STOP Early Journal Content on JSTOR, Free to Anyone in the World This article is one of nearly 500,000 scholarly works digitized and made freely available to everyone in the world by JSTOR. Known as the Early Journal Content, this set of works include research articles, news, letters, and other writings published in more than 200 of the oldest leading academic journals. The works date from the mid-seventeenth to the early twentieth centuries. We encourage people to read and share the Early Journal Content openly and to tell others that this resource exists. People may post this content online or redistribute in any way for non-commercial purposes. Read more about Early Journal Content at http://about.jstor.org/participate-jstor/individuals/early- journal-content . JSTOR is a digital library of academic journals, books, and primary source objects. JSTOR helps people discover, use, and build upon a wide range of content through a powerful research and teaching platform, and preserves this content for future generations. JSTOR is part of ITHAKA, a not-for-profit organization that also includes Ithaka S+R and Portico. For more information about JSTOR, please contact support@jstor.org. DISSIMILATIVE WRITING IN REPUBLICAN LATIN AND UO IN PLAUTUS Bt E. H. Sturtevant Several scholars 1 have in recent years shown quite conclusively that in certain important respects Latin orthography did not accu- rately represent the pronunciation, and they have traced a group of these irrational spellings to the influence of a precept of the schools. During the last century or two of the Republic there was a widespread prejudice against writing vv or n, and consequently it was customary to spell qvom, ingenvos, servos, adicio, inferis 2 to indicate the pronunciations quum, ingenuus, servus, adiicio, inferiis. Niedermann (p. 60) and Anderson (pp. 103 f.) feel that the aversion to vv was at least in part due to a desire to avoid ambiguity; but Kent justly remarks (p. 41) that servvs and the like are not ambiguous to one who speaks Latin. We may add that adicio and inferis are much more misleading than if written with n. Then what was the origin of this bit of Roman pedantry ? So pointless a theory could scarcely grow up except as an attempt to explain some previously existing anomaly, although when once adopted the theory- may well have tended to perpetuate the anomaly. We may safely assume that the facts which we have before us โ€” or some of them, at least โ€” are older than the theory by which the Romans accounted for them. Lindsay, Niedermann, and Kent are quite right in think- ing that pedantry was wholly responsible for the persistence in the last years of the Republic of such spellings as qvom for cum, servos for servus, adicio for adiicio, etc., but they must be wrong in suppos- ing that the Romans chose so perverse a system of orthography for its own sake. As far as origins are concerned, we are driven back upon the older explanation; the spellings with vo and with a single i were phonetic 1 Lindsay, Latin Language, pp. 227, 271; Niedermann, Milanges de linguistique offerts a Saussure, pp. 58-65; Mather, Harvard Studies, VI, 83-151; Anderson, TAP A, XL, 99-105; Kent, TAP A, XLIII, 35-56. * In this paper small capitals are used to indicate spelling, while italics indicate that the pronunciation of quoted words is chiefly thought of. [Classical Philology XI, April, 1916] 202 Dissimilative Writing in Republican Latin 203 to start with, although, as we have just seen, they continued in use after they had ceased to represent pronunciation. In some cases there is no doubt that a pronunciation once existed which would normally be represented by the spellings we have called anomalous. Thus compounds of iacio have an initial short syllable in the early dramatists (e.g., ddiceret, Plautus Poen. 1174), and consequently the spelling adicio was at that time phonetically correct. When i had been reintroduced from ieci, iactus, etc., the old spelling was retained for the new adiicio, and now the pedantic theory about n was called in to explain the anomaly. Similarly pompei originally stood for Pompei and Pompei (both from Pompeiit), and when the analogy of Pompeiius, etc., restored U, the shorter spelling was retained to repre- sent Pompeiii. Inferlis regularly became inferls about 150 B.C., and led to the spelling inferis, which was presently employed for the analogically restored inferlis. So, too, such spellings as socieis at first accurately represented the pronunciation, but when the phonetic distinction between ei and I broke down they required explanation, and so helped to suggest the rule that the letter i could not follow itself. Kent understands adiese, adieset, and adiesent in the Senatus ConsuUum de Bacchanalibus as dissimilative writings for adiisse, etc. But this view is uncalled for if we accept Brugmann's identification 1 of Latin it (perfect of eo) with Sanskrit iy-aya, iy-Mha, a perfect with "Attic" reduplication. Latin shows the diphthong, which