Monogenesis and polygenesis

[bwald]

----------------------------Original message----------------------------
Some replies to responses on my suggestions about polygenesis.
 
In considering my point about what we don't know, Scott DeLancey writes:
 
>...The basic point is that a functioning language--NOT the appropriate
>neural structures, a fully-formed set of parameters, or whatever
>mumbo-jumbo you might want to think of as the necessary biological
>prerequisites for language, but functioning language--would give the
>first population to develop it a significant selectional advantage
>over all others.
 
I don't disagree with the argument that a population having a (functioning)
language has an advantage over a population that doesn't.  The point
remains: how many separate populations developed a functioning arbitrary
lexicon independently of each other?  Once the ability is there, the
adequate arbitrary lexicon might have developed relatively quickly, but not
as quickly as the prerequisite gene spread.  If gene spread was earlier, it
follows that the arbitrary lexicon may have developed more or less at the
same time in independent communities which did not have knowledge of (or
non-sexual interest in?) each other.  If further facts are adduced to the
effect that such a possibility is unlikely, then we are left with
monogenesis as more likely.  However, Scott's suggestion does not offer
such facts.  It only offers a reasonable argument about what happens when a
population with a functioning language actually meets a population,
regardless of ability to acquire language, which has not yet developed a
functioning language.
 
He ends by reiterating:
 
>Thus, in our imagined pre-linguistic situation, once one group has
>developed language, other groups will not have the chance--they will
>either adopt what the first folks off the block have created, or
>they will lose out.
 
and, then he rereiterates,
 
>Once they've developed what they need to exploit their
>language potential, everyone else will join them or be left behind.
 
So I reiterate: OK for "one" group.  But what's to insure that they meet
"other" groups before one or some of them have ALSO developed language.
This is what polygenesis implies.  This really depends on how likely it is
for different populations with capacity for developing an adequate
arbitrary vocabulary, to develop it separately.
 
(In fact, to some extent a somewhat arbitrary or "displaced" set of symbols
may be a more widespread and older property of complex mammals than the
"language faculty", such that such symbols become immediate material toward
developing an arbitrary lexicon once the "language faculty" exists.)
 
  The best clues to an answer are probably available in the independent
in-group development of various sign languages in historic deaf communities
isolated from each other, e.g., Kenyan sign language, Antiguan sign
language, etc.  From what I understand, some deaf groups did develop such
languages without (much?) input from the hearing.  It's a very imperfect
clue because fully functioning languages already existed around them, and
it is difficult to exclude the influence of their speakers on stimulating
isolated deaf speakers or groups to get the idea of using gestures to
develop LEXICON.  BTW many such gestures indicate somewhat NON-arbitrary
sign-meaning correspondences -- but there is still arbitrariness once the
signs become conventionalized, so that other non-arbitrary possibilities
for the same lexical item become excluded, e.g., the sign miming bonnet
ribbons which originally indicated "woman" in the 19th c French sign
language from which AMESLAN developed (if I remember correctly).
 
Next, Adrienne Bruyn writes:
 
"Working in pidgin/creole studies myself, I don't think "polygenesis"  with
regard to pidgins/creoles (PCs) usually refers to their "mixed descent",
even though it can be used in this sense when the problem of
classification is at issue. This, at least on the level of e.g.
 "Is Sranan a Germanic language?"  is something creolists seem to be less
concerned with than historical linguists taking PCs into account (e.g.
Posner) (I'm generalizing)."
 
Adrienne is right.  Originally in P/C studies polygenesis opposed the
"monogenetic" theory that ALL  P/Cs descended from a single historical
pidgin, which some adherents identified with "Lingua Franca", the first
millenium (AD) Mediterranean trade language between the Muslim bloc and the
Christian bloc -- with the possibility that Lingua Franca itself might have
its origins in an older pidgin.  The implication was that all pidgins owe
their origin to a one-shot invention, cf. the monogenetic theory of the
origin of the set of ALL known languages.  The polygeneticists opposed this
idea with the notion that pidgin genesis involves a more general contact
phenomenon which could and no doubt has applied independently at various
times in certain kinds of contact situations.  It was actually that notion
that I was appealing to in raising the issue of P/C polygenesis.  As for
her last observation, this is what early P/Cists, e.g.,  Hugo Schuchardt,
were appealing to in criticizing the tree notion of language
diversification as the ONLY possibility for language diversification and
hence "genetic" classification.
 
