But what you've been chasing in this thread is me knowing aliens exist even though they might not. — Srap Tasmaner

Because I understand "aliens might not exist" as "I am not certain that aliens exist" as opposed to just "I do not know that aliens exist". So the argument would be:

I know that aliens exist
Aliens might not exist (≔ I am not certain that aliens exist)
Therefore, I know that aliens exist and aliens might not exist

Why do we say that we might be wrong? Because the evidence available to us has not proved our belief. Even if we want to say that something like 80% certainty is sufficient for our belief to count as knowledge, it seems strange to say that we could be wrong when we have 79% certainty but can't be wrong when we have 80% certainty simply because that's the – somewhat arbitrary – point at which our belief counts as knowledge (if it happens to be true).

I think the "can't be wrong" is only true when our belief is proved (i.e. it is 100% certain).

p ⊬ □p — Michael

Yes.

Therefore (b) is true if there is a possible world where John is not a bachelor. — Michael

I would say knowledge entails certainty. That is, when one comes to know that John is a bachelor, the alternative possibility is ruled out. From my earlier example, for Bob, the hidden coin's orientation could be heads or tails. Whereas Alice has reduced the possibilities to one - the coin's orientation that she observed.

And if fallibilism is true then knowledge does not require certainty, and so knowledge does not entail certainty. I can know and not be certain. Therefore (c) is true if I am not certain that John is a bachelor. — Michael

Fallibilism means that we are capable of making mistakes, not that we might be mistaken in any particular instance. For example, we're capable of making a mistake when adding two and two. But if we conclude that the answer is four, then we can't be mistaken about that. Similarly, we can't be mistaken if we identify the blue ball as blue. To be certain means to have ruled out alternative possibilities (i.e., we don't doubt).

Suppose Alice says, "I know the ball is blue" or even just "The ball is blue". There is no indication of uncertainty there. Whereas if she says, "I think [or believe] the ball is blue" then that suggests the qualifier, "but I could be wrong".

I would say knowledge entails certainty. — Andrew M

Then knowledge requires certainty. If we are not certain that John is a bachelor then we do not know that John is a bachelor.

The argument I offered was premised on the notion that we can know things even if we are not certain, and so I accept that a rejection of that premise allows one to reject the conclusion.

Whether or not we'd want to reject that premise is another matter, but I see that you are willing.

A related question, then, is what it takes for us to be certain that something is true. My initial view is that we can only be certain that something is true if that thing is necessarily true, and so I can only be certain that John is a bachelor if it is necessarily true that John is a bachelor, although perhaps that's a matter for another discussion.

I thought knowing that p (a proposition) meant that I couldn't be wrong if to be wrong means p is false.

For a contingent truth q, there's a possible world in which q is false.

Is the OP getting mixed up between falsehood (in our world) and contingent truth (could be a falsehood in some other world)?

I would say knowledge entails certainty. — Andrew M

And I again think of the shy schoolboy: I'm inclined to say that he knows the right answer, even if his lack of confidence in himself leads him to doubt that he knows what he does in fact know.

Even if you're right, certainty is a necessary but not sufficient condition for knowledge. We generally believe that knowledge must be arrived at "in the right way" to count, to rule out lucky guesses. And we seem to have the very same problem with certainty. Many people are certain Trump won the 2020 election, but their certainty was arrived at in the wrong sort of way. If we still have to give an analysis of the right kind of certainty to get anywhere, will that analysis differ significantly from an account of the right way to arrive at knowledge? Maybe, but it's not clear to me.

Then knowledge requires certainty. If we are not certain that John is a bachelor then we do not know that John is a bachelor.

The argument I offered was premised on the notion that we can know things even if we are not certain, and so I accept that a rejection of that premise allows one to reject the conclusion.

Whether or not we'd want to reject that premise is another matter, but I see that you are willing. — Michael

:up:

A related question, then, is what it takes for us to be certain that something is true. My initial view is that we can only be certain that something is true if that thing is necessarily true, and so I can only be certain that John is a bachelor if it is necessarily true that John is a bachelor, although perhaps that's a matter for another discussion. — Michael

That would be Cartesian certainty. But in ordinary language, we have at least two or three other uses:

(1) Alice was certain that she left her car keys on the table.

(2) The police ascertained the cause of the victim's death.

Per the first (psychological certainty), Alice can be mistaken and doesn't seem to require any epistemic standard. The second (epistemic certainty) entails success and requires an epistemic standard, but isn't Cartesian certainty.

For another potential use, Descartes says that “moral certainty is certainty which is sufficient to regulate our behaviour, or which measures up to the certainty we have on matters relating to the conduct of life which we never normally doubt, though we know that it is possible, absolutely speaking, that they may be false” (PW 1, p. 289 n. 2). SEP

And I again think of the shy schoolboy: I'm inclined to say that he knows the right answer, even if his lack of confidence in himself leads him to doubt that he knows what he does in fact know. — Srap Tasmaner

Yes that's a good edge case. Though consider whether we would trust his answer if we didn't already know it ourselves. To me, it's like someone wobbling on their bike. Do they know how to ride, or are they about to fall off? Compare also a student who can successfully cram for an exam but soon forgets the answers, or who can parrot the right words, to someone who understands the subject and can reliably use and communicate what they know. Having knowledge seems more like the latter to me.

