On Carrier's epistemology, part 1

When I reviewed Sense and Goodness Without God a month ago, I indicated posts would be forthcomming analyzing the ideas presented in the book. Here's finally getting around to writing the first of those posts. I apologize if I lose anyone here; this may be the philosophically heaviest post I've ever written.

The first major section of Carrier's book (following the introduction) deals with epistemology. It does not discuss radical skepticism in great detail, but seems to contain some rather large concessions to skepticism. On pages 27-28, Carrier says: "the same predictions can be made by wildly different claims, and we have the tough task, through all our lives and in everything we do, of trying to figure out which of several equally plausible of a particular thing is right." A little later, on page 32, he brings up the world of The Matrix, which could be taken as an example of this problem. Carrier argues that there would only be a meaningful difference between such a world and the "real" world if it were not "in every way the same as real life," for example, if it had "glitches, superhuman 'agents,' and groups who can 'wake you up" to the higher reality." However, it is possible to imagine that I am in the matrix-world and that such things exist but I haven't noticed them yet. This is indeed a case of "wildly different claims" making "the same predictions," so how do we figure out which is right?

When Carrier discusses how to do this, he runs into the problem of regress, which he confronts when attempting to rebutt Alvin Plantinga's epistemology. The problem is that if one proposition must be backed up with another, attempts to justify beliefs will result in an infinite chain of propisitions without any basis in anything. If I understand Plantinga correctly, his alternative that there are "properly basic beliefs" which one may hold without any evidence at all, though this does not mean that they cannot be overturned by strong evidence. Plantinga is a Christian philosopher particularly concerned with showing belief in God is properly basic, though I take it that he would not agree with William Lane Craig's position that one ought to believe in Christianity regardless of how strong the contrary evidence is.

Carrier's response to the problem of regress is that "The buck stops with the evidence: which means experience, for there is no other sort of evidence" (p. 45). This response hinges on a subtle distinction between the brute experience and even the simplest statements about experience, such as "there is an experience of me typing now" (p. 46). Carrier thinks that the brute experience of typing is indubitable, but the preceeding statement is not, because it could involve some conceptual confusion. I am inclined to agree.

However, merely having some brute experience to base beliefs on does not solve the problem of regress, because brute experience needs to be interpreted. Any statement about how we ought to go about interpreting experience is open to question, and thus needs further justification, and those justifications demand justification and so on. I do not think that Carrier has escaped the problem of regress.

This point becomes clearer with a few more citations from the book. First, "the only thing we can trust without evidence is what cannot be denied, and the only thing we cannot deny is that certain experiences exist" (p. 45). From this, we can clearly deny the reliability of memory. However, as I pointed out in an early post on skepticism, no line of reasoning can ever get us out of skepticism regarding memory, because in order to reason we must be able to remember the previous steps of the line of reasoning.

A similar problem arises when Carrier tries to formulate specific proposals about what good methodology works. At the beginning of this section, Carrier says, "we need a 'theory of knowledge.' But how do we know ours is correct?... The real test will be its results in practice" (p. 24). Again, when it comes time to list the most reliable methods, he says, "It is reasonable to predict that an accurate method... will exhibit two particular features, which an inaccurate method will not exhibit: predictive sucess and convergend accumulation of consistent results." But how do we know if the results are good? How do we know if predictions are suceeding? We must appeal to facts that are regarded as clearly true. But if one is truly dedicated to considering something a fact when it comes out of rigorous methodology... we are back at the problem of regress.

I have other things I wish to get done tonight, but for now I will say that I think Plantiga has hit upon the only real solution to the problem of regress. Readers of this blog will of course want to know what I think of his attempt to use properly basic belief to defend belief in God; that I will deal with in the next post in this series.