LIBRARY OF THE University of California. OIF^T OF\ Received ffya^Cs , i8g%. Accession No.7/j. § 39. Modern French verse 64 § 40. The origin of French versification 68 § 41. The earliest French octosyllabics 70 § 42. The later development of French octosyllabics 73 § 43. The difference between Latin and Old French verse . ( . . . 75 § 44. Explanation of the difference 77 § 45. Explanation of the change in French verse 83 § 46. French Decasyllabics 87 CHAPTER VI. Latin and French Influence in English verse. § 47. Old English verse .91 § 48. The decay of Old English verse 92 § 49. Development of English verse under foreign influence ... 94 § 50. Chaucer 97 § 51. The syllabic principle in Modern English verse 99 BIBLIOGRAPHY of the principal authorities referred to in the notes. Ambros, A. W. Geschichte der Musik, mit zahlreichen Noten- beispielen und Musikbeilagen. (Vol.11, 2nd ed.). Leipzig 1880. Bartsch, Karl. Chrestomathie de L'Ancien Francais (VHIe — XV e Siecles), accompagnee d'une Grammaire et d'un Glos- saire. Cinquieme ed., corrigee et augmentee, Leipzig, 1884. Bridges, Robert. Milton's Prosody, an Examination of the Rules of Blank Verse in Milton's later poems. * * * * Oxford, 1893. Chappel, William. The History of Music (Art and Science). Vol. I, from the earliest records to the fall of the Roman Empire. London, 1874. Du Meril, E. I. Poesies populaires latines anterieures au dou- zieme siecle. Paris, 1843. II. Poesies populaires latines du moyen age. Paris, 1847. Fetis, F. J. Histoire g6ne>ale de la musique depuis les temps les plus anciens jusqu'a nos jours. Paris, 1869 — 76. Biographie universelle des Musiciens et bibliographie g6nerale de la musique. 2 e ed. Paris, 1873. Gautier, Leon. Les Epopees Francaises. Etude sur les origines et l'histoire de la litterature nationale. 2 e ed. Paris, 1878 a 1882. Julian, John. A Dictionary of Hymnology, setting forth the origin and history of Christian Hymns of all ages and nations, * * *. London, 1892. Kawczynski, Maximilien. Essai comparatif sur L'Origine et L'Histoire des Rhythmes. Paris, 1889. Lubarsch, E. Franzosische Verslehre * * * Berlin, 1879. March, F. A. Latin Hymns, with English Notes. For use in schools and colleges. New York, 1874. Mayor, T. B. Chapters on English Metre. Cambridge, 1886. Meyer, Wilhelm. Anfang und Ursprung der lateinischen und griechischen rythmischen Dichtung. Munich, 1884. VII Monro, D. B. The Modes of Ancient Greek Music. Oxf., 1894. Paris, Gaston. Etude sur le role de l'accent latin dans la langue francaise. Paris, 1862. Lettre a Leon Gautier sur la versification latine rhythmique. Paris, 1866. Schipper, J. Englische Metrik, in historischer und systematischer Entwickelung dargestellt. Erster Theil Altenglische Metrik, Bonn, 1 88 1. Sievers, G. E. Altgermanische Metrik. Halle, 1893. Stengel, Edmund. Romanische Verslehre (in Grober's Grund- riss der Romanischen Philologie, II. Bd., i.Abt., i.Liefg.). Strassburg, 1893. Westphal, Rudolf (and Hugo Gleditsch). Allgemeine Theorie der Griechischen Metrik, (being the 3rd vol. of Die Theorie der musischen Kunste der Hellenen, by Rossbach and Westphal). Leipzig, 1887. CHAPTER I. Introduction. s § i. The principle of parallelism. The simplest known form of civilized verse is that of the Hebrew scriptures. In the poetical parts of the Bible there is present no regular principle of versification except that of the correspondence (or, more technically, parallelism) of one clause with another. This form is probably not only the simplest but the oldest known: — indeed ac- cording to Old Testament chronology it is not long antedated by the creation, for it is found in the song of Lamech to his wives. Hear my voice; Ye wives of Lamech hearken unto my speech : For I have slain a man to my wounding, And a young man to my hurt. If Cain shall be avenged sevenfold, Truly Lamech seventy and sevenfold.(i) This principle of parallelism is of course present in all the more modern forms of verse: it is indeed funda- mental in the theory of aesthetics. From it, probably, are developed all other forms. We have, however, no (i) Gen. IV, 23. The parallelism is sometimes more or less compli- cated ; and in some of the Psalms the use of a recurring refrain seems to indicate a strophic arrangement. Efforts have been made to work out a strict metrical scheme for some of the Hebrew poems, but they have not won much favor. modern poetry of any great importance in which this principle is the only determinant of form: our modern systems of verse differ among themselves in respect to the various refinements to which this principle has been subjected, but it appears in them all only as a theoretical base. § 2. Syllabic verse. Probably the first important ad- vance from this crude stage of art is to be seen in the syllabic verse of some of the oldest parts of the Avesta. Here the principle of parallelism is still seen in com- parative undress, but the parallel periods are all com- posed of equal numbers of syllables. The effect may be shown (for those who, like the present writer, have no knowledge of the original) in an English reproduction of the verse-form.(i) Who was the first of all mortals | to honor thee on earth, Homa ? What reward was bestowed on him, | what honor conferred upon him? Vivaswan was the first mortal | to do me honor upon Earth ; He therefore was so rewarded, | such honor was conferred on him, That he had a son born to him, | the all-powerful King Jima, The most worshipful of mortals, | the deliverer of mankind. Each verse contains 16 syllables, distinctly divided into hemistichs of 8 syllables each. So far as has yet been proved, there is no regular recurrence of long and short, or of stressed and unstressed syllables. In other words, the verse is what we will call purely syllabic. A curious specimen of a similar form of verse is found in a Latin MS. of the 9th century, in a poem on the removal of a Saint's relics.(2) (1) From a German translation from the first part of the Yasna, given by Westphal, p. 41. The German translation is in modern accentual rhythm: \ Wer hat als der Menschen erster dich verehrt auf Erden, Homa?' I have obliterated the rhythm, for the sake of conforming to Westphal's own description, — upon which that in the text is founded. (2) Given by Du Menl I, 162, as from Massmann, Die deutschen Abschworung-s-, Glaubens-, Beicht- und Bet-Formeln, p. 8 n. 1 7. Beatissimus namque Dionysius | Athenis quondam episcopus, Quem Sanctus Clemens direxit in Galliam | propter praedicandi [gratiam, Ibidemque martyrio coronatus | comperitur, et tumulatus, etc. Each verse contains 21 syllables, with a caesura after the 1 2th. No further regularity, either metrical or rhyth- mical, can be perceived. Such a verse could probably not have been written except for music, but the specimen is given here because it is one of the very few available in which the syllabic principle appears alone.(i) In com- bination with other principles, we shall find it of great importance, in English as well as in Latin and French. § 3. Quantitative verse. The second in order of the more modern kinds of verse is that which depends upon the quantity of syllables. Quantitative verse is thought, with much reason, to have developed out of the purely syllabic form.(2) Thus the octosyllabic hemistich of the form XXXXXXXX (where X represents a syllable of indeterminate quantity) was in time supplanted by one of the form XXXXX — ^ — . The verse-end has often proved itself peculiarly sensitive, and liable to change; and if such a change as this could once get a firm hold on the verse-end, it is easy to see how it might quickly extend itself to the whole verse. This is hardly more, however, than speculation, and we must for the present accept the appearance of quantitative verse, in (for example) the earliest known Greek poetry, as one of our ultimate facts. It may have developed from a prehistoric Aryan verse, in which only the syllabic principle was recognized, or it may possibly have been an original creation of the Hellenic race. Its cha- racteristic forms are too familiar to need description here. (1) I. e. without the aid even of rime, in the modern sense. (2) Westphal, p. 45. § 4- Accentual verse. The third and last of the distinct principles of modern verse-structure is that which regulates the verse according to word-accent. This is commonly recognized as the controlling principle of English verse, but it is now generally to be found only in combination with the syllabic principle, or at least marked and defined by end-rime. It is worth noting, however, that in Old English verse this principle relied for its external support upon quantity and alliteration, and that these two, as essential principles of English verse, have passed away as rime and syllabism came in. The accentual principle may therefore be regarded as the sole native base of our modern system of verse, even though it is so seldom to be observed in its native state. So far as the writer is aware, it has never held sole sway in any prevailing form of verse, though isolated examples may be given. Thus in Old and Middle English poetry there are occasional lines in which we find neither alliteration nor rime nor regard tor quantity; but they are exceptions, not types. Perhaps the best specimen of purely accentual verse that can be given is Charles Lamb's "The Old Familiar Faces": I have had playmates, I have had companions, In my days of childhood, in my joyful school-days; All, all are gone, the old familiar faces. § 5. The decay of Greek versification. However un- certain the origin of quantitative Greek verse, we can be sure that it had its justification in the nature of the spoken language. The Greeks wrote in quantity because they spoke in quantity, just as the modern English poet writes what we call accentual verse because his language is in the main an accentual language. On the other hand it is clear that the quantitative system of Greek poetry was in part artificial. It seems impossible that every long syllable should require in ordinary speech just twice the time of any short syllable; and some of — 5 — the phenomena of logaoedic measures are obviously due to arbitrary conventions. In post-classical times, the relations between quantity in speech and quantity in verse seem gradually to have become still more strained. For example the rhetorician Dionysius of Halicarnassus, writing in the first century B. C, says of the famous Homeric line on Sisyphus: "its most striking peculiarity is this: — neither of the long feet which are apt to be found in heroic verse, (i. e. spondees or bacchii), occurs here except in the last place: the first five feet are all dactyls, and that too of the sort that have their second syllables slurred over; so that some of them are not very different from trochees ".(i) In other words, some of the short syllables in avOiq sjceira jiedovds, &c, were markedly shorter than others, in the time of Dionysius if not before. Moreover the character of the Greek accent was also changing. From a mere inflection (as it seems to have been in the time of Aristoxenus,(2) ) it had become in the time of Babrius a marked stress.(3) Quantitative verse had been made easy by the quanti- tative character of the language and by an almost entire absence of accentual stress:— but as the former decayed and the latter came in, quantitative writing became more and more an act of somewhat pedantic affectation; and (i) "0 6h [tafaoxa T(ov aXXcov Oavfid^siv a£iov, gv&iibq ovdelq xuiv (jLaxQ&v, 01 nXrjv enl xfjq xsXevtrjq. oi 6h aXXoi navxeq eiol daxxvXoi xal ovxoi ys TCCcQadsSicoy/ievag e%ovxeq xag aXoyovq, cooxe firj itoXv dia&eoeiv evtovq xdiv xooxalcov. (De Comp. Verb. c. 20, cited by Westphal, p. 16). The meaning of aXoyovq, as Westphal points out, is not clear, as the word is ordinarily applied only to naturally long syllables that are irrationally slurred. Obviously, how- ever, it here designates the second syllable of the foot, and it presumably indicates that Homer's dactyls were understood by Dionysius as cyclic. (2) See, for example Marquardt's ed. of Arist. Fragm., p. 24, 1. 15. (3) As is shown by B.'s peculiar treatment of the Choliambic. the poets had to meet the new conditions of the language by writing a new kind of verse.(i) § 6. Origin and decay of Latin quantitative versi- fication. The later developments in Latin literature were similar, but the beginning seems to have been different. That the earliest known forms of Latin verse were of an accentual character seems now fairly established.^) The quantitative system was not a spontaneous creation of the Romans, nor apparently the natural outcome of any peculiar fashion in their mode of speech, but was an exotic, engrafted upon their literature in the 3rd century B. C. by students of Greek. It could not have thriven at Rome if the Latin language had not been more markedly quantitative (and perhaps less accentual) than, for example, modern English: but it would hardly have been necessary to go abroad for it if the language had not been naturally less prone to it than the Greek.(3) That the application of the Greek system to the Latin tongue must have involved something of a wrench is clear enough,(4) and there is reason for believing that accentual poetry, even through the classical age, kept a place in the ear of the common people. Apparent (1) In the following pages I shall not develop this branch of the subject, because, although the Greek and Latin literatures were so related that a mutual influence in the matter of versification seems very probable a priori, yet in fact the progress in Latin verse seems self-explanatory. It is only with the latter that we are directly concerned, and I have observed no decisive evidence that the Greek verse is even indirectly relevant, except as presenting an interesting parallel. (2) Lindsay, Am. Journal of Phil., Vol. 14 (1893), p. 139. (3) Kawczynski says, (p. 30), "les influences historiques sont plus fortes que les conditions naturelles ", and the phenomena of classical Latin metres seem to support his assertion. I think we shall find, however, that this is an isolated case. At all events, the generalization is un- warranted. (4) Cf. the artificial way in which the Roman poets treated the complexities of logaoedic verse. remnants of it are found, for example, in the song of Aurelian's soldiers, Tantum vini habet nemo quantum fudit sanguinis, and the song of the 6th legion, Mille Francos, mille semel Sarmatas occidimus.(i) It seems not impossible that a keen ear for prosody, — a nice perception of quantities, — may have been something of a rarity even among the upper classes in the Augustan age. But assuming that it was lacking then only among the uncultivated, it is certain that in the succeeding centuries the educated classes lost it too. As early as the beginning of the 5th century the difference between long and short syllables was no more practical to the average Roman than it is now to the average English- man. This is shown by a curious passage in St. Augustine's treatise on music. The treatise is in the form of an imaginary conversation between teacher and pupil. At one point the teacher purposely misquotes Virgil's line Arma virumque cano, Trojae qui primis ab oris. And the student is unable to see that the excellence of the verse is in any way impaired.(2) And the gram- marian Servius, writing probably at about the same time, says most explicitly: — "Quod pertinat ad naturam primae syllabae, longane sit aut brevis, solis confirmamus exem- plis ; medias vero in latino sermone accentu discernimus ; ultimas arte colligimus."(3) In other words even the most highly educated Romans learned the quantity of penults only from the accent, and that of other syllables only from the example of the poets or from established rules. Usage in pronunciation was no guide. Under these circumstances, the composition of quan- titative poetry began in Latin as in Greek to involve too (1) Cf. also Horace, Ep. II. I. 157: " Hodieque manent vestigia ruris". (2) De Musica, II. 2. (3) De Ratione Ultimarum Syllabarum, as quoted by G. Paris, Sur Vaccent lat. p. 30, n. 2. — 8 — much of pedantry : and as in Greece during the Byzantine period, so in Rome during the dark ages, the art of writing syllabic or accentual verse grew gradually in favour, until the old style had been effectually ousted from the field of lyric poetry. In the Romance languages, then in process of formation, one of the new styles was adopted for all kinds of poetry, to the entire exclusion of the old. § 7. The problems of late Latin verse. The object of the first part of this paper will be to trace and explain, so far as possible, the processes of decay just mentioned, The problem may be provisionally divided into two parts, In the first place, was the late Latin verse essentially accentual, or was it merely syllabic ? In the second place, how did the poets acquire the new art? The answer to the first question most favored by contemporary scholar- ship (notably by Wilhelm Meyer) is in substance that the so-called "rhythmical" form of the late Latin hymns was not based upon accent at all; that as quantitative verse passed into " rhythmical ", the element that survived was not the true rhythm of metrical stress, but merely the parallelism that was enforced by uniformity in the number of syllables per verse; and that when there seems to be a strictly accentual rhythm in the later verses, its appearance is in general the result of a happy chance, not at any rate an essential condition of the verse. To the second question the answers have been vari- ous, but we may group the most important of them under three heads. The later "rhythmical" system, according to modern opinion, was derived either (1) from the quantitative system by a natural transition, not the result of external influences, or (2) from foreign sources, either as an entire importation or (according to Meyer) as a sort of graft upon the decayed quantitative system, or (3) from the popular accentual verse of the earlier days of Rome. — 9 — § 8. The theory of a popular origin. The last theory- deserves some attention, although its strongest advocate has withdrawn his support.(i) It has been contended that there existed a continuous literature (if it deserves the name) of accentual poetry, from the earliest to the latest age of the Latin language, beginning with the Saturnian verse, manifesting itself in the classical epoch in the popular songs of which specimens have already- been given, and culminating in the triumph of the accen- tual system. It is unnecessary to repeat the arguments that have been advanced against this contention,^) but, if it should still be regarded as plausible, it is worth while to point out that it does not explain the phenomena that most need explanation in our present study. All the fragments of popular song that have been cited in support of this theory are in a trochaic rhythm. The two verses already quoted in these pages are fairly representative. Now (as will be shown hereafter) the trochaic rhythm was of no direct influence in the development of those Romance verse-forms to which the great body of English verse is indebted. We may fancy that we see in the song of Aurelian's soldiers the direct progenitor of such poems as the anonymous " De Gaudiis Paradisi": — Ad perennis vitae fontem mens sitivit arida, Claustra carnis praesto frangi clausa quaerit anima, Gliscit, ambit, eluctatur exul, frui patria. But we should search French literature in vain for any verse imitated from the latter. § 9. Popular verse of irregular rhythm. There are however certain other evidences which may tend to show (1) G. Paris. Compare his Lettre a M. Leon Gautier, with his note in Romania XV. 138. (2) They are well reviewed by Meyer, pp. 107 — 8 ; — although, as will appear in the next chapter, the argument basjdL^m_£pmmodian's ex- periments is easily refutable. — IO — the continuous existence of a popular Latin versification of an accentual character. The popular songs already mentioned were perhaps merely sporadic, — ignorant imi- tations of a form of quantitative verse heard at the Roman theatres. Those which we are now to consider, on the other hand, seem to have no connection with any quantitative verse. There is preserved from the 7 th century a collection of curious letters that passed between Bishop Frodebertus and a person styled Im- portunus. They are written in a sort of hap-hazard accentual rhythm, not much better than their latinity, as an example will show: Amas puella bella De qualibet terra, Pro nulla bonitate Nee sancta caritate. Bonus nunquam eris, Dum tale via tenes. Per tua cauta longa, Satis est, vel non est?