gone stale," cf. ef-fa-tus), which signifies "nasty," referring, in the first instance, to the smell, and, by a natural transition, to whatever is disagreeable: thus we speak of "a nasty accident," &c. Verbs in I are derived from nouns of the I declension. we have Thus mollire vestire = When we seem to have an exception to this rule, we can always find, on looking into the question, that the crude form of the noun, from which the verb in i is derived, does involve this letter. Thus we have sepire from sepe, which is really an i noun; punire is from pœna, but the Greek Town=πoi-vy-a, and the adjective impuni-s, show that the form ends in i; moliri comes from moles mole-is; sortiri from sors=sor-ti-s, gen. pl. sorti-um; and blandiri is referred to blandus, which is really the participle of bl[a]o=fl[a]o, "to breathe or blow gently" (cf. μαλακός, μαλθακός, &c.); such phrases as blandus prece vel hostia, "soothing with prayer," or "sacrifice" (Hor. Ep. XI. 1, 135. Carm. III. 23, 18), whence we have blanda preces (id. Carm. IV. 1, 8. A. P. 395), still retain the participial meaning; and this is presumed in the adverb blanditer (Plaut. Asin. I. 3, 69), so that the true form is blan-ti-s, whence bland-i-ri. Verbs in u, when this amounts to u-ya, are derived from nouns in u. Thus we have This may be regarded as a singular case; for no contraction is possible in a derivative verb of this kind. § 11. B. Composition. Discrimination of Compound Words. The proper distinction between a compound word and the apparently compounded form consists in the fact, that the former is an union of two or more words, of which the last only is inflected, so that the preceding crude forms remain in a construct or subordinate state, whereas the mere juxta-position, or = apparently compounded term, is made up of separable elements, the inflexions of which are retained. Thus in such words as magnanimus, ædifico, we have entirely new compounds; for the former is an adjective made up from the ablative of quality, so that magnanimus is qui magno animo est; and the latter is a derivative from a compound adjective ædificus, which involves the whole predication ædem facio. On the other hand, the composition is only apparent in res-publica, "the commonwealth,” jusjurandum, "an oath," juris-peritus, "a lawyer," animadverto = animum adverto, "to pay attention to," "to take strict notice of," "to punish," &c. That these are not compounds, but merely juxta-positions of separable elements, is clear from the fact that, in those which are in direct agreement, both parts may be inflected throughout, as rei-publicæ, jure-jurando, and all may be separated by particles, as in res vero publica, juris legumque peritus. There is no doubt, however, that these parathetic structures may pass into regular compounds, in the course of long usage. Thus from the phrase sesque for as semisque, one and a half," we have the compound sesquipes, “a foot and a half," and its derivative adjective sesquipedalis. Again, when the first part of a real compound is an indeclinable word, it may be separated by a tmesis from the inflected part of the compound; thus we have inque salutatus for insalutatusque, and per mihi mirum videtur for permirum. In such forms as nihilo-minus, dum-taxat, ut-pote, vide-sis, sodes=si-audes, scilicet, &c., the two words are merely written in continuity to show their hasty utterance in the flow of conversation. Sometimes it requires a careful analysis to prove that the word is really a compound. Thus annus or anus seems at first sight to be necessarily a simple word; but it is proved by philological dissection (p. 163) to be a shortened form of de-vós = deì veóμενος (cf. οὐρανός and ὠκεανός, according to the old notion of a wide superincumbent firmament, and a swift stream flowing round the earth), and the idea attached to the word is that which is expressed in Virgil's lines (Georg. II. 401): Redit agricolis labor actus in orbem, Atque in se sua per vestigia volvitur annus. Then again it is an etymological discovery that præs, custos, opu-lentus, vio-lentus, &c., are not merely derivative forms, but real compounds (above, pp. 298, 393); and the same remark applies to the verbals in -bilis and -bundus, which involve the verb of becoming (fio), and are not to be explained, like the derivatives in -bulum, as vena-bulum, by a mere reference to the pronominal formations. § 12. Classification of Latin Compounds. If we consider the Latin language only, we may conveniently distribute all the compound words into four classes. (a) Determinative compounds are when the first part of the word defines the second; such are the prepositional compounds: cognomen, dedecus, interrex, semideus, injuria, nefas, consul, collega, pronepos, &c., where the prefix qualifies the meaning of the whole word. In some cases the meaning is defined by an involved epithet, as in: cav-ædium, lati-clavis, lati-fundium, quatri-duum, &c. = (b) Syntactical compounds are when the first word is governed by the second, whether the regimen is that of a noun dependent on another noun, as in galli-cinium, "the crowing of a cock," opu-lentus, "loaded with wealth," stilli-cidium, “a falling of drops;" or, what is much more common, that of an accusative case governed by a verb, as in: agri-cola=qui agrum colit, brevi-loquens breviter loquens, male-dicus = qui maledicit, signi-fer qui signum fert; and in the verbs derived from such compounds, whether the intervening noun is still extant or not; as: æqui-paro = æquum paro, castigo-castum ago, purgo = purum ago, &c. To the same class belongs aurigo from auriga or aureax = qui aureas