Even if you're right, certainty is a necessary but not sufficient condition for knowledge. We generally believe that knowledge must be arrived at "in the right way" to count, to rule out lucky guesses. And we seem to have the very same problem with certainty. Many people are certain Trump won the 2020 election, but their certainty was arrived at in the wrong sort of way. If we still have to give an analysis of the right kind of certainty to get anywhere, will that analysis differ significantly from an account of the right way to arrive at knowledge? Maybe, but it's not clear to me. — Srap Tasmaner

:100:

To your last comment, from the same SEP article as above:

... there is generally “a reluctance to allow the contextually set standards for knowledge and certainty to diverge” (Williamson 2000, p. 254). That is, the standards for what counts as knowledge and as certainty typically match one another. Nevertheless, in some contexts we can pull them apart. — SEP - Certainty

To me, it's like someone wobbling on their bike. Do they know how to ride, or are they about to fall off? Compare also a student who can successfully cram for an exam but soon forgets the answers, or who can parrot the right words, to someone who understands the subject and can reliably use and communicate what they know. Having knowledge seems more like the latter to me. — Andrew M

This is descriptive of something -- Ryle (and of course Wittgenstein) says similar things -- about how we judge another's understanding and ability. And, as you say, it does seem to capture something about understanding what you know and being able to apply it, and so on.

But it's too strict, isn't it? I can ask someone to remember a telephone number for me, and they needn't understand which part is the (American) area code, which the exchange, and so on. They needn't even know it's a telephone number or what a telephone number might be. They either know the digits by heart or they don't. As long as there's no guessing, they know it. They need to be able to recite it back to me, or to reconstruct it if they chose some odd mnemonic, so there's a still an ability-style test, but it's nothing so broad as really "getting" telephones and their numbers.

We know perfectly well that the sort of person who tends to know stuff, and the sort of procedure that tends to produce knowledge, can fail. (Hence this thread.) And we know just as well that an unreliable person who has an unreliable approach to knowledge is sometimes dead right. We might reasonably prefer the former as an approach to rationality, but we'll miss the boat on what knowledge is.

But it's too strict, isn't it? I can ask someone to remember a telephone number for me, and they needn't understand which part is the (American) area code, which the exchange, and so on. They needn't even know it's a telephone number or what a telephone number might be. They either know the digits by heart or they don't. As long as there's no guessing, they know it. They need to be able to recite it back to me, or to reconstruct it if they chose some odd mnemonic, so there's a still an ability-style test, but it's nothing so broad as really "getting" telephones and their numbers. — Srap Tasmaner

I agree. Alice can know the phone number qua a ten-digit number. But if when asked she says, "I think it's <number>", then that raises a question as to whether she really does know it. If she gets it right, we're probably inclined to say she did know it after all. However, given her qualification, she wasn't certain that she knew it, and thus not certain what the number was.

So in that case we could say that she didn't know that she knew it. But with reflection on her (perhaps repeated) success at remembering it, she could come to know that she knows it. To relate this back to the OP, knowing everything would also require knowing that one knows in each case.

We know perfectly well that the sort of person who tends to know stuff, and the sort of procedure that tends to produce knowledge, can fail. (Hence this thread.) And we know just as well that an unreliable person who has an unreliable approach to knowledge is sometimes dead right. We might reasonably prefer the former as an approach to rationality, but we'll miss the boat on what knowledge is. — Srap Tasmaner

In the former case, when there is a failure, we just say that she didn't know it after all. So knowledge claims don't preclude that possibility. As you note, we're not infallible and our procedures aren't perfect. However, in the unreliable person's case, I would attribute that to luck and not be inclined to say they know it. Even a stopped clock displays the correct time sometimes, but it isn't connected to the world in an appropriate way.

Even a stopped clock displays the correct time sometimes, but it isn't connected to the world in an appropriate way. — Andrew M

Yes, absolutely. Someone who always thinks it's 3 o'clock will be right twice a day, but we couldn't say that they know it.

On the other hand, here's a variation on an old story, called "The Boy Who Thought Wolf": suppose our young shepherd is a nervous sort, and every time he hears a rustling in the bushes he concludes "wolf"; we note that his procedure is unreliable, and conclude that even when he's right, he doesn't know there's a wolf. Like the clock. But suppose one night a wolf comes striding out from the bushes. Now he knows there's a wolf.

We can imagine the criterion here (seeing the thing) being made into a procedure, and say that if it were a procedure it would be reliable, and so we consider this case knowledge as a sort of courtesy.

But that's all backwards. The truth is that we talk of his actual wolf encounter as a potential procedure precisely because we recognize that in this case he knows there's a wolf. Making a procedure of this just relies on what we already know to be productive of knowledge, namely, this kind of situation. The issue of process is entirely derivative.

This is the main point I was trying to defend: judging whether someone has knowledge is a very different sort of thing from judging ability and understanding, that sort of thing. Surely not completely separate! But still noticeably a different kind of thing. Knowledge can attach to discrete, one-off events in a way that many things just don't. (I've sidestepped talk of the senses, among other things, because I'm not even trying to provide some explanation or account or analysis of knowledge, just see how it works when it does. Roughly.)

So in that case we could say that she didn't know that she knew it. But with reflection on her (perhaps repeated) success at remembering it, she could come to know that she knows it. — Andrew M

I think, as a general matter, we should preserve both sides of the coin here, not just our fallibility -- the cases where we think we know and we're wrong about that -- but also where we have misplaced doubt, and do know something despite thinking we don't. Even forgetting and remembering has a place here: you can claim, honestly, not to know where Mike is today, and then remember that he has work -- that is, remember that you do know where he is.

