The vocalization apparatus of birds is completely different from ours. Yet,
some bird species are not only capable to produce a wide range of different
sounds, but also to imitate human speech. According to what I was ble to
learn from Wiki, some parrot species even have basic sematic rules: at
least, the scientists claim to find something like nouns, verbs and
adjectives. Singbirds of the same species tend to have different dialects.

Therefore, I've been wondering, if someone has studied sound changes in
birds vocalization. Well, I anticipate the problems of such a research, but
still, maybe someone has done it.

In fact, there are two other questions, more general in nature, which, I
think, answer the first one to a great extend:

1. To what extend can we guess how the sounds of speech would change if we
only know the construction of the vocalization apparatus and possibly the
spectra of sounds it produces?

2. Where from comes the data for the phonetical changes in human languages.
To be more precise, I presume that most of the morphological changes comes
from the data attested in real languages. However at least some of them may
be and are explained from the anatomical point of view. But this is mostly
done post factum, at least to my very limited knowledge. Does anyone,
however, try to deduce the development of phonology basing purely on the
anatomy? Or the data on human languages is big enough to make these things
'alhorithmically', like processing the sequences of symbols representing
the phonemes? Maybe a stupid question, though.