Please visit the new home of Majikthise at

« Voter Due Dilligence | Main | Friday argument blogging »

October 22, 2004

Am I being fair to Jaegwon Kim?

Jaegwon Kim's What is "Naturalized Epistemology?" is probably the most influential critique of Quine's Epistemology Naturalized. At this point, I'm trying to make sure that I understand Kim's arguments against naturalized epistemology. What follows is my precis of Kim's argument. I'm writing a defense of Quine against Kim, so I want to be absolutely sure that I'm giving the opposition a fair shake. Any feedback would be greatly appreciated.


Quine is advancing an alternative approach to epistemology. We can accept Quine’s critique of classical epistemology without endorsing his alternative program. We should only endorse his alternative program if we are satisfied that his approach is an alternative to epistemology, i.e., if we are confident that the Quinean program will give better answers to our epistemological questions.

1. Epistemology is supposed to be about knowledge
2. Knowledge is justified true belief
3. Justification is the distinctively epistemic feature of knowledge.
4. Justification is a normative concept. A belief has the property of being justified if it stands in the correct logical relationship to other beliefs held by a cognizer. Justified beliefs are those that we ought to accept and/or those that we would be epistemically irresponsible to deny. Justification has something to do with praise and blame.
5. Belief also has a normative dimension. Belief attribution requires belief evaluation. We cannot directly observe beliefs, we must attribute them through interpretation. We are forced to make certain normative assumptions about other people in order to recognize their behavior as evidence of a belief. For example, we cannot recognize a belief in a creature unless we assume that it is basically rational and responsive to evidence. We have no right to attribute a belief unless that belief strikes us as logically coherent and/or (in some sense) justified by evidence available to the subject.
6. Quine wants to overthrow the entire framework of justification-centered epistemology. We know this because he urges us to study the causal connection between sensory input and belief-formation in the human subject to the exclusion of the evidential/justificatory relationships between beliefs. Quine argues that it is better to watch and see how science is actually developed and learned than to attempt to rationally reconstruct science from first principles.
7. Quine is only interested in the development of theories. He wants epistemologists to study which inputs produce what beliefs. It is “none of the naturalized epistemologist’s business” whether sensory inputs justify belief outputs.
8. If epistemology rejects justification, then it can't be about knowledge. This might be a reason to dismiss any non-justification centered "epistemology" out of hand. However, the skeptics would have us believe that no one actually has knowledge anyway. If they’re right, an epistemology of belief might be the next best thing.
9. However, even the scientific study of belief requires minimal normative assumptions, i.e., we must presuppose certain basic kinds of justification in order to attribute beliefs to subjects.
10. Quine’s naturalized epistemology is supposed to be the scientific study of belief-formation. But if we repudiate normativity, we have no grounds to attribute beliefs. Therefore, a naturalized epistemology that repudiates normativity will fail on its own terms.
11. A putative naturalized epistemology that addresses neither belief nor knowledge is not worthy of the name “epistemology.” Therefore we should reject Quine’s proposal to naturalize epistemology.

Jaegwon Kim. What Is "Naturalized Epistemology?"Philosophical Perspectives, Vol. 2, Epistemology. (1988), pp. 381-405. [JSTOR]


TrackBack URL for this entry:

Listed below are links to weblogs that reference Am I being fair to Jaegwon Kim?:


This seems dreadfully bad - I mean, Kim's argument seems dreadfully bad, assuming you've summarized it accurately. (5) and (6) seem especially weak. Grant, for the sake of argument, that we want a theory of justified true belief. I frankly have much more confidence that findings in, e.g., condensed matter physics consist of justified (approximately) true beliefs than I have confidence in any epistemological first principles I might try to employ in a reconstruction of cond. mat. That is, I have more confidence that cond. mat. is justified than in my ability to give a good account of justification. If that's so, why not use actually existing science, which we understand, to learn about the normative principles it embodies but which we don't understand? (Larry Laudan says something similar in Science and Values.)

Cosma, thanks for your comment. I'm glad you think (5) and (6) are the weak points. Those are the ones I'm attacking in my paper.

