| |
For perhaps the majority of linguists, and other researchers interested in language, computers have come to be seen as a valuable analytical tool. Traditionally, language-related topics, such as the study of grammar, have relied on a corpus of sentences which, though constantly increasing, is nevertheless pointedly small in comparison to the sheer quantity of authentic language data which can now be scrutinized. Undoubtedly there will continue to be valid reasons for focusing on a limited and carefully selected body of data, but equally clearly, new insights on the very nature of language are emerging simply through the study of vast corpora of language samples.
A major branch of computational linguistics has been involved with the attempt to produce computer systems which can effectively simulate the way we both understand and produce natural language. However, it is one thing to programme a computer so that its language outputs resemble those of natural human language. It is quite another to devise a computational system which processes language in the same way as human beings do, and it is this latter aim to which many researchers aspire. A fundamental problem in this approach has been how to accommodate the common phenomenon of breakdowns in the natural process of communication. If we want a computer to be able to interact effectively with people, then it will have to be able to cope with the multifarious range of minor interruptions, reworkings, signs of incomprehension and so on which characterize natural linguistic interaction.
Computers have also become influential in the preparation of dictionaries. It is becoming increasingly common for the initial stages of dictionary-making to be based on the computational analysis of enormous collections of authentic language samples, comprising several million words, both written and spoken, from a wide variety of sources. The guiding concept has been the use of concordancing, in which an enormous language corpus is scanned for every example of a particular word, along with its linguistic context. (The computer highlights several words on either side of the desired word.) The great advantage of such concordancing is that it allows the lexicographer to establish which words are consistently associated together. Thus, our knowledge of collocations has been vastly improved and is now beginning to provide many new insights on the inner workings of lexical and syntactic organization.
Without doubt, the age of the computer has opened up new vistas for research on a wide range of language-related topics. Machine translation, computer-assisted language learning, speech synthesis, stylistics, lexicography and lexicology, speech processing, syntax, artificial intelligence are characteristic areas. MS
Further reading R. Grishman, Computational Linguistics: an Introduction. |
|