To relate this back to the OP, knowing everything would also require knowing that one knows in each case. — Andrew M

Maybe omniscience can just keep climbing that ladder, knowing that p, knowing that you know it, knowing that you know that you know it, ad nauseam. That's a lot of of infinities though. On a model like this, omniscience might just be incoherent. Whatever.

Knowledge can attach to discrete, one-off events in a way that many things just don't. — Srap Tasmaner

Yes. In the case of the wolf example, the boy can be asked, "How do you know there's a wolf?" Then we can form our own judgment on the evidence.

I think, as a general matter, we should preserve both sides of the coin here, not just our fallibility -- the cases where we think we know and we're wrong about that -- but also where we have misplaced doubt, and do know something despite thinking we don't. Even forgetting and remembering has a place here: you can claim, honestly, not to know where Mike is today, and then remember that he has work -- that is, remember that you do know where he is. — Srap Tasmaner

Yes. For a different kind of example, consider a scientist couching an imminent risk in highly-qualified and conditional terms which the politician interprets as not needing to worry about it then. So the language use and expectations may change for different contexts.

Maybe omniscience can just keep climbing that ladder, knowing that p, knowing that you know it, knowing that you know that you know it, ad nauseam. — Srap Tasmaner

Perhaps it could be tacit. If no doubt is exhibited in the use of knowledge, or the person would respond that they know something if asked, then that would count as knowing that they know.

Perhaps it could be tacit. If no doubt is exhibited in the use of knowledge, or the person would respond that they know something if asked, then that would count as knowing that they know. — Andrew M

I don't think it's a matter of doubt, just a matter of admitting fallibility. I would say that I know that my housemate is a bachelor, but I also accept that he could be lying to me and have a secret wife that he ran away from. Implausible, perhaps, but not unheard of. Does admitting of this possibility (and not just in the "there is a possible world" sense) somehow entail that I don't know that my housemate is a bachelor (assuming he isn't lying to me)? I don't think so. That I might be mistaken is simply an admission that I am not certain, not an admission of doubt.

So in such a scenario I would say that I know (and perhaps I do), but I'd also say that I might be wrong. Both claims are warranted.

In the case of the wolf example, the boy can be asked, "How do you know there's a wolf?" Then we can form our own judgment on the evidence. — Andrew M

I won't belabor this -- in part because I have so little else to say about it now (!) though I've already been saying something related over here, which no one found interesting except MU, who thought it was stupid.

I do want to say though that I think there's something a little funny going on in imagining judging a sort of canonical case of knowing. (Of the "Well I seen it, didn't I!" variety.) What I mean is something like this: you might take the boy's knowledge claim as inferential, and question him in order to recreate that inference and judge its soundness. So how does that go?

Why do you claim to know there's a wolf in the hills?
-- I saw him.
And how do you know that what you saw was a wolf?
-- I mean, I know what a wolf is.

What else can the boy possibly say? What's odd is the feeling that he would infer the presence of a wolf from his seeing a wolf. That makes no sense, because "see" here is used factively, just as much as "know". "I know it was a wolf because I saw it" is only useful in distinction from "I know it was a wolf because I heard it" or something like that.

I'm not saying that the senses are irrelevant, or that their "testimony" can't be scrutinized. I guess I am a little bit saying something Wittgensteiny -- that there's a point at which your spade is turned and there's nothing more to say. That point regularly involves the use of factive verbs. Either you know or you don't. Either you saw it or you didn't.

And factive verbs find use precisely on the strength of their canonical cases. Seeing something you recognize in adequate lighting conditions and in a normal state of mind counts as knowing, counts as knowing if anything does, and all the more difficult and nuanced cases we deal with (I just caught a glimpse of it, it was new moon and very dark, I had been crying, etc. etc.) are dealt with holding just this sort of case as the standard.

And counterfactual descriptions of knowledge seem to circle around the same idea. "If any of you had been there and seen what I saw, you'd be saying 'wolf' too." (And further afield: if we would judge, of anyone in such a situation, in any accessible world, that they would know there was a wolf, then you, in this situation, in the actual world, are properly said to know there's a wolf.)

I'm going to keep mulling over this "situation" business. I've always wanted to say that a key element of knowing is being in a position to know, despite the evident circularity. I might find a way to make that do some work.

Someone who always thinks it's 3 o'clock will be right twice a day, but we couldn't say that they know it. — Srap Tasmaner

I'm going to keep mulling over this "situation" business. I've always wanted to say that a key element of knowing is being in a position to know, despite the evident circularity. I might find a way to make that do some work. — Srap Tasmaner

I think that's an interesting example that gets at something quite basic about epistemic certainty being required for knowledge. If you add the condition that the person who always thinks it's 3 o clock also has a disorder which makes them see the display as 15:00, they would have the psychological certainty. they'd have a paradigm example of epistemic connection to the knowledge through their own perception (if an arbitrary person saw that, they would know), but we'd still want to reject that they know that it's 3 o-clock when it's 3 o-clock because their connection to reality itself is also evincing the claim that it's 3 o-clock by accident.

One of the issues this may create for an account regarding communal standards of epistemic certainty is that we'd still need to be able to preclude instances like the above, in which certain people would be excluded from counting as knowing things due to the facts about them precluding an appropriate access to reality (in some collection of scenarios).

You also couldn't establish the reason that their knowledge isn't knowledge if you happened to assess them at 3 o'clock using the same clock as them without other knowledge, since a non-disordered observer would be able to make claims of the same standard about the perceptual event regarding the clock. They'd both see it, they'd both be certain, the only difference is their position in the broader web of knowledge having norms.

