Featured Video Play Icon

Knowledge-How, Abilities, and Questions


Joshua Habgood-Coote
 (Universities of St. Andrews and Stirling)

[Jump to comments]

 

Introduction

Knowing-how seems to be a distinctively practical kind of knowledge. Yet according to the standard semantics for knowledge-how ascriptions, to know how to do something is to stand is some relation to a set of propositions about how to do it. Intellectualists about knowledge-how take their lead from the semantics of knowledge-how ascriptions and claim that knowledge-how is a species of propositional knowledge. As a consequence they have trouble explaining the practical properties of knowledge-how, usually appealing to the somewhat obscure notion of a practical way of thinking.[1] By contrast, Anti-Intellectualists give priority to the practical properties of knowledge-how, claiming that knowledge-how to V[2] is a kind of ability or capacity to V.[3] Since abilities are generally relations to activities or action-types rather than propositions, they have the parallel problem of making their view compatible with linguistic theory.

In this paper, I explore a novel compromise between these two views, which I call the Interrogative Capacity view. According to this view, knowing how to do something is a certain kind of ability to generate answers to the question of how to do it. This view combines the Intellectualist thesis that knowledge-how is a relation to a set of propositions with the Anti-Intellectualist thesis that knowledge-how is a kind of ability. I argue that this view is uniquely well placed to defuse the tension between semantic theory and the practicality of knowledge-how, and that it is able to elucidate the relationship between knowledge-how, propositional knowledge, and abilities.

The plan of action is as follows. In the first section, I set up the debate between Intellectualists and Anti-Intellectualists, and sketch out the tension between the linguistic evidence, and the idea that knowledge-how is a distinctively practical kind of knowledge. In the second section, I distinguish between weak and strong versions of both Intellectualist and Anti-Intellectualist claims. Distinguishing these claims opens up space for a compromise position which takes knowledge-how to be both an ability and a relation to a set of propositions. In the third section, I consider one such view — which I will call the Interrogative Capacity view — according to which knowledge-how is an ability to answer a question on the fly. I show how this view defuses the tension set out in §1, and elucidates the relation between knowing how to V, propositional knowledge, and the ability to V. In the fourth section, I consider whether the Interrogative Capacity account might be extended to other kinds of knowledge, and contrast this view to the accounts of skill offered by Dickie (2012) and Stanley and Williamson (2016).

§1

The distinction between knowing-how and knowing-that seems to be a platitude of folk epistemology. However, as Stanley and Williamson (2001) point out, semantic theory provides a good case for thinking that knowledge-how is a species of propositional knowledge. In English, knowledge-how ascriptions involve an interrogative complement – ‘how to V’ – which according to standard models of the semantics of interrogatives should be understood as denoting a question, which we can understand to be the set of propositions which are possible answers[4] that question. For example, on the standard account the interrogative phrase ‘how can one open the door?’ denotes a set of propositions about how to open the door, such as one can open the door by jiggling the key in the lock. That standard semantic treatments treat the complement in a knowledge-how ascription as denoting a set of propositions seems like good evidence that knowledge-how itself is a relation to a set of propositions. It seems like a short step from this claim to the idea that knowing how to do something is a matter of having standing knowledge[5] of answers to the embedded question, making knowledge-how a species of propositional knowledge.

One theme in the Anti-Intellectualist fightback to this argument is the idea that knowledge-how is a distinctively practical kind of knowledge, with an epistemic profile distinct from that of knowledge-that.[6] The practicality of knowledge-how has three central features:

  1. Necessity: knowledge-how is a necessary condition for intentional action (Setiya 2008, 2011), (Stanley and Williamson 2001: 415-16, 432-3), (Stanley 2011 188-90), (Hornsby 2016: 8-10);[7]
  2. Directness: knowing-how is exercised directly in intentional action, not via some intermediate act of mental contemplation (Ryle 2009: 17-20), (Stanley 2011: C1);
  3. Flexibility: knowing-how involves an ability to react flexibly to a wide range of possible situations (Hornsby 2011 89-95), (Stanley 2011: 181-5), (Wiggins 2012 §1-7), (Fridland 2013).

Exactly how to formulate these conditions is a tricky question, but some version of each idea appears to be common ground between Intellectualists and Anti-Intellectualists.[8] To see why, consider that the subject-matter of the knowledge-how debate is not just all of the knowledge picked out with ‘knows’ and ‘how’ (no-one thinks that knowing how coffee smells is a kind of know-how (Hornsby 1980: 84), (Rumfitt 2003: 166), (Glick 2011: 426-9)). This set of practical properties provide a neat non-linguistic way to pick out the kind of knowledge which Intellectualists and Anti-Intellectualists disagree about.

The fact that knowledge-how has these practical properties poses a challenge for Intellectualists. Other types of propositional knowledge do not seem to have these properties, so the burden of proof is on Intellectualists to show that a propositional theory can explain the practicality of knowledge-how.[9]  For example, if there is some metaphysical connection between knowledge-how and intentional action, then an Intellectualist ought to be able to explain why the relevant kind of propositional knowledge has this special connection with intentional action (Setiya 2011: 290-3). To explain the practicality of knowledge-how Stanley and Williamson appeal to the idea that first-person thought has a special connection with action, and claim that knowledge-how involves a distinctive practical mode of presentation (Stanley and Williamson 2001: 429-30), (Stanley 2011: 109-10, 182-3). Critics of Intellectualism have contended that the notion of a practical way of thinking is at best obscure, challenging its ability to explain the practicality of knowledge-how (Glick 2015).[10]

Anti-Intellectualists take the practical features of knowledge-how as their starting point, claiming that in order to explain the practicality of knowledge-how, one needs to identify knowledge-how with some kind of capacity or ability. Since capacities to act intrinsically have a tight relationship to action, the hypothesis that knowledge-how is a kind of ability seems well placed to explain the distinctive practicality of knowledge-how. Consider swimming: the ability to swim is exercised directly in swimming, is a necessary condition for intentionally swimming, and can manifest in a wide range of different kinds of swimming. This means that being able to swim looks like a good candidate for being equated with knowing how to swim. However, identifying knowledge how to V with the ability to V creates a serious problem when it comes to offering a semantics for knowledge-how ascriptions. Abilities to act are relations to activities or action-types rather than propositions — the ability to swim is a relation to swimming — so the Anti-Intellectualist view looks to be incompatible with the standard semantics for knowledge-how ascriptions.

Anti-Intellectualists who equate knowledge-how with the ability to V also face the challenge of isolating a kind of ability which is plausibly necessary and sufficient for knowledge-how. There appear to be many situations in which it seems right to say that those who know how are not able and vice versa (see Bengson and Moffett 2011b for an overview). In order to offer an account of knowledge-how in terms of ability, Anti-Intellectualists need to offer an account of the kind of ability which is identical to knowledge-how.[11]

We find ourselves at an impasse. Anti-Intellectualists can deal with the practical properties of knowledge-how, but struggle to offer a plausible semantics for knowledge-how ascriptions. On the other hand, Intellectualists can offer a plausible semantics, but have difficulty explaining the practical properties of knowledge-how. In the next section, I articulate a compromise between these views, and offer a preliminary argument for this view based on habitual knowledge-how ascriptions. In the third section I offer a more systematic argument for this view, showing how it can defuse the tension between linguistic theory and the practicality of knowledge-how.

§2

In thinking about the relation between Intellectualism and Anti-Intellectualism we need to carefully distinguish claims about the object (or relatum) of knowledge-how from claims about the nature of the knowledge-how relation. Claims about what knowledge-how is knowledge of are often confused with claims about what kind of epistemic relation is involved in knowing how. Reflecting on this distinction helps us to clarify the disagreement between Intellectualism and Anti-Intellectualism, and opens up space for a compromise between these two positions.

Ephraim Glick (2011 §4) points out that the Intellectualist claim that knowledge-how is a species of propositional knowledge can be understood in two ways:

Weak Intellectualism: Know-how is knowledge that has a proposition as a relatum.

Strong Intellectualism: Know-how is theoretical knowledge.

Weak Intellectualism is just the claim that the object of knowledge-how is propositional. By contrast, Strong Intellectualism encompasses both the claim that the object of knowledge-how is propositional, and the claim that the relation involved is the same theoretical knowledge relation which is found in knowing that p (i.e. JTB plus Gettier proofing).[12] Weak Intellectualism is compatible with a number of views about the knowledge-how relation. For example, one might think that knowledge-how has as its object the set of propositions which answer the question ‘how to V?’, but think that it is a distinctively practical knowledge relation to that set of propositions (Glick 2011) (Cath 2015).[13]

This distinction is crucial for understanding the significance of the linguistic evidence. The linguistic argument which I sketched in the previous section was focused on the semantics of the complement in knowledge-how ascriptions. Glick points out that this evidence only establishes that the object of knowledge-how is propositional, getting us to Weak Intellectualism. One needs to have a further argument to establish the Strong Intellectualist claim that knowledge-how also involves the theoretical knowledge relation.[14]

The Anti-Intellectualist claim that knowledge-how is not a species of theoretical knowledge can also be understood in two ways:

Weak Anti-Intellectualism: know-how involves a relation other than theoretical knowledge.

Strong Anti-Intellectualism: know-how has a non-propositional relatum, and involves a relation other than theoretical knowledge.

Weak Anti-Intellectualism is the claim that the knowledge-how relation is something other than theoretical knowledge. Strong Anti-Intellectualism endorses both the claim that the knowledge-how relation is a non-theoretical one, and the claim that knowledge-how is a relation to something other than a proposition. Just as Weak Intellectualism is compatible with a number of different views about the knowledge-how relation, Weak Anti-Intellectualism is compatible with various views about the object of knowledge-how. Weak Anti-Intellectualism is compatible with the claim that knowledge-how is a relation to a set of propositions, so long as that relation is something other than theoretical knowledge. For example, the view that knowledge-how is a practical knowledge relation to a set of propositions comes out as being Weakly Anti-Intellectualist on this framework.

Just as the linguistic evidence supports only Weak Intellectualism, the idea that knowledge-how is a distinctively practical kind of knowledge supports only Weak Anti-Intellectualism. The argument I sketched above rested on the idea that one can explain the practicality of knowledge-how by identifying it with a kind of ability, since abilities have the right kind of practical properties to explain the practicality of knowledge-how. At this point it is indubitably tempting to claim that knowing how to V is just the ability to V, which is a Strongly Anti-Intellectualist view. However, the view that knowledge-how is an ability is compatible with various views about what kind of ability it is. For example, Edward Craig (1990 C17) suggests identifying knowing how to do something with the ability to teach others, which is quite some distance from being able to do the activity oneself. In fact, the majority of Anti-Intellectualists steer clear of identifying knowledge how with the plain ability to act. Setiya argues that knowledge-how is the ability to enact intentions to V (Setiya 2008, 2011), and Wiggins suggests that knowing how is a sui generis intellectual capacity relating to V-ing (Wiggins 2012). Even Ryle contrasts knowing how with the ability to do something habitually, suggesting that he doesn’t take just any kind of ability to be sufficient for knowing how (Ryle 2009: 30-3), (Hornsby 2011: 81-82). Because there are various views about what kind of ability knowing-how might be, the argument from practicality only entails the Weak Anti-Intellectualist claim that the knowledge-how relation is abilitative, leaving open a range of possibilities about what the object of that relation is.

Let’s take stock. We’ve distinguished two kinds of disagreement about the nature of knowledge-how: disagreement about whether the object of knowledge-how is propositional or non-propositional, and disagreement about whether the relation is theoretical knowledge or something else. Putting these disagreements together gives us the following picture of the logical space in this debate:

Object
 

 

 

Relation

Propositional (i.e. that w is a way in which one can V) Non-Propositional (i.e. Activity, way of acting, predicate)
Theoretical knowledge (i.e. JTB+) Strong Intellectualism
Something other than theoretical knowledge (i.e. practical knowledge, understanding, ability) Weak Intellectualism; Weak Anti-Intellectualism Strong Anti-Intellectualism

Table 1: Logical space in the knowledge-how debate.

