Avian Learning Experiment
Irene Pepperberg conducted an experiment with aims of investigating whether a parrot could vocally demonstrate a symbolic understanding of the concepts ‘same’ and ‘different’. This experiment was a case study involving one African grey parrot, Alex.
Alex had been involved in previous research on communication for many years, hence, he already had significant vocabulary. He could name colours like yellow, green, red, blue and grey, several shapes and different materials e.g wood, metals, cork and paper.
The purpose for training Alex was to teach him to reply with a categoric label rather than just describing the appearance of the item presented. He was involved in training for about two to four times weekly. The method of training implemented was known as the “Model/Rival Approach”. This is a concept where the trainer would present objects to another person, then ask questions about the item. When desirable answers were produced, rewards were given. This was done in the presence of Alex who saw the interaction. The roles of trainer/model were then reversed.
The beginning training sessions implemented the concept of continuous reinforcement in order to create a close association between the object and its label. When Alex responded with a correct label, he was rewarded by being given the object. If he answered incorrectly, he was scolded by having the object taken away.
After training, Alex was tested by secondary trainers who he was unfamiliar with. In every trial, Alex was shown two items that could differ in either shape, colour or material. Alex was asked “What’s same?” or “what’s different?”. The objects presented to him included the familiar items used during training as well as new items which he had not yet encountered during training. The principal trainer was always present during each trial but sat facing away from Alex, unable to see objects. This trainer would repeat Alex’s responses and if he was correct, he was praised and given the item. Getting the answer correct the first time contributed to the ‘First trial’ results. However, when incorrect, Alex was scolded and had the object removed. This correction procedure was repeated until the desirable answer was given and the number of errors was recorded.
Results showed that when tasked with familiar objects, Alex responded correctly to 76.6% of trials which included correction procedures. For first trials only, he responded correctly for 69.7% of trials.?Results for novel or unfamiliar objects proved to be slightly better than the first. He answered correctly 85% of trials which included correction procedures and 82.3% for first trials. This proved that Alex could generalise his understanding of the concepts of ‘same’ and ‘different’ to new situations.
In closure, the researchers concluded that parrots are capable of understanding symbolic concepts of ‘same’ and ‘different’ and that they can also learn to respond to verbal questions to vocalise labels based on categories.
This study has heavily impacted psychology as it shows us how non-primate animals can be trained to communicate using modelling and reinforcement. This has taught us that other animals are also able reason about abstract categories and apply these concepts to novel situations.