Hekatæus, Herodotus, and Thucydidês,[368] all believed that there had been an ante-Hellenic period, when different languages, mutually unintelligible, were spoken between Mount Olympus and Cape Malea. However this may be, during the historical times the Greek language was universal throughout these limits,—branching out, however, into a great variety of dialects, which were roughly classified by later literary men into Ionic, Doric, Æolic, and Attic. But the classification presents a semblance of regularity, which in point of fact does not seem to have been realized; each town, each smaller subdivision of the Hellenic name, having peculiarities of dialect belonging to itself. Now the lettered men who framed the quadruple division took notice chiefly, if not exclusively, of the written dialects,—those which had been ennobled by poets or other authors; the mere spoken idioms were for the most part neglected.[369] That there was no such thing as one Ionic dialect in the speech of the people called Ionic Greek, we know from the indisputable testimony of Herodotus,[370] who tells us that there were four capital varieties of speech among the twelve Asiatic towns especially known as Ionic. Of course, the varieties would have been much more numerous if he had given us the impressions of his ear in Eubœa, the Cyclades, Massalia, Rhegium, and Olbia,—all numbered as Greeks and as Ionians. The Ionic dialect of the grammarians was an extract from Homer, Hekatæus, Herodotus, Hippocrates, etc.; to what living speech it made the nearest approach, amidst those divergences which the historian has made known to us, we cannot tell. Sapphô and Alkæus in Lesbos, Myrtis and Korinna in Bœotia, were the great sources of reference for the Lesbian and Bœotian varieties of the Æolic dialect,—of which there was a third variety, untouched by the poets, in Thessaly.[371] The analogy between the different manifestations of Doric and Æolic, as well as that between the Doric generally and the Æolic generally, contrasted with the Attic, is only to be taken as rough and approximative. [368] Hekatæ. Fragm. 356, ed. Klausen: compare Strabo, vii. p. 321; Herod. i. 57; Thucyd. i. 3,—κατὰ πόλεις τε, ~ὅσοι ἀλλήλων συνίεσαν~, etc. [369] “Antiqui grammatici eas tantum dialectos spectabant, quibus scriptores usi essent: ceteras, quæ non vigebant nisi in ore populi, non notabant.” (Ahrens, De Dialecto Æolicâ, p. 2.) The same has been the case, to a great degree, even in the linguistic researches of modern times, though printing now affords such increased facility for the registration of popular dialects. [370] Herod. i. 142. [371] Respecting the three varieties of the Æolic dialect, differing considerably from each other, see the valuable work of Ahrens, De Dial. Æol. sect. 2, 32, 50.