SB E53 
 
 
w 
 
 Division 
 
 Range 
 
 Shelf. 
 
QUANTITATIVE PRONUNCIATION 
 OF LATIN. 
 
PRACTICAL HINTS ON THE 
 
 QUANTITATIVE PRONUNCIATION 
 OF LATIN. 
 
 FOR THE USE OF CLASSICAL TEACHERS AND LINGUISTS. 
 
 ALEXANDER J. ELLIS, B.A., F.R.S., F.S.A., F.C.P.S., F.C.P. 
 
 Past President of the Philological Society, 
 
 Formerly Scholar of Trinity College, Cambridge, 
 
 Author of "Early English Pronunciation." 
 
 17 ^kv yap 7reV} Ae|ts ovfievbs ovr 3 6v6jjLaTos, ovre pharos, 
 jSm^ercu rovs xpfoovs, oi5e fjLTarld7]O'tv' dAA' o'las irapi\<ri<j) 
 rf) (f>vffi rots ffv\\afias ras re /xa/epcta ftal rds j8pa%eias, 
 roiavras ^fAarrei. 
 
 DION. HAL. First Century, B.C. (See Art. 41 
 
 LIBRA It v 
 
 - j UNIVERSITY OJ 
 
 ACMILLAN AND, C.C- 
 
 ,8,4. \UALIFORNU. 
 
 [ The Right of Translation and Reproduction is Reserved. ] 
 
LONDON : 
 
 JR. CLAY, SONS, AND TAYLOR, PRINTERS, 
 BREAD STREET HILL. 
 
j LI BU A Li Y 
 
 UNIVERSITY OF 
 
 ((.CALIFORNIA. 
 PRELIMINARY NOTICE. 
 
 PROF. ROBINSON ELLIS, of University College, London, 
 at the close of the Report on Latin Philology which he 
 was kind enough to contribute to my farewell presi- 
 dential address to the Philological Society (delivered on 
 May 15, 1874), introduced the subject of Latin Pro- 
 nunciation as follows : 
 
 " The question of Pronunciation forms one of the sub- 
 jects discussed at the Schoolmasters' Conference this 
 year. The meeting which was held early in 1871 at 
 Oxford, to take steps for the reformed pronunciation of 
 Latin, and the conference of schoolmasters held about 
 the same time, resulted in the combined Syllabus of 
 Latin Pronunciation drawn up by Professor Munro of 
 Cambridge and Professor Palmer of Oxford. This was 
 at once introduced into several of the larger schools in 
 England, at least in the higher forms. I myself adopted 
 it for the use of my classes in University College, and a 
 very similar scheme of pronunciation was only last year 
 printed by Professor Key for University College School. 
 Independently of this, a reformed pronunciation has 
 
vi PRELIMINAR Y NOTICE. 
 
 been adopted in various educational establishments in 
 this country ; and it is no uncommon occurrence in my 
 classes to find students on their arrival already trained in 
 the new method, with such slight differences (and they 
 are really slight) as the divergence of opinion on par- 
 ticular points makes unavoidable. It would be prema- 
 ture at present to express any opinion as to the eventual 
 success of this experiment ; it can hardly be said as yet 
 to have been adequately tried in schools, or properly 
 seconded in the Universities. At Oxford, when I exam- 
 ined viva voce as Classical Moderator in 1872, I was the 
 only examiner who used the reformed pronunciation, and 
 those who then came before me for examination did not 
 generally seem familiar with it. Even now the old use 
 predominates, and it is to be feared that even those 
 trained by the Syllabus at school, e.g., at Shrewsbury, 
 Marlborough, Liverpool College, Christ's Hospital, Dul- 
 wich College and the City of London School, are in- 
 duced to give it up, or, at least to suppress it, when they 
 proceed to the Universities. On one point there seems 
 to be a very general agreement : wherever it has been 
 introduced, it has been adopted without difficulty by 
 students of all ages, even by the youngest boys, from ten 
 years old upwards. It is obvious that if it is to be really 
 successful, it should be taught alone ; at present the old 
 pronunciation is allowed to linger on side by side with 
 the new. This would not happen if boys were trained 
 at the outset on the new system, and if it were an under- 
 stood thing that no other pronunciation was permitted. 
 But the lamentable fluctuation of opinion exhibited by 
 the schoolmasters in their conference of this year is a 
 
PRELIMINAR Y NO TICE. vii 
 
 clear proof, if any proof were wanted, of the difficulties 
 which invariably attend any real reform ; and it becomes 
 doubly incumbent on institutions, which, like University 
 College, and Owens College, Manchester, represent 
 enlightened opinion irrespective of denomination, and 
 which for that reason necessarily work independently, 
 to devote their best energies to the successful carrying 
 out of this apparently small, but in my judgment really 
 important, detail." Transactions of the Philological 
 Society, 18734, pp. 398 9. 
 
 These remarks induced me to add some observations 
 (ibid, pp. 399 407) in which I dwelt on the practical 
 difficulties of making the change, and stated candidly 
 that though I utterly abhorred the old English habit of 
 Latin pronunciation I had been taught to use at Shrews- 
 bury and Eton (1826 1833), yet that if schoolmasters 
 sought success only by written examinations, it was 
 hardly worth their while to make the change. And, I 
 remarked, there was no use in disguising the fact that 
 the change to be effective must be troublesome. There 
 were really no difficulties for English speakers, I ex- 
 plained, in adopting the new sounds proposed for vowels 
 and consonants. " The real trouble of the new pronun- 
 ciation begins," I emphatically stated, "just where no 
 trouble is suspected in accent and quantity." I then 
 shewed that though, as an Eton boy, I had been taught 
 to feel a holy horror for " false quantities," yet also as 
 an Eton boy, I had been perpetually making false 
 
viii PRELIMINAR Y NOTICE. 
 
 quantities in common with all the Eton masters them- 
 selves. In sic vos non vobis mdificutis aves, it was usual, 
 in my day, to pronounce sic like English sick, non to 
 rhyme with on, vobis with the last syllable like the first in 
 biscuit, mdificatis with the syllables as far as the a like 
 English nid" ifica' tion, and aves with av as in aviary, so 
 that five false quantities were made in one short line. 
 Of course opus op'eris, so'lus sol'itudo, and the like, fur- 
 nished thousands of others. Since the place of the 
 Latin accent is dependent on the quantity of the last 
 syllable but one, in words of more than two syllables, if 
 the accent were placed right there, the speaker was held 
 to have made no " false quantities ; " and if in his verses 
 he followed the laws in his " gradus " which were at utter 
 variance with the custom of his speech, he was also held 
 to have made no "false quantities." That he did not 
 pronounce a single vowel correctly by intention, that he 
 did not understand the nature of long and short vowels 
 or syllables, or the rhythm that they made in verse 
 (except as by " gradus " aforesaid), that he had no con- 
 ception of what the nature of Latin accent was, and 
 that Latin as he uttered it (not as he saw it) was pure 
 vox et praeterea nihil, sound without any sense at all to a 
 Roman's ears, of this he had no conception whatever, 
 though in his ignorance he did not hesitate to laugh at 
 a Frenchman's or German's English, which, however 
 poor, would be at any rate properly intended, and at 
 least intelligible. 
 
PRELIM1NAR Y NO TICE. ix 
 
 Having, in view of opening this subject in the address 
 already cited, made an arrangement with the College of 
 Preceptors, to give them a paper on Latin Pronunciation 
 at their monthly meeting in June, I purposely went into 
 this particular question of Latin Accent and Quantity 
 (and especially the latter as determining the former), in a 
 practical paedagogical paper. The chair was occupied by 
 the Rev. G. C. Bell, Head Master of Christ's Hospital, 
 the largest school in England which, as previously men- 
 tioned by Prof. Robinson Ellis, has adopted the new 
 sounds of the letters, and I had an audience of classical 
 teachers, who during an address of unexampled length 
 (nearly two hours and a half) listened with that attention 
 which only great practical interest in the subject could 
 command. 
 
 In revising this paper for separate publication, I have 
 found it necessary to expand many parts, and in especial 
 to add and interpret many passages from Cicero and 
 Quintilian (the only authorities of any value), and to put 
 in a much more complete form the arguments which have 
 induced me to treat the (so-called) elided vowels and 
 final M in the manner here advocated. But not to con- 
 fuse the order of the exposition, I have relegated much 
 of this accessory matter to footnotes. Finally I have en- 
 deavoured, as well as it was possible on paper, to convey 
 a conception of the mode in which I read the examples, 
 which form an Appendix at the close of this tract. I 
 wish it had been possible for me to freeze up my 
 
x PREL1MINAR Y NOTICE. 
 
 utterances into some Munchausen's postboy's horn, so 
 that my readers might have only had to hang it up in 
 the ingle] and hear the very sounds which I have, I fear 
 often vainly, tried to convey on paper. But I have given 
 such ample practical directions for self-practice that I 
 hope classical teachers and classical scholars, and lin- 
 guists in general (for whom, and not for pupils, my book, 
 and especially its notes, have been put together), who 
 feel interest enough in the subject to undertake the 
 labour, will be able to make Latin live again in them- 
 selves, and breathe its magic through their own lips into 
 the souls of their hearers. 
 
 A. J. E. 
 
 June 27, 1874. 
 25 ARGYLL ROAD, KENSINGTON, W. 
 
CONTENTS. 
 
 Preliminary Notice, pp. v. x. 
 
 I. Preliminary Assumptions, Art. I 14, pp. I 1 1. 
 
 Art. I, 2. Introductory, p. I. 
 3> 4- Worthlessness of current pronunciation, p. 2. 
 
 ?j 5_S. Pre- Augustan, Augustan, post- Augustan periods ; Augus- 
 tan pronunciation selected, pp. 2 4. 
 
 ,, 9, 10 Pronunciation of Letters assumed, pp. 4 8. 
 ,, II. Latin accent consisted in change of pitch, p. 8. 
 ,, 12, 13. Latin quantity, the foundation of rhythm, pp. 9, 10. 
 
 ,, 14. Latin had no Force-accent. Note on Length, Force, 
 Pitch, and Ictus, pp. 10, n. 
 
 II. Elementary Exercises on Quantity, Art. 15 21, pp. II 15. 
 
 Art. 15-17. Division of time, pp. n, 12. 
 ,, 1 8. Syllabication, p. 13. 
 
 ,, 19. Pitch and Force as independent of Length, p. 13. 
 ,, 20, 21. "Position," pp. 14, 15. 
 
xii CONTENTS. 
 
 III. The Artificial Rhythmical Unit of Latin Speech. 
 Art. 2230, pp. 1523. 
 
 Art. 22, 23. Long vowels in long Syllables, pp. 15 17. 
 
 ,, 23. Doubtful vowels in long syllables, p. 17. 
 
 ,, 24-26. Doubtful syllables, pp. 17, 18. 
 
 ,, 27. Short syllables, pp. 18, 19. 
 
 ,, 28. Artificial character of the assumption that one long syl- 
 lable is equal to two short syllables, pp. 20, 21. 
 
 2 9> 3' Mode of counting and marking long and short syl- 
 lables in this tract, pp. 21 23. 
 
 IV. Metrical Feet in Latin Words, Art. 31 39, pp. 23 26. 
 
 Art. 31. Simple and Compound Feet, pp. 23. 
 
 ,, 32. Exercises on Pyrrhics, p. 23. 
 
 ,, 33. On Iambs, pp. 23, 24. 
 
 34- O n Trochees, p. 24. 
 
 ,, 35. On Spondees, p. 24. 
 
 ,, 36. On Dactyles, pp. 24, 25. 
 
 ,, 37. On Anapests, p. 25. 
 
 ,, 38. On Molossi and Choriambs, p. 25. 
 
 ,, 39. Examples of all kinds of feet, p. 26. 
 
 V. Elementary Notions of Verse Rhythm with both Accent and 
 Quantity Hexameters, Arts. 40 52, pp. 26 35. 
 
 Art. 40. Temporary Elisions, pp. 26, 27. 
 
 ,, 41. Marking change of Pitch, and extract from Dionysius of 
 Halicarnassus on Pitch Accent, pp. 2731. 
 
 ,, 42-48. Effects in Hexameters, pp. 32, 33. 
 
 ,, 49. Effect of Caesura, p. 33. 
 
 ,, 50-52. Final Cadences, pp. 34, 35. 
 
CONTENTS. 
 
 VI. Slurred Vowels, Arts. 5363, pp. 3543. 
 
 Art. 53. Open vowels, coalescent, and gaping, pp. 35 37. 
 ,, 54. Spanish and Italian analogies, pp. 37, 38. 
 55> 5^- Examples of slurred vowels in Italian singing, pp. 
 
 38, 39- 
 
 ,, 57. Unslurred vowels in medieval Latin, p. 39. 
 ,, 58. Practical Rules for Slurring Vowels, pp. 39, 40. 
 
 *> 59} 60. Accentual separation of words connected by Slurred 
 Vowels, pp. 40, 41. 
 
 ,, 61. Supplement to the Practical Rules, pp. 41, 42. 
 ,, 62. Slurring in Prose, pp. 42, 43. 
 ,, 63. Mode of Practice, p. 43. 
 
 VII. Treatment of Final M, Art. 64 98, pp. 43 73. 
 
 Art. 64. The Two Facts to be accounted for, p. 43. 
 ,, 65. Double Latin pronunciation of M, p. 44. 
 
 ,, 66, 67. Analogues to the custom of retaining one spelling 
 with variable pronunciation according to environ- 
 ment, in French and Dutch, pp. 44 46. 
 
 ,, 68. The neglect of M between two vowels was not a pho- 
 netic necessity to Latins, p. 46. 
 
 ,, 69. Disappearance of final -en in Southern German and 
 English, p. 46. 
 
 ,, 70. Quintilian's "lowing" final M, p. 47. 
 ,, 71-74' Treatment of final M in inscriptions, pp. 47 50. 
 ,, 75, 76. Results as to the sounds of final M, pp. 50, 51. 
 ,, 77. Modern analogues and contemporary information, pp. 
 51, 52. 
 
xiv CONTENTS. 
 
 Art. 78. Almost invariable omission of final M in Spanish and 
 Italian, p. 52. 
 
 ,, 79, 80. Italian omission of Latin final T, D, and treatment 
 of the open vowel thus created, pp. 52, 53. 
 
 ,, 81. Indication of omission of Latin final T, D, &c. in Italian, 
 by doubling the following consonant in writing, pp. 
 53, 54- 
 
 ,, 82. Indication of omission of Latin final T, D, &c. in Italian, 
 by "energising" or doubling the following consonant 
 in speech when this is not done in writing, pp. 54, 55. 
 
 ,, 83. Probability that the treatment of final M in Latin re- 
 sembled that of final T, D in Italian, p. 55. 
 
 ,, 84-86. Contemporary evidence of assimilation, Cicero's cun 
 nobis and Quintilian's cun notls, and the latter's pho- 
 netic theory for the assimilation, pp. 55 58. 
 
 ,, 87. The half M of Verrius Flaccus, pp. 58, 59. 
 
 ,, 88. Critical Examination of Quintilian's account of final M 
 before a vowel, pp. 59 63. 
 
 ,, 89. Result, final M omitted before a vowel in Augustan 
 Latin, p. 63. 
 
 ,, 90. Monosyllables in M no exceptions, pp. 63 65. 
 
 ,, 91. General Rules for Treatment of final M, pp. 65 67. 
 
 ,. 92-95. Mode of Practice, pp. 6770. 
 
 ,, 96. Necessity of the Teacher " setting patterns," p. 70. 
 
 ,, 97. Writing exercises on final M, pp. 70 72. 
 
 ,, 98. How is the presence of final M to be indicated in saying 
 declensions and conjugations ? pp. 72, 73. 
 
 VIII. Elegiac and Lyric Verse Rhythm, Arts. 99106, 
 PP. 73-78. 
 
 Art. 99. First practice of new metres, p. 73. 
 ,, 100. Choriambics, pp. 73, 74. 
 ,, 1 01. Pentameters, p. 74. 
 
CONTENTS. xv 
 
 Art. 102. Sapphics, pp. 74, 75. 
 ,, 103. Alcaics, p. 75. 
 ,, 104-106. Iambics, pp. 75 78. 
 
 IX. Prose Rhythm, Arts. 107 112, pp. 79 85. 
 
 Arts. 107-109. Cicero and Quintilian on Prose Rhythm, pp. 
 79-81. 
 
 ,, no. Duty of Teacher to set a pattern of reading prose, 
 pp. 81, 82. 
 
 ,, in. Mode of treating beginners, pp. 2, 83. 
 
 ,, 112. How to deal with the English Force Accent. Note on 
 Third Century Latin with Force Accents, pp. 83 85. 
 
 X. How to Read Late Latin, Art. 113, 114, pp. 85, 86. 
 
 XI. Final Method of Reading Latin, Art. 115119, p. 8689. 
 
 Art. 115-118. Precautions and Practice, pp. 86 88.. 
 ,, 119. How to utilise "repetitions," pp. 88. 89. 
 
 XII. How to Read the Poetical Examples in the Appendix, 
 
 Arts. 120 131, pp. 90 108. 
 
 XIII. How to Read the Prose Examples in the Appendix, 
 
 Arts. 132 138, pp. 108 112. 
 Art. 139. The Italian Example, p. 113. 
 
 XIV. Conclusion, Art. 140, pp. 113 115. 
 
CONTENTS. - 
 
 Appendix of Quantitative Examples, pp. 117 129. 
 
 Note, p; 117. J * '> .J > '-, 
 
 A, B, C. Virgil's Aeneid, p. 118, 119. 
 
 D. Horace's Satires, p. 120. 
 
 E, F. Horace's Odes, p. 120, 121. 
 
 G. Propertius's Elegies, p. 121, 122. 
 
 H. Ovid's Heroides, p. 122. 
 
 I, K, L. Horace's Odes and Epodes, p. 122, 123. 
 
 M, N, O, P, Q. Cicero's OOrator, p. 124127. 
 
 R. Cicero's De OOratore, p. 128. 
 
 S. Tasso's Gerusalemme Liber ata, p. 129. 
 
 Index of Authors Cited, pp. 130 132. 
 
 ERRATA. 
 
 p. 8, /. 4 front bottom of note, read: slackening strings 
 
 p. 27, /. 14 from botiom of note, readl nova-tut] 
 
 P. 46, /. i read: af*xr a nd 
 
 /. 46, /. 10 and n, read : svave svdvent 
 
 P- 5> I- S/ rom Bottom, read: those considerations 
 
 P. 51, /. 5 from top, read: that m 
 
 p. 66, /. i, read : inaudible 
 
QUANTITATIVE PRONUNCIATION 
 OF LATIN. 
 
 I. Preliminary Assumptions. 
 
 ART. i. The object of this book is entirely practical. 
 It is therefore necessary to make several assumptions, 
 requiring for their justification much detailed consi- 
 deration for which reference must in general be made 
 to other treatises. 1 
 
 1 Not only the immense and, for scholars, indispensable treatise : Ueber 
 Aussprache, Vokalismus und Betonung der Lateinischen Sprache von W. 
 Corssen, 2te Auflage, Leipzig, 1863, Vol. I. pp. xvi. 819; Vol. II. pp. 1086, but 
 the following more accessible English treatises : 
 
 Syllabus of Latin Pronunciation drawn up at the request of the Head Masters 
 of Schools. Cambridge and Oxford, 1872. 8vo. pp. 7. A few remarks on the 
 Pronunciation of Latin, with a postscript by H. A. J. Munro. Cambridge, 
 1871. 8vo. pp. 36. A Grammar of the Latin Language, from Plautus to Sue- 
 tonius, by Henry John Roby. Part I. ist edition, 1871, sm. 8vo. pp. xcv., 476, 
 of which at least 150 are devoted to pronunciation. 2nd edition, 1872, in which 
 these remarks are enlarged, with replies to Prof. Munro's pamphlet just cited, 
 etc. The Public School Latin Grammar, for the Use of Schools, Colleges, and 
 Private Students, by Benjamin Hall Kennedy, D.D. 2nd edition, 1874, pp. 
 xxix., 599, of which at least 68 pages are devoted to phonology and its applica- 
 tions. Elements of Latin Pronunciation for the Use of Students in Language, 
 Law, Medicine, Zoology, Botany, and the Sciences generally in which Latin 
 words are used, by S. S. Haldeman. Philadelphia, 1851. 8vo. pp. 76. Latin 
 Pronunciation, an inquiry into the proper sounds of the Latin language during 
 the classical period, by Walter Blair, Professor of Latin in Hampden Sidney 
 College, Virginia. New York and Chicago, 1873. Svo. pp. 136, an extremely 
 useful Iktle work reviewed in the Southern Magazine, October, 1873, by Pro- 
 fessor S S Haldeman, who refers also to J. F. Richardson's Roman Orthoepy, 
 
2 PRELIMINARY ASSUMPTIONS. [Art. 2 5. 
 
 ART. 2. First, I assume that sufficient interest has 
 been recently excited both theoretically and practically in 
 the question of Latin pronunciation, to make teachers 
 and scholars willing to take a good deal of personal 
 trouble to gain some insight into its nature as a living 
 reality and not merely as a paper fiction. 
 
 ART. 3. I assume that the utmost value which can be 
 attributed to the current English pronunciation of Latin 
 is, that it serves rather roughly to recall to Englishmen, and 
 no one else, the letters with which the words are written 
 in ordinary printed books. 
 
 ART. 4. I assume that this same current English 
 pronunciation is positively injurious even to Englishmen, 
 who wish to understand the nature of Latin linguistically, 
 its flexional and historical relations either as descendant 
 or ancestor, and its rhythmical structure either oratorical 
 or poetical. 
 
 ART. 5. I assume that by Latin pronunciation we 
 mean that current among the principal men of eminence 
 as statesmen, philosophers, historians, writers, orators, 
 and poets during the first century before Christ, the pro- 
 nunciation of Julius and Augustus Caesar, of Maecenas, 
 of Cicero, Virgil, and Horace, that is, the court and 
 literary as distinct from the popular and rustic pro- 
 nunciation. 1 
 
 a Plea for the restoration of the True System of Latin Pronunciation, New 
 York, 1859, PP- 1J 4 ' Dr. L. Tafel and Prof. R. L. Tafel's Latin Pronuncia- 
 tion and the Latin Alphabet, 1860, pp. 172, based on Corssen ; and G. K. 
 Bartholomew's Grammar of the Latin Language, Cincinnati, 1873, who adheres 
 closely to the ancient grammarians. 
 
 1 Compare Cicero's multitude in (N, 2 4), that is lines 2 to 4 of example N in 
 the appendix to this tract. Ultra-Roman pronunciation of course is not to be 
 regarded here, although of considerable philological interest. The names of 
 Roman writers which are anglicised, as Virgil, Horace, Ovid, receive their 
 English sounds, and in accordance with them, we must when speaking English, 
 
Art. 6.] PRELIMINARY ASSUMPTIONS. 3 
 
 ART.. 6. I assume that this Augustan pronunciation, 
 as it may be briefly termed, differed at least as much 
 from that of the preceding century (or pre-Augustan) as 
 the English pronunciation of Queen Anne did from that 
 of Queen Elizabeth ; and that it differed from that of 
 the second and third centuries afterwards (post-Augustan 
 and transitional) at least as much as, probably much 
 more than, Queen Anne's from Queen Victoria's. 1 
 
 talk of Cicero as Sis'ser-oh, and C&sar as Seize-ker, &c. just as we necessarily use 
 Rome, Naples, Venice, and Florence as English words, and even without 
 changing the spelling, give completely English sounds to Calais, Paris, Ver- 
 
 'Cgard LU -L/ami names itiiu J..AUIII wurus citiu pmubcs miiuuuucu IULU j-jngu^n. 
 
 sentences, as we now adopt for French. But just as we should not venture 
 to introduce the English Paris, &c. into a French sentence, so should we never 
 think of reading one of the superscriptions of Cicero's letters-, as if written in 
 English letters Sis'ser-oh Seize-her-eye (Cicero Csesarl). 
 
 1 It will be convenient to remember the following dates in reference to Latin 
 pronunciation, 
 
 Pre-Augustan i Plautus, B.C. 254 184, Ennius (a Calabrian Greek) B.C. 239 
 169, Cato Censorinus, B.C. 234149, Terence (an African freedman), B.C. 195 
 159, C. Gracchus (the younger tribune, see R, 3), B.C. 154 122, Lucilius, B.C. 
 148 103. 
 
 Augustan', Cicero, B.C. 106 43, Julius Caesar, B.C. 100 44, Lucretius, B.C. 
 9552, Catullus, B.C. 8747, Sallust, B.C. 8634, Virgil, B.C. 7020, Horace, 
 B.C. 65 8, AUGUSTUS, B.C. 63 A.D. 14, (Phaedrus. dates uncertain, was his 
 freedman), Livy, B.C. 59 A.D. 17, Tibullus, B.C. 54 18, Propertius, B.C. 51 ? 
 Ovid, B.C. 43 A.D. 18. 
 
 Post-Augiistan: Pliny senior, A.D. 23 79, Silius Italicus, A.D. 25 100, Lucan, 
 A.D. 39 65, Quintilian, A.D. 40 118, Tacitus, A.D. 60? 118? Statius, A.D. 61 
 96, Pliny junior, A.D. 61 105? Juvenal and Suetonius (close of first century?) 
 
 Transitional, second and third century : Aulus Gellius, A.D. 117 180, Teren- 
 tianus Maurus. 
 
 Late, fourth, fifth, and sixth century: Macrobius, Servius, Priscian (gram- 
 marians). , 
 
 It is thus seen that the Augustan period comprehends the most esteemed Latin 
 authors. To Cicero the pre-Augustan writers were antiquated. To Quintilian 
 the Augustan writers were antiquated. A difference so apparent in the style of 
 writing would naturally be accompanied by a difference ot pronunciation. 
 Cicero and Horace are our only real authorities for Augustan speech. Even 
 inscriptions are not sufficiently safe. Quintilian is the next best. Terentianus, 
 
 B 2 
 
4 PRELIMINARY ASSUMPTIONS. [Art. 7 9. 
 
 ART. 7. I assume that this Augustan pronunciation 
 differed in almost every characteristic point from the 
 Victorian pronunciation of English, and therefore also of 
 Latin, and that Englishmen will consequently have to 
 pay close attention and take much trouble to obtain an 
 approximate conception of its nature, in theory and 
 practice. 1 
 
 ART. 8. I assume also that we are not in a position 
 to obtain more than a very rough conception of its 
 details, 2 but even the small results that we can reach 
 are useful in helping us on the road, and are of 
 special pedagogical value as well as of linguistic in- 
 terest. 
 
 ART. 9 I assume that so far as the mere pronun- 
 ciation of the isolated letters are concerned, the syllabus 
 set forth by Professors Palmer and Munro, would give 
 a result nearly as intelligible to Cicero, as an Italian's, or 
 German's, or Frenchman's attempt to pronounce English, 
 when only taught by books which gave keywords in 
 their own languages, would be to an Englishman ; that is, 
 
 Maurus, and Aulus Gellius are to be regarded with suspicion, and the late 
 grammarians knew nothing of the older pronunciation but by a tradition which 
 they could not realise. 
 
 1 Classical teachers seem to have hitherto acted upon Dogberry's principle 
 that "to write and read comes by nature," (Much Ado, 3, 3, 16). French and 
 German masters have to bestow much time upon mere reading. Are not Latin 
 and Greek also foreign languages to Englishmen? 
 
 2 I cannot too strongly insist upon this point. Even after months spent in 
 Paris with French in the air all round, very few Englishmen are able to obtain 
 more than a rough conception and indifferent execution of a pronunciation so 
 utterly different from their own. Latin was at least as different, and yet we have 
 to grub it up from passing remarks made by writers two thousand years old to 
 others who, owing to their own habits of speech, knew .what they meant by a 
 mere allusion, and to piece these remarks together into some sort of a practical 
 and practicable whole. The wonder is, not that our results are rough, but that 
 they are complete enough to be usable. 
 
Art. 9, 10.] PRELIMINAR Y ASSUMPTIONS. 5 
 
 not quite so good as a Scotsman's English in a Lon- 
 doner's ears, but still an infinite advance on the current 
 English pronunciation of Latin, for that would have un- 
 doubtedly been mere gibberish in the ears of Augustan 
 Romans. 
 
 ART. 10. Being obliged to assume some kind of a 
 pronunciation, I shall therefore give the vowels a, e, i, o, 
 u, their sounds as heard in London in father, th^re, 
 mach/ne, b0re, n/le, but keep them pure when they be- 
 come short, exactly as in modern Spanish, and never 
 allow the long sounds of e, o to become diphthongal by 
 the appendage of an i and u sound, as in English say, so, 
 often called sei, sou. The diphthongal forms ae, oe, I shall 
 render by a broader e, like the German a in sprdche. 
 The diphthong eu will be like that heard for ow in Kent, 
 and often in London, albw me now; beginning with e in 
 there, while au will begin with a in father, as in German 
 haus ; both end with u in rule. And ui will be much 
 like ooi in cooing, but monosyllabic. 1 The letters /, u, 
 when forming a consonant, I shall treat as English y, and 
 German w. The latter is produced by sounding v 
 without allowing the lower lip to touch the upper teeth. 2 
 
 1 The relative qui = qu + / has become chi (pronounced ki) in Italian, but cni 
 = c-ui, or qu-ui, ancient QVOEI, has remained cut (nearly cu-i) in Italian. Did 
 Quintilian distinguish the sounds (i, 7, 27) ? 
 
 2 Using if for this German sound, TV, v for the English sounds, and u for th 
 pure Latin vowel u run on as a diphthong to the next vowel, any one who wishe 
 to arrive at a conclusion respecting the Latin consonantal v, must learn t 
 pronounce and distinguish readily, the four series of sounds : ua ue ui uo, wa w 
 wi wo wu, v'a v'e v'i v'o v'u, va ve vi vo vu. These sounds were pronounced t 
 the audience when the paper was read on which this tract is founded. Observ 
 that iiu was impossible as Quintilian observes, who assumes the Eolic digamm 
 as the sound: in his seruus et uulgus Aeolicum digammon deslderatur i, 
 4, 8) ; nostrl praeceptores seruum ceruumque U et O litterls scripserunt, quia 
 subjecta sibi vocalis in unum sonum coalescere et confundl nequlret ; nunc U 
 gemina scrlbuntur ea ratione quam reddidl ; neutro sane modo vox, quam sentl- 
 
6 PRELIMINARY ASSUMPTIONS. [Art. 10. 
 
 The letters c, g will always have their so-called " hard ' ; 
 sounds in sceptic, get* H initial will.be taken nearly as 
 in English, especially when sung, that is with its jerk 
 without its hiss. 2 After a consonant as in th, ch, pit, h 
 will be simply neglected. Qu will be treated pre- 
 cisely as in English, that is, as a single letter, phonetically 
 a labialised k. T, D will also receive their English sounds, 
 
 mus, efficitur ; nee inutiliter Claudius Aeolicam illam ad hos usus litteram adje- 
 cerat (i, 7, 26). My own belief is that in the oldest form of the language 
 z', it were always vowels, which initially diphthongised with the following vowel as 
 la.) ua, and that this stage was recognised, at least in writing, as long as zz, 7/?/, 
 we e replaced by z, ui, but that when zz', uu, (or as we usually write, and as will 
 here be always written, ji, vu] were employed, the sounds were those of German 
 /, w. I have more doubt about / than v. It is probable that the real Augustan 
 pronunciation hovered between uo, v'u, but as Quintilian adopts the latter (if, as 
 many think, the digamma which he knew, to say nothing about Homer's, was z/'). 
 I shall use it a.i being nearer to our old habits. I do not think English w was 
 ever used In modern Italian : uomo, uovo, have uo not ivo. The vowel 
 character of z', u was perhaps never lost after a consonant. Thus in the Benares 
 pronunciation of Sanscrit although y, v, initial have their regular English values, 
 (v with the teeth), yet after consonants they remain the pure Latin vowels /, u, 
 though still written as y, v, to shew that they do not form distinct syllables. 
 Such double forms as : ab-jete abi-ete, fluv-jo'rum fluvi-o'rum, could scarcely 
 have arisen from other habits. They are quite similar to ; a-erri, ag-rl, hereafter 
 considered. 
 
