representative of the third element; therefore, the endings b+ must be divergent representatives of the same original F + r = F+n, to which in point of fact they come quite as near as t + r (above, p. 355). The same analysis may be applied to the nouns in -e-tum, -i-le, and -a-rium, as arbor-é-tum, ov-î-le, gran-á-rium; for if arbor-eus is formed by an adjunct of the second element under the form ya, arbor-é-tum must extend the same form by an addition of the third element, and a similar explanation will be required by the long î= ii and â = ea of ov-i-le and gran-á-rium, to which the land r terminations are appended. We see then that all nouns expressing agency, or the place, means, and occasion of agency, are formed by adding a combination of the second and third pronominal elements-and this is what we should à priori expect for the idea of agency is that something, i. e. a doing, proceeds from the subject, who by the nature of the case is presumed to be near, and passes on to an object, which by the nature of the case is presumed to be relatively more distant. But we observe that the same sort of endings are used to form ordinary adjectives derived from nouns and not from verbs; thus from rer we have reg-ā-lis = reg-ya-lis, from Roma we have Rom-ānus = Roma-yă-nus, from consul we have consul-a-ris consul-ya-ris, from civis we have civi-lis civi-ya-lis, from asinus we have asin-i-nus = asin-yanus, &c., which fully correspond to the forms ov-i-le, gran-â-rium, &c.; and there is also a class of diminutives in -cu-lus, which exhibit the same termination as the verbal nouns veh-iculum, &c. = = = It will be easy to show that the combination of elements in these cases is as consistent with their primitive signification as in the class previously examined. To begin with the diminutives. As there are objective nouns in -tum, -lum, -rum, as well as nouns combining this affix with one belonging to the second element, so there are diminutives in -lus and -leus, as well as those exhibiting the compound termination now under consideration. Thus we have libel-lus liber-u-lus, filio-lus, &c., as well as pisc-i-culus, homun-culus, &c. The origin of the diminutive expression, or vπокópioua, is to be sought in the tendency to speak of a darling object, as, at the same time, little. Whether this has or has not any connexion with a mother's fondness for a child is doubtful. But it is a universal practice to speak of a petted object as a glycerium, yλvképιov, or "dear little thing." In classical Latin the diminutive puella =pueru-la is invariably used instead of the original word. Now in these terms the feeling of personality becomes evanescent, and that of mere objectivity takes its place. With a view to the expression of this idea it seems to be a matter of indifference whether we merely append the objective ending -lus, Greek -v, or connect this with the main verb by some possessive affix derived from the second element-in Latin -c, Greek -2. For example, we may form the secondary noun juven-cus from juvenis without any change of meaning; and pul-lus, catu-lus, &c., will be just as good diminutives as juven-cu-lus. The other derivatives, mentioned above, must be regarded as extensions of the case in -i-na or i-n (p. 276). Thus Român-us= Roma-in-us is a man who lives "at Rome," Roma-i[n]. This is shown still farther by the relation between these nouns and their extensions in -en-sis. These derivatives are either formed directly from their primitives, as praten-sis, "that which belongs to or grows in the meadow" (prato-en=prat-in, in a heavier form prat-en), or else they involve some noun already formed upon the locative, as Roman-ien-sis from Romanus. "In genere," says Ruhnken (ad Suet. Cæsar. § 37, p. 58), "adjectiva, quæ in -ensis exeunt, designant res hominesque, qui sunt in aliqua regione, sed aliunde originem habent. Romanus, qui Romæ natus est; Roman[i]ensis, qui Romæ degit: Siculus, qui in Sicilia ortus est; Siciliensis qui incolit Siciliam, aliunde ortus: v. Fest. v. Corinthiensis et Intt. ad Vellei. Paterc. II. 51. Idem discrimen apud Græcos in Ιταλὸς et Ιταλιώτης, Σικελός et Σικελιώτης, &c. : v. Ammonium in his vocibus et ibi Cl. Valckenar." This is a correct statement of the fact, but it does not explain the formation of the secondary nouns in -ensis. As 'Iтaλwτns, &c., are formed from nouns in -a (New Crat. § 259), so we always find that, if there are co-existing derivatives in -nus and -en-sis, there is an intervening form in -ia. Thus from Hispanus we have Hispania, and from this again Hispaniensis as from the locative Hispania-in. Accordingly, we may infer that Romaniensis, which is the true form, comes from an intervening Romania as the country of the Romani. The permanence of this rule of secondary derivation is shown by the practice of our bishops, who call themselves Cantuariensis, Dunelmensis, &c., to show that they are temporary incumbents, rather than hereditary peers. A comparison of these nouns with the equivalent Greek forms in -w-Tns, -in-tns, -i-tns, teaches us that the termination -sis, attached to the locative -en and belonging to the second element, is identical with the similarly derived -τns. We shall therefore not be surprised to find it also under the forms -tis and -ter. This is the fact when the locative, to which it is attached, is plural, as in the case of those nouns, which express an extended region rather than a definite locality. As we say, in agris, in campis, in sylvis, in terris, rather than in agro, in campo, &c., it is natural that we should find, as we do, agrestis, campes-ter, sylves-ter, terres-ter, rather than agren-sis, &c., which do not occur. At first sight we might feel disposed to refer eques-ter and pedes-ter, rather to the substantives eques, pedes, than to the locatives equis, pedibus. But the omission of b in queis for quibus, &c., shows us how pedeis might be a locative, and we have a passage in Virgil, which actually places the locative equis on a parallel footing with the derived pedes= ped-it-s; Æneid. VII. 624 : Pars pedes [i. e. pedibus iens] ire