--- In
cybalist@yahoogroups.com, "tgpedersen" <tgpedersen@...> wrote:
> > Ekwall has <Bræmbeleg> ca.1000, <Brembeleg> 1200, 1219,
> > <Brambele> 1235.
> >
>
> All one -l- short.
Just because Ekwall's data are incomplete. Here's more evidence,
collected by J.E.B. Gower et al. (1942. _The Place-Names of
Middlesex_, in the English Place-Name Society series, CUP):
<Bræmbelege> AS Charters, ca. 1000, with variants found in texts from
ca. 1128 to 1536: <Brambeley, -legh(e, -ley(e, -le(e>, <Brembeleg'>
1226, <Brameley> 1135-54, <Bramleia> 1201, <Brombele, Bromlegh> 1274;
<Brambelestrate, Brembeleye Strate> 12th c. and later = Bromley Street.
_and_ forms with double "l":
<Brembellee> 12th c., <Brembelley, -leye, -le> 1282-1345,
<Brombelleg'> 1251.
The simplification of a double consonant at a morphological juncture,
especially in the unstressed part of a compound or derivative was a
common phenomenon already in Old English.*) Here it merely reflects
the gradual obscuration of the place-name.
The first element shows the same range of variation as the OE
appellative <bre:mel, bremb(e)l-, bræ:mel, bræmb(e)l-> and the second
element is surely OE le:ag/le:ah 'clearing, glade, meadow', again with
the same kind of orthographic variation as in the examples above:
(-leg(h, -leie, -le:, etc.).
Piotr
*) Also in Modern English <wholly> and <solely> usually keep their
/ll/'s, but after an unstressed syllable (or a stressed vowel
resulting from contraction) the cluster is reduced at least
phonetica[l]y, as in <finally, really>, or even orthographically, as
in <subtly, simply>.