 1 The following well-known passage, justly insisted on by Prof. Blair, seems to 
 have been too much overlooked* in the warm controversy on this subject : it 
 surely conclusively shews that Quintilian, at least, pronounced C with the same 
 sound before all vowels : nam K quidem in nullis verbis utendum puto, nisi qua; 
 significat, etiam ut sola ponatur ; hoc eo non omisT, quod quldam earn, quotiens 
 A sequatur, necessarian! creclunt, cum sit C littera, quae ad omnes vocales vim 
 suarn perferat, (r, 7, 10). C of course determines G. As to S, the English 
 z sound has only been "developed" in French and Italian, and has not yet 
 touched Spanish. It was unknown to Greek, and though in modern times the 
 f, (originally zd ?) has drifted into z, a remains as s. 
 
 2 This effect is produced by bringing the vocal chords together so that they 
 should be ready to emit the vowel sound immediately, and should not allow un- 
 vocalised breath to precede, by not holding them too tightly together, and by 
 driving the breath through them at first with a little impulsion of the muscles of 
 the lungs. All this is easily imitated by opening a valve and jerking slightly the 
 bellows of an accordion or concertina, so as to make the commencement of the 
 note louder than what follows. Of course it is impossible on such an instrument 
 to produce the effect of H in any other way. See (Art. 51) note. 
 
Art. 10.] PRELIMINARY ASSUMPTIONS. 7 
 
 though it scarcely admits of doubt that these were as un- 
 known to the Latins, as they are now to most Europeans. 
 But these sounds are most easy to English organs, and 
 Englishmen sometimes reside for years on the continent 
 without perceiving the difference, or live years in India, 
 (where both sounds are used, and help to distinguish 
 words,) without knowing anything about it, as probably 
 Cicero would have known at once. 1 S will also be 
 always " hard " like ss in hiss. R will always be strongly 
 trilled ; to which un-English but common Scotch 
 fashion, great attention must be paid. Final M will 
 be fully treated later on. The foreign y, I shall 
 generally treat as /", perhaps Cicero made it the 
 French u 9 for he seems to have been fond of a bit of 
 Greek (O, 9), and he may have called z either dz or 
 zd, but this is a mystery. Had I occasion to use it in 
 Latin I should say dz, for that is still heard in Italian, 
 but the letter occurs only in Greek words, where, owing 
 to the Greeks' apparent absence of power to say TO and 
 their fondness for or, the combination zd appears more 
 probable than dz. 
 
 It is not to be concluded that I consider these sounds 
 to be perfectly correct or even justifiable by any authori- 
 ties which can be cited. On the contrary, it is probable 
 that many exceptions occurred, and many slight distinc- 
 
 1 All Indians recognise the English T, D, as their own cerebrals, and the 
 French, Italian, and general European T, D, as their dentals. For the dental, 
 the tongue is pressed against the teeth as for the two English TH. In Cum- 
 berland, Yorkshire, and Derbyshire, at least, if not elsewhere in England, these 
 dental T, D, are heard in connection with R, as TR-, DK-, or -TER, -DER, 
 and are often confounded with the two TH by Southerners. The Irish version of 
 
 speec 
 
 English Pronunciation, Part IV. (in the press). 
 
8 PRELIMINARY ASSUMPTIONS. [Art. 10, 1 1. 
 
 tions prevailed among Augustan literary men of which 
 we can form no adequate conception. But these 
 rough approximations will piobably suffice for all school 
 purposes. 
 
 ART. ii. I assume that the Latin accent consisted 
 solely in raising or depressing, or in first raising and then 
 depressing the musical pitch of the voice at which a given 
 syllable was uttered, and that this use of the acute, grave, 
 and circumflex inflexions of voice respectively was quite 
 independent of the loudness of tone, and (except as regards 
 the third or circumflex accent) of the length of tone, 
 although the determination of the precise syllable on 
 which it immovably occurred in every word of more than 
 one syllable, did depend on the length of that syllable or 
 on the length of the adjacent syllables, as' we shall see. 
 The law of this relation was, however, different in Latin 
 and in Greek, and there is nothing exactly corresponding 
 to it in any European language. The nearest approach to 
 the sound of the Latin accent is now found in Swedish 
 and Norwegian, but the laws of its use are quite differ- 
 ent. I assume, therefore, that the relative pitch of every 
 syllable in every Latin word, was fixed, so far as higher 
 and lower was concerned, that is, that it never varied in 
 position in whatever part of a sentence the word was 
 used, and that this position depended on the length of 
 the syllables (M. 3 5). 1 
 
 1 This is the only result at which I can arrive after very careful study. The 
 expressions of Cicero seem all to point to " acuta vox " meaning a high pitch of 
 voice, and "gravis vox" a low pitch, independently of loudness. The terms 
 "intendens, contentio," from tightening strings, and "remittens, remissio," from 
 slackening springs, point the same way. The extreme confusion that prevails in 
 the use of the words " raising and lowering the voice," as speaking either at a 
 higher and lower pitch, or else with greater and less loudness, renders almost 
 all that is usually said ambiguous. Thus when Mr. Roby says : " Accent is the 
 
Art. 1 2. ] PRELIMINAR Y ASSUMPTIONS. 9 
 
 ART. 12. Next, reserving the nature of length for 
 close consideration presently, I assume that long and 
 short syllables were invariably and carefully heard and 
 distinguished, so that any error was instantly felt and 
 detected, as it would often make nonsense as : mala 
 mala, just as a wick pool would probably not be under- 
 stood if said for a weak pull that it was the main-stay of 
 rhythm both in prose and verse, so that no rhythm was 
 possible to an Augustan Roman which did not observe 
 it, and that no rhythm was pleasant which did not make 
 the relative positions of long and short distinct. (N. 1 
 O. P. generally). 
 
 elevation of the voice with which one syllable of a word is pronounced," either 
 kind of elevation might be meant, though mere increase of loudness is rendered 
 probable, not certain, by what follows, " in comparison with the more subdued 
 tone with which the other syllables are pronounced," (Grammar 296). Corssen's 
 remarks on Latin "betonung" (n. 794 892), are unhappily one mass of con- 
 fusion from this cause. Hence I have been extremely careful in my own lan- 
 guage. It is I find almost impossible to get any intelligible account of accentua- 
 tion in any living language from any living speaker, so little has the subject been 
 studied ; nor, although I have thought upon it much and have observed speakers 
 often, am I yet able to come to any definite conclusion upon the nature of what 
 is usually called accent even in French and English. The later Latin gram- 
 marians, such as Priscian, had no real knowledge of what was meant by quantity 
 and pitch accent, which had become quite as obsolete and traditional in their 
 day as in our own. For details on this subject, see my paper on "the Physical 
 Constituents of Accent and Emphasis," Transactions of the Philological Society 
 for 187374, pp. 113164. 
 
 1 In my paper on Accent and Emphasis (just referred to) I had originally read 
 una syllaba in (N i) as a nominative. Mr. Roby kindly pointed out to me that 
 it is probably an ablative, and should be read itnd syllaba. After re-reading the 
 whole context, . I entirely agree with him. Cicero in 170 173, is speaking 
 about rhythm generally, and mentions Aristotle's rule, "is igitur versum in 
 oratione vetat esse, numerum jubet," 610 pv#p&v 6el exeiv TOV Xofuv. /uerpoi/ 3e /i>/' 
 Troika ^ap carat, and after referring to Theodectes, and Theophrastus, proceeds 
 to say : " who would put up with pe.ople that disapprove of what these authors 
 say, unless\they were ignorant that these authors had said it? Now if this is the 
 disc, (and I think it is) because they are not influenced by their own senses, have 
 they no feeling of imperfection, of want of finish, of mutilation, of lameness, of 
 redundancy? Why [here the extract in N begins] the whole audience in a 
 theatre would cry out if a verse were shortened or lengthened by a single 
 
io PREL1MINAR Y ASSUMPTIONS. [Art. 13, 14. 
 
 ART. 13. I assume, however, that rhythm did not de- 
 pend solely on the length of syllables, though the laws 
 of versification have apparently no other basis, but that 
 the position of the pitch accent or highest pitch of the 
 voice, in a word, was also operative, and that it was to 
 secure variety here that the laws of caesura in Latin verse 
 were gradually developed. 
 
 ART. 14. Lastly I assume that the Augustan Romans 
 had no force accent, that is, that they did not, as we do, 
 distinguish one syllable in every word invariably by 
 pronouncing it with greater force, that is, with greater 
 loudness than the others, but that the force varied ac- 
 cording to the feeling of the moment, or the beat 1 of the 
 
 syllable. And yet the common people know nothing of feet, nor observe any 
 rhythm, nor have any idea what offends them, or why, or in what particular it 
 offends them. But nature herself has put into our ears the power to appreciate 
 (judicium) all long and short quantities in sounds, and rising or falling inflec- 
 tions in speech." This is the only consistent interpretation which I can put on 
 the last sentence, and this fully bears out what I have said in the text. 
 
 1 Barred music (quite a recent invention, arising from the discovery of poly- 
 phony, or the power of singing several melodies by different voices at once 
 without creating confusion, a thing unknown to the ancients) depends upon 
 exactly equal intervals of time regulated by the beat of the conductor's baton 
 or foot, and it is a rule that the note played to the first beat in a bar, should be 
 slightly louder than that on the third, and this again louder than that on the 
 second (in both triple and common time), which should itself again be generally 
 slightly louder than that on the fourth (in common time). But this comparative 
 loudness which forms the undercurrent of rhythm and helps to " mark the 
 time," though very conspicuous in dance music and marches, is constantly over- 
 ruled by the laws of forte and piano, staccato and legato^ just as the precision of 
 the length of the intervals is overruled by the directions accelerando and len- 
 tando. Now there may have been such an ictus in Roman verse, but it was 
 often not at regular intervals as we shall see (Art. 105) and it certainly never over- 
 ruled, (as in our music also it never overrules) the musical pitch and length, or 
 even the emphatic loudness of individual notes. Let the English reader then 
 who knows music, remember that ictus can only apply to the conductor's beats or 
 the interval between two of them ; that such intervals were not necessarily of the 
 same length (owing to introduced lentando's) : that q^lantity answers to the 
 time of duration of any note (as crotchet or quaver) ; and accent, in Latin, to 
 the alterations of pitch by ascending or descending the musical scale, this being 
 
Art. 14, 15.] EXERCISES ON QUANTITY. n 
 
 timekeeper in singing, and was used for purposes of 
 expression ; just as with us, musical pitch is free, that is, 
 just as we may pronounce the same word with different 
 musical pitches for its different syllables, and in fact are 
 obliged to vary the musical pitch in interrogations and 
 in replies. The fixity of musical pitch and freedom of 
 degrees of force in Latin, and the freedom of musical 
 pitch and fixity of degrees of force in English, sharply 
 distinguish the two pronunciations even irrespective of 
 quantity. 
 
 II. Elementary Exercises on Quantity. . 
 
 ART. 15. Having by these assumptions cleared the 
 way for work, I shall endeavour to develop the feeling 
 for Latin quantity or syllabic length, and Latin accent or 
 comparative pitch by suggesting a graduated series of 
 exercises which shall lead up finally to reading Latin in 
 its own rhythm, with due regard to sense. I shall not 
 actually give all the exercises at length but rather hint at 
 the paedagogical process than carry it out, as any teacher 
 will be immediately able to do so for himself. But as an 
 additional assistance I shall state the intention of each step, 
 which of course must remain a secret to the pupil. The 
 
 determined by the number and length of the syllables of the word sung (just as 
 if when several were combined in one word, and the last but one was a quaver, 
 the others being semi-quavers, the quaver should have the highest pitch) ; while 
 force answers to the forte and piano. Hence all four, ictus (or beat), quantity 
 (or length), accent (or pitch) and force (or loudness) are naturally independent 
 of each other, although habitually they may be made dependent one on another, 
 or several of them upon something else (as a musical or intellectual conception). 
 Latin accent consists in fixing the syllable of a word which should have the 
 highest pitch ; English accent consists in fixing the syllable of a word which 
 should have the greatest force. 
 
12 ELEMENTARY EXERCISES [Art. 1517. 
 
 case presumed is that which is at present most common, 
 where a school teacher who wishes .o instruct his higher 
 class of pupils that can already translate Latin with 
 tolerable ease, but have hitherto used the English bar- 
 baric pronunciation, or, having accepted the new method, 
 have not sufficiently realized the action of varied length 
 and varied musical pitch, upon the values of the letters. 
 When the pupil is altogether ignorant of Latin, the case 
 is slightly different, and perhaps easier, for he has no bad 
 habits to unlearn, and will be considered hereafter. 
 
 ART. 1 6. To develop the feeling for division of time. 
 Use a pendulum swinging about four times in a second. 1 
 Class say la ! la ! &c. once to each swing in chorus, then 
 la ! lala ! la! lala ! one la ! for the first, and two for the 
 second swing. Then la-a ! la ! la ! la-a ! la ! la ! keeping 
 the la-a ! two swings, and so on. Try the same exercise 
 with all the vowels in succession (sounds as in Art 10, of 
 course), and without any consonant. Shorten string of 
 pendulum to about 3 inches and try again. 2 
 
 ART. 17. Take the vowels in any line of Virgil 
 and pronounce the short (one swing) and long (two swings) 
 without consonants ; thus A, 2 gives eioeiieeodii 
 add, where the line over the vowel signs indicates a long 
 vowel. Repeat the exercise without pendulum. Make 
 
 1 Take a yard of sewing thread and tie the ends ; pass it through and over the 
 loop of a common key. This makes a pendulum which answers every purpose. 
 By shortening the string when held between the fingers, any degree of rapidity 
 can be obtained ; between 9 and 10 inches of length gives quarter seconds. 
 
 2 More complicated divisions of time, as in music, are not needed. The method 
 of giving instruction in that case has been well developed from a French model 
 by the Tonic Solfaists, see : The Standard Course of Lessons, and Exercises in 
 the Tonic Sol-fa method of teaching music by John Curwen, edition of 1872, pp. 
 160, see pp. 7, 8, 18 20, &c. under taa-tai-ing. 
 
Art. 1719.] ON QUANTITY. 13 
 
 the class feel at least an appreciable difference of length. 
 Continually increase speed. 
 
 To shew the slight effect of an initial consonant, try 
 the same exercises with different initials /, b, /, d, c, 
 g,f, s, r, I, m, 11, j, v, h. Take care in saying pe, pi, po, 
 pe, pu, to run the consonant on to the following vowel, 
 and avoid running the vowel on to the following con- 
 sonant. 
 
 ART. 1 8. The nature of syllabication has now to be de- 
 veloped. This consists in the running of a consonant on 
 to a vowel, by a glide, and of the vowel by another glide 
 on to the following consonant. Both glides may be 
 indicated by proximity of the letters. In saying pe pi, 
 guard against pe pi (long vowels,) /*/z (second vowel long, 
 second p gliding on to both vowels), pepi (both vowels 
 short, glide as before). Develop the feeling of each as 
 distinct from the other. In pepi observe that even less 
 time is lost between / and// than \&pepi. 
 
 ART. 19. To develop feelings of pitch and force as in- 
 dependent of length. Represent raised pitch by an acute 
 accent ('), increased force by a turned period () ; lowered 
 pitch and diminished force are not specially represented. 
 The reader is particularly requested not to confound the 
 acute accent (mark of raised pitch) with the turned 
 period (mark of increased force), and to remember that 
 they are totally distinct (Art. 14). In the well-known 
 song non piu andrai* farfallo*ne amoro*so, the loudest 
 and longest notes are given to the syllables preceding the 
 *, but each of these is lower in pitch than the pre- 
 ceding syllable. Distinguish/^// (uniform length, pitch, 
 and force), pe pi, pe pi, (both, uniform force), pe' pi, 
 pe pi', pc pi. pe pi', (all with short vowels, and without 
 
14 EXERCISES ON QUANTITY. [Art. 19, 20. 
 
 running the vowels on to the consonants). Vary as p'e i, 
 pet, e pi, epi, p'e i, e'pi, &c. Then distinguish pe'pi, 
 pe'pi, pe'pi, pe'pi, and so on, where p glides on to both 
 vowels. Vary with all the consonants. This exercise 
 requires great practice to obtain ease and certainty of 
 execution. Write on blackboard and make whole class 
 follow. 1 The final rising pitch does not occur when 
 both syllables belong to the same word in Latin, but it 
 constantly occurs when the syllables are distributed 
 between two words. 
 
 ART. 20. To develop the effects of "position" First 
 take doubled consonants. Distinguish accurately be- 
 tween pepi (as in pepper), and peppi (as in dee/ /ool). 
 The form p'epi is not Latin but English, peppi is rare in 
 English, except in compounds or between words and 
 frequent in Latin, even within the same words. Com- 
 pare English pity, city, silly (only one /heard), and pu/ fo 
 it, se/ /o, the grea/f foe is grea/er, we /ie, (one / ), we'// /ie 
 (two /s), ti// eight (one /), ti///ate (two /s), pe;//vnfe (two 
 ;zs), pe;/;;y (one ;/). Thou slee/est, thou slee/-/est! Missy 
 mu-^ent: u;//;/own and unowned ! Latin a;^-;zus, to/-/ it, 
 e^e ter-ra. The effect is the same as that of doubled 
 consonants in Italian. It is difficult for an English organ 
 to acquire when not occurring in compound words. The 
 doubled r as in terra is particularly difficult for English 
 speakers, and should be well practised. No relaxation 
 of the organs, no puff of wind or grunt of voice should 
 intervene between the two parts of a doubled consonant, 
 
 1 The following direction is of the utmost importance. " The teacher never 
 sings [speaks] with his pupils, but sings [utters, reads, dictates] to them a brief 
 and soft pattern. The first art of the pupil is to listen well to the pattern, and 
 then to imitate it exactly. He that listens best, sings [speaks] best." Curwen's 
 Standard Course, p. 3. See also (Art. 96). 
 
Art.20 22.] ARTIFICIAL RHYTHMICAL UNIT. 15 
 
 which should more resemble separated parts of one arti- 
 culation than two separate articulations. 1 
 
 ART. 21. Next take two consonants not forming an 
 initial combination. The Latins had not the power we 
 have of combining many consonants. Final consonants 
 were always well developed, and there was sensible rather 
 than perceptible pause 2 before beginning a new syllable 
 with a second consonant. Make this clear by a hyphen 
 in writing, as is uniformly done in the Appendix. Try 
 imaginary words as pep-si* pep-si' (varying force and 
 pitch). Initial Latin combinations consisted of /, &, t, 
 d, c, g, followed by either r or / (//, dl are not used in 
 Latin, and dr is rare) and sometimes n and s or preceded 
 by s, or rarely both. In all these cases the vowel might 
 (within the word) be always run on to the first consonant, 
 as pcp-ri, or not run on as pe-pri.* 
 
 III. The Artificial Rhythmical Unit of Latin 
 Speech. 
 
 ART. 22. The preceding studies are on natural quan- 
 titative rhythm, the following are on the form they have 
 assumed in Latin itself. A Latin syllable is to be con- 
 sidered to end in its vowel when this is possible, and is | 
 then called open. The length of time occupied by the 
 preceding consonants is disregarded, even when there are 
 
 1 Duplication of consonants is consequently regarded simply as the energetic 
 utterance of a single consonant, as will be subsequently explained. It plays a 
 great part in all quantitative languages as Sanscrit, Persian, Arabic. 
 
 2 That is, one rather felt by the speaker than the hearer, not quite amounting to 
 Quintilian's unbecoming pause (Art. 85). 
 
 3 Of course the vowel might be either long or short in this case, as pefi-ie, 
 But the former usage seems confined to poetry (Art. 24) and the latter is doubtful. 
 
1 6 ARTIFICIAL RHYTHMICAL UNIT. [Art. 22. 
 
 two or three as sta-, scrl-. The syllable ending in a 
 vowel is considered to be long or short according as the 
 vowel it contains is long or short, and as this length is 
 not marked in Latin books, a knowledge of it has to be 
 acquired for each individual word. This is a task ab- 
 surdly difficult for learners, and all Latin books now 
 printed, whether for school or other purposes, and not 
 merely dictionaries and graduses, ought to have the 
 vowels which specialists (often with great difficulty) have 
 ascertained to be long, properly distinguished by the sign 
 of length. When the schoolboy makes a mistake here, 
 it is the editor, not the boy, that should be punished. 
 The editor ought to have known, and the boy trusted the 
 editor. But all of us, big or little, are foreigners (Art. 
 7), and have no knowledge of the word but what its 
 letters give us, and hence we should be always properly 
 treated by editors. The orthography used in printed 
 Latin books is notoriously not that of their writers. 
 Hence there is no objection to this little improvement. 1 
 
 1 Quintilian as a Roman knowing words by ear, finds that it is below a gram- 
 marian's dignity to determine what letters should be used in a word except in 
 rases of doubt: recte scribendi scientiam cujus ars non in hoc posita est, ut 
 noverimus quibus quasque syllaba litteris constet (nam id quidem infra gram- 
 naatici ofncium est) sed totam, ut mea fert opinio, subtllitatem in dubiis habet 
 (i, 7, i). Hence he finds it supremely foolish to put the long mark on all 
 long syllables, but admits that it is necessary to know an appletree (malus) from 
 a bad man (malus), a post (palus) from a marsh (palus), [by the bye, if the reading 
 is correct, (it can be easily altered,) Horace says palus, with both short, in : regis 
 opus ; sterilisque diu palus, aptaque remls, A. P. 65, where of course he might 
 have written : sterilisque palus quondam, or even : sterilisque palus diu, with a 
 long vowel shortened in hiatus, as in Cicero's, Q, 13] and nominatives from abla- 
 tives : ut longls syllabls omnibus opponere apicem ineptissimum est, quia 
 plurimae natura ipsa verb! quod scribitur patent : sic interim necessarium, cum 
 eadem littera alium atque alium intellectum, prout correpta vel producta est. 
 facit, ut malus arborem significet an hominem non bonuin, apice distinguitur 
 palus aliud priore syllaba longa [palus] aliud sequent! [palus] significat ; et cum 
 eadem littera nominative casu brevis, ablatlvo longa est, utrum sequamur. 
 plerumque hac nota. monendi sumus (i, 7, 23). For foreigners who have 
 
Art 23, 24.] ARTIFICIAL RHYTHMICAL UNIT. 17 
 
 ART. 23. When on account of two or more con- 
 sonants following which cannot begin a word, the syllable 
 is bound to terminate in one or more of them, then, 
 whether the vowel is long or short, the syllable is con- 
 sidered long. This occasions great difficulty in deter- 
 mining whether the vowel is naturally long or short. 1 
 As a mere matter of convenience, having no Augustan 
 ears to discover the error, or Augustan sense to de- 
 termine what is right, and guided by many other con- 
 siderations, I shall usually treat such vowels as short. 
 
 ART. 24. Even when the consonants are capable of 
 beginning a syllable as in a-gro, the vowel is sometimes 
 run on to the consonant, and the first syllable made long 
 as ag-ro, (see ag-ricolam, D, 9) but we may readily sup- 
 pose that this occurred only in poetry. 2 
 
 not heard and used the word over and over again till it is ingrained, but who 
 have to gather it by the eye from time to time, sometimes once in a year, the 
 ambiguity which Quintilian even as a Roman felt for sense, remains in all cases 
 f jr pronunciation. " Position " is determinable by easy rules, and hence, except 
 to warn beginners (as in the Appendix), need not be marked. But the natiirally 
 long vowel should be invariably marked, as it is throughout this tract. 
 
 1 There is an old bad custom of putting a long mark over the vowel in a syllable 
 which is long by "position," as: agrestem. We must distinguish between the 
 long vowel and the long syllable. Of this word agrestem, Quintilian says, 
 a brevis, gres brevis, faciet tamen longam priorem [syllabam], (9, 4, 86). 
 Cicero says : inclitus dlcimus brevl prlma littera, tnsdnus products, inhumdnus 
 brevl, Infeltx longa, et, ne multls, quibus in verbls eae pnmae litterae sunt, quae 
 in sapiente atque felloe, producte dicitur ; in ceteris omnibus, breviter : itemque 
 composuit, consuevit, concrepuit, confecit : consule veritatem, reprehendet : refer 
 ad aures, probabunt : quaere, cur? ita se dlcent juvari: voluptati autem aurium 
 morigerari debet oratio (Or. 159). This shews clearly that the length of the 
 vowel did not depend upon " position," as the bad notation alluded to, seems to 
 imply. 
 
 2 Quintilian says : evenit, ut metri quSque condicio mutet accentum, ut pecudes 
 pictaeque voltic-res ; nam voluc-res media acuta legam, quia etsi natura brevis, 
 tamen positione longa est, ne faciat iambum quern non recipit versus herous 
 (i, 5, 28). This shews that he would naturally say vdlucres as an anapest, and 
 that voli4C-res was merely poetical. Similarly : quSe flunt spatio, slve cum 
 
 C 
 
i8 ARTIFICIAL RHYTHMICAL UNIT. [Art.25-27. 
 
 ART. 25. When the vowel was short the consonant 
 was necessarily dwelt upon, and even a slight, and 
 when the consonant was mute, quite perceptible pause, 
 was probably made between the consonants as in ac-tus 
 (Art. 21). 
 
 ART. 26. In words ending with a consonant there 
 was always a possibility of a following consonant to 
 lengthen the last syllable, and even when the word 
 ended with a vowel there might follow a troublesome 
 initial combination, creating position. Hence perhaps 
 it arose that the length of a final syllable having a short 
 vowel was unsettled, and a short syllable might be used 
 for a long one. At the end of a clause, a syllable, no 
 doubt, was often lengthened, and Cicero repeatedly tells 
 us that length was indifferent in such cases (P. i 3). 
 
 ART. 27. Syllables ending in a short vowel, not run 
 on to the following consonant, were taken as short. 
 Even where it seems that the vowel must be run on to 
 the consonant; if a following vowel allows the separation 
 of the consonant from the preceding vowel, I think, 
 that this medial consonant was probably attached in 
 speech to the following only and not to the preceding 
 vowel as well. This was almost certainly the case in the 
 middle of a word, and was probably the case at the end 
 thus me-di-u-ses-t = medius est* The words were run on 
 
 syllaba correpta producttur, ut Italiam fdto profitgus, seu longa corripitur, ut 
 iinius ob noxam etfurids, \unius for uriiiis\ extra carmen non deprehendas ; sed 
 nee in carmine vitia ducenda sunt (i, 5, 18). 
 
 1 Mr. Roby (Grammar, p. 87, 272, 273, and preface, p. IxxxiiL, 2nd edition, 
 adopts the English habit, as in critical, where the first syllable ends with the 
 glide of the first i on to t, and the second begins with the glide of t on to the 
 second z. Notwithstanding the reasons he has adduced in the passages cited, I 
 incline to think that the Latins did not speak thus. So far as I can judge, 
 modern Italians do not. When a consonant occurred between two vowels as in 
 
Art. 27.] ARTIFICIAL RHYTHMICAL UNIT. 19 
 
 in all cases very closely together, without any gaps. In 
 fact, as Cicero says, Latins were not allowed to divide 
 their words (Q. 8). This is the practice in almost all 
 languages. We complain of French use in that respect, 
 
 fames, there was, I think no glide of the first vowel on to the consonant, as of a 
 on to m, but only of the consonant m on to e. The glide from vowel to consonant 
 occurred only when at least two consonants followed, and was even then not 
 compulsory if these two consonants could form an initial combination. In other 
 words no such glide occurred without "forming position." At the end of a 
 clause this was always assumable. In position then the vowel glided on to the 
 first consonant, and then if the consonant was mute, silence ensued ; if not mute, 
 the consonant itself sounded a very short time. The second consonant glided 
 only on to the following vowel. If we use the minus sign ( ) to shew absence of 
 glide, and the plus sign (-f) to shew presence of glide, it seems to me that Latin 
 fames = f + a m + e + s, annus = a + n n + u + s, agro a g + r 
 + o, erunt = e r + u + n + t; whereas English famine = f-fa + m + i+n. 
 The question is exceedingly difficult. Frenchmen as a rule assert that their own 
 medial consonant belongs to the second vowel only. The late Mons. A. C. G. 
 Jobert, who spoke English admirably, and was a man of science as well as a 
 teacher of languages, could only hear the glide on to the first consonant (Col- 
 loquial French or the Philosophy of the Pronunciation of the French Language, 
 1854, pp. 191). I had long conversations with him, but could not get him to feel 
 that he also glided -the consonant on to the following vowel, as I heard him dis- 
 tinctly pronounce. M. Tourier (Model-book, 1851) had also noticed the glide 
 from the preceding vowel to the consonant, but not so fully. M. Favarger, a 
 living French teacher, who has carefully studied pronunciation, after for a long 
 while refusing to recognise the first glide on to the consonant, in a recent conver- 
 sation with me stated that further observation constrains him to admit it. These 
 facts serve to shew the great difficulties in the way of the investigation even in 
 living speech, and to explain the hesitation with which I speak respecting Latin, 
 especially when it is impossible to enter into the numerous little reasons which 
 collectively make me incline to the opinion here expressed. This opinion may be 
 made the ground-work of a practical pronunciation, but when the vowel is thus 
 separated from the consonant Englishmen will be apt to lengthen it, and this is 
 certainly a worse error than running it on to the consonant ; fa' - mae and fa me 
 must be kept quite distinct. The pitch accent must be carefully separated from 
 the force accent, and then much of the difficulty will be overcome. If however 
 fa + m -f e is said, care must be taken not to fall into fa + m m + e as in 
 flammae, which perhaps an Augustan Roman would hear in the English sound ; 
 that is, hearing a new sound he would refer it naturally to that most familiar to 
 his own organs. At least such is the habit of all moderns. These remarks 
 apply especially to the pronunciation of both cane and Cannae like the Scotch 
 canny, and so on in other words, as virl like English 7>irile, homo like English 
 hommage, tremor like English tremorir or tremble, &c. which I believe to be 
 mere English expedients to keep the vowel short, but might have served to 
 lengthen the syllable to a Roman's ears. 
 
 C 2 
 
20 ARTIFICIAL RHYTHMICAL UNIT. [Art. 27,28. 
 
 the French complain just as much of English use. It 
 is only the foreigner who breaks up a sentence into un- 
 connected words. 
 
 ART. 28. Now the most artificial part of the Latin 
 and Greek quantitative rhythm, consisted in taking 
 a short vowel, or syllable as the unit of length, and 
 supposing that it was always of the same length, and 
 that the long vowel or syllable was of exactly twice 
 that length. Nothing of this kind is likely to have 
 occurred in speech or declamation, but may have oc- 
 curred in chanting, and must have occurred in simul- 
 taneous chanting. Cicero found Greek lyrics entirely 
 wanting in rhythm when the music was absent, and 
 had great difficulty in following some of the comic 
 metres when the piper was not present to mark the 
 time (O. 9 14). Hence the artificiality is apparent. 
 Still to begin with, this artificiality must be aimed at, 
 because we have nothing like it in English except in 
 barred music, with crotchets and quavers (Art. 14, 
 note). In English singing the consonants are reduced 
 to nonentities, and the short vowels lengthened on 
 long notes. Later Greek and Latin chanters did play 
 such mad pranks occasionally, (p. 28, note) but the 
 older rhythms were very much simpler, the music was 
 merely to steady the voice, and it was important that 
 the words should be intelligible. That they read 
 the verse in a semi chant, if not a full chant, is scarcely 
 to be doubted. That even declaimers did so some- 
 times, the story of Gracchus's piper told by Cicero leads 
 ? us to think (R. 2 7). But from what Cicero him- 
 1 self says, I think that he did not chant much more than 
 \many of our own public speakers, especially when they 
 
Krt.2%,29.} ARTIFICIAL RHYTHMICAL UNIT. 21 
 
 indulge in orotund 1 - delivery. Even declaiming without 
 a vestige of chant, is not a century old in English, and 
 many English readers always chant poetry, or read in 
 a peculiar style totally different from their prose habits, 
 just as their prose reading differs from their ordinary 
 speech. In English, however, the inflexion of the voice 
 is free, except in the final cadence. In Latin it was 
 fixed for every word. 
 
