@Banno@Janus@Count Timothy von Icarus @Ludwig V @Sam26
Thanks, and coincidentally, I also have to be offline for 2 weeks, as I'm going out of the country. Appreciate the conversation and look forward to chatting with everyone when I get back.

JTB seems to be saying, "You can only know something if it's true." Or wait . . . maybe it's saying, "You can only know something if, right now, you are sure it's true." Which is it?
— J

Why not both at the same time? — javra

Because they aren't asserting the same thing, or at least we need an argument to show that they do. The first speaks of the truth of a statement ("something") irrespective of whether I know it to be true. The second brings me into the picture, insisting that I have to be sure it's true.

Otherwise, it's all circular. "I know X" becomes the same thing as "I know that X is true." But this presupposes that "knowing X" involves a definition of knowledge that include knowing that X is true. Isn't this what JTB was supposed to demonstrate?

Filling in with an example:

1 - "I know that I live in Maryland"
2 - "I know that 'I live in Maryland' is true"

What is our warrant for claiming that 1 and 2 assert the same thing? Doesn't it involve a stipulation or presupposition about what it means to know something -- specifically, a stipulation involving the term "true"? Statement 2 talks about what is true, statement 1 does not. But isn't this the very thing JTB is supposed to give us? -- a reason to include the truth of a statement as part of the knowledge claim?

But also, see my previous comments about the inquiry into "What is knowledge?" JTB wants to pin down the correct use of "I know"; I'm suggesting that it might be more profitable to look at the ways we actually use "I know." I don't think they correspond to JTB. There are many things I believe I know, but am not certain they are true. JTB would argue that, therefore, I'm using "know" incorrectly. Whereas I'm saying that it's JTB that needs correction, not me. This latter position lacks punch, of course, unless the "me" can be turned into "us" with sufficient frequency. We need a fairly widespread agreement on the faults of JTB in order to claim that it doesn't capture our common practice.

EDIT: I agree that it isn't possible to claim 1 without also claiming 2, and vice versa. Perhaps that's all you mean by "Why not both?" If so, it's fine. My argument above is that they are nonetheless different claims. And consider a 3rd statement: "'I live in Maryland' is true." This can be the case even if I don't know I live in Maryland.

As I and others have pointed out in previous posts, ontological truths occur, i.e, ontological correspondence/conformity to that which is, was, or will be actual do occur. Implicit in every belief is an assent to that which is true. — javra

Before I reply in any detail, let me be sure I understand you. Are you saying there are ontological truths about the future? That is, the future exists now in such a way that statements about it are, at this moment, either true or false? (I think this is what you mean by an ontological truth?)

If one assumes that JTB must be absolutely devoid of any possibility of being wrong, then we all communicate all the time via beliefs which we don’t know to be true.

How would this not then result in a societal chaos of sorts wherein most all trust goes down the drain? — javra

I'm not sure that societal chaos would follow, but I agree with your point about JTB. That's part of why I'm hesitant to accept it as a good description of knowledge.

Sometimes it helps to pull back from the intricate details and ask ourselves, What are we trying to say about knowledge and truth? JTB seems to be saying, "You can only know something if it's true." Or wait . . . maybe it's saying, "You can only know something if, right now, you are sure it's true." Which is it? There seems to be support on this thread for the former construal: All that matters is that Statement P is true, not whether I can know that to be the case.

I think that, for JTB to be worth using, it ought to take us closer to the second construal: My claiming that P is true ought to say something about what I actually do have some surety about. And this is not a binary judgment. Our justifications will vary in strength. How strong does a J have to be in order to cross the "sure" barrier? I don't know if that's answerable. It's a bit like the old "heap" problem. Is there some line we can name, below which I'm not quite sure, and above which I am? (Notice that I'm using "sure" instead of "certain" or "have knowledge," because I don't want the potential circle to confuse this question.)

For example, it would be odd for a typical westerner to say "though I believe it, I don't know whether I will eat anything tomorrow". — javra

A good example of how different people work with "know." I would in fact say just that, perhaps precisifying it: "I strongly believe that I will eat tomorrow, though there is a very slight chance that I won't." Would I also claim knowledge? This is where it really starts to get murky. According to JTB, I can't, since I don't (yet) know if "I will eat something tomorrow" is true. It may be true, in which case my claim at T1 is knowledge -- the problem is, I can only be aware of that in retrospect. In practice, however, and leaving aside the somewhat bizarre (to me) requirements of JTB, I'd rate the statement pretty low on the knowledge scale. Any number of things might prevent me from eating tomorrow, sickness especially. Whereas "The sun will rise tomorrow" gets an enthusiastic thumbs-up from me as a piece of knowledge, despite the fact that it too is not certain -- there are defeaters, as I proposed to @Janus (who wasn't impressed!).

