I know that if I have evidence that you have a sibling then you cannot be an only child. That doesn't mean that I know that you're not an only child. I need to actually have such evidence and recognise it for what it is. If I meet your brother but don't know that he's your brother then I don't know that you're not an only child. — Michael

In that case, meeting my brother (or meeting my mother) is simply not an instance of you "having evidence that X is my brother/mother", because the conditional (a) doesn't hold here, and so it's not a counterexample (because you don't know that you've met my brother even if you did, while (a) says that if the event of "meeting my brother" is to be regarded as evidence for you that the guy is my brother, then you must antecendently know that the occurrence of the event of meeting that guy would entail the fact that you've met my brother - but in your story you don't know that. The point is that you don't have to suppose the stronger conditional (b) but (a) is sufficient to handle your stories).

Your conditional account of knowledge doesn't work. You need to actually have such evidence and recognise it for what it is. The problem is in recognising the evidence for what it is. How you do recognise that the experience you have is evidence that you're not a brain in a vat? The sceptic claims we can't. — Michael

I agree that you have to recognize that you indeed have the evidence when you do have them (and this is taken care of by (a)), but it doesn't follow that you capacity to recognize the evidence must be infallible.

If I'm seeing a tree in a waking state, and believe that I do, then I do recognize my evidence for what they are (and it's not a matter of coincidence). The question whether I can detect some possible cases of illusion is irrelevant to my ability to recognize the evidence in the veridical cases.

Recognising some evidence as ruling out the possibility that you're wrong, as per your opening post. If I can't recognise some evidence as ruling out the possibility that the tree I see is just in my head – that I'm a brain in a vat or being deceived by an evil demon – then I don't know that the tree I see isn't just in my head. — Michael

You have to distinguish between two senses of recognizing your evidence: a. knowing that if you have evidence of type E then you cannot be a brain in a vat and b. infallibly knowing that you indeed have evidence E and not merely seeming to have them. I agree that knowing (a) is necessary for knowledge, but not knowing (b).

It means knowing that the tree you see isn't just in your head. That you're not dreaming or imagining or a brain in a vat or being deceived by an evil demon. — Michael

You just repeated the world "know" here, but I asked what are the relevant conditions (according to the skeptic) for knowing such a thing? (which we allegedly fail to meet)

If you don't know that the tree you see is an external world tree then you don't know that there's an external world tree, even if the tree you see is an external world tree. Just as if I don't know that the blond woman I see is your mother then I don't know that your mother is blond, even if the blond woman I see is your mother. — Michael

On my account there's no difference between knowing that something is a tree and knowing that something is an "external world tree", so I don't understand what you mean.

The relevance is that simply having a veridical experience isn't enough to claim that one has knowledge of the external world. You need the periphery understanding that your experience is veridical, but that kind of understanding is impossible (according to the sceptic). — Michael

Again, we've been over this already. I don't know what you mean by "knowing that your experience is veridical". As it stands it just means nothing, and so it is no objection.

In my example I see your blond mother, but I still don't know that your mother is blond. It's not enough that I exercise a reliable detecting capacity in the right sort of environment. I need periphery understanding (in this case, that the person I see is your mother). — Michael

I didn't say that having this sort of capacity is sufficient for knowledge, that's a different question. Obviously you have to be responsive in the right sort of way to the deliverances of you perception - I agree that not all instances of e.g. directly seeing a tree (de re) are instances of knowledge, but I don't see the relevancy here.

Your approach seems to just be the truism that if the belief is true then it cannot be false. — Michael

But that's not what I'm saying. Compare the case of believing that there's a tree outside because you seem to see that there's a tree outside, and just guessing correctly that there's a tree outside. In the first case you are basing your belief on a capacity that within certain parameters (that is, in the right sort of environment) is an extremely reliable tree-detector (meaning it can detect real trees as opposed to fake trees, but not real trees as opposed to extremely vivid hallucinations). While in the second case of guessing you are not exercising any such a capacity, but merely depending on pure chance. But if you are exercising your tree detecting capacity in the right sort of environment, then it's not a matter of chance that you seem to see a tree whenever there are real trees (and not seeing trees when they are none), so there's an important difference between relying on perceptual experience, and merely guessing correctly.

How can you claim that knowing that p consists of anything other than excluding the possibility that p is false? If you can know that p, without excluding the possibility that p is false, then what does "knowing that" amount to? — Metaphysician Undercover

I didn't actually deny what you said that I denied (that knowing p means ruling out p's falsehood) - and I agree with you on this. I only disputed the claim that knowing must also entail being able to detect (from the subjects point of view) all cases of p's falsehood, which is something else.

What must rule out the possibility of falsehood is your objective evidence. So for example, if you are in a waking state perceiving a tree in full daylight, you cannot possibly be in this very same state and still be wrong about the tree. But what you can be in error about is whether you are really in such a state or merely seeming to be; so it might seem to you that you are looking at a tree in a waking state, when you are actually asleep, or something like that.

