The Classical Review http://journals.cambridge.org/CAR

Additional services for The Classical Review:

Email alerts: Click here Subscriptions: Click here Commercial reprints: Click here Terms of use : Click here

The Relation of the Resolved Arsis and Resolved Thesis in Plautus to the Prose Accent

Charles Exon

The Classical Review / Volume 20 / Issue 01 / February 1906, pp 31 - 36 DOI: 10.1017/S0009840X00993569, Published online: 27 October 2009

Link to this article: http://journals.cambridge.org/abstract_S0009840X00993569

How to cite this article: Charles Exon (1906). The Relation of the Resolved Arsis and Resolved Thesis in Plautus to the Prose Accent. The Classical Review, 20, pp 31-36 doi:10.1017/S0009840X00993569

Request Permissions : Click here

Downloaded from http://journals.cambridge.org/CAR, IP address: 130.102.42.98 on 24 Jun 2015 THE CLASSICAL REVIEW. 31 On a small tablet found near the Propy- the resettlement of Lesbos after the revolt; laea * in the writing of the latter part of the but the former opinion that we should fifth century we find record of an Athenian restore es 'Bpsrpiav might find favour again, diroucia h *Ep . . . .; the stone is broken at a if the interpretation of our new document tantalising point: the accepted restoration is which I have proposed were to be accepted. now « "Eptaov and refers the inscription to LEWIS R. FARNELL. 1 G.I.A. 1, 339.

THE RELATION OF THE RESOLVED ARSIS AND RESOLVED THESIS IN PLAUTUS TO THE PROSE ACCENT.

No satisfactory attempt has ever been admission of dactyls, anapaests, and proce- made to account for the existence of the leusmatics to the dialogue metres. As these strict rules observed by Plautus, and the rules serve purposes so very similar, they early Roman dramatists generally, in the use naturally overlap sometimes; but it will be of dactyls, anapaests, and proceleusmatics in convenient to consider them in the form in dialogue verse. Why, for instance, is a dac- which they are usually stated. It is demon- tylic word not a good substitute for a trochee, strable that they are all particular cases, or seeing that an anapaestic word is a perfectly at least corollaries of a single law, which good substitute for an iambus ? Or why is may be stated thus :— an anapaest divided as in hds\tibus 6m\nibus Except in the first foot of a colon, a re- forbidden, while a dactyl divided as in qude solved arsis or thesis must, in dialogue metres, uo\Us faci\am omnia is perfectly legitimate 1 begin in a syllable which would bear in prose It is true that explanations have been sug- the primary or secondary word-accent, or a gested of one or two particular rules, but sentence-accent? they are far from convincing. Thus the rule As regards the proceleusmatic, it has al- forbidding caesura in a resolved thesis in ready been shown by A. W. Ahlberg that dialogue verse is explained by Prof. Lindsay both the arsis and the thesis of that par- as due to the desire of the Roman drama- ticular foot must begin in syllables accented tists not to add to the length of the 'irra- in prose (De Proceleusmaticis, etc., Lund, tional ' thesis by ' the pause that inevitably 1900, p. 36. Cf. Ritschl, Prol. Trin. p. 289). attends upon the conclusion of a word.'* His method was the collection and classifica- But it is not the fact that a pause necessarily tion of all the proceleusmatics occurring in or usually occurs between words in a spoken the dialogue metres of Plautus and Terence. sentence, the division of sentences into words Such a method would be hardly practicable being logical, not phonetic. Further, caesura for the resolved arses and theses generally; is just as strictly forbidden in the resolved but I hope, by a different method, to show arsis of the same feet, which arsis is not that every resolved arsis or resolved thesis ' irrational.' It is therefore hard to see why found in Plautus is constructed in obedience such an explanation is offered. All these to the same law. The one seeming excep- rules have for their object the determining tion— the case in which a resolved arsis of the conditions under which certain feet seems to begin in a short final syllable— may be admitted to trochaic and iambic will be separately discussed below. It should verse, and these feet resemble one another be observed that, like the proposed law, the in having ' irrational' theses, and resolution particular