Once again, though, we have to remember that a philosophical question such as "What is knowledge?" or "What counts as knowledge?" can be taken in at least two different ways. We can be asking, "What is the correct way to understand what knowledge is?" If we answer that, then we can go on to ameliorate the incorrect understandings and usages. The other way is to ask, "How is the word 'knowledge' used? What range of situations and applications does it cover?" If we answer this, we're no longer trying to say which (one) of the usages is correct. In fact, if it turns out that many people use "knowledge" in a manner, or in situations, that don't fit a proposed correct understanding of what knowledge is, this may, and should, give us pause. It may suggest to us that a "one size fits all" construal of knowledge is misguided. This doesn't mean that Total Chaos is now rampaging. It just means that the question is nuanced, and often depends on interpretation.

Sorry if I wasn't clear. Satisfying the JTB criteria is how we know a sentence is true, supposedly. What makes it true would be some version of Tarski-truth. I was saying that one could point to a JTB-verified sentence and say, "Well, since it's a JTB, I know it's true." But this kind of knowledge doesn't involve any meaningful justifications on my part. It's a kind of reliance on authority, the authority of having passed the JTB criteria.

Possibly I also haven't been clear about why the PoV matters -- who is doing all this. The phrase "point to" is meant to raise this question. If I am the one who declares a sentence to be a JTB, then presumably I have satisfied myself, as best I can, about the T part, and provided my own justifications. But if you tell me it's a JTB, I haven't. All I can do is accept the "deduction" that, if it is indeed a JTB, then it must be true. I think you've been assuming, in this discussion, that a single person is taking all these steps, but there's nothing in JTB that requires that. We don't ask, "Have I verified that this sentence is true?" but rather "Is this sentence true?"; we don't ask "Have I provided good justifications?" but rather "Are there good justifications?"

Is this what JTB is for?

And thanks for your patience with this.

But of course, we could reply here that you "know it to be true" just in case you have a justified belief that it is true, and it is true. I don't think that answers J's question though, because we still have to assume the "it is true" part. — Count Timothy von Icarus

Yes. Or if not assume, at least spell out some criteria that don't merely repeat the J criteria.

Well, no. If S is some sentence that satisfies the criteria JTB, then by that very fact it is true — Banno

Surely not. This is the absurd "deduction" I was addressing above. Satisfying the JTB criteria is not what makes a sentence true. It's not the "very fact" we're looking for. What makes a sentence true will be, let's say, some version of Tarski-truth.

Or to put it another way, what makes a sentence true is satisfying T; you don't need to bring in J and B at all. The question is, Can we imagine a situation in which T would be apparent to me -- not to a hypothetical anyone, but to me, the user of the JTB criteria -- on other grounds than the J?

The circularity, so far as there is one, is in your then asking "But is it true?" — Banno

Perhaps the right question, then, is "Who knows it to be true?" Does the person applying the JTB criteria have to know this? Or is it sufficient for it merely to be the case, with no one knowing it? This leads back to my concern about the use of all this.

we can deduce, from the fact that we have JTB of X, that X is true.
— J
That's circular. You can only satisfy the JTB if you know that X is true. — Ludwig V

Of course it's circular. But doesn't it follow? If "My aunt lives in Denver" is a JTB, it must be the case that my aunt lives in Denver. No further verification is required. My point is precisely that this is absurd. To avoid the circularity, you have to posit X as true without knowing it to be true, whether on the grounds of pragmatism or T-truth or grammar or something else. That's the move I'm still considering.

Again, there is a difference between P being true and it being established that P is true. J still hasn't taken this to heart. — Banno

I feel like one of the blind guys that's got a different part of the elephant! The difference is completely clear to me. What isn't clear is what JTB is supposed to be used for. As I asked above, "I come back to the question, What is JTB for? Is it a theoretical, criteriological account of what it would mean to know something? Or is it supposed to actually help us evaluate a given piece of putative knowledge?" If you've addressed this already, my apologies, but could you say again?

Truth is a logical device, setting out the move between a sentence and what it says.

The "T" in JTB is that move. — Banno

Yes, this has to be correct, it seems to me, with the stipulation that the result will be some true sentences and some false sentences. A great deal of the conversation here centers on how certain we can be, or have to be, about the status of T.

