Epistemic Value

Tuesday, August 15, 2006

Armchair luck

I have been reading Duncan’s Epistemic luck, and working on the case of a priori domain. Here is my question to Duncan

I assume that your characterization of epistemic luck for contingent proposition is the best worked-out available. And I tend to agree with your general idea that what makes an event lucky, is that it occurs in the actual world but does not occur in a wide class of the nearest possible worlds with same (or maybe very similar—my addition) relevant initial conditions as in the actual world. I want to discuss two closely related questions about how to make the general idea precise. You propose to do it in terms of safety concerning related worlds, but leaving the agent, and her belief in question as they are. I agree with the basic inspiration, i.e. to use counterfactual reasoning, but the question is how to do it. I have two worries, and I would like to discuss the first one in this post.

First, your proposal is explicitly limited to a posteriori cases. For candidate necessary a priori propositions there are no variations in worlds available. Here is the worry: the safety proposal does not smoothly generalize to other domains with minor changes of parameters. So, for armchair knowledge it has to be replaced by something else. It might seriously LACK GENERALITY. Let me illustrate: think of two mistakes in calculation or proof that cancel each other resulting in the correct solution. Suppose they are extremely hard to detect, so that the thinker is justified in trusting her calculation. She is Gettier lucky in her final belief.

Obviously, Jane’s justification is not up to the task. Nothing to do with the structure of external world(s), with counterfactual instability of environment. The problem lies in the thinker: her thinking is the locus of trouble. The counterfactual picture would have to be focused upon variations there, e.g. what would have happen it there hadn’t been a second mistake that killed the first one. So the proposal would have to be along the following lines: had the cognizer’s ways of thinking (or even her capacities) been slightly different, she would not have managed to arrive at the same true belief as in the actual world. She might have ended up with believing the negation of the target proposition, or with agnosticism about it. Let us concentrate upon ways of thinking. This would yield the following proposal:

Procedural veritic luck

It is a matter of luck that the procedure used by the agent has resulted in true belief.

The agent's belief is true and has been justifiably arrived at in the actual world, but in a wide class of nearby possible worlds in which the relevant initial conditions are almost the same as in the actual world—and this will mean, in the basic case, that the agent at the very least forms her belief in the sufficiently similar way as in the actual world—the agent arrives at a false belief (or no belief at all).

The corresponding positive requirement would be a kind of stability of the cognizer:

If an agent knows a priori a (necessary) proposition p, then, in most nearby possible worlds in which she forms her belief about p in a slightly different way or with slightly changed cognitive apparatus as in the actual world, that agent will also come to believe that p.(Agent Stability).

The formulation needs working out, unfortunately. For instance, if both mistakes are very, very subtle, then Jane would continue to make both of them in close possible worlds. My question for the moment is: is it on the right track?


  • At 1:04 PM, Blogger Duncan Pritchard said…

    Hi Nenad,

    Thanks for your interesting post about my book. You're right that I don't try to apply the framework that I develop in Epistemic Luck to necessary propositions, and the reason for this is that dealing with contingent propositions is challenge enough. (More precisely, my view is limited to what I call “fully contingent” propositions--i.e., propositions which are not necessary in *any* sense).

    My hunch as to how the framework would apply to necessary propositions was essentially along the lines you suggest here, though there is a key difference. The reason why the framework can't be directly applied to these propositions is because it targets the worlds where the proposition concerned (p) is false and, of course, there are no such worlds in this case. My thought was thus to focus on the worlds in which the agent forms a belief as to *whether* p, where this includes not just those worlds she believes p, but also those worlds in which she believes not-p. In cases of lucky true belief where the target proposition is necessary, one would thus expect there to be some near-by possible worlds in which the agent, forming her belief in the same way as in the actual world, ends up forming a false belief in not-p. (For example, suppose I think that tossing a coin can settle mathematical questions, and this method happens to result in a true belief in a necessary mathematical proposition. Although there is no near-by possible world in which I believe what I do on the same basis and what I believe is false, intuitively there is a near-by possible world where I form a belief concerning this proposition on the same basis and the belief is false--i.e., the world where the coin-toss leads to the wrong result).

    This way of approaching the matter will deal with lots of cases, and it's close in spirit to your proposal. They are different, however, since it makes no appeal to procedures that the agent might employ. I can see a rationale for appealing to such a constraint on knowledge, though I'm not sure that it should be thought necessary for knowledge that this constraint is met (as opposed to it merely being epistemically desirable that this constraint is met). Do you have an example in mind that would be dealt with by your proposal but not by mine?

