112. 
 
 UC-NRLF 
 
 CNi 
 
 O 
 CI 
 
 o 
 
 l>- 
 
* 1* 
 
 The Cursus in 
 
 r 
 
 fiediaeval and Vulear Latin 
 
 By Albert G. Clark M,A. 
 
 (A Paper read to the Oxford Philological 
 Society on February 1 8 , 1 9 1 o) 
 
 Price Two Shillings ?tet 
 
 Oxford 
 
 At the Clarendon Press 
 
 1910 
 
GIFT OF 
 JANE Ko^JATHER 
 
The Cursus in 
 Mediaeval and Vulgar Latin 
 
 By Albert C. Clark M.A. 
 
 (A Paper read to the Oxford Philological 
 Society on February 18,1910) 
 
 ) > • 5 > ' '■ ' , ^ 
 
 O > ' , V 
 
 ,. ' •.< %,; i\^> *i o. ^' 
 
 Oxford 
 
 At the Clarendon Press 
 
 1910 
 
HENRY FROWDE, M.A. 
 
 PUBUSnBK TO THE UNIVERSITY OF OXFORD 
 LONDON, EDINBURGH, NEW YORK 
 TORONTO AND MELBOURNE 
 
 • ; . ; • • . . • . 
 •v. .*;/.* :Tc- /.. 
 
PREFACE 
 
 The first part of this paper contains an account of the 
 mediaeval cursus drawn from various works, most of which are 
 mentioned in \\^^V>\k:X\ogx2iy^\iy to v^y Pontes Prosae Numerosae 
 (Oxford, 1909). I have also used with profit Rockinger's 
 collection of documents in Quellen z. bayr. it. deutsch. Geschichte 
 (Munchen,i863),and Gabrielli's valuable dissertation, ZV/ij/^/^ 
 di Cola di Rienzo e Tepistolograjia medievale in Archivio della 
 R, Soc. Rom. di Storia patria^ xi (1888). I need hardly say- 
 that this portion of my paper does not claim to be more than 
 a popular account of the results at which others have arrived, 
 and I should not have thought of printing it but for certain 
 suggestions about the influence of accent in earlier prose, 
 which occurred to me when I was writing, and which, so far 
 as I know, are new. As I wish to obtain the opinion of others 
 upon these points, I venture to publish my paper, with a few 
 corrections and additions. 
 
 I should like to take this opportunity of remedying a very 
 unfortunate omission in the Bibliography previously mentioned, 
 viz. De clausulis Ciceronianis^ by J. Wolff, a pupil of Skutsch 
 (1901). I was well acquainted with this very valuable work, 
 which I have mentioned elsewhere {Classical Review, 1905, 
 p. 165), and the omission was due to accident. I also regret 
 that Bornecque's elaborate and learned work, Les clatisules 
 nietriqnes latines^ was unknown to me. Among other writings 
 which have since come into my hands are Skutsch's important 
 article on Sprache^ Metrik und Rhythmus der Romer in ihren 
 Wechselbeziehtmgen in Zeitschrift f. d. Gymn.- Wesen, 1 909, 
 pp. 67-74, and four works by pupils of his, viz. K. Munscher, 
 Die Rhythmen in Isokrates Panegyrikos (1908), V. Miinch, 
 De Claustdis a Valerio Maximo adhibitis (1909), J. Gladisch, 
 
 238.164 
 

 • ••••••• • •<••■ 
 
 De Clausulis Quintilianeis (1909), and the valuable edition of 
 Firmicus Maternus by K. Ziegler (1908). I would also mention 
 Lehnert's exhaustive Bericht iiber die rhetorische Liter atur 
 bis 1906 in Bursian's Jahreshericht 1909 Band CXLII. 
 
 I have also to correct a serious misprint in the Preface to 
 the same work (p. 6). Zielinski's three Forms, if the molossus 
 base is admitted in 2 and 3, come to 60 per cent., not 
 40 per cent. 
 
 ALBERT C. CLARK. 
 
THE CURSUS IN 
 MEDIAEVAL AND VULGAR LATIN 
 
 The sophist Thrasymachus is chiefly known to us from the merciless 
 caricatures of Plato. In the Phaedrus he is depicted as the * Chalce- 
 donian giant who can put a whole company of people into a passion 
 and out of it again by his mighty magic : and is first-rate at inventing 
 or disposing of any sort of calumny on any grounds or none'. In the 
 Republic he appears as a mere child in the hands of Socrates, and 
 resorts to insolence in order to cover his discomfiture. No one could 
 suspect that he was one of the greatest inventors in the field of litera- 
 ture that the world has seen, who laid down the lines upon which 
 prose was to develop for nearly two thousand years. 
 
 We learn from Cicero, whose information is drawn from Theophras- 
 tus, that Thrasymachus was the first person who deliberately introduced 
 metrical cadences into prose for rhetorical effect. Previous writers had 
 used such numeri by accident only : the pleasing result was noticed 
 and certi cursus conclusionesque verhorum (Orat. 178) now became 
 a part of rhetorical technique. We also learn from Suidas that 
 Thrasymachus irpcoros TreptoBov koI kwXov KarcSct^c Koi Tov vvv rrj^ 
 prjTopiKrj^ rpoTTov ilcrTjyrjcraTo. Greek writers use the term Ko/x/xa for 
 a section of the kG^Xov. Cicero translates KOfx/xa by t'na'sum, kwXov by 
 memdrum, and TrcptoSos by ambitus^ comprehension &c. There is a close 
 connexion between this articulation of the sentence and the use of 
 numeri. The theory of the period is that of a rise and a fall. Rough- 
 ness is permissible as the sentence storms its way onward, but har- 
 mony must be restored at the tcXcvtiJ. Later Greek writers use 
 the suggestive term /cporo?, i. e. * beat ', for the end of the clause or 
 sentence. The Latin word is clausula. Wherever the voice halts, a 
 musical efi'ect is required. The numeri coincide with the beats and 
 reveal the secret of ancient punctuation. 
 
 The invention of Thrasymachus was supplemented by Gorgias, who 
 introduced his well-known a-xnpxiTa, viz. laoKwXa, di/Tt^cra, and 6/aoio- 
 TTTwra. These are considered by Cicero to form a part of the 
 
oratorical numerus?- The pvOixoi of Thrasymachus were combined 
 with the arxrjixaTa of Gorgias by Isocrates, of whom Plutarch says that he 
 spent a lifetime on his Panegyric, pasting together antitheses and 
 balanced clauses with similar cadences, smoothing his periods with 
 chisels and saws and making them musical (pv^/xt^wv).^ Demosthenes 
 makes more sparing use of the a-xvf^oLTa than Isocrates, in whose prose 
 they are so constant as to become monotonous, but greatly developed 
 the use of pvOfxoL Thrasymachus had relied for effect upon the paean, 
 a foot which was thought most suitable for prose, since it was little 
 used in verse. DemxOsthenes preferred the cretic, which, as ancient 
 writers point out, is metrically equivalent to the paean (i.e. — ^^ — is 
 equal to — w wo or ww ^ —). The beginning of the De Corona tois 
 ^cots €vxofj.ai Traa-L Kai became proverbial, and Dionysius of Halicar- 
 nassus remarks that the proem in this, his most carefully written speech, 
 is founded on the Kpr/rtKos pvO/xos.^ He also continually uses a cretic 
 followed by a trochee or spondee (the final syllable being regarded as 
 anceps by ancient metrists) and /xrySc to^cvt; is quoted by Quintilian as 
 an example of his severa composition Other favourite endings with 
 him were a double trochee or double spondee. Cicero held him up as 
 an example to those who object to the use of numeric saying that his 
 lightning was so brilliant because it was winged with numbers.' 
 
 Subsequent writers, especially those connected with Asia, further 
 developed the practice of Demosthenes. The double trochee now 
 became the favourite pv0/x6<s and was used with monotonous regularity : 
 so much so that it became characteristic of the school. The Romans 
 adopted the use of numeri from Asiatic teachers before the Ciceronian 
 era. Thus in 90 b. c. Carbo, addressing the people, said : • 
 
 Quicumque eam violassent, ab omnibus esse ei poenas persolutas. 
 Patris dictum sapiens temeritas filii comprobavit. 
 
 Cicero notes the double trochee, persolu/as, and says of the second 
 double trochee, cdmprdbdvit\ 
 
 Hoc dichoreo tantus clamor contionis excitatus est ut admirabile 
 esset. Quaero, nonne id numerus effecerit ? verborum ordinem immuta : 
 fac sic, comprobavit filii temeritas, iam nihil erit. 
 
 Cicero was at great pains to set forth the theory of the subject in 
 his Orator, a work dedicated to the Atticist Brutus, and intended to 
 convert him to the use of fiumeri. While his account .of the origin 
 
 1 Orator, § 202. ^ ^ « De Gloria Athen. ch. 8. 
 
 ' Tlepl avvdeoecus uvofidrojv, § 204. * ix, 4. 64. 
 
 "* Orator, § 234. « ib. 214. 
 
and object of the practice is admirable, it cannot be said that his rules 
 are lucid. He does not appear to have grasped the secret principles 
 by which he was guided. These were first ascertained by the patient 
 analysis of modern scholars. 
 
 Zielinski, following in the path of previous inquirers, notably 
 W. Meyer, Norden, and Wolff, a pupil of Skutsch, has shown that in 
 Cicero the three favourite clausulae are 
 
 1. — w — — Z/ 
 
 2. — w — — w U 
 
 3. — w — — y^ — v^ 
 
 It will be noticed that the first of these is identical with the /xrySc 
 roievr) of Demosthenes, the second- is a double cretic and is identical 
 with Tots ^coTs evxofJML, the third consists of a double trochee preceded 
 by a cretic as in the stock example quoted from Carbo, Jiln compro- 
 havit. Zielinski explains the three forms as consisting of a cretic base, 
 followed by a trochaic cadence of varying length. It may be remarked 
 that this is stated in so many words by Terentianus Maurus (r. 190 
 A. D.), who assigns to the cretic a sedes heata just before the end of the 
 period, when the clausula is completed by a trochee or dactyl. 
 
