21 Apr Language Inherited
How do children acquire language? How do they learn about grammar and productively apply its general rules to creating new utterances? Some attempts to explain this phenomenon have suggested that a set of grammar rules is innate. Advocates of this theory (Chomsky 1968) point to grammatical similarities, or universals, across languages and aver that a set of grammar rules is hard-coded into the amino-acid chains of our genetic inheritance. The a-priori or “given” structural knowledge approaches, such as the X-bar Theory, stating that all phrases are endocentric, say it at least begins in our genes. Language acquisition involves testing input against the set of all a-priori knowledge and seeing how the language of the house fits the parameters that come as standard equipment. New, non-universal parameters are added as they are encountered, and the union of the given and learned constitutes an adult grammar. The adult grammar is treated as an end point. According to this theory, language acquisition has a beginning prior to conception, and an end, presumably at some time in young adulthood.
|Understanding Context Cross-Reference|
|Click on these Links to other posts and glossary/bibliography references|
|Prior Post||Next Post|
|Learning by Doing or Active Osmosis||Breaking Down Language Structure and Function|
|language acquisition||Linguistics Net on X-Bar Menyuk 1988|
|innate grammar||Chomsky 1968 Chomsky 1986|
|structural knowledge||Bobrow 1975 BU on X-Bar|
A-priori Linguistic Capabilities
Several possible explanations for cross-linguistic similarities exist. Environmental factors such as the procedural constraints of communication and functional constraints of language as an expressive symbol system could account for some syntactic universals. It is also possible that diachronic factors in language such as common ancestry and interchange by exposure over time account for some similarities.
Chomsky’s X-bar theory supports the premise that language knowledge is biologically inherited. This theory is based on observations that suggest certain characteristics or syntactic patterns exist in all languages in one form or another. If we accept the existence of language universals, those intrinsic similarities between languages that presumably developed in total segregation from one another, it is natural to presume that, rather than being learned knowledge (nurture), language skills are biologically inherited knowledge (nature). This is an overly simple articulation of the raging debate on nature vs. nurture. Whatever the case, the innate explanation is unsatisfying from a scientific standpoint because of the lack of corollaries and hard evidence other than the arguably universal characteristics shared by all known languages.
Furthermore, even if we espouse innate assumptions about syntax, much of language acquisition is left unexplained. Therefore, it seems a safe and supportable assumption, whether or not it is true, to say that children start life with a clean slate, knowing no specific language and nothing about language structure. Children clearly have brains capable of symbolic thought, pattern recognition and learning, and all the faculties for speech, auditory and visual discrimination and writing. Perhaps this has no bearing on the formalism, but it explains why many (including this author) assume the beginning of language acquisition follows concept learning, and the end-point is either senility, death, or some other terminal mental state.
Can computers learn language without having a prior set of knowledge? Probably not! We’ll have to build some things into computers, like parsers, dictionaries, thesauri and ontologies, to give them a foundation from which they can begin to learn languages. There are likely to be completely different algorithms or heuristics to replicate the capabilities in Broca’s and Wernicke’s areas. Once we do, however, the capacity to learn new languages will increase with each new language computers learn. So, in some respects, the question: “Is language inherited, or do we build it upon a foundation of exeriences in the real world?” may have little or no impact on our models.
|Click below to look in each Understanding Context section|
|4||Perception and Cognition||5||Fuzzy Logic||6||Language and Dialog||7||Cybernetic Models|
|8||Apps and Processes||9||The End of Code||Glossary||Bibliography|