the speaker S utters (7)

/will The earthquake tis supposed to } be here in three hours

S presupposes that (18)

There is a prediction of an earthquake

and (19)

There is (as of now) no earthquake here 23

and what S says is given by

428 (20)

ROD

BERTOLET

According to the prediction,/ the earthquake will be here The prediction says that J in three hours.

The explanation for this will exactly parallel that given for the relevant utterances of (1). What is important is that this gives us a way to understand apparent talk of predicted, anticipated, feared or expected future events or things without resorting to the claim that these already exist in some sense or other. Such talk is talk about certain predictions, anticipations, fears, or expectations, and no more. This may be thought to be a bit too quick. Is the parallel with the explanation of what I say by uttering (1) all that exact? It might be claimed that there is a significant difference between (7a) and (7b), between, that is, (Ta)

The earthquake will be here in three hours

and (Tb)

The earthquake is supposed to be here in three hours.

For it seems that (7a) entails, whereas (7b) does not, that the prediction is correct - that there will indeed be hn earthquake in three hours, and not merely that such is predicted. This seems correct; at any rate I grant it as an assumption. But then, the objection continues, suppose that no earthquake comes: won't (7a), and what the speaker says by uttering it, be false? And won't it be false because no earthquake comes? And doesn't this show that with predictions, unlike perhaps the Pegasus myth, we are dealing not just with what the prediction is, but with its accuracy as well - whether the world is, or comes to be, as it claims? Doesn't the parallel then break down, since whether what one says by uttering 'Pegasus is a winged horse' is true depends only on what the story says, whereas whether what one says by uttering 'The earthquake will be here in three hours' is true is not settled by the content of the prediction alone? Well: If what S says by uttering (7) is that an earthquake will take place, then of course whether there will really be an earthquake is relevant. But in that case, (20) does not represent what S says, at least not fully. He says something rather stronger; he endorses that prediction by saying there will be an earthquake. If in this case S is still talking about a prediction (and he might not be), he is saying not what it says, but that it is right.

REFERENCE,

FICTION,

AND FICTIONS

429

Perhaps this is the most natural thing to say by uttering (7a). Still, it is not the only thing we might say by uttering it. A speaker might also say what is represented by (20) by uttering it, and if he does - if he does just say that some prediction is that there will be an earthquake in three hours - then (a) the explanation of this is exactly parallel to the one given for a certain utterance of (1), and (b) whether what he says is true does depend only on whether the prediction plumps for an earthquake in three hours. There is nothing unusual about this sort of claim. Consider the past American elections: the sentence uttered by the television commentator, 'Ford will win Virginia', or the sentence 'Reagan wins' flashed on the screen. In the 1976 election, the CBS news staff utilized sentences such as 'Ford will win Virginia', whereas NBC's personnel adopted the more explicit 'According to our projections, Ford will win Virginia'. But CBS also periodically added disclaimers, pointing out that really, these were fallible predictions. Despite appearances to the contrary, all were saying 'Our analysts predict that Ford will win Virginia'. Similar constraints restrict what was meant (if not understood) by 'Reagan wins!" flashing on the screen in 1980. The general point here is that there are a variety of speech acts which might be performed by using one and the same sentence. Not merely that one • may assert or question or give a command with the same sentences, but that various different assertions are possible in different contexts. This is not exactly news, but it is quite important, and it will come up again later. So I agree that a speaker may predict an earthquake, or agree with such a prediction by saying that there will be an earthquake, by uttering (7a). I also agree that if this is what happens, then whether there is an earthquake certainly matters; more is relevant to the truth of what is said than the content of the prediction. Still, these facts leave untouched the point that one might also say that it is predicted that an earthquake will come in three hours by uttering 'The earthquake will be here in three hours' in the right circumstances, which is all that I have claimed. So this sort of apparent talk about predicted, etc., events really is no more than talk about certain predictions, etc. This point is quite important. An oft-repeated truism about language and truth goes something like this: Truth is a matter of whether the world is as we say it is; what a speaker says by uttering a sentence is true just in case the world is as he contends it is by uttering that sentence.

430

ROD

BERTOLET

Sometimes the utterance of sentences such as (1) and (7) seems to threaten this truism, since Pegasus and the earthquake are not part of the world. Where on earth, literally, could we look to see whether what we say when we utter them is true? If I am correct, the threat is only imagined; whether what we say in these cases is true also depends exclusively on how the world is. For what determines whether what I say by uttering (1) is true is what the myth says, and surely the myth is part of the world; what determines whether what I say by uttering (7) is true is what a certain prediction is, and that too is a fact about a certain part of the world. So far as the utterances I have discussed are concerned, my view conforms to this truism about truth. This is the view I wish to defend, and some pleasant consequences of it. There are still some loose ends to consider, however.