 Optimus pes et melodis et pedestri gloriae : 
 Plurimum orantes decebit, quando paene in ultimo 
 Obtinet sedem beatam, terminet si clausulam 
 Dactylus, spondeus imam, nee trochaeum respuo. 
 Bacchicos utrosque fugito, nee repellas tribrachyn. 
 Plenius tractatur istud arte prosa rhetorum. 
 
 These three forms without any licence of any kind account for 44 per 
 cent, of the clausulae in the speeches, 17,902 in number. He also 
 recognizes a strong form of the base in Forms 2 and 3, viz. a molos- 
 sus instead of a cretic.^ The addition of these brings the total up to 
 60 per cent. Various licences are permitted, the chief of which is that 
 any long syllable may be replaced by two shorts, e.g. esse videdtur. 
 Also certain variations in the base are allowed, e.g. a choriambus or 
 epitriton (/r.) in place of the cretic, as in exiguo tempore, publice sub- 
 scribttur ; or the trochaic cadence may be prolonged by another 
 syllable, thus creating a Form 4, e.g. spiritum perttmescerem. This 
 accounts for another 26 per cent., raising the total to nearly 87 per cent. 
 The remaining forms need not detain us now, with the exception of 
 some striking cases in which, for rhetorical effect, a spondee is used in 
 the cadence instead of a trochee, e.g. career em condcmndti, where the 
 light beat is replaced by a stroke from the hammer. 
 
 * I denote the strong forms by Roman numerals (ii, iii, &c.). 
 
8 
 
 Zielinski has since been occupied in investigating the nurneri of the 
 colon. In a letter which I received from him a few days ago he tells 
 me that he has now gone through the speeches and tabulated the cola, 
 130,000 in number, and hopes soon to publish important conclusions. 
 Meanwhile the general theory is clear from a passage in Quintilian 
 (ix. 4. 70), who points out that some clausulae are halting and limp, if 
 left to themselves, but are propped up by the context. He quotes 
 
 • Non vult populus Romanus obsoletis criminibus accusari Verrem/ 
 durum si desinas, sed cum continuatum sit iis quae sequuntur . . . 
 * nova postulat, inaudita desiderat/ sah-us est cursus. 
 
 The harsh numeri are comparatively frequent in the cola, while har- 
 mony is restored in the clausula. It may be noticed that the period 
 frequently falls into strophes, recalling the measures of lyric poetry, 
 which, curiously enough, Latin writers considered not to have a strict 
 metrical system. This arrangement by strophes enables us to express 
 by symbols what Cicero termed concinniias. 
 
 Cicero's system is uniform throughout his writings, with the excep- 
 tion of the letters to Atticus. Many of the letters ad Familiares are 
 just as metrical as his speeches. The CommeJitariolum Petitionis 
 ascribed to Quintus is written in the same style. The strict Atticists 
 of his day, Sallust and Caesar, do not resort to numeri, and 
 Brutus was not convinced by the Orator, which Cicero dedicated to 
 him (Att. xiv. 20. 4). It is interesting to notice that the Letters of 
 Brutus, the genuineness of which has been doubted, are not metrical. 
 Among subsequent writers I find no use of numeri in Cornelius Nepos. 
 The practice of Livy is exactly opposite to that of Cicero. He shows 
 a marked partiality for the forms which are least common in Cicero, 
 notably for the spondaic ending. In this connexion Cicero has an 
 illuminating remark (Orat. 212), ^ cur sum contentiones magis requirunt, 
 exposiiiones rerum tarditatem,' i.e. History prefers a spondaic measure. 
 The compositio of Tacitus was intentionally harsh. He was the 
 Wagner among Roman composers and produced music out of dis- 
 cords. With these few exceptions, and some technical works for 
 which such ornament was unsuitable, all subsequent Latin authors 
 were possessed by the spell. 
 
 Just as the later Greek authors tend to monotony, as compared 
 with the free measures of Demosthenes, so in Latin the tendency is to 
 uniformity of type. This is already visible in the prose of Cicero. 
 Zielinski points out that in the early speeches the molossus base, in 
 Forms 2 and 3, is almost as frequent as the cretic, while in the later 
 
speeches the cretic is considerably more frequent. Quintilian actually 
 includes the. strong form of 2 among the bad clausulae (ix. 4. 104), 
 and it is notably rare in the younger Pliny. So in the fifth century 
 Martianus Capella speaks of the vitium in Cicero's ending ad meas 
 capsas admisero (Div. in Caec. 51). The highly metrical prose of such 
 writers as Minucius Felix (r. 162), St. Cyprian (200-57), and Symma- 
 chus {350-420) recognizes few nuvieri except Forms i, 2, 3 (the cretic 
 base in 2 and 3 being predominant), a Form 4, and the esse videaiur 
 clausula (i^). These are used in all the cola as well as in the clausula 
 proper, the effect being very monotonous. 
 
 The excess to which these writers carried the use of numeri may be 
 seen from a passage in Symmachus {in Valent. ii. § 24-6). I add in 
 brackets the symbols employed by Zielinski : — 
 
 Ventilemus prisca moniimenta (i^), scrutemiir annales (i) : invenies 
 vetustatem paene ignaram fluminum qiiae tenetis (3). Ipsi illi vates 
 exoticis nominibus licenter ornati (i), cum ad Indicum Gangen at 
 Borysthenen Scythicum carmen extender int (2), Nigrum parem maxi- 
 mis ignoratione slluerunt (i^). Nunc primum victoriis tuis externus 
 fliivlQs publkatur (3 /r.): gaudeat servitiite (3), captlvus innotiilt (2). 
 Elementis quodam modo ditatam cerno rempublicam (2), cui omnes 
 incogniti manclpantiir (3), longinqua terra coniungltiir (2). Dum 
 orbis terminos quaeris (i), gentium siistulTsti (3). Parum quiddam 
 naturae superest quod adhuc populus Romaniis inquirat(i). Relahi 
 credis imperium nisi semper accesserit (2). Quid si impetus tuos 
 institutio moenlum non teneret (3) ? nescis transTre desertS, (i). 
 
 We are in presence of a phenomenon similar to that by which the 
 Greek Hexameter tends to become more and more dactylic in 
 character. The luxuriance of Homer gradually dries up and finally 
 in Nonnus (f. 500 a. d.) we find that 25,000 verses yield only nine 
 types of hexameter, while 35 per cent, of his lines consist of dactyls 
 only. 
 
 Sidonius Apollinaris (r. 431-82), Bishop of Auvergne, carried on. 
 the tradition of Symmachus, whose rotunditas he praises and sets 
 before himself as a model. The prosody of these accomplished men 
 is correct, as is to be expected in persons of their culture. The only 
 deviation which Symmachus makes in the passage quoted above from 
 the prosody of Cicero is that he appears to scan cerno. Other writers 
 long before this time wrote in a very different style. Foreigners and 
 barbarians found it extremely difficult to master the quantity of Latin 
 vowels. On the other hand the theory of the Latin accent is very simple, 
 the chief point being that words in which the penultimate is long are 
 
 B 
 
lO 
 
 paroxytone, e.g. perfrigit, and those in which it is short are propar- 
 oxytone, e.g. vincula (Orator, § 58, Quintil. i. 5. 29). We soon find 
 strange scansions, such as those of the Spaniard Juvencus {c. 330), who 
 has Idborum, ldtictbus,/uiurum, and ends the hexameter with, e. ^..prae- 
 sentai, Aegypti, &c. Still earlier we have the astonishing hexameters 
 of the Syrian Commodianus (r. 238), who begins his Instructtones with 
 
 Prima praefatio nostra viam errant! demonstrat 
 Respectatque boni cum venerit secuH metra. 
 
 The first line of his Apologeticum is 
 
 Quis poterit unum proprie deum nosse caelorum ? 
 
 Teuffel says of him that * his hexameters in defiance of metre and 
 prosody merely follow the ear and the accentuation of every-day 
 pronunciation '. 
 
 It is only natural that the prosody of prose-writers should exhibit 
 similar defects. In order to appreciate this point it is necessary to 
 consider the relation of accent to ictus in the clausula. In Cicero 
 these as a rule coincide, as in — 
 
 (i) — w — ^ o, e.g. vincla perf regit. 
 
 (2) — «-» — ^ v^ -, e. g. vincla perf regerat. 
 
 (3) — s-» — ^ «^ — w, e.g. vinculum /regerd?7ius. 
 
 It will be observed that in the last instance I have placed a minor 
 accent over the first syllable oifregeranius. This is in conformity with 
 all modern writers, who show that long words cannot be pronounced 
 without the help of minor accents. (Cf. Lindsay's Latin Language^ 
 p. 161.) 
 
 While agreement is normal, there are certain cases in which there is 
 a conflict, as in Zielinski's second Form where a molossus is base and 
 there is no caesura, e. g. vincentem fre'gerat, or when the clausula ends 
 with an iambic dissyllable. It is doubtless due to the tug of the 
 accent that Martianus Capella censures Cicero's ending consul videt 
 (Cat. I. 2). 
 
 The result of the enfeeblement of quantity and the stress of the 
 accent was to produce what some writers have called a cursus mixtuSy 
 a very convenient term which means that some of the clausulae are 
 metrical, while others follow the accent without regard to the quantity. 
 All that is necessary is to have the accents in the right place. The 
 result is that the metrical prose of St. Cyprian, Symmachus, and 
 Sidonius gives way to accentual or rhythmical prose. Form i is 
 succeeded by a rhythmical equivalent consisting of five syllables, e.g. 
 
