Tidman's "Conceivability as a Test for Possiblity": Classic Papers in Modal Epistemology Series
Notes on Paul Tidman's “Conceivability as a Test for Possibility”, American Philosophical Quarterly 31(4), October 1994: 297-309.
Summary of the paper:
Tidman’s paper consists of two main parts: the exposition and critique of conceivability theories of modal epistemology, and a brief exposition of an intuition-based account of modal epistemology.
In the first part, Tidman critiques four main construals of conceivability: conceivability as picturability, as understandability, as believability, and as entertainability.
According to the picturability construal, P is conceivable iff one can form a mental image that represents P being the case, where ‘mental image’ is construed broadly enough so as to include the whole gamut of phenomenal imagery: visual images, sounds, tastes, textures, etc. Tidman argues that the picturability account is inadequate for several reasons. First, many possible states of affairs aren’t picturable, yet we seem to have justified beliefs about the modal status of propositions that express them, in which case picturability isn’t a necessary condition of having justified modal beliefs. There are two kinds of unpicturable states of affairs: those that are in fact unpicturable (e.g., there can be and are colors, sounds, etc. that we cannot picture, since they don’t exist, or we can’t experience them), and those that are in principle unpicturable (e.g., God’s being omnipotent, our having souls, 2+2’s equaling 5, etc.) Second, the account admits of counterexamples. For example, one can picture Escheresque impossibilities. Third, mental images don’t, all by themselves, represent things. Rather, human intentions are essential to the determining of what images represent. The last two problems lead to a fourth -- what Tidman calls the “What Counts?” objection: since not any old mental image counts as a picturing of a genuinely possible state of affairs p , we need criteria to determine whether something is a legitimate case of picturing p. However, the task of formulating such criteria is fraught with apparently insurmountable difficulties.
According to the understandability construal, P is conceivable iff P is understandable, i.e., iff one can understand the sentence that expresses P. Tidman raises two objections to the understandability account. First, it admits of counterexamples, for we can understand necessary falsehoods. Second, the inability to understand a proposition is only indicative of our cognitive limitations, and not of the modal status of certain states of affairs.
On the believability construal, something is conceivable iff it is possible that it be believed. Tidman has two objections to this account. First, it admits of counterexamples, for we seem to be able to believe necessary falsehoods. Tidman’s examples here are of two main types: (i) identities that are only knowable a posteriori (e.g., I believe that Raquel is a woman and not my best friend Kevin, even though Raquel is just Kevin in disguise). (ii) Propositions that are either necessarily true or necessarily false, but are such that some people believe that they are true, and others that they are false (e.g., that a necessarily existent individual exists). Second, even if one can give a compelling account that one can only believe what is possible (as Stalnaker’s account of belief implies), this leads to a major problem for the believability account. For it certainly seems as though we sometimes believe impossible things (e.g., that Hesperus is distinct from Phosphorus). But if so, then if we insist that, although we seem to believe that such cases are possible, we really don’t, then this implies that we are often unable to discern what we believe. Now even if one finds nothing implausible about this, one is left with the problem that we can’t tell if we’ve satisfied the believability condition, in which case we can’t tell when we’ve conceived of a state of affairs that is truly possible.
According to the Entertainability construal, p is entertainable iff it p imaginable, where something is imaginable iff one can tell a coherent story according to which p is true. Richard Swinburne’s version of this account is taken to be the most detailed version of it, so his is the version on which Tidman focuses. According to Swinburne, P is coherent iff one can conceive of P, and any other statement entailed by it, being true. So, for example, if one can imagine that the moon is made of green cheese, then one can conceive that the moon is made of green cheese, and one can conceive that every statement entailed by this is true. Tidman raises two objections to this account of conceivability. First, he raises Van Inwagen’s objection that the account is too demanding. For one can’t get before one’s mind all of the entailments of any statement, in which case no state of affairs is conceivable for human beings. Relatedly, since anything entails all of the necessary truths, and since there are infinitely many of them, the account entails that nothing is imaginable for us. Second, there is the “Hidden Entailments” problem: One can never rule out that any given putatively imagined state of affairs entails something that is incompatible with that state of affairs obtaining. For example, consider Putnam’s “cat automata” case. Suppose it turns out that cats are really automata. Then this entails that if cats exist, then automata exist. But if so, then if one imagines a scenario according to which cats exist and no automata exist, one unknowingly imagines an impossible state of affairs. To generalize the point and draw the moral: either such cases count as conceiving that p or they don’t. If they do, then the account admits of counterexamples. But if they don’t, then we can never be sure that we have succeeded in imagining a coherent state of affairs, in which case we can never be sure about whether some state of affairs is possible.
