how the calendar was managed by a Sabine priesthood, such a result seems almost inevitable. The old Latin numerals are likely thenceforward to have been called Oscan.

As regards the Greek which has entered the Latin language, a very delicate problem is presented to us. So consistent and positive are the accounts of Greek colonies on the coast of Latium, before the historical age; so clearly Greek is the worship of Hercules 36 by the hereditary priesthood of Potitii and Pinarii at the Ara Maxima; so distinct is Pliny's testimony of a town called Antipolis on the Janiculum, when another called Saturnia was on the site of Rome;-so manifest is the powerful influence of Greece on Southern Etruria;-that I cannot doubt there was an Hellenic element in the oldest Roman population. From these may have come a positive importation of words into the Siculian. On the other hand, many words which the Greeks and Latins have in common, might properly be said to have been given by the Latins to the Greeks, viz., by such Siculian tribes as in their westward movement fell short of Italy. In fact, if the Pelasgians were Siculians, they must no doubt have imported to the Hellenic no inconsiderable dash of Latin words. At the same time these very words may have come to Greeks and Latins alike from a northern people, say Celtic. For instance, hos, аλáμη; levis, palma, must be identified with the Welsh, Llae, an expanse, Llaw Gaelic Lamh, a hand; which shew us the common source. Bouxoλos is the Gaelic Buachaille, cowherd; Erse, Buachail; which O'Brien derives from Cal, to keep safe and there seems no doubt that the word is not formed in the Greek. Túpavvoc also, and 'Qxsavòs, are probably Celtic. In fact, as the law of movement was from north to south, and from east to west, it is reasonable ceteris paribus to look to the north as the origin, except where we find the same words in Asia also.

Learned Greeks flattered their own national pride in teaching the Romans that their language was derived from the Hellenic; and to this day they have propagated the unfounded notion, that whatever words are both Greek and Latin, must necessarily have come from Greece into Latium. Of this preju

36 Why Niebuhr and others are so incredulous about this, I cannot imagine. Haec sacra Romulus una ex omnibus

peregrina suscepit, Liv. i. 7. There is no greediness here to refer every thing to the Greeks.

dice we must disabuse ourselves, if we are to trace the currents of language correctly.

F. W. NEWMAN.

POSTSCRIPT.

Since the above was out of hand, I have been able to collate O'Brien's Irish Dictionary, which suggests the following additions:

The Erse has the form Teallúr (earth), nearer to the Latin than Talamh: also Galia, a helmet, Astas or Astal, a spear; [Welsh, Aseth, a sharp stick; Asethol, pointed like a dart, cf. Hasta, Hastile ;] Tall, a cooper's adze; Talla-im, I cut; whence not only Tall, theft, robbery, despoiling, but Teól, a thief; Teallam, to steal; Talca, bravery; Talam, feats of arms. In Talla, I think, we plainly see the Celtic root of the Fr. Tailler and of the Anglo-Saxon Till (Qata', which in Arabic means to cut, in Berber is to fight, rob, or steal; so in Erse, Gur, sharp, also means brave: cf. öğùs, Ooos.) It is then not improbable that Talamh, (in Erse, Talamh, Tealla, Teallach, Teallur), is derived from tilling; and the English Steal also is seen in Teallam. There are also allied roots Tola-im, to pierce through; Tollam, to bore or penetrate.

Gall, a cock or a swan, is Erse; although the language likewise has Coileach, a cock. This justifies my regarding the two words as unconnected. Carr (quiris, a spear) and Curadh (warrior, Quirite) are also Erse. As the correlatives of Jus and Jur, the language exhibits Deas, order, propriety; Dior, meet, proper, decent; Diorach, equitable, &c. As for Lex, the Erse has only Dligh-im, I separate (pivw); Dlighe, Dleacht, law; Dligheach, lawful. Either the similarity of Dlighe with Lex and Lagh is accidental, or we must judge Dlighe parent of Lex, since we know of no tendency to change L into Dl. In Gaelic, the root Dligh means to owe, not to separate.

