Monday, March 26, 2012

Ryle, Stanley, Bach: Overintellectualizing Intellectualism

Kent Bach has published a review of Jason Stanley's Know How.  It's a generally positive review, though with a strongly critical bent.  Bach is not sold on Stanley's variety of intellectualism, but he does not want to object prematurely.  He finds a lot worth taking seriously, but he also finds a lot of problems that need to be resolved.

In pointing out one of these 'loose ends', as he charitably calls them, Bach suggests a small amount of sympathy for Ryle.  The concern relates to Stanley's claim that, while learning how to do something entails learning a fact, learning how to do something better does not entail learning a new fact.  Bach's fear is that Stanley might be inviting "a Ryle-style regress problem."

Though I haven't read this part of Stanley's book yet (still don't have a copy), Bach's concern is plausibly justified, though not clearly expressed.  We might raise a general question against Stanley:  If knowing how to do something entails knowing a fact, then why wouldn't learning how to do something better entail knowing additional facts?  In at least some cases, learning how to do something better entails learning how to do more tasks.  For example, if I know how to play basketball at a low level and then I learn how to perform additional tricks and shots, I will have learned how to play basketball better.  I will also have also learned how to do new things.  Perhaps there are other cases where learning how to do something better does not entail learning how to do new things, though this is not obvious.  Learning how to do something better may, as a rule, entail learning how to do something one could not previously do.  So Stanley's position is not so intuitively obvious.  Yet, this does not necessarily mean a Rylean regress is looming.  I'm not sure what particular Rylean objection Bach has in mind when he calls attention to this part of Stanley's work.  In any case, I'm curious to see what motivates Stanley's discussion of this issue and how he attempts to resolve it.

The issue I would like to address more deeply concerns intellectualism. Bach does not defend Ryle.  He claims that Stanley has "debunked Ryle's arguments" against intellectualism.  I've recently disputed much of the way Stanley interprets Ryle (1, 2, 3, 4, 5), thereby calling into question how effectively we can say Stanley has critiqued Ryle's arguments.  What I want to focus more specifically on now is Bach's claim, following Stanley, that Ryle "overintellectualizes" intellectualism.  And since I usually defend Ryle by appealing solely to his 1949 work, The Concept of Mind, I'll take this opportunity to discuss Ryle's earlier paper, "Knowing How and Knowing That" (1946, reprinted in Ryle, Collected Essays: 1929 - 1968, 2009, pp. 222 - 235; all page references are to this volume), where he first develops his arguments against intellectualism.

Here is what Bach says:

Ryle used various regress arguments against intellectualism to motivate a dispositionalist account of knowing-how. In Chapter 1, "Ryle on Knowing How," Stanley refutes them. The main one assumes that for knowledge of a fact to play a role in action, it requires the prior performance of the distinct action of bringing that knowledge into consideration. But performing that action requires know-how in its own right, and that knowledge must in turn be actively brought to bear; and so on, ad infinitum. Stanley shows that Ryle in effect overintellectualizes intellectualism. Ryle assumes that factual knowledge (whether or not it includes knowing-how) is inherently "contemplative" and isolated from action until actively brought into play. But why assume that? As Stanley argues, knowledge-that can play a direct role in action, just as it plays a direct role in reasoning (otherwise, something like Lewis Carroll's famous paradox of "What the Tortoise Said to Achilles" would arise). Knowledge does not require action to be activated.

Bach is perhaps thinking of Stanley's claim that Ryle makes the following claim:  If intelligent performance is guided by a rule or criteria for that performance, it must be preceded by an anterior acknowledgment of that rule or criteria.  I've already shown that Ryle in fact makes the opposite claim in The Concept of Mind (1949).  Stanley has gotten Ryle's view backwards.  This is also clear if we look at "Knowing How and Knowing That," where, for example, Ryle (1946, p. 228) writes:
When a person knows how to do things of a certain sort (e.g., cook omelettes, design dresses or persuade juries), his performance is in some way governed by principles, rules, canons, standards or criteria. (For most purposes it does not matter which we say.) It is always possible in principle, if not in practice, to explain why he tends to succeed, that is, to state the reasons for his actions. It is tautology to say that there is a method in his cleverness. But his observance of rules, principles, etc. must, if it is there at all, be realised in his performance of his tasks. It need not (though it can) be also advertised in an extra performance of paying some internal or external lip-service to those rules or principles.
Ryle's point, both here and in The Concept of Mind, is that propositions can be about the rules we follow in our intelligent behavior, but the fact that we follow them in our intelligent behavior does not mean we must form propositions about them.  It does not mean we must acknowledge them.  Rules and criteria are followed, but they are not necessarily attended to, in intelligent behavior.  Thus, Ryle says, there is a difference between 'knowing that X is a rule' and 'knowing X.'  In Ryle's view, to know that X is a rule is to be in a certain relation to propositions about X as a rule, whereas to know X is to be in a different sort of relation to X.  Moreover, Ryle warns against trying to reduce the knowing-how relation to the knowing-that relation.  Our knowledge of rules cannot be reduced to the knowledge that our rules are rules.