For the purpose of defining knowledge, we can assume that we have a concept of truth and worry about what it is on another occasion. — Ludwig V

A little too breezy for me! But I see what you're saying; perhaps I'm just being stubborn in wanting to get a preview of what the concept of truth must be, in order for JTB to work. Or see above: Maybe we're simply not sure what the work of JTB is.

So you accept knowledge based on authority. I'm a bit surprised - it is quite unusual for philosophers to accept that. They usually, if only by implication, seem to believe that only first-person verification is satisfactory. That's a very strict criterion and cuts out most of what we (think we) know. — Ludwig V

Indeed. I can only say that, in practice, we use "know" rather differently than that. Philosophers can recommend ameliorating our less rigorous usages, of course. Then "know" becomes a sort of technical term. Do I not know that, say, general relativity is true, because I can't personally verify it? I'd contest that. I feel a great deal more certain of general relativity than I do of many of my own apparent perceptions!

My concerns with JTB are all about how the truth of P is supposed to be established
— J

I would think it isn't. We just act like it is true until we are prompted to reconsider. — Count Timothy von Icarus

I have trouble with that; surely the justifications matter? Can we act like P is true -- that is, assert that we have the T for JTB -- if the justifications aren't strong? I come back to the question, What is JTB for? Is it a theoretical, criteriological account of what it would mean to know something? Or is it supposed to actually help us evaluate a given piece of putative knowledge?

To go toward the mirage is Justified True Belief (if one is not familiar with modern day science). And who knows, it might lead to water. Eventually. — Outlander

I'm not clear why this would be JTB. Even the ancients knew about mirages, judging from classical literature. And, in the unlikely event I had the presence of mind to ask myself, there in the desert, whether "That is an oasis" is a JTB, I would answer no; I don't have a good-enough justification, or an independent fact-check, to include it as a T in the formula. But a guy can hope!

But we don't know that X is true via JTB, but via whatever the truth conditions are for X. — Ludwig V

Well, but there's the rub -- we do. There are two ways of knowing that X is true, on this construal of JTB. We can verify the truth conditions of X (and remember, this a convenient phrase that contains its own puzzles and disagreements), or we can deduce, from the fact that we have JTB of X, that X is true. If you tell me, "I know X, because I have JTB of X," and I believe you, then I know, or at least believe, that X is true, without knowing anything about its truth conditions. Can this work in the first person? Can I myself have JTB of X without knowing the truth conditions of X? This puts us back to justification, and what counts as a good one. Is personal verification of the truth conditions the only truth-guaranteeing justification? Or, if "guaranteeing" is too strong, the only good-enough justification?

If say I am certain that something is the case, then I mean that there cannot be any doubt about it. Then I would say I know it to be the case. If I think something is the case but there is any possible doubt it, then I would say that I believe it to be the case, but do not know it to be. — Janus

I see why this is attractive. "Possible doubt" is the question, though. Is it possible that the sun will not rise tomorrow? (and of course I mean "sunrise" as shorthand for what actually occurs). Well, yes. An alien civilization inimical to ours might choose tonight to destroy our solar system. That is not impossible, or incoherent, or against the laws of physics, etc. Yet I, and I think all of us, would be happy to say that "The sun will rise tomorrow" is a piece of knowledge. As I say, I'm sympathetic to why you'd want to tighten up "knowledge" so it equates to "certainty" but is that really what we mean when we say we know something? Or would you want to argue that solar death by alien attack is impossible? On what grounds?

it's on us to show why we think there needs to be something of the sort where "P is *really* true," and that we must be able to assert that this is so, or even "know" it, and how exactly that is supposed to work, since it seems one could function "pragmatically" whilst only speaking to one's own beliefs without "knowing" that any other beliefs exist. — Count Timothy von Icarus

Excellent. My concerns with JTB are all about how the truth of P is supposed to be established. @Banno recommends just starting with that truth, which seems similar in spirit to the pragmatic approach you describe. I'm still thinking it over.

In the Matrix scenario there is no skepticism about the real world—in fact that is what those who see through the virtual illusion are trying to get back to. If Descartes considered this he would still be faced with the question of being able to doubt the purported real world just as much as he can doubt the virtual world of the Matrix. — Janus

I don't think so. Descartes' skepticism is not about the real world. It's about whether my experiences are veridical. He's not saying that, if these experiences are not veridical, then there is no real world. He's saying we can be deceived. Presumably the Evil Demon can be undeceived, just as the Lords of the Matrix can be.

