On 6 September 2017 at 11:54, PETER BLEACKLEY <[log in to unmask]
> wrote:

> In Old English and Middle English, many words ended with an e, which has
> been muted in modern English. Function words are generally unstressed,
> which makes the word boundary before them weaker, so the rule "Fricatives
> are voiced between vowels" spread to the þ in function words. However,
> since there is a following consonant in "through", þ could not be
> interpreted as intervocalic there.

As I think Alex noted, at some point this purely phonological phenomenon
(whose character you describe) must have become transformed into a rule
sensitive to word-class, since only this would account for how come initial
/D/ is entirely absent from normal (i.e. 'cromulent') open-class words and
for how come initial simplex /T/ is (almost) entirely absent from
non-open-class words (the synchronic exceptions are _thir-(ty/teen)_, which
presumably arose metathetically subsequent to the death of the th-voicing
rule). The question then is whether, once this word-class-sensitive rule
arose, the failure of /Dr-/ to ensue was merely accidental (as Alex had
seemed to suggest) or rather (as I countered) because original /Tr-/ never
satisfied the phonological conditions on the application of the
word-class-sensitive rule (-- the aetiology of those phonological
conditions being as per your description).


> ----Original message----
> From : [log in to unmask]
> Date : 06/09/2017 - 11:42 (BST)
> To : [log in to unmask]
> Subject : Re: Why doesn't English permit tl or dl as true consonant
> combinations?
> On 5 September 2017 at 21:24, Alex Fink <[log in to unmask]> wrote:
> > The change in _thus_ is irregular, in that its conditioning is not just
> > phonological; it only occurs in function words.  Even if this change had
> > started out as a regular change conditioned by the prosodic weakness of
> > words like _the_, it seems to have undergone analogical extension to
> yield
> > _thither_ and _therefore_ and so on.  (That this didn't yield /Dr Dl/
> > clusters is in some sense totally accident.
> I'm not sure what our definition of accident is, but _through_'s failure to
> become /Dr/ shows us that the analogical change of initial TH voicing did
> not override the constraint "no voicing contrast in (most sorts of)
> clusters".
> > I can imagine a nearby parallel universe where e.g. English did like
> > German and ended up with "thrin" /DrIn/ instead of "therein".)
> >
> It may be accidental that this never happened ("may" = concessive "is"),
> but if it had it would have led to a more drastic alteration to
> phonotactics than the general initial TH voicing change.
> --And.
> <
> utm_source=link&utm_campaign=sig-email&utm_content=webmail>
> Virus-free.
> <
> utm_source=link&utm_campaign=sig-email&utm_content=webmail>
> <#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>