So my point is this - when you know that p, then your evidence must objectively entail the truth of p - but this is compatible with your subjective inability to detect all instances of p's falsehood (but you should be able to detect at least some of them, but not the ones that the skeptic says that you should - e.g., you would notice if someone cut down the tree for example, but not if you were a brain in the vat in a treeless world etc.).

You've phrased this very ambiguously. Let's say I meet a woman and that, unbeknown to me, she's your mother. In one sense it is correct to say that I know your mother, but in another sense it's wrong to say that I know that she's your mother.

The sceptic is saying something about the latter sort of knowledge. I can't know that the experience I'm having is an experience of an external world. — Michael

We've been over this already. There's no much that I can add.

It has everything to do with that. The sceptical hypothesis is that we can't know that we're not brains in a vat, for example. No amount of examining cats is going to help us answer that. Unless there's some known feature that "real" cats have and simulated cats don't. — Michael

Again, you are just assuming here that knowing that p requires the ability to detect every conceivable possibility of p being false, while I saying that such an assumption is unwarranted. Again, there's no much that I can add to what I already said on this point.

I think that's just true by definition. We say that some object is "external" if it exists when not being experienced. — Michael

But when you actually do experience the object, then it is not "outside" you experience in the sense that the object doesn't make any difference to how things appear to you, since it surely does (- if this is what you meant to deny in your last comment).

So having a veridical experience doesn't prove that the experience is veridical. And the sceptic's claim is that we can't know that our experiences are veridical. — Michael

The skeptic can claim this, but as I said many times, it doesn't follow from this that we cannot know anything about the external world.

So in this case we have some means to test the accuracy of the report. We can check to see if things are as the report says. But what can we do to test the veridicality of an experience? How do we check to see if our experiences correspond to some external world? This is where we need to be able to recognise some feature that veridical experiences have and non-veridical experiences don't. — Michael

As I said, in most cases it is pretty simple thing to do. If you want to make sure that what you are seeing is a cat and not e.g a rock or a dog, you can approach it and examine it more closely and so on. So there are actually many distinguishing features by which we can ordinarily tell whether an experience corresponds to the world or not (and they are not arbitrary) - but of course such tests are not infallible and don't rule out all possible cases of error, but the point is - so what? Relying on perceptual experience is not simply a matter of making guesses (as you claimed) because we do have all sorts of reasonable standards for determining whether an experience warrants a certain belief, and so there's something very weird in the skeptics' claim that all those standards are completely worthless unless we can first of all rule out a possibility of some sort of crazy error. Surely the relevant question to ask whether I know that something is a cat is whether what I see appears as a cat - and the question whether I'm dreaming or something like that, has nothing to do with it and hence simply irrelevant.

The fundamental problem with the notion of an external world experience is that it places the very thing that defines it as being an external world experience outside the experience. The brain in a vat has no way of knowing if his simulated world is an accurate representation of the external world. — Michael

I don't agree that the external world is "outside" our experience. And surely you have to know first whether there is actually an external world, in order to know whether it is outside our experience, don't you think?

First, what even is this evidence you could have of things being as they seem? Second, knowledge in this context means 100% certain with absolutely no possibility of it being false in any hypothetical or theoretical situations, no matter how inplausible. So no, unless you can distinguish any deception or illusion, the claim is not palusible. — BlueBanana

It all depends on what one calls "evidence" here. My proposal is to distinguish between having evidence for something (which is an objective matter of how your mental state is connected with the real world), and being able to subjectively detect various cases of error, which is something else. And I think that once you have this distinction clearly in mind, the skeptics' argument loses much of its plausibility. The skeptic after all doesn't really challenge the actual reliability or veridicality of your perceptual states (= the evidence available to you), but he's just changing the subject by talking about remote possibilities of error, which don't really undermine the objective evidential status of your experience, but merely challenge your entitlement to rely on them. But then why should one really care about some imagined scenarios of deceiving demons etc.? After all, the skeptic doesn't even claim that the possibility that you are dreaming is remotely plausible, but only that it is "possible" in some abstract sense, and therefore it seems to me that we can safely ignore this possibility.

No, but that's not the point. It proves that it's a possibility, even if the odds of that are one in infinity, and that there's no absolute proof of you being awake. — BlueBanana

Yes, but the question is how the absence of such proof is relevant to my knowledge of the world.

Simply having an accurate report isn't evidence that the report is accurate, and simply having veridical experiences isn't evidence that the experiences are veridical. — Michael

Well yes, having an accurate report doesn't prove by itself that your report is accurate, but having an accurate report is having good evidence for the thing which is reported (you've just called it "a good report" after all). And from the fact that you don't have some independent checks for the reliability of your report it doesn't follow that it is epistemically imprudent or irresponsible to rely on a report which looks to be a reliable report as far as one can tell. And so if it is indeed a good report then it seems to me reasonable to describe this as a case of genuine knowledge of the thing which is reported.

