Even at the risk of starting another boring thread with a lot of miscommunication about knowledge and truth and all things related, I would discuss the justification criterium in the definition of knowledge as justified (true) belief. (Truth is in parentheses, because there are some that would omit it, and I do not particularly crave for the discussion about that at the moment.)

Basically the problem stems from one thought I had a while ago: how can a proposition that is justified be false? If it is false, why would we say that it is justified? If we hold on to this view, how should justification be understood?

I will start my considerations from Gettier, who has famously presented arguments against knowledge as justified true belief. Without going to further details, I think that Gettier's counter-examples to the classic definition is knowledge does not concern knowledge as justified true belief per se, but the ambiguity of justification. The Gettier-examples for instance construct a case where we would grant justification for one's belief, yet it still would not be knowledge - even in the case where it was true.

If I do not problematise the truth-part, nor the belief part (I think both are fairly clear and unambiguous), I am left wondering: why would we say someone is justified in believing x?

It crossed my mind, that justification has been misapplied. Let us consider a simple example: John struck Peter because Peter made fun of John's grandmother's death. Now I think it is fairly obvious that this does not constitute a justified act - it is, I believe, so that in court John would be considered as guilty and obliged to, for example, pay Peter. But when someone asks "Why did John strike Peter?" and I explained the situation, that someone might well sigh that "well, it was understandable". Understandable, but not justified, eh?

What does understandability mean? Basically it seems to mean that someone considers it so that he projects himself into the shoes of John, and believes that he might have done the same thing. Even if it was not justified, we could easily imagine ourselves doing the same thing in the same situation with the same state of mind and same beliefs as John had.

What is the relation between understandability and justification? What about this: understandability is in question when we consider it so that the person whom we try to understand is seen as believing that he has a justification (there are, of course, some problems in this in relation to the above example, because we might consider that John wasn't thinking at all, but reacting, and thus had no beliefs about justifications - he might even consider it unjustified and still do it. But because I will steer this to more epistemological matters after this easy-to-understand example, I think it is not relevant to consider this problem here.) Basically, what I am saying is, when we think it is understandable that x did y, we think that from the point of view of x, it seemed justified for x. That is, x believed he was justified - the fact that he wasn't would then be explained by added premises that x was not aware of (something that the attempted solutions of Gettier-examples are trying to do.)

Can something that we believe be true and justified, yet not amount to knowledge? I do not think this is possible. I think that the problem we face here is that justification has been confused with understandability. This is linked with the fact that justification is considered an intersubjective thing - we consider, for example, science to give adequate justification for Earth's roundness. But because this is a convention, it is always easy to step outside the bounds of these conventions and show that either we were not justified after all, or that this justification (of a true belief) did not in fact amount to knowledge (this is the little hole the skeptic tries to squeeze himself through.). This is what the Gettier-examples do: they assume a more general point of view (the point of view of the narrator, who knows all the relevant details in the situation) that is beyond the beliefs of the person in question, and shows that what we consider as justified true belief (for the person) is in fact not knowledge. (The Gettier-examples always start with a picturing of a situation, where the reader is led to believe that justification exists (which it does, in that point of view), and then given additional knowledge that renders the knowledge questionable (because we step to another system.)

I think this is confused, because it should read: what is justified and true belief for x, is not a justified and true belief for y <=> what is knowledge to x, is not knowledge to y. Y is of course the narrator in Gettier-examples, which allows him to seriously question x's knowledge, even if he accepts his belief's justification and truth. The Gettier examples, simply, change the point of view in the middle, which causes the paradox or contradiction.

Therefore I claim that we should speak of understandability instead of justification. Let me give an example.

