An analysis of knowledge is either an analysis of what the word 'knowledge' means - how we use the word, or an analysis of what the word ought to mean - how it would make most sense in some particular context, to use it.

In neither case must it mean 'true belief justified by true premises'.

In the former case you can see from everyday use we simply don't use it that way because we don't stop to check either beliefs or premises are true before we refer to those beliefs as knowledge. In every day use, knowledge is most often simply a category of belief we have a high confidence in - "I know my keys are around here somewhere!"

In the latter case, you can't analytically separate the two interpretations. They are underdetermined, the analysis supports either. There's nothing conceptually wrong with either the model that knowledge is justified true belief (but has Gettier-like exceptions), or that knowledge is justified true belief with true premises. It's not a problem solved, it's an alternative proposal submitted to our preferences. In supporting it, you'd have to explain its usefulness over the former. In what cases might the latter have some advantage.

The most obvious problem I can see with the latter model is that you'd never be able to specify all the premises. In most Gettier-like problems, there's some premise which is false, but it is often a hidden premise (the clock is actually working). Think of all the hidden premises we use (gravity pulls me down, events have a cause, the external world is real...). If all knowledge is one very long Ramsey sentence, your definition would have to see all the premises as true. A robust definition, sure, but a useless one as very little would ever be categorised as such.

I think "knowledge" cannot rest on false information, it should always result in true to be knowledge.

according to wikipedia X should be true:

A subject S knows that a proposition P is true if and only if:
P is true, and
S believes that P is true, and
S is justified in believing that P is true

your example is an example of stale information rather than knowledge.

I've always found Gettier "problems" to be non-issues. Take the following classic example.

You're in a field. You see something that looks like a cow (it's actually a bed sheet waving in the wind). You go there's a cow in the field. Now, it so happens that there is a real cow in the field.

According to Gettier you got it "right" (there is a cow in the field) by fluke (fake knowledge).

However, before we get our knickers in a twist trying to find a solution, let's just make sure that we have a problem to begin with.

The bed sheet is the cow to you. See?

An analysis of knowledge is either an analysis of what the word 'knowledge' means - how we use the word, or an analysis of what the word ought to mean - how it would make most sense in some particular context, to use it. — Isaac

"Knowledge" and so many other words are like "obscenity" in one of the trials in the 60's: The judge couldn't define it, but he famously "knows it when (he) sees it".

We make these distinctions easily enough, without knowing how we do it. What, if anything, is the underlying logic? This is the task of philosophy as I see it, in answering these "what is" questions.

In every day use, knowledge is most often simply a category of belief we have a high confidence in — Isaac

But this does not cut it, even by the standards of every day use. Sure, if you have a strong conviction, you might claim to know something. But if you had said, "I know my keys are around here somewhere", I can ask, "In retrospect, did you really know it?"

• If in fact the keys were in the car, you did not know it.
• If you knew it because you are a Pisces, you did not know it, even if they were around here somewhere, and you are in fact a Pisces.
• If you knew it because you remember leaving them on a table, when in fact that memory was from yesterday, but they did fall out of your pocket here anyway, you did not know it.

We make these intuitive judgements independently of how strongly you happened to hold the conviction that your keys were around here somewhere.

What, if anything, is the underlying logic? This is the task of philosophy as I see it, in answering these "what is" questions. — hypericin

What if there isn't an 'underlying logic'? I mean there's no intrinsic reason why there need be. what if 'know' as in "I know my keys are around here somewhere!", is different in meaning to 'know' as in "she knew where her keys were". The former expressing a confidence on one's belief, the latter expressing a relationship between a third person's beliefs about the world and our own (she believed the keys were on the table, but I can see them here in the car).

you might claim to know something. But if you had said, "I know my keys are around here somewhere", I can ask, "In retrospect, did you really know it?" — hypericin

I have trouble with this idea that saying "I know" is merely a claim to know. I don't see how it distinguishes the utterance from any other. What else could it be? "that's a tree" is merely a claim that that's a tree - it doesn't change the meaning of the term 'is' to something other than the claim.

