Turtle recall

I’m no expert in animal cognition, so please excuse what’s probably a poorly informed question, but why do so many considerations of self-awareness in animals seem to turn on the mirror test?  You knock a chimp out, paint a spot on her forehead, wake her up and put her in front of a mirror.  The line of thinking seems to be that you can learn something about her sense of self by seeing whether or not she tries to wipe the spot off.  Lots of people study animal cognition by administering this test. For the record, primates other than great apes are rubbish at it, as are pigs.  Some aquatic mammals do OK, exactly one elephant has seemed hip to the self, and a few birds have passed the test too.  At 18 months, about half of human children see the spot as a spot on them. Is it just me, or is this a really weird (and I would have thought largely unhelpful) test of self-awareness?

So what might be a better test? I thought at least a number of researchers would have gone Lockean and considered memory as a mark of self-awareness.  (I know I’m jumbling self-awareness and personal identity, but to remind you of Locke’s idea:  ‘“as far as [a] consciousness can be extended backwards to any past action or thought, so far reaches the identity of that person; it is the same self now as it was then; and it is by the same self with this present one that now reflects on it, that that action was done”.)  A brief poke around online assures me that there are lots of studies of animal memory.  Some of it’s remarkable. Creatures which hide stuff to eat later have amazing spatial memories (some birds and squirrels seem to remember thousands of locations). It turns out that slugs have a long term memory of one month. I stopped wondering about how researchers worked this out when I read the title of their paper, ‘Behavioral analysis of internal memory states using cooling-induced retrograde anmesia in Limax flavus’.  Somewhere, someone is advancing the horizon of human knowledge with a refrigerator full of confused slugs.

I think my question is, why not think that memory is a better mark of self-awareness than scrubbing a spot off in a mirror?  I know there are good objections to the idea that personal identity can be cashed out in terms of memory, but isn’t memory a sign of self-awareness?  Then again, maybe memory is necessary for a sense of self (see Clive Wearing), but just having a memory is not enough for a sense of self.  Maybe all this turns on a more general question, what’s the connection, if any, between memory and self-awareness?

Leave a comment ?


  1. The conscious knowledge of humans, as opposed to the sense knowledge of animals, can be comprehended but not defined or explicated. There is, however, a metaphysical explication of self-consciousness: Self-consciousness means turning in on yourself and catching yourself in the act of your own existence.

    The experiments done with animals and babies and the scientific definition of self-consciousness has nothing to do with metaphysics.

  2. “The experiments done with animals and babies and the scientific definition of self-consciousness has nothing to do with metaphysics.”

    This is NOT true. There is no universally accepted definition of consiousness/awareness. I think asking for better quantification of self-awareness is a legitimate
    scientific question, as well as metaphysical one. It is similar in anestesia, how can you quantify anestesic depth so the level of consiousness? There is a large literature on that. Concerning memory argument, I think “Mirror Neurons are key in quantfying self-awareness from memory perspective.

  3. The Lockean analysis is good – but only for human self awareness (i.e. awareness of the Self). The difference is between having a memory that functions to influence behaviour (as in the nut-retrieving skill of squirrels) as against being self-aware of that memory (as in much of ours).

    It’s clear that many animals can be trained to complete complex tasks that require memory, as well as their being able to execute analogous tasks ‘naturally’ (i.e. without explicit human training). But that phenomenon is a long way from allowing that any self awareness is equivalent to – or required of – that process. It can even hold in humans learning – and then conducting – the same task.

    A good example is seen in advanced motor skills such as walking (universal), typical ball-sport skills, knitting, writing or, say, touch-typing (obviously culture-dependent). These functions clearly require memory, yet they are generally conducted entirely (or at least very largely) unconsciously. Indeed, bringing such complex learned behaviours to conscious awareness and control usually disrupts their smooth operation very considerably. So we have elements of memory-driven behaviour of which we are not self-aware.

    We cannot just assume that the squirrel has the least ‘idea’ where the nuts are in any form of conscious ‘knowledge’. Rather, because of everything else we know about squirrels, it is sufficient to assume that they function in this task ‘un-self-consciously’. That we would need to ‘think about where the nuts were buried’ doesn’t mean the squirrel does.

