Monday, May 19, 2014

Non-Normative Epistemology?

I've previously cited epistemic nihilism as an example of an intellectual black hole -- a view that rational agents must reject in order to preserve their capacity for rationality.  But the normative nihilist might seek to avoid this implication by offering up a non-normative account of rationality (and epistemology more broadly). One could just stipulate a certain extensional account of "rationality" -- perhaps consisting in conformity to norms of parsimony, logical validity, certain inductive/abductive norms, etc. -- without requiring these norms to be backed by irreducibly normative properties, or to give rise to categorical "oughts".  This modest nihilist says, in effect, "Here are my recommendations if you wish to join me in the game of truth-seeking.  But if truth is not your goal, I have no grounds on which to criticize you."

My main concern about this move is that it isn't clear how the nihilist can consistently regard her own preferred epistemic norms as more truth-conducive than any others (at least so far as non-deductive norms are concerned). I've previously noted that anti-skepticism requires us to regard some possible worlds as (a priori) objectively more likely than others, and that to explain rational induction (e.g. the projectability of "green" but not "grue") requires positing objective structure (something empiricist-inclined nihilists may also regard as unacceptably mysterious or metaphysically "queer"). If all possible epistemic norms are metaphysically on a par, as the epistemic nihilist seems committed to, then why regard any particular set of norms as more likely to lead to truth than any other?

Normative realists consider their preferred norms to also be metaphysically privileged in comparison to other possible epistemic norms and practices.  The counter-inductivist and the grue-speaker are, we think, objectively irrational.  They may consider themselves rational, but they are mistaken.  The nihilist cannot think this.  If "rational" in their own mouth picks out some stipulated set of norms that they just happen to endorse -- rational_green, say -- then "rational" in the mouths of the grue-speakers just picks out a different stipulated set of norms: rational_grue.  Rational_green and rational_grue are on a metaphysical par, according to the nihilist; it's just that some communities use the term "rational" to pick out one rather than the other, but that's just an arbitrary semantic fact of no epistemic significance.  The normative realist, by contrast, posits not just a semantic but also a metaphysical difference between the two sets of norms.  Rational_green is privileged in a (language-independent) way that rational_grue is not.  The realist may be wrong about this, of course, but my point is just that insofar as they believe this, it is not unintelligible that they continue to use rational_green norms and consider them more truth-conducive than rational_grue norms.

The key point then, is that the nihilist's continued attachment to the rational_green norms seems entirely arbitrary by their own lights. This renders their epistemic practices unintelligible, in contrast to the realist (who may or may not be correct, but at least is internally coherent).

* My thanks to grad students BK and SS, whose opposing papers on this topic prompted me to revisit my thoughts on the issue.  Let me also mention that SS elaborates on the "arbitrariness" objection to the nihilist in a much more sophisticated way than I have in these rough thoughts!


  1. Andrew Sepielli9:34 pm, May 19, 2014

    So there are two types of metaphysical posits mentioned in this post, and two explananda, and I worry that you're explaining what I take to be your target explanandum with the wrong posit. I can understand the claim that we need to posit objective structure to explain why projecting "green" is more likely to yield true predictions than projecting "grue". I don't know if I'm quite willing to buy it, since I'm quite resistant to the idea of objective structure, but still, I can see the force of the argument. And I can understand the claim that we need to posit properties of epistemic rationality (but not epistemic counter-rationality) to explain why rationality is objectively privileged over counter-rationality. This is the McPherson/Enoch argument, if I'm reading you right. Now, again, I don't buy McPherson/Enoch's argument at the end of the day, but I can see its force.

    What I guess I don't understand is why you would think that we need to posit irreducibly normative epistemic properties (or any other normative epistemic properties) to explain why projecting green is more likely to yield true predictions. Can't I just be, like, an epistemic quietist -- heck, even an epistemic nihilist -- but still say, just on the basis that green is more natural than grue is, that projecting green is more likely to yield true predictions?

