Skip to main content.
December 15th, 2006

Category Mistakes

Here is the text of the Wikipedia entry on category mistake.

A category mistake, or category error is a semantic or ontological error by which a property is ascribed to a thing that could not possibly have that property. For example, the statement “the business of the book sleeps eternally” is syntactically correct, but it is meaningless or nonsense or, at the very most, metaphorical, because it incorrectly ascribes the property, sleeps eternally, to business, and incorrectly ascribes the property, business, to the token, the book.

The term “category mistake” was introduced by Gilbert Ryle in his book The Concept of Mind to remove what he argued to be a confusion over the nature of mind born from Cartesian metaphysics. It was alleged to be a mistake to treat the mind as an object made of an immaterial substance because predications of substance are not meaningful for a collection of dispositions and capacities.

This seems to be quite misleading to me in several ways.

As Ishani just pointed out to me, that a is necessarily not F doesn’t mean that saying a is F is a category mistake. If I said that 3 was half of 5, I’d be ascribing to 5 a property it necessarily doesn’t have, but the mistake is arithmetic, not mathematical.

And it isn’t clear that predication is at all central to category mistakes. Ryle’s introduction of the concept largely involves people asking mistaken questions, not making mistaken assertions.

The last sentence is also odd. It is a little tricky to say just what Ryle’s argument is here against Descartes. On the one hand, Ryle does end up concluding that to have a mind just is to have the right kinds of dispositions. And he does say that we shouldn’t identify the having of those dispositions with any kind of substance, either physical or non-physical. But is that how he argues against Descartes? I would have thought that the identification of the mind with the dispositional properties was after the conclusion that Descartes had made a category mistake, so this wasn’t Ryle’s argument that Descartes had indeed made a category mistake. But what then is the argument against Descartes here?

Ryle at times in chapter one of Concept of Mind says that he is outlining his argument, not making it, so maybe I’m going wrong in reading the order of Ryle’s conclusions. Perhaps, that is, he really does conclude first that mental talk is talk about dispositions, and from that infer that Descartes made a subtle category mistake.

I think this is all rather interesting, because even those philosophers who are neither Rylean nor Cartesian (i.e. most philosophers) should still be interested in Ryle’s claim about logical grammar. Is Ryle right that Descartes makes a category mistake in treating the mind as a substance? Jack Smart thought he was not – if the mind is the brain then the mind is a substance. So the question cuts across physicalism/dualism lines to some extent. Perhaps some property dualists could agree that Ryle is right to accuse Descartes of a category error, while insisting that mental properties are distinctive. (Actually, Ryle’s relation to property dualism is complicated I think, because of his avowed anti-reductionism. But that’s for another post or two.)

Anyway, it would be good if all this stuff about category mistakes could be sorted out, then we could fix the Wikipedia entry!

Posted by Brian Weatherson in Uncategorized


This entry was posted on Friday, December 15th, 2006 at 7:05 pm and is filed under Uncategorized. You can follow any responses to this entry through the comments RSS 2.0 feed. You can skip to the end and leave a response. Pinging is currently not allowed.

11 Responses to “Category Mistakes”

  1. Matt Weiner says:

    If I said that 3 was half of 5, I\‘d be ascribing to 5 a property it necessarily doesn\‘t have, but the mistake is arithmetic, not mathematical. [Should the last word be \‘categorial\’?]

    Maybe this case can be taken care of by saying: A category mistake is an ascription of a property to something when nothing of that kind can possibly have that property. 3 can\‘t possibly be half of 5, but something of the same kind as 3 — another number — can possibly be half of 5.

    That case needs another refinement, though, to handle something like this: \“Five is a positive real solution to 2x + 1 = 0.\” Nothing of the same kind as five, or of any kind whatsoever, can possibly be a positive real solution to 2x + 1 = 0. In order to handle this we\‘d either need a notion of possibility that\‘s weaker than metaphysical possibility — which might not be coherent — or we\‘d need to say something like \“A category make is an ascription of a property to a thing that, in virtue of the kind of thing it is, cannot possibly have that property.\” Of course \“in virtue of\” and \“kind\” need more explanation, but maybe that further explanation belongs to the question of which things are category mistakes rather than to the definition of category mistakes.

