maîtresse. The word, however, is an exceptional one; and, for practical purposes, the combination -tr- may be treated as accidental. The main fact connected with the words in -tress, is that their analysis is -t-r-ess, their origin the -tricis, -tricem, &c. in words like genitricis, &c., and their masculine -tor-tor- with an o, as auctor, actor; which in French becomes eu-auteur, acteur.

But the -r-, as a sign of agency, is English as well as Latin. However, the English termination is -er-never -or. We say fact-or rather than fact-er; but bak-er rather than bak-or.

All these and many others can be added-give us both English roots and English affixes; to which it may be added that the root is a verb. It is a verb, even where it looks most like a noun; as in harp-er, hatt-er, glov-er, where harp, hat, and glove = play on harp, make hats, make gloves. It is a verb and an English verb. Let, however, the verb in question be of foreign origin, yet treated as if it were English. In this case we get words like governor, which are neither English nor French.

§ 286. Hybridism, and the inaccuracies of spelling to which it leads, are the chief points that command our attention with Feminines in -ess, and their corresponding Masculines. The minor details are of less importance. Duch-ess, count-ess, baron-ess, peer-ess, poet-ess, lion

ess. Here -ess is attached, at once, to the main word, and the idea is that of a state, or condition, rather than action.

Empress.-Here one of the r's in Emper-or is omitted. Emperor itself, however, is an anomalous word. The Latin is Imperator. Has the -t been lost? Or is the word an improper formation from empire? This is a point of French, rather than English, philology. Meanwhile, Imperatrice is direct from Imperatrix.

§ 287. The Masculine, in respect to form, is not always the correlative of the Feminine-thus Marquis will not give Marchioness, which comes from the Low Latin Marchio.

$288. In seam-str-ess and song-str-ess we find instances of hybridism, and something more. At present, however, it is enough to say that they are treated according to the analogy of master and mistress.

§ 289. Individually, I consider that hybridism is a malum per se, and that it ought to be discouraged; though, at the same time, I must admit that it is, sometimes, all but necessary; and also that some hybrids are better than others. When this is the case there is generally some combination of sounds which makes the word look more unilingual than it really is. In witticism (for instance) we have so close a parallel to criticism that the same analogy appears to apply to both. The classical scholar knows that it does not. He also knows that w is an impossible initial in a Greek word. Still, the word is better than many others. Again, let an English Verb end in -t. Let -er be added. Let a Feminine in -ess be required. The result will be a regular form in -tress. Hence, such a word as waitress (though beginning with w) is better than foxess, or sheepess.

§ 290. The diminutive termination -et.-Add -et to lance, and the result is lanc-et small lance-a legitimate form,

=

=

because both the root and the affix are French. Add -et to sword, and the result (sword-et little sword, is a specimen of hybridity. Still, there are many of these hybrid words which keep their ground, especially when the -et is preceded by l, as in streamlet.

§ 291. Words like penetra-ble and penetra-bility.-These are not only possible, but actual Latin words-penetrabilis, penetrabilitas. So are possible and possibility. So are legible and legibility. But readable and bearable, with their opposites, un-readable and un-bearable, are hybrid, and (to say the least) exceptionable.

§ 292. The terminations -ize, -ist, and -ism.-These are Greek, and in words like ostracize and ostracism they find a fit and proper place. In words of English origin they are exceptionable.

§ 293. Individually (to repeat what has been already stated), I consider that hybridism is a malum per se. It is often difficult, however, to avoid it. Many scientific terms err in this respect: exhibiting the heterogeneous juxta-position of more than one language. Nor is this, in all cases, an accident. Occasionally it occurs through inadvertency: occasionally, however, it is defended. In a few cases it is the lesser of two evils. It is least blameworthy in words. like the ones just quoted; words ending in -ize. It would be difficult to dispense with such words as moralize, civilize, and some others: however much the former part may be Latin, and however much the latter part may be Greek. Again-to words like botanic, where the ic (like the botan-) is Greek, we may add the Latin -al. As such a word was possible in the Lower Empire, where such words as zgwτovoτágios were common, we may call these (after the fashion of the architects) Byzantine formations. This, however, is only naming our tools. The mixture remains the same. At the same time one of the conditions required in the introduction of new words is

complied with. There exists a language in which they are possible. Generally, however, the actual occurrence of the whole word is impossible. Part comes from Language A part from Language B: whilst in Language C, they are tacked together-something (as in words like botanic-al-ly) with additions.

§ 294. Change of accent.-A change of accent converts a Noun into a Verb. Walker has referred this

to the action of the Participle.

§ 295. Diminutives.-Compared with lamb, man, and hill, the words lambkin, mannikin, and hillock, convey the idea of smallness or diminution. In hillock there is the simple expression of comparative smallness in size. In doggie and lassie the addition of the ie makes the word not so much a Diminutive as a term of tenderness. The idea of smallness, accompanied, perhaps, with that of neatness, generally carries with it that of approbation. Clean in English, means, in German, little = kleine. The feeling of protection, which is extended to small objects, engenders the notion of endearment.* Sometimes, a Diminutive is a term of disparagement; as lordling and hireling.

§ 296. Next to knowing that in some Diminutives there is something more, it is useful to know that in others

As klein is to clean in German and English, so is petitus (sought, desired) to petit (small) in Latin and French.