In a recent combox discussion, I was informed that intuitions were not generally reliable (in response to a claim I had made that intuitions are generally reliable). I thought it would be better to do a blog post about it. So, what does it mean for something to be generally reliable? It occurs to me that the objector probably thought I was claiming, “Of the set of all intuitive beliefs, enough intuitive beliefs have turned out to be true so that we can claim intuition is justified,” kind of like saying “75% of this person’s claims are true, so we can believe the next one is probably true.” That’s not quite what I mean. Stipulatively, I am using “generally reliable” to refer to the source of belief. I am referring to the faculty of intuition itself. I believe rational intuition is itself a properly functioning, truth-aimed belief source (in the normal situation, hence the term “general”).
Taking this definition, what follows if intuition is not a generally reliable source of belief? That is, what if it is an unreliable source? Well, first, and really most relevantly, we must say that whatever belief is generated from an unreliable source must also be itself unreliable, unless or until it can be justified using solely reliable sources of belief or beliefs that are generated themselves from reliable sources of belief. Call this, for fun, the “Transitivity of Unreliability Principle” (TUP).
So, what if we grant TUP? Does anything follow? Coupling this with the premise that intuition is not a source of belief, we get the following consequence: whatever belief is generated from intuition must be itself unreliable, unless or until it can be justified using solely reliable sources of belief or beliefs that are generated themselves from reliable sources of belief. So far, so good.
But then notice an interesting predicament. Scientific beliefs, such as the belief in induction or the uniformity of nature, are beliefs generated from intuition. This is demonstrated in the fact that there is no known scientific way to justify the belief in inductive inferences, even in principle. This is a philosophical judgment, and one that will proceed in a rational, a priori-like manner. If one attempts to prove or justify induction in an a posteriori manner, she will find that she is reasoning in a circle. Further, any attempt to justify it using science will appeal to scientific premises, and, by our hypothesis, will also be unreliable. So, by logical inference from TUP and the given premises above, we will arrive inescapably at this bizarre conclusion: scientific inferences must also be unreliable. This means that all scientific conclusions are unreliable.
So what to do? The only assailable principle, as far as I can tell, is that intuition is not a generally reliable source of belief. At least, that’s the one non-inferential belief that seems most likely false. In fact, it is rational intuition that grounds science.
 Presumably, being unreliable is what it means to be not reliable, after all.