    1. Hi Andrew, a couple of thoughts on that. One was that a major motivation for nihilism (namely, radical empiricism) seems likely to also lead one to reject objective structure. (But you're right that it's then the rejection of structure rather than of epistemic normativity that is doing the actual work.) My second thought was that even given the existence of objective structure, it's a substantive epistemic norm that claims that we should expect more natural predicates to be more projectable.

      To bring out this second point, suppose that it actually turns out that emeralds (etc.) are grue, not green. Come 2020, all newly observed emeralds (etc.) will appear blue to us. Since projecting green turns out not to have yielded true predications, in what sense was it previously "more likely" to do so? The only available sense, I think, is a kind of epistemic probability, i.e. what credence was antecedently substantively rational. But, the thought is, epistemic nihilists lack the resources to make sense of any such epistemic notion of probability, and hence to make sense of our inductive practices (of the sort either Hume or Goodman drew attention to).

  2. Hi Richard. On your first point: Well, again, I reject both objective structure and (non-minimal, explanatorily potent) properties. But in my limited experience, I find that most people reject, e.g., Ted Sider's view because they find the notion of structure unintelligible; they understanding what he's trying to do with it, and so they can think of structure as "whatever plays roles X, Y, and Z" but they don't have an independent grasp of it. By contrast, most of these same people would not, I suspect, have trouble understanding the notion of a normative property. (Or at least, they would not *take themselves* to have trouble understanding the notion.)

    On your second point -- good, I was going to say that, while "we should expect more natural properties to be projectable" is a substantive epistemic norm, "projections of natural properties are more likely to yield true predictions" is not epistemic -- it's metaphysical. But then you say that the probability mentioned here must be epistemic, and you explicate epistemic probability in terms of epistemic rationality. I guess I'm inclined to resist here, for reasons that relate (believe it or not) to some discussions we've had about moral uncertainty. For I think that while, yes, there is a kind of probability deserving of the label "epistemic" that reduces to facts about epistemic rationality, there is also a kind deserving of the same label that does not. This is epistemic probability, statements about which express (rather than report) credences. This has been discussed a lot by philosophers of language and formal epistemologists (e.g. Seth Yalcin, Sarah Moss, and Eric Swanson), and I've found the notion invaluable in discussions of moral uncertainty.

    1. Oh yeah, that was me, Andrew.

  3. Hi, Richard,

    A couple of points:

    1. I'm not convinced by the characterization of “normative nihilist” for some of the people holding views such as the ones you describe, since they may well hold that there are norms – even if they do not agree with the irreducibility of normative properties.

    For example, someone might consider that the property of being water is reducible to the property of being H2O, but that would not make them nihilists about wateriness, and someone who is what is often called a moral naturalist is may well believe that there are moral norms – even if they are reducible.

    I don't have a good word for someone defending the sort of view you describe, but I will call them ENR (for “epistemic normativity reducibilitist”), just to give them a name and avoid the implications of nihilism.

    Side note: By the way, a person might reject reducibility (either in the epistemic or the moral case) if they also reject reducibility of water to H2O, etc. (based on a different analysis of the word “property” and/or how ordinary language describes the world, for example), without positing anything that, say, queer in the sense in which Mackie would use the word.

    2. On the issue of whether the attachment of the normative “nihilist” to the rational-green norms is arbitrary by their own lights, I don't see why that would be so. It seems to me there are a number of potential objections. Similarly, there seem to be a number of potential objections to I will have to what the ENR would say – she may make “ought” claims of epistemic normativity, just as a moral naturalist can consistently make moral “ought” claims.

    I will have split the posts, because explaining the potential objections takes a bit of room.

    1. I think I should have said "reductionist" instead of "reducibilist", but the point is the same.

      Also, it seems to me she may consistently say you're being irrational if you're coming to believe things by a process that fails to meet the conditions of whatever epistemic normativity or rationality reduces to - you rationally ought not to be a grue speaker.

      I do not see any inconsistency there; I see this as similar to a position that is usually called “moral naturalist”, which seems consistent afaik.

      So, I don't think she has to limit herself to the invitation you mention.