.

One further question concerns comparisons between so-called fictional objects, or between fictional objects and real objects. Perhaps there is a truth to be expressed by uttering (21)

Sherlock Holmes is considerably brighter than Inspector Clouseau

(22)

E m m a Bovary is craftier than Eliza Bennet.

or

These comparisons can be handled easily enough by transforming them into comparisons of what the relevant stories say; thus, the Conan Doyle stories attribute a greater degree of intelligence to Holmes than Blake Edwards's movies do to Clouseau, Flaubert attributes greater guile to Bovary than does Austen to Bennet, and so on. Where nothing of this sort can be said, it seems to me that there is no comparative truth to be expressed. In a related way, we can compare the qualities and degrees of qualities that real objects actually have with the qualities a story says someone or something has, so that (perhaps) we could say something true by uttering (23)

Bellerophon was a better horseman than Shoemaker.

Another issue requiring attention is whether the approach suggested in this paper will allow us to construe all comments 'about fictional

REFERENCE,

FICTION,

AND FICTIONS

431

objects' as being about works of fiction, and thus as having their truth conditions determined by features of the world. A possible counterexample is what is said by uttering (24)

Dragons are not much feared anymore

which one might insist is as much about dragons, and thus true only because of certain facts about nonexistent objects, as (25)

Nixon is not much feared anymore

is about Nixon, and true because of Nixon's relation to the populace at large. Somewhat different is how to deal with what someone expresses by a sentence adapted from John Woods: (26)

A certain European philosopher is in love with Molly Jeavons.

The general problem, I suppose, is that we speak as if we can stand in what have unfortunately come to be called de re relations to things which do not exist, which makes it hard to see how we can do without such things. 24 Now I do not think that the sort of view pushed in section 4 above will be of much help here, since I do not find it at all plausible to claim that speakers who utter (24) are really saying something about what a story says. But, neither do I see a need for the intrusion of fictional entities here. Fully justifying this claim would require introducing yet another creature of legend, another paper; what I hope to do here is indicate that these objectiox~s can be deflected. It is doubtless true that we speak in these ways, but this merely shows that our folk theories, which are reflected in the way we speak, allow for the possibility of such relations, and we may well want to abandon those theories. Clearly, the sentence 'Dragons are not much feared anymore' is semantically (and transformationally) related to the sentence 'Not many people fear dragons anymore', which is not obviously about dragons. (Similarly, 'Not many people fear Nixon anymore' is not clearly about Nixon; it makes a claim about the distribution of such fears.) And the truth that we try to get at by (24) can, it seems to me, be characterized nicely without any appeal to nonexistent dragons. Supposing 'People used to live in dread of dragons' to be true, we can explain its truth by appeal to dragon legends - what caused those fears was dragon legends, which people mistakenly took to be true stories. And this is not to say that what they were afraid of was dragon legends.

432

ROD

BERTOLET

In our folksy psychological stories, it is dragons they were afraid of, dragons which were the objects of their fears. But in truth, it seems to me, there is nothing at all such that they were afraid of it. To insist otherwise is to insist that whenever the content of a psychological state • is specified by some noun phrase, 25 there must be some genuine (if not real) object picked out by that noun phrase which is the object of that state. Short of going back to confusing meaning and naming, I cannot see any justification for this insistence. None of this is meant to deny the power or reality of the psychological states suffered by those 'afraid of dragons': compare veridical perception and hallucination. I may suffer an (extended) hallucination in which Nixon is named my department head which will be all too real and which will send me scurrying for Jobs for Philosophers, but the 'perception' is for all that vacuous, without any (de re, as we s a y ) object. Suppose further that we were ignorant of the good philosophical and medical reasons for distinguishing veridical perception from hallucination, using the blanket term 'perception' for all cases. Then we would quite likely balk at my claim, insisting that my perception was just as much of Nixon as was Lyndon Johnson's at the 1968 inaguration. This would be confused. So too, I suggest, is the claim that we stand in de re relations to nonexistent objects; this confusion is born of taking our ordinary claims about what our fears are fears of for far more than they are worth. Some psychological states, fully laden with content, are without a so-called de re object. It is somewhat less clear what to say about the philosopher mentioned in (26), 'A certain European philosopher is in love with Molly Jeavons', and (26) itself. (And we need at any rate be careful here; Woods notes that there cannot be a consummatable love for Molly, "and Freud's psychoanalysis of Gradiva was required to be technologically circumspect.., there could be no question of house-calls".)26 My own inclination is to say that anyone who utters (26) seriously and literally is doomed to utter a falsehood; nobody really loves Lady Molly. Woods is sympathetic to this, butadds that "it would be arch to charge him with insincerity",z7 Perhaps so, but this is not particularly relevant; he may be granted sincere belief but denied accurate belief. That is, we need only deny that he is correct in his sincere belief that he is in love with Jeavons, and this because he is not the arbiter of the best description of his psychological states, this because he is not the arbiter of the best

