a parallel instance of the equivalence of the diphthongs. The etymology from "juger" is, I think, only supported by the word "gaugir" in Roquefort, which he explains by "juger." Jaille, jale, as a measure, might also put in a claim to paternity.

From what I have said, it may be concluded

1st. That the labials b, p, v, followed by i, become frequently represented by the soft French g.

2nd. That the same occurs sometimes when they are followed by e and a vowel, as in cavea, cage (caveola) cajoler; alveus, auge, rubeus, rouge.

3rd. That the labials are represented by g soft in some cases where they are not followed by i, or e and a vowel; as in plumbare, plonger, volvere, bouger, serva, serge, charrub, carouge, ævum, áge; and lastly,

4th. That the same mutation, when occurring initially, has a tendency to run, in North French and English especially, into the hard g; as in palet, jalet, galet; pal, jalon, gale; vaccaria, jachère, gachère; piatta, jatte, gatte.

Of the above four classes of cases, I am inclined to make a marked distinction between the first two and the last two. In the first two, I believe the labial simply disappears; it is the last two only which seem to me (one instance excepted) to exhibit a true mutation.

The principle governing the first class, and subsidiarily the second, is that of the equivalence of i and j, or in other words, of the optional consonant sound of the Latin i before a vowel. I must say I can entertain no doubt as to this fact, any more than as to that of the similar equivalence of u and v, to both of which every old vocabulary bears witness, and which can be traced with the utmost facility throughout the Romance languages: in the i sound of the Italian j; in the correspondence of io-je; not to speak of "āriete crebro," and the like metrical forms, which, I would venture to say, none can overlook who is not determined to do so. Rabies would then, I take it, be rabjes, rabje; and as the labials have a constant tendency in French to disappear, at least in sound, as soon as brought before any other consonant but a liquid

(doubte, doute, compter, conter, &c.), the soft guttural alone would remain. So in one instance of the 4th class of cases; piatta, (pjatta), jatte. Great light is thrown upon this formation by the archaic spelling of chambger or chambgier for cambiar; in which evidently the g represents the Italian i; and again by such words as simia, singe, sturio, esturgeon, somniare, somnium, songer, songe, in which no other equivalent for the g is imaginable.

I can see only one argument of any apparent weight against this view, which is, the existence, in the French archaic forms,—in neige, and in such forms from other languages as raiva, nieve, lieve, &c.,—of an i before the last consonant, which seems like the Latin one transposed. But such instances as Pierre, Pietro, from Petrus, show this at once to be a fallacy. Equally fallacious would be the opposite argument from the i after the g in the archaic forms abrégier, enraigier, légier, allégier, or the Ital. saggio, piccione, leggiero, alleggiar, roggio, &c. That the latter French archaic i belongs to the consonant, and has no substantive value, is clear, from its occurring at the same period both after the g soft and the ch, in such words as étrangier from straniero, marchié from mercatum, &c. Had it not been inserted, I suppose either g or ch would at this time have been sounded hard. The case may be different in Italian, where the forms in gg have mostly beside them synonyms closer to the Latin,-savio, pippione, lieve, alleviar, robbio; compare also aggia with abbia; so that the i takes more the shape of a permanent element in the word. At the same time-it is needless to discuss the point -I should still be of opinion, that the gg represents mainly the i consonant, and that the i following belongs vocally to the gg, so that alleggiar is allevjar with the v left out, or replaced by, not changed into, the first g.

I have said that the same principle, of the consonant value of the Latin i before a vowel, governs subsidiarily the second class of cases, those in which a short e occurs in the Latin. For this has a most evident tendency to run into i, and thence into j or g soft, as in cavea, Ital. gabbia; rubeus, Sp. rubio; habeam, It. abbia; extraneus, Sp. estraño, estranjero,

Ital. strano, stranio, straniero, Fr. étrange, étranger (arch. estrange, estrangier), Eng. strange, stranger; granea, grange; cereus, cierge; aculeus, Ital. aguglia, Sp. aguja, Port. agulha, Fr. aiguille*; the Ital. gl, Sp. ñ, Fr. ll, and Port. lh, all containing a sound of i, and answering to it etymologically.

There will remain, therefore, as I said, only the last two classes of instances (jatte excepted) in which a true mutation of the labials into g soft or j takes place†.

In conclusion, I cannot help pointing out that practical French philology, so far as I know, remains almost an untrodden field. French philologers are found for the most part looking only to the ultimate roots of their language. Unless

* Diez denies this last etymology, which seems to me most evident; preferring, as I understand Mr. Key does, acicula from acus. But Diez himself derives aiguillon from aculeus. See Postscript.

+ I have confined myself in this paper to the soft g. Its next of kin, the French ch (sh), seldom behaves in this manner; except in sepia, sèche, arch. seiche; coucher, seemingly from cubare; and compare also roche, rocher with rupes; sapiam, sapiens with sache, sachant. Mr. Key in his Alphabet refers to this case.

