Language splits and bundled isoglosses

[bwald]

----------------------------Original message----------------------------
Roger Wright responded to my invitation to clarify what he meant by the
difference between "language diversification" amd "dialect diversification"
as follows:
 
> if the speakers involved are
>geographically contiguous, we could use the following criterion: cognate
>*languages* will have many isoglosses bundled along the same line (which
>might coincide with a political boundary, or might not); dialect
>diversification *within the same language* is likely to involve a
>dialect continuum, with many isoglosses, but which rarely coincide
>geographically in this way. (The emergence of Literary standards can
>lead to bundling, rather than being caused by them.)
>        This is a suggestion, not a dogmatic view.
 
I appreciate the tentativeness with which Roger makes the suggestion.  I
would separate two points in the suggestion.  (1)  I would maintain that
the distinction between "language dv" and "dialect dv" as a label for
distinguishing bundling of isoglosses vs. areal ordering of isoglosses is
potentially misleading, since "language" is a label applying to the social
construal of a set of dialects (meaning that the set has a common
historical origin). Thus, there will be a confusing mismatch between
"language" as bundling of isoglosses amd as a socially determined areal
domain where the sociopolitical boundary does NOT coincide with bundling.
(2)  The emergence of literary standards, or conceivably some other kind of
central influence, can lead to bundling.  That is an interesting idea, and
an empirically testable claim with regard to coincidence with a
sociopolitical unit.  I would assume that whether or not there is such
coincidence would depend on the relative strength of the sociopolitical
boundary vis-a-vis  cross-boundary influences, e.g., whether speakers at
the boundary are more influenced by dialects of "another language"  across
the boundary or by the "central / prestige" dialect possibly at a greater
distance but within the sociopolitical boundary.  Either case could be
characterised as a case of "diglossia", "bilingualism" in the case of cross
boundary influence but "bidialectalism" in case of influence from a
prestige dialect of the same language.  NB the "bilingualism" in some cases
could require less difference between the varieties than "bidialectalism".
 
No doubt in most, if not all, cases there is some of both kinds of
influence on border dialects.  It remains to empirical investigation to see
how such border dialects actually behave.  One problem is the selection of
features to observe bundling of isoglosses, since there are potentially one
bundle which aligns the focus dialect with a neighboring dialect of the
language across the sociopolitical boundary and an opposing bundle
distinguishing these two and aligning the focus dialect with the prestige
dialect in the same language.  Offhand, I am not aware of intersecting
bundles, only of the possibility that a neighboring dialect from the other
language may penetrate more "deeply" into the focus dialect, a change from
"below" (or the "side"), while the prestige dialect may remain more or less
distinct in its use by speakers of the focus dialect, but eventually filter
down and have diachronic consequences as an example of change from "above".
Phonological and syntactic change may differ in that syntactic change may
be more susceptible to prestige influence and phonological change to
neighboring influence.  In any case, diachronic interpretation of a bundle
depends crucially on the possibility of distinguishing fragmentation of a
single dialect over space, such that the bundle represents conservative
features of the original single dialect, from spread of a bundle of
features (over time) so that they are innovative in the receiving dialects,
regardless of their status in the source dialect, i.e., in distinguishing
divergence (cf. tree diversification) from convergence (cf. "Sprachbunde")
in such an area.  That is not always easy, and may even be the wrong way to
approach interpretation of the bundle in some cases.
 
If the discussion is to be continued, I invite concrete examples so that
the methodology and decision mechanism for historical interpretation of
bundling can be examined.  Maybe included for Romance could be such
syntactic phenomena as object clitic ordering, among other things.  How
does the boundary between, say, Provencal and Catalan, or Provencal and
"French", display bundling for a number of features (maybe under the
influence of prestige varieties of French and Spanish among bilinguals;
aren't Catalan and Provencal speakers mostly bilingual in Castillian and
"general colloquial" -- if not standard -- French, respectively?)
 
P.S.  I actually think that the best place to look for bundling which tests
adjacency vs. prestigious centers is in the lexicon, since lexical items
approximate the "linguistic" arbitrariness of separating one language from
a closely related one.  But even here I guess that, say, if there is
cross-political boundary agrarianism, then there could be a bundle of
agrarian terms which crosses the language boundary, in contrast to a bundle
of political terms emanating from the prestige center of the same language
that stops at the language boundary.  It remains to be seen how effective
"languages" are at drawing a bundle of isoglosses aorund themselves to the
exclusion of their historically related neighbors.  To be anticipated is
indeterminacy about how to treat border dialects of the low prestige
variety where there is popular sentiment to the effect that it reflects a
"mixed" language or the "other" language has "contaminated" the first.
Then we get into the whole issue of folk-beliefs about language and the
difference between people's idealisations of their "language" and what is
actually happening in the language/s they speak.