    This hasn\‘t addressed the point about questions at all — maybe category mistakes include questions whose answers must be category mistakes as just described above? So if the question is \“What kind of substance is the mind?\” it would be a category error because, in virtue of the kind of thing the mind is, it can\‘t possibly be any kind of substance. (I don\‘t know anything about the specific Rylean point.)

  2. praymont says:

    Here are what I take to be Ryle\\‘s main claims about category mistakes, along with some observations (quoting from The Concept of Mind):

    1. A category mistake \\“represents … facts … as if they belonged to one logical type or category (or range of types or categories), when they actually belong to another\\” (p. 16; ellipses blot out references to mentality in order to abstract away from that particular instance of category mistake);

    2. Questions are handy, but inessential, for indicating category mistakes. After introducing the \\“Where is the University?\\” example, Ryle gets to its point by saying, \\“His mistake lay in his … assumption that it was correct to speak of Christ Church, the Bodleian Library, the Ashmolean Museum and the University, to speak, that is, as if \\‘the University\\’ stood for an extra member of the class of which these other units are members\\” (p. 16). So the error rests in assumptions which can be indicated via questions, but which can just as easily be specified without them.

    3. The above category mistake involving the University was in the assumption that it was correct to conjoin certain terms. This is made explicit by Ryle later when he says, \\“When two terms belong to the same category, it is proper to construct conjunctive propositions embodying them.\\” (p. 22) This seems to be his main rule for individuating categories. He adds (ibid.) that it’s proper to conjoin or disjoin items belonging to the same category.

    4. It’s not clear what exactly one is abusing when one makes a category mistake. In places, Ryle says that an error of this sort rests on an abuse of \\“items in the English vocabulary\\” (p. 17), but for the most part he speaks of a category mistake as a misconception of the \\“logic of … concepts\\” (p. 16) or as allocating \\“concepts to logical types to which they do not belong\\” (p. 17). However, his examples mostly seem to involve ontological mistakes. So the University turns out to be a mode of organization – it is just \\“the way in which all that he has already seen is organized\\” (p. 16); team-spirit also seems to be a mode (\\“the keenness with which\\” tasks are performed (p. 17)); and the pair of gloves seems to be a whole of which the left-hand and right-hand gloves are parts (p. 22).

    5. Ryle seems to indicate that the error he has in mind involves assuming that \\‘exists\\’ is univocal. He speaks of \\“two different senses of \\‘exist\\’\\” (p. 23) and denies that numbers, navies, Wednesdays, minds and bodies may all be said to exist. This seems inessential to his main points. Was Ryle one of the culprits Quine had in mind when he said that while very different types of thing might exist, \\‘exists\\’ is univocal?

    6. Ryle sees several category mistakes in the official, or Cartesian, view; he takes it to rest on \\“a batch of category-mistakes\\” (p. 23). I don’t think the main Cartesian category mistake is one of wrongly taking the mind to be a substance. Sometimes Ryle points to this as the main Cartesian error, but many of his remarks seem instead to identify the central error as the assumption that the mind is a causal system (or machine) alongside the body. His key target, then, is the assumption that \\“the mind must be another field of causes and effects\\” (p. 18); or that \\“mental processes … are different sorts of causes and effects from bodily movements\\” (p. 19); or, again, that we should regard \\“minds as extra centres of causal processes\\” (p. 19); or, finally, that \\“minds are to be described in terms drawn from the categories of mechanics\\” (p. 20). If we take this version of the Cartesian category mistake as the crucial one, then a big part of Ryle\‘s argument against the Cartesian view is given in the \\“Emotion\\” chapter at pp. 88-93, where he argues that explanations by appeals to \\‘motives\\’, or reasons, are not causal explanations.

  3. praymont says:

    Well, I guess that\‘s what happens when you cut and paste from Word. It looks like quote marks and apostrophes went over as \‘a??\’ with a cap over the \‘a\’, and ellipses became \‘a?\’ followed by a pipe.