parat campis: pars arduus altis Pulverulentus equis furit: omnes arma requirunt. The noun seques-ter does not belong to this class. As denoting a functionary, it connects itself at once with magis-ter and minis-ter, and as these involve adverbs, which are of the nature of locatives, we must derive seques-ter, not from sequor with the old grammarians (for then we ought to have secu-tor), but from secus-sequis (cf. sequior)=exas, and thus sequester, which means a mediator, umpire, or other indifferent party, will naturally imply one who stands apart from both the litigants; for quod secus est is opposed to quod interest (Plautus, Trin. I. 2, 93). The patronymics in -ilius must not be referred to the same class with the nouns in -ile, -inus, -arius, &c. As it is known that in this case li-di (compare Acilius, Epilius, Hostilius, Petilius, Pompilius, Popilius, Venilius, with their original forms Acidius, Epidius, Hostidius, Petidius, Pompedius, Umbr. Pumperius, Popidius, Venidius), we must refer these words to the same class with the Greek patronymics in -dns, where the second pronominal element appears under the form of an approximate dental sibilant (New Crat. § 262). § 9. (y) The third Pronominal Element compounded with others and reduplicated. The most common extension of the third pronominal element is its reduplication under the forms t+n, or n+t, the latter combination being by far the most usual. With regard to other forms into which the pronoun enters under the type t, it is not always easy to say whether this is a corruption of ty, or merely the expression of the objective word. Thus we have seen that in t+r there is something more than the third element extended by the addition of r. It is probable, however, that in such affixes as -timus and -ti-nus we have merely the third element in the first syllable; compare the Sanscrit punya-ta-mas, hya-ta-nas and nú-tnas, in which the dental appears unaffected by any foreign element, with ex-ti-mus, legi-ti-mus, cras-ti-nus, hes-te-r-nus ; and taci-tu-r-nus with the passive participle taci-tus. We come to a similar conclusion by comparing the older spelling of the affix, as in op-tumus, with the change in τύπτ-ο-μεν = τύπτ-ο-μες, volu-mus, dic-i-mus, whence it appears that the u is not a vocalized consonant, but a mere change of articulation for an original o = a. In this inversion, it really matters very little, so far as the meaning of the affix is concerned, whether the dental syllable is referred to the second element or the third. This has been shown in the analysis of the third numeral, which admits of a similar explanation, whether we consider it as made up of ta+ra, or regard it as a corruption of an original tva-ra (New Crat. § 157). Be this as it may, there can be no doubt that the combination n+t, which plays so important a part in Latin derivatives, is a reduplicated form of the third pronominal element, expressing objectivity in its vaguest signification. Hence we find this combination (resolved into ă1) as the neuter plural of all nouns; and either unresolved, or in various forms of assimilation, in the third person plural of verbs, in the active participles, and as a further affix to nouns corresponding in meaning and often in origin to the perfect passive participle of the Greek verb and to obsolete Latin participles. In all these usages it denotes collective or vague for 1 A curious collateral proof of this resolution is furnished by Diáσtos ovσios=λioÚvτios from Mov[T]s: vide Steph. Byz. s. v.; and from this we may see that dinλáotos is of participial origin. objectivity in the neuter plural, a heap or mass of objects (like the Hebrew, Maskil le-Sopher, p. 14); in the third person plural, an action performed by an indefinite number regarded as an aggregate; in the participles, a mere notion of doing or being done. In the present instance we are concerned only with the participial forms and the nouns connected with them; and here we find in Latin not only forms in -nt, as aman[t]s, or in m+n or m+nt, as car-men, ver-tu-mnus, car-men[t]s, but also elongations of both in -ntu-s, -ndu-s, and -mentu-s. Thus we have qua-ntus by the side of Tâ[vT]-s, ama-ndus by the side of ama-n[t]s, and the names of places, which, in Greek end in -evt-s = -cis, -ovt-s = -ovs, or -avтs = -as, generally appear in Latin under the form -entum; thus Acragas, Crumoeis, Maloeis, Pyxus, Taras, become Agrigentum, Grumentum, Maleventum, Buxentum, Tarentum. Similarly, we have ar-mentum, orna-mentum, &c., by the side of σώμα[τ] = σώ-μεντ, &c. These extensions have occasioned some difficulties in Latin etymology; it will be sufficient here to take the two interesting examples supplied by fundus and pondus. The former is obviously, on the principle just mentioned, an extension of fun[t]s or fon[t]s, the participle of fuo, "to pour out," which is involved in the agglutinate form fu-n-do (cf. per-do, cre-do, &c.), and in the frequentative fu-to. The nouns fon[t]s, "a fountain," i. e. "that which pours forth water," and fundus, properly "the bottom of a vessel for pouring out," hence the lowest part or basis of any thing, the solid part or foundation of a man's property, his estate or rò üñapxov, exhibit the formation under discussion, without any additional elements. But pondus, gen. ponder-is, leads us to the same class of words as opus, operis, and these, as we have seen (above, p. 299), are terminated by the softened dental, as an additional mark of objectivity. The ablative pondo, however, shows that there must have been a word pondus, pondi, corresponding to fundus, fundi, and the synonymous ablative sponte, "by the weight or inclination," proves that the participial noun pons, pontis (in old Latin abbreviated into pos, Varro, L. L. V. I. p. 3, Müller) originally referred to a weight laid down, or poured forth, such, for example, as an embankment, a mass of earthwork, or separate stones thrown into the water (yé-pupa), which was the primary notion of a bridge, as the means of crossing a stream: for we need not go far to prove the antiquity of stepping |