NS Internet column

Written 01 November 1999 for the New Statesman

Enough. Chicken sexing already. The whole point of chicken sexing is that no one knows how to do it: even the trained chicken sexers, though they often get it right, can't explain how they manage it. So their performance is extremely interesting to anyone who cares about how humans actually think, which includes all serious artificial intelligence researchers; and many of them hang out on a wonderful set of mailing lists called "psyche-L, "psyche-b" and similar names. These are almost the last places I know of on the Internet that still feel as the whole thing did ten years ago, as if it were the largest University bar in the world: the bar of educated world opinion. Students and grand professors mingle wildly.

A couple of months ago there was a flare-up of interest in chicken sexing there which showed these lists at their best. A researcher in Los Angeles started it by asking whether it was true that chicken sexers did it without consciousness. Someone in Sheffield replied with a reference to a paper in the Journal of Experimental Psychology called "sexing day-old chicks - a case-study and expert systems-analysis of a difficult perceptual-learning task." This is apparently related to other tasks which we can learn to do without ever explaining how, such as discriminating between different vintages of wine; sadly no one provided references to papers on the perceptual mechanisms involved in wine tasting.

At this point., Joe Jeffrey, a professor of computer science, objected that skills of this sort are not at all uncommon. In fact "this  is simply another example of a very common, everyday, fact about persons:  A person's ability to do things _far_ exceeds his/her ability to describe what he/she does.  I don't know why one would invent a term such as "implicit cognition" for the phenomenon, implying by contrast that more "ordinary" or "non-implicit" cognition involves the ability to describe what one does or sees.

So far this might be any discussion of perception and artificial intelligence; what I loved was what happened next: Stan Klein, a Berkeley psychologist announced that he had grown up on a chicken farm, and thus knew more than anyone else about the subject on hand. The point, he said, was that chicken sexers clearly had specialised knowledge inaccessible to other people: when his parents bought a batch of 500 chicks as egg layers, only three or four would turn out to be cockerels: "Now for the other part of my expertise, my psychophysics hat. In all my years of playing with lots of stimuli I have never seen such a high discrimination ability (99%) without the subject's being conscious of what the decision is based on … In summary, based on my experience as a chicken farmer and a psychophysicist, I believe the chicken sexers had conscious awareness of the cues that they used."

As any academic knows, it is good to have first-hand knowledge of the subject at hand, but it is even better to have read the literature, and this point was driven home finally by Bruce Mangan, a Berkeley philosopher, who had actually read the chicken-sexing paper form the Journal of Experimental Psychology. This turned out to contain enormous amounts of information about the history of chicken sexing in the USA: the modern method had been introduced from Japan in 1933:

"By showing their audience numerous physical examples of male and female genitalia augmented by verbal descriptions of the many different kinds of sub-type features (passed laboriously through an interpreter) the Japanese were finally able to transfer their knowledge. Schools of chicken sexing sprang up immediately, and the Japanese technique soon became the standard method for chicken sexing, at least on this continent."

This is the sort of knowledge that changes lives. I don't think I can ever consider a  lecture is boring again, knowing that some people have acquired their professional qualifications by studying chicken genitalia with a commentary in simultaneously translated Japanese.

Bruce had even remembered a newspaper story about one of their original pupils who had gone on from these lectures to sex 55m chickens over his lifetime and was prepared to talk at length about the discoveries he had made. There is a further twist. The original researchers had decided there must be something simpler; and examining the blown-up photographs of chicken genitalia which are chicken sexers stock in trade they'd noticed that there is in fact a visible difference between boy chickens and girl chickens, even if the professionals could not articulate it.

They taught complete amateurs to look for this difference, and in not time they were doing nearly as well as the professionals. Fortunately, no one took any notice of their results except academic psychologists. I can't think of any discussion on the Internet which has had quite such a satisfying conclusion.