Far better to just stick with "experience of an external world" and "experience of an imaginary world". So the sceptical hypothesis is that we can't (or just don't) know if our experiences are of an external world or an imaginary world, and more strongly that the experiences that we ordinarily consider to be of an external world are actually of an imaginary world. — Michael

I think that actually your terminology is way more confusing then mine, but I won't argue about that. I just wanted to show that there's a natural understanding of "evidence" (and in particular the evidence provided by perceptual experience), on which all this talk about distinguishing between dreaming and reality is irrelevant to our actual state of knowledge of the world. And so there's really no good reason to think that our inability to detect some crazy possibilities of dreams should undermine our confidence in sense experience.

All that the skeptic can tell us is that either our experience is wildly misleading in a very systematic way because of some ad hoc coincidence, or that it is very reliable as we usually take it to be. But it should be noted that the skeptical scenario is in some sense 'parasitic' on the intelligibility of the ordinary scenario that we normally believe in (since it is essentially defined by contrasting it with the normal case); and so if those are our only two choices (and there's really nothing in between), then it is not clear why there should be even a prima facie reason to think that the skeptical scenario is even remotely relevant to our ability to know the world.

But how do you know how things are? You can't know that, you can only ever know how they appear to you. — BlueBanana

Well, if you have evidence that entails that things are the way the they seem, then I think it's very plausible to call it knowledge.

How so? If you are not, then it's possible that case of deception or illusion is true, and so you can't be sure things are how you think they're instead of you being deceived, ie. you don't know for sure. — BlueBanana

But what if cases of illusion or deception are not actually possible, given the state the I'm right now in? (i.e., of not dreaming but being awake) Maybe if I were dreaming, then I couldn't recognize that I'm dreaming, but what are the odds the I'm really dreaming right now? Surely my not being able to recognize all cases of dreaming, don't prove that I'm actually dreaming right now! So you have to distinguish between subjective and objective reasons, so to speak.

But simply having a veridical experience isn't evidence that it's a veridical experience, just as simply having the real painting isn't evidence that it's the real painting. — Michael

But this analogy doesn't work. If you are in a waking state your experience is strongly correlated with how things are in your environment (otherwise it would not be a waking experience by definition), and hence it is correct to regard your perceptual states as good evidence. But in the case of paintings, unless you are an expert, nothing in your experience is correlated with the fact whether the painting is genuine or not, and so your experience shouldn't be considered as evidence in favor of either possibility.

The trick here is in treating "waking experience" and "dreaming experience" as two different kinds of state, and this distinction is epistemically relevant because having a waking experience is by its very nature to perceive reliably how things are in the world, which is not the case with dreams.

The evidence must allow you to rule out the possibility that you're wrong, and for that you need to recognise some distinguishing feature, as with the painting. — Michael

The question here is which possibilities do we have to rule out. Suppose that in the case of the paintings, I can detect the "distinguishing features" of some forgeries, but not others which are more sophisticated. Does the fact that I cannot rule out more sophisticated forgeries somehow undermines my ability to detect the less sophisticated ones? Surely not (at least in the sense of that I can know when something is a forgery if it is not very sophisticated). So I grant you that we cannot distinguish all conceivable cases of perceptual error (such as in the radical skeptical scenarios), but we do have the capacity to rule out some cases of error. And after all, the skeptic doesn't claim that we cannot e.g. distinguish dogs from cats in normal lighting conditions etc., he's having in mind much more exotic cases. And thus the idea is that the inability to distinguish some hypothetical cases of dreaming from waking, doesn't prove that we can never distinguish dreaming from waking, since in a straightforward sense we often can do just that.

But the first premise doesn't say that in order to know p you have to recognize states of having evidence for p according to some distinguishing feature, it merely says that it's enough if you just have the right sort of evidence, but doesn't mention recognition. You have to add a further premise here to make the argument valid.

Being in a state that is logically inconsistent with being asleep and dreaming does not imply there is perceptual evidence of being in any state, — BlueBanana

What do you mean by "perceptual evidence"? According to how I understand evidence, something is evidence for p, if its presence entails the existence of p. Now if having a waking experience defined as a state such that things necessarily match the way they appear to me, then if I'm having a waking experience I'm in a state which entails the presence of whatever that I perceive - and this is just a case of having evidence according to my definition.

There is a hypothetical scenario where the other state does not entail that information, of which you'd be completely unaware of because it's your only source of how the things are. — BlueBanana

But this is irrelevant to the question whether I can know how things are. The mere possibility of some state in which I would be dreaming that I'm awake, doesn't prove that I cannot know anything when I am awake.

And how does it logically prove that I don't know that I'm having a veridical experience from the fact that I can't recognize such a feature? — Fafner

If you can't recognize it, how do you claim you know it? — BlueBanana

Perhaps because the possibility is irrelevant to my knowledge. There's no reason to assume that whenever I know something, then I must be able to recognize all possible cases of deception or illusion. Perhaps I need to be able to recognize only some.

