I once saw a painting by a Nordic artist that depicted the sun as a woman and the moon as a man. In most Germanic languages that have masculine and feminine (including German and Nynorsk), ‘sun’ is indeed feminine and ‘moon’ is masculine.
On the other hand, in Romance languages the situation is reversed, and ‘sun’ is masculine and ‘moon’ is feminine.
So when I described the painting to a native Romance speaker, she was shocked.
Ever since that episode, I’ve thought that you couldn’t fully discount the Sapir-Whorf hypothesis, even if that placed me in the minority among linguists (I even wrote an essay about Whorf in my second year at uni).
I therefore read this article in the New York Times with great interest. You always have to be extremely cautious about believing any piece of linguistics published in mainstream media, but I didn’t spot any obvious errors.
It’s good that it quickly shows why the simplistic versions of linguistic relativism cannot be true, and I wasn’t aware of the existence of geographic languages, but the description sounds fairly convincing.
I guess I’d better get hold of a grammar of Guugu Yimithirr or Tzeltal to learn more!