Subjective and Objective Justification

Here’s a scene lived by all of us much too often. Some student tries to avoid some epistemological conclusion by distinguishing what is objectively true from what is subjectively true (e.g., that you can’t know that the earth is flat because that claim is false). I used to cringe at the response that purported object of knowledge is subjectively true, at least on the inside, but have become somewhat inured to it because it is so common. It’s bad enough, though, that I want the cringe response back.

Then I find epistemologists talking the same way about justification. Should we cringe here as well? Let’s see.

When one distinguishes objective and subjective justification, notice what one is not doing. Some theories of justification or rationality are subjective: Lehrer’s and Foley’s, for example. Some cringe at such theories, but I think we should rejoice! In any case, calling a theory subjective or objective is perfectly fine. But the distinction between objective and subjective justification is not the distinction between kinds of theories of justification. Instead, the purveyors of this distinction mean something different. The might mean the following: there are standards for justification, and a belief is subjectively justified when the person in question takes the belief in question to have met the standards.

There’s nothing grammatically wrong with qualifiers that block the following qualifier-dropping inference:
X is F-ly Y, or, X is an F-modified Y.
So, X is Y.
We have lots of instances of this phenomenon: decoy ducks, former senators, previously pristine landscapes, fictional entities, etc. But the inference is enough of a “quick and dirty” rule that careful philosophy ought to find a way around language for which the inference would be a mistake.

The reason such care is important is this: taken as above, subjective justification is no more a kind of justification than subjective truth is a kind of truth. So if this is what the distinction comes to, then subjective justification is not a normative property of a belief. So it can’t sneak in a different kind of normativity regarding belief when one’s favored approach to the normative status of belief turns out to be subject to objection based on subjective factors that the approach doesn’t take into account.

A specific example may help. I’ll use as an example a theory that nobody believes (I think): a theory that identifies justified belief with true belief. Call this theory, uhhhm… errr …, the SIMPLE theory of justification and its defenders SIMPLEtons. Counterexamples to SIMPLE abound, but SIMPLEtons may try to dance around the problems by pointing out the distinction between subjective and objective justification, holding that the distinction allows a defense of their theory. How does it do that? I can’t really tell, but I think it can only work if our sense of the normative acceptability of beliefs that runs contrary to SIMPLE is handled by the notion of subjective justification deployed. But, if subjective justification is nothing but the taking of a belief to be justified, it is not a normative property of the belief in question, but merely a further psychological state with the target belief as content. So it is not itself a kind of justification and is not itself a normative property of any belief, and thus doesn’t explain the normative acceptability involved in the counterexamples to SIMPLE.

All of this is predicated on the idea that the distinction between objective and subjective justification is the difference, roughly, between perceived and actual justification. There are permutations to be found here, however. Perhaps, for example, the distinction is between actual justification and normatively adequate perceived justification. In such a case, subjective justification is the justified taking of a belief to be (really) justified. That is a normative property of a belief when it obtains, but notice that it can’t obtain on the SIMPLE view above, for that would require that the taking in question is true. Versions of this problem plagues weaker, reliabilist theories as well–this point is just a generalization of the new evil demon problem.

So there is a dilemma here for the objective/subjective distinction. To the extent that the distinction can address the permutations of the objection that are easy to construct, it can’t itself be a normative property and won’t be a kind of justification. To the extent that it is a normative property, it will be easy to see how to adapt the objection to make the problem re-appear. That leaves, as far as I can tell, nowhere to go for the distinction to help at all.

So my bottom line is this: ‘subjective justification’ is either shorthand for “a subjective theory of justification” or it is a normative feature a belief can have only to the extent that it cannot block certain kinds of objections to a theory. What’s left for those who still want the notion to do more philosophical work is to join the company of those who talk of subjective truth as if it were a kind of truth.


Subjective and Objective Justification — 31 Comments

  1. As far as I can tell, G.E. Moore was a kind of simpleton. He didn’t think that justified belief = true belief, but he did hold that the only possible justification of an action is that by it the greatest amount of what is good would be realized. So when it came to justified action, Moore identified justified and right action. Ross was a simpleton in a sense too. He accepted the equation of justified and right action disagreeing with Moore about the list of characteristics that make right acts right. Rightness was not, as Moore thought, simply a matter of the production of consequences, there were duties besides the duty of beneficence. Still, for Ross, justifiedness was rightness.

