Lecture14: Thought & Language

One of the best comments on the relationship between thought and language that I ever heard was reported in the New York Times "Metropolitan Diary" section. In a coffee shop someone overheard one elegantly dressed woman say to another:
"Thank goodness for the word 'muffin'. Otherwise, I'd be eating cake for breakfast every morning."
This New York lady was under the delusion that Whorf's linguistic relativity hypothesis is still valid.

In the 1920s cultural anthropologists and linguists were making contact with languages strikingly different from their own. Edward Sapir, a distinguished Yale linguistic anthropologist, expressed his conclusions thus:

"Human beings do not live in the objective world alone, nor alone in the world of social activity as ordinarily understood, but are very much at the mercy of the particular language which has become the medium of expression for their society. It is quite an illusion to imagine that one adjusts to reality essentially without the use of language and that language is merely an incidental means of solving specific problems of communication or reflection. The fact of the matter is that the 'real world' is to a large extent unconsciously built up on the language habits of the group. No two languages are ever sufficiently similar to be considered as representing the same social reality. The worlds in which different societies live are distinct worlds, not merely the same world with different labels attached."
In other words, the particular language you speak affects the ideas you can have: the linguistic relativity hypothesis. Benjamin Whorf, an amateur linguist and fire inspector for Hartford Insurance Company, studied with Sapir at Yale and was deeply impressed with his mentor's view of thought and language. Whorf extended Sapir's idea and illustrated it with examples drawn from both his knowledge of American Indian languages and from his fire-investigation work experience. The stronger form of the hypothesis proposed by Whorf is known as linguistic determinism. This hypothesis has become so closely associated with these two thinkers that it is often 'lexicalized' as either the Whorfian hypothesis or the Sapir-Whorf hypothesis.

Languages differ in two major ways: lexically and grammatically. Both of these areas have been explored in great detail in efforts to investigate the intuitively appealing linguistic relativism hypothesis.

Lexical Differences

Color Naming

The color spectrum is continuous, but we categorize it into discrete areas by our use of color names. Languages contain different numbers of color names and carve up the spectrum somewhat differently. The linguistic determinism and linguisitic relativity hypotheses would hold that people's perception of color will be influenced by the color terms available to them in their language.
Berlin&Kay 1969
The cognitive anthropologists Brent Berlin and Paul Kay established that there are 11 basic color terms. Basic color terms were defined as those color names that fulfilled all of the following criteria:
  1. consist of only one morpheme
  2. not contained within another color word
  3. not restricted to a small number of objects (e.g., blond)
  4. common and generally known
They discovered that there were regularities in the pattern of color names selected in different languages that were constrained by the physiology of color vision. Two levels of retinal representation of color are important here. There are three cone or photoreceptor types that vary in their sensitivity to wavelength. They are known as the short, middle and long wavelength cones, or more mnemonically, but not quite accurately, the blue, green and red cones. At a later neural level of the retina these cone types are combined to form opponent color processing cells thus:

When a language contains fewer than 11 basic color terms the terms consist of unions of the basic categories. Black and white, or light and dark, is the minimal set. Red is the color that will be added next, hence the term "laundry sorting theory". If there are between 4 and 6 terms those terms will be composed by adding yellow, blue or green. At this point all of the 'primary' or basic opponent colors are present. The next color to be added will be brown (a mixture of yellow and black). Then come purple (red and blue), pink (red and white), orange (red and yellow), and gray (black and white). Certain possibilities are ruled out by the spectrum and physiology; for example, no language has this set of color terms:

black/white
brown
orange
The idea of basic or focal colors that provide the best exemplar of a category falls quite naturally out of the opponent processing representation. The focal blue response is given when the B-Y cells are giving their most possitive response and the R-G cells are neutral. The focal purple repsonse occurs when the B-Y cells and the R-G cells are firing maximally.
Rosch's Experiments with the Dani
The cognitive psychologist Eleanor Rosch took the bold step of doing some fieldwork. She did some experiments with the Dani people of Papua New Guinea whi have two basic color terms in their language: 'mili' for cold, dark colors and 'mola' for warm, light colors. A group of Dani and a group of Americans were given a recognition memory task involving 40 colors. A sample color was briefly presented, and half-a-minute later the participant selected the sample from the full set of 40 colors. The American group made fewer recognition errors, but both groups made very similar kinds of errors. The color spaces derived from a multidimensional scaling analysis of the errors were highly similar. This suggests that there is a common, underlying conception of color relationships that is due to physiological rather than linguistic constraints. This view is further supported by further experiments on learning color names. Rosch found that the Dani could more easily learn a set of color names (based on their complex kinship terminology) for the focal rather than the non-focal colors. In other words, the prototypical or "best exemplar" colors have significance for the Dani even though they are not marked linguistically or 'lexicalized'.

