|Title:||Re: 15.3096, Mundurucu and Piraha Counting|
|Description:||Re: Linguist 15.3096 (http://linguistlist.org/issues/15/15-3096.html)
Dan Everett writes:
Again: the fact that your research shows that an analysis in
Language A is different from an analysis in Language B cannot be
taken to have 'cast doubt on' the analysis of Language B, especially
when the only connection between them is that they are 'only'
separated by about 1000 miles of Amazon rain forest.
I've been following the published debate with interest, and I must say that
this doesn't strike me as a fair response to Pica et al.'s paper. Even if
there are differences between the linguistic representations of number in
Piraha and Munduruku, the fact that Munduruku lacks words for numbers above
5 makes it a reasonable object for comparison with Piraha. The claim that
numeracy is affected by a lack of words for numbers does predict that a
lack of words for numbers above 5 will affect numeracy with respect to
amounts above 5. Indeed, this prediction is supported by the Pica et al.
That said, I'm not sure that either the Pica et al. or Gordon studies bear
on whether culture is the determining factor in the availability of words
for counting. The obvious alternative (that the determining factor is
genetic) wasn't strongly addressed by these studies. Pica et al. found
poor results on exact arithmetic for bilingual, educated Munduruku
speakers, but much more information about their education would be needed
to draw strong conclusions from this. I take the central claim of Gordon's
study to be instead that language (i.e., the availability of words for
counting) is a determining factor in the ability to do exact arithmetic.
As far as I can see, Pica et al.'s article supports this conclusion, with
the added caveat that it's the speaker's native language that matters.
Another population yet to be investigated is one that has native words for
numerals, but hasn't been educated in arithmetic. If the exact-arithmetic
skills of such a population parallel those of the Munduruku and Piraha,
this would be compelling evidence that education, not native language (or
genetics) is responsible for the effects on numeracy. If, on the other
hand, their performance parallels that of educated speakers of the same
native language, this would suggest that native words for numerals play a
significant role in triggering humans' innate ability to extend recursive
computation to arithmetic. The genetic hypothesis would then be quite hard
to test, unless there's an open tradition of adopted Piraha or Munduruku
infants acquiring a first language that has words for numbers!
Dr. Martha McGinnis, Assistant Professor
Linguistics Department, University of Calgary