Knowing-How and Knowing-That

Transcription

1 Philosophy Compass 3/3 (2008): , /j x Knowing-How and Knowing-That Jeremy Fantl* University of Calgary Abstract You know that George W. Bush is the U.S. president, but you know how to ride a bicycle. What s the difference? According to intellectualists, not much: either knowing how to do something is a matter of knowing that something is the case or, at the very least, know-how requires a prior bit of theoretical knowledge. Anti-intellectualists deny this order of priority: either knowing-how and knowing-that are independent or, at the very least, knowing that something is the case requires a prior bit of know-how. Much of the dispute centers on the relationship between knowing how to do something and having an ability to do it. If having an ability is necessary and sufficient for knowing-how, this is thought to provide comfort for anti-intellectualists. This paper traces the place of ability in the knowhow/know-that debate from Ryle s seminal statement of anti-intellectualism through Stanley and Williamson s more recent defense of intellectualism. Contemporary epistemology distinguishes among three kinds of knowledge: propositional knowledge, knowledge by acquaintance, and practical or procedural knowledge. 1 Propositional knowledge is what is expressed by sentences relevantly similar to Alex knows that George W. Bush is the U.S. president. Knowledge by acquaintance is what is expressed by sentences relevantly similar to Bill knows Beatrice. Practical knowledge is what is expressed by sentences relevantly similar to Callie knows how to ride a bicycle. 2 We are concerned with practical knowledge knowing-how and propositional knowledge knowing-that. It is natural to think of the differences between these two states in terms of their different relations to abilities. To know how to do something suggests an ability to do it, whereas to know that something is the case does not immediately suggest any corresponding ability so thought Gilbert Ryle, the philosopher most responsible for bringing the relationship between knowing-how and knowing-that to the forefront of epistemology. But there are complications. A major portion of the debate is about how those complications manifest themselves in the relationship between knowing-how and knowing-that. After briefly mapping out some possible positions one might take on that relationship and, in section 1, discussing Ryle s view, I will turn, in section 2, to the role ability has played in the debate over the last 50 years The Author

2 452 Knowing-How and Knowing-That I will then, in section 3, look at Jason Stanley and Timothy Williamson s recent and influential response to Ryle s view (and related positions) before closing, in section 4, with a discussion of how normative concerns might figure more prominently than they do. There are three stances you might take on the relationship between knowing-how and knowing-that: (1) knowing-how reduces to or is a species of knowing-that; at the very least, knowing how to do something importantly requires a prior bit of propositional knowledge; (2) knowing-that reduces to or is a species of knowing-how; at the very least, knowing that something is the case importantly requires a prior bit of know-how; (3) knowing-how and knowing-that are independent kinds of states: neither is a species of nor reduces to the other, nor does either importantly require the other. The first stance is intellectualist in that it gives priority to the intellectual state of knowing-that, as opposed to the practical state of knowing-how. The second and third stances deny this order of priority and so are anti-intellectualist. 3 The second stance is strongly anti-intellectualist, because it makes knowing-that importantly dependent on know-how. The third stance is weakly anti-intellectualist, because it allows for an autonomous state of knowing-that. Intellectualism and weak anti-intellectualism or, at least, the bare denial of intellectualism claim the most defenders in the past century, so our focus will be on those more prominent views. 4 Strong anti-intellectualism does have representation, however, and it is worth discussing briefly before turning to the more canonical debate. Ryle is usually interpreted as arguing only for weak anti-intellectualism. Nonetheless, there is some evidence Ryle viewed knowing-how as at least logically prior to knowing-that, which might suggest a stronger antiintellectualism. In The Concept of Mind, Ryle hints at this relative priority, but in his earlier Knowing How and Knowing That, he is more explicit. Knowledge is not mere belief; it requires establishing (224). But establishing is an intelligent operation: it is an act that can be performed well or poorly and, when so performed, is correctly described as clever or stupid, respectively (among other adjectives) (213). Therefore, you have to know how to establish in order to know that anything is the case: knowing-that presupposes knowing-how. Further, you can know a fact only if you can intelligently exploit it (224), which is itself a way of knowing how to do something. John Hartland-Swann is committed to an even stronger anti-intellectualism: all cases of knowing that can ultimately be reduced to cases of knowing how ( Logic Status 114, original emphasis). In general, to know that p is to be able (know how) to state correctly what is the case to correctly answer the question whether p. 5 A more recent proponent of strong anti-intellectualism is Stephen Hetherington, who argues that to know that p is to have the ability the know-how to respond, to reply, to represent, or to reason accurately that p (77). Strong anti-intellectualism is most plausible if knowing that something is the case is essentially