    Incidentally, Ram Neta and Avram Hiller have argued in print that a proposal of this sort won’t work (this is in the forthcoming Synthese special issue on Epistemic Luck). Here’s the relevant passage note that they are dealing with contingent propositions, but this doesn’t matter for our purposes):

    “It might be supposed that Pritchard’s account only needs slight modification. Rather than claiming that for a believer to have knowledge that p, in all (or nearly all) nearby worlds in which the believer forms a belief that p in the same way, p is true, it might be claimed that for a believer to have knowledge that p, in all (or nearly all) nearby worlds in which the believer forms a belief as to whether p in the same way, the belief is true. For one to have knowledge on this account, there can be no (or few) nearby worlds in which one forms (in the same way as in the actual world) a false belief that p or a false belief that not-p. Indeed, that one cannot form a false belief that not-p does handle many of these cases⎯for example, if the thermometer had read 60°C rather than 50°C, Sakari would have believed that it was 60°C and would have at least had a false tacit belief in the claim that it is not 50°C in the box. However, not-p is just one of a large number of propositions that are relevantly similar to p such that, if one is to know any of those propositions to be true, then there are no (or few) nearby worlds in which one forms a false belief in one of those relevantly similar propositions, and forms that belief in the same way as one does in the actual world. There are other examples which show that consideration of the relevantly similar beliefs must not be limited to the negation of the proposition in question. For example, let’s say that the thermometer is actually correct and is accurately measuring the temperature inside the box. But, the thermometer is just a portion of a combination temperature/pressure/volume/mass/time gauge the display of which cycles through all the different measures, and, as it happens, all the other indicators aside from the temperature indicator are haywire. Had Sakari looked a moment later, it would have displayed a radically wrong pressure reading. So, she doesn’t know that the temperature is 50°C, even though there are no (or few) nearby worlds in which she has a false belief about whether it is 50°C in there. It is because there are (more than a few) nearby worlds in which she has a false belief concerning the pressure or the mass or the volume that she does not know the temperature in the box.”

    I haven’t yet decided what to say about this. My instinct is to think that I can hold the line, but perhaps there are even more compelling cases around the corner with the same structure.

  • At 10:02 AM, Anonymous nenad_miscevic said…

    thanks a lot for a very prompt and kind reply. I have two reactions, a short one and the longer one. The short one is that your nice coin-toss example dovetails with a problem of how exactly to characterize Jane’s reasoning. It is not like coin-tossing: the two mistakes she made are subtle (we need this in order for her to be justified), so the incorrect procedure she followed is the one that seems quite compelling to a good mathematician. Why then believe that in nearby possible worlds the procedure is changed at all? Well, we need some change, in order to capture the intuitive idea that it a matter of luck that the theorem is true relative to Jane's justification. So, we need to look at the procedure.
    The longer reaction comes as a new post. Thanks again

  • At 11:09 AM, Blogger nenad said…

    The Right Rpproach (the common denominator of our approaches here in the previous blog) introduces variations in believer’s states and/or doings-beliefs at least, if not also procedures. But this brings in its wake interesting consequences. In order to make them clear, let me re-use the old metaphor of cognitive “hitting the target”, and contrast hitting it by luck as opposed to skill. In the case of luck, the shooter might have easily missed the target. When the target is moving, like in hunt, it is the instability of the target that often accounts for the ease of missing. But many targets are quite immovable. To dramatize, imagine a dialog between an envious Göring and a proud Werner von Brown in the wake of first V-2 bombings of London. Von Brown is bragging, and Göring is downtrodden, since the rockets seem to be much more destructive than the bombs thrown by his ace pilots. “Oh, Werner, it’s a sheer piece of luck that your rockets hit London”, he says. Obviously, he doesn’t mean that London could have been elsewhere, say, far away in China. The target is fixed and it is the rockets that are alleged to be imprecise, or swerve, or whatever. They could have hit pastures instead of hitting the city.
    The case with necessary truths is like this imagined Göring case. They don’t “move”, don’t change across world, but stay the same. There is no moving target. So, in the case of armchair luck, it is the thinker that could have been wrong about the topic. It is her mind that could easily have swerved, “hitting” the negation of the theorem instead of the theorem itself. It is not the immovable necessary truth of a theorem that accounts for possibility, but the unstable human mind with its limited capacities.
    But once we notice the symmetry between the moving target and the “moving” shooter, we understand that exclusive focus with modal instability of truth(s) is unwarranted even in the a posteriori cases. It is often the thinker that could have been wrong about the topic, not the truth i.e, the world- that might have “shifted”. The unstable human mind with its limited capacities is often the main focus of our discontent, not any particular modal shiftiness of truths.
    Let us return to the usual Evil Demon topics, bearing in mind the symmetry between the mind and the world (i.e. world-) as potential contributors to cognitive disaster. Indeed, the skeptical problem for our knowledge of the external world is mind-centered: it is the manipulability of our minds (the Demon), and the poverty-underdetermination of the phenomenology that produce skeptical worries. The phantasy of a different, Demon-ruled world (world-, of course) is just a graphic device to tell the reader something about us and our cognitive weakness, rather than about demons, or about modal instability of truths.
    If this holds, then the characterization of in terms of safety is not only limited and not sufficiently general, but it can be seriously misleading, focusing the philosopher’s attention upon the wrong subclass of mismatches. Both external world luck and armchair luck have a common root, not capturable by considerations of safety, and a definition of epistemic luck is needed that would capture this common root.

