often represented by the French ill; as in cornicula-corneille, apicula-abeille, auricula-oreille, ovicula-ouaille,-the Latin u probably disappearing in familiar pronunciation-cornícla.

But the Latin li or le is at once the most obvious and the most usual equivalent for the French ill, Ital. gl; malleusmaglio-maillet; milium-miglio-millet; Julius-Luglio-Juillet; melior-migliore-meilleur ; filia-figlia-fille; valeam-vaglia-vaille; palea-paglia-paille, &c.

The change of gender cannot of itself be a reason for supposing a departure from the rule. Such changes from the Latin are frequent in the Romance languages, above all in French. Thus the masc. flos becomes the French fem. fleur, with feminine equivalents likewise in Spanish and Portuguese. The whole class nearly of Latin masculine nouns in or, not denoting persons (with a very few exceptions, such as labor, honor-s), become feminine in French, calor-la chaleur, pallor-la páleur, color-la couleur, rumor-la rumeur, pavor-la peur, sudor-la sueur, candor-la candeur. The exceptional Latin feminine arbor becomes, on the contrary, the French masculine arbre. And the Latin hora becomes eventually masculine in the compounds le bonheur, le malheur. It will be seen at once from how narrow a circle all the above instances are chosen.

I should indeed be fully prepared to admit that the persistent feminine termination of aiguille-aguglia, agulha-aguja, indicates an intermediate aculea at some time or other. Such a form may indeed have been original, though not preserved by any author, and may have underlived in vulgar use the correct aculeus; just as olea for oil, occurring, I believe, only in Catullus, must certainly have underlived in Gaul the received oleum, to reappear at last in the French fem. huile (arch. huille)-a "Nebenform," as the Germans say, to aiguille. But at any rate, to the barbarian invader the masc. aculeus, from the fem. acus (masc. only as applied to a particular kind of fish), must have seemed the anomaly, and he would naturally have replaced it by aculea.

For it must never be forgotten that what is termed the corruption of a language under the influence of a new race,—

especially of a conquering one,—in reality only clothes it with new life. The conqueror is sure to bring with him two tendencies, one disruptive, the other restorative, which in the long run, if left to work undisturbed by new invasions, nearly countervail each other. He has on the one hand a tendency to think his own modes of thought into every other language; he has on the other an unfailing human instinct of symmetry and regularity, which will often stamp itself upon the language of the conquered in exact proportion to his own ignorance and insolence.

There is no mystery about these tendencies; we may see them at work any day around us. The irregularities of a language are always what we learn last. A Frenchman, speaking English, will say sheeps, gooses, I comed, he bringed. He will think his own idioms into English, saying of a manher mouth (sa bouche); just as I have heard Englishmen in French use the fem. sa in speaking of any number of masculine objects belonging to their wives or daughters; and an English lady even, probably much preoccupied with the need of asserting her sex, ask, amid the tittering of shop-boys, "Où est ma gant?" or a German persist in employing the French amour (die Liebe) as a feminine; or an Italian couleur (il colore) as a masculine. A more curious process indeed sometimes takes place. We think into the new language, not the idioms of our own, but that of some other foreign language with which we are acquainted. A friend of mine, a good Italian scholar, talks French with more fluency than correctness by simply Frenchifying his Italian; la fumée, for instance, is for him le fume (il fumo). If we had to learn foreign languages, not as visitors in foreign countries, but as brawny Teuton conquerors, with small fear of God or man, lording it over hosts of Latinized Gaulish or Italian slaves, who could doubt but that we should be prone to treat with supreme indifference the formal irregularities, the idiomatic peculiarities, of the language of the conquered?

Now, what attaches to the Romance languages so deeply * The feminine is still poetic-especially in the plural, amours."

66 mes chères

human an interest is, that in them we can discern, more clearly, I think, than in any other, the working of these tendencies,their "glottomorphic" influence, if I may coin the word. We have a considerable, though not a perfect, knowledge of the Latin groundwork of those languages; whilst the Teutonic influence which mainly built them up out of it, lives to this day in a family of languages, many of them, on the whole, singularly true as yet to their origin. Thus the German alone is sufficient to explain almost any instance above adduced of change of gender in the French. Nadel would alone explain the French feminine aiguille; Blume the French feminine fleur; Baum the French masculine arbre. The whole family of physical qualities and conditions are feminine in German (Wärme, Farbe, &c.), with the single exception of Schweiss, sweat, la sueur. This French feminine may however admit of a double explanation. On the one hand, as soon as it began to be the rule that French nouns in eur, not denoting persons, were feminine, there would be a constant tendency to eliminate exceptions. On the other hand, the existence of the French name Le sueur* clearly

