The Arabic natural language is one of the most sophisticated natural languages, and is very rich in morphology and vocabulary. It is classified as a morphological language, compared to English natural language which is synthesis language. Sophistication in Arabic natural language is due to morphology, syntax and semantic. So this paper consider different grammar formalisms, different algorithms for representation of a sample of Arabic natural language sentences categories and comparison among them based on the results obtained from the implementation of their computerized model for each algorithm such as the response time of execution of each model, as a common parameter.
There is a variety of grammar formalisms used to obtain the formal specification of the structure of Arabic natural language such as recursive transition networks, context free grammar, logic grammars ect… we will concentrate here on the last two grammatical formalisms. A variety of parsing algorithms may be used for each grammar formalism, we will use top down depth first, bottom up (left corner technique of implementation), modified bottom up (using linking table), mixed mode parser algorithm (Chart parsing). In each algorithm we measure the response time of their execution, compare among them, also state their advantages, disadvantages from the technical point of view and give the recommended approach for Arabic natural language representation.