Re: THEORY: Difthongization across Europe

John Vertical skrev:
 >> John Vertical skrev:
 >>> A thing I've been wondering. I've noticed that some
 >>> quite similar vowel shifts occur in several European
 >>> languages around the beginning of the Middle Ages...
 >>> Most prominently, this bunch: Romance - E: O: > je wo
 >>> West Slavic - o: > u / wo SW Germanic - e: 2: o: > i@ y@
 >>> u@ <ie üe uo> Baltic - e: o: > ie uo Finnish - e: 2: o:
 >>> > ie y2 uo Northern Sami - E: O: > ie uo which does not
 >>> look co- incidental at all...
 >
 >> Clearly there may be areal influence at work, as has been
 >> suggested e.g. for Old French and Old High German E: O: >
 >> iE uO at about the same time (5th- 10th century -- in OHG
 >> the progeression in spellings
 >> e > ea > ia > ie can actually be observed in manuscripts
 >>     thru the 6th to 10th century!)
 >
 > Hmm, that progression partially messes up my hypothesis -
 > clearly it couldn't've been a part of the same M: > HM
 > difthongization wave then.

Not so sure about that: if Labov is right (and I see no
reason ATM that he wouldn't) Low Mid and Low would seem to
form one Low category and High Mid and High forms one High
category for diphthongization purposes, and that L:
dipthongization always passes thru the stages L@ > M@ > H@
while the > HM/HL stage is optional (as when the language
doesn't want stressed [@], as seems to be the case with
Romance; even in modern French /'@/ merges with /'2/! That's
why the result of O: diphthongization can be any of uO/uE/u9
(the latter probably > y9 > 2: in Old French!) since the Mid
element is always a 'peripheralization' of [@]. I'm sure OHG
_ea ia_ are spellings for [E@ e@ i@], and perhaps _ie_ was
[i@] too initially. Mind you the Latin alphabet had no good
sign for [@]! BTW Germanic "/e:/" and "/o:/" may have been
as low as [&: Q:] even in early OHG; they certainly were in
Common Germanic which had a square

|    i  i:     u  u:
|                        &i &:i Qu (Q:u)
|    &  &:     Q  Q:

vowel system -- probably because PIE *h2 and *h3 had merged
alteady back when in the dialects leading up to Germanic and
Balto-Slavic!


 > (Possibly influenced by it, still.) The Finno-Samic
 > and Romance changes might still be linked thru Slavic
 > & Baltic... I should relocate that document recently
 > mentioned by Amanda, I don't remember if there's a
 > date there.

Don't sweat over it: diphthongization of long vowels on
the patterns

- H: > @H > MH > LH
- L: > L@ > M@ > H@ > HM > HL

Where High-Mid counts as High and Low-Mid as Low, so you get

- H-M: > MH > LH
- L-M: > HM > HL

is a spontaneous and automatic consequence of the difficulty
of maintaining the same tung position thruout a long
monophthong, and can happen -- *will happen* -- whenever
long monophthongs arise, except that H: is more stable than
L:. Ironically any V@ can become V: given chance -- or
perhaps because of a tendency to avoid ['@] which causes [@]
to become some or other peripheral vowel, which may then be
identical to the preceding V in a V@ diphthong. The
alternative is to let the @ become L. In Romance and Finnish
you find H@ > HM/HL, in German you find V@ (_ie_ is not just
a cutesy spelling for /i:/ -- MHG actually has an M: > M@ >
H@ > (HM) > H: > @H > MH > LH cycle!) and English has both
(cf. a@ > A:) Some languages simply have V: > V@ or L:/M: >
VH regardless of the height of the V. Lots of Swedish
speakers realize /V:/ as [V@], which various subsequent
peripheralization of [@]. I have

|  - i: > ie      - y: > y2             - u: > u8
|  - e: > eE             - u\: > u\3\   - o: > oQ
|  - E: > E&      - 2: > 23\     - 3\: > 3\6
|but
|  - a: > a&                            - Q: > QO

and only in closed syllables; in open syllables I still have
long monophthongs, and there is no appreciable difference in
quality between these and the corresponding short
monophthongs. Go figure!

 >> Common Scandinavian A: had merged with Q:. Only one very
 >> old text -- the so-called First Grammatical Treatise
 >> distinguishes the two.
 >
 > Is this a document with Pre-Norse which still actually
 > distinguishes the two phonemes or a 1800sy diachronic
 > analysis? I would assume the latter, except I can't see
 > why would older analyses distinguish more vowels than
 > newer, if the language under analysis is still far older.

It's a 13th cy copy of a 12th cy work. A mean 12th cy
phonologist at work there BTW; he even used commutation
tests. interestingly the running text of the doc is in the
usual 13th cy 'orthography', using the proposed
'phonological' spelling only in expositions. Gaad knows if
the 12th century original MS used the spelling it proposed
in its running text. Prob'ly not!


/BP 8^)>
--
Benct Philip Jonsson -- melroch atte melroch dotte se
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
   "C'est en vain que nos Josués littéraires crient
   à la langue de s'arrêter; les langues ni le soleil
   ne s'arrêtent plus. Le jour où elles se *fixent*,
   c'est qu'elles meurent."           (Victor Hugo)