* I need hardly point out the value of sirnames as perpetuating archaic forms, spellings, and designations. Thus, in French, Monnier, Monier, Lemonnier, mark at least one stage in the long journey from molendinarius to meunier. Letellier indicates a time when tela must have been tele and not toile. Granier stands between granum and grainier. The very common Lefebvre, Lefèvre, Lefeuvre, lead us through Fabvre, Favre, Fabre, to faber, and explain orfèvre (auri-faber). Compare our Smith" and "Goldsmith." Leverrier would now be Le-vitrier; Ladvocat, L'-avocat ; Lhuillier, L'-huilier, had he not long since sold his stock to the " épicier," as Tavernier to the "aubergiste." Carpentier stands between charpentier and our "carpenter"; sabatier is nearer than sabotier or savetier to the Span. zapatero. Couturier has survived men-milliners, Cottier has outlived the cotte (coat); Levasseur, Levassor, Levavasseur, remain to testify of the age of vassals and vavasours, as Larcher and Fléchier of that of the "bowyers" and "fletchers,”—the former still lingering with us as a London Livery, under the shade of Gog and Magog. Lherminier, Lerminier, indicate a time when the ermine, abundant still in European forests, gave a name to the furrier's trade.

So with us, Boucher, Boutcher, give the pedigree of butcher; Boteler helps that of butler; Taylor, Tayleur are older forms of tailor; Dyster is, perhaps, the fem. of dyer; Baxter (not A.S. bacestre), Bagster, of bagman ;

implies at one time a tautologous sueur, from sutor (found in Roquefort), essentially masculine, which would lead to the differencing sueur, sweat, by the gender.

At the same time, that last singular tendency to which I have alluded, of introducing the terms and forms of one foreign tongue into another, may account for many words of distant origin which turn up in every language-like boulders carried far from their original seat, and which are too often supposed to indicate the actual presence in the land at some time or other of the race to which they originally belonged. The Hun on his long journey from the Chinese frontier, if he took the trouble to learn a certain number of strange words from the nearest tribe whose territory he crossed, would certainly not take that of changing them from region to region— least of all those which first expressed for him some new object; and thus by the time he reached the Rhine, though his language would more or less be a chart of the whole course of his wanderings, yet the most distant stages would probably be the most clearly marked.

The lesson from all which is, never to forget the human element in philology; and whilst acknowledging that words have their own special laws as well as chemic substances or plants, to remember also that those laws are themselves subordinate to the laws of the human minds by which the words are used, and to human wills, which now seem to work with, and now to counterwork, all other laws whatsoever.

Note. The circumstances of the barbarian conquest may serve to explain many other Romance peculiarities. For instance, I have expressed the opinion, that the three words manducare, minuere, and mendicare respectively, all tended ultimately to the form manger or menger, but that the two latter remained differenced from it in manquer and mendier, for the avoidance of tautology. But why did manger win the race? Evidently because words expressing bodily wants, such as eating, would be the first which the

Hoste preceded host. Leach, Leech, show that the early surgeons left a numerous progeny. Crozier recalls the day when to sell crosses was a man's livelihood. Bowman and Spearman have survived their occupation. The French chantre lingers among us in Chanter, veneur in Venour; whilst the Saxon Chapman stands at bay in the bankruptcy court, his last refuge.

66

conqueror would get hold of, the first to take new shape at his hands. And why did mendier keep so close to its original? Evidently because begging" could not be, for centuries, the concern of the conquering stock, but that of the conquered, in whose hands the old form would remain much longer unaltered.

May I here point out a blunder which has crept into a former paper of mine, inserted in the Society's volume for 1854, p. 130, line 4 from bottom, where égarer is explained "to be away from the gare"? It should be "to lead away," the verb being purely an active one. Another error of the press, p. 128, line 17 from bottom, "l'apprêt l'une étoffe,”—for "d'une," corrects itself.

I may also add, that the words "vassor" and " vasseur," treated in that paper as hypothetical, appear from Roquefort and the Supplement to the "Dictionnaire de l'Académie" to be real.

III.-ON THE TERMINATION OF THE NUMERALS ELEVEN, TWELVE, AND THE EQUIVALENT FORMS IN LITHUANIAN. BY HENSLEIGH WEDGWOOD, Esq.

[Read February the 5th.]

At a late meeting of the Society I indicated the Finnish lika, excess, overplus, as the origin of the final element in Lith. wieno-lika, dwylika, which are themselves generally recognized as identical with Goth. ain-liba, twa-liba, eleven, twelve. It being thought, however, by some members that the identity of this element with Sanskr. dasa, Gr. Seka, had been so conclusively established by Bopp as hardly to leave the question open to argument, I am desirous of supporting my position by further analogies, and a balance of the probabilities in favour of the two derivations.

"After what has been stated," says Bopp, i. 439, "I think no one can any longer doubt that in our eilf and zwölf, strange as it may at first glance appear, a word is contained expressing the number ten, and identical in its origin with dasan, Seкa, and ten." The theory opposed to that of Bopp has hitherto not had fair play, it being supposed that Lith. wieno-lika, dwylika, and Goth. ain-liba, twa-liba were derived in the one case from a root lik, and in the other from a root