II 
 
 £/ntiS humdnuj?i, Form 2 to one of six, e.g. bona remidia, and Form 3 
 to one of seven, e.g. fdciuni meritSruin. We also find accentual 
 equivalents for other metrical clausulae, e.g. Dictor r^diiHrus = esse 
 vjdeatiir (i^), and (fxceU)e'fitiae ve'sirae scribere =■ opimo zure con- 
 Ugit (4). 
 
 This mixture of metre and rhythm is found in various writers during 
 the latter half of the fourth century. Some of these are Pagans, 
 e.g. Ammianus Marcellinus and Vegetius, others Christian, e.g. 
 St. Ambrose, St. Jerome, St. Augustine. 
 
 I quote the following specimen from St. Jerome (a.d. 331-420), 
 Epistles xxiii. I mark the clausulae by accents, except in the case of 
 two, viz. angeloruvi choris and palmatum consulem, which are metrical, 
 and do not give a clausula by the other method. In several cases 
 accent and metre agree, others are purely accentual. 
 
 Nunc igitur pro brevi labore aeterna beatitiidine fruitur (a), 
 excipitur angelorQm choris (2), Abrahae sfnibus c6nfovdtur (3) et 
 cum paupere qu6ndam Ldzaro (4), dlvitem piirpurdtum (3), et non 
 palmatum consulem (ii) sed atratum stillam digiti minoris c^rnit 
 inqufrere (2). O quanta r^rum mutdtio (2) ! Ille qui ante paucos 
 dies dignitalum omnia culmina praeced^bat (3), qui quasi de subiectis 
 h6stibus triumpharet (3), Capitolfnas asc^ndit arces (3), quem plausu 
 quondam et tripudio populus Romanus exc^pit (i), nunc desoldtus 
 et nudus (i) non in lacteo cdeli palatio {2), ut uxor mentftur inf^lix 
 (i), sed in sordentibus t^nebris cbntin^tur (3). 
 
 I do not find any systematic use of clausulae in his translation of 
 the New Testament. 
 
 St. Augustine (350-430) says of his own practice *ego in m^o 
 el6quio (2) quantum modeste fieri arbitror (2) non praetermitto istos 
 niimeros clausularum' (3). In another passage, when commenting 
 upon the Epistle to the Romans xiii. 14, he says that if the Latin trans- 
 lator had stooped to such artifices he could have produced a clausula 
 numerosior^ but that he preferred to keep the order of the original. 
 In the course of his remarks, he mingles rhythmical forms, e.g. 
 locutus apostolus (2), cur r ere numerose (3), with metrical, e.g. lege 
 compdndt (i), affirmdre non audeo (2). 
 
 The prose of the Syrian Ammianus Marcellinus (330-400) is much 
 more accentual than that of Jerome, as the following specimen will 
 show (xvii. 5. 3) : — 
 
 Rex regum Sapor pdrticeps slderum (2), frater S61is et Lunae 
 (i), Constantio Caesari fratri m^o (3) salutem plurimam dfco (i). 
 Gaudeo tandemque mihi placet ad optimam vlam te r^vertfsse (3) 
 
et incorruptum aequitatis agnovi'sse suffragium (2) rebus ipsis ex- 
 p^rtum (i), pertinax alidni cupiditas (2) quas aliquotiens ediderit 
 strages (i). Quia igitur veritatis oratio (2) soluta esse ddbet et 
 Ifbera (2) et celsi6res fortiinas (i) idem loqui decet dtque sentfre 
 (i), propositum mdum in pauca c6nferam (4), reminiscens haec quae 
 dictunis sum me siepius replicasse (3). Ad usque Str^mona fliimen 
 (i) et Maced6nicos ffnes (i) tenufsse maiores m^os (3) antiquitates 
 quoque v^strae testdntur (i), haec me c6nvenit flagitare (3), ne sit 
 ddrogans qu6d adflrmo (3), splendore virtutumque insfgnium sdrie 
 {2) vetustis r^gibus kntistantem (3). 
 
 The Sermons of Leo I (Pope 450 a.d.) are rhythmical: so also is 
 the Leonine Sacramentary, which, whatever its authorship, is of great 
 interest as the oldest collection of this kind. This may be seen from 
 one of the first Collects : — 
 
 Tuas enim Domine virtutes tuasque vict6rias kdmiramur(3), quoties 
 in Eccl^sia Tua (i) horum dierum f^sta celebrantur (i'), quos 
 insignes confess6rum tu6rum (i) et mdrtyrum palmae (i) ad per^n- 
 nem mem6riam (2) solemn^mque laetftiam (2) fidelibus p6pulis skcra- 
 vdrunt {3). 
 
 The prose of Boethius (480-524) seems to me of special interest. 
 His theological writings are in the usual cursus mixtus. Thus in the 
 de Trinitate I find such endings as nUmerum fdcit (1), pluralitatemque 
 perficere (2),/dciunt meritSrum (3), unitdium repetitio (2''). On the other 
 hand his Philosophiae Consolatio, where he is modelling himself upon 
 Cicero, is metrical. This is interesting as showing that it was possible 
 for one writer to use both methods. The metrical clausulae pre- 
 dominate in the Mythologicon of Fulgentius (480-550), but I find 
 accentual endings such as citius obiiurum (3), collegio nSn donetur (3). 
 Cassiodorus (480-575) writes in a cursus viixtus complicated by 
 assonance and rhyme. Gregory of Tours (538-594) writes accentual 
 prose in which the metrical element is recognizable. 
 
 The letters of Gregory the Great (540-604) are considered to mark 
 the full development of this mixed style. I quote the following speci- 
 men (i. 10): — 
 
 Supplicaverunt nobis Hebraei Tarracfnae deg^ntes (i) ut locum 
 quem synagogae hdctenus habuerunt (3), eum illis nostra quoque 
 auctoritate esset hab^ndi licentia (2). Sed quia pervenit ad nos 
 quod locus ipse sic vicinus dsset eccMsiae (2) ut etiam ad eam vox 
 psall^ntium p^rvenfret (3), scripsimus fratri et coepiscopo nostro 
 P^tro (3) ut si esset ut vox de eodem loco in eccMsiam resonaret 
 (3), ludaeorum celebrationibus privar^tur (3). 
 
13 
 
 Various writings of this period, e.g. the Sermons of St. Innocent I, 
 St. Boniface, and St. Gaudentius, the Pro/essio Fidei of Bachiarus 
 Monachus, &c., are equally rhythmical. 
 
 It is the opinion of the specialists who have worked upon this branch 
 of the subject that rhythm in prose fell into abeyance about this time. 
 Thus the Benedictines of Solesmes say that ' a partir de saint Gregoire 
 le Grand le rythme semble sexiler pour quatre siecles de la prose lit- 
 Uraire\ The statement is one of great importance for liturgical 
 students, since it enables them to fix the date of old rhythmical texts. 
 I have not attempted to check it, since this would involve much labour, 
 and I feel that the question should be left to experts. I would only 
 remark that a Decree passed at the Synod of Toledo in 653, some 
 fifty years after the death of Gregory, is quoted by W. Meyer as an 
 example of accentual prose. Also, I should have thought it likely 
 that the old tradition would survive in some places. As we shall see 
 shortly, it was revived four centuries later by teachers proceeding from 
 Monte Cassino, and I should have thought it more likely that the art 
 had lingered on there, than 'that it had been wholly lost. If once 
 forgotten, it would in all probability have been lost for ever. I also 
 notice in the letters of Servatus Lupus, who was educated at Fulda 
 and was Abbot of Ferribres from 841 to his death about twenty years 
 later, a cursus which I cannot distinguish from that current in the 
 seventh century. I quote the beginning of his first letter : — 
 
 Diu cunctatus sum desideratfssime hominum (2), auderem necne 
 excellentiae v^strae scrfbere (4). Et cum me ab hoc officio aliae 
 rationabiles causae (i), tum etiam eo maxime deterr^bat (3), quod 
 posse id contfngere videbatur (3), ut, dum vestram cuperem amicftiam 
 cdmparare (3), off^nsam incurrerem (2). Scilicet quod praepropero 
 et inusitato prorsus ordine (iv tr.) ab ipso familiaritatis miinere 
 inchoaverim (4), qui nee primordia notftiae c6ntigissem (3). 
 
 Here all the clausulae appear to be accentual with a single excep- 
 tion, which, however, admits of a metrical explanation, being identical 
 
 with a variety of Zielinski's Form 4 (i.e. an epitriton — ^ is 
 
 admitted into the base in place of a molossus). I should, therefore, 
 like to see some further discussion of this point. 
 
 We now pass on to the revival of the cursus in the eleventh century, 
 when it was adopted by the Roman Curia and was the subject of 
 elaborate rules. The prose of this period was largely epistolary. By 
 this I mean not merely private letters, but elaborate and courtly com- 
 positions sent to ecclesiastical dignitaries, and diplomatic documents 
 proceeding from the Papal Chancery. The usual term for such com- 
 
positions was didamen, writers were called didatores, their art was 
 known as ars didaton'a, and handbooks giving the rules were styled 
 summa didaminis. These treatises dealt with a variety of subjects, 
 e.g. the proper method of addressing a correspondent, the choice 
 of subjects and their arrangement, rules for the exordium, narratto, 
 condusio^ and general information of an encyclopaedic character. The 
 didatores were very long-winded : thus a usual phrase for ' listen ' is 
 vestrae prohitatis agnoscat discretio. They also use words sola ornatus 
 et bonae sonoritaiis causa. The Pope is to address various persons 
 according to their status. To ordinary persons he sends salutem et 
 apostoh'cam benedidionem, to the excommunicated he says spiritum 
 consilii samSris, to Jews and heretics viam agnoscere ve'ritatis. Rules 
 for the rhythmical order are given in several of these works, and the 
 terms cursus planus, tardus, velox, corresponding to Forms i, 2, 3, 
 come into use. It may be observed that 'planus' is used in the sense 
 of * ordinary ', just as cantus planus denotes plain song in the liturgy 
 as opposed to elaborate compositions. 
 