    If the problem with the simpleton’s view is that it identifies justified belief with correct or right belief, it isn’t in terrible company. It isn’t just that there were some prominent writers who have tended to identify the justified and the right, it is that it is obscure what justifying something amounts to if it isn’t a matter of showing that the thing is right. So, if we think that the aim of belief is truth, there is a not particularly difficult argument for one part of the simpletonian view: you cannot carry out a successful justification of something that isn’t right. Only true beliefs are right. J –> T.

    So maybe the problem with the simpletonian view is that it accepts T –> J. Moore and Ross both seemed to think that in carrying out a course of action one couldn’t reasonably assume to be right, this opened the agent up for certain kinds of criticism that did not carry over to a criticism of the action (i.e., the agent could be blamed for carrying out a course of action that ‘turned out’ to be right). They seemed to endorse the move from right to justified. I don’t see why a simpletonian shouldn’t do the same. Anyway, if rightness entails justified and truth entails rightness or correctness, then T –> J.

    Maybe what is wrong with the simpletonian is that she assumes that correct or right belief just is a matter of true belief. Insofar as this isn’t an uncommon assumption, I don’t see that the simpletonian view is as silly as one might think. One could try to defend it by appeal to this idea of subjective justification, but the simpletons I’ve read in ethics were quite concerned to get rid of this silly idea of subjective justification. Those who are willing to accept the equation of right or correct belief and true belief but don’t want to end up simpletons I think have to show that in trying to show that a belief is justified one is trying to show that it is right, but what then is the aim of trying to justify a belief? Why think justification doesn’t ‘go’ with permissibility, rightness and the rest?

  2. if we think that the aim of belief is truth, there is a not particularly difficult argument for one part of the simpletonian view: you cannot carry out a successful justification of something that isn’t right. Only true beliefs are right. J implies T.

    Well, the argument may not be difficult to produce, but there are lots of bad arguments like that. I won’t bother to repeat all the counterexamples you’ll have to swallow to accept the argument; I’m sure you know them. In light of them, however, no one ought to grant that you can’t have a justification for something that is false. So the argument, even though not difficult to formulate, is in desperate need of a defense of its first premise. Furthermore, running the analogy with justification in matters of moral action is unpersuasive. First, we needn’t say the same thing about epistemic justification that we say about moral justification. Second, moral correctness is more plausibly taken to have a “for us” aspect to it than is truth in general. That’s why antirealism is more plausible in ethics than in metaphysics (though I’m not here embracing or rejecting the view in either domain).

  3. The counterexamples I assume you allude to include the intuitions underwriting the new evil demon argument and these do clearly elicit the intuition that the subject is justified in accepting a claim. Externalists tend to think that it doesn’t show much. I myself tend to take the intuition seriously, but that’s why I think externalists must try to say something in response to your argument that purports to show that personal justification carries with it doxastic and propositional justification. We should all agree that there is a sense in which ‘S is justified in believing p’ is true in virtue of facts about S’s subjective states. Without breaking the link between the justification of the agent and the attitude, I think their view is sunk by the counterexamples you mention. But I happen to side with Bach and Engel and think the argument can be countered so perhaps we can bracket that for now.

    The claim is that a _successful_ justification of a claim involves (among other things) establishing that it is right or true. It does appear that the agent’s intention in justifying a claim is to show that it is true. That should count for something. This is supposed to be on par with the claim that Ross held that a _successful_ justification of an act involved showing not only that in performing the action, the agent met the demands of some prima facie duty but rather that the agent met the demands by doing what there is overall reason to do. For Moore, it involved showing that the act would in fact produce the right consequences. These writers did not hold that one could not give a partial justification for something that fell short of rightness so I can’t see why the simpleton should be saddled with the claim that there is not the slightest bit of justification that goes with believing a false claim.