If we say that "I know" is a statement about one's mental state, then it has nothing to do with the actual state of affairs, right? So we say that "I know" is a claim about one's mental state and the world. it's meaning therefore has to relate both (something like my mental state matching the way the world is). This is the origin of the idea that "I believe my keys are on the table, and my keys actually are on the table, hence I know my keys are on the table" (I'm leaving out the convention of 'justification' or a minute).

The problem is, that seeing the keys on the table is just another justification for believing the keys are on the table, so all you have is a better justified belief. And if everyone you speak to agree that they keys are on the table, that's just even better justification for believing the keys are on the table.

So in your example...

If in fact the keys were in the car, you did not know it. — hypericin

...the 'in fact' bit can only ever mean that even better justifications exist for believing the keys are in the car (I've seen them there, my friend has too, I started the car with them...)

This list of 'even better' justifications becomes the Ramsey sentence I mentioned above, listing all the beliefs with their preceding justifications. The word 'know' would never be used if used according only to the principle of true facts with true premises.

When I claim "I know the pub is at the end of the road" I simply mean that if you walk to the end of the road, you will find the pub there. So if the pub I thought was there had been knocked down, but later replaced by another, I don't see a problem with saying that I 'knew' there was a pub at the end of the road, since, if you walk to the end of the road, you will, indeed, find a pub there.

Is there any merit in distinguishing between what we call knowledge in the quotidian world and what we call knowledge when we seek to make a claim about what we might call reality?

In ordinary life, epistemology is of little consequence - in picking a partner, choosing a home or selecting a car, working out what university degree to do, or which job to take, what shopping to buy - we do not worry about the problem of induction, or the correspondence theory of truth, or philosophy in general.

But as soon as we start talking about whether we believe in god/s or not, or what we ought to do as human beings, we generally enter into some form of argument with evidence and justifications. It's here that what we call knowledge becomes interesting. Is this a fair observation?

Is this a fair observation? — Tom Storm

Yes, that's something of what I was trying to get at distinguishing what 'to know' does mean and what it ought to mean. Things like justifications are just habits of thinking. They're often sprinkled liberally into analytical propositions, but what constitutes a 'justification' is no more than a gentleman's agreement among peers, an obligation. If I said "I know my keys are on the table because they're made of metal" I've provided a 'justification'. If my keys were in fact on the table, we could say I have justified true belief. But something's off, my justification didn't make any sense. So why not? Well there's no link between being made of metal and being on the table. So we bring in all this background 'knowledge' into what counts as a justification - hence my invocation of Ramsey sentences.

I think there's considerably more than two such uses as well. Consider "you're never going to make it to the meeting you're already late" - "I know that!", or "I just knew it!"

So yeah, I think you're right that there's a use case where "I know X" ought to mean something like 'if you act as if X is the case you'll get the expected results and I've done the generally agreed upon due diligence to make such a claim among my peers in this context'.

(You'll note I've not used the word 'true' as I don't see any need for the word 'true' in there at all, but that's another whole keg of worms)

What if there isn't an 'underlying logic'? I mean there's no intrinsic reason why there need be. what if 'know' as in "I know my keys are around here somewhere!", is different in meaning to 'know' as in "she knew where her keys were". — Isaac

I'm not sure that's true, but if this is the case, that's totally fine. I'm not committed to the meaning of words being rigid. Natural language is allowed to do that. Then, I'm only interested in the latter sense here.

The word 'know' would never be used if used according only to the principle of true facts with true premises. — Isaac

No no no, you are confusing truth condition with condition of use. The truth condition of "to know" is nontrivial and very debatable. But the condition of use is both variable between people, and might be as simple as a feeling of confidence that something is so. These are totally disjoint things. And this question of "what is knowledge?" is here asking about the truth condition, not about the conditions of use.

All these debates and claims we make on this forum are complex, with very complex truth conditions, if we were confined to making true claims only, we would be paralyzed, and say nothing. And even then, we could only limit ourselves to making claims we felt were true with absolute certitude. We would still be wrong 95% of the time.

