Boas Said Inuit Has Four Words for Snow; Sixty Years Later the Number Was 100

In 1911, the anthropologist Franz Boas published his Handbook of American Indian Languages. In a brief passage about morphological compounding, he gave four root words used in West Greenlandic for different states of snow — aput (snow on the ground), qana (falling snow), piqsirpoq (drifting snow), and qimuqsuq (a snow drift) — to make a methodological point about how Eskimo languages glue affixes onto roots in ways that English uses separate phrases for. The roots themselves weren't surprising. The point was the morphology.

The number escaped its enclosure. By the 1940s, the linguist Benjamin Lee Whorf was citing seven words to support a stronger claim about how language shapes thought. By 1978, popular essays had it at fifty. On February 9, 1984, The New York Times ran a piece reporting "100 Eskimo words for snow." By the early 1990s, the number was a Trivial Pursuit answer.

The linguist Geoffrey Pullum wrote a 1989 essay he eventually titled "The Great Eskimo Vocabulary Hoax," pointing out that Inuit and Yupik languages are polysynthetic: they form words by stacking suffixes onto a small set of roots, in such a way that you can produce arbitrarily many "words for snow" — or for anything else — without saying anything English-speakers couldn't say with adjectives and prepositions. Counting words across languages with that different a structure, he wrote, was a category error.

The more recent and embarrassing footnote is that some Inuit dialects do, in fact, have many distinct snow roots when carefully counted. Igor Krupnik's 2010 work documented Central Siberian Yupik with about 40 and Nunavik Inuktitut with at least 53. The hoax was the inflation; the underlying observation about lexical detail was, in a more limited way, true.