 ART. 29. To make all long syllables of the same 
 length or nearly so we shall have to take liberties with 
 the lengths of their vowels and consonants. The 
 rhythmical relations are, however, best studied in com- 
 bined syllables called feet. In all the following ex- 
 aminations of length, set your pendulum to the length of 
 time which you wish your short syllable to occupy, and 
 reckon one single swing for a short, and two single swings 
 (or one double swing) for a long syllable. Instead of 
 a pendulum, the teacher may use rapid finger taps, if he 
 can trust himself for making them sufficiently isochronous. 
 Then the fall of the finger, or the rise is a short length, 
 and the rise and fall together a long length. For some 
 feet, spondee, dactyle, anapest, choriamb, this answers 
 well ; but difficulties arise for trochees, iambs, paeons, 
 
 1 Adopting the favourite elocutionist's adjective made from ore rotundo, com- 
 pare Quintilian : sit autem in prlmls lectio [of verses] virllis et cum suavitate 
 quadam gravis, et non quidem prosae similis, quia et carmen est et se poetae 
 canere testantur : non tamen in canticum dissoluta nee plasmate (ut nunc a 
 plerlsque fit) effeminata : de quo genere optime C. Caesarem praetextatam adhuc 
 accepimus dixisse : si cantas, male cantas; si legis, cantas (i, 8, 2). This 
 affected plasma was evidently something approaching to a high pitched oro- 
 tundity, for he compares it to harmonics on a pipe : nee verba in faucibus patietur 
 [hlc magister] audlrl, nee oris inanitate resonare, nee, quod minime sermon! 
 puro conveniat, simplicem vocis naturam pleniore quodam sono circumlinlre, quod 
 Graecl KaTa7re7rA.ao-/u.eVoi/ dlcunt. Sic appellatur sonus tibiarum, quae praeclusls 
 quibus clarescunt foraminibus, recto modo exitu graviorem splritum reddunt 
 (i, ii, 6. 7). 
 
22 ARTIFICIAL RHYTHMICAL UNIT. [Art. 29, 30. 
 
 epitrites, &c., which will require the interval between 
 two taps to be taken as the standard short, and hence 
 involve very rapid tapping. The exact equality of divi- 
 sion is not ultimately of much importance. Its primary 
 use is to destroy our own modern Western habits in 
 which quantitative rhythm is not known. When a 
 teacher can hear a native pundit read some lines of 
 Sanscrit, or a native Arab or Persian literary man read 
 some Arabic or Persian poetry, where the rhythm is still 
 entirely quantitative, and observe how grand and marked 
 the long syllables stand out from the short ones, although 
 the latter are not hurried over, he will have a better 
 notion of quantitative rhythm as a reality, than he can 
 educe by any amount of mere reading and imagining. 
 
 ART. 30. The length of syllables may be marked by 
 the teacher as in this tract, and in the quantitative ex- 
 amples. The diphthongs " ae, oe, au, eu, ui," are always 
 long. The vowels " a e I 6 u " are also always long and all 
 other vowels are short, so that the short mark becomes 
 unnecessary. A short vowel followed by another vowel 
 or by a single consonant or an initial combination of 
 consonants, ends a short syllable. A long vowel under 
 the same circumstances ends a long syllable. Any vowel 
 long or short or any diphthong followed by a consonant 
 with a hyphen after it, occurs in a long syllable. After 
 a very little practice it will be found unnecessary to mark 
 "position," by putting this hyphen after the consonant 
 which ends the syllable. But it should be added in 
 exceptional cases, as, " ag-ro voliic-res." The invariable 
 place of highest pitch in each word should be marked by 
 an acute or circumflex according to rules subsequently 
 given, (Art. 41) but of course well known to the teacher. 
 
Art. 3033.] METRICAL FEET. 23 
 
 The variable position of force should be marked by a 
 turned period, or by underlining on the black board. 
 Pupils should be encouraged to mark the naturally long 
 vowel, and that only, in their books. 
 
 IV. Metrical Feet in Latin Words. 
 
 ART. 31. There are an immense number of feet, of 
 which, with few exceptions, only those of two or three 
 syllables need be noticed. Quintilian, differing from 
 Cicero, considers all others to be compound (9, 4, 79). 
 
 ART. 32. Pyrrhic of two short syllables, as: memor, 
 suus, meus, mala, bona, friior, cane, pater, jovis. Here 
 the difficulties consist in giving the highest pitch to the 
 first syllable, in letting the voice fall on the second, in 
 not running the vowel of the first syllable on to the 
 following consonant, and in placing the force or loudness 
 sometimes on the first, and sometimes on the second 
 syllable, when the latter ends in a consonant, without 
 lengthening either. This exercise requires great practice, 
 but it is fundamental. 
 
 ART. 33. Iamb, one short and one long syllable, as : 
 para, ama, aman-t, regun-t, siios, fores, plagas, dels, chori. 
 Here again the same difficulty with regard to the first 
 syllable occurs, and the length of the second will occasion 
 some trouble to pronounce it without stress. Recollect : 
 "scatter her enemies," in God save the Queen, where 
 "scat*errr" is sung, with the first syllable very short 
 yet with force, and the second syllable very long and yet 
 without force. In Latin, sometimes the first, sometimes 
 the second syllable had force, but the vowel was never 
 run on to the consonant, as in " scat'errr," so that re'gun-t 
 
24 METRICAL FEET. [Art. 3336. 
 
 must not fall into reg'itnt, which might be heard as a bad 
 reg-gun-t, and then would not be an iambus at all (p. 19, 
 note). 
 
 ART. 34. Trochee, or choreus, one long and one short 
 syllable, as: ro'ma, ve'sa, me'ta, pal-ma, ri'sus, mo'tus. 
 This is altogether an easy foot, especially as the force 
 need not be laid on the last syllable. 
 
 ART. 35. Spondee, two long syllables, as : re'ges, 
 re'gi, vir-tus, vdn-tos, mtil-tos, cur-run-t. The difficulty 
 here is to keep both syllables long, and to practise 
 giving the force to either syllable without losing the 
 length or high pitch of the other. There will be a diffi- 
 culty in keeping the spondee distinct from the iambus 
 and trochee. Decline an adjective as : mag-nus, mag-na, 
 raag-ni, mag-no, mag-nos ; and : citus, cita, citi, cito, citos, 
 and mark the differences of length clearly, avoiding the 
 common school trick of putting great force on the 
 variable final syllable and making the preceding syllables 
 almost inaudible, (see Art. 112). Remember that in: 
 mag-na, mag-nse, &c. the first syllable is fully as long as 
 the second, and that although in: cita, citae, the first 
 syllable is short, it has the highest pitch and is not at all 
 obscured or hurried over. This exercise is really very 
 difficult to English pupils. Such English words as : turn- 
 pike, muletrack, primrose, highway, tollbar, penknife, 
 made to follow the phrase "we can see the " in an 
 affirmative (not interrogative) sentence, may help to give 
 Englishmen some conception of a Latin spondee both 
 in quantity and pitch accent. 
 
 ART. 36. Dactyle, one long and two short, as : ver- 
 tere, rum-pere, cae-saris, l'n sula. The difficulty here is 
 to prevent the last syllable from becoming long, as in : I'n- 
 
Art. 3638.] METRICAL FEET. 25 
 
 sulas, forming a cretic, (also called amphimacrus) or one 
 short between two long. 
 
 ART. 37. Anapest, two short and one long, as : pdpuli, 
 memores, propera. Great care has to be taken not to 
 make the first syllable long, and when the force is given 
 to the first syllable not to make it long by running the 
 first vowel on to the following consonant, (giving a cretic 
 again,) or not to do this and also not to make the last 
 syllable short, converting the anapest into a dactyle ; or 
 not to make all three short. Great care is necessary to 
 distinguish po'pulus, populus, po'pull, populi, memoris, 
 memores. 
 
 Feet of Three and Four Syllables. 
 
 ART. 38. Molossus, three long, as : in-gen-tes, In-fan- 
 dos, sub-mi t-tun-t, and Choriamb, (that is choreus and 
 iambus), two short between two long, as : op-positls, myr- 
 midones. These two feet occasion great difficulty to 
 keep them clear and distinct Their length is the 
 same, but their rhythmic effect, which depends not 
 merely on the length but on the number of syllables, 1 is 
 very different. As these feet are often distributed among 
 different words, the position of the highest pitch varies 
 considerably. If the force coincides with the beat of 
 the line, then in both molossus and choriamb, the first and 
 last syllable have generally the stress, and in putting it 
 there care must be taken not to make the second syllable 
 of the molossus short. The usual bad habit is to read 
 a molossus as one long between two short making instantis 
 in position (K. 3) like aml'ca, which is an amphibrach, 
 or one long between two short. 
 
 1 Cicero (Orator 194) quoting from Aristotle, says : "Ephorus vero ne spon- 
 deum quidem, quern fugit, intelligit esse aequalem dactylo, quern probat [Aris- 
 toteles], syllabls enim metiendOs pedes, non intervallls existimat." 
 
26 VERSE RHYTHM. [Art. 39, 40. 
 
 ART. 39. The following examples of all the feet 
 which are usually spoken of by name are taken from 
 p. 164 of W. Ramsay's "Elementary Treatise on Latin 
 Prosody; (Glasgow and London, 1837, pp. 304), which 
 is extremely useful for the numerous examples it con- 
 tains, each with an exact reference : 
 
 DISSYLLABIC. Pyrrhichius casa, Spondaeus reges, Tro- 
 chaeus Roma, Iambus parens. 
 
 TRISYLLABIC. Tribrachys anima, Molossus Roman!, 
 Dactylus car-mina, Anapaestus populos, Amphibrachys 
 arnica, Amphimacer ap-pull, Bacchius can tare, Anti- 
 bacchius catones (palimbacchius in Quintilian 9, 4, 82). 
 
 QUADRISYLLABIC. Proceleusmaticus habilior, Dispon- 
 daeus Maecenases, Choriambus Romulidae, Antispastus 
 Clytem-nestra, Dliambus Corinthii, Dltrochaeus or Dlcho- 
 raeus (P. i) dimicare, lonicus a majore Lavlnia, lonicus 
 a minore Diomedes, Epitritus primus venenatis, E. se- 
 cundus con-ditores, E. tertius heroic!, E. quartus in- 
 vltamus, Paeonius primus Caecilius, P. secundus Horatius, 
 P. tertius Menedemus, P. quartus profugien-s. 
 
 V. Elementary Notions of Verse Rhythm with 
 both Accent and Quantity Hexameters. 
 
 ART. 40. After a feeling for the rhythm of these feet 
 has been produced, we must proceed to verse treating 
 at first the so-called elided syllables, (of which the precise 
 nature is to be considered afterwards,) as absolutely non- 
 existent. Thus in the examples in the Appendix, where- 
 ver w is used, skip the vowel before it entirely, and where 
 a small m is also written, skip that as well. Where a regular 
 final m is used with a hyphen after it, read it like the 
 
Art. 40, 41.] VERSE RHYTHM. 27 
 
 next following consonant. The object is not yet to teach 
 how to read verse, but only how to understand the action j 
 of feet in producing rhythm. Hence we reduce a verse \ 
 to a mere skeleton of sound, independent of sense and 
 of rhetorical alterations of sound. These are the muscles 
 and nerves to be laid on afterwards. 
 
 ART. 41. But the place of the raised pitch must be 
 strictly observed, and for this purpose the verses had 
 better be first read in a kind of sing-song, the high 1 
 pitched syllables being all of one pitch and the low I 
 pitched syllables being all of one pitch also, but about 
 a musical "fifth" lower than the other, as if the latter 
 were sung to the lowest note of the fourth string of 
 a violin, and the former were sung to the lowest note of 
 its third string. 1 
 
 1 Regarding the musical nature of Greek accent, there is a most instructive 
 passage in Dionysiusof Halicarnassus, Trepl crut'^ecrews ovo/j,a.T(t)v, Chap, xi., which, 
 on account of its musical technicalities, I here annex in English, giving the prin- 
 cipal Greek expressions in parentheses. This writer, who was born between B.C. 
 73 and 54, and died soon after B.C. 7, lived 22 years in Rome, where he probably 
 taught as a rhetorician. He was no doubt acquainted with the best literary men 
 of the Augustan epoch, and though his remarks apply to Greek and not to Latin, 
 there is no reason to doubt that the meaning of acute, grave, and circumflex 
 accent, in Latin, coincided with r) d^cta, >) /Sapeta, and r\ Trepio-Trw/xen? Trpoo-wfita in 
 Greek, of which the Latin words were mere translations. The two first words were 
 however universally applied in Greek to raising and lowering pitch in music. The 
 following passage serves to shew that not only the names, bjat the things signified 
 were identical : " Music (MOVCTIKT}), and the science of pubUc speaking (ij TUV TroAt- 
 TIKMV \6yaiv eTrtcrTTj/ar/), differed from that used in songs and on instruments (r7s v 
 (uSais /cat opyaVots) in quantity (TO> 7roo-<2), not in quality (TO) TrotoJ). For in the 
 latter [public speaking] words (Aeets) have also melody (/xe'Aos), rhythm (pvdf*.bv), 
 modulation (/xeTajSoArJi') and propriety ( rrpenov). In speaking, then, also, the ear 
 is delighted with the melody, is impelled by the rhythm, welcomes the modu- 
 lations, and especially longs for propriety (7ro0et 6' 771 TTOLVT^V TO oc<cetov). The 
 difference is merely one of degree (*cara TO /LtaAAoi/ /cat TJTTOV). The melody of 
 speech, then, (StaAeKTOv /aeAos), is measured by a single musical interval (evi 
 Siao-Ti^om) which is as nearly as possible (oJs e-yytcrra), that called a Fifth 
 (TO) Aeyo/u-eVo) Stci Trei/Te). It does not rise in pitch (eTrtTCtVeTat evrt TO 6u) beyond 
 three Tones and a half (irepa r<Zv jptcoi/ rovuv /cat ij^uTOvt'ov, which is the measure 
 of a Fifth in the musical scale as from C to G), nor is it depressed in pitch 
 (ai/ierat ^TT! TO /3api) more than this amount (TOU x^ptou TrAetoi/). But the whole 
 
28 VERSE RHYTHM. [Art. 41. 
 
 The place of raised pitch is not marked, except in the 
 first example A, but may be easily found from the 
 
 of a word (anaaa 77 Ae'fts), considered as made up of parts (17 naff ei> fj.6pt.ov Adyov 
 raTTO/aeVij), is not spoken (Aeyercu) at the same pitch (rrjs avr^s raVeoj?), but one 
 part in an acute pitch (oeias), one in a grave pitch (jSapetas), and another in both 
 pitches (a/x$oti/, of course, successively). Those words of one syllable which have 
 both pitches, have a low pitch imperceptibly blended with the high (avve^Bap- 
 Hevov c^ovo-i T<*> 6et TO /3apu, the word implies a mixture by the dying off of one 
 into the other ; the low is probably placed first because it was the longer and 
 final effect which grew, as it were, out of the high), and these we call " cir- 
 cumflexed" (7repto-7rw|u.eVa?). But those words which have pitch upon different 
 syllables separately (/ erepo) re /cat ere'pco xwpts) keep its proper nature (TI]V 
 oiKelav <f>vaiv} for each. In dissyllables there is nothing intermediate between 
 high pitch and low pitch (ovSev TO Sta (j.e<rov x <a P^ ov /SapvTrjTos /cat ofvTijTos). But 
 in polysyllabic words, of all kinds, there is but one syllable which has the high 
 pitch among many which have the low pitch (r; TOV bvv rovov e\ovaa /u.ta ev 
 TroAAat? jSapeiats eveanv). On the other hand, the music of song and of instru- 
 ments uses a greater number of intervals (Stcurnj'/u.acri ir\eioaiv), and not only 
 the Fifth, but beginning with the Octave (curb TOU did iraa<Zv c*pa/u,eVj), it per- 
 forms (/aeAwfiet) the Fifth, the Fourth (TO Sid Tecraapajf), the whole Tone (TO 
 didrovov) and the Semitone (TO yfjuroviov), and as some think, even the Quarter- 
 tone audibly (rrjv Sieaiv atcr^rjTw?, the diesis was originally the same as a semitone, 
 whence the modern French diese for sharp, but the word was afterwards used 
 for the later enharmonic division of the semitone, so that the interval B to C 
 had an nserted transitional sound, which we may write 6B, thus giving B 
 6B C, where the lengths of the string sounding the notes were in the proportion 
 32- 31 30; the English Quartertone is a very loose term, but it is generally 
 used n this case ; see the Greek Scale explained in Smith's Classical Dictionary, 
 and more fully in Helmholtz's Tonempfindungen, Chap, xiv., of which my English 
 Translation is in the press). But this [vocal and instrumental] music does not 
 hesitate (au) to subordinate words to the air (TO.? Ae'ei? TOI? jueAecrtv vnroTaTTei^), 
 instead of the air to the words, [as in the present day, but let us hope that this 
 did not happen till the degenerate period of Greek music in which Dionysius livedj. 
 This is especially evident, among many others, in the airs of Euripides, which he 
 has made Electra sing when speaking to the chorus in his " Orestes" [v. 140 3]. 
 
 alya, alya, Aev/cbv i^yo? a'p/SuAij? 
 
 TiOeire, /utry KTVirelre 
 
 aTroTrpo/Sa/r' e/ceiV ciTroTrpofli KOITO.S. 
 
 [The lines are here printed from the t^xt of Tauchnitz's stereotyped edition ol 
 Dion. Hal., and this, it will be seen, was clearly his reading. But Dindorf 's text 
 
 alya, alya, \eirrw i^^ 05 ap/SvArj? 
 
 Ti'0ere, /txr) i^o^eire, jixjj VTW KTUTTO?. 
 
 ctTTOTrpb /SctT* e/cetcr' dnoTrpo /not /cotVa?. 
 
 These are the first lines of lyric poetry in the tragedy. It is not to be supposed 
 that the music here criticised is that used when the play was first performed B c. 
 408, that is, more than three centuries before these remarks were written, when 
 Greek music had probably different rules] . In these lines the words utyo. alya 
 
Art. 41.] VERSE RHYTHM. 29 
 
 quantity. It is never on the last syllable, except of course 
 when the word has been artificially shortened for these 
 exercises by omitting the vowel before ^. If the last 
 syllable but one is long, it is spoken with a raised pitch, 
 which is maintained throughout if its vowel is short, as : 
 ven-tus, ven-tos, or if the last syllable is long, as : fa'mae ; 
 but sinks immediately, if its own vowel \ long, and at the 
 same time the vowel of the last syllable short, as : fama, 
 to be distinguished from : fa/ma. When the word qne is 
 added on, the preceding syllable even if short has 
 a raised pitch (not greater force) and this in the Ap- 
 pendix is marked in all the examples, as : metaque* 
 because it is constantly forgotten by English readers. 1 
 
 AevKoi/ are set (/ueXtuSetTcu) to a single note (<' ei>bs <>0oyyov) although each of the 
 three words has both high and low pitches (jSapeta? re raVeis e\et /cat ofetas, 
 observe that raVei? is here used for Trpoo-wStas). And the word apjSuArj?, has the 
 third syllable of the same pitch as the second (kit\ /UCOTJ <ruAAa/3r} rrjv TpCrrjv oju.6- 
 rovov ?x e O though it is impossible (dfutf)(avw ovros) that one word (ev ofojma) 
 should have two high pitches. In the word Tifleire although the first syllable is 
 made lower (/Sapurepa), the two that follow have both the same high pitch (bi-vrovoL 
 re KCU 6ju,6$wi/oi). The circumflex (6 7repto"7racrju,bs) has vanished <>j</>ataoTai) from 
 KrvTreire, for the two [last ?] syllables are spoken at the same pitch (/otta raVei). 
 And aVoTrpojSaT' does not receive the acute accent belonging to its middle syllable 
 (TY)V r>7? /me'orjs o-vAKajSrys TrpocrwSiai/ 6eiai>, where observe the use of the word 
 TrpocrwSCav accent as synonymous with rao-is, pitch, or degree of tightening of 
 the string), but the pitch of the third (v) Ta<rtf TTJ? Tptrrjs, here again TOCTI? is 
 used for 7rpoo-a>5ta), has descended (KarajSe'jSrjKei/) to the fourth syllable. Rhythms 
 are treated in the same manner. For prose neither forces nor interchanges the 
 lengths of any noun or verb (r) /uter -yap TrecJV) Ae'is ov$ei/6? OVT' bvo/jLaTOS, ovre 
 pjfjuiaTO? /Sta^erat TOVC \p6vovs, ov6e /aetari^Tjcrtv), but preserves long and short 
 syllables as it has received them by nature (a'AA' ot'a? Trapei'Arj^e TT) ^>ucret ra? 
 truAAa/3as ra's re /ixawpa? Kai ras jSpaxeia?, rotavra? ^vAaTTet). Yet rhythmical and 
 musical art change them, shortening and lengthening (/aeiovam ical auou<rai), till 
 they are often reversed (wore 7roAAcx*fi? ei? rdvavTia /u,Taxwpeii/), for they do not 
 rectify the times by the syllables, but the syllables by the times (ou yap rais crvA- 
 AajSais dnevQuvovai TOV? xP vov *> <*^<*- TO?? xpo*' 01 ? Ta? o-vAAa/3as)." The great 
 value and importance of this passage, which is seldom referred to, its explicit 
 identification of Greek accent with pitch, and its clear assertion of the strict 
 observance of quantity in prose, have induced me to give it at length. 
 
 1 Mr. Roby says (Gram. 2nd ed. p. Ixxxiii.) : "I confess to entertaining some 
 doubts as to a short syllable, when followed by an enclitic, receiving the accent, 
 
30 VERSE RHYTHM. [Art. 41. 
 
 In other cases if the last syllable but one is short, what- 
 ever be the length of the adjacent syllables, the last 
 
 e.g. primdque. As the Romans would not have accented pr^maq^le on the 
 penult, if it had been one word, I do not see why the z should have lost the 
 accent by the addition of the enclitic." But shifting the place of the accent was 
 to be expected in order to shew that primaque is not one word, and probably 
 Prl'maque might have puzzled a Roman at first, just as some of our English 
 accents puzzle the French (and conversely). Prof. Key (Tr. Ph. Soc. 1873 4, 
 p. 50, note) says : "I have ventured to place an accent on the first syllable of 
 vl'taque, [in the last line of the Aeneid] rather than on the second, because I 
 utterly reject the doctrine of the grammarians, who contend for mtdque, and who 
 seem in this respect to have obtained the consent of Mr. Munro, see his words : 
 The enclitics qne, ne, ne t attract the accent to the syllable (word ?) immediately 
 preceding, whether long or short, annaque as well as armi'sqne. Thus in p. 389 
 he gives us Lavlnaque, Tiberinaque ; and in p. 390, templaque mentis." Now 
 no one disputes the shifting of accent in armis armi'sque, because the last is re- 
 garded as a single word, and then the law of accent would require the shifting of 
 accent as in vdlucres voluc-res, (p. 17, note,) and as is quite common in ordinary 
 inflections, as Cfcero Cicero'nis Ciceroni. On the modern force accent principle 
 the thing is not impossible, for it occurs in modern music but rarely, and then 
 only to produce a peculiar effect, as when in six-eight time the stress is laid on 
 the second and fifth notes. On the other hand, on the theory of pitch, the shifting 
 of the accent on to a short syllable (which is equivalent to making the middle note 
 of a triplet the highest), is very common indeed, and may be found constantly in 
 the Duet AlF idea in the JBarbiere, thus the loth and nth bars have the lowest, 
 or a descending note in the middle of the triplets, while the i2th bar has three 
 examples of the highest note in the middle of a triplet, as (in the bass), using s 
 for sharp : e a gs, a c' b, c' e' d'.v, e' c' a, g. I presume that, as in English we 
 naturally glide the vowel on to the following consonant when we give it force, 
 Prof. Key and Mr. Roby said : metac'we, vitac'we, to the utter destruction of 
 any remnant of metre left by putting force accents for pitch accents in the other 
 words, and that both (as I have personally heard in the case of Prof. Key), 
 regulated the rhythm to their ear by force accents, instead of quantities. .The 
 result of so doing showed itself clearly in the third century, even in Italy, see 
 Art. 113 ; it had nothing to do with Augustan Latin. There is absolutely no 
 difficulty to English organs in saying: me'taque, vl'taque, and it is a good exer- 
 cise to repeat such a combination many times in succession, beginning slowly, and 
 increasing speed gradually. Observe also that Latin prepositions, &c. added on 
 to following words (as in Quintilian's circumlFtora in the next note), did not 
 change the position of the accent, because they were not proclitics, but did form 
 a single word with the following, whereas the shifting of the accent for enclitics 
 pointed out the double character of the word. Observe also that the law of 
 accent in case of enclitics applies to Greek as well as Latin, and that in Greek it 
 even allowed two syllables with raised pitch in one word, as o-a/xaTa re, or trw/aa re 
 This was not possible in Latin, where omniaque vltaque would be said, the first 
 raised pitch becoming lost. But the principle is the same, and whatever theory 
 is thrown out must apply to both cases. The theory that accent consisted simply 
 
Art. 41.] VERSE RHYTHM. 31 
 
 syllable but two has the raised pitch which is maintained 
 throughout, as : Insula, insulas, riim-pere, Irmite, H'mites, 
 solida. Monosyllables have a raised pitch, as : tune, ars, 
 which, if the vowel is long, falls immediately, before the 
 vowel is concluded, as : hi. But prepositions, relatives, 
 and some unimportant vocables, as <?/, aut, &c., acted 
 apparently as if they were parts of the following word 
 and have no raised pitch. In (A.), the raised pitch is 
 marked by an acute, as : ad-sum, and the low pitch is 
 left unmarked, except where it occurs after the high pitch 
 on the same long vowel, and then the circumflex accent 
 marks both length and change of pitch, as : ora. Some 
 doubt attaches to the absolute generality of these rules. 
 But they are the only rules sanctioned by any persons 
 who were accustomed to hear the Latin we wish to 
 imitate, and consequently it is far better to err possibly 
 by carrying them out strictly, than to err certainly by 
 adopting any conjectures of later Latin grammarians, or 
 the notions of merely modern critics who have never 
 noticed pitch accents in their lives. 1 
 
 in altering pitch does so apply. In Italian when the pronouns, &c., are added on 
 to a verb, they do not change the place of the original force accent, as agitan'do, 
 agitan'domi, invia're inviar tele, anda'te anda'tevene, &c., although, as in the 
 last instance the accent is thrown back further than would be allowable on an 
 unincreased word. Hence Italian has no proper enclitics. I do not indeed know 
 of any real modern analogues. As regards merely the use of a raised pitch upon 
 a syllable of little force, while the preceding syllable is long and of a low pitch, 
 I need merely mention that this is precisely the way in which a Swede taught 
 me to pronounce skona, that is sko'nd, where a is a mere unemphatic inflectional 
 syllable as in German schone, which, except in accent, the word much resembles. 
 Yet this is the same as vi'ta(que), Greek oi>Se &c. 
 
 1 The classical rule is given by Quintilian : in om.nl voce acuta intra numerum 
 trium syllabarum continetur, slve hae sunt in verbo solae sive ultimae, et in his 
 aut proxima extremae aut ab ea tertia. Trium porro de quibus loquor, media 
 longa aut acuta aut flexa erit ; eodem loco brevis utique gravem habebit sonum, 
 ideoque positam ante se id est ab ultima tertiam acuet. Est autem in omm voce 
 utique acuta sed nunquam plus una nee unquam ultima ideoque in dlsyllabis prior. 
 
32 VERSE RHYTHM. [Art. 42, 43. 
 
 ART. 42. Begin with hexameters, because the time of 
 each foot is there most easily measured. (A, B, C, D,) 
 are examples of hexameters. The great difficulty to 
 contend with, on account of our English habits, is the 
 due expression of those long syllables which are not 
 under the beat of the verse, supposing that beat to fall 
 on the first syllable of the foot. 
 
 ART. 43. First take lines having four or five spondees, 
 as A. 24; B. 2; C. i, 2, 10 ; and D. i, 3, 4, 6, 12. 
 
 Praeterea nunquam in eadem flexa et acuta, quoniam eadem flexa et acuta, itaque 
 neutra claudet vocem Latlnam. Ea vero quae sunt syllabae unlus, erunt acuta 
 aut flexa, ne sit aliqua vox sine acuta (i, 5, 30. 31). But he had already given 
 an exception : cum dlco circum litora, tanquam unum enuntio dissimulate, dis- 
 tinctione [that is, speaking as if there were no separation of the words], itaque 
 tanquam in una voce una est acuta, quod. idem accidit in illo Trojae qul primus 
 ab orls (i, 5, 27). This last was taken to be Trojae'qtd or rather Trojae 'qui, with 
 a short vowel, and great increase of force, by Prof. T. Hewett Key, when I heard 
 him apply it. Those who know Prof. Key's views on Latin accent, as laid down 
 in his papers: A Partial Attempt to reconcile the Laws of Latin Rhythm with 
 those of Modern Languages (Trans. Philolog. S0c, 1868 9, pp 311 351), and : 
 Accent a guiding Principle not merely in old Comic Metres, but generally in Latin 
 Poetry, and first of Virgil's Hexameters (Ib. 1873 4, pp. 35 52), will see that I 
 hold altogether different opinions. I may mention that where Prof. Key writes 
 an acute accent I always heard him read with rather an exaggerated increase of 
 force, irrespective of pitch. For myself I think Quintilian meant: Tro'jae, see 
 also the passage from Quintilian quoted in (p. 17,) note With these strict inflexible 
 rules, Quintilian thus contrasts the Greek freedom : sed accentus quoque, cum 
 rigore quodam, turn similitudine ipsa, minus suaves habemus : quia ultima syllaba 
 nee acuta unquam excitatur [raised, excited], nee flexa circumducitur [waved], 
 sed in gravem vel duas graves cadit semper. [This is quite opposed to the later 
 Grammarians]. Itaque tanto est sermo Graecus Latino jucundior ut nostri 
 poetae, quotiens dulce carmen esse voluerint, illorum id nominibus exornent (12, 
 10, 33). The fact was Greek was a foreign language which Quintilian had some 
 difficulty in learning to pronounce, and he therefore esteemed its sweetness too 
 highly, just as the Englishman, who in the fourteenth century, wrote those dia- 
 logues in old French which M. Paul Meyer has reprinted (from Harl. MS. 3988, 
 in the Revue Critique for 1870), speaks of " doulz frangois, qu'est la plus bel et 
 la plus gracious language et plus noble parler, apres latin d'escole, qui soit ou 
 monde," with more in that strain. As the nature of the accent had probably 
 entirely altered before the fifth century, next to no weight is to be attached to 
 what the later grammarians say on the subject. They were nearly as incapable 
 of understanding the nature of pitch accents as we are. Prof. Key has justly 
 pointed out that they always speak of the laws of accent in the past tense 
 'habuit, not habet. (Tr. Ph. Soc. 18734, p. 36). 
 
Art. 43 49.] VERSE RHYTHM. 33 
 
 Read at first with the pendulum or tap of the finger. 
 Place stress at first on the first syllable of each spondee 
 and dactyle, afterwards vary it much, and see that it does 
 not disturb the quantitative rhythm. 
 