Ryle (1946) makes the point on page 225 as well:
Philosophers have not done justice to the distinction which is quite familiar to all of us between knowing that something is the case and knowing how to do things. In their theories of knowledge they concentrate on the discovery of truths or facts, and they either ignore the discovery of ways and methods of doing things or else they try to reduce it to the discovery of facts. They assume that intelligence equates with the contemplation of propositions and is exhausted in this contemplation. 
Interestingly, Bach suggests a similar point in his review of Stanley.  He says Stanley "doesn't exclude the possibility that in some cases what one knows in knowing-how is a way of doing something rather than a fact about a way of doing it."  Thus, like Ryle, Bach distinguishes between knowing a way to do something and knowing a fact about a way to do something.  Perhaps Bach is harboring a Rylean skepticism of Stanley's intellectualism.  He is not embracing Ryle, but he is not sold on intellectualism, either.

Against Ryle, Bach claims that knowledge does not require action to be activated.  I'm not sure what this means.  Perhaps the idea is just that knowledge can be employed without conscious intent or reflection.  One can then be guided by knowledge without attending to it.  Ryle (1946, p. 229) presents such a view.  He says intelligent action is like seeing through glasses, whereas contemplation is like taking off the glasses to inspect them:  "In some ways the observance of rules and the using of criteria resemble the employment of spectacles. We look through them but not at them.  And as a person who looks much at his spectacles betrays that he has difficulties in looking through them, so people who appeal much to principles show that they do not know how to act."  Intelligent action is guided by rules or criteria without us having to acknowledge them as such.  So it is not clear that Ryle makes the mistake Bach is indicating.

The question remains:  Does Ryle take knowing-that to be inherently contemplative?  Does Ryle suppose knowing-that only plays an indirect role in action, via thinking?  As I've argued in the past, there is no evidence for such a view in The Concept of Mind.  Can we find evidence in Ryle's (1946)?

Ryle (1946, p. 223) makes the following argument:
If the intelligence exhibited in any act, practical or theoretical, is to be credited to the occurrence of some ulterior act of intelligently considering regulative propositions, no intelligent act, practical or theoretical, could ever begin.
 Ryle's interest is thus in the relationship between acts of considering regulative propositions and acts of intelligence.  His thesis is that the latter cannot be dependent upon "ulterior" acts of the former kind.  Stanley (and, I presume, Bach) agree with Ryle here.  Intelligent acts do not require ulterior acts of considering regulative propositions.

Ryle is also observing that intelligence can be exhibited in both practical and theoretical acts.  The intelligent consideration of regulative propositions is just that:  an intelligent consideration of propositions.  It thus entails knowing-how, and not just knowing-that.  It is important to keep this in mind, because it means that Ryle does not present the knowing-how/knowing-that distinction as a distinction between practical and theoretical knowledge.  Knowing-how is evident in both practical and contemplative activities.  It would be an egregious error to assume that, because knowing-how underlies practical behavior, then knowing-that must be limited to theoretical acts.

So why, then, do Stanley and Bach claim that Ryle thinks of knowing-that as inherently contemplative, as belonging solely to the province of theoretical acts?

Perhaps the accusation stems from a careless reading of the passage I quoted above, from page 225:
Philosophers have not done justice to the distinction which is quite familiar to all of us between knowing that something is the case and knowing how to do things. In their theories of knowledge they concentrate on the discovery of truths or facts, and they either ignore the discovery of ways and methods of doing things or else they try to reduce it to the discovery of facts. They assume that intelligence equates with the contemplation of propositions and is exhausted in this contemplation. 
It might look like Ryle is equating 'knowing that something is the case' with 'the contemplation of propositions.'  Yet, careful analysis suggests otherwise.  Ryle is clearly identifying knowing-that with "the discovery of truths or facts," and he is identifying knowing-how with "the discovery of ways and methods of doing things."  And Ryle warns against trying to reduce the latter kind of discovery to the former kind.  This is clear enough.  What is crucial, then, is how we understand the last sentence of this passage.