But I think you may be getting into an unnecessary tangle because you (seem to be) focused on the special case of "I know that I know.." — Ludwig V

I can see how you would think that. I probably could have expressed it better. But the "know that I know" issue comes up within JTB itself. If it's right that we can't know X is true via JTB (since it's an element of JTB, not a result, and would require a previous demonstration of knowledge), then we might never know whether we know a given X, since we wouldn't know if X was true. So "know that I know" is really meant to express "know that JTB is satisfied."

That is very helpful - it helps me understand much better Kant's connection of time with number and space with geometry. — Wayfarer

Goes for me too, thanks.

The facts that help you decide on your attitude are irrelevant to whether p is true or not. — Banno

Whether I believe that p and on what grounds is a matter that is entirely distinct from the question whether p is true. — Ludwig V

That's the part I'm questioning. What does it mean to you that something is true? I'm guessing it would be some version of Tarski truth. So how could that possibly be independent of the grounds of justification? That seems to be setting up truth as some quality or property that just is. But we all know that's not right: truth is a property/attribute/quality/judgment/or whatever of propositions, not objects.

So, is the idea that we can possess knowledge (i.e., possess beliefs that are justified and true) but we can never know that we possess knowledge (unless perhaps the object of knowledge is our own beliefs or experiences)? — Count Timothy von Icarus

Not quite. I'm saying that this would be the unwelcome conclusion if this way of construing JTB is adhered to. To avoid this conclusion, I'm suggesting we alter or abandon JTB, not our confidence that we can know we have knowledge.

The Matrix Hypothesis I think is absurd, because it posits that there is a real world in which the virtual world we inhabit is sustained, and this means the need for explanation is just pushed one step further back. — Janus

But Descartes' doubt isn't about explanation. He believes it's possible to doubt whether my experiences are veridical -- that is, of the things they appear to be of. He's not questioning experience in general. The Matrix hypothesis would represent such a doubt.

any verdict I give on the truth or not of the information is inescapably only what I know or believe. — Ludwig V

This is the nub, I think: It can never be what you know, only what you believe. Never, that is, without raising the specter of the vicious circle. Because to already know that the piece of information is true, that knowledge would have to have been verified via JTB. (This all assumes you think JTB is a good yardstick for knowledge, of course.)

I think it is true that we can equally say that Macbeth is seeing something that isn't there or Macbeth thinks he sees something that isn't there. — Ludwig V

Good. Seems that way to me too.

We've talked about this in the context of Williams' book on Descartes. I think you're being too harsh.
— J
Oh dear! My memories of that are, I'm afraid, a bit vague. Perhaps I am being too harsh. — Ludwig V

Harsh on Descartes, not me, I hasten to say; your forum manners are impeccable. As for vague memories . . . don't get me started. Aging is a fog, obscuring near and far.

Dunno if any of this helps or not, — Mww

I appreciate it a lot, thanks.

The differences in the text is so subtle.
….In the Aesthetic, we have intuitions which are given as “the matter of objects”;
….In judgement of mathematical cognitions, we have “….exhibition à priori of the intuition which corresponds to the conception…” for which the matter would be irrelevant;
….In judgement of philosophical cognition we have conceptions which conform to the intuition insofar as “…the intuition must be given before your cognition, and not by means of it.…”. — Mww

Clearly these are differences, as you say. I'm focused still on the discussion in the Prolegomena, where Kant says:

In one way only can my intuition anticipate the actuality of the object, and be a cognition a priori, namely, if my intuition contains nothing but the form of sensibility, antedating in my mind all the actual impressions through which I am affected by objects. [Kant's italics] — Prolegomena 282

How do you interpret this? How might it apply to 7+5 and the use of fingers?

I think that the justifications are mostly the same sorts of facts that would show whether X is true or false. But there can be justifications to the effect that I am in a position, have the skills, to know - which are of a different kind or level. — Ludwig V

Yes, I think so too. Let's see what @Banno says, and then I'll try to show where I'm going with this.

We can loosen that requirement, and say that "X is true" is pre-JTB and therefore not a knowable instance of truth. This seems to resemble more closely our actual practice.
— J
It's true that we rarely consciously and specifically apply the JTB. It's a formalization of what (normally) we actually do in a messy, informal way. I don't understand what it would be for something to be "pre-JTB". — Ludwig V

By pre-JTB I mean that we would enter the "JTB situation" already believing that X is true. Our belief in X is not a result of what is about to happen if we successfully apply JTB, hence not knowledge. The difference between believing and knowing is important here. If, in trying to determine whether I possess a piece of knowledge, I ask myself, "Is X true?" (a JTB requirement for knowledge), I can only reply, "I believe so." I can't say, "I know it is," because this initiates the vicious circle.