 

The argument from linguistics pushes us from the right column to the left column, suggesting that the object of knowledge-how is propositional, whereas the argument from practicality pushes us down from the top row to the bottom, suggesting that knowledge-how involves some kind of ability. Putting the two arguments together suggests that knowledge-how involves an abilitative relation to the propositions which answer a question of the form ‘how to V?’.[15]

What would such a view of knowledge-how look like?. There are several ways to be related to a proposition (or a question) by an ability. For example one might be able to remember that Wayne came for dinner last week, or able to ask in German where the toilet is. Of course, these abilitative relations to propositions are implausible as accounts of knowledge-how. However, there is an abilitative relation which does provide a plausible account of knowledge-how: the able-to-answer relation (where answering is understood as correctly answering). According to this view knowing how to V is standing in the able-to-answer relation to the question how to V?, where one exercises that ability by generating propositional knowledge which answers the question how to V?. Let’s call this view the Interrogative Capacity view of knowledge-how.

We can get a grip on the idea of a capacity to generate answers to a question by considering habitual knowledge-how ascriptions. Habitual sentences – like ‘Jane runs’ – express a generalisation about an agent fulfilling a certain verb across a range of situations. This kind of sentence is to be distinguished from a sentence like ‘Jane is running’ which makes a singular claim about Jane being engaged in running right now. Typical knowledge-how ascriptions are not habituals. Consider:

1)         Bilal knows how to spell ‘DOG’.

This sentence makes the singular claim that Bilal has standing knowledge of the answer to the question ‘how does one spell “DOG”?’. However, there are some knowledge-how ascriptions which are plausibly habituals.[16] Consider:

2)         Elsa knows how to calm people down.

Sentence 2) is naturally understood as saying that in situations in which someone needs to be calmed down, Elsa knows how to calm the relevant person down. On this reading, rather than ascribing a single piece of standing knowledge with a general content, this ascription makes a generalisation about Elsa’s possessing different pieces of situation-specific knowledge in different situations.[17] What might make this generalisation true? Well, it might be that Elsa has some sneaky way to acquire knowledge of how to calm people down (maybe she has an app). However, the simplest hypothesis is that the sentence is made true by Elsa’s ability to generate situation-specific knowledge in the necessary situation. If the generalisation is made true by this kind of epistemic capacity, then the sentence is made true not by the agent’s standing knowledge of answers to the questions but by her ability to answer the question.

We should be careful not to overstate the significance of this linguistic evidence. The slipperiness of habituals means that it is difficult to get a clear sense of which knowledge-how ascriptions are habituals, but I think that it is likely that habitual knowledge-how ascriptions are a rather limited class. Rather than providing straightforward evidence for the Interrogative Capacity view, the significance of this evidence lies in demonstrating the structural point that the standard semantics for interrogative complements is compatible with knowledge-how being a certain kind of ability.

§3

In the last section I suggested that we might take knowledge-how to be an ability to generate knowledge of answers to the questions of the form how to V?.  However, knowledge-how is not identical with just any kind of ability to answer a how-to question. Someone who has read a book on skiing is in a sense able to answer the question of how to ski, but we don’t want to say that they thereby know how to ski. Equally, a skilled skier might be unable to articulate their knowledge in linguistic terms, meaning that there is a sense in which they are not able to answer the question of how to ski. Just as the simple ability account faced the problem of isolating a kind of ability to V which is plausibly identified with knowledge-how, the this account faces the problem of isolating a kind of ability to answer questions which is plausibly identified with knowledge-how. In this section, I address this challenge by isolating a kind of interrogative capacity which is plausibly identified with knowledge-how, which I will call the ability to answer a question on the fly.

I take the an ability to answer a question on the fly to involve three ideas: an account of the kind of answers which are generated by this ability, a picture of kinds of situations in which the ability is exercised, and an account of the way in which these answers are generated. Let’s take these ideas in turn.

First, let’s consider the kinds of answer which are associated with knowing how to do something. There are many ways in which one may answer a question like ‘how to swim?’, ranging from coarse-grained propositions like ‘by splashing about in the water’ to fine-grained propositions which specify an exact technique for swimming in a particular situations (Fridland 2013). In order to respect the connection between knowledge-how and a capacity to react to the particular situation, I will take an ability to generate answers on the fly to be an ability to generate fine-grained answers to the question how to V? which specify a method which is sensitive to the particular situation.

Secondly, we need to get clear on what kinds of situations the ability relates to. There are also many different kinds of situations in which one might be able to generate practical answers: a skiing teacher may be able to generate answers to the question of how to perform various moves in the classroom, but not out on the slopes. Since knowledge-how is a kind of practical knowledge, I will take an ability to answer on the fly to involve the ability to generate answers in practical situations.

Finally, let’s consider the manner by which the relevant question is answered. Here I want to take the idea of answering on the fly seriously. Rather than the paradigm being generating an answer before action begins, I want to pick out a kind of answering which goes on throughout the process of acting. The kind of process I have in mind is one Ryle describes by saying that a skilled mountaineer walking in difficult conditions “is concomitantly walking and teaching himself how to walk,” (Ryle 2009: 30). Working out how to do something need not be something which one can only do in one’s head. It can involve working through the activity, picking up relevant propositional knowledge as one goes along and setting relevant feedback loops into action. To get a grip on this idea, consider the way in which one might solve a difficult maths problem. Although sometimes it might be possible to just ‘see’ what the method for solving a problem is, the more usual way is to work out how to answer the problem as one goes along. Depending on the problem, one might either split it into sub-problems and work on them in turn, or one might just try out different techniques and see what sticks. I suggest that we should think of the problem-solving strategies involved both in cases in which one just sees the solution, and in case in which one works out how to solve the problem as one goes along as involving the exercise of the ability to generate knowledge of how to solve the problem.

Putting these three ideas together, we have the idea that knowing-how involves an ability to generate fine-grained propositional knowledge in practical situations through the process of acting.

It is important to be clear that the ability to generate answers on the fly comes apart from the ability to express those answers and communicate them to others. The kind of answering which we are interested in is a semantic notion, and not the speech act of answering. There are a number of reasons why someone who has the ability to answer a question on the fly might not be able to articulate it in explicit linguistic terms. For one thing, the agent may not be in the right kind of practical situation to generate an answer to the question. A climber who has learnt to scale a difficult wall might be unable to generate knowledge of how to climb the wall without having the holds in front of her as a prompt. Even in the relevant kind of practical situation, an agent may only be express their knowledge by employing indexicala (Stanley and Williamson 2001: 428-9). I might only be able to express and communicate my knowledge of how to tie a Cat’s paw knot by making one and saying ‘this is the way to tie it’. I want to take seriously the idea that using this kind of demonstrative can express knowledge of a method for making the knot, but I want to treat it in a rather different way to Stanley and Williamson. Stanley and Williamson take this kind of demonstrative to pick out a general method for swimming involving a method-type (2001: footnote 29). On the view we’re considering, since answers are fine-grained propositions, we can say that the way picked out is a method-token. This is good feature of this view because whereas it is clear how we can secure demonstrative reference to method-tokens by pointing at them, it is difficult to see how to secure reference to a method-type by pointing at a token which instantiates many different types.[18] Even allowing that one can express knowledge-how via demonstratives, there might be cases in which an agent who is able to generate answers is not able to express those answers. It is hard to point at yourself while you are swimming.

With these clarifications in place, we can formulate a version of the Interrogative Capacity view of knowledge-how:

S knows how to V if and only if for some contextually relevant class of practical situations {F1, F2 …} S has the capacity to generate fine-grained answers to the question ‘how to V in Fn?’ in the process of acting.[19]

I will call this view the Interrogative Capacity view, but we should bear in mind that strictly speaking it is only one version of the Interrogative Capacity view. There might be other views which claim that knowledge-how is a different kind of capacity to answer a question. On the framework above, this view endorsed both Weak Intellectualism, because it claims that knowledge-how is a relation to the set of propositions which answer the question ‘how to V?’, and Weak Anti-Intellectualism, because it claims that the knowledge-how relation is an abilitative one, rather than a theoretical knowledge relation.

The Interrogative Capacity view has two key philosophical benefits. First, it resolves the tension between semantic theory and the practicality of knowledge-how, which led to an impasse between standard Strong Intellectualist and Strong Anti-Intellectualist views. Secondly, this view elucidates the relationship between knowledge-how, propositional knowledge, and the ability to do

In the first section, I argued that an account of knowledge-how ought to respect both the standard propositional semantics for knowledge-how ascriptions, and the intuition that knowledge-how is a distinctively practical kind of knowledge. The Interrogative Capacity view meets both criteria.  This view takes knowledge-how to be a relation to a question — the question how to V? — and understands that question as being identical to a set of possible answers. Hence, it is quite compatible with the standard semantics for knowledge-how ascriptions, which treat the interrogative complement as denoting a set of propositions. However, on the interrogative capacity view, the relation involved in knowledge-how is not standing knowledge, but a certain kind of abilitative relation. This means that the view is also able to vindicate the intuition that knowledge-how is a distinctively practical kind of knowledge. Above, I split this idea into three parts: directness, necessity, and flexibility. Let’s take these ideas in turn.

The idea that knowledge-how is exercised directly in action is explained by the fact that answering a question on the fly involves answering a question by actually engaging in the relevant activity. One of the distinctive features of answering a question on the fly is that the process of doing and the process of answering are intertwined. On this view, if I know how to dance, then I exercise the capacity to answer questions about how to dance directly in dancing. Since this view posits no intermediary between knowledge-how and action, it seems well-placed to explain the idea that knowledge-how is exercised directly in action.

The interrogative capacity view also neatly explains the connection between knowledge-how and intentional action. If we think about doing something intentionally as involving answering a practical question, then it would be natural to think that just as the standard case of acting involves the agent having the ability to act, the standard case of answering a question involves the agent having the ability to answer that question. There might be tricky cases in which someone answers a question by luck without possessing an ability, but this just flags up the issue that it is unclear whether the connection between knowledge-how and intentional action is strict necessity or something weaker (see footnote 7).

We can also explain the flexibility of knowledge-how by appealing to the fact that an interrogative capacity can produce different propositional knowledge for different situations. Whereas Intellectualists identify knowing how to do something with a fixed body of propositional knowledge, on the Interrogative Capacity view knowledge-how is identified with an ability to generate an expanding set of situation-specific propositional knowledge.[20] Someone exercising an Interrogative capacity is both doing and learning. If we think about this kind of capacity as being keyed in to the particular features of the practical situations at hand, then it will be natural to think about this expanding body of knowledge as being fairly heterogeneous, generating different knowledge to meet the needs of different situations.

The Interrogative Capacity view can also naturally explain the connections between knowledge-how, propositional knowledge, and the ability to do. Although the universal claims that knowledge-how always entails propositional knowledge or the ability to do are controversial, it is uncontroversial that in many cases knowing how is connected with both being able, and knowing facts about the activity in question. It would be good to have explanations of these connections. According to this view, knowledge-how will produce propositional knowledge as a product of its exercise. When someone who knows how to swim exercises their ability to generate answers to the question ‘how to swim?’, the result will be a piece of propositional knowledge about how to swim. This propositional knowledge may be temporary, and need not ever be consciously articulated, but it is plausible that at least some of the time it will make its way into an agent’s standing knowledge. This explains why knowledge-how seems to be so often associated with propositional knowledge of methods. This view also predicts that knowledge-how has a close connection with a certain kind of ability to act successfully. Since generating an answer to a practical question on the fly will involves successfully engaging in the relevant activity, an ability to generate practical answers will (assuming relevant enabling conditions are in place) entail an ability to successfully pull off the relevant kind of action.

§4

In this final section, I compare the Interrogative Capacity account to some related views that appeal to abilities in order to understand knowledge. I first consider whether we might think that other kinds of knowledge might be constituted by abilities to answer questions. I then compare the Interrogative Capacity view to the views of skill offered by Dickie (2012) and Stanley and Williamson (2016), and argue that it is preferable to both of these views.

Having established that it is possible to view knowledge-how as an ability to answer questions, one might wonder whether one might hold similar views of other kinds of knowledge. There are some proposals in the literature which understand knowledge-that in terms of ability (Hyman 1999) (Hetherington 2011). For Hyman knowledge-that is the ability to employ p as a reason for action, and for Hetherington it is a concatenation of abilities that are associated with knowing some fact, such as the ability to remember that p, to assert that p, and to explain accurately in p-related ways. Although there is a kind of explanatory unity to be had by positing that both knowledge-how and knowledge-that are abilities, I don’t think that the supporter of an Interrogative Capacity view about knowledge-how need be committed to thinking this. One might just think that knowledge is a heterogeneous kind that in some cases picks out a kind of ability to answer questions, and in other cases picks out a belief state.