      I will address a couple of issues in more detail my next reply.

    2. Regarding whether different people mean different things by “rational”, an ENR may hold that the properties that norms of epistemic rationality reduce to are the same for all humans linguistic communities that ever existed on planet Earth.
      She may further posit (though she needn't account for it to be consistent) that this is a feature of human earthling psychology that resulted from the evolutionary process, given that – it turns out – such features were more conducive to reproductive success in the ancestral environment, precisely because they tracked truth in such an environment.

      She might also further suggest (even if, perhaps, more tentatively) that sufficiently intelligent alien speakers would also have a word picking either the same property – whatever epistemic rationality reduces to -, given how the universe (as a matter of contingent fact) is and was, and given that (this particular ERR claims) different ancestral environments, while different on many different issues, were not so different in those respects.

      Still, she might leave room for aliens picking other properties, especially if the universe happens to be infinite (i. e. to have infinitely many galaxies, stars, planets, etc.)

      So, if some human earthlings were to be grue speakers, they would be – on an account of this sort - irrational in their assessments, and they would be mistaken if they were to consider themselves rational.
      Now, if, say, as a result of some quantum freak event, a community of grue-speakers came to existence on a distant planet (ala Boltzmann brain, but the whole thing rather than just the brain), then they would still be irrational, and if they use some word that plays a role similar to “rational” in our language and call themselves “rational”, it may well be that they would be mistaken, but it might not be (it depends on how similar it is, and what the requirements of such similarity might be; the ENR needn't take a stance on that in order to be coherent).

      She would still say that beings of that sort, in a universe like ours would more frequently be wiped out of existence (the sense of “frequently” might need to be relativized to finite but growing volumes in the universe, if the latter is infinite), precisely because – given how our universe is – they would happen to fail to track truth more frequently.

      If the ENR also a similar reductionist with respect to moral norms, she may hold that in that case, plausibly there is much greater variation in norms across alien species in our universe, given relevantly different ancestral environments.

      Regarding truth-conducive norms, she may hold that, as our universe happens to be, our human earthling norms of assigning probabilities is conducive to truth more frequently, and we ended up with them in the first place, and that the same happens on other planets (except, perhaps, for some freak events if the universe is infinite or finite but sufficiently big, or very infrequent conditions).
      However, if there is not only a universe (finite or infinite) but a multiverse with, say, infinitely many parallel universes (perhaps some of them finite, other infinite, or each of them finite or infinite), it may well be that in some parallel universe, other norms are conducive to truth more frequently.
      Even if there is no multiverse, in other possible worlds, other norms would be more conducive to truth seeking.
      Of course, she grants that not any set of norms will do; for example, norms that systematically lead to contradiction would not work anywhere.

  4. Richard,

    Now I'm not sure I misunderstood how much a person needs to reject in order to meet your criterion for epistemic nihilism.

    I was under the impression that denying irreducible normative properties was enough but now after further thinking about this and a couple of previous posts of yours, I'm not sure I'm on target.

    Could you clarify, please?

    At any rate, as long as she holds that epistemic rationality reduces to some properties, but that being rational is more frequently conducive to truth - and it's so as a contingent matter in our universe, but the standards are the same for our species and perhaps all other intelligent species -, that may avoid any inconsistencies by her own lights.

    However, that still requires acknowledging a shared human standard (of course, some humans might reject such standards explicitly, but even then would have to have the same basic standards instinctively, except perhaps for severe mental illness), so rational persuasion would be possible in principle.

    If your nihilist is committed to holding that the standards may vary from individual to individual so that rational persuasion is not doable (even leaving aside mental illness or things like that), then I agree that their epistemic practices appear incoherent.

    1. By 'nihilist' here I mean someone who explicitly denies that there is any genuine normativity. That is, a Mackie-style "error theorist" who agrees with the non-naturalist about what it would take for there to be genuine normativity (it would take something over and above the familiar natural properties), but who denies that the world actually contains any such thing. The nihilist endorses the claim that "there are no categorical oughts" -- only hypothetical imperatives, which they interpret non-normatively as simply highlighting a means-end relationship.