    I can’t tell whether you think epistemologists should reject: (a) the idea that right or correct belief entails true belief; (b) to succesfully justify is to show right or correct; (c) to attempt a justification is to attempt to show right or correct; (d) all of the above.

  4. We can grant, if you wish, that a successful defense of a claim shows that it is true, so there can’t be a successful defense in the absence of truth. The proper response, then, is to deny that successful defenses in the sense are necessary for justification.

    The counterexamples could be as exotic as evil demon ones, but I was thinking of much more mundane examples. There was a time when it was rational or justified to think that the earth is flat; there was a time when it was rational to think it an open question whether alchemy was a legitimate science; there was a time when it was rational for me to think that I was going to be a professional baseball player, etc. (That last one was a joke, by the way…)

  5. A quick question.

    taken as above, subjective justification is no more a kind of justification than subjective truth is a kind of truth. So if this is what the distinction comes to, then subjective justification is not a normative property of a belief.

    Why not? Mightn’t there be other normative properties of a belief other than justification? And mightn’t those be of interest in certain sorts of evaluations (perhaps the sort that lends intutive support to whatever counterexample is in question)?

  6. Ian, on the understanding of ‘subjective justification’ in question, what it means to say a belief is subjectively justified is that the belief is believed to (really) justified. Such a meta-belief is not itself a normative property of any sort.

    It may be, however, that it may be the ground or basis of some normative property, but it would take an argument to show that such is true.

  7. “Perhaps, for example, the distinction is between actual justification and normatively adequate perceived justification. In such a case, subjective justification is the justified taking of a belief to be (really) justified. That is a normative property of a belief when it obtains . . .”

    Except in this case “subjective” justification is a kind of objective justification. Or, so it looks. ‘Subjectively justified in taking’ I take it means subjectively justified in believing. So to be subjectively justified in believing that p is justified is just to have the actual justified belief q that p is actually justified. But that’s about as far as it goes. If justified belief were closed under T, JJp–>Jp (it pretty clearly isn’t) we’d have subjective justification for p entail objective justification for p. But T licenses the inference prohibited above, viz. that whatever you’re justified in believing is true. Still, subjective justification for p (of the sort here) is the objective justification that p is justified. So it just is a kind of objective justification. But I think that is just to say that there isn’t any subjective justification.

  8. Mike, that’s exactly right. There would still be a distinction between justification for p and justification for think that p is justified, but that’s probably not what anyone was after in distinguishing subjective from objective justification.

  9. Well, I should have read the rest of the post 😉 The above comments are exactly the sort of normative property I had in mind. So maybe I can elaborate. Here’s two different properties we might identify:

    1. S justifiedly believes that p
    2. S jusifiedly believes that (1)

    As properties of belief p, (1) and (2) are distinct. Mike is right that (2) is objective justification for believing that belief in p is justified, but it makes some sense to call it subjective justification for belief that p (esp. when (1) does not obtain). Here’s a case where a reliabilist might invoke subjective justification to explain a counterexample:

    S uses an extremely unreliable method of calculating circle areas, but looks perfectly justified in using that method (S learned it from an expert, S is normally a very good method-pickers, whatever), and S takes himself to be so justified. I take it many would want to claim that S is justified in his circle area beliefs, and so much worse for the reliabilist who claims he isn’t. But here the reliabilist might claim that S IS subjectively but NOT objectively justified in his circle area beliefs.

    The reliabilist might further claim that subjective justification, while not as desirable as objective justification, is still a very valuable property for a belief to have. And it’s often the only way we can evaluate the epistemic status of someone’s belief: we often don’t have time to determine the reliability of some new method of belief formation, but if we have prima facie reason to trust the person who chose the method, we can regard it as justificatory. And, since the reliabilist takes justification to be truth conducive, subjective justification has value WRT the goal of truth: if one is justified in taking oneself to be justified, then it is likely that one is justified, in which case it is likely that the proposition in question is true.

    But maybe this is a real abuse of the way people typically invoke subjective justification?