    (I’m reminded of the exquisite routine by comedian Eddie Izzard where he is posing as a squirrel eating a nut in its characteristic way. He pauses, the mimed ‘nut’ in his ‘paws’, to ask himself ‘did I leave the gas on?’. Eddie’s joke plays exactly against the anthropomorphic tendency we can all share when observing such creatures. His punch line is the squirrel telling itself: ‘Of course not, I’m a f****ing squirrel!).

  4. Memory is a prerequisite for sense of self, but memory is useful without a sense of self.

    In order to acknowledge that you have a self, an identity, you need to be able to recollect events in time, and you also need to be able to see that these events all relate to a particular object.

    Recognising specific objects in the world as existing over time as objects that are distinct from other parts of the world is also a requirement. There are grey areas. For example, when a river meets the sea at an estuary it is difficult to say where the sea begins and the river ends. Fresh water and salt water ebb and flow with the tide. Arbitrary landmarks shift over time as the river banks and the sea coast erode and deposit in different places.

    We recognise a mountain when we see one. But is a falling rock the mountain, part of the mountain, an ex-part of the mountain, no longer part of the mountain? When does a mountain become a hill? As an aside I don’t think Sorites paradox as expressed using sand grains is a paradox, but instead simply a matter of definition.

    In animals we know there is a continuous exchange with the external world, of food, some of which becomes part of the animal, and excrement, some of which was part of the animal. Animal bodies are reformed from new elements all the time. Even the general pattern of the animal, its morphological shape, changes with time as the animal ages.

    But still, there is a sense of continuity in both inanimate objects and living objects, even if sometimes a little vague.

    So, for an entity, mostly humans but also some other animals, to identify self it needs both this means of identifying objects, even though they change, and it needs memory to recognise this continuity in time as well as space, to be able to recollect how long the entity has been an object. But finally it requires a mode of consciousness that is self-referential, being able to see that this specific object is me.

    There are many instances where there is memory and no sense of self. Any simple mechanical system that is time dependent has memory. A clock has memory. The ratchet is a second-by-second memory component. The hand positions are the longer term accumulator of each ratchet action. A typical mechanical clock has only a twelve our memory. Clocks with calendar elements have a longer term memory. Electronic clocks and other computer systems can have very complex memories. But so far we have not managed to instil a sense of conscious self into these mechanisms, even though they have memory.

    Perhaps the spot test is better at identifying fairly well when an appreciation of self is present, but not quite so reliable at identifying when it is absent. As is generally the case it is difficult to be too conclusive about the absence of evidence – i.e. it’s not necessarily evidence for absence.

  5. The feeling seems to be that memory is needed for self-awareness, but it’s not enough for a self-awareness. True, I think. I hadn’t thought of unconscious memory, or anyway what I would have called ‘muscle memory’, but that’s as clear a case as I could have wanted of memory without awareness of the memory (maybe not a case of memory without self-awareness). I think. Because now I wonder if doing habitual things like typing or driving without awareness really does involve memory at all — there’s a temptation to say it’s not memory unless it’s recalled.

    I think my confusion here is probably the result of playing around with a lot of terms much too casually. But thanks for the comments, which have helped a lot.

  6. I think you’re right the mirror test is unfair to many species. It only makes sense for species as visual as we are. It’s quite possible that a dog has self-awareness when he/she sniffs a scent and reacts to it as “mine” or “other”. Penguins could hear their offspring’s vocalizations and have a “mine” reaction. Failing the mirror self-recognition test does not mean an animal has no self-awareness. The trouble is, nobody knows how to test for other modes of self-awareness.

    As to memory as an indicator of self-awareness…I think not. Even if memory is the heart of identity over time, it doesn’t follow it’s the heart of self-awareness. An animal could be the same animal over time (numerically identical), based on memory-based continuity, yet lack the capacity for self-awareness. Or so it seems!

    Sorry–didn’t read all the comments above. It’s very possible I’m repeating what others have said.

