Monday, August 16, 2010


As Brian Leiter reported The New Times has recruited Timothy Williamson for its online blog, the Stone. In a recent entry (perhaps his first?) he writes about the role of the imagination in science:

The main point of the entry is revealed in its closing paragraph. It is to answer unnamed "Critics of contemporary philosophy" who "sometimes complain that in using thought experiments it loses touch with reality...Once imagining is recognized as a normal means of learning, contemporary philosophers’ use of such techniques can be seen as just extraordinarily systematic and persistent applications of our ordinary cognitive apparatus."

I offer four observations:
1. First, Williamson makes it easy on himself by simply asserting without evidence that contemporary philosophers’ use of imagination can be seen as just extraordinarily systematic and persistent applications of our ordinary cognitive apparatus. The blog clearly implies that if the imagination is good enough for science it is good enough for philosophy. But Williamson makes no effort to show that contemporary philosophers systematically constrain the use of the imagination in the manner that scientists (perhaps?) do. He just asserts philosophers' systematicity and persistence. (The piece ends a line later.) This is an argument from authority.

2. Nevertheless, my reason for blogging about this entry is not to continue to harping about the tendency of leading analytic philosophers to claim the mantle of science when it suits them. Rather, it is to note the surprising (to me!) impact of recent (well, post-Kuhnian!) history and philosophy of science on Williamson's thought in at least two ways. First, Williamson takes the context of discovery very seriously. It is what grounds his appeal to the authority and use of the imagination. Second, he asserts that even in the context of justification the imagination plays a very important role, and this is a good thing.

3. So, perhaps philosophers of science can engage Williamson on these two previous points in constructive fashion? The recent methodological turn of my leading (and young) analytic metaphysicians should be an opportunity in this respect.

4. I end with a historical note. Williamson's position is a rediscovery of David Hume's and especially his friend's Adam Smith's understanding of science. In Smith's "The History of Astronomy," the imagination plays a positive constructive and justificatory role in natural science and philosophy: "Philosophy, therefore, may be regarded as one of those arts which address themselves to the imagination." As Smith writes, "For, though it is the end of Philosophy, to allay that wonder, which either the unusual or seemingly disjointed appearances of nature excite, yet she never
triumphs so much, as when, in order to connect together a few, in themselves,
perhaps, inconsiderable objects, she has, if I may so, created another
constitution of things, more easily attended to, but more new, more contrary
to common opinion and expectation, than any of those appearances themselves."
(IV.33, 75)


  1. Maybe I'm missing something, but I'm not exactly sure how Williamson supports his claim that "imagination plays a vital role in justifying ideas". His support would seem to be his claim that "we must develop the consequences of competing hypotheses with disciplined imagination in order to compare them with the available evidence." My take on this is that the justificatory work is done by the comparison with the available evidence. Drawing out consequences is necessary for justification, but those consequences, on their own, don't seem to be doing any justificatory work. If this is all Williamson has in mind, then I agree, but I got the sense that he's driving toward a stronger conclusion here.

    It may well be that the imagination plays an important role in the context of justification (maybe for figuring out whether the available evidence confirms a hypothesis?). But it doesn't seem to me that Williamson has established (what I took to be) his point.

  2. Jonathon (if I may?),
    I agree with you that Williamson's argument is a bit fast. (But this is the blogosphere!) His claims seems to be that the imagination has roles to play that are a necessary condition for the very possibility of justification. Surely we can design an example in which it requires real imagination to work out the relevant regressors or alternative hypothesis (or even null hypothesis), etc.

  3. I also think that taking thought experiments seriously needn't involve considering the context of discovery at all. Thought experiments can be confined to the role of testing hypotheses. (The kind of 'imagination' going on here may very well be grounded in experience; I think of what Mach says, for instance, in his _Science of Mechanics_. I think of Stevinus's great thought experiment.)

    As for whether imagination plays a role in the context of justification, it is hard to see why it ought to on a simple probabilistic model of confirmation or corroboration. One presumably will have a view, for instance, on the value of P(e,hb)-P(e,b) - or more convincingly, on how to rank competing theories by applying such a measure - without the need for any significant imaginative acts!