Is There Something More Than Knowing-how and Knowing-that?

As this is my first blog, please forgive me if it is not consistent with the standard style of common blogs. What is posted below is potentially the first of two installments (i.e., if the second one will ever come through).

Gilbert Ryle has made the famous distinction between knowing-that and knowing-how. Knowing that is the knowledge that something is the case, and knowing how is the knowledge about how to do something. With this distinction in mind, it has been common among students of Chinese philosophy in general and Confucianism in particular to think that Confucians advocate knowing-how in contrast to knowing-that.

Given the fact that Confucians are primarily concerned with moral knowledge and they have undoubtedly put great emphasis on moral self-cultivation, it is clear that knowledge in Confucianism is not merely propositional and theoretical knowing-that. Indeed, if knowing-that and knowing-how have exhausted all possible types of knowledge, then I would certainly also agree that moral knowledge in Confucianism is knowing-how rather than knowing-that or at least is more knowing-how than knowing-that. However, what I would like to challenge is precisely the assumption that knowing-how and knowing-that have exhausted all types of knowledge, and consequently I would also like to challenge the characterization of Confucian moral knowledge as knowing-how. In contrast, I would like to claim that it is other (or more) than knowing-how and knowing-that. I shall use Wang Yangming as an example. 

In Wang Yangming, this moral knowledge is called liangzhi 良知, literally “good knowledge” or, simply, “moral knowledge.” However, it is not the knowledge about the “good” or “moral”; rather it is the knowledge that itself is good or moral. To understand the unique feature of Wang’s liangzhi, it is important to relate it to the well-known neo-Confucian distinction between knowledge of/as virtue (dexing zhi zhi 德性之知) and knowledge from seeing and hearing (wenjian zhi zhi 聞見之知), first made by Zhang Zai 張載 but more fully developed by Cheng Yi 程頤. Knowledge from hearing and seeing is perhaps equivalent to Ryle’s knowledge-that, but knowledge of/as virtue is not equivalent to Ryle’s knowledge-how. What is unique about knowledge of/as virtue is that it inclines its possessor to act accordingly: “When knowledge is profound, action will be thorough. No one ever knows without being able to act. If one knows without being able to act, the knowledge is superficial. Because they know the danger, people do not eat poisonous herbs when hungry, and do not tread on water and fire. People do evil things simply because they lack knowledge.”

It is precisely this aspect of moral knowledge that Wang Yangming wants to emphasize with his idea of liangzhi: “there has never been one who knows and yet does not act. To know and yet not to act is without knowledge” (Wang: 5). Such knowledge that inclines one to act Wang calls liangzhi. Thus he states, “one knows what is good and yet does not act according to this good knowledge (liangzhi), or one knows what is not good and yet does not refrain from doing it according to this good knowledge, only because one’s good knowledge is clouded and one does not make an effort to recover this knowledge.”

To illustrate this type of knowledge that inclines one to act accordingly, Wang uses the famous analogy from the Great Learning: “The Great Learning shows us what are genuine knowledge and genuine action. It asks us ‘to love the good as we love the beautiful color and to hate the evil as we hate the bad odor.’ Here seeing the beautiful color belongs to knowledge, while loving it belongs to action. However, at the very moment one sees the beautiful color one has already loved it; it is not the case that one decides to love it only after seeing it. Similarly, smelling the bad odor belongs to knowledge, while hating it belongs to action. However, at the very moment one smells the bad odor, one has already hated it; it is not the case that one decides to hate it only after smelling it.” In this analogy, since seeing the beautiful color and loving it (or smelling the bad odor and hating it) are not two separate mental states, so knowing what is good and desiring to do the good (or knowing what is evil and not desiring to do it) are not two separate mental states. In other words, liangzhi, good knowledge, will naturally incline a person to act accordingly.

With such an understanding of moral knowledge in Confucianism in mind, let us revisit Ryle’s distinction between knowing that and knowing how. Suppose that we know that we ought to love our parents. As knowing-that in Ryle’s distinction, such a propositional knowledge is entirely compatible with our not actually loving parents. This is what Neo-Confucians call knowledge from seeing and hearing. Now suppose that, in addition, we also know how to love our parents (for example, by keeping them warm in the winter and cool in the summer). As knowing-how in Ryle’s distinction, it is also perfectly compatible with our not actually loving our parents. In other words, neither knowing-that nor knowing-how, as Ryle defines them, inclines us to act according to our knowledge. Even if we know that we ought to love our parents (having the knowledge-that) and know how to love them (having the knowledge-how), we may still fail to love our parents.

However, as we have seen, moral knowledge in Confucianism, knowledge of/as virtue or liangzhi, is precisely the knowledge that inclines us to act according to such knowledge. If our knowledge that we ought to love our parents is the knowledge of/as virtue, and not knowledge from seeing and hearing, then we will not only naturally learn how to love our parents (i.e., seek knowledge-how), if we don’t know how to love them yet, but will take delight in loving them and feel pain for our failure to do so. It is in this sense that I claim that moral knowledge in Confucianism cannot be simply characterized as either knowing-how or knowing-that: it includes both but has something additional; it is more than knowing-that and knowing-how. What is this additional thing? As we have seen, it is the inclination or disposition that a person with moral knowledge has to act according to such knowledge.

68 replies on “Is There Something More Than Knowing-how and Knowing-that?”

  1. I also like Dan’s suggestion “knowing to.” It seems like Aristotle’s phronimos not only knows how, but knows to do what is virtuous — in the right way, toward the right person, at the right time, etc.

    A different issue — I’ve always wondered how exactly Russell’s knowledge by acquaintance vs by description fits into the knowing that and knowing how picture.

    • I’ve been thinking how to characterize such a knowledge in an equally simple way, and Dan’s “knowing to” is indeed a better one than any that I have thought of.

      Do knowledge of acquaintance and knowledge by description both belong to knowing that?

  2. Gilbert Ryle said that in English, “know” can refer to knowing-that or to knowing-how. I wonder whether he or anyone else proposed that “know” can also refer to a third more general thing that is the genus of those two, and that those two species exhaust that genus – so that it would be fair to say that Ryle thinks those two are kinds of knowledge, and are the only two kinds of knowledge. Has anyone held that view?


    Although or because I think Bertrand Russell’s distinction is a confused mess, I think Manyul’s question about Russell is a very helpful and interesting one in the context of Yong Huang’s concerns (as I’ll explain), and so worth addressing at some length, though incompletely.

    Russell was writing at a time when the metaphysical positions of materialism, idealism, neutral monism, etc., were central to philosophy, and in such arguments one discussed the requirements things must meet in order to be objects of knoweldge. One spoke of knowledge of things.

    Two places Russell discussed his distinction are in this paper
    and in The Problems of Philosophy.

    On the question here raised, the latter text is a little clearer, though internally inconsistent. In Chapter IV he says,

    “Again, it is by no means a truism, and is in fact false, that we cannot know that anything exists which we do not know. The word ‘know’ is here used in two different senses. (1) In its first use it is applicable to the sort of knowledge which is opposed to error, the sense in which what we know is true, the sense which applies to our beliefs and convictions, i.e. to what are called judgements. In this sense of the word we know that something is the case. This sort of knowledge may be described as knowledge of truths. (2) In the second use of the word ‘know’ above, the word applies to our knowledge of things, which we may call acquaintance. This is the sense in which we know sense-data. (The distinction involved is roughly that between savoir and connaître in French, or between wissen and kennen in German.)”

    But in Chapter V he says:

    “IN the preceding chapter we saw that there are two sorts of knowledge: knowledge of things, and knowledge of truths. In this chapter we shall be concerned exclusively with knowledge of things, of which in turn we shall have to distinguish two kinds. Knowledge of things, when it is of the kind we call knowledge by acquaintance, is essentially simpler than any knowledge of truths, and logically independent of knowledge of truths, though it would be rash to assume that human beings ever, in fact, have acquaintance with things without at the same time knowing some truth about them. Knowledge of things by description, on the contrary, always involves, as we shall find in the course of the present chapter, some knowledge of truths as its source and ground.
    “We shall say that we have acquaintance with anything of which we are directly aware, without the intermediary of any process of inference or any knowledge of truths. …”

    The former passage says all knowledge of things is by acquaintance; the latter says some knowledge of things is by description.

    In the article linked above, he writes,

    “In order to make clear the antithesis between ‘acquaintance’ and ‘decsription’, I shall first of all try to explain what I mean by ‘acqaintance’. I say that I am acquainted with an object when I have a direct cognitive relation to that object, i.e. when I am directly aware of the object itself. … to say that S has acquaintance with O is essentially the same as to say that O is presented to S. …”(108)
    “… the first and most obvious example is sense-data.” (109)

    “The object of the following paper is to consider what it is that we know in cases where we know propositions about ‘the so-and-so’ without knowing who or what the so-and-so is. For example, I know that the candidate who gets the most votes will be elected, though I do not know who is the candidate who will get the most votes. The problem I wish to consider is: What do we know in those cases, where the subject is merely described?” (108)

    “I shall say that an object is “known by description” when we know that it is “the so-and-so,” i. e. when we know that there is one object, and no more, having a certain property; and it will generally be implied [by my phrase ‘knowledge by description’] that we do not have knowledge of the same object by acquaintance. ‘We know that the man with the iron mask existed,a nd many propositions are known about him; but we do not know who he was.” (113)

    My speculation is that Russell is confused partly because his KD-KA distinction reflects two underlying distinctions he has not distinguished for himself. The first distinction is between indirect and direct knowledge, i.e. knowledge that is and knowledge that isn’t by way of signs or representations. The second distinction is between knowledge by of language and knowledge not by way of language. OR to put the point more briefly, it looks to me as though Russell is overlooking that there are non-linguistic signs that importantly mediate knowledge.

    And that helps him overlook (or so I speculate) a distinction between two kinds of indirectness of knowledge, by overlooking a distinction between two kinds of mediation/signing/representing, which I’ll call Articulation and Evidence. Smoke mediates our knowledge of the presence of fire in the latter way only, but non-linguistic signs commonly help Articulate our knowledge as well, especially in those cases that the word “acquaintance” is likely to suggest to us. I mean the real word, not the Russell-word. By tasting several, I know how the next one will taste; I know that it will taste like that. I know it not so much by inference as by projection. Inference suggests propositions; projection suggests things.

