麻豆传媒

THE opening pages of The Stuff of Thought present the same arguments on the nature of language as Steven Pinker鈥檚 earlier book, The Language Instinct. Language and grammar, Pinker says, are hard-wired into our brains. He presents the Chomskian notion that 鈥渃hildren must be equipped with an innate universal grammar; a set of plans for the grammatical machinery that powers all human languages鈥�.

According to Pinker and Noam Chomsky鈥檚 adherents, children do not learn the particular syntactic processes that characterise the grammar of their native language. Instead, the rules of syntax for every language on Earth lie dormant in their brains. The details of syntax are software that has been preloaded into our brain鈥檚 hard drive by evolution.

Now The Stuff of Thought extends the software licence from syntax to words themselves. According to Pinker, we don鈥檛 learn the meanings of particular words, but instead have a store of innate 鈥減rimal concepts鈥� that are activated when we hear a word. These concepts 鈥� such as cause, motion, space and time 鈥� comprise the elementary building blocks of language and thought. Presumably, when a child sees her mother point to a dog and hears her say 鈥渄og鈥�, innate primal concepts like animate, animal, safe and perhaps edible are activated. Innate concepts, says Pinker, give language its power. 鈥淚f meanings could be freely reinterpreted in context, language would be a wet noodle and not up to the job of forcing new ideas into the minds of listeners.鈥�

Certainly there are limits to the range of meanings a word can take, but to me it seems clear that we actually learn the meanings of most words through experience, or through secondary sources such as books, film and television. Everyday words like fire, table or dog have multiple meanings, and the same words can mean different things to different folks. A short session with the Oxford English Dictionary confirms that a word鈥檚 meaning can change over time.

The 10-tonne gorillas missing from The Stuff of Thought are neurophysiological studies showing how the brain deals with words and their meanings. Studies such as those reported by Alex Martin and Linda Chao (Current Opinion in Neurobiology, vol 11, p 194) show that a word activates the same areas of the brain that are involved in perceiving the thing to which the word refers, as well as pertinent motor areas. These findings point to 鈥渇uzzy鈥� representations of a word鈥檚 meaning acquired through life鈥檚 experiences, rather than hypothetical innate primal concepts.

Ultimately, Pinker neglects the single biological truth that rules out innate detailed knowledge of language: genetic variation. If Pinker鈥檚 hard-wired elementary concepts actually existed, genetic variation would ensure that some people lacked a concept or two. We would encounter people who could not acquire the meanings of entire sets of words or think or act upon these concepts. A child missing the gene coding for the primal concept kinship would never be able to understand words such as family, cousin, mother and so on. Certainly humans possess an innate capacity to acquire language, but as I pointed out in my 1984 book The Biology and Evolution of Language, genetic variation rules out theories that posit innate knowledge of language鈥檚 details.

鈥淕enetic variation rules out Pinker and Chomsky鈥檚 theories鈥�

Pinker鈥檚 discussion of the linguistic analysis of verbs and its relation to meaning is a lucid introduction to formal semantic theory. Jokes and rambling commentaries on topics from Kant to swearing provide breathing space for readers overloaded with the minutiae of theoretical linguistics. But readers who wish to learn more about current research on the manner in which language influences the way we think and act will have to continue their search.