 The first work on XhQ Ars dtdandi was the Breviarium de didamine 
 of Albericus, written at Monte Cassino (1075-ino). He does not 
 give rules for the cursus, but employs it throughout, and says that 
 prose should be sofioram et dtsiindam, id est quasi cur rent em. 
 
 When Urban II was elected Pope in 1088 he went to Monte 
 Cassino, where Albericus was teaching, and chose as his secretary 
 Johannes Caietanus in order that 
 
 antiqui leporis et elegantiae stylum in sede apostolica iam pene 
 omnem deperditum sancto dictante spiritu lohannes Dei gratia 
 reformaret ac Leoninum cursum lucida velocitate reduceret. 
 
 The phrase Leonine has here been interpreted as meaning in the 
 style of Leo the Great, but more probably it refers to the rhythmical 
 nature of the prose in question. Thus another writer uses the word 
 as a synonym for consonantia. 
 
 Caietanus subsequently became Pope in 11 18 under the name of 
 Gelasius II. The cursus is found fully developed in his own letters 
 as well as those which he had written for Urban II, and was adopted 
 by his successors. The rules were published by Albertus de Morra, 
 who became Pope in 1187 under the name of Gregory VIII. From 
 this time onwards the term stylus Gregorianus is used to denote that 
 used by the notaries of the Curia. 
 
 Gregory VIII adopts the mediaeval view that every dissyllable, 
 whatever its quantity, is a spondee, e. g. mare, prdebe, amd, Roma are 
 
15 
 
 equivalents. A trisyllable, if paroxytone, e. g. timdris, is a spondee 
 and a half, if proparoxytone, e.g. negligens ox famulus^ is a dactyl. 
 A long word like excommunicationis consists of four spondees. Rules 
 are given for the composition of a sentence. It is best to begin with 
 a spondee followed by a dactyl, e. g. Deus omnium. It is allowable to 
 employ continuous spondees, but not dactyls, e. g. negligens famulus 
 dliquis, since this makes the movement too rapid. The last word in 
 the sentence should be preceded by a dactyl, e. g. gdudia pervenire, 
 a canon which is obviously identical with the rule of Terentianus 
 Maurus, who gives that place to the cretic, the quantitative predecessor 
 of the dactyl, e. g. gdudium pervenire. The examples given by 
 Gregory belong to the cursus planus and velox. Rules for the cursus 
 tardus are added by his contemporary Transmundus, Vice- Chancellor 
 of the Roman Church. 
 
 A large number of treatises based on that of Albericus were com- 
 posed in Italy and elsewhere. Hugo of Bologna (c. 1124) replies to 
 those who Alherici monachi viri eloquentissimi librum viciant, probably 
 referring to Albertus of Samaria and others who denounced those 
 * scholasticos seu grammaticos dictatores qui dictamina more metrorum 
 seu rithmorum cursitare sive claudicare coegerunt '. Bonum of Florence 
 {r. 1 2 10) wrote a work called Candelabrum, the object of which was 
 to givQ populo dudum in tenebris ambulanti lucidissimam dicendi peritiam. 
 Boncampagnus, also of Florence (r. 1220), wrote an encyclopaedic 
 work, the various books of which had fanciful names such as Oliva, 
 Cedrus, Myrra, Rota Veneris. Thomas of Capua, Chancellor to 
 Gregory IX (1227-41), wrote a Summa dictaminis secundum Curiam 
 Romanam. A minor work by Galfridus de Vino Salvo deserves men- 
 tion for his verses modelled on the beginning of the first Georgic : — 
 
 vobis referam quo sidere vestrum 
 dictamen lucere queat, quo clausula possit 
 lascivire gradu. 
 
 They found imitators in other countries, r Thus in Germany we 
 have the Saxonica summa prosarum dictaminis, in which we are told 
 praecipua sit industria dictaturo ut siilum qui Romanus dicitur teneat 
 et observet. Ludolphus of Hildesheim gives mnemonic rules for the 
 three forms of the cursus. Conradus of Mure de arte prosandi dis- 
 tinguishes prose from verse in the following lines : — 
 
 Vultque venire metrum tanquam domicellula compto 
 
 Crine, nitente gena, subtili corpore, forma 
 
 Egregia. 
 
 Prosaicus versus, res grossior. 
 
i6 
 
 The Gallican Church had the audacity to develop a rival school to 
 the Roman, viz. that of Orleans. We have an Ars dictandi Aurelia- 
 nmsis with letters for the use of Bishops of Orleans, and Poncius 
 Provincialis (1200-50), who lived at Orleans, is a writer belonging 
 to this school. The Italian dictatores carried on a fierce polemic 
 against their brothers of Orleans. Boncompagnus wrote, he tells 
 us, in order that viri scholastic i per falsam et super siitiosam Aure- 
 lianensium^ id est Gallicorum, se?itentiam, tanquam naufragantes . . . 
 formam Sanctorum Patrum, Curiae Romanae et imperialis aulae stilum 
 prosaico dictamine studerent imitari. Thomas of Capua speaks in 
 similar terms. I do not profess to have grasped the exact points, 
 where the Orleans school fell into heresy. Bonum of Florence says, 
 sed hoc aliter ah Aureliafiensibus, aliter a sede Aposiolica observantur. 
 Aurelia7ienses enim ordinant dictiones per ymaginarios dactilos et 
 spondeos. . . . Nos verum secwidum au^toritatem Romanae Curiae pro- 
 cedemus quia stylus eius cunctis planior invenitur. This, however, is 
 not at all clear, since we find that Gregory VIII founds his Forma 
 dictandi upon imaginary spondees and dactyls, I should suspect that 
 the differences were invisible except to contemporary experts, whose 
 acuteness may be seen from a reply of Innocent III (Pope 1198- 
 1216) : — 
 
 Litteris ipsis diligenter inspectis ipsi rescripsimus eas tam ex 
 dictamine quod a stylo cancellariae nostrae discrepant omnino falsas 
 esse. 
 
 Probably there was some professional jealousy in the matter. We 
 find scribes from Orleans exercising their art in Rome itself: thus 
 a Johannes Aurelianensis was Secretary to Alexander III (Pope 
 1 159-81), and the Italian dictatores may have objected to the 
 'dumping' of foreign goods on their market. Also there was the 
 perpetual friction between the Gallican Church and the Papacy. 
 
 One of the most interesting writers of this period is Johannes 
 Anglicus, who wrote in Paris {c 1270). There seems to me no reason 
 for connecting him with the school of Orleans, since there was a feud 
 between it and the University of Paris. He distinguishes between 
 three kinds of stylus^ viz. Gregorianus, Isidorianus, and Hilar ianus. 
 Of the first he says * in hoc stilo considerantur pedes spondei et dactili, 
 id est pedes cadentes ad modum spondeorum et dactilorum. In stilo 
 Tulliano 71071 est observa7ida pedum cadentia set dictionum et sententiarum 
 color alio'. The stylus Isidoria^ius is said to be based on the Soli- 
 loquies of Isidore, in which the clausulae correspond secutidum 
 
17 
 
 leoninitaUm vel consonanttam. This style, he says, is valde motivus 
 adpteiaiem, i. e. useful for sermons. He gives as an example 
 
 prius legunt quam sillabicent^ prius volant quam humi currant. . . , 
 Prius monies scandunt quadrivii quam per valles incedunt trtvii, 
 volant ad astra nee pennas posstdent. 
 
 This series of balanced antitheses reminds us of the o-xiyftara of 
 Gorgias. It is singular to find the methods of the great Sicilian 
 suddenly revived in the thirteenth century. 
 
 The stylus Hilarianus is so called as modelled on the hymn 
 
 ascribed to Hilary : — 
 
 Primo dierum omnium 
 
 Quo mundus exstat conditus. 
 
 In this each verse consists of four iambics, or, as Johannes Anglicus 
 scans it, of 2\ spondees and a dactyl. He gives as an example 
 a letter to be sent to the Archbishop of Paris by an Archdeacon who 
 failed to attend a synod. This runs as follows : — 
 
 Cum essem in itinere^ tendens ad vestram sinodum^ caput meum 
 infirmitas oppressit ita sUbito, quod despero resHrgere portumque 
 vitae tdngerCf nisi Dei dementia me visitare mdrcidum condignetur. 
 Quare pater mitissime, vestra dignetur gratia infirmo mihi cdmpati 
 meque languentem hdbeat excusatum. 
 
 How different from the laconic excuses with which we are now so 
 famihar ! 
 
 The orthodox stylus Gregorianus became universal in Papal Bulls, 
 letters, privileges, dispensations, indulgences, and excommunications. 
 It was used in sermons, prayers, collects, chants, and graces. We 
 also find it in non-ecclesiastical literature, as I shall show shortly. 
 
 Before passing on to other subjects it will be convenient here to 
 state the definitions of the cursiis given by mediaeval theorists and to 
 explain the terminology employed by them. 
 
 Gregory does not himself employ the terms planus^ tardus^ velox^ 
 but gives examples which fall under the planus and velox, Trans- 
 mundus speaks of three kinds, viz. planus, velox, and ecclesiasticus. 
 The last of these was styled tardus by his successors. The examples 
 given by Gregory which belong to the cursus planus are audiri com- 
 pillunt and confide'nter aude'bo : those which fall under the velox are, 
 gaMia pervenire, dgere nimis diire, sufficiant ad voldtum. His general 
 theory is that ' finales dictiones semper debet quasi pes dactilus ante- 
 currere '. It will be noticed that this exactly corresponds to the state- 
 ment of Terentianus Maurus about the sedes beata assigned ^ paene in 
 
 c 
 
i8 
 
 ultimo^ to the cretic. He explains gaUdia pervenire as a dactyl 
 followed by a paroxytone tetrasyllable, which may bfe replaced by 
 two dissyllables (nimis dUreY or a monosyllable and paroxytone 
 trisyllable {ad voldtum), Transmundus gives as examples of the 
 ecclesiasiicus or tardus the forms operdri iusiitiam dirig^ntur in exitus. 
 These he explains as a tetrasyllable (or monosyllable and trisyllable) 
 with a proparoxytone accent preceded by a word with a paroxytone 
 accent. Gregory's account of the two forms which belong to the 
 planus, viz. audit icompe Hunt, confidenter audebo, is somewhat confusing. 
 He explains these as a final trisyllable (without mentioning that this 
 is paroxytone) preceded by a trisyllable or tetrasyllable. As a matter 
 of fact, however, the number of syllables in the preceding word is not 
 the determining point, but the paroxytone accent. Thus nomen 
 assHmenSf sinit int/rdum equally belong to the cursus planus. 
 
