A New Solution to the Problem of Easy Knowledge
I originally wrote this as a response to Chris's comments on my post "Easy Knowledge and Millianism". Something Chris said suggested this as a plausible solution to the puzzle. I thought it might be good to post this independently and see what everyone thinks. However, this post relies on some background information that can be found in my previous post and in the comments to that post.
This new solution is inspired by the Hawthorne/Stanley view that something is known only if it can be "used as a premise in practical reasoning". I am beginning to be attracted to something like this Hawthorne/Stanley position. I will try to spell out a bit of the view below.
Let's let JTB be a three place relation that bears between an individual, a proposition and a way or guise. We can say that someone, S, has a justified true (non-gettierized) belief that P just in case S stands in the JTB relation to a proposition under some guise or other. Further we can say that S knows that P just in case S stands in the JTB relation to P under some guise or other and that belief can be used under that guise as "a premise in practical reasoning". I know this last bit is a little bit vague. But I hope you'll still be able to follow me.
Now, we have gotten rid of the problem of easy knowledge. Ned the navigator does not justifiably believe the truth that that the oldest tree in Washington is at L under a guise that can be used in practical reasoning. (Although he does justifiably believe that truth). So, he does not know that the oldest tree in Washington is at L.
Moreover, we can still connect up knowledge-wh to knowledge-that by saying things like "S knows where x is just in case S knows an answer to the question 'where is x?'". Since we got rid of the problem of easy knowledge, we don't have a problem of easy knowledge-wh.
This solution matches our intuition that you can't have easy knowledge. It also matches with our intuitions about when a person knows-wh. It might also fit better with our ordinary use of the word 'knows'. Moreover, it does not invoke the weird acquaintance relation.
One thing that is slightly counter-intuitive is that even though Ned introduced the name 'L' by saying "let 'L' name the location of the oldest tree in Washington" and even though he can easily reason his way to the conclusion that the location of the oldest tree in Washington is at L, he does not know this. He has a justified true (non-gettier) belief because he JTBs that proposition under some guise or other. But, he doesn't know that proposition because he can't use that justified true belief as a premise in practical reasoning.
I kind of like this view. Does anyone see anything immediately wrong with it or does anyone have a suggested amendment?
This new solution is inspired by the Hawthorne/Stanley view that something is known only if it can be "used as a premise in practical reasoning". I am beginning to be attracted to something like this Hawthorne/Stanley position. I will try to spell out a bit of the view below.
Let's let JTB be a three place relation that bears between an individual, a proposition and a way or guise. We can say that someone, S, has a justified true (non-gettierized) belief that P just in case S stands in the JTB relation to a proposition under some guise or other. Further we can say that S knows that P just in case S stands in the JTB relation to P under some guise or other and that belief can be used under that guise as "a premise in practical reasoning". I know this last bit is a little bit vague. But I hope you'll still be able to follow me.
Now, we have gotten rid of the problem of easy knowledge. Ned the navigator does not justifiably believe the truth that that the oldest tree in Washington is at L under a guise that can be used in practical reasoning. (Although he does justifiably believe that truth). So, he does not know that the oldest tree in Washington is at L.
Moreover, we can still connect up knowledge-wh to knowledge-that by saying things like "S knows where x is just in case S knows an answer to the question 'where is x?'". Since we got rid of the problem of easy knowledge, we don't have a problem of easy knowledge-wh.
This solution matches our intuition that you can't have easy knowledge. It also matches with our intuitions about when a person knows-wh. It might also fit better with our ordinary use of the word 'knows'. Moreover, it does not invoke the weird acquaintance relation.
One thing that is slightly counter-intuitive is that even though Ned introduced the name 'L' by saying "let 'L' name the location of the oldest tree in Washington" and even though he can easily reason his way to the conclusion that the location of the oldest tree in Washington is at L, he does not know this. He has a justified true (non-gettier) belief because he JTBs that proposition under some guise or other. But, he doesn't know that proposition because he can't use that justified true belief as a premise in practical reasoning.
I kind of like this view. Does anyone see anything immediately wrong with it or does anyone have a suggested amendment?
posted by Joshua at 11:45 AM
9 Comments:
How won't this view have the same problems that we find with (e.g.) attempts to avoid Frege's puzzle by supposing that relational attitudes are 3-place and take modes of presentation as arguments?
