The Truth Connection

Here’s an idea I want to try out on the problem of the truth connection. By way of background, there are several approaches one might take toward this problem that are problematic. One might hold, for example, that justification entails truth, but that approach founders on the possibility of justified false beliefs. One might also hold that justification makes for objective likelihood of truth, but that approach has difficulty with the new evil demon problem.

So why not just say the following? When a total body of evidence justifies one in believing a proposition p (in the way required for knowledge), that body of evidence also justifies one in thinking that the belief in question is the appropriate belief to hold in the interests of getting to the truth and avoiding error. Put in shorthand, the total body of evidence not only justifies believing the claim in question, but also justifies believing that the belief is epistemically justified. (As an aside, I think there are hints of such a view in Chisholm’s 2nd ed. of Theory of Knowledge.)

This is enough to distinguish epistemic justification from other kinds of justification, such as practical or moral justification. It also ties epistemic justification to truth in a way that belief itself is not tied to truth (Some think that to believe a claim is to believe it is true, and so, too, to justifiably believe a claim is to justifiably believe that it is true–but here the connection to truth merely piggy-backs on a purported connection between belief and truth.)

Some will remain dissatisfied that there this connection doesn’t imply that the belief in question is objectively likely to be true. That I admit, but the advantage is that this view faces no new evil demon problem. So, my question is whether anyone sees additional problems for the view, besides the obvious lack just noted.


Comments

The Truth Connection — 15 Comments

  1. Jon,

    Your proposal might run into problems when it comes to subjects who, while intuitively able to have justified beliefs, aren’t so sophisticated as to have the concept of epistemic justification. If they don’t have the concept of epistemic justification, it is arguable that they are thereby not in a position to justifiedly believe or know any proposition to the effect that the belief that p is epistemically justified (for me).

  2. I think it is correct to say that if you don’t have the concept C, you can’t justifiably believe a proposition that has C as a constitutent. But I see no reason for thinking that you can’t have evidence for such claims. It’s just that you can’t use that evidence to come to justifiably believe the claim (since you can’t believe the claim).

  3. Jon, I have a general worry about any account motivated in the way that this one is.

    If we suppose that there is some epistemic property that we know apriori (on the basis of rather elementary thought experiments like those involved in setting out the new evil demon problem for reliabilism) to be modally seperable from any properties which raise the chances of our beliefs being correct, don’t we know that the fact that such a property obtains can’t give us a reason to believe any of the beliefs that happen to have that property? If so, don’t we know that this property can’t be the property in virtue of which a belief is justified?

    I think it is one thing to say that the properties in virtue of which our beliefs are justified are such that we don’t know whether the fact that our beliefs have them raise the chances of our beliefs being correct, but if we claim to know such properties fail to raise those chances, we can’t take the obtaining of those properties to be what grounds the fact that we are justified or have a reason to believe something. At least, it seems like that if we think that when one is given evidence for thinking that what they know doesn’t raise the chances of some claim, they have thereby lost justification for believing the proposition in question. I think this is a problem for the view you are sketching as well as any view of justification which is motivated by its appeal to the sort of thought experiments internalists commonly deploy to undermine the claim that one’s beliefs could be justified in virtue of that which raises the objective likelihood of truth.

    Anyway, that’s the worry.

    Clayton

  4. I have evidence that this laptop battery is fully charged. (The little icon indicates it is.) Is that evidence that I am epistemically justified in believing that the laptop is fully charged? Wouldn’t the latter have to include evidence about what I’ve seen — evidence about my psychological states — which doesn’t seem intuitively relevant to the question whether the laptop battery is fully charged?

    If I have the evidence that the laptop is charged, and I’m not abnormally lacking in self-awareness, then it seems likely (maybe even guaranteed?) that I have evidence that that belief is justified for me. But don’t they still seem to involve different evidence?

  5. Clayton, if you know or have evidence that your reasons for belief do not increase the chances of the belief’s being true, that is prima facie evidence that the belief is not appropriate if your goal is to get to the truth and avoid error. So I don’t see how the worries you raise cast doubt on the proposal.

  6. Steven, the proposal is not that the very same evidence that justifies the belief also justifies something else as well. The proposal is that if your total evidence justifies a belief in the way necessary for knowledge, then your total evidence justifies something else as well. The evidence for the two claims might indeed be different.

    Note that I don’t claim that this is a feature of ordinary justification. It may be that there is a kind of justification, ordinary kind, on which the proposal is false. The proposal is meant only to apply to the kind of justification necessary for knowledge (I’m assuming that there is some such kind).

  7. Jon,

    Let me say a bit more to see if it motivates the worry I still have. If you grant that as soon as we (you and I) know as a conceptual truth that all and only the facts relevant to epistemic justification are modally seperable from the facts that raise the likelihood of believing what is true, then won’t we now have a prima facie reason to think our beliefs aren’t appropriate for our epistemic goals? That seems pretty troubling to me.

