• Aucun résultat trouvé

Explanation of Evans’s argument

Dans le document Disagreeing about fiction (Page 94-99)

Direct arguments against the modal account

CHAPTER 4. DIRECT ARGUMENTS AGAINST THE MODAL ACCOUNT

4.3 Evans against Lewis

4.3.2 Explanation of Evans’s argument

Lewis’s storytelling

For Lewis, getting the fictional “truth” of a fictional statement implies that we first recognise that the statement is fictional, i.e. the result of an act of storytelling. Lewis then says:

Storytelling is pretence. The storyteller purports to be telling the truth about matters whereof he has knowledge. He purports to be talking about characters who are known to him, and whom he refers to, typically, by means of their ordinary proper names. But if his story is fiction, he is not really doing these things. Usually his pretence has not the slightest tendency to deceive anyone, nor has he the slightest intent to deceive.21 This way of describing what storytelling is motivates Lewis’s Analysis 0 which amounts to paraphrasing (3) and (4) into (5) and (6). The fictional operator “In

21[Lewis1978], p. 40.

CHAPTER 4. DIRECT ARGUMENTS AGAINST THE MODAL ACCOUNT

Back totable of content Page 92 of369

fiction F, ...” is thus developed into “Were F told as known fact rather than fiction”.

This paraphrase was explained in §2.1.2.

Consequently, the counterfactual analysis of fictional statements requires that one considers the closest possible worlds in which the act of storytelling that happens in our world is a truthful testimony. So the two speech acts (the real act of storytelling and the possible act of truthful testimony) differ only in their intention: one is pretended, the other is serious. In other words, the two speech acts must have the samesemantic content, i.e. contain the same propositions, namely those we recognise as the fictional “truths” in the real world.

It is important that the second speech act (the possible act of truthful testimony) has a serious intention. Indeed, there can be another world in which the second speech act is also an act of storytelling (i.e. pretence), albeit distinct from the original one. Such a case is illustrated by Pierre Ménard in Borges’s eponymous short story.

Ménard is writing fiction in the early 20th century. By a very rare combination of factors, Ménard’s fiction happens to match word for word Cervantes’sDon Quixote.

It is explicitly fictional that Ménardindependently wrote a story which is identical to Cervantes’s original story. However, the aesthetic value of the two works, so Borges argues, are completely different. Ménard is writing in a language and style which is very remote from the spoken Spanish of the early 20th century, while Cervantes uses his own everyday language. So Ménard, Borges concludes, produces a work whose originality is far greater than that of Cervantes, albeit the same work.

Crucially, Ménard intends to produce a fiction, i.e. he is pretending. So when reading Cervantes’sDon Quixote, one should not pick up the possible world in which Ménard tells the story. Otherwise we should then go to the closest possible world in which Ménard’s story is told as known fact, which is, by definition, a different possible world. Ménard does not inhabit the same world as Don Quixote, Sancho Panza and Rocinante. Instead, one should go directly to the world in which Cervantes’s story is told as known fact.

Evans proceeds and denies that there is such a possible world in which a story is told as known fact and has the same semantic content. In other words, the second speech act cannot have the same semantic content as its corresponding act of storytelling.

Evans’s argument

In order to show that the two speech acts (the real pretence and the possible truthful testimony) cannot have the same semantic content, let us, first, focus on names.

Names contribute to the semantic content of the speech act in which they occur

by providing a referent, or failing to do so. In other words, the semantic content of a name is either a referent, or not a referent. By definition, a fictional name has no referent. So a fictional name does not contribute to the semantic content of the speech act in which they occur by providing a referent.

But the name, say “Hamlet”, is used both in Shakespeare’s speech act and in the possible truthful testimony associated with Shakespeare’s pretence. According to Lewis, it must be that the fictional name in the actual world should become a real proper name in the possible world in which the story is told as known fact.

This entails that, in the targeted possible world, “Hamlet” has a referent. Hence, by definition, “Hamlet” contributes to the semantic content of the second speech act (the possible truthful testimony) by providing a referent.

Consequently, the first speech act and the second speech act do not have the same semantic content. Indeed, they differ in semantic content at least as far as the name “Hamlet” is concerned. So there are no possible worlds in which Hamlet can be told as known fact. In other words, the possible truthful testimony associated withHamlet does not have the same semantic content as Shakespeare’s actual story.

Pretence is powerful: it secures the fact that a fictional name is necessarily fic-tional. Since a fictional name isnecessarily fictional, there are no possible worlds in which it is not fictional. This is, in essence, Evans’s argument against Lewis.

Note that Lewis does not agree with direct reference about names, he rather holds the view that names have descriptive content. So Lewis’s general picture is not affected by Evans’s argument. Evans’s argument should rather be thought of as an incompatibility result: it is simply impossible to combine direct reference with the counterfactual analysis of fictional statements.

Kaplan and Kripke’s arguments: The same argument was made independently by Kaplan and Kripke, though it was not directed at Lewis’s analysis of fictional statements.