Frankly, I'm worried that I'm misconstruing Kim. I have immense respect for "What is Naturalized Epistemology?" It's a very subtle and beautiful work of philosophy. On the other hand, I've read the paper about 12 times and I still think this is more or less his case against Quine. My emphasis in the paper is primarily on Kim's interpretation of Quine's "Epistemology Naturalized." Quine never actually says that we shouldn't worry about justification. On the other hand, he says a lot things that seem like an aggressively descriptive agenda. I think Quine just rejects the kind of foundationalism that you're criticizing in your comment. Quine would say that if we want to know about human knowledge, we should start with putative good examples of knowledge, like our best theories of condensed matter physics, and try to find out what general properties of evidence or reasoning or causal connectedness separate successful theories from unsuccessful ones. Quine also thinks that we should investigate human perceptual and inferential abilities in order to understand the foundations of science. It's important to know whether we can actually make the kinds of distinctions that our theories would require us to make in order to justify them. We can be more confident in our scientific verdicts if we establish empirically that we are equal to the task (or that we are able to compensate for our weaknesses).

On reflection, I think I'm going to ammend (6) to emphasize that Kim doesn't think that Quine wants to reject justification-centered epistemology just because Quine recommends that we study belief-formation naturalistically. Kim also thinks that Quine errs by emphasizing causal relationships between stimulus and belief to the exclusion of evidential/justificatory relationships.

I've always read Quine as a sort of Reidian particularist. Instead of starting with a theory or method of justification (ala Descartes or Hume), we start with a list of things we know. Quine replaces mature science for common sense. That we have justified beliefs is not in question. We are justified, for instance, in believing in our best science. That's a starting point. And there's no point in not being dogmatic about it in the face of skeptical challenges. Because we're going to be dogmatic about it, there's no point in torturing ourselves about WHETHER we're justified. Let's just go out in the world and see HOW we are.

Now, if we want to know how it is that we are justified in believing in our best science, we've just got to use our best science to find out what it is we're doing when we gather scientific knowledge. So the descriptive task is first. We find out what it is that's actually going on in our heads and in the social relations of scientists, etc., that leads to scientific belief. Then we tease out of the descriptive account exactly what it is that causes this sort of belief formation to do a better job at truth-tracking than the alternatives. And then we formulate a normative theory: we say that if a process of belief-formation approximates the sort of process that we've discovered to be best in gathering scientific belief, then it's justified. No repudiation of normativity, just an explanation of normativity.

The problem with Kim as I see it is that he's determined to see justification as something that you can discover from the inside. We do attribute belief on the basis of a notion of jusitification or rationality, and so our pratice of attribution has an inherently normative element. But this is not to say that we're not systematically mistaken about the correct normative standard, or in the correct application of that standard. Only naturalized epistemology can tell us if we are or not--whether our practices of attribution actually tend to track the truth about other people's beliefs and such.

Sorry. This is cryptic. And I'm imposing truth-tracking talk onto Quine which isn't there.

I do not understand what the word "true" adds to the definition in #2. Suppose we think a belief is justified, and that no other belief can also be justified. In this case, the word true adds nothing to the idea of justification. If we think that some other belief can be justified, then we have no way of deciding which is true, at least based on knowledge. In this case, we may weigh justifications to see if one might be better, or we might hazard a guess, but until there is more information, there is no way to know. In this case also, the word true adds nothing.

I also do not understand this idea: "A belief has the property of being justified if it stands in the correct logical relationship to other beliefs held by a cognizer." This seems to mean that a belief is justified if the cognizer has constructed an internally consistent set of beliefs. But surely the test is not the possibly idiosyncratic beliefs of the cognizer, but the coherence with other known data.

But what do I know? I'm a lawyer.

masaccio, the use of "true" in #2 stipulates that the belief must be about a true state of affairs. we cannot be said to have knowledge about something that does not exist.

in response to your second point, perhaps you are misconstruing what is meant as your interpretation (one of coherence) is consistent with the original statement. the correct logical relationship to other beliefs may be characterized as membership in a set of coherent beliefs.

Imagine my surprise when I just did a search for "quine belief attribute" and the almighty Goog turned up an old post from a blog I regularly read, which is written by a friend's older sister, and which is normally about American politics. This gives me hope that my blog can similarly evolve alongside my mental life as I settle into life as a graduate student.

The comments to this entry are closed.