David has a more intuitive argument against this sort of strategy, though I think the more technical problems it has are reason enough to abandon it. The intuitive idea is this: If knowledge has modes as arguments, then when I talk about someone's knowledge, I talk about modes. But modes are theoretical posits that are supposed to do a certain job, rather than deliverances of common sense. It's hard to get people to understand what they are, etc. So it's implausible to suppose that people go around talking about these sorts of things. It may be the case that there is this 3-place relation that one is in whenever one is in the 2-place relation. But then the reply to problems involving the 2-place relation are just familiar ways-millian-y replies rather than more radical Hawthorne/Stanley/Spencer ones.
Fwiw I know Andy has done some work resisting this argument of David's. And it seems that other Millians agree with Andy (I'm thinking of Michael Nelson, esp. in his SEP piece on propositional attitudes).
I can see some pull for thinking we talk of ways, but I think at best it's conversational and not logical. So I would still resist relativizing to ways even if we talk about them sometimes.
This comment has been removed by the author.
I wasn't thinking that knowledge would be a three place relation on this view. My thought was that there are certain propositional attitudes that correspond to basic three place relations. The three place relations obtain between individuals, propositions and guises. A person stands in the two place propositional attitude to a proposition when he stands in the three place relation to which the attitude corresponds under some guise or other. For simplicity, I'll just say that the two place relation obtains in virtue of the three place relation obtaining. (we don't need to worry right now about whether this is an appropriate use of 'in virtue of'.
So far, this is just what Salmon says about the belief relation. The new bit comes in when we combine the Salmonian position with traditional analyses of knowledge from epistemology. The pre-gettier epistemologist might have said that S knows that P iff S justifiably, truly believes that P. Now, we know that the belief relation obtains in virtue of the three place BELxyz relation obtaining between S and P and some guise or other. Similar things might be said about the attitude of being justified in believing. We might represent the three place relation that corresponds to justified belief as JUSxyz.
Now we have a few choices about how to rewrite the traditional view using the three place relations rather than the two place ones. We might way this:
S knows that P iff S JUS P under some guise or other and S BEL P under some guise or other and P is true.
Unfortunately, this probably won't do since S may be justified in believing that Superman can fly under the Superman-guise and yet believe that superman can fly under the clark kent guise and not under the superman guise. It doesn't seem that such a person is knows that clark kent can fly. his belief doesn't mach up with his justification in the right kind of way.
But we can solve this problem by simply saying this:
S knows that P iff there is some guise g such that S JUS P under g and S BEL P under g and P is true.
Something like this sounds right. Now, we know that this has to be modified a bit (because of gettier cases). My suggestion is that it should also be modified a bit because of easy knowledge cases. Let's say that S PRAC P under G means that S can use P in practical reasoning under G. Now we can write our analysis as follows:
S knows that P iff there is a guise g such that S JUS P under g and S BEL P under g and S PRAC P under g and P is true.
Now, we still need to spell out what PRAC is. But, intuitively, there is no guise under which Ned the navigator PRAC's his belief about the location of the oldest tree in Washington. If this is correct and the view above is true, then he doesn't know that the oldest tree in washington is at L. This seems to solve the problem of easy knowledge.
So, it is a familiar ways Millian-y reply but with a (sort of) Hawthorne/Stanley twist that helps to solve the problem of easy knowledge.
1) I don't know the Stanley/Hawthorne stuff very well. Can you say more about what it takes to PRAC a proposition under a guise? Do I PRAC a proposition, p, under a guise, g, just in case I can (properly? legitimately?) use p, when taken under g, as a premise in some bit of practical reasoning? I worry that that makes it too easy to PRAC a proposition under a guise. Suppose Ned wants to find L. If all that he knows about L is that the oldest tree in Washington is located there, why can't he use that claim as a premise in a bit of reasoning about how best to go about finding L?
2) As, I think, you point out late in your post, the view you're suggesting seems to entail that some sorta plausible closure principle is false. That seems sorta bad.
3) I'm inclined to think there are lots of cases in which I know some claim, but the way in which I know it considerably restricts the ways in which I can use the claim in practical reasoning. Suppose, for instance, that Greg (a reliable testifier) testifies to me that Yao Ming is a basketball player over seven feet tall. I think I can come to know that proposition thereby even if I've never heard of Yao Ming before Greg's testimony. But, knowledge of that proposition doesn't get me much in terms of practical reasoning.
4) Where easy knowledge is knowledge gained by the sort of language trickery in question, I don't have the intuition that you can't have easy knowledge. That, though, might be a result of my right and proper intuitions being perverted and beat down by you guys.