Catapulta, (which seems to have been only borrowed by the Greeks, -thing and name,) may be fairly explained from the Erse as Cattapalt, battle-sling; from Cat, battle, Tabhal, sling. Cf. Caterva. An arrow in Erse is Saighead, Saighiot, or Sciot; parents alike of the English Shot and of the Latin Sagitta? Milk is Laith, Lachd, Bleacht, (Gael. Bliocht; with Bleagh, to milk;) all which, with yadakтa, aμeλyw, bleach, and Xevкos, may be referred to a Welsh (?) root Llug or Lluc, whiteness and brightness. Not a step of the process is defective. Fionns is the Erse for a fountain, nearer to Fons than Fynnon is; if indeed the final s of Fionns does not move suspicion that the word is ecclesiastical. Eoin is a bird, which shews that the termination -wvos has not been added to oiwvòs within the limits of the Greek language. Indeed Eun, gen. Eoin, is Gaelic likewise.

Remarkable in the Erse is the group Torc, a hog (W. twrch), More or Muc, a pig; (W. Moch, swine), Pórc, a pig, Orc, a young pig. Torc and Morc remind one of Tata and Mama. At any rate, this family seems native to the Irish.

The three words Porcus, Aper, Verres, are probably modifications of the same root; which will denote confusion of tongues as certainly as Lawful, Loyal, and Legal. There is Lat. Porcus, W. Porch, A.S. Berga or Bearh, gen. Bearges, Lat. Verres, Eng. Barrow-pig, Boar, Dutch Beer, Germ. Eber, Lat. Aper. As Beer to Aper, so is Bee to Apis. F. W. N.

XXV.

ON THE RELATION BETWEEN THE CONSONANTAL SYSTEMS OF THE ENGLISH AND SANSCRIT LANGUAGES.'

PART I.

AFTER all that has been done, in comparative philology, for several languages of the Gothic stock, particularly in shewing their relations to the oldest and well-known branch of the whole Indo-European family, the Sanscrit, or simply to the Greek or Latin languages, and in determining the laws or rules, according to which those characteristic changes of sounds, more especially of consonantal sounds, take place from the older to the younger branches,—it is still an interesting study to undertake a separate comparison of the English with the Sanscrit. The English language will thus obtain the same advantage which, for instance, the Gothic and the High-German have had; it can be shewn, without any intermediate comparison of related languages, in what shape the Sanscrit presents itself in English forms, what changes have taken place in the roots, after three thousand years, and after wandering from the Ganges and Indus to the Thames and Forth, and by what laws those

1 The mode of expressing Sanscrit letters in this Essay is according to Wilson's Sanscrit Grammar, 2d Ed. 1847. The expressions Anlaut ("on-sound," i. e. "beginning-sound," letter or sound at the beginning of a word,) Inlaut, ("in-sound," letter or sound between two other letters of the same word,) Auslaut, ("out-sound," i. e. "endsound," letter or sound at the end of a

word,) belong to Grimm's Terminology. I have retained them; they are short, expressive, and, says Chev. Bunsen, in his discourse read before the Ethnological section of the British Association for the Advancement of Science, at Oxford, June 1847-(See the Report of that Association, p. 262,)--they have been found of decided use in the remotest parts of Ethnological inquiry.

changes are regulated. If it were necessary to prove the relationship of the English to the Sanscrit, we should be obliged to begin by establishing the identity of origin of their grammatical forms, structure, and construction; this relationship is now undoubted and fully proved. We, therefore, shall give our whole attention to the secondary inquiry, which relates to the material of the two languages, the roots and primitive words. In these roots, the consonants are the principal vehicles of their meaning. Our inquiry must therefore chiefly relate to the consonantism of the two languages.

