[Spellyans] dictionnaire de l'Académie française

[Michael Everson]

On 30 Jan 2011, at 02:37, Owen Cook wrote:

> Michael made a legitimate point. In French, even though a ç is a type of c, it still needs to be
> written as ç.

Yes, and I made that point to show that Eddie's comment was not a convincing argument to make diacritics "optional" in KS. 

> With that said, Eddie's substantive proposal -- suggesting that there
> might be circumstances in which some or all diacritics might be
> optional -- might be worth discussing. Nothing, however, actually
> compells you to use diacritics in informal contexts anyway. Who cares
> if the spellchecker would mark as incorrect what you'll never run past
> spellchecker in the first place?

We are developing a spellchecker for KS. 

> 
> Personally, I agree with Christian; while the principle of using diacritics in general seems indispensable to me, I don't see that we need to mark y/e alternation with a diaeresis any more than we need to mark s/j alternation.

The s/j alternation affects a rather small class of words, nouns in [ɛnzə]~[ɛndʒə] and a set of verbal forms.

The y/e alternation affects a very large class of words, and overlaps with two other classes of words (also large) which do not have an alternation. I am quite sure that if all the diaereses were stripped from Enys Tresour, readers of both dialects (since both dialects are represented in the book) would have a tougher time recognizing words.

> Long y is not, in itself, any great anathema; KS's current distribution of i and y may need to be revisited in the course of the SWF review.

Throughout the rest of KS anomalous length is marked. We mark "stât" because by the rule the vowel should be short. We distinguish "mès" [mǝz]~[mɛz] from "mes" [meːz]. So right from the start, even ignoring alternation with ë, you have a problem with "bys" [bɪz] and "bÿs" [biːz]. 

You might want to write "bỳs" with a grave, but I GUARANTEE you that there tens of thousands of fonts out there in the real world that people want to use which won't have "ỳ" in it. Unless Cornish users are to be happy being restricted to two dozen web fonts, that's not a runner. And even if we used "ŷ" with a circumflex for the long vowel. In Wales they have problems with ŷ ý ỳ all the time. ÿ on the other hand is found in 

> (Don't get me wrong, I like the existing distribution, but if it needs to be adjusted for the sake of
> consensus, so be it.)
> 
> Besides circumflexes and graves to indicate anomalous vowel length, I do think that ù as in pùb and â as in brâs are really helpful as they indicate vowel qualities that would otherwise be unpredictable.

It seems to me that the very large class of bÿs/bës words is systemically unpredictable which is why we have marked it.

Regarding the s/j distinction: if one wanted to ignore font limitations entirely, one might write ŝ/ĵ (using the Esperanto consonants) for these in which case one could write bês/bŷs for the alternating words. Two things militate against this: (1) every time I mention it no one takes any notice, and (2) the font limitations would inevitably bite back anyway undermining whatever good the ŝ/ĵ/ŷ choice might offer. 

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