  10. Ian, you’re missing the point of my post, I think. The question isn’t what someone might say, but what they might mean when they say that a belief is subjectively justified. If you look at your last comment, you’ll see that you never give an answer to that question. To this point, I’ve suggested two options: they mean “S believes that his/her belief is justified,” or “S is justified in believing that his/her belief is justified.” Neither one will do the worked needed, and if you think there is some other way to interpret the language of subjective justification, you need to say what it is.

    There is, of course, a third way I mentioned: you might mean that there is a theory of justification which is subjective, and it’s an implication of that theory that a belief is justified. You can’t go this route, however, without telling us how the theory works. This is my preferred route, and then on my value-driven approach to the discipline, we should be asking value questions such as you ask in your last paragraph. My sense, however, is that those who distinguish subjective from objective justification have not wanted to provide a theory and have to defend it.

  11. I should qualify my last remark, because John Greco is a clear example of a person who tries to say what the theory of subjective justification is. That approach has two virtues: it avoids the analogy with subjective truth, and it takes on the burden of clarifying and defending the theory.

  12. That helps, I think. I was trying to identify the meaning of ‘subjective justification’ with (2), and give an example of how (2) might help a theorist. So I suppose what’s unclear to me is why it won’t do the work needed (forgive me if I’m being obtuse here!) — it looks like it’s helping in the example I mentioned. I can see how it wouldn’t help a reliabilist with the new demon scenario… but why should one distinction have the burden of explaining all putative counterexamples? Maybe the new demon scenario is particularly relevant because it’s a case where we clearly want to say that, at the very least, people (so long as they’re behaving as they actually do) would have many subjectively justified beliefs?

    Thanks for the post, this is interesting!

  13. Ian, I see better now what you’re thinking. You’re right that different problems may call for different solutions, but the one in question in the example I used is someone trying to defend SIMPLE by appeal to the notion of subjective justification. That won’t work, and wouldn’t work if we replaced SIMPLE with a very simple reliabilist account. For other problems, such meta-justification may be a good solution, but I can’t, off the top of my head, think of any that it would work for. In the circle area example you use, for example, it isn’t clear that the meta-belief (that the first-order belief is justified) is justified. What you say is that the choice of method “looks justified” and that the person “takes himself to be so justified.” But these claims are compatible with the meta-belief being unjustifed, and hence with the belief not being subjectively justified in the sense in question.

  14. What you say is that the choice of method �looks justified� and that the person �takes himself to be so justified.� But these claims are compatible with the meta-belief being unjustifed, and hence with the belief not being subjectively justified in the sense in question.

    Fair enough. Let me try to be more clear. The pre-theoretical intuition about the case is that S is justified. The reliabilist says S is not justified in his belief that p because the the process that led to the formation of p is not reliable. Then why is our intuition that S is justified? Because S also believes that the belief that p is justified, and because S reliably formed this (meta) belief. S reliably formed this (meta) belief by, for example, accepting expert testimony to the effect that the method (process) in question is a good one for calculating circle areas. So:

    1. S believes p on the basis of an unreliable process
    2. S believes that S’s belief that p is justified on the basis of a reliable process (testimony or whatever)

    Given our simple reliabilist thesis that a belief is justified if and only if it was formed on the basis of a reliable process, we get:

    1*. S’s belief that p is not justified
    2*. S’s belief that S’s belief that p is justified is justified

    And we explain the intuition by saying that S is subjectively justified, taking that to mean what (b) says. We can then acknowledge the importance of subjective justification to epistemic evaluation as I suggested, which explains why our intuitions might have confused it with (objective) justification — an obviously valuable, but distinct, property.

    This sort of case doesn’t really seem that uncommon to me, and it does seem to pose a problem for reliabilism. But I think there’s still some unclarity in my analysis of it, and it may be that once its sorted out better, the belief in question may not be subjectively justified.
    Here’s a schematic and explicit way to state the case:

    Process M is an unreliable method of forming circle area beliefs
    Process MM is reliable belief forming method (accepting expert testimony)
    Process R is some sort of reliable combination of memory and reflection
    Process I is deductive inference, a reliable belief forming method

    Belief x = a particular circle area belief
    Belief y = a belief that M is a good method
    Belief z = a belief that x was formed on the basis of M
    Belief z* = a belief that x is justified

    1. S believes x on the basis of M
    2. S believes y on the basis of MM
    3. S believes z on the basis of R
    4. S believes z* on the basis of I (taking y and z as input)

    From 1 (and the reliabilist thesis), it follows that S’s belief that x is not justified. But from 2, 3, and 4 (and the reliabilist thesis), it follows that S’s belief that z* is justified. Which is just to say that S is objectively, but not subjectively, justified.