To recognise some feature that veridical experiences have and non-veridical experiences don't. — Michael

And how does it logically prove that I don't know that I'm having a veridical experience from the fact that I can't recognize such a feature?

Being in a state that is logically inconsistent with being asleep and dreaming does not imply that one can distinguish the two states, because the inconsistency does not stem from the difference in the mental state but from the fact whether one is awake or asleep. — BlueBanana

I didn't say anything about being able to distinguish between the two, I only said that if you are awake then you have perceptual evidence which rules out the possibility that you are dreaming (which if true, proves that premise (3) in the argument is false).

To be able to determine if I'm having a veridical experience or a non-veridical experience. — Michael

And what does it mean to "determine"?

But given that I can't distinguish between the real painting and the forgery, in the case that my belief is true I'm just lucky. — Michael

And what does it mean to being able to 'distinguish' between the two? I formulated the condition for knowledge in terms of having evidence that favors the one possibility over the other (and by 'favors' I meant having evidence which presence entails the truth of what is believed). But it seems to me that you have something else in mind here, and so as I said, this objection seems to me irrelevant to the argument that I'm considering.

(and good look formulating a deductively valid argument which derives from the assumption that one cannot distinguish (in some possible circumstances) between p and not-p, the conclusion that one cannot know that p - since the conclusion doesn't logically follow from this premise alone).

I said that if "dreaming" is defined as being of an imaginary world, and if the experiences which we claim to be waking experiences are actually of an imaginary world, then those experiences aren't actually waking experiences but dreams. — Michael

But claiming something to be a waking experience is a different matter from it actually being a waking experience, so what you are describing is not an example of a waking experience that "falls under the umbrella of "dreaming" ", but just a plain instance of dreaming.

or they can claim that they are defined as being the experience of an external and imaginary world respectively, in which case both types of experience fall under the umbrella term "dreaming" (even though they have other properties to distinguish them). — Michael

How does it follow? Surely being awake doesn't fall under the umbrella of "dreaming" (because it isn't dreaming).

The claim is that if the experience isn't veridical then your belief is false and that if your experience is veridical then your belief is just lucky. So the sceptic doesn't need to claim that there are or aren't any trees. He just argues that either way there isn't knowledge. — Michael

But how is skeptic supposed to prove that if my experience is veridical then it is lucky? I'm not quite sure what you mean by "luck" here, but at least in the literature it means roughly something like basing your belief on an epistemic policy which doesn't reliably track the truth. But what makes a policy reliable is its relation to the environment in which is is exercised, so that means that you must know some facts about the world in order to assess whether my perceptually based beliefs are reliable. If my capacity to visually distinguish trees from non-trees (in the environment in which I in fact live) is reliable, then my beliefs which I form on the basis of experience cannot be lucky when true - and the mere possibility the this very same policy could misled me in some other worlds (in which there are no trees) is irrelevant to the question whether my believes are in fact lucky.

To be able to get beyond luck to actual knowledge you must somehow know that your experience is veridical, which according to the sceptic isn't possible. — Michael

It seems to me that you are changing the argument (of course you are welcome to formulate a different argument). Premise (1) in the argument as I formulated it, doesn't say that you must know that your experiences are veridical but something else. That would be a different argument.

I've already addressed it. It's epistemic luck, not knowledge. Unless you can distinguish between a veridical and a non-veridical experience (or between a real painting and a forgery) then you can't know that your experiences are veridical (or if the painting is real). — Michael

The problem with the epistemic luck reply is the same. You cannot say that a belief is an instance of epistemic luck as opposed to knowledge without antecedently knowing the objective background against which the belief was formed. So you cannot know that my belief that I see a tree is merely lucky if you don't know whether in fact there is a tree in front of me (and perhaps also whether my ability to detect trees in general is genuinely reliable). But of course the skeptic doesn't claim to know that there are in fact no trees (if he wished to prove that my believe is merely lucky), so it would be of little help to appeal to epistemic luck on his behalf.

There's no mention of waking and dreaming states, simply the observation that things appear real to us in dreams, yet are not, so, likewise, the reality of things that appear to our senses may be doubted. This is what leads directly to the famous declaration cogito, ergo sum. — Wayfarer

As I said in another comment, I'm not interested in a textual exegeses of Descartes, the argument is only inspired by some things that he says, but it doesn't mean that it fits 100% the text (and also you are looking at the wrong section of the text, and the cogito argument has nothing to do with this topic, but let's put Descartes aside).

And you can make that distinction on the basis of the nature and quality of experience - that the experience of waking and dreaming is qualitatively different. — Wayfarer

But this is not what the skeptical argument says. The whole point is that there is no subjective differences between waking and dreaming states, otherwise, how appeal to dreams supposed to prove skepticism?

