Abstract
What sorts of things are the intuitions generated via thought experiment? Timothy Williamson has responded to naturalistic skeptics by arguing that thought-experiment intuitions are judgments of ordinary counterfactuals. On this view, the intuition is naturalistically innocuous, but it has a contingent content and could be known at best a posteriori. We suggest an alternative to Williamson’s account, according to which we apprehend thought-experiment intuitions through our grasp on truth in fiction. On our view, intuitions like the Gettier intuition are necessarily true and knowable a priori. Our view, like Williamson’s, avoids naturalistic skepticism.
Similar content being viewed by others
Notes
For a sampling, see the various papers in DePaul and Ramsey (1998).
Williamson (2007), ch. 6 puts it thus: Someone could stand in the relation described in the Gettier story to some proposition. We characterize Williamson’s GC relation in terms of texts of Gettier stories, which is clearly what Williamson has in mind, because we believe there is an important difference between stories and texts. See Sect. 4 below.
Williamson is trying to formalize something like (CF) “if a thinker were Gettier-related to a proposition, he/she would have justified true belief in it without knowledge.” It can be surprisingly challenging to formalize such ‘donkey anaphora’ counterfactuals. See §4 of Williamson (2007)’s ch. 6, and Appendix 2. Our objections to Williamson here will transcend the question of whether this formalization faithfully captures the English counterfactual—they will apply to (CF) and (2cf) alike, even if they are not equivalent.
Williamson recognizes this worry in his (2007), ch. 6, §5. Incredibly, he seems willing to bite the bullet on this point, and to admit that in such cases, our thought experiments are defective. We find this result implausible, both on its face and in light of the particular objection mentioned in the main text below.
Richard Heck has proposed (in conversation) another move on behalf of Williamson. Instead of a counterfactual, one could render premise (2) as a generic like ‘Fire engines are red’, which is true not because all fire engines are red (we’ve seen some green ones), but because fire engines typically are red. Premise (2) then becomes: in cases that satisfy the Gettier text, people (typically) have justified true belief without knowledge. The conclusion we arrive at, then, is that typical cases that satisfy the Gettier text are ones in which people have justified true belief without knowledge. The generic version of (2) would seem to avoid the objections that plague counterfactual renderings while keeping the spirit of Williamson’s counterfactual approach.
Lewis (1978). This is (paraphrased) Lewis’s Analysis 1, the simplest of several theories of truth in fiction he suggests.
But the views would still not be identical. Consider the modal status of the propositional content of the Gettier intuition. On the Lewis account of truth in fiction, the account above would imply that in order to reach the Gettier conclusion, one relies on knowing (q*): necessarily, if the fictional truths (for this actual fiction) were true, Joe would have NKJTB. On the Williamson, one relies on knowing (2cf): if some x were to stand to some p as in the Gettier text, then anyone who stood to any proposition in a way matching the text would have NKJTB. Williamson’s (2cf) is false in possible worlds where nearby Gettier worlds include bad cases. But q* may be true in all worlds, as ‘actually’ rigidifies the embedded counterfactual. A given story has the same fictional truths in every possible world; a given text picks out different fictional truths depending upon the circumstances in which it is told.
In an earlier version of this paper, Ichikawa defended a view related to the one explored in this section. That account was: (1) It’s possible for someone to be in a position like Joe’s position in this fiction I’m actually engaging with. (2) Necessarily, anyone in a position like Joe’s position in this fiction I’m actually engaging with would have NKJTB. (3) Therefore, it’s possible to have NKJTB. The rigidification of the fiction was to allow for us to misinterpret authors and engage with our own private Gettier fictions, which would underwrite our fictions. On this view, the Gettier conclusion here will be a priori if one knows a priori that Joe’s position in the fiction one’s actually engaging with comes to justified belief without knowledge. If (and only if) introspective knowledge is a priori, this is plausible. Our eventual account will not depend on the apriority of introspection.
Premise (1p) isn’t quite right. Joe is a fictional character, and as Kripke says, he may be essentially so. (Although it is possible for there to be some person that matches the description of Sherlock Holmes, we might doubt that any such possible individual would be Sherlock Holmes.) If this is right, then it is not possible for all of the members of STORY to be true, for it is not possible for any person to be Joe. The solution to this worry is technical and peripheral to the larger project, and the simpler version we give above is clearer and more intuitive, so we present the more accurate version in this footnote. Although it is not strictly possible for fictions involving characters picked out by proper names (who are not inhabitants of the actual world) to be true, there are available modified versions of the fictions which are possible. Professor Muthos’s story began with the metaphysically impossible, “Joe had left his watch at home.” A modified, metaphysically possible version of the story would insert, at the start of the story, something like the computer programmer’s variable declaration: existential introductions to all the elements in the story that would have had proper names, along with instructions as to how to refer to them while engaging with the story. So we’d have something like: “There was some guy, whom we will refer to as ‘CHARACTER1’. CHARACTER1’s name was ‘Joe’. Joe had left his watch at home….” If every object with a proper name is introduced thus, then the story is metaphysically possible. On the more technical account, then, we replace the story with a new story (whose truths describe possible worlds that are qualitative duplicates to those described in the old story), taking care to introduce each character in the way described. Then we can reason as described above, using the set of truths in the new fiction. See ch. 6, §2 of Williamson’s (2007) for a similar treatment.
