Modal fictionalism: a response to Rosen.
The deflationist strategy in question has fittingly been coined "Modal Fictionalism", juxtaposing it nicely with Lewis's own Modal Realism (Lewis, 1986). Rosen's ingenious proposal is to exploit an analogy with standard accounts of a general analysis of truth in fiction. Statements like "There is a ghost that haunted Heathcliff on the windy moors" and "There are possible worlds other than our own" are true but not true simpliciter. They must implicitly be understood as elliptical for longer sentences explicitly about a story. The reason we assent to their truth is because the context of utterance is such that they can only be interpreted as sentences beginning with a silent sentential operator, "according to such and such fiction...". Thus, we are not committed to the existence of any entities embedded therein, despite an appearance to the contrary.
Once the operator has been recognised, Rosen claims that we are in a position to see how useful it is. By treating the realist's thesis as a fiction, the fictionalist can analyse a wide variety of modal claims cashing in on the realist's analysis. For example, a fictionalist can say that it is possible that there are blue swans iff according to the hypothesis of a plurality of possible worlds (PW), there is some world containing blue swans. To be sure, we can, as Rosen does, put the proposal more generally:
... let P be an arbitrary modal proposition. The modal realist will have ready a non-modal paraphrase of P in the language of possible worlds; call it P*. The realist's assertions about possible worlds are guided by explicit adherence to the schema P iff P*. The fictionalist's parasitic proposal is therefore to assert every instance of the schema: P iff according to the hypothesis of the plurality of worlds [PW], P*. Like modal realism, the theory would seem to provide truth conditions for modal claims in a systematic way. (Rosen 1990, pp. 332-3) Thus, it would appear, the fictionalist can help himself to the possible worlds framework without presupposing an ontology of worlds. Nice work, if it is consistent.
Rosen considers, in some detail, a number of problems for the fictionalist thesis. It is my contention, though, that there is an additional problem which has not been considered, a problem that is far more damaging to the programme.
The main motivation for going fictionalist about possible worlds, as we have seen, is ontological simplicity. The fictionalist allegedly has the obvious advantage of not presupposing an ontology of possible worlds. It is my contention, though, that Rosen seems forced to admit that possible worlds do in fact exist. In order to see why, consider first the proposition that 2 + 2 = 4. The proposition is necessarily true and the fictionalist will give the truth conditions in the following way:
It is necessary that 2 + 2 = 4 iff according to PW, at all worlds, 2 + 2 = 4. We should notice two things:
(i) The fictionalist is committed to the view that the proposition on the left hand side of the formula affirms a modal fact, a real modal fact. This is because the proposition on the right hand side of the formula affirms a real fact, a fact to the effect that a story has a certain content. Futhermore, if we use the standard modal axiom "[unkeyable]p[unkeyable]p" we can derive the non-modal fact that 2 + 2 = 4.(1)
(ii) The propositon on the left hand side of the formula does not quantify over worlds. Quantification over worlds does occur on the right hand side of the formula, but the quantifier occurs within the scope of the fictionalist's prefix and therefore is not existentially committing.
The trouble comes when we consider a proposition like "There is a plurality of possible worlds". The fictionalist should say that it is false. The realist, on the other hand, will not only say that it is true, but that it is true necessarily. Let's see why.
(i) Most of us agree that there are other ways the world might have been. The realist, though, says something more: every way the world might have been is a way that some world is.
(ii) Because the accessibility relation between worlds is both transitive and symmetric, according to PW, the proposition that there is a plurality of possible worlds is true at all worlds.
The fictionalist, then, has a problem. Consider the following:
It is necessary that world exist iff according to PW, it is true at every world that worlds exist.
It has already been established that the right hand side of the formula is true. Thus, if the fictionalist's analysis is correct, the left hand side of the formula will also be true. This is not a good consequence for the fictionalist. By using the axiom "[unkeyable]p[unkeyable]p" we can derive the non-modal proposition that worlds exist. The fictionalist, it seems, is committed to an ontology of worlds.