    To illustrate that point about Articulation by non-linguistic signs, I’ll quote from myself quoting myself in our June discussion of these issues, which is here:

    Here’s the quote:

    “The classical Chinese verb zhī 知 most commonly takes as its object a simple or complex noun. Gong-sun Chou asks Mencius, “What do you mean, you ‘know words’ (zhī yán 知言)?” Mencius replies, “From biased language, I know wherein the speaker is misled (zhī qí suŏ bì 知其所蔽) …” (2A2). Similar constructions are familiar in English. First, sometimes the English “know” has a fairly simple noun complement: you know clouds, wine, Paris, Susie, your tools, math, or the taste of potato chips. You are familiar with it, you understand it, or you know what things of that type mean; in general, you know enough about it for the purposes in question. Second, very often “know” has a complex noun complement easily tranformed into an indicative sentence if we add the main element. For example, I might know the capital of France; for I know that the capital of France is Paris. But similarly I might know where the hips go in the foxtrot, the style of your clothes, the shape of the thing I am looking at, what paisley looks like, how potato chips taste, or who Susie is. Here the missing elements are harder to supply, and sometimes the only practical way to communicate the content of the knowledge involves showing something, such as Susie, or a picture, model, diagram, or sample. Paisley looks like – this. [FN here: Hansen casually lumps all such cases together with knowing how and knowing to, as non-propositional knowledge (Hansen 1983, 66; Hansen 1994, 8, 44). Perhaps that is because his attention is more on belief than on knowledge. One cannot “believe” the taste of apples or the capital of France.]”

    That’s a quote from a draft of my paper “Confucianism and Moral Intuition,” corresponding to p.219 in the published version in Ethics in Early China: An Anthology, edited by Chris Fraser, Dan Robins, and Timothy O’Leary, HKU Press, May 2011.

    Consider the following familiar picture of knowledge of a thing – say, a table. You present a table to me. I get a sensory image which amounts to an articulation of a cognition of the table (perhaps by means of some kind of supposed resemblance). That’s a very familiar picture especially in the context of discussions of metaphysics focusing on sensory experience. This image goes with the idea that knowledge of things is commonly by way of non-linguistic signs. The sense-image in the picture functions at once as Articulation and Evidence. I think Russell’s KD-KA distinction confusedly conflates my relation to the table (in this picture) with my relation to the sensations, by overlooking the role of non-linguistic signs.

    How might signs mediate knowing-how and knowing-to? I’ll say something briefly about the latter.

    Here’s a similar picture to the table picture: I imagine what it would be like to be on the receiving end of the things I’m thinking of doing, and the image is unpleasant. The unpleasantness of the image represents the unpleasantness for you if I were to treat you in the way I’m contemplating. That’s a kind of knowledge that is somewhat artificially tied to motivation by the deliberative protocol that is the Golden Rule, but it’s based on something fundamental that is not artificial: unpleasantness tends to represent unpleasantness, not pleasantness, because what Peirce called iconic signs are basic. When I feel your pain, that’s by having some pain of my own, not by having pleasure or red, though some highly inefficient sign systems of the latter sort might be set up with great effort.

    The paper of mine that I mentioned is about how early Confucianism used this fact about natural representing to construct motivating moral knowledge.

  3. I tend toward eliminativist responses, and just mention that “knowing” is present in plenty of folk psychology, but may not refer back to any delineations that any one culture has coined. The episteme/techne distinction distinguishes certain clear-cut cases, but there hasn’t been any empirically rigorous reason to conclude that this is the extent of all of the components to “knowing,” or that this is really a distinction that helps clarify how we anatomically input, process, and utilize stimuli.

    For instance, it’s clearer that the episteme/techne distinction delineates knowing my home address and knowing how to tie my shoes, but what about knowing that I exist (self-awareness). I know my phone number because I took a bit of time to “memorize” it (which is another for-grabs verb on its own). I know how to tie my shoes through practice and refinement through repetition (which in many ways relates to memorization). However, neither side alone tells me that I am me and not the clothes that I’m wearing. If we attempt to answer the empirical question in terms of the pre-made semantic categories that we’ve made, we’re going about the enterprise backwards.

    I don’t know about that whole “moral knowledge” thing. I worry on a practical level because it at first blush seems to imply that immorality is some kind of epistemic problem, and that all sociopaths really need is education on right and wrong or are really like other mentally challenged people (people who can’t acquire episteme or techne knowledge). Then I worry on a factual level because it seems like it relies on historical coinage rather than empirical evidence for support.

    • Joshua, thanks for your post. Regarding your last point, on “moral knowledge,” I think there may indeed be a question about whether there is such a kind of thing. I shall try to address this issue in a follow up post. Here I simply wand to say that, if we define moral knowledge, as in Confucianism, as something that includes inclination, then it it not merely an epistemic problem, unless we redefine the epistemic problem so that it also involves the issue of inclination (but then we would have to take out the word “merely” in our original sentence, as it is a quite complcated problem). Understood in this way, sociopaths are not like mentally changed people but are “xinly” (I don’t know whether, just like “mentally” is an adverb of the noun “mind”, there is an adverb corresponding to the noun “heart” as the [partial] translation of the Chinese character “xin”) challenged people.

    • There’s heartily in English, meaning “with gusto and without reservation,” so it may not exactly be where you were aiming.

      However, we do have a word for lack of inclination to act, regardless of knowledge — apathy. Economic psychological research has shown that we’re actually more apathetic in practice, even with full knowledge of the results of a given action, at least as demonstrated by willingness to increase altruistic donation based on the gravity of the assistance that such a donation would provide. This holds true even when the factors are thousands of lives or more.

      I have little doubt that sociopaths are apathetic (or rather, “sympathetically challenged”), as are other psychopaths. However, whether we need to introduce it as an epistemic matter is still worrisome for the reasons that I mentioned in #4. I think the research on this is the sympathy is not so often taught as it is untaught.

  4. I really enjoyed this posting. A few years ago, my husband attended a lecture on Wang Yangming at his company in Tokyo (You gotta love the big Japanese corporations– for no matter how bad the economy, they continue offering activities for employee self-cultivation). The professor who gave the lecture discussed the dangers of knowledge based on presuppositions (or knowledge ex nihilo). And to illustrate the point he used a famous painting of which scholars had been quite mistaken about the content but never really realized that they didn’t know what they didn’t know. When we then apply this to moral knowledge, 良知, I think Wang Yangming (and, of course, Yong Huang) make a very interesting and good point. I thoroughly enjoyed this and was able to re-write an old blog post on the painting and 知行合一 here:
    Obviously too, in a world where there is a radically de-emphaised mind-body duality, all knowledge would be embodied in some way, leading I think naturally to knowing-to.

  5. I’m interested in mining “knowing to” a bit. One possible source of reduction is to say that knowing-to is just “knowing that one ought to…” and hence is just a species of knowing that. I want to resist the reduction, but I’m not sure how. Any suggestions?

    • That might depend on idiom. There’s at least: “He knows to do that, but he won’t” which sounds normal to me. It’s not necessarily weakness of will, but knowing-to doesn’t guarantee action. It would be interesting if knowing-to tends toward refusals to act whenever it (i.e. knowing to) breaks down. That might capture the “accedia” aspect of “Chinese philosophy” (sic) preoccupations with action breakdown that Nivison wrote about.

    • This is an interesting discussion. What I would like to emphasize is regarding Confucian moral knowledge, liangzhi or dexing zhi zhi, is that it involves an inclination to act according to such a knowledge. However, this does not mean that one will necessarily act. On one particular occasion, one may have more than one inclination and one can only act on one of them, just as there are often more than one thing that one ought to do on a particular occasion, but one may only be able to do one of them. This is a separate issue. I will be satisfied with Dan’s suggestion of “knowing to” to characterize liangzhi as long as it can indeed include the inclination and is not simply an abbreviation of “I know I ought to” as Manyul is concerned with.

    • Hi Manyul – Your comment here looks interesting, but I’m not sure I get it. Does this paraphrase capture what you’re thinking?:

      One sign that knowing-to is thought of as motivating is if situations are discussed in which at first someone was going to do something she knew to do, but then she stopped knowing to do it (she forgot or something), and just because she no longer knew to do it, she was no longer going-to-do-it.

    • Hi Bill; my original thought was that (A1) ‘He knows he ought to Φ’ and (A2) ‘He can’t get himself to Φ’ go together as traditional weakness of will, but (B1) ‘He knows to Φ’ sounds more like it goes idiomatically with (B2) ‘He refuses to Φ’ than with A2. But I’m not at all sure I have any good reason to think that.

    • Oh! I was way off. Hmm – Now that I understand your thought, I adopt it exactly, including the qualification.

  6. Hm. Manyul, good question! At first I was thinking of it simply as a linguistic question about a phrase, and this was my answer:

    I can’t think of any strong reason to think knowing to do X isn’t just knowing one ought to do X, beyond some sense that the construction of the phrase somehow suggests it. One needs more reason than that. (I don’t have such a sense about knowing how, and yet, now you mention it, I find myself more comfortable with the idea that knowing-how is irreducible to knowing that than with the idea that knowing-to is so irreducible.)

    After all, we generally suppose that people have some concern to be moral, polite, reasonable, etc. If I assume you want to do what you ought, and if by “knowing to” I just mean “knowing that one ought,” I’ll think that in your case knowing to will be motivating, even if I don’t think it’s inherently motivating.

    (By the way, Yong Huang, it seems to me that in the conversations as reported in the Analects and the Mencius, Confucius and Mencius generally make pretty strong assumptions of that sort about their interlocutors, so that even if I set aside my inclination to deny that they were primarily concerned about moral knowledge, and that Mencius was concerned about moral self-cultivation, the opening argument of your post seems to me not to work regarding them: “Given the fact that Confucians are primarily concerned with moral knowledge and they have undoubtedly put great emphasis on moral self-cultivation, it is clear that knowledge in Confucianism is not merely propositional and theoretical knowing-that.” But I suppose you don’t have Confucius or Mencius in mind there?)

    But then I thought, maybe that’s the wrong approach to the question. Maybe the right approach is to ask whether there’s a kind of intelligence that is a kind of being-on-track morally (I’m alluding to Nozick’s “tracking the truth” but talking about action) but that isn’t reducible in any obvious or easy way to propositional knowledge about what one ought to do. If there is, then presumably the phrase “knowing to” is about it, I guess?