Next, Tidman raises two problems that are applicable to all accounts of conceivability as a guide to possibility. First, the Hidden Entailments problem generalizes to a problem for all accounts of conceivability as a guide to possibility, in which case they are all pretty much worthless. Second, he raises what he calls “The Stand-off Problem”: On any account of conceivability, whatever state of affairs p one can apparently conceive, one can almost always, and just as easily, conceive that not-p. But if so, then it appears that they cancel out each other’s epistemic force. But if so, then unless and until one comes up with an account that doesn’t result in conflicting conceivabilities, one must take all conceivability appeals to be epistemically worthless.
In the final portion of the first part of the paper, Tidman argues that one can’t accept the previous criticisms and still hold, without further justification, that conceivability provides prima facie, defeasible justification for modal claims. For it seems that conceivability and possibility are completely different and unrelated subject matters: facts about what we can do with our minds vs. facts about how things could be. But if so, then it’s not obvious what the connection between the two is supposed to be.
In the second part of the paper, Tidman briefly discusses Van Cleve’s account of modal epistemology. The basic ideas behind Van Cleve’s account of modal epistemology are: (i) that modal knowledge is based on intuition, or simply seeing, that certain states of affairs are possible. This is analogous to, or a special case of, our ability to intuit or simply see that certain statements are true, e.g., that 2 + 3 = 5, etc. (ii) Van Cleve distinguishes between weak and strong conceivability. Something is strongly conceivable iff one can intuit or simply see that it is possible, and something is weakly conceivable iff one cannot intuit or simply see that it is impossible. With this distinction in mind, one can avoid objections to conceivability, such as that both the truth and the falsity of Goldbach’s Conjecture is conceivable. One can just say that such cases are instances of weak conceivability, and that only strong conceivability is a guide to possibility. Tidman asserts that the way out of the objections to modal epistemology is to endorse Van Cleve’s modal epistemology – both (i) the core notion that modal knowledge is based on modal intuitions (not imaginability or believability or…), and (ii) the weak and strong conceivability distinction.
1. Introduction
1.1.1. The conceivability thesis: (Hume) whatever is conceivable is possible
1.1.2. The inconceivability thesis: (Hume) whatever is inconceivable is impossible
1.1.3. The inconceivability thesis is implausible on its face, since inconceivability seems largely to be a matter of ignorance and epistemic limitation than a constraint on reality.
1.1.4. So, the focus will be on the conceivability thesis
1.1.5. Appeal to conceivability as a guide to possibility is widespread in philosophy (e.g., Kripke on mind-brain identity; Plantinga and the “It’s possible that I have an alligator body” example…)
1.1.6. Core claims of the paper:
1.1.6.1. No matter how one construes conceivability, it is implausible to think that conceivability entails possiblity.
1.1.6.2. Furthermore, it isn’t even a defeasible source of modal knowledge.
1.1.6.3. However, there is such a thing as modal intuition and it is a reliable guide to possibility.
1.1.7. Two constraints on any adequate account of conceivability:
1.1.7.1. We must be able to readily tell when a state of affairs is adequately conceived.
1.1.7.2. Determining whether a state of affairs is adequately conceived must be knowable prior to knowing whether that state of affairs is possible
1.1.8. These constraints thus tell us that no acceptable account of conceivability entails that conceivability arguments are epistemically circular (in Alston’s sense: can’t defend all of the premises without already relying on the truth of the conclusion)
1.1.9. We will see that many accounts of conceivability violate these constraints: they make conceivability arguments epistemically circular.
2. Mental Images and the “What Counts?” Objection:
2.1. Hume: something is conceivable iff one can form a mental image of it (where the scope of ‘image’ is broad enough to include the whole gamut of phenomenal imagery, such as tastes, sounds, smells, textures, etc.).
2.1.1. Example: I can form a mental image of a golden mountain; so, golden mountains are possible.
2.2. Problem 1: There are many possibilities of which we can’t form a mental image, in which case conceivability has severely limited scope and use.
2.2.1. “In fact” cases: there can be and are colors, sounds, etc. that we cannot see
2.2.2. “In principle” cases: God’s being omnipotent, our having souls, 2+2’s equaling 5, Gold’s having an atomic number of 79…
2.3. Problem 2: The “What Counts?” Objection: the previous objection leads us to wonder: what counts as an image of a state of affairs? We need criteria to determine whether something is a legitimate case of picturing x.
