the boundaries of language

I'm working on a short paper discussing the claim that knowing, for some way w, that w is a way to X is sufficient for one to know how to X (ignoring for now complications such as the need for one to grasp this proposition under the right mode of presentation, or for one to minimally understand w). But rereading John and Marc's paper has got me thinking again about the associated claim that possession of the corresponding ability is not necessary for knowledge how.

There are now a number of stock cases designed to show that knowing how to X does not entail being able to X. These have tended to fall into broadly three classes. The first are cases in which, as Snowdon puts it, we 'describe cases in which the subject can show, teach, or tell (or otherwise convey to) us how to do something, and hence must be credited with knowing how to do it, but is for some reason or other unable to do it' (Knowing How and Knowing That: 9). Jeff King's example of the ski instructor who is personally unable to perform the stunts is a case belonging to this first class. The second class of cases are one's in which the subject uncontroversially previously both knew how and was able to X, but is now unable to X. The tragic concert pianist who has recently lost his arms in a car accident is a case of this kind. The third class consists of cases where performing X is intuitively 'more of the same' of what it takes to perform Y, where the subject in question uncontroversially both knows how to Y and is able to Y. John and Marc offer the example of Irina, the figure skater who knows how to perform a quadruple Salchow, and is able to. They suggest that intuitively she also knows how to perform a quintuple Salchow, even though she just can't seem to land one. Further cases of this sort have been offered by Jonathan Ellis in unpublished work.

I incline towards thinking that these cases don't yet show that ability is unnecessary for knowledge how, and towards thinking that Noë and Hawley offer responses to these cases that are more or less on the right track. According to Noë, if one knows how to X, and certain 'enabling conditions' are satisfied, one will be able to X. So, of the tragic concert pianist, he writes:

'We judge she knows how to play even though she is now unable to play, because we think of the loss of her arms as comparable (in the relevant sense) to the loss of her piano; as we tell the story, it is reasonable to think that the accident brings about the failure of a necessary enabling condition to be satisfied.' (Anti-Intellectualism: 283)

Hawley suggests that satisfying a counterfactual success condition is necessary for knowing how; if one's knows how to X, then: if one tried to X, under normal circumstances, one would successfully X (Success and Knowledge-How: 22). The ski instructor and the concert pianist meet this condition, since were they to attempt the stunts or playing the piano under conditions which are normal for those activities, they would succeed (23). Their lack of actual success does not impugn their knowing how.

I rather suspect these proposals are best treated as variations of the same idea; one should replace the notion of 'normal' conditions for X-ing with the notion of all of the relevant enabling conditions for X-ing having been satisfied. This kind of response puts tremendous pressure on the notion of an enabling condition, and it's far from clear it can bear the strain. There's a real danger of trivialization here (as John has stressed to me in conversation - there's some relevant discussion in Hawley's paper), since we can't allow that having learned to play the piano is an enabling condition for playing the piano; we don't want to count as knowing how to play the piano someone who would be able to play were it not for her failure to have learned to play. Similarly, having practiced sufficiently cannot count as an enabling condition for playing the piano, nor can having lived in Moscow for 5 years be an enabling condition for speaking Russian; one does not know how to speak Russian even if one would be able to speak Russian if one lived in Russia for 5 years (this latter example is Hawley's). We clearly need a much better grip of the notion of an enabling condition than we currently possess.

One option that Hawley mentions in passing (22) but doesn't pursue is that cases in which one knows how to X but is cannot actually successfully X are cases in which 'physical limitations' prevent one from X-ing. This would mean that satisfaction of enabling conditions would amount to the absence of such physical limitations. Of course, this notion of a physical limitation is itself very slippery, but that doesn't show that no progress has been made. We can make what seem to be intuitively correct classifications: lack of limbs is a physical limitation, being unfit or inflexible are physical limitations, whereas not having learned to play the piano and not having lived in Russia for 5 years are not.

What surprised me reading John and Marc's paper again is that they seem to put quite considerable weight on a distinction between being unable to X right now and being unable to X generally. They hold that putative counterexamples to the necessity of ability for knowledge how fail if they are merely cases in which the subject is unable to X right now (Know-How and Concept Possession: fn5). In contrast, they argue that the ski instructor and ice-skater cases succeed in showing that ability isn't necessary for knowledge how.

They also invoke this distinction in response to an objection later in the paper. They argue that a gap between knowing how and being able does not open up with regards to certain activities, such as carrying out simple mathematical and logical operations. But they consider the objection that a small gap can open up even here:

'For instance, Irina might know how to add, but be unable to do so because she presently lacks pen and paper, the use of her fingers, an abacus, etc. (which , owing to a poor short-term memory, she needs to perform even the simplest mathematical calculations). On the face of it , this seems to be a case of knowing how to add absent the corresponding ability. But in fact it is not. For in such a case, Irina is able to add; she is just unable to do so right now.' (36, emphasis in original)

However, this is misleading. It's not just that Irina is unable to add right now; as the case is described she's unable to add generally in the absence of pen and paper (or an abacus, etc.), and it just so happens that she presently lacks these things, which explains her present inability to add. If we want to maintain that Irina knows how to add quite generally, what can we say to prevent her general inability to add under certain conditions from calling into question the necessity of ability for knowledge how even in John and Marc's restricted class of cases?

I want to suggest that John and Marc have implicitly acknowledged the need for more or less the kind of distinction that the anti-intellectualist seeks to lean on in offering a response to the ski-instructor/ice-skater/tragic pianist cases, but they've mishandled it. In taking it to be a distinction between not being able to X right now, and not be able to X generally, they've thought that it's a distinction which is of no help to the anti-intellectualist. They seem to have seen things as follows. Cases in which a subject is unable to X right now do not show that an ability to X is unnecessary, whether X here is performing a quintuple Salchow or subtracting 2 from 4. However, there are cases in which a subject intuitively knows how to perform a quintuple Salchow but is generally unable to land one, while there are no cases in which a subject intuitively knows how to subtract 2 from 4, but is generally unable to. But the case in which Irina is unable to add without paper and pen is just that; it's a case in which she's only able to add under certain conditions. The anti-intellectualist line I attributed to Hawley and Noë above suggests that just as Irina, idiosyncratically, is only able to add under certain circumstances, people more generally successfully X, for any X, only under certain circumstances; perhaps those are best thought of as circumstances in which the relevant physical limitations are absent - and perhaps not.

This suggests the story about the link between knowledge how and abilites/successful action will be enormously complicated, since we must factor in not only physical limitations or the like, but also idiosyncratic limitations like that displayed by Irina in the simple arithmetic case. But I don't find the conclusion that our story must be complicated a disappointing or objectionable one; it strikes me as one thing we should have all learned already from the ongoing debate between intellectualists and anti-intellectualists.

(Thanks to John Bengson for discussion of these issues over the past year or so).

Labels: Epistemology, Knowledge How