 ART. 44. Next take lines beginning with two 
 spondees, and hence with five long syllables, as A. 3, 6, 
 19; B. 5; C. 9. 
 
 ART. 45. Then lines with a spondee followed by 
 dactyle, as A, 4, 10, 14, 17, 18, 20,, 22 ; B. i, 3, 6 ; C. 5, 
 14, 16, 18; D. 8. 
 
 ART. 46. Then those which have a dactyle and 
 spondee, and hence begin with a choriambus, as A. i, 
 9, n, 13, 16; B. 8, 9; C. n, 12, 17 ; D. 5, 9. 
 
 ART. 47. Then those with two dactyles, as : A. 2, 5, 
 
 7, 8, 12, 22, 25 ; B. 7 ; C. 4, 6, 7, 8, 15 ; D. 2, 7, n. 
 ART. 48. Observe the difference of rhythmical effect 
 
 in each of the cases (Art. 43) to (Art. 47). Observe the 
 mode in which lines differ which begin with the same 
 feet, owing to the different division of the words, and 
 hence the different positions of the raised pitch among 
 the quantitative feet. 
 
 ART. 49. Observe particularly the effect of breaking 
 the third or fourth foot by a caesura, by means of which 
 a low pitch is secured for the beginning of the third or 
 fourth foot. Examples : Break of third foot A. i, 4, 5, 
 
 8, 10, ii, 13, 14, 15, 17, 18, 19, 21, 22, 24. Break of 
 fourth foot A. 2, 3, 6, 23. Half break, not at the end 
 of the first long syllable of the foot but at the first short 
 syllable, A. 9, causing a markedly different distribution of 
 the pitch of the voice. The case of A. 7, 12, 16, 20, 
 where the break occurs on a vowel which is omitted in 
 the present exercises, really belongs to this class. 
 
 D 
 
34 VERSE RHYTHM. [Art. 50, 51. 
 
 ART. 50. Observe the constant form of the final 
 cadence, almost always dactyle and spondee, with the 
 high pitch on the first syllable long of each. The only 
 exception in (A., B., C.,) is (B. 2). Observe the re- 
 markable effect of pitch in that line. This is very rare 
 in Virgil's poetry, and seems always to have been in- 
 troduced for a purpose. Only some 42 examples occur 
 in his works, but such cases are commoner in Lucretius. 
 Characteristic cases are 
 
 Turn variae illudunt pastes; saepe exigiuis mils. 
 
 Ceo. I. 181. 
 Illic, ut perhibent, aut intempesta silet nox. 
 
 Geo. I. 247. 
 Prima vel auctumnl sub frigora, quum rapidus sol. 
 
 Geo. n. 321. 
 Ipse ruit, dentesque Sabellicus exacuit sus. 
 
 Geo. in. 255. 
 Dat latus, insequitur cumulo praefuptus aquae mons. 
 
 Aeu. I. 105. 
 Vertitur interea coelum et riiit oceano nox. 
 
 Aen. ii. 249. 
 Sternitur exanimusque tremens procumbit hum! bos. 
 
 Aen. V. 481. 
 
 There can hardly be better illustrations of the 
 effect of pitch accent in Latin verse. The ordinary 
 method of reading produces hideously unrhythmical 
 results. 
 
 ART. 51. The following are Virgilian examples of 
 p, different cadence : 
 
 Cara deiim soboles magni Jovis incrementum 
 
 EC. iv. 49. 
 
 Aut le'ves ocreas lento diicunt argento. 
 
 Aen. VII. 634. 
 
Art. 5153.] SLURRED VOWELS. 35 
 
 And the two are united in : 
 
 Cum- patribus populoque, pena'tibus et magnis dis. 
 
 Aen. vill. 679. 
 
 ART. 52. Having thus attuned the ears of pupils to 
 quantitative rhythm as modified by variable pitch, we 
 can approach the consideration of the real treatment of 
 final vowels followed by other vowels, and of the final m 
 before vowels and consonants. 
 
 VI. Slurred Vowels. 
 
 ART. 53. Take the case of a vowel ending one word 
 and a vowel beginning the next, when the two vowels 
 are rhythmically reckoned together as one syllable. With 
 this case consider that where the second word begins 
 with H., because this H. never made position, and had 
 probably no hissing effect, so that it readily died out. 1 
 
 1 The chommoda hinsidias, and hwnios for commoda insidids and lonios, 
 mentioned in Catullus (Ixxxii. or Ixxxiv. ) may have had a iully hissed h, because 
 they were mere detects of utterance arising from an explosive manner with which 
 we are very familiar in English, especially from uneducated speakers who try to 
 do their bitter best (et tune mirifice sperabat se esse locutnm, Quum, quantum 
 Poterat, dixerat ^insidias), and of course Catullus had no means of expressing 
 the two varieties. In fact only one h occurs in writing, and phonologists are 
 driven to great straits to express the many varieties with which they are now 
 acquainted. (See my paper in the Academy for 17 January 1874, on a Physica 
 Theory of Aspiration). The words of Cicero and Quintilian, and the slurring of 
 vowels, which would be disturbed by the interposition of vocalised breath, lead 
 me to consider that the Augustan Latin h was merely a forcible, jerked utterance 
 of the following vowel, without any unvocalised breath, exactly as now in India 
 for the combinations bha, dha. gha, &c., produced in the way explained on p. 6 
 n. 2, Cicero says : quln ego ipse, cum sclrem ita majores locutos esse, ut nusquam 
 nisi in vocall adsplratione uterentur loquebar sic, ut pulcros cetegos triumpos 
 Kastdginem, dlcerem : aliquando, idque sero, convlcio aurium, cum extoita 
 rnihi veritas esset, usum loquendl populo concessl, scientiam mihi reservavi. 
 Orclvios tamen et matones, otones, Caepiones % sepulcra, coronas, lacrymds dici- 
 nuis, quia per aurium judicium semper licet (Cic. Or. 160). Quintilian writes : 
 
 D 2 
 
36 SLURRED VOWELS. [Art. 53, 
 
 Cicero (see Q.) objects strongly to open vowels, which 
 he considered to gape (Jiidre), or io meet with a shock 
 (concur sus, conjungere). Yet in his writings there are 
 constant cases of open vowels; thus in (Q.) itself: 
 legendo oculus (i), ne extremorum (2), si inconditis (5), 
 nemo ut (7), illae ipsae horridulae (io), qul ut (n), saepe 
 hiabant (n). It is evident, therefore, that these cases 
 were to him quite different from : sclpio invicte (12), and 
 etesiae in (13), where the open vowel made a distinct 
 syllable, and that it must have been this habit of allowing 
 open vowels to form a distinct syllable which he found 
 so offensive in Greek, and which he declared that Roman 
 ears could not endure if frequently repeated (Q. 6, 13), 
 and it must have been this separation which he contem- 
 plated when he said that no Roman was allowed to pull 
 his words asunder (distrahere voces, Q. 9). To Cicero 
 therefore two' vowels thus situated formed one syllable 
 in prose, as in verse. There is not even a hint that the 
 first vowel was dropped. Whenever this occurred in 
 Greek the vowel was not written, a habit followed by 
 
 ilia vero nonnisi aure exiguntur, quae flunt per sonos ; quanquam per asplra- 
 tionem, slve adjicitur vitiose sive detrahitur, apud nos potest quaeri an in scripts 
 sit vitium ; si H litterae st, non nota. ['* If H is a letter," i.e. represents a separate 
 sound, and "not a mark," i.e. represents an initial modification of sound ; this is 
 I believe the real distinction meant by those many orthographers who since 
 Quintilian's time have disputed whether H is or is not a letter}. Cujus quidem 
 ratio mutata cum temporibus est saepius. Parcissime ea veteres us! etiam in 
 vocalibus, cum oedos ircdsgue dlcebant, difi deinde servatum, ne consonantibus 
 asplrarent, ut in Graecls et in triumpis, erupit brevl tempore nimius usus, ut 
 choronae, ckenturwnes, praechones, adhuc quibusdam inscriptionibus maneant, 
 qua de re Catulli nobile epigramma est [just cited]. Inde durat ad nos usque 
 vehementer et comprehendere et miki, nam mehe quoque pro vie apud antlquos 
 tragoediarum praecipue scriptores in veteribus librls invenimus (i. 5, 19 21). 
 It is evident that Quintilian's h was very small indeed, the precursor of its 
 French, Italian, and Spanish evanescence, where it is really merely a diacritical 
 sign, or, as we shall find Quintilian says of w, merely, inter duas vocales velut 
 nota est, ne ipsae coeant (9, 4, 40). 
 
Art. 53, 54.] SLURRED VOWELS. 37 
 
 modern writers generally in Italian, German, and English. 1 
 French, however, does not cut out the mute vowel except 
 in monosyllables, because of a general rule of pronun- 
 ciation. In French verse, except in very few cases, no 
 open vowels at the end of words, not even open nasalised 
 vowels are permitted. In the middle of a word, how- 
 ever, open vowels occur in French, making two syl- 
 lables, and this was also the case in Latin ; compare : 
 conticu-ere (A. i), aene-as (A. 2), eru-erint (A. 5), fu-i 
 (A. 6), incipi-am (A. 13), e-a (A. 17), pri-aml (A. 22), 
 stati-6 (A. 23), abi-isse peti-isse (A. 25), and frequently. 
 But: svadent (A. 9), abjete (A. 16), were exceptions, 
 which I have indicated by using consonantal forms /, v, 
 without being certain that the true consonants were 
 spoken. There was no necessity therefore for a Latin 
 tongue to connect concurrent vowels into one syllable, 
 but it habitually did so when the two vowels belonged 
 to different words. It was a habit, not a necessity of 
 speech, but a habit on which versification reposed. 
 
 ART. 54. Now in Spanish and Italian, the two 
 Romance languages most like the Latin, we find the same 
 ability to separate vowels internally, and the same habit 
 of connecting them between words. In (S.), containing 
 the two first stanzas of Tasso's Jerusalemme, 28 instances 
 of vowels thus connected (marked by J) occur in 16 
 lines. In (S. 3) : senno^, e, this connection occurs with a 
 pause after the first vowel, and in (S. 5) : s'oppose^, ejn- 
 vano, not only is there a pause after the first vowel, but 
 
 1 In older English when the final e was still a distinct syllable before a con- 
 sonant, there is reason to believe that it was entirely omitted before a following 
 vowel, and generally even before a following h, and although it was still re- 
 tained in writing, as indeed it still is in most cases, it has long ceased to be 
 sounded in any way. This is not the case even in modern high German. 
 
38 SLURRED VOWELS. [Art. 5456. 
 
 there are three vowels, e -e -i. reckoned to form one 
 syllable. On the other hand there are real omissions of 
 vowel before vowel in: l(e) armi, e(i)l (i), che (i)l (2), 
 l(o) inferno, s(i) oppose (5), d(i) asia (6), s(e) adorno 
 (15), d(i) altri (16). But the principal omissions are 
 before consonants, as gran(de) sepolcro (2), popol(o) 
 misto (6), ciel(o) gli (7). And there are other omissions 
 so common as to be unwritten, as, e for ed, a for ad, col 
 for collo which again stands for con lo. 
 
 ART. 55. Now it is easy to hear what Italians actually 
 say in verse, and especially in singing, on account of our 
 Italian operas, but it is necessary to distinguish between 
 an Italian singer and a singer of Italian, who is frequently 
 a foreigner, and hence of no authority for pronunciation. 
 It will be found that such singers let all the written 
 vowels be heard, but yet bring them on to one single 
 musical note, which may be itself very short. The 
 following are cases from well-known pieces of music, 
 (technically called "numbers") in the Nozze di Figaro. 
 The hyphen is used to shew the vowels forming one 
 syllable in two words, and the superior figures shew the 
 relative lengths of the syllables as indicated by the 
 music, and as invariably observed by the native singers. 
 
 ART. 56. Nozze di Figaro, No. 2, se-a 1 ca 2 so T , che 1 
 vuol 2 mi-il 1 pa I dro T ne 1 , se-u'dir* brasmi-il 1 res 2 to 2 , di T scac 2 - 
 cia-i 1 sos I pet 2 ti 1 . But with the pause between two lines 
 of verse : se-il 1 mat I ti 2 no T il 2 ca 2 ro T , il 1 ca 2 ro j il 1 ca 2 ro x . 
 
 No. 3, very quick time . Tar 2 te-a I do I pran 2 do 2 . 
 
 No. 4, quick : e-un3 pia*cer 2 ser 2 ba3to-ai r sag'gi 1 . 
 
 No. 7, quick, to shew the effect of the pause : (sepa- 
 rated) tos3to-an I da 2 te 2 e3 scac3cia 2 te 2 il 2 se3dut r tor 2 ; 
 (connected) tos 6 to-an 2 da 4 te-e 3 scac'ciaste-il 1 se 3 dut T tor 2 . 
 
Art. 56-58.] SLURRED VOWELS. 39 
 
 No. 9, non3 piu-anMrai* fai^faHo* ne-a3mo T ro 4 so 4 . 
 No. 12, ve 2 ni3te-in I gi3noc 1 chia I te T vi 2 . 
 ART 57. This practice seems to have preserved 
 classical traditions better than the medieval Latin hymns, 
 which allow open vowels to form syllables, and know 
 no quantity, as in the following prayer of St. Bernhard 
 (H. A. Daniel's Thesaurus Hymnologicus, vol. iv. p. 228, 
 Leipzig 1855), where the '"marks open vowels, the () the 
 louder syllables. For- the pronunciation see (Art. 114). 
 
 *~V . x*. 
 
 Dilata're* aperi're A^J 
 
 Tarrquam ro'sa fra* grans mi're A 
 
 Cor'di me*o te conjuirge ^ >^V 
 
 Un'gue* il'lud et compuirge ' ,O^ S^ *^ 
 
 Qui* a "mat te', quid pa'titur?,^ 
 
 Vi'va cor'dis vo'ce cla'mo, ^^V ^*V ^^ / 
 
 Dul'ce cor, te nam'que* a 'mo ^^ /S^ ^^ 
 Ad te* o'ret, ad te plo'ret ^^ ^ 
 
 Te* ado "ret, te* hono'ret. ' 
 
 And so on. All of this, if it could fefe.r?a!d 
 
 Jr 
 
 Augustan Roman according to his own habits, wouj^f 
 sound horrible to him, and shew a most monstrous 
 ignorance of versification. 
 
 ART. 58. The practical rules hence deduced for Latin, 
 are as follows : 
 
 a. When a vowel ends one word, and a vowel, (pre- 
 ceded or not by H,) begins the next, pronounce both 
 vowels, quite distinctly and audibly. 
 
 b. When there is no pause between the words, run 
 on the vowels closely together, and make the time 
 occupied by the two sets of vowels in the one syllable 
 no- longer than is required by the laws of the verse. 
 The length of both will therefore have to be altered. 
 
40 SLURRED VOWELS. [Art. 58, 59. 
 
 Thus syllabise (B. 8) hae 2 -ti r -bie I -run 2 '-tar 2 -tes 2 -pa 2 -cis 2 - 
 queim 2 - poMie'-re 1 - m6 2 -re 2 , so that bie has only half 
 the length of queim. Similarly : Pal 2 las 2 tehoc 2 vul 2 - 
 ne T re z , PaHas 2 , AEN. xii. 948. 
 
 c. When there is a pause between the two words, 
 the conjunction of the two vowels, and the accuracy 
 of the time, becomes (like the Italian syllable) rather 
 a matter of " faith " or practised acknowledgment, 
 than of real audition, but generally one of the two 
 syllables should be unduly shortened to indicate the 
 effect. Thus (A. 25) nos 2 a I bi I is 2 se 1 ra r (tl, et) 2 ven 2 to 2 , 
 &c., ti, may be i% and et only \ or tl being nearly 
 fully 2, et will be nearly evanescent, (D. 8) m6 2 men 2 
 (to aut) 2 ci'ta 1 mors 2 , the to and aut may each be 
 only i. But in all cases the effect of long and short 
 vowels and diphthongs would be indicated by a prac- 
 tised speaker, and was no doubt felt by the poet. 
 As this practice is similar to the slurring of notes in 
 music, I call it by the same name. See also (Art. 61). 
 
 ART. 59. It is evident that when two words are 
 closely united by slurring their final and initial vowels, 
 they might be difficult to separate by the ear. The 
 mode in which words are connected in all languages 
 causes a difficulty of the same kind to foreigners, which 
 no native feels when he hears them spoken, because he 
 would feel any other way of pronouncing them in a 
 phrase so " unnatural," that is unusual, that he might fail 
 to catch the sense (p. 19). In modern writing it is 
 customary to assist the eye by separating the words 
 without indicating the mode of connection. In Sanscrit 
 however the mode of connection is always written, and 
 the native commentators have laid down rules for discover- 
 
Art. 59-61. ] SLURRED VOWELS. 41 
 
 ing the separate from the connected form. In some lan- 
 guages as English and French, the position of the greatest 
 force varies under different forms of combination, and 
 hence great difficulties arise to foreigners when they hear 
 the languages spoken, which entirely vanish when they 
 see them written. 
 
 ART. 60. In Latin, however, the strict laws of the 
 position of the highest pitch must have clearly separated 
 the words, however closely they were run together. Every 
 word (with the exceptions already noted) had one syllable 
 spoken in a higher pitch than the rest, and only one, 
 and that syllable was never the last (in polysyllables) 
 and never further off than the last but two (Art. 41). 
 Hence con-ti-cu-e-reom-nes (A. i), re-ge-reim-pe-rio (B. 6), 
 pro-pi-ahaec (C. 4, where ae marks a circumflex on a 
 diphthongal form, or digraph), fa'-taas-pera (C. 15) &c. 
 were clearly separated as two words by their two raised 
 pitches. Observe the combinations -reom-, -taas- in the 
 first and last instances. Here the first vowel is in a low 
 pitch, and the second in a high one, so that we might 
 write -reom-, -taas-, the effect being a wave up, or a re- 
 versed circumflex, which is a wave down (haec). In 
 some cases, where the word on to which the slur was 
 made had a low pitch, this would appear to fail ; but in 
 reality such a word was habitually treated as part of the 
 next, thus : deserto in li'tore (A. 24) were really felt to be 
 divided : deserto inll'tore, that is, they were treated as 
 two words, not three. When slurred they became : de- 
 ser-toin-lf-to-re. 
 
 ART. 6 1. Perhaps then we may supplement the rules 
 in (Art. 58) by the following practical usages. 
 
 If the first of the two slurred vowels is short and also 
 
42 SLURRED VOWELS. [Art. 61, 62. 
 
 low in pitch, make it still shorter, so that the full length 
 of both vowels makes up only the un slurred length of the 
 second vowel, thus in quae-queip-se (A. 5), ei takes up 
 the time of one ordinary short vowel only. 
 
 If the first vowel is long, whether high or low in pitch, 
 and the second short and low in pitch, the second vowel 
 is most shortened and may almost entirely vanish, as de- 
 ser-toin-lf-tor-re (A. 24). But if the first vowel is low 
 in pitch and the second high in pitch, the second vowel 
 though short must be made long enough to bring out the 
 effect of its higher pitch, as : 
 
 Quam lepide lexeis compos-taePut tesseru-laeom-nes, 
 cited from Lucllius by Cicero (Or. 149), where lexeis 
 may be read le'xls. 1 This is probably a rare case, and 
 no example of it occurs in the Appendix below. 
 
 If both vowels are long, perhaps that of highest pitch 
 might have had the greatest length in the slur as, sub 
 anti'-qual'-lice (L. 9). 
 
 ART. 62. If Cicero is not exaggerating (Q. 8) similar 
 usages must have been even more strictly observed in 
 
 1 Of course it is not easy to say what were the sounds of pre-Augustan ei, oi, 
 at, u>7, ou, and I do not enter upon the question here. But Quintilian seems to 
 have considered ei to be a mere digraph for J, and the passage is noteworthy 
 because it also seems to shew that he pronounced the Greek e< as z also. Semi- 
 vocales geminare diu non fuit usitatissiml moris, atque e contrario usque ad 
 Accium [born B c. 170, and lived to a great age, so that Cicero as a young man 
 conversed with him, if the words of Cic. Br. 107, are to be taken in this sense], 
 et ultra porrectas syllabas gemims, ut dixf [referring to I. 4, 10, see (Art. 88) 
 note], v5calibus scripserunt. Diutius diiravit, ut EI jungendls eadem ratione 
 qua Graecl ei uterentur ; ea casibus numerlsque discreta est, ut Lucllius prae- 
 cipit: Jam piie-rei venere, E postreimun facito atque /, Ut puert pluresflant ; 
 ac deinceps Idem: Menddcl jurlqrie addes JE, cutn darefurei Jiisseris. Quod 
 quidem cum supervacuum est, quia I tarn longae quam brevis naturam habet, 
 turn incommodum aliquando. Nam in ils, quae proxima ab ultima litteram E 
 habebunt et I longa terminabantur, illam rationem sequentes utemur E gemina, 
 qualia sunt haec aursl, argentel et similia (i. 7, 1416). We shall therefore be 
 as right as Quintilian (though he may have been wrong), if we pronounce Latin ei 
 and Greek e< as I. 
 
Art. 6264.] FINAL M. 43 
 
 prose. Of course exceptions occurred in actual usage, 
 but in our present state of ignorance we cannot do 
 better than strictly carry out an intelligible rule which 
 most probably held in the great variety of cases. 
 
 ART. 63. The next step is therefore to read the 
 examples in the Appendix in the same sing-song manner 
 as before, with the. same strict regard to quantity, but in 
 place of leaving out the vowels preceding w , slurring 
 them on to the following vowels, paying great attention 
 to the alteration of pitch. The little m between two 
 vowels must be entirely neglected, the vowels being 
 slurred as if it did not exist, but the m- final must at 
 present be pronounced as the following consonant. The 
 proper treatment of this last case forms the next step. 
 
 VII. Treatment of Final M. 
 
 ART. 64. The two facts to be accounted for are, that 
 in all Augustan verse (and consequently in all literary 
 Latin verse of a later period, because it is a mere imita- 
 tion of the Augustan) : 
 
 a. Final m, did not prevent the preceding vowel of 
 its own word, and the following vowel of the next, 
 from being reckoned as one syllable, precisely as 
 would have been the case if no m had intervened. 
 
 b. Final m, followed by a word beginning with a con- 
 sonant, invariably " made position," that is, made the 
 syllable which it terminated long. 
 
 Both cases might occur to the same word, even when 
 a monosyllable, and in the same line, as to dum in 
 
 Jam satis est ! dum aes exigitur dum mula ligatur. 
 
 HOR. Sat. I. 5, 13. 
 
44 FINAL M. [Art. 65, 66. 
 
 ART. 65. The conclusion is inevitable that m had 
 a different effect in the two positions, and that if the 
 letter ;;/ were preserved in writing for both cases it was 
 simply from etymological reasons, to assist the eye, the 
 alteration of sound proceeding by a rule known to all 
 Augustan Romans, though a matter of difficult inquiry 
 2000 years afterwards. We have several similar in- 
 stances in Quintilian (i. 7,) such as obtinuit written for 
 opinuit said, and inmunis for immunis. 1 It is quite 
 clear from inscriptions that a revision of orthography 
 took place at the Augustan period, and was continued 
 in later periods. We know that cum in particular was 
 quum as a conjunction, and cum as a preposition, and 
 ad was the preposition, at the con-junction. 2 With 
 regard to the alteration of cum, in in composition, 
 some orthographical varieties occur, but that in all 
 cases the final m, n was accommodated to the following 
 letter there is no doubt, as : compono, conficio, corrumpo, r 
 colludo, conduce, coeo, cohaero. 3 
 
 ART. 66. The same custom of writing the final con- 
 sonant in the same way, in order that the eye might re- 
 cognize the word, although in speech various combina- 
 tions changed its sound and the ear always readily 
 
 1 Quaeri solet, in scrlbendo praepositiones, sonum quern junctae efficiunt, an 
 quern separatae, observare conveniat [this precisely applies to all other juncturae], 
 ut, cum dico obtinuit, secundam enim b litteram ratio [knowledge of verbal de- 
 rivation], poscit, aures magis audiunt p', et immunis, illud enim quod veritas 
 exigit, sequentis syllabae sono victum m gemina commutatur. Quint, i. 7, 7. 
 We see then that "ratio" was allowed to lord it orthographically over "aures" 
 even in Quintilian's time. 
 
 a I'la quoque servata est a multls differentia, ut ad, cum esset praepositio, d 
 litteram, cum autem conjunctio. t acciperet : itemque cum, si tempus significant 
 per q>, si comitem per c ac duas sequentes scriberetur (Quint, i. 7, 5). 
 
 3 Prof. Blair (pp. cit. p. 95), refers to "Lachmann on Lucretius, p. 136, touching 
 the forms ccxrpertuS) cocoieretiir, coicere coventionid, connbium, cojnovis&e, cog- 
 nomen &c." 
 
Art. 66, 67.] FINAL M. 45 
 
 recognized it under its various forms, 1 is common to 
 many languages. Thus in French the final consonant is 
 generally written though almost invariably omitted in 
 speech before a consonant; compare un peti(f) cheval, 
 with un petit dne, which is precisely contrary to the appa- 
 rent Latin usage with regard to m. In other French 
 words, as a for Latin habet, the final consonant is so 
 usually lost that it is not written unless when accidentally 
 pronounced, so that the moderns regard it as a mere 
 euphonic introduction, as il en a, en a-t-il? 
 
 ART. 67. There is a singular usage in Dutch where 
 before a word beginning with b and d, any unvoiced 
 letter as/, k,f, s, is voiced, thus zee/bak, o/doen, strij- 
 bout, stie/broeder, mudaal are written, but the italic/, k, 
 /, s, are pronounced as English , g, v, z 9 nearly as 
 zeb'bak, ob'dun, streig'bout, stivbrirder, miz'dad in Latin 
 letters, and singularly enough this is the only case in 
 which the sound of English g can occur in Dutch. On 
 the contrary z>, z, g, become /, s, ch (guttural) after all 
 preceding consonants except r : thus in voe^rouw, stie/^ 
 soon, a^rund, the Italic letters are to be pronounced, as 
 tf> f s > fch ( cn guttural) respectively, as vuffrou stif'son 
 
 1 Writing can express but a small part of speech, leaving much to be supplied 
 by the reader's habits. The tone of voice, the rising and falling of pitch and 
 loudness, the pauses of speech, and so on, which form so many little com- 
 mentaries upon the meaning of the words, are yet so difficult to indicate, that 
 writing has seldom grappled with them, except to a very small extent. Hence 
 the speaker, who knows the concrete effect, is generally ignorant of its com- 
 position. Writing therefore seeks by various contrivances to let him know the 
 word intended, leaving him to alter it for the occasion, as he does usually 
 "by nature," that is, by a habit acquired in childhood and confirmed by the 
 practice of all around him. Horace's maxim (A. P. 180), 
 
 Segjiius irritant animos demissa per aurem, 
 
 Quam quae sunt oculls subjecta fidelibus, 
 
 which should never be separated from its context, does not apply to the case of 
 speech and writing. 
 
46 FINAL M. [Art. 6769. 
 
 af'xrund, in Latin and Greek letters. Dutch is a language 
 in which the orthography has been revised within the last 
 hundred years. Hence we must not be surprised at 
 a single written final form in -;// being retained in Latin, 
 on a revision of orthography, for cases where "junctura" 
 alone constantly altered its effect. 
 
 ART. 68. Now in Latin the neglect of m between 
 two vowels, was not a phonetic necessity of the language. 
 Thus if the m belonged to the second word there was no 
 difficulty : suave marl was a choriamb, not a cretic, as 
 suavem ard would be. 1 If the m occurred in the middle 
 of the word as, maxima, amet, there was also no slurring 
 of the vowels. The loss of m must therefore have been 
 merely a habit of speech. 
 
 ART. 69. And in this respect we may observe that 
 the termination -en so frequent in German and old 
 English has disappeared in South Germany, and in 
 literary and even most dialectal English. The first step 
 was to neglect n simply, and then to pronounce e, and in 
 this stage we frequently find the transition still in Ger- 
 many. In old English the -e final thus left was neglected 
 before a following vowel in verse, and consequently we 
 have, in an older state of our own language, an exact 
 counterpart of the Latin neglected m. 
 
 1 This fact I think conclusively disposes of Prof. Blair's suggestion. Referring 
 to Priscian's obscure final m, in such a phrase as nunqiiam ego, he says " the 
 final ;;z is almost nothing (paene nullius vocis), as far as nunquam is concerned, 
 \s\\\. Passes over in a 'weak utterance to join the e, of ego closely following^ thus : 
 iiunqua' mego, with which we may compare English d in the connection baffled 
 investigation, if pronounced baffle 1 dinvestigation" (op. cit. p. 99). But in the 
 first place I never noticed any Englishman so pronounce, and should consider it 
 accidental if I had observed anything of the kind, and in the second, the d in 
 English baffled is quite distinct and may not glide on to a following vowel. Blair 
 founds his notion of this "sort of mumbled connection," as a "passing over" of 
 the in to the following vowel on an expression in Quintilian, " ut in earn trans- 
 Ire possit " (9, 4, 40), which will be considered hereafter. See (Art. 88). 
 
Art. 7072.] FINAL M. 47 
 
 ART. 70. Quintilian, when comparing Latin and 
 Greek sounds to the disparagement of the former and 
 glorification, of the latter, speaks of the Latins termi- 
 nating so many words with the " lowing " letter m which 
 was never used in Greek, where it was replaced by the 
 " pleasant ringing" n. T But Quintilian we know from the 
 context was exaggerating, and his " lowing " m was merely 
 an eye-sore, while the " ringing " nature of n as distinct 
 from ;;/, is as we know from the difficulty felt in English 
 printing offices to distinguish the very names of the 
 letters en, em? at the least strongly imaginative. The 
 passage is not one on which we can rely, and is in fact 
 opposed to other more careful dicta of the same author. 
 
 ART. 71. Now the question is, how was this final ;// 
 treated in both cases by Latin contemporary writers? 
 We have no manuscripts in which the orthography is of 
 the slightest antiquarian value. We are driven therefore 
 to inscriptions, (of which the orthography was very 
 formal,) and to the occasional references to older prac- 
 tices in later writers. 
 
 ART. 72. A large number of cases in which m was 
 omitted in accusative, and genitive plural, and a rather 
 larger number of cases in which it was written, in in- 
 scriptions, have been given by Corssen (I. 267 271), 
 
 1 Quid? quod pleraque nos ilia quasi mugiente littera cludimus ;;*, quanullum 
 Graece verburn cadit : at illi n jucundam, et in fine praecipue quasi tinniemem 
 iillus loco ponunt, quae est apud nos rarissima in clausulls (12, 10, 31). 
 
 2 When in the years 1847 9, I had a printing office of my own for the purpose 
 of printing English phonetically (issuing among other things, the Phonetic News, 
 a weekly newspaper), and had the three types for the sounds of/;/, n, ng, in sum, 
 sun, sung, my compositors found it absolutely necessary for the prevention of 
 confusion, to give them names beginning with different vowels as, am, en, ing. 
 I wonder whether Quintilian would have admired the "ringing " Chinese terminal 
 ng, in place of m, or the French and Portuguese nasality. Hardly. Greek was 
 fashionable in Rome. 
 
48 FINAL M. [Art. 72, 73. 
 
 who has also collected many cases, really more valuable 
 than formal inscriptions for determining colloquial usage, 
 where m is omitted in the scribblings of the walls of 
 Pompeii (of course not later than A.D. 79) although these 
 form a decided minority in comparison with the number 
 of cases in which m was retained (I. 272), which is not 
 surprising, as the omission of ;;/ was then cacographical, 
 although it may have been orthoepical, in the sense of 
 representing real usage. But after the third century A.D., 
 with which we are not specially concerned, the m final 
 was very commonly omitted in all words, and the stone- 
 cutters seem to have applied the m at random even to 
 ablative cases, (I. 273-6). Corssen, however, confines 
 himself to giving the instances without reference to the 
 initial letter of the following word, which is most im- 
 portant for our inquiry. 
 