It is also worth noting that one can make a good case for thinking that at least some knowledge-wh consists in an interrogative capacity, a view noted by Masto (2010), Michaelis (2011), and developed by Farkas (2016b). One piece of evidence for this view comes from habitual knowledge-wh ascriptions. Consider:

3)         Paula knows where the best clubs are.

Sentence 3) plausibly has a reading which says that Paula has the ability to generate knowledge of where the best clubs are in various locations.[21] If one takes habitual ascriptions seriously, this seems to be good grounds for taking at least some knowledge-wh ascriptions to be made true by an interrogative capacity relating to the relevant wh-question.

Turning back to knowledge-how, let’s compare the version of the Interrogative Capacity view which I sketched out in §3 to the accounts of knowledge-how offered by Dickie and Stanley and Williamson. Both views appeal to the idea of abilities generating propositional knowledge, but identify this capacity with skill rather than knowledge-how. Although at first this looks to be a merely verbal disagreement, I will argue that these views are substantively different, and both Dickie and Stanley and Williamson face problems which do not arise for the Interrogative Capacity view.

Dickie (2012) sets out to defend the claim that skills are explanatorily prior to propositional knowledge. She endorses a Strongly Intellectualist picture of knowledge-how, but wants to understand this knowledge as the by-product of the exercise of a skill. On her view, a skill at V-ing is a capacity to select reliable ways to V, and an act of V-ing counts as skilled just in case the agent Vs because she recognises that acting in that particular way is a reliable way to V. Although on this view a skill is strictly a capacity relating to V-ing and thus not an Interrogative Capacity in my sense, Dickie argues that such a capacity will involve the generation of propositional knowledge as a by-product. The basic idea here is that the recognition involved in a skilled action can be thought of as a belief of the form w is a reliable way to V, and that the reliability of a skill at V-ing provides a kind of practical justification for that proposition, thereby making that belief into knowledge.[22] One way to think about this view is as a kind of virtue-theoretic account of knowledge-how, on which knowing how to V is a cognitive achievement grounded in a skill at V-ing.

There’s a lot to unpack here, but this sketch of Dickie’s view should be enough to show her how view differs from the Interrogative Capacity view developed in §3. The crucial difference is that whereas Dickie claims that skill is a capacity which generates propositional knowledge (which she calls knowledge-how), on the Interrogative Capacity view, it is knowledge-how which is a capacity which generates propositional knowledge. Although this can seem like a mere verbal disagreement, I think that there are important issues that hang on this choice. For one thing, Dickie’s view entails that knowledge-how and skill are tightly related, since for her knowledge-how is the product of the exercise of a skill. By contrast, the Interrogative capacity view doesn’t need to make any commitments about the nature of skill, and is compatible with there being cases of knowledge-how without skill, and vice versa.

Dickie’s equation of knowledge-how with the product of skill also causes problems for making sense of the temporal profile of knowledge-how, and threatens to make knowledge-how epiphenomenal. Knowing how to do something is a standing epistemic state. Someone who has learnt to V knows how to V throughout the time after learning, unless something goes awry and they forget. Dickie’s view seems incompatible with this intuitive point. It is clear that for her skill is a standing epistemic state: someone who has learnt the skill of V-ing will be able to select reliable methods for V-ing at any point afterwards, even when they are not actually exercising that capacity. However, on Dickie’s view knowledge-how is a by-product of the exercise of skill, generated at the time of action. There doesn’t seem to be any reason to suppose that this situation-specific knowledge will be maintained after action has finished. On Dickie’s picture, knowing how to do something looks a lot like knowing what you had for breakfast — one can’t fail but know what you are having for breakfast while you are eating it, but this knowledge is lost soon afterwards, simply because there is no reason to keep hold of it. The problem is that if knowledge-how is transient in this way, it is difficult to make sense of the idea that knowledge-how is a standing epistemic state which persists when the agent isn’t acting.[23] If a skilled swimmer is sitting on the sofa, then their capacity to generate answers will be dormant, meaning that Dickie seems committed to thinking that they do not have knowledge how to swim. The Interrogative capacity view has the consequence that the propositional knowledge which is the product of an interrogative capacity will be transient, but since this knowledge is not identified with knowledge how, its transience doesn’t undermine the idea that knowledge-how is a standing epistemic state.

Dickie’s account also has trouble making sense of the explanatory significance of knowledge-how. It is a plausible Rylean thought that knowledge-how is the kind of state which one can use to explain intelligent action. The standard way for Intellectualists to understand this idea is by appealing to the idea that knowledge-how guides an agent toward successful action in virtue of representing a way to act successfully (Bengson and Moffatt 2011a: 20-24), (Stanley 2011: 175-7) (Stanley and Williamson 2016). Dickie explicitly repudiates the idea that knowledge-how guides intelligent action, instead explaining intelligent action in terms of skill. However, she doesn’t offer any other account of the explanatory significance of knowledge-how, which leaves knowledge-how looking like a mere epiphenomenal by-product of skill.[24] If the connection between knowledge-how and intelligent action is meant to be constitutive of the state of knowing-how, then we should identify knowledge-how with whatever explains skilful action. If Dickie is right, and the interrogative capacity is the explanatorily significant state, then we should identify this state with knowing how.

Stanley and Williamson (2016) offer a view of skill which is closely related to Dickie’s. Like Dickie, Stanley and Williamson claim that skill generates propositional knowledge, and they identify this propositional knowledge with knowing how. However they treat a skill as a capacity to know, rather than as a capacity to do, meaning that they explicitly treat skill as an interrogative capacity. On their view V-ing generates propositional knowledge about V-ing (relating to both how, and other wh-questions), and this propositional knowledge then guides action, meaning that skill only manifests indirectly in action. For example, if I am a skilled conversationalist, I possess a capacity to generate knowledge of the answers to range of practical questions  — when to speak, how to be polite, and so on — and while speaking, I will generate knowledge to these answers which will in turn guide me in what I say.  One way to think about this view is as a kind of particularism about intelligent activity (see Stanley 2011: 181-5). This view allows that intelligent actions require situation-specific judgements which cannot be subsumed under general rules, but it still claims that these situation-specific judgements are expressed in propositional knowledge and that intelligent action is guided by propositional knowledge.

Although Stanley and Williamson’s updated view is even closer to the Interrogative Capacity view than Dickie’s, I think that there remain important differences which are grounds to prefer the Interrogative Capacity view. They face the same worry as Dickie about explaining the temporal profile of knowledge-how, and their account of the relationship between knowledge-how and intelligent action relies on the controversial idea that intelligent action is guided by propositional knowledge.

Stanley and Williamson’s view shares with Dickie’s the commitment to thinking that skill is a standing epistemic state, whereas knowledge-how is propositional knowledge produced to meet the demands of the situation. As we saw above, this means that they face the problem of explaining the intuition that knowledge-how is possessed at every time after learning, rather than just at the times at which an agent is actually exercising her skill. I cannot see any way for them to resolve this problem, other than to endorse the implausible claim that the propositional knowledge produced by skill is always retained after the time of action.

One respect in which Stanley and Williamson’s account does better than Dickie’s is that they do have an account of the way in which knowledge-how relates to intelligent action: they can appeal to the standard intellectualist idea that intelligent action is action guided by propositional knowledge. Although Stanley and Williamson seem happy to endorse this claim as a means of displaying their robustly Intellectualist credentials, it remains extremely controversial. By contrast, because the Interrogative Capacity view identifies knowledge-how with the interrogative capacity which generates propositional knowledge, it can remain neutral about the explanatory significance of the propositional knowledge produced by this capacity. This means that it can allow both cases in which an agent generates knowledge of a method which guides their action, and cases in which the agent muddles through and only generates knowledge of the method once action is finished.[25] Consider again the various ways in which one can work out how to solve a maths problem. Sometimes it is possible to just see what the method for solving a maths problem is, but in other cases one needs to work through various parts of the problem in order to work out how to solve it. On the Interrogative capacity view it is possible to say that whereas the agent in the first case is being guided by her propositional knowledge, whereas the agent in the second case is not guided by propositional knowledge, but is rather intelligent just in virtue of the exercise of her capacity to generate answers to the question of how to solve a certain class of maths problems.

§5

The standard debate about knowledge-how has contrasted understanding knowledge-how as a kind of ability and understanding it to be a relation to a question. In this paper, I have shown that it is possible to hold a compromise position, which understands knowledge-how to be both an ability and a relation to a set of propositions. I have also developed a specific version of this kind of position — the Interrogative Capacity view — which equates knowing how to do something with an ability to generate answers to the question to how to do it on the fly. I have argued that this account is compatible with the standard semantics for knowledge-how ascriptions, can vindicate the intuition that knowledge-how is a distinctively practical species of knowledge, and illuminates the connections between knowledge-how, propositional knowledge and abilities.[26]

 

 

References

Bengson, John, & Marc Moffett, (2011a) Two Conceptions of Mind and Action: Knowing How and the Philosophical Theory of Intelligence. In John Bengson & Mark Moffett (eds.), Knowing How: Essays on Knowledge, Mind, and Action. Oxford: Oxford University Press. 3-58.

Bengson, John, & Marc Moffett, (2011b) Non-Propositional Intellectualism In Bengson, J., & M., Moffett, (eds.). Knowing How: Essays on Knowledge, Mind, and Action. Oxford: Oxford University Press. 161-195

Braun, David (2006). Now you know who Hong oak yun is, Philosophical Issues 16 (1):24-42.

Braun, David, (2011) Knowing How and Knowing Answers In Bengson, J., & M., Moffett, (eds.) Knowing How: Essays on Knowledge, Mind, and Action. Oxford: Oxford University Press. 244-260

Brogaard, Berit. (2011): Knowledge-how: A unified Account, In Bengson, J., & M., Moffett, (eds.). Knowing How: Essays on Knowledge, Mind, and Action, Oxford: Oxford University Press. 136-160

Cath, Yuri, (2011) Knowing How without Knowing That In Bengson, J., & M., Moffett, (eds.) Knowing How: Essays on Knowledge, Mind, and Action. Oxford: Oxford University Press. 113-13

Cath, Yuri (2015). Revisionary intellectualism and Gettier. Philosophical Studies 172 (1): 7-27.

Dickie, Imogen (2012). Skill Before Knowledge. Philosophy and Phenomenological Research 85 (3):737-745.

Dreyfus, Hubert. (2002). Intelligence Without Representation – Merleau-Ponty’s Critique of Mental Representation and the Relevance of Phenomenology to Scientific Exploration. Phenomenology and the Cognitive Sciences, 1 (4) 367-383.

Farkas, Katalin (2016a) Know-Wh does not reduce to Know-that, American Philosophical Quarterly 53 (2):109-122

Farkas, Katalin (2016b) Practical Know-wh Noûs doi: 10.1111/nous.12152

Fridland, Ellen (2013). Problems with intellectualism. Philosophical Studies 165 (3):879-891.

Friedman, Jane (2013). Question‐directed attitudes. Philosophical Perspectives 27 (1):145-174.

Glick, Ephraim, (2011). Two Methodologies for Evaluating Intellectualism. Philosophy and Phenomenological Research 83 (2): 398-434.

Glick, Ephraim (2012). Abilities and Know-How Attributions. In Jessica Brown & Mikkel Gerken (eds.), New Essays on Knowledge Ascriptions. Oxford: Oxford University Press

Glick, Ephraim (2015). Practical Modes of Presentation. Noûs 49 (3):538-559.

Hawley, Katherine. (2003) Success and Knowledge-How, American Philosophical Quarterly 40: 19-31.

Hawley, Katherine (2011). Knowing How and Epistemic Injustice. In Bengson, J., & M., Moffett, (eds.) Knowing How: Essays on Knowledge, Mind, and Action. Oxford: Oxford University Press: 283-99.

Hetherington, Stephen (2011). How to Know: A Practicalist Conception of Knowledge. Wiley-Blackwell.

Hornsby, Jennifer, (1980). Actions. London: Routledge and Kegan Paul.