2.3.1. Example: Would picturing an old bearded man on a throne in space, accompanied by the picturing of objects popping into existence at his audible command, count as adequately picturing God’s omnipotence? Presumably not. But if not, then how do we sort out the legitimate cases?
2.3.2. Example: I can picture the refutation of Goldbach’s Conjecture by just associating the modal statement with some scientists standing in front of a computer printout and saying, “I can’t believe it! Goldbach’s Conjecture is false!” Does this count? Presumably not. But if not, then what does count – i.e., what are the criteria of adequate pictureability?
2.3.3. If we demand that the image be perfectly lucid, then we can’t meet this criterion
2.3.3.1. Mental imagery is inherently vague and ill-defined, even in simple cases. Mental imagery inherently lacks detail and determinacy.
2.3.3.2 Mental imagery is inherently limited.
2.3.3.2.1. You can’t imagine, e.g., the whole surface of the earth, let alone very small and simple objects.
2.3.3.2.2. You can’t imagine all of the perspectives/vantage points from which it’s pictureable. But if not, then it’s hard to say which perspective (and what level of magnification) is the adequate one for determining the truth-value of a modal claim in any given case – perhaps we’d see that P is impossible if we were to picture it from some vantage point we neglected to entertain.
2.3.3.3. We can form mental pictures of impossible states of affairs (e.g., Escher’s ever ascending stair); so even if they were perfectly lucid, this still wouldn’t be a sufficient criterion
2.3.4. Objection: such mental pictures don’t represent impossible states of affairs. Rather, they merely look that way from certain perspectives. For example, if we looked closer at our image of Escher’s stairs, we would see that some of the stairs that appear to go up are really going down, thus eliminating the appearance of an impossible state of affairs.
2.3.5. Reply 1: Escher intends the staircase to portray an impossible state of affairs. To view it otherwise is to view it improperly.
2.3.6. Reply 2:
2.3.6.1. Part of what makes an image a representation of something is the intention of the author or the imaginer. For example, consider Wittgenstein’s “duck-rabbit” picture. Whether it represents a duck or a rabbit is not fully determined by the features of the image alone – or perhaps not at all. An agent’s intention is also required to determine this.
2.3.6.2. If so, then it looks as though any image can represent anything. Just tack on an intention to refer to x, and that image now represents x.
2.3.6.3. This results in a dilemma for the mental picture account of conceivability:
2.3.6.3.1. If conceivability is to be construed in terms of pictureability, then either we allow our picturing impossible SOAs to count as picturing or we don’t.
2.3.6.3.2. If we do, then conceivability doesn’t entail possibility.
2.3.6.3.3. But if we don’t, then we can’t tell when we’ve pictured a possible SOA.
2.3.6.3.4. Therefore, if conceivability is to be construed in terms of Pictureability, then either conceivability doesn’t entail possibility, or we can’t tell when we’ve pictured a possible SOA.
2.3.6.4. The root of the “What counts?” objection is the element of intentionality involved in mental pictures. If so, then it’s up to us what the picture represents, in which case it’s arbitrary what they pick out, and there’s a problem of criteria for adequate picturing of a possible SOA. For then it seems that pictures don’t intrinsically refer to any particular thing. And if they don’t intrinsically refer, then it's not at all clear how we are to tell which, of the infinitely many things they can refer to, are the possible SOAs they picture.
2.3.6.5. Notice that the point isn’t that Pictureability is at best a defeasible method of determining possibility; it’s that it’s of no help at all in determining possibility –i.e., we have no way of telling whether what we picture is possible. For pictures are arbitrarily connected to SOAs, and they’re also connected to impossible SOAs. Thus, there’s no relevant epistemic link between the former and the latter.
2.3.7. The upshot of our exploration of the “what counts?” objection is that not only is picturability a fallible guide to possibility, but it provides no justification at all for modal claims.
3. Conceivability as Understandability:
3.1. Reid: something is conceivable iff one can understand the sentence (or proposition or…) that expresses it.
3.2. Objection 1: we can understand necessary falsehoods
3.2.1. This fact is presupposed in utilizing reductio ad absurdum
3.2.2. It’s not as though statements expressing necessary falsehoods are meaningless gibberish (e.g., ‘2 + 2 = 5’)
3.3. Objection 2: inability to understand something is only indicative of our cognitive limitations, not modal facts.
4. Conceivability as Believability:
4.1. Mill: Something is conceivable iff it is possible that it be believed
4.2. Stalnaker: on his account of belief, we are able only to believe what is possible.
4.3. Objection 1: it seems that we can believe the impossible/necessary falsehoods
4.3.1. Identity cases: unaware that a=b:
4.3.1.1. The “Raquel and Kevin” case:
4.3.1.2. Other necessary truth/falsehood cases: The “I can believe that God does/doesn’t exist (assume it turns out that God is either metaphysically necessary or impossible)” case:
4.3.2. Objection 2:
4.3.2.1. But suppose a good case can be made that one can only believe what is possible.
4.3.2.2. Still, it seems like we sometimes believe impossible things (e.g., Hesperus is distinct from Phosphorus).
4.3.2.3. But if we insist that we really don’t believe them, then this must mean that we can’t tell whether we believe certain things.
