Tuesday, November 16, 2004

Are Evolved Perceptions Reliable?

Here's something I've been meaning to comment on for ages. A while ago I argued that moral properties mustn't be directly observable, for evolutionary reasons:

We would evolve to be attracted to objects or actions that promote evolutionary success, regardless of their 'intended' (by who?) moral status. Moral properties would be just another bit of data (alongside photons and soundwaves) for us to work with and interpret for evolutionary ends. Again, there's no reason at all to think that we would learn to distinguish which properties were intended to mean 'good' and which 'bad'. If intrinsically 'bad' things increased our evolutionary fitness, then we would come to (mistakenly) interpret them as 'good'.

So unless one wants to conflate 'good for my genes' with 'morally good' (not a good idea!), we must deny that objective moral properties exist independently, 'out there' in the world, waiting to be perceived by humans.

To which Brandon of Siris made the interesting response:
But suppose someone were to substitute 'perceived properties' for 'moral properties' and modify everything else accordingly? The conclusion would be that, while there are real facts of the matter, and while our cognitive functions are such as to promote evolutionary success, we have no reason to think our cognitive abilities map the world.

But note that I never suggested we wouldn't be able to accurately perceive (the existence of) moral properties. Perceiving them would be a prerequisite for our being able to use them as information for evolutionary ends. Rather, what we must be blind to is their 'intended' status.

Evolution implies that we should be skilled at perceiving various 'descriptive' properties out there in the world, but utterly blind to any normative ones. Or, to put it another way, we should have fairly reliable 'existence detectors', capable of distinguishing various physical phenomena, but any deeper 'meanings' should be inaccessible to us (at least by direct observation). If there were intrinsic moral properties out there, we might detect their existence (being a descriptive aspect), but not their goodness (the normative/'meaningful' aspect). We instead 'attach' the normative attitudes ourselves, having developed pro-attitudes towards objects that improve our evolutionary fitness, and con-attitudes towards predictors of evolutionary harm. (Thus the relevance of evolution to aesthetics - why we love a lush countryside but are repulsed by mould and faeces.)

What Brandon calls our 'perceived properties' (I assume he means things like colour, shape, distance, etc.) don't have any 'intended' normative status. (Or if they do, then we can't directly perceive it). But we should be able to detect the purely descriptive aspects of these properties just fine, according to my argument. So I don't think there's a slippery slope to skepticism here after all.

Perhaps Brandon is pointing out that what we see must not be what things "really" look like. (Qualia would seem to be a deeper 'meaning', rather like normativity.) To which I would entirely agree - the universe doesn't have any mind-independent visual properties. It doesn't intrinsically "look" like anything. But we can detect the existence of various wavelengths of light, and then interpret this in an evolutionarily beneficial way - giving us the appearance of colours (for example).

So yes, my argument implies that our qualia will not map to the objective reality - the way things ought to seem (look/taste/smell etc). But that's okay, because there is no "way things ought to seem". There's just the way things are - the detection of which is surely in our evolutionary interests.


  1. As I noted in my blog I dont think there are absolue moral properties - various moral principles will contrdict at various times and it will be entirely
    subjective which should rule.

    Pure theories like "live has intrinsic value - dont kill anything" or "you are responsible for all your actions and their most remote indirect effects" reach a point where evolution / or some other effect has to try to deny us the ability to live by them.

    Evolution will tend to work with those theories that are most consistant, though, because in a sense that is what is offered to evolution to chose from even though evolution will make use of self deception to cut them into a useful design. certain aspects may be easier to manipulate by these methods for example normative aspects? particularly when not atached to other important genes.

    Moral truths are if anything even less real than the colours we percieve and the normative asociations we push on them are side effects largely of habit (and of course a litle genetics) but deep analysis still forces consistancy so you can push closer to consistancy and remove hypocracy gradually by debate
    which is what moral philosophy presumably does. 