(i) This is not much better than prose, and it really seems unnecessary to believe that the author had ever seen any rhythm of the kind before : — to have invented it out of whole cloth would have required no great effort of ingenuity. But there are those(2) who have little belief in any natural penchant of ignorant men for rhythmical expression, and seek more or less confidently for pre- cedents for all such phenomena. It is certainly not im- possible that Frodebertus and Importunus may belong to an undiscovered order of popular Latin versifiers, with an unbroken file of predecessors reaching back to the earliest times : their verse certainly resembles the Saturnian as closely as does that of the soldiers' songs. In that (1) Boucherie, Cinq Formules Rhythmees et Assonancees du VII. Siecle, (Montpellier 1867), p. 26. (2) Notably Kawczynski. See ante, § 6, note, and his " Essai" y passim. — u — case, the following rhythmical invitation to dinner, from a ioth century MS.(i) may belong to the same family of verse. Jam dulcis arnica venito Quam sicut cor meum diligo; Intra in cubiculum meum, Ornamentis cunctis onustum. Ibi sunt sedilia strata Et domus velis ornata, etc. It should be remarked however, that these verses, like the trochaic songs before mentioned, are not par- alleled by the ordinary forms of French verse. A striking peculiarity of the Latin lines just quoted is that in them the number of syllables per verse is altogether irregular.(2) It happens, to be sure, that the earliest extant specimen of French verse, the song in honor of St. Eulalia, is in- deed written in a rhythm which at first glance seems vaguely similar to that employed by Importunus: — Buona pulcella fut Eulalia Bel avret corps, bellezour anima. Voldrent la veintre li Deo inimi, Voldrent la faire diavle servir. Elle non eskoltet les mals conseillers, Qu'elle Deo raneiet chi maent sus en ciel, etc. But the Eulalia verse is known to be of very different origin 5(3) — and it is, moreover, unique in early French literature^) However numerous, therefore, compositions (1) Given by Du Meril, II. 196. (2) In the "Invitation" it varies from 8 to 10: in the letters it varies still more widely. (3) It is clearly modeled upon a Latin sequence. (4) Unless we possibly should class it with such defective verses as those in the Enseignements Salomon and certain Anglo-Norman poems, which require to be scanned by the number of accents rather than syllables. Of these, however, the former, if not corrupt, are doubtless simply bad verses ; and the latter are due to the retro-active influence of English verse. Even these Anglo-Norman verses are regarded by Stengel as merely unskilled work, for as he points out, (p. 11), Gower and Frere Anger wrote correctly. In any case, to connect them with the Eulalia would be quite wild. 12 of this character may have been in the dark ages, they left no permanent traces in later literatures. The only verses, whether Latin or French, which concern the present inquiry, are those in which the syllabic principle is maintained, — either with or without an accentual rhythm. § 10. Conclusion. It is clear therefore that the con- trolling influences for which we are searching are not to be found in Latin popular poetry.(i) The other branches of the problem will require more extended discussion. The chapters immediately succeeding will state certain reasons for believing that the later "rhythms" grew out of quantitative verse by a gradual and natural process ; and in the course of the investigation an effort will be made to show that the somewhat old-fashioned belief that these "rhythms" were really rhythmical (i. e. accentual), has been too hastily discarded. If Meyer's theory of the nature of the late Latin verse is the wrong one, then his theory as to its origin need not detain us long, for the two are inter-dependent. (i) Stengel's and Blanc's theory of a lost popular Latin rhythm, the parent of the French io-syllable verse, may be disregarded: — for as will appear hereafter, that verse is not in the direct line of descent from Latin to English. The specimens of irregular verse that have just been cited are doubtless mere slovenly imitations of the regular rhythms to be examined later. CHAPTER II. Commodian's Verse. § ii. The importance of the subject. One of the most important questions to be determined is whether, as the Roman poets lost their ear for quantity, the feeling which remained uppermost was the feeling for rhythmical run and stress, or only the feeling for parallelism and uniformity in the counting of syllables. This is indeed but another way of stating the main question at issue, for if the rhythmical ictus survived, then the later versi- fication would of course be accentual as well as syllabic : — otherwise it would be syllabic but not accentual. It has been ably urged that whatever there may have been in the nature of ictus, in the classical quantitative verse, it was not separable from the quantitative system, but perished with it: — and strong confirmation of this belief has been found in the apparently unrhythmical character of the verse of Commodian, a writer ot the transition periocT This poetT~aboutr the middle of the 3rd century of our era, introduced the practice of writing hexameters of a barbarous kind, which not only were not strictly quantitative, but also seemingly failed to attain any smoothness of accentual rhythm. His so-called "Carmen Apologeticum" , a poem of some 1060 lines, begins as follows : — Quis poterit unum proprie Deum nosse coelorum, Nisi quern is tulerit [longe] ab errore nefando? — i 4 — Errabam ignarus spatians, spe captus inani,(i) Dum furor aetatis primae me portabat in auras. Plus eram quam palea levior: quasi centum inessent In umeris capita, sic praeceps quocumque ferebar. The question for us to determine is whether these lines were meant to ±>e read and scanned like quantitative hexameters, thus: Quis pote | rit u | num l| propri | e Deum | nosse coe | lorum Nisi quem | is tule | rit )| Ion | ge ab er | rore ne | fando '.(2) or whether (as Meyer contends) they should be read with their natural prose ^accents. A definite solution of the problem will go far toward establishing the true theory of the later "rhythms". §12. Meyer's theory. The prose-accent theory seems at first glance particularly plausible. Commodian cannot have written in ignorance of the true nature of the hexa- .meter, or of the laws of quantity, for in certain respects (as will be pointed out hereafter) he observes those laws with care: he was, moreover, a man of wide reading, and it is not easy to see how such a man could scan his lines in the excruciating manner indicated above. The probability is(3) that Commodian deliberately rejected what had already become a highly artificial mode of ; composition, because he thought it unsuitable for the expression of Christian earnestness : — and having rejected it in part, one might expect him to reject the artificial elements of it altogether. Meyer's contention is ably reasoned, and the present writer can add nothing to this side of the argument except a suggestion as to the practical effect which (1) This is indeed a good hexameter, and so are several others of Comraodian's lines: — they are so however, only by accident. In this case the accident is obvious, for the line was certainly designed to be read without elision. (2) Cf. Tennyson's burlesque: "These lame hexameters the strong- winged music of Homer"; — and the Elizabethan hexameters generally. (3) This suggestion was first made, I think, by Du Menl. — 15 — Commodian may have designed. Meyer says (p. 302) " Demnach findet sich bei Commodian nur die eine Riick- sicht auf den Wortaccent, dass er, wie die quantitirenden Dichter, in die funfte Hebung stets eine Silbe riickte, welche den Wortaccent hatte, wahrend es ihm nahe lag, das nicht zu thun. Dies ist der einzige Fall, von dem man sagen darf, dass Commodian sich um den Wort- accent mehr gekummert habe als Virgil oder Ovid": Here Meyer seems to weaken his own case by representing the structure of these verses as comparatively purposeless. If Commodian designed his lines to be read according to their prose accents, there can be no doubt that he was consciously imitating the effect produced by such a reading of strictly quantitative hexameters, and that he wished his lines to sound to his unlearned contemporaries substantially as Virgil's or Ovid's must have sounded. Such a verse, for example, as Commodian's Sub jugo servili ut portent victalia collo (Inst. 39, 16). fairly reproduces the accentual rhythm of Virgil's Aut age diversos, et disice corpora ponto. (Aen. I, 70.) To match every verse of Commodian's with a verse from Virgil would be a forbidding task, owing to the great variety of possible combinations; but if the verses are split into hemistichs the latter can be matched without difficulty. Thus the following hemistichs from Virgil will be found to tally with the passage cited above from the Carmen Apologeticum. Hinc populum late (2i)(i) Latio genus unde Latinum (6) Ipsa Jo vis rapidum (42) Stridens Aquilone procella (102) Italiam fato (2) Meritis pro talibus annos (74) Et soror et conjunx (47) Memorem Junonis ob iram (4) Judicium Paridis (27) Animam hanc effundere dextra (98) Andierat Tyrias (20) O terque quaterque beati (94). (1) The numbers refer to lines in the 1 st Book of the Aeneid. I may add that from the same book I found no difficulty in matching each hemistich in the 39th of the Instructiones (containing 26 lines), having selected that passage quite at random. — 16 — § 13. Objections to Meyer's theory. But plausible as is this theory about Commodian's verse-structure, there is an insuperable objection to it; and the substance of that objection is, in part, so carefully stated by Meyer himself, that it is hard to see how he failed to appreciate its weight. The fact isjthat in certain respects Commodian is strictly attentive to quantity, and it seems almost perverse to accept any theory which reduces this la- borious strictness to "todter Zierrat". Meyer himself points out (p. 291) that of the 490 verses of the Carmen Apol. which end with dissyllables, there are only two in which the penultimate syllable is short; also (p. 296) that the penultimate syllable of the first hemistich is always strictly correct in quantity. The significance of the first of these facts is perhaps not obvious at first sight. It will be observed, however, that when the verse ended with a polysyllable, the penultimate syllable would necessarily be long, for otherwise the desired accentual effect would fail; but in the poet's use of dissyllables one would expect to find him utterly capricious as to quantity, since in such words the accent is necessarily on the penult. Here then Meyer finds indisputable proof not only that Commodian understood quantity but that at the end of each hemistich he was careful to observe it. On page 296 he says "Die Bildung des Schlusses war Commodian die Hauptsache "; in the remaining parts of each verse he finds (with a few exceptions which need not be noted here) nothing but entire indifference. § 14. Commodian's use of dissyllables. But as a matter of fact it can be shown that Commodian's regard for quantity extended much further than even Meyer has observed. The latter's remark upon dissyllables at the ends of verses naturally suggests an inquiry into the use of dissyllables elsewhere; and it will be found that the quantity of their penults is observed as care- fully in all parts of the verse as it is at the end. This — 17 — observance of quantity follows a very curious but simple law. If the verses are read rhythmically, (i. e. with the same rhythmical movement as that of quantitative hexa- meters), it will be found that the thesis generally falls on the first syllable of a dissyllabic word only if that syllable is really long-. — if, on the other hand, the first syllable of such a word stands in the arsis, then that syllable if long may form part of either a dactyl or a spondee, but if short can form part only of a dactyl. This law can easily be demonstrated by an analysis of a passage from one of Commodian's poems. For ex- ample, in the first hundred lines of the Carmen Apolo- geticum there are some 108(1) dissyllables of which the penults will receive the thesis if the verses are read as quantitative hexameters. Of these, 52 are at the ends of lines, constituting in each case, of course, the 6th foot of the line. Of the latter all but two (edunt, 1. 22> and quoque,(2) 1. 41) have the first syllable long. Of the other 56, which are found in the first, second, fourth and fifth feet of the lines,(3) all but three have the first syllable long, the exceptions being nisi (2), datas (27) and bonum (87).(4) This almost perfect regularity is not due to any overwhelming preponderance of trochaic or spondaic dissyllables in the Latin language, for there is no such preponderance: — and it is not due to any whimsical avoidance by Commodian of dissyllables with short penults ; for in this same passage such dissyllables occur to the number of 59, and all except the three just mentioned stand with their penults in the arsis. Furthermore, — and this is perhaps the most curious fact to be observed, — of all the 56 dissyllables with short penults standing in (1) This is without deduction for repetitions. (2) Which Meyer seems to have overlooked. (3) Not in the third, for Commodian puts his caesura there. (4) Perhaps also prius (83) should be added, making a total of 4 out of 57: — but I am doubtful as to the scansion of this line. — i8 — the arsis, only one is so placed that its penult forms part of a spondee. This single exception is statim (12). In all the other cases the penults stand properly for the short syllables of metrical dactyls. If, on the other hand, the penult of a dissyllable is long, then it may be in the thesis or the arsis indifferently, and in the latter case may be treated as either long or short. No system seems to prevail here, except at the close of the first hemistich (as will be pointed out hereafter). Compare for example the hemistich Et lumen offerimus (76) or Si pinguis est opibus (23) with Et nemo scibat (46). § 15. Commodian' s use of polysyllables. This curious system in the use of dissyllables seems quite inexplicable, unless the rhythmical reading of the verses (which we have assumed) was that really designed by Commodian. If that was really the case, it seems a fair working hypothesis (first) that Commodian was writing verse in quantity, so far as quantity was perceptible in his gene- ration, in ordinary speech and to unscholarly ears, and (second) that for some reason the quantity of accented syllables was more marked and determinate than that of syllables not accented. If this were the case, then the same system would probably prevail in the use of polysyllables; — and an examination will show that in fact it does. In the first hundred lines of the Carmen Apologeticum Commodian uses some 255 polysyllabic words. The ac- cented syllables of these words are long in 182 cases, short in 73. Of these 182 long syllables, 140 stand in the thesis, 42 in the arsis. Of these 42 again, 17 are found in spondees and 25 in dactyls. A majority, there- fore, of those that stand in the arsis, are improperly used. Of the 73 short syllables, on the other hand, 67 are used in the unstressed parts of dactyls, while none occur in the unstressed parts of spondees, and only 6 — ig — receive the metrical thesis. These 6 exceptions are found in divitiis (20) and divitias (27), humiles (29) and humilem (92), praeposuit (35), and arbitrio (85) : and they practically reduce to three, for two of them are repetitions, and in arbitrio the penult may well have been regarded as long by position. It is worth noting, also, that in each of these exceptional cases the irregular foot is a dactyl, not a spondee; — and that the same is true where a dis- syllable is irregularly used. Thus in the second line of the passage quoted above, the irregular foot is nisi quern, not nisi alone. Apparently the presence of two other syllables in the foot made it easier to tolerate the im- propriety: — or perhaps, indeed, we should say that Commodian did not regard these feet as dactyls at all, but that he sometimes allowed himself the license of substituting a tribrach. Such a substitution was of course not authorized by precedent: but it was logical and natural enough when all the dactyls were understood as cyclic, and of course Commodian cared little for precedent. There is yet a further peculiarity in Commodian's use of polysyllables which remains to be noted. Where the accent, in Latin, is proparoxytone, the penult is of course always short. In the hundred lines now under examination these penults are never misused. In other words, not one of them receives the thesis, or stands in the arsis of a spondee. This uniform recognition of the quantity of short penults is sufficient to account for one of the facts stated in the last paragraph, namely the absence of short accented antepenultimate syllables from spondaic feet : for if such a syllable could finish a spondee, then a short penult would have to begin the next foot : — but the statement was made as it stands for the sake of completeness, and because, moreover, the appearance of the same phenomenon in dissyllables shows that it would doubtless appear in polysyllables also as an independent fact, without this special necessity. 2* \J — 20 — § 16. The rationale of Commodian's verse. As to Commodian's treatment of the quantity of syllables in general it remains only to say that in monosyllables, and in all parts of other words except accented syllables and short penults, he shows entire indin°erence.(i) But that his practice in the matter of accented antepenults and short penults should be merely fortuitous, or any- thing but deliberate and systematic, is impossible. More- over it is obvious that this practice must be capable of explanation and justification by the phenomena of ordinary speech in Commodian's time. To give such an explanation with thoroughness and certainty, is beyond the scope of this paper and beyond the writer's present ability; but the evidence seems irresistible that Commodian was writing in prosody as it existed in his own^Hay; The common speech of his contemporaries seems to have exhibited a phase of transition between that of Virgil and that of St. Augustine's pupil. The quantity of accented syllables seems to have been appreciated by the unaided ear.(2) If they were long they could be slurred, but if they were short the poet was careful to avoid drawling them. Of syllables which did not bear the accent no particular account was taken, for their quantity, if perceptible, was at any rate not boldly marked. Perhaps only scholars knew whether they were long or short. If so, the use of these syllables by the classical poets must have seemed purely capricious to the unscholarly, and therefore Com- modian might properly use them as convenience dictated.