This is the Champs-Elysée of Budapest: a grand tree lined avenue, framed by glamorous buildings, and with some famous coffee houses. We are sitting outside the classic Café Müvesz, with a splendid view of the Opera: Ildiko Kiraly, Kata Egyed, Chris and me after a nice open air lunch at Liszt Ferenc Tér, not more than two hundred yards away.
-You used your famous “Head-touch” experiment with autistic children. You found in a new experiment that young autistic children, unlike carefully matched children with Down Syndrome, disregarded the experimenter’s intention as indicated by ostensive gestures. This is a strong indication that we cannot rely on what Gergely and Csibra called natural pedagogy when teaching autistic children.
The waitress brought cups of coffee and glasses of water. I took up a previous threat of our conversation:
-Your colleagues made the stunning discovery that the A not B error in young infants was much reduced if the experimenter did not use ostensive gestures. Let me see if I got this right: It is precisely the communicative setting that makes infants perseverate in their error. It is as if they assume the experimenter has taught them to go to a particular hiding place – and this is what they reach for. If the experimenter does not use ostensive gestures, they don’t learn this and consequently don’t make the error. Instead they go for the new hiding place where the object really is.
-Yes, that’s correct, said Ildiko, -and this was the same in dogs, but not in wolves, as shown in the paper with Topal. So natural pedagogy is something that works for dogs too, probably because they have been social companions of humans for millenia.
Chris and I had been extremely interested in studies testing the theory of ‘Natural pedagogy’, introduced by Gergely and Csibra (aka the rockstar Hungarian developmental psychologists). It is a fascinating theory that suggests that humans have a means to acquire culturally relevant information from each other, that makes learning incredibly fast and powerful. The trick that evolution has provided is ‘ostensive communication’.
Chris: – There are two ways of learning from others. They can address you ostensively, perhaps call you by your name, look at you directly, flash their eye brows at you. But you can also learn by just observing them.
When our coffee was almost finished, the conversation strayed towards a Bayesian theme. Chris asked: -Why do we pay so much attention to the information that comes from our own senses? When does this start in development?
Ildiko and Katalin both considered this question. -It seems very possible that at first young infants do not pay so much attention to their own sensory information. Take the A not B error. They follow the object with their eyes, and they therefore know where it has been put. But this information counts for less than the information conveyed by the adult’s communication. So it is another person’s perception that seems to win over their own.
-Interesting! So it is not so obvious that we first and foremost regard our own sensory perceptions when making any inferences about the world. Perhaps we are taught that the evidence of our own eyes is the best.
Chris added: – I have just seen a paper by Jaswal who studied children’s trust in information provided by adults. Toddlers believe what an adult says even though they have just seen something different.
My thoughts strayed to “nullius in verba” -take nobody’s word for it- the motto of the Royal Society.
– So is it only since the enlightenment that we feel we must see for ourselves to believe? It was clearly a huge cultural change that brought about the attitude that we should not put our trust in the evidence transmitted to us by others.
-But then aren’t there lots of pitfalls when we put all our trust into our own senses? And by implication, our own experiments?
– Hmmm, we have a conundrum and this relates to our earlier and rather controversial post “Not to be found in any methods section.”