17 June 2009

Birdbrain Statistics

I'm reading Alex & Me by Irene Pepperberg, which I originally blogged about here, chronicling her research with Alex, the African Grey Parrot who won the hearts of many bird- and research-lovers worldwide over the years. I love reading about her research methods and Alex's antics. He was smart but a difficult student at times.

One recent passage amused and interested me:
I found myself having to prove over and over that Alex had more going on in his bird brain than some mechanical trickery or other. One such challenge was, "Oh, he can produce labels all right, and he sounds convincing, but does he really understand what he's saying? Does he comprehend the noises coming from his beak?"

It seemed quite clear to me from my hundreds and hundreds of hours watching and listening to Alex that he did indeed know what he was saying. A simple example: if Alex said "Want grape" and you gave him a banana, he'd spit it right back at you and repeat insistently, "Want grape." He wouldn't stop until you gave him a grape. If you were dealing with a child, you would accept without question that he or she really wanted a grape, and that banana simply wouldn't do. But that's not science. Science needs numbers. Science needs tests to be done over and over again--actually, sometimes sixty times or more--before the answer has statistical legitimacy, and before scientists will take you seriously. Poor Alex.


On a sidenote, earlier in the book, she talked about how she was trying to teach Alex to say "apple", but after two weeks of training and struggling to make the "puh" sound in "apple", he had other plans in mind. She describes it:
...he looked at the apple quite intently, looked at me, and said, "Banerry...I want banerry." He snatched a bite of the apple and ate it happily. He looked as if he had suddenly achieved something he had been searching for.

I had no idea what he was talking about. So I said, "No, Alex, apple."

"Banerry," Alex replied, quickly but quite patiently.

"Apple," I said again."

"Banerry," Alex said again.

OK, buddy, I thought. I'll make it a bit easier for you. "Ap-ple," I said, emphasizing the second syllable.

Alex paused a second or two, looked at me more intently, and said, "Ban-erry," exactly mimicking my cadence.

We went through this double act several times: "Ap-ple." "Ban-erry." "Ap-ple." "Ban-erry." I was a little ticked off. I thought Alex was being deliberately obtuse. In retrospect, it was quite hysterical. When I told one of my students, Jennifer Newton, about it later, she literally fell off her chair laughing. But Alex hadn't quite finished with me just yet. At the end of that session he said, very slowly and deliberately, "Ban-err-eeee," just as I might do with him when I was trying to teach him a new label. Maybe he was thinking, Listen carefully, lady. I'm trying to make this easy for you.

...No matter how hard we worked to get him to say "apple," he stuck with his label. As far as Alex was concerned, "banerry" it was and "banerry" it was going to stay.

A few days later I was talking to a linguist friend about all this. He said, "It sounds like lexical elision." It's a fancy term for putting parts of two different words together to form a new word. Alex might have thought the apple tasted a bit like a banana. Certainly it looked like a very large cherry (it was a red apple). "Banana" + "cherry" = "banerry."

That passage had me laughing so much my eyes were watering. I've experienced the personality of greys, so I could easily picture the interaction, and it brings me joy.

This relates to the first section I quoted because it illustrates Alex's stubbornness and their general relationship. When it came to repeating the tests enough to gather the statistics necessary for scientific reporting, there were other challenges:
The tests involved putting various of his "toys" on a tray and asking questions such as "What object is green?" "What matter is blue and three-corner?" "What shape is purple?" "How many four-corner wood?" At first, Alex answered correctly most of the time: "key" or "wood" or "wool" or "three," et cetera. But before too long, he started to act up. He would say "green" and then pull at the green felt lining of the tray, hard enough that all the objects would fall off. Or he would say "tray" and bite the tray. Sometimes he'd say nothing and suddenly start preening. Or he'd turn around and lift his butt in my direction, a gesture too obvious to need translation. Once he grabbed the tray out of my hand and flund it on the floor, saying, "Wanna go back," which meant, I'm done with this. Take me back to my cage.

Who can blame him? None of the objects were new to him. He'd answered these kinds of questions dozens of times, and yet we still kept asking them, because we needed our statistical sample. You could imagine him thinking, I've already told you that, stupid, or simply, This is getting very boring. He was like the bright little kid at school who finds none of the work challenging and so passes the time by making trouble.

Sometimes, however, Alex chose to show his opinion of the boring task at hand by playing with our heads. For instance, we would ask him, "What color key?" and he would give every color in his repertoire, skipping only the correct color. Eventually, he became quite ingenious with this game, having more fun getting us agitated rather than giving us the answers we wanted and he surely knew. We were pretty certain he wasn't making mistakes, because it was statistically near to impossible that he could list all but the correct answer. These observations are not science, but they tell you a lot about how sophisticated his cognitive processes really were. Whether you would describe what he did as something to amuse himself or as making a joke at our expense, I cannot say. But he was definitely doing something other than routinely answering questions.

This, of course, led me to wonder how you report, statistically, the responses of an animal which seems to be deliberately refusing to give you the responses you need to support your research. Obstinate little cuss. I love it. I'm enjoying the book muchly.





Alex showing Alan Alda what he can do:


ABC News report when Alex passed away:


A longer video on Dr. Pepperberg's research with Alex and other parrots:

2 comments:

blj1224 said...

I love the Alex and Apache video. When Alex passed away, I cried. He was an exceptional animal, and watching him interact with his trainer was heart warming.

JJ said...

Oh, I couldn't figure out what you were talking about with the "Alex and Apache" thing. I suppose one of the videos links to that one, but that's actually a different parrot. There are a few videos on YouTube of African greys named Alex who aren't the one Dr. Pepperberg researched with. But African greys are entertaining nonetheless. :-)