When I claim "I know the pub is at the end of the road" I simply mean that if you walk to the end of the road, you will find the pub there. So if the pub I thought was there had been knocked down, but later replaced by another, I don't see a problem with saying that I 'knew' there was a pub at the end of the road, since, if you walk to the end of the road, you will, indeed, find a pub there. — Isaac

I disagree. "I know there is a pub is at the end of the road" is distinguishable from the statement of bare fact, "there is a pub at the end of the road". The truth condition of the first is not that of the second. The first adds additional constraints to the truth condition: "There is a pub at the end of the road, and additionally I stand in a knowing relationship with that fact".

Consider, we are in a city we haven't been to in 10 years. You say "I know there is pub at the end of the road." We go to the end of the road. There is a pub with signs of fresh construction, and a "grand opening" sign. You say, "I knew it!". This would be a joke. Because, while there is in fact a pub at the end of the road, you absolutely did not know it.

n ordinary life, epistemology is of little consequence - in picking a partner, choosing a home or selecting a car, working out what university degree to do, or which job to take, what shopping to buy - we do not worry about the problem of induction, or the correspondence theory of truth, or philosophy in general. — Tom Storm

I completely disagree. All of these life decisions are fraught with epistemological and philosophical considerations. It is what makes decisions so hard. If philosophy were a quaint exercise confined to certain abstract questions, it would be utterly uninteresting.

I completely disagree. — hypericin

Cool. We disagree.

It is what makes decisions so hard. — hypericin

I've never found such decisions hard at all.

If philosophy were a quaint exercise confined to certain abstract questions, it would be utterly uninteresting. — hypericin

From what I can see it mostly is. But I am not a philosopher.

Maybe you could take one of these questions, let's say, choosing a home, and demonstrate how philosophy would be applied to this task. :wink:

I've never found such decisions hard at all. — Tom Storm

Hmm, maybe it is the fact that I have always been philosophically inclined that has made these kind of decisions nightmarishly hard for me!

"I like this home. At least I think I do. But do I really know that? What if it is a passing whimsy? How do I distinguish my preference of the moment from a stable preference that will endure 10 years from now. I won't even be the same person by then! So how can I make this decision for a person I hardly know? Do I really even want a house? It is the largest purchase I have ever made, how do I justify spending the accumulated capitol of a lifetime on one? Is home ownership even my preference, or a socially normative one? Why do I really want one? How do I know there is not something drastically wrong with the house. There is an inspection, but is that sufficient evidence? How can it rule out every problem? Is the inspection not an instance of motivated reasoning? And it does not rule out horrible neighbors, a dog that barks at 3am, a wildfire that will destroy it in 5 years. The world is entering a phase of chaotic change, is it rational to tie oneself completely to one location in what might be a new and unpredictable epoch..."

These are the thoughts that will actually, frantically, go through my head, with my exasperated realtor wondering why I am passing on yet another perfectly good house. I have never bought one!

you are confusing truth condition with condition of use. The truth condition of "to know" is nontrivial and very debatable. But the condition of use is both variable between people, and might be as simple as a feeling of confidence that something is so. These are totally disjoint things — hypericin

I'm not 'confusing' them, I'm arguing that they amount to the same thing. That - "Is it true that 'I'm a woman'?" amounts to the same thing as - "have I used the term 'woman' and the grammatical construct 'I'm a...' correctly?"

Regretting my choice of example in today's climate, but pushing on... If I had male genetalia, it would not be true that "I'm a (biological) woman". It would not be true by virtue of the fact that the kind of thing I am is not the kind of thing we use the word 'woman' for. I'd have misused the word.

Same goes for 'know'. If I say "I know my keys are on the table", when in actual fact I haven't a clue, it's not true that "I know my keys are on the table", I've misused the word because the state of mind I have (in relation to the world) is not the sort of thing we use the term 'know' for.

But, using this analysis, "I know where my hat is", when used to describe a high degree of confidence in my belief about the whereabouts of my hat, is exactly the right use of the term, and so it is true that "I know where my hat is", because I used the term correctly. Even if my hat turns out not to be there. (Although, any reflective past tense use would not be true since we don't use the term in the past tense that way).