^) (1) Except, of course, in the special cases mentioned in the next section. (2) Of this exceptional definiteness in the pronunciation of accented syllables there is perhaps a remnant in the law of the preservation of the accented syllables in the passage from Low Latin into French. (3) Of course his treatment of unaccented penults needs no explanation. Such syllables were known to be short by the rules of accent, whether they were so pronounced or not. 21 § 17- Commodian's treatment of the Ccesura. There is yet one other peculiarity in these verses of Commodian's. He observes the strict rules of quantity in one set of cases, (namely in the penultimate syllable of each first hemistich), which cannot be explained as dependent upon the accents of the words at all. Meyer says (p. 296) : "Die Bildung des Schlusses war Commodian die Haupt- sache. Die letzte Silbe der beiden Halbzeilen, in welche er die Langzeile des Hexameters zerlegt,(i) ist von ihm als Zeilenschluss behandelt und frei gegeben". [Perhaps we should say, its quantity is indifferent because it is the final syllable of a word'. — thus in 1. 6 of the specimen given above, capita receives the same treatment as potent in Li]. "Dagegen ist ihm die Bildung der vorletzten Silbe die Hauptsache. Diese ist so gut wie immer quan- titirend recht gebildet". Here Meyer is undoubtedly right. In many cases, to be sure, the penultimate syllable of the first hemistich must be correct in quantity, by the rules already discovered: thus if the caesura is preceded by a polysyllable or dissyllable whose penult is short, then the second foot of the line must be a dactyl, and the quantity of the syllable in question is exactly what it should be. But there are other cases which can be explained only upon Meyer's principle. Thus, for example, in the first hundred lines of the Carmen Apologeticum, the first hemistich is six times ended with a mono- syllable, e. g. Spero reus non est (81): Interdum quod meum est (83): — and in every case the syllable preceding, though final, is correct in quantity. Two explanations of this are possible, first that this attentiveness to quantity at the caesura was merely a bit of pedantic affectation, or second that the quantity (1) Note again that the caesura is always penthemimera}, — 22 even of final syllables was still barely distinguishable in Commodian's time, and that therefore while it could be disregarded elsewhere, it must be heeded in those parts of the line in which grace was most needed. The latter supposition seems more plausible. § 1 8. Concluding remarks. Further support for the opinion that Commodian's system was closely related to his mode of speech, may be found in the pseudo-hexa- metrical poems of the 8th century ;(i) for although they were seemingly made in imitation of Commodian's verse, they do not exhibit the chief peculiarities of his system at all. Thus in the Exhortatto Poenitendi we find such lines as Mens confusa taediis itinera devia carpens (3) Nee casus honoris sed ruinas animae plora (6) The authors of these verses could have thought of Com- modian's poetry only as doggerel. Their scholarship would tell them that he used false quantities ; their instinct could not tell them that he used correctly the prosody of his own century: — so in imitating him they would naturally overlook the very essence of his art. But the reason for Commodian's practice is, so far as our present inquiry is concerned, merely a question for the curious. (JThe one fact of great present importance is that to Commodian the central point of interest and attention in his rhythm was the thesis of the metrical foot. He lived at a time when the native feeling for quantitative verse was all but gone; — but the verse still lived in its rhythm, by the force of its metrical ictus. Commodian has heretofore been regarded as the writer .of an isolated type of verse, but it is now plain that he is directly in the line of our research. The particular path which he opened led nowhere, it is true, because it was based on an ephemeral condition of the language: — (1) See Meyer pp. 276 — 284, and appendices. — 23 — but it gives as perfect a specimen as could be desired of the transition state between metric and rhythmic. The other 7 theory, — that these verses were of the syllabic order, — seemed indeed almost grotesque, for the essence of syllabism is equality, and here there is no equality: — and the writer has never been able to under- stand why the elements that were obvious should be lost and those that were not should be preserved. An interesting corroboration of the foregoing argu- ment is afforded by the passage in St. Augustine already referred to.(i) After the pupil has confessed himself baffled, the teacher says "At hoc mea pronuntiatione factum est, cum eo scilicet vitio quod barbarismum gram- matici vocant: nam primus longa est et brevis syllaba; primis autem, ambae producendae sunt, sed ego ultimam earum corripui ; ita nihil fraudis passae sunt aures tuae ". Then the teacher repeats the verse in both forms, dwelling this time on the long -is in primis, and the pupil cries without hesitation, "Nunc vero negare non possum, nescio qua deformitate me offensum ! " — He did not object to false quantities in the least, but his ear was offended by any hitch in the run of the line.(2) (1) § 6, ante. (2) It will be observed that the exact scansion of some lines, if con- sidered by themselves, is doubtful : e. g. does the I st line of the Carm. Apol. begin with a dactyl or a spondee? But the number of doubtful lines is minimized by Commodian's practice of putting his caesura after the thesis of the 3rd foot: thus if the 1st hemistich contains 7 syllables, it must contain 2 dactyls; if only 5 syllables, 2 spondees: its scansion is doubtful only if it contains just 6 syllables. So in the 2d hemistich there can be no doubt except where there are exactly 9 syllables. I have no doubt that this in the reason why Commodian divided his verses so uni- formly. In doubtful cases we must choose that scansion which agrees with the system : — the system itself could be adequately established by examination of only those lines which are free from doubt. CHAPTER III. The Latin Hymns of Ambrose and his Followers. § 19. Introductory remarks. Commodian's verse was composed at a time when natural prosody still survived, through moribund. By the time of St. Augustine, how- ever, as we have already seen, it was dead. While therefore the experiments of Commodian aimed to pre- serve the old rhythm in the last remnants of the old metrical garb, we should naturally expect later experi- menters to discard the latter altogether, if they wished their verse to conform to the laws of ordinary pro- nunciation, and to find for the rhythm some other support than that of a forgotten prosody. This is exactly what was done, in the fifth and sixth centuries and after, by the writers of the Latin hymns. In the fourth century, Ambrose wrote the hymn beginning Jam surgit hora tertia Qua Christus adscendit crucem; Nil insolens mens cogitet, Intendat affectum precis. Here we clearly have a quantitative metre, a scholarly reproduction of the classical iambic dimeter. Several centuries later, on the other hand, Adam of St. Victor was writing such verses as the following: Jesu, tuorum militum Transcendens omne meritum, Ad laudem tui militis Nos ejus juva mentis. 2 5 ^ Here we at once see that there is no dependence upon quantity. The analogy of the 3rd and 8 th century hexa- meters raises a strong presumption, however, that the apparent rhythm of the lines was designed; and if so, the verse is manifestly of the accentual order. Our task will be to ascertain how far this presumption is supported by the facts, and (if it proves justifiable) by what steps the_new accentual system came jn- The hymns~t1iat7 have come down to us from the dark ages, exhibiting the change from quantitative to unquantitative structure in all its phases, afford more • than enough material for solving the problem; but un- fortunately something more is needed than abundance of material. Most of the specimens of verse in the interesting stages of transition are of very uncertain date. Thus the poem De Gaudiis Paradisi, of which the opening lines have already been quoted, has been variously ascribed to St. Augustine (354-430) and to Damiani (1002- 1 07 2), not to mention divers intermediate conjectures; and consequently, in an investigation to which accurate chronology is all-important, this hymn is practically use- less. As might be expected, the authorship of the earliest hymns (i. e. those of the fourth, fifth and sixth centuries) is the most thickly befogged; and these, as it happens, are just the ones that exhibit the most interesting forms. There are a few, however, of the early hymns, which we can arrange in chronological order with entire con- fidence, and from these, and a few others, we shall be able to get some light upon our inquiry. § 20. ^dmixns.e. About 385 A.D. Bishop Ambrose introduced in the church of Milan the singing of psalms and hymns. Many of these hymns were written by Ambrose himself, and many more, of similar style and metre, have been erroneously ascribed to him. Four are incontestably genuine,(i) namely "Deus Creator omnium", (1) Being mentioned as his by St. Augustine, — 26 — "Aeterne rerum conditor", "Jam surgit hora tertia" and "Veni redemptor gentium". These are all composed in stanzas of four iambic dimeter s, each constructed accord- ing to the common classical scheme, namely - _ u _ i - _T^ ... and all exhibit a fairly strict adherence to the traditions of quantity. In the first-named of the four there are two imperfect lines, namely Te diligat castus amor (15) and Ne hostis invidi dolo; (27) in the second, there is one slight imperfection, in Jesu labentes respice; (25) and similar irregularities may be found in the others; but in g£neraJLc| uantity is res pected, and of course (as the example given in the preceding section shows) there is no appearance of any regard for prose accent. Prom the teaching of Augustine we know that this form of composition involved a considerable degree of pedantry, — that it was governed by tradition rather than instinct ; — and from the example of Commodian we can very well understand that what actually appealed to the ear in the reading of these hymns was not the quantity but the metrical ictus. Commodian avoided putting the ictus on any syllable which his senses felt to be short; Ambrose put it only on syllables which his erudition knew to be long. §21. Sedulius. This poet (b. second half of 4th cent d. first half of 5th) wrote a hymn beginning "A solis ortus cardine ". The hymn now extant with this beginning is thought to be only in part the work of Sedulius, k but the whole hymn may properly be considered here as an early imitation of Ambrose.(i) The hymn is as strictly (1) Although lines 13 — 24 are sometimes attributed to Ambrose him- self: — see Julian's Dictionary. — 27 — quantitative as those of Ambrose himself, the irregularities bemgli>infre~~5ii^^ g. Verbo concepit filium. (16) The peculiarity of the hymn is this, that the authors have apparently sought, while observing quantity with scrupulous care, to attend to the prose accents also. The first three stanzas, for example, can be read accentually with perfect smoothness; — they will be found to contain no inversions except such as are common in modern English poetry. A solis ortus cardine Ad usque terrae limitem Christum canamus principem, Natum Maria virgine. Beatus auctor saeculi Servile corpus induit, Ut carne carnem liberans Ne perderet quos condidit. Castae parentis viscera Coelestis intrat gratia; Venter puellae bajulat Secreta quae non noverat. An accentual reading of the whole hymn will, to be sure, exhibit more violent inversions than these, as for example in Templum repente fit Dei (14) or even Et angeli canunt Deo. (26) But these inversions are decidedly less noticeable than they would be in the hymns of Ambrose, similarly pro- nounced; and it is especially worthy of remark that in this Sedulian hymn the inversions always fall upon dis- syllables. In all the 96 lines of the poem there is no polysyllable in which the metrical stress does not fall upon the accented syllable. That this would in all like- lihood not be the case if it were not designed, is made sufficiently clear (without consideration of antecedent probability) by observation of the shorter hymns of Ambrose. In the first-named of the four already de- scribed we find — 28 — Et noctes exortu preces (10) in the second, Hoc omnis errorum chorus (n) Politique mitescunt freta (14) Et ore psallamus tibi (32) in the third, Qua Christus adscendit crucem (2) Intendat affectum precis (4) Matri loquebatur suae (18) and in the fourth Vexilla virtutum micant (11) Aequalis aeterno Patri. (21) Although the authorship and date of "A solis ortus cardine " cannot be assigned with certainty from external evidence, the curious fact just noted seems to afford ground for a safe conjecture. Not long after the time of Sedulius a complete divorce had been consummated between rhythmic and metric. Thereafter those who wrote " rhythms " felt no scruple in disregarding quantity, and those who wrote in quantity recognized fully that their art was of the past not the present, and so could hardly care to struggle for an ineffectual compromise. This hymn, therefore, seems to have been composed, if not by Sedulius, at least by some one or more of his early followers, when the old system was still in the top of the fashion. They sought to keep it fresh by skilful doctoring. % 22. Fortunatus. There is at least one hymn, how- ever, of unquestioned authorship, which exhibits the same peculiarity as that just discussed. This is the "Vexilla regis prodeunt" of Fortunatus (530-609?). That this hymn was composed with careful regard for quantity is evident from the consistent care with which the 3rd and 7th places in the verse are reserved for short syllables, the only slip being in Dicens in nationibus. (11) This interesting hymn contains several verses which, if — 2 9 — read accentually, show the ordinary inversion of the first foot, — namely lines 10, n, 13, 18, 22, 24 and 31. Three verses show inversions of a more serious nature, namely Fulget crucis mysterium (2) Regnabit a ligno Deus (12) O crux, ave, spes unica. (29) But here, as in "A solis ortus cardine" there is no case of a polysyllable in which the normal ictus fails to coincide with the prose accent. For convenience and brevity, dissyllabic words used with iambic stress have been spoken of as exhibiting "inversions". This is true only if the verses are read accentually, and it is of course not to be presumed that such a reading was ever intended by these early poets. We must, for the present, assume that such lines as Dicens in nationibus were meant to be read with a regular alternation of arsis and thesis, and in them ictus and prose accent did not coincide. They exhibit the phenomenon known as "wrenched accent", rather than inversions of foot. But in the care with which the accent of polysyllables is preserved, there is a curious significance. Even to our modern ears, trained as they are to accentual rhythm, there is much less of the obviously conventional in such lines as those above quoted from Fortunatus, than in the Hoc omnis errorum chorus or the Matri loquebatur suae of Ambrose: and it is evident that while the authors of the two hymns last described were anxious to respect the conventions of prosody they also were especially anxious not to thrust then into the foreground. Further but perhaps more dubious evidence of the intention of these poets is found in the endings of their lines. In the hymn of Fortunatus only one line ends with a dissyllable: — that is to say, only one line fails to show concidence of ictus and accent at its end. In "A — 3 o — solis ortus cardine", with a total of 96 lines, there are 9 dissyllabic endings. In the four hymns of Ambrose, with a total of 124 lines, there are 37.(1). §23. Substitution of accent for prosodical length. But the most remarkable fact about this hymn of Fortunatus has yet to be mentioned. Although the verse is for the most part quantitative, as has been shown, there are two lines in which, in defiance of the rules of quantity, short syllables receive the metrical stress. These lines are Suspensus est patibulo (4) and Praedamque tulit tartari. (20) Here the stressed short syllable is the one that bears the prose accent. Patibulum is of course an impossible word in pure iambics, but the 20th line could have been saved by an easy transposition. The poet chose rather to let a merely accented syllable stand for a long syllable, than in either case to make any sacrifice of rhetorical effectiveness. Fortunatus was of course consciously follow- ing what he believed to be the best fashion in sacred composition. He imposed upon himself, for the nonce, unusual restrictions, in the effort to secure coincidence of accent with verse-ictus. He allowed himself frequently to sacrifice the normal accent in the first foot of the line; that was necessary, unless he would avoid dis- syllabic beginnings, and it was comparatively unobjection- able: but he did not in general allow himself the same liberty at the verse-end, although he had to forego dis- syllabic endings to avoid it, for there it was not un- objectionable. In polysyllables, which could not be treated lightly without a serious wrench, he adhered strictly to (1) I state these facts because they exhibit the most vulnerable point in this argument. There is no reason why wrenched accents should be especially objectionable at the verse-end: cf. our English ballads. I can explain the scruples of Fortunatus only by the consideration that rime and "rhythm" were both new, and he treated them with the tenderness of unfamiliarity. — 3i — the normal prose accent, and in two extreme cases he allowed this accent to usurp altogether the function of prosodical length. § 24. Early hymns of uncertain date. Here we seem to witness the last stage of transition, before the complete transformation of metre into rhythm. The hymns just considered afford at least prima facie reason for believing that the change was a natural and gradual one, that the rhythmical effect upon the unsophisticated ear was the same in each kind of verse, that the later hymns were not merely syllabic, but accentual, and that the change was begun by the device of making the metrical ictus of quantitative verse coincide with the prose accent of the words, and carried a step further by the substitution, at first only tentative, of accented for long syllables. The hymns of Ambrose and Fortunatus are the only ones in iambic metres which can with certainty be as- signed to known authors of this transition period: and for reasons which will appear in the next chapter, the early hymns in quasi-trochaics cannot properly figure in the comparison. The materials are so scanty that our deductions cannot, of course, be regarded as independently conclusive; but it will be seen in the next chapters that their value can be greatly increased by corroborative evidence from other sources. Many of the so-called Ambrosian hymns, (i. e. hymns in stanzas of four iambic dipodies, either metrical or "rhythmical"), and those attributed to Damasus and Hilary, would be of great value to the student of versi- fication, if only he were in a position to speak con- fidently of their authorship and date. Those credited to the two authors named, if known to be authentic, would be of especial interest as being the oldest extant specimens of the Latin hymnology ; but they are pretty certainly not authentic. To Damasus tradition ascribes the celebrated hymn on the martyrdom of Agatha, in dactylic verse: — 32 — Martyris ecce dies Agathae Virginis emicat eximiae, Christus earn sibi qua sociat Et diadema duplex decorat. 