Topic: GENERAL LOGIC
Knowing How v. Knowing That: Some Heterodox Idea-Sketches
by Uriah Kriegel
(Source: Desert Landscapes 7/19/2004)
Since Ryle, orthodoxy had it that knowledge-how is categorically different from knowledge-that. The latter is a form of propositional representation of the way things are, whereas the former is just a capacity. The occurrence of the word "knowledge" in both expressions should not mislead us to think that they have something significant in common. There is no way to reduce Agent's knowledge *how* to ride a bike to some knowledge *that* certain things have to be done.
I have a somewhat different view. On my view, knowledge-how consists in *non-conceptual conative representations*. (...)
By "conative representations" I mean representations with a telic, or world-to-mind, direction of fit (wishing that p, wanting that p, hoping that p, intending that p, etc). These are to be distinguished from cognitive representations, which have a thetic, or mind-to-world, direction of fit (believing that p, expecting that p, suspecting that p, etc.).
The above parenthetized examples of conative representations all have propositional, and therefore (presumably) conceptual, content. But just like there are non-conceptual thetic representations, so we should expect there to be non-conceptual telic representations.
(..)
If my suggestion is on the right track, then although we cannot say that knowledge-how reduces to knowledge-that, since knowledge-that is propositional (because of the `that'-clause), we *can* say that it reduced some sort of representational knowledge.
REACTIONS
Hi Uriah,
In what sense are your non-conceptual conative representations actually representational? One specific worry: what makes it the case that every step in the causal chain leading from my desire to ride the bike to my actual riding of the bike doesn't count as a non-conceptual conative representation of the next step along the chain?
Brad.
Comment by Brad Weslake -- 7/20/2004 @ 12:54 am
Ok, I'm only half getting this. Suppose I know how to pedal, and the way to pedal is to use M16 or M17. Intuitively, it's both possible and quite likely that I don't know that the way to pedal is to use M16 or M17 - I don't even have concepts for M16 or M17. But if I know how to pedal, I *must* have a concept for pedaling.
You say, knowledge-how consists in *non-conceptual conative representations*. But there must be at least a little more, right? Because I have to at least have the concept of the thing I know how to do. And it seems that it's optional for me to have the concepts of M16 and M17 - if I'm a physiologist with a detailed understanding of the mechanics of pedaling, I still (can) know how to pedal.
I'm not quite sure I'm interpreting you right, so I'm going to stop here. Does this sound correct?
Comment by Jonathan Ichikawa -- 7/20/2004 @ 6:41 am
Woo-hoo, my favorite kind of response: everybody's right (except me?).
Christian, I think your formulation of the conclusion in terms of ability rather than knowledge is better than my original formulation. Here's why I was thinking of knowledge nonetheless. The common account of knowledge - as a belief that is true, justified, and Gettier-proof - seems to be tailored to knowledge-that. But once we agree that there is something fundamental in common to knowledge-that and knowledge-how, then we need some wider uderstanding of knowledge simpliciter. My view of knowledge-how isolates only one fundamental commonality with knowledge-that, namely, that the respective states are representational. However, with this may come other commonalities. Representatons have conditions of satisfaction (truth conditions in the case of knowledge-that, and on my view, fulfilment conditions in the case of knowledge-how) and they are answerable to certain standards of representation formation (compliance with which gives "justification," "wareant," or something in the vicinity). So maybe a case could be made for talk of knowledge rather than mere ability. But talk of ability is certainly more cautious.
Brad: I think you're also right that talk of conative states being representational is problematic. Conative states are certainly intentional, they are about something. But do they represent something? Only in a somewhat technical sense. In the regular sense of the word, as I hear it, to represent something is to represent it to be the case, or just to be. In that sense, conative intentional states are not normally representational, because they don't represent the way the world is, but rather the way one would want the world to be. In using the term "representation" I had in mind the more technical sense used in discussions of the representational theory of mind etc. It is common in these discussions to take desires and other conative states to be representational, in the minimal sense that they have conditions of satisfactions - indeed, in the minimal sense that they are intentional or have aboutness.
Comment by Uriah Kriegel -- 7/20/2004 @ 9:21 am
Uriah Kriegel,
Even with representation limited to the standard philosophical usage, I am worried about how much representation there is in your idea of non-conceptual conative representations. If I intend to ride a bike, but fall off, it is clear that my intention has failed to be satisfied; moreover this normative component of intention seems integral to it being an intention in the first place. Similarly, if I see in my schooner (as we say here in Sydney) a particular hue of yellow beer, it seems I could come to believe that this seeming was mistaken, and that the beer is actually some other hue (even if I couldn't articulate the difference beyond saying "it seems different"). What is the analog for your non-conceptual conative representations? My question about the causal chain was designed to get at this - some criteria is needed for sorting out causal links that count as representational from those that do not; it seems that this criteria needs to be normative; and it doesn't seem that the processes that underlie know-how are normative in the right way.
Brad.
Comment by Brad Weslake -- 7/21/2004 @ 10:53 pm
Right, I forgot to address the Brad's problem of fixing the content of those "non-conceptual conative representations." The problem described by Brad is parallel, however, to the *horizontal problem of disjunction* for naturalist theories of cognitive representations. There are two disjunction problems that arise for cognitive representations. The first is: what makes it the case that my cat-thought is a representation of a cat and not of a cat-or-small-dog-on a moonless night? Call this the vertical problem of disjunction. The second is: what makes it the case that my cat-thought is a representation of a cat and not of a cat-or-big-bang? Call this the horizontal problem of disjunction. Similarly, we may ask what makes it the case that a desire to bike is a (conative) representation of biking and not of biking-or-getting-to-the-office? I don't have a good answer to this problem, or any answer really, but I'm comforted by having the partners in innocence I have in naturalist theories of cognitive representations. In particular, Fodor, Dretske, and Gillet had some interesting things to say about this problem at the end of the 80s.
Comment by Uriah Kriegel -- 7/22/2004 @ 3:15 pm
OK,
Seems I am just more sceptical than you of causal theories of content in general (I don't think any of the replies to these sorts of problems ended up succeeding). To expose my own biases, I think the prospects of an explanation of know-that in terms of know-how (via functionalist or pragmatist approaches to content) are far better than those going in the other direction...
Comment by Brad Weslake -- 7/22/2004 @ 9:32 pm
Posted by Tony Marmo
at 13:58 BST
Updated: Monday, 9 August 2004 08:05 BST
