>There is what has been called on the list the "Sonority Contour Principle" or
>something like that. It states that the syllable peak will be the most sonorant
>phone of the syllable, and phones around will go from it with decreasing
>sonority, the order of sounds from most sonorant to least sonorant being:
>vowels>approximants>laterals, trills, flaps, nasals>fricatives>stops.
>
>According to this principle, most beginning consonant clusters will be of order
>less sonorant-most sonorant. So [tS], [sw], [kw], [pr], [dl], [ks], [tlj] would
>tend to appear easily, while [mb], [vd], [zg], [jlk] should not occur. The
>order is the opposite at the end of the syllable.
>
>Of course, languages are quite rebellious little things and tend not to follow
>principles. Thus English and many languages allow initial [st] or final [tS],
>French has quite a lot of words with final [pR], Polish litterally adores
>initial [zd], and I'm not mentioning Georgian ;)))) .
>
>Other things are probably working there: clusters with disagreeing voicing are
>rare (so you expect [ks] but not [gs]), somehow clusters stop+nasal are not
>very common ([kn] has disappeared from English, but is still present in other
>Germanic languages, but not in Romance ones for instance (Although it doesn't
>mean much. Old French seemed not to be able to pronounce initial [w] and
>borrowed Germanic words beginning with this sound by replacing it with [gw])
>
>The problem of consonant clusters is a complex one, and I'm not sure that
>frequency studies have been made cross-linguistically about it (I mean, real
>ones, with representative samples from every known language family :)) ).
>
>Christophe.
>
>
http://rainbow.conlang.free.fr
>
>Take your life as a movie: do not let anybody else play the leading role.
>
>
>