belief

Originally posted by Agerg
This is an interesting discussion I've been reading (and I'm backtracking to this post) but...

Supposing a person identifies themselves as a fallibilist and works on the principle he knows P in isolation from Q, and knows Q in isolation from P, is he forced into having the same degree of "knowledge" when conjoining them? What if he also knows (in a fallibil ...[text shortened]... ))?

I'm having trouble seeing where the acceptance of contradiction is necessary here.

Let's suppose that fallibilism is true. Then, S can know P without being certain that P. So, suppose that S knows P and knows Q, that there is some nominal uncertainty about P and Q, and that neither proposition has any evidential relation to the other. Presumably, his confidence in the conjunction (P&Q) should be less than his confidence in either proposition taken in isolation. There will come a point where, although S knows each proposition in a set, he will not know the conjunction of the elements of that set. Essentially, as the set gets larger, it becomes more likely that (~P v ~Q v ~R....) [i.e., that at least one of the elements in the set is false]. If it seems weird to you that S can know a set of propositions and yet not know their conjunction, then you may feel the pull of the infallibilist position. Unfortunately, this also tends to push you towards either skepticism (we don't know a whole bunch of stuff), or being a phenomenalist (eschewing talk about the external world in favor of talking about the content of our conscious states, to which we putatively have direct introspective access and about which we're incorrigible). And then, essentially, you're back to the Cartesian Meditations.

Originally posted by bbarr
Let's suppose that fallibilism is true. Then, S can know P without being certain that P. So, suppose that S knows P and knows Q, that there is some nominal uncertainty about P and Q, and that neither proposition has any evidential relation to the other. Presumably, his confidence in the conjunction (P&Q) should be less than his confidence in either propositio ...[text shortened]... which we're incorrigible). And then, essentially, you're back to the Cartesian Meditations.

If you claim that knowledge can only be held for things that are infallible then you can't claim to know anything beyond "I think therefore I am" and any constructs you logically build within your mind, (ie you could claim knowledge of mathematics).
You can't claim knowledge of the world at all.

Any claim to know things in the world must by definition be fallible.

As I think that as a concept knowledge is pretty useless if it can't be used for things in the real world then
infallibilism as a universal definition and theory of knowledge is untenable.

However the Infallibilist concept of knowledge can still be used for things (like mathematics) where you both
can't accept fallibility, or have any need to do so.

For everything else you simply have to deal with the consequences of knowledge that is not infallible.

I had to add 3 words to my browsers dictionary for that.... I hope I spelled them right ;-)

Originally posted by bbarr
Let's suppose that fallibilism is true. Then, S can know P without being certain that P. So, suppose that S knows P and knows Q, that there is some nominal uncertainty about P and Q, and that neither proposition has any evidential relation to the other. Presumably, his confidence in the conjunction (P&Q) should be less than his confidence in either propositio which we're incorrigible). And then, essentially, you're back to the Cartesian Meditations.

If it seems weird to you that S can know a set of propositions and yet not know their conjunction, then you may feel the pull of the infallibilist position.
To be honest I don't think it is weird but then that's because I'm also adding in the stipulation that S has the option of claiming to know any collection of elements in that set only in isolation from its compliment.
Indeed, to continue working with your lottery example; I might (supposing I was a fallibilist) claim to know that lottery ticket {1,2,3,4,5,6} won't win, and had I not previously been asked about that one I might instead claim to know that {1, 4, 9, 16, 25, 36} won't win, or ignorant of both of these I might claim to know that {2,3,5,7,11,13} won't win, and so on.... If you ask me about any arbitrary collection of them however (i.e. not in isolation) then surely you've added in a non-trivial feature to this situation that didn't exist when asking me about one ticket (such that I am excused from needing to have the same degree of certainty about their conjunction)?

Originally posted by googlefudge
I think we may have to agree to differ about the terminology... However that is the less relevant or interesting part of the discussion.
so...


The two cases, knowing where my keys is, and knowing my lottery ticket wont win, are not contextually the same.
(for starters you have the logical paradox of buying a ticket you know wont win, If I know my ...[text shortened]... lieving it.
i just think those beliefs should be rationally and evidentially justified.

You can stipulate some nonstandard usage of terms that are used by working philosophers to mean particular things; that is your prerogative. But this is not agreeing to disagree, this is just you departing from the actual meaning of terms. But, whatever, do what you will.

