Monday, October 19, 2009

The Epistemic Role of Suggestive Evidence

This'll be my last post on this particular subject for a while, since this project has to go on the back burner for a while as certain other deadlines loom.

As Eric Winsberg says in reply to my first post,

suggestion is a causal relation, support is not. Whether E suggested H is a matter of historical fact indexed to a particular time. whether A supports B is a very different kind of fact; one which might vary over time, etc.


A dream or a blow to the head might very well suggest a hypothesis, but they are not evidence. the kinds of examples where we are inclined to call the thing that suggests a hypothesis "evidence" are the ones where the thing also supports the hypothesis.

In other words, insofar as suggestion is independent of support, it is purely a causal relationship. Now, I'm disinclined to admit a complete epistemic/causal dichotomy, as I think it is well-established that we rely on all kinds of causal relations in our epistemology. But I also think that suggestion has clear epistemic features independent of actual causal events. As I said above, suggestion is an ability in scientists that is trained, it is something we evaluate agents as being better and worse at. Making good suggestions is a kind of agential epistemic virtue. It's also not entirely the case that "Whether E suggested H is a matter of historical fact indexed to a particular time." Whether E is taken to suggest H is such a historical fact, but so is whether it is taken to support H. On the other hand, we can, looking back at the historical case, verify or evaluate whether E suggests H (this is probably clear to anyone who has had that "detective story" experience when reading the history of science, coming up with the hypothesis "before" the scientist), just as we can do with support. We can also identify alternative hypotheses the evidence suggests that went unrecognized.

Now, I won't deny that it is possible in some degree to re-construct many cases of suggestion as cases of support; rational reconstruction can be a powerful tool for fitting square pegs in round holes. However, I would argue that doing so is not only fairly unilluminating for understanding the practice of science, but also that it provides succor to skeptical arguments. It is my eventual hope to argue that it is precisely such reductive, mono-functional, uni-direcitonal accounts of evidence as support which are responsible for skeptical worries about, e.g., theory-ladenness and the experimenter's regress. Such skeptical worries can (and are) ignored, but the common refrain that skeptical problems are insoluble and irrelevant ignores the fact that the problem is internal to these common accounts of evidence. Compelling versions of anti-skepticism have always been accompanied by alternative theories of evidence in which the "E supports H" relation is supplemented by other epistemic functions. (e.g., Quine's view that treats evidence-statements and theoretical statements as symmetrically related and judged on their functional fitness in accommodating experience, for all its problems, provides an internal reason for saying that skeptical worries ask too much).

1 comment:

  1. I look forward to seeing how the idea develops. Thanks for the posts. I think the core of disagreement can be seen in a sentence like this:

    "Whether E is taken to suggest H is such a historical fact, but so is whether it is taken to support H."

    I think support, unlike suggestion, is independent of how it is "taken" (that's what I take the epistemic/causal distinction to mark.) It is very difficult to see how suggestion is something other than support except by saying that suggestion is "historical" (or existential) and support is "reconstructed" (or rational).

    But perhaps a kind of meeting of minds is possible in this formula:

    "E suggests H" means "E is taken to support H". This would leave open the question of whether E actually (i.e., logically) supports H and therefore also the possibility of E suggesting H without supporting it. It would, however, retain the close (mono-functional) connection between evidence and support since "E is taken to support H" would mean "E is taken as evidence for H". E, as such, could not be considered evidence prior to its analysis as supportive and/or suggestive.

    That is, the claim, "E is evidence for H" would not be in play.