over, biðla, to woo, from biðja, rugla, to turn upside down, from rugga, &c.; and the argument derived from these rests upon two main propositions: first, that the idea of frequency, when it leads to any secondary notion at all, leads to that of pettiness and contempt; and secondly, that these derivative verbs are so closely connected with the corresponding nouns, that the meaning of the verb-termination once determined, that of the substantival suffix follows as a matter of course. Now with regard to the first proposition, I demur to the theory which deduces the idea of diminution from that of frequency, as a general proposition, although I do not doubt that in certain cases of verbs the frequentative termination has engendered the notion of pettiness or contemptibility. There are many instances of frequentative verbs enhancing and intensifying the significations of their primitives instead of diminishing them. In Sanskrit the frequentative form of the verb, where it is not confined to the simple idea of repetition, is always intensive*: thus from 'dip,' to shine, comes 'dedipya,' to shine brightly; from 'subh,' to be beautiful, 'sosubhya,' to be very beautiful; from 'rud,' to weep, 'rorudya,' to weep excessively. In Greek the frequentative forms of the verbs, which after the analogy of the Sanskrit are formed by reduplication, most certainly convey an intensive rather than a diminutive sense, as may be seen by Homer's constant use of such forms as παμφαίνω, παιφάσσω, πορφύρω, ποιφύζω, μαιμάω, ἀμαιμάκετος, ἀλαλκομένηϊς, δαιδάλλω, &c. And in Latin the frequentative flagito outweighs in intensity all its rivals postulo, peto, rogo and posco: compare Cic. de Legg. i. 5. 'Postulatur a te vel flagitatur potius historia;' agito and excito equally surpass ago and excieo, and minitor in like manner is explained by Forcellinus as equivalent to ‘valde minor.' And Ramshorn, in his Lateinische Grammatik, § 84. ii. 5, expresses himself thus: "-To, -ito, -so, -xo, endigen die frequentativa oder iterativa, die ein oft wiederholtes, oder doch ein mit Eifer und Beharrlichkeit verbundnes Handeln, auch ein Pflegen bedeuten." It is clear therefore that the utmost that can be maintained is, that the idea of frequency * Wilson's Grammar, Art. 212.

may in certain cases give rise to diminutives, not that the origin of diminutives in general can be accounted for in this

manner.

With regard to the second branch of the argument, reference is frequently made to the analogy of the Latin for support. That language possesses in addition to its regular frequentatives in -ito a class of verbs, termed in the grammars indifferently frequentatives and diminutives, in -illo, -ulo, -olo, although in the two latter forms the force of the termination is generally insensible. These last verbs correspond to the noun diminutives in -illus or -ulus and -olus, and in fact in most cases are formed directly from them. The argument therefore is, that as a precise correspondence exists between the terminal forms of murmurillum and murmurillo, sorbilum and sorbillo, jaculum and jaculor, lutulum and lutulo, &c., we are justified in assuming a similar correspondence between the significations of the English termination le, when applied to verbs and substantives respectively. The answer is, I think, not difficult. It is not fair to argue from a language containing a class of nominal verbs precisely corresponding in all points with their originating nouns, to another where the verbs in question are much more probably formed independently from the primitive verb than from the diminutive noun. It does not follow because nidulus and nidulor are merely the counterparts of each other, that we can therefore deduce from the pettiness implied in dribble the existence of diminutiveness in fennel or nozzle, because the formation of Teutonic frequentatives is in most instances direct from the principal verb, and not, as in Latin verbs in -illo, -ulo, and -olo, directly from the substantive. The conclusion may possibly be true in certain instances as a matter of fact, but it cannot be assumed as a basis for general reasoning. An example will illustrate this: Grimm*, under the head of Verbal Diminution, gives three instances from the Old Norse of frequentatives in -ka, -blíðka, placare, íčka, colere, and tíčka, solere, to which we may add seinka, to delay, and reika, to wander up and down (a frequentative like our dawdle); and a hypothesis might be raised in *D. G. vol. iii. P. 689.

accordance with the principles I am discussing, that the termination -ka in Old Norse had a diminutive meaning in nouns, fortified as such a supposition could be by the well-known parallel of the Teutonic diminutive in k. But what says Grimm himself a few pages previously*: "Die altnordische feminina blaðka folium, harka durities, grænka viror, haben keine diminutiv-bedeutung;" and the same conclusion would be arrived at if we had compared the larger class of Scandinavian frequentatives in -sa with the nouns in -si. The only real parallel in our language to the Latin, although the order of formation is just the converse, is that furnished by such verbal substantives, or perhaps more correctly speaking, nominalized verbs as 'jumble,' 'rumble,' 'jingle,' 'rattle,' 'muddle,' &c., which we often use as nouns, and which indisputably carry with them the notions of frequency, diminutiveness, or contempt, already impressed upon the originating verbs.