    Posted by geniusNZ

  2. Richard, I don't think you can deal with the problem this easily. Your response here puts all the weight on the 'intended' status; but the notion of 'intended' status doesn't play any role at all in the moral intuitionism you were criticizing. That was an argument that moral properties are perceived and are objective. If your response was just that we can't tell what is 'intended', in a robust sense, then it doesn't seem relevant to the view you were discussing. What does the work in the view put forward by Kiwi Pundit is not 'intended' status but 'objective' status. And this is where I think your response begins to run into problems, because it then becomes an argument not against perceiving intended status but against perceiving objective status. And this seems way too strong.

    The thrust of your argument was that our perception of 'good' and 'bad' as intrinsic or objective properties would be hijacked by evolution. But the argument only appealed to completely general elements of evolution, which, if that were the whole story, would mean that recognition of any objective properties would be hijacked by evolution in exactly the same way. And this would appear to be the same sort of hollow subjectivist to which, you argued, the moral intuitionist falls prey: if we follow the parallel, your conclusions would have to be that there are no intrinsic properties to anything, period. Your attempt to say that we can detect the existence of light and its properties would be as illegitimate as anything else; you wouldn't have any basis for thinking there were any way the world intrinsically or objectively was at all, by your own reasoning. If, however, you make the move you try to make above, then given that the issue was objectivity, there is nothing to prevent the moral intuitionist from making an identical move with regard to moral properties (whose existence we can determine and then interpret in an evolutionarily beneficial way). This is definitely a qualified moral inuitionism, but qualification is a far cry from the conclusion you said followed from your argument. So as far as I can see, you've painted yourself into a corner with this argument, as it stands: if you block moral intuitionism you block all knowledge by perception; if you save some knowledge by perception, you allow moral intuitionism to that same degree.

    There are two other ways it could go: one could deny the original argument, and say that we have no particular reason to think that evolution would hijack, in any serious way, perception of any objective properties, moral or otherwise; or one can refine the argument by appealing to evolutionary fitness in a way that would only apply to perception of moral properties, or at most only qualify other perceptions. But I think either of these would require changing the argument as it stands.

    So the problem with your argument, as I see it, is that to be relevant, and what you claimed it to be, it has to argue against the objectivity of moral properties; but it doesn't give us any reason to think that it wouldn't apply to any objectivity if it applies to that objectivity. As I said, as it stands it seems too strong and swift. 

    Posted by Brandon

  3. nice blog...try tis blog too at http://rock-your-boat.blogspot.com/ 

    Posted by derico

  4. Anytime I run into a discussion of perception and the world, I recommend (sometimes demand, depending on my relationship with my interlocutors) that everyone read "What the Frog's Eye Tells the Frog's Brain." Increasingly, cognition courses are requiring that students read this papers, and chances are if you've taken a course on vision, you've read it. However, I've only met a handful of philosophers who've said they've read the paper, and this is a tragedy. What it says about vision may be more telling than all of the philosophy of vision since Berkeley (with the exception of Gibson's The Perception of the Visual World). I'm being a little hyperbolic here, but only to stress the point that anyone who hasn't read that paper should.

    Here's the message about vision that you get from that paper: we perceive "objective" properties of the world, but only a tiny subset of them, the subset that holds ecological import for us. If we extend this analogy to objective moral properties of the world (assuming there are any), then evolution would imply that we only perceive those objective moral properties that hold ecological import for us as well. This leads to what is, at least for me, the unattractive implication that there are (or at least may be) moral properties of the world of which we are completely unaware (and completely unable to perceive, ever, given our current perceptual capabilities). I don't even know what that would mean, which is why I find the proposition that there are objective moral properties that are discoverable through perception a bit odd. 

    Posted by Anonymous

  5. The anonymous post was from me (sorry about that). I wanted to add that vision science has further demonstrated that our visual systems reliable pick out certain statistical properties of the world (e.g., edges and boundaries, depth cues, color cues, prototypical shapes, spatial frequency, etc.). The real philosophical question is not how reliable what we're getting is, but how much we're missing, and how much that limits the reliability of our overall picture of the world. I imagine the same arguments carry over to the perception of objective moral properties of the world. 