It may perhaps be of some use if I give here a list of the different Latin (and sometimes Greek) equivalents of the French g soft and j, so far as I have observed them :

g:-virga, verge; gemere, gémir; and often with a contraction, as navigare, nager, pagina, page. (The same as to the Teutonic or Italian g or gi: jars, gans, gander; jambe, gamba; jaune, giallo (giallino).)

jor i before a vowel :—juvenis, jeune; jacens, gisant; sturio, esturgeon. hy or hi-hyacinthus, jacinthe; Hierosolyma, Jerusalem; hyoscyamus, jusquiame.

e before a vowel :-extraneus, étrange.

c (or Greek x or κ):—cibarius, gibier; caryophyllum, gérofle, girofle; Kavbos, jante; chamadrys, germandrée (rare).

d, de, di, dec, dic, duc:-raídiov, page; adusque, jusque; viridarium, verger; sedes, siège; hordeum, orge; diurnalis, journal; judex, juge; vindicare, venger; manducare, manger.

ic or ric-carricare, charger; fabrica, forge; filix (filicaria), fougère (rare).

t:-polentarius, boulanger (solitary instance?).

p, v, bi, be, pi, vi, ve :—see instances above.

s:—salire, jaillir or saillir,-compare also clergé through Span. cleresia from clericus (very rare).

z :—zinziberis, gingembre; zelosus, jaloux; zizyphum, jujube.

Kelto-maniacs, they are never satisfied till they have reached backward to Arabic, Pehlvi, or Sanskrit; just as their predecessors two hundred years ago would always jump to the Greek or Hebrew root. Of the gradual development of their own language,-of its relations with its sisters of the Romance family, they tell us little. Whilst, if French is treated of by foreigners, it is commonly only incidentally as one member of that family. I think, however, it will be found,—as soon as the "Dictionnaire de l'Académie" is fairly thrown aside, and the language is viewed as one historic whole, and not as the dialect of a particular age,—that French is by far the most remarkable and peculiar of any of the Romance tongues, and deserves to be the object of special study. The student of modern history cannot fail to recognize that France, since the downfall of the Roman empire, has really been the intellectual centre of continental Europe, and that its vernacular literature has flourished continuously for a longer period than any kindred one; whilst its intellectual supremacy prepared for that vernacular literature a richer soil than anywhere else, in the development of the scholastic Latin of the middle ages, itself, if once looked at in its reality, a very flexible, expressive, and even characteristic language.

The Romanism and despotism of the "grand siècle" have indeed put a check upon the development of the French language, from which it will probably never recover. The French of Rabelais is as fresh as the English of Chaucer himself, though so much later, and is far richer and riper. But compare the French of Racine with the English of Milton, even in his latest poems; it is like a full-dressed doll by the side of a finely-modelled statue. To speculate upon the causes which might have prevented this drying-up of the French language under the parching influences of court patronage, and of that worst of pedants "le bon goût," would lead me too far into historic considerations. Suffice it to say, that it is only in our own days that it has recovered a little of its freshness from reverting to its well-springs, whether in its early literature, or in the homely, racy, speech of its peasants. In the latter respect, no one has done so much for its regeneration

as Georges Sand, especially in those three tales which are known in France as her "idyls," La Mare au Diable,' François le Champi,' and 'La Petite Fadette.'

But I would also point out, that in studying French philology, the group of early English derivatives from the French should be carefully noticed. The graft from the Romance upon the Saxon stock which they represent, is a much earlier shoot than the one which has since fruited in the mother country, and many wholly obsolete French archaisms will be found amongst us in full vitality, such as our specially English comfort, from confortare, now only surviving over the water in reconforter, and bacon, Old French bacon.

Not only is this the case as to words, but as to pronunciation also. Our broad a, so peculiarly English, is largely explained by the Old French forms raige, saige, which I have above instanced, and further by, I presume, the whole list of words in age, Mid. Lat. agium, as homage, hommage, arch. hommaige, homagium; courage, Eng., Fr. arch. couraige, coragium. Palace, palais, afford a similar illustration. In all these cases, I suspect it is the French pronunciation which has altered, not the English. Another such instance is afforded by our cream, certainly nearer in sound to the Greek xpíopa than the French créme*, or in the sense of anointing oil, chrême, arch. cresme, chresme, crisme, chrisme. Possibly, as the s lost its sound, the vowel was changed, by religious scruple, to avoid amphibology with crime. The same religious feeling may have helped to turn criste, from crista, into creste, crête, our crest; though there are other cases in which the change of vowel occurs without any moral reason to account for it.

POSTSCRIPT.

Aiguille,-whether from acicula-acucula, or from aculeus? I am far from denying that cul in the Latin diminutives is * Only imagine the dictionaries deriving crême from cremare! Devonshire cream is certainly scalded; and hot milk and even custards are certainly called crême in French cookery. But it is equally certain that the word applies to cream uncooked in any way, and the circumflex accent might have led men to suspect a suppressed consonant. Xpíopa, it is well known, applies to any fat substance which may be used for anointing.