I recently obtained a copy of Daniel Everett’s pop-science paperback Language: The Cultural Tool (2012) from the Brooklyn Public Library. The chunky fonts of the cover made me think I was about to enter the world of a staunch iconoclast. But what I actually found was a laundry list of what you might call “grievance studies”—if that didn’t already mean something else—against a broadly generativist conception of language.
Everett, once a specialist in languages of the Amazon, does not draw so much from niche fieldwork so much as splashy papers by non-linguists in high-impact pop-science journals like Nature and Science. Thanks to my colleague Richard Sproat, I have seen how those august organizations make their sausage: they either don’t let linguists referee, or if they do, they simply ignore their negative reviews. (Everett, as it happens, has glowing things to say about the latter paper even though it has nothing particular to do with his titular thesis.) In general, the works cited draw from disparate areas that have received relatively little attention from specialists, so while Everett is a decent prose stylist,1 he is tilting at windmills for much of the book.
Everett often substitutes appeals to authority to actual arguments. For instance:2
Michael Tomasello, the Director of Psycholinguistics at the Max Planck Institute for Evolutionary Anthropology in Leipzig, says exactly this. A world leader in the study of cognitive development in canines and primates, including humans, he says simply ‘Universal grammar is dead.’ It was a good idea. It didn’t pan out. (p. 192)
That’s all we get on that point.
The other thing I was struck with were elementary factual errors that would have been cleaned up had literally any other linguist read the book before it went to press. Early on, Everett is discussing definitions of language. After describing the proposed definitions by Sweet and by Bloch and Trager, he quotes (p. 32) a passage from Noam Chomsky (the reference is neither given nor known to me):
A formal language is a (usually infinite) set of sequences of symbols (such sequences are “strings”) constructed by applying production rules to another sequence of symbols which initially contains just the start symbol.
Now, obviously this is not a definition of language as we understand it but rather the start of a definition of the mathematical construct formal language, a notion which predates Chomsky by at least half a century. Everett is either deeply confused or is deliberately misleading his readers.3 The second howler I found is the following passage, now from Everett:
The late Professor George Zipf of Howard University formulated an explanation of the relative lengths of words that has come to be known as ‘Zipf’s Law.’ His law predicts that more frequent words will be shorter than less frequent words. (p. 106)
George Kingsley Zipf taught at Harvard University, not Howard University, and that’s not what Zipf’s Law denotes.4
There are several factual errors. For instance, we’re told that ejectives are not found in European languages, which is only true if we don’t consider Armenian, Georgian, etc. languages of Europe (p. 177). And Xhosa is described as a Khoisan language when in fact it’s Bantu (p. 178).
And there’s the casually racist, classist, and sexist stuff. For instance, Everett posits that Pirahã children lack a theory of mind:
…many Pirahãs used to stare at me (some children still do) and talk about me in front of me—they didn’t believe I had a mind! (p. 165)
Okay. But maybe they were surprised rather than mentally deficient.
Later, Everett tells us:
…for many Ohio factory workers being overweight is less of a moral problem and more of a health problem—they do not value being at the right weight all that highly. (p. 300)
Okay. But the factories pretty much all closed down in Ohio years ago.
We’re told that in Wari‘, a language of the Amazon, the word for ‘wife’, manaxi’, means literally ‘our hole’ or ‘our vagina’. Everett suggests that “some outsiders”—let’s call them “the libs”—might “jump to the facile conclusion that this is a crude and demeaning comparison”. What’s the right analysis, though?
Perhaps to the Wari’ reproduction and the family are such important values that they honor the wife and the vagina as the source of life. So it is the highest form of flattery to call the wife ‘our vagina’, the source of life. Is this a possibfle conclusion? Yes. Is it the right one? I don’t know. No one can known unless they undertake a systematic analysis of Wari’ culture… (p. 195)
Okay. But maybe Everett could have just asked his coauthor Barbara Kern, an anthropologist who lived among the Wari’ for over forty years and who speaks their language fluently.
Finally, we’re told Banawá, another language of the Amazon, uses feminine as the default gender. Everett then proceeds to describe what I would call a (from a non-relativist perspective) brutal and essentializing coming-of-age ritual for pubescent Banawá girls. Are these facts related?
It is exactly by exploring such cultural values that we would try to build a connection between feminine identity and grammar in Banawá and other Arawan languages. I have not yet established such a link, but I am working on this. (p. 210).
Okay.
Footnotes
- Despite his affectation for cheery-dreary Boomer cultural touchstones, that is. In the first few chapters he mentions “Under The Boardwalk”, the music of Cream, the plot of an episode of The Andy Griffith Show, and the murder trial of Phil Spector. Sorry, but I already have a Dad.
- For the record, this also gets Tomasello’s title wrong: he was “Co-director” of the Institute, not “the Director of Psycholinguistics”.
- As a colleague pointed out, Everett himself is a coauthor on a paper (Futrell et al. 2016) that claims that Pirahã, an Amazonian language, can be described by a regular language. This suggests that Everett understands the distinction between human languages, of which Pirahã is an instantiation, and formal languages, of which the regular languages are an instantiation, and is simply being disingenuous here. For what it’s worth, the argument in that paper is incoherent. The authors simply observe that their corpus can be described by a regular language, but so can any finite sample. This is a vacuous observation. That said the study is not totally without value: the appendix contains an annotated corpus of Pirahã sentences.
- Zipf does observe something of the sort in his 1935 book The Psycho-biology of Language (p. 28f.), but “Zipf’s law” does not refer to word length at all.