From the "Transactions of the N.Z. Institute," Vol. XIX., 1886
Art. LXVII.—A Note on Latin Place-names.
In Livy, xxi., 19, we read, "turn maxime Sagunto exciia;" further on, in xxi., 21. we read, "Sagunto capto." The first expression is explained per synesin of "urbe" vith Sagunlum, and the participle is taken in agreement with it. Livy occasionally introduces urbem, vicum, in apposition to the names of towns, in "urn." Hence has arisen a certain perplexity as to the gender of Latin place-names; add to which the influence of Greek place-names, and we have the erroneous statement of our Latin Grammars on this point seemingly justified. But Livy, in using such a sentence as the folloving (among a host of such instances), ii., 63: "Fusi, in primo proelio hostes, et in urbem Antium, ut turn res erant putentissimam acti," is telling us that the enemy fled to Antiun— a town of very great wealth, as the times were then—and uses the plainest way of saying what he has to tell us.
In our Latin Grammars, (two books of this year, 1886, are enough to cite,) the statement runs substantially thus : "Nanes of countries, cities, islands, and trees are feminine." In another Grammar the statement is somewhat guarded: "Most nanes of cities are feminine." Here is a qualification of the previous statement; and it is to be hoped that in time the statement will be further attenuated, so as to represent the facts.
What are the facts? In my copy of Madvig's Gramnar (third edition, an old book), p. 28, the author says very little about the subject; but adds, "of the words in us the names of towns are feminine. These mines are all Greek." The itaics are mine; and the statement is worth noting, because it indicates the natural order of things: that, in the case of one highly-inflected language passing on names into another highly-inflected language the names hear their gender with them. All these Latinized spellings of Greek place-names only go to show that in Greek the names of towns in os are feminine.
But in his "Notes on Latin Word-systems," published in 1844, this great scholar (who has died since this note was can-piled,) goes further: "Not a single name of a place in Lain, irrespective of the nature of its termination, is of the feminne gender." Notwithstanding which dogma of the master, compilers of Latin Grammars for English boys have gone on reiterating the same misleading "rule" with a sort of hide-bound obstinacy.page 508
We find place-names declined according to the scheme of the first, second, and third declensions. I am not acquainted with any belonging to the fourth and fifth declensions, and am inclined to think that as geographical names usually belong to rough speech, these somewhat obscure varieties of declension do not contain any place-names.
Taking suffixes in order, we begin with
All feminine, as the terminations require.
ii.—ii or i.
And, by analogy, Pompeii, together with numerous tribal names, of which, in the case of towns, the suffix ii is a survival. These are masculine words.
—with many others. These are all neuter.
iv.—a (of the plural).
—Greek names, but neuter, as their suffix requires.
There are no Latin place-names with this suffix, which is native, however, to Greek, and brings with it its gender; even in the case of variants,—as e.g., Canopus, Isthmus, Orchomenus, Pontus,—names masculine in Greek are masculine in Latin.
vi.—o (gen. -onis).
—all masculine, as the suffix requires.
—all neuter, as the suffix requires. (But Arelate, a Greek word of 1st declension, is feminine.)
—neuter, as the suffix requires. (Anxur, the mountain, is masculine by analogy with the usual gender of the names of mountains.)
—neuter or masculine. (The indeclinable words are neuter.)
In all the cases quoted above we note that the suffix determines the gender of the place-name; the "rule" is not even traceable. There is, e.g., a well-known suffix -onis, and another -inis. The former is masculine, the latter feminine: hence Narbo -onis is masculine (Narbo Martius), and Carthago -inis is feminine (Carthago Nova).
And even such hybrids as Augustodunum and Cæsarodunun. All these words are neuter.
But the suffix um, or ium, is freely used to reduce to the Latin scheme a very large number of words found amoig subject tribes:—
|(All in Britain)|
—besides words like Trajectum, Durotrajectum, and many others, all neuter, as the suffix requires.page 510
What becomes of the "rule"? As Zumpt seems to have felt, it is so overwhelmed with exceptions that mole ruit earum. Having examined three hundred and fifty place-names, found chiefly in the western section of the Orbis Romanus, I am not able to discern any "rule" applicable to the names of towns. But the influence of the "rule" is very great. Even Lewis and Short, s. v., are misled by it. In order to justify Liv., xxi., 19, cited above, they allege that Liv. used Sazuntus. But Saguntum is in good prose the only form used, cf. Mayor on Juv., xv., 114. Poets and writers like Mela and Florus use Sazuntus. Juv., loc. cit., uses Zazynthus, a thinly-veiled form of Zacynthus.