R E F E R E N C E , FICTION, AND FICTIONS

433

psychological theory. Similarly, I would suggest that someone who utters 'Amis admires James Bond' is either confused or struggling to say that Amis admires the sorts of qualities attributed to the protagonist of the Fleming novels. So ! think that sentences such as (26) and 'Amis admires James Bond' are strictly speaking false; I do not find this any more troublesome than claiming that Jones says something false when he seriously and literally utters 'I see snakes' while hallucinating. I would certainly agree that more ought to be said about all this. But I think that this much suggests that the anti-Meinongian need not be speechless about these problem cases. One final issue. Can we refer to Pegasus, and more generally can we refer to things that do not now exist and have not ever existed? I do not know the answer to this question. Nor do I any longer find it especially interesting, despite the prominent role it has played in various discussions, since I am extremely doubtful that the question deserves so much attention. Whether we can refer to Pegasus or not, this is not what we do when we utter 'Pegasus is a winged horse' in the sorts of circumstances which provoke the question. What we do is refer to a G r e e k myth, not Pegasus. This will grate on what some will call their intuitions; they will report a strong intuition to the effect that of course we can refer to Pegasus, and say something true of Pegasus. This, I now believe, is a mistaken description - really, an attempted explanation - of the intuition that we can say something true by uttering 'Pegasus is a winged horse'. That we can do - in the way outlined above. T h o u g h I shall not undertake to prove this here, I think that the view I have set out can be made to work without the introduction of any syntactic, semantic or pragmatic machinery beyond that needed anyway to provide a satisfactory treatment of indexical expressions. T h a t is one virtue of the view. Others include the ability to avoid the claim that there are significantly different kinds of discourse (about actuality and about fiction) involved, and the preservation of the view that whether what we say is true depends on the way the world is. Finally, what is quite right about the intuitions that led Meinong to claim that the Sein of an object is independent of its Sosein is also kept in place: the way that an object is said to be by some story or other is quite independent of whether there is any such thing. These, I should think, are considerable virtues - whether the view is correct is, as always, another question. 28

434

ROD

BERTOLET

NOTES I See Terence Parsons, Nonexistent Objects, Yale University Press, New Haven, 1980, and his earlier papers referred to there. Some of the difficulties for the view can be found on the first page of David Lewis, 'Truth in Fiction', American Philosophical Quarterly 15, 1978, 37-46. Lewis's own account is a version of the second view noted in my opening paragraph, as is the recent treatment in Chapter 6 of Michael Devitt's Designation Columbia University Press, New York, 1981. I argue against this view in 'On a Fictional Ellipsis', Erkenntnis 21, 1984, 189-194. 2 If even this. There might, after all, be an earthquake going on now, at just the place where our seismologist claims there will be one three hours hence. Still, that one is not the earthquake he predicted, and so there is now nothing existent to which 'the earthquake' or even 'the predicted earthquake' refers, or which satisfies that phrase. 3 There seems to be an additional curiosity here. If the prediction is correct, there is a perfectly real earthquake, and if it is the one which is predicted then this existent earthquake is identical with the nonexistent (and, for some, incomplete) earthquake the Meinongian may be tempted to postulate. 4 Where ' A ' is a dummy symbol replaceable by a name. 5 Strictly, the replacement of ' A ' ; I shall ignore this for ease of exposition. 6 Where a ranges over any singular term. 7 Hereafter, unless otherwise specified, any claim about a sentence should be construed as a claim about a sentence token. 8 Limitations of space preclude me from giving a full explanation of how I expect this story to go; an embryonic version of the account may be found in my 'The Semantic Significance of Donnellan's Distinction', Philosophical Studies 37, i980, 281-288, and 'Referential Shifts', Analysis 40, 1980, 135-138. I hope to spell out the details in a subsequent paper. I should add a word about my use of the term 'say'. I am using this in the sense of the word I think is appropriate to providing indirect speech reports (see 'Referential Shifts', and also 'Context and What is Said', Canadian Journal of Philosophy, Supplementary Volume VI, 1980, 97-109), Now it may turn out, as Kent Bach insists in correspondence it has already turned out, to be useful to avoid the word 'say' here in favor of a technically defined sense of a statement the speaker makes, tying what is said more strictly to the uttered sentence. (See Bach's 'Referential/Attributive', Synthese 49, 198 I, 219-244, for deployment of this suggestion.) Perhaps so,but I am not fully convinced; at any rate my employment of 'say' throughout this paper is not intended to prejudge this issue. 9 This is neither sloppy nor confined to oral presentations. Consider these remarks of P. F. Strawson's: "Briefly, any succession of perceptions is a perception of objective change only if the order of those perceptions is necessary; but the order of the perceptions can be necessary only if the change is necessary, i.e. causally determined. Any objective change which is an object of possible experience for us, i.e. an object of a possible perception, is causally determined. Hence the Law of Universal Causality is valid for all possible experience" (The Bounds of Sense, Methuen, London, 1966, p. 138). Strawson wrote all that but asserted none of it. This is the argument he labels a nonsequiter of numbing grossness, and he was of course explaining Kant's views. This really is all too familiar, and all too often overlooked. lo See H. P. Grice, 'Logic and Conversation', in P. Cole and J. Morgan (eds.) Syntax and