Asking the question "what is a hallucination?" in the sense that you seem to mean it presupposes that a hallucination is an object. — Ludwig V

I didn't mean it to. It can just as well be an activity or an event. We can still ask what it is, taking "is" in one of its many familiar usages. The question was whether there's a "correct way" to describe the activity of hallucinating using the word "see." I'm saying, no, it's terminological; "see" can work either way.

The problem is that he does not consider what actual limitations there are on doubts, and reduces it to the possibility of saying "I doubt that..." in front of almost any proposition. But if we ask what the content, the reality, the significance, of the doubt is, we find nothing. — Ludwig V

We've talked about this in the context of Williams' book on Descartes. I think you're being too harsh. If it should turn out that I am really a brain in a vat or a Matrix-bound person, then my doubt about the objective references of my experiences is well-founded. It's not merely a linguistic construct. Indeed, this possibility seems much more plausible to me than the Evil Demon! -- though still pretty unlikely. In any case, perhaps we've just set different "limitations on doubts," and what seems doubt-worthy to me, doesn't to you. But surely this kind of doubt signifies more than nothing, wouldn't you say? It's a thinkable thought, and not nonsense.

In the case of the conception of a priori itself, Kant did not mean it with respect to time as such, but with respect to placement in the system as a whole. — Mww

Good.

To then say a priori, as it relates to time, is before experience, is not quite right, — Mww

Yes, that's what I was suggesting.

Now we see synthetic judgements a priori are only representations of a very specific cognitive function, a synthesis done without anything whatsoever to do with experience, and of which we are not the least conscious. — Mww

But then why does Kant say:

We must go beyond these concepts by calling to our aid some intuition which corresponds to one of the concepts -- that is, either our five fingers or five points . . . -- and we must add successively the units of the five given in the intuition to the concept of seven. — Prolegomena 268

But we stop dead in our cognitive tracks, when the very same synthesis is just as necessary but for which immediate mental manipulation is impossible. — Mww

Kant notes this in the same section: "larger numbers . . . however closely we analyze our concepts without calling intuition to our aid, we can never find the sum by such mere dissection."

the cognitive part of the system as a whole, and in particular the part which reasons, does something with the two given conceptions… — Mww

I still read this "something" as requiring intuition. Do you not see it that way?

OK. Bear with me. Let's say I'm in a "JTB situation"; that is, I want to find out whether I possess a piece of knowledge. Will the justifications that I cite -- the J in JTB -- for why I believe X refer to the same sorts of facts that, out in the world, would show whether statement X is true or false?

Example: X is "Taurize is a village in France." My justifications for believing X would be, let's say, "I've looked in a reliable atlas, and spoken to someone who's visited Taurize and confirms this." (And we could tighten this up ad infinitum, but you get the idea.) Now suppose I want to find out whether X is true -- whether Taurize really is a village in France. Would the facts necessary to do this be the exact same ones cited as my justifications for believing X? Are the two sets completely congruent?

The problem I have is that he doubts things on the mere logical possibility that he might be deceived by an Evil Demon. — Janus

This may seem like quibbling, but he doesn't actually doubt things. He points out that it would be possible to doubt them. Of course he knows that no sane person is going to doubt most of their own experience, but that doesn't satisfy his Method. He wants the grand prize -- absolute certainty, beyond even the possibility of doubt. I personally feel that we don't need that in order to do metaphysics and epistemology; Descartes disagreed, hence his Method. But we really shouldn't see him as raising "philosophers' doubts" for the sake of skepticism. He detested skepticism and believed he had refuted it. (And we have a perfectly good modern version of the Evil Demon: the "Matrix hypothesis.")

Justification is only for beliefs, not for those things known with certainty. — Janus

OK, that seems like a good way to look at it, with perhaps the caveat that it's reasonable also to ask, "Why are you certain?" or "What makes you rely on this experience?" (similarity to previous ones, presumably). These aren't requests for justifications in the same way that asking for a justification of a belief is, but their aim is to ask for an account, a rationale. I can't simply reply, "Well, I just do." This rarely comes up, of course, but it's worth noting.

Also, you can say that Descartes "feigned" skepticism, but he wasn't trying to fool himself or others. He wasn't just being annoying. He was trying to pursue a method he believed would lead, by elimination, to clear and certain knowledge.

