=

terranean source may chance to resemble genuine Greek material. For example, vos "frankincense" seems to be connected with ouw, Ouμós, etc., and Latin tus is therefore always cited as a Greek loan word (the initial consonant forbids us to call it a cognate of Oúos). Since the genitives turis and Oveos show the treatment of intervocalic s which is regular in inherited words (e.g. generis Yéveos), the borrowing, if that is the true explanation, must have occurred while both Greek and Latin still retained intervocalic s. Contact of Greek and Latin at so early a date is unlikely, and therefore we must rather assume, even at the cost of abandoning a satisfactory Greek etymology, that both got the word from a third source. There is no lack of indication in Greek itself that the base ou in religious words was of Aegean origin. We cannot trace to an Indo-European source such words as θύρσος, θυμέλη, θυμάλωψ. There are also many place names and mythological names with the same element, of which I will cite only these: Overσós, a town in Lydia; Ovéσrns; Ovén, a town in Oenotria; Ovnλai, daughters of Earth, who taught men to sacrifice; Ovývn, a nymph of Dodona; Oviάdes = Báкxides; Ovwvn, an epithet of Semele.

On the other hand, it would not be safe to assume that all terms of civilization common to both the classical languages for which there is no satisfactory Greek etymology were borrowed independently by Greek and Italic from a Mediterranean source; the Greeks undoubtedly carried many a foreign product and its name to Italy.

Neither is it safe to say that a relatively uncommon variation in form between a Greek word and a Latin is proof that the latter does not come directly from the former. For instance Cuny 7 was wrong in holding that Latin words with b, d, g for π, T, K; or p, t, c for ß, d, y (e.g. gubernare for Kußeρvâv) cannot derive from the Greek. F. Forhalle

8

agrees with

7 Rev. des Ét. Anc. XII, as cited by Forhalle. The article is not accessible

here.

8 Mélanges Linguistiques Offerts a M. J. Vendryes (1925), pp. 157–178.

9

Cuny in the main, but he recognizes the necessity of tracing some such words to Greek as their immediate source (e.g. Burrus for Пuppós), and he also notes the similar irregularity in certain Greek loans from the Latin (IIóλos from Publius, 'Apeyavós from Africanus, etc.). These two groups of words he explains by Meillet's doctrine that the stop sounds, both in Greek and in Latin, were weakly articulated and therefore easily misunderstood. This argument proves too much; in general the stop sounds were accurately preserved in ancient Greek and Latin, and even in the modern forms of those languages there is less alteration of the original stop sounds than there is in the Germanic languages, to which Meillet assigns a strong articulation. I am still of the opinion which I expressed eight years ago 10 that we must find the explanation of such forms as Burrus and IIóλtos in differences of pronunciation between Greek and Latin. I expect soon to indicate some modifications of the theory then advanced which will, I hope, meet the objections of my critics.

11

In this paper I shall assume that loan words 11 appearing both in Greek and in Latin were brought to the Romans by the Greeks unless I know of some reason for deciding otherwise. Quite likely some of the words in the following lists will ultimately have to be deleted.

Many of the Greek words which Latin possessed in the time of Plautus must have been part and parcel of the language for a long while. Thus exanclare (eçavrλâv) shows the change of tl to cl, which is familiar in Latin words like pōclum, cognate with Sanskrit pātram, and which was completed in prehistoric times. The pair, oleum and oliva (ëλaɩfov, éλaífa), illustrate the process which before Plautus changed *deivos to deus and *deivi to divi.12

Much more numerous are the Greek loan words which 9 M. S. L. XIX, 163-173.

10 T. A. P. A. XLVIII, 49-62.

11 Most of my material, aside from proper names, has been taken from F. O. Weise, Die Griechischen Wörter im Latein, Leipzig, 1882.

12 See Sommer, Handbuch der Lateinischen Laut- und Formenlehre2, p. 74.

shared in the weakening of medial vowels, another process which was completed before the beginning of the literature. Examples are anquina (àykoivn), catapulta (KATATÉλTNS), epistula (¿ñiσtoλý), Hecuba ('Ekáßŋ), machina (properly macina —μαχανά), paenula (φαινόλης), pessulus (πάσσαλος), placenta (πλακοῦντα), scopulus (σκόπελος), scutula (σκυτάλη), sesuma (ohoapos), Sicilia (ZiKelia), Siculus (Likeλós), talentum (σήσαμος), (Σικελία), (Táλavrov), Tarentum (Tápavra), trutina (Tourávn). The (τάλαντον), change of a to e before r was also prior to Plautus, and hence words like tessera (a short form for reooapȧywvos) 13 and phalera (páλapa) 14 had belonged to the Latin language from preliterary times.