Consider, we are in a city we haven't been to in 10 years. You say "I know there is pub at the end of the road." We go to the end of the road. There is a pub with signs of fresh construction, and a "grand opening" sign. You say, "I knew it!". This would be a joke. Because, while there is in fact a pub at the end of the road, you absolutely did not know it. — hypericin

Ah, you've misunderstood my example (or I've been unclear). In your example, I couldn't possibly justify my statement because I'd never been to the city before. In my example I could justify my statement, and it also turned out, in hindsight, that I did know there was a pub at the end of the road. Justification is part of using the word 'know' correctly. Truth clearly cannot be.

In every day use, knowledge is most often simply a category of belief we have a high confidence in - "I know my keys are around here somewhere!" — Isaac

I agree with this. Even in more formal or consequential situations, such as engineering, it's true. One of the things we have to understand is the consequences of failure. Justification has to be good enough. Justification for knowing where my keys are is less stringent than that required to make sure the bridge doesn't fall down. The simplistic approach described in JTB doesn't reflect how people actually know things or how they should know things. Knowledge is adequately justified belief, whether or not it is true.

But this does not cut it, even by the standards of every day use. Sure, if you have a strong conviction, you might claim to know something. But if you had said, "I know my keys are around here somewhere", I can ask, "In retrospect, did you really know it?"

If in fact the keys were in the car, you did not know it.
If you knew it because you are a Pisces, you did not know it, even if they were around here somewhere, and you are in fact a Pisces.
If you knew it because you remember leaving them on a table, when in fact that memory was from yesterday, but they did fall out of your pocket here anyway, you did not know it. — hypericin

The only meaningful definition of "knowledge" is information adequate to support action. Knowledge doesn't just sit there doing nothing - that's information. Who cares if something is true until I have to make a decision? If I have information justified adequately to support action given the consequences of failure, then I have knowledge. If that's not true, then the word "knowledge" is useless.

The wiki article mentions an early response to this problem, which is also my response: you just have to amend "knowledge is justified true belief" with a condition which rules out false premises: say, "knowledge is true belief justified with true premises". — hypericin

I can't think of any Gettier Problem which survives this kind of analysis, or any other counterexample to:
Knowledge is true belief justified by true premises. — hypericin

Given that Gettier Problems were presented to show that the JTB definition of knowledge is insufficient, having to add a fourth condition to overcome them shows that the JTB definition of knowledge is insufficient.

Given that Gettier Problems were presented to show that the JTB definition of knowledge is insufficient, having to add a fourth condition to overcome them shows that the JTB definition of knowledge is insufficient. — Michael

Fair

The only meaningful definition of "knowledge" is information adequate to support action. — T Clark

Not sure if it's the only one. I tend to disagreements here...

In ordinary life, epistemology is of little consequence - in picking a partner, choosing a home or selecting a car, working out what university degree to do, or which job to take, what shopping to buy - we do not worry about the problem of induction, or the correspondence theory of truth, or philosophy in general. — Tom Storm

There are lots of times in the regular old everyday world when it's important that I know how I know something and how certain I am. If I'm going to dig a hole in my yard, it's important that I know if there are buried gas pipes in that location. If I'm going to paint the wall, I should know if the new paint is compatible with what's there now. As an engineer, if I'm going to dig up the contaminated soil on a property, I need to understand the source and quality of the information I'm using to decide where to dig.

I'm not 'confusing' them, I'm arguing that they amount to the same thing. — Isaac

You are absolutely confusing them.

But, using this analysis, "I know where my hat is", when used to describe a high degree of confidence in my belief about the whereabouts of my hat, is exactly the right use of the term, and so it is true that "I know where my hat is", because I used the term correctly. Even if my hat turns out not to be there. — Isaac

"I know where my hat is" is a perfect exemplar the verb "know". Nonetheless, if the hat is not there, it is an incorrect claim, no matter the degree of belief.

The ancient Greeks did not know the world was the center of the universe. They merely thought they knew, with complete and justified confidence. And they made this claim in absolutely perfect Greek.

Ah, you've misunderstood my example (or I've been unclear). In your example, I couldn't possibly justify my statement because I'd never been to the city before. — Isaac

No, you misread mine:

Consider, we are in a city we haven't been to in 10 years — hypericin

You have a foggy memory there is a pub at the end of the road. The memory was wrong. But by chance, a pub was built there in the last month. You were right that there is a pub at the end of the road. But you were wrong that you knew it.

