Wednesday, November 15, 2006

On the Reliabity Condition for Justification

Alrighty, I've been trying to think of how to formulate an argument for the point I was trying make earlier today at lunch. I'm stuck, but not in a way that is entirely detrimental to my position. Let me explain.

Consider two scenarios. In scenario UNICORN, S sees a horse with a misshapen ear and mistakenly thinks that it is a horse with a horn. He coins the name "unicorn" for its kind, and says, "That's a unicorn." He later comes across several other horses with misshapen ears and each time he thinks that it is a horse with a horn and thus says, "That's a unicorn, too." Now, after this, S discovers that everything he called a unicorn was just a horse with a misshapen ear. In my view, the right thing for him to say here is, "Well, I guess those weren't unicorns after all because unicorns actually have horns." Here, though the property of having a horn was not had by any of the objects S called a "unicorn," still the property of having a horn is essential to being a unicorn. And both the fact that he thought the horses he saw had horns and that he called them "unicorns" on the basis of that perception suggests to me that having a horn is essential to something's being a unicorn, despite the fact that nothing S ever saw actually had this property.

Now consider scenario WHINNY. Here, for the first time S hears the sound of a whinnying horse. He's never seen a horse before this, let's say. Now, let's say that every time he hears the whinnying sound, he finds out that a horse is making the sound. Now, let me just stipulate here that for this illustration, at no point does S come to think that what is essential to being a horse is to make that sound. However, he picks out horses on the basis of figuring out what satisfies the description "makes a whinnying sound." And it works perfectly because there happen to be no other animals that make that sound, let's say. Here, the property of being that which makes a whinnying sound is the basis for identifying instances of horses, but it is not essential to what a horse is (even in S's own mind).

Given the coherence of both types of scenarios (indeed, I think both kinds happen all the time), I'm not sure it's possible to construct an argument from the premise "S recognizes instances of X on the basis of property P" to the conclusion "Therefore, P is an essential property of X." It just seems to me, though, that it's right to say in UNICORN the property of having a horn was a basis for identifying instances of unicorns (or what he thought were unicorns) and that there is a way in which that property was fit to serve that role because it is an essential property of unicorns. However, in WHINNY, it's not right to say that the property of making a whinnying sound, though the basis for identifying instances of horses, is an essential property of a horse.

Now, when it comes to knowledge, my claim is just that when I've claimed to have a justified belief (and I'm using "justification" for the third condition for knowledge, hence it is an essential condition for knowledge by definition), I did so in virtue of thinking that beliefs based on certain mental states/experiences are based on states/experiences which reliably connect to the truth. I identified instances of justified beliefs on the basis of their being formed on the basis of experiences that I implicitly but vitally assumed were reliable. Were a belief formed on the basis of an experience I did not think was reliable, I wouldn't have identified that belief as an instance of a justified belief. (Note, this consideration implies not just that an indicator that a belief was a justified one was only that I thought it was based on reliable evidence, but that it is in fact based on reliable evidence. Whether I was actually in position to know that a belief is based on reliable evidence is beside the point. The point is that I took it for granted that my mental evidence was reliable, and that was a crucial factor in determining which beliefs to call "justified," viz., those that were based on that kind of evidence. To be justified was to be based on states/experiences that were not simply thought to be reliable, but in fact were. And my calling beliefs "justified" reflects that I just took it for granted that I could tell, wrongly or rightly, what kind of beliefs were based on such evidence.)

OK, so it's clear (to me) that the apparent exemplification of the property "based on evidence that reliably connects to the truth" was what I partly based my judgments on as to which beliefs were justified and which weren't. Now, where does this leave us? Well, you could make this analaogous to UNICORN or to WHINNY. You could claim that my identifying certain beliefs as justified beliefs on the basis of their appearing to be based on evidence that was reliable is like S's identifying instances of what he thought were unicorns on the basis of the essential property "being a horse with a horn," or you could say that it's more like S's picking out horses on the basis of the exemplification of the unessential property "makes a whinnying sound." I don't have an argument for why it has to be the case that it is more analagous to UNICORN than to WHINNY. I'm inclined to think there is one that could be constructed, but for the time being it is all I can claim that at least in *my* case, it is clear to me that the right analogue is to UNICORN. And that was demonstrated when I heard Feldman's twin scenario (You and your BIV twin) and found my first reaction to be, "Hmmm... they both are internally rational/reasonable/responsible,etc, but it's not at all clear that they are both knowledge-level justified."