This much for arguments, à posteriori, arising from the criticism and analysis of words as they exist, and the comparison of them with their former selves; let us, however, now in conclusion briefly look at the subject from another point of view. It will probably be conceded that the only channels through which a distinctively Teutonic form would be likely to have entered and become incorporated with our language, are the Anglo-Saxon and the Norse, and the latter only in case such form were common to both the Teutonic and Scandinavian groups of languages. Neglecting the insignificant admixture of Keltic in English, which after all is confined to isolated words, nothing that is integral in the language, and which is not classical, can have entered it through other than one of these two sources. If therefore we find that neither in A.S. nor in the O.N. is the l active as a diminutive, but that it has simply left traces of its existence without vitality or influence, it is difficult to understand how that which was dead and unfelt in the parent tongues can have been warmed into a second life in the offspring, vigorous and luxuriant though it be. As to the Anglo-Saxon, we have Grimm enunciating, and Rask by his silence acquiescing in, the * D.G. vol. iii. p. 676, n.

statement that l as a diminutive is merely 'Ein ungefühltes, nach welchem keine lebendigen diminutiven gebildet werden*, and I need therefore add nothing of my own to such authority. The opinion of the former of these philologists is the same on this point with regard to the Old Norse; but I find it stated in Dasent's Transl. of Rask's Icelandic Grammar (art. 326) that ill forms a kind of diminutive in that language. Now, in the present state of Norse Lexicography, to make anything approaching to a complete collection of words ending in this syllable would be simply impossible, but Grimm has got together about 80†, to which I have added between 20 and 30 more, making a total of upwards of 100. Out of this number 12 have, among other meanings, been appropriated as names of Vikings, such as pvinnill, Ekkill, Endill, Gestill, &c., while of the remainder I do not find that more than the following are stated to possess a diminutive signification:

kistill, a little chest, from kista, a chest. knýtill, a little bundle, from knýtí

which have the same meaning.

I can cite no authority by which to try the significations of the first four of these words (one of which, kistill, is evidently a mere Latinism), beyond Haldorson's Lexicon; but as to the last, although vöndr and vandill both coexist, and the latter is of course derived in some way from the former, so far is it from appearing in actual use as a diminutive, that it occurs in various poetical periphrases, where a diminutive would be obviously out of keeping with the bombastic nature of the circumlocution; thus in Viga Glum's Saga, c. xxvi. str. 1, we have Með veðr-stöfum Viðris vandils for 'with the columns of the storm of the spear (or rod) of Odin,' which in simple English means 'with warriors.' Again, the great falchion of Arinbjorn, which possessed a mythical history of its own, and was given by him to Egill, the son of Skallagrim, is called Dragvandill in the history of the latter (c. 64), and the Skaldskáparmal gives Angrvandill, the rod of anguish,' as a general †D. G. ii. p. 110.

* D. G. iii. p. 675.

I

synonym for sverðr. I think it will be admitted that even a mediæval chronicler would hardly have described Coeur-deLion's two-handed sword as a Gladiolus or an Ensiculus, except in a spirit of jocose irony, which the Northmen do not seem to have adopted in naming their weapons. Skarpheðinn, the son of Njal, was certainly in a sterner mood when he christened his battle-axe Rimmu-gýgr, 'the giantess of contention.' And I cannot help believing that the few real Norse diminutives in this termination which I have cited, may have been originated by the action of the Teutonic forms upon the Scandinavian, in comparatively recent times, after a more familiar intercourse than anciently existed had been established between the two nations, and the Old Norse was being gradually restricted as a spoken language to Iceland. The argument in fact may be clinched thus: if in Old Norse ill really has a diminutive force, and such a word as 'litill,' the derivative of the older litr, be adduced in confirmation of such a view, how is ' mikill,' great, a precisely analogous form, to be explained; or how do the ideas of grandeur and magnitude conveyed by such words as pengill, leader or king, prifill, a strenuous man, ristill, a giantess (poet.), and hypill, a large garment, square with such a hypothesis? I have no doubt that other examples might be produced, but it is almost useless to attempt any complete investigation of such a subject as this, until we are in possession of a Norse Lexicon more adapted to the exigencies of philology at the present day than that of B. Haldorson, which, although an admirable book, considering the state of Norse literature at the time when it was compiled, mixes up the ancient and modern tongue without distinction, and but very rarely cites authority.

The general conclusions arrived at are therefore the following:

I. That let forms two suffixes, one composite, and the other simple.

II. That the former is made up of el and et, and is attached almost invariably to words of classical, i. e. Latin or French origin.

III. That the simple let contains the root of the adjective