Updating, undermining, and perceptual learning
- 280 Downloads
As I head home from work, I’m not sure whether my daughter’s new bike is green, and I’m also not sure whether I’m on drugs that distort my color perception. One thing that I am sure about is that my attitudes towards those possibilities are evidentially independent of one another, in the sense that changing my confidence in one shouldn’t affect my confidence in the other. When I get home and see the bike it looks green, so I increase my confidence that it is green. But something else has changed: now an increase in my confidence that I’m on color-drugs would undermine my confidence that the bike is green. Jonathan Weisberg and Jim Pryor argue that the preceding story is problematic for standard Bayesian accounts of perceptual learning. Due to the ‘rigidity’ of Conditionalization, a negative probabilistic correlation between two propositions cannot be introduced by updating on one of them. Hence if my beliefs about my own color-sobriety start out independent of my beliefs about the color of the bike, then they must remain independent after I have my perceptual experience and update accordingly. Weisberg takes this to be a reason to reject Conditionalization. I argue that this conclusion is too pessimistic: Conditionalization is only part of the Bayesian story of perceptual learning, and the other part needn’t preserve independence. Hence Bayesian accounts of perceptual learning are perfectly consistent with potential underminers for perceptual beliefs.
KeywordsBayesianism Credences Undermining defeat Epistemology Rigidity
Many thanks to Josh Dever, Sinan Dogramaci, Jim Pryor, Miriam Schoenfield, David Sosa, Jonathan Weisberg, audiences at the University of Texas at Austin and the Arché Centre at the University of St Andrews, and an anonymous referee.
- Davidson, D. (1986). A coherence theory of truth and knowledge. In E. LePore (Ed.), Truth and interpretation. Perspectives on the philosophy of Donald Davidson (pp. 307–319). Oxford: Basil Blackwell.Google Scholar
- Goodman, N. (1983). Fact, fiction & forecast. Cambridge: Harvard University Press.Google Scholar
- Howson, C., & Urbach, P. (1993). Scientific reasoning: The Bayesian approach. Open Court.Google Scholar
- Jeffrey, R. C. (1983). The logic of decision. Chicago: University of Chicago Press.Google Scholar
- Jeffrey, R. C. (1992). Bayesianism with a Human Face. In Probability and the art of judgment, (pp. 77–107). Cambridge: Cambridge University Press.Google Scholar
- Pryor, J. (2013). Problems for credulism. In C. Tucker (Ed.), Seemings and justification: New essays on dogmatism and phenomenal conservatism. oxford: Oxford University Press.Google Scholar
- Williamson, T. (2000). Knowledge and its limits. Oxford: Oxford University Press.Google Scholar