5 Stirpe decens, elegans specie, * * 21 Jam renitens quasi sponsa polo, etc. And to Hilary the hymn beginning "Lucis largitor splendide" is ascribed by his biographer Fortunatus, as well as by a fairly consistent tradition. This hymn (written for the most part strictly in quantity) contains two notable lines, namely Ne rapientis perfidi (19) and Haec spes precantis animae (29). The hymn of Damasus seems to reveal a comparative disregard of quantity in the arsis : for though the blunder in the 5 th line might be purely accidental, the false quantity in renitens (or renidens according to other MSS.) cannot have been unnoticed, as the long penult controls the accent of the word. But though both these hymns (like divers others of similar character), could obviously be fitted into our theory very conveniently, the un- certainty of their true dates makes it unsafe to rely upon them. Both are later than the time of their supposed authors, but we cannot tell how much later except by reasoning in a circle. It is to be hoped that theological, literary and historical scholarship may before long clear up some of these questions, and so make considerable additions to the material now available for our present inquiry. § 25. Ambrosian hymns of Adam of St. Victor. Before considering the more corflplex features of the problem, it will be useful to examine the rhythm of the Ambrosian hymn in its last stage of development. Adam of St. Victor, the skilful (if hardly great) poet of the 12th century, was the most perfect master of Latin "rhythmical" com- position, and has left many specimens of his work. — 33 — Among these are seven Ambrosian hymns. If these hymns are read accentually^ they exhibit a^stnKngly large number of initia l inversions , eittier^smgle (i. e. of the^first footJ~as m "Jesu, tuorum militum", or double (i. e. of the first two feet together), as in "Felix ortus infantulae". The frequency of these inversions can be conveniently shown by the following table. First line of hymn Single inversions Double Total no. of lines Lux illuxit Dominica IO 7 40 Genofevae solemnitas 12 16 52 Laudemus omnes inclyta 17 20 60 Aeterni festi gaudia 12 22 60 Supernae matris gaudia 7 15 52 Aurora diem nuntiat 6 5 23(1) Jesu tuorum militum 7 4 23 Total 71 89 310 In somewhat more than half of the verses we find initial inversions, either single or double. But in all the 310 lines there are only two possible instances of inversion in the 3rd or 4th foot. These are both in the hymn "Aeterni festi gaudia", viz.: — Quae vellet potest mens pia (32) and Qua praefulget Augustinus (51) In the former of these cases the inversion is obviously no inversion at all ; the line is to be read with a wrenched accent. This is shown clearly enough by the fact that pia rimes with omnia, scientia and caetera. The latter case is doubtful. The whole stanza is as follows:— Datur et torques aurea Pro doctrina catholica: Qua praefulget Augustinus In summi regis curia. In as much as the line in question does not rime, (contrary to the uniform rule of this hymn), it seems likely that the inversion is real, and that the poet de- (1) The last line of this hymn, "Sancto sit spiritui", is defective, and I have not included it in the table. The same in true of the next hymn. 3 — 34 — signedly introduced an entire trochaic line into this stanza : but of course the passage may be corrupt. The frequency of initial inversions in such hymns as these seems to give color to the theory that the verse does not depend upon accent at all, — that it is in theory only syllabic. But the fact that the inversions are practic- ally always initial— (there is no instance of inversion in the 2nd foot alone, unless possibly in 1. 37 of the first hymn in the table, "O mors Christi mirifica) — seems strongly to negative this theory. It is perfectly true, of course, that final inversions would have prevented mas- culine rime; and the principle of systematic rime is the one principle that is admitted, in addition to that of syllabism, by those who uphold the theory that these verses are not accentual.(i) But this is no sufficient answer, for it is not clear how the principle of rime can be scientifically separated from that of accentual rhythm. If words like Maria and maria had been systematically treated as of equal fitness for the verse-end, we might regard the two principles as quite separate; but they in fact were not. Further confirmation of our view will be found in the next section: but for the present it may merely be remarked that the syllabic theory is that (for the most part) of French and German scholarship, and that our ears are perhaps readier to perceive the true cha- racter of these rhythms than those of the French and the Germans, because the latter are accustomed in their own poetry to a much more regular alternation of thesis and arsis than is found in ours, while the former, from the nature of their language, are hardly trained to ap- preciate an accentual rhythm at all.(2) (1) See for example Kawczynski p. 117 et seq. It should perhaps be noted that the use of final monosyllables would have enabled the poet to invert the third foot alone without marring the rime, had he cared to do so. (2) See for example a remarkable comment of Kawczynski's (p. 44). "Les poetes allemands qui imiterent le vers octosyllabique des romans — 35 — It has been assumed in the foregoing paragraphs that these inversions in the poems of Adam are real in- versions, not wrenched accents. That this is the case is clear enough from the fact that they occur only at the beginnings of lines. For that in accentual poetry an inverted foot is positively agreeable in the beginning, and in general positively disagreeable at the end of a verse, is clear enough; but no such assertion can be made as to wrenched accents. If a line like Jesu dulcis memoria, read with regular stress on the even syllables, was satis- factory to the ear, so would be such lines as Te nostra vox primum sonet,(i) Fortunatus to the contrary notwithstanding. And the comments of the grammarians, (see § 36, post), leave no doubt that these verses were read accentually. § 26. Trochaics of Adam of St. Victor. The opinion uged in the last section, as to the theoretical presence of the accentual principle in Adam's Ambrosian hymns, receives positive confirmation from his treatment of trochaic rhythms. Kawczynski says(2) (after arguing for the purely syllabic character of Latin "rhythms" in general) : — "II faut toutefois accorder une large exception au rhythme trochai'que. La langue latine ne possedant pas d'dxytons polysyllabiques, et les proparoxytons n'y 1 francais ne parvinrent pas a donner a leurs vers un nombre de syllabes determine. Cela les g£nait trop probablement. En voici la regie exprimee par un rimeur du quatorzieme siecle: Ouch ich diss getichniss rim Uef die zal der silben zune, Sechse, sibene, achte, nune. II rime son Scrit, dit-il, sur le nombre des syllabes, dix, sept, huit ou neuf. Le seul rhythme qu'on donnait d ces vers consistait done dans la rime. (!) (1) From the "Aeterne rerum conditor" of Ambrose. (2) p. 119. 3* - 36 - etant pas tres nombreux,(i) la cadence trochai'que est tres frequente et toute naturelle aux mots latins. II y a done beaucoup de vers latins rhythmiques dans lesquels elle est presque regulierement maintenue par les accents, mais dans les meilleures pieces, meme de ce genre, on trouvera toujours quelques vers ou elle est brisee, bien qu'une legere transposition des mots eut suffi pour la retablir. Elle n'etait done pas obligatoire, elle n'entrait pas dans la notion et dans la definition des vers ryth- miques". But an examination of the hymns of Adam, which are certainly among "les meilleures pieces", will show that the regular trochaic cadence was obligatory and did enter most essentially into the poet's conception of the rhythm. The exceptions referred to by Kawczynski are of that rare sort which really do prove the rule. To make the character of this verse clear a careful examination has been made of some of Adam's hymns. A number were selected at random: — namely the seven hymns on the Nativity, which are given first in the recent editions. The results of this examination may be conveniently tabulated as follows: — First line of hymn No. of inversions Total no. of lines Potestate, non natura 3 56 In excelsis canitur 8 58 In natale Salvatoris 4 70(2) Lux est orta gentibus 2 12 Jubilemus Salvatori 2 24 Nato nobis Salvatore 2 30 Splendor Patris et figura 7 68 Total 28 318 Here we see that while inversions occur in more than (1) M. Kawczynski has evidently not resorted to the mechanical device of counting them! (2) There are 76 lines in the hymn, but 11. 49, 50 and 54 — 57 form a sort of interlude in a different kind of rhythm. Parts of this hymn are repeated in the next two, and I have therefore only reckoned those parts of the latter which are new. — 37 — one half of Adam's iambic lines, they are found in less than one eleventh of his trochaics. A typical specimen will illustrate the character of the inverted trochee. The second hymn begins as follows: — In excelsis canitur Nato regi gloria, Per quem terrae redditur Et coelo concordia. Now the comparative infrequency of these inversions might be partly accounted for by Kawczynski's explanation of the trochaic character of the Latin language, though it is hard to see how that could effect a discrepancy, such as we find here, of 83 per cent. But there is a curious uniformity in the character of these inversions which his theory does not account for at all. The seven hymns in question consist in the main of lines of 8 and 7 syllables. The octosyllabics are decidedly the more common, Adam's favorite stanza-form being a combination of four or six of the longer lines with two of the shorter. The former have a feminine ending, the latter a masculine, as the first lines of the hymns in the table show. Now there is obviously no reason in the supposed trochaic character of the Latin language why such inversions as occur in the first two feet of these lines should happen to be more common in one form than in the other; but as a matter of fact all the 28 inversions shown in the table do occur in the first two feet, and of these all but two are in the lines of 7 syllables.(i) (1) I doubt if these two are real inversions after all. They occur in the hymn "In natale Salvatoris", and are as follows: — Harmonia diver- sorum (4), and Post Deum spes singularis (60). The latter is easily ac- counted for as exhibiting a wrenched accent, — which as all students of the subject know, was particularly common with the word Deus t as with most proper names. The former line may indicate that harmonia was accented on the penult after the analogy of the Greek ctQixovLa. This sort of accentuation was very common, and the fact that Adam uses harmonia as a proparoxytone in 1. 2 of his familiar hymn "Aeterni festi gaudia" determines nothing; for the Greek word ei6o)Xov for example was repre- sented in Late Latin by either idolum or idolum indifferently. - 38 - The true reason for the comparative infrequency of inversions in these trochaic lines is not hard to find. It lies not in the character of the language, but in the different characters of the two rhythms. The iambic admits inversions easily, the trochaic does not. It is well understood that in English verse, for example, there is a natural tendency toward the iambic run, and if it is desired to keep a measure trochaic in the main, it must be kept so throughout : while an iambic measure can be frequently interrupted by trochees without serious disturb- ance. Any one who is not familiar with this fact can satisfy himself of it by counting the number of inversions in any trochaic English poem (as for example Locksley Hall), and comparing his results with the number found in any passage of good blank verse. The particular type to which all these inversions in Adam of St. Victor's trochaic hymns belong, is, as it happens, almost unknown in English verse, and at first glance it seems more divergent than it really is, from the normal. If a number of them are examined without their context, as Et coelo concordia, De carne puerperae, Coelesti praeconio, Sub noctis silentio, they look as if they ran with a sort of anapaestic move- ment, (or perhaps like dactylic dimeters with anacrusis), and so understood they threaten seriously to mar the general harmony of the verse : but if they are read care- fully as proper trochaics with only the first foot inverted, the effect will generally be found far from displeasing, even to our unaccustomed ears. Thus instead of Codes | ti praeco | nio, or Coe | lesti prae | conio, read Coeles | ti prae | coni | o, — 39 — and the line will fit its context. The effect is substantially reproduced in English in a stanza from The Vision of Sin : Trooping from their mouldy dens The chap-fallen circle spreads: Welcome, fellow citizens, Hollow hearts and empty heads! § 27. The invention of the initial inversion. The analogy, then, of Adam's trochaics removes all doubt as to the presence of the accentual principle in his iambics. Such of his lines as are strictly regular must be viewed not as happy accidents, but as exhibiting the normal verse-form: those which show initial inversions, on the other hand, are not to be regarded as equally represen- tative of a supposed merely syllabic type, but are inter- ruptions to the smooth flow of the ideal rhythm. They occur often enough to prevent monotony; but constant recurrence to the type keeps the latter always in the reader's mind.(i) But while it cannot be doubted that these late Latin rhythms were read generally according to their accents, with frequent initial inversions, it is almost equally clear, as we have seen, that the irregularities in the first ten- tative efforts at accentual verse must have been regarded not as inversions but as wrenched accents ; - for the ideal rhythm of the quantitative iambic verse was perfectly regular in the time of Ambrose, and that ideal rhythm was the element of the old verse that suggested the invention of the new. Inversions of foot, in the sense (1) The same inversions occur in English octosyllabics, but they are in comparison with the Latin very rare. In our heroic verse they are of course common. I have not yet been able to satisfy myself positively as to the reason for the difference between our octosyllabics and the Latin, but it must probably be found in some practical difference in the two modes of speech. A suggestive clue may perhaps be found in the general regularity of the German decasyllabic verse, as compared with ours. Com- pare, however, the note on the pronunciation of Latin, § 44, post. — 40 — in which we have been using the term, were unknown to the writers of quantitative iambic dipodies, and when they were first introduced they must have seemed a very radical innovation. The history of the transformation of metric into rhythmic cannot be regarded as fully as- certained until the time and manner of this innovation have been determined: but it must be admitted, though with reluctance, that with the present dearth of material no conclusive evidence on these points can be obtained. Clues are to be found, to be sure. For example, in the quantitative pseudo-Hilarian hymn "Lucis largitor splendide", the lines Cujus admota gratia, (15) Nostra patescunt corpora (16) and Ne rapientis perfidi (19) exhibit interesting peculiarities in quantity: — for not only is the second syllable short in each case; — but the first syllable is long. And moreover in the first two cases, [and perhaps in all three(i)], the initial feet are not only quantitative trochees, but accentual trochees also. If phenomena of this sort could be assigned to a definite period, (say between the dates of Sedulius and Fortunatus) a solution of the problem might be obtained from them : but unfortunately they cannot be assigned to any date. Another clue, perhaps not quite so faint, is found in the treatment of polysyllables in the anonymous early Ambrosian hymns. The writer has made, for this purpose, an examination of six(2) of these hymns: namely "Jesu nostra redemptio", "Hymnum dicamus Domino", "Christe (1) A slight rhetorical emphasis on ne will of course obliterate the secondary accent in rapientis. Compare, however, the view of this line suggested in § 24, ante. (2) I have of course made the selection quite at random, so far as fitness for the argument is concerned. For convenience of reference I chose those six of the Ambrosian hymns given in March's "Latin Ifymns" which are clearly not intended to be quantitative. — 4i — qui lux es et dies", "Aurora lucis rutilat", "O Rex aeterne Domine", and "Mediae noctis tempus est". In the first of these hymns, which is probably the oldest, (having been persistently, though of course erroneously, ascribed to Ambrose himself), there are no polysyllables in which the regular accent does not coincide with the normal metrical stress. In the last named of the six, which is probably also the latest,(i) there are 8. In all the hymns together there are 18 such polysyllables, in a total of 236 lines. A decided majority of these (13) occur in the first halves of their respective lines, so that that those lines, if read accentually, present only the familiar initial versions, single or double. E. g. in the second of the hymns named we have Oscula petit Dominum (14) and Innocens et innoxius (19) but also Fallaces Judaei impii (25) and in the third we have Oculi somnum capiant (13) Dextera tua protegat (15) and Famulos qui te diligunt (16). It looks very much as if these polysyllables, where their accent affects only the first foot, or the first two feet, were meant to be read as in prose, — otherwise (perhaps) to be forced into the regular rhythmical scheme. If this can be believed, it would seem that at a very early date, — as soon as the accentual principle was fairly grasped, — the propriety and beauty of occasional in- versions came to be appreciated. We should then explain the verses of Fortunatus somewhat differently: for his irregularities could be called ordinary inversions. His (1) See March's note: he assigns it to the 6th century. I should think, from the irregularities in the numbers of syllables (11. 33, 34) and the intermixture of trochaic lines (11. 46, 47) that it must be as late as that. It could hardly have been written before the accentual idea was firmly grasped. — 4 2 — hymn would then be regarded not as written in quantity, with careful observance of prose accents, but as written accentually, with all convenient regard for the rules of quantity. This, indeed, seems the more plausible view, but in the present state of our knowledge it is hard to see how any certain inferences can be drawn. CHAPTER IV. Early Church Music. Syllabic verse. § 28. Introduction. The opinion advanced in the foregoing pages as to the origin and nature of the late Latin rhythmical versification is of course too obvious to be wholly novel,(i) but lately, in the search for more remote explanations, its obviousness has become obscured. The purpose of this chapter will be to explain away one of the most serious misunderstandings by which students have been led away from the plain path. There_jrejn_^ existence several specimens of late Latin verse of a strictly syllabic character, in which even the most strenuous upholder of the accentual theory can find no trace of the accentual principle. These have been regarded as proving (by obvious analogy) the absence of that principle from other forms of verse '.