I believe that pi is irrational. I think on this for a while, wondering why this is so, and come up with a proof. But the problem with the proof is, unbeknownst to me, that it is faulty. Say that it in one point misses a single tiny detail, and even though it does lead to the irrationality of pi, and I believe it, the proof itself is false. Now according to the standard view, I am nonetheless justified in believing what I do. The proof is good enough, it is rational and well argumented but for the one detail, so I have every reason to assume that it is correct, and I can most probably explain every detail of it. But wait a minute! How can it be justified, if it is an erroneous proof? I contend that it is not, in fact, justified. What we mean by "he was justified in his belief" is really that "it was understandable that he believed so" - that I believed I had justification, and that you consider it so that if you were me, you would too have believed this (that is, I didn't miss out anything that I shouldn't have missed - the only reason I did that was because I didn't have the proper knowledge to pull the proof through).

My point is: justification is always justification from the point of view of x. Justification is given by science, because it is scientific justification. We give science the ability to give justification because we have noticed that it works. It works for us. Saying that knowledge is justified true belief is simply too ambiguous, because it fails to recognize the fact that justification is always a relational term: x is justified in y, where y is some system of beliefs or truths, such as science.

Therefore, I conclude, there is no such thing as justified true belief in general. This is why the Gettier-examples are so powerful: only our imagination is the limit in considering more and more general cases where the narrator knows something the protagonist doesn't, and therefore can either claim that there was no knowledge, or there was no justification. What we mean when we speak of justification in these cases is really understandability.

What is understandability's relation to justification exactly? Basically this: consider x and y, and their relative belief-systems A and B. We can now assume that the statement p is true. Now let us say that x believes that p. x's belief that p is understandable to y insofar as y can relate to A (instead of B) and agree that if A is assumed, then p is justified. That is, y will say "x's belief is understandable" if he considers it justified to believe p in A. Note that the crux of the matter is that y might still consider it unjustified for x to believe in p, because the belief-system B holds some information about A that would render the justification null - such as that my proof was actually faulty.

Justification in general is simply understandability from the point of view of an omniscient being. Or, perhaps more precisely, the system of the omniscient being can give us justification in general, because there is no greater system that could affect the premises in the system and render it merely understandable.

This has been a rather vague explanation of an idea. It is vague simply because the matter is very complicated, and perhaps also because I have not given it enough thought as of yet. This survey can be seen as presenting two lines of argumentation: one showing that there is a problem with the justification-criterium because of its ambiguity, thus calling for a need to define justification, and another trying to solve the problem by presenting a distinction between justification and understandability. Because of the lack of proper definitions for justification and understandability both (B's understanding of A is surely something that is hard to define strictly), the presentation is somewhat vague and general.

Therefore I would ask some opinions on the whole approach in general as well as propositions to better define the terms. Pointing out any ambiguity is surely useful, even though some of them might be beyond reach at this moment. Please bear in mind that the Gettier-examples, however nuisance they are considered, are still unsolved in many cases and show a real problem in the definition of knowledge as justified true belief. My suggestion, in general, is that instead of adding demands for knowledge, we should instead labour to define the used terms better. In this, the relationality of justification plays an important part.

In fact, for those to whom it says something, I consider at least some of the Gettier-examples to be paradigm cases of the fallacy of arguing from one set of axioms to another. This goes unnoticed by both parties, if the relationality of justification is not properly attended to.

This text caused many questions, and in this following post I will try to answer those problems that arose among the forum-members.

Okay, boys, been a way for a while, but I will now try to give some sort of, probably rather long, response to the concerns and ideas expressed here. (And indeed, it is rather long. If you don't want to read the whole of it, just read the part that is directed at you, as expressed by the names (except one part where I note that jaoman says something in between of my explanation to Machiveli. My thoughts to jaoman are in the end.)

First and foremost, Let's try to keep the discussion within the bounds of Knowledge as Justified True Belief (Thank you, Reformed Nihilist!). The only chance we have at ever working out something functional is to keep some variables fixed and see how changing one affects the whole picture. Changing them all at once will only result in disaster. There are threads about truth, and I suppose there are some about the nature of belief too. To some extent of course we will have to take these things into account, but I should think it proper to assume that truth is something that expresses how things are, and belief is something that a person has. If I go on to define things even a bit more profoundly, then there will just be a hundred people disagreeing with my definitions.

