Theory of Knowledge
Fall 2006
Coherentism
I. The Basic Idea of Coherentism
There are no beliefs that are foundational or basic. Every justified belief is justified in virtue of its relations to other beliefs. So, this rejects the step of the regress argument that rejects circular justification. But the idea is not that simple circular reasoning is ok. Rather, a belief is justified in virtue of the way its fits with all the believer’s other beliefs.
Three main sorts of reason for turning to coherentism are: a) the belief that there is no decent way to fix up foundationalism to deal with the problems we talked about; b) the idea that only beliefs can justify other beliefs (we’ll talk about this next week); c) the idea that all our justified beliefs do seem to fit together with at least some other beliefs and that unjustified, at least typically, do fail to fit with some others.
Go over examples from text.
II. Simple Formulations of the Coherence Theory and Questions About It
CT. S is justified in believing p iff p coheres with S's system of beliefs
Two questions for coherentists:
C1) What counts as S's system of beliefs?
C2) Under what conditions does a belief coherence with a system of beliefs - what’s coherence?
Consider (CT1) from text. We can prove, for any arbitrarily selected belief, that it is justified:
1. S believes p [Assume]
2. So, p is in S’s system of beliefs [From (1), def. of “system of beliefs”]
3. So, p follows from S’s system of beliefs [(2), simplification]
4. So, S is justified in believing p [(3), CT1]
But this is absurd. Every belief, no matter how preposterous, is justified according to this theory.
Consider (CT2) from text.
1. (CT2) [Assume]
2. Either (a) S’s system of beliefs is coherent or (b) S’s system of beliefs is not coherent.
3. If (a), then for all p, if S believes p then S is justified in believing p (i.e., if (a), then all S’s
beliefs are justified.) [1]
4. If (b), then for all p, if S believes p, then S is not justified in believing p (i.e., if (b) then none
of S’s beliefs is justified. [2]
5. Either all S’s beliefs are justified or none of S’s beliefs is justified. [(2), (3), (4)]
But this result is absurd. For any person, some beliefs are justified and some are not. [Remember, that’s what the standard view says.] So (CT2) must be wrong. This is the “All or None Objection”.
Simple formulations don’t seem to work very well.
III. A Version of Coherentism
Consider next (CT3) from the text. We’ll proceed as if we have some idea what coherence is.
See text for problems of detail:
1) dropping individual propositions will always lose some coherence, given other things a person believes. So it’s not obvious that a system is ever made more coherent simply by eliminating one belief.
2) whenever there is, intuitively, a conflict between 2 beliefs, you can get an increase in coherence by dropping either one. So, by CT3, neither is justified. But sometimes there is one bad guy among the 2.
An issue to think about: there are interesting real life cases of people who have beliefs that are like the hair-loss case: the imposter syndrome, paranoia. In each case, people single out themselves for beliefs that make them exceptions to general principles that they otherwise accept. Must there be something incoherent about their systems?
IV. BonJour’s Theory
See his article in the anthology. BonJour says that coherentism does not really rely on circular justification. Rather, it takes justification to be “holistic” and “non-linear”. See p. 391 for statement of main claim - “beliefs are justified by being inferentially related to other beliefs in the overall context of a coherent system.”
In BonJour’s spelling out of his theory, some of which is not included in our book, he makes a number of points worth considering. First, he distinguishes a belief’s being arrived at by inference and its being justified by inference. Roughly, he thinks that cognitively spontaneous beliefs about the world, e.g., “that’s a book” are not arrived at by inference. But they are justified by inferences of the sort stated on p. 393-4. So here he seems to have dropped some of the stuff we found worrisome about cartesian foundationalism - there aren’t all these introspective beliefs.
Maybe also we are supposed to think of one’s system of beliefs as including things such as the one’s in his example on 393-4 which one does not consciously consider but would believe if one thought about it (or something like that).
In a section not included in our book, he says a number of sensible things about what goes into coherence:
1 - logical consistency
2 - probabilistic consistency. (Discuss probabilistic inconsistency a little - just believing that an unlikely thing has happened is not unreasonable.)
3 - inferential connections
4 - lack of unconnected subsystems
5 - lack of unexplained anomalies
This seems to fall victim to the objection to (C2) - the “all or none” objection. But CT3 can make use of this account of coherence to help explain the view. It avoids this objection.