But we can't distinguish between a veridical experience of a tree and a non-veridical experience of a tree (or so the sceptic claims), and so the analogy holds.

I just don't know if right now I'm being deceived by an evil demon. — Michael

It doesn't follow, and I've shown this already. Mere inability to distinguish on subjective grounds all non-veridical states from veridical proves absolutely nothing about ones knowledge of the world. Again look at the original argument. If you have a proposal how to fix it then I'll be glad to hear it, but I think that as the argument stands, the transition to the third premise is obviously unsound, and hence the final conclusion doesn't follow.

the way things appear to you when you are awake usually matches very closely the way they really are.... — Fafner

But the sceptic will say that this simply begs the question, i.e. assumes what it sets out to prove. If one's life were a perfectly well-ordered and consistent dream state, then the correlation between experience and the objects of perception could likewise be perfectly consistent and empirically verifiable. Even if fundamental physical constants were actually part of an illusion, provided they were consistent, then they would still make accurate predictions. — Wayfarer

It doesn't contradict what I said in the quote. You just gave another example of a dream state, but my question is, what distinguishes dream states from waking states? And the point is that you cannot draw this distinction without reference to how the world is like.

You're not comparing like for like here. Simply having a veridical experience is comparable to simply being shown an Arabic word. Just as the latter isn't the same as understanding Arabic, the former isn't the same as knowing that the experience is veridical. — Michael

It all depends on what one means by "knowing that the experience is veridical". My point is that there is something confused in the way the skeptic thinks that we ought to know this. The skeptic thinks that our knowledge should be ultimately grounded in subjective states which don't themselves assume anything about how the world is. But I was trying to show that there's no coherent conception (by his own lights) of "subjective states" which could be described independently of how the world actually is. So to insist that one ought to prove that his experience is really veridical by the skeptic's rules is to miss the point. And I tried to undermine his argument not by responding to his challenge, but by rejecting the assumption which lead to it.

Of course one cannot prove that he "knows" that his experience is veridical in a non circular way (that is, without appealing to the experience itself), but my point is - so what? If we reject the assumption of the skeptic that mental state could be individuated without reference to the external world, then there would be no justification to keep clinging to the skeptics demand the we must distinguish veridical from non veridical experience in a way that would satisfy him.

To offer another example, let's say that I'm shown a painting, which may or may not be a forgery. You seem to be suggesting that if it's the real painting then I know that it's the real painting, or if it's the forgery then I know that it's the forgery. But that's just not right. The fact that it is or isn't the real painting isn't sufficient to claim knowledge that it is or isn't the real painting. And so the fact that it is or isn't a veridical experience isn't sufficient to claim knowledge that it is or isn't a veridical experience. — Michael

The difference is this: a non expert cannot visually distinguish forgeries from non-forgeries, while we all can in most cases distinguish trees from non-trees (and such like). So I'm not saying that whenever you believe that p, and that p happens to be true, then you know that p; what I'm saying is that you can know that p, if you have the capacity to perceive that p. You don't need a justification to believe that you have the capacity to perceive that p which is independent from the deliverances of that very same capacity; but this is not the same as saying that you can know that p whenever p happens to be true (and in your example you've assumed that the observer lacks the capacity to visually detect forgeries, so it doesn't apply to my account).

So knowing Arabic requires more than just being presented with Arabic words. Somehow I need to learn that these symbols are in fact Arabic words. And so knowing that our experiences are veridical requires more than just being presented with veridical experiences. Somehow I need to learn that these experiences are in fact veridical. — Michael

But my point is that knowing that something is an Arabic script is inseparable from the ability to understand Arabic (that is, you cannot describe someone as being fluent in Arabic without presupposing that he can as a matter of fact read Arabic); and analogously, knowing that you are having a veridical experience is nothing but just having it as a matter of fact. You know that you are seeing a tree just by seeing a tree (and perhaps some other conditions like the absence of fake tree facades in the vicinity, re the Gettier problem), and I claim that it is unreasonable to require something else in addition. (and also notice that I'm not trying to persuade the skeptic in a non question-begging way that I in fact do know that what I see is a tree (see the end of this reply), I'm only trying to give a reasonable characterization of what it is to be in a state of knowing that one is looking at a tree - see the end of my reply).

But, again, at best your argument is "if our experiences are veridical then we can know that our experiences are veridical", but given that the skeptic questions the antecedent, your argument would seem to beg the question. I could even turn your argument around and argue that because our experiences are not veridical we know that there isn't an external world (or at least none that we see). — Michael

First, I didn't assume, nor was trying to prove, that we do know that there is an external world; I was merely trying to block the skeptical conclusion that we don't know - but of course it doesn't prove by itself that we also do.