    If you grant that much, it seems we also have to say that others who we would have otherwise taken to be justified in believing what they believe are really under some illusion thinking that their beliefs are appropriate for the goal of believing the true which would vanish as soon as they realize the conceptual truth that we’ve realized having thought through the new evil demon thought experiment. And since I’m not at all tempted to think that this is so, I’m tempted to think that there is something faulty with that thought experiment and to think that the modal separability of the justification conferers and chance raisers isn’t a conceptual truth at all.

  8. Clayton, you say,

    “If you grant that as soon as we (you and I) know as a conceptual truth that all and only the facts relevant to epistemic justification are modally seperable from the facts that raise the likelihood of believing what is true, then won’t we now have a prima facie reason to think our beliefs aren’t appropriate for our epistemic goals?”

    This conditional is not plausible at all, I think. It has as an antecedent a modal claim, and as a consequent a non-modal claim. In general, such conditionals are false: I have reason to believe my wife is faithful even though I know that her being so is far from a necessary truth, etc. And here the conditional I use is even stronger than the one you use, since mine has the same claim in both antecedent and consequent, whereas yours has different claims. Yours I don’t find any more plausible than the claim that because our evidence doesn’t guarantee truth, it can’t justify belief at all.

  9. Suppose part of your total evidence is E, and that on its own E would justify you in believing P to a high degree, say 0.95. Suppose that another part of your evidence is the testimony T of some philosophical authority, who tells you (falsely) that (prior to his testimony) you’re only justified in believing P to degree 0.2.

    I assume that T would to some degree undermine the justification E gives you to believe P. But it need not WHOLLY undermine it: additional defeating evidence could always undermine it more. Let’s say that on the basis of T and E together, you’re justified in believing P to degree 0.8.

    On Jon’s proposal, having evidence E alone would give you some justification to believe you’re justified in believing P. I suppose that justification would still be there when you acquire evidence T; it’s just that now the T-evidence that you aren’t justified and the E-derived evidence that you are justified have to be balanced against each other. I see no reason to insist that the balance point will also be 0.8. Why couldn’t your second-order justification, on balance, be that you’re only justified in believing P to degree 0.6?

    This picture would allow first-order justification to generate some second-order justification; and it would also allow second-order justification (from other sources) to affect your first-order justification. But it doesn’t insist that your degree of first-order justification and your degree of second-order justification always coincide.

    Now, in his original proposal Jon uses all-or-nothing-justification talk. And there’s much obscurity about how to translate between degrees-of-justification talk and all-or-nothing-justification talk. But naively, it seems like, in a particular case, 0.8 degree of justification for P could be enough to count as “justified in believing P” but 0.6 degree of justification for JB(P) not be enough to count as “justified in believing JB(P).” So it looks to me like you could be, all-things-considered, justified in believing P but fail to be justified in believing that you’re justified in believing P. Even granted all the necessary conceptual resources.

  10. Jim, I don’t think your case will run through if we are considering ‘total’ evidence as Jon was assuming. If the total evidence includes T and E, then the degree to which p is justified and the proposition that p is justified appear to be the same.

    Jon, I see 2 potential worries. First, the worry as to whether evidentialism will extend to a priori justification. Second, if seeing the connection between the body of evidence and the propositions it justifies is necessary for justification (see Plantinga on Defeaters), then it is possible to have one of the two beliefs justified and the other not due to the failed existence of the contingent mental state that is ‘seeing’.

  11. Jon,

    I say yea and amen. But I’ve already shown my hand on these issues, I guess. Some of the things I say about connecting beliefs in my response to Clayton is very similar to your main claim here. I’ll be interested to see how you think your claim and my claims connect.

    Matt

  12. Jim, your case is quite interesting, and I have two comments about it. The first supposes that the case goes through as you describe it. If so, then the proposal I gave must be modified (or abandoned), but I’m thinking it could be modified as follows. Epistemic justification constitutes prima facie justification that the belief in question is the right one to hold in one’s pursuit of the goal of getting to the truth and avoiding error. In short, just change the connection between object-level and meta-level justification from an ultima facie to a prima facie connection. Such a prima facie connection would still distinguish epistemic justification from other sorts, and yield a connection to truth that doesn’t piggy-back on the connection that believing a claim has to believing that it is true.

    Maybe that will work, but I’m not sure it is necessary, since I’m not sure about your example. I think the crucial question about your case is how the authority is perceived by the epistemic agent. As you describe the case, I think you intend the authority to be taken as such. If so, however, the agent has an all-epistemic-things-considered reason to assign a probability of .2 to the proposition in question.

