Wolfgang Schwarz

Our view isn't the relatively innocuous view
that knowing-how entails knowing that. It's the more dramatic view that knowing how just is knowing that.
As we note in our paper (footnote 42, p. 437), German isn't a good language to conduct this inquiry. For German doesn't allow infinitives in embedded questions. All of the following are ungrammatical auf Deutsch:

(1) * Hans weisst wie zu schwimmen.
(2) * Greta weisst wen anzurufen.

undsoweiter...

As we write in our paper, "As a consequence, there is no direct German translation of [knowing how to F constructions]...This distinction between English and German in no way indicates some deep conceptual difference between the English 'know how' and the German 'wissen wie'. It merely reflects the brute syntactic fact that German embedded questions cannot occur in untensed clauses, no matter what the question-embedding verb may be."
All you've got in German is "konnen", which is a translation of "can", and not "know how". It clearly begs the question against us to suppose that "konnen" is the translation of "know how"!
(Btw: the example you cite from Rumfitt's paper is our example (31b), which Rumfitt claims is ungrammatical -- which he has to, given his view that knowing-how isn't knowing that).

I will check out the Kemmerling piece. I've always been curious to see a German discussion of Ryle. In English, Ryle's claim is the dramatic, and on the surface, rather implausible claim that "John knows how to ride a bicycle" expresses the same proposition as "John can ride a bicycle". Obviously, for the reasons mentioned in the previous post, the drama of Ryle's claim is not translateable into German. It's no surprise to anyone to be told that "Hans kann Fahrenfahren" expresses the same proposition as "Hans kann Fahredfahren".
In German, it is grammatical to say "Hans weiss wie man Fahred faehrt". But this is a very awkward construction (as is the English construction "Hans knows how one rides a bicycle"). It also isn't synonymous with the desired reading of "Hans knows how to ride a bicycle" (see our discussion on pp. 424-5). The desired reading of the latter ascribes knowledge of one proposition to Hans, and the desired reading of the former ascribes knowledge of another.

"John knows how to swim" and "John kann schwimmen" are as synonymous as translations usually get. I don't think this begs any question. At any rate "John weiss wie man schwimmt" is hardly the correct translation, except in very extraordinary (but clearly specifiable) contexts.

On the other hand, "John weiss wie man zum n?chsten Ort kommt, wo es Bier gibt" ("John knows how to get to the nearest place selling beer") is not at all awkward. Here using "kann" would be very odd, again except in extraordinary but clearly specifiable contexts. So it is a linguistic datum that "knowing how" has two translations in German, and that there are systematic rules for when each translation is the correct one. Moreover, the distinction drawn in German between these two translations closely matches the distinction in French between "savoir comment faire" and "savoir faire". From the point of view of a native German, the English "knowing how" seems clearly ambiguous or indeterminate, as it can't distinguish the ordinary "John kann schwimmen" from the extraordinary "John weiss wie man schwimmt" with their different truth-conditions. (The latter requires that John knows some non-trivial proposition about how to swim, the former doesn't.)

My main point however was that Ryle's distinction makes perfect sense in German, with "knowing how" not translated as "wissen wie" but as "k?nnen".
I don't think it is of any importance to Ryle whether "John knows how to ride a bicycle" expresses the same proposition as "John can ride a bicycle". His main thesis is that not all intelligent (or competent or rational or reasonable) behaviour is grounded in propositional knowledge. This is not a claim about the semantics of "knowing how".

You have not succeeded in raising a single worry for our position.

First, you are correct to say that "John knows how to swim" is not correctly translated as "John weiss wie man schwimmt". But the point is irrelevant. "John knows how to swim" is not correctly translated (in English!) as "John knows how one swims" (read our paper). So this is irrelevant.

Second, your claim that "John knows how to swim" is correctly translated as "John kann schwimmen" just *is* Ryle's claim that "John knows how to swim" is synonymous with "John can swim". That's what we're challenging. So it's a mystery to me why you think you can appeal to it.