  7. Habit is a form of memory use, whereby the brain-body system has learned what is essentially a program for performing a task that doesn’t need too much conscious oversight.

    Recall, as we normally use the term, is the bringing into consciousness of memories that we are not currently conscious of. In computer systems it is something like fetching specific data from long term memory into RAM where a program processed the recalled data. In an order entry system you can build an order using the order entry program and save the record to the database. Later, by knowing only the order number, you can get the order entry program to recall all the details of a specific order, which you might then view or change.

    The odd thing about biological brain memory recall is that it seems to incorporate some reconstruction. So when we remember an event we are also re-enforcing the memory, but also possibly altering it. On top of that the recall activity may not be the same as the memorising activity. If you recall the face of someone you know, and recall each of the features of their face, then you will recall those features individually to fill in your recalled picture. But if you repeat this recall again a minute or so later you might recall the features in a different order, or to different degrees of detail. You might, on separate occasions, recall quite different examples. I can now recall my wife’s face from what I saw her this morning looking fresh, or from last night when she looked tired, or from the weekend when she was glammed up for a night out with friends. I don’t think we have the control over recall that we think we have.

    Try something more precise. Perhaps the value of pi. How many digits are you recalling? How are the digits appearing in your mind? Do they have colour? Recall at this level is difficult to study in detail.

    Erik Kandel has been one of the leading researchers on memory. This is a good video of a lecture on memory: http://www.youtube.com/watch?v=K0cnyqzqgkQ. But try his book, In Search of Memory: The Emergence of a New Science of Mind. It’s worth it.

  8. Well, for one, can linear memory be all that indicative of anything important? Take the case(s) of people who bore you almost literally to death with long accounts of their experiences or memories that have no particular insight or interest: aren’t such people lacking something to do with self-awareness? The manner in which they affect others?
    Self-awareness must have something deeper than this kind of memory.
    Just wondering.