    If the agent doesn’t do this, then the agent must have some reasons for distrusting the authority in question (even if the reasons are only Moorean ones). If so, however, then it is not clear that the testimony in question needs to carry any weight at all.

    But it could still carry some weight, as you say; there are intermediate cases between the all or nothing cases I just described. So suppose we have some function on E and the testimony in question that yields probability X for p for the agent S. We then ask S how likely it is that s/he is justified in believing p; does the answer have to match X? I agree with you that there’s no obvious reason why, but if S’s total evidence confirms that S is not justified in believing p, then, as the case is described, it would have to be on the basis of some strong authority granted to the testifier, an authority that remains unrebutted by further considerations. If so, however, I don’t yet see how probability X itself could be high enough to yield epistemic justification.

  13. Christian–I think on the question of evidentialism and a priori justification, you must be thinking of Plantinga’s woeful notion of evidence. The more interesting versions, e.g., Feldman and Conee, find a place for experience, and would certainly include a priori intuition if there be such.

    On the second point, note that the proposal I made doesn’t involve actual beliefs at all. You can have justification for a proposition you don’t believe, and having that does not require seeing the connection between the evidence and the belief. If one comes to believe the claim in question, it may be that this belief could only be properly based if one sees the relationship (though this is controversial), but my proposal was about justification itself, not the further complex property of holding a belief justifiably.

  14. Jon wrote:

    but I’m thinking it could be modified as follows. Epistemic justification constitutes prima facie justification that the belief in question is the right one to hold in one’s pursuit of the goal of getting to the truth and avoiding error.

    Yes, I agree your proposal can be modified in that way and still retain bite. I wanted to push you in that direction.

    but I’m not sure it is necessary, since I’m not sure about your example. I think the crucial question about your case is how the authority is perceived by the epistemic agent.

    This may help. As before, suppose that E on its own justifies you in believing P to degree 0.95. Suppose it also gives you p.f. justification to believe that you’re justified in believing P to degree 0.95. In the absence of any further evidence, that’s what you’d be justified in believing tout court about your epistemic state.

    The authority comes along and tells you that you’re only justified in believing P to degree 0.2. You trust this authority to some degree, but not entirely. His testimony T will undermine your justification to believe P by some degree, but not all the way down to 0.2. Let’s consider how T and E interact with respect to the second-order question, how justified you are in believing P. Imagine that, as a result of combining T and E, you’re justified to degree 0.001 in believing that you’re justified in believing P to degree 0.2; justified to degree 0.67 in believing that you’re justified in believing P to degree 0.7; and justified to degree .329 in believing that you’re justified in believing P to the original degree 0.95. This is a simplification, of course, but it’ll make my point.

    Now the “center of gravity” of your different degrees of justification concerning how justified you are in believing P is somewhere around .78. (That is, the x such that (x-.2)*.001 + (x-.7)*.67 = (.95-x)*.329 is approximately .78.) I’m prepared to accept that your first-order justification concerning P will coincide with _this_ value. That is, the net result of adding T to your evidence E will be that your justification for believing P goes down from 0.95 to 0.78.

    But 0.78 is not the degree to which you’re justified in believing you’re justified in believing P. In the simple example I gave, you’re justfied to degree ZERO in believing you’re justified in believing P to degree 0.78. In a more realistic example, you’ll have SOME degree of justification for each x between 0.2 and 0.95, to believe you’re justified in believing P to that degree x. But, just as in my simple example, the center of gravity of all that justification need not be located at the same places as its peaks. In my example, the peak is at 0.7, and the center of gravity is higher, at 0.78.

    In such cases, I think you’d be justified to a degree higher than .7 in believing P. Perhaps you’d be justified to degree .78. In a given situation, that may translate into your being justified tout court in believing P. However, the degree to which you’re most justified in believing you’re justified in believing P _is_ 0.7. In a given situation, that may translate into your being justified tout court in believing you’re not justified in believing P. If that’s right, then you can be justified tout court in believing P while being justified tout court in believing you’re not.

    In any event, I think we’ll have to allow some divergence between your justification concerning first-order questions and your justification concerning second-order questions. That’s why I think the modification Jon gestured towards is called for.

  15. On Jim’s example: in the prior comment, I conceded that the two values need not coincide (the object level claim and the meta level claim). I wondered whether the difference needed to be reflected in the cutoffs between justified and not justified. Your example is a coherent set of probability assignments, and it reminds me to honor the degree of subjectivity I want in a theory of justification. So even though I can’t yet imagine a scenario in which the role of testimony works as this example specifies, I have no grounds on which to deny its possibility. So I think the right way to go here is to make the connection between justification and meta-justification a prima facie one.

    Maybe some still think there are problems, though?

Leave a Reply

Your email address will not be published. Required fields are marked *