Song Learning in Birds

The auditory template model is the basic model for explaining song learning in birds, though there are many bird species who deviate from it. The model was developed by Peter Marler through experiments on chaffinches. He observed that chaffinches were only capable of learning song when they were hatchlings and during their first spring, and that they ceased to be able to learn when they reached a fully crystallized song or had high levels of testosterone. Additionally, the chaffinch hatchlings wouldn’t learn songs from heterospecifics, even when they hadn’t heard any other kind of song before. Lastly, when males were deafened they produced songs far more distorted than the ones produced by males without a tutoring song. From these data, Marler hypothesized that birds must possess some kind of crude template, and that they match this crude template to the songs they listen to until they hear their conspecific song, which then causes the template to crystallize into an exact template. Later, when chaffinches reach their first spring and testosterone levels begin to rise, they begin producing their own songs. With normal hearing, the birds can listen to themselves sing and match the song they produce to the template they’ve remembered. This explains the babbling phase most birds go through before reaching a crystallized song, as they need time to match the sounds they produce to the template sounds. Eventually, they reach a fully crystallized song that is a more or less faithful version of their species’ song. Though the data from several playback experiments seem to fit this model, there are many exceptions to it. For example, white-crowned sparrows will learn conspecific song played from a speaker, but will not learn heterospecific song learnt from a speaker. However, when the sparrow hatchlings are placed in social contact with heterospecific tutors they are able to learn heterospecific song, which seems to go against the idea of a crude template that filters heterospecific song. Another problem with this model has been highlighted in European robins which have an open-ended learning system. European robins have wide repertoirs spanning many hundreds of songs, each different from the last, and have no defining species specific song. In swamp sparrows, until shortly before full song crystallization, they produce 12 elements, which are then reduced to the number needed for their full song. These findings have led to the rise of another model known as the action model. In the action model, rather than having an auditory template, birds have pre-encoded elements that they must learn to put in the correct order to produce their species’ song. This model would explain the reduction in the number of elements used in swamp sparrow song. As for species with open-ended learning, one can only hypothesize that they have the capacity to learn throughout their lives and that they can learn as many elements as their syringeal muscles allow.