March 16, 2004

Panting up the Hill of Significance

Steve at Language Hat writes today:

"That poor meme is panting and sweating, but it just can't drag the load up the Hill of Significance."

A nice way to describe the widespread misapprehension that a culture's degree of interest in a topic can be accurately measured by counting the number of words it has for relevant concepts. Steve is critiquing a New York Times Op-Ed piece by Simon Montefiore on Russia, which applies this idea to the concepts of "law" and "personal connections". As Steve shows, the vocabulary metric indicates that "Russia must be twice as legal a culture as any English-speaking one", and that in the "personal connections" area the two cultures are more or less in a dead heat, thus refuting Montefiore's thoughtless little sentence: "There are few words in Russian for the Western concept of 'law,' but there are legions of words for connections, helping people from one's neck of the woods."

One reason for this meme's resilience is that it's based on a valid generalization: people develop the concepts that their lives and their livelihoods require, and find or invent words or word-senses or fixed phrases to express those concepts. As a result, for example, the Carrier have lots of beaver words, and the Somali have lots of camel words, and not vice versa. This is neither arbitrary nor surprising.

Why is the Eskimo snow-words meme so often wrong, then? Because it's usually a cheap rhetorical device, masking facile cultural stereotyping and little or no actual linguistic analysis.

As an expert in Soviet history, Montefiore doubtless knows that Russians have plenty of experience with legal codes, both civil and ecclesiastical; and as a writer with six or seven books from American and British publishers, I bet he knows that English speakers have plenty of experience with social networking. There may well be differences between the cultures in these areas -- there are certainly profound historical differences in the civil societies -- but it's naive to think that these differences will show up in word counts, or even in the counts of culturally identified concepts. Montefiore wants to draw attention to the role of personal networks in Putin's (and Stalin's) Russia. Fair enough, but it's one thing to assert that personal power and patronage play a bigger role in Russia than in (say) France, and a completely different thing to believe that Russians make a larger number of conceptual and lexical distinctions among types of personal and social influence. It might be true -- though I doubt it -- but some evidence should be required, not just a facile assertion that relies on the reader being too ignorant or too uninterested to check.

And evidence is just what we almost never get. Such claims are usually asserted without any careful comparative linguistic analysis, and often without any linguistic analysis at all. The infamous Eskimo snow words meme is spread by people who don't know about Inuit or any related language. Montefiore doubtless knows quite a few Russian words for the form, contents and use of social networks, but he clearly has not tried to compare this part of the Russian lexicon systematically -- even by the blunt instrument of counts -- to the comparable vocabulary of other languages. His rhetoric needed a piece of evidence about the role of relationships vs. laws in Russian politics, and this assertion about numbers of words is brief and clear. And probably wrong, at least in its implicit comparison to languages such as English and French, but you can't have everything...

Posted by Mark Liberman at March 16, 2004 09:00 PM