Stanley’s writings (Stanley & Williamson 2001; Stanley 2011; Stanley & Krakauer 2013) on the topic are suggestive. However, he seems to mistake evidence in favour of the insight (that understanding is present in perception and action, as well as in the setting of explicit deliberative thought) with support for intellectualism itself (for the view that judgement governs action and perception). And, on top of that, he may commit the fallacy of conceiving the whole genus on the model of one of its species; like thinking that every dog is a cat because, well, they are mammals, or that seeing is a way of touching because, after all, they are both forms of perception. In this case it is the fallacy of thinking that knowing how must be a form of knowing that because, after all, it is form of knowledge.
Let’s turn to this last point first, briefly. Stanley (2011) notices that we use “to know” both for propositional knowledge and also for practical knowledge (know-how). Contrary to what he suggests, however, there are cognate languages where this is not the case. For example, we don’t express knowing how in German using the same verb that we use to express propositional knowledge (Stanley 2011, pp. 36-37). We use können, which means can; we don’t use wissen (as in wissen wie).
But in any case, the more important point is, so what? How dispositive are facts like this supposed to be? It is common ground, I would say, that know-how is a form of knowledge, an achievement of understanding. The question is whether it is a form of knowledge of the same type as propositional knowledge, the sort of knowledge that gets expressed in judgement. Crucially, all the evidence in the world that it is a form of knowledge doesn’t add up to evidence that it is propositional knowledge.
Now, as a matter of fact, we know that knowing how to do something is not merely knowing that a proposition is true, for any proposition you might care to think up. For knowing how to do something implies that you have the ability to do it (and vice versa), whereas the corresponding propositional knowledge has no such practical entailments.
Stanley would deny this (Stanley & Williamson 2001; Stanley 2011). You can know how to perform a stunt but be unable to perform it (because you’ve been injured, say); so, he claims, possession of know-how cannot be equivalent to possession of an actual ability. But this is unpersuasive. Of course it is true that you can know how to do something even though you are unable to do it. But this is because your being unable to do it is not, in the relevant sense, evidence that you can’t do it! Consider: you can’t swim if there’s no water, even though you can swim. You can swim but you can’t swim. Far from showing that know-how and ability part ways, this sort of consideration reminds us that they move along the same rails.
So knowing how to do something isn’t possession of propositional knowledge: it doesn’t consist in being in a position to make certain judgements. This is a point that Stanley and Williamson accept, if only implicitly, for they provide a different analysis of the cases precisely to account for the critical link to action in the case of know-how. Knowing how to do something, on their view, consists in grasping a true proposition, yes, but it consists in grasping it in a distinctively and irreducibly practical way (making use of practical modes of presentation).[11]
Again, it is worth noticing that to deny, as I do, that knowing how to do something consists in knowing the truth of a proposition, is not to deny that, as a matter of fact, knowing how to do something may put you in a position to make certain judgements, or may require you to appreciate the truth of certain propositions.
This brings us to the first point above: the confusion of evidence for the insight with evidence for the thesis. I am assuming that know-how, like propositional knowledge, is a form of knowledge. This common ground is already secured by the insight: our understanding, our knowledge of concepts, is put to use in both cases. So we can readily agree with Snowdon (2004), cited approvingly by Stanley (Stanley & Williamson 2001), that knowing how and knowing that go together—that where you have one, you have the other. In general, as Snowdon observes, if you know how to do something—say, how to get home from here—then you’ll know that all sorts of things are true, such as, for example, that you need to turn left here, that you aren’t already home, etc. And vice versa. Knowing how and knowing that, in this sense, commingle and cooperate. These considerations are adduced by Stanley, and by Snowdon, I think, to suggest that Ryle was mistaken in believing that the propositional and the practical are disjoint and disconnected (1949); in fact they operate together and in support of each other. This is an important point and one I endorse. And this is exactly what one should expect given the intellectualist insight. After all, understanding operates in both spheres: the practical and the judgemental or propositional. Crucially, however, the fact that the practical and the propositional mutually entail each other in this sort of way lends no support to the intellectualist idea that one of these, the propositional, is foundational in respect of the other; indeed, it weighs against that very idea. Why press on and insist on this thesis when, it would seem, the insight on its own is enough to capture the phenomenon at hand?
Stanley’s motivations seem fairly clear. He wants to break with the idea that propositional knowledge is detached and, as he puts it, behaviourally inert. He wants to insist that it’s wrongheaded to think that athletes and clowns and craftspeople are skilful zombies, whereas philosophers and mathematicians and physicists are intellectual workers whose actions exhibit authentic brain-power. It may be, even, that he thinks this is a point of political significance.
