John Cowan wrote:

> Marco Cimarosti scripsit:
>
> > - The sound, /u/, is the same as the sound Greek Upsilon
> presumably had when
> > Cyrillic was invented;
>
> That's the difficulty: upsilon had by that date shifted past /y/ and
> was well on its way to /i/, as in Modern Greek. Your other arguments
> are cogent, though.

Well, they don't seem so cogent anymore when you look at this table of Old
Church Slavonic Cyrillic (I saw it after my reply to Richard Wordingham):

http://www.omniglot.com/writing/ocslavonic.htm

It seems that, at that time, all I said about Cyrillic U (sound, numeric
value, position in the alphabet) actually applied to the OY ligature.

The thing most similar to modern Cyrillic U is Izhitsa (the last letter in
Omniglot's chart), which has the correct numeric value, *nearly* the correct
shape (no descender), *nearly* the correct sound (it was /y/, not /u/), but
certainly not the correct place in the alphabet.

So, on a second thought, I'd say that Richard was right, at least for an
early stage of Cyrillic. Probably, it was originally the Omicron+Upsilon
ligature was adopted for /u/; later on, when it became clear that no letter
was necessary for noting the Greek sound /y/, the Omicron part of the glyph
was dropped.

--
Marco