(2) — but it is believed that they can easily be explained as not inconsistent with the accentual theory. Most of the early "rhyth- mical" verse was meant to be sung, and sufficient attention has not been paid — in the investigation of the "rhythms" themselves — to the exigencies of the music which accompained them. It is believed that a con- sideration of the diverse styles of music in use in the early churches will materially aid the present investigation. (1) See, for example, Gautier, I. 281 et seq., where similar con- clusions are reached, though by somewhat questionable reasoning. (2) Cf. Meyer's treatment of Augustine's Psalm. — 44 — Our information about early Christian music is scanty. "However persevering" says Fetis(i) "may be the histo- rian's efforts to collect information about the music of the Roman church during the first four centuries, they must result only in a conviction of their futility, for no trustworthy document on that subject is now extant. A few words of Tertullian, dealing in generalities too vague to be of any use to us, are all that we can cite". It would seem, therefore, that most of what we can learn about the music of this period must be by inference from what we know of the state of music before the Christian era and of the state of music after the fourth century: but we shall find that there is some contemporary evi- dence which, though indirect, will be of no mean value. §29. Greek music. Fortunately the aspect of the subject that is most nubilous is one which does not concern our present inquiry at all, — namely, the question as to the modes and keys employed in the early churches. The various modes in vogue at different times seem to have had none but the most indirect effect upon the time and rhythm of the melodies. These latter are the elements with which versification is most intimately connected, and they can be studied independently, so far as they need now be studied at all. The most striking characteristic of the religious melodies composed by the Greeks immediately before our era was this, that they followed closely, in time, the quantities of the syllables to which they were sung.(2) A long note was not sung to a short syllable, nor vice versa-. — and in general each syllable was represented by a single note in the melody. A few measures from one (1) Vol. 4, p. 147. Tertullian's valueless testimony is given on p. 5 of the same volume. (2) The contrary view was formerly held, on the authority of a passage in Dionysius of Halicarnassus : — See Chappell, I. p. 172. But cf. also Monro, p. 117. — 45 — of the musical inscriptions recently exhumed at Delphi will make this clear. ^m 5T3I V w w r i-va il, I. 120, I noted four exceptional lines, two of 1 7 and two of 1 5 syllables each. [Of course the synizesis in Evangelium (4 syllables) and Ecclesiam (3 syllables) presents no difficulty]. Several verses, however, are made regular only by Du Menl's conjectural emendations. The careful examination of the Benedictine text made by Meyer (q. v. p. 20 et seq.) shows 21 irregular lines:— but this number is too liberal, as it includes several cases in which an obvious syncope is all that is needed to make the line normal. — 5° — ends with a paroxytone; so also do most of the first hemistichs: but there is no further accentual regularity. Each line ends with the vowel e, thus riming crudely with all the rest. By the frequent recurrence of the first line the whole poem is divided into stanzas, which are further distinguished by the alphabetical consecutiveness of their initial letters. Each of these stanzas contains an even number of verses, (either 10 or 12), and at the end there is a concluding passage of 30 verses.(i) § 33. The music of Augustine's Psalm. Now as Augustine was clearly not striving for poetical effective- ness, these various elements of regularity must have been introduced for the sake of the music; — and they seem to show pretty plausibly what the general style of that music was. There must, in the first place, have been a very thorough correspondence of note for syllable and syllable for note, or the syllabic structure of the verse would hardly have been maintained so rigidly; — it was certainly of no intrinsic value. Secondly, the music must have been neither prosodical nor strongly rhythmical ; in other words it was probably so slow and sustained that it made no difference, as the notes dragged along, whether the syllables that accompanied them were long or short, accented or unaccented. Thirdly, the music, though so unrhythmical in general movement, must have taken the form of some sort of tune, or melodic progression, which was repeated or at least imitated after every second verse. And lastly, there seems to have been something like a rhythmical cadence, involving the last two or three notes of each bar, which relieved the melody of its creeping monotony. (I) These facts are stated in substance by Du M6ril and in detail by- Meyer. The careful analysis given by the latter makes it unnecessary for me to state more than the general results of my own examination of the psalm. - 5i — That this was the character of the music is perhaps not conclusively proved by the versification of the psalm: — but the probability amounts to more than a reasonable hypothesis, for no other hypothesis affords a satisfactory explanation of the verse. The unrhythmical nature of the music in general is all the more probable because of the care with which verbal rhythm is secured at the verse-end, where the music would naturally, in any case, be somewhat rhythmical. A cadence-effect would indeed be produced at the end of the bar by the mere occurrence of a pause after the hemistich, for the ear in such a case is forced to dwell on the last two or three notes and make a cadence of them, even when they are actually equal in time and stress ; and that such a pause occured here is clear from the author's practice in the matter of elision: but the rime at the verse-end, and the almost uniform feminine endings of the hemistichs show that in fact the verse-end was still more strongly marked, and of course the music corresponded. It is a significant feature of the verse that where verbal rhythm was de- manded it was secured not by quantity but by accent. (See the short accented penult in mare 1. 4 of the extract.) To Augustine, as to Sedulius and Fortunatus, verbal accent was the natural concomitant of rhythmical ictus, wherever such ictus needed verbal reinforcement at all. Other compositions in verse of the same character as that just described are best explained in the same way. As a single additional specimen of such verse it may be interesting to refer to the ancient hymn in honor of St. Patrick.(i) It begins as follows: — Audite omnes amantes Deum, sancta merita viri in Christo beati, Patrici episcopi; quomodo bonum ob actum similatur angelis, perfectamque propter vitam aequatur apostolis. (1) Found in an 8th cent. MS. :— said in Julian's Diet, to have been composed perhaps about 458, but I do not know upon what authority this conjecture rests. — 52 — Beati Christi custodit mandata in omnibus, cujus opera refulgent clara inter homines, sanctumque cujus sequuntur exemplum mirificum, unde et in coelis patrem magnificant Dominum. Trie two hemistichs of each line are of 7 and 8 syllables respectively, and are neither quantitative nor accentual, — except that the poet shows a preference for a feminine ending in the first hemistich and for a masculine ending in the second. The hymn is divided (according to the letters of the alphabet) into stanzas of four lines each. Elision does not occur at all in the two quatrains given above, and it is rare in the rest of the poem. § 34. The two schools of church music. It has been shown that antecedent probability, and the evidence of two distinct kinds of verse in the early hymns, favor the theory that there were in the 4th century two kinds of music in common use. That which prevailed at Milan perpetuated the traditions of Greek culture. It was taken up by the church at a time when Christianity was no longer vulgar ; and we have seen from Augustine's auto- biography that the congregation included at least one woman of exceptional culture. This music seems to demand a verbal rhythm in whatever verses were set to it. But at the same time, there seems to have been another style, prevailing elsewhere, which made no such demand. But the existence of two styles of music in the early church is a matter of history; and the evidence of tra- dition, as well as that of later musical development, points to just such a distinction between them as has here been drawn. The two styles are commonly known as the use of Milan and the use of Rome, and are associated with the names of the two great reformers, Ambrose and Gregory, respectively. Without reviewing in detail the a posteriori arguments of the historians of music, it remains only to cite their testimony as to the general results. — 53 — In Fetis's Dictionary we find the following :(i) — "La prosodie et le rhy thine paraissent avoir disparu de la langue latine chantee an temps de Saint Gregoire: on croit qu'il acheva de 1'effacer, et que, dans son anti- phonaire, toutes les syllabes etaient notees a temps egaux". And again 1(2) "La distinction entre le chant gregorien et Tambroisien consista done ordinairement, d'une part, en ce que celui de saint Ambroise etait la tradition du chant de 1'eglise grecque, avee ses ornements et 1'usage de certaines suites de sons chromatiques, * * * tandis que la reforme de saint Gregoire fit disparaitre ces successions de sons etrangeres au chant diatonique; d'autre part, le chant ambroisien etait rhythmique, et le gregorien ne l'etait pas. Mais par la suite des temps, ces differences essentielles ont disparu, et depuis plusieurs siecles on n'apercoit plus de distinction saississable entre ces formes du chant ecclesiastique." The obvious explanation of the so-called reforms of Gregory is well stated by Apel as follows :(3) — "Was so- gleich in die Sinne fallt, dass namlich der accentuirte Gesang, der sich in Hauptmomenten bewegt, weit mehr geeignet ist von grossen Volksmassen gesungen zu werden, als der quantitirende, weil jener ungebildeten Stimmen zu Hilfe kommt, die sich bloss dem kunstlosen Naturgefuhl von Arsis und Thesis zu iiberlassen brauchen, und iiberdies grosse Tonmassen sich allezeit anstandiger und wurdevoller in gleichen Zeitraumen fortbewegen als in ungleichzeitigen : dieses bemerkte auch Gregorius und begriindete auf diese Bemerkung seinen Plan zur Re- formation des Kirchengesanges." (1) Fetis Diet. s. n. Gregoiie. (2) lb. s. n. Ambroise. This is curiously inconsistent in some respects with dicta in F£tis' history : — but there is such a general unanimity on the main point att issue, among tfee best authorities, that h has seemed not worth while to re-open the questions involved. (3) II. § 498, as quoted by Ambros, II. 60, — 54 — And finally the most scholarly of all the recent treatises on this subject contains the following :(i) "Man pflegt, wie gesagt, den Unterschied zwischen dem Am- brosianischen und dem Gregorianischen Gesange wesent- lich darin zu suchen, dass jener Langen und Kurzen unterschieden, dieser die unterschiedlos gleiche Dauer aller einzelnen Tone eingefuhrt habe. Richtiger hiesse es vielleicht : dass der Ambrosianische Gesang wesentlich auf der poetischen, der Gregorianische auf der musika- lischen Metrik beruhte. * * * Die eigentliche Bedeutung der Gleichdauer der Bewegung des Gregorianischen Ge- sanges liegt aber nicht in dem taktmassigen, gleichlangen Aushalten jeder Note, sondern (im Gegensatz gegen die metrischen, d. i. die prosodische Eigenschaft jeder Silbe zur Geltung bringenden Gesange) darin, dass an sich alle Sylben ohne Riicksicht auf Prosodie fur vollig gleichbedeutend, fur isometrisch genommen werden, und daher nach den Bedurfhissen des Rhythmus die proso- disch lange Silbe auch in der Geltung einer kurzen ge- nommen werden kann und umgekehrt, und bloss die Ge- setze der naturlichen Declamation zu berucksichtigen sind." Here it is to be observed that there are some striking differences of opinion as to matters of detail. Ambros by no means believes that the music of Gregory's ce- lebrated antiphonary was wholly unrhythmical: — and Fetis thinks that the style of Milan, though strictly pro- sodical, was distinguished by a much more lavish use of musical ornament. But as to these essential facts their conclusions are the same: — that the Ambrosian music was rhythmical, and necessitated verbal prosody, but that the Gregorian, whether rhythmical in itself or not, made no such demands upon the verse. § 35. The influence of the church music on versi- fication. Ancient tradition has it that Gregory's system (1) Ambros, II. 59-61. — 55 - ■ was not a new invention, but largely a compilation and revision of systems that had prevailed in the early churches : his object was not to introduce a new system, but to reduce to uniformity the diverse practices of different communities.(i) As to the most distinctive feature of his system, it is evident from what has gone before that this tradition must be correct; for St. Augustine's psalm shows that this feature, though perhaps in a much cruder form, was familiar as early as the 4th century. As it was apparently popular in Augustine's time, it may well have existed from the beginning of our era, or even earlier. The significance of all this is apparent enough. During the 4th, 5th and 6th centuries, which as we have already seen witnessed the transformation of metrical into rhythmical verse, two radically different styles of music were in vogue. One style was at first strictly prosodical, and always strongly rhythmical. The other was so unrhythmical (or, as Ambros would say, so inde- pendent of verbal rhythm) that even in the days when prosody was not yet a thing of the past, it could and did disregard it. The former style (the Ambrosian) was strict in its demands upon the verse, for it was at first only by a prosodical correspondence between words and music that the difficulties of the music could be over- come: hence we find a class of hymns in which, as prosody died out, regular accentual versification took its place. The old prosody could be discarded, for the new rhythm answered the same purpose better. Later still, as the unified system of Gregory began more and more to assert itself throughout the church, the stricter iambic rhythm might, so far as the music was concerned, be dispensed with, and the poet would need to practice only what he had learned about euphony of verse, without (1) Ambros, II. 43, 44, and esp. the authorities cited p. 43, n. - 56 - reference to musical exigencies: — and it is probably at this time that the wrenched accent yielded to the in- verted foot. The Gregorian style, on the other hand, or rather that style which afterwards was associated with the name of Gregory, made no such demands on the verse. Whether or not it possessed some sort of slow and stately rhythm, it was at least easy for choirs to sing it without verbal prompting, and so adequate verses could be written for it without much regard for either prosody or accentual rhythm. The Gregorian style required that the verse should be strictly syllabic, for note and syllable must correspond; but it did not, like the Ambrosian, require the verse to be rhythmical also. We are now in a position to discern the error into which students of this subject have fallen. Meyer, having failed to discover the true principle of Commodian's verse, was led to believe that in the change from metric to " rhythmic " only the syllabic principle survived. He asserted that this was the case even in the so-called rhythmical hexameters, where upon his own showing the first hemistich contains either 5, 6 or 7 syllables, and the second either 8, 9 or 10. Next he observed that all the Latin "rhythmical" verses (or nearly all) exhibit the syllabic principle in rigorous application, while some of them show an apparent accentual structure and others do not. His natural conclusion was that the syllabic principle was the essential thing in all, and that where an accentual rhythm seemed to show itself it was merely accidental. The true explanation is already too clear to need re-statement. Meyer is not content, it is true, with the assumption of mere survival of the syllabic principle from the clas- sical verse-forms. He contends that the Latin poets adopted from the Oriental literatures the discovery that this principle was sufficient in itself to build verses upon. — 57 — The present writer is not familiar enough with this branch of the subject to criticize the argument in detail,(i) but it is certainly unnecessary. St Augustine, for example, needed no foreign model. As we have seen, he apolo- gized for not writing any real kind of verse. The kind of verse which he thought might have been expected of him was doubtless a form of the trochaic tetrameter: it would have been both quantitative and syllabic. As, however, Augustine was not writing poetry for its own sake, he dropped the quantitative element altogether; but he kept his lines strictly syllabic because he had to. Those of the Ambrosian hymns which contain complex inversions might possibly be regarded as imitated from foreign models; or, perhaps more plausibly, they might be regarded as due to the influence of such syllabic verse as that of Augustine's psalm: — but, as we have seen, it is much easier to regard them not as evidences of the naked presence of the syllabic principle, but rather as mere euphonious varieties of the accentual type. § 36. Contemporary accounts of the Latin rhythms. Kawczynski, in his very interesting treatise on the Latin rhythms, has followed Meyer to very similar conclusions. Only one branch of his argument needs to be touched here. By copious quotations from early grammarians and rhetoricians he maintains the thesis that his (and Meyer's) conception of the Latin rhythms was also the conception of the scholars of the period under discussion. There is something, perhaps, of scholastic nicety in Kawczynski r s argument, but it nevertheless seems at first sight formi- dable. It is clear, however, that a form of verse which arose after the middle of the 4th century, cannot be satis- (1) Except that as to the prominent part which he assigns to the Hebrew poetry in the Latin movement Meyer must certainly be wrong. The latest results of Hebrew scholarship seem to negative the idea of any syllabic structure in the Jewish psalmody. - 58 - factorily explained by the definitions of grammarians who were then dead. Thus when Diomedes, in the 3th cen- tury, says " Rhythmus est pedum temporumque junctura cum levitate sine modo",(i) we are inclined to doubt whether this definition was framed after actual ob- servation of such verses as we are considering; we are tempted to explain it rather as an unthinking repetition of a subtle distinction made by the Greek and Alexan- drine theorists : and in any case, it tells us no more than we have already learned from Augustine's psalm. The only real difficulty to be encountered in Kawczynski's quotations, — and that not a serious one, — is the persistent application, by theorists after the time of transition, of the same old word " rhythmus " to the new phenomena of what we now call rhythmical verse. Thus one of his quotations is from the venerable Bede (675 — 735 A. D.), viz :— " Videtur autem rhythmus esse metris consimilis, quae est verborum modulata compositio non metrica ratione sed numero syllabarum ad judicium aurium examinata, ut sunt carmina vulgarium poe- tarum ".