Ernest Sosa has suggested in conversation that these two premise argument accounts of Gettier reasoning don’t do justice to the simplicity of the Gettier intuition. Really, in having the Gettier intuition, one just comes to see that one way to satisfy the text of the Gettier story involves someone having justified true belief without knowledge. Sosa’s suggestion may amount to the proposal that people directly apprehend (3). Moreover, while (3) is no doubt necessarily true, necessity doesn’t enter into its content. (Hence, the discussion in Sect. 10 is irrelevant.) The propositional content of the Gettier intuition, then, is much simpler in form than our (2p). After some thought on the issue, we’re not confident that this suggestion represents a genuine alternative to our view. After all, we need some account of how the thought experiment generated (3). Surely, it did so through guiding one to consider a fairly determinate possibility—something like the possibility in which p is true. But it’s not enough merely to have in mind this determinate possibility; one has see that if this possibility obtains, then there would be an instance of justified true belief without knowledge. Of course, this process is very close to (if not identical with) what we’ve suggested. We don’t mean to insist that such a process be entirely explicit, conscious, or deliberate, but it’s hard to see how thought experiments work without some such process. See also Williamson’s (2007), ch. 6, §2.
For some psychological studies on the way humans think about fictional worlds, see Skolnick and Bloom (2006a, b). Weisberg (née Skolnick) and Bloom are currently working on a more directly-related question: how children and adults come to recognize fictional truths that are not explicitly stated. Results are not yet published.
Williamson (2005) also emphasizes that Gettier cases need not be fictional (p. 15).
For a few attempts to spell out this conceivability, see Chalmers (2004) and Yablo (1993). For a very different kind of strategy, see Bealer (2004). We do not commit to the success of any of these attempts here; we merely express confidence that there is a relevant kind of conceivability according to which our p is conceivable.
See Yablo (2004) for one interesting discussion of conceptual possibility.
Brian Weatherson (2003) argues that given David Lewis’s theory of meaning (the “reference magnetism of natural properties” view—see Lewis (1984)), we might just discover that ‘know’ and ‘believe truly and justifiedly’ are co-intensional. We find this suggestion implausible and will ignore it. (If Lewis’s theory of meaning does entail as much, all the worse for his theory of meaning.)
Thanks to Richard Heck for helpful discussion here.
The only recent dissension we know of from this truism is Bonjour (1998).
For a good discussion of a particular sort of case in which we may go awry in distant worlds, see Gendler (2002). Gendler’s suggestion that we are subject to certain kinds of systematic mistakes with distant thought experiments is, of course, consistent with our invocation of our general reliability.
There are numerous ways that a correct metasemantic account could explain disagreement. An easy explanation is provided if the relation between concept and conceptual role is one-to-many. Different conceptual roles could lead to different judgments even if they coincide when it comes to more quotidian cases, and concepts might apply indeterminately in such cases. Alternatively, if for every concept there is one conceptual role, it might be that people possess different but very closely related concepts.
See Gendler (2002).
□A = def ∀p (p□→A). Note that, given that one can define metaphysical necessity in terms of counterfactuals, the existence of certain counterfactual knowledge is also sufficient to show there must be some modal knowledge, and thus, that some explanation as to how we come to have modal knowledge must ultimately succeed. See Williamson (2005), Williamson (2007) ch. 5, Hill (2006).
While they have not necessarily made the a priori/a posteriori distinction in this way, a number of prominent philosophers have tried to defend the a priori by showing that one’s ability to entertain the proposition in question is sufficient for having the inclination to believe it. See, for example, Bealer (1999), Boghossian (1996) and (2003), Peacocke (1993), and Sosa (2004).
(2007), ch. 5.
Incidentally, we are not so fully convinced of (T), but we assume it for the sake of argument.
The flexible individuation of cognitive/inferential skills is probably essential in that regard. We can think of the same person both as simultaneously having the skill of drawing valid instances of modus ponens and having the skill of drawing some more restricted set of the valid instances of modus ponens. This flexibility should allow one to isolate completely those skills which contribute towards approximating the idealized set of cognitive/inferential skills associated with a concept.
Williamson (2004), p. 152.
References
Bealer, G. (1999). A theory of the a priori. Philosophical Perspectives, 33, 29–55.