Third, your claim that "John weiss wie man zum n?chsten Ort kommt, wo es Bier gibt" is correctly translated as "John knows how to get to the nearest place selling beer" is also problematic. More precisely, though there is a reading of "John knows how to get to the nearest place, where there is beer" which translates this, it is not the natural reading of "John knows how to get to the nearest place that sells beer" (again, read our paper -- in particular, the part about PRO).

Fourth, you don't need to read Ryle to know that there is a distinction between knowledge (wissen) and ability (koennen). So this can't be Ryle's interesting claim. Ryle's interesting claim is that there is an sense of ability that is deeply related to knowledge; intellectual ability, if you will. His claim that "know how" expresses this sense of ability is evidence for this claim.

So: of course Ryle's distinction makes perfect sense in German, with "knowing how" not translated as "wissen wie" but as "k?nnen". The distinction between knowledge and ability also makes sense in English -- it's the distinction between knowing and being able. You really don't need to appeal to German to make the distinction between knowing (wissen) and being able (koennen).

Fifth, perhaps one of the conclusions Ryle wanted to reach is that not all intelligent (or competent or rational or reasonable) behaviour is grounded in propositional knowledge (though this claim is awfully vague -- what does 'grounded' mean?).
But, insofar as I understand the claim, we argue that all intelligent (and competent and rational and reasonable) behavior *is* grounded in propositional knowledge. We do so by giving an analysis of knowledge-how that reveals it as a species of propositional knowledge, and then show how this analysis explains all of the phenomena that Ryle discusses.

Sixth, about your claim that "It is a linguistic datum that 'knowing how' has two translations in German, and that there are systematic rules for when each translation is the correct one." I'm not sure how to respond except to say that linguistic data do not come in this form.

Wie wir sehr deutlich in unserum Aufsatz geschrieben haben, gibt es keine deutsche Uebersetzung fuer "John knows how to ride a bicycle". So es kann nicht sein dass diese Satz *zwei* deutsche Uebersetzungen hat!

I'm obviously not a linguist, so apologies if I don't know the rules of the game. On my common-sense understanding of translation, "John knows how to ride a bike" can be adequately translated into German, and the correct translation is almost always "John kann Fahrrad fahren".

I agree that Ryle's interesting claim, let's call it claim 1, is that there is a kind of intellectual ability. And his claim that "know how" expresses this sense of ability, let's call this claim 2, may have been taken as evidence for claim 1.

But claim 2 is not claim 1 itself. So it seems to me that even if you could demonstrate that the English construction "know how" expresses a species of propositional knowledge, this would at best undermine a superficial reason for his interesting claim, claim 1. (It also still seems to me that "know how" is ambiguous and only expresses a species of propositional knowledge on one reading, but I don't know anything new to say in support of that.)

You say you showed that intelligent behavior is always grounded in propositional knowledge by giving an analysis of knowledge-how that reveals it as a species of propositional knowledge. But what you give is merely an analysis of the English "knowing how". That's why your argument would be a non-starter in languages like German and French where nothing like "knowing how" would be used to state Ryle's claim 1. Why couldn't Ryle just drop all usage of "knowing how" and replace it by phrases like "can", "intelligent ability", etc.?

All fair comments. On the translation point. The reason you think one can often translate "know how" in terms of ability ("koennen") is that one of the reasons we say someone knows how to do something is that we want to communicate that they can do it. So the conversational effect of ascribing know-how can often be achieved by ascribing ability. It is, after all, rare that someone knows how to swim, without being able to swim. Similarly, it is rare that someone knows *that* a good marinade for tuna involves white wine without being able to make a good marinade for tuna.

One of the distinctive differences between "knowing how" and "koennen" (that is, between "knowing how" and "ability") is that you can know how to F, even though you can't F. For example, a skilled ski instructor may know how to execute a certain maneuver, without being able to do it herself, or (to take Carl Ginet's classic example) a piano teacher who loses her arms in a car accident still knows how to play the piano. But she isn't able any more. This distinctive feature of "knowing how" is what sets it off from ability, and is why "koennen", though an apt translation for conversational purposes, is not a genuine translation -- ohne Armen, kann man das Klavier nicht mehr spielen!

