21 August 2004

Language and thought

New Scientist has an article about a recent study of a Brazilian tribe and how they differentiate between different numbers of items. The tribe, the Piraha, have a word for "one," "two," and "many" but not for any other values. Members of the tribe were shown a group of four items and then a group of five items. They could not tell the difference between the two groups.

This suggests that language defines, or can define, the limits of what we are able to think.

This is a controversial conclusion, first expressed as the Sapir-Whorf hypothesis in the first half of the 20th century, and is still very much open to debate.

The Piraha language itself has several interesting features:

  • It has only 150 speakers,
  • It consists of only 10 phonemes (basic units of sound), the least of any language,
  • Its typology is agglutinative (see below),
  • It uses five "discourse channels," where the words can be spoken, whistled, hummed, yelled, or sung

Language typology describes the grammatical features of a languages. Agglutinative describes Piraha's location in a continuum of language types from those that isolate all parts of speech into separate words (analytic languages such as Chinese) to those that bundle all parts together into one word per sentence (polysynthetic languages such as those spoken by some American Indians).

This is one small study of many, so there's still no clear consensus. As the researcher suggested in the article, the limitations of language on possible thought probably hold for some types of thought and not others.

[ posted by sstrader on 21 August 2004 at 1:34:25 PM in Science & Technology ]