 Recent writers have noticed other forms to which the dictatores did 
 not give names. Grospellier* mentions the following, which he 
 describes in terms modelled upon mediaeval theory : — 
 
 (i) tr. = trispondaicus, e.g. d6na sentiamus. 
 
 (2) dd. = dispondeus dactylicus, e. g. virtutis operdtio. 
 
 (3) o. = octosyllabicus, e. g. fldtibus supplicdntium. 
 
 (4) m. = medius, e. g. ( iugiter p6stulat, 
 
 - pr^cibus n6stris, 
 , fov^mur m^ritis. 
 
 The mediaeval method seems to me very cumbrous and it has led 
 to much misconception. The results can be stated more clearly 
 by the help of the symbols which Zielinski has invented for the 
 metrical clausula out of which the accentual is developed. He 
 considers the Form as a metrical whole and distinguishes the types 
 
 by means of the caesura. Thus Form i (— w \j) has, five possible 
 
 types, viz. iudicabatur, non oportere, mprte vicistis, civitas possit, restt- 
 iutus sit, which he terms fa, i fi, ly, i 8, i c. The other Forms he 
 treats in the same way. Some types are much more frequent than 
 others. His statistics show that in Cicero y is the favourite type of 
 Forms i and 2, while 8 is characteristic of 3. In the middle ages 
 these types have become predominant. The cursus planus, e.g. 
 sinit int/rdum, is identical with 1 y, the tardus, e. g. cri^cis affixiOf with 
 2 y, and the velox, e. g. gaUdia pervenire^ with 3 S. 
 
 If we apply this method to the less frequent types collected by 
 Grospellier, the first of these, dSna sentiamus, is identical with Cicero's 
 
 * This form contravenes Meyer's law (p. 22). =* Santi, II ' Cursus ', pp. 25-26. 
 
19 
 
 esse videdtiir, which indeed Grospellier quotes among his examples of 
 this type. This is Zielinski's i^ y (L e. the second long syllable in 
 Form I is replaced by two shorts). So mrtHiis opndtio corresponds 
 to esse vtdedmml, i. e. 2^^ y : flitibus supplicdntium is metrically as well 
 
 as accentually identical with Zielinski's Form 4 (— w w — w/ i=^). 
 
 The three varieties of the so-called medius cursus belong to the 8 
 type. Thus tUgiter postulat = cdllidefecerit, i. e. 2 8, precibus nSstris 
 = dvitds pdssti, i. e. i S, \\h\\t/6vemur meriiis appears to correspond 
 to creddtts pdsiulo, i. e. ii 8 (I denote the ' strong ' variety of Forms 2 
 and 3, i. e. with molossus base, by Roman numerals). 
 
 We are now in a position to draw up a list of endings permissible 
 in the cursus of this period, which will, I think, be more intelligible, 
 viz. : — 
 
 I y e. g. audfri comp^llunt {planus). 
 
 i2 y e. g. d6na sentiamus. 
 
 1 8 e. g. prdcibus n6stris. 
 
 2 y e. g. operari iustitiam {tardus). 
 2^ y e. g. virtutis operatio. 
 
 2 ^ ^.g. iiigiter p6stulat. 
 ii 8 e. g. fovdmur mdritis. 
 
 3 h t.%. gaudia pervenire {velox). 
 
 4 8 e. g. fl^tibus supplicantium. 
 
 The variety of the tardus given by Transmundus, viz. dirig^ntur in 
 /xitus, has a minor caesura, and is strictly 2 y 8, but of course in exitus 
 is treated as one word. So the varieties of the velox, viz. sufficiant ad 
 voldtum (where the same explanation applies) and dgere nimis diire, 
 are 'strictly 3 8 e and 38^. The presence of the 8 division is the 
 important point. 
 
 I do not wish to imply that the cursus was always employed 
 strictly according to the rules laid down by the notarii of the Roman 
 Curia. It is, however, unmistakably present in a vast body of litera- 
 ture such as the Policraticus of John of Salisbury (iiio-ii8o), 
 the writings of St. Thomas Aquinas (i 225-1 274), and the corre-. 
 spondence of H^loise and Abelard. I add an extract from a letter of 
 H^loise : — 
 
 Cum me ad temporales olim voluptdtes exp^teres (2), crebris me 
 epistolis visitabas (3), frequenti carmine tuam in ore omnium Helofs- 
 sam pon^bas (i). Me plateae omnes, me domus singulae rbsonabant 
 (3). Quanto autem rectius me nunc in Deum quam tum in libidinem 
 excitares (3). Perpende, qua^so, quae d^bes (i), att^nde quae 
 postulo (2) et longam epistolam brevi fine concliido (i). 
 
20 
 
 The style of sermons may be illustrated by St. Bernard of Clairvaux : 
 
 Quam gloriosi revertuntur vict6res de pradlio (2). Quam beati 
 moriuntur mart^res in pra^lio (2). Gaude, f6rtis athl^ta (i), si vivis 
 et vfncis in Domino (2). Vita quidem fructuosa et victoria gl6ri6sa 
 {3), sed utrique mors sacra iiire praeponitur (2). Nam si beati 
 qui in D6mino m6riuntur (3) non multi magis qui pro Domino 
 m6riuntur (3) ? 
 
 It is stated by the experts who have worked upon this subject that 
 the cursus began to decline under Nicholas IV (1288-92) and that it 
 disappeared from Bulls in the fourteenth century, except in the case of 
 formulae which were reproduced from earlier documents. This, again, 
 is an assertion which I must take on trust, since I have no time in 
 which to verify it. In any case the cursus survived in literature in its 
 most exact form, e.g. in the Latin works of Dante (i 265-1321), the 
 letters of Petrarch (1304-13 74), those of his friend Cola di Rienzo, 
 and the letters and de Mulieribus Claris of Boccacio. I cannot refrain 
 from quoting the famous letter of Dante to the Florentines, which 
 derives additional interest from the fact that W. Meyer has made 
 a certain correction in the text upon rhythmical grounds : — 
 
 O miserrima Faesulan6rum propdgo (i) et iterum iam punita 
 barberies (2). An parum timoris praelibata incutiunt (2) ? Omnia 
 vos tremere drbitror vigildntes (3), quamquam spem simuletis in facie 
 verb6que menddci (i), atque in somniis expergfsci plerumque (i), 
 sive pavescentes infusa praesagia (2), sive diurna consflia rbcoMntes 
 (3). Verum si merito trepidantes (3) insanisse p^nitet c6ndoldntes 
 (3), ut in amaritiidinem pbnit^ntiae (4) metus dolorisque rivuli confluant 
 (2), vestris animis infigdnda sup^rsunt (i) quod Romane rei baiulus 
 hie divus et triumphdtor Henrlcus (i) non sua privata sed publica 
 mundi c6mmoda sftiens (2), ardua queque pro nobis aggrdssus est (2) 
 sua sponte penas n6stras partkipans (2), tanquam ad ipsum post 
 Christum (i) digitum prophetic propheta dir^xerit Isaias (3) cum spiritu 
 dei revelante predixit (i), ' vere languores nostros ipse tulit et dolores 
 nostros ipse portavit' (i). Igitur tempus amarfssime penit^ndi (3) 
 vos t^mere presumpt6rum (3), si dissimulare non viiltis (i), addsse 
 consplcitis (2). 
 
 Punka harharies was formerly printed in the first line. This reading 
 yields a very doubtful sense and offends against the laws of the cursus. 
 W. Meyer emends to punita, referring to the previous destruction of 
 Florence by Totila. It is very difficult to distinguish between / and c 
 in fourteenth-century MSS., but Mercati, who has examined the 
 Vatican MS. which is the sole authority for this letter, is of opinion 
 
21 
 
 that it really gives / not c. Meyer's correction is, of course, absolutely 
 
 convincing. 
 
 With the dawn of the Renaissance the knowledge of quantity 
 revived, and the cursus was abandoned as barbarous. Thus Coluccio 
 Salutati says to a correspondent who had written to him a letter in 
 which it was avoided : — 
 
 Cum omnia michi placeant, super omnia michi gratum est quod 
 more fratrum ille sermo rhythmica lubricatione non ludit. non est ibi 
 syllabarum aequalitas quae sine dinumeratione fieri non solet, non sunt 
 ibi clausulae quae similiter desinant aut cadant. Quod a Cicerone nostro 
 non aliter reprehenditur quam puerile quiddam quod minime deceat 
 in rebus seriis vel ab hominibus qui graves sint adhiberi. Benedictus 
 deus quod sermonem unum vidimus hoc fermento non contaminatum 
 et qui legi possit sine concentu et effeminata consonantiae cantilena. 
 
 Coluccio and the other Humanists failed to observe that the 
 accentual system was based upon one of quantity, and the art of 
 numerosa compositio was lost, only to be recovered gradually during 
 the last few years. 
 