Hornsby, Jennifer, (2011) Ryle’s Knowing-How, and Knowing How to Act, In Bengson, J., & M., Moffett, (eds.). Knowing How: Essays on Knowledge, Mind, and Action. Oxford: Oxford University Press. 80-100.

Hornsby, Jennifer (2016) Intending, Knowing how, Infinitives, Canadian Journal of Philosophy, 46:1, 1-17

Lewis, David (1999) What experience teaches in Papers in Metaphysics and Epistemology Cambridge: Cambridge University Press: 262–90

Masto, Meghan (2010). Questions, answers, and knowledge-wh. Philosophical Studies 147 (3):395 – 413.

Michaelis, Laura, (2011) Knowledge Ascriptions by Grammatical Construction In Bengson, J., & M., Moffett, (eds.) Knowing How: Essays on Knowledge, Mind, and Action. Oxford: Oxford University Press. 261-83

Parent, Ted. (2014). Knowing‐Wh and Embedded Questions Philosophy Compass 9 (2):81-95.

Pavese Carlotta. (2013) The Unity and Scope of Knowledge. Ph.D. Dissertation Rutgers University.

Pavese, Carlotta (2015). Practical Senses. Philosophers’ Imprint 15 (29): 1-25.

Ryle, Gilbert, (2009): The Concept of Mind. Abingdon: Routledge

Setiya, Kieran, (2008). Practical Knowledge. Ethics, 118: 388-409.

Setiya, Kieran (2012). Knowing How. Proceedings of the Aristotelian Society 112 (3):285-307.

Snowdon, Paul (2004). Knowing how and knowing that: A distinction reconsidered. Proceedings of the Aristotelian Society 104 (1):1–29.

Stalnaker, Robert (2012). Intellectualism and the Objects of Knowledge. Philosophy and Phenomenological Research 85 (3):754-761.

Stanley, Jason (2011a). Know How. Oxford: Oxford University Press.

Stanley, Jason, (2011b). Knowing (How). Noûs 45 (2): 207-238.

Stanley, Jason (2012). Replies to Dickie, Schroeder and Stalnaker. Philosophy and Phenomenological Research 85 (3):762-778.

Stanley, Jason, and Krakauer, John (2013) Motor Skill Depends on Knowledge of Facts, Frontiers in Human Neuroscience, (7) 503, 1-11 http://dx.doi.org/10.3389/fnhum.2013.00503

Stanley, Jason & Williamson, Timothy (2001). Knowing how. Journal of Philosophy 98 (8): 411-444.

Stanley, Jason & Williamson, Timothy (2016). Skill, Noûs Early Access http://onlinelibrary.wiley.com/doi/10.1111/nous.12144/abstract (accessed 23rd May 2016) DOI: 10.1111/nous.12144

Weatherson, Brian (2016). Intellectual Skill and the Rylean Regress, Philosophical Quarterly Advance access, (accessed 8th August 2016) DOI: 10.1093/pq/pqw051

Wiggins, David (2012). Practical Knowledge: Knowing How To and Knowing That. Mind 121 (481): 97-130.

Williamson, Timothy (1990/2013). Identity and Discrimination. Oxford: Blackwell.

 

Notes

[1] (Stanley and Williamson 2001: 429-30, 33-34), (Stanley 2011 C4), (Glick 2015) (Pavese 2015)

[2] Throughout, I will use ‘V’ as a variable for activity types.

[3] I will use these two notions interchangeably. Some Anti-Intellectualists also claim that knowledge-how is a kind of disposition, but this view is difficult to square with the idea that the exercises of knowledge-how are under intentional control.

[4] From this point, I will use answer to mean a correct — i.e. true and complete — answer.

[5] I take standing knowledge to be knowledge that is retained over a longish period of time.

[6] For reasons of space, I will pass over critics of Intellectualism who appeal to empirical evidence. For an overview see, (Glick 2011: 400-2) (Stanley 2011 C7).

[7] There is considerable controversy about whether to interpret Necessity as:

N1: If S is intentionally V-ing, then S knows how to V

Or the weaker:

N2: If S is intentionally V-ing, then S knows how to V, or else she is doing it by W-ing, and she knows how to W.

For discussion, see (Setiya 2008, 2012), (Stanley 2011: 188-90).

[8] See (Stanley 2011: 1-35, 181-5, 188-90).

[9] A possible exception is directness. Stanley stresses that on his conception of propositional knowledge, all knowledge is exercised directly in action (2011: C1) (see: Stalnaker 2012)

[10] See (Pavese 2015) for a defence of the explanatory power and practicality of practical modes of presentation.

[11] For an account of a kind of ability which is plausibly identified with knowledge-how, see (Glick 2012).

[12] Of course, there is considerable disagreement about what these properties are. To complicate things, Stanley (2012) suggests that establishing Intellectualism might motivate endorsing a more practical conception of theoretical knowledge in general. See also (Stalnaker 2012).

[13] See also Cath (2011), which suggests that the relation might be a seems that relation.

[14] Stanley accepts this this point (2011: 148-9), and it is natural to read chapters 7 and 8 of Know How as a defence of the idea that knowledge-how involves the theoretical knowledge relation to propositions.

[15] Are there any views which occupy the top right hand box in this table? One might think that knowledge-how is a theoretical knowledge relation to something other than a proposition. See (Brogaard 2011) for something close to this view.

[16] (Pavese 2013), (Stanley and Williamson 2016).

[17] Katherine Hawley and Matthew McKeever have pointed out to me that one can also find a similar reading in ascriptions with a quantifier in the complement. Consider:

1) Jonah knows how to please everyone.

Although this sentence has a singular reading (Jonah knows one method for pleasing everyone), it also has a generalising reading (for every person, Jonah knows a method for pleasing that person), which might be made true by an ability to generate knowledge of how to please people.

[18] This also explains why demonstratives are not an effective way to transmit general knowledge-how. What the demonstrative gives is a hold on a specific manifestation of knowledge-how, whereas genuine know-how requires a general ability to produce many different methods for doing something.

[19] I introduce the notion of a contextually relevant set of practical situations to explain the shiftiness data introduced by Hawley (2003: 22). Hawley points out that knowledge-how ascriptions seem to be indexed to a contextually supplied set of situations or tasks. For example, in a US context in which only driving an automatic car is salient then someone who can only drive an automatic car will count as knowing how to drive, whereas in a UK context in which both driving a manual and driving an automatic are conversationally salient, then only people who can drive both types of car will count as knowing how to drive. For discussion of the context sensitivity of knowledge-wh see (Braun 2006, 2011), (Parent 2014)

[20] Stanley claims that first-personal knowledge can involve a disposition to acquire new propositional knowledge (2011: 182-3), but it is not clear that there are any non-know-how examples which display this phenomenon.

[21] (Pavese 2013), (Stanley and Williamson, 2016). Farkas (2016a) also gives cases in which it seems felicitous to assert that an agent knows-wh in cases in which they can easily access the answer to the question. For example, I might say that A knows what his partner’s phone number is even if he hasn’t memorized it, so long as he can easily look it up. Although this phenomenon is suggestive, it seems like a good case for a pragmatic explanation. The typical situation in which one would ascribe to A this knowledge would be one in which one wants to find out his partner’s phone number, so we might well think of the ascription as strictly false but pragmatically conveying that A will be able to tell you what the answer is.

[22] For a similar idea about justification, see (Brogaard 2011: 147-8)

[23] See also Farkas (2016b).

[24] Note that Dickie’s account of what it is for an action to be skilled nowhere appeals to knowledge-how (2012: 739).

[25] Note that this is rather different to the standard criticism of the idea that intelligence is guided by skill, which is to suppose that intelligent action is never guided by propositional knowledge. See (Dreyfus 2002).

[26] Thanks to Mark Bowker, Matthew Cameron, Lucy Campbell, Katalin Farkas, Katherine Hawley, Matthew McGrath, Matthew McKeever, Mona Simion, Fenner Tanswell, and Caroline Torpe Touborg.

Published by

Josh

I am a third-year PhD student in Philosophy based in St Andrews in the joint St Andrews/Stirling programme. I work mainly on knowing how to do stuff (swim, play the trumpet, run a marathon), and its connection to a range of debates in epistemology and philosophy of action.

9 thoughts on “Knowledge-How, Abilities, and Questions”

  1. The paper is packed with useful thinking. Thanks very much to the author for sharing it. I particularly admire the sustained development of the impulse to overcome what seems to be an impasse on the question of the fundamental character of knowing-how. (Hereafter I will simply address the author directly and informally!) And I also think that you are correct in holding that exercising one’s know-how typically involves learning some previously unknown specific truths bearing on the doing.

    Your proposed middle way between the intellectualist school and the anti-intellectualist school is to appeal to what you call the Interrogative Capacity view (ICV). Since this view makes basic appeal to a capacity or ability (the ability to answer a practical question on the fly) in giving an account of knowing-how it would seem to belong in the anti-intellectualist camp. But this view also counts as weakly intellectualist in that it identifies knowing-how to V with standing in relation to a set of propositions, viz., a set of good answers addressing the question of how to V in Fn. To put the general lesson very simply, your view is that knowing-how is as propositional as any other sort of knowledge and that it essentially involves an ability relating to the relevant propositions. Fair enough.

    Yet It seems to me that there is plenty of room for some further clarity as to what exactly the view comes to. (My comments will focus on the last few pages of §3 of the draft.)

    In your canonical formulation of the ICV, (several pages into §3) you identify knowing how to V in a contextually relevant class of situations {F1, F2, …} with the “capacity to generate fine-grained answers to the question ‘how to V in Fn’ in the process of acting”.

    I have two worries about this formulation of the ICV. First, arguably, one who does not know how to V but is confident and willing to experiment with methods may well have the capacity to, given enough time, figure out ways of V-ing across a range of situations. Your formulation seems blind to an important distinction here, and (arguably) gets this case wrong in attributing knowledge-how where there is initially merely some willingness and ability to figure it out. What do you say to this?

    My second worry needs more elaboration to motivate.

    To develop it: on the same page and shortly after introducing this canonical formulation, you are seeking to illustrate the philosophical virtues of the view. In the course of that you identify the question ‘how to V in Fn’ with the set of answers (where we are to view this set as that self-same set of propositions). I quote: “This view [the ICV] takes knowledge-how to be a relation to a question—the question ‘how to V?’—and understands that question as being identical to a set of possible answers.” Call this the equation thesis (ET), which you embrace in the course of explaining the strength of the ICV while arguing that it is compatible with a standard story in the formal semantics of embedded questions.

    Are you really committed to this thesis? If so, the canonical formulation you give of the ICV would be rendered unintelligible, assuming that we can substitute in co-referring expressions.

    To spell this out:

    The ET is the view that the question ‘how to V in Fn’ = the set of answers concerning how to V in Fn (i.e., the relevant set of true propositions).

    The ICV: Knowing how to V is the “capacity to generate fine-grained answers to the question ‘how to V in Fn’ in the process of acting”

    If we can substitute co-referring expressions, an alternative way to express the ICV is this:

    The ICV*: Knowing how to V is the “capacity to generate fine-grained answers to the set of answers concerning how to V in Fn in the process of acting”.

    But this latter (ICV*) is obviously nonsense since the idea of generating answers to the set of answers makes no sense. That seems to be a problem of some kind and this is the core of my second worry. So what do you say to that?

    On the next page (and a half) I find that things get a little more mysterious and very interesting! I am particularly interested in your discussion of directness and what follows. More precisely I would like to hear more about how “the process of doing and the process of answering are intertwined”. Your view seems to be that there are are, strictly, two processes going on when an agent Vs intentionally, drawing on her knowledge of how to V. Do I have that right? One process is that of doing (call that the intentional action proper) and the other is a process of learning. And then there are two achievements, if things go well: the relevant doing gets done and some previously unknown propositional truths enter the agent’s grasp.

    I find this interesting. But it seems to me that the explanation of knowledge-how that you give makes the relevant intentional action only accidentally connected to knowing-how, since you treat it as essentially a capacity to generate the knowledge of specifics. I think that, in the paradigm cases, one acts intentionally out of one’s knowledge-how (to V) and that this is revealing of the real nature of the phenomenon. There will ordinarily be some specific knowledge generated, but it seems mistaken to treat knowledge-how as properly constituted merely as a capacity for generating a grip on some specific previously unknown truths (answers) in attempting to answer a question. Yet that is just what the ICV does.

