|Title:||Response to McGinnis on Mundurucu/Piraha counting|
"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. study."
I was not responding directly to Pica's paper. I was responding to the
claim that their research cast doubt on the Piraha analysis that I
proposed, namely, that culture plays a crucial role as one of many
determining factors in the architecture of the grammar. For this
reason, numeracy is a small part of my paper, which also includes
studies of kinship, lack of embedding, lack of quantification words,
lack of color words, etc. It is the conjunction of these facts which, I
argue, is in need of a unifying explanation. I argue that culture, not
innate knowledge, provides this.
Martha's reply ignores this point. But it is crucial. It is the
particular constellation of properties in Piraha culture and grammar
that give the Piraha case its force. And that particular set of
properties is not found in Mundurucu, at least it hasn't been shown
that they do. So the Pica, et. al. paper, while very interesting for
studies of numeracy, is pretty much irrelevant to the claim in my paper
that culture can exercise an architectonic effect on grammar, affecting
even the so-called core grammar.
By the way, the most comprehensive study of numbers and counting in
Amazonian languages that I am aware of is by Diana Green and is, I
believe, available though the SIL Brazil website.
My paper addresses the genetic factor as well. Martha appears not to
have read that paper. What the paper says is that there are many cases
of Piraha raising children of Brazilian fathers. There is no difference
in their behavior relative to numeracy or any of the other facts I
discuss in my paper. This shows that genetics is highly unlikely to
play a role in the matter.
Moreover, in a recent field visit, Keren Everett was able to design
some more culturally appropriate counting experiments and these
preliminary studies do show that some Piraha adults seem to be able to
learn to count. This is predicted by my cultural account (among
others), but not by the Whorfian account in any clear way.
Rochel Gelman (Rutgers Psychology) has agreed to help me design
additional experiments to test numeracy among the Pirahas. Ted Gibson
(MIT Brain and Cognitive Sciences) has also agreed to discuss
experiment design for testing my claims on the lack of embedding. So
additional studies are planned.
I agree strongly that Whorfian relativity plays no explanatory role in
the Piraha facts. In fact, my proposal that it is culture affecting
language is exactly the opposite of a Whorfian account.
Daniel L. Everett
Professor of Phonetics & Phonology
Linguistics and English Language
University of Manchester