On the Latin Language

...from the Antiquities. It will be taught more accurately below that Verrius—from whom Varro drew in grammatical matters—also depended upon these, although his methods were otherwise entirely different from Varronian ones. It is consistent that Remmius Palaemon, the leader in treating Latin grammar with a new method, read these books, even though we lack certain traces of Varro in the fragments. Indeed, one would be hard-pressed to say whether Varro devised or merely propagated the grammatical terms 1 through which he agrees with him, having used terms long since invented. It seems uncertain whether the words of Suetonius (p. 217)—by which we are taught that he called him a "pig"—pertain to these books or to something else. This epithet is conjectured by Norden (Ancient Artistic Prose, p. 195) to denote a rough style of speaking. Perhaps the things Quintilian presents not without ridicule regarding Varro's etymologies in Book V (Institutes of Oratory I 6, 37; cf. testimonies at 12, 6), like other things already established in the same book, return to his teacher Remmius Palaemon, although Spengel, in that commentary which we shall soon use frequently (p. 469), assumed that Quintilian himself had read Varro. What Columella reports as Varronian should be referred to the Antiquities, and to the traces of Dionysius of Halicarnassus and Plutarch, to which they are certainly for the most part to be recalled. How Valerius Probus used Varro's books we do not know; the same must be said of Pliny, although certain testimonies of later grammarians will not rashly be traced back to Pliny or Probus as the primary author. More certain traces are found in Gellius, whom we have no reason to doubt read Varro's books themselves. We have used Gellian readings more than once in correcting Varro's words (cf. 78, 1 ff.; 190, 29; and elsewhere). Grammarians and scholiasts of the fourth and fifth centuries depend on older sources; but Priscian (p. 52, 2 ff. = GL. III 410) or his author drew from Varro himself, as no one else preserved a larger fragment. By his testimony, the errors of the Florentine codex were frequently able to be corrected (see p. XXIII, XXIV bis). Isidore, who mentions Varro in his Etymologies in not a few places, transcribed most things from Servius's commentary, and we do not think it probable that any of the passages whose source lies hidden were drawn from Varro himself. Haupt's opinion (Hermes vol. I, p. 44) that Rhabanus read Varro in On Computation ch. 3 is false, as those words should be attributed to Priscian, as Spengel correctly noted in Preface 2, p. II, note. What Osbern presents is owed partly to Isidore, partly to Priscian, and partly to Macrobius. It is entirely certain that no writer of the Middle Ages drew from Varro himself, except that in the eleventh century someone propagated the Florentine codex, and in the following century...

1 Cf. L. Jeep, "On the History of the Doctrine of Parts of Speech," p. 12 ff.