rules which are deducible from it of the arsis or of the thesis or of both. as corollaries do not apply to the first foot What is needed therefore is a comprehensive of a colon.3 Also the few iambic and trochaic explanation. The true reason for the ex- lines occurring in the Cantica should not be istence of any one of these rules will prob- ably be the reason for the existence of all 2 It need hardly be said that this law did not at first occur to the writer in the form in which it is the rest. To show that such a comprehensive here presented. It emerged in the course of a explanation is possible is the purpose of the minute analysis of many hundreds of lines, under- present paper. taken with a view to discovering the exact means by which the dramatists avoided metrical ambi- Writers on Plautine metric usually recog- guity. It is here demonstrated in what seems to nize some five or six rules regulating the be the briefest possible form. 3 See O. Seyffert in Berl. Phil. Woch. 1891, 1 Introduction to Captiui, 1900, p. 68. p. 926. 32 THE CLASSICAL REVIEW. included in the dialogue metres. I take the a\micus — or falling — cdrus a\micus: in rules seriatim. either case the law would be broken.* 1. A dactylic or tribrach word or word- 6. A long syllable accented in prose may ending cannot receive the ictus on the middle not be shortened in dialogue verse under the syllable.1 law of Breues Breuiantes. This follows directly from the general law. In order that such a syllable might be For the tribrach word this rule is not relaxed shortened, it would be necessary that it for the first foot or for anapaestic metre. should be preceded by a short syllable, and 2. A dactylic word or word-ending can- that an ictus metricus (cancelling the word not be a substitute for a trochee or an accent) should fall either on the preceding iambus. short syllable, mdUstus, or on the following As a substitute for a trochee, a dactylic syllable, mdlgst&s. In the first case a re- word would contain a resolved thesis whose solved arsis, in the second case a resolved first syllable would not be accented in prose. thesis would offend against the proposed law. As a substitute for an iambus, it would re- This rule, like the others, is suspended in ceive the ictus on the middle syllable, which the first foot of a colon and in anapaestic is forbidden by the first corollary, and there- metre, as has been proved by the careful fore by the general law. It is significant studies of A. W. Ahlberg.5 that a word that becomes dactylic by elision As has been already remarked, there is is admissible as a substitute for either a one case of resolution of the arsis which trochee or an iambus (eriper(e) dculos, or seems to contradict the proposed law, the eriper(e) omn6s).2 Such a word so placed case in which the resolved arsis (not thesis) obviously fulfils the requirements of the seems to begin in a short final syllable. If law. this case be reserved for the present, the 3. The first two syllables of an anapaest above demonstration proves the law not for may not be the ending of a polysyllable. every instance of resolved arsis or thesis in Two such syllables must constitute either Plautus, but for every instance falling within a resolved arsis or a resolved thesis—pec\t6ra the scope of these six generally recognized mul\c6nt, or pic\tora mul\cent—the law being rules—that is to say, for the overwhelming broken in either case. majority of instances. We are quite justi- 4. The first syllable of an anapaest may fied therefore in entertaining the presumption not be the last syllable of a word of more that the law is general, and in seeking to than one syllable.3 prove that it is so. To be more precise, these A syllable so placed would be the first of six rules prove that the law is observed a resolved arsis or of a resolved thesis— whenever the resolved arsis or thesis begins mag\nd nolup\tas, or mdg\na uoMp\tas. The in one of the last three syllables of a word law is obviously broken in either case. of more than one syllable. There are three 5. The first two syllables of a dactyl may cases in which they fail to prove that the not form one word, or the ending of a law is observed: word. 1. When the resolved arsis or thesis begins A dactyl so formed must be rising—carus in that part of a polysyllable that precedes the chief accent—cdlamitdtem, sapiintia; 1 As these rules are generally