I'm not sure I understand your claim 1. But I can tell you what we do in our paper. In the first part of our paper, we give a semantic analysis of "knowing how" (I should say that I do hope that the analysis extends to French -- one would need to argue that "Il sait nager" is really, despite appearances an embedded question, however, and that would take some syntactic research). In the second part of our paper, using our semantic analysis, we apply it to all the cases that might motivate one to think that one needs to postulate some non-propositional sense of knowing, and show that one can explain these cases in terms of propositional knowledge (that is, the propositional knowledge that emerges from our analysis of knowing-how).

But we're certainly not denying that people have abilities. Sure, people are able to do lots and lots of stuff. If Ryle is only claiming that people have abilities, sign us up as Ryle supporters #1 (and #2, speaking for Tim). But I actually didn't think one needed to read Ryle to know that people have abilities!

Yes, the difference you mention between "know how" and "can" is a problem for my claim that "know how" is indeterminate between "k?nnen" and "savoir comment". The ski instructor's knowledge-how would most naturally be translated as "savoir comment" (and "Wissen wie"), so it's no big surprise that it does not entail the corresponding ability. The disabled pianist's knowledge is more tricky. It would be better for my view if there was an interpretation of "he knows how to play the piano" on which it is no longer true of a pianist who has lost both arms. At any rate, it is not determinately false to say "er kann Klavier spielen" in such a case. That's in fact the natural thing to say if the injury that makes it impossible for the pianist to play is only temporary. So perhaps I have to say that "know how" is indeterminate between a particular precisification of "k?nnen" and "savoir comment".

If "know how" really never means "k?nnen" but always what your semantics says it means, I find it mysterious why in German we hardly ever say any of

"er weiss wie er Klavier spielen soll"
"er weiss wie man Klavier spielen soll"
"er weiss wie er Klavier spielen kann"
"er weiss wie man Klavier spielen kann"

even though these are grammatical and correspond closely to your interpretations of "he knows how to play the piano". Why don't we use any of these on the occasions where English speakers use "he knows how to play the piano"? Why do we always say "er kann Klavier spielen" instead? And why do translators usually translate "know how" as "k?nnen" and not as "wissen wie", which is much closer to the original meaning on your semantics?

Anyway, I don't really mind if your interpretation of "know how" turns out to be always correct. I still believe that this does not threaten Ryle's main thesis. Of course this thesis is not that people have abilities. It is that having an ability does not necessarily require having corresponding propositional knowledge of something like a rule which one follows when one displays the ability. You're right that this is vague, but Ryle's target really is a vague cluster of positions.

You write that:

"Ryle's main thesis...is that having an ability does not necessarily require having corresponding propositional knowledge of something like a rule which one follows when one displays the ability."

Well, if this is Ryle's main thesis, then if we're correct about knowing how, Ryle is wrong.

As we argue in our paper (pp. 442-3, in the discussion of Lewis's response to the knowledge argument), when one manifests one's ability to F, and F is an intentional action, then one manifests one's knowledge how to F [Ryle, by the way, is committed to this -- see our reconstruction of Ryle's argument on pp. 413ff.]. It follows that if we're correct about knowing-how, then in manifesting an ability to F, where F is an intentional action, one is manifesting knowing how. Hence, having an ability to F, where F is an intentional action (that is, any action of interest in this discussion), does entail possession of knowing how, and hence does entail possession of propositional knowledge.

I'm not sure about about the rule following stuff -- probably employing one's propositional knowledge doesn't require following a rule. I very rarely follow rules, but I certainly regularly employ my propositional knowledge.

So your argument is something like:

1) If x has the ability to F then x knows how to F.
2) If x knows how to F then x has propositional knowledge.
Therefore
3) If x has the ability to F then x has propositional knowledge.