3 Knowing-How and Knowing-That 453 dispositional if knowing that something is the case is a matter of having a disposition or capacity to ϕ (for some ϕ). 6 And perhaps you can have the disposition to ϕ only if you know how to ϕ. If so, then you can know that something is the case only if you know how to ϕ. But if knowing that something is the case is not essentially dispositional, then it seems rather implausible for knowing-that to be reduced to or a species of knowing-how. The claim that knowing that something is the case is essentially dispositional is controversial, and usually advanced only after some anti-intellectualist argument is already in place. If intellectualism hasn t been dispelled, then know-how might very well require some prior bits of non-dispositional knowledge-that. Therefore, strong anti-intellectualism seems like it requires a prior argument against intellectualism. And this is born out by the historical record. Hartland-Swann and Jane Roland argue for their versions of strong anti-intellectualism by assuming a dispositional account of knowingthat an assumption they take as legitimate only because they see Ryle as having successfully undermined intellectualism. So, a denial of intellectualism seems required before any stronger version of anti-intellectualism can be successfully defended. It is unsurprising, then, that this issue whether intellectualism is true has been the more prominent historical concern. It is the debate over intellectualism that will concern us most and to which we now turn. 1. Ryle s Anti-Intellectualism That knowing-how and knowing-that are distinct kinds is the view normally attributed to Ryle, who distinguished between the two in order to undermine what he called the intellectualist legend, the view that in intelligently performing an operation, you first run through a series of rules or statements about how it should be done, and then go ahead and perform the action: when we describe a performance as intelligent, this does not entail the double operation of considering and executing (Concept of Mind 29ff). Without the legend, we lose motivation for attributing a mysterious ghost in the machine, a distinct (and, to others, unobservable) mental agent related to the body and bodily movements as a puppeteer (if I may mix metaphors) is related to the puppet and its movements. If you believe in the ghost, then it is natural to believe that there are occult processes running in tandem with observable, behavioral processes. But if intelligent action is distinguished by observable dispositional tendencies, then we have no need for the ghost. Ryle s primary argument for the distinction is a regress argument common to the more cited chapter in The Concept of Mind and the earlier Aristotelian Society paper. There are not two processes in intelligent action, a doing, and a mental contemplation of the right way to do, preceding (or accompanying) the doing. If there were, then this mental contemplation

4 454 Knowing-How and Knowing-That would have to be itself a doing an act of mental contemplation. But this act had better also be intelligently done. This will require a further act of intelligent mental contemplation another doing. There seems no way to break through the regress. 7 This regress argument has been subject to the most criticism in later work. 8 According to Carl Ginet, we can manifest our knowing-that in intentional action without a further act of mental contemplation. For example, we manifest our knowing-that turning the knob will open the door simply by turning the knob and opening the door. We don t have to run through a further intelligent mental act of contemplation. Merely establishing that exercising one s know-how doesn t require further intelligent mental acts, then, doesn t entail that knowing-how is a distinct kind from knowing-that, nor does it entail that knowing-how doesn t have implications for our knowledge-that. Knowing-how might not require a prior act of mental contemplation, even if it does importantly require a prior bit of knowledge-that. So, perhaps the intellectualist can agree with the stated conclusion of Ryle s regress argument that there is no mental act of contemplation that necessarily precedes intelligent action. But there are at least two other arguments in the earlier paper that have received less attention. 9 In terms of evading Ginet s objection, the second is most important, though it shows up only briefly and unannounced in the middle of the paper. Ryle asks us to imagine a clever chess player who imparts to his stupid opponent so many rules, tactical maxims, wrinkles, etc. that he could think of no more to tell him (215). Despite all this additional knowledge this knowledge-that the stupid player might very well continue to play stupidly, might fail to know how to play chess well. He might fail to be able to intelligently apply the maxims (215). In general, according to this second kind of argument, knowing-how is not just knowing-that because two people who both know all the same propositions might differ in what they know how to do. Of course, the intellectualist can insist that we can t fully imagine the stupid chess player knowing all the propositions that the clever chess player knows. Ryle responds by saying that even if the stupid chess player does know all the propositions known by the clever chess player, he would need to know how to execute the relevant maxims, and that this would require further bridge principles, leading to the regress. But the intuitive point that it s possible to have identical knowledge-that while differing in know-how remains: a fool might have all that knowledge without knowing how to perform (217ff). The advantage of this argument is that, if sound, it shows more than just that there are not two separate processes in intelligent action a contemplation and a doing which is all that Ryle s straightforward regress argument shows. This, Ginet claims, the intellectualist might be able to agree with. But the intellectualist cannot agree with the conclusion

5 Knowing-How and Knowing-That 455 of the intuitive argument that two subjects might have identical knowledgethat without having the same know-how. If this argument is sound, it shows the denial of a supervenience thesis: no change in know-how without a change in knowledge-that. This thesis seems closer to the heart of intellectualism than a principle to the effect that there aren t two separate processes in intelligent action. Even if knowledge-that isn t a process in intelligent action even if it s just manifested, not actively contemplated or considered if we can change in know-how without changing in our knowledge-that, knowing-how can t just be a species of knowing-that. Ryle s position has both a negative component that knowing-how is distinct from knowing-that and a positive component a theory about what knowing-how is. According to the positive component, know-how is a capacity for or a disposition to a set of behaviors for example, the behaviors constitutive of the activity that one knows how to do. This makes it seem as if Ryle views knowing-how simply as an ability. 10 But there are abilities that, for Ryle, do not amount to know-how. For example, he contends that the ability to give by rote the correct solutions of multiplication problems differs in certain important respects from the ability to solve them by calculating (Concept of Mind 42) This makes rote answers more an exercise of habit than an exercise of an intelligent capacity an exercise of genuine know-how. In most of Ryle s examples of genuine know-how (knowing how to play chess or to speak a language) the relevant capacity or disposition is a capacity or disposition to act out of a set of rules. Of course, to act out of a set of rules in the relevant way, it is crucially not required that one know that the rules apply or that one in any sense theoretically avow (46) them. But it is required that the rules play a crucial role in one s action: Knowing how, then, is a disposition... Its exercises are observances of rules or canons or the application of criteria (47). If we want an interpretation of Ryle that makes know-how a matter of ability, it will have to be a certain kind of ability an ability to act out of the relevant rules (or canons or criteria) governing the intelligent execution of the activity known how to do. Nonetheless, much of the subsequent literature has concentrated on a much simpler claim: that S knows how to ϕ just in case S has the ability to ϕ. After all, if S knows how to ϕ just in case S has the ability to ϕ, then it seems that knowing-how must be distinct from knowing-that. It is to a discussion of this claim that S knows how to ϕ just in case S has the ability to ϕ that we now turn. 2. Knowing-How and Ability First, consider worries about the claim that to have an ability to ϕ is sufficient for knowing how to ϕ. David Carr considers a dancer who performs a dance he recognizes only as A performance of Improvisation No. 15. Carr continues:

