Conceivability and possibility

Abstract

We often decide whether a state of affairs is possible (impossible) by trying to mentally depict a scenario (using words, images, etc.) where the state in question obtains (or fails to obtain). These mental acts (broadly thought of as ‘conceiving’) seem to provide us with an epistemic route to the space of possibilities. The problem this raises is whether conceivability judgments provide justification-conferring grounds for the ensuing possibility-claims (call this the ‘conceivability thesis’). Although the question has a long history, contemporary interest in it was, to a large extent, prompted by Kripke's utilization of modal intuitions in the course of propounding certain influential theses in the philosophy of language and mind. The interest has been given a further boost by the recent two-dimensional approach to the Kripkean framework. In this paper, I begin by providing a detailed examination of a most recent attempt (due to Chalmers) to defend the thesis and argue that it is unsuccessful. This is followed by presenting my own gloss on Kripke's explanation of the illusions of contingency and I close by raising a general problem intended to undermine the prospects for a successful defense of the thesis.

Keywords:

• conceivability
• possibility
• modal intuitions
• primary intension, secondary intension
• two-dimensionalism

Acknowledgements

I would like to thank Brian McLaughlin, Chris Gauker and an anonymous referee of this journal for helpful comments on earlier drafts of this paper.

Notes

1. Kripke (1980), first published as an article in Davidson and Harman (1972).

2. See various articles in DePaul and Ramsey (1998).

3. This is more or less how Bealer (2002) defines epistemic possibility—he calls it the ‘possibility-of-qualitative-evidential neutrality’. Yablo (2002) also defines it in essentially similar terms, though he calls it ‘conceptual’ possibility. Thus P is conceptually possible iff some world w is such that it would have turned out that P, had w turned out to be actual. He picks out ‘Hesperus is not Phosphorus’ as an example of a conceptually possible proposition.

4. For some of the difficulties in circumscribing the notion of conceivability see the ‘Introduction’ in Gendler and Hawthorne (2002).

5. This strategy appears, often in different guises, in Chalmers's works. I shall, however, focus on Chalmers (2002a) as it provides the most fully developed account of this approach.

6. A somewhat similar claim about the dependence of conceivability (imaginability) on our state of knowledge has also been made by van Inwagen (2001) though the similarity is only apparent. Van Inwagen distinguishes between basic modal knowledge—which he thinks we possess—and substantive modal knowledge, ‘remote from the concerns of ordinary life’—which he thinks we lack. Examples of the former include such knowledge as the table I am sitting behind could have been two feet to the left while the latter includes such possibilities as the possibility of an extra phenomenal color and sheets of solid iron transparent to visible light. He rejects the suggestion that conceivability might provide substantive modal knowledge on the ground that our present state of knowledge may not allow us to conceive of the pertinent scenarios. However, he makes an additional claim that undermines his argument. He thinks that while we have basic modal knowledge, he ‘regard[s] much of this knowledge as mysterious … What is the ground of “basic” modal knowledge? I do not know how to answer these questions’ (2001, 250). But if van Inwagen is willing to concede that conceivability might provide an epistemic route to the space of possibilities, then the obstacle of lacking detailed scientific knowledge would no longer be present in basic cases (involving, say, the table's location). Van Inwagen's stance makes it more likely that either he does not take the conceivability thesis seriously or else his remark about conceivability being dependent on our knowledge state was not intended in the sense advocated in this paper.

7. Note that the question of the a priority of conceivability here pertains to the context of its formation rather than that of its justification.

8. Two-dimensionalism was introduced by, among others, Davies and Humberstone (1980), Kaplan (1989), Jackson (1993) and Stalnaker (2001).

9. To see this, consider how the reference of ‘water’ is fixed in the actual world: the description ‘the dominant clear drinkable liquid in the lakes and oceans’ yields an actual-world extension for ‘water’. This same reference-fixing description yields a different extension for worlds where the liquid in question is, say, XYZ rather than H₂O. Accordingly, the primary extension of ‘water’ is a function from each possible world to the extension generated by the actual-world reference fixing description within that world. The secondary intension of ‘water’, on the other hand, picks out the water in every counterfactual world. Given that ‘water’ turns out to refer to H₂O in the actual world, its secondary intension picks out H₂O in all worlds. The primary and secondary intensions of ‘water’ therefore differ. By contrast, the primary and secondary intensions of ‘H₂O’ coincide because the associated actual-world reference fixing description for ‘H₂O’, namely ‘the stuff made of molecules with two hydrogen atoms and one oxygen atom’, picks out H₂O in all possible worlds. See Chalmers (1996, 57).

10. See Yablo (2002) and Bealer (2002) for a discussion of this question. I might however mention that Chalmers's way of introducing secondary intensions is somewhat ad hoc. In Chalmers (2002b), having offered epistemic (primary) intensions as playing the role of Fregean senses, Chalmers goes on to defend his account against some of the familiar objections to Frege including Kripke's modal argument. The gist of Kripke's modal argument is that names and descriptions behave differently in modal (subjunctive) contexts (i.e., contexts or scenarios acknowledged not to be actual). Consider, for example, the following pair of sentences: (a) It might have been that Hesperus was not the evening star, and (b) it might have been that the evening star was not the evening star. Kripke plausibly argues (and Chalmers concurs) that (a) is true but (b) is false and concludes that names are not synonymous with descriptions. To absorb Kripke's point without giving up on senses (intensions), Chalmers associates an additional, subjunctive (secondary) intension with every expression beside the familiar epistemic (primary) intension. While an epistemic intension governs the application of an expression to epistemic possibilities, a subjunctive intension governs its application to counterfactual possibilities, i.e., ways our world might have been but (probably) is not. Thus, while the epistemic intensions of ‘Hesperus’ and ‘the evening star’ are similar, their subjunctive intensions are distinct which is why the truth-values of (a) and (b) in Kripke's argument differ. This, I think, brings out very clearly the ad hoc nature of his approach.

11. Here I differ from Brueckner (2001) in regard to the dialectical thrust of the objection although I agree with his criticism of Chalmers's argument.

12. Here, all I need is the so-called ‘weak underdetermination’ (also known as ‘methodological underdetermination’) as opposed to strong underdetermination where one is supposedly in possession of all the relevant evidence for a particular theory.

13. See, for example, McGinn (1989).

14. Note that although a sentence like ‘water is H₂O’ is said to express a false proposition under epistemically possible circumstances (as described earlier), this proposition is different from any false proposition about, say, twater for, by hypothesis, we are still talking about water, rather than another substance (twater), under the aforementioned conditions.