Like @frank, I'm not sure I get this. "That is a prime number" is true (or false) regardless of what John thinks about it. The question is, How confident can he be that he knows which is which? Of course there are fuzzy cases, but let's just consider this straightforward one. Isn't he trying to bootstrap himself into a JTB?

Maybe this is a good way to frame the problem in terms of JTB:

Do I have to know that X is true in order to use it as the T in a JTB statement? Is that knowledge the result of a previous application of JTB? etc. I know you think this can be dissolved, but I still don't see how. Truth is certainly foundational to the whole set-up, I agree, but the question seems to be whether I have to know that X is true. We can loosen that requirement, and say that "X is true" is pre-JTB and therefore not a knowable instance of truth. This seems to resemble more closely our actual practice.

I cannot justify that I have that knowledge to you, if you believe me you take it on faith. — Janus

Agreed. I think we're speaking of self-justification here. Can you justify to yourself that "I am thinking X" is necessarily true? This starts to become merely verbal, depending on what sort of thing you think a justification is. Self-evidence, on some accounts, requires a justification, or at least an explanation. On other accounts, it's the very thing that obviates justification. Does it much matter which construal we pick? What matters is the concepts in play, not our terms for them.

But your description is excluding the "straightforward" answer that the drunk is hallucinating a pink elephant. — Ludwig V

That was kinda deliberate. For what is a hallucination, and how do we talk about it? Is there an obvious consensus? Some would describe hallucinating as "seeing something that isn't there"; others would describe it as "thinking you see something that isn't there". Is there a meaningful difference, apart from choice of terms?

In a sense, of course, it just kicks the can down the road, — Ludwig V

Precisely, as I just described. But the can isn't important, in my opinion, so this shouldn't bother us.

I think your knowledge of Kant is deeper than mine, so please say if you don't agree with my interpretation of these passages in the Prolegomena.

A priori means “prior to experience.” If you tell me you have seven beers in the fridge and I bring to another five to give you, I can know you have twelve beers without opening the fridge door. That’s a trivial example, but it illustrates the point: the truth of 7+5=12 doesn’t depend on checking the fridge. — Wayfarer

Right, that's the standard interpretation, but think about it: Prior to how much experience? Can I know about the 12 beers if I don't know what beer is? Can I know it without knowing about counting? Can I know what 7 or 5 or 12 anythings are without lived experience? So where do we imagine the "a priori judger" standing, so to speak, when they make their judgments? (BTW, you can see immediately that this is yet another place where Rodl's important questions about propositions surface.)

Indeed. Hectoring rather than conversation.

It’s a perfect case of the synthetic a priori . . . — Wayfarer

The debate about this often centers on how "prior" the a priori is supposed to be. What is the ideal situation in which an a priori judgment is imagined to take place? Prior to what, exactly, can we know that 7+5=12? Prior to what can we know that antimatter exists? Prior to (or independent of) observations, perhaps, but prior to any experience of the world whatsoever? Priori to knowing how to count? "A priori" and "a posteriori" have conventional interpretations, and a better Kantian scholar than I could perhaps tell us precisely what Kant envisaged, but the division doesn't feel like a "natural kind" to me.

Yes, it's just terminology, as I said. I certainly don't feel strongly about it. We can use an ambiguous term like "see" any way we want to stipulate, as long as everyone knows what that is!

hallucinations and mirages are not introspections (aka, self-examinations of one’s own being, thoughts, etc.) … but imaginings (such as can occur in daydreams) seen with the mind’s eye — javra

That sounds right -- but it also means that we can't say the drunk saw a pink elephant. Seeing with the mind's eye is a metaphorical extension of what it means to see something.

I can't help feeling that applying the description "pink elephant" to whatever I am seeing is not immune from mistake. — Ludwig V

I'd say the mistake is in the use of "see".

But in any case, this is about choice of terminology. We could say to the drunk, "No, you didn't," and mean either "You saw nothing" or "What you saw wasn't a pink elephant." Neither one is obviously correct, apart from pedantry. But we're all three dividing up the conceptual territory the same way. (And I think @sime is getting at this too, with their A and B analyses of beliefs.)

Oh good lord, sorry, I meant the Prolegomena. :grimace:

Synthetic a priori = adds new content, but is knowable independently of experience.