Secondly, we must consider justification. There seems to be some variation in the way justification is understood to begin with. That is, of course, a problem when trying to define what the justification in JTB stands for. If someone takes justification to be subjective, I don't think it fits the point in the definition. One who has belief to something most probably also thinks that he has justification for that. That is NOT the point - the point is to ask whether one actually has a point. This is why justification criterium is brought forth in the first place. Probably the easiest and most compact way of explaining what JTB stands for is to say that belief is the subjective, justification is the intersubjective and truth is the objective ingredient of knowledge. Knowledge is then something that links a subjective assumption through generally accepted ways to something actual - grounding the belief in the reality. Justification is needed to carry the belief to the actual - a mere true belief is true "by accident", which I believe means that there is no direct epistemological link from the subjective states of affairs to the objective. It is a mere coincidence that they match.

Of course, the nature of intersubjectivity is open for debate, but justification ought to be distinguished both from the subjective intensional state of believing and from the objective state of something being like something. There are theories of truth that omit the truth, but I think this is only accomplished through careful analysis of that which is directly accessible to us: beliefs and justifications. If the definition of justification leads to omitting truth, then it does, but let us assume that such a thing exists. If it doesn't, then call it an indirect proof - that would make any mathematician proud ;).

The reason this is vague is that it just came to my mind. I will try to formulate my idea better.

So, we have two belief-systems: A and B. It doesn't matter what those include, but they are the relevant things that x and y consider true, respectively. Understandability requires something I would call axiomatic or philosophical empathy, that is, the ability to set oneself into the belief system of another, for the sake of argument, so to speak. It is related to normal empathy in that if one lacks empathy, one cannot understand others - and if one lacks axiomatic empathy, then one does not consider any action understandable; she considers it only either justified or not justified.

We also have a statement p, which can be either true or false, which is something we do not concern ourselves with here. We also assume that the belief-systems A and B are accessible to both x and y - something that is of course ideal, but we are trying to define justification, not forming criteria to when the assumptions of justification are false; we can do that only if we know in what situation one would be justified.

Thus understandability of x's proposition from the point of view of y is dependent only on x's belief-system A. The justification of it is dependent on y's system of belief B. Note that when we switch the aspect, we will see that something that is justified for y might be merely understandable to x. The justification of x's proposition (for y) on the other hand is only dependent on B, not A. Thus if we fix the point of view to y's, then:

Now you might be concerned about the excess-usage of justification here. What makes something justified in B? That is another matter that we will return to later, but for now I only try to explain the difference between understandability and justification as well as I can. Note that y doesn't have the final word here - what y considers justified for x to claim might in turn be merely understandable to z and so on.

Now to the matter of knowledge versus certain knowledge.

When are there such belief-systems? We would have to have a criterion for a belief-system, but we do not. Let me then at least say that the belief-system must be rational. The rationality is, once again, undefined, but that is the case in philosophy in general. We cannot here handle the problem of "rationality", but we will have to keep that again fixed somewhere. This is of course the exact problem with skepticism: one can always doubt, but can this doubt be rational? This is something that is simply too big to handle here. If we assume some criterion for rational doubt, then we accept some criterion for rational belief-system, and we can all, I think, agree that a piece of knowledge is certain if it cannot be rationally doubted - after all, this is all I am saying, but I am trying to formulate this in the language of justification and understandability.

Jaoman suggested that my criterion is too strong. I think it is not, because the demand for certain knowledge is very strong to begin with. It only formulates the idea of certain (undoubtable) knowledge in the language of justification. I agree that it is difficult to attain such heights of justification, but it is equally hard to attain such certainty of knowledge too - my definitions can be considered correct insofar as the difficulties raise in parallel. That is, all things being equal, demanding more of justification means demanding more of knowledge, and vice versa. Moreover, it doesn't, I think, render it wholly impossible to attain certain knowledge. I do think that "all bachelors are unmarried" expresses a certain piece of knowledge (please do not take this claim too seriously, because it is a sidepoint: I do not wish to derail the discussion to the certainty of analytic propositions). It is also true, I think, that this is at piece with my definition: we cannot rationally imagine such a belief system that would render this statement merely understandable (meanings of words being equal, of course, otherwise we are not speaking of the same proposition). Even more defence for my system is gained when one thinks what would render this piece of knowledge uncertain: it would do that when someone imagined a case where it could be false. Exactly as I would have it.

