Irene Pepperberg, of the University of Arizona, has been studying the complex cognitive abilities and sophisticated communication capacity of a nonhuman animal for several decades. However, if you are not familiar with her work, the species under scrutiny may surprise you: African grey parrots (Psittacus erithacus). Pepperberg described her work in her presentation "In Search of King Solomon's Ring: Studies on the Cognitive and Communicative Abilities of Grey Parrots."

In Search of King Solomon's Ring: Studies on the Cognitive Abilities of Grey Parrots

Irene Pepperberg, University of Arizona

Most previous work in this strand of comparative cognition has used animals with large brains or a developmental history similar to humans (e.g. bonobos, chimpanzees, or dolphins). However, communication in parrots displays striking parallels to human speech development. For example, both humans and parrots must be exposed to the sound patterns of their group within a critical developmental period in order to communicate, and both species go through a babbling phase. Furthermore, parrots are clearly flexible in learning different sounds, unlike other birds whose songs are more or less genetically fixed. This is also similar to humans in their capacity to learn a wide variety of languages (organizations of sound), depending on what they are exposed to in their culture.

Most animal language research goes through a series of false starts before appropriate and effective methods emerge. Earlier researchers attempted to condition parrots using various learning theories, or simply exposed them to tapes. These studies were generally failures, except they showed that birds learn almost nothing this way (an important discovery in its own right). These "non-findings" highlight some significant differences. First, "bird brains" are structured quite differently: they lack the cerebral cortex which contributes so much to mammalian intelligence. However, the structure isn't all that counts - it's what the brain lets you do.

Pepperberg reasoned that these birds might learn more effectively in a social situation, because they regularly engage in spontaneous vocalization, and tend to "duet" in a call-and-response manner in the wild. To model the training technique on what she presumed to be a more natural form of learning of the parrots, she developed the "model-rival" paradigm. In this technique, one experimenter serves as a trainer, while the other acts as a model (engaging in distinctive interactions with the trainer) and simultaneously a rival (for the trainer's attention). This method, inspired by observational learning, is exceptionally effective, says Pepperberg.

Alex, the first grey parrot Pepperberg trained, learned to identify objects in his environment quite accurately when trained in the model-rival paradigm. Even the errors Alex makes are indicative of understanding. Common among the errors are "generic errors," in which Alex responds with a more general category instead of the specifics of the item the experimenter has in mind. For example, Alex would sometimes respond an identity question about a green key by saying "key" instead of "green key." These errors, which are accurate in a sense, may indicate that Alex understands the concept of category labeling. In general, Alex is quite adept at answering questions about color and shape, abstract properties of compound stimuli he is exposed to.

Ambitious experiments have also demonstrated that Alex can understand abstract concepts. Pepperberg has been able to replicate some of Premack's work on understanding of the "same-different" concept, originally demonstrated by Premack with common chimpanzees (Pan troglodyte). She has successfully taught her parrots to differentiate between same and different stimuli, and furthermore, has increased the difficulty of the task. Alex can accurately indicate which properties of different stimuli are dissimilar on 75 percent of different trials. In a follow-up test examining the transfer of this concept to another class of stimulus (toys), Alex demonstrated capability for even higher accuracy. In an aside, Pepperberg suggested that he is more intrinsically motivated to pay attention and respond correctly in an experiment involving interesting stimuli that he might wish to play with.

Some critics feel that Alex doesn't understand the responses he's making, but instead has been trained to emit a particular response when confronted with a given stimulus, in the manner of a conditioning experiment. Pepperberg could not address this criticism while confined to studying only one parrot. When her "subject pool" was expanded to include two more parrots, she was able to train them initially under different regimes to see if a conditioning approach would lead to similar responses. African grey parrots trained with basic video- or audio-only techniques performed worse in tasks in which Alex is successful. Pepperberg controlled for the degree of reward associated with each of the training tasks, so it would seem that operant conditioning alone cannot account for the success of the model-rival technique.

Pepperberg's research with Alex and the other African greys is nothing if not prolific. Too many interesting discoveries have been made with this species to cover in one article or even for Pepperberg to cover in her address. Fortunately for curious readers, her book, "The Alex Studies," presents the information on cognition and communication in African grey parrots in one volume.

Although much has been learned about this fascinating species and their cognitive capacity, we still don't know everything about how their minds work. There are still many unanswered questions, some of them fundamental. For example, how do parrots use these abilities in the wild? Why and how have these remarkable capabilities evolved? Although they occupy a very different ecological niche, are there conceptual similarities between their evolution and primate evolution? These questions are still being formed and explored, so we can look forward to learning much more from these interesting and intelligent avians. - Christopher J. Anderson

To learn more, please visit www.media.mit.edu/~impepper/