That last category was Kant’s unique insight. Mathematics is built around it — “7+5=12” is not analytic, because “12” isn’t contained in “7+5,” but it’s still a priori. — Wayfarer

Let's slow down on this one. Kant doesn't speak about "content" in the [Prolegomena] (where the 7+5 example is discussed). He says that the concept of "12" is not the same as the concept of "7+5". According to him, we need an "intuition" ("perception" would probably be our way of saying it today) of the physical in order to discover "12". (He suggests that our five fingers, and then seven fingers, would do the trick.) "Hence our concept is really amplified by the proposition 7+5 = 12, and we add to the first concept a second concept not thought in it." What Kant regards as analytic here is the judgment that 7 and 5 must add up to some number -- but this does not tell us what particular number.

The place where this can be challenged, I think, is the reliance on intuition. If this is truly the case, don't we have to question whether the judgment is indeed a priori? Kant addresses this in Sec. 281: "How is it possible to intuit anything a priori?" I don't want to take this any farther, except to say that the case for math as a series of synthetic a priori judgments, even on Kant's own terms, is far from closed.

But Wittgenstein disagrees with Moore’s depiction of this form of certainty as a kind of empirical knowledge. — Joshs

Yes, that's a good link to Moore and Witt. In this context, I'm not so concerned to ask whether there is such a thing as direct knowledge -- or rather, I take it as given that there is. The problem is to explain what we mean by that, and what can count as a justification for our confidence in it. Witt, as I understand him, is raising a doubt not about the "knowledge" part, but about whether it is empirical. That's also what I was moving toward, in saying "If 'I am having thought X' needs a justification, we'd have to look elsewhere." Hinges and forms of life are good candidates.

I agree that when it comes to claims of knowledge, justification is required. On the other hand I know many things with certainty that require no justification simply because they are directly known―in these cases justification just doesn't enter the picture. — Janus

And this resembles the "A or ~A" case, where it's difficult to see it in terms of justifications. Still, I think the conclusion we ought to draw from this is that we're not quite sure what a justification is. What sorts of reasons may play a part in justification? (We noted earlier that a "good justification" is very unclear, in many cases.) If you ask me for my justification in believing "I am having thought X right now" and I reply, "I am directly observing this occurrence as we speak," have I offered a justification? Perhaps so; that's one way of understanding what reasons count as justification, though I'd probably also need to say something about the previous reliability of my direct observations. Or we might conclude that "directly observing" and "having" are two ways of saying the same thing, so no actual reason has been offered. Then, if "I am having thought X" needs a justification, we'd have to look elsewhere.

The problem I think you see is of your own creation. Or so it seems to me. — Banno

You may well be right. But I haven't yet satisfied myself one way or the other. Your input, as always, is appreciated.

That bird looks sad. Is he a cousin of the fly in the fly-bottle? (I also notice that he could leave the cage anytime he wanted to.)

it could have been better written. — javra

No worries, I'm not always a model of precision myself.

the proposition of “the planet is physical and roughly spherical” is taken to be an instance of knowledge, thereby being [a] JTB claim — javra

OK, good, so on this construal, JTB becomes a reasonable description of our ontic and epistemological requirements. We're not saying that "The planet is physical and roughly spherical" is known to be true, pre-JTB. Nothing is known to be true, in that non-fallibilist sense.* Rather, we bring our belief and our (presumably pretty impressive) justifications, and fill in the "T" with some such formula as you used above. The result is knowledge. This should make us suitably modest about what knowledge actually consists of, but at the same time extremely confident that the target statement is knowledge if anything is.

* The monkey wrench is logical and other putatively analytical truths. I think such statements require a different series, and type, of arguments to explain what knowledge of them would mean. Historically, was JTB proposed as being useful for determining that kind of knowledge? I'm used to thinking of it as an attempt to evaluate knowledge claims about the world, but I don't know the back-story. There's something odd about asking whether "If A, then not (~A)" is a belief, or how we might justify it. But I'll leave that for others.

You seem to be expecting something from the JTB account that it does not provide. It's not a theory of truth. — Banno

That's true (sorry!), but it's a theory of knowledge that includes, as one of its criteria, that a statement be true. So if JTB tells us that X is known, it must also tell us at the same time that X is true.

But see my response to @javra, below.

This is a very helpful analysis. It sharpens the question, Are we meant to take the knowledge guaranteed by JTB as infallible or certain (@Janus's question)? You argue convincingly that we can't -- that this can't be the point of JTB. So if we're happy to substitute "has no reason to suppose that one’s 'belief that X' is not in fact ontologically true" as a synonym for "knowledge", then all is well. I don't know whether every proponent of JTB would be happy with this, though.