  • At 11:53 AM, Blogger Duncan Pritchard said…

    Thanks for this Nenad. Just to clarify: are you suggesting that a simple modification of the view so that it talks of forming a belief *whether* p rather than simply a belief in p wouldn't suffice to resolve the problem (this isn't clear from your post)? If so, then what sort of example do you have in mind as posing problems for this sort of account?

    Incidentally, there's a parallel debate about epistemic luck taking place over at JanusBlog that readers of this thread might be interested in. Click here.

  • At 8:46 AM, Blogger nenad said…

    If we replace “that” with “whether” we create a space for VARIATION IN BELIEF, and not only in world (since there are no variations there anyway, when it comes to necessary propositions). No, once we do this, the FOCUS CHANGES, and we are now focused upon the thinker herself. It is here that a virtue epistemologist might again see a chance for stressing the qualities of the thinker, as opposed to mere hospitality of the world.

  • At 9:10 AM, Blogger nenad said…


    I would like to raise another issue. I have been quite impressed with the idea of reflective luck, and with the view that it is unavoidable in skeptical matter , all in Duncan’s Epistemic Luck . Then I read his “McDowellian Mooreanism” and started worrying.
    Question: What is exactly the character of beliefs involved in reflective justification? The question turns out to be very difficult, and much more should be said about it. Here is a beginning.
    Duncan defines Reflective epistemic luck
    Given only what the agent is able to know by reflection alone, it is a matter of luck that her belief is true. (EL:174)
    A natural and charitable reading, is internalist: the reflecting agent just looks “into her own head”; this reading yields the result Pritchard wants, i.e. that some reflective luck is unavoidable: indeed, I cannot, just by introspecting, find whether there is an external world.
    However, Pritchard has been moving towards disjunctivism recently, and has noted that this move raises an equivalent of McKinsey problem. He points out, that
    on the disjunctivist view gaining reflective knowledge of one’s factive empirical reasons is not a purely introspective affair<, in contrast to the classical internalist view. Thus, recognizing that one is in possession of a factive empirical reason, such as that one sees that such-and-such is the case, is itself a partly world-directed activity. And he reasonably concludes that the drawing of relevant consequences from one’s reflective knowledge in this regard will result in empirical knowledge. (McDowellian Neo-Mooreanism:19)
    I don’t see how one can defend the unavoidability of reflective luck on this disjunctivist conception of reflection (“disjunctivist reflection” or d-reflection):
    1 I assume that a McDowellian is introducing the whole disjunctivist apparatus to block the skeptical attack. Then, of course, the relevant kind of reflection for debate about skepticism is the d-reflection, and this seems confirmed in the discussion of “McKinsey problem” by Duncan himself. However, in the case of d-reflection, its factive ties are accessible to the armchair thinker. She can draw anti-skeptical consequences from what she is able to know by d-reflection alone. But then, it is NOT a matter of luck that my anti-skeptical belief is true. But in this case, the McDowellian Neo-Moorean is bound to claim that there is no significant non- disjunctivist reflection, and also that there is no disjunctivist reflective luck. Bad luck.
    2. If our McDowellian Neo-Moorean allows for some internalist reflection (-kind), it only introduces new complications. Since this kind leaves one helpless against the skeptic, so, if there are two kind of reflection available, one should turn immediately to d-reflection.
    I conclude for now that disjunctivist conception of reflection is incompatible with there being an important element of reflective luck in our anti-skeptical beliefs.