6 456 Knowing-How and Knowing-That To the astonishment of a member of his audience who just happens to be an expert on communications, the movements of the dancer turn out to resemble an accurate (movement perfect) semaphore version of Gray s Elegy, though the dancer is quite unaware of this fact. ( Logic of Knowing 407) It seems clear that, while the dancer is able to bring about a semaphore version of Gray s Elegy, the dancer does not know how to bring it about. 11 Likewise, in Carr s Knowledge in Practice, he notes that a novitiate trampolinist, for example, might at his first attempt succeed in performing a difficult somersault, which although for an expert would be an exercise of knowing how, is in his case, merely the result of luck or chance. Since the novice actually performed the feat one can hardly deny that he was able to do it (in the sense of possessing the physical power) but one should, I think, deny that he knew how to perform it. (53) Early anticipation of this sort of example can be found in Robert Ammerman s and Hartland-Swann s exchange over Hartland-Swann s attempted strongly anti-intellectualist reduction of knowing-that to knowinghow. According to Hartland-Swann, to know that p is to have the capacity to answer correctly the question whether p: in short, in answer to your question, I am able (know how) to reply correctly ( Logical Status 115). In reply, Ammerman points out that, if we are able to answer a question and, by luck, we are correct, we will count as knowing the answer to that question: we will be able, by luck, to answer correctly. Thus, Hartland- Swann s reduction seems to obliterate the distinction between genuine knowledge and right guessing or true belief (31). In further response, Hartland-Swann contends that we don t have the ability to answer correctly unless we have some evidence that the answer is correct. As he says, I can only claim to know or to have the capacity to state correctly that Hitler is actually dead, or actually alive, if I have had access to evidence generally regarded as both relevant and sufficient in this context. ( Knowing That 70) 12 We have here a distinction between a weak and a strong conception of what it takes to have an ability (here synonymous with capacity). On Ammerman s weak conception, you have an ability to ϕ just in case it is possible that you ϕ (e.g., you have an ability to answer correctly whether p just in case it is possible that you answer correctly whether p). On Hartland-Swann s stronger conception, to have an ability to ϕ, it is not enough that it is possible that you ϕ (e.g., you have an ability to answer correctly whether p only if you have sufficient evidence). Abstracting from the specifics of their debate, there seem to be a number of things that could be required to be added to the weak conception. For instance, it could be required for you to have an ability to ϕ that (in some relevant context) you can ensure (to some sufficient degree) that you ϕ. Or it could be required for you to have an ability to ϕ that certain counterfactuals are true of you that in all close possible worlds in which you try to ϕ, you ϕ.

7 Knowing-How and Knowing-That 457 Depending on the requirements loaded into having an ability to ϕ, it looks like Carr s dancer and trampolinist examples might fail to be counterexamples because the relevant subject, contra Carr, lacks the respective ability. If the dancer can t ensure that a semaphore version of Gray s Elegy is displayed, it matters not that it is possible for the dancer to display it. If the novitiate trampolinist does not execute the difficult somersault in sufficiently close possible worlds in which he tries, then it matters not that the novice executed it. Neither has the ability to do what Carr claims. There is a legitimate question whether these stronger requirements square with our ordinary use of the term ability. We do, I think, have the intuitive reactions that the dancer is able to do a semaphore version of the Elegy and that the trampolinist is able 13 to execute the somersault. 14 Nonetheless, we can perhaps specify a technical sense of ability and claim that having an ability to ϕ in that sense is sufficient for knowing how to ϕ. As more requirements get loaded into what it takes to have an ability, it will get easier to resist counterexamples to the proposition that the ability to ϕ is sufficient for knowing how to ϕ. But it will get correspondingly more difficult to resist well-trodden counterexamples to the claim that having the ability to ϕ is necessary for knowing how to ϕ counterexamples neatly summed up by Brown: Ability to V is a sufficient but not a necessary condition of knowing how to V...There is an abundant supply of people who cannot do things but know how to do them. They can be found among aging athletes, neurotic coaches, actors with stage fright, architects, and male experts on natural childbirth. ( Reply to Brett 303) A fair number of similar examples involve loss of or damage to crucial appendages: Stanley and Williamson offer up the example of a master pianist who loses both of her arms in a tragic car accident [who] still knows how to play the piano (416); Paul Snowdon shares with us the equally tragic case of Raymond Blanc, the world s greatest chef, [who] knows how to make an excellent omelette. He loses his arms in a car accident, and is no longer able to make omelettes. However, he retains his knowledge how to make omelettes. (8) 15 And Carl Ginet notes that If a violinist cannot be said to have forgotten (or otherwise lost his memory of) what all the subtle right moves are for fingering a certain piece, but he is no longer able to make them all simply because of damaged fingers, then he still knows how to finger that piece. (8) 16 The strategy for resisting counterexamples to the sufficiency condition that having an ability to ϕ is sufficient for knowing how to ϕ was to strengthen the concept of having an ability (by adding further requirements

