How birds learn their songs
An experiment with chaffinches
One way to approach the problem would be to take some newly-laid birds' eggs, incubate them separately in soundproof chambers, hand-rear each young one (also in individual and acoustic isolation) and then see as each bird grew up what sounds it produced.
At Cambridge, England, some chaffinches were reared in these exacting conditions. Even when nearly a year old they sang very simple songs, representing as the experiment intended, the inborn component of the song. In the wild, a young bird would add the finer details during the first few weeks of its life, having learnt them from its father and other cock chaffinches within hearing; and again in the following year when the wild chaffinch's first breeding season approached, it came to sing in competition with neighbouring territory-holders.
These recordings of the impoverished and the full songs illustrate the difference, exposing what is apparently the embroidery that has to be learnt, though bird behaviourists are increasingly reluctant to adhere to the “nature vs nurture” dichotomy because new research is showing development to be very complex.
No-one knows the relative importance of inheritance and learning in each and every one of the world's 9,000 species of bird. Comparatively few have been studied. But among the typical songbirds it seems that learning partly by imitating elders as in the chaffinch example is probably quite widespread.
Tree pipit raised by a chaffinch
Song of a tree pipit raised by a chaffinch, recorded in England by William Thorpe.
The mynah's neural template
There are also species whose song (and indeed whose entire vocabulary) is handed down culturally, generation by generation. One such species is the Indian hill mynah, the glossy black bird with the yellow face wattles, so commonly kept in captivity for its talking (i.e. learning) ability.
Intrigued by this skill, a British ornithologist went to live in Assam in north-eastern India to study the bird's vocalisations and in particular their origin. He found that the entire repertoire is learnt by ear from its elders by each season's crop of youngsters. If the only companion a young caged mynah has is you or me, then it is you or me he is going to copy, hence the “talking” mynah. If, however, he is in a natural state then he copies his own kind and normally no others. Wild mynahs copy none but their kin. But since there are many others he could copy how does the bird choose? The suggestion is that each learning species has an innate predisposition to attend only to the language of its own species; it employs a “neural template”.