Noctes Atticae

What the Greeks understand by a)nalogi/a, and, on the contrary, by a)nwmali/a.

IN the Latin language, just as in Greek, some have thought that the principle of a)nalogi/a should be followed, others that of a)nwmali/a. is the similar inflection of similar words, which some call in Latin proportio, or

regularity.

)Anwmali/a is irregularity in inflection, following usage. Now two distinguished Greek grammarians, Aristarchus and Crates, defended with the utmost vigour, the one analogy, the other anomaly. The eighth book of Marcus Varro's treatise On the Latin Language, dedicated to Cicero, maintains [*](viii, p. 146, G. & S. ) that no regard is paid to regularity, and points out that in almost all words usage rules.

As when we decline,

says he,

lpus lupi, probus probi, but lepus leporis; again, paro paravi and lavo lavi, pungo pupugi, tundo tutudi and pingo pinxi. And although,

he continues, "from ceno and prandeo and poto we form cenatus sum, pransus sum and potus sum, [*](That is, pransus, potus and cenatus are used in an active sense; see Cic. pro Mil. 56, adde inscitiam pransi, poti, oscitantis ducis, and Priscian (ii. 665. 17, Keil) ut cenatus sum . . pro cenavi.) yet from destringor and extergeor and lavor we make destrinxi and extersi and lavi. Furthermore, although from Oscus, Tuscus and Graecus we derive the adverbs Osce, Tusce and Graece, yet from Gallus and Maurus we have Gallice and Maurice; also from probus probe, from doctus docte, but from rarus there is no adverb rare, but some say raro, others rarenter." [*](Charisius (i. 217. 8, Keil), cites rare from Cicero, Cato and Plautus, but the modern texts do not admit the form.) In the same book Varro goes on to say:

No one uses

v1.p.211

senior and that form by itself is naught, but almost everyone says adsentior. Sisenna alone used to say adsentio (I agree) in the senate, but later many followed his example, yet could not prevail over usage.

But Varro himself in other books wrote a good deal in defence of analogy. Therefore his utterances on the subject are, as it were, common-places, [*](Haec argumenta quae transferri in multas causas possunt locos communes noininamus. Cic. De Inv. ii. 48; cf. Brut. 46. and Quintilian passim.) to cite now against analogy and again also in its favour.