and especially in participles and adjectives (above, p. 301). This is particularly the case with the forms in nt-s or nti-s, and we may compare the affix -tis or -tus, in pes-tis, "a destroying," ves-tis, "a covering," po-tus, "a drinking," spiri-tus, "a breathing," with the Greek nouns in -σs, -Tis, and -Tús, as πρᾶκ-σις, “a doing”=πρακτύς, φάτις, “a speaking,” &c. Το the same class we must refer the participial adjectives in -dus, as cupi-dus cupient-s, candi-dus-candens, &c. (New Crat. § 265). The nouns in which the termination Fa assumes the form v, are much less numerous in Latin than in Greek. We have, however, the following: al-vus, ar-vum, cal-vus, cer-vus, ci-vis, cla-vis, cla-vus, cor-vus, cur-vus, eq-vus, fla-vus, ful-vus, fur-vus, gnavus, lae-vus, ner-vus, par-vus, pra-vus, sæ-vus, sal-vus, ser-vus, tor-vus, vul-va. If we compare cer-vus, cla-vis, cur-vus, gnavus, la-vus, with the Greek ké-pa-Fos, кλñƑis, yûp-Fos, yevvaîos, λa-ios, we shall see that the v in the former cases corresponds to a digamma in the terminations of the latter; ner-vus and par-vus compared with veupov and Taupos suggest the possibility of a metathesis in the latter analogous to that in ἐλαύνω for ἐλα-νύω; ci-vis compared with the Oscan ce-u8 brings us back to the root ke (above, p. 125); tor-vus contains the same root as τορ-ός, ταρ-βεῖν, ταῦρος, trux, trucido, trucu-lentus, and we must assign ă-trox to the same class, the initial being one of those prefixes, which we find in a-výp, d-oτýp, &c.; and eq-vus compared with the Sanscrit aç-va refers us to the root âç-u, "swift," Greek wkús, Latin acer. All the words in -vus, which have been mentioned, join this termination immediately to the root; but in some few, to which incidental allusion has been made above (pp. 146, 195), the v immediately follows an r; thus from the roots ac- and cat-, both signifying "sharp," we have the derivatives ac-er-vus, cat-er-va denoting a pointed, pyramidal heap, or a crowd following its leaders. Similarly, we have Min-er-va, from the root min-, "to think," and in the Arvalian chant we find lu-er-ve[m] for luem. In these instances we may suppose that the affix -v- is attached to a lengthened crude form, just as hones-tas, onus-tus, tempes-tas, venus-tas, involve something more than the mere root of the word, and there need be no doubt that the r in the cases just cited is merely the usual substitute for an original s. The termination -vus is also appended to certain adjectives derived from and

expressing the verbal abstraction; thus we have ac-ti-vus from the supine ac-tum (ac-teFos), by the side of ac-tio, na-ti-vus from na-tum (na-te Fos), by the side of na-tio, &c.; and we know that the verb-forms in -se, -tum are the same in effect with the abstract nouns in -sis, -tis, -Tús=TEFOs, &c. (New Crat. §§ 254, 410 (3)).

§ 6. (c) Forms with the third Pronominal Element only.

The most common forms, under which the third element appears as an affix, are t- and n-. The former must be carefully distinguished from the participial ending in -dus, and those nouns in -tus, -tús, which, as has just been mentioned, belong to a formation connected with the second pronominal element: the latter must be identified with the endings in and including the diminutives in -lus, -leus (New Crat. § 266). For the original identity between t and n it may be sufficient to refer to such forms as re-ple-tus compared with ple-nus, cas-tus compared with câ-nus, &c. The simple use of the third element, under the forms t or n occasionally softened into s=r, is confined to those neuter nouns which express a mere object, as capu-t, cor-pu-s; or to those nouns and participles which express a thing done, or the passive result of an action, as digi-tus, "that which is pointed," compared with dignus, "that which is shown," liber-tus, "he who is freed," compared with dô-num, "that which is given," lec-tus, "that which is gathered" (of leaves and the like), compared with lig-num, "that which is bound" (of a faggot), and the forms in -nus and -tus mentioned above; to which may be added fa-num compared with fa-tum, and reg-num, pug-na, pen-na, pœ-na, va-nus and pa-nis, compared with rec-tum, cris-ta, lacer-ta, vi-ta, sagit-ta, &c. The identity of the affixes -tum and -num with -rum and -lum may be seen by comparing Béλo-s=ßéλo-7, "a thing cast or thrown" (jac-tum), with jacu-lum, do-num with dw-pov, sac-rum with ȧy-vór, and tem-p-lum with réμe-vos, whether as Téμevos ailépos (Æsch. Pers. 357), or as the portion of land cut off and set apart for divine uses.