We may be reasonably sure that a word was current in the Latin of Plautus's time if it appears in his plays in a form inconsistent with his regular method of transliterating Greek, What Plautus's ordinary practice was may be inferred from his treatment of the Greek names of his characters, excluding only the mythological names Alcumena and Amphitruo, which may have received their Latin form at a much earlier date. The other names of Plautine, and also of Terentian characters are treated as they would have been in the Ciceronian age, with the following exceptions: (1) All feminines of the first declension show final a in the nominative singular and am in the accusative. (2) Terence usually treats masculines of the first declension in the same way (Geta, Phaedria). Plautus follows this practice in names ending in las, éas (Sosia, Saurea, Dinia, Simia); otherwise he prefers Charmides, Apoecides, Lyconides, Theopropides, etc. (but note Leonida). (3) There is somewhat more variation in the declension of Greek names than there is in classical Latin. (4) In Plautus's time the letters y and z were not in use, their place being taken in Greek loan words by u and s(s). (5) Greek aspirates were not indicated by h. Such names as Lyconides, Philolaches,

13 See Bréal, M. S. L. vi, 5, and Walde, Etymologisches Wörterbuch s.v. Bréal is wrong in calling tessera "un example de mot d'origine savante."

14 This word itself is not quotable before Cicero, but the derivative phaleratus occurs in Terence, Ph. 500.

and Theopropides exhibit the work of modernizing redactors or copyists.

In other respects Plautus transliterated Greek names as Cicero did, and it is safe to say that he followed the same system whenever he took any other word directly from his Greek model. These words were of course modernized by redactors and copyists in the same way as the personal names, but any other variation in our manuscripts from Plautus's system of transliteration indicates that he found the word in question already current in Roman speech. Such a word is aurichalcum (ópeíxaλkov), which owes its initial diphthong to a popular etymology. Other examples are Alcumena ('Aλkμývŋ), balaena (páλaiva), calx (xáλış), conger (vóyypos), drachuma (Spaxμn), gubernare (κvßepvâv), lanterna (λaμπтýр), mina (μvâ), peller (παλλακίς), resina (ῥητίνη).

Many of the Greek words in Plautus are shown to have been already naturalized at Rome by the existence of derivatives with Latin suffixes. Plautus uses both gubernare and gubernator; both opsonium and opsonari, opsonator, opsonatus; balineum (Badaveîov) and also balineator; elleborus and elleborosus; pausa (Tavσis) and the denominative verb pausare; machina and machinari; moechus and moechari; zona and zonarius. Scarcely less valuable as an indication that a word was in common use at Rome is frequency of occurrence in the early literature, as in the case of dapsilis (dayıλńs), hilarus (ιλαρός), mina (μνᾶ), drachuma (δραχμή).

It is probable that most of the other Greek words in Plautus were familiar to his audience, but in some cases we cannot get beyond this general probability. Such a word as apologus, Stich. 538, 541, 544, 570, may have been taken directly from the Greek original. There are besides several words which are more likely to have been borrowed by the dramatic poets than by anyone else. Technical terms of the theatre, such as scena, prologus, choragus, must have been unknown at Rome until there was a Roman drama. There are also in the comedies several Greek words which could not have found a

place in Roman speech because of differences between Greek and Roman society. No one in early Rome except the authors of fabulae palliatae was likely to need the words gynaeceum, parasitus, and ephebi.

As to the precise Greek community from which an early loan word was taken, we are usually left quite in the dark. It has been thought that nominatives in a, like poeta, pompa, scena, indicate derivation from a dialect other than Attic or Ionic; but no such inference can properly be drawn. Plautus and Terence took their characters from Attic plays, and yet they gave them such names as Scapha, Philumena, Ampelisca. The noun bibliotheca can hardly have been borrowed before the Ciceronian period or from another dialect than the Attic Kown, and yet it usually shows final a in the manuscripts. In these instances the analogy of native Latin nouns fully accounts for the final a.

There are, however, a few words which contain medial long a in place of Attic-Ionic 7, and their source must, of course, be other than Attic or Ionic. One of these words is machina, which may well have been borrowed in the sense of "block and tackle." 15 If it is a plausible guess that this and other early terms of seafaring and commerce came to Rome from Syracuse, the reasons are exclusively historical and archeological. From the same source may come such words as balaena, gubernare, ancora, citrus, mina, drachuma, lanterna, paenula. Similarly Argivus, Achivus, and oliva must come from dialects or a dialect that retained ; but precisely which dialects remains uncertain.

One reason for the representation of Greek v by u in early Latin, as in Burrus (IIvppós), gubernare, etc., was that u was a pure u in many of the Greek dialects of Southern Italy and Sicily, including not only Syracuse and the other Doric colonies but also Ionic Cumae (not Cymae!). Still no argument as to the source of a word can be drawn from this usage; for even Attic v was as near to u as to any Latin sound. When, 15 Cf. Horace, Carm. 1, 4, 2: trahuntque siccas machinae carinas.