agit, according to Festus (p. 8): aureax, auriga. Aureas enim dicebant frenum quod ad aures equorum religabatur; oreas quo ora coercebantur" (cf. pp. 27, 4, 182, 23). If this interpretation is not sufficient, we must consider the aures or avruyes of the chariot as referred to in the compound; for as this term is applied to the side-pieces of the plough, which Virgil terms a currus1 (Georg. I. 174), it may have been also a designation for something corresponding to 66 1 Modern editors read cursus, but it is difficult to see why cursus should be applicable to a plough, when the same word with an assimilation is considered inapplicable. It appears to me that the secondary word is more suitable to the metaphor than the direct verbal. Besides, it is clear from the verb torqueo that the plough itself, not its motion, is here alluded to in "currus a tergo torqueat imos.” these side-pieces in the wheeled vehicle. I may remark, in passing, that the oriel window, in Gothic architecture, was undoubtedly so called from its projecting like the human ear from the side of a building. The old spelling shows this. Thus we find in an ancient MS.: "The Lords always eat in Gothick Halls, at the high table or oreille (which is a little room at the upper end of the hall where stands a table,) with the folks at the side tables;" in accordance with which we find in Matthew of Paris (ap. Ducang. s. v.): "ut non in infirmaria, sed seorsim in oriolo, monachi infirmi carnem comederent." Now it is well known that oreille is a representative of auriculus. So that the oriolum or "oriel" is the "ear-window" or projecting chamber used for privacy and retirement. (c) Auxiliary compounds are when two verbs come together, and the second helps the former either in a predication of time, or by introducing a modification of meaning or reference; thus we have: ama-vi = amare-fui, ven-do-venum do, ven-eo = venum eo, arcesso = accedere sino, treme-facio tremere facio, &c.; and to the same class belong all the tenses in -bam and -bo, -vi and -veram. = (d) Possessive compounds are when the first part denotes the manner, in which the thing, denoted by the last word, is possessed by the subject, to which the whole compound is referred either as predicate or epithet; thus we have: aheno-barbus, alti-sonus, crassi-pes, magn-animus, in which the first part is a declinable word; and affinis, concors, nefastus, beneficus, inermis, bimaris, elinguis, in which the first part is an uninflected particle: in both cases the possessive adjective, constituted by the whole compound, involves a determinative compound, which is made moveable, so as to agree with different substantives. Among these nouns, we must take care to distinguish between those in -ceps from caput, as bi-ceps, gen. bicipit-is, and the syntactical compounds involving -ceps from capio, as muni-cep-s, gen. muni-cip-is, &c. Although this classification of the compounds is sufficient for all practical purposes, so far as the Latin language alone is concerned', it is convenient, with a view to comparative philology, 1 Livy remarks incidentally (XXVII. 11) that the Latin language was inferior to the Greek in the power of forming compound words. to inquire how far these composite formations admit of arrangement in accordance with the system of the Sanscrit grammarians. As I have compared the six classes of the samâsa with the Greek compounds (New Crat. § 309), and as Bopp has subsequently adapted this arrangement to his more general purposes (Vergl. Gramm. pp. 1427, foll. VI. Abtheil. 1852), it may be as well to place the Latin formations under these heads. The six classes of the Sanscrit samása are designated by names some of which describe and others exemplify the nature of their construction; and they are arranged by Vopadeva in the following order: (1) The first are described by the term dvandva, i. e. "two and two," "pair," or "doubling," and consist of mere aggregations of words which might be written separately and joined by a copulative conjunction, as agnê-çómâu, “Agni and Soma," in the dual; brâhmana-kshatriya-viť'-çûdras, the four Indian castes, in the plural; &c. ; (2) the second are exemplified and named by the compound bahu-vrihi, "that which has much rice," and therefore consist of compound epithets; (3) the third are called karma-dhâraya, “ that which comprehends (dhârayati) the object (karma)," and include such words as mahá-rájah, "a great king," where a substantive is defined by an uninflected epithet prefixed; (4) the fourth, exemplified by tat-purusha, "the man of him," comprises compounds formed of two or more nouns, the first set being in some oblique case governed by the last, which may be a substantive, adjective, or participle in -ta, as raja-purushah, "the king's man;" (5) the fifth, called dvigu from dvi, "two," contains compounds of which the first part is a numeral and the second a noun, as chatur-yuga-m, "the four ages of the world;" (6) the last class is called avyayî-bhava, or "adverbial," and is made up of indeclinable words, the first part being some particle, and the last a noun in the neuter gender, as a-san çaya-m, "without doubt," ati-mâtra-m, "over the measure." It appears from this enumeration that classes (3) and (5) are determinative, class (4) is syntactical, class (2) is pos sessive, and class (1) is merely an aggregation of terms. The following examples will suffice, so far as the Latin language is concerned. (1) There are no Latin dvandva, unless we recognise such a form in su-ovi-taurilia = suile+ovile + taurile. But the Latin language, especially in its oldest form, abounds in examples of |