    [I\‘ve fixed these now – it seems both Word and WordPress try to do smarttype, and they get in each others way. Brian]

  4. clayton says:


    I have a worry about your proposal. While I’m convinced that the actual origins of a thing are necessary, I’m not convinced that the assertion that things could have had origins that differed from their actual origins is a category mistake although things that come from things, being what they are, couldn’t have come from different things.

  5. clayton says:

    And by ‘Marc’ I obviously meant Matt Weiner.

  6. Matt Weiner says:

    Hi Clayton,
    Let me try to get clearer on your worry. I don’t think the proposal has any trouble with “Michelangelo’s David might have been carved from this tree”; the relevant kind for David is statues, and statues can be carved from trees. So that doesn’t come out as a category error. Your worry is “David might have been made from something different than it actually was”; “being made from something different than it actually was” is a property that David couldn’t possibly have had, in virtue of the kind of thing (thing made from something else) that it is?

    I’m tempted to say, “Well the property involved here is ‘being made from something different from this block of marble,” since ‘what it actually was’ rigidly designates the block of marble. But that seems like at best a cheap kludge.

    [Also, it seems to me that Ryle’s definition as quoted in praymont’s point 1 is probably best. “Where is the University?” doesn’t seem to count as a category mistake on my definition, since the University can and does have a location. Maybe the idea is that the question suggests that it has a location different from its (actual) constituent parts, though I’m not sure that this would be compatible with my kludge; so much the worse for the kludge, probably.]

  7. Ofra says:

    A few quick comments:

    1.The Wikipedia entry cited is indeed hopelessly confused. To mention one other problem, note the claim that the sentence in question is a category mistake ‘because it incorrectly ascribes the property, sleeps eternally, to business’ – presumably, the sentence ‘Bill Gates is poor’ incorrectly ascribes the property of being poor to BG, but that in no way shows that the sentence is a category mistake. I should also note that the claim that category mistakes are ‘syntactically correct but nonetheless meaningless’ is very popular but far from uncontroversial. In fact, both conjuncts of this claim have been contested in the literature (I personally defend the former but reject the latter).
    2. Matt, here’s one problem with your alternative definition. You say ‘A category mistake is an ascription of a property to a thing that, in virtue of the kind of thing it is, cannot possibly have that property’. One could argue that the number two belongs to the kind ‘even numbers’, and that when we ascribe the property of being odd to the number two, we are ascribing to it a property that it cannot possibly have, in virtue of belonging to the kind ‘even numbers’. Yet the sentence ‘Two is odd’ is not a category mistake. (I guess you can respond by denying the claim that ‘even number’ counts as a kind, but that opens a whole other discussion…).
    3. Regarding the discussion of Ryle. I am mostly in agreement with praymont above. Ryle’s discussion of category mistakes in the concept of mind is very confused and hard to understand. (He has a much clearer discussion of the topic in his 1938 paper ‘Categories’, though he does not explicitly use the term ‘category mistake’ there). But as far as I can tell, Ryle’s complaint in the concept of mind is not that ‘The mind is a substance’ is a category mistake. Rather, he seems troubled by the following two concerns. The first is that application of mechanistic-like predicates to mental entities is a kind of category mistake (so, for example, talking of mental causation is a category mistake). The second is a kind of ‘double counting problem’. For example, on page 22, Ryle claims that it is a kind of category mistake to say ‘I bought a left-hand glove, a right-hand glove and a pair of gloves’. His idea is that it would be inappropriate to use this sentence to describe my buying of exactly one pair of gloves, and his analysis of this confusion is that it is wrong to mix talk of the categories ‘glove’ and ‘pair of gloves’ in this way. (Of course, this analysis of the inappropriateness is no good – using the sentence ‘I bought a left glove and a left glove’ to describe my buying of exactly one left glove would be equally inappropriate). At any rate, he thinks that, by analogy, saying ‘I have a body and a mind’ is a kind of category mistake, because it involves the same kind of double counting problem. (As far as I understand, his famous ‘colleges/university’ examples are mainly intended to demonstrate this point).
    4. ‘Anyway, it would be good if all this stuff about category mistakes could be sorted out, then we could fix the Wikipedia entry!’ – Absolutely. I have been working on sorting this stuff out for some time now (finishing up my DPhil dissertation on this topic and plan to turn it into papers/a book soon), but it would be good if someone fixed the wiki entry to begin with :-).