The fictionalist could of course modify his anlysis so as to rule out an analysis of propositions which prove difficult. For example, the fictionalist might say the following: P iff according to PW, P*, except where P, and therefore P*, embed a proposition about worlds. Such a proviso certainly seems ad hoc. Perhaps the analysis could be defended from this charge. After all, propositions about worlds cannot be seen as part of our ordinary language. They are introduced only as instrumental devices and therefore are not, and should not, be analysed in the usual way.
Once we have introduced the notion of a world into our analytic discourse, though, sentences of the form "Necessarily world exist" will presumably have some meaning; it's just that the fictionalist will want to say that they are false. According to the fictionalist's analysis, the truth conditions for modal statements about worlds, unlike other modal propositions, will not depend on the content of PW.
Such a restriction would, however, seriously undercut the analytic advantage of the program. It's easy to avoid any ontological commitment to worlds; just adopt a primitive modal language. The alleged advantage fictionalism has over primitive modal languages is its reduction of a wide variety of modal primitives to one.(2) Once such ad hoc restrictions are placed on the program, though, the primitives are multiplied. One is left wondering if it would not be better merely to take the more conventional modal operators as primitive.
To illustrate, consider two further arguments:
(i) Consider the subjunctive conditional "If it were the case that p, then it would be the case that q" (p [unkeyable] q). According to a realist like Lewis, this is true at a world i if and only if some p [unkeyable] q word is closer to i thn any p [unkeyable] q world (if there are any p worlds). Assume further that p is any proposition true at the actual world, and that q is the proposition that world exist. The actual world, then, is a p [unkeyable] q world according to the realist. Because the actual world is closest to itself, the throughgoing fictionalist--that is, one who exploits the full resources of modal realism--should say that the subjunctive is true. Thus, the proposition that worlds exist is a consequence of the fictionalist's analysis of subjunctives.
(ii) Consider the following strict implication: p [unkeyable] q. According to the realist this is true if and only if all p worlds are q worlds. Assume again that p is any true antecedent and that q is the proposition that worlds exist. Because all worlds are q worlds according to the realist, if the antecedent is true, the thoroughgoing fictionalist should say that the implication is true. Thus, the proposition that worlds exist is a consequence of the fictionalist's analysis of strict implication.
We could go on, but the point should be clear. If the fictionalist were to leave modal propositions about worlds analysed, he would be left with a remainder--a series of primitives, like those who adopt a primitive modal language. As it is not clear that the fictionalist's language is complete without the requisite ontological commitment, it seems that if he wants the benefits, he had better pay the cost.
Rosen's paper is a provocative one. What I hope to have shown, however, is that the analysis, as it stands, will not do the work that Rosen hopes it will. The deflationary motivation is, of course, admirable. We should not want to commit ourselves to a realm of possibilia unless it is essential to an analysis of modality. The deflationist, though, cannot rely on the tools provided by Rosen.(3)
(1) It might be though that the fictionalist has a problem countenancing this inference. It is difficult to see how a fiction would imply anything about actuality. Rosen's fiction, though, is a fiction of many worlds, among which the actual world is one. What is true at all worlds in the fiction, then, must be true at the actual world.
(2) Rosen allows that his analysis inherits a primitive modal components. The fictionalist's operators--according to PW--is classed as a modal operator. Unlike the realist, then, the fictionalist has not "analyse(d) all modal locutions in non-modal terms ... what he has done ... is reduce a wide variety of modal notions to this one. Now this may not seem as impressive as a thoroughgoing eliminative reduction. But it remains a non-trivial analytic advance. No similar claim can be made, for example, by those who take the logican's standard modadl operators as primitive" (pp. 344-5).
(3) I am indebted to Nicholas Agar, Frank Jackson and Mark Sainsbury for helpful comments and discussion.
Lewis, D.K. 1986: On the Plurality of Worlds. Oxford: Basil Blackwell.
Rosen, G. 1990: "Modal Fictionalism". Mind, 99, 395, pp. 327-354.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Gideon Rosen, Mind, vol. 99, p. 327, 1990|
|Date:||Jan 1, 1993|
|Previous Article:||Constitution is identity.|
|Next Article:||Ramachandran on restricting rigidity.|