§ 7. (a) Terminations compounded of the first and other Pronominal Elements.

Many nouns exhibit in their affix a combination of the first element with the third, under the form m+n, which is often

strengthened by a repetition of the objective affix under the form t, so that the whole affix is m+nt. Of nouns in -môn corresponding to the Greek nouns in -uwv, we have only three, ser-mó[n], pul-mó[n], te-mó[n], which may be compared with γνωμών, πλεύμων, &c.; we have also a limited number of nouns in -mnus, corresponding to the Greek passive participle in -μevos, such as auctu-mnus, da-mnum, vertu-mnus, alu-mnus, colu-mna, æru-mna, &c. A comparison of alu-mnus, "the person nourished," with al-mus, "the nourisher," shows that the combination m+n completes the agency and carries it on to

the object acted on. As in Greek we have -μovn=-μo-v-ia by

the side of -uwv, so in Latin we find an extended termination -monia, in such words as acri-mōn-ia, ægri-mōn-ia, ali-mōn-ia, cere-mōn-ia, casti-mōn-ia, parsi-mōn-ia, sancti-mōn-ia, all of which express a quality or abstraction inferred from an act done. The force of these words is best shown by a comparison between these and the nouns in m+nt, which have a repetition of the third element instead of an addition of the second. These words, which agree with the Greek neuters in -μaт=-μevт, either omit the final t, as in car-men, cri-men, legu-men, stra-men (above, p. 299), or, which is more common, exhibit the lengthened form -mentum, as in ali-mentum, ar-mentum, arma-mentum, aug-mentum, orna-mentum, pul-mentum. Now all these words express an action proceeding from the subject (m), but become objective (n), and exhibited in its results (t). Thus car-men[t] = ποιή=Sanscrit kar-man means a thing made," with (μεντespecial reference to the maker. But cere-monia, which contains the same root (cere-, cre-, kri-), calls attention by its affix to the doing or process. Similarly, al-mus is "a nourisher," alu-mnus, "a person nourished," ali-mentum, "a thing for nourishing," but ali-monia, "the process of nourishing."

ματ

§ 8. (B) Terminations compounded of the second and other

Pronominal Elements.

Of these combinations the most usual and important are the forms in which the second element, vocalized into i, is prefixed to the third element with or without further extensions. A very large class of nouns end in -io[n], and express, if one may say so, a fixed or consolidated abstraction. These nouns, which are

always of verbal origin when the noun is feminine, are formed either by affixing -io[n] to the simple stem, as in leg-io, opin-io, reg-io, rellig-io, and this is always the form in the masculine nouns, as cur-io, centur-io, scip-io, &c.: or by adopting the t=s of the supine, as in man-sio, ses-sio, con-fu-sio, quæs-tio, col-lec-tio, dis-trac-tio, dubita-tio, pulsa-tio, con-jura-tio, positio, ad-moni-tio, erudi-tio, &c. The masculine nouns generally denote a person or thing belonging to that from which the noun is derived; thus cur-io[n] is the man of the cur-ia, centur-io[n] the man of the centur-ia, &c. And as the genitive ended originally in -ion=-sion, we must consider these nouns in -iōn=ion-s as extensions of the genitive case. The same explanation will apply to the nouns in -o[n], as epul-o[n] from epul-a; for there is reason to believe (above, p. 295) that these forms have lost or absorbed an i. As the termination -ia, -is, -sis, -tis is particularly appropriated to verbal nouns expressing the action of the verb, we must conclude that the verbal nouns in -io, -sio, -tio, are also derived from the genitive of nouns in -ia, -sis, &c. And this will lead us to the meaning already suggested, namely, that these words denote the result of an abstraction which has become fixed and objective.