 The development of the Latin cursus cannot be fully understood 
 without some consideration of a similar phenomenon which appears 
 in late Greek prose. Here also in the fourth century a. d. accent 
 becomes the dominant factor, and the rhythm depends upon the 
 interval between the accents. This discovery is due to the acumen 
 of W. Meyer, who shows that from this time onward until the fall of 
 the Byzantine Empire it became the fashion in prose for at least two 
 unaccented syllables to come before the last accent in the kwXov 
 or period. 
 
 This explains the remark of Psellus (1018-1078), who states that 
 the rhythm of prose depends upon the interval between the accents 
 (cCTTt SiaXcKTOV fitXos TTOtov Tt SLOLCTTrjiJia Iv ScacfiopoLS (TvAAajSats jSapv- 
 TovovpLivov 17 o^vTovov/Acvov). The typical clausulae are avras <ro(f>6s, 
 tro^iav ri/xa, a7rdvTU)v dvdptoTruiVy SiaXeyovrat avdpwTroi. In the case 
 of a long word like KjAXoTLfiov/xevo^, there is a minor accent on the 
 third syllable before the last accent. The first two Forms, viz. aTras 
 croc^os, ao^Lov rifia, have no analogies in Latin, since in it the last 
 syllable is not accented : the second, viz. dTravrwv dvOpwiraiv, is identical 
 with the cursus planus {^genus humdnum), and the third, viz. hiaXi- 
 yovrai dvOpniTrot, with the tardus {bona remedid). This last Form, the 
 double dactyl, corresponding to the old double cretic, e. g. <fiaiSp6n)a-i 
 XdfiTTov&av, became very frequent. Meyer quotes a sermon written in 
 634 A. D., in which all the clausulae belong to this Form, so it must 
 
22 
 
 have been' considered valde moHvus ad pietatem. The Byzantine and 
 the Latin cursus therefore exemplify the same principle, viz. a ' binary 
 movement', in which there are at least tvi^o unaccented syllables 
 between the last two accents. In the cursus planus and tardus there 
 are two syllables unaccented, e.g. genus humdnum, bona remedta, in 
 the velox, e. g. fdciunt merii^rum there are four. When the first 
 unaccented syllable ends the last word but one, there may be an 
 interval of three unaccented syllables, e. g. vincla reserdre, correspond- 
 ing to the three short syllables in the metrical Form esse videdtur (i'^). 
 When such a word as reserare does not occupy this place, but comes 
 after a proparoxytone, e.g. vinculum, it is treated as the successor 
 of the ditrochaeus, e. g. vinculum reserdre, its first syllable receiving 
 a minor accent. 
 
 The two articles of Meyer from which I take these observations are 
 most profound and illuminating. He was the first to show the 
 connexion between the Greek and Latin cursus and the common 
 principle on which both were based, to state that metrical prose is 
 founded upon the cretic, and to demonstrate the existence of a cursus 
 mixtus in writers like Ammianus and Vegetius. 
 
 The statement hitherto accepted is that this cursus mixtus first 
 appears in the fourth century a.d., e.g. in the writings of St. Augustine. 
 I have recently arrived at the conclusion that it came into existence 
 much earlier, and was in fact characteristic of vulgar or colloquial 
 Latin, as opposed to the sermo urhanus. 
 
 While I was writing this paper, it occurred to me to glance at some of 
 those authors whose prose is confessedly not metrical, or only partially 
 so. I began with Petronius, thinking that it would be interesting 
 to see how far the bas-monde conformed to metrical rules. In this 
 inquiry I had been preceded by M. ^i^mile Thomas, who found that, 
 although a number of the clausulae were metrical, others were not, 
 and arrived at the conclusion that the textual criticism of the author 
 was not likely to be advanced by such a method of study. I looked 
 first at §§ 70-7 1 and found that, though several clausulae were correct 
 from the metrical point of view, others were recalcitran t, e. g. 
 
 familia. occiipaverat, super me positum cocum, contentus fiiit 
 recijmbere, contentlone Trlmalchlo, testamento meo manumitto, meam 
 heredem faclo. 
 
 Others, though defensible, were harsh, e. g. Circensihus primdm 
 pdlmdm, vicensimam it lectHm stratum. 
 
 It then occurred to me to look at the acce"nts. The result was 
 
23 
 
 surprising, since, where the quantities are wrong, the accents are 
 correctly placed. In order to show this, I add the passage with the 
 accents marked, and insert in brackets the Forms to which the 
 clausulae belong : — 
 
 Paene de \6ci\s dei^cti sumus (3), adeo totum triclinium famflia 
 ^ccupaverat (4). Certe ego notavi super me positum c6cum (i), qui 
 de porco anserem fi6cerat (2), muria condimentlsque fetdntem (i). 
 Nee contentus fuit recumbere (2), sed continuo Ephesum tragoedum 
 c6epit imitari (i^) et subinde dominum suum sponsi6ne provocare (i'*) 
 * si prasinus proximis circ^nsibus prfmam pdlmam ' (3). Diffusus hac 
 contentione Trimdlchio (2), ' amici ', inquit, ' et servi homines sunt et 
 aeque unum lactem bibdrunt (i), etiam si illos malus fatus oppr^ssit 
 (i). Tamen me salvo cito aquam liberam gustabunt {S 2). Ad 
 summam omnes eos in testamento m^o manumftto (i'). Philargyro 
 etiam fundum lego et contuberndlem silam (3). Carioni quoque 
 insulam et vic^simam et lactam stratum ' (3 /r.). 
 
 The only clausula which does not suit the accentual method is 
 liberam gustabunt, which, however, gives the metrical Form ^ 2 (i. e. 
 a spondee takes the place of the trochee in the cadence). As hiatus 
 seems normal in Petronius, I take -cisimam et to correspond to the 
 metrical Form -v^u-, i.e. a choriambic base (/r.) before a double 
 trochee. 
 
 I add another passage, § 103 : — 
 
 Non est dildta fallacia (2) sed ad latus navfgii (2) fiirtim proc^ssimus 
 {2), capitaque cum supercfliis (2) denudanda tons6ri praebuimus (2). 
 Implevit Eumolpus frontes utriusque ing^ntibus lltteris (2) et notum 
 fugitivorum epigramma (i") per totam faciem llberdli mdnu duxit 
 (iii /r.). Unus forte ex vect6ribus (ii /r.) qui acclinatus lateri nivis 
 (i) exonerabat stomachum nausea gr^vem (i)*notavit sibi ad liinam 
 tons6rem (i) intempestivo inhaer^ntem minst^rio (2), bxecratusque 
 6men (3) quod imitaretur naufragorum liltimum v<5tum (i) in cublle 
 rei^ctust{i). Nos dissimulata nausedntis dev6ti6ne (3) ad ordinem 
 tristftiae rddimus {2), silenti6que comp6siti (2) r^liquas n6ctis h6ras 
 (3) male soporati consumpsimus (2). 
 
 The clausula liber dli mdnu duxit seems to correspond to Zielinski's 
 Form iii /r., in which an epitriton (- o - -) replaces a molossus in the 
 basis. So also forte ex veciSribus corresponds to ii /r., in which the 
 same change takes place. I have printed reiidust {reiectus est MS.), 
 since this seems to have been the popular pronunciation (Lindsay, 
 Latin Language, p. 167). I also give the vulgar form minsterio 
 {minisierio MS.), since this seems demanded by the rhythm. The 
 
»4 
 
 evidence for minsterium in vulgar Latin is given by Lindsay (pp. 173 
 and 201). It is interesting to notice that in § 92 in conspe'ctu min- 
 strdntem (i), and § 117 detrectdtor minst&n (2) the rhythm seems to 
 require the syncopated form. 
 
 I would mention in this connexion that the vulgar pronunciation 
 mult&e, for which Lindsay (p. 164) quotes the line of Dracontius 
 (c. 4-90) :— 
 
 Insuper (t Salamon eadem muliere creatus 
 
 seems to occur in § 9 muliere pugndsii (i). 
 
 Caper quoted by Lindsay (p. 163) mentions that words ending in 
 -r, like adhuc, hac, illuc, formed from the enclitic -ce, which have 
 dropped their last syllable, are an exception to the general rule that 
 the last syllable is not accented. So we find § 57 opp6nerent hdc Hide 
 (2), and § 63 adhtic capilldtus issem (3). I previously mentioned that 
 hiatus was a characteristic of vulgar Latin. That this is the rule in 
 Petronius appears from e. g. § 44 dendrios milk atireos (3), § 66 
 excelMnte Hispdnum (i). 
 
 I do not wish to imply that all passages yield results quite so 
 satisfactory as the two which I have quoted. The inquiry is as 
 yet in its infancy, and I am content to indicate the principles involved. 
 I would only say that wherever the eye rests one sees forms like cdve 
 contimnas (i), vdlde sucSssi sunt (2), honesiidres cendbant (i), hdmo 
 negdtians (2), literas didici (2), voca cdcum in midio (2), Trimalchidnis 
 topdnta est (2), sframenticium vdvatdnem (3), tengdmenas facidmus (3), 
 maiiisto et dtgnitdsso (3), ietigerat mdla mdnus (3), cantabundus et sUlas 
 niimero (4), which carry their explanation written on them. Since 
 everywhere the tendency is from multiplicity to uniformity, it will 
 probably be found that the" cursus of Petronius is richer in forms 
 than that of writers in the fifth century. 
 
 While the colloquial parts of the author are based on accent, 
 the more elevated passages are metrical. I quote the following, 
 § 2, using Zielinski's symbols : — 
 
 Grandis et ut ita dicam piidlca oratio (ii tr.) non est maculosa nee 
 turgKda (2), sed naturali pulchritiidlne exsurgit (i). Nuper ventosa 
 istaec et enormis loquacitas Athenas ex Asia commigravit (3 tr.) ani- 
 mosque iuvenum ad magna surgentes (i) veluti pestilenti quodam 
 sidere afflavit (i), semelque corrupta regiila (ii) stetit et obmutmt (2^). 
 Ad summam, quis p5stea Thucyd!d!s (ii tr.), quis Hyperidis ad famam 
 processit {S 2)} Ac ne carmen quidem sani coloris enltiiit (i'), sep 
 omnia quasi eodem cibo pasta (i) non potuerunt usque ad senectutem 
 
25 
 
 canescere (ii). Pictura quoque non alium exltum fecit (i), postquam 
 Aegyptiorum audacia (ii /r.) tarn magnae artis compendiariam in- 
 venit (i). 
 