 ART. 73. As I have not the complete collections of 
 inscriptions at hand, I will content myself with reviewing 
 the whole of the words written with final m in the twenty- 
 seven inscriptions given in Vol. I. App. B. of Roby's 
 Grammar, adding the following words, numbering the 
 inscriptions, giving the approximate date, and ordinary 
 orthography. When the inscription is in verse the 
 position is indicated in the ordinary orthography by 
 a hyphen as in the examples of the Appendix below. 
 Evident abbreviations and defective words are not 
 quoted. They are mostly pre-Augustan. 
 
 I. B.C. 270-250. dono dedrot = in ordinary ortho- 
 graphy dono (donum?) dederunt. 
 
 IV. (Date not given), uicesma parti apolones = /// 
 ordinary orthography vicesimam partem Apollinis. 
 
 VI. B.C. 250. oino ploirume duonoro optumo fuise 
 
Art. 73.] FINAL M. 49 
 
 uiro luciom scipione filios Corsica aleriaque urbe 
 aide mereto = in ordinary orthography unum- plurimi 
 bon6ru m optimum- fuisse virum- (if the next word' was 
 virorum as conjectured) Lucium- Scipionem filius Cor- 
 sica" 1 Aleriam-que urbem- (if the next word was pug- 
 nando as conjectured) aedem- merito. 
 
 VII. About B.C. 250. donu dede = /;/ ordinary ortho- 
 graphy donum dedit. 
 
 VIII. About B.C. 250. taurasia cisauna samnio 
 cepit omne loucanam opsidesque in ordinary ortho- 
 graphy Taurasiam- Cisaunam- Samnium- (Samnio, ac- 
 cording to Mommsen) cepit. 
 
 IX. B.C. 189. hastensium Feruei agrum oppidumqu 
 item possidere dum poplus = in ordinary ortho- 
 graphy Hastensium servi agrum oppidumque item 
 possidere dum populus. 
 
 X. B.C. 1 86. The Bacchanal inscription, which never 
 omits the m, nor the 5- nor even the d of the ablative in 
 arf, od, of which this is the last appearance. 
 
 XI. About B.C. 164154. apiceinsigne dial in genium 
 quibus gloriam maior um qua re ingremiu Scipio 
 prognat um publio in ordinary orthography apice m in- 
 insignem dialis ingenium quibus gloriam- majorum 
 quare in gremium- Scipio prognatum- Publio. 
 
 XII. Not later than B.C. 134. donu danunt = in 
 ordinary orthography donum dant. 
 
 XIII. About B.C. 135. magna sapientia multasque 
 quom parva saxsum quoiei honore is nunquam uictus 
 honore queiminus = in ordinary orthography magnam- 
 sapientiam- multasque quum parva saxum cui hono- 
 rem is nunquam- victus honorem- qulminus. 
 
 XIV. About B.C. 135. progeniemigenui maiorurn 
 
 E 
 
50 FINAL M. [Art. 7375. , 
 
 optenuilaudem ut creatum la?tentnr stirpem nobili- 
 tauit = in ordinary orthography progeniem- genul [a 
 doubtful amendment of a clearly defective original] 
 majoru m obtinui laude m ut creatum laetentur; stirpem- 
 nobilitavit. 
 
 XV. Between B.C. 146 and 134. romam reclieit 
 aedem et signu herculis = in ordinary orthography 
 Romam redilt aedem et signum Herculis. 
 
 XVI. Date soon after XV., down to XXVII., "at the 
 end of the republic," never omit m final, or nominative s, 
 or insert ablative d. 
 
 ART. 74. It must be observed that these inscriptions 
 very rarely double any consonants in writing. Even the 
 careful Bacchanal inscription (X.) is full of words like : 
 due/onai baranalibus e^ent habuiVe ue/et ade^e adie.re 
 iouori^ent, with single consonants = in ordinary spel- 
 ling Eellonae Bacchanalibus essent habuisse vellet 
 adesse adiise jiississenr, with double consonants. 
 
 ART. 75. Now the rule with regard to variations from 
 ordinary orthography occurring in the writing of early or 
 unpractised writers in any language, is, that the deviations 
 are always on the side of pronunciation. Cacography is 
 always a surer guide to sound than orthography, because 
 it is due to Quintilian's "aures" and not his "ratio." 
 Orthography depends upon rules which had to be laid 
 down because other considerations outweighed phonetic 
 reasons, and which are consequently ill followed by those 
 who do not appreciate those reasons. The historical 
 character of orthography, also, always makes it more 
 archaic than actual usage. There is no need to go be- 
 yond English and French for proofs of this. A long 
 study of the English usages and examination of older 
 
Art. 7577.] FINAL M. 51 
 
 works on orthography and orthoepy, leaves no doubt in 
 my own mind on this point. 1 
 
 ART. 76. An examination of these inscriptions would 
 therefore lead me to the following conclusions. 
 
 First, that final m had no appreciable sound at a'l 
 when final, or before a following word beginning with 
 a vowel. 
 
 Secondly, that final m, before a following word begin- 
 ning with a consonant, had no sound of m, but became 
 appreciable in speech, either by lengthening the pre- 
 ceding vowel, or by doubling the succeeding consonant. 
 
 Thirdly, that the phonetical omission of the ablative 
 d, (the vowel preceding it being always long), was 
 connected with the orthographical restitution of the 
 accusative and neuter ;;/, to prevent the eye from con- 
 fusing ablatives and nominatives with accusatives, and did 
 not necessarily imply any restitution of the sound of m. 
 
 In the second conclusion there is an alternative hypo- 
 thesis of lengthening the preceding vowel or doubling 
 the succeeding consonant, because the inscriptional or- 
 thography would have been the same in either case. 
 The supposition that orthographers could have actually 
 omitted the final m and doubled the following consonant 
 in writing, at a time when doubling consonants under any 
 circumstances was so rare, is not tenable. 
 
 ART. 77. Two questions necessarily arise with regard 
 to these conclusions. First, is there any analogue in any 
 known language? Secondly, is there any basis for such 
 an hypothesis in Cicero and Quintilian? We need not 
 inquire further, for it is self-evident that these conclusions 
 
 1 See numerous instances in my Early English Pronunciation 'with especial 
 reference to Shaksjere and Chaucer, 1869 71. 
 
 E 2 
 
52 FINAL M. [Art. 7780. 
 
 would explain both the facts (a, b} of Augustan versi- 
 fication with which we started (Art 64). 
 
 ART. 78 Now the two languages which most closely 
 resemble Latin are Spanish and Italian. Spanish has 
 lost the habit of doubling consonants altogether, but it 
 almost invariably omits this final m in the words taken 
 from the Latin. Italian, however, does double con- 
 sonants. It also almost invariably omits the final m of 
 the Latin. 1 
 
 ART. 79. Italian, however, also frequently omits other 
 Latin finals, as /, d. In this case there are two treat- 
 ments, one orthographical and the other orthoepical, 
 which bear a striking analogy to the conclusions just 
 drawn from the inscriptions, and which, in point of time, 
 first led me to conceive this solution of the difficulty. 
 
 ART. 80. First, the omitted Latin letter is never 
 written in Italian when it is not pronounced. Thus ad, 
 et are in Italian a, e generally, but occasionally ad, ed 
 when required for metre. The bare a, e are slurred, as 
 e invano (S. 5), ai = ad i (S. 7); but ad, ed make 
 syllables as : 
 
 miro* tut'te co'se, ^ed'vn. Sori'a. 
 
 TASSO, Ger.i. 8, i. 
 re 'co ad\a al'ta^origina'ria foirte. 
 
 *Ib. i. 30, 5. 
 
 1 The qualification "almost" before " invariably" is necessary because there 
 are a few words, chiefly monosyllabic, in which a trace of nasality remains, Latin 
 cttm, sum, spent, are Italian con, sono, spene($) ; Latin quern, tarn, are Spanish 
 guietr, tan; Latin rent is old Spanish ren, French rien; and on Roman inscrip- 
 tions con, qnen, tan, occur. ButjVzw has lost its m everywhere. The in of the 
 accusative has quite vanished, except in mon, ton, son, for meum^ tuum, suum. 
 These and other particulars may be seen in Diez. The object here is to arrive 
 at general conclusions, not particular exceptions. It is impossible to attempt a 
 list of exceptions in Augustan speech, where even the general rules are reached 
 with difficulty (see Art. 8). 
 
Art. So, 81.] FINAL M. 53 
 
 The employment of ad, ed, is rare and archaic, and 
 confined almost to poetry. It is comparable perhaps to 
 the occasional use of m final to preserve the preceding 
 syllable in Lucretius, 1 if indeed the latter are not, as I 
 believe, really cases of unslurred open vowels which occur 
 much more frequently (Q. n). See also p. 65. 
 
 ART. 8 1. Next, in certain cases, words, in which the 
 omitted Latin letter is not written in Italian, are con- 
 nected orthographically with a following word, beginning 
 with a consonant, and that consonant is then doubled to the 
 eye as well as the ear, thus replacing the omission by 
 assimilation. Thus gia che = Latin jam quod, become 
 giacche; gid mai = Latin jam magis, become giammai ; 
 con lo = Latin cum illo, become col (S. 3, and compare 
 the preserved con la in the same line for the usual colla); 
 gia sia cib che Latin jam sit ecce-hoc quod become 
 giassiadocche ; do che = Latin ecce-hoc quod become 
 generally cioccJie; a dio = Latin ad deum, become addio ; 
 a fatto = Latin ad factum, become affatto ; a fine che = 
 Latin adfinem quod become affincke; a lato Latin ad 
 latus become allato ; a le arme = Latin ad ilia arma 
 become allarme, whence our alarum, alarm, and a le is 
 generally alle\ a mano Latin ad manum become 
 ammano ; e bene, e poi, e pure Latin et bene, et posted, et 
 pure become ebbene, eppoi, eppure; da vero = Latin de ad 
 verum become davvero ; o vero, o pure Latin aut verum, 
 aut pure, become ovvero oppure; si bene, si fatto = Latin 
 sic bene, slcfactum become sibbme, siffato ; il dio =. Latin 
 ille deus become Iddio the one God. Again the perfect 
 tenses ending on ~b, -I, when a lo, vi, &c. is added on, 
 
 i Nam quod | fluvidum | est e | levibus J atque ro j tundTs. u. 464. 
 Sed dum a | best quod a | vemus id | exsupe | rare vi | detur. in. 1094. 
 
54 FINAL M. [Art. 81, 82. 
 
 become -olio, -ovvi, &c., as : ei dimostrollo [for dtmostrb 
 lo\ a lungo (Tass. Ger. i. 29, *]),~amovvi for amo vot, 
 and so on. 
 
 ART. 82. Thirdly, when, as is most frequently the 
 case, words from which a final Latin consonant has 
 been dropped in passing into Italian, are not written in 
 connection with the following consonant, they are still 
 spokeji in connection with it. For this observation I am 
 indebted to Prince Louis Lucien Bonaparte. He con- 
 siders the following consonant in this case to be " ener- 
 getic" (Art. 20) and to preserve this character at the 
 beginning of a phrase. 1 In the first stanza of (S) all the 
 words preceded by a hyphen are thus energetic, according 
 to a copy which the Prince was good enough to make 
 for me himself, and so distinct is the phonetic effect, that 
 when I was reading out the stanza to him, he corrected 
 me in every case where, from want of habit, I neglected 
 
 1 The following is a translation of the rules prefixed to Prince L. L. Bonaparte's 
 Sassarese Sardinian translation of St. Matthew (1866), with the exception of 
 two omitted at his own request. The Italian text may be seen in my Early 
 English Pronunciation, p. 799 note. "Simple consonants are often pronounced 
 like double consonants in Italian, according to the following general rules: 
 i. When they are initial and commence a phrase, either at the beginning of a 
 sentence or short clause, or after a vowel : 4. When the preceding word, although 
 ending in a vowel, has the force-accent on the last syllable, or is a monosyllable, 
 in both cases derived from a Latin word ending in a consonant which has been 
 suppressed in passing from Latin to Italian. Thus the preposition a from Latin 
 ad, the conjunction e from et, and si from szc, ne from nee, and truncated words 
 as amo from amdvii, pote from potuit, have all the property of making the initial 
 consonant of the following word energetic. Hence though we see the written 
 words : a Pietro, e voi, si grande, ne questo ne quello, amo molto, pote poco ; 
 we really hear nothing but: appie'tro, evvo'i, siggran'de, necques'to nec- 
 quel'lo, amom'mol'to, potep'po'co. In other cases the consonant remains weak 
 [single]. Thus in : di Maria, i doni, la mente, le donne, mi dice, ti lascia, si 
 gode, ama molto, pote' poco, molto largo, the initial consonants are spoken 
 as written, either because the preceding words are the Latin tie, Hit, ilia, 
 illae, me, te, se, /Wz, which end in vowels, or because in the case of a' ma 
 mol'to, mol'to lar'go, the words a'ma and mol'to are not accented on the last 
 syllable." 
 
Art. 8284.] FINAL M. 55 
 
 to make the consonant " energetic." 1 Taking only the 
 cases where this " energy " supplies the place of an 
 omitted consonant, we find practically : aHiri (S. 5) ; 
 ec-con (S. 3), ed'di (S. 6), es'softo (S. 7); liberod'di 
 (S. 2), oproc'col (S. 3), armod'da'sia (S. 6), dieffavo're 
 (S. 7), soffrin-nel (S. 4). Observe that two syllables have 
 the force accent in certain cases, distinguishing two 
 united words from a single word, as the Latin pitch 
 accent was supposed to act in Art. 60. 
 
 ART. 83. Taking account of Cicero's emphatic de- 
 claration of the necessity of connecting words in Latin 
 speech (Q. 9), and remembering that Italian is the de- 
 scendant of at least some form or forms of Latin speech, 
 (not necessarily or probably of the literary Augustan form, 
 and hence in some cases shewing usages not traceable 
 to that form, 2 ) is it too much to suppose that an Italian 
 usage, exactly conformable to what we may assume as 
 the colloquial forms indicated by old inscriptions, was 
 also the old Augustan usage for final m ? To me it 
 seems that very strong and direct contemporary evidence 
 would be necessary to lead to any other conclusion, and 
 hence I proceed to examine the second question of Art. 77. 
 
 ART. 84. What is the contemporary evidence? It is 
 neither much, nor clear. Cicero's consists simply in 
 stating that cum nobls, was not said, but either cum autem 
 nobis, or nobiscum, to avoid the sound of cun nobls.^ To 
 
 1 This happened a year before this paper was written. This is mentioned to 
 shew how the hypothesis here developed originated. 
 
 2 To these might be attributed the rare preservation of nasality as sono for both 
 sum and stint, so that an original confusion of these words is possible. 
 
 3 The reason he gives and its application, like most old pieces of linguistry, 
 are almost impossible to credit: quid illud? non olet unde sit, quod dlcitur, cum 
 illis ? cum autem nobtsnon dlcitur, sed noblscum ? quia si ita dlcere ur, obscaenius 
 concurrerent litterae, ut etiam modo, nisi autem interposuissem, concurrissent. 
 
56 FINAL M. [Art. 84, 85. 
 
 this must be added Quintilian's cun notls, where, however, 
 he endeavours to give a phonetic explanation of the 
 assimilation. And this explanation requires attention, 
 because, although valueless in itself, it seems to imply 
 that final m in his time was considered so difficult to 
 pronounce before a consonant, that it had to be either 
 assimilated to it, (as in the Italian examples), or entirely 
 omitted. If we may credit him, Roman organs could 
 not pronounce 1 their final m purely, without making 
 a pause after it, and before the following consonant. 
 
 ART. 85. Quintilian's words in reference to cum notts, 
 are : quia ultima prioris syllabae littera, quae exprimi nisi 
 labrls coeuntibus non potest, [that is, if m is expressed 
 at all, the lips are closed, but so must they be for the 
 common , /,] aut intersistere nds indecentissime ["most 
 unbecomingly," this does not refer at all to the meaning 
 of the words, but to the unbecomingness of the hesi- 
 tation] cogit, aut continuata cum insequente in naturam 
 
 ex eo est mecum et tecum : non cum me et cum te, ut esset simile illls I'dbiscum 
 atq ae noblscum, Cic. Or. 154. Such as it is, however, Quintilian echoes it for 
 cun notls hominibus, which he therefore puts in the order cum hominibus notls, 
 and he gives several other instances, much more difficult to enter into, where 
 accidental combinations of words may be twisted into dirty senses, until Quin- 
 tilian is forced to exclaim : quod si recipias, nihil loqul tutum est (8, 3, 
 4447)- 
 
 * Any unusual combination of even usual sounds creates a difficulty to the 
 speaker. An Irishman, when I remarked to him on the inconsistency of the 
 Irish pronunciations of machine fatigue as rhyming to English seen plague re- 
 spectively, made the extremely just observation, that if you were to ask an Irish 
 peasant, he would tell you it was much "az'sier" to speak so than to make them 
 rhyme to either English seen league or rain plagrie. The case is that of habit 
 only. Prof. Blair (ibid, p. 97), quotes from "Servius in II. Donati editionem, 
 ap. P. p. 1797 : nemo enim dicit cum me, cum te, propter cacophaton," which 
 in view of dum me, ditm te, is simply absurd. But Servius had always heard 
 mecum tecum, and hence anything else sounded wrong, as in English I goed 
 (really the good old yode) for / went. Italians, who usually say me'co te'co 
 sometimes indulge in con me'co, con te'co. It's well that Servius can't hear 
 them ! 
 
Art. 85, 86.] FINAL M. 57 
 
 ejus corrumpitur [" broken down, assimilated," probably 
 not implying a censure on what was an exceedingly 
 common Latin custom]. Quint. 8, 3, 45. Now with 
 regard to shutting the mouth, it must be remembered 
 that the lips and tongue are in exactly the same position 
 for /, , m ; and that they are also in the same position 
 (though different from the last) for /, d, n. Now does 
 Quintilian mean to say that- he could not distinguish 
 abnuit from annuit (which had absolutely contrary mean- 
 ings), without an "unbecoming pause" after ab in abnuit , 
 to prevent the b from being " broken down " into n and 
 thus making annuity He would probably have been 
 very much surprised at the question. But his phonetic 
 reasoning applies strictly to this case, which he never con- 
 templated, because he was thinking only of his habits 
 respecting final m, and these, as I gather from his words, 
 were so ingrained, that he could not pronounce without 
 considerable effort and an " unbecoming " hesitation, that 
 " lowing " sound m, whose presence at the end of words 
 he found so offensive in Latin when he wanted to de- 
 preciate his own language in respect to Greek (12, 10, 
 31) ; see Art. 70, note i. Such inconsistencies are very 
 common with persons who write on subjects they have 
 not studied scientifically, and very few indeed have con- 
 sidered it at all necessary to attend to the real nature and 
 science of speech sounds. Quintilian always speaks as if 
 the matter were so elementary that it was beneath his 
 notice, and consequently writes hurriedly, inconsistently 
 and insufficiently, which is a great loss to modern 
 investigators of ancient habits of speech. 
 
 ART. 86. These are really all the indications which 
 I can find of the use of m final before consonants in the 
 
5 FINAL M. [Art. 86, 87. 
 
 Augustan and post-Augustan centuries. So far as they 
 go they confirm the conclusions from the inscriptions and 
 from Italian usage. It is of no use referring to Priscian 
 and Donatus, except to note that Priscian says that m 
 final had an "obscure" sound, 1 because we know for 
 certain that for at least a century before his time, final 
 m could not have been pronounced at all, and hence 
 this use of "obscure" enables us to estimate the mean- 
 ing of Quintilian's "obscuration" of m, before a vowel, 
 which we shall have to examine presently. 
 
 ART. 87. Before proceeding to this observation of 
 Quintilian, I will refer for a moment to Verrius Flaccus, 
 his contemporary, whose work on orthography might 
 have been of much service in this respect. The following 
 statement is made on the authority of Velius Longus, 
 & contemporary of Macrobius, who belongs to the late 
 
 1 Prof. Blair (op. cit. p. 93), and Corssen (I. 263), quote Priscian (the first as 
 I- 555> tne second as T. 38, H, I have not verified either), thus: " M obscurum 
 in extremitate diet onum sonat, ut 'templum; ' apertum [as if there could be an 
 open mouthed M !] in principle, ut 'magnus ;' mediocre in medils, ut 'umbra.'" 
 Now all phonologists know that initial consonants have generally less sound than 
 that which it is possible to give to final consonants, and that consonants which 
 are followed immediately by other consonants in the same syllable or adjacent 
 syllable, may be variously altered. In w ^at sense Priscian meant his apertum 
 and mediocre, it would have been probably very difficult to determine, even in a 
 viva voce examination (judging from much experience), but what he meant by 
 "obscure" is very clear; it was simply that m was written and not sounded. 
 ''Obscure" implied no more than this. Donatus ad Ter. Adel. II. i, 53 (as 
 quoted by Prof. Blair, ibid. p. 93), says: mussitdre dictum a miitd, vel ab nil 
 quae littera est nimium pressae vocis ac pene nulllus [really, when final, ab- 
 solutely mute, in his time, he lived in the fourth century and was the master of 
 St. Jerome, the writer of the Vulgate, ] adeo ut sola omnium quum inter vo- 
 cales inciderit, atteratur atque subsldat [is ground up and settles down, like grit 
 in water?] We know that m is so far from being naturally " nimium pressae 
 vocis ac pene nulllus," that it can actually be sung upon, not only in ordinary 
 humming, but as a singing part in Mozart's Flmito Magico. Had Donatus 
 any notion of "energising" the following consonant? His expressions at least 
 do not contradict such a theory. That the practice may be common without 
 being observed, I at leat know from having lived a year in Italy without ob- 
 erving it. 
 
Art. 87, 88.] FINAL M. 59 
 
 grammarians. He says in his orthography (Blair, p. 94, 
 Corssen I. 26; P. p. 22, 38): nonnulli synaloephas 
 [junctures of Quintilian] quoque observances circa talem 
 scriptionem existimaverunt, sicut Verrius Flaccus, ut 
 ubicunque prlma vox m littera finiretur, sequens a vocali 
 inciperet, M non tota, sed pars illms prior IV tantum 
 scriberetur, ut appareat expriml non debere. That is, 
 Verrius Flaccus wished to shew by partial defacement 
 that an M though written, was not pronounced, just as 
 "mute letters " are underdotted, italicized, or printed in 
 skeleton type, in various systems of teaching to read 
 English. Here only the case of final m preceding 
 a word commencing with a vowel is mentioned, for if 
 a consonant followed, Verrius Flaccus would probably 
 have considered that the m was pronounced, though assi- 
 milated, as the effect would have been different had no m 
 been there at all. 1 It must be remembered, however, that 
 we have not the precise words of Verrius Flaccus but 
 only a report of them furnished by a writer in whose 
 time this m was certainly not pronounced at all, and 
 that probably much more existed in the original. 
 
 ART. 88. The classical passage in reference to final 
 m before a vowel is in Quintilian. He has been speaking 
 of juncturae, and of the disagreeable effect of the con- 
 currence of certain letters (Q. 2, 3) which " in commissura 
 
 1 I have myself been told that gh in English sight and e in English site were 
 "pronounced," because if they were left out, the words would be szV, and have 
 an entirely different sound ! We are, I believe, indebted to Mr. B. H. Smart 
 for a distinction, important in reference to Latin final m, that the italicised letters 
 in the English words, meal, charcoal ; flu's, fo^s, du^s ; pazn, ez'ther, suz't ; broach, 
 door; soul, bowl; play, kejj/; Messiah; mate, bath<?, paste, age, s.ce ; gwess, 
 guide, plagz/e, &c. "are mute, though in general significant." (Pronouncing 
 Dictionary, 1836, Art. 171). This amounts to distinguishing marks of separate 
 sounds, from marks used diacritically to modify the other marks, that is, "lit- 
 terae " from "notae," according to Quintilian, as we shall presently see. 
 
60 FINAL M [Art. 88. 
 
 verborum rixantur." He notices ars studiorum, and the 
 Lucilian omission of s final in Aes^rninu(s) fuit, dignu(s) 
 locoque, which was evidently so strange to him that he 
 could never have noticed it in speech, and he goes on to 
 say : inde belligerare, pomeridiem, et ilia censoril Catonis 
 dice hanc, aeque M littera in E mollita, [the letter M 
 "softened"' to E, "in the same way" aeque; the phrase 
 must have meant diem hanc, and the double ee probably 
 stood for a long e, and shewed that there was no elision ; x ] 
 quae in veteribus libris reperta mutare imperiti solent, 
 et dum librariorum Insectarl volunt Inscientiam, suam 
 confitentur. 2 atqui eadem ilia littera, quotiens ultima 
 est et vocalem verb! sequentis ita contingit, ut in earn 
 translre possit, [so that it can " pass over into it," that is, 
 be " softened " into it, as in diee for diem just cited, and not, 
 " be pronounced with the second vowel instead of with 
 the first," see Art. 68, note; see also the expression 
 dissimulator in the passage referring to the same practice, 
 quoted in Art. 137], etiamsl scribitur, tamen parum 
 exprimitur, [is " little " expressed, euphemistic for minime 
 = "not at all," the word really means, "not enough," 
 "not so much as it ought to be considering that it is 
 written,"] ut multum ille et quantum erat: adeo ut paene 
 
 1 Quintilian says in the passage referred to by him in the quotation on p. 42, 
 note : in ipsis vocalibus grammaticl est videre quae ut vocaJes junguntur aut 
 unam longam faciunt, ut veteres scripserunt, qui geminatione earum velut apice 
 [as a mark of length] utebantur, aut duas (i, 4, 10). Hence dice would = die. 
 This would allow us, in the absence of proper types, to use double vowels for the 
 capitals of initial long vowels, as OOrator = orator, a practice which would be 
 quite unambiguous. When a word is entirely in capitals, a small letter inter- 
 posed would mark length, as OoRAaTOR, LAaOCOOoN. In the examples of 
 the appendix, capitals have been entirely avoided, but in general quotations, and 
 in modern editions of classical works, the above antique practice, slightly varied 
 to prevent ambiguity, might be adopted. 
 
 2 Editors of old texts, please to copy ! 
 
Art. 88.] FINAL M. 61 
 
 ["almost" the usual orthographical saving clause], cu- 
 jusdam novae litterae sonum, [really " sound"? see what 
 follows] reddat, neque enim eximitur [probably merely, 
 it is not omitted in writing, with a feeling that it could 
 not be left out in general spelling because of its action 
 before a consonant], sed obscuratur [" made inaudible/' 
 compare Priscian, as quoted in Art. 86,] et tantum 
 aliqua inter duas vocales velut nota [a mark, not 
 a sound by itself, see on ^, Art. 53, note] est, ne ipsae 
 coeant. Then he goes off at a tangent on the necessity 
 of being able to divide words rightly : videndum etiam 
 ne syllabae verb! prioris ultimae sint primae sequentis, 
 [instancing most singularly cases in which no difficulty 
 could have occurred], id ne quis praecipl miretur, Ci- 
 ceroni in epistolis excidit : " res mihi invisae visae sunt, 
 Brute : " et in carmine : 6 fortunatam natam me consule 
 Romam. (9, 4, 39 41). Now the points here are that 
 the m almost acts as another letter, (like that of Ver- 
 rius Flaccus?) to which Quintilian indeed attributes a 
 sound (sonum), although he owns almost in the same 
 breath that it only acted as a mark of separation, like the 
 stonecutters (.) ! Now, years ago when I read this passage, 
 I was misled to think that the "parum exprimi" im- 
 plied that m was expressed in speech "somehow," if 
 not " enough," and to think that the " neque eximi- 
 tur" referred to its not being altogether left out in 
 speech instead of to its not being omitted in writing, 
 and that the " obscuratur/' and the u paene cujusdam 
 novae litterae sonum" referred to that nasalising of 
 the previous vowel which is to this day expressed by a fol- 
 lowing m or n in French, and m in Portuguese, so that m 
 is thus applied to a particular means of modifying vowel 
 
62 FINAL M. [Art. 88. 
 
 sounds. 1 But in the first place nasalisation is strange 
 to Italy, even to its dialects (with rare exceptions). 
 Next in French there is every reason to suppose that 
 nasalisation of vowels did not exist till at least a century 
 after the Norman Conquest, and was not complete till 
 the seventeenth century. 2 Again this nasality, when it 
 existed, did not affect the Latin final m, except in such 
 rare cases as to have no weight, for the whole of the 
 case-endings in -am, -em, -//;;/, had disappeared before 
 the nasalisation of French had commenced. I do 
 not know the history of Portuguese nasalisation, but I 
 find independent nasalisation arising in Southern German 
 (where certainly it sometimes implies an omitted n) and 
 in American English, while it probably exists widely in 
 extra-European languages. Also I find the German -en 
 
 1 This suggestion of mine (Phil. Soc. Trans. 1867, supp. p. 20) having been 
 referred to by Mr. Roby (Gram. Pref. p. Ixxxiii. 1. 10) in relation to this passage 
 of Quintilian, I wish particularly to state that I have entirely withdnwn it. 
 Prof. S. S. Haldeman, from whom I differ on phonetic points with great re- 
 luctance, adopts the nasality unreservedly (ofi. ci. in Art. i, p. 27), referring to 
 the nasal anusvara in the Sanscrit originals of Latin final in, to the passages from 
 Priscian, and Verrius Flaccus, just quoted, and to an inscription copied by 
 Manutius "in which a small curved line (~) is used (at least by him) to repre- 
 sent M, N, and N (ng\" This last mark was probably merely a cursive m or n 
 as in modern Portuguese and Spanish, frequently degenerating into a single 
 straight line, of which the mathematical sign = in for minus, is a well-known 
 example, and which is of constant occurrence in all medieval (even English) 
 MSS. where no nasality need be suspected. Prof. Haldeman says explicitly 
 (op. ci. Art. 105): "the Latin nasal vowels are i, E, A, o, u, as em m , dece m , 
 ta m , flovio, tu m ," both the nasality and long quantity are doubtful to me. 
 What the anusvara was originally, I do not know ; it has become in Indian pro- 
 nunciation of Sanscrit a simple English ng in singing, and its mark (a mere dot) 
 seems to be the affixed dot by which alone the devanagari character for this ng 
 was distinguished from that of the cerebral d (which is the English </). This 
 seems to point to an early use of this sound, which certainly replaced final m 
 under certain circumstances in Sanscrit, th^t apparently have nothing to do with 
 Latin. A very early Latin final m pure, is not disputed. 
 
 2 See my investigation of the subject for the sixteenth century, with M. Paul 
 Meyer's results for the earlier period, in my Early English Pronunciation, 
 pp. 825828. 
 
Art. 8890.] FINAL M. 63 
 
 disappearing into -e on the Rhine, without nasalisation. 
 Also certainly the same change occurred in English itself 
 without even a tendency to nasalisation. I know of no 
 passage which even slightly supports the nasalisation 
 theory, except that just cited, and this (which really says 
 nothing of nasalisation directly), may I think be inter- 
 preted as above, in a manner more consonant with the 
 expressions of orthographers and orthoepists. 
 
 ART. 89. The result is that (excluding compound 
 words) we have one undoubted case of assimilation of 
 final m to following n in Cicero and Quintilian ; we have 
 also Quintilian stating that this assimilation could not 
 be avoided without an unbecoming pause, shewing 
 evidently that his organs were not accustomed to avoid 
 it ; we have a recorded though archaic case (diee) of final 
 m being omitted and the preceding vowel lengthened (if 
 indeed " diem " ought not to have a long vowel always, 
 as being of the e declension, so that it is after all a simple 
 case of omitted m) and we have an account of final M 
 generally before a vowel implying that it was really not 
 heard at all. Hence our. examination of contemporary 
 authority rather confirms than weakens the conclusion 
 drawn from the usages of pre-Augustan and post-Au- 
 gustan inscriptions and of modern Italian. 
 
