On Everett & Piraha: "history holds the key"

[Daniel L. Everett]

On Apr 23, 2007, at 3:21 PM, Jagdish Jain wrote:

> Hi Funknet members,
>
> I have read the buzz created by Dan Everett's claim that PirahaN is  
> an exceptional language. He also claims that it poses a challenge  
> to Comsky's linguistics. I have a few observations to make.
>
> 1. I assume that PirahaN people are cognitively modern human  
> beings, not chimps or bonobos or rhesus macaques. Their brains (as  
> physical organs) are like ours. They are genetically endowed with  
> modern human capabilities --- cognitive recursiveness, metaphoric  
> mappings, metonymic mappings, etc. Their expression for a foreign  
> language, according to Everett, is "crooked head." A wonderful  
> expression! The "head" (rather than "tongue") stands for language  
> through metonymy, and "crooked" is metaphorically mapped to "bad."  
> This is very similar to the 19th century British imperialists'  
> attitude to the languages of India - they were vulgar and inferior  
> to English. The PirahaN people are good imperialists!
>

We all are. Yes.


> 2. Chomsky has identified the following two traits of language  
> design as very important:
>
> (i) Discrete infinity: We can use a small number of discrete  
> elements (e.g. 8 consonants, 3 vowels, a few tones as in PirahaN to  
> generate an infinite number of utterances. Dan Everett has given no  
> evidence to challenge Chomsky on this point. Nor has anybody else.  
> This is now a noncontroversial point.
>

This is not a Chomskyan principle. Just a fact about combinatory  
principles that has been around forever. Languages are not infinite  
though, not in practice, so this is to some degree a metaphor. But  
these issues will be discussed at the Recursion Conference this week  
here at ISU.

> (ii) Recursiveness: This trait is AVAILABLE to all languages. If a  
> language does not exploit this trait in one linguistic construction  
> (e.g. a clause-within-a-clause construction), it may do so in some  
> other construction (e.g. a NP within a NP, as in "my brother's  
> son's wife's sister"). It is possible that PirahaN does not use  
> clause embedding as exemplified by the English sentence, "I know  
> (that) he lied." They might say, " I know (it). He lied."  In Hindi  
> we cannot embed a small clause as we can in English, " He kicked  
> the door open." In Hindi we have to say, "He kicked the door. The  
> door opened." We need to examine other constructions where PirahaN  
> may use recursion. If we do not find any recursion in any  
> construction, the only thing we can say that PirahaN has not  
> exploited this feature of language design. It would be a surprising  
> fact but it will not disprove the Chomskyan hypothesis that this  
> trait is AVAILABLE to all languages.
>

This says nothing. Facial recognition is available to all languages  
too. The fact that something is available to languages could either  
follow because it is part of Universal Grammar or because it is part  
of general human cognitive abilities. That is the question. The lack  
of recursion in Piraha syntax alongside the clear evidence for  
recursion in Piraha thought and discourse interpretation and  
compositional semantics indicates that it is very strange, at the  
least, to call recursion a fact about grammar or the Faculty of  
Language, whether FLB or FLN. In fact, Herbert Simon noted years ago  
(1962) that recursion characterizes all information processing  
systems, human or not. This cannot be so easily linked to UG, etc.  
And if I am correct that recursion is absent in Piraha (experiments  
are on-going) then the simplest hypothesis is that recursion is a  
fact about brains and not about language. It adds nothing to any  
debate to say that it is 'available'. Available where, how, and why?  
In the brain generally due to greater intelligence or in a  
specialized language compartment, language organ, etc? The evidence  
suggests the former over the latter. And if I am also correct (maybe  
not, I grant) that recursion is absent for cultural reasons, then  
this is culture affecting core grammar in ways that are very  
difficult to reconcile with the view of recursion as part of the  
biology. Culture doesn't affect whether hair grows, for example, only  
how that growth is managed.


> 3. Dan Everett confuses "language" with "communication." Language,  
> especially syntax, is a cognitive object involving computation  
> (merging, adjoining, moving,etc.) with word-sized units. cultural  
> meanings enter the language through its lexicon, metaphors,  
> metonymies, conceptual blends, etc., NOT through syntax.  
> Communication involves exchange of ideas, emotional states, etc.  
> between two parties; it can be done without language, as it seems  
> that the PirahaN people communicate with each other by prosodic  
> means only ( humming without using any vowels and consonants of  
> their language, using nasal whines, popping or flipping their  
> lips,etc.)


The syntax is in fact the claim I have made - the culture affecting  
the syntax. We all know it can affect the language.

>
> I am afraid I do not understand this excitement about Dan Everett's  
> "exceptional findings" about PirahaN.


I do not either, since I believe that on closer examination many  
languages will be found that show similar characteristics. The  
excitement is not about whether Piraha is exceptional, but whether it  
and many other languages show that Universal Grammar is an  
unnecessarily baroque and empirically inadequate hypothesis.

Dan