One final point. Maybe I could at least say this in favor of an analogue to UNICORN. In WHINNY, if I told S that I found a horse that doesn't whinny, there'd be no incoherence in saying that to him, since whinnying is not an essential property of horses, even to him. But in UNICORN, if I told S I found a unicorn that has no horn, he'd say, "That's impossible." Similarly, my first instinct when you tell me that a person could have a justified belief based on mental states/experiences that doesn't reliably connect to the truth is to say, "Huh? That's incoherent. Internally rational, yes. But third-condition-of-knowledge-justified? That can't be." Hence, what I was calling "justified belief" had as an essential quality that the basis for the belief was a reliable basis. (But of course, I could be the weird one in all this, though I honestly don't think I am relative to the masses.)

6 Comment(s):

  • You know, after all that I wrote, having thought about it a bit more, I think the basic point is just this. We’d agree (I presume) that if there is a property P such that anything that is an X must have (i.e., have in all possible worlds in which X exists) P, then P is an essential property of X. The property of being a horse which has a horn is an essential property of being a unicorn such that anything that doesn’t have that property is not a unicorn.

    In my UNICORN scenario, the fact is that S would not have called something a unicorn (rightly or wrongly) unless he thought it was a horse with a horn. For S, being a horse with a horn was not just a co-extensive description which happens to pick out unicorns. Rather, it is part of the intension of “unicorn” such that by definition, to be a unicorn is, in part, to have a horn (among other things). It the kind of property that was especially well fit for picking unicorns out because it constitutes the essence of what it is to be a unicorn.

    A good indicator that you have a property like that is when it just can’t be the case to you that an instance of X is had apart from an exemplification of P in any possible world. Note, there are certain other descriptions which could happen to allow you to identify rightly instances of X though they aren’t essential features of X (e.g., the property of whinnying in WHINNY), but those are properties which you, upon reflection, would say, “Oh, that’s not an essential property because there are possible worlds where X is had apart from exemplifying P.”

    OK, now, what about the reliability of one’s evidence as a condition for justification? It’s clear to me that when I form the belief that the lights are on based on the perceptual experience of the light’s being on (i.e., being appeared to lights-onishly), that strikes me as a justified belief. Why? Well, in large part it’s because I implicitly assume that my perceptual experience is reliably connecting me to the truth. Now, is that assumption merely just a co-extensive description that helps pick out instances of justified beliefs? I think not. Rather, it is the kind of assumption which if not made would prevent me from claiming my belief is a justified one (in this world or any other). Reliability is a without-which-not kind of property. Those kinds of properties are essential properties.

    So here’s the argument:
    1) P is an essential property of X iff the absence of P (or an exemplification of P) entails the absence of X.
    2) The absence of reliable evidence entails the absence of a justified belief.
    3) Therefore, reliable evidence is an essential property of justified belief.

    Now, before you go hooting and hollering about how (2) begs the question we’re trying to answer, note that my argument for (2) is that for every individual, were they not to believe (at least dispositionally) that their, e.g., perceptual experience was not reliably connecting them to the truth, then they would not consider the belief based on that perceptual experience justified. This is no different than how S would not call something a unicorn unless he believed the object had a horn. Now, note that this doesn’t mean that in order for an object to be a unicorn requires S’s belief that it has a horn. The point is that the fact that S would reject something’s being a unicorn unless it had a horn indicates that having a horn is an essential property of the kind of thing S has in mind when he calls something a unicorn.

    Similarly, the fact that I wouldn’t have called a given belief justified unless it was based on experiences that I thought were reliably connecting me to truth is an indication that what a justified belief is, in part, essentially involves having reliability.

    By Blogger kdfkwak, at 11/16/2006 9:49 AM  

  • John,

    This is (of necessity) going to be very quick, but here are my few initial comments.

    You suggest a number of times that you take your beliefs to be justified in large part because you implicitly assume that your evidence for your beliefs is reliably connected to the truth. I think I agree that we do tend to implicitly assume that, but I think that I also deny that this is a large part of why we (or at least I) take our (my) beliefs to be justified, for all of the (pretty strong, or so it seems to me) reasons that follow. It seems to me taht I take my beliefs to be justified because they are the right, or fitting, responses to my evidence. Given everything that I have to go on, my beliefs are justified when they are "adequate" to these very things that I have to go on (in some sense, I couldn't do any better). Whether those things that I have to go on then additionally connect (in a reliable manner) to the truth is another matter entirely. Certainly, if my evidence does not connect to the truth in that way, then there is something deficient about my epistemic state—I at least don't know many things (given that knowledge entails truth). But again, this is different from justification. And, indeed, I think that most people are with me on these intuitions; these are, in fact, the very intuitions which get the New Evil Demon problem for reliabilism started in the first place. The point of that problem is that the demon-worlder intuitively has justified beliefs, but his evidence does not reliably connect to the truth. I take the widespread reaction to this problem—indeed, even many reliabilists must see that there's something wrong with saying there's no justification here, because they go on to modify their accounts specifically in response to this problem—as substantially strong evidence that (contrary to what you seem to be saying here) being a response to evidence that is reliably connected to the truth is just not an essential property of justified beliefs. So, if you take it to in fact be an essential property, then I think you're going to have to do some massive denying of intuitions (or so it seems to me).