 ART. 90. A little uncertainty may still prevail as to 
 the monosyllables in -;;/, where as we have seen (Art. 78) 
 an m or ;/, or nasality is sometimes found in romance 
 languages. But the following examples from Horace's 
 Satires (which are selected especially because although 
 " pede certo," so that the effect can be measured without 
 hesitation, they are " sermonl propius" and thus little 
 likely to resort to " poetic licenses," and to have been 
 
64 FINAL M. [Art. 90. 
 
 effective must have reproduced the speech of the day, 
 very slightly twisted to force it into metre), will shew 
 that monosyllables were not excepted from the law of 
 silent m before a vowel, but were treated precisely like 
 other monosyllables which had no /;/. 
 
 DUM . . du m w ex parvo. HOR. Sat. I. I, 52. 
 
 du m w aes exigitur. I. 5, 13. 
 
 natus du m w ingenuus. I. 6, 8. 
 
 nee du m w omnis abacta. II. 2, 44. 
 TUM . . tibi tu m w efficient res. I. 2, 97. 
 
 tu m ^immund6. I. 5, 84. 
 
 tu m w inlecto. n. 8, 77. 
 NUM . . pocula; nu m esuriens. I. 2, 115. 
 
 nu m w ignota. I. 6, 36. 
 CUM . . cu m ^ / ux6ribus. I. 2, 57. 
 
 nugari cu m w illo w et distinct!. II. I, 73. 
 QUUM . quu m w exiret. I. 2, 30. 
 
 quu m w est jussa venire. I. 2, 122. 
 
 quu m w adsectaretur. I. 9, 6. 
 
 quu m w est Lucilius ausus. II. I, 63. 
 
 quu ra w llionam w edormit. II. 3, 61. 
 
 quu m w immerit5s. II. 3, 211. 
 SUM . . quidquid su m w ego quamvis. II. I, 74- 
 
 quid su m w ego? nempe. II. 7, 80. 
 NAM . . na m w exemplo v _ / est I. I, 33. 
 
 na m w ut ferula. I. 3, 120. 
 
 na m w ut multum, nil moror. I. 4, 13. 
 
 na m w ut quisque^insanus. I. 6, 27. 
 
 na m w inquiram. 1 1. 3, 41. 
 QUAM . qua m w ex hoc fonticulo. I. I, 56. 
 
 qua m w ex re decerpere. I. 2, 79. 
 
 qua m w aut aquila^aut serpens. I. 3, 27. 
 REM . . re m w imperito w ac si. n. 3, 189. 
 
 committes re m w omne m w et vita m w et 
 cum- corpore fama ra . II. 7, 67. 
 
 There is however one passage, where, if the reading is 
 correct, num forms a short syllable before a vowel : 
 
Art. 90, 91.] FINAL M. 65 
 
 num vesceris ilia 
 quam laudas pluma? cocto num adest honor idem? II. 27-8. 
 
 where as before, and as even in drcumago circumeo I 
 should incline to read as an hiatus, cocto nu adest, circu 
 ago, circu eo. It is possible, however, that the correct 
 reading is " cocto nu m w et adest honor idem," the et is 
 wanting to the sense at least as much as to the metre. 1 
 
 ART. 91. The general rules for the treatment of final 
 m which this discussion leads to are as follows. Ex- 
 ceptions must be admitted, but the only safe course is to 
 carry out the principle strictly, for there is absolutely no 
 means of determining what and where the exceptions 
 may be. 
 
 a. Final m, followed by a word beginning with a vowel 
 in the same clause, on to which, had there been no m, 
 the preceding vowel could have been slurred by the 
 regular rule, is totally inaudible and ineffective, and the 
 vowels are slurred just as if no m had been written. 
 
 b. Final m, at the end of a sentence, is totally inaudible 
 and the preceding vowel, like all final vowels (P. 2) is 
 indifferently long or short, but had better be made long 
 to indicate the excision of m. 
 
 c. Final m, before a pause, however brought about, is 
 
 1 Horace is talking of the extravagance of eating a peacock, and says that 
 people would prefer it to a common fowl, if- they had the choice, because it is 
 dear, and rare, and with a splendid plumage, as if that had anything to do with 
 the matter: "do you eat the plumage that you laud? has the cooked bird still 
 the same glories?" For adest, we should in prose most probably have had 
 manet, but this would not scan, and the sense is preserved by using et adest. 
 I don't know of any authority for this conjecture, but the syllable num before 
 a vowel is I think so impossible a reading for Horace, that I cannot quote the 
 passage without noting that it is manifestly incorrect. Even the line amator 
 Sxclfisus qul distat, agit- ubi secum, eat an non (Hor. Sat. ii. 3, 260), is suspicious 
 in lengthening agit- by a kind of position before ubi, although less so than the 
 archaic num with pure m Compare also //#.$ formates, (Art. 22) note, which 
 is also suspicious. Corrections are easy as : agit quum, agit quT. 
 
 F 
 
66 FINAL M. [Art 91. 
 
 totally inaudible, but if the following word commence 
 with a consonant the vowel before m may be lengthened, 
 and in any case the succeeding consonant must be 
 uttered with more force, to indicate the excision of the 
 m, (see Art. 82, and Art. 92). 
 
 d. Final m before a word commencing with any con- 
 sonant but j or z/, is to be rendered effective by omitting 
 the m, and pronouncing the consonant as if it were 
 doubled, keeping the vowel before m short, but running 
 it on to that doubled consonant, 1 and the pitch of the 
 voice must be raised upon as many syllables as it would 
 have been had the words been separately pronounced. 
 The final m before enclitics quc, ne, &c. is to be treated 
 in the same way. Use cqu for doubled qu, even in 
 fiamquc, dumque =. ndcque, ducque. 
 
 e. Final im followed by /is to be pronounced as I, and 
 final urn followed by v as u. No other pronunciation 
 could have well prevailed. It may be best to pronounce 
 mj as ij\ and mv as uv in all other cases, so that the 
 /* u which replace m, form diphthongs with the pre- 
 ceding vowel. But as a compromise, to ride over 
 a difficulty which there is no authority to settle, 2 the 
 
 1 It is probable that the sight of m will occasion many teachers to object to 
 this rule, and to wish m to be made at most n before /, d, n, /, z/, and n adul- 
 terinum (ng English) before c, g, qu. I long wished it myself. But I could not 
 find that this would explain Augustan habits. So I let it go. Cases like, conficio, 
 convTcium, belong of course to an older stage of the language. The fondness 
 for n before ./j V, is quite local In Scotland Banff is written and am f said, 
 which shews a precisely opposite feeling, and one more in accordance with the 
 positions of the organs of speech. 
 
 2 If the m is neglected then j'amj'am would be equivalent to j'aja. Now if 
 this is treated as in ajacem, cujus, ejus, we should say ja'ja. But it may be 
 questioned whether jaij a, aijacem, cuijus, eijus, &c. would not be more correct. 
 This seems at least to have been the opinion of the later grammarians (Corssen 
 I. 301 3), but of what value is that for Augustan Latin? It is possible indeed 
 that Quintilian's words imply Cicero's usage to have been the same : sciat etiam 
 
Art. 9193.] FINAL M. 67 
 
 reader may always pronounce am, em, im, om, um as 
 simple a, e, I, d, u before both j and v. The final m 
 before enclitic ve is to be treated in the same way. 
 
 ART. 92. Mode of practice. Rules a and b require no 
 new practice, as the pupil is already familiar with slurred 
 vowels. Rule c requires practice in giving force to the 
 following consonant. If that is /, /, c or b, d, g, this 
 force is produced by making the closure of the organs 
 tighter, so that the following sound comes out more ex- 
 plosively, and there is the same delay between closing the 
 organs and opening, as if pp, it, cc or bb, dd, gg occurred 
 before the vowel. But if the following consonant be 
 f> s ij> v i m i n > I* r > which have sounds of their own that 
 can be made both long and loud, then prolong them 
 sensibly and firmly before the vowel. This may be in- 
 dicated on the black board by doubling the initial of the 
 following word. Thus (A. 13) incipia, ffractl; (F. 8) 
 carpe die, cquamminimu, ccredula postero. 1 
 
 ART. 93. The rule d is also easy to carry out when 
 once the effect of doubled consonants between vowels 
 has been properly grasped. The difficulties arise first, 
 from thus connecting together several distinct words, and 
 consequently several syllables with raised pitch ; secondly 
 from the fact that the doubled consonant occurs in 
 a syllable of low pitch, and often little force, which has 
 
 [puer] Ciceroni placuisse aiio matmmque. geminata I scrlbere quod si est, etiam 
 jungetur ut consonans, (i, 4, u), that is, if so, Cicero must have said aijo maija. 
 As the plan is very reasonable it may be adopted without much hesitation, so 
 that we should read (L. 20) as, jaijaffutu'rusru'sticus. 
 
 1 Thus I find in a curious Neapolitan version of the Aeneid by Giancola Sitillo, 
 (Naples, 1784, lent me by Mr. Hodgson), in a translation of (A.), the following 
 initial reduplications : che mme, lo ccomme, uocchie ssi, la parte cchiu, a mme, 
 io line, de la cchiu, a ffunno. These seem to imply this energy, but I am not 
 sufficiently acquainted with the dialect to judge, having only heard very little 
 of it more than thirty years ago. 
 
 F 2 
 
68 FINAL M. [Art. 93-95. 
 
 nevertheless to be made sensibly Ion? ; and thirdly, from the 
 necessity of distinguishing, for example, tantuS'|sperat, with 
 two /s, from ta"ntum sperat = tantussperat with double s. 
 Hence it will be necessary to practise detached combina- 
 tions of words, still without regard to sense and in the 
 singsong manner, until they can be executed with ease 
 and precision, just like a troublesome passage in music. 
 There is really no difficulty for an Englishman in any one 
 of the sounds to be produced, because they are all 
 existent in his own language. The sole difficulty exists 
 in their being combined in an unusual manner. 
 
 ART. 94. Practise first, final words, as : dolo're (A. 3) 
 labo're (A. n), fiirtl (A. 18), cre'do e'quide (B. 4) which 
 offer no difficulty except in remembering that the m has 
 only to be regarded as a mark to shew that the preceding 
 vowel is long. 
 
 Next, final words "energising" the next consonant, as 
 already cited. 
 
 Then words with enclitics ve, as dolopuve (A. 7). and 
 especially with enclitic que, as uteriicque (A. 20), le- 
 giicque (D. 9), sylvanicque (I. i), certcque (I. 10) pavi- 
 diicque (L. 17). The difficulty here is simply in not 
 inserting the customary m. 
 
 ART. 95. Lastly, the very common cases of joined 
 words must be practised till they run with perfect ease 
 from the tongue. Little phrases must be practised 
 separately, or otherwise no good result will be obtained. 
 Thus infdndurregina (A. 3) et quo'ruppars magna, myr- 
 midonuddolop^ve (A. 7), jannox (A. 8), (quite different 
 from jannox, in which the voice falls on the second syl- 
 lable), ducto'res danau ttot jallabentibus annis (A. 14), 
 instar mdntis equu ddivlna palladis arte (A. 15), (as this is 
 
Art. 95.] FINAL M. 69 
 
 a common example in grammars, it will be very difficult to 
 correct the old bad habits of reading it), vo'tuppro reditu 
 (A. 17) delecta vim ssortl'ti (A. 18), dives opu, ppriami 
 (A. 22), nunc tantussinus (A. 23), quo fe'ssurrapitis (B. i), 
 ingentellu'ctu nne quae're, or inge'ntellu'ctunne quae're 
 (C. i, practise both ways), hunc tanturTata (C. 2), nimiu 
 vobis (C. 3), quantos ille viru mmagnammavortis adurbe 
 (C. 5), quuttiimulu ppraeterlabere or quuttumuluppraeter- 
 labere (C. 7, practise both, the last will be found difficult 
 owing to the number of syllables necessarily kept in 
 a low pitch), quisquaddege'nte (C. 8), intantusspe toilet 
 aVos (C. 9), se tantutte'llus (C. 10), seu quiippe'des iret 
 (C. 13), purpiireos spargafflo'res animacque nepotis his 
 salte^accumuleddoms. 
 
 Also try (HoR. Od. 3, i, 5,) 
 
 re'guttimendo- ruin- propios greges, 
 r'eges in ipsos imperi- uest- Jovis ; 
 
 out of which, by practice, a really majestic effect can 
 be obtained. Dwell on the do and run the ru very 
 briefly on to the /;/, which, without being unduly em- 
 phasised or raised in pitch, must be fully lengthened. 
 The uest following ri, will be found difficult at first. 
 
 The well-known line Aen. 3, 658, is merely a case of 
 omitted m and slurred vowels, and is not likely to 
 occasion difficulty, except perhaps in giving the high 
 pitch to the first syllable in mgens, thus 
 
 mon-strdfhor-ren-dijin-for-m^n-gens cul-lumen ademptu. 
 
 The following are also useful exercises : 
 
 tantus sperabat, tantus speraverat her5s ; 
 tantum sperabam, tantum speraveram et ipse. 
 
 Read the second line : tantusspera/ba, ttantusspera/ve- 
 
70 FINAL M. [Art. 9597. 
 
 raet-ipse, and keep the tusspe, bd ttan, quite distinct from 
 the tus spe, and bat tan of the other line, where the words 
 will distinctly separate s s and / / as two letters, from the 
 double letters ss and //. 
 
 Distinguish, qul musam vidit, musa vidit musam ; as : 
 qul mu'sau vi'dit, musa vi'dit mu'sa, which shews that the 
 mu'sau means musam and the first musa therefore must 
 be musa the ablative (" by means of the muse "), so that 
 the last musa must be the accusative musam. 
 
 ART. 96. These examples will serve to shew how the 
 passages may be written on the black board by the 
 teacher, for pronunciation by the pupil, after he has 
 pronounced them himself several times to the class 
 (Art. 19 note). All pronunciation is acquired by imi- 
 tation, and it is not till after hearing a sound many times 
 that we are able to grasp it sufficiently well to imitate. 
 It is a mistake constantly made by teachers of language 
 to suppose that a pupil knows by once hearing unfamiliar 
 sounds, or even unfamiliar combinations of familiar 
 sounds. When pupils are made to imitate too soon, they 
 acquire an erroneous pronunciation, which they after- 
 wards hear constantly from themselves actually or men- 
 tally, and believe that they hear from the teacher during 
 the small fraction of a second that each sound lasts, 
 and hence the habits of these organs become fixed. Jn 
 the present case both teacher and pupil have been prob- 
 ably for years accustomed to attribute a different sound 
 to m final, to give it the same sound in Latin jam as in 
 English jam, and the utmost care and attention will be 
 required to overcome this habit. 
 
 ART. 97. It would be not a bad exercise to require 
 pupils to write out Latin passages after the model 
 
Art. 97.] FINAL M. 71 
 
 of the preceding, indicating the words that are run on 
 together by writing them together, marking the acute and 
 circumflex accent always, together with the naturally long 
 vowel, writing also the proper sound heard in place of 
 m when it is sounded at all, and omitting it otherwise. 
 Thus the passage (C.) would be thus written : 
 
 6 nate ingentellu'ctu nne quaere tuo'ru, 
 
 ostendent terris hunc tantuffata, nequeiiltra 
 
 esse sinent, nimiu vo'bis romanapropa'go 
 
 visa potens, super!, propriahaec sido'na fuissent. K 
 
 quantos ille viru, mmagnammavortis adurbe 
 
 campus aget gemitus, velquae, tiberine, videbis, 
 
 funera, quuttumuluppraeterlabere recente ! 
 
 nee puer iliaca quisquaddegente lati'nos 
 
 intantusspe toilet avos, nee romula quonda 
 
 ullo se tantuttellus [or tantu ttellus] jactabit ; 
 
 heu pietas, heu prisca fides, invictaque bello 
 
 dextera ! non ill! se quisquaimpune tulisset 
 
 obvius arma'to, seu quiippedes Tret inhoste, 
 
 seu spumantis equi foderet calca'ribus armos, 
 
 heu miserande piier, siqua fataaspera riimpas 
 
 tu marcellus eris. manibus date H'lia ple'ms, 
 
 purpiireos spargafflo'res, animacque nepotis 
 
 his salteaccumuleddo'ms, et fungar ina'm 
 
 mu'nere. 
 
 The writing will fix the sounds and accents in the 
 minds of the pupils. Then several pupils should be 
 called on in succession to read out what one has 
 written, and the others should watch to check them, by 
 ear only, without looking at the writing, which is best 
 effected perhaps by turning the back of the blackboard 
 to the listeners. The master should be just as much on 
 the watch for corrections to be made, as the pupils to 
 make them, and should not be satisfied unless attention is 
 drawn to the error, and the right sound is given, by some 
 
72 FINAL M. [Art. 97, 98. 
 
 of the pupils. Once reached, the right pronunciation 
 should be repeated over and over again correctly, to 
 overcome the former vicious habit. A special study 
 should be made of monosyllabic slurs and assimilations, 
 as : tuintendens turremittens (M. 10), suadmiratus (R. 9). 
 Few perhaps would not hesitate at first over the line 
 quoted from Horace in Art. 90 ; committes reomneet 
 vl'taet cuccorpore fa'ma, as it must be read at first for 
 mere metre. Afterwards the efs will be more separated 
 as : reomne, et vita, et cuccorpore. 
 
 ART. 98. A purely scholastic question, of some im- 
 portance however, arises from teaching grammar to boys. 
 How is the presence of the unpronounced m to be 
 marked ? Should not boys be made to pronounce it, to 
 shew the teacher that they know it to be present ? If 
 a boy repeats amaba, how is the teacher to know that the 
 boy knows the word to be written dmdbam ? Let the 
 classical master remember English. How does the pupil 
 know that thou art, you are, are spelled as they are, when 
 he speaks the ou in thou and you quite differently, and the 
 ar without any e in art in the same way as the are in 
 you are, which has an e ? If the commonest " school 
 board" children and teachers get over this difficulty, 
 what is to be expected from " classical school " children 
 and teachers ? Again let the classical master consult his 
 colleague the French teacher. When a boy repeats in 
 French, je sui, tu e, il e, nou som, vouzet, ilson ; how 
 does he know of the existence of the numerous con- 
 sonants, which on occasion may be very active, as je 
 sidr ici, il son/ ici? Or let him ask himself how, on the 
 present plan of barbaric English Latin pronunciation, 
 the boy knows that vitium has /, convidum has c, 
 
Art. 98100.] L YRIC VERSE RHYTHM. 73 
 
 Elysium has s, when he hears English sh in all? The 
 boy learns Latin by eye chiefly, and hence knows the m 
 from the first, and his main difficulty consists, not in for- 
 getting its presence, but in recollecting it too well. How- 
 ever, the master has an easy plan in speech. Let the 
 boy say : amabam ego, amabam te, amabam vos, amabam 
 matrem. If he uses the same amdba in all cases, he does 
 not know that the word ends in m. But if he says : ama- 
 baego, ama'batte, ama'-bauvos, ama'bamma'tre, (not 
 ama'bam ma/trem) he may be trusted to have a right 
 mental vision of the spelling. It is right, however, that 
 teachers should remember that for this one new difficulty, 
 which is felt as a difficulty only because it is new, thou- 
 sands of facilitations, orthographical, grammatical, ety- 
 mological and metrical, would be introduced by the 
 quantitative system of pronunciation. 
 
 VIII. Elegiac and Lyric Verse Rhythm. 
 
 ART. 99. By this time the pupils should have over- 
 come all rhythmical difficulties of pronunciation so far 
 as length of syllable and pitch accent is concerned. 
 But it will always be safest to exercise the pupil in the 
 singsong fashion and with the pendulum, to acquire any 
 new metre, slowly at first to make sure of each syllable, 
 and then rapidly. The following remarks apply to the 
 acquisition of new metrical systems. 
 
 ART. 100. After hexameters proceed to choriambic 
 verses. These have generally a spondee or iambus to 
 begin and an iambus to end. (E.) and (F.) are very 
 instructive. Observe in (F.) how the division of the 
 
74 LYRIC VERSE RHYTHM. [Art. 100102. 
 
 words varies the three choriamb! in each line, except in 
 (F. 5,) and note the consequent monotony of that line. 
 
 ART. 1 01. Pentameters are best studied as cho- 
 riambic, consisting of two halves, composed (as I view 
 them) each of a dactyle and choriambus, varying into 
 a spondee and molossus in the first half only. Observe 
 also the great difference of the pitch rhythm in older 
 writers, as Propertius (G.), and the stereotyped Ovidian 
 cadence in (H.), where the pentameter must end in a dis- 
 syllable, of which the last syllable comparatively seldom 
 ends in a short open vowel. Nothing serves to shew the 
 great difference of feeling in Greek and Latin rhythm more 
 than the treatment of the final half of a pentameter line. 
 
 ART. 102. Both the Sapphic and Alcaic stanzas, in 
 (L) and (K.), also sound to me principally choriambic. 
 In the Sapphic, the choriamb is preceded by a compound 
 foot, called a second epitrite, consisting of a trochee and 
 spondee, and as a general rule a word ends with the 
 first syllable of the choriamb (which would therefore have 
 a low pitch), though sometimes it ends with the second 
 syllable, as in (I. i, 9, 10.) The common English 
 "swing" with which these verses are read is simply 
 atrocious. 1 Even the final short line gives me the feeling 
 
 other pitches. It is not possible to give the English sounds in Latin letters, but 
 I can fancy Maecenas on waking, after 1900 years' sleep, in an English school, 
 attempting to write the English pronunciation of (I. 2 8), somewhat in this 
 "'"-' ; the final italic m indicating that it was heard : 
 lu' - sidew sl'lai df'kesoco le'ndai 
 
 way 
 
 sem'peret k'ol'tai, de' tiquipri ke''mer 
 
 tem'pori se'kr5. 
 
 quo''sibi lai'nai mdn'niulri ver'sius 
 ver'ziniz lec'tas piu'er5squi cas tos 
 dai'squaibes se'ptew plac'iulri kdl'llz 
 dai'siri l5"dls. 
 
Art. 102104.] LYRIC VERSE RHYTHM. 75 
 
 of the central choriamb, with a syllable over, rather than 
 of a dactyle and spondee. 
 
 ART. 103. The same choriamb or its equivalent 
 molossus, which were both very common combinations 
 in Latin speech, seem also, to me, to give their own pecu- 
 liar character to the Alcaic stanza. Thus in (K.) the 
 choriambs : propositl, prava juben- | tium, men-te quatit-, 
 tur-bidus had- | riae, mag-na manus-, im-pavidum-, and 
 the molossi : in-stan-tis-, il-labatur, are (to me) the pith 
 of the rhythm, the rest is accessory. This of course is 
 far from being the usual view. 
 
 ART. 104. Iambic verses come next to prose, and 
 may be studied in Horace, or Phaedrus, to escape the 
 ancient comic writers, Terence and Plautus, where the 
 rhythm was so difficult to seize even by Augustan 
 Romans, that Cicero felt the necessity of a piper to 
 make him feel it (O. 13-15), and Horace could not scan 
 Plautus, even on his fingers, as we learn by his saying to 
 the Pisones (Ars. Poet. 270 274): 
 
 at vestri proavi Plautlnos et numeros et 
 laudavere sales, nimium patienter utrumque, 
 ne dicam stulte mirati ; si modo ego et vos 
 sclmus inurbanum lepido seponere dicto, 
 legitimumque sonum digitis callemus et aure.- 
 
 Of course he would have no conception that the words were those of his old 
 friend Horace, in a metre invented by Sappho, and it would be real cruelty to 
 enlighten his darkness. The remarkable parts of this method of reading are, the 
 foot of four syllables, the first having both the force and pitch accent, and the 
 last three just audible, and the uniform singsong of the pitch. Such feet and 
 such swing occur only in English sapphics, as in this stanza from Canning's well- 
 known Needy Knifegrinder. 
 
 I'' should be | glad' to | drink' your honour's | health" in 
 A ' pot of | beer, if | you' would give me | six' pence, 
 But* for my | part', I I nev'er love to | med' die 
 
 With' poli | tics', sir. 
 
 Here English accents are magnificently procrusteanised to fit into the school- 
 boy's rhythm. " O, reform it altogether ! " 
 
76 L YRIC VERSE RHYTHM. [Art. 105. 
 
 ART. 105. The putative resemblance of the Latin 
 Iambic to the ordinary English verse occasions much 
 difficulty to an Englishman, because it constantly mis- 
 leads him into violations of the laws of quantity. 1 It is 
 also difficult to mark the time in Iambic verse, because 
 the length of each foot is generally not constant. In 
 Catullus we have indeed examples of pure Iambic verses 
 in the poems 
 
 iv. phaselus ille quern videtis, hospites 
 
 ait fuisse navium celerrimus. 
 
 xxvii. (xxix.) quis hoc potest videre, quis potest pati 
 nisi^impudicus et vorax et aleo ? 
 
 But generally a spondee was allowed in the first, third, 
 or fifth foot. The effect of these spondees requires 
 careful consideration. That it was very marked and im- 
 portant is clear from the words of Horace who calls the 
 iamb a " rapid foot," and says that originally every foot 
 in the senarius was an iamb, but that " not long since,*' 
 the verse admitted u steady spondees " to make the lines 
 " slower and more solemn," but that Accius seldom used 
 this liberty, although Ennius either from carelessness or 
 
 1 English verse is not regulated by the length of the syllables it contains, as 
 Latin verse is ; although length, of syllables frequently produces rhythmical 
 effects, which shew the master hand, just as pitch-accents embellish Latin 
 rhythm. If we consider that the names of Latin feet refer to combinations of 
 long and short syllables exclusively, tjien it will be seen how inappropriate they 
 are to English measures, or combinations of strong- and -weak syllables ex- 
 clusively (using these adjectives as expressing greater and smaller amounts of 
 force). If we persist in using an iamb in English for the measure of a weak 
 syllable followed by a strong one, as awa'ke, and appropriate the other names in 
 the same way, let us at least prefix such a term as force, and speak of awa'kf 
 being a force-iamb ; slee'per a force-trochee, ver'ity ^force-dactyle, and so on. 
 Or, to avoid repetition, preface any such use of the names of feet by the notice : 
 "names of quantitative feet are here employed solely for force-measures'* 
 There is more than a mere question of name here. The application of the old 
 name to the new case arose from an absolute confusion of ideas, which the con- 
 tinuance of the custom perpetuates. 
 
Art. 105, 106.] LYRIC VERSE RHYTHM. 77 
 
 ignorance of his art allowed his lines to be " oppressed ;> 
 by spondees. The whole passage is of great importance 
 as shewing the sensible effect of spondees as contrasted 
 with iambs. 1 The iamb itself, by merely resolving its 
 long syllable into two short, might be replaced by a tri- 
 brach, in all places but the last, and the spondee by 
 a dactyle or anapest. This occasions the difficulty in 
 counting time, for the number of short lengths in a foot 
 was variable. In (L. i, 7) we have pure iambics, the only 
 cases in this example. One spondee occurs in (L. 2, 3, 
 4, 6, 8, 9, 10, n, 12, 14, 16, 18, 19.) Two spondees 
 occur in (L. 5, 13, 15, 20, 21, 22.) A tribrach is found in 
 (L. 8) jace | re modo, and (L. 17) pavidum | que lepo | re m . 
 A dactyle occurs in (L. 14) aut ami | te. And two anapests 
 as well as a tribrach and a slur occur in (L. 17) pavidum | 
 que lepo | re m et ad | venam | laqueo | gruem, and one 
 of these occurs very unusually in the fifth foot. 
 
 ART. 1 06. Phaedrus allows iamb, tribrach, spondee, 
 dactyle or anapest in any place, except the last. The 
 comedians are even more dreadful. The effect to me is 
 
 1 syllaba longa brevl subjecta vocatur iambus, 
 
 pes citTts; unde etiam trimetrls accrescere jussit 
 
 nomen iambeis, quum senos redderet ictus 
 
 primus ad extremum similis sibi : non ita prtdem 
 
 tardior ut paulo gravidrqut veniret ad aures, 
 
 spondeos stabiles in jura paterna recepit 
 
 commodus et patiens, non ut de sede secunda 
 
 cederet aut quarta socialiter. hie et in Acci 
 
 nobilibus trimetrls appdret rdrus, et EnnI 
 
 in scenam missos cum magno pondere versus 
 
 aut operae celeris nimium curaque carentis, 
 
 aut ignoratae premit artis crlmine turpl. 
 
 A. P. 251262. 
 
 Observe here the use of ictus as denoting a metrical interval, and the implica- 
 tion that the length of these intervals varied when the spondee was introduced, 
 with which compare Quint. 9, 4, 51, who speaks of intervals between the beats 
 (percussiones) of four and five units (o-n^ela), so that classical time-keeping was 
 very different from our conductor's baton. See (Art. 14). 
 
78 LYRIC VERSE RHYTHM. [Art. 106. 
 
 then merely prose spoiled with neither the melody of 
 verse nor the fine flow of prose, reminding me forcibly 
 of my own boyish attempts at blank verse, when I con- 
 sidered it to be simply prose chopped up into lengths of 
 ten syllables. This may be very unscholarlike, but it is 
 a comfort to err on such a point with Cicero and Horace. 
 If the older writers had a really intelligible metre (as we 
 cannot but believe they had), then their language and its 
 pronunciation were entirely different from the Augustan 
 Latin (which is otherwise extremely probable), and we 
 have not as yet acquired a sufficient insight into it. The 
 English scansions I have heard attempted viva voce, 
 were simply impossible and intolerable, and emendations 
 proposed upon the hypothesis of their correctness are to 
 me by that very hypothesis discredited. There was the 
 same feeling at one time respecting Chaucerian metres, 
 till a key to his pronunciation was found. We must 
 wait for a similar key to Plautus and Terence. At least 
 I have not found it in Corssen. At present then I can 
 only recommend these older writings to be read in the 
 Augustan pronunciation as semi- versified prose, knowing 
 indeed that this must be considerably wrong, but feeling 
 that it cannot be so dreadfully wrong as our present 
 habits. Reading Phaedrus rhythmically may be a useful 
 introduction. He certainly felt the rhythm himself, or he 
 could not have commenced by saying 
 
 Aesopus auctor quam materiam repperit 
 hanc ego polivl versibus senariis ! 
 
 The first line, with its dactyle in the fourth place 
 (whence Horace refuses to oust the iamb) is a study in 
 itself. The number of feet and the final iamb are the 
 only remaining marks of the iambic trimeter. 
 
Art. 107, 108.] PROSE RHYTHM. 79 
 
 IX. Prose Rhythm. 
 
 ART. 107. Prose is the most difficult thing to read in 
 any language. We allow much to the swing of verse, but 
 if the prose reader does not bring out the eloquence of 
 the original, and make all the points as he goes on, 
 pronouncing purely, accurately, and distinctly, and pre- 
 serving the national custom of intonation, at the same 
 time, we think little of him. The passages quoted from 
 Cicero in the Appendix (M., N., O., P., Q., R.), will serve 
 to show the extreme importance which he attributed to 
 prose rhythm and to its clear separation from verse 
 rhythm. 
 
 ART. 1 08. It is curious to see him in (M.) finding the 
 Latin system of intonation the only natural one. The 
 Latins and Greeks indeed agreed in not placing the 
 highest pitch at a greater distance than the third syllable 
 from the end. but the Greeks allowed it on the last syl- 
 lable, (as in the word in O. 9) and the Latins, at least 
 in Quintilian's time, did not. 1 Still the absolute fixity of 
 the Latin custom is proved by Cicero's view that it was 
 natural. In (R.) we have it confirmed by the story of 
 Gracchus's piper, in (N.) the same steady observance of 
 quantity is shewn, and in (O.) and (P.) it appears that in 
 prose as well as in (at least iambic) verse, rhythm was 
 
 1 See the decisive passages quoted from Quintilian (Art. 41 note). When 
 Priscian admits the accent to be placed on the last syllable " discretions causa," 
 his accentus was no longer Quintilian's tenor, and Cicero's acuta et gravis vox. 
 Hence all the indications cited by Corssen (II. 808, et sqq.) are inapplicable to 
 Augustan pronunciation, that is, to our present investigation. There is nothing 
 so unscientific in historical investigations of pronunciation, as the confusion of 
 periods. No attempt should be made to venture on this dangerous ground in 
 our imitations of Augustan Latin. 
 
