August 26, 2004

The Straight Ones: Dan Everett on the Pirahã

I was co-editor of the volume in which the first full description of the Pirahã language appeared (Desmond C. Derbyshire and Geoffrey K. Pullum, eds., Handbook of Amazonian Languages, Volume 1, Mouton de Gruyter, Berlin, 1986). Dan Everett's 200-page chapter on Pirahã is a highlight of the volume. (The Wikipedia article on the language is currently an unedited mess from which you can't even figure out the phoneme list, so I won't link to it; don't go there.) Dan is now a distinguished specialist on Amazonian languages and professor of phonetics and phonology at the fine Department of Linguistics at the University of Manchester in England. I know him well and respect him greatly. And I thought he might like to respond to some recent suggestions to the effect that Pirahã is just too strange to be true. So in this long post I include a statement that he supplied at my invitation.

It was Peter Gordon's recent publication in Science ( here, if you subscribe) that led to the long overdue recent discussion of Pirahã language and culture in various forums. The focus has mainly been on their innumeracy and its possible linguistic roots. Now, I should point out that I actually believe Dan may be overstating things in saying (as he does in a recent paper) that they are the only human group ever to have been found to have no numeral system; my understanding is that many Australian aboriginal languages, Warlpiri being one example, have no native number vocabulary; the speakers can and do learn to count, and simply borrow the number system of English in order to do so. That alone suggests there is little plausibility to the totally Whorfian spin that The Economist puts on Pirahã innumeracy ("At least in the field of maths, it seems, Whorf was right"). In what I've learned, I see no support for the usual vulgar Whorfian claim that everyone seems so besotted with (that your language determines the thoughts you can have); Mark Liberman makes the counterargument beautifully with his imaginary no-throwing culture, and as you'll see below, Dan agrees with Mark.

But Whorfian spin aside, the Pirahã language itself has many fascinatingly unusual features. We owe most of what we know about it to Dan and Karen Everett. In an email to Mark Liberman (quoted here), the entangledbank author appeared, in a damning-with-faint-praise sort of way, to be really skeptical about the Dan Everett's work on Pirahã: "None of it reads as obviously loony, but I have to wonder whether he's some Borgesian fantasist, or some Margaret Mead being stitched up by the locals." These suggestions — that either Dan (and Karen too?) made up aspects of Pirahã, or the tribe colluded to pull (without having any knowledge of linguistics) an intricate linguistic confidence trick on two skilled linguists, sustained over a quarter of a century and never revealed despite visits by a distinguished phonetician like Peter Ladefoged and a fine psycholinguist like Peter Gordon — are pretty ludicrous when you think about it. But it's logically possible, I suppose. There have of course been anthropological hoaxes in the past; think of the Tasaday, thought to be an isolated group of stone-age survivals until their language was shown to share 85% of its vocabulary with Cotabato Manobo and its members were discovered to have been manipulated into play-acting by a Filipino official.

So let's have Dan Everett speak for himself. Here's what he emailed to me about whether he was a hoax victim.

I started working with the Pirahãs in 1977, not knowing what I was in for. (By the way, they do not call themselves Pirahã, which is not even a word of their language. They refer to themselves as the Hi'aiti'ihi'. The ' indicates high tone on preceding vowel, no mark indicates low tone. Literally the name has four parts, glossed as 'his+bone+straight+Nominalizer'. A rough English translation is 'the straight one[s]'.) I have lived in all their villages for an aggregate time now approaching 7 years. For the past 27 years I did think I was a 'Borgesian fantasist'. I worried about this when I first published on their unusual stress system in Linguistic Inquiry in 1984, a squib which resulted in letters from well-known phonologists to me, a new PhD, to the effect that I was likely incompetent and telling me what stress meant. Years later, Peter Ladefoged came to work with me on the Pirahã language (and three others). When I met him at the airport in Porto Velho, he immediately said "Hello. By the way, I am very skeptical about your stress claims." Peter Ladefoged did in effect what Peter Gordon did. He heard about some weird claims and went to check them out. Both of them came to agree with me (more or less).