Here it is important that "knows how" is not ambiguous between having an ability and having some kind of propositional knowledge -- contrary to the semantic intuitions of at least some English speakers (Rumfitt for one) and contrary to what is suggested by its translations into many other languages. For otherwise the argument seems to rest on an equivocation. -- Don't you find it worrying that even though Ryle's position can easily be translated into German, every way of translating your argument against Ryle into German makes it look quite unpursuasive? Doesn't this hint at the possibility that your argument depends upon a peculiarity of English that is not essential to Ryle's position?

Anyway, let's stipulate that "know how" is always understood in the way that makes (2) true. Then I wonder how you can be so sure about (1). It's true that (1) doesn't sound too controversial. But as you acknowledge, "the conversational effect of ascribing know-how can often be achieved by ascribing ability". So maybe (1) sounds okay only because people confuse the information often conveyed in conversations with real truth-conditions.

In fact, on your understanding of "knowing how", it seems that (1) isn't plausible at all. As you agree in your paper, in some cases it is natural to say both that "Paul has the ability to suck" and that "Paul does not know how to suck". If that's true for sucking, I see no reason why it shouldn't also be true for swimming, riding a bicycle or playing the piano.

So to make your argument work, you have to restrict it to a rather narrow class of "intellectual abilities". If that class is sufficiently narrow, your claim is certainly true, but not very interesting anymore. For example, (1) is probably true for the ability to pass a physics exam. But nobody ever doubted that propositional knowledge is required for this kind of ability.

You say the class of abilities for which (1) holds comprises the class of intentional actions. I'm not sure if that is so (still given your interpretation of "know how"). If Paul can be able to suck without knowing how to suck, why can't he also be able to want milk without knowing how to want milk? At any rate, I can't find any argument for this claim in your paper. Moreover, even if the claim is true, it's not clear that this undermines Ryle's position. As I said, Ryle certainly doesn't claim that *no* kind of ability is grounded in propositional knowledge, and he may well grant that your "intentional abilities" belong to the class for which the intellectualist legend is correct.

Oops. In my previous comment I misunderstood what you mean by "intentional action". From the paper it's clear that you mean actions performed intentionally. Wanting milk can presumably be performed unintentionally. Still, I think you need an argument why on your reading of "knowing how" (1) holds for all intentional actions. (Maybe Ryle himself needs that in one version of his circularity argument. If so, this version of the argument may fail. But that doesn't imply that the conclusion of the argument is false. Moreover, in some versions of the circularity argument nothing like (1) is required. These include all versions that can be translated into German.)

I also don't think it's clear that only intentional actions are "of any interest in this discussion". Ryle talks about "intelligent acts", and "operations that can be performed intelligently". It's not obvious to me that this coincides with intentional actions. For instance, a spontaneous decision to do F may be intelligent without being intentional in any ordinary sense, without e.g. being the result of a decision to decide to do F.

Again, to repeat, for entirely accidental grammatical reasons, "John knows how to ride a bicycle" has no direct German translation. So Ryle's view (in German) is the non-startling view that "koennen" means koennen, together with some more grandiose stuff about what "koennen" means. In English, Ryle's view is the startling claim that "knows how" means what "can" means.

Now, in English, we have a construction "John knows how to ride a bicycle", that expresses propositional knowledge. As we argue, when someone employs an ability to F, and F is an intentional action, that entails that they have the propositional knowledge expressed by a "know-how" ascription. Hence, manifesting an ability entails propositional knowledge, and Ryle is wrong.

You say that Rumfitt has the "semantic intuition" that "know how" means ability (I suspect that Rumfitt wouldn't put it this way). Be that as it may, as he admits, Rumfitt's view entails that the following sentences are ungrammatical:

(1) John knows which movie to go to and how to get there.
(2) Politicians know when and how to leave Iraq.

(as well as the examples discussed by Brian).

These intuitions are intuitions about grammaticality -- they carry more weight, because they are less theoretically biased. Our theory explains why such sentences are grammatical, your theory entails that they are ungrammatical.