8o PROSE RHYTHM. [Art. 108, 109. 
 
 chiefly regulated by the quantity of the few syllables 
 towards the close of a clause. 
 
 ART. 109. In other places Cicero cautions the 
 speaker of prose from falling into verse. Thus quoting 
 Aristotle approvingly with a slight change, he says (Or. 
 194) : 
 
 Iambus enirn et dactylus in versum cadunt maxime ; itaque ut 
 versum fugimus in oratione, sic hi sunt evitandi continuati pedes. 
 Aliud enim quiddam est oratio, nee quidquam inimicius quam ilia 
 versibus. Paeon aurem minime est aptum ad versum ; quo liben- 
 tius eum recepit oratio. 
 
 And then, from himself he says (Ib. 195) : 
 
 Ego autem sentio, omnes in oratione esse quasi peraristos et con- 
 fuses pedes. Nee enim effugere possemus animadversi5nem, si 
 semper isdem uteremur. Quia neque numerosa esse, ut poema ; 
 neque extra numerum, ut sermo vulgi est, debet esse oratio. Al- 
 terum nimis est vinctum, ut de industria factum appareat : alterum 
 nimis d'ssolutum, ut pervagatum, ac vulgare videatur : ut ab altero 
 non delectere, alterum oderis. Sit igitur permista et temperata 
 numeris, nee dissoluta nee tota numerosa. Paeone maxime (quo- 
 niam optimus auctor ita censet) sed reliquis eliam numeris quos 
 ille praeterit, temperata. 
 
 The paeon has three short syllables and one long, 
 differently distributed, and Cicero really dissents from 
 Aristotle as to its use (P. 9, 10). No doubt this de- 
 pended on the difference between Greek and Latin 
 intonation. The great value of these passages to us is 
 to shew that quantity was the only recognized guide to 
 rhythm in prose and verse. Quintilian thoroughly 
 agrees with Cicero's view, saying : 
 
 In compositione orationis certior et magis omnium aperta servan 
 debet dimensio. Est igitur in^pedibus, (9, 4, 52. ) Ratio vero pedum in 
 
Art. 109, no.] PKOSE RHYTHM. 81 
 
 oratione est multo quam in versu difficilior : prlmum quod versus 
 paucis continetur, oratio Iongi5res habet saepe circuitus [never 
 ' circuwitus,' see p. 65]: deinde quod versus semper similis sibi 
 est et una ratione [unaltered measurement] decurrit, orationis 
 compositio, nisi varia est, et offendit simultudine et in affectatiSne 
 deprehenditur. Et in omni quidem corpore totoque (ut ita dixerim) 
 tractu numerus insertus est. Neque enim loqui possumus nisi syllabis 
 brevibus ac longis, ex quibus pedes fmnt (9, 4, 60. 61.) 
 
 ART. i ic. To teach a person to read prose well even 
 in his own language is difficult, partly because he has 
 seldom heard prose well read, though he is constantly 
 hearing prose spoken around him, intonated, but un- 
 rhythmical. In the case of a dead language, like the 
 Latin, which the pupil never hears spoken, and seldom 
 hears read, except by himself or his equally ignorant and 
 hobbling fellow-scholars, this difficulty is inordinately 
 increased. Let me once more impress on every teacher 
 of Latin the duty of himself learning to read Latin 
 readily according to accent and quantity ; the duty of his 
 reading out to his pupils, of his setting them a pattern / 
 of his hearing that they follow it, of his correcting their 
 mistakes, of his leading them into right habits. If the 
 quantitative pronunciation be adopted, no one will be fit 
 to become a classical teacher who cannot read a simple 
 Latin sentence decently with a strict observance of that 
 quantity by which alone the greatest of Latin orators 
 regulated his own rhythms. We have by this time also 
 probably learned to acknowledge that the introduction 
 of a pitch accent, that is, the elevation of the pitch of 
 the voice on the so-called accented syllable, and its de- 
 pression on the other syllables, even in interrogative 
 sentences, is quite as essential to the feeling of that part 
 
 1 See Art. 19, note, and Art. 96. 
 
82 PROSE RHYTHM. [Art. 110, in. 
 
 of Latin rhythm which depends npon caesura, or the 
 division ot words ; and that the adoption of our English 
 freedom of pitch, or rising inflection in questions, must 
 have been as disagreeable to an Augustan Roman, as the 
 Scotch or French intonation of English is to a Londoner. 
 Think how Shakspere's lines would fall from the mouth 
 of a Frenchman ! There are some Frenchmen (as 
 Fechter) that give our English rhythms far better than 
 we English can hope to give the Latin rhythms ; but even 
 in them we at once detect the foreign intonation which 
 destroys the genuine roll, as we like to hear it though 
 certainly our own modern speech would have been very 
 thin and poor, effeminate and affected, in Shakspere's 
 own ears. But only fancy a Frenchman declaiming 
 Shakspere with his own values of the vowels, his own 
 curious use of the force and pitch accent and emphasis, 
 his own treatment of quantity, and his own intonation ! 
 Would not every Englishman stop his ears and flee? 
 And this is but a faint shadow of the atrocious manner in 
 which we have hitherto dared to treat Virgil and Horace 
 and Cicero ! 
 
 ART. in. My remarks have been directed to the 
 case of a transition, where the pupils have been hitherto 
 accustomed to our vile English pronunciation of Latin, 
 and I have also taken into consideration the difficulties 
 which the teachers themselves have to overcome in 
 unlearning the old and learning the new. When the 
 teachers are able to read with instinctive fluency, and 
 begin with young boys, making them read by quantity 
 and pitch from the first, the whole matter is much sim- 
 plified. Care should be taken that all long vowels are 
 properly marked in the school books. There is no 
 
Art. ill, 112.] PROSE RHYTHM. 83 
 
 occasion to mark unpronounced or assimilated ///, or the 
 short vowels ; and I cannot sufficiently reprobate the 
 usual custom of marking a vowel as long, when all we 
 know is that the syllable containing it is long, owing to 
 a concurrence of consonants (see Art. 23, note). The 
 teacher must carefully read out the words to be learned. 
 He has to become a teacher of reading, and must recog- 
 nise the responsibilities and difficulties of that office. Of 
 course, such teaching begins with prose and very prosy 
 unrhythmical prose ; but the length and pitch of each 
 syllable can be scrupulously observed even in declining, 
 musa, mu'sae, and conjugating, amo, amas, amat. Indeed 
 if a pupil can regularly and securely mark the pitch 
 accent and quantity in every form of the usual paradigms 
 given in grammars, he will have very little difficulty in 
 what follows. But when boys are allowed to say (using 
 Latin letters to express our English pronunciation) : 
 emo ! emas' emaf eme'mes emetis eman't, emebanr 
 emebas* einebaf am-ebe'mes am-ebetis emeban't, and so 
 on, the master is laying up a store of difficulties for the 
 future (see Art. 35 ; see also Art. 98.) 
 
 ART. 112. The great question for an Englishman is, 
 what shall he do with his own force accent? The 
 answer is pretty much the same as for French Put it by, 
 and say as little about it as possible. An Englishman 
 cannot avoid, and has no occasion to avoid, speaking 
 some syllables forcibly and emphatically ; but he must 
 never allow that force, as in his own language, to alter 
 the relative length or pitch of the syllables, or the purity 
 of the vowel sounds. Taking these precautions, at- 
 tending most scrupulously to these points, he may do 
 pretty well what he pleases with the force accent. How, 
 
 G 2 
 
PXOSE RHYTHM. 
 
 [Art. 112. 
 
 > o 
 
 when the feeling for quantity dimmed, pitch accents be- 
 came gradually converted into force accents, does not 
 concern us at present, for we must suppose that quantity 
 and pitch accent are in full force, and must make them 
 live in our imitation of Augustan speech. 1 It will be 
 
 1 The feeling for quantity seems to have gone first, while the consciousness of 
 the accented syllable remained. The nature of the accent then became indifferent. 
 Possibly raised pitch and increased force had for some time gone regularly 
 together, and as their combination required greater exertion, this very effort 
 assisted in impairing the feeling for quantity. The modern Greek, the modern 
 Italian, and modern Spaniard, seem not to know the meaning of fixed quantity. 
 In their languages quantity is now as variable as pitch, but force is fixed upon 
 certain syllables. Corssen (II. 942) quotes some Latin hexameters of the latter 
 part of the third century, in which it is evident that all feeling for quantity had 
 died out. In my paper on Accent and Quantity (Philol. Trans. 1873 4, p. 153), 
 I have compared these with lines of Virgil which have almost the same rhythm 
 of force accents as they would have been read with at Eton when I was there 
 (1830 3). The same result remains if we use the Augustan pronunciation, 
 neglecting distinctions of quantity, but preserving the place of the force accent. 
 I add the comparison here, representing the force accent by a turned period as 
 usual, and omitting all marks of quantity and pitch accent. And I also add 
 some English lines (of about the same calibre, that is, nearly "nonsense verses,") 
 but of precisely the same rhythmical construction, which may facilitate the com- 
 parison. The hexameter lines, I. to V. are subdivided because the force rhythm 
 is thus better exhibited. The numbers i, 2, 3, refer to the original, the English, 
 and the Virgilian lines : 
 
 I. i. Praefa'tio nos'tra 
 
 2. Irra'tional doc'trines, 
 
 3. Excu'tior sonvno, et 
 
 ^ i. vi'am erran'ti demon 'strat, 
 
 2. held '-by believers as-per'fect, 
 \^ ^M* 3. sum 'mi fasti 'gia lecti. Ae. 2, 302. 
 
 H * II. i. Respectum'que bo'num, 
 
 ^ 2. Are-but-arrant foHies 
 
 3. Praesentem'que vi'ris 
 
 1. cum-ve'nerit sae'culi me'ta, 
 
 2. to-those'-who-have stud'ied-the sub'ject. 
 
 3. inten'tant om'nia mor'tem. Ae. i, 91. 
 
 III. i. Aeter'num fi'eri, 
 
 2. Misfortune teach 'es-us 
 
 3. Carpe'bant, hy'ali 
 
 1. quod-discre'dunt in'scia cor'da. 
 
 2. that-the-best'-are of'ten-in error. 
 
 3. sat'uro fuca'ta colo'ri. Geo. 4, 335. 
 
 tc 
 
Art. 112, 113.] LATE LATIN. 85 
 
 found that much effect is often given by a force accent 
 on a syllable with depressed pitch ; and at least as much 
 variety can be produced by judicious variations of force 
 in Latin, as we give in English by judicious variations of 
 pitch, 
 
 X. How to Read Late Latin. 
 
 ART. 113. A question arises as to the proper method 
 of reading late Latin, written and pronounced by the 
 writers without attention to pitch accent or to quantity. 
 The answer is very simple. Only one pronunciation can 
 be taught in schools, and that should be the best imita- 
 tion we can obtain of the Augustan Latin. If we ad- 
 mitted the principle of using the pronunciation of the 
 later writers, we should probably have to learn a mul- 
 titude of different pronunciations at least one for each 
 century, and one for each native country of the writer. 
 
 IV. i. E'go simiHter 
 
 2. E'ven nobil'ity 
 
 3. Qua'lis popu'lea 
 
 1. erra'vi tern 'pore mul'to 
 
 2. of-heart '-and sour-may-be cheat 'ed. 
 
 3. mae'rens philome'la sub um'bra. Geo. 4, 511. 
 
 V. i. Fa'na prosequen'do, 
 
 2. Sense 'less adora'tion 
 
 3. Hu'jus odora'to 
 
 1. paren'tibus in'sciis ipsis. 
 
 2. of-all '-that-is old'-is-the keystone. 
 
 3. radi'ces in'coque Baccho. Geo. 4, 279. 
 
 It is evident that it is our ratio and not our aures which, on the old plan of 
 reading, would find the first lines of these triplets full of false quantities, and the 
 third lines proper hexameters. But read them by the rules of Augustan pro- 
 nunciation, and Virgil's lines have a regular appreciable rhythm, while the 
 others become a mere jumble. It is evident then that we can learn nothing of 
 Augustan usages from the tertiary strata which produced the above hideous 
 fossils. In the words of the accentual hexameter I, " viam errant! demonstrant," 
 they shew the way ycu're not to go ! 
 
86 FINAL METHOD OF READING. [Art. 114, 115. 
 
 ART. 114. As far as prose is concerned, there is 
 evidently no objection to using ab near an approach to 
 Ciceronian speech as we can compass. As regards all 
 literary verse, it is written on Augustan models down to 
 yesterday's school exercises all over the world, and cannot 
 be read rhythmically except in Augustan pronunciation. 
 It may indeed not be possible to read it rhythmically at 
 all, but that is the fault of the author's skill, not of his in- 
 tention. On the other hand, medieval hymns (Art. 57) 
 and other Latin verses, like Father Prout's 
 
 Quam pul'cra sunt o'va 
 Cum tosta et no'va 
 
 E stabtilo sci'te leguirtur, 
 Et a Margery bel'la, 
 Quae festi 'va puel 'la ! 
 
 Pirrguis lar'di cum frus'tris coquun'tur, 
 
 may be read with force accents only, without regard to 
 quantity or slurring, and with English final m, but other- 
 wise with the pronunciation of the letters here assumed. 
 It is only in merely farcical mixtures of English and 
 Latin, or some other modern language and Latin, that 
 the pronunciation of the modern language has to be 
 adopted, but with these we have really nothing to do. 
 
 XI. Final Method of Reading Latin. 
 
 ART. 115. After the pupil has thus acquired the full 
 feeling of the rhythmical construction of any passage by 
 learning to feel the length and musical pitch of its 
 syllables, he has to clothe this bare skeleton with living 
 flesh and make it talk Augustan sense in Augustan 
 
Art. 115-117.] FINAL METHOD OF READING. 87 
 
 speech. Of course a perfect comprehension of the 
 passage and its bearings is necessary for this purpose, 
 and some theory must be formed of the Augustan 
 method of intonating phrases as well as words, and 
 of general delivery (actio, pronuntiatio). On this point 
 there is much to read in Cicero and Quintilian with 
 which a teacher should become familiar. But through 
 all varieties of tone (omnes sonorum gradus, M. 9) the 
 reader must bear in mind relative quantity and relative 
 pitch so far as the syllables of a single word are con- 
 cerned. He must be, as an Englishman, particularly 
 careful not to raise his voice on the last syllables of ex- 
 clamations and questions as in : e'ja ! quid sta'tis ? (D. 18, 
 19) which he will be very apt to read : e'ja ! quid stasis ? 
 
 ART. 116. It is recommended that short pieces 
 should be gradually practised perfectly. The younger 
 pupils, who have not the power of mind or experience to 
 originate for themselves, should follow patterns set by the 
 teacher (Art. 19), and committed to memory by the 
 pupil. So far as verse is concerned, a power of repeating 
 from memory the examples in the Appendix will be of 
 great service. I have found that by constantly repeating 
 them, either mentally or aloud (as in country walks), 
 with the best methods of delivery I could call up, I have 
 gained more knowledge of the possible life of Augustan 
 poetry, than I ever possessed during my school days at 
 Shrewsbury and Eton, or my college days at Cambridge. 
 And I have thus been able to convert dead signs into 
 real living sounds, such as may have moved the hearts 
 and thoughts of men of old. 
 
 ART. 117. The prose part of the examples is deficient 
 in passages of varied feeling and declamation ; it is 
 
88 FINAL METHOD OF READING. [Art. 117119. 
 
 almost entirely didactic. My wish was to give the most 
 important words of the most important writer who had 
 treated of prose rhythm, and to shew by them the com- 
 plete subservience of Augustan prose rhythm to quantity. 
 But the reader will find no difficulty in turning up 
 passages in Cicero's orations or Livy's histories, which 
 involve the highest declamatory powers, while the fami- 
 liar letters of Cicero may be read as actual communi- 
 cations with his friends, and his " Terentia et Tulliola 
 duabus animis suls." 
 
 ART. 118. In reading the paper on which the present 
 tract is founded, I delivered the whole of the following 
 passages in the most varied way I could compass, not, as 
 I explained, for the purpose of shewing off my own 
 powers of reading, but to enable the classical teachers 
 then present, to realise, if but for a short time, my notion 
 of the nature of living Augustan Latin in its various 
 phases, in its various moods, tones, pitches, qualities of 
 voice, in short, to bring back, so far as I could conceive 
 it, the old ring of Augustan speech, and to shew that my 
 rules were not dead pedantry, but a living breath. I 
 think, therefore, that it may be convenient to add a few 
 remarks on the way in which I endeavoured to read 
 these examples, as a guide to the teacher who wishes to 
 use them as patterns for his own pupils. 
 
 ART. 119. And first I would observe, that, at least in 
 my day, a schoolboy's repetitions were " gabbled ;" that 
 the master's sole interest seemed to be that the boy should 
 shew he remembered the words, without any regard to 
 sense or style of delivery ; a falsely -placed accent (styled 
 a " false quantity ") might be observed and punished, but 
 nothing else. The dreary drone, the hesitation, the 
 
Art. 119.] FINAL METHOD OF-READING. 89 
 
 repetition of words, the humming and hawing, the cast- 
 down eyes, the depressed figure, the weary tone, are 
 dismal recollections of my youth. Now we do not want 
 to make actors or orators of the boys, but we should 
 make them read decently, intelligibly, with the just sound 
 of each syllable both in length and pitch, and if possible, 
 with an indication of the characteristic quality of tone 
 by which joy and grief, rage and joke, argument and 
 feeling, are distinguished. More has to be done for 
 "speech days " of course, but I am talking of ordinary 
 repetition, which should be made a lesson in distinct 
 utterance, and for that purpose the speaker should be 
 placed as far off as possible from the master, not close 
 to him, as so often happens when a boy is " called up," 
 a custom partly due to our inherited habit of hearing 
 many classes at once in one large hall, instead of in 
 separate rooms, so that the master can carry out 
 Quintilian's recommendations : 
 
 Imprimis vitia si qua sunt oris [of pronunciation] emendet, ut 
 expressa sint verba, ut suis quaeque litterae sonis enuntientur. Ciira- 
 bit etiam, ne extremae syllabae intercidant ; ut par sibi sermo sit ; 
 ut recta sit facies dlcentis, ne labra distorqueantur, ne immodicus 
 hiatus rictum discindat, ne suplnus vultus, ne deject! in terram oculi, 
 inclmata utrolibet cervix. Nam frons pluribus generibus peccat. 
 Vidi multSs, quorum supercilia ad singulos vocis conatus allevaren- 
 tur, aliSrum constricta, aliorum etiam dissidentia, cum alterum in 
 verticem tenderent, altero paene oculus ipse premeretur (i, n, 4. 8. 
 9 ii). 
 
 Boys we see have not much changed since Quintilian's 
 time. I have myself seen in English schools every fault 
 he mentions, and no doubt every classical teacher will 
 recognise the picture. But he should do more. He 
 should paint it out. 
 
9 POETICAL EXAMPLES. [Art. 120, 121. 
 
 XII. How to Read the Poetical Examples in 
 the Appendix. 
 
 ART. 120. Bring out the character of each foot and 
 each line distinctly, but gently, with a slight flexion of 
 voice, never degenerating into a regular chant, and never 
 interfering with the sense. Study a few lines at a time, 
 and repeat them till you are quite perfect. This is es- 
 sential at first, as otherwise bad habits will be acquired 
 which can scarcely be thrown off. The grand sonorous 
 swing of the Virgilian hexameter, the prettiness of the 
 Ovidian elegiac, the playfulness or intenseness of the 
 Horatian lyrics, the colloquiality of the satirical hexa- 
 meter and iambics, must all be characterised, and all be 
 well distinguished from prose by their unmistakable 
 "pede certo." 
 
 ART. 121. Example A. The two first lines are prose 
 in sense and verse in form. They must be spoken 
 rhythmically, but levelly, quietly, very distinctly, and 
 without the least haste, as an evident prologue to what 
 follows. In (i) conticuereomnes, the con- must be dis- 
 tinctly long, the pitch of the voice must be kept low till 
 e and then allowed to rise and fall again. Two English 
 faults must here be guarded against ; no sound of /, or 
 indistinct English u should follow e, or be inserted 
 before r, which should come clearly down upon e in re 
 continuing the descent of the voice, but the syllable re 
 is held for a very short time, and the pitch rises at once 
 for 6m. The nes should be full, clear, low, and strong, 
 and there should be no degradation of s into z. The 
 
Art. 121.] POETICAL EXAMPLES. 91 
 
 following molossus intenti must have the two first syl- 
 lables kept at a low pitch, so that the line is made very 
 quiet by the three long low syllables nes inten. Observe 
 that the pitch rises and falls on the last syllable ti of the 
 spondee, and in the syllable qued, it has first a low, then 
 a high, and finally a low pitch. 
 
 The rhythm of (2) is brought out by dividing : inde- 
 toro pateraene'as sic orsusabalto. Be particular in the 
 opening choriamb, giving full length to the first and last 
 syllables. Run the pater shortly and clearly on the 
 aene'as, and be very careful in keeping pa short and ae 
 long. We are so accustomed to talk of pe'tzririi'ds 
 (where 9 indicates our peculiar vowel in final -er\ that 
 much practice is necessary to overcome the difficulty. 
 Mind that sic is English seek, not English sick. 
 
 Lines (3 13) are an excuse. /Aeneas is full of painful 
 reminiscences.) He begins slowly, in a dull voice : in- 
 fandurregl'na ; mind the opening long in, and the as- 
 similated m. This double rr is a difficulty to be much 
 practised. Mind the long re; there is a great tendency 
 to shorten it. In (4) mind the molossus trojd'nas, and 
 the anapest utopes^ and then the three long low syllables 
 et lamen with the mournful rise on to! (be sure not to 
 say lam or tab). The choriamb and anapest which open 
 (5): enierint danai, will occasion difficulty; they must 
 be quite mournful, with the long e well brought out ; the 
 voice rises for an instant, on ru da, and the main expres- 
 sion is obtained by sinking the voice on the final long J. 
 The opening mournfulness continues to this point. 
 Then, Aeneas thinks of himself and his voice becomes 
 firmer. In the end of (5), ipse and miserrima have to 
 be especially marked, the last most ; in (6) mdgna is the 
 
92 POETICAL EXAMPLES. [Art. 121. 
 
 chief word, but be careful with the assimilated m, the 
 trilled r, and the hard s in : quc'ruppars. In ful the 
 quantitative iamb must be well brought out without any 
 tendency to a force-spondee, and the I made long and 
 low, without particular force. 
 
 In (6 8) another chord is struck. The tale of woe 
 is so sad that even Myrmidons and Dolopes, and harsh 
 Ulysses' men would weep over it. Bring out the re- 
 ference to these points, and especially emphasise dun 
 (7). Line (8) begins with a dactyle and choriamb, it is 
 much like (4), the voice sinks on the mis of Mcrymis, 
 not rises, although interrogative. The a must be kept 
 long and low, without any force. 
 
 Lines 8 and 9 merely continue the excuses for not 
 telling the story which is forthwith continued through 
 two books of the Aeneid. The tone implies "and 
 you see that it's so late to-night, there will not be time." 
 Care must be taken with jdnnbx, with the long prae in 
 praecipitat, and the pure / in cadentia. 
 
 In (10, n) we have a broken sentence, the verb is 
 missing. Begin with sed, si, as quite distinct syllables, 
 bringing out the initial spondee ; tdntus is the chief 
 word j in (n) breviter (with careful er in erring not zr, 
 and well trilled r\ must have a little force. Take 
 breath after trojae to allow of uniting sup-re 'muaudire- 
 labo're, with a little emphasis on supre'mu. Observe 
 that the slurred vowels uau are all in low pitch. In (12) 
 bring out the hbrret, making its second syllable, as well 
 as the following luctu, very long, and lead up to refugit. 
 Take breath after animus, and horret. Make a little 
 pause after refugit, and speak the incipid, with the tone 
 of a victim to circumstances, making n in in very long, d 
 
Art. 121.] POETICAL EXAMPLES. 93 
 
 clearly high pitched but short, and the a rather indif- 
 ferent. Then pause. 
 
 Begin the narrative (13) in a quiet and altered tone. 
 Aeneas has now made up his mind to unbosom, and 
 does so without stint. The two first lines (13, 14) are 
 very quiet therefore, but observe ffrdctl to compensate 
 for omitted m, and oppose distinctly, but very quietly 
 the two sentiments, ffrdctlbello (keeping I and o quite 
 long) and fatl'sqiierepulsl (fa and tls both quite long). 
 In (15) bring out mbntis, but let equu be quiet, while the 
 three last words ddlvi'na pdlladis drte (with dd) are 
 parenthetical, and spoken with a tone implying that the 
 Greeks could not have done it without such assistance. 
 In (16) keep the first ae quite long, and observe dbjete 
 for abiete, the position of high pitch being changed with 
 the change of ie into je for the sake of the verse, com- 
 pare Quintilian on voluc-res in Art. 24, note. 
 
 Line (17) is also parenthetical. The first clause is 
 spoken with a kind of contempt for the trap into which 
 the Trojans fell. The eafama vagatur, is a mere obiter 
 dictum, and, though important to the subsequent narra- 
 tive, must be passed over lightly. 
 
 Lines (18 20) are simple narrative. Observe viru, 
 and guard against English vir-, also ss in ssorti'tl to 
 compensate for omitted m. Be careful with the molossi, 
 includunt (19), and ingentes (20). 
 
 In (21) a completely new subject is taken up. The 
 first words are as quiet as a guide book, and the diffi- 
 culty consists in giving the many spondees their full 
 length, when they convey so little matter. But just for 
 that reason I consider this line to be one of the most 
 useful quantitative exercises here given, and it should 
 
94 POETICAL EXAMPLES. [Art. 121, 122. 
 
 be well studied. The short te with high pitch is very 
 trying, because it begins a significant word, and the 
 chief importance must be given to its final low-pitched 
 but long syllable dos, with short vowel. In (22) mind 
 divesbpu, and the following parenthesis, which must be 
 spoken with a special reference to ancient splendor, 
 and thus contrasted with (23) which shews the mournful 
 results of its wreck. Pause after nuric, make tdntum 
 important, and sinus short and distinct, without separa- 
 tion, tdntussinus, except by the raised pitch. The long 
 o in stdtio gives this word its chief force. 
 
 Lines (24) and (25) require no notice beyond what they 
 have already received (Arts. 47, 49, 58*:, 60.) They join 
 on the description of Tenedos to the following narrative. 
 
 ART. 122. Example B. Although these are consecu- 
 tive lines in the original they belong to two distinct 
 subjects, and have been introduced especially for the 
 sake of (2). Anchises has been passing in review a 
 large number of the future heroes of Rome, and had 
 been asking how he could omit naming this one and 
 that, when he sees the Fabii, and asks how far they 
 would lead him in his weariness, so that he merely 
 notices Fabius Maximus Cunctator. In reading, observe 
 the fessurrdpitis, the low pitch on i in fabil, (where we 
 English use a high pitch,) the emphasis on tfi, the treat- 
 ment of Maximus as a name, as there had been several 
 of that name, but he was the only one, unus (keep the 
 us long) who by such an act as cunctdndo (which should 
 be very significantly spoken as quite contrary to Roman 
 elan) would save the state (re). The qulno'bls should be 
 long and even, as the voice must be prepared for the 
 cunctdndo, before which it is best to take breath. The 
 
Art. 122.] POETICAL EXAMPLES. 95 
 
 two last words form the difficulty. The rhythm has 
 been already alluded to in Art. 50. The res must be 
 quite long, and low, but strong ; the sti high and short, 
 descending rapidly to the es, when, the voice unexpectedly 
 and unusually rises in re, only to fall again in the same 
 syllable. 
 
 The last seven lines are a kind of Roman anticipation 
 of " Rule Britannia ! " They must have been repeated 
 over and over again with extreme unction 1 by warlike 
 political Romans, who despised the versatile scientific 
 Greek. They contain the glorification of a system of 
 repression, conquest and government, as opposed to all 
 that was lofty in art and science, and they must be 
 spoken as if the speaker fully entered into the sentiment. 
 The dlil are of course the Greeks, whom Anchlses does 
 not even deign to name. This word dlil must have a 
 chief emphasis, which can be principally produced by 
 lengthening and reinforcing the final /, and as this is 
 quite opposed to our English e'liai^ it will require care. 
 The opposition to dlil is ///, which must be brought out 
 with great emphasis, and breath should be taken after 
 it. The other important words are Me and dries in (8). 
 The tone of the first lines is given by the credoequide (4) ; 
 "of course, who disputes it? but what matter? be it 
 so. they are smarter and 'cuter, but we can thrash 
 'em ! " Hence lines (3) to (6) have to be spoken in a 
 sort of depreciatory high and hollow tone. At the same 
 time great care must be taken with the opening molossl, 
 (excu'dent, ora'bunt, descrl'bent). In beginning (7) the 
 tone of voice changes, and becomes round, solemn and 
 assertive ; tu " this is what you have to ; " (take breath 
 
 1 In anything but the Roman Catholic sense of the words. 
 
96 POETICAL EXAMPLES. [Art. 122. 
 
 here,) regereimp'erio pbpulos. The high-pitched syllables 
 are all short, the whole force of the sentence depends 
 on the length and delivery of the three long syllables 
 im, o, os. The last words of the line are light, because 
 they are so obvious that the listener supplies them at 
 once, though they are still solemn, as befits an injunction. 
 Fancy how a whole forum of Romans would rise at the 
 words, and speak them with their feeling. Then comes, 
 " these are to be your arts," in English the "your" would 
 be emphasised, and considerable meaning would be 
 thrown into the long but un emphatic arts. In Latin the 
 dissyllable dries supplies the place of both English syl- 
 lables. The strong tu in the preceding line, renders tibi 
 weak, and allows it to be slurred on to erunt, so that the 
 whole force falls on dries, " these are your substitutes 
 for the arts, you are governors, men (viri) not mechanics 
 (homines)." Then the idea is further developed, and 
 the three special arts alluded to are described : to im- 
 pose your own laws upon others, to spare (we know hovf 
 Romans spared) those that submit, and to put your 
 heel on those who resist. This is the spirit to be in- 
 fused into the utterance. The effect is obtained by the 
 pronunciation of impo'nere, with very long im, and 
 steady, decided, unhurried, but very emphatic po'nere; 
 by making pdrcere quiet, but with a kind of tone, im- 
 plying, " oh yes ! of course, we spare them," and 
 throwing the chief force on subjectls, where sub must be 
 long, and the rise and fall of pitch with the length of 
 the Us be so managed as to give the feeling "yes, if 
 they give up everything and surrender unconditionally ; " 
 and then, gathering strength on the three following long 
 and low syllables etdebel, rise with a kind of quiet exulta- 
 
Art. 122, 123.] POETICAL EXAMPLES. 97 
 
 tion on la and finish by bringing out the per in superbos 
 with a savage emphasis which shews that every one who 
 ventures to oppose Rome has a proud spirit which must 
 and shall be humbled. 
 
 It is very difficult to describe these points in language 
 at all, and hence the schoolboy slang which I have used 
 must be forgiven. The reader must endeavour to realise 
 them by repeating these wonderful lines till they are 
 perfectly familiar to him, and he will have then learned 
 more of Latin pitch-accent and quantity, with freedom 
 of force, and the great variety of oratorical power which 
 this allows, than by any other means that I know. 
 