8 458 Knowing-How and Knowing-That that must be satisfied in order to have an ability). If it s harder to have an ability, then counterexamples to the effect that S has an ability to ϕ but doesn t know how to ϕ will be harder to come by. What of the necessary condition the claim that having an ability to ϕ is necessary for knowing how to ϕ? A similar strategy here would involve weakening the concept of having an ability. If it s easier to have an ability to ϕ, then counterexamples to the effect that S knows how to ϕ but lacks the ability to ϕ will be harder to come by. But, two points: first, it seems like it will be difficult for having an ability to ϕ to be both necessary and sufficient for S to know how to ϕ. This would require simultaneously weakening and strengthening the concept of having an ability. Second, the weakening required to handle these cases will be somewhat extreme: it will have to involve allowing S to have an ability to ϕ even if it is not possible (in a suitably restricted sense) for S to ϕ. Is this move plausible? Alva Noë (283) argues that it is, by distinguishing between having an ability to ϕ and being able to ϕ. It does not follow from the fact that one has an ability to ϕ that one is able to ϕ. For example, one might have an ability to play the piano, but not be able to play it (be unable to play it) because there just happens to be no piano around. Contingent facts about one s context can affect whether one is or is not able to ϕ without affecting whether one has an ability to ϕ. Some of these contingent facts can be facts about one s bodily constitution; loss of limbs might affect whether one is able to ϕ without affecting whether one has an ability to ϕ. 17 So, more argument is needed to show that, in the counterexamples, the subjects lack an ability to ϕ, despite knowing how to ϕ. 18 Carr offers a more principled argument against the claim that know-how is a matter of ability. You can sometimes interchange different descriptions of a single act successfully in know-how contexts. Sometimes, it seems, you cannot. To return to Carr s dancing/semaphore case, we cannot substitute a semaphore recital of Gray s Elegy for a performance of Improvisation No. 15 in the context The dancer knows how to bring about... while preserving truth. We can, however, substitute a performance of prestidigitation for a display of conjuring tricks in the context [T]he magician knows how to bring about... and preserve truth. What s the difference? According to Carr, there are two options: either know-how contexts sometimes do and sometimes don t allow for substitution. Or, in those seeming impermissible instances, the actions in question are not identical, because the actions are individuated by the purposes associated with them: when the dancer intends to perform Improvisation No. 15, the resulting act is not the same intentional act as a semaphore recital of Gray s Elegy. It would only be legitimate to substitute descriptions of one act for another if the acts thereby picked out were one and the same. But if the acts are individuated by the purposes associated with them, then the acts picked out are not one and the same.

9 Knowing-How and Knowing-That 459 A theory that unifies know-how contexts seems preferable. So, know-how contexts must posit relations between agents and intentional actions must take objects individuated by whatever is essential to intentional actions (for example, purposes and intentions). But ability contexts can take any old events as their objects, and they need not be individuated by intentional features. If an agent causes a state of affairs to obtain, then the agent causes any state of affairs co-extensional with that state to obtain as well. (Obviously, this is not so for intentional action.) Because know-how contexts take only intentional acts as their objects, but ability contexts can take any events as their objects, know-how is not a matter of ability. This, however, should be no solace to the intellectualist. For suppose that, when one knows how to do something, one has an attitude that essentially takes an intentional action as its object. When one knows that something is the case, on the other hand, one has an attitude that essentially takes a proposition as its object. Therefore, the attitude one has when one knows how to do something essentially takes on different kinds of objects than does the attitude one has when one knows that something is the case. This is evidence that knowing-how and knowing-that are distinct kinds. It is possible that, in essentially taking an intentional action as its object, an attitude ipso facto essentially takes a proposition as its object. 19 But this is a controversial claim. If it s false, then the difference in their essential objects entails a difference in kind between the attitudes themselves. And this is how Carr is standardly interpreted. 20 So, even if we reject the view that know-how is a matter of ability as Carr does there is still room for a distinction between knowing-how and knowing-that. 3. Stanley and Williamson s Intellectualism What then of our first option: the intellectualist option according to which knowing-how reduces to knowing-that or is a species of knowingthat? The most discussed contemporary representatives of this intellectualist view that knowing-how is a species of knowing-that are Jason Stanley and Timothy Williamson. According to Stanley and Williamson, the antiintellectualist majority relies too greatly on surface syntactic differences between know-how and know-that contexts. While it is true that sentences of the form, S knows how to ϕ differ from sentences of the form S knows that p, the differences should not be overstated. There are only two relevant syntactic differences between know-how contexts and knowthat contexts: know-how contexts contain embedded questions, and they are untensed. 21 But the first can be said for all of the following: (1) S knows whether p. (2) S knows when E happened. (3) S knows what water is composed of. (4) S knows who shot the clerk.

