Underdetermination Arguments for Skepticism

I think Rich Feldman thinks they're the most powerful kind of argument and Duncan Pritchard makes crucial use of them in _Epistemic Luck_.

Duncan argues that appreciating the force of underdetermination arguments leads one to see that no externalist condition on knowledge truly engage the skeptical challenge. I agree with Rich and Duncan on the significance of this kind of argument, but I just wanted to point out that Brian Weatherson has strongly demurred. (see discussion on TAR here)

I think Carrie Jenkins and Matt Wiener put him some tough questions. I don't know if Brian has anything like a worked-out theory of evidence or if he endorses a specific theory, but I'm very interested in the nature of evidence and would like to explore the plausibility of non-phenomenal theories of evidence.

Weatherson says in the comments on that discussion

"if evidence is everything we have to go on, then why shouldn't all the causal influences on me count as things I have to go on?"

It would be odd if every event in my backward light-cone counts as my evidence, so surely he means to restrict to certain kinds of causal influences. This also seems to indicate a non-propositional theory of evidence since, I would assume he thinks, events are the relata of the causal relation. But perhaps he agrees with Mellor that facts are the causal relata.

One common type of theory of evidence is a doxastic theory like that of Swinburne who holds that our evidence consists in the contents of our basic beliefs. All this suggests something like the following first-stab at a taxonomy of theories of evidence.

I. Propositional Theories
A. Doxastic Theories (Swinburne)
B. Fact-based Causal Theories (Fumerton?)
II. Non-propositional Theories
A. Phenomenal Theories of Evidence (Chisholm, Feldman?)
B. Event-based Causal Theories (Weatherson?)

Perhaps direct acquaintance theories of evidence will be of type IB.

[Update: While reading on the Sellers Dilemma yesterday it struck me that the taxonomy must make clear how the evidential theories of McDowell and Williamson relate to the others.]

[Update update: I got onto the reading on McDowell and Williamson after reading Sect. 9.2 of Duncan's Luck book and posted the update (apparently) after he'd posted the comment. So Duncan managed to overdetermine my considering the point all by himself!]