@Andrew M

My usual rambling reply. Apologies.

we'd still need to be able to preclude instances like the above, in which certain people would be excluded from counting as knowing things due to the facts about them precluding an appropriate access to reality (in some collection of scenarios). — fdrake

I think that's pretty clear, and why I included "in a normal state of mind" as part of the canonical situation. Normal for who? Normal for me, with my strange disorder? More like, normal for us, really.

There are a couple of things I'm after here: one is thinking of knowledge in a rather old-fashioned way, the way we use the concept ordinarily, and the way we still teach young people, namely, that it is distinct from opinion and from guessing. I don't want to lean too hard on the way English happens to distribute its verbs, but everyday usage lumps "belief" in with opinion and friends; if knowledge shouldn't be in there, then knowledge is not some sort of super-belief, isn't at the top of an axis marked "confidence," isn't on that axis at all. And that dovetails also with the cases mentioned above, that it may very well be clearer to someone else whether I know something or not than it is to me. (See Robert Burns.) We can know that we know something, and we can have beliefs about our knowing, and these don't attach automatically to instances of knowing, so we're not automatically in a privileged position compared to others with respect to whether we know or not.

Forgetting and remembering have all kinds of cases. I can know someone's birthday and not remember it until really pushed by someone else; but I can also have really, completely, irrevocably forgotten something that someone else wants to insist I know because they learned it from me. (I expect that will become a more and more frequent experience for me, alas.)

Coming back to your issue, fdrake, I tried to gesture at another of those ordinary ideas about knowledge with my first counterfactual: if any of you had seen what I saw, then you'd know what I know. We want cases of knowledge not to be cases of gnosis! It should be accessible to almost anyone, and the knowledge I acquired, by virtue of the position I was in, is just the knowledge almost anyone would acquire. (This is a vaguely science related notion, that it shouldn't matter who makes the observation, that you can freely substitute one observer for another, so long as they follow the same procedures, that sort of thing.) But we have to say "almost" because there are things you have to know, and maybe things you have to be able to do, to acquire certain sorts of knowledge even in ideal conditions. If you don't already know what a wolf is, you won't know when you've seen one. And here again, you might report that there's something up in the hills, looked like a dog but different somehow, and someone more knowledgeable could correct that to "You saw a wolf, your first."

So it turns out the canonical "situation" is not just the environment but involves quite about you, whether you have the capacity to acquire the knowledge available, whether you are receptive to it, and so on. Whether you were paying attention -- that one matters quite a bit. All of that goes into what we can't help but keep calling "being connected to reality the right way" to acquire knowledge. Or we could say that there are ways of interacting with your environment that are knowing ways and ways that aren't. Conducting surveillance is putting yourself in a position to know, and conducting experiments is creating situations where you can be in a position to know. Some of the difficulty of carrying off the acquisition of knowledge is not knowing enough to design those situations; you have only your current capacity to rely on in making the design, and if that's inadequate you might get an interesting result but not know it (the CMB story), or you might force the results to conform to your pre-existing knowledge, misinterpreting rather than simply missing the novelty.

Getting pretty far afield. I just want to capture the sense of saying something like, "Dad knows where the Easter eggs are, because he's the one who hid them," and that sense holds even if Dad forgets where he put a few of them. Similarly, "I know there's one in the flowers because I saw him put it there." These are cases of knowledge if anything is. They give the word "know" meaning. They are the sorts of cases you reach for to say what's missing when someone only has an opinion or an educated guess or a belief or a hunch about where the Easter eggs are. If Dad says, "I think I put one on the mantelpiece -- no, wait, I remember I was afraid it would roll off," that's less like switching from one belief to another and more like switching from one kind of state -- believing, opining, guessing -- to another kind altogether, knowing. You can still be wrong about whether you know, but the state you want to be in is not just a state of having different and better beliefs. That's the idea.

So it turns out the canonical "situation" is not just the environment but involves quite about you, whether you have the capacity to acquire the knowledge available, whether you are receptive to it, and so on. Whether you were paying attention -- that one matters quite a bit. All of that goes into what we can't help but keep calling "being connected to reality the right way" to acquire knowledge. Or we could say that there are ways of interacting with your environment that are knowing ways and ways that aren't. Conducting surveillance is putting yourself in a position to know, and conducting experiments is creating situations where you can be in a position to know. Some of the difficulty of carrying off the acquisition of knowledge is not knowing enough to design those situations; you have only your current capacity to rely on in making the design, and if that's inadequate you might get an interesting result but not know it (the CMB story), or you might force the results to conform to your pre-existing knowledge, misinterpreting rather than simply missing the novelty. — Srap Tasmaner

I think that's quite perceptive. The only contributions I have here are muddying the waters further and joining you in rambling. I agree with you that it's the case that a 'canonical situation' is required. As a tentative estimate of what a canonical situation is, I think it's an implicitly known set of context appropriate rules for generating a specific type of knowledge when it is engaged. If someone was in an such a relation to some knowledge item, they can be deemed to have known it. It is like participation in a rite.

I also think you're right with the claim that engendering a situation where something can come to be known (and what might block that situation) is a skill in itself. A kind of epistemology of contextual knowledge production, rather than one of linking statements to conditions of satisfaction through idealised argument. I might be able to design those situations for code - like finding a bug in a system I know well enough - but not others - like finding confounding variables in an experiment in neurology. I can perform some rites but not others.