should appear in the strong forms, in red-ieit (CIL, I, 541), inter-ieisti (CIL, I, 1202), ad-iese (i.e., if from iet), and svb-M, etc., in the poets, while iero (Plautus Capt. 194), lerant (Terence Ad. 27), etc., and Umbrian iust preserve the I (from H- of the weak forms) which appears in Sanskrit ly-us, ly-dthus. If Kent is right (p. 47) in interpreting abiegnieis, aescvlnieis (CIL, I, 577) as abiegnels, aesculnMs, the pedantic aversion to double vowel signs must here have operated as an impelling cause, and it may indeed have done so at any time after it gained acceptance as an explanation of other phenomena. It is, however, still possible to interpret these words as abiegnls and aesculnls, as I formerly did. 2 ยป Grundriss*, II (1913), No. 3, pp. 34 f. ' Contraction in the Case-Forms of the Latin "io-"and " ia-" Stems and of "deus," "is," and "idem," p. 35: 204 E. H. Sturtevant It is certain that the spellings qvom, ingenvos, and servos once represented the pronunciation; but here there is a difficulty. If, as Anderson seems to think (p. 105), servos became servus at the time *dolos became dolus, we have no explanation of the survival of the old spelling servos beside the newer dolvs; as we have already seen, the pedantic aversion to doubling a vowel sign cannot be the cause of the phenomena which that theory was invented to explain. It is, then, extremely probable that the pronunciation servos persisted after *dolos had become dolus. This phonetic difference may have lasted only a short time; once a distinction in spelling has established itself it tends to persist after all reason for it has vanished โ€” as witness a hundred anomalies in the traditional English orthography. It is possible, therefore, that servos was pronounced servus even as early as Plautus' time; whether it actually was so pronounced or not is a question that will claim our attention later. At present we are merely contending that the pronunciation servus was of somewhat later origin than the pronunciation dolus. The same reasoning goes to show that the change of voltis to vultis must have been somewhat later than the change of *moUus to multus. That this actually was the history of the sounds we have fairly conclusive evidence. Anderson (pp. 102 f.) holds that no 5 before I became u until after Plautus, but the evidence he cites scarcely justifies so wide a generalization. His argument is based upon the following: (1) consolvervnt, cosoleretvr, tabolam, oqvoltod in the Senatus Consultum de Bacchanalibus of 186 B.C., (2) moltare in Eph. Ep., II, 298, and moltai in CIL, XI, 4766, (3) co(n)sol in various early inscriptions, and (4) four passages in early authors where assonance makes the pronunciation vol for the later mil seem probable, e.g., Plautus Amph. 114, volt voluptatem. The evidence of the fourth group concerns the words voltis, volt, and volta (=voltus), and it is therefore quite in harmony with the current opinion that the early Latin orthography vo was phonetic. Although evidence from puns and plays on words must be used with great cau- tion, we may grant that Anderson has given us a valid reason for pronouncing early Latin wZ+consonant as we have been accustomed to do. Dissimilative Writing in Republican Latin 205 His first three groups of evidence, however, can scarcely stand against the great host of words with ul from earlier ol which occur in our manuscripts of Plautus and Terence in the same form as in classical Latin. Such words are multus, culpa, consultus, pulvis, fabula, stdbulum, vestibulum, tetuli, postulo, and others far too numer- ous to name. It is granted that one form on a stone is more valuable than many forms in manuscripts, since the orthography of the dramatists may have been largely modernized; but it is after all scarcely credible that so far-reaching an alteration as this should have been made in Plautus' spelling without leaving some traces of the original state of affairs. We must, therefore, inquire whether Anderson's epigraphic instances of ol may not be archaistic spellings of ul. In the first place, it is well known that the legal phraseology of the Romans was in general very conservative, and furthermore that the laws of the Twelve Tables were so thoroughly learned by every schoolboy that they had a powerful influence upon the development of the Latin language. We cannot, then, lay much stress upon the archaic spell- ing of a legal word like molto or molta, or of a publicist's word like co(n)sol. There remain the four forms in the Senatus Consultum de Bacchanalibus, a document which is known to contain archaisms (e.g., the ablatival -d). Some of the forms themselves are rather suspicious: the phonetic interpretation of oqvoltod is too uncertain to serve as the basis of an argument; the verb consulo was popularly connected with consul (Varro LL. 5. 