New research just published in the Proceedings of the Academy of Sciences shows that at least two bird species are able to learn abstract grammar rules. Birds have long been known for their language abilities but grammar has largely been thought of as a human skill. Scientists Michelle Spierings and Carel ten Cate have now been able to show that zebra finches and budgerigars are capable of learning grammar rules and sentence structures.
Budgerigars, also known as budgies (or “parakeets” in the United States), are small, brightly colored parrots capable of mimicking sounds and speech. Zebra finches are small songbirds, commonly kept as pets or lab animals. The researchers kept individual birds in isolated environments, teaching them to recognize sentence triplets. They did this by playing sounds in a certain pattern, XYX or XXY. Upon hearing the correct sound pattern, the birds had to peck a key in order to get a food reward.
Later, the birds were retested but this time, the test strings contained a mixture of known sounds from earlier and new sounds that the birds had never heard. Both bird species were able to identify the correct “sentence” patterns. The two species used different strategies, however. The zebra finches compared the test strings to the patterns they learned in training, using the ordinal positions of sounds to discriminate between the two sound pattern types. The budgies showed more abstract thinking. They were able to correctly discriminate between XYX and XXY patterns even when completely novel sounds were used.
The authors concluded that birds are capable of detecting grammar rules in sound strings. Budgies seemed to have an especially deep level of understanding and were capable of applying what they learned to completely novel, artificial sounds. The concept of grammar has generally been thought of as uniquely human but this research shows that at least some bird species are capable of recognizing sentence structures.
Spierings, Michelle J., and Carel ten Cate. Budgerigars and zebra finches differ in how they generalize in an artificial grammar learning experiment. Proceedings of the National Academy of Sciences (2016).