And secondly I don't agree that the argument begs the question. The skeptic claims that my having an experience say of seeing a tree, isn't a good evidence that there's a tree in front of me, since my perceptual state doesn't rule out the possibility that I'm actually dreaming. In other words, he says "I don't know whether there is or isn't a tree in front of you - you may be right about that - but what I do know is that you mental state is consistent with the possibility that there is no tree in front of you", and I have argued that he can't actually say this since he can't know whether my perceptual experience is consistent with the possibility that the tree is absent, because it conflicts with his own (implicit) characterization of the difference between waking and dreaming states. He cannot know what kind of state I'm in unless he also knows whether there is in fact a tree in front of me.

This is why I called it a "version" of his argument, and my aim wasn't to correctly represent his actual philosophical views. And in any case, my argument does show that we can (in some sense) acquire certainty about the world, since being in a waking state does actually logically entail that you cannot be in error about what you perceive.

Then let's change my example slightly. I am given a piece of paper that either has random symbols drawn onto it or Arabic writing. In either case I have cognitive "access" to the evidence; I can see it right in front of me. But I don't know if it's random symbols or Arabic writing. I need some second order understanding of how to distinguish the two. And with the case at hand, I need some second order understanding of how to distinguish a veridical and a non-veridical experience, which the skeptic claims we do not have. — Michael

In this case the "evidence" that you need is simply to know Arabic (or at least being able to reliably identify Arabic writing). And this is not "second order" evidence in my sense, because for you to know that you know (say) Arabic, all you need is simply the mere ability to read Arabic, and you don't need in addition some further ability to identify your own fluency in Arabic which is distinct from you just being fluent in the language.

I need to also be able to recognize them for what they are, but according to the skeptic we can't. — Michael

But this is an absurd demand, since it generates a regress. If something is evidence only by virtue of having a second order evidence in its favor, then nothing can ever count as evidence, since that second order evidence would in turn require having additional third order evidence and so on. And I don't think this requirement has much plausibility if we consider what we usually call "having evidence" in the normal sorts of cases outside of philosophy.

I think it is very reasonable to think that having evidence in favor of something simply consists in a. having in some sense a cognitive "access" to the evidence (and notice that in your example this requirement isn't met since you cannot read French, while in the perceptual case you can directly perceive your environment without further ado) and b. the evidence itself should in some sense guarantee the truth of what it is evidence for (and this requirement is met when you are having a waking experience, since in that case your cognitive state is very reliably correlated with the facts that you perceive).

(Also notice that this understanding of evidence is strongly 'internalist' in the epistemic sense, and so it is not an externalist account in the style of Nozick or Sosa)

This 'running together' is not simply nonsense, but is in fact a result of an unwarranted mixing together of 'kinds' of (concepts of) freedom, not all of which can be spoken about in the same breath without causing issues with conceptual inconsistency. — StreetlightX

I can agree with this formulation, though it still leaves open the question of how we ought to identify whenever a word/concept is used in the same or a different sense within a given context. And this is the point at which many philosophers (even some 'followers' of Wittgenstein, like P.M.S Hacker or Paul Horwich) fall into the trap of attempting to come up with semantical or metaphysical theories to explain how words mean what they mean. The challenge here (at least if you are a Wittgenstenian) is to avoid this sort of theorizing, and still have a clear method of providing a philosophically illuminating analysis of meaning or uses of language.

One can of course 'come up' with a new, novel meanings for every apparently mismatched pair of words, but not without paying a certain semantic price. — StreetlightX

But what prize?

And also notice that in the examples that I described we do not come up with a new meaning, but rely on the 'old meaning' which is extended to new cases that no one thought about before.

Of course I agree that if someone talks about measuring the brightness of a sound by using lumens, without providing any concrete explanation of what he means, then it would be a good prima fecie reason to suppose that he is talking nonsense. But my point is that you cannot know this just by looking at the sentence which he utters (that is, only from the particular words from which it is composed and their combination). There's nothing intrinsically erroneous in this or that combination of words so that you could have an easy or quick philosophical method for identifying 'category mistakes'.

Thus, as I said in my other response to you, it is not clear to me what is the philosophical utility of the term 'category mistake', if you don't mean it to be understood either semantically or metaphysically. If you simply use it to be synonymous with "words that lack sense" then by calling it this name, it doesn't explain anything about why some words happen to be nonsense (in this or that context) nor does it tell you how to identify whether something really is nonsense.

(and also I should note that this is not how 'category mistakes' are commonly understood by philosophers, so I think the way that you use the term is pretty misleading, and so is your explanation of the term in the OP)

Hah, I was waiting for the synaesthesia response. But then, one has provided a context by which one could make sense of such a question. And part of my point was that is just what is needed: sense-making can be understood as simply another way of saying 'context-providing': of showing how a difference makes a difference, of elaborating the stakes behind any one question. — StreetlightX

But if any sentence can be made sense of in a suitable context, then what's the point of talking about 'category errors' in the first place? If identifying a "confusion of kinds" (as you put in your OP) is not a sufficient condition for rendering a certain sentence nonsensical, then I don't see any philosophical utility in this idea (to condense the main point of my previous lengthy post).