    • I would like to joyride on your syntactic piece a bit, first by pointing out that there are a handful of likely elliptical constructions for “x knows to p,” so we don’t want to rush out the door with Manyul’s first intuition. Here are a few that come to mind (with the ellipses in square brackets).

      x1. “x knows [that he ought] to p.”
      x2. “x knows [that he is [expected / required / commanded …]]” to p.”
      x3. “x knows [that it would be better [for him]] to p.”
      x4. “x knows [the [ideal / best …] way [how]] to p.”
      x5. “x knows [what it is] to p.”

      These all seem to cover some basic aspects of prescriptive speech acts. Their illocutionary acts are recommendations or commands. They address the subject x as an agent with assumed control over his actions.

      Also, note that x4 allows for a knowing-how reading.

    • Yes — I don’t know why, but it seems to me the common phrase is not elliptical (in any of these ways).

    • The common phrase is syntactically correct as-is insofar as there is a transformation from sentences to “infinitive phrases,” and then the use of that construction as an argument of a sentence, as is the case with many transitive verbs.

      The problem with that route and the hunt for something particular about sentences invoking “…knowing to…” is that if you’re not adopting elliptical constructions, you may already be conceding to a knowing-that, as is common in many parallel sentences (in which the pronoun ‘it’ invokes the previous sentences as its argument):

      “He eats alone. He likes it.” –> “He likes to eat alone.”
      “He eats alone. He wants it.” –> “He wants to eat alone.”
      “He eats alone. He seeks it.” –> “He seeks to eat alone.”
      “He eats alone. He intends it.” –> “He intends to eat alone.”
      “He eats alone. He knows it.” –> “He knows to eat alone.”

      “He eats alone. He likes it.” –> “He likes that he eats alone.”
      “He eats alone. He wants it.” –> “He wants that he eats alone.”
      “He eats alone. He seeks it.” –> “He seeks that he eats alone.”
      “He eats alone. He intends it.” –> “He intends that he eats alone.”
      “He eats alone. He knows it.” –> “He knows that he eats alone.”

    • You haven’t said, but it seems to me that the argument you’re making just here is that

      1. In the first four lines of each list, the sentence on the right is synonymous with the second sentence to its left (in the context of that sentence’s antecedent companion).
      So 2. Each sentence on the far right is synonymous with the far-right sentence using the same verb on the other list.
      So 3. The general fact of the systematic character of language suggests that the same thing is true of the sentences with ‘know’: that the two far-right sentences are synonymous, so that knowing-to sentences can be accurately paraphrased with “know that.” (2)

      But it seems to me obviously false that “He knows to eat alone” can mean anything like “He knows that he eats alone.” (Are you saying below that you are not sure it’s false?)

      So I think I may have misunderstood in some very broad way what argument you’re making here.

    • That actually was my claim, yes. I was saying that I was not so sure that it’s false. My #10 was my failure to find the correct “Reply” button.

      Actually, I think I found the solution in WordNet, the massive lexicographical project at Princeton.

      Sense 10 (“knowing to (1)”): be able to distinguish, recognize as being different
      – “The child knows right from wrong.”

      Sense 1 (“knowing that (1)”): be cognizant or aware of a fact or a specific piece of information; possess knowledge or information about; cognize, cognise
      – “I know that the President lied to the people.”

      So, Bill, you are right that my analogy is incorrect, because I’m switching senses in these two cases. It doesn’t help the normativity issue, however, and the examples below (in #10) don’t suffer this error, and they draw a picture much more like x2. WordNet, however, does not have a “know to” sense that exactly connotes “have an instinct to act a certain way.”

      The closest that it gives is this:

      Sense 7 (“knowing to (2?)”): have fixed in the mind

    • Joshua, I don’t understand your quotes from WordNet. The accounts of senses 10 and 7 each seem to me as obviously false as saying that “knowing how” means “asked when.” Maybe the problem is just that WordNet is employing special terminology that you know and I don’t. I don’t know what the numbers in parentheses indicate.

    • WordNet 3.0 lists eleven senses of the verb to know.

      Under Ryle’s model of “knowing that” and “knowing how,” there are actually a few senses that WordNet offers that fit them, which is why I offer possible mappings from Ryle’s distinctions to a pretty authoritative lexicographical database. The numbers in parentheses indicate that one sense appears to map to one common use of the term as Ryle describes them, while others may better match them as we’ve been discussing “knowing to.” Lexicographers aim for their collection of definiens as separated by terms’ senses (or “synsets’ senses” in WordNet) to generate an exhaustive mapping of the real-world uses of the terms (or “semantic nets”) that they’re defining.

    • Bill, regarding the parenthetical question in your post, perhaps I’m not entirely clear about what you mean, as I think what I say in the sentence you quoted is also applicable to Confucius and Mencius: the knowledge they are primarily concerned with is also moral knowledge. In an article on Confucius on Moral Education published in the Journal of Moral Education early this year (, I just intend to make this point.

    • Hi! Thanks for your reply. Here’s what I was thinking. You had written near the beginning of the original post (I’ve added numbers):

      Given the fact that (1) Confucians are primarily concerned with moral knowledge and (2) they have undoubtedly put great emphasis on moral self-cultivation, it is clear that (3) knowledge in Confucianism is not merely propositional and theoretical knowing-that.

      At first I couldn’t see how premises (1) and (2) could seem to lend any support to (3), much less make (3) clearly true. After all, a person can be interested in more than one thing. Eventually I decided the reading that made the best argument would have to begin like this:

      A) Basically the only thing Confucians are interested in is moral knowledge. (premise)
      B) If Confucians are interested in anything other than moral knowledge, that’s because they think it’s the same as moral knowledge, or a part of moral knowledge, or a means to moral knowledge, …. (from A)
      C) Confucians are very much interested in moral self-cultivation. (premise)

      And I think (A) is clearly false of Mencius and Confucius.

      But your reply today shows me that you never meant to say (A) at all; your (1) doesn’t mean anything like (A). Rather, now it appears that by (1) you only meant “The main kind of knowledge Confucians are interested in is moral knowledge.” So I’m completely in the dark about what line of thought your argument (1)(2)(3) is trying to express. How would the premises lend support to the conclusion? What’s the argument?

    • As my conclusion is a very loose one, “Knowledge in Confucianism is not merely propositional and theoretical knowing that,” it can really be supported by a whole lot of different set of premises. As a matter of fact, given the strength of the my premises, the conclusion is perhaps unnecessarily too weak. As long as Confucianism has any any interest, even not primiary one, in moral cultivation, I can already say that knowledge in Confuicanism is not merely theoretical. So I still don’t see what is the problem in my original passage. In any case, however, what you see as the conclusion is really not one of the main points that I try to make. It is merely one of my concessions.

    • You don’t say what statement you mean by “what you see as the conclusion.” From my comment above, plainly I think your conclusion is (3). I didn’t suggest any other. Possibly you mean (C).

      What I said is that (C) is a premise in an abandoned interpretation of the beginning of your argument.

      I agree that your conclusion can be supported by various sets of premises. Obviously, any conclusion can be supported by an infinite number of different sets of premises. What I cannot see is how your conclusion could possibly seem to be supported at all by the premises you actually offer. Can you explain what the line of thought is that might connect the premises with the conclusion?

      As long as Confucianism has any any interest, even not primiary one, in moral cultivation, I can already say that knowledge in Confuicanism is not merely theoretical.

      How could that premise seem to support that conclusion? I just don’t see it. I apologize for being so blockheaded.

  7. Thanks for those, Joshua! I’m sympathetic to keeping the elision issues up front so there’s no “cheating” in terms of intuitions about the “knowing to” language. What Bill alludes to (two Bill comments up), and what I think the knowing-that/knowing-how issue is really about, are associated categories of mind, mental states, intelligence, or some other appropriate type of thing that those syntactical differences might track.

    Maybe some syntactical uses of “know to” track some kind of practical norm sensitivity or readiness for action, or something of the sort?

    • Manyul, my stand is that “knowing to” doesn’t track much of anything special, but is just another example of a common syntactic move in English. I’m not sure, but I think you are semantically priming yourselves to expect moral terminology, ellipsis, and illocution when they may not be there generally, and are not tracking any particularly moral “mental state” or a “moral knowledge.”

      I think a good clarification of that would be some examples where “x knows to do p” is not already assuming an agent and (at least evidently) does not appear to be a prescriptive, but is instead just giving some sort of plain-old statement of a much more regular sense of “knowing p.” Then replacing p with another prescriptive sentence, we produce something that invokes the kind of intuitions that “knowing to” is generating for you. From x-drop frequency alone (that is, the rate at which terms in the English language are dropped from the complete syntactical construction of a sentence), I would bet my chips on my x2 above. That may just be my own inductive bias, as well, so I’d call it a hypothesis which could stand to a test like this.

      But I’ve laid out my cards earlier, which are not the most charitable to tracking as you describe it. The first is from my agreement with a position that ethical language reduces to factual arguments in which the statement of “x ought p,” is the declared prediction p from assumed premises about preferences, motives, and desires (pitting me in a sort of ethical naturalist camp), plus I’m (in #4) laying out my eliminativist skepticism about the project of tracking via syntax, as well. These posits don’t even let Ryle’s distinction off of the ground, and they fly in the face of ontologically loaded “良知.” (But you all know that I’m characteristically hardcore about those sorts of things.)

      This is relevant reading, but gives a good shot at what I think would be the biggest issue with the approach taken to elucidate “knowing to” on “moral knowledge”: Starting on the bottom of p.8, “The Argument from Syntax:,” I would also level the responses here, only replacing “factive” with “evaluative.”

      On another note, you all speak handfuls of languages. Are there going to be analogous “tracks” and morally loaded “knowings-to” in those languages? If not, are we going to say that some people just don’t get certain “mental states” because of the languages that they employ? If so, that’s a treacherous route to outdated linguistics.

    • On another note, you all speak handfuls of languages. Are there going to be analogous “tracks” and morally loaded “knowings-to” in those languages? If not, are we going to say that some people just don’t get certain “mental states” because of the languages that they employ? If so, that’s a treacherous route to outdated linguistics.

      I agree with the latter point about outdated linguistics. But whether there are analogously tracking turns of phrase in other languages is partly an empirical question and partly a question of interpretive maneuvering. The mental states may be multiply expressible and social contexts might be the things that determine the interpretation. So, I don’t know that we have to shy away; I’m not sure the route is so treacherous.