    In short, I think this over-intellectualizes the matter in treating answering on the fly by generating a grip on some propositions as central and acting intentionally to V out of one’s knowledge-how as a sort of parallel process. At the very least this is, in my view, a mistake of emphasis. I also think that you are mistaken in leaving no room for standing knowledge in your account. Surely at least some of what knowing-how to V is, is a standing rational power to enact an intentional effort to V at will in relevant conditions. This does not preclude generating some fresh knowledge. But that is not, though your formulation suggests it is, the central story.

    Thanks again! EPR (August 2016)

    1. First of all, I want to thank all of the commentators for their really helpful comments. One of the big issues I’ve been facing in trying to give a novel view is to get clear on what one’s view actually is, and these comments have taught me a good deal about the view I’m trying to develop!

      I’m going to split my comments up by commentator, which should hopefully make navigating between threads of the conversation a little easier. There are a couple of overlaps, which I’ll try to make clear.

      Response to Evan Riley

      Evan raises three really interesting questions in his comments. I don’t have proper answers to any of these questions, but I’ll try to develop the problems, and point toward the shape of some answers.

      i) How can a view which identifies knowledge-how with the capacity to answer a question distinguish between someone who has standing know-how, and someone who is just really good at figuring things out?
      ii) Is it really true that a question is identical with a set of propositions? They do not seem to be intersubstitutable in extensional contexts.
      iii) Doesn’t thinking of the exercise of knowledge-how as involving dual processes of generating knowledge and acting on that knowledge over-intellectualise knowledge-how?

      Working it out

      Evan raises the issue of how the interrogative capacity view is to distinguish someone who knows how to V from someone who is quick at learning, but doesn’t know how to V. I agree with Evan that this is a hugely difficult issue for the kind of view that I’m trying to articulate. It is easy to set up cases in which someone who is ignorant about how to engage in some activity can work out how to do it via their general problem solving skills. Someone who has no idea how to fix a flat tyre on a bike might nonetheless fix one by exercising their general engineering know-how. We might think that such a person does have the relevant kind of capacity to answer the question ‘how to fix a bike tyre?’

      The way I’d like to describe these kinds of cases is this: the agent who has general learning capacities, exercises some very general piece of know-how on some rather specific problem, in so doing coming to know some specific way-proposition, and also (perhaps) coming gaining a capacity to work out how to solve this kind of problem in the future. In order to give this kind of description of the case, I’d need to have theoretical resources for distinguishing between exercising some general know-how in V-ing, and exercising specifically knowledge how to V. I have some half-baked ideas about how to do this, but I don’t have anything like a fully worked out story.

      In my defence, I think that standard Intellectualist accounts face a similar problem. Any activity will be associated with many way-propositions of different levels of generality, and plausibly only some of those way propositions will suffice for knowledge-how. Which? (see the Touch-typist example in Hornsby 2011) http://www.oxfordscholarship.com/view/10.1093/acprof:oso/9780195389364.001.0001/acprof-9780195389364-chapter-3. I think that currently Intellectualists don’t have a worked-out story about this issue either. This doesn’t mean that my account is in the clear, just that lots of people in this debate face this kind of problem.

      (It’s also worth noting that Stanley and Williamson’s account of skill will face parallel worries, since their account looks like it predicts that someone who is good at working stuff out counts as being skilled at activities which she is in a position to quickly work out how to do.)

      Are questions sets of propositions?

      Evan notes that I endorse an equivalence thesis:

      ET: a question Wh-F? is identical to the set of propositions {p1, p2, … pn} which answer that question.

      (Or maybe, even better, I endorse the claim that the semantic value of an interrogative is just the relevant set of propositions, so that the set of propositions gives the meaning of an interrogative).

      (Note that it’s controversial exactly which propositions are included in set of answering propositions –- is it just true ones, are they complete or partial answers — but for the moment let’s put these issues to one side, and come back to this in Jay’s comments).

      Now, Evan points out that if ET is true, then the following pattern of substitution ought to be acceptable:

      1) Knowing how to V is the capacity to generate fine-grained answers to the question ‘how to V in Fn’ in the process of acting.
      2) Knowing how to V is the capacity to generate fine-grained answers to the set of answers concerning how to V in Fn in the process of acting.

      But, according to Evan, 2) is nonsense. Why is this? One diagnosis would be that the idea of standing in an abilitative relation to a proposition is flawed in principle. I don’t think that this diagnosis is right. Instead, I think that this phenomenon fits into a wider pattern of cases in which substituting an interrogative phrase with a set of propositions which purport to be its semantic value leads to nonsense.

      Notice that we can find similar cases for other verbs like ‘wonder’ that embed interrogative phrases:

      3) Paula wondered whether Dave Weir will win a Gold
      4) Paula wondered the set of propositions {Dave Weir will win a Gold; it is not the case that Dave Weir will win a Gold}

      (I’m assuming that whether p? denotes the set {p; not-p}, although we get the same weirdness if we assume that whether p? denotes {p}).

      Intuitively, we can wonder into questions, but not into propositions (except in the sense of ‘wonder’ where it means something like ‘marvel’). Should we conclude that questions are not identical to sets of propositions? That would be one way out. An alternative would be to take these failures to be examples of the general phenomenon of substitution problems (which should not be confused with Frege puzzles!) (I owe the idea that these kinds of cases might be examples of substitution puzzles to Luke Burke).

      Here’s an example of this kind of problem. It is popular to think that ‘that p’ denotes the proposition that p, and that propositional attitudes should be thought of as relations to propositions (and that propositional attitude ascriptions pick out these relations). However, there are many examples of sentences in which the substitution of ‘the proposition that p’ for ‘p’ yields a false or meaningless sentence (for examples, see Moltmann 2013: C4 §3.1 link).

      Consider:

      5) Dai hopes that Hannah Cockcroft wins a Gold in Rio
      6) Dai hopes the proposition that Hannah Cockcroft will win a Gold in Rio

      I don’t know what the solution to these kinds of substitution puzzles is, but I presume that the phenomenon is ultimately a rather general one. The point being that sometimes substituting a term for its semantic value doesn’t lead to an acceptable sentence. Any successful solution for ‘that’-clauses ought to be generalizable to ‘wh’-complements. Maybe, not very satisfying, but hopefully that points toward a possible answer (or at least a bigger and more interesting problem).

      Two processes or one?

      Evan worries that my account pulls knowledge-how apart from intentional action because it claims that knowing-how is a capacity to generate propositional knowledge, which makes it one step removed from intentional action. I’m not so sure about the worry about over-intellectualisation, but I agree that the kind of dual process picture which Evan describes (where working how to V, and V-ing are distinct processes) would cause problems for explaining directness. (incidentally, I think that Stanley and Williamson’s view of skill is a good example of something like this dual process view)

      However, what I was trying to go for with the quote from Ryle was the idea that one process can at the same time be a process of acting, and a process of learning a proposition. I guess I wanted to reject a dual process story, and endorse something like a dual aspect picture, where one process is at once acting, and generating knowledge. The idea is that we can see intentionally acting as being a process of answering a how to question, with the idea being that the upshot of that process is both an act, and knowledge of the relevant proposition. As Carlotta points out there’s a lot that’s mysterious about this picture, but this is the kind of account that I’m aiming for.

  2. In Knowledge-How, Abilities, and Questions, Joshua Habgood-Coote’s proposes an account of knowing-how which is intended to diffuse the tension between the Intellectualists and Anti-Intellectualists positions in the debate regarding what is required to know how to do something. Habgood-Coote’s view, which he calls the Interrogative Capacity view, claims:

    S knows how to V if and only if for some contextually relevant class of practical situations {F1, F2 …} S has the capacity to generate fine-grained answers to the question ‘how to V in Fn?’ in the process of acting.

    While this isn’t an entirely implausible view to hold, I argue that, as stated above, the Interrogative Capacity view is unsatisfactory.

    First, Habgood-Coote does not specify that in order for S to know how to V, S must be able to generate correct answers. That is, even if one satisfies all of the conditions for knowledge-how on the Interrogative Capacity view, we are not given reason to think that the answers S gives are true, or informative in a useful way. It seems plausible that someone can ask me how to make a souffle and I can give them fine-grained answers as I’m performing actions which I believe will make a soufflé, while having never made a soufflé before, never learned how to make a soufflé, and while actually failing to make a soufflé in response to the question. While I would satisfy Habgood-Coote’s conditions for knowledge how to make a soufflé, we wouldn’t want to say that I know how to make a soufflé in this case. Therefore, Habgood-Coote needs to add a truth condition to his account of knowledge-how.

    Assuming Habgood-Coote takes the advice above, a second and larger problem still remains for Habgood-Coote’s Interrogative Capacity view. Whereas Dickie and Stanley and Williamson’s accounts of knowledge-how identify capacity to V as an acquired skill to V, Habgood-Coote supposes his view benefits by removing this skill requirement. Habgood-Coote proposes that “the process of doing and answering are intertwined” (p. 13) and that we should take the relevant sense of ability to V to be the ability to V intentionally. However, by removing a skill requirement and instead requiring that the action is done intentionally, Habgood-Coote’s view allows for S to know how to V in situations which we normally wouldn’t attribute knowledge-how.

    For instance, suppose Greg, who has never bought a lottery ticket in his life, claims that he knows how to win the lottery. When asked how he knows this, Greg goes to the local party store and, while saying to his inquirer, “This is how you win the lottery,” Greg buys a lottery ticket with the intention of winning the lottery. Suppose he then wins the lottery. On the Interrogative Capacity view Greg does in fact know how to win the lottery. In the contextually relevant and practical situation of buying a lottery ticket from this particular party store, Greg not only has the capacity to generate answers to “How do you win the lottery?” but Greg actually exercises this capacity while performing the action. Furthermore, since Habgood-Coote’s account considers performing action V in the relevant class of practical situations to provide sufficiently fine-grained answers to the question ‘How to V?’, Greg’s act of winning the lottery must be sufficiently fine-grained and, therefore, Greg meets the knowledge-how requirements of the Interrogative Capacity view. Since we wouldn’t consider Greg to actually have knowledge of how to win the lottery in this case (supposing it was a fair lottery and Greg did not have any tricks up his sleeve), the Interrogative Capacity view seems to lack some type of requirement which ensures that we don’t consider people like Greg to know how to win the lottery. What is needed is some sort of condition which guarantees that the capacity is not only exercised once, but repeatedly and consistently in a non-random way. This is exactly what you would gain by adding a skill requirement, which Habgood-Coote wants to do away with.

    Lastly, even if Habgood-Coote made the adjustments I’ve suggested, his solution to the debate between Intellectualists and Anti-Intellectualists regarding knowledge-how would not ease the tension between the two sides. The importance and simplicity of the debate between Intellectualists and Anti-Intellectualists can be (perhaps overly) simplified when asking what it means to say that you “know how to swim.” Whereas Anti-Intellectualists claim the appropriate way to demonstrate this knowledge is by swimming, the Intellectualists claim the appropriate answer contains propositional content such as “kick your legs, move your arms like you’re crawling, and keep your head above water.” Though both sides of the debate contain more intricacies than I have just suggested, the tension between the two sides is clearly palpable. Although Habgood-Coote claims that his view diffuses this tension by combining the two views, I argue the tension remains.

    If we assume S can V in a contextually relevant class of practical situations, Habgood-Coote’s Interrogative Capacity view can essentially be understood to claim that S knows how to V if S can answer the question of ‘how to V?’ while Ving. On this account, knowing how to swim would require something like the ability to answer questions of ‘how to swim?’ while swimming. This account clearly combines both the Intellectualist propositional demands (by requiring that S can answer the question ‘how to V?’) with the Anti-Intellectualist abilitative demands (by requiring that S has a certain capacity). However, putting these two demands together does not diffuse the tension between them.

