Here's Givon text: Dan Everett on Piraha and Universals

[Suzanne Kemmer]

Dear Funknetters,

Funknet doesn't take email attachments (to avoid spreading viruses  
and for other technical reasons),

so I got the following text from Tom and will disseminate it for  
him.  I am leaving the page breaks in,

in case people want to paste it back to a Word document.

--Suzanne





RE:  DAN EVERETT ON PIRAHA AND UNIVERSALS


                                                            T. Givón

                                                       White Cloud Ranch

                                                          Ignacio,  
Colorado



When Dan Everett came out with his original  peer-review article  
("Cultural constraints on grammar and cognition in Piraha" , Current  
Anthropology 46.4, 2005), I  was disappointed with the peer  
reviewers' response (or lack thereof) to one major area where he  
stakes his claims, so-called 'recursivity'.  My disappointment  
centered  first on the fact that  Dan, while  going after Chomsky  
(cf. Hauser, Chomsky & Fitch 2002) with the relish of a recent  
convert, let Chomsky–once again, alas--dictate both the agenda and   
terms of engagement. But I also thought that few of the peer  
respondents did  justice to the issues of syntactic complexity   that  
were either raised or implied in Dan's work. My disappointment   
persisted through the eventual discussion earlier this year on  
Funknet.  So I thought it would perhaps be useful to  raise  the  
topic  once again, and if  nothing else  give both Dan and y'all a  
second go at it.



1. Chomsky's 'recursivity' as a framework:



It is indeed sad to see the Hauser et al. (2002) article being  
allowed to frame our discussion, indeed dictate its terms. That  
article was one more attempt, undisguised, to resurrect 'competence'  
as the bastion of pure innateness. This was presumably achieved by  
singling out one important property  of human grammar(s)--syntactic  
complexity achieved through the embedding of clauses inside other  
clauses--as  the  real  essence of human  grammar, and then claiming  
that this unique property is not the product of adaptive-selective  
evolution. There are a  zillion issues wrapped up in this tortuous   
intellectual  exercise. First, there are many other  major and minor   
functional and structural  properties of  human grammar  that  are  
just as species-unique--declarative speech acts, multi-propositional  
coherence, referential displacement,  multiple  grammatical   
morphologies and syntactic constructions (and their attendant  
semantic-pragmatic correlates). Why single out 'recursivity'?  Never  
mind that  recursive-hierarchic structure is actually the hallmark of  
complex, automated biological  processing in motor-control, memory,  
vision and other complex  behaviors, all predating homo sapiens by  
healthy margins.



Next, by lumping up all complex/embedded  constructions under the  
rubric of  'recursivity', one  tends  to obscures the fact, well- 
known from both child-language and diachronic-syntax research, that  
different complex  constructions  arise  in  different communicative  
contexts, under different adaptive  pressures, at starkly different  
ages in children (Diessel 2005),  and  through distinctly different   
diachronic pathways of  grammaticalization. By lumping them all  
together into 'recursivity', and by insulating  'recursivity'  from  
both its adaptive (communicative) motivation and its developmental  
(diachronic, ontogenetic) sources, one guts the issue of syntactic  
complexity. All that is left is the abstract  computational aspect  
(Simon 1962).


2/everett.07









Further, Chomsky's Generative  approach to universals has always been  
distinctly Platonic and patently non-biological. Only  properties of  
language that are attested in 100% of human languages are universal.  
A single 'exception' shoots down the rule.  But neither  biology nor  
behavior nor language are amenable to this pristine view of  
universals. They are  notoriously  subject   to variation and  
gradation, and those are in turn introduced and mediated by the  
developmental processes that shape both structure and behavior-- 
evolution, ontogeny, diachrony. Indeed, in order to achieve his  
pristine  universals, Chomsky has found in necessary to make  
universals more and more abstract and increasingly removed  from the  
visible phenomenology of language.



In one sense, however, Chomsky has had the right intuition about  
universals: They are not the mere list of visible properties manifest  
in extant (or extinct)  languages. Rather, they are principles that  
'govern'--or in our terms,  explain--the  visible  phenomenology. And  
they manifest their 'government'  through the developmental  
processes--acquisition and diachrony–in much the same way as   
universals of  biology  manifest  themselves through phylogeny  
(evolution) and ontogeny (embryology). In both biology and language,  
the reason why variation  and  'exceptions' are both possible and   
necessary is because the underlying adaptive principles are often in  
conflict ('conflicting motivation'), and pull structures  in  
different directions, often in a see-saw fashion. Thus, in phonology,  
articulation & perception are at constant loggerhead with each other.  
And likewise in grammaticalization, similarity to related source  
construction pulls in one direction (analogical extension and  
synonymy), while communicative distinctness pulls in the other  
(specialization and differentiation).  But Chomsky, much like  
Saussure,  was not about to subject his  principles and parameters to  
the messy vagaries of acquisition and diachrony. The entire  
Generative enterprise rests on this rejection of  the relevance of  
change and variation. This is where  'competence'  rides in to the  
rescue.



2. The diachrony of complex/embedded constructions



Embedded constructions in both the verb-phase and the noun-phrase   
fall under the unified intonation contour of  the host  main-clause.  
This is the most general structural feature of embedding.   
Invariably,  however, these syntactic constructions start their life  
as paratactic ones--two clauses under separate intonation contours  
but already performing the communicative function of embedded   
clauses. The  condensation--or merger--under a single intonation  
contour is the earliest structural indication of the   
grammaticalization of complex  clauses (Mithun 2006, 2007). The  
literature  on  this parataxis-to-syntaxis trend is massive, and some  
of it not all that recent (Givón 1971, 1979; Dahl 2004;  Heine and  
Kouteva 2007; Givón 2006, 2007; inter alia). It involves a two-step  
developmental  trend:

(a)   Parataxis to syntaxis

             (b) Syntaxis to morphology/lexis

Each of those steps contributes to the rise of complex-hierarchic  
structure--syntactic and lexical, respectively.


