Ruhlen doesn't claim that every language would have the same sound. He
notes that given the infinitude of possible sounds that could be selected
by all languages for water (in Mandarin it is shui pronounced shwi) a
surprising number selected a aqwa type form. From this he suggests that
since there is no similar widespread contender for water, that the
original word was some form of aqwa.
This type of analysis is precisely what linked sanskrit to English when
Jones originally proposed the indoeuropean language family. All Ruhlen and
Greenberg are doing is extending that type of analysis to older language
divisions. Indoeuropean may only be 12,000 years old. The splits he is
talking about are earlier.
The Australian family of languages has probably been separate from the rest
of the world's languages for over 60,000 year. From the time that
Australia was settled, we have some indication of the minimum time that
mankind has had language. By the way, in general the pattern of languages
and language families around the world, matches the pattern of genetics .
Since Australians are a separate group genetically and linguistically, it
is reasonable to assume they brought their language with them 60,000 years
Adam, Apes and Anthropology
Foundation, Fall and Flood
& lots of creation/evolution information