  15. Jon,
    I’m not sure quite how to generalize the following (and you might well be tired of talking about it) but there do seem to be a range of cases in which JJP entails Jp. So the inference from “I’m justified in believing that I am justified in beliving p” to “I’m justified in believing p” does hold for a range of cases. But I am not sure how to specify the limit in any nice way.
    So, suppose that I will select an item from bin B1 and I am justified in believing that I will select a red item from B1 just in case the probability that I will select a red item from B1 is greater than .5. Suppose there are two bins B1 and B2 and the total number of red items in B1 and B2 is equal to the total number of white items in B1 and B2. Case #1 (trivial) I can assign a probability = 1 to the proposition that B1 contains more red than white items. It follows that I am justified in believing that I am justified in believing that I will select a red item from B1. And it does follow from that that I am justified in believing that I will select a red item from B1. Case #2. The probability that the distribution of items is such that B1 contains greater than .7 red items is .8. In this case I am justified in believing that I am justified in believing that I will select a red item from B1. But also the probability that I will select a red item from B1 is .56. So I am also justified in believing that I will select a red item from B1. Obviously, JJp will not entail Jp unless the probability= x of Jp and the probability= y of p both exceed .5. And JJp does seem to entail Jp just in case xy > .5.
    Of course these cases are simple and they do assume independence and so on. I suppose they assume too (implicitly) that we are assigning probabilities to objective chances. Things become interestingly more complicated in cases where both probability assignments are just credences or where there is no objective chance.

  16. Mike, very nice examples here, and I don’t know when I’ve ever gotten tired of talking about epistemology…

    I’m going to post something on levels confusion in epistemology, in part because of concerns of this sort. In short, a good theory of justification has to respect some interaction of levels, and one way to do this is to try to describe the restrictions on the entailment you mention. In general, the entailment would be surprising, since justified false beliefs are possible (so it looks like one ought to able to be justified in believing that one is justified in believing p, and simply be wrong, leaving JJp, but ~Jp).

  17. Ian, I think what you’re after here in the notion of subjective justification is what Sosa talks about under the heading of reflective knowledge (vs. animal knowledge). I agree that a reflective grasp of one’s epistemic position is valuable, but I don’t think this reflective grasp is explained in terms of the reliability of a meta-belief concerning the reliability of a first-order belief. The reflective understanding we want requires a more systemic feature than that; it requires something more akin to what coherentists think about when they describe the nature of justification.

    “All of this is predicated on the idea that the distinction
    between objective and subjective justification is
    the difference, roughly, between perceived and
    actual justification.”

    Do you mean “perceived” as (a) “merely perceived” in the sense of opposed to “actual”, or (b) does “perceived” mean “merely perceived OR actual”?
    If (a) then cringing is OK. Being “merely perceived as justified” then ENTAILS not being justified. It is not non-normative, but negative normative!And the mistake is like the student mistake with “subjective truth”. But not much interesting follows, as far as I can see, since we know in advance that “subjective” in this sense entails lack of justification.
    If (b) then further distinctions, like Ian Evans’s become relevant. but I don’t think (b) is what you meant?
    Am I being wrong?

  19. Hi Nenad, no, (a) isn’t what I meant, since it is not the relevant analogue of subjective truth. It is true that students start talking about subjective truth when the claim in question is false, but when quizzed, they assent to using the language of subjective truth whether or not the claim in question is really true.

  20. “The[y] might mean the following: there are standards for justification, and a belief is subjectively justified when the person in question takes the belief in question to have met the standards.”