      So my target here isn't the naturalist who thinks that they can capture genuine categorical normativity without needing any further ontology (beyond what the natural sciences demand). [Though it is true that I'm independently dubious of that view, and so do think that metaphysical naturalism ultimately entails normative nihlism, but I'm not arguing for that here.]

    2. Okay, thanks and sorry I got that wrong.

      So, how about the following nihilist variant? (or a similar one).

      1. Alice is a nihilist, she offers a non-normative account of rationality - as you proposed - and invites you to be rational - in that sense - if you wish to join her in the game of being rational.
      Alternatively, she may say that such an account is doable, but very difficult, and leaves the matter for future research.

      2. She holds that the term “rational” picks those properties, and does so in all human earthling societies, and (perhaps) all societies of sufficiently advanced aliens in the universe. We humans care about rationality (i. e., about those particular properties), because being rational (i. e., following those standards) happens to be truth-conducive in our universe (or any universe with its causal structure, or something like that), and that was overall conducive to reproductive success, so natural selection favored those who were rational in their assessments, all other things equal.

      Still, natural selection also favored other stuff (e. g., maybe sometimes conformity to some of the group's prevalent beliefs was better for reproductive success than truth-seeking, if the members of the group generally cared a lot about having those particular beliefs) , so humans aren't in fact always rational. But as long as you want to seek truth, rationality is the way to go, and she invites you to be rational.

      Also, he holds that if grue-rational speakers were to pop into existence on a distant planet via a quantum freak event, they would be bad at seeking truth.

      She may also claim that our language involving “ought” (whether moral “ought”, or “ought” practical or epistemic rationality) ended up (for whatever reason, she may say it's one of the cases evolution got it wrong and gave us misleading preliminary intuitions) having ontological commitments that result in an error theory, since no such normativity exists, or is even possible.

      It seems to me that her approach avoids the objection you raised, because while she's not claiming anything “queer”, or any metaphysically privileged norms, she does hold that, given the way contingently is, there is a way of assigning probabilities and/or forming beliefs – rationality, as extensionally defined, or in some similar way if her particular definition is mistaken; she needn't have an actual, specific definition, and may leave it to future research – that happens to be more conducive to finding truth in a universe causally like ours – but not in all possible worlds, hence no metaphysical privilege -, and we got a generally good enough but not perfect way to make those assessments, due to our evolutionary history.

    3. "being rational (i. e., following those standards) happens to be truth-conducive in our universe"

      My worry is: What grounds does Alice have for this claim? According to some alternative possible epistemic norms, given her evidence (current experiences and apparent memories), she should instead believe that she is a Brain in a Vat who popped into existence 1 minute ago and has been fed false memories and sensory experiences.

      Of course, we think those alternative norms are objectively unreasonable, and hence aren't inclined to take them seriously. My worry is, how can the nihilist dismiss them, since she cannot judge them unreasonable, and to judge them "unlikely to be truth-conducive" would seem to arbitrarily privilege the norms she just so happens to have started with. (The alternative norms are self-ratifying, after all: According to themselves, they are the most truth-conducive norms you could find!)

      To make matters worse, suppose that Alice really is a BIV. So in fact our standard inductive norms are not truth conducive. Still, I think they are clearly the rational norms to follow -- they are what one a priori should expect to be truth-conducive, whether reality turns out to cooperate or not. But I don't see how the nihilist can make any such claim.

    4. Nice points. I think that's a very interesting challenge to nihilism.

      I'll try to argue from the nihilist's perspective to play the devil's advocate in the next post (because it's a bit long).

      Also, I'll try to improve a bit the credibility of her case for the sake of making it more interesting - earlier I was just going for internal coherence -, but just to be clear, I'd like to say I'm not a nihilist myself.

      In fact, I think the hypothetical Alice ought not to take that position, if not for other reason, because I think some version of naturalism that is similar to her views but without the additional claim about the ontological commitments of “ought” language, is more probable than nihilism (i. e., should be assessed to be more probable).