I actually believe that there's no such thing as 'category mistakes'. I subscribe to the idea that nonsense arises only when we fail to provide a clear meaning to our words, and never because some punitively nonsensical string of words kind of means something, and by virtue of this meaning they generate a 'mistake' or a logical 'clash' of meaning. And I want to claim that on the first understanding of nonsense (where nonsense is simply the lack of sense) the kind of philosophical criticism that you envision is not available (there's no apriori rule that allows you to determine which combinations of words are nonsense), and the second view is simply mistaken, if not incoherent (that is, the view that nonsense is a species of 'defective meaning').

(I should also mention that my views here are largely derived from Cora Diamond's reading of the Tractatus - the so-called 'resolute' understanding of nonsense, which I believe can also be found in W's later philosophy (e.g., see PI 500: "When a sentence is called senseless, it is not as it were its sense that is senseless. But a combination of words is being excluded from the language, withdrawn from circulation."), so I believe that you are wrong to ascribing to W' any such view)

I think there are two common ways for understanding what a 'category mistake' is supposed to be. On the first, category mistakes arise when we have a sentence that attempts to combine words that refer to things that somehow necessarily cannot be combined or put together in reality. So to take your example, one might argue that something in the nature of color and ideas preclude the predication of color to ideas. But this I believe is an incoherent position. If we are talking about the nature of the things in the world to which we are referring by our words, then it doesn't seem plausible that we can know apriori how things can and cannot be 'combined' in reality (and it is not clear in the first place what 'combined' is supposed to mean here anyway). For example we don't know apriori which substances chemically react and which don't, and for this reason we must perform experiments to know this (and there's no reason to think that the case of color and ideas (or whatever) is any different). But if this is an empirical question, then it must make sense to ask about any two substances whether they are or aren't chemically reactive, since even if they are not, it would be false and not nonsense to say that they are (and if it was nonsense, then chemical experiments would've been superfluous).

A different way to understand the idea of category mistakes would be to say that certain sentences are nonsense because of the pre-existing meaning of the words that compose them. So to take your example again, one could say that the claim that ideas have color is nonsense because the meaning of 'color' and 'ideas' somehow precludes their combination in the sentence; the sentence is in some sense semantically or grammatically 'ill formed'. And this claim I think is plainly false, because words in natural language don't fall neatly into 'categories' as you say; it seems to me that whatever 'categories' you would assign to any word, it will be always easy to come up with counterexamples.

So to illustrate, I think it is very easy to come up with an examples where your alleged 'categorically mistaken' sentence would make sense--without changing the meaning of the words, or using them in a 'metaphorical' or 'non-literal' sense. Suppose that the physicalist identity theory turns out to be true, and every time a person entertains an idea, there happens to be a brain state with a particular color (e.g., the neurons change color or something of that sort). Another possibility: imagine a synesthetic tribe of people that always see colors whenever they think about something, and so they classify their thoughts according to color. Is anything about the 'conventional meaning' of 'color' or 'ideas' entitles you to rule out such uses of language as somehow necessarily nonsensical? I think that would be indefensible thing to say. It's true that we usually do not use color terms to talk about ideas, but we do use other similar adjectives like "dark ideas" or "bright ideas".

Or think about the case of sounds or musical notes: we call certain notes 'high' or 'low', and also low sounds 'dark' and high sounds 'bright' (is a high note 'high' in the same sense that a mountain is high? yes and no). Clearly nothing about the 'literal' meaning of high/low or bright/light (in the sense of high/low places and bright/light colors etc.) determines whether it makes sense to talk about sound or pitch in this way; there's nothing logically/grammatically/semantically necessary about these terms that tells you that it is either permissible or impermissible to apply these predicates to sounds. It so happens that we can make sense of such talk for various sorts of reasons, but not by virtue of some 'metaphysical' or 'semantic' fact that we've discovered about the predicates.

So the moral is that the idea of 'category mistakes' is either false, or philosophically useless. You cannot demonstrate that some sentence doesn't make sense just by inspecting the words from which it is composed and assigning a logical/semantic/metaphysical categories to which they belong. Language simply doesn't work this way, since meaning is a very plastic thing that can develop and extend in various unpredictable ways. Here's a final example: some axioms of euclidean geometry used to be considered logically necessary in the strongest sense of the word, until people constructed new types of geometry where it suddenly did make sense to deny them, and this happened without either changing the meaning of the words or using a metaphorical talk. So the point is that even the meaning of technical terms in mathematics and geometry is not 'fixed' in advance, so that we are entitled to say that once and for all such and such sentences must be false or nonsense no matter what.

