Some years back when I read Paul Kay and Willett Kempton’s article What is the Sapir-Whorf hypothesis and they presented the evidence for the naming strategy, which is what they dubbed the phenomenon that people categorize by linguistic criteria instead of physiological criteria in a specific task, one thing hit me.
Let me start by putting this in context. Kay and Kempton were studying 5 native English speakers and 4 native Tarahumara speakers, in itself not exactly a mind-blowing number of subjects, and the subjects were to look at three Munsell color chips of which they were to “pick the odd one out”. These chips would all be close to the linguistic blue-green boundary of English and be arranged such that if chips A and B would be called green in English, chip C would be called blue, but A and C could easily be closest in terms of wavelength while still on each side of the lexical blue-green boundary of English. In Tarahumara, however, there’s no lexical distinction between blue and green. They use one word to cover the entire blue-green spectrum.
What Kay & Kempton found was that the Tarahumara speakers appeared to discriminate in terms of wavelength — i.e. they would pick the physiological “odd one” regardless of what an English speaker would name the color. The English speakers would pick the “odd one” by using the naming strategy — i.e. if two chips would be called blue, then the odd one would be the green one.
While I do think this experiment is very interesting, and definitely one of the better (and simpler) to be caried out in the name of categorization, there’s still one thing that bothers me. How can we be sure that English speakers aren’t simply choosing a logical way to discriminate? Remember, they’re asked to “pick the odd one out”, which necesarrily means that there must be an odd one and language seems to be a pretty logical — and obvious — choice in this situation. What would happen if they were asked to “pick the odd one out” of, say, an F-14 fighter, a can of beans and an elm tree? I’m pretty sure the subjects would find some way to do it. Maybe not uniformly, but they’d definitely be able to.
Now, a later study by Gilbert et al. (involving Kay himself), has shown that the naming strategy in fact primarilly occurs when the “odd one” is in the right visual field, but not the left. This supports a Whorfian interpretation when things are processed in the left hemisphere, but won’t that be exactly where the processing is taking place if you conduct the experiment as Kay & Kempton did back in 1984?
Maybe someone has already raised these issues, maybe they’re already dealt with or maybe “zere iz no zpöön” and I’m just looking at this wrong. Either way, I urge anyone with a little knowledge of this to comment and maybe drop a reference so I may be enlightened.
: Kay, P. & Kempton, W. (1984), What is the Sapir-Whorf hypothesis, American Anthropologist 86:65–79
: Gilbert, A.L. et al. (2006), Whorf hypothesis is supported in the right visual field but not the left. PNAS 2006 103:489-494
[Note: In spite of the heated debate on Danish language policy, I’m choosing to publish this in English in the hope that it’ll reach a larger amount of readers, thus contributing further to the development of linguistics as a science than it would if published in Danish.]