07-12-03 21:05, Miguel Carrasquer wrote:
> 1) -VDV- > -V (D=voiced stop) LATE
> 2) C1C2 > C2 (C1,C2=stop)
> 3) CN > N (C=stop, N=nasal)
> 4) NT > ND
> 5) NS > S
> 6) VnV > VrV (Tosk) LATE
> 7) an > ën (Tosk) LATE
> 8) *ty, *k^y > s; *d(h)y > z
> 9) *kWe > se; *g(h)We > ze
> 10) merger voiced asp./unasp.
> 11) *k^ > th; *g^(h) > d(h)
> 12) *k^w > s; *g^(h)w > z
> 13) *k^l- > q; *g^l- > gj-
> 14) sk > h EARLY
> 15) s > sh LATE
> 16) s- > gj-
> 17) -s > 0; sd(h) > dh; sw > w
> 18) sp > f
> 19) sw > d
> 20) r > rr
> 21) l > ll; sl > ll
> 22) -nj > -j
> 23) r. > ri
> 24) wl. > ul
> 25) m. > a
> 26) y- > gj-
> 27) -y- > 0
> 28) w- > v-
> 29) -w- > 0
> 30) i > e before a:
> 31) u > y before i
> 32) i: > i
> 33) u: > y
> 33) -u: > -i
> 34) e > je
> 35) je > ja
> 36) je > i
> 37) o > a EARLY
> 38) a > e before i
> 39) e: > o
> 40) a: > o
> 41) o > ua/va/vo
> 41) o: > e
> 42) ei > i
> 43) ai/oi > e
> 44) eu > e
> 45) au/ou > a LATE
> 46) au- > ve-
2) EARLY. E.g. Lat. kt --> ft, while inherited *kWt/k(^)t > t. Beekes
doesn't mention *-tt- [-tst-] > *c^ > s, also EARLY
4) LATE: centum, parente- > qind, prind
5) Hm... I wish I knew
8) LATE. It was still productive in Latin loans (tj > *c^ > s, etc.), as
were the other pre-*j palatalisations, including 22). So was the
palatalisation of velars before front vowels (producing modern <q> and
<gj>: centum --> qind)
9), 10), 11), 12) EARLY. Modern <s> and <z> reflect Proto-Albanian
palatoalveolar affricates *c^ and *3^ (their diachronic behaviour
suggests labialisation -- a frequently found accompaniment to
palatoalveolars), and <th> and <d(h)> reflect some other type of
affricate articulation, perhaps simply *c, *3
13) The dispalatalisation *k^l, *g^(H)l > *kl, *gl is old, but the
change of the latter into <q>, <gj> is only dialectal in Albanian.
16) *s- > *z^- > gj- (in stressed syllables) is EARLY, since it doesn't
affect Latin loans (one possible counterexample, which might suggest a
very early loan, is <gji> 'breast', if from Lat. sinus rather than
inherited). 19) is probably EARLY as well -- anything so odd needs a lot
of time and some intermediate stages :-)
18) EARLY. In Latin and later loans sp- > shp- (spatha --> shpatë)
20), 21) I suppose a detailed discussion (and more refinement) would be
necessary. These points sum up various processes occurring at different
times
23), 24), 25) have to be considered EARLY. Beekes doesn't mention *l. >
li, for which a couple of examples could be given. Before laryngeals the
development is different: *R.h > aR
26) LATE: Lat. iu:dica:re --> gjykoj
28) is a trivial phonetic change, hard to date, but most likely LATE
(see 41a: the original prothetic consonant was likely *[w]); 27), 29)
are LATE, like other medial losses: lava:re --> laj 'wash', etc.
33a) LATE: de:bitu:ra, *padu:lem > detyrë 'duty', pylle 'forest'
33b) This final unrounding/fronting is earlier than the general fronting
of *u: (see the recent dicussion of Alb. dy), but hard to date in
absolute terms.
34), 35), 36) may represent an old diphthongisation of stressed *e, but
note that Balkan Romance *ie < E < Lat. e, ae also becomes Alb. <je>:
medicus --> mjek
39), 40) EARLY, but not as early as Doric loans into pre-Albanian:
ma:kHana: > mokën/mokër 'millstone'. The merger of *e: and *a: (as *a:)
is surely much older than the retraction to <o>.
38) LATE and productive in Latin loans: gall-i: --> gjel
41a) LATE: orphanu- --> vorfën/varfër 'poor'. Inherited *e:, *o: were
first merged, retracted and shortened to /o/ before they received a
prothetic consonant, cf. *h2ah1-tr-ah2 > *a:tra: > *otër > vatër 'hearth'
41b) Lat. o: and PIE *o: develop differently, so the fronting must be
EARLY (despite the relatively late date [or longer productivity!] of *u:
> y). I think o: _was_ fronted in Doric loans, but I can't remember any
sure examples (<pelë> 'mare' may be inherited or <-- Gk. po:los with a
secondary feminine suffix).
42), 43), 44) If 45) is LATE, the other monophthongisations probably
aren't very ancient either, though Latin evidence is not really decisive
either way.
46) looks completely spurious to me. <vesh> 'ear' comes much more likely
from *h2o:(u)s- than from *h2aus-
Piotr