(2) Here the word is certainly applied to the kind of verse that we have been discussing. Does its use signify that Bede saw nothing more in it than is set forth in the definition of Diomedes ? Here we might fairly take refuge in the belief that novelties in art can be better explained, in their theo- retical aspect, by the students of a later age than by contemporaries. With a dozen centuries of accentual poetry to study, we are less likely to misunderstand the sub-conscious purposes of its originators than were those who had the simple advantage of living with them. But we are by no means driven to rely entirely upon this rather bald assumption of superiority. Let us add to (1) Gram. Lat. I. 473, as quoted by Kawczynski, p. 115. (2) Kawczynski, p. 116. — 59 — Kawczynski's excerpt the rest of Bede's paragraph. " Metrum est ratio cum modulatione ; rhythmus modu- latio sine ratione; plerumque tamen casu quodam in- venies etiam rationem in rhythmo non artificis mode- ratione servatam, sed sono et ipsa modulatione ducente, quern vulgares poetae necesse est rustice, docti faciant docte: — quomodo et ad instar iambici metri pulcherrime factus est hymnus ille praeclarus: — [0](i) Rex aeterne Domine, rerum creator omnium, qui eras ante saecula, semper cum patre filius, et alii Ambrosiani non pauci. Item ad formam metri trochaici canunt hymnum de die judici per alphabetam ; — Apparebit repentina dies magni Domini, fur obscura velut nocte improvisos occupans."(2) Here Bede seems positively to confirm our argument. It is evident that such lines as Rerum creator omnium, with their apparent inversions of foot, must have been a serious puzzle to the theorists. Bede (as Kawczynki shows) borrowed most of his metrical science from the earlier grammarians, and nothing of this kind had been explained by them. Such verses were agreable enough when weighed by the " judicium aurium ", but who could divide them into feet ? They seemed to have no method in their irregularity. But it is evident enough from the words "instar iambici metri" and "ad formam metri (1) Note that as Bede quotes the first line of this hymn its syllabic regularity is destroyed, while a certain accentual rhythm remains. MSS. vary as to the presence or absence of the interjection. Similar irregularities, by the way, are found in lines 53 and 59 of the same hymn, but they did not prevent Bede from praising it. (2) De Arte Metrica, § 24. - 6o — trochaici" that the practical effect of these verses was to Bede's ear just what it is to us : — and he thought that the imitation of the classical iambics had been " most beautifully done ". The difficulty that perplexed Bede, and drove so many of Kawczynski's authorities to deny the existence of any " modus " in these verses> was a purely theoretical one. The same difficulty is encountered in modern English verse. Theoretical investigators are not now agreed as to the essential principles which control poets of the 19th century: they do not know whether our blank verse is divisible into feet or not, nor whether it requires in each line three, four or five accents, or indeed any determinate number. If the Latin grammarians could have had revealed to them a few lines of Paradise Lost, they would doubtless have described it substantially as they did their own accentual verse-forms ; for in both, the underlying principles of internal structure are sub- stantially the same. Their definitions are not, therefore, to be cited as authoritative, but should be set down as due to an interesting bewilderment, which they could hardly have escaped without superhuman insight. § 37. Conclusion. It is unnecessary to follow in detail the later history of the rhythmical versification. The latest developments of the Ambrosian hymn, which we have already examined, were but the logical con- sequences of such work as we have found done by For- tunatus. From Ambrose himself to Adam of St. Victor, the line of descent is direct. It was doubtless in the 8 syllable iambic verse that the change from metric to rhythmic was effected; and this was done under the fostering influence of the Ambrosian music; — but how large a share of credit must be assigned to the latter, we of course cannot tell precisely. The new principle spread rapidly to other forms of verse, and at an early, though now uncertain date, hymns in trochaics began — 6i — to show an equally marked accentual rhythm. The hymn "Ad perennis vitae fontem", of which a few lines have already been quoted,(i) shows what we may per- haps regard as a development from the style of verse employed by Augustine, evolved under the influence of the new idea. Mone thinks this hymn was written about a century after Augustine's time. Perhaps it would be safer to put it a little later, but we cannot be exact. It is very interesting to note that the Ambrosian hymn, which began with perfect iambic regularity, developed in the middle ages into a form so irregular as to make modern scholars call it purely syllabic ; while the trochaic rhythm, which seems to correspond with the purely syl- labic verse of Augustine, became almost perfectly regular when the accentual principle took possession of it. Here we have a most striking refutation of the view that historical influences are stronger than the natural con- ditions of language and the fitness of things. The proof that the Ambrosian verse was governed by the accentual principle, is clinched by comparison of it with the work of its French imitators. Consideration of these latter, however, is more conveniently postponed to the next chapter. Evidence many also be found in iambic hymns other than octosyllabic. One of the earliest of these is the "Aurea luce et decore roseo ", ascribed to Elpis. If this hymn was really written by the wife of Boethius, it must have been as early as the first quarter of the sixth century; but here again we have to rest in doubt. The hymn contains 28 lines of 12 syllables each, and there are only two instances of inversions other than initial. Something has been made of the verse of the mystery play on the foolish virgins, in which the Latin verses, though so late, are plainly not accentual. Part (1) § 8, ante. — 62 — of the verses are Latin and part French, and in both the structure is for the most part merely syllabic. It is sufficient to say of this, however, that the verses were obviously written for musical recitation, and the Latin metre was designed to correspond with the French. The bad quality of the Latin shows that French was to the author the more familiar language. It will be made clear in the next chapter that a Latin imitation of a French form of verse can prove nothing as to the nature of true Latin verse. CHAPTER V. Early French Verse. § 38. Introduction. It has already been intimated that the accentual rhythm of mediaeval Latin poetry ex- hibits a fair parallel to the ordinary rhythm of modern English verse. Thus in the lines Ad perennis vitae fontem mens sitivit arida, and Aurea luce et decore roseo, we find the same rhythms that are so familiar in Comrades leave me here a little, while as yet 'tis early morn, and This is my son, mine own Telemachus. It has been seen that the licenses which the poets of the two languages have allowed themselves have been in some respects different: but their verse-forms in general have been controlled by the same essential principles. Now Latin rhythmical hymns have been familiar to English poets of every generation since the time of Bede, and it seems not unlikely that our verse would have assumed substantially its present forms, if it had had no other foreign model to imitate. In that case this investigation might fitly end here. But in point of fact our most important verse-forms, so far as they have - 64 - been borrowed at all from abroad, have been borrowed from the Latin largely by an indirect process. The Latin influence was felt, during a critical period in the history of our versification, almost solely through the medium of the French. It is one of the remarkable facts in the history of the subject, — one of the facts tending most strongly to negative Kawczynski's theory of the supreme power of external influences, — that though the French system itself was widely different from the Latin, yet when it came in contact with the English it had the effect of moulding the latter into substantially the Latin form. The remaining chapters of this essay will briefly review the development of the French system out of the Latin, and its influence upon the English. § 39. Modern French verse. The essential principles of modern French versification are most clearly discern- ible in the Alexandrine. A fairly typical passage is subjoined, to illustrate the main features of this verse. Quoi, madame! parmi tant de sujets de crainte, Ce sont la les frayeurs dont vous etes atteinte? Un cruel (comment puis-je autrement l'appeler?) Par la main de Calchas s'en va vous immoler; Et lorsqu'a sa fureur j'oppose ma tendresse, Le soin de son repos est le seul qui vous presse! On me ferme la bouche! on l'excuse, on le plaint! C'est pour lui que Ton tremble; et c'est moi que Ton craint! Triste effet de mes soins! est-ce done la, madame, Tout le progres qu'Achille avait fait dans votre ame? iphigenie III, 6, 73 — 82. Here we have verses of twelve(i) syllables each. The 6th and 12th syllables are always tonic, and there is a fixed caesura after the 6th.(2) There is no other syl- (1) The 13th syllable in the feminine rimes need not be counted, as it is not pronounced. (2) The modern romantic forms of the Alexandrine need not concern us in the present inquiry. - 65 - lable in the verse which must be tonic, or which can be said even to be generally so; and on the other hand there is no place in the line which must be filled with an atonic syllable. The 5th and nth are commonly atonic, but not necessarily so. Owing to the weakness of the word-accent in French, — the fact that it tends continually to lose itself in rhe- torical emphasis, — the scientific principles which underlie French versification have not been easy to discover. Even now so keen an observer as Stengel suggests(i) that such verses as those above quoted should be re- cognized as divisible into feet of equal length ; that they move with a definite accentual rhythm, and that the frequent conflict of accent with ictus is one of the chief beauties of the form. This however is not the prevailing opinion. There is in reality no conflict of word-accent with verse-accent, because the latter, in the ordinary sense of the word, does not exist at all except at the verse-end. Of the six syllables in each hemistich, one, two, three or four may be tonic; and provided one of these is in the 6th place, the others arrange themselves in obedience to no other laws than those of euphony. It frequently happens, of course, that a French Alexan- drine gives substantially the rhythm of the corresponding English verse. Thus we find this rhythm in Et lorsqu'a sa fureur j'oppose ma tendresse: and such a line as Quoi madame! parmi tant de sujets de crainte may with difficulty be fitted to the English scheme:— compare Spenser's As the God of my life? Why hath he me abhord? F. Q. I. 3. 7- But this is an extreme case of the English Alexandrine, (1) p. 8, § 13. — 66 — and the parallel even here is forced. Such verses more- over, as On me ferme la bouche! on l'excuse, on le plaint! Cest pour lui que Ton tremble; et c'est moi que Ton craint! cannot be regarded as possessing any rhythm except an anapaestic one. Their rhythmic parallels in English are such lines as Byron's For the Angel of Death spread his wings on the blast. The Destruction of Sevnacherib. The only apparent element of regularity, then, in the internal structure of this style of versification, is the uniform number of syllables. That a Teuton like Stengel should regard the ordinary accentual reading of French poetry as destructive of its proper character, and should contend for an ideal iambic rhythm, is therefore not to be wondered at, for the Germanic ear is not satisfied without a more or less regular ictus.(i) It is true that even to the Frenchman, with his habit of approximately equalizing his stresses in ordinary speech, the rhyth- mical movement of French poetry is far less of a sen- suous gratification than that of English poetry is to us. Its nearness to prose is a serious disadvantage,^) and (1) For example Koster {Schiller als Dramaturgy p. 96) regards a trochee in German blank verse, even when initial, as a violation of law, justifiable only in special cases. German poets are of course much more chary of inversions than the English. (2) For example, " Les vers ou l'on traite des sujets familiers et simples ne sont que de la prose mesuree, ou sont obliges de s'ennoblir a l'aide de la p£riphrase et de l'emploi du mot noble en place du mot propre " — G.Paris, JJ Accent Latin, 125. The comparatively low value of mere syllabism, from an aesthetic point of view, is well illustrated by a recent incident of French criticism. " The French Academy has the bestow- al of a prize, — le prix Archon-Desperouses, — for the ' most notable volume of verse during the year'. This year the prize was voted to the 'Maison de VEnfance * by M. Gregh. But after the decision a horrible discovery was made. Some of M. Gregh's Alexandrines had thirteen or even fourteen syllables: — some lacked the regular caesura; some were loosely rhymed. - 67 - for the elements of supreme beauty the French poet has to resort to other devices than those of rhythm. As compared with English or German verse, then, that of the French poets is almost purely syllabic. Praise- worthy efforts have been made(i) to analyze it by feet, and the laws governing the various verse-forms have been ascertained by forcible and (as it seems) sound reasoning. But these laws are in reality the laws partly of chance and partly of euphony, — not strictly speaking the laws of French poetry. French verse, to adapt the description of Bede, "est verborum modulatio non rhyth- mica ratione, sed numero syllabarum ad judicium aurium examinata": and the question what will and what will not satisfy the " judicium aurium " is in a sense a general one, belonging to the whole study of aesthetics rather than to French versification in particular. The structure of other forms of French verse is the same in principle as that of the Alexandrine. Without considering details it will suffice to point out that all forms are, in general, syllabic and not rhythmical. Thus the sonnet of Alfred de Musset's beginning J' at perdu ma force et ma vie is written throughout in octosyl- labics ; — but it contains lines which could not be matched, as to accentual rhythm, by lines from any one English poem. For example, the line Dieu parle, il faut qu'on lui reponde is like On Linden when the sun was low: and Ici bas ont tout ignore The poems might be interesting, but how could Academicians overlook their irregularities ? They finally compromised by adopting a minute which practi- cally said, 'As poets we approve of M. Gregh, as Academicians we con- demn him". N.-Y. Evening- Post, June 7, 1897, P« 6. (i) Notably by Lubarsch. s* — 68 — may be matched by a line from Tennyson's "In the Garden at Swainston " With a love that ever will be. But the latter poem contains no eight-syllable iambic verses, and Hohenlinden contains no verse like the last quoted: while in the French poem the two lines given above, though so different in rhythm, are used and re- garded as equivalent. The same is of course true of the io-sy liable verse, — the only other form which we need consider. The be- ginning of Beranger's beautiful Chant du Cosaque illu- strates the ordinary form of this verse. Viens, mon coursier, noble ami du Cosaque, Vole au signal des trompettes du Nord. Prompt au pillage, intrepide a l'attaque, Prete sous moi des ailes a la Mort. In this, as in the Alexandrine, there is a fixed caesura. The verse thus resolves itself into hemistichs of 4 and 6 syllables respectively. But these hemistichs, with the exception that the last effective syllable must be tonic, are purely syllabic. Thus while the 4th line of the extract seems like a line of English heroic verse with the ordinary initial inversion, the 26. line corresponds rather with our 4-accent verse. Such lines as Beat to the noiseless music of the night! and That a calamity hard to be borne? (both from Maud), though each contains just 10 syllables, are ordinarily, of course, felt in English to have very different rhythms. § 40. The origin of French versification. The dis- cussion as to the origin of French versification has chiefly centred about the decasyllabic verse. This has been variously derived from the Phalaecian, the Sapphic, the - 6o - iambic trimeter, the trimeter scazon, the dactylic hexa- meter, the dactylic trimeter hypercatalectic, and the Old High German long line of four accents. Finally Stengel derives it from a hypothetical popular Latin verse of 14 syllables, which in turn was directly descended from the Saturnian ; — thus assuming a continuous stream of verse, 13 centuries long, all of which is lost.(i) In criticism of this varied assortment of supposed ex- planations there is little to add to the excellent remarks of G. Paris.(2) "Avant d'essayer de montrer comment s'est constitue la systeme de la versification francaise, il faut etudier comment s'est etabli, a l'epoque anterieure, le principe de la versification rhythmique en regard de la versification metrique. Une fois ce principe constitue, les differents vers en sont naturellement issus, sans que chacun d'eux ait un rapport direct avec une des formes de la versification metrique, d' origin e grecque, de venues toutes, pour le peuple, incomprehensibles avec le prin- cipe meme de cette versification." This seems to assume that the mediaeval Latin rhythms, like the modern French, were not accentual; that the two systems are fundamentally the same, and that therefore, when the origin of the former is found, the origin of the latter is found along with it. (3) Now we have already seen that the Latin rhythms were (1) See Stengel, pp. 14 — 20 for this derivation, and for the biblio- graphy of most of the others. Substantially the same theory as Stengel's was proposed as early as 1844 for the Italian hendecasyllabic by L. G. Blanc, Gramm. der Ital. Spr. p. 706. (2) Romania XV, 1 38 (Quoted also by Stengel). (3) It is especially easy lor a Frenchman to overlook the accentual character of the Latin rhythms, as the important part played in them by the secondary accents of Latin words can hardly appeal strongly to his ear. I trust it does not seem arrogant to admit a certain preference for the judgment of our English senses in these matters. Our own training happens to be better adapted to the purpose than that of either the French or the Germans. — 7o — accentual: — but it is also true that the earliest French verses were accentual also, and the substance of M. Paris's remark is therefore still good. We shall find that the French versification in its earliest form was a mere imi- tation of the Latin rhythmic system, as the latter has been explained in the foregoing pages. The change to the system now prevailing in French took place after the period of this imitation was past, and was presumably due to causes inherent in the language itself. Our search, in obedience to M. Paris's precept, should be for general principles rather than for single parallels: — but as it happens we shall find that the familiar verse of the Ambrosian hymns, and its French offspring, exhibit the true principles most clearly. § 41. The earliest octosyllabics. That the Eulalia sequence was an accentual poem has been mentioned already, but we are not directly concerned with it now because it is unique in its peculiar style. It is important here only as indicating, perhaps, that the natural ten- dency of the earliest French verse-makers was toward accentual composition. The next in order of the early monuments of French verse are the Passion-poem (516 lines) and the Life of St. Leger (240 lines), both of the 10th century, and the Fragment of the Alexander Romance ascribed to Alberic of Besancon (105 lines), which dates probably from the first half of the nth century. All these poems are in 8-syl- lable lines, and the rhythm of all is unquestionably ac- centual, with a marked iambic movement(i) A few lines from the first-named poem will make this clear.