    • I think at least part of the reason that it is difficult to find a term in English that names a state of heart/mind that includes both knowledge and desire to act according to this knowledge is not that people who speak English don’t get such “mental state” as Joshua suspects, but that they these two elements (knowledge and desire) as two different mental states. The Confucian idea of the single mental state, liangzhi, would be “besire” (beleive + desire, a term coined by Altham), which is regarded as something bizzare. Again, this is the topic of my follow up post (now it seems that I got to do it and to do it soon, as I have announced it too many times).

    • Yong, my point is that you’re treading on a path to linguistic determinism, which is dead wrong, in this line of thinking.

      Or you’re on a path to linguistic relativism, which isn’t yet proven dead wrong.

      This notion of heart/mind or believing/desiring are, to me, highly flawed attempts to interpret them in a way that still has a chance of being correct in light of contemporary empirical knowledge about the workings of the heart, mind, believing, and desiring. The matter seems more befuddled by mixing and matching natural language terms, not clarified.

    • Well, while I am not fully clear about what you mean by all these “isms”, if at all, what I say is precisely against linguistic determinism: the very fact that English, unlike Chinese, does not have a word to describe the single mental state of “besire” does not mean that there is no such a mental state in English speaking people. Of course, simply because Chinese has a word, liangzhi, for such a state does not mean that the state therefore must exist (and only exist in Chinese speaking people). After all, in both English and Chinese, we have many words for things that don’t exist. When I claimed that such a mental state of besire exists as Confucians claim, I didn’t say this is because Chinese has a word for it. So I don’t see why I’ve to be a linguistic determinist in saying what I say. But I will say more about “besire” in the separate post when I have some free time, whether it is befuddling or enlightenming.

    • From x-drop frequency alone (that is, the rate at which terms in the English language are dropped from the complete syntactical construction of a sentence), I would bet my chips on my x2 above.


      x2. “x knows [that he is [expected / required / commanded …]]” to p.”

      Ignoring that this involves a “that” clause, for now, couldn’t this track what I call a “practical norm sensitivity or readiness for action”?

    • Couldn’t you get just the same tracking with, “x is [expected / required / commanded …] to p?”

      I’m guessing that our main difference here is that I don’t think that “practical norms” track any isolated mental sensitivity, but are a mutually acknowledged and trained rhetorical games among interlocutors (with shameless self-linking here: Three-year-olds usually outmaneuver this rhetorically because they appear to realize earlier that only acknowledging mutual practical concerns, namely shared desires, empowers prescriptive language.

  8. Joshua, is it your thought that as the first four items on your first list show each right-hand sentence synonymous with its pair of left-hand sentences, so does the fifth?

    • Sorry, I mean “… roughly synonymous with the second sentence to its left taken in the context of the first …?”

    • I think that following syntactic transformations from the two sentences on the left generate the sentence on the right. Think of “–>” as “transforms into” or “derives”.

      However, my bet is really on x2 because of its x-drop likelihood, parsimony, and a similar prescriptive illocution that follows with, “x is to p.”

    • I understood that you were talking about transformations. But it seems to me that what you have done is shown that there is a transformation that (a) arguably generates the familiar phrase “know to” in a way that (b) unmistakeably fails to suit the cases where we would use “know to.” I don’t see what the exercise is meant to show.

    • Joshua, I feel I need an answer to my #9 question in order to know what you’re driving at.

    • (Sorry to have re-asked this twice, Joshua! I thought the first one didn’t post.)

  9. I’m not so certain that it fails, but violates a few intuitions about the use of the verb with prescriptive sentences as arguments. I admit that it’s counter-intuitive compared to the others, but that may not be a basis for rejecting its actual function in the language.

    But the exercise is meant to show these things: We’re interested in knowing what the initial sentences are that transform to “x knows to p.” Three possible syntactic solutions are that there is an elliptical construction in the sentence, there is a straightforward transformation, or we’re posing that a certain structure is primitive and a constituent to a more-or-less atomic sentence.

    The third just assumes that the thing exists without any constituents, and so is not really a conclusion, just a supposition to a whole “moral knowledge” folk theory. Vacuously, then, the initial sentence is just the atomic sentence, “x knows to p,” but since we’re trying to establish a clear domain for its distinction and set it apart from Ryle’s “knowing how” and “knowing that,” it’s mere assertion fallacy to reify it as primitive in this way.

    The second means that “He eats alone. He knows it.” –> “He knows to eat alone.” is analogous to another transformation. One acceptable, but uncommon one is this: “He eats alone. He knows it.” –> “He knows that he eats alone.”

    The first poses a number of elliptical constructions.

    Based on your previous rejections, I think that you’re aiming for the third solution, while I’m pretty set on the first.

    Here may be another explanation, but it is again elliptical in nature.

    This one is more on par with auxiliary uses of verbs.

    “She has [an obligation / a command / etc.] to do something.”
    “She knows [an obligation / a command / etc.] to do something.”
    “She is [obliged / commanded / etc.] to do something.” (This one is in line with x2.)

    But I think there is another counter-intuitive move that carries over as a transformation, and is still true, but affects how we see certain uses of terms in our language.

    “He eats alone. He gets it.” –> “He gets to eat alone.”

    The issue here is that we often think that the second argument of “to get” must be a simple noun phrase, when it could reasonably be a {sentence –> noun phrase}. Something may similarly occur with “to know” when you import sentences that repeat the subject.

    But it could equally well be elliptical:

    “She gets [a chance / permission / etc.] to do something.”

    But let me give you some examples of “knowing to” that imply nothing prescriptive, but just make a statement about instincts:

    “Baby cuckoo birds know to push the other eggs out of the nest.”
    “We know not to cry out when we’re hiding from the killer.”
    “The natives have long known to build shelters for themselves to protect against the elements.”
    “Superior businessmen, like tacticians, know to keep their friends close and their enemies closer.”

    • Hi Joshua, sorry, I still feel I don’t know how you’d answer my question of #9 (head comment with my immediate follow-up).

    • Thanks Joshua. Your last list of examples here is really helpful. Regarding your statement about them just before:

      I don’t think anyone’s idea was that the direct point of “Smith knows to φ” is to prescribe, nor that the sentence is any more prescriptive in its implications than “Smith knows she ought to φ.” I’m not sure whether either of those ideas is the idea you’re attacking here.

      I think you’re right to point out that “knows to” is not just moral. I was being too narrow above. I think that like moral language, the phrase applies over a wide range of different kinds of norm or value, such as prudence and the as-if normativity that we commonly speak in terms of when we’re talking about instinct and biological function. I think that revised claim fits the examples you list.

    • I was thinking that people were talking of, “x knows to p,” as though “knowing to” were some sort of tracker for “moral knowledge.” My understanding of normativity is that if statements are going to be normative at all, they will have to prescribe that something be true. If they don’t or can’t, I don’t see how they would differentiate themselves from descriptive claims. I’m not saying that “knows to” sentences are not just moral. I’m saying that they’re not moral at all, and that they are used as descriptions of something entirely separate from normativity in general.

      In “x knows to p” as, “x knows [that she ought] to p,” only p is prescriptive, as it assumes that x perhaps does not act according to p. If everyone acts according to p or if []p, it would be hard to settle a norm for it, since prescriptive language assumes that ◊~p is the state of affairs. The case with my latter “knowing to” examples is that the more complex sentence, “x knows to p,” does not contribute to that normativity, but is rather a description of instincts, heuristics, commonalities, etc. Normative language, I thought, always implied a claim that one sought, desired, preferred, insisted that p be the state of affairs.

      For example, “Baby cuckoo birds know to push the other eggs out of the nest,” does not imply that the bird knows he ought to push the eggs out, or that it knows pushing eggs out is prudent, or anything like that. That doesn’t enter the mind of a cuckoo. I would concern myself with trivializing normativity if I allowed myself to regard descriptions of instincts as if they were separate normative domains, because a lot of normative language prescribes acting against their instincts, biases, common heuristics, etc.

      The only values I see “x knows to p,” range over are plain-old truth-values.

      Much of the rest is my attempt to explain why we might be led to believe that we’re focusing on something that’s not actually elucidating to the matter of the existence of “moral knowledge” or some inborn intelligence about morality, since (a) the syntax of a natural language is not going to provide solid evidence for it and risks self-confirming a folk theory, but (b) we are prone to affirming the existence to phenomena that we can more easily contrive within our own syntax and semantics (a psychological trick, that it’s easier to target something with a label, even if that thing isn’t there). Especially concerning, though, is my sense that (c) you (as professional ethicists) have trained yourselves into a sort of hypersensitivity to common invocations of moral language — sort of how some feminists “spot” sexism in every remark between differently gendered people, or how some communists can always “find” proletariat abuse into every transaction — that you sort of greedily took the periphery of prescriptive p’s “x knows to…” to itself be of ethical import, when it does not appear to be that way at all. It was more cherry-picking than discovery, is what I mean here.

  10. I was thinking that people were talking of, “x knows to p,” as though “knowing to” were some sort of tracker for “moral knowledge.”

    Yes, but broader than that; that is, it tracks moral knoweldge no more than “ought” tracks “morally ought.”

    My understanding of normativity is that if statements are going to be normative at all, they will have to prescribe that something be true.

    I think it would be more correct to say “…be the case” or “… be done (or omitted).” But setting that aside –

    Suppose Smith thinks “knowing to φ” is a kind of moral knowledge (a kind of knowing of morality or of some part of morality) that inherently tends to issue in right action φ. It doesn’t follow that Smith thinks “Jones knew to φ” is a prescriptive statement, a statement whose central use is to prescribe. Still, Smith will think it has prescriptive implications. Compare: “Jones knows that Paris is the capital of France” and “Jones knows how to uncover the fact that Paris is the capital of France.” These are mainly statements about Jones, not about geography, though they have implications about geography, viz. that Paris is the capital of France.

    In “x knows to p” as, “x knows [that she ought] to p,” only p is prescriptive, as it assumes that x perhaps does not act according to p.

    Two replies to that:

    1. Suppose for a moment, for the sake of argument, that “x knows to p” means “x will p because she knows that she should,” and “x knew to p” means “x ped because she knew that she should,” etc. On that supposition, “know-to” sentences would imply prescriptions but would not assume that x perhaps does not p. Right? So that kind of combination must be possible in principle, yes?