    This means I'm denying premise (2) of your argument. Again, for the reasons I've just cited (and for intuitive reasons), that seems pretty plausible. Sure, maybe an essential property of really, objectively good evidence is that it reliably connects to the truth in some way. But beliefs are just (supposed to be) responses to that evidence, whether the evidence tracks the truth or not, and such beliefs are justified when they fit that evidence. At least, those are my (quick) thoughts right now.


    By Blogger Jason Rogers, at 11/16/2006 2:17 PM  

  • I've (unexpectedly) got some more time, so here's a clearer way (maybe) to see it.

    Take the UNICORN example. There, you seem to say that subject S takes object X to be a unicorn because S implicitly assumes that X has property P, where P is something like "being horse-like and having a single horn on its forehead." Further, you contend that any object which does not have P is just not a unicorn. You get this latter claim by pointing to S's (and our) intuitions regarding what is and is not a unicorn in some possible world(s)—presumably the intuition is supposed to be that there is no unicorn that does not have P.

    Now take justified beliefs. There, you seem to say that subject S takes object X to be a justified belief because S implicitly assumes that X has property P, where P is something like "being based on evidence (or experience) that is reliably connected to the truth." Further, you contend that any object which does not have this P is just not a justified belief. You (attempt to) get this latter claim by pointing to S's (and our) intuitions regarding what is and is not a unicorn in some possible world(s)—presumably the intuition is supposed to be that there is no justified belief that does not have P.

    It is this intuition I deny, for the reasons I've already given. Further, I take it that the widespread reaction to the New Evil Demon problem for reliabilism is strong confirmation that most others deny this intuition as well.

    By Blogger Jason Rogers, at 11/16/2006 2:50 PM  

  • Sorry for the comment flood, but I need to point out that the word "unicorn" in the last sentence of the 3rd paragraph of my last comment is supposed to be "justified belief."

    By Blogger Jason Rogers, at 11/16/2006 2:51 PM  

  • I'm inclined to think you're taking a property of knowledge and spreading it to justification.

    I also think your taking a concept of evidence that applies to superstructural beliefs and spreading them to basic beliefs.

    We are accustomed to thinking of "evidence" as positive correlations between signs and what they indicate. This is probably because we typically take our basic beliefs for granted and don't think much about their reliability. We probably just assume that in some way.

    However, the concept of evidence is more broad than that. Evidence is what makes evident, or clear. Smoke makes pretty clear there's fire, it purports to *reveal* to us that there is fire.

    But our experiences purport to reveal the world to us. If our world is a normal world--as we hope it is--then our experiences *do* reveal the world to us, make some of its properties clear to us.

    So in this world, or evidence is reliable. Epistemizing evidence is always reliable. However, if, through no fault of our own and in a way totally indiscernible to us, those signs are misleading, we remain justified in following those signs.

    Some people have found this analogy helpful. Some street signs are *wholly* conventional, like an orange triangle signaling a slow-moving vehicle. But other street signs are more like what used to be called "formal signs" in that the are intrinsically connected with what they signify, they where their intelligibility on their sleeves, like a slippery road sign with the little picture of the car going side to side (or maybe the men and women's symbols on the sign for bathrooms).

    Evidence for basic beliefs are like this latter kind of sign, the mental states which constitute our basic evidence are an intelligible match with what they evidence.

    Evidence for superstructural beliefs are based on these and consist of correlations which don't depend on intelligible match, like the correlation between smoke and fire.

    In fact, the intelligibility match between basic evidence and basic beliefs is so close that some have had circularity worries about basic justification.

    The fundamental principle of common-sense epistemology (of which Chisholm was champion--is that an experience as of there being an F provides a reason for thinking there's an F.

    By Blogger Trent_Dougherty, at 11/18/2006 12:21 PM  

  • Interesting blog you got here. I'd like to read something more about that theme. Thnx for sharing this info.
    Sexy Lady
    English escorts

    By Anonymous Anonymous, at 12/15/2009 8:54 AM