      So, the devil's advocate:

    5. Alice:

      I start by trusting my faculties. I have no other way. I cannot jump out of my brain.
      Granted, I can challenge of my faculties from the perspective of other, intuitively stronger ones. But that's as far as it goes. I can't challenge them all. You and I are on the same boat in that regard, so that's not a problem for me, as it is not for you.
      Now, by my own lights, looking at the empirical evidence, reasoning, etc., I conclude that humans have, as a result of the evolutionary process, what we may call human-rationality, which is usually conducive to finding truth. That system is not perfect, but it's generally good enough, and it's also self-correcting to some extent, given sufficient time to gather info and pondering the matter.
      For example, if we find it intuitive that some object would probably behave in some way, but we see that objects of that kind (e. g., subatomic particles) regularly do something counter to our pretheoretical intuitions instead, we change our assessment about what objects of that kind probably will do next – using still our human epistemic intuitions to assess probability based on new evidence.
      So, what about smart aliens from other planets?
      Well, aliens evolve in the same universe, so due to evolutionary considerations, one would expect systems that are similar to rationality. They might not be the same, since their preliminary theoretical intuitions may well be considerably different in some cases, but if they're self correcting, it seems probable that at least frequently, alien rationalities will generally converge to beliefs similar to those human rationality converges to.
      Still, we shouldn't expect – and here I'm correcting my previous views a little – that, say, human rationality is exactly the same as the rationality as that of all species that evolved, or the best. Maybe, say, evolved species#18436433-rationality is slightly better than rationality at truth seeking in our universe.
      However, rationality is also pretty good, and it's the best we have; perhaps, it's slightly better at that that, say, evolved species#85534-rationality, even if the latter is generally pretty good too. On the other hand, freaks that result from quantum events likely to have faculties that are much less conducive to truth.
      That is a conclusion that I reach from my own standpoint.
      But my own standpoint tells me that being such a freak is much less probable in our universe.
      So, I coherently claim that those norms are the rational norms to have, and that they are in our universe generally conducive to truth.
      I can't step out of my brain/mind, of course, but on the other hand, I can at least give an evolutionary account of the general reliability of my pattern-recognition/truth-seeking faculties.
      But what if I can't?
      What if the Boltzmann-brain objection, given our present-day limited knowledge of the universe, keeps me from giving an evolutionary account supporting the reliability of my own faculties?
      In that case, you would be in the same situation as I am in that regard.
      Moreover, you say that our standard inductive norms are what one a priori should expect to be truth-conducive. But then again, that's your epistemic probabilistic assessment, from the perspective you contingently happen to have. Other possible agents say the same about their inductive norms, which are different from yours.
      So, metaphysically, you just happen to contingently have the norms you have, and they're not more metaphysically frequent that any other (what would that mean?)

      What happens with other possible agents seems no more of a challenge to my perspective than it is for yours; in particular, while some possible agents endorse other self-validating norms, for that matter some possible agents endorse other self-validating norms that they claim to be metaphysically privileged.

    6. Interesting response. In essence, where I point out that realists can hold their epistemic norms to be intrinsically metaphysically privileged, as the "a priori objectively reasonable" ones, you point out that even nihilists can at least consider their norms to be extrinsically privileged, in the sense of being "actually truth-conducive in a universe like (we take) ours (to be)". The latter judgment is unavoidably question-begging, sure, but so is the former that I'm happy to attribute to the realist, so what's the difference?

      I'm thinking there are two important differences.

      (1) We know that actual truth (concerning contingent empirical matters) is something beyond our direct epistemic grasp, and so the "leap" involved in just directly assuming that our norms are (actually) truth-conducive is one that should seem suspicious / unwarranted by our own lights. By contrast, any a priori truths of reason are, in principle, within the direct epistemic grasp of rational agents. So while it's true that we must (in a sense "blindly") trust our own epistemic faculties in judging that our preferred norms are "objectively rational", this assumption needn't seem "unwarranted by our own lights" because as a putative truth of reason it is precisely the kind of truth to which normative realists believe rational agents do have direct epistemic access. We then, indirectly, conclude that our norms are extrinsically truth-conducive, on the basis of our grasp of their intrinsic merit (rather than directly assuming this extrinsic claim on no basis whatsoever, as the nihilist must).