Remember, Ryle is talking primarily about intellectualism.  "They" in the last sentence refers not only to philosophers in general, but more specifically to intellectualists.  Intellectualism is what Ryle takes to be the dominant strain in philosophy, and it is that -ism which fails to properly consider the relationship between knowing-how and knowing-that.  According to intellectualism, all outwardly intelligent behavior is inherently contemplative, or "intellectual", and therefore must be the result of internal acts of contemplation.  This is the point Ryle is making in the last sentence of the passage above.  Thus, because they take all kinds of intelligent behavior to reduce to acts of contemplation, they take the discovery of ways (knowing-how) to be the discovery of facts (knowing-that).  Therefore, according to intellectualism, all intelligent behavior is an employment of knowing-that.  This is not because knowing-that is essentially contemplative (Ryle has never claimed that it is so), but because contemplation requires knowing-that.  (What would one be employing in contemplation, if not propositions?)

In sum, Ryle would accept the following proposition:  Acts of contemplation employ knowing-that.  This does not mean that all employments of knowing-that entail acts of contemplation.  It certainly does not mean that all employments of knowing-that entail prior acts of contemplation, since that would lead to the very sort of regress Ryle is warning against.


To emphasize the point, let's look at what Ryle says about the relationship between knowing-how and knowing-that a few pages later (Ryle 1946, p. 233):
When a person knows how to do things of a certain sort (e.g., make good jokes, conduct battles or behave at funerals), his knowledge is actualised or exercised in what he does. It is not exercised (save per accidens) in the propounding of propositions or in saying ‘Yes’ to those propounded by others. His intelligence is exhibited by deeds, not by internal or external dicta. A good experimentalist exercises his skill not in reciting maxims of technology but in making experiments. It is a ruinous but popular mistake to suppose that intelligence operates only in the production and manipulation of propositions, i.e., that only in ratiocinating are we rational.
Knowing-how is identified with successful performance, while knowing-that is identified with making or affirming statements.  In other words, employments of knowing-that only seem to require apt linguistic behavior, whereas knowing-how is identified more broadly with intelligent, rule-governed performance.  Factual knowledge is not essentially contemplative, nor does the employment of knowing-that necessarily entail contemplation--unless we were to assume that Ryle takes the mere recitation of maxims to entail an act of contemplation.  That would hardly be a charitable assumption.

Ryle's point is that acknowledging propositions is more the sort of behavior we identify as "knowing-that," whereas intelligently following rules is more the sort of behavior we classify as "knowing-how."  Sometimes employments of propositional knowledge involve knowing-how, and sometimes they don't.

This might look confusing if we forget that Ryle takes knowing-how and knowing-that to be logically different sorts of things.  They cannot simply be added together, like different fruits in a salad.  They are not simply two different states of the brain, or two different regions in the brain. It's not that sometimes our propositional knowledge states are employed without our non-propositional knowledge states, while other times they are employed together.  Combining propositional knowledge with non-propositional know-how does not result in two distinct knowledge states working together.  While manifestations of knowing-that can involve knowing-how, this is only to say that sometimes propositional knowledge is directly involved in intelligent action.  At least, it would be highly uncharitable to take Ryle to be saying anything more than that.  Ryle has not clearly overintellectualized intellectualism.

It might be asked:  If Ryle believes intelligent action can involve propositional competence directly, without requiring prior acts of acknowledging propositions, then why does Ryle claim that intelligent action cannot be reduced to propositional competence?

Stanley addressed a similar question to me via email a couple of years ago.  He did not like my reply, which was more or less as follows: Ryle's view is that only some kinds of intelligent action directly involve propositional competence.  Namely, it is those kinds which require the use of propositions. Since Ryle takes propositions to be utterances, then propositional competence is demonstrated only with the use of (and interaction with) language.  But contemplation is neither the only nor the most basic manifestation of linguistic competence.  So it is wrong to suppose that Ryle takes all employments of knowing-that to involve acts of contemplation.