Bealer, G. (2004). Modal epistemology and the rationalist Renaissance. In T. Gendler & J. Hawthorne (2004).
Boghossian, P., & Peacocke, C. (Eds.). (2000). New essays on the a priori. Clarendon Press.
Boghossian, P. (1996). Analyticity reconsidered. Noûs, 30.3, 360–391.
Boghossian, P. (2003). Blind reasoning. The Aristotelian Society Supplementary Volume, 77.1, 225–248.
Bonjour, L. (1998). In defense of pure reason. Cambridge University Press.
Chalmers, D. (2004). Does conceivability entail possibility? In T. Gendler & J. Hawthorne (2004).
Currie, G. (1990). The nature of fiction. Cambridge University Press.
DePaul, M., & Ramsey, W. (Eds.). (1998). Rethinking intuition. Lanham: Rowman & Littlefield.
Fodor, J. (1998). Concepts. Clarendon Press.
Gendler, T. (2002). Personal identity and thought experiments. The Philosophical Quarterly, 52(206), 34–54.
Gendler, T., & Hawthorne, J. (Eds.). (2004). Conceivability and possibility. Oxford University Press.
Greenberg, M., & Harman, G. (2006). Conceptual role semantics. In E. Lepore & B. Smith (Eds.), The Oxford handbook of philosophy of language. Oxford University Press.
Hanley, R. (2004). As good as it gets. Lewis on truth in fiction. Australasian Journal of Philosophy, 82, 112–128.
Hill, C. (2006). Modality, modal epistemology, and the metaphysics of consciousness. In S. Nichols (Ed.), The architecture of the imagination. new essays on pretense, possibility, and fiction. Oxford: Oxford University Press.
Kitcher, P. (1980). A priori knowledge. The Philosophical Review, 89.1, 3–23.
Kitcher, P. (2000). A priori knowledge revisited. In P. Boghossian & C. Peacocke.
Kripke, S. (1980). Naming and necessity. Harvard University Press.
Lewis, D. (1978). Truth in fiction. American Philosophical Quarterly, 15, 37–46. (Reprinted in Lewis, D. (1983)).
Lewis, D. (1983). Philosophical papers (Vol. I). Oxford University Press.
Lewis, D. (1984). Putnam’s paradox. Australasian Journal of Philosophy, 62, 221–236.
Peacocke, C. (1993). How are a priori truths possible? European Journal of Philosophy, 1.2, 175–199.
Skolnick, D., & Bloom, P. (2006a). The intuitive cosmology of fictional worlds. In S. Nichols (Ed.), The architecture of the imagination. New essays on pretense, possibility, and fiction. Oxford: Oxford University Press.
Skolnick, D., & Bloom, P. (2006b). What does Batman think about SpongeBob? Children’s understanding of the fact/fiction distinction, Cognition, 101(1), B9–B18.
Sosa, E. (2004). Reliability and the a priori. In T. Gendler & J. Hawthorne (Eds.), Conceivability and possibility. Oxford University Press.
Walton, K. (1990). Mimesis as make-believe. Harvard University Press.
Weatherson, B. (2003). What good are counterexamples? Philosophical Studies, 115, 1–31.
Williamson, T. (2003). Understanding and inference. The Aristotelian Society Supplementary Volume, 77.1, 249–293.
Williamson, T. (2004). Philosophical ‘intuitions’ and scepticism about judgment. Dialectica, 58.1, 109–153.
Williamson, T. (2005). Armchair philosophy, metaphysical modality and counterfactual thinking, Presidential address. Proceedings of the Aristotelian Society, 105, 1–23.
Williamson, T. (2006). Conceptual truth. The Aristotelian Society, Supplementary Volume, 80 (forthcoming).
Williamson, T. (2007). The philosophy of philosophy, (forthcoming).
Yablo, S. (1993). Is conceivability a guide to possibility? Philosophy and Phenomenological Research, 53, 1–42.
Yablo, S. (2004). Coulda, woulda, shoulda. In T. Gendler & J. Hawthorne (Eds.), Conceivability and possibility. Oxford University Press.
Acknowledgements
We are grateful to Tamar Szabó Gendler, Richard Heck, Kelby Mason, Joshua Schechter, Peter Godfrey-Smith, Alvin Goldman, Deena Skolnick Weisberg, Ernest Sosa, John Turri, and Timothy Williamson for invaluable comments and discussion. An earlier version of this paper was presented at the 2006 Harvard-MIT Graduate Philosophy Conference; we are grateful for helpful discussion from the conference participants.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Ichikawa, J., Jarvis, B. Thought-experiment intuitions and truth in fiction. Philos Stud 142, 221–246 (2009). https://doi.org/10.1007/s11098-007-9184-y
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11098-007-9184-y