Dixon's "The rise and fall of languages"

[bwald]

----------------------------Original message----------------------------
Roger Wright responds to my criticism:
 
>Yes, of course. I wasn't referring to dialect diversification (that is,
>*language-internal* variation of a normal kind) but to actual splits
>between languages. Am I really the only person on the List to find
>actual language *splits* (as opposed to obvious and normal dialect
>diversification) inherently unlikely, and thus in need of some kind of
>non-linguistic explanation?
 
>From further exchanges involving AMR and Roger, I see that Roger makes a
distinction, which seems to be arbitrrary, between "dialect" and
"language".  In fact, by "language" (in context) he means a socially
determined collection composed of historically related dialects.  AMR
points out that this view is arbitrary, and gives many examples, e.g.,
Scandinavian, Serbian/Croatian, Hindi/Urdu, Provencal/Catalan, etc.
Meanwhile, historical linguists sub-classify reconstructed languages
without this social implication, although there is clearly some kind of
social implication for any form of diversification, e.g., they do not
insist that Low Germanic and High Germanic were ever "languages" as opposed
to "dialects" of West Germanic.  (Well. they don't all insist.  Some assume
that for each historical dialect they reconstruct they are also
reconstructing the birth of a "tribe".  But this is illogical for any
*linguistic* split used as criterial for sub-classiifying of West Germanic,
etc, since similar splits occur within "languages".  Liverpool English, for
example, is reproducing part of the High German shift, i.e., voiceless stop
-> affricate/fricative.)
 
Various "languages" evolved from either (and maybe in some areas BOTH) the
Low and High Germanic "dialects", e.g., English, Dutch from Low, German and
Swiss-German from High.  Their status as languages is social, and only
linguistic to the extent that the varieties united in one language have all
evolved from at some time-depth from a single major source.  Most often
there is a reference dialect for the "language", far from always associated
with a dialect cultivated for literacy in that "language".  But the rise of
literary dialects, call them "languages" if you want,  seems to be what
Roger fixes on.  Literary or not, it seems, however, that plenty of
diversification has taken place before reference dialects achieve the
status of "language", and it is purely social that the same label appplies
to most other dialects, and stops applying to them at an arbitrary point --
not the point of mutual unintelligibility, for example.  (Mutual
intelligibility is a continuum quite independent of the continuum between
dialect and language for any closely related "languages"; obviously it
applies quite well to languages which are only distantly related, or,
pardon the expression, not related at all.)
 
I invite Roger to clarify what his concern is in distinguishing "language"
diversification from dialect diversification.  Is it the emergence of
literary dialects, or what?  From an older exchange between him and Miguel.
I think one of his concerns was the influence that the reference dialect
has in controlling the direction of change of the other dialects socially
subordinated to it in some way.