    2. From several of your choices of expression here it seems clear that you’re reading “p” in “x knows to p” as a prescriptive sentence (e.g. “Eat!” or “Push the other eggs out of the nest!”). I’ve always been hearing it rather simply as a verb (e.g. “eat” or “push the other eggs out of the nest”).

    • Ah, I see what you think I mean. This is a problem of attempting to phrase something I am understanding formally in a natural language. Maybe these will help out.

      First, when I mean that “perhaps not p,” I’m strictly speaking modally, so the actual situation can’t tell us whether she could not do p, or that she doesn’t necessarily do it. That’s a matter of strict logical entailment and tautology. I’m not on the Lewisian-Kripkean boat about modality, and I would only accept something tautological as a necessity, and since we don’t assume in ethics or philosophy that everything is a logical predetermination from prior states of affairs, we can’t legitimately claim that a mere happening of an event implies the necessity of that happening.

      In the second, the matter is probably seen more clearly in this way.

      To know is a two-place predicate. K(x,y).
      We’re pretty clear that x is some sort of agent that can know things.
      However, in y, we have a few options (O).

      O1: We can insert simply noun phrases, which we treat as terms and as arguments. K(x,y) remains a WFF of first-order logic.

      O2: We transform a sentence φ into a noun phrase p, via a rule that transforms propositions into terms for use as arguments. K(x,y), y = p, φ –> p |- K(x,p). This must allow a higher-order logic.

      I go with O2, and then from there, can ask, “What is p?” Well, p is just a name for φ, which is some sort of sentence. My wager is that it’s something like, “z is [obliged / commanded / ordered …] to act / do something / …” which is another C(z,y), y = q, ψ –> q |- C(z,q). ψ is really immaterial here, but for notational ease, I’ll just assume that it’s a first-order “z does w.” (X(z,w)).

      However, no matter the ellipsis of φ (or p), we can only ask about what that proposition means. p must name a well-formed sentence.

      My exact challenge is your reply (2), that p (as I’ve described it here) is not moral at all. I think that, among the possible elliptical pieces in line with x2, one of them is [… / compelled / …].

      If we unpack it, we have something like K(x,C(x,X(x,w))). The knowing (K(…)) isn’t prescriptive or normative. The being commanded / compelled / coerced (C(…)) isn’t prescriptive or normative. The doing something (X(…)) isn’t prescriptive or normative. x is just an agent. Where, then, is the ethical claim or “moral knowledge” hiding?

    • I should be clearer. The normative and prescriptive statement isn’t really there until you stipulate that it’s there. You can do that by arguing for its elliptical placement on φ, but that still won’t make the predicate K(…) normative, moral, prescriptive, etc.

  11. Maybe the reason I think “know to φ” implies “actually φ” is that this fits the common usage of some other, much more common phrases: “remember to φ”, “say to φ”, “tell to φ”.

    —Did you remember to send a note?

    That reply implies that the person did send the note.

    (One might object that remembering to send a note doesn’t imply sending the note, because the following reply is also intelligible: “Yes, but I didn’t send it.” – – – But I think that that objection’s evidence isn’t conclusive. Here’s why. Suppose remembering to send it does imply sending it. Still, answering simply “No” to the above question would imply that you forgot to send it, while what you really did was to remember about sending it but decide that you shouldn’t, and so not send it. Answering “No” would then be misleading (and “I remembered that I should send it, but decided not to” would be false). So one answers with a very minor paradox: “Yes, but I didn’t send it.”)

    Interestingly, forgetting to send a note seems to imply not sending it.

    • I’ll interest you with a feasible case of the opposite. This one will sound a bit more like Gettier. (I think the second one is better, personally.)

      Case 1:

      You have a stack of envelopes. One of them is the note. You don’t know that the note is there, but pick it up with the stack of mail and drop them in the mail slot.

      Later, your friend asks you, “Did you remember to send the note?”
      You think to yourself, “No, I don’t think I did,” remembering a place where you left it before you put it with the stacks. You tell him, “No, I’ve forgotten,” thank him for the reminder and check where you last remember leaving it. It’s not there, and, as far as you can remember, it’s now “lost” to you.

      The recipient of the note calls you a week later to thank you for the note. You, up to that point, were still thinking that you had lost it and had not remembered to send it.

      Case 2:

      You suffer from anterograde amnesia, and thus you remember nothing after a certain period. As you’re walking to the mailbox with the note in your hand, that period passes. You’re by the mailbox, but don’t remember anything about your life after a date well before there was any scenario that involved said note. You ask yourself, “What am I doing out here with this note, envelope, etc. by the mailbox?” You know how postage works, so you surmise that you should send it. You send it and that period passes. Later someone asks you, “Did you remember to send the note?” You explain your condition and go on with your day.

      Later, someone thanks you for the note that you sent them. You still have no idea that you sent it.

    • When I said that the reply “Yes, I remembered to send it” implied that one had sent it, I didn’t mean the obviously absurd claim that

      (A) the fact that one has made the statement entails that one has sent the note.

      Rather I meant that

      (B) the statement itself implied that one had sent the note.

      Compare: the statement “I have eaten a pomegranate seed” implies that one has eaten a pomegranate seed, but the fact that one has made the statement does not imply that one has eaten a pomegranate seed. And again: the fact that I sincerely say that I know that the capital of Morocco is Fez does not imply that the capital of Morocco is Fez, while the statement that I know that Morocco is the capital of Fez does imply that the capital of Morocco is Fez. That is, my knowing that Morocco is the capital of Fez implies that Morocco is the capital of Fez.

      My guess is that your Case 1 is meant to challenge a claim like (A), but about forgetting. I don’t have any idea what your Case 2 is supposed to challenge.

    • The cases offered counterexamples to the following: “Interestingly, forgetting to send a note seems to imply not sending it.” They came more easily to me.

      There are two cases where it’s possible that a person forgets to send a note, but sends a note. The statement itself would be true, and yet the action in question was not done (that is, the note was not unsent).

      I’ll see if I can dream up some counterexamples for the claim: “Remembering to do something implies that one has done something.”

      However, my first ideas (mad scientists, etc.) corrupt the subjective feeling of remembering something over a more legitimate/actual remembering something, where one actually remembers implies that one actually does as he remembers.

    • Oh, I see — sorry!! Then:

      Case 1

      In case 1, it seems to me that the person sent the note but not as an intentional act, so I’m not sure that counts.

      Case 2

      I’m not sure I understand the narrative in Case 2. Here are my two best guesses (the latest of many), though neither fits your words:

      2a. Maybe by “anteretrograde” you mean instead “retrograde,” and maybe by “remember nothing after a certain period” you mean “remember nothing after a certain period (from that period).” Thus the scenario you have in mind is that when the person starts for the mailbox she has no kind of amnesia, but on her way there she acquires an amnesia that blocks memories from, say, at least the past month.

      2b. Maybe you do indeed mean “anteretrograde” but by “period” you don’t mean “period”; rather you mean “length of time” (2011 is a certain period; one year is a certain length of time), and by “remember nothing after a certain period” you mean instead “remember nothing (from after the date of acquisition of the amnesia) for longer than a certain interval,” some length of time long enough to allow coherent activity over time such as appropriate note-making and packing, but not long enough to span the walk to the mailbox. I think that’s what you mean.

      As the close of your narrative more or less acknowledges, the question “Did the person forget to send the note? Yes or no?” is somewhat inappropriate to the case. (Compare: “When did you stop beating your wife?”) Saying that she forgot to send the note would be misleading. So I don’t think this is a good counterexample to show that forgetting to send a note doesn’t imply not sending the note.

      Aside from the fact that the person has a condition normally not contemplated, there’s also a problem about time.

      Suppose I want to shoot down the claim that “S knows that P” implies “S believes that P,” and I adduce a case where S knew that P but then forgot, and so stopped believing that P. The counterexample would be bad, because the original claim only meant that knowledge implies belief at the time of the knowledge. As we might say, there’s a time constraint on counterexamples.

      There also are time constraints on counterexamples to the claim that (A) remembering to φ implies φing, the claim that (B) forgetting to φ implies not φing, etc. For example, if φing is something that can be done today or tomorrow, and I forget to do it today but remember tomorrow, and do it only tomorrow, that doesn’t look like a counterexample to (B). Your Case 2 looks like it might be of this kind. But I’m not sure what exactly the time constraints are on counterexamples to those two claims.

      Incidentally, I am not and was not sure whether (B) would tend to support or oppose the idea that the workings of the other words (“remember”, “say”, “tell”) suits and supports the claim that ‘know to φ’ implies φing (or at least tending to). That’s why I didn’t list ‘forget’ at the beginning of my #12. So I’m not sure whether shooting down (B) would support or opposes my more central point.

    • There is a real condition called anterograde amnesia, in which people fail to be able to retain new memories, often due to brain trauma. People with retrograde and anterograde amnesia suffer global amnesia, like this one here: Think of the anterorade amnesiacs as those who have no means of transferring their data from running RAM active running to hard disk space for later retrieval, while the retrograde amnesiacs just suffer a data retrieval error.

      “Intentionally forgetting,” seems a bit odd, but my guess is that you’re indicating a sense that better matches what film mafiosos invoke when they say, “Forget about it!” However, it means actually remembering some issue, and then not intervening on it. I didn’t intend to address that sense, but am assuming a sort of mutual exclusivity between remembering and forgetting, wherein “S forgets something if, and only if S does not remember that thing,” for any simultaneous duration/period.

      I should say two things about the tense issue:

      In the first, forgetting and remembering are tensed issues specifically because their domain of interest is some cognitive retention of past events. There is no remembering the future, just remembering that we expect that there will be an anticipated future in which we will be able to act. People are said to remember something only if their recall of that thing is an accurate reproduction or retelling of that thing. People who fail at this have forgotten that thing.

      Because of this, we need to consider exactly what our agent remembers or forgets, that is, what φ is, regardless of the time at which he remembers or forgets φ. If φ is something like, “S sends this note at some duration,” then it’s clear that S forgets to do so in both cases, and yet has accidentally done so.

      Second, neither cases suffer the error that, say, avoiding tense issues in, “S knows something,” implies, “S believes something” could produce with your said bad counterexample, because your claim is actually a universal one. It would read like this: “For all durations (or periods of time) d, if S knows something in that duration d, then S believes that thing in duration d.” My cases are actually counters to a claim of the same form, because no matter the duration d, both examples show that, “For all durations d, if S forgets something in that duration d, then S doesn’t do that thing in duration d.” That’s false because people with anterograde amnesia forget in mid-action, and yet manage to complete them. If you’ll look at that video, Clive Wearing doesn’t even remember that he’s ever dreamt, or eaten, or survived his childhood. Of course, he’s done those things, but he simply does not remember. It’s also false because we often do things by accident or when we’re not paying attention, which we simply do not retain because, as it appears, memories very much require active association of events. Forgetting is actually easier.