10 460 Knowing-How and Knowing-That And both can be said for: (5) S knows whom to call. All are clearly amenable to treatment in terms of knowing-that. 22 Treating know-how contexts accordingly delivers an initial (paraphrased) formulation of Stanley and Williamson s proposal: S knows how to ϕ just in case there is a way, w, such that S knows that w is a way to ϕ (425). This is only an initial formulation because it is, as stated, incomplete. They supplement the view to take into account two additional issues. First, S might know that w is a way for someone else to ϕ but not know how to ϕ, because S doesn t know that w is a way for her to ϕ. Furthermore, S might know that w is a way for her to ϕ without knowing how to ϕ, because S hasn t been presented with w in the right way. For example, I might know that one way for me to ride a bicycle is by moving my legs in a circular fashion, while keeping steadily balanced, hands on handlebars, etc.: in short, a complete physical description of the motions involved in riding a bicycle; call it Descriptive List L. But I still might not know how to ride a bicycle (in the relevant sense). To know how to ride a bicycle by knowing that w is a way for me to ride a bicycle, I must know that w is a way for me to ride a bicycle under, in Stanley and Williamson s term, a practical mode of presentation. That there is such a mode of presentation is clear from the fact I might be moving in exactly the way described by Descriptive List L, might know that Descriptive List L picks out a way for me to ride a bicycle, but not know that what I m doing is a way to ride a bicycle. I might fail to know this because I might fail to know that how I m moving is exactly what is described by Descriptive List L. What is the practical mode of presentation? It certainly seems that if w is presented to me by my actually instantiating w, then w is presented to me under the practical mode. But isn t this cheating? In order to be presented with w in this way, I must already know how to instantiate w. Here, Stanley and Hawthorne would, strictly speaking, be right: if I know that w is a way for me to ride a bicycle in this way then I do indeed know how to ride a bicycle. But it s not because I know that w is a way for me to ride a bicycle (under a practical mode of presentation) that I know how to ride a bicycle. Rather, I know that w is a way for me to ride a bicycle (under a practical mode of presentation) because I already know how to instantiate w. Know-how has ended up to be an irreducible part of the analysis, because the explication of the practical mode of presentation requires that I have a certain bit of know-how. As John Koethe puts the point, It appears, then, that entertaining the proposition that w is a way for one to F under a practical mode of presentation involves knowing how to instantiate w one s self. But, if so, then Stanley and Williamson s account of knowing how appeals to the very notion it seeks to explicate. (327)

11 Knowing-How and Knowing-That 461 It s not at all clear that to be presented with w under the practical mode, I must actually know how to instantiate w. It should be enough that I actually go through whatever operations are specified in w. As antiintellectualists are fond of pointing out, I don t know how to beat my heart, even though my heart can beat. However, when it comes to ways I might ride a bicycle, it looks like, in order to be presented with those ways under a practical mode, I will have to know how to instantiate those ways. Is there another way around this difficulty? In order for me to know how to ride a bicycle, I must not only know how to instantiate w, I must know that w (under the practical mode) is a way for me to ride a bicycle. For, suppose I know how to instantiate w, and I know, under the practical mode of presentation, that I am instantiating w, but I do not know that w is a way for me to ride a bicycle. That is, suppose I perfectly well know how to move my feet, hands, and body in such a way that, were I situated on top of a bicycle, I would successfully ride that bicycle (I wouldn t fall). I still might fail to know that moving this way is a way to ride a bicycle. If that is the position I am in then, it seems, I do not yet know how to ride a bicycle. Before I can know how to ride a bicycle, I must know that this way of moving a way I know how to move is a way to ride a bicycle. The worry was that, if being presented with w under the practical mode requires knowing how to use w, then Stanley and Hawthorne are strictly speaking correct knowing how to ride a bicycle requires knowing (under the practical mode) that w is a way to ride a bicycle. But they are incorrect that this makes know-how a species of knowing-that. The response on Stanley and Williamson s behalf is that all know-how will still require a corresponding bit of knowledge-that, if the following principle is true: if you know how to instantiate w, but don t know that w is a way for you to ϕ, then you don t yet know how to ϕ. So, if S knows how to ϕ, then S knows that w is a way for her to ϕ. This principle has some plausibility, for the same reason it s plausible that, if you know what it s like to be you, and it s also the case that you re so food-deprived that you re two minutes away from death (but you don t know this latter fact; you have no idea that it s true), then you don t really know what it s like to be two minutes away from death. If you were asked, What s it like to be two minutes away from death? you d respond, I have no idea what that s like. The principle is not by any means unquestionable. We can surely say, after finding out that what we ve been doing all along is a way to ϕ, Oh! I guess I knew how to ϕ all along! (Or, as you expire from hunger, Oh! I guess I knew exactly what it was like to be two minutes away from death! ) If we can truthfully say this, then it will be enough to know how to ϕ that we can be presented with what we re doing under the practical mode: we will not have to know in addition that what we re doing is a way to ϕ. If this is the right conclusion, then there must be some

