Critique of Behaviorism: Noam Chomsky
On Chomsky and Nativism
The Chomskyan revolution has had a great impact on the way many linguists study language. Chomsky's formal theory focuses on linguistic competence, the capacity of the human mind to generate "all and only the sentences of a given language." Nativism is a view of language acquisition, which claims that language development is innate. In short, contra Skinner, children are not born with a tabula rasa.
Chomsky proposes a universal grammar, which is a genetically “prewired” capacity for language with certain principles and parameters that shape language acquisition. In this view, children acquire language because of this innate language faculty. One piece of evidence for this claim: every child learns to speak grammatically well-formed utterances fairly quickly and without receiving explicit instruction. Further evidence: the temporary ‘developmental’ errors children make while they are in the process of (unconsciously) acquiring the rules of the language. (“I goed to the store” shows that they are testing the past tense rule and are not aware of the exception to the rule; this testing happens “out of awareness”). Moreover, all children acquire language by going through the same stages or milestones—babbling (around 6 months), one word (around 12 months), two words (around 18-24 months). After these stages, children are on their way through a kind of telegraphic speech that quickly adds other building blocks such as inflections and lengthier more complex utterances. Stephen Pinker calls this time of the child’s linguistic development a "grammar explosion."
In terms of second language acquisition, there is a debate about whether a second language is learned in the same way one learns one’s native language. Most research in this area provides evidence for a kind of sensitive or critical period (depending on the age of the learner) for language acquisition, after which it becomes difficult to achieve native-like competence, particularly for mastery of the sound-system.
The pedagogical implications of this theory can be enormous. If, in fact, a language is acquired without being taught, and a second language cannot be mastered fully even with instruction, what is the role of language educators for learning a second or foreign language? Is it impossible to master the grammar of a new language? Not impossible; research has shown examples of adults who have achieved a near-native proficiency. Second and foreign language educators know that heavy exposure to the language and meaningful interaction are important to learning.
What about computers? Computers are being used to create models of language acquisition. Computers can analyze massive amounts of language data. Researchers have recently begun to contribute language data to a project called CHILDES (Child language Data Exchange System). This data base of child language corpora now includes contributions of second language learners.
There is an entire field known as computational linguistics, which tests various theories and models of language. Developers of computerized language models have used nativist theory to ‘teach’ computers to generate grammatical structures. With a well developed model of language structure, a computer program should be able to generate all and only grammatical strings (sentences) in a given language. For an applied-linguistic example, Gerard Dalgish has used Chomsky’s notion of “phrase structure rules” in developing a program that generates sentences randomly in English. Further, researchers in speech sciences develop programs to synthesize speech from written text, and linguists have worked with computer scientists to develop speech recognition software, which turns speech into writing.