recognized, it is 2. when it begins in the first syllable of a not, perhaps, necessary to refer to places where word, or word-ending, of the measure of they may be verified. The rules are given in Prof. muliere; 3. when it begins in a monosyl- Lindsay's Capliui 2, Introd. iii. Most of them are also to be found in Klotz's AltrSmische Metrik. In lable—quid agis, quid id 6st. some cases Klotz was driven by the exigencies of In case 1 the assumption that the syllable his theories to deny generally admitted, and indeed under the ictus would have borne the obvious facts. Thus he states (p. 274) that the ac- secondary accent in prose is obvious, and it centuation _ i, „ (of a dactylic word) is permissible in any part of the line; but of the fifty-two in- is strengthened almost to the point of cer- stances there adduced, forty-five are in the first tainty when we find that Plautus and foot, and the remaining seven are in Cantica. Terence observe strict rules in their metrical Maurenbrecher also questions some of these rules treatment of such words. Thus words of (Hiatus u. Verschleifang, pp. 25 ff.); but succeeds in proving nothing more than that in the text, as the measure of calamitatem normally re- we have it, there are some exceptions. Which was ceive the ictus on the first and fourth to be expected. 2 Cf. Lindsay, I.e. p. 74. 4 Klotz (p. 263) seems to admit _ •,, „ ... ; but 3 Klotz, to suit his theory, allows ... | i, „ _ | . all his examples are in the first foot, except Aut. But all his instances are of the type of Ptrs. 653 : 139, which is in a Canticum. m&is pater ubi me sciet, where he unnecessarily 6 De Correptione Iambica Plaut., Lund, 1901, p. scans ubi{p. 257 ff.). THE CLASSICAL REVIEW. 33 syllables, not on the second and fifth— Case 2 offers no difficulty. It is now cdlamitdtem, but not caldmitaUm. Accord- generally admitted that words ending in ing to Prof. R. S. Radford,1 words of that four short syllables were accented on the measure occur nearly 200 times in Plautus fourth syllable from the end in the time of alone; but there are only four or five places Plautus, muliere, pu&ritia. The accentua- in all Plautus and Terence where they seem tion muliere is normal in Plautus, muliere to be stressed otherwise than on the first occurring only in arguments (e. g. Pseud. and fourth syllables. The offending instances Arg. I. 7), prologues, or in lines otherwise are Gist. 3 apSritistis (in a bacchiac line), suspect.3 where we should restore apertuistis; Heaut. In case 3, when the resolved arsis or 906 dp&rilere, where restore operiuere (cf. thesis begins in a monosyllable—quid agis, Men. 550); Andr. 941 r&Ugtone, where scan quid hdc quod—it is evident that the law relligione; Mil. 562 n&n m&Mttose twmen, was obeyed if the monosyllable was accented where the right order is possibly mdlitidse in prose more prominently than the word or ndh tam&n. The stressing cdlamitdtem may words that followed it, that is to say, if it therefore be said to be invariable. The bore what is called a sentence-accent. The accent of cdlamitas, and words and word- question of sentence-accent arises whenever endings of that measure, in republican Latin, two or more words occurring in a sentence seems to have become the secondary accent are grouped together for the purposes of of cdlamitdtem; and it seems to be generally accentuation, with the result that the true that, where the position of the accent proper accent of one or both or all the varied in the different parts of an inflexional words forming the group is modified. For system, a syllable which lost the principal example, if in the spoken sentence quid agis 1 accent acquired a secondary accent. This the two words of which it is composed were is further indicated by the metrical treat- grouped together, so that either the proper ment in Plautus of words of the measure of accent on the first syllable of agis was commoditatem. In such words the ictus suppressed or some other modification of the falls either on the second and fifth syllables, proper accent of one or both words followed, commdditaUm (e. g. Poen. 916, Ter. Andr. then the accent (or accents) of the group 569), or on the first and fourth syllables, might be called a sentence-accent (or accents).