 I next looked at Vitnivius, whose Latinity is known to have aflBnities 
 with the sermo vulgaris. Some of his linguistic peculiarities are so 
 striking that Ussing attempts to show that the work de Architectura 
 cannot have been written during the Augustan period, but during the 
 decadence of the Latin language and its transition to the Romance 
 tongues. I took at random a passage, ii. 9. lo-ii, alnus . . . hdbent 
 palos. As this appears to be rhythmical, I mark the accents, and add 
 in brackets Zielinski's symbols for corresponding metrical Forms : — 
 
 Alnus autem quae proxima fluminibus rfpis procredtur (i^) et 
 minime materies utilis videtur habet in se egr^gias rationes (3). Est 
 enim aere et igni plurimo tbmperata (3), non multum terr^no um6re 
 paulo (3). Itaque in paliistribus 16cis (i) infra fundamenta aedificiorum 
 palati6nibus cr^bre fixa (3), recipiens in se quod minus habet in 
 c6rpore liquoris (? i^) permanet immortalis ad aet^rnititem (3 /r.) et 
 sustinet immania pondera structurae et sine vltiis consdrvat (?). Ita 
 quae non potest extra terram paulum t^mpus durare (i), ea in umore 
 obnita permanet dd diuturnitatem (3). Est autem maxime id con- 
 siderare Rav^nnae (i), quod ibi omnia opera et publica ^t privata (3) 
 sub fundamentis eius gdneris habent palos (3). 
 
 The passage is certainly not metrical, and seems to exhibit the 
 same phenomena as those visible in the prose of Petronius with the 
 exception of the clausula vitiis cons&vat, which does not fall in ^vith 
 any system unless we consider it a rhythmical equivalent to Zielinski's 
 jP I (- o o - - o). I should not, however, be inclined to accept 
 Ussing's theory concerning the date of the treatise, but should prefer 
 to suppose that it is a specimen of accentual prose as written by plain 
 folk in the Augustan era. 
 
 Another author whom I examined was Frontinus. The works 
 attributed to him are a treatise De aquis urbis Romae, written in 97 a.d., 
 and a book on military stratagems {Strategematicon), of which the 
 fourth book is possibly a composition of a later date. I give the 
 following citation from the first book of the Siraiegematon, ch. 6 : — 
 
 Ventidius Parthico bello adversus Pacorum r^gem (i) non ignarus 
 Pharnacem quendam nati6ne Cyrrh^stem (i) ex his qui s6cii vide- 
 bantur {3) omnia quae apud fpsos ager^ntur (i^) nuntiare Parthis 
 perffdiam barbari (2) ad militates suas conv^rtit (i). Nam quae 
 maxime fieri ciipi^bat (3), ea vereri se ne acciderent qu^ie timdbat (3), 
 ea ut evenirent optare simulabat (i'). Sollicitus itaque ne Parthi ante 
 
 D 
 
* 26 
 
 translrent Euphrltem (i) quam sibi supervenfrent legi6nes (i*) quas in 
 Cappadocia trans Taiirum hab^bat (i), studiose cum proditore egit, 
 uti sollemni perfidia Parthis suad^ret (i) per Zengma traicerent ex6r- 
 citum (ii /r.), qua et brevfssimum iter est (2) et demisso alvep 
 Euphrdtem decurrit (i): namque si Ilia venfrent (i) adseverabat se 
 opportunitate collium usurum ad elud^ndos sagittdrios (ii^), omnia atitem 
 ver^ri (i) si se infra (per) patentis campos pr6iecfssent (ditrochaeus). 
 
 Much of the treatise de aquis is so technical that it can hardly 
 be used for this purpose. I quote, however, the following from 
 ch. Ixxxviii :— 
 
 Sentit banc curam imperatoris piissimi Nervae prfncipis siii (i), 
 regina et domina 6rbis in dfes(i), quae terranim d^a conslstit, (i) cui 
 par nihil et nfhil secdndum (i), et magis sentiet salubritas eiiisdem 
 aet^mae drbis {3), aucto castellorum, operum, munerum et Idcuum 
 niimero (2) nee minus ad privatos commodum ex increment© bene- 
 ficiorum efus diffiinditur (2). 
 
 Armed with these clues I finally looked at Cicero's Letters to 
 Atticus, which are considered by all critics not to exhibit those 
 metrical clausulae which characterize his other works. I selected 
 a letter which has always seemed to me the most private in the 
 collection, viz. iv. 5, in which he unbosoms himself concerning the 
 painful incident of the TroAtvwSta after the Conference of Luca. My 
 surprise was great when I found that the clausulae appear to be 
 accentual. 
 
 Sed valeant recta, vera, hon^sta consflia (2), non est credibile quae 
 sit perfidia in fstis princfpibus (2), ut volunt esse et ut essent si 
 quicquam hab^rent fid^i (i). Senseram, noram inductus, relictus, 
 proi^ctus ab lis (i), tamen h6c eram dnimo (2) ut cum iis in re piiblica 
 c6nsentlrem (3): iidem ^rant qui fiierant (2). Vix aliquando te 
 auct6re resfpui (2). Dices ea tenus te suasfsse ut facerem (2), non 
 ^tiam ut scrfberem (2 ir.). Ego me hercule necessitatem mihi volui 
 imponere huius n6vae coniiincti6nis (3), ne qua mihi liceret labi ad 
 lllos (i) qui etiam turn cum misereri mei debent non d^sinunt inviddre 
 (3). Sed tamen modici fiiimus viroOiati (3) ut scripsi : erimus 
 uberiores si et ille lib^nter acclpiet (2) et ii subringentur qui villam 
 me moleste f^runt habere (i) quae Catuli fiierat (2), a Vettio me 
 emisse non cogitant (2) : qui domum negant oportuisse me aedificare 
 (i), vendere afunt op6rtufsse (3). Sed quid ad hoc, si quibus sententiis 
 dixi quod et ipsi probarent (i), laetati sunt tamen me contra Pompei 
 voluntatem dixfsse (i). Finis. Sed quoniam qui nihil possunt ii me 
 n61unt amdre (i), demus operam ut ab iis qui possunt diligamur (? 3). 
 
%1 
 
 Dices 'v^llem iam prfdem' (i). Scio te voluisse et me asinum 
 germanum fufsse (i). 
 
 Here the only clausula which gives any difficulty is qui possunt 
 
 diligamur, which, however, is metrically correct ( \j-\jue. iii) or 
 
 the ditrochaeus may be regarded as sufficient in itself. In h6c eram 
 dnimo I allow hiatus as in vulgar Latin, and regard eram as unaccented 
 {cf, Lindsay, p. 167, who says that erat, erit, &c., were unaccented 
 or accented according to the caprice of the writer). 
 
 Other letters at which I looked appeared to confirm this analysis. 
 Of course Cicero continually drops into his metrical style. Thus in 
 I. 16, a very private letter, of which he says in § 8, in ea praesertim 
 ipistula quam nolo dliis legi he goes on to refer to a lost speech and 
 uses metrical clausulae. 
 
 I add the following passage from a letter of Caelius, Fam. viii. 6. 3 :— 
 
 Pompeius dicitur valde pro Appio Ikborare (3), ut etiam putent al- 
 terum utrum de filiis ad te missurum (i). Hie nos omnes abs61vimus 
 (2) et hercules consaepta omnia fodda et inhon^sta sunt (4). Con- 
 sules autem habemus siimma diligdntia (2^) : adhuc s. c. nisi de feriis 
 Latinis nullum facere pbtudrunt (3) Curioni nostro tribunatus con- 
 gldciat (2). Sed dici non potest quo modo hie 6mnia iaceant (2). 
 Nisi ego cum tabernariis et aquariis pugnarem (?) veternus civitdtem 
 6ccupasset (3 /r.). Si Parthi vos nihil calflciunt (2), nos non nihil 
 frfgore frig^scimus (2 /r.). Tamen quoquomodo p6tuit (2) sine 
 Parthis Bibulus in Amano nescio quid cohorticularum amfsit (i). 
 
 The only clausula here which calls for comment is aquariis pu- 
 gnarem. This may be regarded as equivalent to Zielinski's P i 
 ^ w w — — — , a rare Form in which a dactyl appears in the base in 
 place of the initial trochee. For the accent of quoquSmodo cf. 
 Lindsay, p. 169. 
 
 It would be a long task to examine the prose of Cicero's corre- 
 spondents and determine whether they write metrically, rh)rthmically, 
 or in neither style. The letters of Plancus are of course metrical like 
 the elaborate compositions of Cicero himself when writing to Lentulus. 
 A short letter of Pompeius at which I have glanced, Att. viii. 11. A, 
 appears to be rhythmical from such clausulae as iter habere, profictsci 
 Cor/inio, cohSrtibus subsequi, Luceriam ve'nias, tutissimo pviio fSre, 
 I find neither metre nor rhythm in the hurried dispatch of Caesar to 
 Oppius, Att. ix. 13. A, or in the elaborate letter of'Antonius, Att. 
 xiv. 13. -4. 
 