It is generally taken for granted, that the English stands, in its consonantal structure, on the same level with the Gothic,2 and that those rules which, in this respect, relate to the Gothic, may with equal force be applied to the English, in determining its relations and position among the whole tribe of the IndoEuropean languages. This conclusion may be considered true, when we speak in general, without distinguishing languages of the same stock, or the consonants of the different organs. But it is far from true, when one particular language, e. gr. the English alone, is held up for comparison with the Sanscrit, and when the consonants of each organ are considered separately. What, in a general consideration of a great subject may appear an exception, may be the rule in some particular section of it. J. Grimm himself acknowledges this to be the fact in our case. It is but right to remember that he never intended to give, in his grammar, a detailed account of the consonantal relations of the language we are speaking of. His own words are (1. 590), die lautverschiebung folgt in der Masse, thut sich aber nie im einzelnen rein ab (the transformation of consonantal sounds is only observable en masse, but not so in all particular cases.) He then mentions several classes of exceptions, as Lat. dis, and Goth. dis, do: Lat. dies, and Goth. dags; Lat. sedula, and Old Sax. sëdel; Gr. títoŋ, A. S. tit. But these exceptions really corroborate his law. For the Lat. dis has its regular corresponding t in other languages of the Gothic stage, as in Saxon tô, Engl. to, which, in its last regular development, becomes

"Grimm's well known law stands thus, (D. Gr. I. 584,) according to the shorter table,

High-German zu, zer. In this case we see the English keeps to the law for the Gothic stage even more strictly than the Gothic itself. As for Lat. dies, the English does not present the required form "tag," but it is not more irregular than the Gothic dags. In the third example, the O. Sax. sedel is irregular, as is also the Old German sëdel; the Goth. and E. sitts and seat, and the Nhg. seszel, agree perfectly with Grimm's law. In Gr. títoŋ; A.S. tit and E. teat, Ohg. tutto; the tenuis in anlaut has remained unchanged in all three stages, and it is only in the Nhg. zitze, that organic life is visible again.

But let us take the consonants of another organ, the Labials for instance. There Gothic and English p ought to correspond with a Sanscrit, Greek, or Latin b. When J. Grimm attempts to prove this by examples, he feels himself compelled to remark, (1. 585) "für den anlaut weisz ich keinen beleg, zur bestärkung meiner ansicht" (as to the anlaut, I do not know any example to corroborate my law.) For the inlaut he gives xávvaßıç, cannabis; Oldn. hanpr, (Engl. hemp); Ohg. hanaf, Nhg. Hanf); and three other examples which still seem doubtful to him, and which, therefore, cannot prove the rule; yet might corroborate it, if it were proved otherwise: Lat. turba, Goth. thaurp (AS. and Engl. thorpe); Ohg. dorof (Nhg. Dorf); Lat. stabulum, Oldn. stöpull; (AS. stapel, ol, ul; Engl. staple); Ohg. staphol, al. (Nhg. Staffel, Stapel);-Lat. labi; Goth. hlaupan (Engl. leap); Ohg. loufan (Nhg. Laufen). The fact is, that the Sanscrit, which we shall prefer now to take as the representative of the classic languages or the first stage, has not one single b in anlaut, that would develope itself into a Gothic or Engl. p, and Hg. f; but all such Sanscrit b, though few in number, remain unchanged throughout the three stages. From this particularity of the labial media, and from other examples, like those above given by J. Grimm as exceptions, Graff3 draws a rule very different from Grimm's, viz., that the Sanscrit media remains unchanged throughout all the cognate languages. And moreover, while Grimm (1. 558,) pronounces those words to be of suspicious connexion and organization, in which only one consonant follows his law, and the other differs from it, Graff, (1. Pref. xiv.) by his own independent researches, has come, in opposition to Grimm and Bopp, to the conclusion, that in perfectly organic formations one cannot conclude from a

Althochdeutscher Sprachschatz, vol. I. pref. ix. Berlin, 1834.