    To demonstrate how combining these conflicting accounts does not diffuse the tension between them, suppose I have decided that I should invest my money where it will result in the most money in the future. I could invest in option A which will result in big returns in 5 years but very little after that or I could invest in option B which will result in me having minimal returns for the first few years but will lead to larger returns in 20 years. There is clearly a tension between which option is best and this tension will be the result of how far in the future I should be concerned with the returns on my investment. However, to say that I should invest in both plan A and plan B does not diffuse the tension at all. True, it is one solution to the problem, but it doesn’t make it clearer how forward-looking I should be. Similarly, the Interrogative Capacity view may be a plausible view that incorporates important features from both the Intellectualist and Anti-Intellectualist camps, but this does not do anything to elucidate which option is superior nor does it relieve any tension between camps since they still are opposing positions. Therefore, while Habgood-Coote’s Interrogative Capacity view has problems which are curable, the view fails to resolve the disagreement between Intellectualists and Anti-Intellectualists about knowledge-how. Nonetheless, it remains possible that the Interrogative Capacity account of knowledge-how could still be useful and preferred to either of the current camps since it does incorporate important features of both.

    1. Response to Jay Spitzley

      Jay raises three pressing worries for my picture:

      i) He recommends that I understand answers that figure in the interrogative capacity view to be true propositions,
      ii) He worries that divorcing know-how from skill means that the Interrogative Capacity view cannot explain the idea that knowledge-how requires reliability,
      iii) Finally, Jay questions whether we should think of the interrogative capacity view as being able to resolve the tension between Intellectualist and Anti-Intellectualist views.

      What’s an answer?

      Jay quite rightly observes that the Interrogative Capacity view is only plausible if it claims that knowledge-how requires a capacity to generate true answers to a question. There are cases where it is easy to generate false answers to a question. I have a capacity to generate false answers to the question how can one prove the four-colour theorem?, but this doesn’t mean that I know how to prove the four-colour theorem.

      I think that I was being a little sloppy on this point. My intention was to make sure that what I said was neutral between different accounts of the semantics of questions. There is a debate in the semantics literature about whether interrogative phrases denote true or merely possible answers, and whether those answers need be complete, or can be merely partial. My hope was to not make any commitments on this debate. However, when it comes to the idea that knowledge-how is a capacity to answer a question, I absolutely agree with Jay that we should focus in on true answers to a ‘how to’ question. I think that this is just a feature of the relevant kind of ability: one way to put the point is that in the relevant sense of being able to answer, ‘answer’ is a success-verb.

      Can we have Reliability without Skill?

      One salient difference between my view of knowledge-how and those offered by Dickie, and 2016 Stanley and Williamson, is that my view is neutral on the relation between know-how and skill. I think of this as a positive of the view, since I haven’t been convinced that skill is a necessary condition for knowledge-how. (My suspicion is that ‘knows how’ and ‘skill’ are both context-sensitive terms, and that in certain contexts they can be closely connected, but that in others they can come apart). Jay questions whether this feature of the view is a positive by contending that one cannot explain the idea that knowledge-how requires reliable success without a skill condition. To this end Jay appeals to a case in which an agent luckily generates an answer to the question of the question of how to win the lottery (buy this ticket, when this ticket happens to be the winning ticket), without knowing how to win the lottery.

      The important question is whether the kind of account I like can explain the idea that knowledge-how requires a reliable disposition to success. I think that it can. Note that in the official statement of the view I appealed to a contextually supplied set of practical situations, and claimed that knowing how requires being able to generate the answers to ‘how to’ questions about all of those situations. In cases in which this contextually supplied set is rather large, knowledge-how will require a reliable ability to generate situation-specific way-propositions. With this idea in place, I think that I can say that in the lottery case, the agent doesn’t have a suitably reliable capacity to generate propositional knowledge. My line is that I think that I can secure the relevant kind of reliability by appealing to a context-sensitive ability condition.

      Does the Interrogative Capacity view really resolve the tension?

      Finally, Jay questions whether my view really resolves the tension between Intellectualism and Anti-Intellectualism.

      I’m not sure I quite understand the worry here. I didn’t take my goal to be to argue that Intellectualism and Anti-Intellectualism are the same view (at least in their strong varieties they certainly are not). Rather the idea was that we could take a kind of compromise position which incorporates elements from both views (the respective weak conditions). One upshot from this compromise was to hopefully resolve the tension between the two central kinds of evidence in this debate, which on the face of it seem to point toward diametrically opposed accounts. Is there something more to this worry, Jay?

      1. Joshua, I apologize for any confusion. If your goal was only to show that a compromise position which incorporates elements from both views is possible, then this is not worrisome for you.

        However, I was under the impression that you were claiming your view somehow eliminated the conflict between the two views. Though your view may demonstrate a possible compatibility regarding aspects of these two views, I was attempting to point out that the two views still take opposing stances and have conflicting demands. Therefore, while a certain compatibility may be operable, there remains an important and significant tension between the demands of these two views. If your view is supposed to entirely resolve this tension, then I would like to hear more about exactly how that works. Nonetheless, if you agree that this tension between the demands of these two views persists and you are instead only making the more modest claim that your view can operate in some kind of middle-ground between these positions, then I don’t think we are disagreeing about anything substantive. I hope this makes things clearer.

        1. No worries Jay – this is a really helpful point about set up, and thanks for raising it!

          I think you’re right on – I’m only arguing for the possibility of a compromise position (the modest claim), not trying to argue that the tension between Intellectualism and Anti-Intellectualism goes away.

          Thanks again for the commentary!

  3. Thank you very much to the organizers of the Minds Online Conference for the invitation to comment and to Joshua for his essay.

    The essay starts off with a dilemma for existing views of know how. Joshua observes that anti-intellectualism, by identifying know how with an ability, fails to do justice to the way in which know how is ascribed (at least in English). Intellectualism, on the other hand, by identifying know how with propositional knowledge of an answer to the question “How to F”, does justice to the linguistics of know how ascriptions but fails to account for the distinctively ‘practical properties’ of know how — its being distinctively a kind of practical knowledge. Joshua thinks that, in order to account for such practical properties, intellectualists have to appeal to a ‘somewhat obscure’ notion of practical modes of presentation. That motivates Joshua to develop a third view — the interrogative capacity view — which is supposed to improve on both intellectualism and anti-intellectualism.

    My comments are structured in two parts. In the first part, I will take issue with Joshua’s claim that intellectualism cannot account for the distinctive practical properties of know how. In the second part, I will question Joshua’s claim that the view developed by him in this essay — the interrogative capacity view — can do better.

    1. Intellectualism and the practical properties of know how
    As pointed out by Joshua, intellectualists aim at explaining at least some of the practical properties of know how by appeal to practical modes of presentation. Now, it is true that many have complained that the notion of practical modes of presentation is obscure. But this charge of mysteriousness and obscurity to the notion of practical modes of presentation is no longer legitimate in the light of the recent debate. Although earlier version of intellectualism (Stanley & Williamson (2001); Stanley (2011)) have not told much about the nature of practical modes of presentation — and so maybe charges of mysteriousness were legitimate towards those versions of intellectualism — more recently a lot of work has been done by intellectualists to vindicate the distinctively practical aspect of know how and to clarify the notion of practical modes of presentation. In my own work, I develop a detailed view of practical modes of presentation (Pavese (2015b)). I propose we think of them along the lines of operational semantic values for programming languages. Thinking of practical senses along the lines of operational semantic values affords the intellectualist a response to the mysteriousness challenge and gives us independent reasons to think that things like practical modes of presentation do exist.

    Now, the view I have developed in “Practical Senses” and in another paper of mine (“Knowing a rule”, (2015a)) has all the resources to explain each of the three features which Joshua identifies with the ‘practicality’ of know how. According to Joshua, three features are distinctive and in need of explanation:

    i. Directness: knowing-how is exercised directly in intentional action, not via some intermediate act of mental contemplation;
    ii. Necessity: knowledge-how is a necessary condition for intentional action;
    iii. Flexibility: knowing-how involves an ability to react flexibly to a wide range of possible situations.

    Let us start with the last, flexibility. My understanding of practical senses along the models of programs can explain flexibility very naturally, for we have independent evidence that machines can be programmed to be flexible. As an illustrative example, think of the Mars rover, which has been programmed to autonomously deal with a great variety of unpredictable situations on Mars’ surfaces — the properties of which situations the programmers cannot predict in advance. The Mars rover does it by implementing a procedure that is computed on the basis of an updatable and ever-changing model of the environment (data). (As I explain in my essay, a program is the combination of a procedure and the data.) This program tells the Mars rover what to do when by means of algorithms on the basis of which the Mars rover can calculate novel responses to new circumstances as they come along. I would recommend looking at the behaviour of the Mars rover and at the program, as it is studied by Rutgers Phd Bethany Leffler’s dissertation http://www.research.rutgers.edu/~bleffler/thesis/dissertation.pdf).

    Now, my point is that the Mars rover can respond autonomously and flexibly to a wide range of possible situations. It genuinely can come up with new solutions to new tasks (navigating on never-met-before kinds of surfaces, for example). But it does it through the computation of a program. Note that the program does not need to anticipate all the novel situations the Mars rover can find itself in. Nonetheless it enables the Mars rover to calculate responses ‘on the fly’ on the basis of an ever-changing model of the environment.

    When I elaborated my view of practical senses, the flexibility problem struck me as serious and not sufficiently addressed. But it also struck me as a point in favour of my view of practical senses that it provided an explanation of the flexibility of know how for which we have independent evidence from computer science and robotics.

    Of course, there is more to be said about this and it is part of my current work to use my view of practical senses more explicitly to deal with flexibility from an intellectualist point of view. For example: one might object that it might be hard to envisage what program could explain more complex flexible behaviour, such as the flexibility of a Jazz composer, for example. But examples such as the Mars rover or Google cars suggest that this is an epistemic limit — having to do with the kind of programs we can now envisage given the advances and the degree of sophistication of programming techniques — rather than a principled problem with the proposal.

    What about Directness? In “Practical Senses,” I explain that operational semantic values are ability-entailing: if one grasps them, one is enabled with the ability to follow a certain rule/method. That is so because they are structured in such a way that each of their parts and their modes of combination are defined in terms of rules that the agent can follow. I also argue that the ability to perform a task and the ability to perform the task intentionally come apart. In “Knowing a rule”, I make the case that only knowledge of a proposition that has a practical sense as component can explain the ability to intentionally perform a task that goes together with knowing how. My argument hinges on consideration of cases where a subject can perform a task through a certain method, but because she has false beliefs about what task that method is for, the subject fails to bring the method to bear intentionally and relevantly. So while grasp of a practical sense enables one with the ability to perform an action, only a propositional attitude can explain the ability to intentionally perform an action. (Incidentally: the idea that Directness cannot be explained by intellectualism is encouraged by a view of propositional knowledge as divorced from action that it is quite far from how today many epistemologists and philosophers of mind think of it and that is, at any rate, not at all to be taken for granted. As Stalnaker (2012) puts it, knowledge is the ability to use information for the purpose of action. Any view of propositional knowledge as merely a contemplative attitude is doomed to miss out on the important role played by knowledge in guiding action.)

    That leads us to Necessity. According to Necessity, know how is necessary for intentional action. The alleged burden for intellectualism is to show that a propositional attitude such as propositional knowledge is also necessary for intentional action. As Joshua observes, know how goes together with the ability to intentionally F. But as I argue in (Pavese (2015a) and Pavese (201?)), the ability to intentionally F should not be taken as an unexplained primitive and is arguably best explained in terms of a propositional attitude of belief (and maybe knowledge. See Gibbons (2001)). That is in line with many prominent views of intentional action. For example, Goldman (1979) has elaborated a view on which a belief about how to perform an action is necessary for being able to perform it intentionally. And even action plan-based theories of intentional action impose a belief constraint on intentional action (See Bratman (1999)).

    Setiya (2011) has argued that belief is not necessary for intentional basic action. His main example is that of a subject previously injured in their hand who tries to clench their fist, but they are uncertain about whether they have healed. Setiya’s example is interesting but I think the debate here is still very much open. As Setiya acknowledges, his objection may be dealt with by appealing to degrees of belief: the subject would not clench their fist intentionally unless the subject has a non-zero degree of belief in their success. And as argued by Moss (2013), degrees of belief can be knowledge too. Even without appealing to degrees of belief, the claim that the subject clenches intentionally in the circumstance described is not at all uncontroversial. If intentionality goes together with planning, then one cannot possibly perform a basic task intentionality if one does not believe that one can perform it, for that basic task could not play the role that basic actions are supposed to play in the subject’s plans (for one cannot plan anything from an action that one does not believe one can perform).

