On 14/01/2014 08:28, Christophe Grandsire-Koevoets wrote:
> On 14 January 2014 09:12, R A Brown  wrote:
>> If Esperanto can take a diacritic whose original use
>> was solely on long vowels or diphthongs (and, indeed,
>> often marked a contraction of vowels into a single
>> sound)
> Not necessarily two *vowels*: in French,

I know - but I was talking about the _original_ use when the
circumflex was devised by the Alexandrian grammarians of the
3rd cent. BC        ;)

> most cases (if not all) of use of the circumflex come
> from an original vowel+ 's'! (the 's' became silent, but
> had resulted in a change of quality of the vowel before
> it. Eventually, the 's' itself was removed from the
> orthography, and the circumflex was used to mark that
> change in quality).

Yep - but that seems to me a reasonable re-use of the old
Greek circumflex, and it is used only above vowels.  Indeed,
the use of diacritics in French is clearly very much
influenced by their original Greek use.

In Welsh the circumflex is used only over long vowels in a
way not dissimilar to their use in Sindarin   :)

Applying the circumflex to consonant symbols was an
innovation in Esperanto.  But if such a _radical_ re-use of
diacritics has become acceptable in that conlang, then IMO
anything goes in a conlang.  Indeed, Sylvia's proposed use
of ï and ë seem to me very modest (almost conservative) in
comparison to Dr Zamenhof's      ;)

If /ni/ can change into /ɑ/, then practically
anything can change into anything.