  • At 12:22 PM, Blogger Duncan Pritchard said…

    Thanks for this Nenad (and apologies for taking so long to get back to you, but I've been out of the office a lot lately). You've identified a big issue for me that I've been thinking a lot about lately. What I think is that the McDowellian thesis, while complicating the distinction between reflective and veritic luck, can nevertheless be made to be compatible with it. What we need to do is distinguish between two types of reflective luck--one that deals with non-factive grounds and one that is not so restricted. McDowell's proposal is therefore still able to allow reflective epistemic luck in the narrow sense. I actually think this is an advantage to the view, since we need to accommodate the intuition (which even McDowell has) that our knowledge is hostage to luck in some way and this is one way to do it consistent with the factivity of (some) reflectively accessible grounds. If McDowell isn't receptive to the distinction, then it starts to look spooky that he's able to get such a definitive response to the sceptic (indeed, as I say, I think he doesn't want such a definitive response since he's happy with the idea that our knowledge is in some sense hostage to luck).

    In the book, I argued that the sceptical issue was really an issue not about knowledge possession simpliciter but rather about the possession of a certain kind of subjective assurance which is only gained by eliminating reflective luck. If a subjective assurance is gained simply via reflectively accessible grounds, then it would appear that the McDowellian picture allows for such subjective assurance and so deals with this more fundamental type of scepticism as well. In distinguishing between different types of reflective luck, however, we also need to distinguish between different types of subjective assurance. The sceptical issue that is aimed at reflective luck focuses, I would argue, on the subjective assurance that one gets from non-factive reflectively accessible grounds since such grounds do not already entail the denial of scepticism. I suspect that McDowell simply isn’t interested in this form of scepticism--his target is solely on what it takes to know. Indeed, he may even agree with me that our epistemic standings are unavoidably subject to epistemic angst of this sort.

    What do you think?

  • At 5:55 PM, Blogger nenad said…

    I am impressed by the idea, and wish it a good luck. If it can be made to work, it would reconcile things that have to be reconciled. Here is a worry.
    The skeptic might insist that once we have distinguished factive from non-facctive reflection, we are back at square one. Here is an analogy. / I am following McDowell in his Criteria paper/ in drawing parallels between other minds and external world issues.
    Consider the following scenario, reminiscent of Bergman’s movies.
    Ingmar has just lost his ( de facto ) loving wife. He is tormented by suspicions an jealousy: did she really love me, was she really faithful. He finds her diary, that ( in fact ) makes manifest that she did faithfully love him. But he is worried: what if this is just the last and most perverse deception? What if she wrote it in order posthumously to deceive him? Now, the FACTIVE EVIDENCE Ingmar has is excellent; his wife is really opening her heart the best way a human being can do. Her love is making itself manifest to Ingmar, as McDowell would interpret it. If Ingmar could only bring himself to trust her, he would COME TO KNOW THAT SHE FAITHFULLY LOVED HIM. And he would have REFELECTIVE ACCESS TO A HIGHT GRADE FACTIVE EVIDENCE. However, his NON-FACTIVE EVIDENCE is, of course, not conclusive; nothing is.
    Now, consider your intuitions about the case, the skeptic proposes. Isn’t Ingmar epistemically (and, in consequence, humanly) in a miserable position? And isn’t his whole epistemic situation GOVERNED BY HIS NON-FACTIVE EVIDENCE REFLECTIVELY ACCESSIBLE to him? So, the skeptic might argue that between the two kinds of evidence, it is the non-factive one that epistemically wins. In other words, once the disjunctivist has admitted both kinds of evidence, the game is over, and old doubts are back. The promise of disjunctivism was precisely to get us out of the predicament; if it can’t deliver, why bother with it?

    PS. I am very happy that disjunctivism is being discussed in this way. In McDowell’s hands it was original but hard to grasp; with Tim Williamson it came closer to the usual style. With present discussions things are getting crystal clear.

  • At 10:36 AM, Blogger Duncan Pritchard said…

    Hi Nenad,

    The case you describe is very helpful, and in fact I think it neatly illustrates what I was trying to say. The kind of scepticism in play here is not, I think, scepticism about knowledge. After all, the agent concerned does indeed possess great (indeed, factive) evidence in favour of his belief in his wife's fidelity (I'm reading the case such that he does believe this, even though he has concerns). But the angst remains, and is not affected one iota by the presence of the sort of reflectively accessible factive grounds that McDowell has in mind. Thus (bringing the sceptical concern back into play), epistemic angst is unavoidable even if one has reflectively accessible factive grounds, but this is consistent with knowledge possession (one of the themes of my book). I'm inclined to think that even McDowell would grant this, since his says in a few places that we should learn to live with luck even while recognising that there is no sceptical argument about knowledge possession worth worrying about.


Post a Comment

<< Home