    Much more could be said about this (I do discuss many related issues in my work). My point about Necessity is that the debate is wide open and it is not at all clear that intellectualism is losing it. So I think the Joshua is far too quick when claiming that Intellectualism cannot provide a satisfactory explanation for Necessity. And as I have explained, it is simply not true that Intellectualism cannot provide an explanation of Flexibility and Directness.

    II. The Interrogative Capacity View and The Practical Properties of Know how
    According to Joshua, the interrogative capacity view “is uniquely well placed to defuse the tension between semantic theory and the practicality of knowledge-how.” I worry Joshua has not done enough to support this claim. In fact, I think that in order for the kind of view that Joshua is defending to be able to capture the “practical aspects” of knowing how, it will also need to appeal to something like practical modes of presentation.

    According to the interrogative capacity view, knowing how to F is an ability to come to know situation-specific answers to questions “How to F?” [Brief parenthesis: At the beginning of the paper, Joshua tells us that know how is an ability to generate answers to questions “how to F?” But what does it mean to generate an answer to a question? Later, we are told more:

    However, there is an abilitative relation which does provide a plausible account of knowledge-how: the able-to-answer relation (where answering is understood as correctly answering). According to this view knowing how to V is standing in the able-to-answer relation to the question ‘how to V?’, where one exercises that ability by generating propositional knowledge which answers the question ‘how to V?’. Let’s call this view the Interrogative Capacity view of knowledge-how.

    This formulation of the view is a little problematic. As the author acknowledges later, answering is a speech act (p. 11). But it seems one may know how to do something without being able to perform any speech act. I recommend that Josh formulate his view from the very beginning in terms of ability to know — and so more explicitly along the lines of Stanley & Williamson’s view of skill.

    The natural question is: could one not be able to know situation-specific answers about how to F without necessarily being able to act on them — without being able to F? At p. 9, Joshua considers this objection to his view. He considers somebody who has read many ski books, and because of that is able to know answers to the questions of the form “How does one ski” in all sorts of different circumstances without nonetheless knowing how to ski.

    To this objection, Joshua concedes that for one to know how to F, it might not be enough that one be able to know situation-specific answers to questions “How to F.” One must in addition be able to know those answers on the fly. (Notice here the close resemblance between the cases that motivate Joshua’s appeal to the notion of “on the fly” and those that motivated the intellectualists’ appeal to practical modes of presentation.) But what does it mean to know those answers “on the fly”? Here I think Joshua has not told us enough and he has not convinced me that the notion of “on the fly” is in better standing than the notion of practical modes of presentation as employed by earlier version of intellectualism.

    Joshua imposes three conditions on what it takes for one to know an answer on the fly. The first is that the answers ought to be about situation-specific methods; the second is that they need to be about practical situations. Finally, we are told that there is something distinctive about the way such answers are generated (and, I guess, known).

    First off: What is a practical situation? What distinguishes a practical situation from a non-practical situation? Is a practical situation practical under a description but not under another? Joshua does not tell us much of anything here. He just invokes an intuitive understanding of practical situations. But if the notion of a practical situation is supposed to do theoretical work in explaining the metaphor “on the fly,” we should expect a much more worked out view of what a practical situation is. Secondly, how specific must a suitable-specific method be? Just saying that they ought to be situation-specific methods is not very explanatory. Finally, it seems to me that the interesting question is: how does one get to have the ability to come to know situation-specific answers to questions “How to F”? (I will return to this issue later.)

    Thirdly, Joshua’s account of the ‘distinctive way in which answers are to be generated and known’ is much less than ideally worked out. Joshua does not really give a general story here. He merely hints at one. He tells us (p. 10) that the answers may have to come about as one acts:

    Rather than the paradigm being generating an answer before action begins, I want to pick out a kind of answering which goes on throughout the process of acting. The kind of process I have in mind is one Ryle describes by saying that a skilled mountaineer walking in difficult conditions “is concomitantly walking and teaching himself how to walk,” (Ryle 2009: 30). Working out how to do something need not be something which one can only do in one’s head. It can involve working through the activity, picking up relevant propositional knowledge as one goes along and setting relevant feedback loops into action.

    And he draws some analogy with the case of math solving problem (p. 11):

    To get a grip on this idea, consider the way in which one might solve a difficult maths problem. Although sometimes it might be possible to just ‘see’ what the method for solving a problem is, the more usual way is to work out how to answer the problem as one goes along. Depending on the problem, one might either split it into sub-problems and work on them in turn, or one might just try out different techniques and see what sticks. I suggest that we should think of the problem-solving strategies involved both in cases in which one in which one works out how to solve the problem as one goes along as involving the exercise of the ability to generate knowledge of how to solve the problem.

    This paragraph is suggestive but falls well short of a fully worked out account of what it means to come to know an answer on the fly. So, it does look as if Joshua is invoking something at least as obscure as the notion of practical modes of presentation (probably more obscure, as “on the fly” is a metaphor). In fact, talking about the “distinctive” way in which answers come to be known sounds like a less explicit way to invoke practical modes of presentation, which on a Russellian construal are exactly that: ways in which propositions can be known.

    Joshua probably should invoke practical modes of presentation. Else, there is nothing in Joshua’s discussion that ensures that coming to know a situation specific answer in a practical situation will be ability-entailing in the way desired and promised by Directness. For could not one figure out situation-specific answers in the right practical situations, and also as the practical situation develops, thereby satisfying all of Joshua’s three conditions, without being able to act on those answers? I am quite worried that Joshua’s view cannot properly explain Directness unless it also appeals to practical modes of presentation.

    The final issue on which I want to draw attention is Joshua’s alleged explanation of flexibility. According to Joshua, the interrogative capacity view of know how is uniquely placed to account for flexibility — our ability to react flexibly to a wide range of possible situations. How does the interrogative capacity view account for flexibility? Well, we are told that if knowing how to F is a matter of being able to come to know situation-specific answers to the question “How to F,” then “we can explain the flexibility of knowledge-how by appealing to the fact that an interrogative capacity can produce different propositional knowledge for different situations” (pp. 13-14).

    Stanley and Williamson (2016) identify skill with a disposition to know and put forward the same explanation of flexibility that Joshua proposes. So the considerations that follow apply equally to Joshua’s view of know how and to Stanley and Williamson’s view of skill.

    I want to take issue with the claim that the interrogative capacity view or Stanley and Williamson’s view provide an explanation of the flexibility of know how/skill. One thing is for a view to predict flexibility; another is for that view to explain it. When it comes to flexibility, I take it that what is to be explained is our coming up with situation-specific solutions across a variety of different cases. Joshua wants to explain that in terms of our ability to come up with situation-specific solutions. But as an explanation, that leaves a lot to be wished for. For it simply traces our successes at being flexible across situations back to some mental property (an ability) of the subject of which we are told close to nothing. In other words, Joshua is explaining something (know how), in terms of something else (the ability to know situation-specific answers) of which we are not given any real explanation. Moreover, unless something more is said about what makes such an ability possible — about what grounds such an ability — it is not even clear we have not gotten into an explanatory circle. For on most theories of ability (see Manley and Wasserman (2008)), the ability itself will be thought of in terms of reliable successes (in this case, successes at coming up with flexible answers across possible and actual situations). A really explanatory view of flexibility would be one that explains what grounds such an ability and such flexibility — what makes such ability possible. That is the kind of explanation of flexibility I aim to give in my work.

    In my opinion, there is where the very interesting questions about know how and skill stand. But those are precisely the questions which views such as Joshua’s or such as Stanley and Williamson’s are not really equipped to answer. (I think these dispositionalist views like Joshua’s and Stanley and Williamson’s are not explanatory in other respects too but I will not go into that here. I discuss other issues with Stanley and Williamson’s dispositionalism in a blogpost about my work at https://politicalphilosopher.net/2016/04/15/featured-philosop-her-carlotta-pavese/ and in my forthcoming piece in Phil Compass (Pavese (2016).)

    Conclusions
    I argued that intellectualism, properly developed, has all the resources to account for the practical properties of know how. In contrast, it is less clear that Joshua’s interrogative capacity view provides a satisfactory explanation of those very same properties, unless he avails himself of practical modes of presentation. If so, Joshua’s main criticisms against intellectualism threaten to strike back hard at the heart of his proposal.


    References

    Bratman, Michael (1999). Faces of Intention: Selected Essays on Intention and Agency. Cambridge University Press.
    Gibbons J. (2001). Knowledge in action. Philosophy and Phenomenological Research 62(3): 579-600.
    Goldman, Alvin I. (2015). Theory of human action. Princeton University Press.
    Glick, E. (2013). Practical Modes of Presentation. Nous, 49(3):1–22.
    Hawley, K. (2003). Success and Knowledge How. American Philosophical Quarterly, 40(1):19–31.
    Manley, D. and Wasserman, R. (2007). A gradable approach to dispositions. The Philosophical Quarterly, 57.226:68–75.
    Moss S. (2013). Epistemology formalized. Philosophical Review 122(1): 1-43.
    Pavese C. (2015a). Knowing a Rule. Philosophical Issues. A Supplement to Nous, 25(1): 165-188. DOI: 10.1111/phis.12045
    Pavese C. (2015b). Practical senses. Philosophers’ Imprint, 15(29): 1-25.
    Pavese, C. (2016), forthcoming. Skill in epistemology, part II: Skill and know how. Philosophy Compass.
    Pavese, C. (201?), forthcoming. Know how, intentions, and abilities, Synthese.
    Ryle, G. (1949). The Concept of Mind. Chicago University Press, Chicago
    Setiya K. (2011). Knowing how. Proceedings of the Aristotelian Society 112(3): 285-307.
    Stalnaker R. 2012. Intellectualism and The Objects of Knowledge. Philosophy and Phenomenological Research 85(3): 753-761.
    Stanley J. (2011). Know How. Oxford University Press, Oxford.
    Stanley J, Williamson T. (2001). Knowing How. Journal of Philosophy 98 (8): 411-444.
    Stanley J, Williamson T. (2016). Skill. Nous Online first: pp. 1-14

    1. Carlotta’s rich and challenging comments raise two kinds of issue. First off, she argues that some versions of Intellectualism are in a position to explain the practicality of knowledge-how. She then raises a series of important challenges for the Interrogative Capacity view. I’ll consider these two sets of issues in turn.

      Intellectualism and Practicality

      In the paper I claimed that the main challenge for Intellectualism was to account for the practicality of knowledge-how, which I split into three central properties (directness, necessity, and flexibility). Both Intellectualists and Anti-Intellectualists discuss these conditions, which makes them seem common ground in this debate, but I thought that it would be helpful to be a bit more explicit about what the rules of the game are.

      I claimed that Intellectualism faces a challenge in explaining these properties, since other kinds of propositional knowledge do not have these properties. I wasn’t super clear about this in the paper, but I didn’t take myself to show that Intellectualism couldn’t explain these properties, only that explaining these properties is a challenge for the view. I’m extremely happy to have a debate about whether Intellectualism can meet this challenge, and I don’t consider this debate by any means closed.

      In her comments Carlotta faces this challenge head on, explaining how her account of knowledge-how addresses these three conditions. I think that Carlotta’s spin on Intellectualism is probably best positioned to explain the practicality of knowledge-how, and she’s certainly been the most explicit about the challenge of explaining these properties. In particular, her account of practical senses looks to be extremely well-positioned to explain the flexibility which is associated with knowledge-how. Her view is extremely rich, and I I’d recommend that anyone who’s interested in this debate check out her papers, especially the super-interesting Practical Senses link.

      I’d be absolutely up for having a conversation below about Carlotta’s version of Intellectualism and whether it can explain the practicality of knowledge-how, but I won’t address these issues in this comment (it’s already pretty long!).

      Worries about the Interrogative capacity view.

      In the second part of her comments Carlotta raises a set of challenges for the supporter of the Interrogative capacity view. She contends that although Intellectualists need to appeal to Practical modes of presentation (PMPs), the interrogative Capacity view is going to have to end up appealing to a similarly obscure concept (if not PMPs themselves) if it is to explain the practicality of knowledge-how.

      Do we understand Interrogative Capacities?

      Carlotta identifies three elements in the Interrogative Capacity view which are less that fully developed:

      i. The notion of a practical situation
      ii. The fineness of grain of the propositions that are produced by the exercise of knowledge-how.
      iii. Answering ‘on the fly’.

      If these notions are just as obscure as PMPs, then presumably the Interrogative Capacity view is no better off than an Intellectualist account which appeals to PMPs.

