Mark Davis wrote:
> Ah, now I understand the 'guilty' part.
OK, now I feel relieved...
> The UTC decided that rather than change the base rules in #29, it
> would provide a prominent example of how those rules would be tailored
> for French and Italian, citing those rules in that section. So for
> that section, the only requirement is the set of letters X for
> French/Italian specifically that would break in the sequence LETTER
> APOSTROPHE X -- it doesn't have to be all vowels.
>
> I am doing the other list for something unconnected to #29, but we do
> need that set X also. However, if I and U are sometimes consonants in
> Italian, it does make the #29 apostrophe rules trickier -- unless
> those cases are such a small percentage that it is not worth
> contorting the rules for.
For the limited purpose of tailoring UTR#29 to Italian, consonants /w/ and
/j/ (spelled "u" and "i") can be considered as vowels, because they behave
pretty like vowels as far as elision is involved.
E.g., although "uovo" (/ˈwɔ vo/ = "egg") phonetically starts with a
consonant, a preceding word is elided as if it started with a vowel:
"quest'uovo" (/kwes tˈwɔ vo/ = "this egg"). In fact, sequence /wɔ/ is
considered a diphthong in Italian grammar.
But if the purpose of the vowels list is something different -- e.g. finding
syllable boundaries or hyphenating -- then handling "u" as a consonant would
not work for Italian in this case, because the two-syllable word "uovo"
would be segmented as "u-o-vo". Similarly, it wouldn't work for English
"quo-ting" or Spanish "bue-no", both cases where "u" represent consonant
/w/.
In general, as many others have said, it is not possible to say which
letters represent vowels and which don't, especially if the language of the
text is not known.
Knowing what the list is for may help imagining which degree of
approximation can be acceptable.
_ Marco
This archive was generated by hypermail 2.1.2 : Mon Sep 09 2002 - 12:02:10 EDT