* c&mmoditdtem (e.g. Men. 140, Mil. 1383). We have little certain knowledge of This double stressing is explained if we sentence-accents in Latin, but we have suppose the (principal) accent of commd- some ; and if, where we have certain know- ditas to have become the secondary accent of ledge of a sentence-accent, we find that commoditatem. When the ictus fell as in Plautus has regard to it in forming resolved commdditaUm, the syllable that bore the arses and theses, that fact will increase the secondary accent began a resolved arsis : probability that in some other cases also, when it fell as in c&mmoditdtem, the same where its existence is less certain, such syllable began a resolved thesis. Similarly, an accent was heard and played its part iiie accent of dliquis became the secondary in Plautine metric. It is known, for in- accent of dliqudntum, which explains the stance, that a preposition and a follow- stressing praeterea dliquantum animi caiissa, ing noun ' governed' by it were accented, Trin. 334. In this way inflexion would and sometimes written, as one word (Quint, cause a minimum of accentual disturbance; i. 5. 27). This inforo, inurbe, anUpedes, thus detinui, dettnuisti (probably not, in antededis, dnteUmen, prdpediem, ergdpatrem, prose, dbtinvAsti). This conclusion is further ergamdtrem, propUreos, and so forth. In supported by the evidence of the Romance such cases the proposed law is obeyed by languages. Thus the doubling of I in the Plautus and Terence. Thus we find such Italian scellerato (beside scelesto) is an effect stressing as dnte ped&s, Mer. 780 (where of the secondary accent of Latin sc&lerdtus, the first syllable of the resolved thesis would which had the same position as the (prin- 2 bear the chief accent in prose, anUpedes : cf. cipal) accent of sdlerat, scSlere, etc. It is, the metrical stressing diuitids beside the at the very least, a significant fact that we prose pronunciation di/uAtias); intir eos, Trin. do find in Plautus the stressing commddi- 623; inter eds, Cas. 561; prdpter amorem, taUm, but do not (normally) find a stressing Mil. 1284 (where the first syllable of the caldmitaUm. 3 Cf. Lindsay, op. cit. iii. 5 (2), p. 158. 1 'The Latin Monosyllables in Relation to Ac- 4 1 should perhaps apologize for defining a term cent and Quantity.' Trans. Amer. PhU. Association, so generally used; but if it is an error, it is an vol. xxxiv. 1903, p. 66 ff. error on the safe side, and the term is not always 2 Cf. Lindsay, Lot. Lang. iii. 7, p. 160. used strictly. NO. CLXXIV. VOL. XX. 34 THE CLASSICAL REVIEW. resolved thesis would bear the secondary ac- not be proper for me to assert that aedibus cent in prose, propteram&rem, the secondary has its final syllable metrically lengthened accent in this case occupying the position of in this verse, unless there is independent the principal accent in propte'reos, etc. : cf. evidence for that view. I am aware that commdditas : commoditdtem); per ann&nam Ritschl, with, it must be confessed, more of caram, Stich. 179 (in prose pe'r&nndnam). It emphasis than of argument, pronounced is certain, again, that interrogative pronouns against this metrical lengthening in Plautus,4 and adverbs were prominently accented.1 and that he has been followed by other When therefore we find that such stressings as scholars; but the question is to be decided quis homo, quid agis, quid agdm are normal solely by an appeal to facts. in Plautus and Terence (invariable, I think, at The most striking fact in connexion with least in those parts of the line to which the the occurrence of short syllables in arsis in proposed law applies), we are entitled to Plautus is that, by a rule strictly observed, claim that here too regard was had to a the succeeding syllables must be of such a sentence-accent. Of all the monosyllables kind that lengthening of the short syllable found beginning a resolved arsis or thesis, quid in arsis shall be at least possible. First, the is by far the commonest; but there remain next following syllable may be long; as in a number of instances in which other mono- As. 250 fingeri falldciam, Pseud. 