Skip to main content

Our Many Words for Snow

Researchers tested the infamous claim that native Alaskan languages have a gazillion more words for snow than English. Turns out, it’s sort of true.

By Nathan Collins


At the mouth of the Matanuska Glacier, in Alaska. (Photo: Dhilung Kirat/Flickr)

In Alaska, there’s snow in pretty much every direction. Falling snow, snow on the ground, blowing snow, re-frozen slush snow, avalanche deathtrap snow—the list goes on and on. In 1911, anthropologist Frank Boas realized that languages native to the region had distinct names for all those varieties of snow, while English has only a few, like, well, snow. Boas’ observation led others to conclude that languages are profoundly shaped by their physical environments—that native Alaskan languages, for example, have so many words for snow because there’s so much of it laying around.

Of course, nothing about that presumption is exactly correct. For one, Alaska is not just one big blob of snow. And while Eskimo-Aleut languages do have many words for snow, so does English—hoarfrost, groupel, firn, powder, penitente, and sastrugi, to name a few. And Boas’ initial argument has, over the years, become so grossly exaggerated that linguists and psychologists rarely take it seriously.

Temperatures were five to 10 degrees Celsius colder in places where there were separate words for snow and ice.

Yet there’s a certain logic to the basic idea, argue psychologists Terry Regier, Alexandra Carstensen, and Charles Kemp: The more something dominates our environment, the more we’ll notice subtle differences. The more we notice those differences, the more we’ll want and even need to communicate them. Hence, we invent names like “penitente,” the word for thin blades of snow that, to somebody at some point, resembled a person praying.

Regier, Carstensen, and Kemp decided to give that idea a proper test, the results of which they’ve published this week in the journal PLoS One. In particular, the team wanted to know whether there was a connection between a language having separate words for snow and ice, on one hand, and the likelihood people speaking that language would encounter snow, on the other. For the former data, the researchers looked up translations for ice in snow in dictionaries and other references works, including the Intercontinental Dictionary Series. For the latter, they derived local temperatures—a good proxy for the frequency of encountering snow—from the University of East Anglia’s Climate Research Unit’s data set.

Analyzing that data, the team found that temperatures were, on average, five to 10 degrees Celsius colder in places where there were separate words for snow and ice—evidence that people who have more experience with frozen water also have more words for the stuff. An additional analysis of Twitter data found that people in colder climates are more likely to use their languages’ words for snow and ice in their tweets, suggesting that cold environments contribute to a “communicative need” to differentiate between snow, ice, and their cousins.

“Our results support the claim that local communicative needs can leave their imprint on category systems across languages,” Regier, Carstensen, and Kemp write. “We hope that our results will help reclaim for analytical study an important topic that has been to some extent lost to popularization. We also hope our results will help to rehabilitate a widely criticized yet sensible way of thinking about variation in semantic systems across languages.”