Re: partial letter replacement in languages?

Philip Newton wrote:
> On Thu, 9 Dec 2004 02:48:04 +0200, Rodlox <Rodlox@...> wrote:
> >
> >  is there a term for when a language is evolving/being changed, &
> > replaces
> > one letter with another (ie, /d/ becomes /t/) in nearly all
> > instances...yet
> > there are still words in the resultant language  which retain (to
> > continue
> > the example) /d/ ?
>
> I don't know a term for it, but just wanted to note that some
> instances of this come when there are two (nearly) concurrent sound
> changes such that, say, /d/ becomes /t/ while, say, /D/ becomes /d/ --
> so all or most original /d/'s disappeared but there are still /d/'s in
> the resulting language that used to be a different sound.
Yes; that sort of thing is explicable, and depends on rule ordering, which
presumably reflects sequence in time. Compare:

1. d > t
2. D > d

vs.

1. D > d
2. d > t

Otherwise, the situation Rodlox describes is truly the despair of the
Historical Linguist.  If "ALL" d > t, then any observed d's must be 1.
irregular ~failed sound change (but why?) 2. later borrowings 3. "dialect
mixing" and/or "substrate influence", which in the absence of any real
evidence are simply fancy, less honest, ways of saying "inexplicable"."

Sometimes analogy (paradigmatic pressure) is the culprit-- e.g. in Buginese,
*d fairly regularly > r, but there was also /r/ < *r. Verbal forms had
C-final prefixes (modern ma?-, maN-) so you get alternations like:
(original *d) base /r..../, prefixed /ma?-d.... ~man-d..../ VS
(original *r) base /r..../ pfx. /marr... ~manr..../

Then for some perverse reason speakers decide that some random forms
with -Cd- must have a base with /d.../ --or vice versa, some forms with base
/r.../ < *d have prefixed forms as if they were original *r.

The old Neogrammarian dictum "sound change proceeds without exception" has
so many exceptions as to be little more than a suggestion.

More recent theories hold that "sound change proceeds word by word";
consequently some very frequent (or perhaps very infrequent) words become
exempt.  This seems to explain such things as (assuming that same spelling =
same sound) Shakesperean (or earlier?) "great, meat" no longer rhyme (I
could be wrong on details here, but not on principle IIRC)

> Watch the Reply-To!
Ooh, caught it just in time............