355 syllables, such as et, ut, hie, si, occur in that pr&meri possum domo, Men. 842 Idmpadibus position. That those monosyllables bore a ardintibus: cf. Men. 327, 506, 900; Most. sentence-accent is less certain than that 402, 849, 1118; etc. In these cases the interrogatives did so; but seeing that in lengthening of the short syllable in arsis must other cases, where we have certain know- be admitted, unless the text be declared ledge of a sentence-accent, we find resolved unsound. But farv more significant than arses and theses beginning in syllables that such instances is the fact that, wherever a bore it, it is at least extremely probable syllable can be said to remain short in arsis, that in the comparatively small number of it must be followed not merely by one, but resolved arses and theses beginning in such by two short syllables, capable of forming a words as et, ut, Me, si, the proposed law resolved thesis. The rule is, that, when the holds good.2 This disposes of the three final short syllable in arsis is followed by a cases in which the preceding demonstration word beginning with a short syllable, that fails definitely to prove that the law was short syllable must bear the primary or observed. secondary prose accent (Klotz, I.e. p. 255 f.); Nothing remains therefore but to consider from which it follows that the next fol- the one apparent exception, the case in lowing syllable must be either short, or which a resolved arsis seems to begin in a shortened under the law of Breues short final syllable, e. g. Most. 402 tamquam Breuiantes. Thus, dicer6 uolui tibi would si nitus natus nemo in dedibus habitat.— be a good verse-ending; but dicer6 modestd licet.3 If the final syllable of aedibus modd would be impossible. Even filius in remained short in this verse, the proposed exircitum (cf. Am. 102), though uncommon, law was broken: if it was metrically is quite possible, because the scansion Ine'xe'r- lengthened, the law was not broken. It is citum is not excluded (in- having a secondary tempting to a theorist to assert anything accent); but quam filius amabat tamen which supports his own theory; but he must is not possible. There are many hundreds be on his guard to make no assertion that is of instances in Plautus of the placing of a short final syllable in the arsis; but there is not supported by evidence independent of 5 the theory he is advocating. Consequently, not one exception to this rule. I do not though there is nothing incredible or even see how it is possible to resist the conclusion unusual in the lengthening of a short final that in all these cases the poet intended the syllable by the power of the ictus, it will syllable placed in arsis to be metrically lengthened, so that the two following short 1 syllables could form a resolved thesis. Klotz For the testimony of the grammarians on this point, see Lindsay, op. cit. iii. 12 a (5). perceives to what conclusion the facts point; 2 Cf. Radford, op. cit., who, using an entirely nevertheless he rejects it on the a priori different method from that here followed, comes to grounds that such lengthening is not proper the conclusion that, wherever a short monosyllable causes shortening of a succeeding syllable by the law of Breues Breuiantes, sid dpstnlisti, it a me, 1 Prol. Trin. p. 185. the metrically stressed monosyllable would have 6 For Poen. 1194, Stick. 55, 737, see Klotz, p. 265. borne a sentence-accent in prose. The last-mentioned occurs in a scene which is really 3 For a large number of instances, see Klotz, a Canticum: observe the abnormal ictus amabilis, Altr. Metr. p. 257 ft. causing abnormal shortening of the preceding a. • THE CLASSICAL REVIEW. 35

('eigen') to drama, and does not occur in riuenidt sen&x, Cure. 86 ricipidt mark, Phor, Greek drama (I.e. p. 265). 507 rdtinedm scid, Men. 550 operuit fords. Another significant fact is this, that a It cannot be an accident that in five of these final short syllable so placed is allowed in instances we have a verb compounded with any foot of the line, except the last but one re(d)-. Such verbs had variable scansion in verses requiring an iambic close. Thus of the initial syllable in republican Latin. satis agitdt or dlceri uolui are impossible We have reuortimini in Am. 689, reccludo verse-endings where an iambic close is re- in Gap. 918, redducem, ibid. 923, Bud. 909, quired. Why? Apparently because they rSducem in Gap. 931. Where the metre would have been recited satis | agitdt, etc., decides the quantity, redduco is normal in with an anapaest in the last place.1 Plautus and Lucretius (Munro on Lucr. i. We have another means of testing the 228; Sonnenschein on Bud. 909). It is question whether Plautus regarded a short needless to multiply instances, which are final syllable placed in arsis as metrically well known. In each of the above cases we long, in the special case in which the may scan the first syllable long. There phenomenon is associated with an apparent remains the solitary dperuit foris. This is breach of Luchs's well-known law. That one of several places in which irregularities law, not always rightly stated, forbids an are associated with the perfect of operio or iambus in the penultimate foot of verses aperio. In Gist. 3 ap&ruistis, in Heaut. 