There is an interesting link from this to the problem of necessity, that I will now examine in brief. p is necessarily true iff it is true in all possible worlds. p is only possible, that is, not necessarily true, iff it is false at least in one possible world (but not all). Now consider this in connection with my justification-criterium, and you will see that:

This means, I think, that my definition would lead to a parallel kind of consideration regarding epistemology as the possible worlds semantics does regarding truth. (I believe this is also applicable to the third kind of certainty, that is: subjective certainty as an attitude towards one's beliefs: one is incapable of coming up with any believable proposition that would render the subjective certainty faulty - subjective possible worlds of beliefs, so to speak...)

Now there is still something to add. My terminology is far from conclusive. The word "justification" itself is related to its common use in epistemology (it might be that it should be revised instead of redefined), and understandability is introduced only to express the difference between unconditional justification and conditional justification. Instead of "p is understandable" I might as well say that "p is c-justified (for conditionally justified)" or something like that. I just think the understandability-criterium expresses in concrete terms what I am trying to say, so that the IDEA could be expressed - even if the formulation should in the end be different. I want you to say to yourself "Ah, THAT's what he is trying to say" - only after that we can consider whether different expressions should be used. If I started using conditional justification or some such, I think it would lead people astray: most things that have too much relativistic connotations tend to do that.

On this thread, we can agree to that. With a minor addition: the belief is held true by an individual - or a group of individuals. When we say "Finns believe that...", we mean that they share the same belief - not that there is a conscious entity, Finns, that believe something collectively. This point can be argued, though, but perhaps it is easiest not to, at least here. I agree that the words can be used in such an such a way, but this is not what I mean when I speak of beliefs in my justification-criterium.

So, you are saying that because beliefs have reasons, then if we assume that the group x has a belief (the people comprising the group believe the same thing), the group has its reasons, and these reasons are then intersubjective, and therefore able to constitute justification - or make justification redundant because the intersubjective reasons are enough? I guess you can do that, but then you are not speaking of the same thing. The problem is that the reasons I might have for believing whatever are mostly not epistemological reasons in any way: my belief that there went a car past my window is based on the direct perception of the sound of a car, and then on some implicit non-cognized reasoning. Only after someone asks "WHY do you believe that", I can start poking into myself to see why I in fact did. Often a perception is behind it all, sometimes we can only shrug helplessly and say "I don't know - I just thought that it was so for some reason".

Here we see a difference between the subjective grounds for believing something (the reason - it might even be a causal reason, mind you!) and the intersubjective grounds for believing something. I have no justification for my belief that it is day outside, but I can formulate a justification by referring to my perception, perhaps even showing someone that it is bright outside. The justification is something that is uttered, or formulated, or communicated in some other way. I have a host of beliefs (probably millions all the time) that are not epistemological - but can be made such when questioned. This is the difference between reason and justification. The reason is no justification. Justifications are reasons that are held to be good enough to hold - that is, we give those to people who question something. This is the point in knowledge in general: knowledge is subjective dispositions taken to objective/intersubjective world through communication, and perhaps through somehow comparing those dispositions to reality.

I do not think a group can have reasons in the same sense as you say individual's beliefs have reasons. It is important to understand that the reason behind a belief might be a mere causal procedure (my reason for believing a perception is that this perception was caused by a firing neuron in my head), but this causal procedure is no justification. If it was so, then everything we did was justified just because everything we do is done for a reason. Causal or otherwise.

Simply because having a reason does not amount to having a justification.

Justification for a belief I think is not the same thing as justification OF a belief. That is, I may justify my belief in many ways, but it doesn't follow that it is justified. Justification is something, at least here, that can be measured, that the action of justifying can fail. It is the same as trying to prove a theorem and actually proving it. The point is that this proof can be evaluated by others.