    I should think that would be approximately right, although I’d phrase it in terms of the satisfaction of the epistemic standards of a particular epistemic community. And I should think that that in some sense really isn’t justification at all, since it doesn’t bridge the seems/is gap. It just says, “Well, the connection between belief x and state of affairs y *seems* to be there.” It doesn’t say anything about whether or not the connection really *is* there. That would be the job of objective justification. The assessment of objective justification is a metaphysical assessment, while the assessment of subjective justification is an epistemic one.

    Subjective assessment is what is used to generate Gettier counterexamples, since it is what people really have in mind when they say that justification can be mistaken. If Smith sees Jones driving a Ford, and has always known Jones to drive a Ford, and has just yesterday seen Jones’s driver’s registration for a Ford, and yet it later turns out that Jones had sold his Ford and was just renting the one Smith saw him driving, is Smith justified in believing [and I *really* prefer the term “tentatively accepting”] that Jones owns a Ford? Of course not–not objectively. Smith’s evidence increases his epistemic likelihood that Jones owns a Ford, and the increase, apparently, suffices to satisfy readers of Gettier’s article, so that Smith is considered subjectively justified in his belief [tentative acceptance] (or perhaps Gettier’s readers, getting his point, simply grant that the increase in epistemic likelihood is sufficient to render Smith subjectively justified); but objectively–read “metaphysically”–such justification just doesn’t exist.

    This, for me, is the difference between objective justification and subjective justification. Am I misreading your question?

    Keith Brian Johnson

  21. Keith, your use of objective/subjective here isn’t what I have in mind here. As epistemologists use the distinction, Smith is objectively justified, on the basis of perceptual and background information, in believing that Jones owns a Ford. Your use of objective and “metaphysical” suggests you wish to identify objective justification with truth, though I may be wrong about that. In any case, if that’s not right, then I don’t know what you intend here.

    The path you take in your first paragraph is the respectable path of offering a theory of some epistemic notion that counts as appropriately subjective. The theory you offer, however, has obvious counterexamples to it. For example, it suffers from the “reformer’s objection” that plagues simple versions of moral relativism. A given person can know that the epistemic standards of a particular community are wrong, and can knowingly violate them in forming and holding beliefs. In such a case, there’s nothing wrong with the beliefs in question.

  22. What it sounds like you have in mind by “objective justification”–and I could very easily be wrong, of course–is justification that even an omniscient observer would say counted as justification, even though it might be mistaken. He’d say yes, you had good reason for your belief–however he might define “good reason.” Somehow, he’d evaluate the relation between your perceptual information and background and your beliefs and would say, “Yes, I agree, you were justified.” What I’ve had in mind is the further stipulation that you have factually sufficient reason for belief–that you be in no way in error. I don’t (yet) understand what else objective justification could be, really. How is it decided whether Smith’s perceptual and background information justify him in believing that Jones owns a Ford, if it isn’t either by reference to the actual state of affairs, on the one hand, or to subjective standards, on the other?

    (Perhaps it would help to have the motivation behind the distinction I make: justification comes in degrees. You can be really, really well justified in believing something, or you can be really, really badly justified. As long as there’s the possibility of error, your justification cannot be perfect–i.e., you might turn out to be right, but your justification still only provides partial support for your knowledge claim. Your justification confers epistemic likelihood, not knowledge. Only perfect justification would *fully* support a knowledge claim. What degree of justification supports a knowledge claim in the first place–what degree of justification will count *as* justification–will be determined by the epistemic community of which one is part; normally, that’ll be the best we can do, and our epistemic community’s determination of epistemic standards will then tell us what is [subjectively] justified. The only way to be objectively justified, rather than merely subjectively justified with respect to epistemic standards, would be to have perfect (non-partial) justification. But perfect justification would seem to entail a match between the actual state of affairs and the knowledge claim made. So it would *entail* truth. Exactly what perfect justification would require–whether, for instance, to be perfectly justified would require that your justification logically imply your knowledge claim–I haven’t quite decided. I’m not sure how well I’ve put this, but this is at least part of the motivation behind my characterizations of objective and subjective justification.)

