Latin languages

[Arnold M. Zwicky]

have we had this one before?

yesterday (5/20/04), in a Fresh Air interview, Samir Khader of
Al-Jazeera (or Al Jezeera, etc. etc.) maintained that the word "martyr"
doesn't mean the same thing in Arabic that it does in English and other
Latin languages.  now, there's a word/thing problem and a
language/religion problem in this, but here i'm drawing attention to
the expression "Latin languages", which clearly means 'languages
written in scripts based on the Latin alphabet'.

a google search throws up lots of uses of "Latin languages" as a
synonym of "Romance languages", which i'd known about and is in
standard dictionaries, plus a fair number of computer citations for
"non-latin languages" (always or almost always with the negative "non-"
and with a lower-case "l" in "latin") 'languages written in scripts not
based on the Latin alphabet', which i assume has already been collected
by the sharp-eyed lexicographers among us.

both the computer usage and khader's show a language/script confusion,
but they come at it from different directions: the computer usage is
intended to pick out scripts (namely, the ones that are troublesome for
the  most common encoding schemes) by referring to the languages that
use those scripts, while khader's aims to pick out languages (namely,
the ones not associated with the religion islam or with arab culture)
by referring to the scripts those languages use.  a subtle difference,
perhaps, but i think a very important one.

lexicographers: would this difference merit two subentries for LATIN
LANGUAGE?  if so, how would the two subentries be distinguished in
their definitions?

arnold (zwicky at csli.stanford.edu)