Re: Futurese

Ray:
> At 7:16 pm -0400 29/4/02, Javier BF wrote:
> [snip]
> >> Several years ago, The University of California, Los Angeles Phono-
> >> logical Segment Inventory Database, made a famous linguistical sur-
> >> vey (known in short as the UPSID survey) in which they analyzed the
> >> phoneme systems of 317 languages, chosen so as to include one
[...]
> [snip]
> >> for the liquids (/l/ and /r/), 96% of the languages used at least
> >> one, 72% used more.
>
> Not helpful.  Somewhere about a quarter of the world's population speak a
> language with only one.
I hadn't been following this thread, but I would like to butt in to
point out that the reasoning from UPSID is bogus phonologically as
well as (as observed by Ray) statistically. UPSID -- at least as
reported in Maddieson 1984 -- gives no solid information about the
phonetic realization of the phonological segments, and nor does it
give any indication about how 'secure' the phonemic analysis is.

It is quite possible for a language to contrast (say) [l] and [r]
yet for no /l/:/r/ contrast to be reported. For instance, [l] or
[r] might be a realization of /d/.

--And.