Dear Language Nerd,
My teacher once told me that the Eskimo have forty-five words for snow, but recently I read that they have over two hundred – a pretty big difference! How many is it really?
[name redacted by Language Nerd]
NOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
I suppose it was bound to happen sometime. From the moment I had the (now, it suddenly seems to me, terrible) idea of writing about language on the internet, the snow words debacle has been creeping inexorably towards me. The whole Eskimo-snow-words concept was thoroughly discredited by Laura Martin a solid thirty years ago, in her paper “‘Eskimo Words for Snow’: A Case Study in the Genesis and Decay of an Anthropological Example.” And that paper was soon after popularized in one of the all-time great linguistics essays, Geoffrey Pullum’s “The Great Eskimo Vocabulary Hoax.” Yet, for reasons that escape me, this myth continues to whirl around our society like some fearsome tempest, so I guess all I can do is add my voice to those hollering against the wind.
All right. I’ve girded my loins. Let’s do this.
Problem #1: “Eskimo” is not a language.
There are five Eskimo languages. The two with the most speakers are Inuit and Central Alaskan Yup’ik. (And, although the word “Eskimo” is sometimes disliked, the idea that it means something insulting like “raw meat eater” is not true — it’s the term used by the different Eskimo groups to refer to the whole, as opposed to, say, just the Nunavut.)
Problem #2: Inuit and Yup’ik are agglutinating languages.
Agglutinating, besides being possibly the most entertaining word in linguistics, is the term for languages that like to smush things together. Concepts that would be expressed in whole sentences in English get expressed as one reaaaaally long word. For example, Anthony Woodbury gives us “waterboots” as one word, “ivruci-t.”
Then he lists “they definitely don’t have anyone to make them waterboots” as… also one word, “ivruci-li-ste-ngqer-sugnait-uq.”
What this means for us is that, if you have a vague enough definition of “word,” then yeah, there are hundreds of words for snow, because “they definitely don’t have any snow” might get counted as one word. But by that method, there are also hundreds of words for roosters, cell phones, Batman, existentialist crises, and your mom. What we need to find out about are root words or lexemes, the small chunk of those long words that actually means “snow.”
Problem #3: There are two root words for snow.
This number is from Martin, who says that that’s the max for any of the Eskimo languages for words that “refer to snow itself (and not, for example, to drifts, ice, storms, or moisture).” She names the two from West Greenlandic, aput “snow (on the ground)” and qanik “snow in the air; snowflake.” Woodbury casts a larger net (he includes the word for to “have clinging debris” because that debris can be snow as well as lint or dirt), and he finds fifteen root words. But using a similar wide net on English, he finds over twenty snow words. Point being, no matter how you count them, there are not an unusually large number of snow words in the Eskimo languages.
Problem #4: The guy who first wrote about snow words listed four.
So, since we’ve seen that the languages don’t support the wild linguistic claims that get made about them, where did this whole thing start? The guy who accidentally set off the pebble that started this avalanche1 was Franz Boas. In 1911, he wrote that “Eskimo” (he studied Inuit) had four different words2 for snow, which he translated into four short English phrases that had the word “snow” in them. The point was that we say the word “snow” four times, while they say four different words, but that languages have to make distinctions and they’re largely arbitrary.
Problem #5: The rest is just made up.
So where where where did the whole idea that Eskimos couldn’t possibly stand our measly English words for snow and had to have a bazillion of them because they thought about snow all the time come from? Finally, we can finger the culprit: Benjamin Lee Whorf, who Pullum describes as a “Connecticut fire inspector and weekend language-fancier,” who unlike Boas had never been within a hundred miles of an Eskimo.3 As an extreme skeptic of linguistic relativity, I blame Whorf for basically everything, including Firefly getting cancelled.4 But this one definitely gets laid at his door, for in 1940 he took Boas’ research and, well, embellished it. Here’s the pivotal quote, as requoted in TGEVH:
“We have the same word for falling snow, snow on the ground, snow packed hard like ice, slushy snow, wind-driven flying snow – whatever the situation may be. To an Eskimo, this all-inclusive word would be almost unthinkable; he would say that falling snow, slushy snow, and so on, are sensuously and operationally different, different things to contend with; he uses different words for them and for other kinds of snow.”
That sure is an eloquent portrayal of Eskimo psychology that you just pulled out of your ass, Whorf.
This completely made up image gets picked up by mostly, to our undying shame, linguists, who pass it on to the rest of the world, and at every turn the number of words attributed to the imaginary single Eskimo language gets higher… and higher…
But hey, a cranky blog post should be able to turn this misinformation train around, right? Right?
The Language Nerd
1Avalanche, a special word just for lots of snow moving down a mountainside? Man, those wacky English-speakers must care a lot about differentiating between types of snow! </lampshading>
2He was not clear about whether he was talking root words or just “words,” thereby also invoking problem #2.
3Probably? Full disclosure, I have not actually traced Whorf’s exact distance from Eskimos across his lifetime.
4I only learned while researching this article that Whorf came up with the concept of the allophone, one of the cornerstones of modern phonology. So he may deserve more credit than I give him. Still full of shit about snow, though. And we’re not even gonna start on Hopi time.
Update: I came back and edited section 4 after doing further research on Boas, because my original phrasing here was too strong. Boas does say that he thinks that the use of related or different words can be based “to a certain extent” on if the language-users need to “distinguish a certain phenomenon in many aspects.” However, he is definitely not drawing the kinds of conclusions that Whorf does. In fact, the example he gives right before the snow one is that English has the different words “river” and “dew” and “lake” (as opposed to, say, phrases like “fast water” and “morning water” and “big flat water”). Somehow this failed to touch off a century-long discussion of how obsessed English speakers are with the nuances of water types.
Got a language question? Ask the Language Nerd! email@example.com
Or: Twitter @AskTheLeague / facebook.com/asktheleagueofnerds
My two main references for the life and times of this little drama are probably clear, but let’s go ahead and link them once more, because they are excellent reads: Laura Martin’s “‘Eskimo Words for Snow’: A Case Study in the Genesis and Decay of an Anthropological Example” and Geoffrey Pullum’s “The Great Eskimo Vocabulary Hoax.” That’s also the title of one of Pullum’s books, which I sadly do not own. I do keep up with his blog, though, where as recently as this January he was continuing to fight the good fight: “Notice that English has a huge range of words for fools (blockhead, dope, dunderhead, idiot, moron, twit … the thesaurus has dozens), but that betokens no deep interest in fools, nor any native expertise in discriminating different types of dimwit.”
Actual info on Eskimo languages, particularly Yup’ik, comes from Anthony Woodbury. He kindly put together “Counting Eskimo words for snow: a citizen’s guide” back in 1991 when TGEVH started getting around, and looks at inflection in “The word in Cup’ik,” from Word: A Cross-Linguistic Typology, where I got the “waterboots” example. (Cup’ik is a dialect of Yup’ik.)
“Raw meat eater” debunking from Mark Israel, from back in the day when newsgroups were where all the hot linguistics info was at.
The Whorf quote comes from his article “Science and linguistics,” in Technology Review, which I can’t get to a direct site for, but which is easily found in Google scholar.
Franz Boas’ first four little words are found in his introduction to the Handbook of American Indian Languages.
P.S. – Hope y’all are ready for the debut of the awesome Art Nerd later this week, because I know I am super excited.