      I don’t think that any of these notions are especially obscure, but I’ll say a little more about each of them to address the concern that they are.

      To start off, I must say that I don’t have a general story about what makes for a practical situation. One aspect of the Interrogative view which I perhaps underplay in the paper is that it is Contextualist, in the sense that the set of situations which figure in the way propositions is fixed by conversational context. This means that the general answer to the question ‘which kinds of situations are the practical ones?’ is ‘the contextually salient situations which are apt for V-ing’. That’s not a very informative answer, but it might be the best that I can do.

      As I noted in my response to Evan’s comments, way-propositions can come at various different levels of generality. Carlotta challenges me on this issue, asking how specific the way-propositions which are produced by the exercise of knowledge-how are. On a straight Intellectualist story, way-propositions will need to be somewhat general (at least if a piece of knowledge-how is identified with knowledge of just one proposition). By contrast, on the interrogative capacity view the way-propositions can be arbitrarily fine-grained, since a given way-propositions only needs to apply to one situation. So the answer is that way-propositions are arbitrarily fine-grained, describing a method that is sensitive to a particular array of action-related variables in a particular situation.

      Lastly, Carlotta challenges whether I give a clear notion of what it is to answer a question ‘on the fly’. I think that she’s absolutely right to push on this point: in the paper I definitely don’t say enough about this notion. However, I do think that there is an interesting phenomenon in this vicinity. Here are two routes to understanding the phenomenon I’m interested in:

      Consider cases of practical problem solving. This is the strategy I pursue in the paper by pointing toward the example of maths problems. There are myriad other examples of practical tasks which I think nicely characterise the kind of practical answering process that I’m interested in, and I think that thinking about those examples can help to illustrate the kind of phenomenon. For some especially well-observed description of working out how to do things on the fly, I’d point to Matthew Crawford’s descriptions of fixing engines in Shop Craft as Soulcraft (published in the UK as The Case for Working with your Hands). (There’s a taste of this here [link (http://www.nytimes.com/2009/05/24/magazine/24labor-t.html?_r=0)).
      Appeal to Partial Plans. A more systematic way into understanding the idea of answering ‘on the fly’ is to appeal to the connection between knowledge-how and intentions. I’m attracted to the idea that there is a close connection between the exercise of knowledge-how, and Bratman’s idea that Intentions should be understood as partial plans (Bratman 1976 http://philpapers.org/rec/BRAIPA). As I understand Bratman, his point is that an intention for the future fixes some features of an action-plan, whilst leaving other aspects of that plan open to be decided when the time of action comes. I think that it’s pretty plausible that this process of ‘filling-in’ a partially specified plan should be understood as working out how do the relevant activity on the fly. This connection needs more development, but it’s a gesture towards a more systematic account of what it is to answer a question ‘on the fly’.

      Do Interrogative Capacities require PMPs?

      The second part of Carlotta’s challenges concerns whether the Interrogative Capacity view is adequate without appealing to PMPs.

      One issue which Carlotta raises is whether the interrogative Capacity view can secure the relation between knowledge-how and the ability to act. One might think that it is possible to have an ability to generate propositional knowledge, without being able to act on that knowledge. I think that this worry is closely connected to Evan’s worry about the dual process picture of knowledge-how (one needs to think of acting and answering as seperable processes to think that the abilities can come apart). Because I take the activity of answering on the fly to be a single process of acting and answering, an ability to answer on the fly just is (a certain kind of) ability to act. One caveat is that ‘ability’ talk is highly context sensitive, and there will be cases in which an agent has an interrogative capacity, but cannot exercise it due to external conditions, so there will be a sense in which we can say that someone is unable, despite having an interrogative capacity.

      A second issue which Carlotta raises is whether the Interrogative Capacity view in fact explains flexibility, or merely predicts that knowledge-how involves flexibility. I think that Carlotta is absolutely right that just saying that knowledge-how is a capacity to generate flexible performances doesn’t explain how knowledge-how produces flexible performances. I think that Carlotta’s right to point out that an Interrogative Capacity basically builds in a capacity for flexibility, rather than explaining flexibility

      Carlotta also helpfully observes that an explanation of flexibility will plausibly come from an account of the grounds of Interrogative Capacity. (For the distinction between debates about the nature and grounds of knowledge-how see, the introduction to Bengson and Moffett 2011 link). I absolutely agree: although I have argued that knowledge-how is a kind of ability to generate knowledge, but I have left open the question of what grounds this kind of ability. Here I am tempted to be fairly permissive: perhaps knowledge-how is sometimes grounded in standing propositional knowledge, other times in certain kinds of ability (perhaps in the case of basic actions), and maybe other times in knowledge of the kinds of practical sense which Carlotta discusses.

      The background issue here is whether identifying knowledge-how with Interrogative Capacities explains the nature of knowledge-how, or just leads us in an explanatory circle. I don’t have a reductive account of the relevant kind of Interrogative capacity on hand (I don’t have the resources to reductively explain ‘on the fly’ for example), and I don’t think that identifying knowledge-how with an interrogative capacity explains how knowledge-how disposes us to flexible and intelligent acts. However, I don’t think that this is a fundamental problem for the account. I am attracted by the Williamsonian thought that we can make progress in understanding mental states by situating them within the mental economy, without reducing them to simpler mental states. So, although I agree that the Interrogative Capacity view offers a reductive account of knowledge-how, I don’t think that it’s uninformative: it situates knowledge-how in the mental economy, fixing its properties in place, whilst leaving open the important question of what the grounds of knowledge-how is.

  4. First of all, I want to thank all of the commentators for their really helpful comments. One of the big issues I’ve been facing in trying to give a novel view is to get clear on what one’s view actually is, and these comments have taught me a good deal about the view I’m trying to develop!

    I’m going to split my comments up by commentator, which should hopefully make navigating between threads of the conversation a little easier. There are a couple of overlaps, which I’ll try to make clear.

    Strong Response to Evan Riley

    Evan raises three really interesting questions in his comments. I don’t have proper answers to any of these questions, but I’ll try to develop the problems, and point toward the shape of some answers.

    i) How can a view which identifies knowledge-how with the capacity to answer a question distinguish between someone who has standing know-how, and someone who is just really good at figuring things out?
    ii) Is it really true that a question is identical with a set of propositions? They do not seem to be intersubstitutable in extensional contexts.
    iii) Doesn’t thinking of the exercise of knowledge-how as involving dual processes of generating knowledge and acting on that knowledge over-intellectualise knowledge-how?

    Strong Working it out

    Evan raises the issue of how the interrogative capacity view is to distinguish someone who knows how to V from someone who is quick at learning, but doesn’t know how to V. I agree with Evan that this is a hugely difficult issue for the kind of view that I’m trying to articulate. It is easy to set up cases in which someone who is ignorant about how to engage in some activity can work out how to do it via their general problem solving skills. Someone who has no idea how to fix a flat tyre on a bike might nonetheless fix one by exercising their general engineering know-how. We might think that such a person does have the relevant kind of capacity to answer the question ‘how to fix a bike tyre?’

    The way I’d like to describe these kinds of cases is this: the agent who has general learning capacities, exercises some very general piece of know-how on some rather specific problem, in so doing coming to know some specific way-proposition, and also (perhaps) coming gaining a capacity to work out how to solve this kind of problem in the future. In order to give this kind of description of the case, I’d need to have theoretical resources for distinguishing between exercising some general know-how in V-ing, and exercising specifically knowledge how to V. I have some half-baked ideas about how to do this, but I don’t have anything like a fully worked out story.

    In my defence, I think that standard Intellectualist accounts face a similar problem. Any activity will be associated with many way-propositions of different levels of generality, and plausibly only some of those way propositions will suffice for knowledge-how. Which? (see the Touch-typist example in Hornsby 2011) http://www.oxfordscholarship.com/view/10.1093/acprof:oso/9780195389364.001.0001/acprof-9780195389364-chapter-3. I think that currently Intellectualists don’t have a worked-out story about this issue either. This doesn’t mean that my account is in the clear, just that lots of people in this debate face this kind of problem.

    (It’s also worth noting that Stanley and Williamson’s account of skill will face parallel worries, since their account looks like it predicts that someone who is good at working stuff out counts as being skilled at activities which she is in a position to quickly work out how to do.)

    Strong Are questions sets of propositions?

    Evan notes that I endorse an equivalence thesis:

    ET: a question Wh-F? is identical to the set of propositions {p1, p2, … pn} which answer that question.

    (Or maybe, even better, I endorse the claim that the semantic value of an interrogative is just the relevant set of propositions, so that the set of propositions gives the meaning of an interrogative).

    (Note that it’s controversial exactly which propositions are included in set of answering propositions –- is it just true ones, are they complete or partial answers — but for the moment let’s put these issues to one side, and come back to this in Jay’s comments).

    Now, Evan points out that if ET is true, then the following pattern of substitution ought to be acceptable:

    1) Knowing how to V is the capacity to generate fine-grained answers to the question ‘how to V in Fn’ in the process of acting.
    2) Knowing how to V is the capacity to generate fine-grained answers to the set of answers concerning how to V in Fn in the process of acting.

    But, according to Evan, 2) is nonsense. Why is this? One diagnosis would be that the idea of standing in an abilitative relation to a proposition is flawed in principle. I don’t think that this diagnosis is right. Instead, I think that this phenomenon fits into a wider pattern of cases in which substituting an interrogative phrase with a set of propositions which purport to be its semantic value leads to nonsense.

    Notice that we can find similar cases for other verbs like ‘wonder’ that embed interrogative phrases:

    3) Paula wondered whether Dave Weir will win a Gold
    4) Paula wondered the set of propositions {Dave Weir will win a Gold; it is not the case that Dave Weir will win a Gold}

    (I’m assuming that whether p? denotes the set {p; not-p}, although we get the same weirdness if we assume that whether p? denotes {p}).

    Intuitively, we can wonder into questions, but not into propositions (except in the sense of ‘wonder’ where it means something like ‘marvel’). Should we conclude that questions are not identical to sets of propositions? That would be one way out. An alternative would be to take these failures to be examples of the general phenomenon of substitution problems (which should not be confused with Frege puzzles!) (I owe the idea that these kinds of cases might be examples of substitution puzzles to Luke Burke).

    Here’s an example of this kind of problem. It is popular to think that ‘that p’ denotes the proposition that p, and that propositional attitudes should be thought of as relations to propositions (and that propositional attitude ascriptions pick out these relations). However, there are many examples of sentences in which the substitution of ‘the proposition that p’ for ‘p’ yields a false or meaningless sentence (for examples, see Moltmann 2013: C4 §3.1 http://www.oxfordscholarship.com/view/10.1093/acprof:oso/9780199608744.001.0001/acprof-9780199608744-chapter-5).

    Consider:

    5) Dai hopes that Hannah Cockcroft wins a Gold in Rio
    6) Dai hopes the proposition that Hannah Cockcroft will win a Gold in Rio

    I don’t know what the solution to these kinds of substitution puzzles is, but I presume that the phenomenon is ultimately a rather general one. The point being that sometimes substituting a term for its semantic value doesn’t lead to an acceptable sentence. Any successful solution for ‘that’-clauses ought to be generalizable to ‘wh’-complements. Maybe, not very satisfying, but hopefully that points toward a possible answer (or at least a bigger and more interesting problem).

    Strong Two processes or one?

    Evan worries that my account pulls knowledge-how apart from intentional action because it claims that knowing-how is a capacity to generate propositional knowledge, which makes it one step removed from intentional action. I’m not so sure about the worry about over-intellectualisation, but I agree that the kind of dual process picture which Evan describes (where working how to V, and V-ing are distinct processes) would cause problems for explaining directness. (incidentally, I think that Stanley and Williamson’s view of skill is a good example of something like this dual process view)

    However, what I was trying to go for with the quote from Ryle was the idea that one process can at the same time be a process of acting, and a process of learning a proposition. I guess I wanted to reject a dual process story, and endorse something like a dual aspect picture, where one process is at once acting, and generating knowledge. The idea is that we can see intentionally acting as being a process of answering a how to question, with the idea being that the upshot of that process is both an act, and knowledge of the relevant proposition. As Carlotta points out there’s a lot that’s mysterious about this picture, but this is the kind of account that I’m aiming for.

Comments are closed.