Intellectualism isn’t necessary to secure any of this, however. The insight has already done that.
In fact, intellectualism, as Stanley develops it, threatens to distort the nature of the cognitive achievements that are put to work in our practical, perceptual, and personal engagements. This comes out in the discussion of skill. Stanley & Krakauer (2013) defend Aristotle’s claim (from Metaphysics 1046b) that we can only speak of skilful action, as opposed to mere habit, or brute capacities, where we can speak of rational control of action, and also where we can speak of teaching, learning, practicing, getting better, or achieving expertise. They defend Aristotle’s claim that it is a mark of skilfulness, that you can voluntarily choose to perform what you can do skilfully badly.
This last point seems unlikely. I can’t choose not to understand what you say, or to see writing as mere squiggles, or words as composed of bits I need painstakingly to sound or spell out. A guitarist cannot choose to experience the instrument in his hands as strange or unfamiliar. At best, maybe, I can pretend I am unable to do these things.
Is this because talking and reading and playing guitar are not really skilful at all, that they are mere habits outside the range of rational control? Hardly! They’re expressions of skilful competence, rational understanding and knowledge if anything is. The mistake is to think that a performance is only rational if control is exerted in the mode of judgement, as if from outside. The understanding that is put to work in our talk and play, as in our thought, is native to these various styles of engagements themselves.
Stanley and Krakauer make a lot of the demand that skill depends on knowledge of facts. It’s worth noticing, yet again, that insisting, as I do, that skilfullness does not consist in the exercise of concepts in the judgemental mode does not entail that there can be skilfulness in the absence of the ability to exercise them in that mode. It may be, as a matter of fact—this is related to the Snowdon point above—that only someone who is sensitive to all sorts of facts, for example, about how something is done, will in fact know how to do it. This doesn’t show that knowing how is a kind of knowledge of the facts. It shows rather that our distinct conceptual capacities may be interdependent.
Stanley and Krakauer try to draw a line between true skills, which are, in their sense, governed by rationality, and others—for example perceptual and linguistic skills—that are too basic, or too simple to qualify as skills in the fuller rational sense.[12]
One problem with this suggestion is that it is not so easy to draw a sharp line between skills and supposedly brute abilities. Take colour vision, for example, which is innate in humans. Despite this, it turns out that children find it very difficult to recognize and discriminate colours long after they’ve mastered the names of familiar objects, people, games, etc. As Akins (unpublished manuscript) has argued, this is probably because colours are not simple, as our phenomenology, or rather, our conventional wisdom about our phenomenology, leads us erroneously to believe. Getting blue or yellow or red is to develop a sensitivity to suites of constancies and variations—to ecological variation in what I have called colour-critical conditions—that takes time and learning, and allows for criticism and reflection. Is colour vision basic? Or is it skilful? It may be both.
This is not a special case. Because seeing is saturated with understanding, it is very hard to find features of our ability that are not modulated by knowledge and context. Granted, the ability to discriminate line-gratings of different densities is fixed, at its limit, by the resolving powers of the eyes; yet our discriminations are likely to be sensitive to task and motivation, to attention and distraction—that is, very broadly, to our engagement with the meaningful world. So where does skill stop and brute ability begin? I am skeptical that learnability, teachability, or rational control provide an interesting or valuable demarcation. The most basic reason for this is that perceiving is never merely registration. It is a matter of knowledgable access (Noë 2004, 2012).
There is a second important issue as well. Consider language. Linguistic misunderstanding doesn’t stop language in its tracks, ejecting you and sending you back to the grammar, written, as it were in advance, by those responsible for setting up the language. Rather, coping with misunderstanding—dealing with not getting how someone is using words, or how we should use them, or with not knowing how to use them—is one of language’s familiar settings. We adjudicate and teach and learn and improve and criticize and define and formalize and evaluate within language, not from outside it. Language, contrary to the claims of Chomskyan linguistics, is not a rule-governed activity. It is a rule-using activity. And we make up the rules as we need them and for our own purposes. This may be controversial. But here’s why I insist on it: according to the logician’s or the linguist’s picture of language, first you assign values to primitives, then you set up rules governing the construction of well-formed formulas. If you think of language this way, then it looks like you need judgement—the application of rules to cases—to secure the meaningfulness of what would otherwise be mere marks and noises. But we don’t need judgement—we don’t need understanding in the judgemental mode—to secure meaning. We don’t need guidance from the outside.
The opposition between habit and skill is a false one; and it is a mistake to think that what marks the opposition is that habit is below or before understanding whereas skill is the deliberate exercise of understanding.