At this point it is irrelevant that Schuchardt's citicism was premature.
Most historical linguists were not about to acknowledge Schuchardt's
insight, while the tree method had still not been exhausted.  And we're
still arguing about how far the method can be pushed -- but we have pushed
it beyond current consensus, cf. Altaic -- why is this such a problem after
all this time? I don't see how it can simply be "bad" scholarship somehow
bungling "tried and true" methods.  I recognize that there are more
languages than there are competent historical linguists competent in all
these languages, but, still, what the heck is the problem?
 
Adrienne goes on to discuss how the P/C monogenesis and polygenesis
theories got sorted out.  Monogenesis is about reconstructable historical
facts, both language and social.  Polygenesis is about more general
principles of language contact, mixture and the like.  She concludes:
 
"In this sense then "polygenesis" appears similar to the way it is used in
the discussion on the origins of language -- and also raises questions and
debate. "
 
Again she is right.  The principle stands that NOT ALL current languages
can be traced back to a single ancestor in any sense that has relevance to
the present discussion.  This is already generally accepted.  Historical
linguists do not contest this principle; instead they ignore it in their
quest for the proto-language -- as if it were irrelevant to the resources
they have for succeeding in their quest.  And so it gets lost in the debate
about monogenesis or polygenesis of an "original" language.  My suggestion
was simply to extend the polygenetic argument to more distant ancestors of
current groups of languages, esp with an eye to groupings for which more
polemic than progress is encountered.
 
(Indeed, in their zeal there was a brief period when some P/Cists were
speculating about ALL problematic groupings and various other languages,
e.g., English, being ultimately P/C in origin, as noted say in the Kaufman
& Thomason book -- but that's another matter, stemming from a time when
P/Cists were still getting a grip on what their own primary subject matter
and principles were.)
 
With regard to Sranan, if knowledge of all "other" Germanic languages were
to disappear, historical linguists would have little trouble recognizing
most of the lexical material of Sranan as Indo-European, and would even be
able to discover Grimm's law in the Sranan data.  They might be puzzled by
the lesser irregular correspondences which represent the difference between
the English vocabulary (major) and the Dutch vocabulary (minor), but they
would surely classify the language as Indo-European, of the Srananic (=
Germanic) branch.
 
Now, if all knowledge of Germanic and Romance were to disappear, the
creoles Saramaccan and Djuka would present much greater problems to
reconstruction, because their vocabulary is almost equally English (<
Germanic) and Portuguese (< Romance).  The brighter reconstructionists
would eventually propose that two distinct Indo-European languages were
responsible for this situation, but for some words they would not be able
to decide which of the languages was responsible for which words -- or
indeed if either of the languages was responsible for the words without
wider Indo-European cognates.  And in a few instances they'd be right,
because the words are not Indo-European, but probably in even more
instances they would be wrong, assigning a non-Indo-European origin to a
Germanic or Romance word that doesn't have (surviving) cognates elsewhere
in Indo-European.  So, by the way linguists use arbitrary lexicon to
classify linguistic origin, Sranan is statistically Germanic -- it is even
English (or "Angloid", if you care), but Saramaccan and Djuka are not; they
are Germanic-Romance -- but there is no node in the Indo-European tree
which separates Germanic-Romance from other branches.  That would surely
confound historical linguists, who BTW are not even sure whether Saramaccan
"liwa", or something like that, comes from English "river" (ironically from
Latin ripa etc.) or Port/Spanish "rio" -- or both.