To try to reduce possible confusion, how this works in practice: “I know that the planet is physical and roughly spherical,” is a claim of JTB. — javra

Here, I wonder whether you misstated your target sentence. Are you talking about a JTB claim for "The planet is physical and roughly spherical" or for "I know that the planet is physical and roughly spherical"? A great deal depends on this, so I'll wait until you reply before going on.

inside reasoning is non meta reasoning. And must be used to determine truth of an argument generally. Rather than using a meta lens like psychology or sociology or genetics.
— Jack2848

Yes, that's right. Typical 'outside' claims, of the type Nagel is criticising in that essay, are claims that attempt to justifiy reason based on evolutionary biology. — Wayfarer

And I would add that such claims help themselves to terms like "justify" or "explain" as part of their discourse about why reason can be reduced to biology! This would seem to be a performative contradiction, as Nagel says. Or else our entire understanding of what it means to justify or explain something has to change radically.

Q1a. Yes.
Q1b. Yes.
Q1c.Yes - follows from Q1a: if you believe it, you believe it to be true. — Banno

But Q1c was not about belief, but rather truth. Yes, it follows from believing something that I also believe it to be true, but that's not a reply to Q1c, which asks "Is it true?" Nothing I believe can supply the answer; it depends on the facts.

JTB is supposed to help us evaluate knowledge claims -- keep us epistemologically honest. And on this construal, it can't.
— J
It doesn't tell us if they are true or not, so much as if they are known or not. — Banno

That's what I don't see how to separate. We both agree that only true things can be known. So if JTB tells us that X is known, it must also tell us at the same time that X is true.

I'm good with all that. Just wanted to make the case that almost anything we claim to be true requires some (potential) justification.

If something is true by definition or if something is logically self-evident, or if the proposition concerns something being directly observed, then I would say we need no further justification. — Janus

I agree. But all three of these things -- truth by definition, logical self-evidence, and the reliability of direct observation -- are ways of demonstrating justification. To understand this, imagine explaining any one of them to an intelligent child. They all involve steps, cogitation, judgment, insight. We don't simply see why they are true, or at least not usually. In fact, as you know, the reliability of direct observation can be challenged, and the challenge is precisely for a justification as to how such observations lead to truth.

JTB sets out criteria for a sentence to count as knowledge. It is not a method for determining the truth of some sentence. — Banno

That's right. So, anticipating your investigator image, using JTB would go something like this:

Q1. Do I have knowledge of X (a proposition)? Depends on . . .
Q1a. Do I believe X?
Q1b. Can I justify my belief in X?
Q1c. Is X true?

Q1b is problematic, as we've noted, but let's allow that we understand what a "good enough justification" would look like. So we can go down the list:

Q1a. Yes.
Q1b. Yes.
Q1c. ????

What can we say about Q1c that doesn't involve an appeal to knowledge? We want to say, Yes, X is true. But we would have to know that, just as we know the first two answers, and we can't, because that begins the vicious circle.

I think what you're suggesting is that instead we should say, "I don't know if X is true. Such knowledge is impossible without circularity. But if it's true, then I know X. And if it isn't, then I don't." So, as you say, a strictly descriptive or criteriological formulation. However, I maintain that this is not only useless, but contrary to the spirit of JTB. JTB is supposed to help us evaluate knowledge claims -- keep us epistemologically honest. And on this construal, it can't.

That the sentence is true is one of the criteria for the sentence being known. This says nothing about how we determined if the sentence is true. — Banno

That's right. But then what should we use JTB for?

There is a difference between "P is true" and "J determined that P is true". JTB specifies that the sentence must be true, not that the sentence must be "determined to be true".

This seems to me to be the source of your confusion. — Banno

Well, I think it's JTB that's confused, not me. Because as I said to @Srap Tasmaner, what use is JTB if it can't show us how to tell whether we know something or not? Being told, "Well, you would know it, if it was true" is pretty thin gruel and, as I said, not to the purpose of JTB as I understand it.

You seem to have an image of an investigator looking at a sentence and saying "ok, Criteria one: I believe this sentence; criteria two: this sentence is justified by such-and-such; but criteria three: how can I decide if the sentence is true?" But that's not how the idea would be used - there's an obvious circularity in such a method, surely. If you believe the sentence (criteria one), then you already think it to be true and criteria three is irrelevant. — Banno

Good, this is a helpful image. And if we set it up like this, then it's another way of showing how JTB is faulty. Because it's absolutely right that the T criterion becomes irrelevant; that's what I've been arguing, using a different approach, all along. I would emphasize the justifications rather than the belief, but it comes to the same thing.