There is no logical paradox in buying a ticket you know won't win. If there was a logical paradox you could derive a contradiction from the conjunction of the propositions 'S knows lottery ticket X won't win' & 'S buys lottery ticket X'. You can't, so there's not. There is certainly practical irrationality here, but that is the problem with gambling (unless it's just done for fun).

So, you admit that it is not probability that determines contextual differences in knowledge attributions. Fine. But, as far as I can tell, you've simply reiterated that there is a contextual distinction between the lottery case and the keys case without explaining what is epistemically different between cases. You can give reasons for thinking your keys are here or there. You can also give reasons for thinking your lottery ticket won't win. It is not the availability of reasons that makes a difference. If you tell me you know your keys have not been spirited away by ninjas, I'd think it as odd (and as reasonable) as you claiming that you won't be hit by a plane engine, or that you won't win the lottery with your recently bought ticket.

And, look, one unlikely event is that your keys will be spirited away by ninjas tomorrow. But if I take you seriously, you're claiming that you can't claim knowledge that unlikely events will not happen. So, presumably, you can't claim knowledge that your keys won't be spirited away by ninjas. But you claim that you can claim knowledge that your keys will be where you left them. So, you claim you can know P, but can't claim to know some proposition entailed by P. If you know where your keys are, then you know they haven't been spirited away. This is a basic problem with your contextualist account. There is overlap between contexts that yields conflicting standards of justification.

Originally posted by bbarr
You can stipulate some nonstandard usage of terms that are used by working philosophers to mean particular things; that is your prerogative. But this is not agreeing to disagree, this is just you departing from the actual meaning of terms. But, whatever, do what you will.

There is no logical paradox in buying a ticket you know won't win. If there was a ...[text shortened]... overlap between contexts that yields conflicting standards of justification.

No it isn't the availability of reasons but the type and context of the reasons that is different.

I am not sure admit is the right word, I wasn't ever claiming pure probability as my reason for claiming something as knowledge.


(sorry I wasn't claiming a proper paradox for the lottery, more an attempt at humour. )


I am not sure your ninja example works.
I think you are using an infallible definition of knowledge to get the conflict.

Lets say I specify 'knowledge of where I put my keys' as 'soft' knowledge.
It is (in my specified instance) likely to be true, and is reasonably justified, but admits the possibility to be wrong.
With soft knowledge knowledge that P does not necessarily give knowledge not-P.
However I can claim to know P and strongly believe not-P.

For example.

To claim any knowledge of the world you have to accept (on faith if you like) that it actually exists (in some form or another)
so that you can have knowledge about it.

You could have some statement P, where P is infallibly true, unless the claim the world exists is wrong (ie it is the only case in which P
is not true)
Thus if P is true the universe exists, if not-P is true then it doesn't.

I would in that instance claim to Know P, and believe not-P.

P is fallible knowledge because I can't prove the world exists.
But practically I can't function without the assumption (born out by everything I observe about it) that the universe is actually real.

If I also state that P is useful information for operating in the world.
Then by believing P I gain a benefit.
If I believe P, and it is only not true if the universe is not real, and I have already decided to discount that possibility, then I Know P.


If I try to as much as possible believe things that are true, or, true and useful, then those things I can justify strongly enough become knowledge.

I believe I know where my keys are, and I have a good reason to justify that, so I claim I know where my keys are.

I don't believe they will be stolen by ninjas, but that is not the same as actively believing they wont be stolen by ninjas.
It is not particularly useful to believe my keys are not going to be stolen by ninjas. So I don't go to the effort of actively believing that
ninjas are not going to steal my keys.
So I don't claim I Know they wont be stolen by ninjas (or all the other possibilities you could conjure for why they are not there).



However for the lottery I can describe the system perfectly with mathematics which I can know infallibly.... so why muddy the waters adding in
fallible knowledge where I don't need it.

I know the likelihood of the ticket winning and thus don't play.

I understand probability and statistics, so where I can describe the world accurately and as completely as needed with those, why use anything else?

However trying to do that for something as messy and mundane as where I left my keys.... It would be far to time consuming to work out the probabilities
assuming you could ever actually do it with any kind of accuracy and rigour.

Originally posted by googlefudge
No it isn't the availability of reasons but the type and context of the reasons that is different.

I am not sure admit is the right word, I wasn't ever claiming pure probability as my reason for claiming something as knowledge.


(sorry I wasn't claiming a proper paradox for the lottery, more an attempt at humour. )


I am not sure your ninja ex ies
assuming you could ever actually do it with any kind of accuracy and rigour.