      (2) As a way of further bringing out this asymmetry, notice that the realist's fundamental claim (that our epistemic practices are objectively reasonable) can survive regardless of external circumstances. Even if we're all (unknowingly) BIVs, it's nonetheless objectively reasonable to believe that the external world exists as we take it to be. But the nihilist's fundamental claims (of actual truth-conduciveness) are falsified in such circumstances. I think this brings out that the normative realist's primary assumptions are, in an important sense, more "modest", and less "suspicious" (or difficult to coherently maintain on reflection), than the nihilist's. They are consistent with the full range of empirical possibilities consistent with our evidence, whereas the nihilist's claims can only get off the ground if reality is more or less just as they take it to be. So the realist can acknowledge that, while we may be mistaken to believe in the external world, we are at least certainly reasonable in doing so. But a nihilist cannot offer any such robust endorsement of her own beliefs: the only epistemic virtue available to her is "actual truth-conduciveness", which -- by her own lights -- her beliefs may well fail to possess (e.g. if we're actually BIVs, which is a possibility compatible with her evidence).

    7. The nihilist might reply as follows:

      (1). a. She can say that, in the colloquial sense of the word “objective”, there is an objective fact of the matter as to which norms are truth-conducive in our universe. There is even an objective fact of the matter as to whether some norms are truth-conducive in any possible world. For example, those that lead to contradictions all the time are necessarily not truth-conducive, whereas some other norms – like the priors we humans start with – are contingently truth-conducive, but still there is an objective fact of the matter as to whether they are truth-conducive in our universe, and so in the usual (colloquial) sense of the term, they're objectively truth-conducive.

      (1). b. She can then grant that our norms are objectively rational. They are not objectively species#1873437613-rational, perhaps. That other species has slightly different priors. And quantum freaks may come with vastly different priors. Also, there are possible species that evolved on other possible worlds would have wildly different priors too, which are conducive to truth – or not, depending on the possible world – in their world.

      So, her objection is not to the claim that our norms are objectively rational. She may also say that it would be irrational not to hold the priors rational agents have to be likely truth-conducive.
      For example, (though this is not needed), she may say that rational agents are (say) agents that are consistent, Bayesian-updater, and start with some set of priors that we humans got.
      On the other hand, species#1873437613-rational agents are consistent, Bayesian updaters that start with an incompatible set of priors.

      (1). c. She can argue that having rational priors is likely truth-conducive, but species#1873437613-rational agents hold that having #1873437613-rational priors is likely truth-conducive, and other agents may similar claims. All of them are internal claims, and so is hers. But there is no problem from her perspective in making likelihood assessments, as long as there is no claim of metaphysical privilege – other than ruling out, say, rules that lead to contradiction, but if ruling them out is a problem for the nihilist, then you can use that.

      (1). d. As for the claim of “intrinsic merit”, she may concede that our rational priors have greater intrinsic merit that the #1873437613-rational priors of species #1873437613, but point out that “intrinsic merit” is an assessment of rationality, not #1873437613-rationality.
      So – the nihilist holds – while rational priors have greater intrinsic merit than #1873437613-rational priors, #1873437613-rational priors have greater intrinsic #1873437613-merit than rational priors. That's unsurprising – she holds -, given that intrinsic #1873437613-merit is an assessment of #1873437613-rationality, rather than an assessment of rationality.

      (2). Her claim that our practices are objectively rational – but objectively not #1873437613-rational – can also survive different circumstances. She may say that if we're BIV, believing in the external world would still be rational. It would not be, perhaps, freak#1835-rational, but as a human, she cares about rationality, not about freak#1835-rationality.
      Her claims about truth-conduciveness do not survive BIV circumstances, but neither does your claim that our norms are actually truth-conducive.

