First of what might be a few posts on “Bill Lycan’s paper”:http://www.unc.edu/%7Eujanel/Gettier.htm. I don’t agree with the following claim.

bq. It is well to remind ourselves that no effort of analytic philosophy to provide strictly necessary and sufficient conditions for a philosophically interesting concept has ever succeeded.

First a philosophical comment, then some examples.

To succeed at a project, you don’t usually need to be widely recognised. If I set myself a project of watching baseball for fourteen hours straight (with the help of my DVR and multi-channel feed) and I start watching at 3 and finish at 5 the next morning, I’ve succeeded, even if no one else believes that I really did that. So I think that to succeed at giving an analysis just is to give the correct definition, and _maybe_ to motivate it. Widespread acceptance is not the issue. Having said that, here are some cases of what I think are successful analyses.

_Supervenience_: The work starting with Jaegwon Kim I think did a very good, and successful, job of cleaning up what was a bit of an intuitive mess to start with.

_Vagueness_: Here everyone seems to think the analysis of what it is to be vague in terms of borderline cases is correct – well everyone except all the experts. In this case I think the experts are wrong and the conventional wisdom is correct, but if that’s wrong I suspect one of the alternative proposed analyses is correct.

_Consistency_: I think model-theoretic definitions of consistency are basically successful at clarifying a pre-technical notion of consistency. Now there are disputes about the details here, but I think we’ve basically got it right, at least as an account of logical consistency.

_Counterfactuals_: The nearby possible worlds analysis of counterfactuals is very widely accepted, and may even be correct. To be sure, this isn’t an example of analysis into uncontroversial parts, but neither was the JTB analysis. (And to be fair, Bill Lycan has an entire book on why this one is wrong, so it’s more than a little question-begging to use it against him.)

_Possibility_: And I think the analysis of _X is possible_ as _X is true at a possible world_, or in certain contexts as _X is true at an accessible possible world_ is correct. Again, it isn’t maximally reductive, but I don’t think that’s a criteria on being an analysis.

_Computability_: As Turing computability.

And though there’s no single term it is a smooth analysis of, we shouldn’t ignore the success of the probabilist tradition at giving an account of consistency for partial belief.

Many or all of these are controversial, but what’s controversial is whether or not anyone has succeeded. Which is to say that it’s controversial (at worst) whether we’ve ever produced a successful analysis.