Knowing how to

3. An epistemology of knowing how to?

The line of thought sketched in §2 has been challenged (see for example Rumfitt 2003). But for epistemologists the outcome of that debate may not be of any great importance. From the thesis that knowing how to is a subspecies of knowing that, nothing follows as to whether it does or does not have a distinctive epistemology. (Nobody takes the fact that a posteriori and a priori knowledge are both classes of knowing that to show that there is no fundamental epistemological difference between them (see A PRIORI; A POSTERIORI).) True, the notion of a ‘practical mode of presentation’ has an epistemological ring. But for the moment we know little of what it means, and it may turn out to signify something about what the subject S knows about W, or about the way W promotes Φ-ing, rather than anything characteristic about how S knows it. To close in on these questions calls for more about just what is known by someone who knows, under a practical mode of presentation, that W is a way to Φ. What we know is that such a someone is in whatever state a subject S is in when S knows how to Φ. So what state is that?

We at once find ourselves hampered. First, intuitions amongst competent English speakers about the applications of ‘knows how to’ are not strikingly uniform. There are those (prominent in this context is Gilbert Ryle; see below) who sometimes use it interchangeably with ‘can’, not drawing much of a distinction between, for example, ‘knows how to swim’ and ‘can swim’ or ‘is able to swim’. Others (including the present writer – see Craig (1990)) are uncomfortable about using ‘knows how to’ unless at least some degree of reflective awareness is present; they would not say that the boy of Ryle’s example (see Ryle 1949: Ch. 2, §6, where the practical linguistic abilities of a young native speaker are contrasted with the largely theoretical knowledge of a hypothetical foreign grammarian) knew how to speak grammatical English but only that he could.

Second, the English ‘knows how to’ does not translate very reliably – some languages have an analogous construction which on closer inspection turns out not to do quite the same work. Hungarian, for example, has (know (tudni) + infinitive) – which is its standard expression for ‘can’; German has (know + infinitive) also, but it is not very common and always indicates a particular aptitude on the part of the subject. Russian has a very similar construction which is semantically very close to the English ‘knows how to’, but which unlike English uses a different verb from the one it uses when the object is a proposition. Faced with this situation the best thing for the epistemologist must surely be to forget linguistic detail and try to survey the whole range of phenomena that these expressions cover. For the second point suggests that it can without significant loss of communicative efficiency be carved up in different ways, while the first suggests that we are not unanimous even about how English carves it up.

The range of phenomena in question covers two areas. On the one hand we have cases in which a subject is aware (whether they can put what they are aware of into linguistic, propositional form or not) that certain features of a performance make it more likely that it will be an instance of Φ-ing, or of particularly high-quality Φ-ing, or of Φ-ing in a way appropriate to the circumstances, and such like. On the other hand we have subjects with abilities to do things (change the way things are), spanning both the mental and the physical, from occurrences that barely count as intentional actions at all to the exercise of the most refined and intricate skills.

The first area does not look promising. Awareness of this kind calls for recognition of the features in question and their correlation with Φ-ing, good and bad, appropriate and otherwise. Certain particular values of Φ may raise their own problems, but from our present level of generality no need for any distinctive new epistemology is visible – just the fact that in some cases expertise, sensitivity and detailed knowledge of social convention will be required.

The second area looks better. Here the suggestion seems to be that abilities themselves are cognitive in their own right, which sounds more exciting and may well be what has caught the attention of some of Ryle’s readers. (And perhaps also of POLANYI – see Polanyi (1958).) But if abilities are to be declared cognitive this should be based on something more solid than just the fact that some languages speak of them using terms that are quite certainly cognitive when used in other contexts. What is needed is some connection or analogy between possession of abilities and other cases uncontroversially taken to be cognitive. So we might ask whether notions such as justification or warrant get any purchase. But the notion of being justified in possessing an ability (supposedly parallel to being justified in holding a belief) seems to make little sense. One might be said to be justified in possessing an ability if the payoff from it would recompense the trouble of learning it and keeping it in good repair, but that seems to parallel the idea of a belief’s being worth the bother of acquiring and sustaining, or of a car’s being worth buying and maintaining, rather than that of epistemic justification.

Are prospects brighter if we forget about justification and look at the ways in which abilities are acquired? No doubt there is an interesting story to be told about our remarkable facility in imitating what we see and hear done by others, on which our capacity to pick up further abilities so heavily depends – a facility as remarkable in its own way as that of generating, as we constantly do, representations of our environment from perceptual stimuli. But it will probably best be told by experimental psychologists and researchers in the field of artificial intelligence rather than philosophers.

It would give us a novel link back to more familiar epistemology if we could regard an individual act of imitation, such as frequently occurs in learning new abilities, as a kind of representation of the action imitated. Perhaps we can, but there are grounds for caution. The point of such imitation is not to represent what it imitates, but to capture those features of it which made it an act of Φ-ing, that is of whatever the imitator is trying to learn to do. Unlike a representation, its success is judged not by looking back to the original but by looking forwards to its own consequences.

The hope for a new epistemology connected with knowing how to begins to look over-optimistic, but this field is under-researched at present. The above is meant only to indicate lines that such research might take, and some of the difficulties it may encounter. Whether they are insuperable or not remains to be seen.