Monday, September 19, 2005

Self-Recognition and Awareness

According to Gallup's famous "mirror test", you can test whether an animal is self-aware by whether it can learn to recognize itself in a mirror. In particular, if you secretly mark the animal's forehead, and it sees this mark in the mirror and responds by touching its own forehead, then this shows that the animal recognizes itself in the mirror. So, the argument goes, the animal must have a concept of self, and be self-aware. But despite the initial plausibility, this conclusion doesn't follow at all.

All living organisms draw some sort of distinction between self and other. For example, the job of our immune system is to identify and destroy alien cells, whilst leaving our own ones untouched. But no-one would thereupon conclude that our immune system is self-aware. Building on this sort of idea, our lecturer pointed out what I think is a devastating criticism of Gallup's test.

Suppose experimenters marked the animal's arm instead, in a clearly visible spot. Now, we surely wouldn't be surprised if the animal then touched the mark on its arm. It can identify its own arm, but that doesn't entail full-blown self-awareness. But why does bringing a mirror into the picture make any difference? A mirror is merely a perceptual tool - it allows you to see things that you might not otherwise see. In particular, it allows an animal to see its own forehead, in addition to its arms and such. Most animals can't make use of this -- they can't work out that the mirror gives them a view of themselves (they either think the image is another animal, or else ignore it entirely). Some apes are more intelligent in this sense -- they can make use of the mirror as a perceptual tool. But it still doesn't imply anything new about psychological self-awareness.

What we really want to know is whether animals know that they have minds. Do they have a concept of self (as opposed to simply being able to recognize themselves)? Can they think about their own beliefs and desires, recognizing that they have thoughts, and that they have a life, and are the same creature today that they were yesterday? These are the important questions, but they're questions that the mirror test is silent on. Despite all the hype, it doesn't really say anything interesting about self-awareness at all.


  1. I wonder how the ape would respond to a dot that is placed on the mirror itself. (Are animals generally interested in dots?) If the ape wasn't interested, it wouldn't prove much, I suppose, except that dots close by (on the arm, for example) are more interesting than dots far away.

    You could perform a slightly different experiment to reach the same conclusion as your lecturer. Set the ape in front of the mirror. Dangle a banana above its head. Fix the ape so that it cannot see or otherwise sense the banana without deliberate movement. If the ape reaches towards the banana, we would not say that it has self-awareness: we would say, firstly, that it can recognise correspondences between things in the mirror and things outside it; and, secondly, that it is interested in bananas. Likewise for the dots, I suppose.

    Why is it important that we know whether or not animals can recognize thoughts etc.? Surely such things are matters of degree, anyway, and if we did discover that animals were vaguely aware of some of their most basic thoughts, it would not transform our attitude towards the cuddly wee things, would it?
    Who was the lecturer?

  2. insightful post...

    the test certainly not clear what it really measures... it's so vague and definitely brought out alot of controversies and further studies are needed to know for sure...

    as humans beings, we all have different level of self-awareness...

    here's an article of National Geographics regarding this test.

  3. One criticism made of mirror tests is that they are biased in favour of a sense (ie, sight) which is strong in man, but may not be as strong in other animals. Thus, dogs, for example, typically fail the mirror test, but can pass a similar test based on smell.

    Of coure, any dog-owner could have told the scientists that dogs have self-awareness. Dogs playing jokes on their owners would seem to indicate a pretty advanced sense of self and of other.

  4. I don't know that pet owners are the best judges of their pets' cognitive abilities. ("Common sense" tells us little enough about how humans work, let alone other species -- just ask a cognitive psychologist.) There are almost always lower-order interpretations of complex behaviour that doesn't appeal to meta-cognition (thoughts about thoughts) at all. It's a very contentious issue whether we're justified in ascribing higher-order thoughts to animals in light of these alternative explanations.

    Mike, the lecturer is Derek Browne, from the Phil 238 Cognitive Science paper. I think metacognition is of intrinsic interest, but then I tend to be curious about such things. For those of a more pragmatic bent, I suppose it might have ethical implications (since self-aware creatures are presumably more morally significant). But the main thing is just that it would improve our knowledge of animal minds, and like all scientific/theoretical advances, this is surely worth knowing even if it doesn't have immediate pragmatic payoffs. I agree that it's a matter of degree -- other animals obviously don't have anything near the understanding of minds that we humans have, but it's still a contentious question whether they have any such understanding at all.

  5. I don't know, Richard; your response sounds a little too Cartesian to me. If the monkeys are identifying their arm as their arm, or their forehead as their forehead, that sounds rather significantly self-aware to me. I take it that precisely the point of the mirror is that it is only 'merely' a perceptual tool for those who already are able to use it as such; and to be able to recognize facts about oneself not merely from glancing down at oneself but also from looking at a reflection show that you have something that has been abstracted and applied beyond mere body-awareness. Now, I think there's lots and lots of room to argue about what exactly this something is; but certainly there's something going on beyond mere perception, and that something is very probably a further level or kind of self-awareness. I think, for instance, that the tendency to label this as a 'theory of mind' is unfortunate (if the phrase is taken literally), since the term suggests rather more sophisticated linguistic associations than are usually required even in ordinary human self-awareness. But an ability to identify oneself out of all the other things in the world is surely significant, since it shows a real ability to make sign-based inferences about oneself.

    On the issue of whether there's merely a recognition of correspondences going on, my understanding is that there's a rather substantial body of work investigating the issue; and that they've shown that being able to use the mirror as a tool in this way doesn't suffice for being able to pass the mirror test -- many animals that can use the mirror in ways that show that they can make such correspondences fail the test. And this is surely plausible; it is one thing to recognize that a mirror mirrors non-mirror events, but it is another thing to recognize that the image in the mirror is just your own face.

  6. I note that animals like cats seems to quite often treat mirrors as "pieces of wood with glass on them" as opposed to "another cat".

  7. One reason that a mirror could be important, is that it makes a representation of an individual that is completely disconnected with the viewer. Thus knowing that you are the figure on the other side of the mirror is sort of like knowing a contradiction.

    The reason we know that pointing at a dot on an arm is not that insightful is because there is a spacial connection with the pointer and the arm. That doesn't exist when looking at a mirror.

    Still, it is true that self-recognition does not entail self-awareness.

  8. I suggest almost all animals higher than a plant can differentiate between self and non-self (although maybe some can't for some vague evolutionary reason).
    Also I think most social animals have some sense as to others being similar to them and having felings like them and most animals that can plan and remember decent periods of time can imagine "if this then that".
    As a result even a merecat wil from time to time be thinking if I die then... and all those other thoughts we might like to reserve for humans.

    But just like a human out in the wild it probably wouldn't go much further than that (if you have no historic or religious data you will tend to have no sophisticated concept).

    I think there is a tendancy to want to over complicate these things because many peopel "want" self awareness to be the focus of evolution, when instead they emerge fairly naturally out of very basic concepts.

    So I can imagine potentially that a lizard might have evolved self awarenes superior to us (for some hard to fathom evolutionary reason) and then lost it again when not required.
    (note more self aware not more intelligent - they wouldn't have had the brain cels to do the latter).

  9. Genius --

    Even plants are able to distinguish self from non-self, according to this recent research report:


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.)

Note: only a member of this blog may post a comment.