In appendix XI of [Kaplan 1973], Kaplan warns us about a surreptitious change of language. Focusing on the worlds in which the myth of Pegasus is “truthfully told” (the “M worlds”), Kaplan remarks that the name has changed its meaning:

But beware the confusion of our language with theirs! If w is an M world, then their name “Pegasus” will denote something with respect to w, andour description “thexsuch thatxis called ‘Pegasus’” will denote the same thing with respect to w, but our name “Pegasus” will still denote nothing with respect to w. Also, in differentM worlds, different possible individuals may have been dubbed “Pegasus”; to put it another

CHAPTER 4. DIRECT ARGUMENTS AGAINST THE MODAL ACCOUNT

Back totable of content Page 94 of369

way, our description “thex such that x is called ‘Pegasus’” may denote different possible individuals with respect to differentM worlds.22

In [Kripke 1972], Kripke makes explicit that a metaphysical consequence of his causal theory of reference is the following:

I hold the metaphysical view that, granted that there is no Sherlock Holmes, one cannot say of any possible person that he would have been Sherlock Holmes, had he existed.23

Lewis’s “half-perception”

Lewis indeed “half-perceives the problem”, namely that is there is no possible world at which the story is told as known fact and the two speech acts have the same content. He explicitly discuss the problem of names like “Sherlock Holmes” changing their semantic contribution. As he puts it:

At those worlds where the same story is told as known fact rather then fiction, those names really are what they here purport to be: ordinary proper names of existing characters known to the storyteller.24

His way of responding to the worry is to suggest that the name “Sherlock Holmes”

as used in our world is not an ordinary proper name. On the contrary he suggests the following picture:

[...] the sense of “Sherlock Holmes” as we use it is such that, for any world w where the Holmes stories are told as known fact rather than fiction, the name denotes at wwhichever inhabitant of wit is who there plays the role of Holmes. Part of that role, of course, is to bear the ordinary proper name “Sherlock Holmes”. But that only goes to show that “Sherlock Holmes” is used at w as an ordinary proper name, not that it is so used here.25

As we have just seen, Kaplan, Kripke and Evans precisely deny that this can happen.

For there is no world at which the name “Sherlock Holmes” is the same and picks up an individual. As Kaplan puts it, the only way one can do this is to change one’s language.

22[Kaplan1973], p. 507.

23[Kripke1972], p. 158.

24[Lewis1978], p. 40.

25[Lewis1978], p. 41.

So far Lewis is alright, for he does not hold that fictional names are rigid des-ignators. But in footnote 9 he considers the objection that one can use “Sherlock Holmes” without knowing the descriptive content of “Sherlock Holmes”:

Many of us have never read the stories, could not produce the de-scriptions that largely govern the non-rigid sense of “Sherlock Holmes”, yet use this name in just the same sense as the most expert Baker Street Irregular. There is no problem here. Kripke’s causal picture of the conta-gion of meaning, in “Naming and Necessity”, will do as well for non-rigid senses, as for rigid ones.26

So Lewis fails to see the incompatibility between the causal theory of names and his counterfactual analysis of fictional statements. Lewis’s failing to perceive this is difficult to understand.27

Generalisation of Evans’s argument

What is true about fictional names actually generalises to all kinds of semantic information, as Evans “would have argued”. He holds that information (and misin-formation) in general (hence semantic inmisin-formation) is individuated by causal origin.

That is, information enters the cognitive system from the outside world through perception and testimony seen as causal mechanisms.

Hence, only the information originating in the outside world and causally related to the agent can be interpreted by the cognitive agent. In the case of fiction, pretence makes it the case that the information received is not veridical. Pretence is a case of deception, albeit a case where no harm is done. As a result, what is represented in fiction does not hold in reality. Indeed Hamlet is not a human being in reality, although the information contained in the fictional proposition expressed by the fictional sentence “Hamlet is a human being” is definitely part of the fictional content of Hamlet. The real-world origin of this fictional information is clearly identified: it is Shakespeare’s playHamlet, this sophisticated real-world object with which we can causally relate by reading it (or watching it on stage).

Were the complex fictional information contained inHamlet veridical, everything is changed. The origin of the information in this case would clearly not be the play

26[Lewis1978], p. 41.

27Kripke’s above quote comes from the addenda to the second edition ofNaming and Necessity, published in 1980, though. And Lewis was working with the first edition of 1972. So either Lewis really misunderstood Kripke’s lectures, or Kripke was definitely cryptic during his lectures. As for ignoring Kaplan’s version of the argument published in 1973, I guess Lewis has no excuse...

CHAPTER 4. DIRECT ARGUMENTS AGAINST THE MODAL ACCOUNT

Back totable of content Page 96 of369

(the real-world object, written by Shakespeare), but the events that it truthfully tells the story of. However, if such information could enter one’s cognitive system, then it must be that one is causally related to it. There is no world in which we, ourselves, are causally connected to Hamlet. We, ourselves, could only possibly be causally related to individuals in the actual world (and, unfortunately, only very few of them).

So there is no sense in which the fictional information we can gather by reading fictions and going to the theater could be gathered otherwise, in a causal manner.

The same incompatibility result generalises: one cannot hold both a causal theory of semantic information and Lewis’s counterfactual analysis of fictional statements.28 Application on the mud-pie game

The mud-pie game example is structurally identical to Hamlet, but the causal rela-tions are much simpler to describe. Some children play with mud, forming globs of mud, and pretending thatthese mud globes are cakes.

Manipulating the globs of mud in such and such a way in the real world would correspond to such and such “truths” about the corresponding cakes in the pretence.

For instance, adding some gravel on top of one glob of mud is likely to suggest that

“this is a cake with chocolate chips on top of it” is “true” in the pretence, and so on.

But these globs of mud they manipulate and pretend about could not possibly be cakes. If they were cakes, then they would not be globs of mud. The information the children gather about these globs of mud originates in the causal relations they have with them, especially through manipulation. It is just impossible that they could gather the same kind of information, were these globs of mud to be real cakes.

In other words, there is no possible world in which these globs of mud arecakes.

Consequently, any counterfactual statement whose antecedent is “if these globs of mudwere cakes” is bound to run into trouble.29

Dans le document Disagreeing about fiction (Page 94-99)