80: "consul nominatus qui con- suleret populum et senatum, nisi illinc potius unde Accius ait in Bruto 'qui recte consulat, consul ciat'"), and therefore ol in con- solvebvnt and cosolebetvr may be due to the traditional spelling co(n)sol (note the omission of n in the second word). As far as oZ-f- consonant is concerned, we have in Plautus' sultis for si voltis indubitable proof that the change to ul (except after v) was earlier than that author's day. Since v is lost only between like vowels, sultis must be due to the analogical proportion, vis: sis = voltis ix, 1 which could yield nothing but *soltis. This *soltis changed before Plautus' time to suliis, which occurs with this spelling in the 1 Cf . Lindsay, The Captivi of Plautus, p. 230. 206 E. H. Stubtevant Ambrosian manuscript in Stich. 65, and many times in the other manuscripts, and whose u is guaranteed by such w. II. as stultis (bcd) in Rudens 820. In this word a modernizing of Plautus' spelling is out of the question, since suttis is not quotable from any later author. The contrast in pronunciation between voliis and multus, which we found ourselves bound on theoretic grounds to assume for some early period, is shown by sultis to have belonged to the period of Plautus. That voltis was still pronounced with an o in Caecilius' time is Ander- son's inference from Aethrio 5 (cited by Diomed. 1. 386 K.) : "Actu- tum, voltis, empta est; noltis, non empta est." Anderson argues that before Plautus' time uo and vo became uu and vu in all circumstances under which Imperial Latin shows the change, excepting only in the position before I. Although Kent (p. 41) accepts the conclusion, the present writer is still unconvinced. The first argument is this: " It seems certain . . . . , in view .... of the fact that vowel weakening normally came later in accented than in unaccented syllables, that servos began to be pronounced servus at a considerable interval before volnus began to be pronounced vulnus." We may admit the validity of the reasoning, merely sub- stituting the word "likely" for the word "certain," but as long as we do not know when volnus became vulnus such a consideration can scarcely be said to "establish" 1 servus in the plays of Plautus. The only other evidence adduced that applies to words of the type of servos is the spelling flavs in CIL, I, 277 (an early coin inscription), which Anderson interprets as Flavus. 2 It seems more likely that it is a phonetic spelling of the form which regularly developed from Flavos. 3 Flaus then stands in the same relation to Flavus as deus to divus, oleum to olivum, and Gnaeus to naevus. The words with quo>cu form a special class, since they exhibit two changes which may or may not have taken place at the same time. The confusion between the preposition cum and the conjunc- 1 Kent, loc. cit. 2 Anderson's citation of Ritschl's theory (Opusc, IV, 488) about this form is incorrect, but, as Ritschl's discussion of the matter is now antiquated, we need not consider it here. * See Solmsen, Studien zwr lateinischen Lautgeschichte, pp. 37 f . ; cf . Sommer, Handbuch 1 , p. 162. DlSSIMILATIVE WKITING IN REPUBLICAN LATIN 207 tion, which is proved by the use of qvom 1 for the preposition in CIL, I, 34 (about 150 B.C.), shows that the conjunction had come to be pro- nounced cum shortly after Terence's death. Although it is possible that the influence of turn made quom become cum before equom became ecum, it is probably safer in the absence of further evidence to assume that all words with unaccented quo suffered the change at about this time. Anderson's argument that since u after q was a weak sound it could not have prevented equos from becoming equus at the time *dolos became dolus is not convincing. There seems to be no cogent reason for doubting the phonetic character of the Plautine and Terentian spellings qvom, eqvos, etc. The conclusions which Kent draws from his own paper and those of Mather and Anderson (Kent, pp. 55 f.) require one modification. The use of i for ii and Hi and the use of vo for uu and vu was due in each case to the retention of an old orthography to denote a changed pronunciation. The resulting anomalies were explained in the schools as due to the impropriety of doubling vowel signs, 2 and the satisfac- tion given by this explanation perpetuated for varying lengths of time some of the unphonetic spellings. Anderson's conclusions need revision in two points. His attack upon the traditional interpretation of vo in Plautus is not successful, and he is wrong in dating after Plautus the weakening of ol (except when v precedes) to ul. On the other hand, he supplies a useful prop for the current theory in regard to the pronunciation of voltis, etc., in Plautus. He shows that the change of unaccented quo to cu took place about 150 B.C., a hundred years earlier than had been sup- posed, and he makes it probable that the spelling vo for imperial w was no longer phonetic in Cicero's time. Columbia Untvtebsity 1 A less satisfactory explanation of this form is suggested by Sommer, Handbook*, p. 158. 1 Was this explanation due to the same grammarians who rejected Accius' plan for denoting vowel length by doubling ?