No, it is not an objective fact that either the world is this way, or it is not this way. The concept of "the world" and the existence of the world, as understood by human beings, is supported by the concept of matter. Aristotle demonstrated that matter is necessarily exempt from the law of excluded middle, which you are employing to produce your so-called "objective fact". This refutes your argument. — Metaphysician Undercover

Now you are simply appealing to authority. Some famous philosopher said it, therefore it must be true... It seems to me that you've ran out arguments, so I'm out.

All we have to refer to, as "the way which the world is", is how the world appears to us. This is our interpretation of the supposed objective reality. And how the world appears to us, may or may not be a true representation of the way which the world is. Both sides, (a) and (b) are subjective. — Metaphysician Undercover

You don't see it, but what you said here actually proves my point. If the world appears to you in a certain way, then it is an objective fact that the world is either the way that it appears to you, or that it isn't. So having a mere appearance of reality already makes your appearance objectively true or false. So for example if you have an appearance of seeing a cat on the sofa, then it is either objectively true that there's a cat on the sofa, or objectively false. Nothing can be an appearance unless it already objectively represents reality to be in a certain way (and "objectively represent" is not the same as representing truly - a false representation is still objective in this sense, since it represents what is not the case, but could have been).

So ironically, interpretation is precisely what grounds objectivity.

Now you've gone back to claiming that "known fact" is necessarily the way that the world is. Clearly this is not the case, because what is referred to as known fact is often proven wrong. — Metaphysician Undercover

All of this is just irrelevant to my definition (and the rest of what you say in the comment). It doesn't matter what we say or believe about the facts. My definition merely states the conditional that if someone knows that P, then P is a fact. If P is not the case, then by definition the subject cannot known that P (and it doesn't matter if he himself is aware of this). I'm not claiming that we actually know the facts, it is only a definition of what it means to know something.

It seems to me that you either don't understand what definitions are, or what conditionals mean (or both), because your objection simply makes no sense.

First thank you for the very detailed and informative reply.

However, when suitably conjoined with a contextualist account of knowlege, disjunctivism would not render a unique verdict in this case, as it indeed shouldn't. The mistake that must be avoided is the idea that there is a unique objective probability of the perceptual experience being an experience of a real barn independent of the characterization of the epistemic power being exercized. How might this probability rather to be evaluated? What contextual range of counterfactual circumstances is it that might relevantly be taken into consideration for purpose of determining whether or not your belief that there is a barn count as knowledge? I'll let you think about it a little before I propose my own suggestion. — Pierre-Normand

And I agree that disjunctivist could possibly respond by giving some sort of contextualist account - but this was my point, it doesn't look that the disjunctivist has any inherent advantage over other accounts simply by virtue of being a disjunctivist. He still needs some pretty complicated story to tell in order to explain the difference between ordinary cases of knowledge and the Gettier cases, and that story would probably be as complicated as any of the proposed non-disjunctivist solutions to the problem (like those that appeal to 'sensitivity' or 'safety'), and hence no less contentious. I think the disjunctivict would've had a real advantage if he could handle the Gettier problem without the need of any "fine tuning" to specifically address the Gettier cases (that is, if the solution was built-in from the start).

And besides, I think that many non-disjunctivist epistemologists (at least those with a contextualist leaning - which I think nowadays is the majority view) would agree with most of what you say, and some I'm sure would even agree that we need, as you said, to reject "the idea that there is a unique objective probability of the perceptual experience being an experience of a real barn independent of the characterization of the epistemic power being exercized.". So my point is that it doesn't appear that there are any special resources which are available to the disjunctivist qua-disjunctivist to handle the Gettier problem, which are not also available to other epistemologists.

The sentence means what it means, without being interpreted? I give up. — Metaphysician Undercover

Sure, you can assume here anything you want about interpretation, but it doesn't matter because you have (b) as well that grounds its objective status.

The sentence only makes sense to a person interpreting it. Without a person interpreting it, it makes no sense, and therefore cannot be true. — Metaphysician Undercover

In my story I assume that the sentence is interpreted, because after all you understand what the sentence "extraterrestrial life exist" means. But the point is that interpreting the sentence is not sufficient either to guarantee its truth or to guarantee that it is known, because interpretation can only give you the meaning of the sentence.

We can't tell the difference between a known fact, and something believed to be a known fact, because they both appear to be known facts. So we call them both known facts. Since we can't distinguish between a known fact and what appears to be a known fact, or just believed to be a known fact, then it cannot be incorrect to call the thing which is believed to be known fact, by this name, "known fact", unless you want to ban the use of "known fact". Therefore your definition of "known fact" is untenable, rendering it always incorrect to use "known fact", because we would never know whether it is a known fact or not. However, if it is acceptable to refer to the thing which appears to be a known fact, as "known fact", then your definition is wrong. So your concept of "known fact" is actually useless. — Metaphysician Undercover

Again, you are begging the question. You just assume that knowledge (in my sense) is impossible without an argument. You wrote: "Since we can't distinguish between a known fact and what appears to be a known fact" - I don't accept this and I don't see any argument to support this claim.