In some respect, though, it doesn't matter that I can do those things, what would make the produced facts, claims, knowledge etc seems also to need to be generic and or/generalisable. You can't 'just know', even if you really truly know. The working needs to be able to be shown. I think we often take on trust that the working could be shown if needed. Like if I tell you I went to Lidl today, you'd probably take that as a fact and not even wonder if in fact I'd gone to Aldi, or even think about how I knew it was a Lidl. You can doubt whether I have performed the rite or not, but since most rites are taken on trust, you will take it on trust that I've performed the rite.

In that regard, it seems 'the collective' becomes acquainted and takes as a given a collection of rites of knowledge production, which trigger in certain contexts. You'd have no reason to doubt that I went to Lidl instead of Aldi today, unless I told you another time that I get the two confused all the time - and in that manner you'd be able to sensibly doubt that I was following the right rule of knowledge production to know I went to Lidl. I may have forgotten that I went to Aldi instead of Lidl, but I could sensibly be deemed to have known it under the trust that I had the capability to perform the appropriate rite.

Do you think something similar is going on with your dad hiding the eggs scenario? Insofar as dad performed a rite (placing the eggs) that makes him deem-able to know where the eggs are. Even if you ask him later and he forgot.

When someone justifies a knowledge claim like "I know where they are because I saw him put them there", what makes the "because" function as a justification is ultimately a (trusted) appeal to a connection between location knowledge of objects and sight of the person, which is a common rite of knowledge production regarding the location of objects.

When a person seeks to know something, I think you're right that they'll try to enter 'a state' in which they can come to know it. I think the production of that state is the successful performance of an implicitly sanctioned rite.

As an aside, I don't think the rites themselves can be true or false, only more or less accurate, more or less fit for task. It isn't like "I know where they are because I saw him put them there" has an easy parsing in terms of logic In that regard, the connection to reality which is ascribed to genuinely productive states of knowledge is effectively a sanctioning of that rite through (again publicly deemed) sufficient accuracy/reliability/fitness for task.

I don't think the rites themselves can be true or false, only more or less accurate, more or less fit for task — fdrake

Well that's the thing. Some of what you say in your post has the feel of the "rites" (clever choice, that) underwriting knowledge production -- a bit like what Austin says about how only in specific circumstances does saying "I name this ship the Queen Mary III" make it so that the ship is now named "Queen Mary III."

But it's evident that we can judge whether a given candidate for a rite is knowledge producing. "You don't find out how many we have in the store by checking the receiving logs; you have to go and count them." What's going on there? I could claim that we are relying on a pre-existing understanding of knowledge to judge whether a rite works -- but it also looks like I'm proposing an alternative rite already known to work.

There's circularity here that leads to a bootstrapping problem. I have to know what knowledge is to know whether a rite candidate works; but all I have for an understanding of knowledge is pointing to rites known to produce it. How could I ever get from not having a rite that produces knowledge to having at least one I can use for reference? If I don't know what knowledge is, how can I possibly find out?

That bootstrapping problem infects every attempt at "explaining" knowledge -- for instance, if we take the talk of rites here as an explanation.

You can't 'just know', even if you really truly know. The working needs to be able to be shown. — fdrake

Now that's a biggie. For something to be a rite, we must be able to set out the steps in detail and teach those steps to the novitiate.

Is it true? It's at least true that if you follow the steps then you will acquire knowledge. But do you know because you followed the steps? Do the steps constitute knowledge acquisition? Is there maybe one step where we say, "Here, here's where the knowledge comes in"? Again, I think any such claims will be circular. How could you possibly come to know such a thing? So whatever the status of these rites, I don't think they can be an account or an explanation of knowledge.

One thing I think I'm resisting here is the suggestion (derived from Sellars) that "I know ..." is not really a factual claim at all, but an offer to defend or to justify my claim, to enter the space of reasons. In "I know X because Y," I'm not taking Y as being my justification or my warrant for claiming that X. I'm thinking of X and Y as being more intimately related than that. If I lack one justification, I might have another. You can swap out Y's. Reasons are things you can "come up with". The Y I'm interested in is not something like the basis for an inference, but more like an explication of what sense in which I'm using the word "know".

So the sense in which the steps of the rite must be capable of being made explicit, that could be that you must be able to say in what sense you meant the word "know". (Is there really more than one sense? Need to come back to that.) And since we do also make inferences based on evidence, can we tell the difference between distinguishing senses of "know" and offering justifications? "It was crowded and I didn't get a good look at him, but I heard him laugh and I'd know that laugh anywhere. He was there alright." Here's where I would start: one of the absolutely central elements of a knowledge claim like this is "I was there."

"I was there" is powerful. Imagine a vet listening to some guys at a bar, talking big about what we should have done in Vietnam or in Afghanistan. "You don't know what you're talking about," he says. "And you do? You some kind of expert?" "I was there." End of debate.

But again (and this is also, I understand, a key point for Williamson) knowing doesn't automatically mean you know that you know. (Knowing is not "luminous.") You can think you know, because you were there, but you weren't paying attention at the crucial moment, or you didn't recognize the significance of what you were seeing, and so on. We need there to be something definitive in the canonical situation, something automatic, but there are so many ways to fall short of that we have to be sensitive to.

One last bit on justification and "just knowing" without reasons. If we start from some position, with knowledge of some facts, say, and reason from there to something else we are prepared to count as knowledge, something we intend to rely upon, that's a bit like a "save point" in a video game. Calling it knowledge means precisely that you don't have to go back before that, and you can even jettison the reasons you relied on and just keep the conclusion. Knowledge of this sort is detachable from the reasons supporting it. When questioned, you have to check to see if you kept the original reasons; if you did, you have to reconstruct the inference, and if you didn't then you have to reconstruct the whole thing. Maybe it'll turn out your reasons weren't solid, or your inference was faulty. That happens. But in treating, let's just say it, such a belief as knowledge, you're in a way committed to not needing reasons for it anymore. It's a new save point you can treat as as-far-back-as-I-need-to-go.