I'm sorry, but your post is too muddled to get through. I'm presenting arguments, and you're writing yourself in circles. So, I'll just let this go. But, if you're interested, you should pick up an introductory textbook on epistemology. A former colleague of mine, L. BonJour, has a great book: Epistemology: Classic Problems and Contemporary Responses.

Originally posted by Agerg
[b] If it seems weird to you that S can know a set of propositions and yet not know their conjunction, then you may feel the pull of the infallibilist position.
To be honest I don't think it is weird but then that's because I'm also adding in the stipulation that S has the option of claiming to know any collection of elements in that set only in isolation ...[text shortened]... I am excused from needing to have the same degree of certainty about their conjunction)?[/b]

I have no idea what you're saying here.

Originally posted by bbarr
You're not paying attention.

More likely, I haven't studied the subject, so I am not understanding you.

You can revise the notion of fallibilism, or revise your agglomerative principle
I think I would revise both from the way you are defining them. You seem to take the claim to "know" under fallibilism to be infallible, which to me makes no sense.

And 'Fallibilism' is the view according to which we can know propositions without being certain of them.
Well I interpret it as 'saying we know something when in fact, we can not be certain'.

Originally posted by bbarr
I think that's right. It's the agglomerative principle that needs to be rejected. It's just like the 'Author's Preface Paradox'. But what a strange implication. We have sufficient reason to believe each of a set of propositions but not to believe their conjunction. The interesting question, I think, is just what sort of revised agglomerative principle we'l ...[text shortened]... erivation, and in fact (((P1 & P2 &...Pn) & F) > C), I may not be able to know C.

I have not really studied the lottery example or the author's preface paradox before, but I have to say that I find them fascinating on first inspection.

If I understand the basic idea, it can be summarized roughly as follows. Suppose S believes individually each of P1, P2, …, Pn on the evidential bases of e1, e2, …, en, respectively. But suppose also that S believes ~(P1 & P2 & … & Pn) on the evidential basis of E. We can envision cases (such as in the lottery and author's preface examples) where all these listed beliefs seem justified in virtue of their respective bases; but at the same time these listed beliefs are obviously inconsistent.

So this is a case where it seems that all individual beliefs listed are justifiably held and yet together they form an inconsistent set. I guess one thing that it is relevant to note is that it does not follow from the above that S holds contradictory beliefs. I think I am right in saying that because from the supposition that S believes individually each of P1, P2, …, Pn it should not follow that S thereby believes (P1 & P2 & … & Pn). I guess to relieve the tension (all individually justified but together inconsistent), one could try to deny that they are all individually justified or else deny that they are together inconsistent. However, I would think that, for instance, considering the lottery example, neither of these approaches would work. The denial that all listed beliefs are individually justified seems absurd, since each Pn is overwhelmingly probable (credence level can be taken arbitrarily close to 1 for each) and since ~(P1 & P2 & … & Pn) is also basically guaranteed within the structure of how this lottery operates. For the latter approach (denying that they are together inconsistent), I suppose one could argue that what S actually believes in the lottery case is rather each of probably P1, probably P2, …, probably Pn and ~(P1 & P2 & … & Pn). In that case, there may be no inconsistency of the beliefs. But the general idea that "S believes P on the basis of e" translates to "S believes probably P" seems false to me. The reason why it seems false to me is that it would confuse two distinct ideas: (1) that probability factors into the evidential basis e for S's belief and (2) that probability factors into the actual content of S's belief. Do you have any ideas here?

So to me, it seems that the lottery example does show that there can be cases with genuine tension where all the beliefs are justifiably held but together are inconsistent. I do not see a way to break this tension; again, I would deny that justified beliefs are closed under entailment. As you say, this is a rather strange implication. I guess I would also have to deny that if S holds inconsistent beliefs it need follow that S is doing something wrong or that it is incumbent upon S to give one or more of these beliefs up. Still, though, if some set of beliefs are inconsistent, then one or more of them must be mistaken. So, if S holds inconsistent beliefs and S is aware of the inconsistency, maybe it still follows that S has some epistemic duty to endeavor to figure out, as far as reasonably achievable, which belief(s) is false.

Thank you for describing "The Complex Proof Paradox". I will have to give it some thought. By the way, is the complexity supposed to relate also to C itself; or does the complexity just relate to the complicated nature of the proof demonstration (owing to the fact that n is large, that the inferential trajectories are many and deeply involved, etc)?