906 requiring an iambic close, when the end of opiruere are abnormally stressed, as has been that foot coincides with the end of a word. noticed above. These various irregularities 4 5 disappear if we restore the archaic perfects If, in the line, uolt te n6uos erus operam operiui, aperiul: cf. posiui, saliui, etc. Thus these six instances, so far from proving that dare | tuo ueteri domino, the final syllable Plautus tolerated the ending i „ | „ _/_, | _ _/_, of erus was short, the penultimate foot was afford very strong evidence indeed that he an iambus, and the law was broken; if, did not. The dramatists never write such however, it was metrically lengthened, the verse-endings as muUeris domtim, fidicinde penultimate foot was an anapaest, and dedit, fdcilius tibi. How convenient it the law was not broken. It has been said would have been to do so, if it had been that Plautus tolerates the line-ending metrically possible, is proved by the com- I ^ I « j-> I w _£-• If that can be proved, parative frequency with which modern then lengthening of the final syllable of erus editors have introduced such endings into in the line-ending erus operdm dare" is the text (e.g. Bud. 1247: cf. Luchs, I.e. neither proved nor disproved ; but if it can p. 14). be shown that in no other case he tolerated that ending, then that fact is very strong In type (b), in alium diim, lengthening of evidence for the postulated lengthening. a monosyllabic preposition in arsis is not to Excluding verse-endings of the type erus be thought of, inalium being as much one operdm dare", all other seeming breaches of word as ineo. But it is a necessary con- Luchs's law are divisible into three types 2 : dition in this type of verse-ending that the (a) r&ueniunt domum, (b) in alium di&m, preposition shall be preceded by synaloepha. (c) ddb(o) operdm tibi. At first sight, the There are only seven certain instances instances of type (a) seem to prove conclu- (including those in which ut or at takes sively that Plautus did tolerate the verse- the place of the preposition). They are all of the type of milit(i) in alium diem (Aul. ending iu|uii|ui_. There are five instances of this type in Plautus, and one 530: other instances in Luchs, I.e. p. 16, in Terence: Am. 188 r&ueniunt domum, where, however, some correction is necessary). Poen. 27 riuenidnt domum, Most. 57 That this never absent synaloepha is not 1 accidental, but has a prosodic effect, is I have maintained, in Hermathena xxx. (1904), shown by the conditions found in the p. 140 f., that an anapaest is occasionally found for an iambus in the verse-close. That opinion remaining case in which Luchs's law seems I still hold; but I had not then observed the re- to be broken, namely, in endings of the type striction to which that licence is subject. It is (c), ddb(o) operdm tibi. Omitting one or allowed only when the verse ends with a quadri- two places where the text is not certain, syllable of the measure of perueniant: e.g. atigurid (Ace. Trag. 624 B), diuilids, iuenidnt, sirrupuit. there are nineteen instances of this ending Veniant would be impossible in place of the final in Plautus, and three in Terence (Luchs, pp. iambus; and those editors who postulate peruenant, 13, 15). The necessary conditions are that and the like, give no instances of verse-closes in there shall be synaloepha before the ana- which the simple verb is required in a form *uenant. paestic word, and that either a pause (e.g. a See A. Luchs in Studemunds Studien, i. p. 13 ff. at change of speakers) or a long syllable D 2 36 THE CLASSICAL REVIEW. shall precede the entire sequence—jdctumst, introduced by Plautus that they can always ddb(o) operdm tibi. The presence of this be scanned long, but cannot always be long syllable (an 'inner' thesis), or of scanned short; further, the scanning of such the pause that may take its place, in- syllables as long involves no metrical diffi- dicates that the need was felt of