Perhaps color terms were not the most fortunate choice for testing the Whorfian hypothesis since they are clearly embodied in that the focal colors are strongly determined by biology. Color categories result from the spectral composition of light reaching the eye plus human biology plus a cognitive mechanism that has some of the characteristics of fuzzy set theory plus
WORLDplusHUMAN BIOLOGY plusCOGNITIVE MECHANISMplusCULTURE SPECIFIC CHOICE
visible spectrumopponent color mechanismscombination of outputs of
opponent color mechanisms		naming of groups of focal colors
basic categories

The Great Eskimo Vocabulary Hoax

Franz Boas, linguistic anthropologist reported that just as English uses separate roots for a variety of forms of water (liquid, lake, river, brook, rain, dew, wave) that might be formed by derivational morphology from a single route in another language, Eskimo apparently uses four distinct roots for 'snow': aput 'snow on the ground', gana 'falling snow', piqsirpoq 'drifting snow', and quimuqsuq 'a snow drift'. Whorf picked up Boas' example and used it, vaguely, inflating the number of terms to at least seven, in a 1940 article on "Science and Linguistics". From there, popular accounts of Whorf's idea have inflated the number to hundreds. The claim is false: think of snow, slush, sleet, blizzard, and avalanche.

Lexicalization

There are many concepts that have not been lexicalized in English, for example, 'the interior surfaces of a room', but are easily apprehended. Whether concepts are 'lexicalized', that is denoted by one word, has to do with the frequency of use of the concept: if we want to talk about a concept fairly often we will coin a new word for that concept, for example 'television' from 'tele' meaning 'far' and vision. It is perfectly possible, however, to have concepts without lexicalizing them. Bearing this concept of 'lexicalizing' in mind puts the claims about words controlling thoughts into perspective. Even if the Eskimo words-for-snow observation were not a hoax, it would not indicate that speakers of English were less capable of thinking about or observing the differences among varieties of snow, they would simply have to use more words to express the same concept. The complexity of expression does not preclude the ability to conceive of the distinctions.

Grammatical Differences

Thinking of the Future

Consider the speech of the fictional 'river people' in Carter's novel 'The infernal desire machine of Doctor Hoffman':
The speech of the river people posed philosophical as well as linguistic problems. For example, since they had no regular system of plurals but only an elaborate system of altered numerals for denoting specific numbers of given objects, the problem of the particular versus the universal did not exist and the word 'man' stood for 'all men'. This had a profound effect on their societisation. The tenses divided time into two great chunks, a simple past and a continuous present. A future tense was created by adding various suffixes indicating hope, intention and varying degrees of probability and possibility to the present stem.
On first reading this seems logical: if the river people have no simple future tense this should affect their conception of the future as a simple counterpart to the past. The striking thing about this example is that it is an accurate description of the English tense system. We are obligated to use modal auxiliaries that express hope, intention and varying degrees of probability and possibility when we refer to the future.
We say, "I might walk", "I will walk", "I should walk" and "I shall walk". Whereas in the past or present tense it is not necessary to express hopes and intentions:
We say simply, "I walked" and 'I am walking".

Do we want to claim that English speakers have a different conception of the future than French speakers who have a "proper" future tense?

Bloom's Counterfactual Hoax

In English the subjunctive is used to express counterfactual statements: events that are known to be false but are entertained as hypotheticals. Chinese lacks a subjunctive, thus counterfactuals must be expressed circuitously. Bloom claimed to find evidence for the Whorfian hypothesis by demonstrating that Chinese speakers had more difficulty with counterfactual statements than English speakers. Bloom presented 120 native Chinese speakers and 55 native English speakers with a story about a European philosopher named Bier. He tested their understanding of a counterfactual statements of the form 'If Bier had known about the technique to mnaster the Chinese language, he would certainly discover the different attitudes between the Chinese and European philosophers.' 54 of Bloom's American subjects understood the implications of this sentence, in comparison to only 8 of his Chinese subjects.

Terry Kit-Fong Au, a native Chinese speaker and psychologist at Harvard, did not take kindly to this linguistic slight of his presumed powers of reasoning. He repeated Bloom's experiment with one crucial change: he asked Chinese bilinguals to translate an idiomatic Chinese version of the story into English. With this translation his results were in the reverse direction from Bloom's. Only 60% of American high school students who read the nonidiomatic versions understood the counterfactual, whereas 97% of Au's monolingual Chinese subjects who were given an idiomatic Chinese version grasped the significance of the counterfactual.

Bad translation may also have been at the root of Whorf's claims about Hopi concepts of time.

Sources