I have long hesitated to write this stuff up because even to me it sounded so weird. I am aware of the criticisms Pinker has, among others, leveled at Whorf for his fanciful sounding glosses, etc. But as I wrote (what people should indeed read if they are interested in this is the article cited by Mark Liberman and available on my website, on Pirahã culture and grammar, not just the number abstract, which I will be taking off the website), I realized that, yes, this is my best understanding of all of this stuff. I sent the article for comments to Steve Sheldon and Arlo Heinrichs, two SIL members (I am no longer a member of SIL) who each lived for many years among the Pirahãs and who speak/spoke the language fairly well. Both of them agreed with the account in that paper.

I disagree, as Mark points out, with Gordon's conclusions. I don't think that this case requires an appeal to Whorf. In fact, as I try to argue in the larger paper, just the opposite seems to be the case (Mark summarizes this all very well, no doubt better than I am doing here).

A well-known MIT linguist with an interest in morphology distributed throughout the grammar asked me why, if I am correct, Pirahã would be the only case known like this. Another linguist from the Northwestern US who has held high office in the LSA asked me if I thought the Pirahãs were genetically different.

To answer the first question, I think that it is hard to hazard analyses that go so strongly against the grain. It took me 27 years to work up the courage to say these things and I am still called a 'Borgesian fantasist' (and have been called much simpler things, like 'stupid'). There just aren't that many linguists with that kind of time on a language so isolated from Western civilization. Therefore, I am not surprised that there are so few claims. I do believe, however, that many analyses of number and grammar in the literature, on similarly primitive societies (in some technological sense of primitive) are likely 'overanalyzed', e.g. that there are likely to be other languages without embedding, where juxtaposition has been taken to be embedding without much thought given to the matter.

To answer the second question, Pirahã women occasionally have children with Brazilian traders passing through, children raised as Pirahãs. These children don't show any difference I can see from other Pirahãs on these cognitive skills or language facts. I don't think genes, retardation, or other such suggestions are useful or appropriate here.

My own view then is that the case of Pirahã illustrates, perhaps as well as any example ever discussed in the literature, the kind of bi-directional causal relationship between language and culture that Boas and Sapir would have expected us to find.

There is a problem for universal grammar in all this, though. That is the non-trivial one of setting the boundary between culture, grammar, and cognition in light of examples like this where previous boundary lines have been shown to be potentially illusory.

I just left the Pirahãs a few days ago. They are oblivious to all of this attention, yet doing well as a people. However, I have heard the very disturbing news that an electric power company is thinking of using their river, the Maici, to generate power in some way. If any outside company enters their reserve (which I helped demarcate, with support from Cultural Survival, 20 years ago), this could be the end of the Pirahã people. So I hope that this attention on them right now can be used to generate some support for their survival. Examples like Pirahã illustrate very clearly the loss inherent to knowledge of our species, if such a language were to cease to exist without having been studied. It also shows, I hope, that some studies take a LONG time, perhaps the length of an entire career.

-- Dan

Daniel L. Everett
Professor of Phonetics and Phonology
Postgraduate Programme Director
Department of Linguistics and English Language
University of Manchester
Manchester M13 9PL UK
Fax: 44 161 275 3187
Phone: 44 161 275 3158

Here's a map of where the Pirahã live. And for some pictures of Pirahã efforts at drawing and writing (they really do seem to be utter beginners at both), take a look at this drawing of a cat and this drawing of a tapir (with a few numerals added; they seem to be for decoration). Dan has also supplied this picture of two Pirahã women busily engaged in learning how to use pencil and paper.

[Update 4/9/2007 -- links to other Language Log posts on the Pirahã and related topics can be found here.]

Posted by Geoffrey K. Pullum at August 26, 2004 01:48 PM