By justifying my belief I give you the reasons for believing it. The belief is justified only when these reasons are deemed good enough (this is rather vague, of course, because justification depends on who is doing the evaluation).

Vindication would indeed include the truth of a statement. You are then saying that I am confusing justification for vindication, is that it? I think that claim is warranted, but not entirely true. Note that even my certain knowledge does not include truth - even a certain piece of knowledge can be proven to be false: the criterium is that no one could think of a way for it to be false. It is true that the first question I presented: "how can a proposition be justified and yet false" indicates this confusion, but it is only a question that leads to the definition of justification.

This is a good point, though, because I had not considered the concept of vindication.

This is true. However, a question: when is this vindication "justified"? That is, let me claim that JFK was assassinated by CIA. I may not have sufficient justification for this, but it might then turn out that someone finds a document holding conclusive evidence that this was so: my belief was vindicated, right? But this vindication may yet prove out to be false: the document was forged. I would then be deprived of my vindication. So the vindication is linked to justification: there is no such thing as something turning out to be true - we may always ask for the justification for assuming that that something indeed is true.

It seems to me that vindication is then tied up with justification - and might indeed be reduced to analysis of understandability and justification in my system: vindication happens when some system of belief that would justify my claim turns out to be acceptable, even though my belief-system is not enough. In short: something comes up that fills my belief-system so that the belief becomes justified. This does not implicate that there could be yet another system of belief rendering this justification insufficient.

In a sense, yes. Understandable is c-justified (conditionally justified), and justified would be unconditionally justified. But as I laboured to show in the beginning of this moster-of-a-post, this justification is again only c-justification from the point of view of another belief-system. The real justification is always linked with certainty: we can consider that which is justified in all systems to be justified irrelative to those belief systems. But as I also said just above, this does not yet link to truth - a certain piece of knowledge may be false! But vindication here includes a relation to truth. Vindication too is conditional, and in fact it is highly problematic: "look, there is a sheep there after all!" may be an error again, and my vindication may be short lived. Just because we gain more support for our justification, and thus make the merely understandable belief justified, it doesn't mean that it may yet again be rendered merely understandable. Vindication would then be simply the rendering of an understandable or c-justified belief into a justified belief. But this is again just c-justified to someone else.

Your analysis is very sharp-sighted, and I will think on this more. But the connection of vindication to truth makes me doubt that this is what I am saying.

True to some extent. This is analogical to the problem between knowledge and certain knowledge. Just because someone can doubt my knowledge it does not follow that I do not know. It does follow, that (if the doubt is rational) it is not certain, but that is entirely another matter.

I am merely conforming to the fact that something that x considers justified may not be considered justified by y. Thus to speak of justification as unrelated to the justifying person is misleading. It is like saying that the number 27 is great. It is great only in relation to some number that is smaller than it. I think my actions are, then, justified ;).

Yes, but the justification criterium in JTB is not such a justification as expressed here. It does not mean that the proposition "pi is rational" is justified just because someone thinks that it is. This is the crux of the matter: the 'justification' as expressed in JTB is not something merely subjective, and it seems to me that it is not something merely objective either - it is something intersubjective. It is then to be asked: when does something constitute justification? We do not consider that x is justified in believing that "pi is rational" just because he thinks that irrational numbers cannot exist in nature and pi exists in nature. He surely has reasons, and those reasons are expressible and even evaluatable, but he has no justification - or so we think. Basically: the justification I am trying to understand is that justification in JTB, which is not subjective justification. The difficulties associated with justification are, then, the exact problems of determining how to distinguish real justification from assumed justification.

I agree that there is a sense of justification that amounts to your points, but I am not speaking of that - and I claim that neither are epistemologists when they speak of knowledge being justified true belief. The justification here is indeed linked with the social ruling, so to speak. But this is not merely so: it is not arbitrary to say that science can justify things better, but it is grounded (in a sense, again justified) on experience and the results of science.