      Given the above, I don't see why her claim is less modest than yours. You're claiming metaphysical privilege of sorts. She's denying metaphysical privilege. Her claims and yours seem to survive radical circumstances just as much as far as I can tell.


    8. Clarification: when I said “On the other hand, species#1873437613-rational agents are consistent, Bayesian updaters that start with an incompatible set of priors.”, I meant a set of priors incompatible with the set of priors rational agents have. Their set is internally consistent too.

      That aside, and regarding the likelihood that some set of priors is truth-conducive, nihilist Alice holds that the rational assessment is that that rational priors are - either generally, or always; either alternative seems compatible with her nihilism – truth-conducive. By “generally”, she may mean that it's at least very frequent that they're truth-conducive in the world in which the rational agent lives, or something along those lines.
      She may or may not hold – either alternative seems compatible with her nihilism – that, a priori, the rational assessment is that rational priors are [or are likely] generally more truth-conducive than any other priors.
      If she does hold that the rational a priori assessment is that rational priors are [likely] generally more truth-conducive than any other priors, she may also add that, a posteriori and given our knowledge of our universe, evolution, etc., the rational assessment still is that rational priors are [likely] generally truth-conducive, but not that they're likely more truth-conducive than any other priors – they're a lot more truth-conducive than freak-priors, but they might or might not be more truth-conducive than species#1873437613-rational priors, for some species that evolved or will some day evolve on some other planet, perhaps on some other galaxy.

      In any case, I don't see any obvious problem with this part of her views. What seems to be weak is her claim that “ought” statements are all false. A naturalist – for example – may hold a view similar to that of this nihilist in most respects, but without adding the extra condition that our “ought” language is committed to some metaphysical queerness that is impossible. Such claim places the nihilist in the difficult situation of having to both defend that semantic claim, and explain why evolution would give us those systematically false beliefs in this particular case. She may still consistently address both issues I think, but the naturalist would have a much easier time at defending the plausibility of some of her views, in my assessment.

  5. If all possible epistemic norms are metaphysically on a par, as the epistemic nihilist seems committed to, then why regard any particular set of norms as more likely to lead to truth than any other?

    Yes, but why critique the epistemic nihilist if any particular set of norms as more likely to lead to truth than any other? That's the meta of nihilism.

    Of course you might get some sort of dogmatic 'nihilist' who commits dogmatically to a set of norms - but I use scare quotes because they might not be nihilists at all, just religious or dogmatic (if I may distinguish the two)

    1. It's true that a thorough-going nihilist could grant my point that they lack a basis for thinking their preferred inductive norms to be truth-conducive, and just say "so what?" But their view is then going to be very unappealing to those of us who are committed to engaging in rational inquiry. Insofar as they previously hoped to make their view seem at least tolerably attractive to the rest of us, it would be significant if they must ultimately give up on that ambition -- if, at the end of the day, the most they can say in defense of their view is, "I don't care what you think!" (Just repeating "So what?" over and over seems unlikely to win them many converts!)

  6. Richard, it seems that once we try to think about how we could get a property such as grue, it would need to be a much more complicated property than green. Certain other independent variables would have to obtain in order for the property grue to be instantiated (would the wavelength reflected by said objects have to suddenly change? In order for that to happen, the energy levels of the electrons in the atoms comprising grue objects may have to change as well. Or alternatively, the way our eyes and brain perceives colour may have to change instead. Grue is a property that entails a change in subjective experience after some point t. Any story we tell about grue must logically entail such a change in experience. I may have to work on the details, but it seems that any story we tell which involves grue instead of green would have to be less probable than stories about green simply because they involve more independent factors (which logically could have been otherwise) In order to deny that, someone would have to deny deductive validity more generally in a way that would really be the sort of black hole that you pointed to earlier.


Visitors: check my comments policy first.
Non-Blogger users: If the comment form isn't working for you, email me your comment and I can post it on your behalf. (If your comment is too long, first try breaking it into two parts.)