(2) (1) This was first pointed out, I think, by G. Paris, in his remarks on the St. Leger, in Romania I, 295 ; but its significance does not seem to have been duly appreciated. (2) I take these from the beginning of the excerpt given in Bartsch's Chrestomathie, which is probably the text most easily accessible. — 7i — 117. Christus Jhesus den s'en leved, Gehsemani vil' es n'anez. Toz sos fidels seder rovet, Avan orar sols en anet. 121. Granz fu li dols, fort marrimenz. Si condormirent tuit ades. Jhesus cum veg los esveled, trestoz orar ben los manded. Here the rhythm is an obvious reproduction of that of the Ambrosian hymns, even the division into four-verse strophes being retained. There are frequent inversions, — as in the first foot of the 121st line, — but the general iambic movement is unmistakable. Indeed the iambic movement seems more marked than it does in the Latin hymns, because of the fact that initial inversions are generally only single, not double. Thus a tonic syllable is almost always found in the 4th place. The verse is thus characteristically distinguished from that of modern French, and in this respect it differs slightly even from that of the Latin hymns. The degree of regularity with which the 4th syllable was made tonic is worth noting. Of the 100 lines of the Passion which include and follow the above excerpt, there are 73 with a final tonic syllable in the 4th place (e. g. Toz sos fidels), and 25 with a penultimate tonic syl- lable there (Si condormirent). (i) The other two lines have each a final syllable at the fourth place, but that syllable is atonic. These lines are Melz ti fura(2) non fusses naz (151) (1) I have separated these two types of verse, lest the figures should seem to suggest that the early octosyllabic had a fixed caesura. This has sometimes been fancied, but is obviously not true. (2) It is barely possible that in these pluperfects the accent was thrown forward, contrary to the general principles of O. F. etymology. If so the line is regular. In 1. 168 we have Quar sua fin veder voldrat, where the assonance and syllabic structure both show that voldrat must be oxy- — 72 — and Mult lez semper en esdevint (210). Here of course the exceptions are so few that they do not even threaten the rule. In the poem on St. Leger the verse is substantially the same. Of the first 100 lines, 81 have a tonic final syllable in the fourth place, 13 a tonic syllable not final, and 6 a final atonic syllable. These last are the following: — Que il auuret ab duos seniors (8) Quandius visquet del reis Lothier (49) II lo presdrent tuit a conseil (61) A nuil omne nol demonstrat (78) Quant ciel ire tels esdevint(i) (79) Paschas furent in eps eel di (80) In the Alexander fragment there are 73 lines with a tonic final syllable in the fourth place, 26 with a tonic syllable not final, 4 with a final syllable not tonic, and 2 with a syllable neither final nor tonic.(2) The lines belonging to these last two classes are as follows: — En tal forma fud naz lo reys (54) Que altre emfes de quatro meys (57) Que altre emfes del seyentreyr (75) A fol omen ne ad escueyr (78) and Toylle sen otiositas (6) Ne ad emperadur servir (43) tonic; — but I should prefer to regard this as an instance of wrenched accent. Cf. 1. 134, Zo lor demandent que querent, where the wrenched accent is obvious. (1) A doubtful reading: but no plausible emendation normalizes the verse. In one line of this passage, viz (18) rovat que litteras apresist, we must either read letres, with G. Paris, or assume a syncope in the Latin word litteras. (2) Although G. Paris says [Romania I, 296): "Dans 1' Alexandre l'accentuation de la troisieme syllable, reste de la rhythmopo6e latine, a tout-a-fait disparu: le vers est toujours divise en deux moittes egales, accentu6s pareillement sur la derniere ". This I do not at all understand. — 73 — One of these last two lines, and perhaps both, may be regarded as regular if we recognize the principle of secondary accent: — but it is safer to class them as ex- ceptions. Lines certainly do occur with the 4th syllable neither tonic nor final; e. g. in the Passion, Barrabant perdonent la vide (225). § 42. The later development of French octosyllabics. After the middle of the nth century, this ceased to be the regular style of French versification. Examples do occur in which the adherence to the accentual movement is more or less rigid, but they are from Anglo-Norman poems, or from the work of poets exposed to the Anglo- Norman influence, and they are easily explained as due to the retroactive influence of English versification. In pure French octosyllabics, the modern system asserted itself early. This fact is of such importance in the his- tory of English verse that it deserves more detailed con- sideration. Schipper says:(i) "In der mittelalterlich-lateinischen Poesie, sowie auch in der romanischen, ist . . . eine regel- massige Aufeinanderfolge von starker und schwacher betonten Silben oder von Hebungen und Senkungen Gesetz, die beide von gleichem Wert fur den Rhythmus sind". And again, speaking of French octosyllabics, he says : " Zwei Verspaare aus dem Roman de Brut des Wace mogen das Wesen dieses Metrums veranschaulichen : Cordeille out bien escute et bien out en sun cuer note cument ses deus sorurs parloent, cument lur pere losengoent. Wir haben hier ein Versmass von im Ganzen jambischen Rhythmus vor uns ".(2) (1) P. 79. (2) Id. p. 107. This is a common error. E. g. Courthope, Hist. Eng. — 74 — But the rhythm of Wace's poem is by no means iambic. The first word in the passage quoted forms an anapaest (and should be written with a diaerisis, Cordeille), as is shown by its use at the verse-end eight lines before this excerpt: — Adunt apela Cordeille (riming with fille) ; and again 60 lines later we have Cordeille qui fu li mendre (riming with atendre). Indeed the whole passage in the original is quite as nearly anapaestic as iambic, as the ten lines before Schipper's quotation will show: — Mult a ci I dist il | grant amur, ne te sai | demander | graignur; jo te redunrai | bon seignur et la tier | ce part | de m'enur. Adunt I apela | Cordeille qui esteit | sa plus joes | ne fille; pur ce que il | l'aveit | plus chiere que Ragaii | ne la premiere quida | que el | e cuneiist que plus chier | des al | tres l'eust. Cordeil | le out bien | escute, etc. Indeed, of the first 100 lines of the passage in Bartsch's Chrestomathie (from which Schipper quotes), there are only 57 with a tonic syllable in the 4th place; and most of the other 43 are clearly anapaestic. This passage is, to be sure, rather more unfavorable to Schipper's generalization than others which he might have cited. Chrestien of Troyes, for example, exhibits Poetry I, no, says: "When through the genius of Chaucer the French iambic movement was naturalized in the Middle English, the triple move- ment, inherent in the old style, gave way before the new tendency". So also Crow, Zur Gesch. d. Kurzen Reimpaars im M. Eng., p. 6, says: "Im Franzosischen besteht das Schema des kurzen Reimpaars aus zwei durch Endreime gebundenen Versen von je vier jambischen Fussen, meist mit Caesur nach den zweiten ". — 75 — a slightly less strong tendency to the anapaestic move- ment. Thus in the first ioo lines of his Chevaliers au Lion there are some 24 that have a regular iambic rhythm throughout, as in Si s'est de lez le roi levee; (63) and 43 more exhibit the chief characteristic form of the earliest French rhythm, namely the tonic syllable in the 4th place, as in Uns cortois mors qu'uns vilains vis. ($2) But on the other hand there are at least 25 that have a markedly anapaestic movement, — i. e. two anapaestic feet and one iambus, in any one of the three possible arrangements : e. g. Li autre parloient d'amors (13) Si m'acort de tant as Bretons (37) Se feisoient cortois clamer (22) The remaining lines are of hybrid character, and it is not worth while here to attempt an exact classification. Of course, figures of this sort cannot be relied on for absolute precision, for the scansion of many lines depends upon one's individual taste. For example, the verse Et cil fable et man^onge en font (25) will entirely change its character according to the degree of rhetorical emphasis on cil. But the foregoing estimates have been kept purposely conservative: and it must be perfectly apparent that before the time of Wace or Chrestien the change in French verse had become estab- lished. Their verse, far from being "in the main iambic", was as purely syllabic as that of the 19th century. § 43. The difference between Latin and Old French verse. We shall now have to consider the reasons for such changes as we have seen in the octosyllabic metre ; —first the change from Latin to Old French, and second - 76 - the change from Old to Mediaeval French. Stengel, while recognizing the essential facts of the history, seems to give an entirely mistaken explanation of them. He seems to regard the early specimens of French accentual verse as the last examples of a decaying school, — a remnant from popular verse-forms that have perished, not from the well-known Latin rhythms. And he says:(i) "Gewohnte man sich in Anlehnung an den gleichsilbigen Vers der rhythmisch-lateinischen Verskunst fruh daran, statt an vierter ofters an dritter Stelle einen Wortton zuzulassen (freilich anfanglich nur, wenn als vierte eine Wortschlusssilbe folgte), so wurde damit gerade das Gegenteil von dem bewirkt, was die Be- tonung der vierten Silbe bezweckte, der jambische Rhythmus wurde verdunkelt, und damit erschien auch jede weitere Markierung desselben im Innern des Verses uberfliissig ". Here it is implied that the earliest French rhythms were in their accentual character, essentially different from the Latin; and secondly we are told that the change in French rhythms was due to Latin in- fluence. It is clear enough that both these opinions must be erroneous; but we will nevertheless examine them separately in detail. In the first place, is it true that in the Ambrosian hymns it was a common thing to accent the third syl- lable at the expense of the fourth? How far, if at all, was the distinguishing characteristic of early French verse lacking in that of the Latin hymns? For the purpose of answering these questions, a special examination has been made of the six Ambrosian hymns already used for another purpose.^) The results may be conveniently tabulated as follows: — (i) P. 4 6. (2) § 27, ante. — 77 — First line Total lines Accent Accent not on 4th on 4 th Jesu nostra redemptio 20 18 2 Hymnum dicamus Domino 32 26 6 Christe qui lux es et dies 24 23 I Aurora lucis rutilat 44 38 6 rex aeterne Domine 64 57 7 Mediae noctis tempus est 52 42 10 Total: 236 204 32 The lines tabulated as having an accent on the 4th syllable are of course either regular iambic lines, such as Ut ferres nostra crimina, or lines with a single initial inversion, such as Amor et desiderium. The right-hand column in the table includes all lines with double inversion, such as the title line of the first hymn. The table shows that on an average about six- sevenths of the lines in an ordinary Ambrosian hymn are accented on the 4th syllable. In the three passages of the earliest French verse already examined, the total ratio is about nineteen-twentieths. In other words, the proportion of comparatively regular lines is 11 per cent larger in Old French than in Latin. § 44. Explanation of the difference. This numerical discrepancy does not seem sufficient to count for much against the judgment of our ears as to the substantial identity of the two rhythms; but in order to eliminate the element of personal judgment it will be well to show that the discrepancy is in fact of just the sort that we ought to expect. In the first place, it must be remembered that some of the Latin verses which we have counted in the last column of the table, must probably have been brought - 78 - into conformity with the regular type by reading with wrenched accents. In such a passage as the following, from Aurora lucis rutilat, Tristes erat apostoli de nece sui Domini, quem poena mortis crudeli servi damnarunt impii, it is certain that crudeli must have been read as a propar- oxytone, although, as has already been shown, it is an extreme example of the wrenched accent. It therefore seems very likely that such lines as Fellaces Judaeii impii, and Laudes Deo cum cantico, (both from the Hymnum dicamus) were read with a similar license in the placing of the second accent. This is especially probable in the case of dissyllabic proper names, which are well known to have been treated with great freedom. Now in French this license was com- paratively impracticable. The Latin poets wrote in a language that was nearly or quite dead to vulgar con- versational use, and it is probable that they pronounced it in a measured manner, each syllable receiving its due value,(i) so that with the tradition of quantitative verse (i) The exact proportion of wrenched accents to inversions of foot, in the Ambrosian hymns, is so uncertain that I have not cared to enforce this branch of the argument as any thing more than a suggestion. The main reason, however, for my conjecture as to the pronunciation of Latin may be explained by reference to the treatment of rime in Latin and French, as compared with English. In modern French, identical rime (under the name of rime riche) is often desired: in English it is usually intolerable. The chief reason for this seems to be that the tonic accent is weak in French, — in English comparatively strong. Thus such a French rime as tombe, malgre, is almost like the English half-rime candle, pestle. The latter is improved by substituting cradle for pestle ; still more, perhaps, by bundle : and so in French the poet in many cases seeks identity of con- — 79 — behind them it was a small matter to throw forward the accent of a dissyllable. But the French poet had no such traditions to influence him in his use of the ver- nacular, and he used a language in which, while enun- ciation in general may possibly have been as measured as it is now, the final unaccented syllables are known to have been grievously scanted.(i) While therefore he could in Latin say homo, he could not so well say, in French, omne. If, therefore, we could fix the percentage of lines in the Ambrosian hymns which were actually normalized by throwing an accent forward, we could fairly deduct this percentage from the difference which we have to account for, since it would stand for a type which the Frenchman, in the nature of things, could not imitate.(2) In the second place, disregarding the possibility of occasional wrenched accents, if we read the Latin hymns accentually throughout, we find in them many inversions such as the French poet would not be tempted to in- dulge in. The greater number of the lines in the hymns which (under such an accentual reading) do not accent the 4th syllable, have their second feet filled by dis- syllables. E. g. in the Hymnum dicamus we find the following : Laudes Deo cum cantico, (2) Qui nos cruris patibulo, (3) Nullam culpam invenio, (22) Vita mundi suspenditur: (31) sonants before the riming vowel. A similar fondness for identical rime in the Latin hymns seems to me explicable only on the hypothesis of a similar weakness of the tonic accent in the half-artificial pronunciation of late Latin. (1) This, I suppose, will not be doubted. It is shown (1) by the curious variety in the spelling of these syllables, and their phonetic changes in general, and (2) by their subsequent disappearance from ordinary speech. (2) Wrenched accents are found in the Old French poems, as for example in the line cited ante, § 41, note; but they are rare. — 80 — and it is evident that this rhythm must result whenever the poet's convenience led him to begin a verse with two dissyllables. But when a French poet chose to do the same thing, he would find in most cases that the rhythm adjusted itself to the normal scheme, for most of the dissyllabic words in his vocabulary were oxytonic.(i) A third fact, tending further to explain why in- versions are less frequent in the first French poems than in their Latin models, is perhaps less obvious. When, in a line in an Ambrosian hymn, the 4th syl- lable is not accented, the accent is almost always found on the 3rd; — and in these cases, again, another accent is almost always found on the 1st: — in other words, a simple case of double inversion is presented. Now such a verse as Laudes Deo cum cantico, read strictly according to the word-accent, gives an agreeable rhythm which to an English ear does not seriously disturb the general iambic movement of the hymn ; — just as the general iambic movement of Paradise Lost is not disturbed by the double inversion in such a passage as There with my cries importune Heaven, that all The sentence, from thy head removed, may light On me, sole cause to thee of all this woe, Me, me only, just object of His ire. X, 933—936. But in a French verse in which the 3rd syllable was accented, it was less easy to put another accent on the 1 st. Such a verse might begin with a word of 1, 2, 3 or 4 syllables. In either of the last two cases the first syllable of the line could not be heavily enough accented to save the rhythm, for if the secondary accent existed at all in the Old French it was certainly pretty weak. (1) See, for example, the passage quoted from the Passion poem, §41, ante, in which the sixteen dissyllables are all oxytones. — 8i — In consequence, the verse would begin with a marked anapaestic movement, and the homogeneity of the rhythm would be at once imperiled. Moreover, if the line began with a dissyllable, the same result would generally follow, unless the dissyllable was one of the comparatively in- frequent variety with accent on the penult: — for other- wise, according to a principle familiar to students of modern French versification, the accent of the first word would usually be lost in the collision with that of the second.(i) Reference back to the several lines quoted as ex- ceptional from the oldest French poems will show that when the 3rd syllable was accented instead of the 4th, initial dissyllables were apparently not desired (unless, of course, they were paroxytones), and initial polysyllables were carefully avoided. In other words, the poet gene- rally sought those few combinations of words which would secure an accent on the 1st syllable also: but the effect, so far as we can judge, was not always suc- cessful. In Melz ti fura non fusses naz there is a strong rhetorical stress on the first monosyl- lable, and the effect is as good as that of Paschas furent in eps eel di; but in such lines as II lo presdrent tuit a conseil there seems to be a real change to an anapaestic move- ment,— the same that we found in Alfred de Musset's Ici bas ont tout ignore (1) See Lubarsch, p. 40 et seq. This principle of modern verse is due largely to the general weakness and mobility of French accents. This branch of the argument, therefore, loses part of its force if it can be shown that accent was very much stronger in Old French. That it was somewhat stronger, I have no doubt. 6 — 82 — and in Wace's Cordeille out bien escute. Since, therefore, any departure from the system of ac- centing the 4th syllable was likely to result in an ana- paestic movement, and so to disturb the general run of the verse, a French poet with the ideal iambic rhythm in his head would naturally be comparatively slavish in his fidelity to the strict scheme. Finally there remains one fact which may perhaps go farther than anything else to explain the regularity of these French poems. Their authors were not poets at all. The Passion and the Life of St. Leger reveal an impressive religious earnestness, — with a good deal of the sincerest kind of mediaeval asceticism, — but of artistic genius there are no traces in any of the three poems. Now regularity in rhythm is just what the poorest artist is likely to attain. He cannot make his lines smooth or sweet or majestic, but he can and generally does make his accents follow one another with mechanical regularity. These writers, it seems, were able to do this, and not much more. Where inversions do occur in their verse, they were probably introduced less for the sake of giving grace or variety to the rhythm, than because the author lacked the patience or the skill to avoid them. His ideal was more mechanical than his model, and he pursued the former except where he was obliged to resort to such licenses as the latter plainly allowed.