Birdsong: the interface with human language.

Abstract Birdsong is believed to provide the most adequate model for studying the learning process of human language. Songbirds require external song models after birth to learn their songs which contain highly complicated acoustic variables. They memorize their song models as ‘templates’ in their brains during a particular phase (sensitive phase), whereas vocalization starts in a subsequent step (sensorimotor phase). There may be two song templates: one innate and the other learned. A different nucleus in the song control system of the songbird brain may be responsible for each template. These nuclei are probably analogous to discrete cerebral nuclei of the human language system, including Broca's area.