12 462 Knowing-How and Knowing-That relevant differences between know-how contexts and other know-wh contexts, or else know-wh contexts have been wrongly interpreted in terms of knowing-that: the embedded question makes a difference. 4. Normative Arguments Let us distinguish between normative and non-normative arguments. Normative arguments for a difference in kind use relevant normative differences between two items to demonstrate that the two items are of different normative kinds. So, we might argue that keeping promises satisfies a categorical obligation whereas flossing at most satisfies a hypothetical obligation and, therefore, keeping a promise is of a different normative kind from flossing. Non-normative arguments for a difference in kind use non-normative features of the items being evaluated to demonstrate that the two items are of different non-normative kinds. Keeping a promise is quite a different kind of act than flossing: flossing, for example, can t be done without floss, but keeping a promise can (in most cases) be done without floss. There can also be normative and non-normative arguments for sameness in kind. Keeping a promise is of a normative kind with telling the truth, if both satisfy a categorical obligation. Flossing is of a normative kind with thanking one s host, if both satisfy a hypothetical obligation. Likewise, flossing is of a non-normative kind with going to the dentist, because both are activities that tend toward the maintenance of one s teeth. And keeping a promise is of a non-normative kind with redressing a past wrong, because both are activities in which one allows one s current actions to be guided by one s past actions. All the arguments we have looked at so far have been non-normative. We might oversimplify the traditional interpretation of Carr s argument as follows: what you are essentially related to when you know how to do something is only an act. What you are essentially related to when you know that something is the case is only a proposition. Therefore, knowinghow is of a different kind from knowing-that. Another simplistic nonnormative argument might go something like this: when you know how to do something, you essentially have an ability. When you know that something is the case, you essentially have a belief. To have a belief is not essentially to have an ability. Therefore, knowing-how is of a different kind from knowing-that. Both of these arguments make the anti-intellectualist case not on the basis of any distinctively normative features of the relevant kind of knowledge. Rather, the arguments are based on a distinction between the essential objects of the attitudes acts vs. propositions or the essential psychophysical manifestations of the attitudes abilities vs. beliefs. If these arguments work, they show not just that knowing-how does not reduce to knowingthat, but also that knowing-how does not reduce to truly-believing-that (or any other propositional attitude). Propositional attitudes are just not

13 Knowing-How and Knowing-That 463 the right sorts of things to deliver know-how. And this leaves the normative components of knowing-how and knowing-that irrelevant to the argument. It s not because there is a different sort of positive evaluation in the case of knowing-how and knowing-that. It s because the things being evaluated in the two cases are different. Intellectualists argue as you might expect, given the sorts of arguments they re responding to by trying to demonstrate that the non-normative components of knowing-how can be reduced to the non-normative components of knowing-that. So, recall Ginet s argument that I exercise (or manifest) my knowledge that one can get the door open by turning the knob and pushing it (as well as my knowledge that there is a door there) by performing that operation quite automatically as I leave the room; and I may do this, of course, without formulating (in my mind or out loud) that proposition or any other relevant proposition, (7) thus demonstrating that the act of opening a door can manifest belief even if the belief is not actively contemplated. 23 In general, it seems like the most immediate anti-intellectualist case has to do with the sorts of activities that knowing-how and knowing-that involve. But this debate, even if it can ultimately be resolved in favor of one side or the other, may not get to the heart of the matter. The reason is that it doesn t engage with the knowing part of knowing-how and knowing-that; it doesn t engage with the normative components of the respective attitudes. This might be a problem if it turns out that, though knowing-how and knowing-that are distinct non-normative kinds, they are in fact of a single normative kind. (Keeping promises and telling the truth, after all, are of distinct non-normative kinds, but are arguably of a normative kind.) A happy exception to this trend is Katherine Hawley s Success and Knowledge-How, in which she proposes a direct analysis of knowing-how according to which it is a matter of successful action plus warrant. Though she explicitly does not prejudge whether one form of knowledge is a sub-type of the other (20), if such an analysis was more amenable to one side than the other, that would seem of particular epistemological interest. Most normative arguments in the literature favor, if not intellectualism, at least the view that knowing-how and knowing-that are of a normative kind. For example, Ernest Gellner argues that knowing-how is not just the ability to do something, but the ability to do it according to the right rules. Knowing how to argue is not just the ability to draw conclusions from premises, but to do so according to valid inference rules. Knowing-how and knowing-that, then, are of a normative kind (though it might turn out, of course, that the respective rules are radically different when it comes to evaluation of know-how and know-that 24 ). Or, consider Snowdon s recent normative argument: The question is whether we accept that there is a distinction between knowing how and

