Re: Phoneme diversity: questions

> 1) Why some languages, for example, Polynesian, have so few consonants, while Caucasian
> have dozen of them?

Natural variation. The number of consonants can range from few to plenty even within a single language family. Areal patterns occur (already Proto-Polynesian had a rather meagre consonant inventory) but there do not seem to be any overarching geographical motifs (several other Oceanic languages have considerably larger inventories).

> 2) Is there any study about growing or reduction of phoneme diverstiy along the evolution
> of language? Why and how can phonemes be "created"? A people can "add" phonemes
> to its lanaguage, or additional sounds always came from adstratal or superstratal source?

Conditional splits, combinatory reductions etc. create new "indigenous" phonemes all the time. Nasal vowels in French would be a simple example. "Loanword phonemes" can certainly occur but my impression is that it's more common for phonemes found in contact languages to develop on an inherited basis; considering French again, its front rounded vowels developed not in Germanic loans but by internal developments.

> 3) Why it seems that Western IE proto-languages had no palatals?

Why should they? Plenty of languages from all kinds of language families have no palatal/postalveolar consonants other than the basic glide /j/.

> 4) Is there any study about distribution of particular phonemes across the world? For
> example, rounded vowels (ö/ü) or uvular R?

Phonological typology is an thriving field of research. For a starting point, check out the World Atlas of Language Structures: http://wals.info

_j.