But, using this analysis, "I know where my hat is", when used to describe a high degree of confidence in my belief about the whereabouts of my hat, is exactly the right use of the term, and so it is true that "I know where my hat is", because I used the term correctly. Even if my hat turns out not to be there. — Isaac

What do you mean by the right or correct use of the term? Do you mean appropriate? Because it can be appropriate to say something that is in fact false. So it may be that you're equivocating here. That it is appropriate to claim that you know where your hat is isn't necessarily that you know where your hat is.

Justification for knowing where my keys are is less stringent than that required to make sure the bridge doesn't fall down. — T Clark

That's a good point, what we understand by 'I know' varies even within claims about states of affairs.



If I wanted a lecture I'd visit the university. I came here for a discussion. If you can't even be bothered to justify your assertions, then there's no point continuing. Things are not the case simply because they seem that way to you.

What do you mean by the right or correct use of the term? Do you mean appropriate? — Michael

No, I mean 'correct' as in 'to be understood, to make sense'. No different to if I pointed to a tree and said "dog". I'd have just used the wrong word. " Tree" is the correct word.

I'm understood, if I say "I know where my keys are", to be very confident about my belief. I'm not understood to have verified the absolute truth about their location. As such, it seems reasonable to conclude I've used the term correctly, and I do indeed 'know' where my keys are.

The alternative seems really weird to me. That I say "I know where my keys are", I used all the terms correctly, but I don't actually know where my keys are.

No, I mean 'correct' as in 'to be understood, to make sense'. No different to if I pointed to a tree and said "dog". I'd have just used the wrong word. " Tree" is the correct word.

I'm understood, if I say "I know where my keys are", to be very confident about my belief. I'm not understood to have verified the absolute truth about their location. As such, it seems reasonable to conclude I've used the term correctly, and I do indeed 'know' where my keys are.

The alternative seems really weird to me. That I say "I know where my keys are", I used all the terms correctly, but I don't actually know where my keys are. — Isaac

I can understand you even if what you say is false, so this doesn't work either.

I can understand you even if what you say is false, so this doesn't work either. — Michael

It's not about your ability to understand. If I said "pass me the stapler" whilst pointing at the hole-punch, you'd understand me. I've still used the 'wrong' word, haven't I?

'Correct' here means more than being understood because you mentally make up for my error, it's trying to get at an ideal assuming you don't have to.

'Correct' here means more than being understood because you mentally make up for my error, it's trying to get at an ideal assuming you don't have to. — Isaac

So "correct" doesn't mean "understandable" and doesn't mean "appropriate". So what does it mean?

It's not about your ability to understand. — Isaac

But you just said "No, I mean 'correct' as in 'to be understood, to make sense'."

As I said, it's trying to reference an ideal. If I'm way off base here is there some other meaning you'd use for what a 'correct' word is?

If you hear a child refer to a car as "bus", you say "that's not the correct word", what is it you mean by that? Or do you not have the concept of the 'correct' use of a word?

As I said, it's trying to reference an ideal. — Isaac

I don't know what you mean by this.

If you hear a child refer to a car as "bus", you say "that's not the correct word", what is it you mean by that? — Isaac

That the thing they're referring to isn't a bus.

I don't know what you mean by this. — Michael

I mean a state we can imagine but which doesn't ever exist. Imagine a world where everyone was nice to each other all the time - an ideal. Here, I'm talking about a world where people don't cover other people's mistakes by guessing what they really mean.

That the thing they're referring to isn't a bus. — Michael

What's wrong with using the word 'bus' to refer to something that isn't a bus

What's wrong with using the word 'bus' to refer to something that isn't a bus — Isaac

If there isn't a bus and you say "there's a bus" then what you say is false. Or by "wrong" did you mean something other than "false"? Perhaps "inappropriate"? Like with your use of the term "correct" we're at a risk of equivocation here.

If there isn't a bus and you say "there's a bus" then what you say is false. — Michael

But what's wrong with saying something false?