    I looked up “reformer’s objection” on Google but couldn’t find anything. Please forgive my ignorance, but I’m unfamiliar with that term. What does it mean?

    “A given person can know that the epistemic standards of a particular community are wrong, and can knowingly violate them in forming and holding beliefs. In such a case, there’s nothing wrong with the beliefs in question.”

    I don’t know what it would mean for a given person to know that the epistemic standards of a particular community are wrong. He might *think* they were wrong; he might *disagree* with that community’s epistemic standards; but how could he *know* they were wrong? In fact, I conceive of epistemic standards as arbitrary, in some sense: we can choose our standards, and we might be able to give reasons for so choosing, but ultimately, we do simply *choose* them.

    But even if he *could* know that the community’s epistemic standards were wrong, and then knowingly violated them in forming and holding beliefs–well, of course there’d be nothing wrong with his holding the beliefs in question, assuming that he held the beliefs in accordance with the epistemic standards for justification of his own epistemic community (and, of course, he could be an epistemic community of one). This is subjective justification, after all–why should anything be wrong with his applying his own epistemic standards? I guess I’m failing to understand your objections. Sorry to be dense–and if you’re kind enough to explain further, thank you in advance.

    Keith Brian Johnson

  23. Ah, I misunderstood a sentence: “As epistemologists use the distinction, Smith is objectively justified, on the basis of perceptual and background information, in believing that Jones owns a Ford.” But how is Smith fully justified in believing anything more than that Jones probably owns a Ford?

    Keith Brian Johnson

  24. Keith, the reason against inserting a probability notion is this: if you try that here, you’ll have to endorse it everywhere, since the evidence can be made as strong here as it is anywhere. It is simply implausible, however, to think that you’re not justified in thinking that there are other people who exist (but only that there probably are).

  25. But it’s absolutely *not* implausible to think that you’re not *fully* justified in thinking that there are other people who exist. My belief that there is an external physical world is, as far as I can tell, a metaphysical commitment that I make, along with the concomitant epistemic commitment to being able to gain reasonably reliable information about that posited world via my senses. Once I’ve made such commitments, I can then make knowledge claims about that world, but always under the implicit assumption that that world really exists (alternatively, I could restrict myself to knowledge claims about my own mental phenomena [or, to avoid reification, my own mentally phenomenalizing], but the language involved in doing that gets really tedious really fast). Under those assumptions, Smith’s information gives me good reason to think Jones owns a Ford, but it doesn’t guarantee that Jones owns a Ford, and Smith knows perfectly well, if he thinks about it at all, that it doesn’t give him such a guarantee, so if he phrases his claim carefully, he should claim only that it’s epistemically likely that Jones owns a Ford. And epistemic likelihood, I think, isn’t vulnerable to Gettier-type counterexamples, although I haven’t given the matter my complete attention yet.

    If you just use the term ‘justification’ as though people either *are* or *are not* justified, well, then, yes, you’ll run into problems.

    Keith Brian Johnson

  26. Keith, perhaps it will help to note that you don’t really think we know very much. Don’t say we have knowledge on an assumption–that’s not knowledge. And skepticism is the cost of denying that ordinary justification is adequate for knowledge, in combination with true belief, in non-Gettier cases. The choice is simple: we either don’t have knowledge or there’s a Gettier problem. (Put aside the issue of knowledge of mental states here.)

  27. You’re right: I *don’t* think we know very much, where the word ‘know’ is construed in an absolute sense. It seems to me that there are two uses of the word ‘know’: the first is, with respect to propositional knowledge, justified true belief (or, as I prefer, tentative acceptance), but where justification cannot be mistaken (thus denying Gettier’s first premiss); the second is epistemically likely justified tentative acceptance, where the kind of evidence that counts as conferring epistemic likelihood, and the degree of epistemic likelihood required before one labels a claim ‘knowledge’, is a matter to be decided by the epistemic community (one might, of course, constitute an epistemic community of one, and different epistemic communities may make different decisions and therefore count different claims as knowledge claims). Neither of these, I think, is subject to Gettier-type counterexamples, and it seems to me that the two reflect the ways in which the term ‘knowledge’ is used: the former sense is non-provisional and involves actual truth; the latter sense is merely provisional–you’ll change your mind if you find evidence to the contrary–but involves not actual states of affairs but merely our perceptions and beliefs about the way the world works.

