The science of language is a science comprising grammar, rhetoric, and poetry. It is therefore fitting that its definition should be given in a manner that harmonizes with these three branches, and afterwards each of them should be differentiated by its specific characteristic. I say that the science of language is a science which teaches the general methods that may be employed in a language, in a manner fitting the conventionalities of that language. By saying: ‘In a manner fitting’ I wish to convey that the expression must agree with that which is conceived by the mind, according to the linguistic axioms. I call the study of language a science, because the term ‘science’ is indeed more comprehensive than the philosophers think. For they have defined science as an acquisition which the mind attains by true beginnings and causes; by ‘true beginnings’ they mean the beginnings which are known either through the nature of the intellect, as the first principles, or through sense perception, or through experience. According to this, the study of language is no science at all, for its beginnings are conventional, and are not evolved through any of the ways that I have mentioned. In our opinion, however, the term ‘science’ is more comprehensive, for we call science anything that is known through investigation, analogy, or proof, no matter whether the principles and beginnings upon which it is based are essentially true, or arbitrary and conventional. The teacher[226] has already called such sciences ‘conventional sciences.’ Accordingly, the study of language is a science, and the definition of science rightly applies to it; for by it are known, through investigation and proof, the results derived from principles and beginnings which have been laid down in it. The wise man R. Jonah said in the description of grammar that it signifies investigation and searching. This is merely an interpretation of the term, which does not give us the definition of grammar itself.