Once again, close consideration of Ryle shows that he does not overintellectualize intellectualism.  What he does do is restrict the notion of propositions to the domain of linguistic functions.  This move, while perhaps controversial, and certainly contrary to Stanley, should not be so easily dismissed.

Perhaps Ryle does overintellectualize intellectualism somewhere.  Bach mentions Lewis Carroll's paradox, 'What the tortoise said to Achilles?' Ryle (1946) mentions that paradox, too (and Stanley emphasized its importance during our email exchange a couple years ago), so let me address that paradox before concluding my case.

Carroll's presentation is a lot more entertaining than Ryle's, so I'll quote Carroll (1895), beginning with the Tortoise asking Achilles to write down a set of premises:

"Now write as I dictate: --
(A) Things that are equal to the same are equal to each other.
(B) The two sides of this Triangle are things that are equal to the same.
(C) If A and B are true, Z must be true.
(Z) The two sides of this Triangle are equal to each other."

"You should call it D, not Z," said Achilles. "It comes next to the other three. If you accept A and B and C, you must accept Z."

"And why must I?"

"Because it follows logically from them. If A and B and C are true, Z must be true. You don't dispute that, I imagine?"

"If A and B and C are true, Z must he true," the Tortoise thoughtfully repeated. "That's another Hypothetical, isn't it? And, if I failed to see its truth, I might accept A and B and C', and still not accept Z. mightn't I?"

"You might," the candid hero admitted; "though such obtuseness would certainly be phenomenal. Still, the event is possible. So I must ask you to grant onemore Hypothetical."

"Very good. I'm quite willing to grant it, as soon as you've written it down. We will call it
(D) If A and B and C are true, Z must be true.

"Have you entered that in your notebook?"

"I have!" Achilles joyfully exclaimed, as he ran the pencil into its sheath. "And at last we've got to the end of this ideal race-course! Now that you accept Aand B and C and Dof course you accept Z."
   
Of course, the Tortoise does not accept Z, because the Tortoise requires yet another premise, such that if one accepts A, B, C and D, then one must accept Z, and so on, ad infinitum.  The moral: being able to draw a logical conclusion from a set of premises cannot be reduced to the acceptance of that (or any other) finite set of premises, for one must always know how to get from the premises to the conclusion.

The Tortoise responds to Achilles:  "I accept A and B and C and D. Suppose I still refused to accept Z?"  Achilles responds:  "Then Logic would force you to do it! . . . Logic would tell you 'You can't help yourself. Now that you've accepted A and B and Cand D, you must accept Z!' So you've no choice, you see."

The Tortoise does not see, and that is the point--or rather, the question:  What is the Tortoise's problem?  What does "not seeing" comprise in this case?

Ryle concludes that the Tortoise is not missing a bit of propositional knowledge.  No new fact will solve the Tortoise's problem.  Rather, Ryle urges that understanding the logical relation between premises and conclusions cannot be reduced to (or explained entirely in terms of ) knowing facts.  One can know the facts and still not "see."  The capacity to reason from premises to conclusions must include some knowledge of a method or a way of applying our knowledge of facts, and this understanding cannot be explained as the knowing of yet another fact.  (Stanley might say it entails our knowing the facts better, an option which needs to be critically explored, as I've already noted.)

Let's look more closely at what Ryle (1946, p. 227) concludes about the Tortoise and Achilles:  
Knowing a rule of inference is not possessing a bit of extra information but being able to perform an intelligent operation. Knowing a rule is knowing how. It is realised in performances which conform to the rule, not in theoretical citations of it. . . . Principles of inference are not extra premisses and knowing these principles exhibits itself not in the recitation of formulae but in the execution of valid inferences and in the avoidance, detection and correction of fallacies, etc.
Ryle's conclusion entails a distinction between (1) knowing rules and principles of inference and (2) being able to recite or cite such rules.  Being able to reason requires more than being able to talk about or indicate the rules of inference.  It entails being able to employ them intelligently.  So the distinction between knowing-how and knowing-that looks, again, like the difference between being able to apply criteria and being able to repeat, identify or indicate facts.

Ryle is also making another point here:  Principles of inference are not facts.  Knowing a principle is not knowing a fact.  We might know that such-and-such is a principle, in which case we have factual knowledge of a principle.  But knowing the principle as such is not propositional, in Ryle's view.  Propositional knowledge enjoins our knowledge of rules, but it does not comprise that knowledge.

