In Philosophy of Philosophy chapter 7, section 2 Williamson reaches a provocative conclusion about intuitions:
Philosophers might be better off not using the word “intuition” and its cognates. Their main current function is not to answer questions about the nature of the evidence on offer but to fudge them, by appearing to provide answers without really doing so (2007: 220).
The bumper sticker for Williamson’s argument could be: “intuitions are nothing special.” How does he reach the conclusion that intuitions fail to be distinctive and that they constitute “smoke and mirrors” methodology?
Williamson’s argument against intuitions is reductive. He explores various interpretations of intuitions and shows that each account can be reduced to routine forms of reasoning. Further, he argues that the use of intuitions in philosophy is pernicious. The practice is ineffective in answering philosophical questions and makes what philosophers do look like a scandal against rigor.
The context of this argument is Williamson’s argument against Evidence Neutrality (which I discuss here) and Williamson’s 2004 paper on intuitions (which is found in the journal dialectica). I will not rehearse those arguments, but as a rough summary of key considerations: intuitions are often taken to be evidence in philosophy, as evidence intuitions report psychological facts, Williamson is an externalist about mind, he wants to show that intuitions are not truly evidence, thus his favored understanding of evidence (as factive, not psychological) is unfazed by the misuse of the word “intuition” by philosophers. Put simply, intuition-speak is something philosophers engage in when they run out of arguments and want to try to make evidence uncontentiously decidable. Philosophers need to own up to the fact that evidence is always contestable. This last claim, of course, can be placed in the context of Knowledge and its Limits (i.e., one can know something without being in a position to know it). Now, I will raise problems with Williamson’s treatment of intuitions.
I claim that only if certain distinctions are misapplied or ignored can Williamson help himself to his conclusion to dump the use of “intution” in philosophy. Once certain distinctions are brought into focus his reductive project is thrown into relief.
Williamson’s strategy involves setting the target at intuition minimalists like Lewis (1983) and van Inwagen (1997). For them intuitions are mere opinions, beliefs or attractions to belief. There is nothing distinctive about such intuitions because even a field like science relies on some beliefs. The minimalist conception of intuitions is not distinctive of philosophical practice. If other approaches to intuitions can be show to be non-distinctive, then they can be reduced to the minimalist conception which shows that “intuitions are nothing special.”
When considering the rationalist conception of intuitions Williamson misapplies a distinction between the application of concepts and the grasp of concepts. For Williamson, application of concepts requires skills. Simply grasping concepts is insufficient because certain capacities are required to, for instance, correctly respond to the Gettier case. However, for rationalists like Bealer (1998), applying concepts to cases is part of the standard justificatory procedure. Capabilities for applying concepts to cases are not something separate or over-and-above what it means to correctly grasp concepts. Part of what it means to properly grasp a concept is having correctly applied the concepts to cases. A rationalist like Bealer would not claim that concept possession alone is sufficient, as Williamson implies. Such a reading stems from a misapplication of the application/possession distinction regarding conceptual analysis. As understood by rationalists, intuitions about cases are valuable because they clarify concepts of philosophical importance (e.g., knowledge and justice).
Williamson sharpens his attack against rationalists by looking at their interpretation of intuitions, as intellectual seemings analogous to perceptual seemings. Using the Muller-Lyer illusion as an example Williamson concedes that background information can defeat our inclination to view the two lines as incongruent. We can know how the illusion works and not be fooled. Yet, the richness of the phenomenology between intellectual and perceptual cases are not the same. Williamson argues much more appears to us perceptual when something seems the case versus intellectually when something seems the case. There is no seeming beyond our conscious inclinations to believe certain propositions. These inclinations to believe, however numerous, constitute our informational field. They are what we are justified in believing given our informational position. For example, the inclination to believe the Muller-Lyer lines are incongruent can only be reististed if we are in an informational situation that allows us to be justified in believing the lines are actually congruent, which consists in background knowledge about illusions. This brings up the distinction between situational justification and belief justification (Audi, 2003: 2-3), which seems to throw into relief Williamson’s knowledge-first epistemology.
Belief justification is the positive status that attaches to beliefs which warrants us, as rational beings, in thinking the beliefs true. Belief justification is ground on situational justification and knowledge is not possible without justified beliefs. As Audi says:
We cannot have a justified belief without being in a position to have it. Without situational justification we are not in such a position (2003: 3).
So, Williamson cannot impose knowledge from top-down onto beliefs; rather, knowledge must be built-up by having belief justification ground on situational justification. Inclinations to believe can only be stably overridden by justified beliefs. Williamson is correct that feelings or seemings are not analogous in the perceptual and intellectual cases, but in both cases the defeat of inclinations requires an epistemology in contrast with the knowledge-first approach.
There are two other distinctions I think Williamson misses. The first is the distinction is between inferential and non-inferential reasoning. Williamson simply dismisses the possibility of intuitions as non-inferential beliefs. He does this by showing that his intuition examples are inferential in nature. This reduction of intuitions to the inferential variety fails to account for the possibility of intuitions stemming from non-inferential reasoning. Bart Streumer (2007: 3) mentions the following example of non-inferential reasoning:
(Belief:) The weather forecast predicts rain.
(Belief:) The sky is full of clouds.
So, (Belief:) There are reasons to believe that it is going to rain.
It is possible to form a non-inferential belief that there is reason to believe that it is going to rain. One might even intuit that it is going to rain through such a non-inferential process. Williamson seems to discount and dismiss the possibility of non-inferential intuitions without argument.
The last distinction I’ll consider is the difference between having an intuition that P and finding it intuitive that P. Williamson uses the example of revisionary metaphysicians who deny that mountains exist as an example that shows that philosophical practice does not always involve accepting what is intuitive and denying what is not. Such metaphysicians allow the proposition “there are no mountains” into their system because the theoretical upshots outweigh the cost of having to endorse a counterintuitive proposition. Some intuitions, such as complex propositions, may fail to be intuitive. That is, it may require additional reflection to see the truth of the proposition. A proposition might not always invoke a sense of non-inferential credibility, but it may nevertheless be known to be true (see Audi 2008). That some philosophers accept counterintuitive propositions does not show that intuitions fail. It just shows that intuitiveness cannot be equated to something being an intuition.