Wednesday, August 17, 2005

Intelligence as Flexibility

What are minds? Or, as I recently asked, what's the difference between a mouse and a mousetrap, such that 'cognition' only occurs in the former? A fairly natural answer, I think, is to understand cognition as involving a 'hidden layer' of information processing that mediates between (sensory) input and (behavioural) output. Building on this idea, we can say that flexible behaviour is evidence of intelligence.

We should distinguish two types of plasticity here: flexibility in forming representations, and flexibility in acting on them. The former involves using multiple cues to identify properties in the world. We can identify an object not only by how it looks, but also how it smells, sounds, and so forth. There can even be multiple cues within a single sensory modality: if I see a realistic-looking plastic fish, but also see that my cat has no interest in it, then this latter (still visual!) cue might lead me to doubt whether it really was a fish. To adopt Sterelny's terminology, "robust tracking" occurs when internal representations are sensitive to a wide range of diverse stimuli. The benefits of robust tracking are clear: it helps us to identify properties more reliably and thus form more accurate representations (true beliefs).

In addition to robust tracking, another form of plasticity is found in the breadth of responses open to an organism once an identification has occurred. I might have a million different ways of spotting danger, but one would question my cognitive capabilities if all I ever did in response was to run away as fast as I could. More flexible (=intelligent) organisms can tailor their responses to fit the specifics of the situation. They might flee, hide, prepare to fight, or call for help, etc. They are not bound by the "tyranny of the stimulus", unlike the infamous Sphex wasp:
When the time comes for egg laying, the wasp Sphex builds a burrow for the purpose and seeks out a cricket which she stings in such a way as to paralyze but not kill it. She drags the cricket into the burrow, lays her eggs alongside, closes the burrow, then flies away, never to return. In due course, the eggs hatch and the wasp grubs feed off the paralyzed cricket, which has not decayed, having been kept in the wasp equivalent of deep freeze. To the human mind, such an elaborately organized and seemingly purposeful routine conveys a convincing flavor of logic and thoughtfulness -- until more details are examined. For example, the Wasp's routine is to bring the paralyzed cricket to the burrow, leave it on the threshold, go inside to see that all is well, emerge, and then drag the cricket in. If the cricket is moved a few inches away while the wasp is inside making her preliminary inspection, the wasp, on emerging from the burrow, will bring the cricket back to the threshold, but not inside, and will then repeat the preparatory procedure of entering the burrow to see that everything is all right. If again the cricket is removed a few inchies while the wasp is inside, once again she will move the cricket up to the threshold and re-enter the burrow for a final check. The wasp never thinks of pulling the cricket straight in. On one occasion this procedure was repeated forty times, always with the same result. (Woodridge, 1963, p. 82)

Response breadth thus reflects versatility in action, whereas robust tracking reflects versatility in perception (broadly understood). So far as I can tell, these two types of 'intelligence' are completely independent of each other (at least conceptually; there could be some overlap in their implementation - best ask a psychologist). It reminds me of the distinction between theoretical and practical rationality (i.e. beliefs vs. actions), though I'm not sure exactly how strong the analogy is.

Next up: what, if anything, the two types of plasticity can tell us about the intentional content of a representation...


  1. The problem may be that the Wasp's "objective" is not actually survival - it is "to have a cricket at the enterance". and bringing a cricket there may represent quite considerable waspy joy.
    Still I think I might be streching it a bit to suggest this is generally true at the wasp level.

    Having said that - I expect you could catch a human with the same trick.

  2. Any half-intelligent human would avoid getting stuck in such a sphexish rut. That's the whole point of intelligence/plasticity: our behaviour isn't bound by just one simple stimulus-response mechanism. Instead, we can take in lots of various stimuli and reflect on them before deciding how to act. We can, in a word, think. Any animal with the capacity to think or "learn" (even in the rudimentary behaviourist sense of 'conditioning') will avoid this sort of 'trick'. So you'd have to pick some area in which humans do not think, i.e. where we have an automated reflex-like involuntary response to a simulus.

  3. Does this mean that an OCD afflicted individual would be half intelligent? As long as you are willing to alter the results of a tic most individuals with OCD would continue to respond to the stimulus in the same way.

  4. Compulsive behaviour is not intelligent. But agents can be intelligent in some respects but not others. My blinking reflex is not intelligent, but hopefully some of my other behaviour is.

  5. On second thoughts, I suppose my claim that "compulsive behaviour is not intelligent" is somewhat controversial. Perhaps one could argue that the compulsive person just has a really strong desire that the world be a certain way (e.g. tidy!), and they will act in appropriate ways to achieve this goal?

  6. Yes that is what I was thinking - in addition to the other point "where we have an automated reflex-like involuntary response to a simulus."
    However I note most people might consider such a person to be an idiot.

  7. Hmm, I think there are some problems with this argument. For example, consider a person who has been paralyzed in an accident and can only communicate by blinking his eyes in response to questions. Surely you wouldn't claim that such a person was not intelligent, even though they can only respond to stimuli in one very specific way. You might argue, since they can change their eyeblinking behavior based on external stimuli, that this counts as flexibility of response. But then how is that different from a mousetrap, whose response varies depending on whether or not it "perceives" a mouse in the right location? If you argue that the difference is robustness of perception, that's fine, except that you've reduced flexibility of response to a special case of flexibility of perception. I could then argue further about a better mousetrap, which in addition to pressure sensors had computerized cameras and infrared lasers and motion detectors and complex heuristics to determine whether it "saw" a mouse... Or, one could consider a person who is deaf and blind. Reading Helen Keller's memoirs, it seems fairly clear that she developed an internal representation of water based solely on touch.

    It seems like robustness of perception and flexibility of response are desirable characteristics, but neither necessary nor sufficient for intelligence.

  8. You might find this article about measuring the intelligence of AI rather interesting.

  9. The New Scientist link isn't working for me (I'll try again later, in case it's just my computer playing up).

    Covaithe - Perhaps I should have said that plasticity indicates, rather than constitutes, intelligence. It does seem possible to have an intelligent mind trapped in an entirely paralyzed and unresponsive body. Though I find your particular examples unconvincing -- if the person can vary their eyeblinks sufficiently to communicate, then that is very flexible behaviour indeed. I can't imagine a souped up mousetrap doing anything comparable.

  10. Oh yup, the link works now -- interesting stuff.


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.)