3. The Arabs trace their grammatical science, like so many other things, to Ali, to whom is ascribed even Aristotle’s tripartite division of speech. In reality the study began to be cultivated in Basra and Kufa. Its earliest development is involved in obscurity, for in the Grammar of Sibawaih († 786) we have a finished system,—a colossal work—, which, like Ibn Sina’s Canon of Medicine in after times, could only be explained by later generations as the production of many scholars working in collaboration. We are but ill-informed even on the points of difference between the schools of Basra and Kufa. The Basra grammarians, like the school of Bagdad in subsequent times, must have conceded a great influence to Qiyas (Analogy) in the determination of grammatical phenomena, while those of Kufa allowed many idiomatic forms which diverged from Qiyas. On this ground, to mark the contrast between the Basra grammarians and those of Kufa, the former were called ‘the Logic people’. Their terminology differed in detail from that of the Kufa school. Many, whose heads had been turned by logic, in the opinion of the genuine Arabs, must have gone decidedly too far in their captious criticism of the language; but on the other side caprice was raised to the position of rule. It was from no mere accident that the school of Basra was the first to avail itself of logical resources. Generally speaking, it was at Basra that the influence of philosophic doctrines first appeared, and among its grammarians were to be found many Shiʻites and Mutazilites, who readily permitted foreign wisdom to influence their doctrinal teaching. 4. Grammatical science, in so far as it was not confined, to the collecting of Examples, Synonyms &c., when so determined by the subjects specially treated, was affected by the Aristotelian Logic. Even before the Muslim era, Syrians and Persians had studied the treatise περὶ ἑρμηνείας, with Stoic and Neo-Platonic additions. Ibn al-Moqaffa, who at first was intimate with the grammarian Khalil (v. infra), then made accessible to the Arabs all that existed in Pahlawi of a grammatical or logical nature. In conformity therewith the various kinds of Sentences were enumerated,—at one time five, at another eight or nine, as well as the three parts of speech,—Noun, Verb and Particle. Afterwards some scholars, like Djahiz, included syllogistic figures among the Rhetorical figures; and in later representations there was much disputation about Sound and Idea. The question was discussed whether language is the result of ordinance or a product of nature; but gradually the philosophic view preponderated, that it came by ordinance.