When is “different” different from “not the same”?

27 11 2015

I went to the GTA Animal Cognition Group on Monday. Objectively it was a crazy use of my time, since it involved one hour reading and two hours discussing a paper on what humans are doing when they learn a “match to sample” as opposed to a “non-match to sample” rule, plus four hours on the road there and back. Yet, as ever, it just made me insanely happy. To be really intellectually stretched by what is very new material for me, as well as by the insight and logic of people like Jake Beck and Sara Shettleworth (not the full list!) was deeply enjoyable.

It was interesting too: we discussed unpublished work from the Carey Lab, inspired by older work showing that non-match-to-sample tasks are easier to learn than are match-to-sample tasks (some kind of innate ‘win-shift’ strategy, I wondered?). Over three experiments, clever eye-tracking data showed that when we learn to non-match-to-sample tasks, although we could follow a “seek out stimuli that are different from the sample” rule, we actually follow an “avoid stimuli that are the same as the sample” rule.

This in turn – rather more speculatively – suggests that the concept “same” is easier or more basic for us than is “different”, even though of course the two are reciprocals. This odd idea is made more plausible by similar asymmetries detectable in the development of concepts by infants: children apparently understand “more” before they understand “less”, “over” before “under”, and “front” before “back”, even though to us adults these pairs of concepts seem like mirror images.