3/everett.07







Grammaticalization  is  multi-stepped and 'cyclical',  the latter  
meaning that it may move from the earliest  'source'  stage to a  
'mature' construction stage, and eventually  to phonologically- 
induced deterioration and back to zero. Most syntactic constructions  
and morphologies grammaticalize independently of each other. Only in  
post-pidgin contexts do constructions start together from zero. And  
such seeming synchronicity is temporary.  For each potentially- 
complex syntactic  construction  (REL-clause, V-complement, passive,  
cleft, WH-question, etc.), one can catch a language at any given  
point of the developmental cycle. Piraha is not exactly unique in  
showing  an early  paratactic  stage of REL-clauses and V-complement.  
Serial-verb languages all over Africa and Southeast Asia are in the  
same typological bag. Bambara (Mendeic) and  Supyire (Senufu) are  
there. In both, however, the earliest step of clausal complexity-- 
merger of intonation contours--is already discernable, more advanced  
in Bambara, less so in Supyire. The same is true of  Hittite. The  
same of Mohawk. The same in  scores of Southeast Asian languages. All  
you have to do is catch them at the early stage of the  
grammaticalization cycle. But cycles--and their stages-- come and go,  
often gradually, often piecemeal. Is it so unique to catch a language  
at a particular stage?



3. Does culture constrain grammaticalization?



If the diachronic facts indeed hold, as I think they do, then the  
question about the role of culture in constraining grammar must be  
recast as 'the role of culture in constraining grammaticalization'.  
Put another way: Is the fact that we find a language at an early  
stage of the rise--most often renovation--of complex  construction in  
any way correlated to culture?  The challenge  for Everett  here is  
two-fold.  First, of the languages that display,  roughly, the same  
early paratactic stage of the rise of complex clauses, some are small  
hunting-and-gathering societies-of-intimates (Piraha), some are old- 
establish  pre-industrial cultivators with much larger social units  
(Bambara, Supyire, Mohawk), some are  agrarian city-states or even  
empires (Hittite, Han Chinese). What exactly does culture predict here?



And second, grammatical constructions rise and fall. And their  
renovation seems to be motivated largely by communicative need. In  
most language where we have historical or reconstructive evidence,  
one could show two or even three generations of rise-and-fall of the  
same construction. And  most often  no cultural change is correlated  
with such diachronic cycles. Did German revert to a society-of- 
intimates ca. 300-400  years ago when it was renovating its REL- 
clause construction, reverting to parataxis?  And did it then sprint  
back to the industrial revolution when it eventually proceeded to  
well-grammaticalized syntaxis (merged intonation contours, de- 
stressed REL-pronouns)?  Did the ascendant Han empire change from an  
intimate hunting-and-gathering society to a complex  society-of- 
strangers  as it created, one piece at a time, the complex syntactic  
construction of  Mandarin Chinese--in every case starting  from  
parataxis  of clause-chaining? And did Han culture collapse earlier  
on, back to an intimate small society of   hunters-and- gatherers,   
when it expanded imperially, moved east and south, and took over the  
vast Austro-Asiatic


4/everett.07







substratum--and in the process  'regressed'  from the highly-complex  
Bodic-Tibetan Syntax to the near-pidgin parataxis  of the Tao The  
Ching?  Did the Hittite empire change culturally between the  
paratactic REL-clause of Old Hittite and the embedded  REL-clauses of  
Middle Hittite?

If one advocates cultural constraints on grammaticalization, in this  
case on the presence or absence of complex/embedded, one  needs  to  
demonstrate some, hopefully consistent, correlations between cultural  
and linguistic traits.





References



Dahl, O. (2004)  The Growth  and  Maintenance of  Linguistic  
Complexity, Amsterdam: J. Benjamins

Diessel, H. (2005)  The Acquisition  of  Complex  Sentences,  
Cambridge: Cambridge University Press

Everett, D.  (2005) "Cultural constraints on grammar and cognition in  
Piraha", Current Anthropology, 46.4

Givón, T. (1971) "Historical  syntax and  synchronic morphology: An  
archaeologist's field  trip", CLS #7, Chicago: Chicago Linguistics  
Society

Givón, T. (1979) On Understanding Grammar, NY: Academic Press

Givón, T. (2006) "Multiple  routes to clause union: On the diachrony  
of syntactic complexity", Seminario de Complexidad Sitáctica,  
Universidad de Sonora, Hermosillo, November 2006 (ms)

Givón, T. (2007) "Toward a diachronic typology of relative clauses",  
Symposium on the Genesis of Syntactic Complexity, Rice University,  
Houston,  March 2008 (ms)

Hauser, M., N. Chomsky and T. Fitch (2002) "The faculty of language:  
What it is, what it has, and how it evolved", Science, 298

Heine, B. and T. Kouteva (2007) The Genesis of Grammar, Oxford:  
Oxford University Press

Mithun, M. (2006) "Structural  parameters  of clause integration:  
Complementation in Mohawk", Seminario de Complejidad Sintáctica,  
Universidad de Sonora, Hermosillo (ms)

Mithun, M. (2007) "Threads in the tapestry of syntax: Complementation  
and Mohawk", UC Santa Barbara (ms)

Simon, H. (1962) "The architecture of complexity", Proc. Amer.  
Philos. Society, 106.6