August 13, 2004

Can Geoff Pullum rest on his laurels?

There's a new collection of essays called "The Genius of Language: Fifteen Writers Reflect on Their Mother Tongues", edited by Wendy Lesser, in which Amy Tan discusses "Eskimos and their infinite ways to say `snow,' their ability to see differences in snowflake conflagrations, thanks to the richness of their vocabulary." And at least one reviewer -- Philip Marchand in the Toronto Star -- gently but firmly corrects her:

Tan seems not to realize that this old canard about the Inuit having 32 different words for snow, or whatever the number, is pure myth. Apparently, the Inuit have only a few more words for snow than English speakers do. The Sapir-Whorf thesis has generally been oversold in recent decades, and perhaps it is time to give it a rest — it has had dire effects, such as the desperate attempt of disadvantaged groups to come up with new names for themselves, in the belief that such names will magically alter society's perception of them.

This is perhaps not exactly as Geoff Pullum would have put it, and I'm not quite sure what's going on with the "new names for themselves" business, but the rest of it seems pretty reasonable to me. And this is featured right up at the top of Marchand's review, it's not a by-the-way at the end.

Marchand devotes much of his review to a related issue that is apparently also widely featured in the book (which I haven't read yet): how characteristic differences between cultures manifest themselves in characteristically different ways of using language. He quotes from Ha-yun Jung's essay:

"In Korean, the first-person singular is an elusive voice," she writes. A Korean would be much more likely to say, `It would be nice to have an apple,' rather than `I want an apple.'

"Rarely will you hear a Korean speak — or write — consecutive sentences that start with I-this or I-that," she notes. "`I' seems to crawl behind the curtain at the first given moment."

Alas, Marchand introduces this passage in a problematic way: "Another interesting test case for the Sapir-Whorf theory is mentioned by a writer named Ha-yun Jung". I don't see that poor old Sapir and Whorf really have anything at stake in this matter. Their key idea was that differences among languages -- especially in plurality or gender or definiteness or other sorts of morphosyntactic marking -- should have an effect on what people pay attention to. Recently, Lira Boroditsky has been doing some interesting work supporting this idea in the case of gender marking in European languages. But the Sapir/Whorf hypothesis has nothing to say, as far as I understand it, about the case in which a language has a perfectly serviceable piece of morphology that its speakers mostly choose not to use.

Maybe Marchand meant that the low rate of first-person use in Korean tends to argue against the Sapir-Whorf hypothesis, by suggesting that meaningful differences in culture are often not reflected in the (the basic structure of) the resources that each culture's language makes available, but only in how speakers act on their linguistic opportunities? Thus showing that linguistic differences are not a necessary condition for cultural differences in characteristic modes of thought? Though the S-W hypothesis really is that that linguistic differences are normally a sufficient condition for cultural differences in characteristic modes of thought?

If this is how Marchand is thinking, then it's only a small problem in logic. The alternative is that Marchand is one of those who think that the Sapir-Whorf hypothesis means something like "there's some sort of relationship between thought and language, which shows up when you look at how members of different cultures tend to think and talk".

That would be too bad, because the other examples that Marchand cites in this connection are interesting ones, once poor old Sapir and Whorf are off the hook:

A curiously parallel example is provided by the Chilean writer Ariel Dorfman. He recalls an incident in school when he accidentally smashed an object with a hammer in carpentry class. Rather than say, `I broke it,' he told his teacher, se rompio — `It broke.' His teacher had a fit. According to Dorfman, he shouted, "Everything in this country is se, it broke, it just happened, why in hell don't you say, I broke it, I screwed up. Say it, say, Yo lo rompi, yo, yo, yo, take responsibility, boy." Sometimes that `I,' that `Yo,' needs to crawl out from behind the curtain.

Dorfman notes that his acquisition of English helped him to counter the "proliferation of passive forms" his Spanish-speaking friends employed to pass the buck. He is also careful to point out, however, that, "no one language condemns you to laziness or efficiency, mendacity or truth. If you dispose of two languages, therefore, you can lie twice as much — but also have a good extra whack at the truth, if you are so inclined."

I think that the kindest thing to say here is that mistakes have been made. Maybe Chileans tend to express things without ascribing agency; and maybe Dorfman became more sensitive to this point as a result of learning English; but surely this is not because English has inadequate resources, relative to Spanish, for evading responsibility. It seems much more likely that Dorfman was wrestling with issues of moral responsiblity at the same time that he was learning English, and therefore he associated some genuine grammatical differences -- such as the different role of reflexives, which I sincerely hope that Dorfman is not calling "passives" -- with his ethical concerns. If so, that's mainly a fact about Dorfman versus his friends, not a fact about Spanish versus English.

Here at Language Log, we believe that the evidence about the Sapir-Whorf hypothesis, and about broader issues in the relationship between language and thought, is mixed. At least, that's my personal view. The concepts are tricky ones, and the facts are not favorable to those on either side who want to see everything in black and white. However, there's an analogous hypothesis that I hope we can all strongly support. I'll express it by editing a famous passage from Benjamin Lee Whorf. (His original is in blue, my interpolations in black.)

'We dissect nature speech and language along lines laid down by our native languages our initially very faulty understanding. The categories and types that we isolate from the world of [linguistic] phenomena we do not find there because they stare every observer in the face; on the contrary, the world [of linguistic behavior] is presented in a kaleidoscopic flux of impressions which has to be organised by our minds - and this means largely by the linguistic systems in our minds the concepts, terminology and analytic skills (if any) that we have been taught in courses, or developed by careful thought and experiment. We cut nature speech and writing up, organise it them into concepts, and ascribe significances as we do in a coherent and interesting way, largely because only to the extent that we are parties to an agreement that holds throughout our speech community intellectual tradition, and is codified in the patterns of our language our learned techniques of description and analysis. The agreement is, of course, an implicit and unstated one, but its terms are absolutely obligatory; we cannot talk at all [coherently about speech and language] except by subscribing to the organisation and classification of data which the agreement decrees.'

In other words, the writers in Lesser's anthology, as sensitive as they are to issues of language, thought and culture, could all have benefited from a good introductory linguistics course. This would have given them the concepts, terminology and analytic skills to think and write about their own experience more clearly.

[Note: the other reviewers that I've read so far don't pick up on Tan's Eskimo snow blunder, nor on the other language, thought and culture questions that Marchand features. Matt King in the East Bay Express doesn't mention the issues at all in his review, nor does Charles Matthews writing in the San Jose Mercury News, nor does Brian Dolan writing in the San Francisco Chronicle.]


Posted by Mark Liberman at August 13, 2004 09:39 AM