History of Latin Prepositions.
Prepositions are Adverbs, which came to be specially used in connexion with certain cases of the Noun or in composition with a Verb. In the early stage of I.-Eur. languages the cases alone were sufficient to indicate the sense, but as the force of the Case-suffixes became weakened, or as the necessity for clearer definition was more recognized, the Case-suffix was strengthened by the addition of an Adverb. Thus ire monte might mean "to go out of the mountain" or "to go down from the mountain." To indicate the first sense, the Adverb ex was used, ire monte ex; to indicate the second, the Adverb de, ire monte de; ex-ire monte, de-ire monte. Those Adverbs which, owing to their meaning, are most frequently associated with particular cases of Nouns, or are used in composition with Verbs, are called Prepositions; and the process, by which Latin Adverbs became Prepositions, may be seen in operation at various periods of the language. Thus contra, which has hardly passed the Adverb stage with Plautus and Terence, is a Preposition in classical Latin and governs an Accusative Case; coram is not a Preposition till Cicero's time; simul in Augustan poetry and Silver Age prose; retro not till Late Latin (e.g. vade retro me, "get thee behind me").
It is customary now in writing Latin to write the Preposition and the Verb in one word, e.g. exire, but not the Preposition and the Noun, e.g. ex monte, although the Romans usually wrote exmonte, &c., and always pronounced the Prep. and Noun as one word-group (ch. ii. § 11). This close union of the Preposition with its Verb and Noun led at an early time to the syncope of the final short syllable of a Preposition, e.g. indo-gredior became ind-gredior, a form confused with in-gredior (cf. O. Lat. induperator and class. imperator).
In the later stages of a language the use of Prepositions increases more and more. In Latin this culminated in the loss of Case-suffixes, and the use of Prepositions in their place, as we see in the Romance languages (cf. Fr. je vais à Rome with Lat. vado Romam). As early as the first cent. A.D., a grammarian points out that in manus aqua is the phrase in vogue instead of the older da aquam manibus. New distinctions of prepositional meaning were expressed by compounding Prepositions with one another, e.g. de-ex, de-sub, ab-ante (Fr. avant), just as I.-Eur. Prepositions often had a Particle appended to define their meaning. Such Particles were (1) -s(e), Gk. -de, e.g. Gk. οἴκαδε, εἴσω; Lat. ab-s, ex, sus-; (2) -d(e), Gk. δή, -δε (e.g. Lat. prod-eo, O. Lat. postid, antid); (3) -n(e), e.g. Germ. vo-n, Lat. pone for pos-ne. They are related to the Pronominal stems mentioned in ch. v. § 3.
In the earlier stage of every language the Prepositions must have been used with great elasticity, sometimes with one case, sometimes with another, the fixing down of Prepositions to a particular case being a late feature. Thus in potestatem esse instead of in potestate esse is a usage of archaic Latin. Their position too varied in course of time. In I.-Eur. the Preposition seems to have preceded the Verb, but to have followed the Noun, while between the Prep. and the Verb a Particle or Enclitic Pronoun might be inserted (cf. O. Lat. anti-d-eo, sub vos placo, the archaic phrase for supplico vos retained in Latin prayers, ob vos sacro for obsecro vos). In classical Latin a Preposition, especially a monosyllabic Preposition, precedes the Noun (hence "Pre-position"), except in particular circumstances (e.g. metu in magno); but in the older literature often follows it, just as our "in here" was earlier "here-in."
List of Latin Prepositions. (1) Ab.
Ab, from, is I.-Eur. ap (Engl. of, off), a curtailed form of apo (Gk. ἀπό), of which another curtailment was po (Lat. po-situs, pono for po-s(i)no, O. Lat. po-lubrum, a wash-basin, po-lire). The form ap- appears in ap-erio, and was no doubt the pronunciation of the word before an initial p, t, c, &c., e.g. ab templo; ab is due to the same Latin preference of final -b to -p as substituted ob for op (see below). The form abs (pronounced and often written aps, ch. ii. § 10), in which the Preposition is augmented by the Particle -s(e) (Gk. ἄψ), is used in Composition before t, c, e.g. abs-traho, abs-condo, while before p it is, by a law of Latin phonetics (ch. x. § 20), reduced to a, e.g. as-porto for aps-porto, as-pello for aps-pello; it appears also in the O. Lat. phrase absque me (te, &c.) esset (foret), equivalent to "si sine me esset," where que (ch. ix. § 2) seems almost to have the sense of "if" (cf. O. Engl. "an" for "and"); at a later period absque me, &c., was used without the verb, and absque came to take the sense of sine, without. Ā may be another form of ab as ē of ex, e.g. ā-mitto for ammitto, ab-mitto, as ē-mitto for emmitto, ex-mitto (ch. x. § 20). Au- of au-fugio, au-fero, however, represents an entirely different I.-Eur. Preposition awē, which was brought into requisition in these Compounds before an initial f to avoid confusion with the Compounds of ad, e.g. affero. A curious Preposition af, used in Cicero's time occasionally in account-books, with the name of the person from whom money had been received, may be a mere (Greek?) trick of writing, with the symbol F (the Greek Digamma) employed to denote the u- or w-sound.
