TERRY GROSS, host:
How many times have you wished you could talk with your pet and find out what they were thinking? My guest Irene Pepperberg wanted to conduct research into animal thinking, so she bought a talking bird from a pet store, a gray parrot she named Alex. Her idea was to replicate the linguistic and cognitive breakthroughs demonstrated in research with chimps, using an animal that could talk. Alex became her good friend as well as her longtime research subject. As a result of their work together, he probably became the most famous parrot in the world. When he died two years ago at the age of 31, he got an obit in the New York Times headlined, "Brainy Parrot Dies, Emotive to the End." Pepperberg has written a memoir called "Alex & Me." She's now an associate research professor at Brandeis University.
Irene Pepperberg, welcome to FRESH AIR. How many words was Alex capable of saying?
Dr. IRENE MAXINE PEPPERBERG (Associate Research Professor, Brandeis University): It's hard to say because there were some that were just contextually applicable, something like I'm sorry. There was no contrition. It was just something to say when he had done something wrong, and everybody made sure he knew that it was something wrong.
But there was good data on about 50 different object labels, seven colors, five shapes, quantities up to eight just before he died. And then he would combine these to identify, request, refuse, categorize, quantify more than 100 different things in the laboratory. So once he knew block, then he knew green and yellow and orange, and so he could identify the green block, the yellow block, the orange block, things like that.
GROSS: How much ability did he have as a parrot to pronounce the words that you wanted him to say?
Dr. PEPPERBERG: He got most of them right. There are some of them that are really tough, like imagine saying paper without lips, and he actually did that. But final S's always seemed to get him, so he would say things like Alec instead of Alex. If the S was in the middle of a sentence, like what's that, not a problem, or what's saying, not a problem. He would say six by saying sic, and we had to really push him to say sic, sic to get that final S-type noise. So there were some things that were tough.
GROSS: We have lips, and that helps us pronounce. Parrots don't. They have beaks, so that must be an impediment for subtle variations of vowels and consonants?
Dr. PEPPERBERG: Not so much the vowels. His vowels come out looking exactly like ours; they're really quite similar. The change comes on those consonants, those P and V and B sounds that we use our lips, and he uses esophageal speech to do that. So just like people who've had laryngectomies, he will kind of burp those sounds in some way. They don't look that different on the sonograph, but they come up at very different energies.
GROSS: What do you think your work with Alex, and with your other parrots, disproves about preconceptions of the abilities of animals to think and communicate?
Dr. PEPPERBERG: Well, what Alex really did was lay waste to the term birdbrain as something derogatory. He really did show that this creature with a brain the size of a shelled walnut could do the same types of tasks that the apes did and the dolphins did and in many cases, young children could do. It was a major breakthrough.
Before I started my work, which was in the '70s, most people were studying pigeons. They were studying through a procedure called operant conditioning, where you starve an animal down to about 80 percent of its normal body weight, put it in a box with nothing in the box other than a couple of buttons, and you do a lot of tasks using those buttons and trying to see what the animal can do. And obviously, we did things in very different ways. Instead, we treated animals...
GROSS: I should just say, you're talking about starving the animal, that's so it would be receptive to food as a reward stimulus?
Dr. PEPPERBERG: Correct, correct. I'm sorry, I should have mentioned that. Yes. And you would, the rewards would have nothing to do with the task that you were giving the animals. So, if you were trying to do, say, you know, that the animal could do a match to sample, meaning you show it a red light, and then you give it a sample of red and green lights and see if it could learn to hit the red light appropriately. And then, if you change the sample to a green light, could it switch over to hitting the green light. And you give it a little piece of food as its reward. This was not the way to treat an animal if you wanted to get communication.
Another big difference in what we did was, we trained Alex to label objects that he initially wanted. So there was a real incentive for him to learn to say key because that was something he could use to scratch himself, or wood because that was something he enjoyed chewing. And these were his primary rewards, the close correlation of the label and the object to be learned.
GROSS: You know, as you mentioned before, you didn't use the behavioral sciences model of conditioning with your parrots. Can you describe the training models that you used to teach your parrots how to communicate?
Dr. PEPPERBERG: We used something called the model/rival technique and it is very, very simple. We started by finding objects the bird wanted, and we would decide to train him those labels. So the bird would be on a perch. My student and I would have this object, say it was a piece of wood that the bird really wanted to chew. And I would show it to my student, who is the model for the bird's behavior and its rival for my attention, and I'd ask her, what's this? And she'd say, wood. And I'd say, that's right. It's wood. And I'd give it to her, and she'd go, wood, wood, wood. And she'd proceed to break it apart. And the parrot's, you know, practically falling off the perch. Alex really wanted this object, and he was really watching.