V. Traditional Objections to Coherentism
Let’s pretend we understand what it is for a system of beliefs to be coherent and for a belief to cohere with such a system. We’ll pretty much accept the intuitive judgment coherentists would make about examples
A. The Alternative Systems Objection
This is a common objection to coherentism. See quotes from BonJour in text, and others from in Earl Conee’s paper in Phil. Topics, 1995.
See BonJour’s book, p. 107 for his initial discussion of the objection. He suggests that one can see the idea by negating everything. But this is wrong - this won’t produce a consistent system.
But it is true that there are many different coherent systems possible. See example in text.
The Alternatives Systems Objection - 1st Version
1. There can be many different and incompatible internally coherent systems of belief.
2. If coherentism were true, then all these systems (or all the beliefs in these systems) would be
justified. [(1), def. of coherentism]
3. But it is not the case that all these different systems (or all the beliefs in them) can be justified.
4. Coherentism is not true. (2), (3)
[An alternative formulation emphasizes the idea that we can’t identify, on coherentist grounds, which one of these systems is justified.]
But (3) is false. Suppose some 8th century monk had a coherent set of beliefs. Suppose one you does as well. These are coherent systems. They are inconsistent with one another. There’s nothing wrong with the supposition that these systems, or the beliefs in them, are justified. Alternative systems can be justified.
Notice that this formulation just talks about what “can be justified”. You might think that my point about different people having different justified systems misses the point. The idea is that all these different systems are justified for an individual. But that’s not right. What’s justified for you depends upon coherence with your system of beliefs.
There’s another way to formulate the objection. See text, p. 67. Suppose S believes p. It’s something he really cares about. He can, then, pick and choose his other beliefs to make a coherent set of beliefs around it. If he succeeds, then that belief is justified. But that seems wrong. The Alternative Systems Argument - 2nd Version - see text.
A common reply to this: (2) is false. We don’t have that much control over our beliefs. That’s probably right. Also, it’s probably beside the point. Suppose someone did have the relevant kind of control over his beliefs. It’s not entirely clear that (4) is true. Earl Conee (from whom almost the entire discussion of this objection has been stolen) writes:
It is not a flaw in a theory to allow that someone who has some sufficiently pervasive control over his or her own cognitive life (including the power to forget how the resulting cognitive situation came about) might have reasonably believed virtually anything. It cannot be credibly denied that, as a result of various possible renovations in our cognitive condition, we might have quite reasonably believed ourselves to be made out of glass, or to be identical to Napoleon, etc. (“Isolation and Beyond,” Philosophical Topics 23, 1996.)
I’m not entirely convinced by this reply. But the reason to think this might not be a good reply is pretty close to what seems right about the next objection to coherentism. So we’ll consider that objection rather than pursue this one.
B. The Isolation Objection
The key idea of coherentism seems to be that justification depends only on other beliefs. If just beliefs do the justifying, then experiences seem not to matter. And that’s not right. Reconsider the example earlier about the car. Suppose I just add beliefs about how the crunching sound was right only for the junk car, the junk car was right under a branch, etc. Unless there’s some input to my system that justifies these, I don’t get justified. I’ve just woven a larger tall tale. You must somehow account for the “data of experience” and the coherence theory seems omit this.
Examples can make the point sharper, although they have the air of unreality about them. The
Magic Feldman example in text does it. The argument based on this example is formulated in the text, p. 69.
Another way to make the point: if only other beliefs can justify a belief, then, since MF and MJ have the same beliefs, MJ doesn't have anything to justify his beliefs that MF lacks. So, MJ can't be better justified than MF. But he is. So, part of what determines what's justified is the character of one's experiences.
Questions about the example:
1) Is the example possible?
2) Is it true that MF would not be justified in his beliefs?
3) Is it true that there is no way, within the confines of the coherence theory, to fix things up?
I think that the answer is “Yes” to all three questions. A final question is:
4) Can any more realistic examples be used to make the same point?
Again, I say “Yes” using the Example 4.12 on p. 69.
VI. Conclusions on Coherentism
See text, p. 70.
In the case of foundationalism, it seemed to me that the general foundationalist idea was not undermined, even though the details of Cartesian foundationalism were problematic. In the case of coherentism, things seem worse.