I recently wrote on Lingua Franca about my astonishment over Piotr Cichocki and Marcin Kilarski. In their paper "On 'Eskimo Words for Snow': The Life Cycle of a Linguistic Misconception" (Historiographia Linguistica 37, 2010, Pages 341-377), they mistook my 1989 humorous opinion column "The Great Eskimo Vocabulary Hoax" for a research paper, and bitterly attacked it for dogmatism, superficiality, offensiveness, and all sorts of scholarly sins. But there is an additional thing about the paper that puzzled me deeply. It concerns the word "misconception" in the title.
I have read the early sections of the paper over and over again trying to figure out what Cichocki and Kilarski think the misconception is, and I just cannot figure it out.
In their opening sentence they say they will "examine the well-known example of the Eskimo words for snow as a case study on the complex interdependence of linguistics and the discourses of social sciences and philosophy." Then they make some fairly grandiose claims about demonstrating "the influence of social and philosophical theories on linguistic analysis, and the subsequent reflexive impact of linguistic evidence on social sciences and philosophy." (The paper contains nothing that could possibly be thus described, as far as I can see. The excellent work linguists have done on Eskimoan phonology, morphology, syntax, semantics, and lexicon has been no more colored by social and philosophical theorizing than most linguistics; and it is ludicrous to portray the vapid allusions to snow words that are found in other disciplines as influenced by serious linguistic evidence.)
Finally their third sentence mentions "the exposition of the misconception by Martin (1986)"; but what is this misconception? Who misconceived what, exactly, and when? It never gets any clearer.
Cichocki and Kilarski claim they are going to "examine the shortcomings of the exposition of the misconception by Martin, Pullum, and Pinker, focusing on the contradictory and erroneous interpretations, a frequently offensive style of argumentation, and lack of attention to formal detail." There it is again: "the misconception". Clearly a misconception is bad; everyone is against misconceptions. But Cichocki and Kilarski never tell us anything about this one.
Do they mean that the great anthropologist Franz Boas had misconceptions about Eskimoan languages? Certainly not; he traveled for months with an Inuktitut band in Canada and learned something about their language.
Do they mean that Benjamin Lee Whorf had misconceptions about what Boas had said or done? No, Cichocki and Kilarski are firm defenders of Whorf.
Do they mean that the general public developed a misconception about what Whorf had said? That was basically what Laura Martin was saying, and it was the theme of my essay publicizing her work; yet Cichocki and Kilarski seem to be annoyed with both Martin and me. In fact they seem angry that they have to breathe the same air as us.
I find no clues in Cichocki and Kilarski's paper as to what they mean by "the misconception" about Eskimoan snow vocabulary. But if you asked me about misconceptions, I'd say there are lots of them: great big confused clusters of them. People seem to think
- that speakers of Eskimoan languages see far more snow varieties than those of us who are hampered by a Germanic language background, or
- that they fail to perceive the unity of snow that we perceive, or
- that because their language illuminates new conceptual distinctions they experience reality in a different way, or
- that none of us can perceive what they perceive without knowing an Eskimoan language, or
- that because they deal so much with snow they naturally had to develop huge numbers of words for it, or
- that because they have so many words for it they naturally have greater expertise with it…
All sorts of off-the-wall hunches about language and thought seem to lurk in the weird things article-writers and lecturers say about why we should pay attention to the prolific character of snow vocabulary among the nomads of the Arctic. Most of them seem amply silly enough to justify the ridicule that I pour on them, and that Steven Pinker pours on them in his skeptical section on Whorfian determinism in The Stuff of Thought (Penguin, 2007, pp. 125ff).
What Laura Martin drew attention to in her 1982 conference paper and her 1986 research report in American Anthropologist was a widespread but apparently unfounded belief that (i) the set of distinct roots in Eskimoan languages that denote snow types is relatively large, and that (ii) its size is a notable fact, having spinoff importance for anthropology, psychology, sociology, whatever. (I say "whatever" because I heard the factoid asserted by two different lecturers in a five-day University of California academic management course. I have no idea what the significance was supposed to be in a management context.)
The sense in which the belief is unfounded is that no one cites a source for the specific number they give: they just give a number, apparently picking one at random, and assume no one will contradict them. They are apparently making stuff up. That tendency to stray beyond mere credulity into mendacity is what gave me the sense that they were a legitimate target for satire and ridicule in my 1989 essay.
Cichocki and Kilarski show some signs of thinking that in fact there is a certain amount of empirical support for the claim that the size of the set of snow-denoting roots in Eskimoan languages is significantly greater than that of (for example) English. They reproduce some lists taken from 2010 work by Igor Krupnik and colleagues (of which more some other time), as if they thought this would prove their point.
But although they have the nerve to mention "lack of attention to formal detail" when talking about Martin and Pullum and Pinker, they do no formal analysis at all. I am not a specialist in Eskimoan languages, but at least I have discussed this topic in detail with people who are; and they always have etymologically based criticisms of the various lists of alleged snow words that are drifting around — the lists are always padded with items that shouldn't be included. Cichocki and Kilarski follow Krupnik in doing no filtering for this padding effect. For a short and beautifully written reaction of one genuine linguistic expert to the kind of thing that is found in the bloated lists, Lawrence Kaplan's beautiful short paper on the topic would be an excellent place to go.
[Slightly re-edited on the morning of 31 January 2013.—GKP]