REFERENCE,

FICTION,

AND FICTIONS

435

Semantics, Vol. 3, Academic Press, New York, 1975, pp. 41-58. My characterization of speaker's presuppositions is akin to Robert Stalnaker's in 'Pragmatic Presuppositions', in M. Munitz and Paul Unger (eds.), Semantics and Philosophy, NYU Press, New York, 1974, pp. 197-213. It is not the only possible characterization, and it may or may not prove fruitful elsewhere, but I think it is the one we want for the purposes of this paper. While I shall not argue the point here, I think there are good reasons for denying that the presuppositions I point out are semantic. ~1 This accounts for at least part of Richard Cartwright's observation that a speaker says the same thing by uttering 'Pegasus is not real' and 'Pegasus does not exist'. See 'Negative Existentials', in C. Caton (ed~), Philosophy and Ordinary Language, Univ. of Illinois Press, Urbana, 1963, pp. 55-66. lz I am well aware that writers of fiction - though not necessarily other purveyors of myth or legend - do not claim that what they write is true, or suppose it to be true, or anything of the kind. But such writers do, in a sense to be discussed in more detail later, represent something as being the case (the sentences they write do that). Those who come across such works may suppose, correctly or not, that it has been claimed, or at least represented as being the case that, (say) dragons exist, and it is this presupposition which enters into my performance. 13 The possibility operator is another candidate. I say some psychological attitude verbs because others, the so-called factives (e.g. 'remember', 'regret', 'resent'), require the truth of their complement clause. The point is that the relevant verbs are just those which must be ruled out in ordinary discourse, and I am content to turn the taxonomic problem of which these are over to the linguists for future resolution. 14 Authors may be deeply interested in conveying important truths in metaphorical or analogical fashion, but that is, clearly enough, another matter. 15 What Wittgenstein actually wrote is: "To understand a proposition means to know what is the case if it is true. (It is understood, therefore, without knowing whether it is true.)" (Tractatus Logico-Philosophicus, Routledge and Kegan Paul, London, 1961, paragraph 4.024.) And we may add, for these cases, without caring whether it is true. 16 These are the opening sentences of Joyce's Ulysses. 17 This also opens up the possibility of nonlinguistic representation, such as in tapestries, paintings, and the illustrations in mythology books. ~8 The previous paragraphs suggest that what lies at the root of aboutness here is a causal relation, particularly one in the specification of which we appeal to counteffactual relations. For an attempt to work this out, see the closing paragraphs of Dennis Stampe's 'Toward a Causal Theory of Linguistic Representation', in P. French, T. Uehling, and Howard Wettstein (eds.), Contemporary Perspectives in the Philosophy of Language, Univ. of Minnesota Press, Minneapolis, 1979, pp. 81-102. Such a theory has some promise, due to its focus on counterfactuals, of discriminating a story's being about someone from a character being modeled on someone, and of saving us from the disastrous step from the (likely true) observation that all fiction is in some sense autobiographical to the claim that no fiction is fiction. 19 A complication is needed for an utterance of, e.g., 'The Grand Inquisitor does not exist', since he does not figure in The Brothers Karamazov (hence, the analogue of (15a) cannot be what is presupposed, nor the analogue of (16a') what is asserted). But the Grand Inquisitor does figure in a (dream) story within that novel, and I propose to simply appeal to such stories within stories, as we ordinarily speak of the play within "A

