Re: [tied] Re: the glottalic theory

On Fri, 15 Nov 2002, Miguel Carrasquer wrote:

> On Fri, 15 Nov 2002 17:55:56 +0100 (MET), Jens Elmegaard Rasmussen
> <jer@...> wrote:
>
> >On Fri, 15 Nov 2002, Miguel Carrasquer wrote:
> >
> >> [...]
> >> The *b gap and the non-existence in UPSID of ['gW] indicate that if
> >> the glottalic theory is corrrect, the sounds in question must have
> >> been voiceless ejectives, not voiced implosives. That's to say at the
> >> _phonetic_ level, not necessarily the _phonological_ level. We may
> >> have had /t/, /'d/, /dh/ realized as fortis [t:], ejective [t'] (or
> >> preglottalzied [?t]), aspirated [th].
> >
> >That's sophistery! What *can* it mean in sane terms?
>
> I wasn't seriously suggesting such a phonological analysis, but I
> thought I'd better add the qualification in view of : "It does not
> matter what status the two letter with which I write /th/ have, for
> the whole business of "phonological typology" is one of - phonetics!
> Look through the many sound systems given by Ruhlen, they are all
> simply phonetic and not based on any deeper analysis, so phonetic is
> the level the IE stops should be assessed on".
>
> What it *can* mean is something like what we have in English, where
> /t/ and /d/ are realized as [th] and [t] (amongst other allophones).
> But I prefer a diachronic analysis where /t:/, /t?/, /th/ later become
> /t/, /d/, /dh/.

Is such a reanalysis also to be preferred *after* /t:/, /t?/, /th/ have
become /t/, /d/, /dh/? That is the kind of nonsense I am revolting
against. It will end up meaning that even a living system with /t/, /d/,
/dh/ is useless as a parallel since that can be reinterpreted too.

Jens