Language and attitudes

JOHNSON has previously expressed doubts about recent studies purporting to show that the language you speak affects the content of your thoughts in clear and measurable ways. This thesis was most famously voiced by Benjamin Lee Whorf, an American linguist, in the 1930s. Now that a team of psychologists at Harvard University have put their oar in with a paper just published in the Journal of Experimental Social Psychology, it would be remiss of us not to weigh in. Especially as the claim is indeed quite weighty: Oludamini Ogunnaike and his colleagues suggest that language influences speakers' implicit attitudes towards different ethnic groups, including their own.

The team used the so-called implicit-association test (IAT) often used by psychologists to study bias. In it respondents are asked to categorise words which flash on a computer screen or are played through headphones by pressing one of two keys on a keyboard. The team asked their subjects, Moroccans bilingual in Arabic and French, to press one button when they saw either an Arab name (such as Hassan or Fatimah) or a "good" word (like "nice" or "happy"), and another when they saw a French name (Pierre or Marie, say) or a "bad" word (like "hate" or "sadness").

Participants only get a fraction of a second for each word, which leaves them with no time to ponder their options and thus elicits automatic responses. So, if they implicitly associate "good" words with Moroccans and "bad" ones with the French, they ought to be able to perform the classifications faster than if their prejudice were the other way around. (As a control, the test was repeated with Moroccan names and "bad" words sharing the one key, while French monikers and "good" terms shared the other.)

When the IAT was conducted in Arabic, the Moroccans were discernibly quicker off the mark when the Arab names shared a key with "good" words and the French with "bad" than with the key assignments reversed. However, when asked to perform the same task in French, the difference disappeared (though no significant preference for French names was observed).

Mr Ogunnaike then repeated the experiment with Spanish-English bilinguals in the United States, with like results. (A study published earlier this year in Psychological Science by another team found a similar linguistic bias when Arab Israelis bilingual in Arabic and Hebrew were asked to perform IATs with Arab and Jewish names.)

By performing the experiment on bilinguals, the researchers eliminated a slew of confounding variables at one fell swoop. The assumption is that since we are dealing with the selfsame individuals, everything about them ought in principle to remain constant over the brief timespan of the experiment. The language in which the test was administered really does appear to be the only non-random independent variable. And it is shown clearly to affect participants' attitudes.

At first blush, this looks like grist to the neo-Whorfian mill. However, as my fellow Johnson explained in an earlier post:

My guess is that speaking a foreign language puts you, psychologically, in the place of the speakers of that language, including in their culture. Learning a language, you're encouraged to mimic the rhythm and pace as much as the individual sounds. This can lead to a kind of out-of-body experience. Learning Italian recently, I streeetched out the long vowels and bounced up and down staccato-style on the short syllables, picturing an over-friendly Italian waiter. And that's how I felt: garrulous, friendly, loose. Nothing about Italian's grammar or vocabulary did this to me. I was just picturing myself as a little more Italian.

In other words, in these IAT results it isn't the structure of the language that helps shape our judgments, but rather the cultural cues stemming from a particular tongue. Or, to caricature the notion, consider your immediate responses to the following questions: "Do you like eating meat?" and "Do you like eating dead animals?" Clearly, "meat" and "dead animal" aren't perfect synonyms, but then neither are most allegedly equivalent terms across different languages.

An unrepentant sceptic could always retort that recent research in social psychology suggests plenty of seemingly extraneous factors, like the weather or tidiness of the experimental venue, sway decisions. Rather than elevate language to the rank of a crucial shaper of attitudes, he could argue, these findings actually bring it down to earth, and into the ever-expanding list of irrational influences on human behaviour.

Still, there is no denying that at the very least the latest results give the hoary exhortation to "mind your language" a somewhat new meaning.