Other Animals

Inside Dr. Pepperberg’s Lab: When Favored Treats Aren’t Enough

As I’ve noted previously, we always try to give our birds specific, task-related items as rewards when they respond correctly to our queries. So, for example,  if they identify a piece of wood, they get to chew it up. Same for pieces of paper and such. Ditto if they give the correct color or shape—they get to chew on the green ring or the four-corner piece of wool. The point of such rewards is to ensure that the birds understand the one-to-one correspondence between the label they are using and the object that they receive. With this procedure, they quickly see that their labels are meaningful and can be used to obtain objects and alter their environment.

Sometimes, however, we just can’t give them these so-called “referential” rewards. Maybe the birds are so bored with the toys from extended play, or the objects are too expensive to replace. In those instances, we train them to use the label “want,” so that they can request something else as a reward. In that case, we are still maintaining some reference, as we say things like “You’re right! It’s an X! Now, what do you want?”—and then, within reason, we give them exactly what they request. Again, they see that their vocalizations are meaningful.

Occasionally, the birds are actually directly working for a very desirable reward (e.g., finding a hidden nut) rather than labeling something. In those instances, we try to keep the rewards small enough so that we can fit plenty of trials into a given session before they stop working. Usually, treats like bits of raw cashews or organic crackers are such valued items for them, and they seem like bottomless pits in terms of consumption, that we don’t have any problems getting through the tasks.

Griffin Reaches His Limit

Every once in a while, however, the system breaks down, and if the resulting standoff between bird and humans weren’t so frustrating, it would actually be amusing. We had one such instance when a colleague was visiting and needed to record Griffin saying similar labels (e.g., “cork” versus “corner”) in order to analyze how these utterances differed in terms of various frequency patterns called “formants.” (Each human sound that we—and the parrots—produce is made up of energy bands, called formants, and the relative positioning of the bands, particularly the first and second, can be used to identify that specific sound…such as the vowel in the word “hat” versus that in “hit.”)

A long time ago, my students and I did a similar analysis of Alex’s vocalizations (Patterson & Pepperberg, 1994, 1998), in that case looking at labels that were even more closely related—so-called “minimal pairs” like tea and pea that differed by only one phoneme (a basic or “minimal” unit of speech). The idea for the current study is to see whether Griffin uses the same types of strategies to produce his labels as Alex had used, and whether the more advanced types of analyses now available will provide additional information. The problem was that we needed at least 20 instances of each of the labels that my colleague wanted to study, which meant that we were asking Griffin to say the same things, over and over and over, for several days in a row.

Initially, Griffin was fine. The several microphones in his face didn’t faze him, and he labeled the objects and took them as his reward. Then he began to balk. Understandably—he wasn’t making mistakes, and yet we kept asking the same questions! So, we started bribing him with nuts…either very small pieces after every response, or somewhat bigger pieces after several responses. That worked for a while. Then we had to go to the bigger pieces for every answer. Finally, by Friday evening, he had absolutely had enough: Every time we brought out a new object, even if we showed him a nice big piece of cashew, he would turn away from us and start to preen…or he would state “Wanna go back!” (i.e., to his cage). Clearly, at some point, even his favorite treats just weren’t enough to keep him working at what he viewed as a totally boring task. We didn’t take his picture, but anyone with parrot would have recognized the equivalent of the teenager’s eye-roll that tells a parent, REALLY!!?!?! You’ve GOTTA be kidding!

We really couldn’t blame Griffin—he’s a living organism with his own wants and desires, and we had to respect his wishes. In the normal course of events, we would never push our birds to work when it is clear that their interest in a task has waned. This time, however, because my colleague had only a few days to get all the necessary recordings, we were just hoping that bribes would keep Griffin happy. And they did, but only to a certain extent. I guess that all creatures have their limits.

Patterson, D.K. & Pepperberg, I.M. (1994). A comparative study of human and parrot phonation: I. Acoustic and articulatory correlates of vowels. Journal of the Acoustical Society of America, 96:634‑648.
Patterson, D.K. & Pepperberg, I.M. (1998). A comparative study of human and Grey parrot phonation: Acoustic and articulatory correlates of stop consonant

Products You May Like

Leave a Reply

Your email address will not be published. Required fields are marked *