I’ve been thinking about Plantinga’s proper functionalism, and today am perplexed by his latest account of defeat. It looks like classic fodder for those sensitive to the conditional fallacy, but I won’t accuse it of that here (Keith is right that without a specific characterization of that fallacy, the label comes to mean something like “subject to counterexamples that can be constructed relatively easily”).
Here’s Plantinga’s account of defeat:
D is a purely epistemic defeater of B for S at t iff
1. S’s noetic structure N at t includes B and S comes to believe D at t,
2. any person S*
a. whose cognitive faculties are functioning properly in the relevant respects, b. who is such that the bit of the design plan governing the sustaining of B in her noetic structure is successfully aimed at truth . . . and nothing more c. whose noetic structure is N and includes B, and d. who comes to believe D but nothing else independent of or stronger than D, would withhold B (or believe it less strongly). (WCB, p. 363)
Later, Al says something incompatible with the above. He says,
Still, argument is one way to give me a defeater. Is there another way? Yes; you can put me in a position where where I have experiences such that, given those experiences (and given my noetic structure), the rational thing to do is to give up the purported defeatee.” (WCB, p. 367).
This quote calls for a revision of the official account, since it makes no sense to say, where D is an experience, one comes to believe D (as clause d. requires). The required change is obvious though. Just change the beginning of d. to read “one who comes to experience D or believe D…,” and change clause 1. in a similar fashion.
There is one other minor problem here, and a major one. The minor problem has to do with self-locating or indexical beliefs. For example, it is part of my noetic structure that I believe I’m JK and not anyone else, that I’m married to Carol, and that I have two kids. I live in Columbia, MO, my parents are Kenneth and Alice, my ancestors trace to Tore Kvanvig, a kvann farmer on the west coast of Norway near the hamlet of Kvanvik, 17 miles west of Kristiansand, born in 1440 (kvann was a plant used in medicines, and kvann farms are mentioned, e.g., Magnus Lagaboter’s law, the Norwegian king from 1263-1280).
But enough about me. The point is that Plantinga’s account of defeat quantifies over individuals with precisely the same noetic structure as this! I don’t think it is possible to have all of these beliefs and still be sane without actually being me–at the very least, one won’t have such beliefs be the result of a design plan successfully aimed at truth without being me. I assume we should take the iff connective to signal a necessary connection here (if we read it more weakly, the counterexamples are too obvious to mention). So, the above account, in effect, quantifies over all possible individuals (I’ll leave this metaphysically problematic characterization as is; feel free to inoculate the description as desired), the relevant subset for the account being the possible me’s.
So, when we think about defeaters for self-locating beliefs, the best we’ll be able to do is to quantify over possible me’s and ask what response it is possible for me to have to D. Is this a broad enough, representative class on which to base an account of defeat? It looks too narrow: the above account seems to want to quantify over a broad and varied range of individuals, so maybe some change is required here.
If we want a broader and more varied reference class for indexical beliefs, maybe we can purge the indexicality from the beliefs by taking beliefs to be a triadic relation between a person, a proposition, and some mode of access or means of presentation of that proposition. Then when I believe I’m tired, the propositional core can be a non-self-locating claim to the effect that JK is tired. The self-locating part comes in when we consider the mode of access or the means of presentation. This move has the cost of committing the account to a defense of a triadic theory of intentional states, but since I think there are good reasons for endorsing such an account, I won’t complain. So suppose we take a noetic structure to be simply the system of propositional contents plus the particular mental modalities involved, such as beliefs, presuppositions, assumptions, and whatever else we might find here.
Regardless of what a defender of the above account wants to say about this issue, there is a deeper problem. It is the Quine/Duhem problem. Here’s the lesson: if you’re willing to make enough changes elsewhere in your system of beliefs, almost any belief can be maintained in the face of new information. So, when we quantify over all individuals with the same noetic structure, it will be rare indeed when the universally quantified formula is true, given this Q/D result. Some properly functioning systems will drop B and others will make other necessary adjustments to the noetic structure.
My inclination is to think that we should begin anew rather than continue to try to patch up Plantinga’s approach. The new beginning will not begin with noetic structures and beliefs within them, but rather with propositions and what is evidence for them. Then we have “e is evidence for p” and the fundamental form of defeat is where the conjunction of e and the defeater is not evidence for p (I ignore partial defeat here). Then, when put into the doxastic context, you can add a defeater, and yet not lose justification, because your body of total evidence includes a restorer (and nothing else). Moreover, the Q/D problem goes away, for even if d is a defeater of the e/p relation, it need not be a defeater of the e/p&r relation. That allows rational adjustments to a system of beliefs in response to learning d that don’t require abandoning p. All learning d requires (on the assumption that there are no restorers present) is that some evidentially suitable adjustment is made, one of which is abandoning p, but not the only one. This way of proceeding is Pollock’s (though I don’t recall him mentioning the Q/D issue and I’m not confident his own approach takes this issue into account, but this is sheer recall-based, so I could easily be wrong), and it accords well with my own claims that propositional justification is the fundamental form, with doxastic justification derivative on it. From this perspective, the fundamental problem with the above account is that it begins by talking about B as a belief, so that you can’t acquire defeaters for propositions that your total evidence confirmed prior to acquiring the defeater, but where you did not believe said proposition.
But maybe Plantinga’s account can be salvaged? I suspect that the account can’t be salvaged, but my inclination here is mitigated somewhat by my recognition that I think the doxasticism inherent in this approach is completely wrongheaded and hence that I’m not as motivated to look for ways out as others will be who are more sympathetic to such an approach…