  9. To give a worthwhile reply here would take ages in fact I am still trying to assemble, without writing a paper on it, a reply to a similar question posed last week. You may like to consider the fact that say cats and many other animals wash themselves often very thoroughly and at length. This is some evidence that something in the nature of self awareness exists. The mirror test is often, perhaps nearly always, carried out with encultured animals i.e. laboratory animals. Such animals have had their natural behaviour adulterated by their close association with humans so we can never be sure that some of their behaviour is not merely mimicry.
    I am not sure that memory is indicative of self awareness Computers have memories but artificial self awareness seems far in the future, if at all possible. Perhaps not the best analogy but the best I can do at the moment without further thought. I cannot call to mind at the moment any Pathologies where memory functions in the absence of some vestige of consciousness.
    Humans look for goals and meaning in their lives. Regarding what happens as merely a mechanistic process, is not an attractive explanation. It seems preferable to view human behaviour as goal orientated and as the result of free will. May we not therefore view some animals similarly?
    Some things, states, or events in the world possess the property of being about or directed upon other things, states, or events. We can accordingly describe them as having aboutness. Philosophically this is termed as intentionality. So far as humans are concerned intentionality is a capacity of the mind by which it can direct itself on things. Some instances of this are found in hoping, believing, and desiring, which are always directed at something, and thus we speak of intentional states. In view of the foregoing we may describe the human as an intentional system. If some animals behave similarly and observations strongly suggest they do, then like humans it seems possible that the amimals do possess some form of self awareness.
    Our communication with say, cats and dogs, is surely deeply rooted in a shared mammalian nature. To deny mental attributes, purposeful communicatory ability, thought, and any consciousness whatsoever to animals, seems counterintuitive and not the case, although much has yet to be done towards arguing against such viewpoints.
    What, we may ask, is the point of testing analgesics on say rats, if the prior anthropomorphic assumption is not made that they are valid models for human pain? If it is maintained that they have no feelings, no thought, and their pain behaviour is no more than that of an automaton, then surely they are unsuitable models for human pain. What narcotises a rat, narcotises me, what alleviates the pain of burning in me, similarly seems to alleviate the pain in a rat.
    In opposition to Descartes who held that Animals were no more than Bête Machines, automata, with no thoughts, souls, or reason Charles Darwin stated in “The expression of the Emotions in man and Animals” “there is no fundamental difference between man and the higher animals in their mental faculties” and that “ the lower animals, like man manifestly feel pleasure and pain, happiness and misery.”
    I suggest you Google “machiavellian intelligence in animals” and access the article by R. W, Byrne and A. Whiten. It does seem difficult to envisage this Machiavellian scenario, (assuming it is a correct hypothesis, and I believe it is) as one in which animals function only as organisms, which uncomprehendingly modify their behaviour patterns under more and more sophisticated stimuli with no self awareness and no ability to capture any element of a mental state in others. Most certainly such an idea is hardly credible in the case of the human race.
    All these afore mentioned activities are of course apparent in humans; so the question remains if in addition to their knowledge of the behaviour of conspecifics do these lower primates have any knowledge whatsoever of the psychological states of conspecifics as do humans? Thus is it “Theory of Mind” which guides them rather than constant unreflective modification of behaviour patterns?
    There is a substantial body of laboratory work in existence which has purported to demonstrate TOM in Chimpanzees and some other primates.
    This is a fascinating and complex subject to which I have not really done justice here. Notwithstanding it still seems that the old problem remeins to be defeated; as we only ever contemplate our own mental states we can only by analogy, argue that others most probably have the same experiences. This of course eventually leads to the Hard problem of consciousness which may well be insoluable or perhaps as others claim, it is not a problem at all.
    My cat has just entered the room. I knew it as she said Hello to me in Catese. Our eyes met and I stroked her. She then carried out her customery twice daily inspection of the room she then vocally summoned my attention again and our eyes met again. She sat for a moment deciding whether to climb on my lap or inspect another room. She chose the latter alternative. It is accordingly difficult to claim in the light of such en experience that the animal does not have some sort of consciousness and self regard. Trying to keep this as brief as possible we have initially to ask ourselves on what grounds are we justified in thinking ourselves to be exceptions so far as the animal world is concerned. The evolutionary background is fairly clear so far as the development of animals is concerned and characteristics in common are multitudinous. There is a continuity across the species so far as physical attributes are concerned; of course a cat’s heart, is not identical to mine but common ancestorship is obvious My we not make similar claims for mentation, e.g. self consciousness of some kind, in some non-human animals?

  10. David Keith Johnson

    I request assistance as I raise a point. Was it Ruskin who described the self as an array of affinities and revulsions? Whoever it was, the point of self recognition might be related to the idea – “This is me, the person who prefers X and does not prefer Y, etc.” Conscious, non-automatic memory would at least play a part in this, I would think. While we are not always conscious of our reasons for attraction/revulsion, is it not often memory based? (I don’t eat peanuts because I have tasted them, and hate the flavor.) Finally, the mirror test for animals would depend on both the mechanical optical capabilities of the animal – can they even see the spot? – and an assumed mental state that even cares about spots on faces. Difficult not to see it as rank anthropomorphism, and therefore not at all valid.

  11. Many good summarising remarks here (e.g. James Garvey “The feeling seems to be that memory is needed for self-awareness, but it’s not enough for a self-awareness.” and Ron Murphy “Memory is a prerequisite for sense of self, but memory is useful without a sense of self.”).

    Here are two features relating to memory and neurological states that, I think, give pause for philosophical thought about the nature of self, memory and consciousness (never mind more simple human, emotional responses).

    First; we are aware of increasing numbers of unfortunate folk suffering short-term memory loss dementias. This is but one form of neurological deficit that seems to leave self-awareness fully functioning but, as it were, compromised by the memory deficits. It does lend itself to the interacting multi-unit way of interpreting brain (and thus ‘mind’) function. When neurological deficits like these reach extreme levels, we are left wondering what is left of a sense of self for the sufferer. However, it equally challenges the very notion of what conception of a ‘normal’ sense of self can really require of brain function.