And then we exchanged roles of model, rival and trainer. So the bird saw that one person was not only the questioner and the other one the respondent, but it was an interactive process. And she'd show it to me, and I'd go quack, and she'd turn away and go, no, you're wrong. So the bird would see that not any weird new noise would transfer the object. And she'd give me another chance. And I'd say, wood, and I'd get it, and we'd play that game again. And we did this several times. And then we'd show it to the bird. Now, at the beginning, Alex obviously wouldn't just say wood, but he might go something like, ood, a new noise, and we'd reward that. And then over several weeks, we would shape it up into something that sounded like wood.
GROSS: And what about things like numbers? How would you get Alex to understand that three objects meant the number three?
Dr. PEPPERBERG: That was really interesting because for him, you know, wood was wood, so what's this business of all these numbers of woods. And we'd do it using the same modeling technique. I'd show, you know, this is wood, and then I would put a couple of more pieces on the tray, and I'd say, this is three wood. And the student would go, three wood and, you know, get the reward, and then she could play with all three pieces of wood. And Alex, if he said three, could play with all three pieces of wood.
GROSS: Did you think he could count?
Dr. PEPPERBERG: He did. We really had data on counting. And when we did a study on addition with Alex, we would put numbers of things under cups. And there would be like, say, two nuts under one cup and three nuts under the other cup. And we'd lift the first cup and we'd say, look, and after a second, cover those nuts. Pick up the second cup, show it to him for a second and say look, put it down. And then with both cups on the tray covering the nuts, we'd say, how many? And he'd say, five, and he was fine, but he couldn't do five and zero. Five nuts under one cup, and no nuts under the other. And we couldn't figure that out at first. Every time we did that, he'd say six. And then it finally dawned on me that oh, maybe he's doing what humans do. We're not giving him time to actually count.
So we finished the second half of the trials, giving him now maybe five or six seconds, and lo and behold, he could do it. So this was our real evidence that he was literally counting because he needed time to perceive all those things under the cup.
GROSS: What do you consider some of the most advanced things linguistically and conceptually that Alex was able to achieve?
Dr. PEPPERBERG: When we were studying concepts of same and different, we taught him the label none to refer to absence of similarity or difference. So we'd show him two objects and say, what's same or what's different, and he'd say color, shape, matter, or none, if nothing were same or different. Without any prompting, he transferred this when we did concepts of bigger and smaller. So the first time we showed him something that was the same size, and we said, what color bigger? He looked at me, and he said what same? And I said, well, you tell me? And he said none. And then the really exciting part came when we were doing number comprehension. And we gave him trays of different sets of objects with different colors and numbers. So there would be, for example, a tray of three yellow blocks, four purple blocks, and six orange blocks. And the question would be, what color three?
So he'd have to understand the number three, what it meant, find the set of blocks that were three, which would be all mixed up - all the different colors would be mixed up on the tray - and then tell me the color of the set that was three. And he did this for about 12 sessions perfectly well. And then - it's a bit anthropomorphic - but he would get bored, and he would - what he would start doing would be to throw everything on the tray on the floor with his beak, just knock it off or give me colors that weren't on the tray or turn his back to me and say, want to go back, and be very clear that he didn't want to work.
And so you start getting inventive, and you start using things like Jelly Bellies instead of wooden blocks, and that he'd get one of those for his reward. And you're pushing the edge of the envelope a lot. And then one day, I come in and I show him trays, and it was three, four and six things on the tray. And I said, Alex, what color three? And he looks at me, and he looks at me and he says five. And I'm thinking, huh, there's no five things on the tray. And so I say, Alex, come on, what color three? Let's go.
And he looks at me again, and he says five. And this goes back and forth several times. And I'm thinking, what is going on here? He's not throwing everything on the floor. He's not giving me wrong colors. He's saying a different number. And there isn't any of the stuff on the tray. So I finally said, OK, smarty, you know, what color five, not knowing what to expect. And he looks at me and he says, none.
So not only did he transfer this information from that other task to this task, but he was responding to an absence of number, a kind of zero-like concept. Plus, he had figured out how to manipulate me into asking him the question that he wanted to answer, which I think was pretty, pretty sophisticated on his part.
GROSS: My guest is Irene Pepperberg. We'll talk more about the research she conducted with her parrot, Alex, after a break. This is FRESH AIR.
(Soundbite of music)
GROSS: My guest is Irene Pepperberg. We're talking about her work with her now-famous African grey parrot, Alex, with whom she conducted pioneering research into animal linguistics and cognition. Alex died in 2007. Pepperberg's memoir is called, "Alex & Me."