    Posted by Chris

  6. Chris - thanks for the link (I haven't done any cognition courses yet, though I hope to next year).

    Brandon - I'm not sure I follow your argument. I've been suggesting that we can accurately perceive (objective) existence, but any deeper meanings are 'attached' by us for our evolutionary advantage. How is that a problem for our everyday perceptions? How could it *not* be a problem for a directly-perceived objective morality?

    If all we could do was perceive the existence of some moral property or other, without knowing what it meant (e.g. 'good' or 'bad'), then we couldn't use this as a reliable foundation for our moral lives.

    You seem to be suggesting the same problem holds for physical properties, but I don't understand why this is. Lightwaves don't inherently mean anything, it seems to me that the mere perception of their existence is sufficient for our purposes. We can then interpret them however we want.

    "there is nothing to prevent the moral intuitionist from making an identical move with regard to moral properties (whose existence we can determine and then interpret in an evolutionarily beneficial way)"

    Sure, one could do that, but - as I point out in the original post - doing so is to confuse "good for my genes" with "morally good". This strikes me as something we should strongly want to avoid. (Put another way: If moral properties are interpreted qua evolutionary success rather than qua morality, then that is no morality at all!)

    By contrast, I don't think adding meanings (via evolution) to physical properties leads to any such unfortunate conflations. (If you think it does, I'd be interested to hear the specifics!) 

    Posted by Richard

  7. Richard, I don't understand what you mean by 'meanings' or why you think moral intuitionism somehow involves them. All moral intuitionism strictly requires is that we be able to perceive the existence of (at least some) moral properties.

    Also, when you talk about 'good for our genes' and 'morally good' you have provided no reason to think the two would be completely mutually exclusive. It is a fallacy, for instance, to think that because our perceptions of such-and-such are good for our genes it follows that they are perceptions of goodness-for-our-genes. Were that true, the only thing we would ever perceive is goodness for our genes. 

    Posted by Brandon

  8. "I don't understand what you mean by 'meanings' or why you think moral intuitionism somehow involves them."

    I take the 'meaning' of a moral property as whether it is a property of objective goodness or badness. An acceptable moral intuitionism must surely require not only that we can perceive the existence of two forms of moral property, but also that we assign them the correct meanings. If we merely 'saw' various moral properties, but confused their intrinsic meanings (say we thought good was bad and bad was good), then that seems problematic!

    I never said 'good for our genes' and 'morally good' are mutually exclusive (caring for your biological children, for example, is an obvious example of both). But they are obviously not the same thing either! E.g. rape is immoral, no matter how it benefits your genes.

    "It is a fallacy, for instance, to think that because our perceptions of such-and-such are good for our genes it follows that they are perceptions of goodness-for-our-genes."

    That wasn't what I was suggesting. Rather, I think that we will tend to evolve a pro-attitude towards things that were of evolutionary benefit to our ancestors. That is, if something was good-for-our-genes (e.g. ripe fruit), we will tend to see it as desirable (i.e. we will have a positive attitude towards it).

    [We don't directly see things as good-for-our-genes. What would be the point of that? Until recently we didn't even know what genes were!]

    And just to put this debate in perspective a bit: my main objection to moral 'observations' is not this evolutionary argument (though I still think it is a good one), but just that it is overly simplistic. Morality isn't something we directly perceive, it's something we (indirectly) infer. This just seems obvious to me.

    Someone might say: "you can just see that X is wrong", but I don't think that's a very accurate thing to say (unless one is being very metaphorical in their use of 'see' - to mean something like "immediately understand", for example). Instead, I'd suggest that it highlights that the particular judgement in question has become so habitual that they don't even need to engage in any conscious reasoning to reach the judgement. But that doesn't mean the wrongness is something they directly perceived by means of some strange sixth sense. 

    Posted by Richard

  9. Richard,

    I don't understand what there would be to 'meaning' in this case other than the moral property itself - yes, if we actually perceive an objective property, we have to perceive the actual property that is objective, but that's true in any case.