Stanley might object to Ryle's distinction between 'knowing a rule' and 'knowing that such-and-such is a rule,' in which case he presumably would object to a strong distinction between knowing a way and knowing that such-and-such is a way.  It would be consistent for Stanley to take 'knowing a way' to mean 'knowing that w is a way' in a certain way.  This would allow him to regard 'knowing a way' (which he might agree is knowing-how) as a variety of propositional knowledege.  But it also means he cannot be using the term "proposition" in the sense Ryle adopts.

It might look like the dispute is semantic.  Stanley wants to take knowing-that as basic, and to regard all kinds of knowledge as propositional, but this might not be much more than a preference for how we name and organize the chapters in our treaties on semantics and linguistics.  Does it matter if we call knowing-how propositional knowledge, if our understanding of it remains the same?

Stanley must think that our understanding of it does not remain the same.  He believes we can learn something interesting about knowing-how by finding out if it is propositional knowledge.  But what sorts of things does Stanley think he can find out about knowing-how?  This question may, more than any other, get at the difference between Stanley and Ryle.

According to Bach's review, Stanley is open to the possibility that propositional knowing-how is a folk fiction of sorts, and that it does not correspond to any phenomena in the brain or anywhere else.  Bach writes:
Stanley acknowledges that the conclusions he draws from his semantic analysis of ascriptions of knowing-how might not square with findings in cognitive science and that, perhaps, "work [in the cognitive sciences] shows that knowing how is the phlogiston of folk psychology" (p. 149). Assuming his account of knowing-how captures the folk psychological notion, it might turn out that the psychological states we describe as knowing-how either aren't propositional, as Stanley's account requires, or that they are not knowledge states at all.
Stanley's supposition is that folk psychology, in employing talk of knowing-how and knowing-that, is postulating the existence of neurological cognitive states.  Perhaps those states do not exist, in which case our folk psychology is wrong.

Ryle's view is quite different.  He does not suppose we are postulating actual states when we talk about 'mental states.'  He says that this talk is figurative, and that when we ascribe mental states to people we are giving ourselves license to make a wide variety of explanatory-cum-predictive assertions about their behavior.  We are only indirectly implicating any actual states (such as neurological states) in the process.  The mental idiom does not make causal predictions about brain states, and so cannot be disproved through neurological research.

What Stanley seems to be saying (if Bach has presented him clearly) is that, if no actual states correspond to know-how relations between persons and ways, then know-how is a mistaken hypothesis which might be replaced by a better one, just as the concept of phlogiston was made unnecessary by a better understanding of oxygen.  Ryle, in contrast, says that such an attitude is the result of a category error:  it takes entities in the mental idiom to be the same sorts of things described in the idiom of physical entities.  His argument against intellectualism (as developed in The Concept of Mind) is part of his more general argument against this very sort of confusion.

By advocating intellectualism, Stanley is not just claiming that knowing-how is propositional; he is claiming that our mental idiom makes predictions about how our brains cognitive faculties work.  If Stanley is right, then we can explain knowing-how in terms of neurological cognitive states which comprise relationships between persons and propositions.  Ryle would object.  His analysis of propositions (see, for example, his 1929 paper, "Are There Propositions?," also reprinted in the 2009 Collected Essays), and his regress arguments against intellectualism aim to show that this view is illogical.

So far, I don't think Stanley has successfully dealt with Ryle.  Of course I have to read Stanley's book before I can state my definitive opinion, but I think I've amassed enough evidence to cast a substantial shadow of doubt over Stanley's treatment of Ryle.  I also think Bach is right to remain skeptical of intellectualism, though I respect how he manages to approach it with optimism and encouragement.  I think, if Stanley's project is going to work, it will require a view of propositions which is not only very unlike Ryle's, but also very much unlike that which has predominated in philosophical discourse for the last century or so.  I also suspect Stanley is aware of this fact.

Update:  As the text indicates, I've made some changes.  Thanks to Kent Bach for pointing out that he never mentioned neurological states in his review.  I had assumed a common point of view in cognitive science:  that cognitive states supervene on neurological states.  Stanley presumably does not assume this position.  Yet, if mental states are capable of being likened to phlogiston in Stanley's analysis, he must take them to be the sorts of things we could rule out through empirical means.  So they are presumably physical phenomena, in Stanley's view.  If not neurological states, what?  Or is he using the term "states" figuratively, a la Ryle?