    Second; pain is peculiarly fascinating in that it cannot be remembered. We might well remember that ‘it hurt a lot’, and so on, and make great efforts to avoid the experience. But we cannot re-experience pain, despite recalling – often in exquisite detail – the where and when of painful events. Other sensory processes can be more fully ‘conjured up’ and re-run by mental effort (or in response to a suitably evocative stimulus). We can recall or bring to mind smells, sights, sounds, even touch – but not pain.

    This makes the action of the so-called dissociative drugs fascinating. They are not analgesics (ie they don’t deaden the immediate sense of pain) but they block the memory of the painful event. A video of (say) your dislocated shoulder being painfully ‘put back in’ whilst you are under such drugs will be quite disconcerting to behold. It’s you, evidently in great pain, but you don’t remember anything about it. But you still feel like you remain yourself, despite the memory ‘gap’. It’s also intriguing to ponder what it really means to say you felt the pain, or not, if you can’t remember the event.

    So the interplay between memory, sensation and what it is we need in order to construct and sustain our sense of self-awareness is truly complex.

  12. The idea that we can identify a system that acts like a monkey or bird that can be ‘self aware’ in the simple sense of being able to detect changes to the surface of itself doesn’t necessarily imply a consciousness behind it. A consciousness is assumed. To my knowledge, a model of consciousness has not yet been proven so we remain in the dark about at least some of its content and structure.

    Further, in the mirror test, we are asking if the subject critter has the ability to remember its own image before the spot was there; is it aware of its appearance. This requires that it have a memory of its appearance. The mirror test seems to move past the memory test alone.

    Lastly, part of what self awareness struggles with is John Searle’s Chinese Room problem. It suggests that we may not be able to use outside measures to determine self-awareness at all. His example supposes that there is a program that gives a computer the ability to carry on an intelligent conversation in written Chinese. If the program is given to someone who speaks only English to execute the instructions of the program by hand, then in theory, the English speaker would also be able to carry on a conversation in written Chinese. However, the English speaker would not be able to understand the conversation. Similarly, Searle concludes, a computer executing the program would not understand the conversation either.

  13. Computers have memory and do not seem to be self-aware. I think the spot test shouldn’t be viewed as a necessary and sufficient condition, but as a sufficient one. That is:
    Recognizes itself in the mirror -> It is self-aware
    But not
    It is self-aware -> Recognizes itself in the mirror
    Finding a sufficient and necessary condition would of course be ideal, but it doesn’t seem easily feasible since we don’t really know what we’re actually looking for…

  14. Jean Kazez notes above that the mirror test can be unfair to some species, true unless the test is modified. There are examples of experiments that have overcome that issue by making species-friendly substitutions that get at the same problem. See The Dolphin in the Mirror by Diana Reiss, whose results are convincing to me, and my turtle cognition lab’s work with terrestrial turtles. Turtles can’t reach the top of their heads to remove a mark, as in the primate version of the mirror test, which would be a problem for the test in its literal form. Our version then consisted of whether the turtle tried to remove a previously unnoticed piece of food from the top of the head after being shown the mirror. They did, with different solutions. Most interesting was that of the pancake tortoise who looked repeatedly at the mirror between attempts to scrape it off with the narrow bottom surface of the hung mirror. (She did what other species couldn’t be expected to do: pulled her head in to scrape it off.) All animals tested in the various studies, including mine, were first given experience with mirrors. Those who then passed Gallup’s mark test or a species-modified variant had first explored their bodies’ movements and/or parts otherwise unseen. The turtles’ initial mirror reactions depended on their degree of agonism and were the ones whose change in perspective was obvious; box turtle males were either submissive to the mirror or attacked it–clearly a response to the “other” male turtle. When these turtles were lifted to look into the mirror beside my familiar face, submissive and agonistic behavior ceased. Instead, they reacted like the non-agonistic turtles: They looked at me, looked at the image, opened their mouths, nodded, moved their limbs in odd postures, turned over or spun around in my hands, and so on, their own versions of testing. Me or other? Their conclusions seemed clearly to be the former.

Leave a Comment

NOTE - You can use these HTML tags and attributes:
<a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>