Monday, November 29, 2010

Analogy as the Core of Cognition

Douglas Hofstadter has an interesting discussion in a Presidential lecture at Stanford. Skip the introductions and start at 13 minutes 30 seconds to listen to Hofstadter.

Hofstadter believes that analogy making is at the core of all cognition, and what is especially interesting is how frequently analogies seem to occur in everyday experiences and how complex the parallels can be when suddenly we have a flash of insight, "That's just like...(something else)".

It's probably true that we don't think about using analogies as much as we should in education. Analogy-making abilities exist in preschoolers and studies even suggest that the use of analogies in present new concepts to young elementary school children improves learning and retrieval.

Another interesting read is Gentner's Analogy. In it, Dr. Gentner recalls an interesting study by Gick and Holyoak that found 30% of people primed with an analogous story thought to use it in solving a unique problem (10% solved it without the priming analogy); however, test subjects were simply given a hint to think about stories they had heard, 80% were able to solve the new problem (the problem was how to kill a tumor without too much surrounding irradiation; the analogy involved soldiers converging on a fort). So part of the dilemma for getting students (and adults for that matter) to solve problems - may be that they aren't accessing analogous situations from their fund of knowledge or memory. Perhaps some of the inventive souls and creative thinkers (like Hofstadter) are more attuned to analogies in their thinking and this gives them their creative edge.

Analogy: The spark for invention
Hofstadter's Review of Mental Leaps


  1. That really reminds me of something I'm pretty sure you said somewhere in the Mislabeled Child about how our thinking is really just all about our memories -- how we store them, how we use them ... at least, that's the way I remember it, that's probably not exactly what you said.:) But it seems like when we draw on an analogy, maybe what we are doing is somehow accessing a memory that does not seem relevant on a superficial level ... how was that original memory stored that caused it to be triggered by this other situation that does not seem entirely relevant? What makes some people instinctively recall the superficially irrelevant memory (and then note the similarities), while others fail to notice? Is it less a storage thing and more a question of how hard you try to "scan" your memories when you look for a solution? I'm sure it's both. It's really fascinating, and maybe I'm not thinking about this the right way, but I think maybe I can see how this is very close to the core of who we are and how we operate.

  2. I just loved this conference ! I am a french native, so during listening I couldn't help "self-monitoring" myself about all the analogies I had to do trying to "translate" Dr Hofstadter explanations. I am pretty sure the Einstein photon analogy explanation is also an analogy with this great "Analogy as the core of cognition" hypothesis ;) ... And now to develop an analogy based AI. Human cognition might not be so mysterious after all, but might an AI have noticed THIS analogy ?

  3. As a physics teacher, we use analogies constantly to get into new material and problems, as well as more complex problems. Electric current in a circuit is like water in your home plumbing system...or electrical resistance occurs because the electrons are like a ball in a pinball machine...or energy and matter are two forms of the same thing, like steam and ice. And many others. Anecdotally, I agree with the statement that analogies are essential for learning and problem solving, and allow learners to make connections that would be more difficult without something else to compare to.

    Thank you for bringing up this topic.

  4. I always use analogies when I talk about our program to new parents. My wife, Catherine, apparently doesn't feel they are as useful to understanding as you do. She just tells me I tell too many stories . . .