By the way -- Rumfitt would *not* agree that
"know how" in English is translateable by "koennen". Rumfitt of course accepts that for Ryle's theory to be interesting, it must be about "know how", not what is expressed in English by "can". In particular, Rumfitt completely accepts the distinctive difference between "know how" and "can" ("koennen") that one can know how to F without being able to F (as in Ginet's armless pianist example). That's why Rumfitt (whose German is quite good) did not use German as an example. It doesn't support his case at all.

The issue between us and Rumfitt is whether "umyet", in Russian, is more like "koennen" in German (that is, a direct translation of "can", and so not relevant), or is more like "know how", as far as the Ginet-type cases go. Rumfitt is clear about this in his paper -- that is, he tries to argue that "umyet" is etymologically related to knowledge-words. But as Rumfitt and I both agree, more research is needed on this question (Mandarin Chinese is also relevant).

I'm sorry that I seem unable to make my points clear. By "Ryle's position" I meant his rejection of the intellectualist legend, i.e. his rejection of the claim that whenever x has the ability to do F intelligently then x has corresponding propositional knowledge. This position is of course not correctly translated into German as "'K?nnen' bedeutet K?nnen", but rather as "Nicht immer wenn x die F?higkeit hat, auf intelligente Weise F zu tun, verf?gt x ?ber entspechendes Tatsachen-Wissen". This is a non-trivial thesis in German, and as far as I can see it is exactly the same thesis as the original English one. But no argument like yours can be put forward against it in German.

I also didn't want to say that "know how" determinately means exactly the same als the German "k?nnen", nor did I want to attribute this claim to Rumfitt. What I claim (very tentatively!) is that "know how" is ambiguous or indeterminate between attributing mere ability and attributing propositional knowledge.

Before we move around in another circle, could you (or someone else in case someone else is reading this) explain why a theory according to which "know how to ..." always attributes an ability -- which I never said -- entails that sentences like

1) John knows which movie to go to and how to get there.

must be ungrammatical? As I said, I'm not a linguist, so I may just fail to see the obvious reason. That said, I even think I have a couple of reasons why ambiguities need not show up in ungrammaticalities. Here they are:

1. Why is "John knows Paul and how to swim" ungrammatical? Because a) "knows" is ambiguous (let's say between "knows_1" and "knows_2") and b) the two resolutions of the ambiguity ("knows_1" and "knows_2") are not of the same syntactical type. That's why in ordinary contexts, the syntactical embedding of "knows" already makes clear what the relevant reading must be. But surely a word can be ambiguous without the resolutions of the ambiguity belonging to different syntactic types. In this case I see no reason why the ambiguity must show up in ungrammaticalities.

2. There are examples of words that I would classify as ambiguous but that do not give rise to ungrammaticalities (as far as I can see): "fourdimensionalism", "externalism", "realistic", "conscious", "hot", "ounce", "sister", etc.

3. I see no clear distinction between ambiguity and indeterminacy. But indeterminacy usually doesn't give rise to ungrammaticalities, does it? I think of indeterminacy such as that of "mass" between inertial and rest mass, or of "gavagai" between rabbit and undetached rabbit part.

4. Why couldn't a grammar just tolerate some instances of the "ungrammaticality" generated by an ambiguity (zeugmatic constructions, as they are apparently called)? These constructions are in a sense clearly understandable. If we used them all the time, wouldn't they have to be counted as grammatical? But that wouldn't make the ambiguity go away!

Maybe the technical linguistic (or English?) notion of ambiguity is conceptually tied to the grammatical test. To my bewilderment Brian has frequently argued as if it is. Then I have to use another word to formulate my "theory". Roughly, the procedural test for what I mean by "ambiguity" is that in some contexts where an ambiguous word w is uttered it is reasonable to respond by asking "do you mean 'w' in the sense of ... or in the sense of ...?". So if there is a context in which it would be reasonable to respond to an utterance of "I'm looking for somebody who knows how to perform the freestyle dinner roll" by asking "Do you mean somebody who can describe what to do in order to perform the freestyle dinner roll or somebody who can actually perform the roll?" this would show that "knows how to" has the kind of ambiguity my "theory" (I wouldn't dare to call it so!) postulates.