some culties, but the scanning of them as short sort of signal when this ending was used: involves a number of exceptions to Luchs's it breaks the pace of recitation, and allows law. If therefore we accept the guidance the actor a moment's pause. The syn- of the facts, accept, that is, the inductive aloepha always found here evidently had method, we are indisputably led to the con- the same prosodic effect as the differently clusion that short final syllables so placed placed synaloepha in type (b) : it gives (at were lengthened, probably by a pause of one this particular place in the line) weight to mora. The acceptance of that conclusion the properly short syllable which it either would even enable us to introduce a little precedes (b) or follows (c), and allows a order into the scanning of Saturnians. For pronunciation fdc]toimst, dk%6 \ operdm \ tibi example: facile facteis superdses ; Gnaiudd Or mi\lid in \ aliiim \ di6m, though the patri progndtus; insed uonHtum. Inci- hiatus need not have been BO pronounced as dentally, it would also remove the one it is here represented. All that was stumbling-block in the way of Klotz's law necessary was, that the short syllable in that the syllaba breuians and the breuiata Arsis should be in eaeh case weighted a little, must both belong to the same arsis or thesis to allow a momentary pause. It is not to (in cases like pati/r ubi, where Klotz was fee supposed that so abnormal a hiatus, or, compelled by his theory to scan Ubi). rather, synaloepha with such abnormal «fltecte, could be used in any part of the line. If the lengthening of short final syllables It has long been suspected that the point in under the ictus be admitted, the sole seeming the line that immediately precedes the exception to the law proposed in the earlier final iambic dipody had some special char- part of this paper disappears. There is, of acteristics.1 Nor should it be forgotten course, a further step to be taken—to enquire $hat these types of verse-ending are excep- the raison d'itre of the law itself. I hope tional and comparatively infrequent. I to consider that question on a future occa- hesitate to claim that the prosodic effect sion. Meanwhile, thus much may be said. We of synaloepha at this point in the line is do not find its raison d'itre in the supposed, definitely proved; but there is always one the oft-asserted preference of Plautus for unanswerable argument in its favour. If coincidence of the verse-accent (ictus) with Plautus really tolerated a verse-ending the prose-accent. For three reasons : 1. the 6 w I w -£_, I ~ J-> why does he never once, inlaw requires the first syllable of a resolved all bis 19,300 lines of dialogue verse, end a thesis, which has the weakest metrical stress, line with such a collocation as m&lieris to be a syllable that would be accented in domum ? Or why do we never find such an prose (c&nsilitim, intered); 2. the law does ending as sdtis adist tibi 1 Those would be not apply to the first foot of a colon, or to really conclusive instances. But they are anapaestic metre; 3. except in resolved wanting. I conclude therefore that in those arses and theses, and in those cases that seeming breaches of Luchs's law which come come under the dipody-law, there is no under the types r&uenvdnt dom&m, milit(i) evidence in the dramatists of any effort to in alvAm diim, ddb(o) operdm tibi, we have avoid clashing of the verse-accent and the no certain evidence that Plautus tolerated prose-accent. In the limitations of the appli- the ending i „ | „ _/_, | ^ j_. It follows that cability of the law we, naturally, have a clue it is highly improbable that he tolerated it to its raison d'&tre. The law does not apply in lines ending like er&s operdm dare". If to that metre (the anapaestic), or to that he did not, the final syllable of erus must part of the dialogue metres (the first foot) have been lengthened in arsis. which is exempt from metrical ambiguity. In the necessity of avoiding metrical ambiguity, To sum up the evidence in favour of the of providing a sure verse-clue—a necessity in view that a short final syllable under the every form of verse, and in every language— ictus was treated as metrically long by we shall find, I believe, the raison d'&tre of Plautus: we find that such syllables are so the proposed law. 1 Cf. Jacobsohn, Quaestiones Plavtivat: Gottin- CHARLES EXON. gen, 1904.