 I would conclude this investigation by a few remarks upon the 
 nature of the Latin accent and its relation to the Greek. The Greek 
 
accent was one of pitch, at any rate in classical times, so that there 
 was no conflict between accent and quantity. At a later date it 
 became one of stress. Thus accent is taken into account in the verse 
 of Babrius, and in the fourth century a. d., as Meyer has pointed out, 
 it became the regulating principle in prose. In the case of Latin the 
 facts are wholly different. It is clear from a variety of considerations, 
 such as the syncope of unaccented vowels, vowel changes, the shorten- 
 ing of the vowel in unaccented final syllables, and the accentuation 
 of the Romance languages, that the original accent was one of stress. 
 This is also evident in the indigenous Saturnian metre, which was 
 essentially rhythmical. To quote Lindsay's description of this (p. 128), 
 the line consisted of two halves, the first of which contained seven 
 syllables with three accents, one always on the first syllable of the 
 line, while the second half contained six syllables with two accents. 
 He distinguishes between two types, viz. 
 
 w ^- ~~ ~~~ II ~ " 
 
 e. g. ddbunt vidlum Metelli Naevio poe'tae 
 
 w ~~ ~~ ~~~ II ~~~~ ^~ 
 
 e. g. prim{d) incedit Cereris Proserpma puer. 
 
 When the Romans adopted the Greek quantitative metres, the 
 influence of the accent in shortening unaccented syllables is seen in 
 such lines of Ennius as 
 
 Virgines nam sihi quisque donii Romanus habet sas 
 
 and in various features of Plautine versification. 
 
 The Latin language being essentially rhythmical, was ill suited to 
 the quantitative system. Thus Skutsch says ' Die Natur der lateini- 
 schen Sprache ist stark undactylisch ', and has shown the various 
 devices by which the poets adapted it to the Hexameter {Zeiischrift 
 f,d. Gymn.-W., 1909, p. 68). There would appear to have been 
 a perpetual struggle to lessen the conflict between accent and ictus 
 with the result that certain endings to the Hexameter and Pentameter 
 become conventional. There remained, however, constant collision, as 
 may be seen from the first line of the Aeneid. 
 
 Arma virumque cdno TrSiae qui primus ab oris. 
 
 So the disyllabic at the end of the Pentameter, though possibly 
 smoother than some other endings, causes a perpetual conflict. Mr. 
 Robert Bridges considers that the music of Latin verse is assisted by 
 these discords, since, if accent and ictus always agree, the eff"ect 
 becomes monotonous. This may be true, but I would point out that 
 
29 
 
 there is no such conflict in Greek verse, and no one would say that 
 this was less musical than Latin. Certainly this was not the view of 
 the Romans themselves. 
 
 While the internal evidence shows clearly that the Latin accent was 
 originally one of stress, it is equally indisputable that the Latin gram- 
 marians use terms which are inconsistent with the explanation. Cicero, 
 Varro, and Quintilian are unanimous in speaking of pitch {vocts 
 altttudo, vocts fastigium, &c.), and use precisely similar terms when 
 describing the accent in both languages. It is usual to suppose that 
 they borrowed Greek terms without noticing the essential difference 
 between the two systems, but this cannot be considered a satisfactory 
 explanation. 
 
 Prof. F. F. Abbott of Princeton has recently made a suggestion which 
 appears to me most illuminating (Classical Philology, 1907, pp. 444- 
 60). This is that, when the Romans took over the Greek metres, 
 the Greek pitch-accent was also adopted in the sermo urhanus, while 
 the native stress-accent survived in the sermo vulgaris. This would 
 explain the language used by the grammarians, who wrote only for 
 educated people and did not concern themselves with the talk of the 
 vulgar, and the fact that syncope and vowel reduction were chiefly 
 found in the sermo pleheius. In reply to the objection that orators 
 who did not use the popular accent would not be understood by the 
 masses, we may point out that the same objection would apply to the 
 employment of the sermo urhanus itself, which was essentially artificial 
 in vocabulary and construction. Abbott thinks that when Cicero tells 
 us how the whole theatre would break out into an outcry if any false 
 quantity was made by an actor, he is referring to the Senators and 
 Knights by whom he was surrounded, who would be followed by the 
 rest of the audience. It is interesting to notice that a similar explana- 
 tion occurred to Leonardo Aretino in the fifteenth century.* 
 
 It may also be doubted if the difference between pitch and stress 
 was in practice so great as we might at first suppose. Thus the 
 accent in French resembles a pitch-accent, while our own is emphati- 
 cally one of stress. Would the difference between the two methods 
 of pronunciation have been greater than that of two Englishmen, one 
 of whom had been educated in France and had a French accent ? 
 
 On the other hand, this theory, if stated without qualification, runs 
 contrary to certain facts. If there was no conflict between quantity 
 and accent, how are we to explain the attempt visible in poetry to 
 reduce this to the smallest possible proportions by the choice of certain 
 
 * Epp. vi. 8. 
 
3© 
 
 recognized endings in dactylic verse ? So also in prose Zielinski has 
 shown that the ictus of the clausula generally agrees with the accent, 
 and that where there is a conflict the Form, or the type of the Form, 
 which produces this conflict is never common and steadily becomes 
 rarer. I would, therefore, propose a very slight modification of 
 Abbott's view, viz. that in the sermo urhanus the accent tended to 
 become one of pitch, but that the process was not complete. If so, 
 the conflict would be slight, while it was sufiiciently felt to be noticed 
 and on the whole avoided. 
 
 Abbott's paper contains suggestions which appear to me very 
 interesting in view of the results at which I arrived after studying the 
 accentual system of Petronius and Cicero's Letters to Atticus. Thus 
 he says (p. 457), 'When we bear in mind the great differences in 
 pronunciation, vocabulary, and phraseology which are to be found 
 between the Latin of the freedmen in Petronius and the Latin of 
 Cicero or Quintilian, and when we recall the statements which the 
 Romans themselves made concerning the vulgar tongue, we can 
 readily believe that vulgar Latin and formal Latin were distinguished 
 from each other in the character of their accents.' In a note on 
 p. 458 he says, ' It would be quite possible for Cicero to modify his 
 accent slightly in addressing the people, just as he makes his voca- 
 bulary and his phrases suit the topics which he discusses and the 
 people whom he addresses in his letters.' This is practically what 
 my analysis of such a spontaneous and unstudied document as 
 Att. iv. 5 would show him to have done when writing in his most 
 colloquial style to his most intimate friend. 
 
 I end this paper with a suggestion concerning the meaning of 
 a passage in Cic. ad Att. xii. 6. a. Tyrannic is said to have been 
 the first person who introduced the theory of accentuation to the 
 Romans (Lindsay, p. 151). He sent a book on the subject to Atticus. 
 Cicero had arranged to read it with Atticus, but Atticus, to Cicero's 
 annoyance, read it first by himself. Cicero, after expressing his 
 disappointment, says : — 
 
 Te istam tam tenuem OeoipCav tam valde admiratum esse gaudeo. 
 Etsi tua quidem sunt eius modi omnia. Scire enim vis, quo uno modo 
 animus alitur. Sed, quaeso, quid ex ista acuta et gravi refertur 
 ad TcXos? 
 
 In the Dublin edition etsz . . . tcAo? is translated, * The whole bent 
 of your mind is for subtle speculations. You desire knowledge, which 
 is the oiAy pabulum of the mind. But, I ask you, what in that acute 
 end grave treatise has any bearing on the ultimate principle of conduct?* 
 
•• • • • • • * 
 
 31 
 
 •• • •••••••••'• " 
 
 •••••• • •»• • • • •• • • " 
 
 The editors say that there is a play upon. the 'acute and grave' 
 accents which were the subject of the treatise, and Cicero's work 
 De FinibuSj on the tcXos or summum bonum. This seems to be the 
 recognized view. Professor Reid, however, dissents in a note added 
 by the editors, and says, * I feel sure that there is no allusion to the 
 de Finibus! He explains Tk\o% as referring to the end of Atticus, 
 viz. the acquisition of knowledge. It occurs to me that tcXos may be 
 used for the end of the sentence, i. e. Cicero asks, * What is the 
 relation of the accent to the clausula ? ' 
 
r r < 
 
 < e c i 
 
 c, , < < , 
 
 < < <■ < 
 
 OXFORD 
 
 PRINTED AT THE CLARENDON PRESS 
 
 BY HORACE HART, M.A. 
 
 PRINTER TO THK UNIVERSITY 
 
14 DAY USE 
 
 RETURN TO DESK FROM WHICH BORROWED 
 
 LOAN DEPT. 
 
 RENEWALS ONLY— TEL. NO. 642-3405 
 
 This book is due on the last date stamped below, or 
 
 on the date to which renewed. 
 
 Renewed books are subject to immediate recall. 
 
 
 -■»:■ V - 
 
 r:^ 
 
 I^TERLIBRAR 
 
 tNTER-L!BRA>?i r 
 LOAN 
 
 4JN 
 
 CA.^^LJ[kii^ 
 
 YjjQAf 
 
 MR. 
 
 r: 
 
 < r^ '"? 
 
 tx 
 
 W^ 
 
 V. O^ CALfF.. BERK, 
 
 WI\Rga^S^Q/' 
 
 -n 1 3 1976 1 
 
 IRVINE 
 
 APR 15 1988 
 
 INTERLIBRARY LOAN 
 
 
 Due end of WINTER Q-jar 
 
 subject to rocQll Q 
 
 ' -i4 
 
 IIBIKJUN 181907 
 
 ?f- 
 
 -J i 
 
 72 
 
 4ttC& 
 
 tn nTtn9.'^i'73-lPiSf 
 
 UCLA 
 
 IN T ERLIBRARY LOAN'^"'' "" 
 
 7 
 
 JUL 2 9 1975 
 
 REaciR. rra 6 "^8 
 
 LD21A-60to-6.'69 
 (J9096sl0)476-A-32 
 
 MAfV-i^SlSfc, 
 
 fornia 
 
 Berkeley 
 
Cjaulorci ; 
 
 PAMPHLET BINDER 
 
 Syracuse, N. Y. 
 Stockton, Calif. 
 
 u.c 
 
 BERKELEY 
 
 YC 0062L 
 
 UBR^R^ts 
 
 600301^1^