436

ROD BERTOLET

Midsummer-Night's Dream", in specifying the presupposition and assertion appropriate for denials of the Grand Inquisitor's existence. 2o This, for the sort of utterance of the sentence with which I am concerned, e.g., my utterance of it to you. This is not true for my utterance of 'Santa has a sled pulled by flying reindeer' to my gullible niece, who is taken in by the story. But of course, when I tell her the story, I am not talking about the story. Whatever the difference is exactly, to tell her what the story is is not the same as to tell her the story. One difference, at least, seems to be that when I tell her the story I am not referring to it. In order to refer to the story, I would need to presuppose that there is nothing the story is about, and it would be irrational for me to believe that she believes this. When I take on the role of storyteller, I do not refer to or say anything about the story; this fact has a natural explanation within my theory, which I think is another point in its favor. I should also point out that my concern in this paper is with empty names, hence uncontroversially empty names. There are also disputed cases - perhaps 'King Arthur' is one - and these will need slightly different treatment, since those who use such names will not presuppose the analogue of (16a'), preferring to remain agnostic. They will rather presuppose something like this: (i) it is not clear whether there is an individual x such that (a) the relevant story is committed to x by the occurrence of the relevant name, and (b) the story is about x; (ii) whether there is such an x is irrelevant to the present conversational purposes. A more general view than the one I have presented, which attempted to handle the disputed cases along with the names we all believe to be vacuous, would, I suggest, exploit this sort of presupposition. Interestingly enough, this also suggests a way to handle another case put to me by Bill Rowe. Suppose a theologian is known not only to be an atheist but to think that the Bible is fiction on the order of Greek mythology, and that he is teaching a course on the Bible to students he knows to be believers. He will hardly presuppose, in my sense, that the Bible is not about anyone. But appeal to the sort of presupposition indicated in the previous paragraph seems to allow us to explain how he and his students can argue, and resolve arguments, over what Moses, Matthew or God did, by explaining how they might agree to be discussing what the Biblical story is, whatever the truth may be. Strictly, all this lies beyond the scope of this paper, but it provides some hints on how the account might be extended. 21 Keith Donnellan, 'Speaking of Nothing', Philosophical Review 83, 1974, 5-6. 22 There is, though, a further complication here: we need to consider what is implied by a story as well as what is asserted in it, since there seem to be truths to be told concerning the former as well as the latter. (For example, it seems plausible to suppose that we may say something true by uttering 'Sherlock Holmes had male genitalia', despite the fact that Conan Doyle did not deign to mention this.) I have given my solution to this problem in 'Inferences, Names, and Fictions', Synthese 58, 1984, 203-218. 23 As was noted earlier (n. 2), things are more complicated than this suggests. 24 Arguments of this sort appear in Gail Stine, 'Intentional Inexistence', Journal of Philosophical Logic 5, 1976, 491-510, and in Parsons, 'Referring to Nonexistent Objects', and Nonexistent Objects. For why the terminology is unfortunate, see Devitt's critical study of Contemporary Perspectives in the Philosophy of Language, Australasian Journal of Philosophy 59, 1981, 211-220. I have borrowed the characterization of our everyday psychological theory as folksy from Devitt.

REFERENCE,

FICTION,

AND FICTIONS

437

25 Or, in the more usual case in which the fear (etc.) is specified propositionally (...fears that p), the subject term of the embedded sentence. 26 John Woods, The Logic of Fiction, Mouton, The Hague, 1974, p. 136. (Freud wrote a psychoanalytic study of Wilhelm Jenson's Gradiva: A Pompeiian Fantasy.) 27 Woods, p. 135. 28 The numerous stages through which this paper has passed have benefitted from the discussion, advice, incredulous stares, and (occasionally) agreement of a number of people, some of whom I have forgotten and some of whom will have forgotten me. They include Mike Byrd, Alan Code, Lloyd Eggan, Berent En G Bill Gustason, Bob Hambourger, Dennis Henry, Martin Hunfley, Terry Parsons, Bill Rowe, Lilly Russow, Mary Sirridge, Elliot Sober, Morton Winston, and most prominently, Dennis Stampe. I am also indebted to the advice of the referees and editors of Synthese, which has led me to improve the paper by paring a good bit of extraneous material from it. Department of Philosophy Purdue University West Lafayette, Indiana 47907 U.S.A.