Print

Print


On 3.9.2007 Paul Bennett wrote:
 > Taking the purely ON vowels per BPJ:
 >
 > :    i i:    y y:        u   u:
 > :
 > :      e:
 > :
 > :    E E:    9 9:        o   o:
 > :
 > :                    A   Q   Q:
 > :
 > :    Ei  9y  Qu
 >
 > I think my sound changes would be something like...
 >
 >-  i > i\
 >-  y > u
 >-  y: > ju
 >-  i: > i
 >-  e: > e
 >-  E > a
 >-  E: > E
 >-  9 > ew
 >-  9: > jo
 >-  u > @
 >-  u: > o:
 >-  o > wo
 >-  o: > wo:
 >-  Q > o
 >-  Q: > o:
 >-  Ei > aj
 >-  9y > oj
 >-  Qu > Aw
 >
 > Leading to:
 >
 >:   i     i\ u
 >:    e e: @  o o:
 >:     E
 >:      a    A
 >
 > (plus existing /j/ and /w/ semivowels)

I'm sorry, but these don't seem natural at all to me. For
one thing front vowels usually don't become back, so y > u
is unnatural. Also, and more importantly these changes seem
all random, without the parallellism and chain shifts you
usually see in vowel system development. Neither does the
resulting system seem very close to neither an Algonkian or
an Iroquoian vowel system, but like a mixture of both, which
doesn't strike me as very realistic. In fact I doubt that
Iroquoian had penetrated as far north as New York around the
year 1000, so with a so early implantation of ON the
substrate would likely be purely Algonkian. I'd expect
changes like:

:- y        >      i (like Icelandic & Faroese)
:- y: (i:)  >      uj (like Faroese and Scanian)
:- 9        >      e
:- 9:       >      ew (like Scanian)
:- 9y       >      oj (like Faroese)
:- E        >      e
:- E:       >      aj (like Icelandic & Scanian)
:- Ei       >      ej (like Icelandic & Scanian)
:- u        >      o
:- u:       >      iw (like Scanian)
:- o        >      o or @ (much of Swedish has o > 3\)
:- Q:       >      Aw (like Icelandic)
:- Q        >      @ or A (u-umlaut of *A failed to apply
:                  in most of  peninsular Scandinavian, or
:                  was reversed since Q > 9/2 or O/o
:                  in isolated items)
:- Qu       >      @w if Q > @, else ow -- nowhere Aw in Scand.

i, e:, A, o: and perhaps (i:) remain unchanged.

Vr > V: and since Vn > V~ perhaps Vrn > V:~. You should
      probably get rid of r. Merger with l or n initially
      would be Algonkian-like, though some south Swedish has
      #r > zero (with Vr > V:) by way of r > R. Perhaps rr >
      S? certainly hr > S! Other SC may come from sCj.

resulting in a quite Algonkian

:    i    i:
:    e    e:   @    @:    o    o:
:                         A    A:

plus a bunch of Vj and Vw sequences. (jV and wV already
being plentiful in ON -- NB w > v\ was probably late!)

The mergers suggested are no way near catastrophic. NB
that Icelandic and Faroese unround y(:) and 2(:), and most
of Scandinacian simply loses length in i: and u: and
merges o with Q:. I might well have long front rounded
vowels unround, but uj, ew, aj, iw are in style for
Algonkian, and u: > o: would be unlikely (although u: > ew
wouldn't for Scandinavian, since actually some Scanian has
u: > eu and 9: > Eu!)

Besides languages don't avoid mergers; they clear up the
mess afterwards instead.

Note that fC probably > wC too, unless fC > hC. Certainly pp
tt kk would > hp ht hk, and Gt kt > ht. (Icelandic actually
distinguishes tt > ht from kt/Gt > xt!)

This would be scarily close to my own Hwinlenska, but I'll
probably not do anything with that anyway.

/BP