Originally posted by bbarr
I have no idea what you're saying here.

Then I'll try again... You lottery example argues that supposing S knows P_1, then with the same justification he can claim to know P_2 and so on... till we reach P_n. In considering each out of all of them them he then knows their conjunction, whereby you establish a contradiction.
I see no reason why a fallibilist should continue with your argument beyond the first step however since he might only know each P_i in isolation from the others - that is, the most you can say might be S knows P_1 or he knows P_2 or ... or he knows P_n (and those "or"s should be regarded as "exclusive or"s). Your contradiction would not show up in this case.

Moreover in in the real world this is feasible, if I ask a fallibilist if he knows lottery ticket x is a losing ticket then though x can stand for any out of all the possible tickets it still references no more than one ticket when I ask the question. As such with U denoting the set of all tickets he has not yet considered the set U \ x i.e. all the other tickets apart from x.

Originally posted by Agerg
Then I'll try again... You lottery example argues that supposing S knows P_1, then with the same justification he can claim to know P_2 and so on... till we reach P_n. In considering each out of all of them them he then knows their conjunction, whereby you establish a contradiction.
I see no reason why a fallibilist should continue with your argument beyond t l tickets he has not yet considered the set U \ x i.e. all the other tickets apart from x.

That's not the argument.

Again, S justifiedly believes (P1..Pn), (P1...Pn) jointly entail C (that no ticket will win), but S cannot justifiedly believe that P (since some ticket will win). If Infallibilism is true, then S cannot justifiedly believe (P1-Pn), since these will be uncertain. So, no problem. But Infallibilism is nuts. If fallibilism is true, then justified believing isn't closed under deduction. But some sort of agglomerative principle is necessary, so what will it be?

Now, what is your problem? Because, look, your "exclusive or" strategy; essentially saying that S knows some complex disjunction, doesn't make sense. S does justifiedly believe each of the propositions, not just this one xor that one. S doesn't know each of the propositions, because one of them is false. But S does have evidence (as close as you want to establishing certainty with actually establishing certainty), that each proposition is true, and presumably such evidence is sufficient for justification... (and if you think otherwise, I'll need a good reason why we shouldn't take a belief that is 99.99999....% likely to be justified; and remember, fallibilism doesn't require certainty for justification, by definition).

Originally posted by twhitehead
More likely, I haven't studied the subject, so I am not understanding you.

You can revise the notion of fallibilism, or revise your agglomerative principle
I think I would revise both from the way you are defining them. You seem to take the claim to "know" under fallibilism to be infallible, which to me makes no sense.

And 'Fallibilism' ]
Well I interpret it as 'saying we know something when in fact, we can not be certain'.

Revise to your heart's content! But these are not my definitions. These are just the definitions as they are used by epistemologists. And, no, I'm not taking claims to know as infallible. I have no idea what you misread that would give you that idea.

Feel free, though, to use 'fallibilism' any way you want. That's not how I use it. But, since it's apparently the trend in these fora, why not let many semantic flowers bloom?

Originally posted by LemonJello
I have not really studied the lottery example or the author's preface paradox before, but I have to say that I find them fascinating on first inspection.

If I understand the basic idea, it can be summarized roughly as follows. Suppose S believes individually each of P1, P2, …, Pn on the evidential bases of e1, e2, …, en, respectively. But suppose als ...[text shortened]... that n is large, that the inferential trajectories are many and deeply involved, etc)?

I agree with most of this. I'm not sure about the epistemic duty stuff, but I'm interested in a virtue epistemological account's prospects for handling these sorts of tensions. C is just any old conclusion, it's the derivation that's supposed to be complex. OK, I'm off for a vacation with the girlfriend for the weekend. I'll think more about this and write you on my return. All the best...

Originally posted by googlefudge
I think we may have to agree to differ about the terminology... However that is the less relevant or interesting part of the discussion.
so...


The two cases, knowing where my keys is, and knowing my lottery ticket wont win, are not contextually the same.
(for starters you have the logical paradox of buying a ticket you know wont win, If I know my ...[text shortened]... lieving it.
i just think those beliefs should be rationally and evidentially justified.

This is why I am a Christian and not a Hindu.

Originally posted by bbarr
You can stipulate some nonstandard usage of terms that are used by working philosophers to mean particular things; that is your prerogative.

What is the standard definition of fallibilism? I was taking it to mean 'I say I believe something even when I know it is simply highly unlikely (but still possible) that I am wrong.' Clearly that is not what you mean by the word.