The important word relligio will furnish a good exemplification of my meaning. There have been two different opinions with regard to the etymology of this word. For while most modern scholars adopt the suggestion of Servius (ad Virg. Æn. VIII. 349), Lactantius (IV. 28), and Augustin (Retract. I. 13), namely, that the word comes from religare, supporting this view with the quotation from Lucretius (I. 931, IV. 7): "relligionum nodis animos exsolvere;" Cicero makes religere the main verb, and gives the following explanation (de Nat. Deor. II. 28. fin.): "qui omnia, quæ ad cultum deorum pertinerent, diligenter retractarent et tamquam relegerent, sunt dicti religiosi ex relegendo, ut elegantes ex eligendo, tanquam a diligendo, diligentes, ex intelligendo, intelligentes: his enim in verbis omnibus vis legendi eadem, quæ in religioso;" and similarly, in another part of the same work, he says (ibid. II. 3, § 8): "relligio est quæ superioris cujusdam naturæ (quam divinam vocant) curam cærimoniamque affert." This etymology is in accordance with the verse quoted by Aul. Gell. IV. 9: "religentem esse oportet, relligiosum nefas." And there can be no doubt that it is perfectly

true. It is clear from the use of the word that relligio is not derived from religare, "to bind back," but from religere, "to gather over and over again," "to think perpetually and carefully on the same subject," "to dwell with anxious thought on some idea or recollection:" so that re-ligens is nearly a synonym of di-ligens, and an opposite of neg-ligens. The word expressing the abstraction of the verb should end in -ia, but this, as in most of these words in -io[n], is lost, and we have only the derivation from the genitive case expressing the result of the abstraction-the realized ideal. Hence, practically, relligio signifies, (1) "religious worship," considered as scrupulous obedience to the exactions of the conscience, and with especial reference to the act of worship; as (Cic. Verr. II. 4, § 49): "qui sacris anniversariis ac summa religione coleretur;" or to the religious sanctity of an object; as (id. ibid. § 46): "fanum est Junonis antiquum, quod tanta religione semper fuit, ut semper inviolatum sanctumque fuerit;" (2) "religious scruple" or "superstitious fear," considered as something objective and real; as (Cæs. Bell. Civ. III. 72, § 4): "non recordabantur quam parvulæ sæpe causæ vel falsæ suspicionis vel terroris repentini vel objectæ relligionis magna detrimenta intulissent;" and especially in the plural, as (Lucret. I. 109): "relligionibus atque mineis obsistere vatum;" (3) by substituting the cause for the effect, "guilt causing religious scruple or fear," and "the divine curse and consequent remorse or oppression of the conscience caused by a sense of violated religious scruples." In the second and third sense it is used in a curious connexion with violare and expiare in three passages of Cicero, which have never, so far as I know, been compared by any lexicographer or commentator: (a) ad Atticum, I. 17, § 16: "quare et illa, quæ violata, expiabuntur; et hæc nostra, quæ sunt sanctissime conservata, suam religionem obtinebunt." Here, it should seem, religio means "scrupulous observance;" and the maintenance of uninterrupted intimacy between Cicero and Atticus is opposed to the atonement necessary to restore the violated harmony between Quintus and his brother-in-law. (b) Tusc. Disput. I. 12, § 27: “id quum multis aliis rebus, tum e pontificio jure et cæremoniis sepulcrorum intelligi licet; quas maximis ingeniis præditi nec tantâ curâ coluissent nec violatas tam inexpiabili religione sanxissent, nisi hæsisset in eorum mentibus mortem non interitum esse omnia