    I think I'm not understanding your point about 'good for our genes'; this is not the same thing as 'morally good' (in any simplistic way), but no one said they were. So I'm not seeing what problem you thought the moral intuitionist was facing here. 'The object of a pro-attitude' isn't in any simplistic way the same thing as 'morally good', either, so I'm still not sure where the criticism is.

    Moral intuitionism actually admits of considerable sophistication; not all moral properties have to be seen, many can be inferred, most moral intuitionists argue that we can progress in our knowledge of moral properties by building on our perceptions by way of inference and habituation, etc. It's also not uncommon among moral intuitionists to allow that at least most (and for some intuitionists, all) moral perception is not primitive perception, i.e., it's not like perceiving color, but more like perceiving that you are in danger, or that if you were to walk to the door you would have to walk around the chair, or that adding the two apples in your hand to the bowl would make there to be five apples in the bowl, or that the lion is about twenty feet away and ready to come at you. (The exact sort of analogy would vary depending on the version of inuitionism - there are several varities of moral intuitionism: Intellectualist, Moral Sense, Inductionist, Common Sense, Descriptionist, Emotionalist, etc.) I don't see that there's anything obviously wrong about this, nor does it appear to be a particularly metaphorical way of talking, nor does there seem to be any weird 'sixth sense' here. We do, in fact, perceive things like this; trying to deny that these things are really perceived would be excessive finickiness. And moral intuitionism simply claims that we have no reason not to extend this to moral properties as well. It can't be rejected a priori; it might end up failing to fit the facts better than a rival, but this would require a rather extensive examination of details. 

    Posted by Brandon

  10. Hmm, I was thinking that we might be able to perceive a distinction (i.e. realise that two things are different) without having full knowledge of the two things being compared. So we might realise that action X has a property A whereas action Y has a different property B, but we might not realise that A and B are moral properties, let alone tell which is 'good' and which 'bad'. In that case, there is a sense in which we perceive the moral properties A and B, whilst remaining ignorant of what I've called their "meaning".

    You're right that 'The object of a pro-attitude' isn't in any simplistic way the same thing as 'morally good'. But the way I understand moral intuitionism (I may be mistaken here), it seems that moral judgements are (or at least involve) a species of attitude. E.g. we see something that makes us feel an 'icky'/disapproving intuition so we call it wrong/immoral.

    What I see as the problem here, is that presumably these 'intuitions' or attitudes of ours evolved along with the rest of our intuitive heuristics and prejudices, for the sole 'purpose' (so to speak) of reproductive success. That raises doubts, in my mind at least, as to whether such intuitions can be trusted to accurately correspond to the moral reality. Just because it's to our evolutionary benefit to see X as wrong, does that necessarily mean that X is in fact wrong? Surely not. Is it at least a reliable indicator of wrongness? Maybe, but I suspect that reason is a more reliable guide.

    My objections might not apply to the more sophisticated moral intuitionisms you bring up. I don't really know very much about intuitionism (it probably shows!). A theory which allows for inference, etc, and holds that moral perception is not primitive perception, sounds just fine to me. But then, all along I have been arguing merely that moral properties must not be directly perceived. 

    Posted by Richard

  11. Thanks, Richard; that clarifies quite a bit. While I don't think it's quite as clear-cut as you originally made it sound, I do think that the issue, as you characterized it in your most recent comment, does in fact constitute a worry for at least some variants of moral intuitionism. I'm not sure if it's any more troublesome for them than the once-common worry about how biological altruism is compatible with natural selection; but it is, I agree, a worry.

    I think one difficulty in the discussion is what is meant by 'direct perception' - there are (as it were) degrees of directness of perception, and while in the strictest sense we don't even perceive bodies directly (just colors, sounds, etc.), in another sense of 'direct perception' the perception of bodies, or distances, or the like is a fairly obvious case of direct perception, even though it is in some sense derivative. 

    Posted by Brandon


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.)