And that could be one of those cases where we're reaching for a word, "knowledge", because the application here would have some structural similarity to its use elsewhere, even though the cases are actually different. In the situation where I know it was a wolf because I saw it, we are not making an inference and so there's no need to talk of reasons; in the case where we have made what we believe is a successful inference, we no longer seem to need the reasons (we have our save point) and so we call that "knowledge." But they're not really the same sort of thing at all. Knowledge has this strange dual nature, that it can be what you are most able or least able to defend, most willing or least willing to support with reasons.

I don't think it's a matter of doubt, just a matter of admitting fallibility. I would say that I know that my housemate is a bachelor, but I also accept that he could be lying to me and have a secret wife that he ran away from. Implausible, perhaps, but not unheard of. Does admitting of this possibility (and not just in the "there is a possible world" sense) somehow entail that I don't know that my housemate is a bachelor (assuming he isn't lying to me)? I don't think so. That I might be mistaken is simply an admission that I am not certain, not an admission of doubt.

So in such a scenario I would say that I know (and perhaps I do), but I'd also say that I might be wrong. Both claims are warranted. — Michael

Consider what it would take to be certain that your housemate was a bachelor. If it's never possible, then that's a Cartesian standard, not an ordinary standard. In everyday usage, "I might be wrong" qualifies the specific claim in an informative way - that there is some concrete reason why I don't want to fully endorse or commit to the claim. It's not a general claim of human fallibility - it's common knowledge that even the most careful investigations can sometimes lead to mistaken conclusions.

As SEP notes, there is generally "a reluctance to allow the contextually set standards for knowledge and certainty to diverge" (Williamson 2000, p. 254).

Here's the context for that quote:

One might fear that such arguments would prove too much. After all, something is wrong even with the assertion ‘A and I cannot be certain that A’. Does that not suggest that only something more than knowledge warrants assertion? What seems to be at work here is a reluctance to allow the contextually set standards for knowledge and certainty to diverge. Many people are not very happy to say things like ‘She knew that A, but she could not be certain that A’. However, we can to some extent effect such a separation, and then assertibility goes with knowledge, not with the highest possible standards of certainty. For example, one may have warrant to assert ‘A and by Descartes's standards I cannot be absolutely certain that A’, where the reference to Descartes holds those standards apart from the present context. Again, it would often be inappropriate to respond to the assertion ‘A’ by asking ‘How can you be so certain that A?’. The word ‘so’ flags the invocation of unusually high standards of certainty. By ordinary standards you may have had warrant to assert that A even if you could not be so certain that A. — Knowledge and Its Limits, p. 254 - Timothy Williamson

Consider what it would take to be certain that your housemate was a bachelor. If it's never possible, then that's a Cartesian standard, not an ordinary standard. — Andrew M

I don't think it even needs to reach the "Cartesian" standard. It's really just the same point you made earlier (which I missed):

I agree. Alice can know the phone number qua a ten-digit number. But if when asked she says, "I think it's <number>", then that raises a question as to whether she really does know it. If she gets it right, we're probably inclined to say she did know it after all. However, given her qualification, she wasn't certain that she knew it, and thus not certain what the number was.

So in that case we could say that she didn't know that she knew it. — Andrew M

In the case of my housemate being a bachelor, I just have a greater conviction.

I do want to say though that I think there's something a little funny going on in imagining judging a sort of canonical case of knowing. (Of the "Well I seen it, didn't I!" variety.) — Srap Tasmaner

For sure. With the wolf example, I was distinguishing between the scenarios of the boy hearing a rustle in the bushes versus seeing the wolf. If the boy reports the latter then, yes, we should be satisfied that he knows it.

In line with what you're saying, I'd just add that Gilbert Ryle called terms like "see" achievement verbs. To see a wolf entails that there is a wolf there. (Though, of course, one could think they had seen a wolf, but be mistaken.)

I don't think it even needs to reach the "Cartesian" standard. — Michael

:up:

But it's evident that we can judge whether a given candidate for a rite is knowledge producing. "You don't find out how many we have in the store by checking the receiving logs; you have to go and count them." What's going on there? I could claim that we are relying on a pre-existing understanding of knowledge to judge whether a rite works -- but it also looks like I'm proposing an alternative rite already known to work. — Srap Tasmaner

I particularly like the bolded bit. If I understand what you're saying right, we clearly have a somewhat generic ability to assess whether a given rite is appropriate for producing a given item of knowledge. You gotta count to get the number, you don't check the acquired stock.

You gotta follow the bug back from what's producing it, not start somewhere random. You check if the shirt button is sufficiently fastened by inspecting if it's through the button hole, not by pulling your shirt down.

Each of those instances had a failure mode, a break in connection of the proposed conduct with the goal's item. I don't think it's possible to specify generic content of a failure mode that breaks the connection, but it seems to be possible to say that the connection of the rite from the desired knowledge item is severed. In that regard perhaps a rite fails when it is sufficiently severed from its desired knowledge item.

Now why I particularly liked the bolded bit is because it seems like you've caught a productive ambiguity in how this works. In retrospect it appears the rites are given, like they're an a-priori, but they also seem to be modified and passed on in the act of examining the rite. I'd never thought about the example of counting items in a store vs checking just the receiving logs before, but when I read it it's clearly a successful rite. The pre-existing understanding seems to be what semantic and epistemic resources an individual (or collective) can draw on while either creating or enacting a rite; the prior context of interpretation. But when using the prior context for interpretation, it is difficult to tell what is prior context and what is created synthetically in the act. A kind of alchemy of a particular event into an instance of, or failure to satisfy, a candidate practice for producing knowledge.