14 464 Knowing-How and Knowing-That merely being right as to how without in fact knowing how. I think that the answer is that there is such a distinction. We have, for example, the concept of guessing how to do something; and that allows a place for someone being right as to how, without actually knowing how (21). There is no explicit follow-up here to the effect that what might bring us from being merely right as to how to actually knowing-how is the addition of bits of knowing-that. The point seems to be that knowing-how and knowing-that are of a kind with respect to the need for something like justification. And, if this is right, then know-how and know-that are both properly deserving of the name knowledge. These two arguments are in a broadly intellectualist tradition. But there seems no principled reason why anti-intellectualist arguments can t engage with normative concerns. As Daniel Hutto observes, in explicating the anti-intellectualist position, practical knowledge is not representational [in the way propositional knowledge is]... [W]e cannot say, even in principle, how we achieve such feats by articulating the set of tacit rules or maxims followed since there are none. 25 Even so, such knowledge is normative in that our performances are still subject to assessment in terms of achievement or failure. (390, emphasis added) Despite this, an anti-intellectualist might argue for her anti-intellectualism precisely by arguing that knowing-how, unlike knowing-that, is not normative or perhaps, while normative, its normative status differs dramatically from the normative status of knowing-that. Perhaps, though knowing-how has obvious analogues of belief and truth (action and success, respectively), there is no analogue of justification when it comes to knowing-how; or perhaps knowing-how can be Gettiered and remain know-how; or perhaps merely having a true belief that w is a way for me to ϕ (presented under the practical mode) suffices for knowing-how. Some of these arguments might even grant that know-how is ultimately propositional. Such may not be to the liking of many anti-intellectualists. But it will deliver the result that know-how is significantly different in kind in normative kind from knowledge-that. If the debate is about whether knowing-how is a kind of knowing-that, then these arguments should suffice (if sound). As Hawley notes, knowledgehow is interesting qua species of knowledge (19). If the debate was all along about the kinds of psychological states involved in knowing-how and knowing-that, or the proper objects of knowing-how and knowing-that, then the normative arguments won t help. But, then, it s also less clear why the debate is of such epistemological interest. 5. Conclusion Suppose anti-intellectualism is correct, and suppose you know how to ride a bicycle but you lack all of the attendant knowledge-that intellectualists

15 Knowing-How and Knowing-That 465 want to claim you have when you know how to ride a bicycle. But suppose that you come to have all of that attendant knowledge-that anyway (antiintellectualists don t claim that you can t know that w is a way for you to ride a bicycle just that such knowledge isn t necessary for you to know how to ride a bicycle). What should we say happens to your bicycle-riding know-how? Is all your new knowledge-that just irrelevant? Charlotte Katzoff thinks it is not. Commenting on an expert mechanic, she remarks: Although his actual driving performance may be identical with that of his unversed counterpart, in such circumstances, I think, we would be inclined to count his special expertise as part of his knowing how to drive. This would be an instance of his driving being dependent upon that knowledge, albeit not necessarily. (65ff) If she s right, then your attendant propositional knowledge about how to ride a bicycle seems like it would likewise be part of your knowing-how to ride a bicycle. But what does it mean for knowledge-that to be part of know-how? A natural interpretation is that the addition of the propositional knowledge makes it the case that you know how to ride a bicycle better than you did before. Even if it is possible to know how to ϕ without any important attendant knowledge-that, one s know-how is improved with the addition of relevant pieces of knowledge-that. It s not simply that one s total state is superior: one s know-how is better. This suggests that there are various grades of know-how. Let us distinguish them according to how they relate to the other concepts we have here discussed. Consider the following ordering of potential conditions on knowing how to ϕ, most of which have come up in our discussion thus far (the first three are possible specifications of what it is to be able to ϕ or to have an ability to ϕ): S knows how to ϕ only if: (1) it is possible that S ϕ. (2) were S to try to ϕ, S would ϕ. (3) were S to try to ϕ in a suitable context, S would ϕ. (4) S is able/has the ability to ϕ particularly well. (5) S knows that w is a way to ϕ. (6) S knows that w is a way for her to ϕ. (7) S knows why w is a way to ϕ. We can specify grades of know-how depending on which of these conditions are necessary. How many, and which, of these grades of knowhow will square with our ordinary use of the term is open for debate. Nonetheless, a very low grade of know-how might have it that S knows how to ϕ just in case it is possible that S ϕ. A slightly higher grade of knowhow might have it that S knows how to ϕ just in case, were S to try to ϕ, S would ϕ. We might call these low grades of know-how, following Ernest Sosa on low grades of knowledge-that, animal know-how. 26 A maximal

16 466 Knowing-How and Knowing-That grade of know-how might require S have it all that S knows how to ϕ only if, (1) were S to try to ϕ, S would ϕ; (2) S knows that w is a way for her to ϕ; and (3) S knows why w is a way to ϕ. We might call this, borrowing Sosa s label for high-grade knowledge-that, reflective know-how know-how that requires an epistemic perspective on one s own body of knowledge. 27 Sosa thinks that reflective knowledge is a better knowledge than the lower grades ( Reflective Knowledge 422). And, Keith Lehrer points out, epistemologists are not concerned with the sort of knowledge attributed to animals, small children, and simple machines that store information, such as telephones that store telephone numbers. Such animals, children, or machines may possess information and even communicate it to others, but they do not know that the information they possess is correct nor are they in a position to use the information in the requisite forms of ratiocination. (11) When it comes to knowledge-that, epistemologists should care most about the higher, more reflective grades not the knowledge-that of animals and supermarket doors. Perhaps we should care most about the more reflective grades of know-how, as well. Acknowledgment Many thanks to Tamar Gendler, Matt McGrath, Brian Weatherson, and an anonymous referee for Philosophy Compass. Short Biography Jeremy Fantl received his Ph.D. from Brown University and is currently assistant professor of philosophy at the University of Calgary. He works primarily in epistemology in particular, on the a priori and the structure of justification and has publications in The Philosophical Review, Pacific Philosophical Quarterly, Philosophy and Phenomenological Research, Nous, and the Canadian Journal of Philosophy. His projects include a defense of epistemic infinitism, the view that justification is a matter of having an infinite series of non-repeating reasons in support of a belief. He has also published several papers with Matthew McGrath developing an argument that knowledge and justification can come and go with changing stakes, even if one s evidence remains constant (one version of which has come to be known as subject-sensitive invariantism ). Their book on the same subject Knowledge in an Uncertain World is under contract with Oxford University Press and due out in Notes * Correspondence address: Department of Philosophy, University of Calgary, 2500 University Drive, NW Calgary, Alberta, T2N 1N4, Canada. jfantl@ucalgary.ca.

