[Spellyans] <y>, <i>, etc

[Michael Everson]

At 13:47 +0300 2008-07-25, Owen Cook wrote:

>Nicholas said:
>  > We will either have to endure ambiguity on 
>the one hand or diacritics on the
>>  other. There is no other choice.
>
>Agreed. Let's have diacritics. I have no problem with that. (But <res>
>and <rës> are both unambiguous; both must be long.)

Both must be long, yes, by the normal quantity 
rules. The diacritic on <bës> shows that it has 
an alternate pronunciation <bÿs> and the lack of 
diacritic on <res> shows that it has no alternate 
pronunciation.

>  > As far as I understand it the proposal is to avoid having to write ÿ in
>>  words like bÿs, ÿs, etc. by using
>>  y only for the sound that may also be e, e.g. bys/bës; but writing bis
>>  'until' with <i>. And also gwinn 'white',
>>  me/my a vinn, etc.
>
>Nope, that's not the proposal I've been floating 
>since yesterday. I'm a diacriticophile, so I 
>have added <ì> back into the mix. But Dan's 
>original incarnation was diacritic-free. We 
>would have just have bìs and bys/bes where 
>length is indicated normally just as for every 
>other vowel in our system (i.e. depending on the 
>following consonant). As for disentangling res 
>words from rys/res words, we would just have to 
>put up with the ambiguity.

I don't want to put up with any ambiguity of this 
kind in KS. The current proposal doesn't have 
this particular ambiguity; therefore it is, in my 
judgement, superior to this new proposal.

>Not ideal, but this is the same ambiguity that 
>we currently have with s/j. If I'm a RMC user 
>reading a RLC text, I can't automatically 
>substitute <s> in my mind for every <j> I see. 
>Same goes for ow/ew -- I can't automatically 
>substitute <ew> for every <ow>.

That's not a good argument for having unnecessary 
ambiguity between <bës>-class and <res>-class 
words. Maybe s/j is insoluble, but bës/res isn't.

>  > Now, however, it is being suggested that <y> should be long and <i> short.
>>  This is not wise, to say the very least.
>
>KK, UC and UCR always had long <y>. The SWF does too. And so does KS,
>in the form of <ÿ>.
>
>Both the SWF and KS have short <i> in words like tiryow.

Only in words derived from monosyllables with 
<i>. Our rules (assuming the usual vowel length 
rules based on consonant quality):

<i> [i:] in stressed monosyllables and [I] in polysyllables derived from them.
<y> [I] in stressed and unstressed monosyllables.
<ÿ> [i:]~[e:] in stressed monosyllables.
<ë> [i:]~[e:] in stressed monosyllables.

-- 
Michael Everson * http://www.evertype.com