Same and Different

June 1977, a Chicago pet shop. Irene Pepperberg, twenty-eight, has a doctorate in chemical physics from Harvard, finished the year before. She buys an African grey parrot twelve or thirteen months old. She does not choose the bird for a known talent. The name she gives the work is Avian Language Experiment. The bird is called Alex. The shop is not a laboratory. The question is whether a parrot’s speech can be more than mimicry under a protocol someone else can check.

Grey parrots were already famous for repeating household sound. The scientific default was that the sound was empty: a throat copying a throat. The laboratory animals of language in those years were mostly apes. A pet-shop parrot was not a prestige subject. She took him to Purdue, then to Northwestern, the University of Arizona, a spell at the MIT Media Lab, Brandeis, and Harvard. The rooms changed. The work did not become a tour.

The method is called model/rival. It comes, as she records, from Dietmar Todt’s 1975 work on how grey parrots learn vocal patterns socially. Two people sit with the bird. One acts as trainer. The other acts as a student, and as a rival for the trainer’s attention. The student is shown an object, asked its name or color or shape, and praised or corrected in the bird’s hearing. Then the roles reverse, so the word is not one person’s trick. The reward for saying the label is the object itself, to hold or chew, not an unrelated treat. Todt had used a version of this for phrases. Pepperberg tied the sound to a thing in the room, and to a question that could be asked of a new thing.

Same and different, in the 1987 paper, are not yes and no. Alex is asked “What’s same?” or “What’s different?” of a pair, and is to reply with the category: color, shape, or matter. The objects vary in any combination of those. On pairs of familiar objects not used in training, his accuracy sat between about seventy and seventy-seven percent. On pairs whose combinations of color, shape, and material were new to him, it sat between about eighty-two and eighty-five. Further trials were run to show he was answering the question posed, not only a cue on the objects. By the time that paper appeared, ape-language claims were under heavy fire. The parrot work continued.

By 1999 she reports that he could label on the order of fifty objects, seven colors, five shapes, and quantities up to six. He could request. He could refuse. He could, in her reports, say “none” of a missing difference. Later papers take that “none” toward a zero-like absence, which she does not call the human zero. Those are trial sheets.

Critics of this kind of work always have Clever Hans in the room: the horse that tapped answers from a trainer’s face. Pepperberg’s tests tried to separate the person who trained a label from the person who asked, and to put novel combinations where a glance would have to do more work. Whether every session was clean is a live argument in the comparative literature. The argument does not erase the trial record. It also does not close it.

On the night of 6 September 2007, as on other nights, he said the phrases used when she left the laboratory: you be good, I love you, see you tomorrow. He was found dead in the morning. He was thirty-one. Grey parrots in captivity often live into their forties or beyond. A necropsy that weekend found no apparent cause. A later pathology note from the Alex Foundation names a sudden event associated with arteriosclerosis: arrhythmia, heart attack, or stroke, not predicted by that week’s tests. The last words are a human story about a nightly routine. They are not the experiment.

After 2007 the work continued with other greys, Griffin among them. The laboratory did not become a shrine. Asked that week to summarize three decades, Pepperberg told Nature that people think more seriously about the intellectual abilities of these feathered creatures, and that she liked to think being called a birdbrain was now a compliment. That is a 2007 sentence. It is not why a chemist bought a parrot in 1977.