Natural language processing

To ability to communicate in some kind of natural languages, be it Turkish or Tagalog, often seems to be the hallmark of the human race. Computers will not be able to perform many of the tasks people do every day, too, share the ability to use language. Although three-year-old children, who cannot play even a legal game of chess, much less beat an expert, can speak and understand their native language, we have still not produced a computer program whose overall linguistic performance rivals that of people. Processing natural language is hard. It requires both linguistic knowledge of the particular language being used and world knowledge relating to the topic being discussed. But first of all the computer must have a sufficient knowledge base about the natural language that will be processed. This knowledge base consist of word, the functions of words, the structures which we used when making a sentence, and the types of correct sentences. After storing these information in a database, the first phase is the analyzing of the input, second one is semantic analyzing and the last pragmatically analysis of the sentence. The collection of these phases is called "natural language processing". In this thesis, firstly I look at the definition of intelligence and artificial intelligence, then by using these concepts I study on the linguistic knowledge that is necessary for "understanding", just as at how world knowledge can be combined to produce an effective language understanding program. After then, by using the PROLOG language, I developed a language parser for Turkish. I applied this program to 180 couplets of Yunus Emre and determined the distribution of the source and functions of the words used by the poet.