From: Philippe Verdy (verdy_p@wanadoo.fr)
Date: Tue Feb 17 2004 - 17:13:50 EST
From: "Peter Constable" <petercon@microsoft.com>
> So, the expression *hxC(V)- ~ *shxC(V) is saying, in relation to certain
> phoneme sequences known to exist in later varieties, that an earlier
> precedessor to the language(s) in question is believed to have had hC or
> shC, and hCV or shCV (with the vowel colouring on the h unknown or left
> unspecified).
In shorter term you agree that it contains a mix of a mathematical notation used
to write specific regular expressions (alternation noted '~' where regexp uses
'|', C and V being variables), and some plain-text characters (here only "hx"
and "shx").
It true that this looks more like a mathematical notation, but it still contains
phonological/phonetic letters which are the only ones that are addressed by the
proposal (here the subscripted x for undetermined vowel tainting of the
preceding h).
Also the complex expression could be as well written more simply as:
*(s)-hxC(V)-
without the alternation. My opinion is that parentheses are just there for
grouping and that the minus sign operator is used in this notation to denote the
optionality (where regexps use '?')
I see some similarities between the undetermined vowel tainting letter (the
subscripted x) and the leading star in the expression, used to denote an
undetermined infered historic letter. Shouldn't both use the same glyph with
just a distinct positioning? Could it be that the undetermined vowel tainting
letter be shown as a subscripted star ?
This archive was generated by hypermail 2.1.5 : Tue Feb 17 2004 - 17:57:50 EST