      Whether a higher-order sentence implies any of its own arguments at any lower order, “P(…φ…) ⇒ φ,” with additional qualifiers for duration (to meet the tense issue), is very much a case-by-case issue. When it’s about forgetting, it’s pretty clear that the implication doesn’t follow. When it’s about remembering, right now I have no reason to deny it or to affirm it.

      The tense issue is easy to consider in these cases, since we’re looking for a duration throughout which P(…φ…) is true, in which φ is false. I have actually devised a tense logic that gives more rigorous sense to this (because it counts durations/time periods as the continuum between stipulated “beginnings and endings” on a separate domain of discourse called a “calendar”). I’ll hope that you’ll take this as an article of faith, because the actual formalization of this issue is a big fat pain on my fingers.

      However, I think sentences of these sorts might counter a claim that, “S knows to φ,” implies, “φ.”

      “Afterwards, I knew the rules. I understood what I was supposed to do, but I didn’t. I couldn’t. I was compelled to stay, compelled to disobey.” —
      “James knew to stand his ground, but he fled from the bear anyway.” —

  12. There is a real condition called anterograde amnesia

    Yes, I described it above.

    “Intentionally forgetting,” seems a bit odd, but my guess is that you’re indicating a sense that better matches what film mafiosos invoke when they say, “Forget about it!”

    I didn’t mention or suggest anything like intentional forgetting. What I wrote was, “In case 1, it seems to me that the person sent the note but not as an intentional act…”

    I have lots of other complaints like that, about your most recent comment.

    But I also think you say some important and insightful things, calling for an extended substantive reply. I’m not ready with a proper reply yet.

    • I wanted to be sure we were on the same understanding of the condition. The term “anteretrograde” threw me off. Only anterograde amnesia is necessary for the example.

      Case 2 is a bit clearer with a picture of a man who simply loses track of his memories after each period, so we could describe them in succession.

      Duration 1: S is going to a mailbox with a note and intends to send it.
      Duration 2: S forgets Duration 1. It is lost to him forever.
      Duration 3: S pieces together that he is sending a note, or that he might as well send the note, since he’s got it and it’s ready to go. He then sends the note.
      Duration 4: S forgets Duration 3, and it, too, is lost to him forever.

      S has forgotten everything about sending the note, but he has successfully done so. S doesn’t recall some previous events, spoken promises, personal reflections, etc. He just pieces together a relevant response based on some environmental cues. As far as I know, most anterograde amnesiacs get by on their present actions by bare-bones inductive reasoning, not sustained recollection of accounts from a continuous personal narrative.

      I think I may be missing something fundamental to your reply, or just am not reading thoroughly. Of course, if someone forgot to do something, but then did it independently of his remembering to do so, that we would not call his action intentional, as there wasn’t some consistent planning or something of that sort, but that doesn’t seem to disqualify a statement which says that someone’s forgetting to do something can still enable people to do such things despite not remembering to do so. Now, I’m not sure if, “S forgets to do something,” implies, “S does not do something intentionally (that is, opts against the option to do it).” I have my doubts, but is this closer to what you were describing?

  13. This somewhat disjointed comment may be of interest to those who are not following my conversation with Joshua. It benefits from much that Joshua said above, but I don’t try to spell out how.


    Joshua, I agree that knowing to φ doesn’t strictly imply φing. (See the string under comment #6.) I apologize for writing misleadingly about that.

    Similarly, Yong Huang’s claim is not that in Confucianism, moral knowledge (or some kind of moral knowledge) strictly implies action.


    I … am assuming a sort of mutual exclusivity between remembering and forgetting, wherein “S forgets something if, and only if S does not remember that thing,” for any simultaneous duration/period.

    Joshua, I think you might just mean “only if,” in which case I pretty much agree. But I strongly disagree with the “if and only if” view even about remembering/forgetting that Paris is the capital of France, for two reasons.

    (I) one might never have known that Paris is the capital of France, and then one would neither remember nor forget it. Interestingly, when we use the words “remember” and “forget,” we often use them to make points (or ask questions) framed by the assumption that the matter at issue was originally known. If that assumption turns out to be false, then the view we intend to convey by “Smith remembed …” is not accurate, and neither is the view we intend to convey by “Smith forgot …”. Neither term is applicable. But where the assumption is true, the two states are indeed mutually exhaustive—

    except for a further and much finer point: (II) our use of the two words is usually also framed by a much more common and (normally) unstated and unnoticed working background assumption: the assumption that the case at hand does not involve rare pathologies of the memory. Where that assumption turns out to be false, it can turn out that the sort of view we would normally intend to convey by “Smith remembed …,” and the sort of view we would normally intend to convey by “Smith forgot …,” both fail to be accurate to the case at hand. (Joshua, I think your Case 2 presents such a case, where it would be misleading to answer the question “Did she remember to send the note?” with a Yes or a No.)

    I think the fact that our remarks are commonly framed by background assumptions, and that certain kinds of assumption tend to go with certain words and phrases, is very important to keep in mind toward avoiding mistakes in the analysis of those words and phrases. If assumption A is standard for term T, then no mere paraphrase of T can be an adequate analysis of the term, and different analyses of T can each in be correct in a way: “T means P”, “T means A and P”, “T tends to mean P”, etc.

    It might be helpful to keep in mind that these are different:

    (A) “S remembers to φ” tends to imply “S φs”,
    i.e. it is normally used to communicate that S φs.

    (B) “S remembers to φ” implies “S tends to φ”;
    in the sense that her internal state includes a motivation or inclination to φ, though one that might be overridden by another motivation or inclination.

    (C) “S remembers to φ” implies “S is motivated or inclined to φ overall,”
    though of course one can be blocked from acting as one wishes.

    (D) “S remembers to φ” tends to imply “S is motivated or inclined to φ.”


    We often think of remembering or forgetting as something that is done especially at the beginning of the period of remembering or forgetting. “That’s when I remembered”; “I forgot a long time ago.” Hence it seems that “remember” and “forget” can each sometime indicate an event that happens at a particular time, and sometimes indicate instead a continuings state: “I forget where I left my keys.” (“Know” doesn’t have this dimension of ambiguity.)

    But when we speak of remembering or forgetting to φ, we are also focused on the time of φing or at least the time of the occasion to φ, which (depending on the φ) might be some particular moment.

    Can Smith “remember in January to take a walk in April”? Insofar as one can speak that way, surely Smith’s remembering in January does not imply that she will take the walk in April. Smith could forget the whole business in March.

    (This last point might be worrisome at first glance: for if we are interested in moral knowing-to, we are likely to be especially interested in the idea that this can be a state of mind whose possession at one time tends to imply action at a later time. But compare: even the value of knowing-that generally depends on our retaining the knowledge, which retention the fact of the knowledge doesn’t inherently guarantee.)


    I think that in the main common use of the past tense “S remembered to φ” and “S forgot to φ,” the primary point of the remark is to report that S did or did not φ, respectively. Sometimes a further point about S’s mind and/or character is implied. (The usual main purpose in other tenses may also be to say something about whether Smith φs.)

    That point suggests but of course does not require the following theses:

    (1a) “S remembers to φ” means “S remembers that she ought to φ, and S φs.”
    (1b) “S forgets to φ” means “S forgets that she ought to φ, and S does not φ.”

    Those theses are consistent with the claim that our beliefs about what to do (our beliefs that we ought to do this or that) have no tendency whatsoever to influence our action.

    Different from 1a:

    (2a) “S remembers to φ” means “S remembers that she ought to φ, by a special sort of Moral Remembering that inherently tends to imply that she is motivated to φ.”

    Perhaps Yong Huang’s thought is that Confucianism involves an assumption like (2a) about moral knowledge.

    But I think our interest in “knowing-to” is based on the thought that knowing-to might be a state that inherently either motivates S to φ or tends in some other way to imply that she φs, without requiring that she know that she ought to φ.

    And I think the fact that the primary purpose of the past-tense sentences is to report that Smith did or didn’t φ can suggest that “remembering that” and “forgetting that” are less essential to what it is for Smith to remember or forget to φ.


    Toward the end of my long Comment #3 on this thread I mention that in a recently published paper on early Confucianism I sketch a picture of how moral knowledge can be inherently motivating. I mean, the mechanisms. Maximally simple summary view: to appreciate the importance of something you might do is to feel the pain or pleasure at stake in it (not just for you), and that’s motivating. That is, the hedonic states internal to one’s inclinations are representational, as the brown color of a painting or mental image of a forthcoming glass of iced tea represents the tea itself as being brown.

    One might object that the picture only shows that one vehicle of moral knowledge, one mode of representation, may be inherently motivating.

    My reply would be that I suspect it’s a natural fact about us that we can’t have much moral knowledge except by heavy reliance on that sort of vehicle. By far the most effective individual and collective procedures we have for expanding and sophisticating our moral knowledge rely heavily on that sort of vehicle.

    And that sort of vehicle centrally involves nonverbal signs. Putting it into verbal propositions is to some extent a matter of translation, and there’s no guarantee that it’s possible. (Think of translating a treatise into some nonverbal medium.) So it may be helpful, in thinking about moral knowledge, to look past knowing-that.

    • If that claim (1) on Yong Huang’s description of moral knowledge is the case, I don’t know what is going to confirm that Smith actually φ’d because she knew to φ, or that she doesn’t φ despite knowing to φ. In this sense, moral knowledge doesn’t sound very motivational, but sounds like a tertiary consideration. Even ceterus paribus, it doesn’t seem like people would really act on the basis of this moral knowledge, but give it a sort of passing consideration before they did whatever they wanted.

      If we want to treat moral knowledge as an instinctual matter (like the sentence about cuckoos does for non-specific knowledge), then, ceteris paribus, people should strictly act according to those instincts. Recently, Patricia Churchland’s primary target has been to plot neurological correlates to any of the instincts and neural patterns that govern human sociality, but that reduces “moral knowledge” of the sort discussed here to a network of Darwinian-selected survival mechanisms.

      Your claim (2) is legitimate. I should say “only if.”