5. Conceivability as Entertainability:
5.1. D. Lewis, R. Adams, R. Swinburne: something is conceivable iff it is entertainable
5.1.1. Something is entertainable iff it is imaginable
5.1.2. P is imaginable iff one can tell a coherent story according to which P is true.
5.1.3. Swinburne’s is the strongest: P is coherent iff one can conceive of P, and any other statement entailed by it, being true.
5.1.3.1. The “moon made of green cheese” example: one can conceive that the moon is made of green cheese, and one can conceive that every statement entailed by this, is true.
5.2. Objection 1: Van Inwagen: too demanding: one can’t get before one’s mind all of the entailments of any statement, in which case no state of affairs is imaginable for us (requires omniscience to imagine something).
5.2.1. A slightly different point: since anything entails all of the necessary truths, and since there are infinitely many of these, again, the account entails that nothing is imaginable, and so, not entertainable, for us.
5.3. Objection 2: The “Hidden Entailments” problem:
5.3.1. One can’t rule out that A entails B (and perhaps vice-versa).
5.3.2. But if not, then if one imagines A without B, then one will imagine what is impossible.
5.3.2.1. Putnam’s “Cat automata” case: If it turns out that cats are really automata, then the existence of cats entails the existence of automata. But if one can imagine a world with cats but not automata, then one imagines the impossible.
5.3.3. But if so, then since one can never rule out hidden entailments like this, then one can never be sure that one has succeeded in imagining anything.
5.3.3.1. We can see the epistemic circularity problem raising its head in this case: need to rely on prior belief that certain statements are possibly true in order to say that something is imaginable/entertainable.
5.3.4. The “hidden entailments” problem spreads to all accounts of conceivability
5.3.4.1. The Hesperus/Phosphorus case:
5.3.4.2. How do we know such hidden entailments don’t apply to, e.g., Kripke’s “brainless pains” case?
6. The Stand-Off Problem:
6.1.1. On any account of conceivability, whatever state of affairs p one can apparently conceive, one can almost always, and just as easily, conceive that not-p.
6.1.2. But if so, then it appears that they cancel out each other’s epistemic force.
6.1.3. But if so, then unless and until one comes up with an account that doesn’t result in conflicting conceivabilities, one must take all conceivability appeals to be epistemically worthless.
7. Conceivability as a Defeasible Indicator of Possibility: Before one can hold to the weaker claim that conceivability is a defeasible guide to possibility, one must first give a reason to believe that there is any connection at all between conceivability and possibility
7.1.1. They seem to be of entirely different subject matters: facts about what we can do with our minds vs. facts about how things could be.
7.1.2. But if so, then it’s not obvious what the connection between the two is supposed to be.
8. Strong vs. Weak Conceivability: Tidman briefly discusses Van Cleve’s account of modal epistemology.
8.1. The basic contours of Van Cleve’s account of modal epistemology:
8.1.1. Modal knowledge is based on intuition, or simply seeing, that certain states of affairs are possible. This is analogous to, or a special case of, our ability to intuit or simply see that certain statements are true, e.g., that 2 + 3 = 5, etc. (
8.1.2. Weak vs. strong conceivability.
8.1.2.1. Something is strongly conceivable iff one can intuit or simply see that it is possible
8.1.2.2. Something is weakly conceivable iff one cannot intuit or simply see that it is impossible.
8.1.2.3. With this distinction in mind, one can avoid objections to conceivability, such as that both the truth and the falsity of Goldbach’s Conjecture is conceivable.
8.1.2.4. One can just say that such cases are instances of weak conceivability, and that only strong conceivability is a guide to possibility.
8.2. Tidman asserts that the way out of the objections to modal epistemology is to endorse Van Cleve’s modal epistemology – both (i) the core notion that modal knowledge is based on modal intuitions (not imaginability or believability or…), and (ii) the weak and strong conceivability distinction.
Comments
Post a Comment