  8. Matt Weiner says:

    Ofra, thanks. Definitely a problem with my proposed definition is that it’s hard to determine which kinds are relevant. There may be some purposes where we would say that ascribing a property to an even number that only odd numbers can have is a category mistake, though I can’t think of one offhand.

    But analogously, there can be some purposes for which it seems like a category mistake to ascribe a property to a number that it can’t have in virtue of being a non-integer; as if I were to ask whether pi is prime. Though if I were to say that the square root of pi is 2, it seems that that would not be a category mistake even though that’s a property that pi can’t have in virtue of being a non-integer. So if we wanted to work with my definition we’d need to say something like “in virtue of the kind of thing it is for some relevant kind.”

    Of course “for some relevant kind” is all kinds of fuzzy, but I think the concept of category mistake is probably fuzzy in the same way.

  9. clayton says:

    Matt (if that is your real name),

    Yes, the worry had to do with the modal properties a thing couldn’t have such as the property of being such that it could have had its origins in something it happened not to have. Other things of a kind could have had their origins in other things (e.g., statues other than David could have come from this now unused hunk), but nothing of any kind could have had their origins other than their actual origins. I suspect that any modal test will be too course grained.

  10. Matt Weiner says:

    nothing of any kind could have had their origins other than their actual origins.

    Is that true in virtue of what kind they are, though? I think you may be right about the modal analysis, but it seems to me that an ironclad counterexample would have the following properties:
    (1) X is of (relevant) kind K
    (2) Nothing of kind K can possibly have property F
    (3) Something not of kind K can possibly have property F
    (4) “X is F” is intuitively not a category error.

    (3) is necessary in order to ensure that X can’t have F in virtue of being kind K, and it seems to me that this example doesn’t fulfill kind K. OTOH, (3) may be too strong to capture every “in virtue of.”

    Here’s a problem I thought of: We’re arguing about what kind of sculpture David is. I say “David is made of wood.” You say “David is made of marble.” Of course, you’re right, and (given the metaphysics we’re presupposing) David couldn’t have been made of wood; as a marble thing it couldn’t have been made of wood. In fact (1)-(3) are all satisfied. Yet I don’t seem to have made a category error; I’ve made an error about what category David belongs to, but that’s not a category error, it’s just a mistake.

    If we come up with something similar for logical types it seems like Ryle’s definition might have the same problem. I say numbers are sets, you say they aren’t. Should we each accuse the other of a category error? That doesn’t seem useful. Maybe whoever is right will have said something true but not useful, though.

  11. joshparsons says:

    I once made the mistake of trying to teach Ryle. I set the usual selection from Concept of Mind as a reading, and after skimming it, made up some study questions to put on the syllabus. These were “What is a category mistake? Why does Ryle believe that Descartes has committed one?” I never had such an easy time making up study questions. I put off the detailed preparation of the material for the week I would teach it. Afterall, Ryle’s a great writer and a classic analytic philosopher; surely it would be easy to go over his main argument in class.

    Two months later, the time came to prep my lectures and I re-read the reading more carefully. To my horror, I realised that I had no idea what the answer to either study question was. Really none. If I had been a student in my own class I would have felt that I was failing.

    It was as if I’d set my students to read a slab of Nietzsche, and not bothered to read the piece myself until the day before class. It’s clear enough what Ryle thinks, and you feel like you’re on his side, but the argument consists in vaguely poo-pooing Descartes’ views and describing some obviously foolish strawmen who are supposed to be making the same mistake.

    Now I just teach “behaviourism” and knock it down without showing the kids any live behaviourist. One strawman begets another.

Leave a Reply

You must be logged in to post a comment.