5) The Problem of Easy Knowledge, as you introduced it, is a problem about how it is we know these propositions when we're not acquainted properly with the things they're about. That doesn't strike me as a problem. But, as you note, even if we think we can have easy knowledge we might still wonder why it is that some propositions when known are easily used in practical reasoning while others are not. There's a similar sort of question about your view. Why is it some propositions are PRACed and others are not? A natural thing to say, I think, is that a proposition is PRACed (or not) in virtue of the guise under which the proposition is taken. But if that's right, the friend of easy knowledge can explain the difference between regular and easy knowledge in terms of guises as well.
1) I was thinking that someone PRACs P under guise g when he can (properly or legitimately) use P, when taken under g, in practical reasoning. Unfortunately, I think you might be right about the fact that someone can use:
P1: The oldest tree in Washington is at L
in practical reasoning even under the seemingly useless guise. Someone can do so when they want to find L and have other information about the oldest tree in Washington that might help them to find that location. I'm not sure what to say about this problem right now.
2) I was slightly worried about a denial of closure.
3) I like the Yao Ming case and I think you can strengthen it by combining this objection with your first objection and creating a kind of dilemma: (a) If the practical reasoning view is true, then someone can use the proposition that Yao Ming is the tallest basketball player in practical reasoning after merely having heard that this is true by testimony. but (b) if someone can use the proposition that Yao Ming is the tallest basketball player in practical reasoning after merely having heard that this is true, then a person can use the proposition that the tallest tree in Washington is at L in practical reasoning as well.
The point of the dilemma is to show that there is a tension, for the practical reasoning theorist, between his desire to hold on to knowledge by mere testimony and his desire to eliminate so called easy knowledge.
4) I too have trouble with the intuitions. I think I have a slight intuition that something is amiss, but that could be accounted for by simply noting that some bits of knowledge are more useful than others.
5) I think it is right that the friend of easy knowledge can adopt PRAC to help explain why some knowledge is useful and some not. but, if we have an initial intuition that so called easy knowledge is not knowledge, then we have a reason to take the PRAC approach (if it works) to distinguish between knowledge and non-knowledge rather than between useful knowledge and useless knowledge.
Alright avwake, I think I have the beginnings of a response to your first objection. But, it is still very rudimentary.
Consider Ned the navigator again. Here are two claims that someone might find plausible:
1w. Ned does not know where the tallest tree in Washington is.
2w. Ned does know what L is a location of.
Now, these are knowledge-wh claims. But, we use them to argue for a couple of knowledge that claims. If Ned knows that the tallest tree in Washington is at L, then he knows were the tallest tree in Washington is. But, given (1w) he does not know where the tallest tree is. So, it follows that:
1p. Ned does not know that the tallest tree in Washington is at L.
Moreover, let's suppose that (2w) is true because this is true:
2p. Ned knows that L is the location of the tallest tree in Washington.
Now, we can explain the intuition behind your first objection. Although it seems like Ned can use the claim that the tallest tree in Washington is at L in practical reasoning, actually he cannot. Rather, he can use the claim that L is the location of the tallest tree in Washington in practical reasoning. It is this latter claim that he uses when he is trying to find L and not the former claim.
Here is a rough claim. We can know facts about L by introducing the name 'L' in an appropriate way, but we can't come to know facts about other things by such an introduction.
I think this sounds plausible, but it seems to make spelling out PRAC more difficult and might exacerbate the Yao Ming problem.
Joshua,
A couple of questions:
(1) It seems like (a) should entail (b):
(a) L is the location of the tallest tree in Washington.
(b) The tallest tree in Washington is located at L.
You want to say that Ned might know (a), but be unable (given his evidence) to know (b), despite being aware that (a) entails (b)?
(2) Given that the resulting view allows us lots of easy knowledge, is it still supposed to be a solution to that problem?
(3) Given that, on this view, Ned can use (a) in practical reasoning, how do we answer the sorts of questions that (I thought) were motivating the view? That is, why can't Ned take me to the tallest tree in Washington given that he knows (a)?
(4) Why do you say that this view exacerbates the Yao Ming problem?
Andrew,
(1) yes
(2) yes, because there is far less easy knowledge on this view
(3) It has something to do with the fact that you can know facts about a location without knowing where the location is but you can't know the location of something without knowing where the location is.
(4) I don't remember why I said that.
Post a Comment
<< Home