Page 108 - Encyclopedia of Philosophy of Language

Language and Mind
Humboldt and in the twentieth century by Edward Sapir, greatly influenced by the study of non-Eur- opean languages, and culminating in the work of Benjamin Lee Whorf: this viewpoint has been called 'linguistic relativity'. In Whorf's words:
We cut nature up, organize it into concepts, and ascribe significances as we do, largely because we are parties to an agreement to organize it in this way—an agreement that holds throughout our speech community and is codi- fied in the patterns of our language.
(Whorf 1956:213)
Whorf gave lexical examples (the familiar assertion that American Eskimoes have more words than Eur- opeans for snow) and syntactic examples of the differ- ences between languages which he claimed reflected differences in thought. Thus the Hopi, lacking in their language a system of tenses similar to English past, present, and future, would also lack English notions of time and velocity. For Whorf, similarities of gram- matical structure necessarily led to similarities of con- ceptual structure:
The English sentences 'I push his head back' [in Shawnee ni-kwaskwi-tepe-n-a] and 'I drop it in water and it floats' [Shawnee ni-kwask-ho-to] are unlike. But in Shawnee the corresponding statements are closely similar, emphas- izing the fact that analysis of nature and classification of events as like or in the same category (logic) are governed by grammar.
(Whorf 1956:235)
What all these examples, from Port Royal to Whorf, lack is any assessment of thought processes inde- pendent of their expression in language. How does one know that the propositions used in judgment cor- respond to the subject-predicate linguistic structure, that one has assessed the 'profundity' of the British and their language separately, that Shawnee con- ceptual categories correspond to Shawnee gram- matical categories? Listeners are notoriously bad at keeping the ideas expressed in a linguistic message separate from superficial features of the message. Thus the same message, expressed in identical words by speakers with different dialects, is often less favorably evaluated when spoken in a socially less prestigious dialect, e.g., Quebec French. This is not an example of Whorfian linguistic relativity (the words and the grammar are the same), merely a demonstration that it is no trivial matter to assess such things as Jenisch's 'moral essence' and language separately.
2. Experimental Tests of the Whorfian Hypothesis
It is unfortunate that the largest research effort related to the Whorfian hypothesis has involved rote learning of simple colored stimuli. The idea seems appealing: choose a dimension for which we can be sure the sensory information is processed similarly the world over, but where different languages code this sensory information in reliably different ways. However, color
is too tightly related to the physics and physiology of vision, and rote learning is too modest an exemplar of what could be regarded as thought for significant interactions between language and thought to be apparent.
Initial investigations, however, were promising. A study by Brown and Lenneberg (1954) assessed the linguistic 'codability' of colors and showed this was related to the colors' discriminability and memor- ability. Codability was assessed by a number of mea- sures of length and speed of response and intersubject and intrasubject consistency: a color which was given a short response (e.g., red), and which subjects pro- duced rapidly and consistently on repeated pres- entations of the color was regarded as highly codable; a color that might be described reddish purple, which was produced more hesitantly and with less con- sistency would be regarded as less codable. Examples of these colors were briefly presented to American college students (first language English) and after delays ranging from seven seconds to three minutes they were required to point to them on a large chart of possible colors. There were positive correlations between codability and memory performance, and these correlations were larger the greater the delay. A study of Zuni Indians, using similar materials, showed that there was not a complete correspondence between the codability of colors for English speakers and Zuni speakers. In particular, the Zuni do not have a label to distinguish between orange and yellow, and this was related to the Zuni speakers' memory performance,
where they frequently confused orange and yellow stimuli.
Thenotionofcodability,whichprovestobea useful concept, shall be returned to, but first it must be poin- ted out that crosslinguistic studies with color have proved to be more difficult to interpret than was first thought. Rosch's work was prominent in the 1970s. Rosen's starting point was the work of Berlin and Kay in 1969, who had established that color terminology was not arbitrary across languages and, despite dis- agreements between speakers about where the bound- aries between various color terms should be placed, there was good agreement, even across languages, about the identities of certain basic color terms, which Berlin and Kay termed 'focal' colors. Rosch studied the Dani, an agricultural people of West Irian, who have only two color terms. For them, focal colors were not more codable than other colors, but they were more memorable (tested by recognition after 30 seconds) and they were more learnable (tested by pair- ing colors and arbitrary names and testing learning over several days). This suggests that an important influence on performance on these tasks is the precise location of the color in a psychological representation, which is determined by innate and universal properties of the color-vision system, not by language-specific labels.
86