Question: Are birds that mimick human language e.g. parrotts etc. capable of understanding the language? Or are they simply associating a phrase with an action, similar to eccelelia in some autistic children? (Joe suggested i asked you!)

Carolyn McGettigan answered on 17 Mar 2010:

Thanks for passing this on to me! Great question. I think the most fascinating thing about parrots and songbirds (like zebra finches) is that their amazing capacity for sensorimotor behaviours with complex sounds. So, parrots can imitate human speech (and even common sounds in the environment, like ring tones) and songbirds can learn to produce very complex songs given to them by a ‘tutor’ bird during their early development – good evidence for sophisticated input and output processes with sound. Moreover, there is evidence that songbirds can modulate and refine their song according to the social context, so a songbird will sing differently if it is in the presence of a mate than if it is alone. These are certainly sophisticated behaviours. Sticking with songbirds, there is evidence that a gene identified as important in speech development in humans (called FOXP2) is also critical for the development of the neural systems for song in birds. So, there are potential evolutionary commonalities between us and birds in aspects of this sensorimotor learning behaviour.

However, as you rightly point out, being able to hear something and repeat it back isn’t enough for language. Although many animals and birds can use sounds in communicative ways, they don’t have the sophistication of higher-order conceptual processing in order to use abstract language like us. So, after long periods of intensive training, there are cases where a researcher or pet-owner can teach a mammal or bird to respond to simple sentences, for example asking it to fetch a specific object. I watched a BBC documentary about dog behaviour recently and there was one case where the dog could respond to hundreds of different words. But when you consider how quickly a human child learns to talk, develop self-awareness and higher-order concepts like Theory of Mind (attributing mental states and intentions to others), it’s obvious that even other primates are lacking something that we’ve got to help us do that. The mechanism for learning to respond to sentence commands in parrots or dogs may mainly be one of associative learning -‘hear X, do X’ – and when it gets any more sophisticated, this owes a lot to the intensive and lengthy interaction of these animals with humans. In other words, I seriously doubt that even this most basic evidence for language would emerge in the wild.