17 Knowing-How and Knowing-That For classic treatment of the distinction between propositional knowledge and knowledge by acquaintance, see Russell. For the most influential discussion of the distinction between propositional knowledge and practical knowledge, see Ryle, Concept of Mind. For a textbook treatment of all three, see Hospers 39. It is not clear that at least the ancients made the relevant distinctions. For example, according to David Runciman, we should beware of any initial assumption that Plato is clearly aware of a distinction between knowing that, knowing how, and knowing by acquaintance (13). This worry is shared by Gregory Vlastos, though John Gould takes Plato as committed to a conception of knowledge as practical. See Vlastos for a criticism of Gould s position. 2 One application of results in this debate concerns contemporary philosophy of mind and, in particular, the so-called knowledge-argument according to which, because a scientist fully informed of all physical facts would learn (come to know) something new when she experienced red for the first time, what it s like to experience red must not be reducible to any physical facts (see Jackson for a classic formulation of the argument). Some (e.g., Lewis; Nemirow) try to undercut the argument by conceiving of the scientist s new knowledge as practical rather than propositional. Others (e.g., Conee) construe the scientist s new knowledge as knowledge by acquaintance. 3 John Bengson, Marc Moffett, and Jennifer Wright define anti-intellectualism as the positive thesis that know-how is a matter of having some ability. Here, anti-intellectualism is defined as the negative thesis that intellectualism is false. The negative construal does more justice to the traditional belief (see, e.g., Stanley and Williamson 416) that Ryle had two projects: (1) undercutting intellectualism; and (2) offering a positive analysis of know-how. Defining anti-intellectualism positively as a view about know-how leaves unclassified those views that deny intellectualism but do not share the positive view about know-how. Those views should surely be classified as anti-intellectualist. 4 Among the defenders of intellectualism are Bengson and Moffett; Snowdon; Stanley and Williamson; Katzoff; Ginet; Brown, Knowing How ; Reply to Brett ; Vendler. Among defenders of weak anti-intellectualism or, at least, the denial of intellectualism are Noë; Rumfitt; Koethe; Lewis; Nemirow; Carr, Logic of Knowing How ; Brett; Ryle, Concept of Mind; Knowing How and Knowing That. Among defenders of strong anti-intellectualism are Hetherington; Roland; Hartland-Swann, Logical Status ; Knowing That. 5 See Hartland-Swann ( Logic of Knowing Jones ) for an argument that, unlike knowing-that, knowing a person does not reduce to knowing-how (say, to recognize a person). See Ammerman and Hartland-Swann ( Knowing That ) for an early and much-cited interchange on Hartland- Swann s reduction; Roland for an early critique of Hartland-Swann s reduction; and Beck for a critique of Ryle, Hartland-Swann, and Roland. 6 The claim that knowing-that is ultimately dispositional is derived from Ryle and simply assumed by Hartland-Swann ( Logic of Knowing Jones ) and Roland. 7 As Ryle puts the argument in The Concept of Mind, The crucial objection to the intellectualist legend is this. The consideration of propositions is itself an operation the execution of which can be more or less intelligent, less or more stupid. But if, for any operation to be intelligently executed, a prior theoretical operation had first to be performed and performed intelligently, it would be a logical impossibility for anyone ever to break into the circle (30). 8 See, for example, Ginet; Stanley and Williamson. See Noë 281 for a reply to Ginet (and Stanley and Williamson s use of Ginet s example). 9 The first is a second regress argument: if there were two processes, there would have to be some link that gets you from the contemplation (the theory ) to the act (the practice ). But then it would have to be kith and kin ( Knowing How and Knowing That 213) to both theory and practice: it needs to have two aspects, a contemplating kind of aspect and a doing kind of aspect, which means a further process would be needed to bridge those two aspects. The alternative is that one process can both contemplate and execute, and that s just to say that there is not a further theoretical process involved in intelligent action. If Ginet s objection works against Ryle s first regress argument, it works against this one as well. 10 This, for example, is how he is interpreted by Stanley and Williamson. But an ability to what? There are two issues to consider. We might ask whether knowing-how to ϕ requires that you have the ability to ϕ or the ability to ϕ well or the right way. D. G. Brown distinguishes