(2) Ad.
Ad, at, to, I.-Eur. ad (Engl. at) is a different word from the Conjunction at, I.-Eur. at, though often confused with it in Roman spelling. On the old form ar, e.g. arfuerunt, arvorsum, due to the phonetic change of d to an r-sound before f, v, see ch. ii. § 8.
(3) Ambǐ-.
Ambi-, around, on each side, I.-Eur. ambhī (Gk. ἀμφί), a Locative of the same stem as I.-Eur. ambhō, "both" (Gk. ἄμφω, Lat. ambo), appears in Latin compounds in the form am- before a consonant, e.g. am-plector, am-icio for am-jicio (see ch. ii. § 12). This must be distinguished from an-, a curtailment of I.-Eur. ana, "on" (Gk. ἀνά, Engl. on) in an-helus, an-quiro, an-tennae, an-testari.
(4) Antě.
Ante, before, I.-Eur. antī (Gk. ἀντί, opposite, instead of; Engl. answer), a Locative Sing. of some stem connected with Lat. antes, rows (Engl. end) of which Gk. ἄντα, opposite (cf. ἄντην), is another case. In anti-stes, the ī of I.-Eur. antī, not being final, does not sink to ĕ (ch. ii. § 14 n.).
(5) Ǎpǔd.
Apud, which is also spelt aput, seems to be the I.-Eur. Preposition apo (of which Lat. ab is a curtailment), augmented by the Particle d(e) or t(i), and must have been originally apod or apot (cf. Dor. ποτί).
(6) Circum.
Circum, around, the Adverbial Acc. Sg. of circus (Gk. κρίκος, a ring), is the older form. In class. Lat. a by-form circa appears, first found in Cicero, a formation on the type of supra, extra, &c. (ch. vii. § 4), which was originally employed with verbs like esse owing to a feeling that circum was suitable only for verbs of motion, e.g. legatos circum civitates mittere, "to send ambassadors a tour of the states," ire circum urbem, "to go a circuit of the city." Circiter, an adverbial formation like breviter (ch. vii. § 2), came to be restricted to the logical sense of "about," "almost," e.g. Plaut. loca haec circiter, "hereabout." The form circo appears in the Adverb id-circo, as circa in quo-circa, also with logical sense.
(7) Cǐs, cǐtrā.
Cis, citra, on this side, are formed from the I.-Eur. pronominal root ki-, "this" (Gk. -κι of οὐκί, πολλάκι, Engl. he), exactly as their opposites uls, ultra, on that side, from the I.-Eur. pronominal root ol-, "that" (ch. v. § 3), the first by the addition of the particle s(ĕ), the second (an Abl. Sg. Fem.) with the suffix -tero- (ch. xi. § 8). The Adv. citro (Abl. Sg. Neut.) corresponds to citra as ultro (e.g. ultro citroque) to ultra.
(8) Clam.
Clam, an Adverbial Acc. Sg. Fem. from the root kel-, "to hide" (Lat. celo, occulo, &c., ch. vi. § 2), had in O. Lat. a by-form clam-de, whence was formed the Adj. clandestinus. It governs the Acc. (not the Abl.), and has in the Comedians another, apparently a Diminutive form, clanculum, e.g. Ter. clanculum patres.
(9) Cum.
Cum, older com (a form still retained in Composition, e.g. com-es, a companion), is I.-Eur. kom. (On the change of ŏ to ŭ in Latin, see ch. ii. § 4, and on the loss of -m in co-eo, &c., ch. ii. § 3.)