(i) (1) I have in the last few pages selected for detailed discussion only the one difference between Latin and O.F. octosyllabics which seemed to me most likely to mislead. There are two other striking differences of detail, namely the comparative frequency of inversions in the 3rd foot in French, and the use of mere assonance in the latter language after rime had been well developed in Latin. Both these phenomena can, I think, be easily explained by comparison of the exigencies of the two languages, but it did not seem worth while to swell these pages with such matters. The fact that the French verse was borrowed from the Latin seems so - 83 - § 45- Explanation of the change in French verse. This brings us to the second branch of the assertion quoted from Stengel.(i) He believes that after the time of these early poems, Frenchmen learned, by imitation of the Latin rhythms, to avoid the regular iambic move- ment, and to write the almost purely syllabic verse that is so familiar now. This involves a curious misconception as to the way in which the Latin influence worked. The poems in which the modern syllabic system begins to appear are for the most part romances and songs of chivalry. (This is strictly true of the 8 -syllable line, although the Alexis, in decasyllabics, while in the main unaccentual, is anything but chivalric in feeling.) Now it is impossible to believe that the monks of the ioth century, whose business it was to study and expound Latin, should persistently adhere to a system inherited from poets of France now forgotten, — a system peculiar to the vernacular; and that the trouveres of the nth and 1 2 th centuries should be the ones to adopt the system of the Latin sacred song! We have seen, how- ever, that the Latin rhythms were not, in general, purely syllabic; and it will be easy to find an explanation of the change in French verse that will be unobjectionable. The explanation is found in the laws of French accent. Every French word is accented, (so far as its accent is fixed at all), either on the last syllable, or, in the case of feminine endings, on the penult. In the latter case the final syllable is so insignificant that the word is almost always virtually oxytonic : — and this oxy- tonic character is absolute in the prose of ordinary speech, and becomes absolute in poetry wherever elision occurs. If, therefore, the French write strictly accentual rhythms, their verses will of necessity show a general patent, that I fear too much space has already been given in the text to unnecessary demonstration. (I) § 43, ante. 6* - 84 - identity of word-foot with verse-foot; and it is well understood that such identity is far from pleasing in its effect. It was avoided in quantitative verse by both the Greeks and the Romans, and it is avoided in accentual verse by modern poets generally. This fact has led to the great preference that the English have shown for iambic and anapaestic metres, as compared with trochaic and dactylic; — for oxytonic words in English are in a small minority. In Spanish on the other hand, it is said that trochaic rhythms are generally the favorites, and in Spanish we find a comparatively large number of oxy- tonic words. The French poet who attempted to write iambic verse, quickly (though doubtless unconsciously) dis- covered this fact. His verses acquired that choppy cha- racter which is so annoying, for example, in Hiawatha\ though the illustration is by no means adequate, for the tendency to throw accents back in English is far less marked than the tendency to throw them forward in French. Moreover at an early date (though perhaps we cannot say how early) the secondary accent was lost to the French language. The poet could not, therefore, make two proper metrical stresses rest upon a single word. Polysyllabic words were a source of serious em- barrassment to him. Trisyllables, for example, with masculine endings, constitute anapaests, and they would have to be ruled out of the poet's vocabulary altogether, unless the first syllable could be worked in as the last part of an inverted foot. Longer words would in general be quite impracticable: — they would sink the rhythm beyond recognition. Thus accentual verse became an artistic impossi- bility.^) As quantitative verse was equally out of the (i) It has been attempted in modern French, but only experimentally, it seems, or for use with music. Lubarsch, p. 198, et seq. - 85 - question, it is plain that the system actually adopted, whether we regard it as satisfactory or not, was the only one available. To the modern French language it is peculiarly suited, on account of the generally equal value of syllables in the modern speech, and the lightness of the French tonic stress. How far these characteristics had been developed in the French of the middle ages, the writer has been unable to determine ;(i) but it seems not unlikely that they were then present to some extent, and materially aided the introduction of the new system of verse. Stengel's theory as to the manner of its intro- duction is probably correct, if we disregard his explanation of the cause. " Gewohnte man sich (not, however, "in Anlehnung an den gleichsilbigen Vers der rhythmisch- lateinischen Verskunst ") fruh daran, statt an vierter ofters an dritter Stelle einen Wortton zuzulassen ... so wurde . . . der jambische Rhythmus verdunkelt, und da- mit erschien auch jede weitere Markierung desselben im Innern des Verses iiberflussig. " The irregular lines in The Passion and the Life of St. Ltger look as if they were introduced merely for convenience; but those in the Alexander may well have been consciously written as tentative efforts at the new system: the latter seems to be the case in Gormund et Isembard, and after that the sway of the new system was undisputed except where foreign influence came to disturb it. It is possible, however, that the idea of writing syl- labic verse instead of accentual was suggested by a mis- pronunciation of Latin. Latin is now commonly pro- nounced by Frenchmen according to the laws of French accent, — i. e. with the accent thrown forward in all (i) I should conjecture that the equal measurement of syllables was to some extent characteristic of the tongue from the first, because of the way in which the syllables before the accent were preserved, in the passage from low Latin to French, while those after the accent melted away. That, on the other hand, the Old French accent was somewhat stronger — 86 — words.(i) In the middle ages the usage in this respect seems to have been variable.(2) Thus the following lines seem to accent all the proper names on the ulti- mate : — Et Troillus et Eneas, roi Menon et Pollidamas, rois Sarpedon et rois Glaucus et de Lancoine Eufrenus, . . . rois Terepex, rois Adrastus rois Episirox, roi Alcanus, etc. Beneoit, Roman de Troies 9717, et seq. while in the Alexander we have apparently the true Latin accents (including the secondary) in Est vanitatum vanitas et untversa vanitas. In the Passion we have, on the one hand Post que deus filz suspensus fure, and, on the other hand, Jhesum querem Nazarenum. Now it is easy to see that a little indulgence of this ten- dency to throw the accent forward would convert the accentual verse of an Ambrosian hymn into a measure purely syllabic, and the supposition that some such process gave the French poet his first suggestion is at least not violent. To the present writer, however, it does not seem necessary to go abroad for an explanation. Here, as else- where, it is surely not unreasonable to repose some con- fidence in native genius. We have seen the necessity and propriety of the system: — we may well believe that the poets found it out for themselves. than the modern, seems probable from the accentual habit of speech observed among the peasant classes, and from stage tradition. (1) G. Paris, Accent Latin, p. 22, 23. (2) M. Paris thinks we can fix the time at which the correct pro- nunciation of Latin passed out of use, but I am unable to follow his reasoning. - 8 7 - § 46. French decasyllabics. Mention has already been made of the great number of derivations proposed for the 1 o-sy liable verse in French. The excellent precept of M. Paris, however, which has been quoted, directs us to the most straightforward explanation. This verse is probably not an exotic at all, but a natural development from the principles established in the 8-syllable line. There are three difficulties in the way of this common-sense derivation which will be briefly considered: — first, the fact that in the earliest known decasyllabics the syllabic system appears firmly estab- lished, — second, that there is a fixed caesura, usually after the fourth syllable, — and third, that there is often, in the earliest specimens, an extra-metrical syllable at the caesura, or in other words, that the first hemistich often has a feminine ending. The first fact occasions no surprise in view of the date of the appearance of this verse. It is first found in French in the Life of St Alexis ,— %. poem more recent than the Alexander or perhaps even Gormund et Isem- bard, but older than any of the other romances. It would not have been strange, indeed, if the new system had appeared in decasyllabics long before it was estab- lished in the 8-syllable verse;— for the more syllables there were in a line, the more obvious were the dis- advantages of the accentual system; but in fact the Alexis was written just as the new system was introduced in octosyllabics. The second difficulty is no more serious. Given the syllabic system, long lines without caesura were im- practicable^ 1) A little consideration of the possible varieties of such lines will make this clear. If we count (1) They have been written, of course, by the poets of the Romantic School in the 19th century. only those lines which could be made by combinations of iambic and anapaestic feet, we find that the following varieties might occur: ww — vu — w — w WW W — WW w ww — w — w — ww — W WW WW — w — W WW W WW w — w — ww — ww — But the introduction of a caesura after the 4th syllable necessarily eliminates all except the 6th of these varieties, because that is the only one that permits a tonic syllable in the 4th place. The above, now, are only the com- binations of two particular kinds of feet. It is easy to see that, with the variety of other possible combinations, the condition of decasyllabics without caesura would have been chaotic: the fixed pause was necessary to reduce the line even to such regularity as is desirable in French verse. 10 syllables arranged helter-skelter could not be counted by the ear, and the essential prin- ciple of the verse would therefore be obscured; — but 10 syllables arranged in groups of 4 and 6 are comparatively easy to grasp. Substantially the same explanation may be given, perhaps more clearly, in another form. The French poet, we will say, was sensitive enough to feel that long verses were objectionable, in that the number of their syllables could not readily be grasped; — and therefore he did not compose long verses at all: but he found a beauty in the combination of short verses of unequal lengths. The French decasyllabic line can with entire propriety be regarded as made up of two verses, one of 4 and one of 6 syllables. The reason why it is not so regarded is that rime is not found at the caesura, and rime is commonly insisted upon as the indispensable mark of the verse-end in French. It is of course true that French poetry in general needs rime to mark off C \ B R A OF" TBF UNIVERSITY 8 9 — X^C^rh^ one verse from another : — syllabic regularity alone is not enough to distinguish verse from prose : French blank verse is not verse at all. But it is clear that the longer the verse is, the more necessary does rime become, and conversely, alternate lines of 4 syllables might well be left unrimed. There is therefore no logical im- propriety in regarding the 10 syllable long line as made up of two verses. Whether it should now be so re- garded, is a purely academical question, but that it came into being as such seems extremely probable. The early French version of the Song of Songs tends strongly to confirm this opinion, for it is made up of rimed 10 syl- lable lines intermixed with unrimed lines of 4 syllables. The first 6 lines illustrate its verse: — Quant li solleiz conviset en leon en icel tens qu'est ortus pliadon perunt matin, une pulcellet odit molt gent plorer et son ami dolcement regreter, e si Hi dis. The effect of this is obviously that of a series of short verses, those of 4 syllables unrimed, those of 6 rimed. It differs from the ordinary decasyllabic only in this respect, that here the number of short verses exceeds that of the long, instead of being exactly equal. A regular alternation of unrimed 4-syllable lines with rimed 6-syllable lines, would differ from decasyllabics only in the manner of writing; and in a manuscript in which poetry is written as prose even this difference would disappear. The same consideration explains the third difficulty, namely the occasional presence of an extra syllable at the end of the first hemistich ; for if that hemistich was understood as a separate verse, it might, of course, have either a masculine or a feminine ending. Masculine endings were the rule in the first three specimens of — oo — French poetry(i) — as was indeed natural, since they were copied from a Latin verse in which the ending was uni- formly masculine. The feminine verse-end first appears, as a regular form, in the Alexis itself; and that is there- fore exactly the poem in which we should expect to find the extra syllable at the caesura. The new ending being once devised, it was naturally applied to the short and the long hemistichs alike.(2) (i) Isolated examples of the feminine ending are found, as in 11. 127, 128 of the Passion. (2) The argument is strengthened by the appearence in French of two other forms of the decasyllabic line — those, namely, with caesura after the 5 th or after the 6th syllable. They are simply other combinations of shorter metres. The efforts to find a separate classical original for each of the three are painfully superfluous. The only serious difficulty that confronts our theory is in the comparison of the decasyllabics of other romance languages. The Italian endecasillabo, however, is probably not a proper subject of comparison. It has no fixed caesura, and while it has, in a sense, two fixed accents, one of these may fall either on the 4th or on the 6th syllable. While therefore it looks like a combination of two forms of the French line, and as such perhaps threatens our derivation of the latter, it is simpler to regard it as not a composite line at all, but a natural extension of the octosyllabic form. The Provencal poem Boethius is peculiar. Here the 1st hemistich of the decasyllable is frequently feminine, but the verse-end proper is uniformly masculine ; and the structure of the verse is syllabic rather than accentual. Bartsch {Gesch. d. Prov. Lit. § 8) seems to think this poem was written in the 1st half of the 10th century: Stimming (in Grober's Grundriss, II. 2, 44) simply says it dates from the 2nd half of the 10th century, or according to others from the beginning of the nth. If the earliest of these dates is correct, then the poem presents an entirely isolated phenomenon, which I am at present unable to explain by my own theory, or by any other: but if the latest date is to be trusted, then the poem, while singular, need not puzzle us. I am the more content, for the present, to leave this branch of the problem partly unsolved, because the whole subject of French decasyllabic verse is only collaterally involved in our inquiry (as will appear in the next chapter). CHAPTER VI. Latin and French Influence in English Verse. § 47. Old English verse. Old English versification was dependent chiefly upon accent, but also in part upon, quantity . The typical verse was composed of two he- mistichs. The normal hemistich may perhaps most logi- cally be analyzed as composed of four measures. Two of these measures were accented, and consisted in general of single long syllables; but sometimes an accented measure comprised (by a species of resolution) two short syllables, of which the first received the actual stress; and in a special class of cases an accented measure might even consist of a single short syllable. The other two measures in each hemistich were unaccented (either actually or comparatively) and each consisted, generally speaking, of an indeterminate number of syllables, in- differently short or long. The regular accents of the hemistich might fall upon any two of the four measures, except only that one, at least, of the accents must fall upon one of the first two measures. Alliteration bound together the two hemistichs of each verse.(i) Such was the structure of the simplest verses, — the five types announced by Sievers; but complications were common. "Without considering the latter, however, we necessarily feel that the rhythm of the verse was hardly (1) This description is taken from Sievers, p. 23 et seg., with changes only in the form of statement. ? — Q2 — (a rhythm at all, in any modern sense. HgwexeiLs tudiously we accustom our ears to it, there seems still to be a barbaric crudeness inherent in the verse: — and yet the investigations of Sievers have revealed complex laws such as could have been evolved only in a refined form of art. It is evident therefore that while we understand the laws of composition in minute detail, we have not yet perfectly realized the actual manner and effect of the customary recitation of the lines.(i) Perhaps they were delivered in a musical or quasi-musical manner, and in this recitative the time may have been quite as salient a feature as the stress; so that occasional pauses might effect an approximate equality between the measures: — but the details are yet to be established. Certainly the verse-form was well grounded in the character of the language : it changed as the language changed, and some such mode of recitation as we have suggested can pro- bably be proved by examination of the curious phases of decay through which the verse passed, — betraying hopeless efforts to shape the new materials into the old form. That inquiry, however, is not within our present province. § 48. The decay of Old English Verse. This verse fell into decay in two ways.(2) The more conservative poets of the early middle ages probably reproduced the effect of it as well as they knew how. Their verses and those of their followers reveal a loss of the more refined feeling for quantity, a loss of the peculiar Old English distinction of half-accents, and often a tendency to fill up the unstressed measures with multitudes of syllables ; but the real continuity of the verse-form is apparent even down to our own time. We trace it in its decay (1) The necessity of further study of this subject is made clear by Heath, Trans. Phil. Soc. 1891 — 93, p. 375. (2) Schipper, p. 76 and passim. ^3-JLJ^ — 93 — to the doggerel verse found so frequently in the mo- rality plays, and even in Shakspere's early comedies; rarer examples are to be found in the 17th and 18th centuries; and the 19th has resuscitated and beautifully refined such of its essential principles as had any life in then2 I __rn_thp fami%r_C hristabel me tre. A large part of our modern verse has thus descended directly from the ^-*^-*-a-<*j( Old English long line. It counts not the number of