One thing I think I'm resisting here is the suggestion (derived from Sellars) that "I know ..." is not really a factual claim at all, but an offer to defend or to justify my claim, to enter the space of reasons. In "I know X because Y," I'm not taking Y as being my justification or my warrant for claiming that X. I'm thinking of X and Y as being more intimately related than that. If I lack one justification, I might have another. You can swap out Y's. Reasons are things you can "come up with". The Y I'm interested in is not something like the basis for an inference, but more like an explication of what sense in which I'm using the word "know". — Srap Tasmaner

I think I agree with that, and I want to 'yes, and' it. The "I know" when someone says "I know" in a context of justification isn't just an epistemic move in the game; it's staking yourself on the prior epistemic moves. Manoeuvring yourself into a position where you expect to be deemed to know.

I am willing to bet that the expectation there falls on the trust of a previously enacted rite; that it was conducted appropriately and an appropriate rite. But not explicated in those terms, it's taken on trust. I think this is an intimately related phenomenon to the one you reference here:

Knowledge of this sort is detachable from the reasons supporting it. When questioned, you have to check to see if you kept the original reasons; if you did, you have to reconstruct the inference, and if you didn't then you have to reconstruct the whole thing. Maybe it'll turn out your reasons weren't solid, or your inference was faulty. That happens. But in treating, let's just say it, such a belief as knowledge, you're in a way committed to not needing reasons for it anymore. It's a new save point you can treat as as-far-back-as-I-need-to-go. — Srap Tasmaner

Being able to declare a position (analogous to "I expect to be deemed to know (in virtue of this or that rite)") I think is also to declare that the position itself is valid, because its constitutive steps have been ensured by (expecting yourself to be) following a trusted rite.

That trust seems to be irreducibly social - insofar as it pertains to common access to a shared environment. but also irreducibly entity focussed - because the entities are shared in the that environment. It is simultaneously an elevation of environmental patterns into socialised principles of associations and a matching of socialised principles to patterns; for following a pattern to success or failure in a rite.

We have at least a couple different threads here now.

I'm tempted to respond directly to your proposed model here (especially because some of it is close to some things I nearly wrote about earlier), pitch in on refining and extending it, all that, but I'm going to resist that temptation for now.

One thing that bothers me a little is that the model seems very broadly applicable, which may be a strength, but means we might be missing something specific to knowledge. I think I could read most of what you wrote as applying to, say, rational belief. (And possibly to a great number of other things, ethical questions and so on.) Would you say there's a point in here that is specific to knowledge?

One thing I've been trying to capture is that there's something a little arbitrary about knowledge. If I know because I was there, even by chance, and you don't because you weren't, that's just the way it is. If I happened to look up and see the balloon before it went behind the trees, I know there was a balloon and you can only take my word for it or not, even if you were walking along beside me.

I keep reaching for these examples that are clearly matters of perception -- which is suspicious -- but there are also the examples related to remembering, so I don't think I'm actually talking about perception rather than knowledge.

What about that balloon? Could I provide reasons for you to believe it was there? Maybe, maybe not. You didn't even notice me looking up, but if there was a balloon, we could try to find it again. Even if we find a balloon, that won't prove I saw it, might not even be the same one, but that might give you grounds for believing me. And that might be a mistake -- could be I made it up and the universe provided a spurious balloon to join in the fun of having you on. But I just know one way or the other, at least about whether I saw it, and you can only conceivably have reasons to believe. We are up to two completely different things.

Do you see your model as talking about this, or as comfortably absorbing this as a particular case?

In line with what you're saying, I'd just add that Gilbert Ryle called terms like "see" achievement verbs. To see a wolf entails that there is a wolf there. (Though, of course, one could think they had seen a wolf, but be mistaken.) — Andrew M

Yes, absolutely. I had Ryle in mind when I said "know" and "see" are both being used "factively". Point of interest that we might be disinclined to treat the identical construction, "I heard a wolf in the bushes" as factive. That is at least intended as a factive use, but we the audience are reserving some doubt that, say, wolves and foxes rustle bushes so distinctively. (But maybe to an experienced woodsman ...) For a more attractively factive use of "hear", consider how different is the sound of a bird searching for bugs among the dry leaves under a bush from the sound of a chipmunk just passing through. Or, for that matter, "I heard an owl up in the big oak tree last night."

One thing that bothers me a little is that the model seems very broadly applicable, which may be a strength, but means we might be missing something specific to knowledge. I think I could read most of what you wrote as applying to, say, rational belief. (And possibly to a great number of other things, ethical questions and so on.) Would you say there's a point in here that is specific to knowledge? — Srap Tasmaner

I don't know if there's anything specific to knowledge in it. I also don't know if there's anything which distinguishes knowledge from (placehlder) produced by/enacted in a 'successful connection to the world' - because we use 'to know' as a summary of perception, as well as for synthesis of perceptions and judgements (counting), as well as for practical activities (know how to ride a bike).

Maybe something that would make it more specific to knowledge would focus on objectivity (like this thread does here), what is it that makes social practices generalisable, knowledge-productive and binding (deemed to know if enacted)?