You describe a time when you were teaching at the University of Arizona in Tucson, and you were living outside the city. And you had brought Alex home with you, which you didn't typically do. And there were a couple of owls outside the window, which terrified your parrot, Alex. And would you describe the communication that happened after he got really freaked out by the owls?
Dr. PEPPERBERG: Yeah. I mean, normally, when I would bring him home, the first minute or two, you know, he was outside of his carrier and in his trap - this big cage that I would have for him. He would say, want to go back. And I'd say, oh, just calm down. You're fine. And then he would look around the cage, see there was food, there was water, there were toys, and OK, he was fine, and he would settle down.
Well, this time, he was just going on and on - want to go back, want to go back, want to go back. And he's staring out the window. And I finally realized that there were these little screech owls - little, tiny screech owls that were nesting up there. And so my first response was to just close the shade and say, look, they're out there. You're in here. You're safe. But Alex had something called object permanence. And he knew those owls were still out there, and he just kept insisting - want to go back, want to go back. So, I had to take him in the carrier and bring him back to the lab that night. And he never really came back to the house after that.
GROSS: Alex was hospitalized with a life-threatening infection. How did you try to communicate with him when you had to leave him at the hospital?
Dr. PEPPERBERG: Oh, that was - that was so incredibly difficult. He knew phrases like, you know, I'll see you tomorrow, I'll be back, because we would say that every night when we put him in the cage. And so here, we're not leaving him in his normal place. I mean, we're putting him in this little hospital cage, in a strange place with all these people, many of whom he didn't know. He knew the vets themselves, but not the technicians. And as I walk out the door, he looks at me, and he says in this pitiful voice, I'm sorry. Come here. Want to go back. And you sit there and look at him and go, oh, well, how am I going to explain this? And I just kept saying I'll be in tomorrow. I'll see you tomorrow. I promise. I'll see you tomorrow. And finally, he calmed down, and, of course, I made sure that I was there tomorrow.
GROSS: How do you think he knew to say I'm sorry?
Dr. PEPPERBERG: That was - again, it was something we called contextually applicable. When he did something bad, you know, if he bit somebody or if he threw things on the tray and we'd get angry at him, and, you know, we'd say bad boy, you know, don't do that, no. He learned over time that, you know, the phrase I'm sorry was very good. He would say it in this pathetic little voice -I'm sorry. And, of course, you're a little - you go oh, you know. Your heart melts even though you know there's no contrition. So that was something that he had associated. And I guess something in his little bird brain said, oh, they put me in this horrible place because I've been a bad boy. Maybe if I say I'm sorry, you know, things will get better. I mean, I'm just guessing at that.
GROSS: You know, this story of Alex's death is just so sad. You learned about it through an email that you got…
Dr. PEPPERBERG: Hmm.
GROSS: …in the morning at your home from one of the people working in the lab, who delivered the sad news. Were you able to actually find out how Alex died?
Dr. PEPPERBERG: The autopsy did not show a lot. By sort of subtraction, my veterinarian assumed that it was heart arrhythmia because there was nothing obvious. And he did have a little bit of arteriosclerosis, which meant that if there was a heart arrhythmia, things could've shut down, and it could've happened very quickly. And certainly, you know, we gave him the best foods and healthy foods, and he had just had a checkup the week before. I mean, it was just like the middle-aged guy who goes to his doctor, and the doctor does all the tests and says, hey, you're great. You'll live another 30 years. And the guy walks out the door and, you know, collapses. And that's sort of what happened.
GROSS: What do you believe now about the potential of animals to communicate that you weren't sure of when you started your research?
Dr. PEPPERBERG: That that potential is much greater. I think that for many animals, we need to figure out the appropriate channels to use. Obviously, Alex could talk. People who work with apes use computers and sign language. People who work with dolphins also use computers and sign language. It's a matter of figuring out what medium. I mean, to emphasize the communication with humans in some ways is unfair because they have their own communication systems that work wonderfully well in the niche in which they live. And in a sense, pushing them to communicate with us is unfair. But it's one way of our actually getting - as Don Griffin(ph), my mentor, would say - getting a window into their minds to actually determine how they process information, how they think, by giving them these tools.
GROSS: Irene Pepperberg, thank you so much for talking with us.
Dr. PEPPERBERG: Oh, you're so welcome. Pleased to be here.
GROSS: Irene Pepperberg's memoir about her research with her parrot is called "Alex and Me." Our interview was recorded last year, when it was published. Our Animal Week series continues tomorrow.
(Soundbite of music)
You can download podcast of our show on our Web site freshair.npr.org. I'm Terry Gross.