      I think in (3) that you are right to mention that forget as a verb speaks to the period at which one has begun not remembering, while no single verb in English accounts for “retaining a lack of remembering,” though be forgetting of isn’t too bad. We use remember to speak to both initiations and durations, and we use know to speak of the duration, though do not explicitly state when that knowledge was first acquired.

      In your (5): “Maximally simple summary view: to appreciate the importance of something you might do is to feel the pain or pleasure at stake in it (not just for you), and that’s motivating.” This sounds like a simplified simulation theory, and I would agree that a failure to “put oneself in others’ shoes” would make motivation to serve more than one’s own immediate interests very difficult, but that this may very well show that it’s misguided to word “moral knowledge” as an epistemic matter. Instead, we need only describe the anatomical and social bases for human empathy, and not “look past,” but “look away from” the “knowing” part of this. We may very well know that using certain heuristics like: “Think globally, act locally,” or something of that sort makes the problem clearer, and thus at least possible to simulate, but whether we’re innately inclined to use these kinds of reasonings in our ethical affairs is not a question of “knowing to do something” because of some moral heuristic, but of “being compelled to do something” because of some moral heuristic. Those are very distinct questions that “moral knowledge” worryingly rams together.

    • Hi Joshua, thanks for the very interesting replies!

      On (1): I agree that Yong Huang’s brief proposal doesn’t propose answers to all the questions one might ask.

      I don’t know how you read “the knowledge doesn’t strictly imply the action” as implying or suggesting “the knowledge motivates the action not strongly, but only very weakly if at all.” Even very strong motivation to φ doesn’t strictly imply φing.

      Regarding instinct: I wonder whether you’re eliding the idea that we have moral knowledge by instinct with the claim that it is instinctive in us to act on whatever moral knowledge we have? I don’t think either of those ideas was on the table, though.

      On (2): The point ultimately at issue was whether Case 2 (on anteregrograde amnesia) is relevant to forgetting-to- φ. The point we agree on blocks that relevance, yes?

      On (3): Good point, there does seem to be a difference like that between ‘forget’ and ‘remember’. But I think it’s not so simple, since commonly these verbs refer to a time (a moment or a period) that is the occasion for using or doing that which one might remember or forget. For example, ‘I forget’ in the present tense doesn’t mean I am only just now losing track. If you ask me again tomorrow, I may still forget. Similarly: Why is she scurrying about? –She forgets where she put her keys. And if my boss says of me simply that I forgot to send the memo, she isn’t speaking of a past event in such a way as to leave open that I later remembered.

      I agree that ‘be forgetting’ sometimes works; so I think does ‘has forgotten’.

      On (5): I don’t see what your argument is, unless you simply don’t like to call something “knowledge” if you think there’s a naturalistic account of it.

    • On (3): The examples give a good case that I am wrong about forget. We do actually use it to imply a persistent state of not remembering. However, know still does not seem to ever indicate when one begins to know, but I think we have other verbs for that — realize, recognize, and maybe some relevant synonyms.

      On (2): I think we began to alter the question from, “Does Smith’s remembering/forgetting/knowing φ imply Smith’s doing/not doing/doing φ?” to, “Does Smith’s remembering/forgetting/knowing φ imply Smith’s motivation to do/lack of motivation to do/motivation to do φ?” My examples only challenge that former claim. With forgetting, it seems trivially true. With remembering and knowing, it seems demonstrably false.

      On (5): I would want to leave to knowledge what is strictly of knowledge. An acceptable necessary condition for motivation is knowledge, but I would disagree strongly to the idea that motivation is an important condition for knowledge. Knowledge is largely a function of retaining input. It would be a false attribution to claim that the retention, or that the information itself, motivated an action. What matters much more than that is our emotional reaction to that retained information. We also know well that emotional reaction to input is very relevant to its retention in the first place.

      Additionally, I think that it’s incorrect to call whatever information that motivates us “knowledge” on a principle that mere belief suffices to motivate people to action, and that the actions are amoral, moral, and immoral. Misinformation appears to be the culprit in a lot more decision-making than we admit of ourselves, and it motivates benign and heinous acts, too.

      Another issue is that if this moral knowledge is as innate as an instinct is, then it’s not really “knowledge” any more than it is collective bias. And there’s a marked difference between knowledge of facts and consensus on ethical mores.

      I think a final issue, which I raised only briefly with Manyul, is that much of this discourse just presupposes moral realism, that there are “moral facts” to be known. My metaethical stance doesn’t accommodate the moral realist, since I’m the kind of naturalist who argues that moral statements are to be broken down into arguments which aim to alter states of affairs as the moral locutor prefers to have them. I don’t know where I fit in the regular labeling, but it doesn’t count as moral realism because subjectivity is crucial to the dispute that motivates users to use moral rhetoric, since the argument structure I propose incorporates subjective assessments and impersonal facts (that we’re only ethical naturalists insofar as we’re perfectly in touch with all of the relevant facts of the argument that moral discourse abbreviates).

      On (1):

      I don’t know how you read “the knowledge doesn’t strictly imply the action” as implying or suggesting “the knowledge motivates the action not strongly, but only very weakly if at all.” Even very strong motivation to φ doesn’t strictly imply φing.

      I usually run a descent of implications’ likelihoods: strict implication (100% correlation, like a deductive tautology), statistically relevant implication (99.99…% to 50.x% consistent correlation), and statistically irrelevant implication (no charted correlation beyond random chance). Since we have agreed that φing doesn’t strictly follow from knowing to φ, we have only the question of statistical relevance to keep the idea afloat. However, I think that the statistical relevance won’t hold for reasons that we’ve been discussing.

      Regarding instinct: I wonder whether you’re eliding the idea that we have moral knowledge by instinct with the claim that it is instinctive in us to act on whatever moral knowledge we have? I don’t think either of those ideas was on the table, though.

      My position is more along the lines that beliefs and knowledge of any kind are inert until one is motivated to use them. In my view, knowledge is not a sufficient condition for motivation, much less for actual action.

    • On (2): I stand corrected!

      On (3):

      However, know still does not seem to ever indicate when one begins to know

      I saw this, and that’s when I knew you were right. I mean, mostly.

      On (5):

      I would want to leave to knowledge what is strictly of knowledge

      Why? Surely something can be true of knowledge even if it is true of other things too (such as belief or money)? While it’s tempting to object to the claim that I weigh 200lb on the ground that it is only the corporeal thing that weighs 200 lb, not the corporeal animate sentient rational thing, still I think the objection would be incorrect.

      Or have I misunderstood?
      It would be a false attribution to claim that … the information itself, motivated an action … … Misinformation … motivates benign and heinous acts, too.
      I gather that what you mean is that information (belief) is only one leg of motivation, which requires desire as a second leg, so that while we can speak loosely of belief motivating an action, really that’s not quite right.

      I think a final issue, which I raised only briefly with Manyul, is that much of this discourse just presupposes moral realism, that there are “moral facts” to be known.

      Here and in the bit I quoted just before, you seem to be supposing that the proposal on the table with regard to “knowing to” is that knowing that it is right to φ can motivate us to φ. But I think the proposal on the table with regard to “knowing to” is that the phrase “know to φ” might capture a kind of moral knowledge that does not essentially involve knowing that it is right to φ (or some similar proposition). I think what tempts some people to find that proposal attractive is the precedent of “knowing how” – for it seems offhand (though it has been contested) that, say, knowing how to ride a bike does not amount to knowing (i.e. believing+) that some proposition of the form “the way to ride a bicycle is to …” (or even a long list of propositions) is true.

      But of course, even if Smith’s “knowing to φ” doesn’t involve her “knowing that” she ought to φ, still that wouldn’t settle the question whether Smith could “know to φ” if there were no truth “Smith ought to φ.” So you may be right that the issue about “knowing to” arises only within moral realism broadly construed.

    • On (2): I may not have been very clear. I meant the forgetting to do something implies a lack of motivation to do it (at least, on the basis of that initial remembering), while remembering or knowing to do something doesn’t seem to imply that we’re motivated toward it at all. I don’t give money to fund starving children in Africa, but not because I don’t know that they’re there.

      I mean in my claim that “I want to leave to knowledge what is strictly of knowledge,” that we not try to phrase issues about acquiring skills through practice and self-estimation (“knowing how”), or issues behind the collective motivational forces of moral acts (“knowing to” as it involves all of the implied motivations to action) all into epistemic phrasing. In our language and in actual empirical practice, those separate domains are better off rephrased and remodeled (or, since “moral knowledge” is the novelty, not rephrased or remodeled) to have no bearing on the straightforward “knowing-that propositional knowledge” that is focal to epistemology.

      I wouldn’t say, either, that we can rephrase all technical skills into acceptances of beliefs about how we discuss going about them. “Knowing that” should remain about propositions and surveys of acceptance of claims and foundations for acceptability of claims, and the terminology that surrounds it is most suited to that. I know that a way how to catch fish is to cast a baited line, wait for fish, reel in the fish, and so on, but that is strictly a survey about what facts are true about the steps toward catching fish.

      But “knowing how” doesn’t need to be concerned with the epistemological terminology. Sticking to fishing, “I know how to fish,” could be phrased something like: “Given my knowledge of the relevant components and processes of fishing or my experience with fishing, I have a confidence in my success at fishing.”

      I think that it’s easy to mix them up in English, but more formal phrasing would give us (for knowings that), “I know φ,” where φ is “how to fish” (that is, the steps involved in fishing), but could just as well be any proposition that can be true or false. However, that would be different than saying (for knowings how), “I know_how to φ,” is, “I am learned and adept at φ.”

      Certain “knowings that” are necessary for “knowings how,” and vice versa, but that doesn’t mean that they are analogous phenomena.

      My point in the “knowing to” issue has been to show that we could very easily reduce this stuff about “moral knowledge” into the sort of “knowing that” without a loss to the comprehension of the meaning to those sentences, without extra ontological baggage, without appeal to narrower metaethical views (since we could separate knowing from the proposition that is known, and thus have a question over whether one can actually know moral statements as propositions, and then as facts), and without jumbling different value theories into each other (like with 良知, that somehow the knowledge itself, or our possession of it, would be good). I don’t then understand the impetus for the venture, to treat this particular molehill of phrasings like it was its own special mountain, when we don’t have any need (and I see no reason to want) to treat them that way.