One thing I've been trying to capture is that there's something a little arbitrary about knowledge. If I know because I was there, even by chance, and you don't because you weren't, that's just the way it is. If I happened to look up and see the balloon before it went behind the trees, I know there was a balloon and you can only take my word for it or not, even if you were walking along beside me. — Srap Tasmaner

There's something almost arbitrary about a process of knowledge production, yeah. It could be a case of 'garbage in, garbage out'. If you can declare a prior reality as fixed by taking something as knowledge, taking a position in the space of reasons - a 'save game' as you helpfully put it -, if there are errors in the save game, treating them as knowledge has the capacity to make them part of social reality. It can be that a factual error propagates and itself gets ritualised, or a faulty connection to the state of things is sanctified.

I've been envious of people who are prone to certainty in their convictions and actions for a while, to me it seems they have the ability to conjure social reality around them; no matter how distorted it is! Maybe this is the same thing.

I honestly should have said "contingent" where I said "arbitrary" -- but I liked how forceful "arbitrary" sounds, and it captured the peculiar way in which each of us is just blocked from knowing certain things. If you weren't there, you weren't; you can have all the rational beliefs about it you like, but you'll never know. That's just contingency, but it feels arbitrary.

Here's something for you to talk about: what I postponed in looking at your model was not so much the "rites" stuff, but the propagation mechanism, having been deemed to have performed a rite, trust, staking a position, aiming to be deemed, all that business. That's nice, but applies to lots of things, it seems to me. More than just to the performance of rites. Did I miss something there?

As for "rites": rites sound like the sort of thing one would generally perform knowingly and deliberately. That's one way of putting the circularity problem. But, as above, another thing I want to say is that you know a great many things accidentally, so to speak, not intentionally and deliberately. I rely every day on things I happened to have learned that I didn't set out to learn. I didn't look up in order to see the balloon that I didn't know was there until I looked up; I just happened to look up and there it was, and now I have knowledge that you don't. From the perspective of a third party, Bill, I could be said to have taken an action that resulted in knowledge, and if Bill's definition of "rite" is loose enough, then Bill can say I performed a rite. Roughly, I did the same thing and achieved the same result as someone who performed that rite knowingly and intentionally. But the fact remains, I could not have performed the rite intentionally, since I lacked the knowledge required to do so. (And we've circled back to circularity.)

Seem like we also ought to say something about rites as general and specific. If I perform the rite of looking in a box to find out what's in it, there's the general, as it were, "rite schema" and then there's the specific rite, a token of that type, I will perform with this specific box. This "application" step might be a place where we point to additional knowledge required to perform the rite (knowing it, that it applies, how to apply it, etc.), but it's also clearly an area for experimentation. If you wonder what's in there, you might reason, "It looks somewhat like a box; let's see if we can open it (like a box)." That step of analogizing, or of extending the known range of applicability of a rule, is obviously terribly fruitful, and one way to generate unplanned-for, unexpected knowledge, without requiring at least certain sorts of knowledge up front. You don't have to know that the rite will work, or even know that it applies, to try it out. So there's experimentation again, and maybe not quite a solution to bootstrapping, but a clear path for expanding your knowledge base. And that path is deliberate, and intentional, without any knowledge of what the result will be.

I was watching Sean Carroll's monthly podcast 'ask me anything september' and he was answering a question about 'Constructor theory.' I had not heard of it and looked it up on wiki:
As I read it, this thread came to mind and I wondered if constructor theory could be used to assess the plausibility of an omniscient system by means of a possible or impossible 'task' as this concept is employed in constructor theory. I have copied and pasted the couple of paragraphs below from: https://en.wikipedia.org/wiki/Constructor_theory which made me think of this thread:

The fundamental elements of the theory are tasks: the abstract specifications of transformations as input–output pairs of attributes. A task is impossible if there is a law of physics that forbids its being performed with arbitrarily high accuracy, and possible otherwise. When it is possible, a constructor for it can be built, again with arbitrary accuracy and reliability. A constructor is an entity that can cause the task to occur while retaining the ability to cause it again. Examples of constructors include a heat engine (a thermodynamic constructor), a catalyst (a chemical constructor) or a computer program controlling an automated factory (an example of a programmable constructor).
The theory was developed by physicists David Deutsch and Chiara Marletto. It draws together ideas from diverse areas, including thermodynamics, statistical mechanics, information theory, and quantum computation.
Quantum mechanics and all other physical theories are claimed to be subsidiary theories, and quantum information becomes a special case of superinformation.
Chiara Marletto's constructor theory of life builds on constructor theory.

and

According to Deutsch, current theories of physics, based on quantum mechanics, do not adequately explain why some transformations between states of being are possible and some are not. For example, a drop of dye can dissolve in water, but thermodynamics shows that the reverse transformation, of the dye clumping back together, is effectively impossible. We do not know at a quantum level why this should be so. Constructor theory provides an explanatory framework built on the transformations themselves, rather than the components.

Information has the property that a given statement might have said something else, and one of these alternatives would not be true. The untrue alternative is said to be "counterfactual". Conventional physical theories do not model such counterfactuals. However, the link between information and such physical ideas as the entropy in a thermodynamic system is so strong that they are sometimes identified. For example, the area of a black hole's event horizon is a measure both of the hole's entropy and of the information that it contains, as per the Bekenstein bound. Constructor theory is an attempt to bridge this gap, providing a physical model that can express counterfactuals, thus allowing the laws of information and computation to be viewed as laws of physics.

Do you think this has any relevance to your OP? I admit my thought that constructor theory may be a way to assess the plausibility of an omniscient system is based on very tenuous linking on my part as I am basing it on a loose reading of a theory I had not heard of before today.