    What I *haven’t* worked out, and perhaps what is intended here, is what those standards of epistemic likelihood–what counts as good evidence, and how high the bar should be set before claiming knowledge–ought to be.

    Yes, I am a philosophical skeptic with respect to the non-provisional sense. I don’t see how I can really know anything but my own presently-experienced mental phenomena–I can’t even know that my memories are really memories and aren’t just presently-experienced mental phenomena that have “echo” sensations attached–that *seem* like copies. In this strong sense, I can’t even know there’s a past.

    But I’m happy to use the provisional sense most of the time, in ordinary life. Thirty years ago, I would have said, “Jupiter has twelve moons,” and I would have intended it as a true statement about reality. But the fact that it turned out to be false says nothing about knowledge; it only says that I phrased it badly. I should have said, “Jupiter has twelve known moons in 1975.” Similarly, the fact that one can phrase claims in terms of knowledge instead of in terms of epistemic likelihood says nothing about knowledge; it only says that we phrase ourselves badly. Or so it seems to me. Your and my intuitions do seem to differ.

    Keith Brian Johnson

  28. I’ve just now noticed that my dashes, which I type as double-hyphens, are showing up as single hyphens. As are, of course, my intended single hyphens. I hope everyone is able to tell which are supposed to be which!

    Keith Brian Johnson

  29. Keith, here’s some things I know: that I have two kids, that I administer this blog, that today is Friday, etc. When you talk about provisional and absolute senses, I say: never give up the more obvious for the less obvious. It’s more obvious that I know I have two kids than that there is some absolute sense in which I don’t know that.

    There is a truth in the neighborhood, however, that has a connection to knowledge, but only through a bad theory: I’m not infallible about any of this.

  30. Jon–maybe that’s all the skeptic really wants: the admission that we’re not infallible about much of anything; the admission that we could *conceivably* be wrong about the way the world is, even about what may seem obvious. That we could conceivably be wrong doesn’t mean we think it’s likely that we’re wrong, of course, or that the skeptic is going to say that you should stop identifying your wife and kids *as* your wife and kids. From a functional standpoint, you know perfectly well who they are, and the skeptic will agree (I hope). The skeptic, I think, pushes the demand for certainty as far as it will go; what I would say is that virtually everything people say they know, they really only *think* they know.

    I *don’t* think philosophy should lead us to doctrines that defy common sense unless it’s necessary. I can’t prove to my inner skeptic that the outside world exists, in the form of proof my inner skeptic demands; nevertheless, I accept that there is an external world. To my inner skeptic, I put it this way: I make the metaphysical commitment that there is an external world and the corresponding epistemic commitment that I can learn about that external world via my senses. But in ordinary life, I speak the same way you do–I know who my mother is and I know who my father is and I can identify them from photographs, and so on. In ordinary life, I take empirical evidence to be good evidence.

    But I *do* think a lot of confusion arises from different people’s using words like ‘belief’ (the force of religious conviction? tentative acceptance as true on the basis of available evidence? attitude toward living one’s life [pragmatic or functional “belief”]?) and ‘knowledge’ (what you have when you read textbooks or look around you? what you have when you examine your own mental phenomena? what you have when you draw inferences that you know might be mistaken? what you have when there can be no conceivable doubt?) and ‘justification’ (beyond any doubt? beyond a reasonable doubt? by a preponderance of the evidence?; metaphysical [an omniscient entity would “see” the justification] or epistemic?). The usage differences quickly lead to doctrinal differences and to people’s talking past each other, and I really think that if people are to understand each other, they’re going to have to first make very clear what the different usages are and how they’re going to distinguish among them.

    What’s the bad theory you mention?

    Keith Brian Johnson

  31. Add the words “in different ways” at the end of the third paragraph’s first sentence. Sorry!

    Keith Brian Johnson

Leave a Reply

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