Theory of Knowledge – Fall 2006

Reliabilism

A. Set Up

There are lots of ways that reliability could enter into an epistemological theory:

(1)       If you discover your method in coming to believe P is unreliable, that may defeat your justification for P. Evidentialists could accept this thesis.

(2)       You could include some sort of reliability condition as the fourth condition in the TAK, to try to deal with the Getter cases.

(3)       You could scrap the 3^rd and 4^th conditions altogether, and completely replace the idea with justification with reliability. This is similar to what Goldman did with the causal theory we looked at last week.

(4)       You could also keep condition 3, and use reliability as a theory about what that condition amounts to. That is, you could say that reliability is what makes a belief justified. This is the view Goldman defends in ‘What is Justified Belief?” and it is the view we’ll be looking at here.

Goldman wants to give a theory of justification: what is it for a belief to be justified? And, he wants to provide an analysis, not a mere synonym.

Question: what is the difference, and why would Goldman want to give an analysis?

Answer: he wants to explain what it is for a belief to be justified, etc.

Question: he says that certain terms can appear in his analysis of justified belief and others can’t. Why does he include and exclude things as he does?

Answer: discuss descriptive v. evaluative properties, supervenience, etc.

B. Crude Reliabilism (R1)

Question: what descriptive property does Goldman think that justification supervenes on?

R:        S justifiably believes p iff: S’s belief in p is caused by a reliable process.

Discuss: which sorts of processes are reliable, and why. Reliable processes are ones with a good ratio of true to false beliefs

Discuss various processes. (1) visual perception of a nearby middle-sized object. (2) hasty glimpsing. (3) close careful visual and tactile inspection of a close-up object.

Question: how well do our assessments of reliability and justification correspond here? That is, discuss that the more reliable the process, the more justified the belief seems to be. So, the theory seems intuitively plausible.

Question: why isn’t that view good enough?

Discuss: the example using reasoning from false premises.

Discuss: conditional reliability, belief independent and belief dependent processes

C. Sophisticated Reliabilism (R*)

R*       (i)        If S’s belief in p at t results from a belief-independent process that is reliable, then S’s belief in p at t is justified.

            (ii)       If S’s belief in p at t results from a belief-dependent process that is  conditionally reliable, and the beliefs the process operates on are     themselves justified, then S’s belief in p at t is justified.

            (iii)      The only way beliefs can be justified is by satisfying the conditions in (i)

Note: this is similar to MF, but it differs from MF in that it replaces talk of reasons and evidence with talk of reliability. And, it may provide an answer to what justifies the “basic” beliefs: they are the result of a reliable process.

D. Objection #1: The Demon World / BIV

Note: this is not a question about skepticism. We are not now asking “how do you know you are not a BIV.” That question comes later. To distinguish the two arguments, I’ll do the demon version of this one.

Being formed in a reliable way, is not necessary for justification. That is, there can be justification without reliability.

Imagine a person who is a molecule for molecule duplicate of you. But, instead of receiving data from the “real world,” he is constantly being manipulated by an evil demon. All of his experiences are controlled by that demon. It seems like your processes are very reliable, but your duplicate’s aren’t. So, that would have the result that you are justified but the he isn’t . . . even though you have the very same (internal) evidence. But, that doesn’t seem right. You are justified in believing you are in a classroom, and it seems like your twin is justified in believing the exact same thing.

We can formulate this as an argument:

1.         If reliabilism is a correct account of justification, then the victim of the demon has almost no justified beliefs.

2.         But the victim of the demon has lots of justified beliefs. 

3.         Therefore, reliabilism is not a correct account of justification.

How can we support P2? Remember the Same Evidence Principle . . . the demon victim’s “epistemic situation” is the same as ours. So, his beliefs are as well justified as ours are.