On Tue, 30 Sep 2003 23:16:00 -0700, JS Bangs <[log in to unmask]>
wrote:

>Oh, there's more consistency than you think. Consider:
>
>Consonants:
>
>p, t, k, b, d, g are usually equal to their phonetic values, or differ
>only non-distinctively. E.g. English {p} often represents [p_h], but
>aspiration isn't distinctive in English, and Spanish {b} represents [B],
>but [B] and [b] are allophones in Spanish.
>
>f, s always represent at least [f] and [s], though some languages may also
>use them for [v] and [z]. If {s} can have meanings other than [s] and [z],
>it's [S], and usually this is conditioned by surrounding letters.
>
>Vowels:
>
>Vowels are more variable. If you strip away accents, all of the
>following are true in every language I know:
>
>a is [a] or [A]
>e is a mid front unrounded vowel
>i is a high front unrounded vowel
>o is a mid central/back vowel
>u is a high central/back vowel
>
>Vowels may have values other than these, but they at least sometimes have
>these values.
>
>I'm curious to see what inevitable exceptions people bring up.

To give an example from the list's official language :-), there's the Dutch
"g", which in some dialects is [x]. A few other exceptions that I can think
of: "u" is [y] in French (not a back vowel) and a sound traditionally
represented as [9] in Dutch (sounds more like [8] to me; in any case, it
doesn't seem to be a high vowel or a back vowel, and the long "u" is [y]).
In Swedish, "o" is a high back vowel (taking the place of "u", which has
moved forward). "s" is [S] in Hungarian (where [s] is spelled "sz"), and
"d" is [z] in Vietnamese, but of course these two aren't IE languages.