      I would say that belief is just a leg of perhaps many legs that prompt motivation. Desire is an important leg. Another leg is the power alter the world according to those beliefs and desires. There may be others, but I know that our motivation doesn’t come from just believing or knowing that something is the case, ought to be the case, etc. Consider my failure to donate to starving children. If it were really just a problem of information, I would probably donate after I saw just one commercial. But factors of motivation aren’t just of a sort that says, “I know to give money to starving children, and so I’m more motivated to do so.” I may not really care, relatively speaking, about starving children, or I may not be able to spare the expense, or I may be pessimistic about the whole undertaking, etc.

      If we want to discuss the things that trigger actions in agents, there’s already a wealth of folk language and scientific jargon that is not epistemological. But that discussion won’t be particularly moral, either, and if we cover the basis for motivation of action on “believing(+) that,” a few extra legs, and some discussion of instincts, I don’t know why we would want to invoke “moral knowledge” on top of it.

      But I could, in my own perversity, show that we could raise the same kind of murkiness with something like “knowing when.” We know when the planets will next align, but that’s different from knowing when to use corporal punishment on people, isn’t it? The former invokes no propriety of the affair, while the latter appears to ask our knowledge of the time, or of the sufficient and necessary states of affairs that justify an act. But the issue isn’t in the knowing, it’s in the sentences-turned-arguments-for-a-predicate (the φ’s), as I’ve ostended.

      I’m getting too long. I haven’t touched the metaethics chunk in any good detail.

    • Hi Joshua, thanks for the really helpful reply, and I’m sorry I badly misread you again!

      If I understand you know, by “I would want to leave to knowledge what is strictly of knowledge” you mean you want to narrow the reference of the word ‘know’ on the grounds of a view about what the most useful categories are. That is, you grant that English commonly uses ‘know how’ to refer to a condition that might not be reducible to knowings-that, but you think English is making a practical error here and should be reformed. (Or at least, you want it reformed and are making an argument aiming at producing that result.) Yes?

      One reply is that the phrases ‘know that’ gives us (at least, gives intellectuals who have the patience for it) the narrower category you want without sacrificing the wider one.

      Your answer might be that the wider one, including whatever kinds of “knowing” can’t be reducible to knowing that, isn’t a worthwhile category.

      I think that on the question what terms are worth having, natural language speaks with a great deal of authority. I don’t have a clear view of the benefits and drawbacks of the current actual range of ‘knowing’, to set against that authority.


      On realism: I wonder whether the questions about ‘knowing to’ can indeed arise outside of moral realism, on the grounds of some parallel rejection of realism about knowing. What counts as knowing would seem to involve norms of some kind—norms about justification, say—and one might not be a realist about those.

    • That’s a fair take, though I would say that ordinary, natural language is more or less the terminology of a folk theory, but that itself may not be too authoritative on the matter at large. My primary worry with that appeal is the map/territory errors, that we somehow begin to force the empirical facts into an inadequate semantic model, committing category mistakes, and then defending the legitimacy of the model despite those errors because we, ourselves, invented the categories. I think Wittgenstein rightly insists that we do not think, but that we look.

      One reply is that the phrases ‘know that’ gives us (at least, gives intellectuals who have the patience for it) the narrower category you want without sacrificing the wider one.

      Your answer might be that the wider one, including whatever kinds of “knowing” can’t be reducible to knowing that, isn’t a worthwhile category.

      As I’ve made sense of this, “knowing that” and “knowing how” are separate categories, the wider of which would be “mentally involved skills” or something of that sort. This makes it much more amenable to the empirical matter, and we don’t have to wangle one “knowing” into another “knowing.”

      It appears that “knowing to” can be so wangled.

      On moral realism: I would say that these higher-order sentences present “loaded propositions” (in the vein of “loaded questions”), wherein “I know not to drive under the influence,” just assumes that, “I am not to drive under the influence,” is just a fact to be known.

      Are we talking about epistemological realism vs. idealism? I think I’m with the epistemological idealists (that knowledge only exists with knowers, and is strictly a brain-sustained tool), but I don’t see how that informs whether I’m a moral realist or not.

    • Wittgenstein, PI 66:
      “ …Don’t say: “There must be something common, or they would not be called ‘games’ “–but look and see whether there is anything common to all. — For if you look at them you will not see something that is common to all, but similarities, relationships, and a whole series of them at that. To repeat: don’t think, but look!”

      (Carroll, Alice in Wonderland IX:
      “She had quite forgotten the Duchess by this time, and was a little startled when she heard her voice close to her ear. ‘You’re thinking about something, my dear, and that makes you forget to talk. I can’t tell you just now what the moral of that is, but I shall remember it in a bit.’
      “‘Perhaps it hasn’t one,’ Alice ventured to remark.
      “‘Tut, tut, child!’ said the Duchess. ‘Everything’s got a moral, if only you can find it.’ And she squeezed herself up closer to Alice’s side as she spoke.)”

      I would say that ordinary, natural language … may not be too authoritative on the matter at large. My primary worry with that appeal is the map/territory errors, that we somehow begin to force the empirical facts into an inadequate semantic model, committing category mistakes, and then defending the legitimacy of the model despite those errors because we, ourselves, invented the categories. I think Wittgenstein rightly insists that we do not think, but that we look.

      But your proposal was not that we couldn’t find something in common between knowing-that and knowing how, etc. (which anyway doesn’t look hard to do). So the Wittgenstein quote seems to me off point. Rather it seemed to me that your veiw was that English is strictly wrong to use ‘know’ for knowing how. That’s what I wanted to challenge. It looks to me like the sort of thing Wittgenstein was complaining about. He wasn’t, I think, claiming that we should reform the word ‘game’. If you don’t want to make that claim, then I’m happy to withdraw the rest of this comment.

      Was Wittgenstein complaining about thinking or about prejudice? If he was complaining about thinking, I disagree. I would like to know what your actual thoughts are about the disadvantages of having a term ‘know’ that covers ‘knowing that’, ‘knowing how’, and ‘knowing to’.

      Your argument here seems to involve two points: (a) we should be suspicious of category schemes because they might involve errors, and (b) we should be especially suspicious of a scheme insofar as we invented it, because that fact can tempt us to bias in its favor. Or, just possibly, you may also mean (c) we should be especially suspicious of a scheme insofar as we have invented it, because in general, what has evolved naturally is less likely to involve error than what a few people have rigged up artificially.

      (a) is neutral on the point at issue, so long as no particular errors are mentioned. As between (b) and (c), (c) is the one that looks to me as though it might have some weight. It would seem to weigh in favor of natural language. Which side does (b) take? I suppose that depends on which category scheme you and/or I have a greater sense of having invented: the natural language scheme or the one you want to replace it with. I suppose I don’t have a parental feeling with regard to either of them.

      As I’ve made sense of this, “knowing that” and “knowing how” are separate categories, the wider of which would be “mentally involved skills” or something of that sort.

      By As I’ve made sense of this I don’t know whether you mean “My interpretation of what you mean here, Bill, is that” or “My take on the taxonomy of ‘know’ has been that”.

      By “the wider of which” I think you mean “the genus of both.” Me too — the wider category I was referring to is the whole range of things called ‘know’ in English. But I wasn’t looking to give an account of what they have in common. I was defending the legitimacy of using the word ‘know’ for them, or at least asking what disadvantages of doing so are so great that we should say the language is in error about what knowledge is.

      You immediately continue: This makes it much more amenable to the empirical matter, and we don’t have to wangle one “knowing” into another “knowing.”

      I’m sorry, I really don’t understand you here: What makes what more amenable to empirical inquiry?


      I wasn’t talking about epistemological realism v. idealism. I think your first answer hit the mark.

  14. Sorry for the late reply. I am talking about my take on the taxonomy of these terms.

    I’ve read Wittgenstein a bit differently on this point. I always read him as demarcating how we use language, something perhaps between using language to describe / look at the world over using language to dissect / think about the world. It takes a bit of wrangling to try to fit, say, “knowing that” into “knowing how” or the other way around, and so, just as we wouldn’t just assume that we should characterize things that we call “games” in the same way (since the surrounding terms that adequately describe competitive games don’t describe non-competitive ones, for instance), we don’t want to do the same thing with “knowing.”

    With knowledge, I think that we should avoid the claim that there are going to be common features to all of the things for which we use the word “know” because we happen to use it that way in a natural language. The major epistemological terms and demarcations (the epistemological categories that have developed from Plato onwards) are geared toward “knowing that” more specifically, but appear to cause problems when we’re describing these other areas of “knowing.” Also, the empirical data on these topics is markedly distinct, involving different brain regions and different processes toward acquisition. Fewer commonalities remain that the epistemological terminology can cover for both types of “knowing.” This is why I would propose dropping it for “knowing how” and “knowing to” unless something straightforward could cover most or all of the instances.

    Since I’ve treated “knowing to” to mean “knowing that one is to,” I could see that main predicate belonging to epistemological concerns. I can’t see how the argument to it (the “…one is to…” part) would fit epistemological categories. But if we detached the label, even temporarily, and worked on it without thinking of it as “knowing,” we could have a description of “knowing how” that is rather distinct from the “knowing that” which concerns epistemology more generally. Again, maybe not, but the science right now doesn’t lean that way.

    I’ve treated Wittgenstein as challenging a common assumption: “They’ve got similar names and a handful of similar features, so the same category scheme must be adequate to describe the so-named things fully.” In a brief reflection on the history of science and of language, that assumption should be challenged.

    But I’m not for reverse-engineering the language. It’s not the label “knowing how,” for instance, that really bothers me (since it could be called anything we agreed to call it). It’s the use of a whole surrounding theoretical framework that may work well for one phenomenon, but be totally inadequate for another, similarly named phenomenon. Phrased in your way, I think my point is more to say this: We should be suspicious of the idea that one category scheme is suited to describe a homonymous thing.

    But I also make claim (b). The issue with the appeal to natural language is that we’re appealing to things that people made up, maybe assisting in clarification, or maybe masking a defective understanding with an inadequate umbrella term. Wittgenstein’s point doesn’t particularly favor the natural language, or at least as I’ve read it, he doesn’t. It criticizes a perhaps natural compulsion to make our world deceptively neat and tidy via some categorization, regardless of origin.

    But that’s overlong. I’ve been looking for a simple verb to cover this idea. I’ve opted for wangle. But I think that the best idiom we have for this is “square peg, round hole,” where the round hole is the conceptual scheme and the square peg is the empirical information that we have.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.