 ART. 123. Example C. The story of these lines is 
 well known. M. Claudius C. f. C. n. Marcellus, the 
 hope of Augustus and the Romans had died (B.C. 23) 
 the year before they were written, in his twentieth year. 
 All Rome had flocked to his funeral, and Augustus 
 himself had pronounced the funeral oration. 
 
 Aeneas sees the shade of this youth accompanying 
 the great M. Claudius M. f. M. n. Marcellus (B.C. 268 
 208) :- 
 
 Atque hie Aeneas, una namque ire videbat 
 egregium forma juvenem et fulgentibus armis, 
 sed frons laeta parum, et dejecto lumina vultu : 
 Quis, pater, ille virum qui sic comitatur euntem ? 
 Filius, anne aliquis magna de stirpe nepotum ? 
 QuT strepitus circa comitutn ! quantum instar in ipso ! 
 Sed nox atra caput tristi circumvolat umbra, 
 turn pater Anchises, lacrymis ingressus obortis : 
 
 and answers in the words of (C.). 
 
 These lines, recited by Virgil himself to Augustus and 
 his sister Octavia, the mother of Marcellus, must have 
 
 H 
 
98 POETICAL EXAMPLES. [Art 123. 
 
 been, in the poet's eyes, the best that he could make, 
 and are therefore extremely well adapted for an exercise 
 in Latin rhythm. It so happens also that they exem- 
 plify the treatment of the final m in a most remarkable 
 manner, and have been re -written for that purpose in 
 Art. 97. Only a few hints need be added. 
 
 Line (i) is mere sadness. Line (2) depends for its 
 force on ostendent and tdntum, and its great accumulation 
 of spondees. Lines (3) and (4) are a passionate appeal 
 to the gods, and at the same time an outrageous flattery 
 of the Romans. The chief words are nimium (3) and 
 hake (4). Lines (5) to (7) are reminiscences of the 
 mighty pageant of the funeral, the chief words are qudn- 
 tos, virum, qudZ, and the melancholy recentem, which, 
 scarcely in character where it stands, spoke volumes to 
 Augustus and Octavia. Then comes a more cheerful, 
 though regretful view of Marcellus's prowess. Take 
 care of lliaca, on account of the fearful English ilai'acd. 
 The quisquaddegente, intdntusspe, ullo are the important 
 points. Then comes the exclamation of regret for losing 
 one that shewed such eminent filial affection (pietas), 
 bringing back the days of " old faith," (just as the 
 German prides himself on alte deutsche Treue^) and then 
 such wonderful success in war as his youthful feats would 
 lead one to hope. It is on the last of course that most 
 stress is laid. Observe \h&\. pietas and fides will acquire 
 their chief effect from their final long syllables with a 
 low pitch, the descent of the voice allowing of great ex- 
 pression of regret, and the long quantities of the vowels 
 giving abundant opportunity of developing it. Observe 
 especially invictdque ( 1 1) with the high pitch transferred 
 to the short vowel d to mark the enclitic addition. The 
 
Art. 123, 124.] POETICAL EXAMPLES. 99 
 
 contrast in (13, 14) is between pedes and eqid, but after 
 armd'to the two lines may be spoken rapidly. 
 
 The next line (15) begins with quite a new strain, as 
 mournful and tender as possible. Miserdnde, must be 
 well brought out, and the two first syllables must not be 
 those in the English miserable, but clear mi-se-. Ptier 
 may occasion some difficulty, but great tenderness can 
 be conveyed by the fall of the voice, and the lengthened 
 second syllable. In qua the greatest stress is reached, 
 the following fdtadspera rumpas, are tearful, broken 
 voiced, but quiet. Then in (16) we have the soft pene- 
 trating tu marcellus en's, on hearing which Octavia 
 fainted, so that probably Virgil never finished his reci- 
 tation. But I recommend the whole of the rest of the 
 speech to be finished in a low, weak, rather hurried, 
 mournful tone, till it dies off at mu'nere. The assimilated 
 m in (17) and the slur in (18) must be carefully studied 
 and steadily attacked, and the final I in ina'm must have 
 its full effect, for though the tone is that of a man who 
 has lost all heart, yet as the reader is a totally different 
 person, he must be particularly careful not to make the 
 speech ridiculous by ludicrous errors of pronunciation. 
 
 ART. 124. Example D. In (D.) the whole character 
 of the verse is changed. We have still hexameters cer- 
 tainly, but they have no roll, they are verses of society, 
 charming by the way in which they set colloquialisms to 
 metre. The present passage has been selected for its 
 variety in a short space, and the great numbers of 
 spondees it contains. 
 
 Begin with a quiet interrogative tone, "pausing in (i) at 
 fit, maccetnas, a.nd nemo, emphasising qudssibi distinctly, 
 and in (2) bringing out the two seu and contrasting the 
 
 H 2 
 
ioo POETICAL EXAMPLES. [Art. 124. 
 
 ratio and fors (beware of final z). The ratio and dederit 
 will derive their chief force from their long final syllables. 
 After obje'cerit pause again, and then take ilia, making 
 the double / quite distinct, and the a long with the 
 descending pitch. All three lines are to be spoken in 
 a tone of amused argumentative puzzlement. 
 
 In (4) the old soldier speaks, and the words must be 
 in a tone of longing envy, to which the words of four 
 long syllables, with a final low pitch, and very long 
 vowels, add great force. At grdvis dnnls change the tone 
 to Horace's, who says, " there you see, this is why he 
 says it." The grdvis must be short (beware of English 
 gre'vts), and the double n and long I of dnnls must 
 make a strong contrast (beware of reducing the word to 
 English anise). In (6) the words introduce the mer- 
 chant, with a parenthetical explanation of his opinion, 
 "you know, his ships were in a storm at the time," 
 which shews how to bring out the na'vljac-, all the syllables 
 quite long. In (7, 8) we have the merchant's speech. 
 Beware of calling militia like the English milislw. The 
 quidenl for quid enim, had possibly only one syllable with 
 raised pitch. Many instances of enim, quidem, autem, 
 seem to me to come under Quintilian's rule of junctura. 
 But begin ccon- with energetic c. In (8) we have the 
 difficulty of two' long syllables being slurred momenta aut, 
 and there must be a pause between them, see Art. 58, c. 
 For such verses the ear will sufficiently recognise the 
 intention even if neither of them is much shortened. 
 Be very particular with the pyrrhics dta, v'enit, in (8). 
 
 In (9) mind the ag- to lengthen the first syllable, and 
 the two assimilated m. In (10), the subgdlll cdntu must 
 be spoken with a tone which shews that it was only the 
 
Art. 124.] POETICAL EXAMPLES. 101 
 
 unconscionably early hour of rousing which could have 
 led the lawyer to such an opinion. The slur in .consul- 
 toriibibstia pulsat will require some practice to execute 
 neatly. Say bs-ti-a, not bstja. In (is) bring out the Hie 
 (with distinct double /) as referring to the consultor who 
 is an agricola. The ddtis vddibus, being so different from 
 the common English de'tis ve'dibds, wild require care to 
 preserve the high initial pitch and the final long syllables. 
 In (12) much effect can be produced by the wondering 
 longing tone of the farmer's first three words. 
 
 Then in (13) everything is changed. Tiae tone be- 
 comes quick and petulant, the parenthetical ddeo sunt 
 midta,) shews that Horace is tired of enumerating cases 
 which would tire even gossipping Fabius. The long e in 
 ddeo allows of much effect in the tone. In (14), ne 
 te'mbrer is quick, and the last word very rapid. Observe 
 the m's in quo reddedu'ca in (15), and particularly the en 
 ego as distinct from our common English en t'go'. In 
 (16) the assimilated m is important, jdffdciacquid volt is. 
 All this speech of the god (15 to 19) should be read in 
 a quiet ordering tone, but very carefully as to quantities. 
 Line (17) requires great care, as well as mutatis in (18). 
 For eja ! statist (19), see Art. 115. The no' lint (19) 
 belongs to Horace, with an amused tone implying, " they 
 wouldn't, I told you so." Then comes the remark on 
 their refusing happiness when in their power. 
 
 The last three lines are in an indignantly con- 
 temptuous, though still almost conversational tone. The 
 m'erito is the chief word, and owes its weight to the 
 final long syllable. Take breath after jupiter. The 
 whole three lines are in close connection and in one 
 characteristic quality of tone. 
 
fo2 POETICAL EXAMPLES. [Art. 125. 
 
 ART. 125. Example E. We now come to a totally 
 different style of poetry. In lyrics the metrical rhythm 
 must be always well brought out, but in general with 
 very little chant. The (K, F., I., K.) have totally dif- 
 ferent characters, and the chant belongs to (I.) only. 
 Each ode of Horace indeed requires a distinct study. 
 (E., F!) are given for the sake of the choriambs. In 
 (E.) there are two in each line, and the first ends a word, 
 but there is a spondee at the beginning connected with 
 the first, which must be well brought out, and an iamb 
 or pyrrhic at the end connected with the last, so that 
 there is in many lines a false appearance of ending 
 with two dactyles. Many lines are read by English boys 
 . as if they had three final dactyles, thus : 
 
 misi 'lias at "avis ed 'iti redj 'ibas, 
 
 using dj for English j. All that is absolutely absurd. 
 To bring out the choriambs nicely, the last long syllable 
 of the second should be slightly lengthened, and the 
 voice should grow weaker for an instant before proceed- 
 ing to the next syllable, but care should be taken not 
 absolutely to divide the last word into two. This may be 
 indicated by a hyphen before the | which separates the 
 feet ; thus : 
 
 maece'- | nas atavis | e'dite re'- | gibus. 
 
 Line (2) will require considerable care to bring out 
 the slur properly, and so will (7). There must be a 
 slight pause after the second et in (2) and this gives an 
 opportunity for a slight " gush " on dulce. 
 
 In (3) put an emphasis on quos, '-''some like to do 
 this," together with many other things in the lines omitted, 
 
Art. 125.] POETICAL EXAMPLES. 103 
 
 as opposed to the me of lines (7) and (8). Be careful 
 to indicate the raised pitch in metdque (^}, palmdque (5), 
 see p. 29, note i. In (6), the contrast of terra'ruddbminos 
 and deos led me accidentally, when I was reading the 
 lines out, to raise my pitch on the last syllable of deos, 
 and, though I immediately corrected the error, I mention 
 it to shew the difficulty of overcoming national habits. 
 It would be equally wrong to raise the pitch of the voice 
 on ad) and that creates the difficulty. Read : 
 
 terra' | ruddominos | e'vehit ad- | de5s 
 
 and mark the contrast between the two first words and 
 the last by the quality of tone used for the final long 
 os in deos. 
 
 In (7) be careful of / in frdn- | tium, and in (8) and 
 (9) mind the three assimilated m's. In (10) be careful 
 of tl'bias, which as an English word we call tib'tiz. Also 
 remember the initial spondee, and the assimilated m in 
 (12), producing the slight lengthening just mentioned, 
 and the double r, as lesbo'- | urrefugit | . 
 
 Great effect can be given to the two last lines, which 
 close the dedication. The sentiment is, " I am satisfied 
 with all this, and have no such desires of athletic 
 honours, or political office, or wealth or hunting &c. as 
 I have mentioned; but if only you will place me 
 among lyric poets, I shall be as proud as a god." 
 Make short pauses after quod si me come with a 
 joyous "gush" upon ly'ritis, lengthening out the last 
 syllable, and getting immense expression out of the fall 
 of pitch from the short high pitch of the first syllable. 
 In (14) sub-ll- | ml there must be a slight echo of the 
 commencement of the preceding line. Then fe'rias- | 
 
104 POETICAL EXAMPLES. [Art. 125, 126. 
 
 si'dera ver- | tice, first falls down quickly to a slight 
 pause without separation, (of the kind already mentioned) 
 and the voice rises again to a magnificent sl'dera, the 
 last word being comparatively unimportant. 
 
 ART. 126. Example F. Here there are three cho- 
 riambs, the first and last are treated as in E., the middle 
 one is always quite distinct, as, scirenefas, Leuconoe, 
 utme'lius ; seutribuit, debilitat, vinaliques, diilldquimur, 
 quamminimu, and are seen to be much varied in pitch- 
 accent. 
 
 Suppose Horace to be sitting one winter's night before 
 his table with a bowl of Falernian, while on the other 
 side pretty Leuconoe is " telling fortunes " with " Babylo- 
 nian numbers" and other contrivances. Horace looks 
 at her amused for some time, and then bursts out with, 
 " Don't look how long we've to live, Leuconoe, it's 
 not right ! and don't try your Babylonian contrivances. 
 How much better it is to endure whatever happens (with 
 stoic dignity), whether we are to live long or die now. 
 Be wise my girl, and drink, (like an epicuri de grege 
 porcus !) and for the short present renounce the long 
 future. Don't you see, time is flying while we speak; 
 make use of now and a fig for then!" This, in pur- 
 posely unpoetical language, seems the gist of the original. 
 On this view lines (i) to (3) are a kind of paternal 
 rebuke, in a pleasant tone of voice, however, half 
 coaxing. But when the stoic comes out, the voice and 
 manner change almost to that of (K.), the utmelius, 
 quicquid erit, and especially the pdtl (where the descent 
 of tone and final long syllable are capable of great 
 effect)* must be strongly brought out. The next two 
 lines to the beginning of (6), and especially all (5), which 
 
Art. 126-128.] POETICAL EXAMPLES. 105 
 
 is wonderfully, of course purposely, monotonous, may be 
 rapid, but take breath after metre, in order to run on 
 quickly from tyrrhenum on to a loud, bold, commanding 
 and yet jovial sdpias, with an assimilated ;;z, as tyrrhe'- | 
 nussapias | , then the vinaliques is quite in the jolly 
 manner, which is continued in the following words, 
 where brevl must be well emphasised, and in running on 
 spellon- | garreseces | , the last word must come out 
 very strong and decided, with the chief force and highest 
 pitch on re (which must be kept short), and then without 
 running on to s, the energy of speech must expend itself 
 on the low ces (keep c, s as pure /, s). Then make a pause 
 and resume in quite an altered quiet tone, dulloquimur | 
 fu'gerit in- | vida | ae'tas | and then encouragingly car- 
 pedie | , and with an implied smiling shake of the head : 
 cquamminimu | ccre'dula post- | era 
 
 ART. 127. Example G. This and (H.) are introduced 
 for the sake of the ehoriambic pentameters. Observe 
 in (G.) the strictly ehoriambic terminations militiae (2), 
 carminibus (4), indbmina (6), with a pitch-accent rhythm 
 quite different from Ovid's, which is only found in (8). 
 The piece is not particularly good for declamation, but 
 the flattery of (3), with the attestation ita siffe'lix, and 
 the parenthetical good wishes of (4) admit of a little 
 variety, as also the parenthetical utconsv?mus (5). The 
 assimilated final m must be observed in (6) and (7), and 
 care must be taken to preserve in long in (7) and ge 
 short, as : nectantuinginio quantusservire dolo'rl. 
 
 ART. 128. Example H. This is a sample of those 
 monotonous Ovidian " longs and shorts" which so 
 afflicted our school life. It is necessary thoroughly to 
 understand the cadence of the final choriambs : ipsevenl 
 
106 POETICAL EXAMPLES. [Art. 128-130. 
 
 (2), trojafuit (4), a | diilter aquis (6), re | licta dies (8), 
 tela manus (10), which ought not now to occasion any 
 difficulty. Their monotony of pitch-accent is wonderful. 
 Be careful of the /;/ in (5). Take breath after deserto and 
 quaerentl (9) in order to run on jacuisseftrigida (7), and 
 spatio' saffdllere (9), making the sef, saf quite long, but 
 entirely without emphasis and in low pitch. 
 
 ART. 129. Example I. Sapphics put on very different 
 characters in Horace's odes. The present example is 
 devotional and was meant to be sung or chanted. It 
 will be advisable to put a little chant into it therefore. 
 These stanzas are chosen because they give three in- 
 stances (i, 9, 10) of a break in the line opposed to 
 English habits. Chant then solemnly, and bring out the 
 choriamb distinctly, thus : 
 
 phoebe sylva- | rucque potens | dia'na, 
 lu'ciduccoe-' | IT decus, 6 | colendi 
 semper et ciil- | ti, date quae | preca'mur 
 tempore sac- | ro. 
 
 Be careful with assimilated m in (7, 9) and especially 
 (10), the last is 
 
 spebbonaccer- | tacque domur- | reports 
 
 and breath should be taken after que. See p. 74, note. 
 
 ART. 130. Example K. This is completely stoical, 
 lauding the immovability of the "just" man. It is 
 therefore a bold piece of declamation. There must be 
 a pause after et (i), the slur in (2), ndn ci'vi-ucir-dor, 
 requires especial practice to save it from ndn cl'vju-ardor, 
 with an hiatus : but ndn ci'vi war dor would be preferable 
 to this. Practice uardor separately, and ci'viu separately, 
 
Art. 130, 131.] POETICAL EXAMPLES. 107 
 
 and then put them together. Bring out the molossus 
 instantis in (3). The mentequdtit \ sblidd \ must be bold 
 and fine 3 the effect depends, as so often before, on the 
 length and low pitch of the last syllables. In (6) make 
 a point of fulmmdnttS) and run magnajbvis into a distinct 
 choriamb. But reserve force for (7) and (8). There 
 must be quiet preparation first for the explosion on bd 
 (7) after the long syllables ilia, and for the strong impdvi- 
 dufferient, where the effect is produced by the long but 
 low pitched im and duf t The rulnae is a comparatively 
 quiet word because its coming has been so thoroughly 
 anticipated. 
 
 ART. 131. Example L. This is a piece of fun through- 
 out, which is explained by the last four lines. An old 
 usurer is seized with a false sentiment for country life, 
 and gloats over what he fancies would be its amusements 
 to him, but which of course would prove dreadful an- 
 noyances to a man who had lived such a life as his. So he 
 works himself up through some 66 lines of false sentiment, 
 of which only 18 are given in (L.), into a determination to 
 give up his usury and go to the country. Whereupon he 
 calls in all his capital in the middle of the month, and 
 at the end of a fortnight he is so disgusted with his 
 venture that he tries to reinvest the whole. This false 
 sentiment is easy to express, and no harm is done if it is 
 ludicrously exaggerated. Hence the first four lines can 
 be spoken with a wonderful variety of pitch, the high 
 tones rising to a kind of falsetto. \\\ prbcul (i) lengthen 
 the cul greatly, and similarly the final Is in nego'tils. 
 Mind the spondee in (2), contrast the paterna and sills, 
 and bring out the molossus (exercei) in (3), and in (4) 
 mind the long ]m of 6mm, of which much may be made in 
 
io8 PAOSE EXAMPLES. [Art. 131, 132. 
 
 expression. Line (17) will require care on account of its 
 trissyllabic measures and final m. At (19) the voice and 
 manner entirely changes. The writer gives the key to 
 the mystery and dismisses the usurer caustically. Pay 
 attention to the mental effect of jam/am, as well as to its 
 phonetic effect of assimilated m, as ja'ijaffutii'riis ru's- 
 tiais, (p. 66, note 2) and mind the spondee in the third 
 place. In (21) attend both to mental and phonetic effect 
 of bmnerredegit. After this line pause, and give the last 
 line in a quiet cutting tone. See also the remarks on 
 p. 77. 
 
 XIII. How to Read the Prose Examples in 
 the Appendix. 
 
 ART. 132. These cannot be considered in such detail, 
 for I despair of making myself intelligible except by 
 reading the passages. But I will observe upon neces- 
 sarily -unemphatic words, such as, autem, etiam, quidem, 
 enim, &c., and the probability that they also came under 
 Quintilian's rule of juncturae. Thus in (M. i), estautcin 
 or estautein will come in weakly, and lead up to the mo- 
 lossus dicendo which must be clear and decided, while 
 etia will fall weak. Then equldaccantus will be distinct 
 and important, and obstu'rior must end with a clear dac- 
 tyle. Difficulties arise in long words with several long 
 and short syllables mixed in a manner which could 
 scarcely occur in verse, except rarely in lyrics, as oratione 
 (M. .4), modularetur (M. 4), ad auriuvolupta'tessequatur 
 industria (M. 6), multitu'do (N. 2), longitu'dinuet brevi- 
 ta/tulnsonls (N. 4. 5), animadversio (O. 6), nominantur 
 (N. 9), jOdiciussuperbissimu (Q. 6). It is only by study- 
 
Art 133-136.] PROSE EXAMPLES. 109 
 
 ing such separate little phrases that these difficulties can 
 be overcome. It is principally necessary to keep the 
 quantities pure, but care must also be taken to raise and 
 lower the pitch of the voice in the right places. 
 
 ART. 133. Example M. offers no difficulties beyond 
 what have been mentioned. It is a quiet piece of lecture 
 delivery in which every point should be neatly and 
 cleanly rendered without fuss or mouthing. 
 
 ART. 134. Example N. is similar, but (i, 2) relate 
 a fact, which has to be emphasised by tota and u'na 
 sy'llaba, while (2) and (3) states. another fact, advanced in 
 a different tone of voice, and (4) to (6) gives Cicero's 
 explanation of both, and all this should be indicated by 
 the character of tone employed. 
 
 ART. 135. Example O. is also a lecture piece, but (12) 
 and (13) are capable of a little point, by funnily empha- 
 sising the cdmico'rum and especially the abjectl. 
 
 ART. 136. Example P. is a criticism on the rhythmical 
 character of a sentence. Only the last word of the first 
 clause is here given, but it should be read as a de- 
 clamatory terminal, " persolu'tas," with much orotundity. 
 Then comes parenthically the name of the foot (dlchoreus\ 
 in a quiet explanatory tone of voice, followed by remarks 
 (one taken from another section of the book), on the 
 indifferent quantity of final syllables. Afterwards another 
 sentence is taken, which must be delivered with great 
 orotundity, as if at a public assembly, to bring out the 
 final dichoreus, which, in a perfectly quiet altered tone, 
 must be stated to have had such an " admirable " effect. 
 Then a query : " Did this depend on the sense or the 
 rhythm ? Try. Change the order." Then the order is 
 charged and the reader should endeavour to make the 
 
no PROSE EXAMPLES. [Art. 136, 137. 
 
 final paeon temeritas as effective as possible, in order to 
 bring out the justness of the following criticism, janni- 
 hilerit ! said in a quiet rather high off-hand tone, such as 
 is often adopted in similar circumstances on the Con- 
 tinent. The disagreement from Aristotle (see Cicero's 
 original bowing to his author, quia optimus autor ita 
 censet, in Art. 109) must be given with some unction, 
 and the last sentence must be well contrasted and em- 
 phasised. Observe : at eadeu'verba, eades*senten tia, 
 animo'istuc satisest-, aVribus non'sa"tis. The last 
 dactyle must come out well, and we must feel that 
 it is preceded by a cretic. The nature of the feet 
 in the cadence of all clauses (not merely of sentences) 
 were, if we may trust their expressions, always brought 
 out by Cicero and Quintilian. 
 
 ART. 137. Example Q. contains extremely important 
 remarks on slurred and gaping vowels, but is all in the 
 quiet critical style. With this we may compare Quin- 
 tilian : dilucida vero erit pronuntiatio [delivery] primum, 
 si verba tota exierint, quorum pars dlvorare, pars destitui 
 solet, plerisque extremas syllabas non perferentibus, dum 
 priorum sono indulgent, [shewing Latin vicious habits, 
 comparable to our own, and explaining many subsequent 
 changes.] Ut est autem necessaria verborum explanatio 
 [clear utterance;] ita omnes imputare-et velut anrmme- 
 rare litteras molestum et odiosum, [which must be par- 
 ticularly noted, but foreigners are always allowed to 
 speak with more distinctness than natives, without being 
 stigmatised as molesti et odiosi, thus Gaelic and Welsh 
 English is often felt to be " prettier " and " pleasanter " 
 than our own more freely treated language.] Nam et 
 vocales frequentissime coeunt [slur into one another,] et 
 
Art. 137, 138.] PROSE EXAMPLES. in 
 
 consonantium quaedam [he is evidently thinking prin- 
 cipally of m but he is allowing for old s and other ex- 
 amples in Cic. Or. 153 162] insequente vocali 
 dissimulatur ["made to appear something else," this is 
 its literal meaning, but in White's Latin Dictionary it is 
 in this passage translated by a totally different metaphor 
 as "absorbed." Observe how this interpretation supports 
 that given to in earn tramlre on p. 60.] Utrique ex- 
 emplum posiiimus : multum ille et terrls. Vitatur etiam 
 duriorum inter se congressus, unde pellexit et collegit, &c. 
 
 (n, 3, 3335)- 
 
 ART. 138. Example R. The story of Gracchus's 
 piper is here .told as a dialogue and allows of a little, not 
 much, alteration of voice. The parenthesis (3-5) should 
 be marked, for the connection of Gracchus with cum ebur- 
 neola solitus est habere fistula qui staret &c. (3, 5) is not 
 very clear to an Englishman, and will require some 
 management of the voice to bring out. Observe cuccon- 
 tionaretur (6). The observation of Catulus (9, 10) 
 must be made in quite another tone of voice. Then 
 Crassus returns and begins meditating on Gracchus's 
 treason and I have taken advantage of Julius's cutting 
 him short to cut him still shorter. The mitte bbsecro, 
 is a polite stop, and the last words of Julius's speech, 
 cujus ego nondum &c. admit of considerable expression 
 as shewing that he had failed to understand the reason. 
 Now whether the reason given by Crassus is right or 
 wrong it is not very possible to say. I must own I am 
 not satisfied with it, and I think Quintilian's version 
 more likely to be correct, judging from Cicero's further 
 account of the extremely artificial character of Gracchus's 
 oratory : 
 
H2 PROSE EXAMPLES. [Art. 138. 
 
 Quid fuit in GracchS? quern tu, Catule, melius meministi, quod 
 me [Cotta is speaking] puero tantopere lerretur [C. Gracchus was 
 killed B.C. 122, Cotta was born B.C. 124.] " Quo me miser con- 
 feram ? quo vertam ? in Capitoliumne ? at fratris sanguine redundat : 
 an domum? matremne ut miseram lamentantemque videam, etabjec- 
 tam ? " Quae sic ab illo acta [delivered, not ' acted '] esse constabat, 
 oculis, voce, gestii, inimici ut lacrimas tenere non possent. Haec 
 eo dlco pliiribus, quod genus hoc totum 5ratores, qui sunt veritatis 
 ipsms actores, reliquerunt ; imitatores autem veritatis histriones 
 occupaverunt. (Cic. De Or. book iii. 214.) 
 
 It is evident that the whole manner of Gracchus had 
 become tradition al v even in Cicero's time, and that there- 
 fore Crassus's explanation can only be a theory. 
 
 Now Quintilian was of course still farther off, and 
 possibly he took the tale from Cicero whom he is con- 
 tinually quoting, but, speaking of the relation of oratory 
 to music he says 
 
 UUno interim content! simus exemplo C. Gracchi, praecipui 
 suorum temporum oratoris, cui contionanti consistens post eum 
 musicus fistula, quam tonarion [pitch pipe] vocant, modos, quibus 
 deberet intend!, monstrabat. Haec e! cura inter turbidissimas 
 actiones vel terrent! optirnates vel jam timent! pr5fuit (i, 10, 
 27). 
 
 Now the " modos " would seem to imply the cadences 
 which were peculiar to the Greek musical modes, and it 
 appears to me, that, although very possibly Quintilian 
 used the word at a venture knowing little of musical 
 theory, this is a more likely solution than Cicero's, put 
 into the mouth of Crassus, since these different " musical 
 modes " were especially suitable to Gracchus's theatrical 
 oratory. 
 
 At the end of Crassus's explanation, the last lines of 
 advice may be given in a quiet little preceptorial manner, 
 
Art. 138140.] CONCLUSION. 113 
 
 " of course you will understand that the piper is to be 
 left at home, and only the meaning of his piping brought 
 with you to the forum." 
 
 ART. 139.- Example S. is Italian and not Latin at all, 
 but I attempted to recite it when reading my paper, in 
 order to shew the precise nature of the phonetic facts on 
 which I have founded the explanation of slurred vowels, 
 and omitted but assimilated m. It is retained here in 
 order that the reader may be able to verify these facts, 
 by getting Italian natives who have had a literary educa- 
 tion and speak pure Tuscan, to read the passage to him, 
 so that he may really hear and understand for himself 
 that these theories are founded on living realities. 
 
 XIV. Conclusion* 
 
 ART. 140. Notwithstanding the theoretical points 
 which have been touched upon in the arguments raised 
 in favour of the method for the treatment of slurred 
 vowels and final M, here proposed, of which the latter 
 forms the greatest novelty in this tract, yet I hope 
 that I have not belied the promise of my title and of my 
 opening paragraph, but have furnished strictly practical 
 rules for arriving at a feeling for ancient Roman rhythm 
 in verse and in prose, quantitative and accentual. It is 
 perhaps scarcely necessary, to observe that even if my 
 theory of assimilated m be not accepted, through the 
 great prejudice in favour of the use of pure final m, 
 dating from the time when Latin, having broken up into 
 separate languages for common use, was restudied as 
 
 i 
 
114 CONCLUSION. 
 
 a dead language with modern usages, and without a con- 
 sciousness of the ancient treatment, to which the modern 
 form of the derived languages was obviously due, yet 
 even if m be retained, if Quintilian's lament (in him 
 certainly quite alphabetical) over the frequency of the 
 "lowing" final m be justified by scholastic usage, all the 
 rest of the work that I have done is as available as ever 
 for practical use, and will equally well serve to give 
 a feeling for the rhythm of Latin speech which is of 
 course dependent principally on a strict observance of 
 the laws of length and pitch, and independent to a great 
 extent of particular alphabetic usages. But allow me 
 finally to draw attention to the absolute necessity of 
 general phonologic studies to all those who would deal 
 with the intricate questions of classical pronunciation, 
 not merely Latin, which is comparatively easy, although 
 very very far from having been completely investigated, 
 but especially Greek, which presents problems of re- 
 markable difficulty. Those who have hitherto written on 
 the subject have seldom known much more of phonology 
 than they could learn by speaking their own language, 
 without thinking of how they spoke, or of what speech was. 
 But for such investigations as the present, an acquaint- 
 ance with the habits of many natiors is indispensable, 
 and of the historical alteration of sounds. It will have 
 been seen that I was led to my theory of final M by 
 delicate observations on the synthesis of Italian sounds, 
 scarcely known, and seldom even mentioned in Italian 
 books, which I owe entirely to Prince Louis Lucien 
 Bonaparte. It is also necessary to become acquainted 
 with the phonetic facts which have underlain explanations 
 by grammarians who were but roughly acquainted with 
 
Art. 140.] CONCLUSION. 115 
 
 their nature, in order to have a glimpse of the meaning of 
 other grammarians who are presenting unknown facts. If 
 I have in any way succeeded in putting together a "work- 
 ing model" of Augustan speech, with of course all the 
 roughness and incompleteness of a model, I owe my 
 success to my previous phonologic studies for more than 
 thirty years, and to my yet unfinished historical studies 
 during the last ten years, on Early and Existing English 
 Pronunciation, received and dialectal. 
 
 I 2 
 
LI 
 
 > 
 
 U -A 
 
 APPENDf! 
 
 CALIFOil.MA. 
 
 QUANTITATIVE EXAMPLES. 
 
 NOTE. The diphthongs ae, oe, au, eu, ui, being naturally long, 
 are unmarked. Other naturally long vowels are marked a, e, I, 6, 
 u. Short vowels are left unmarked. The consonant which "makes 
 position" is followed by a hyphen, as in ip-se. ART. 22. 30. 
 
 Vowels connected by w are to be slurred together in one syllable , 
 this mark is not used in the prose examples. ART. 58. 
 
 In example A only, a high pitch of voice sustained throughout art 
 entire syllable is marked by an acute accent, as omnes ; when the pitch 
 of the voice rises and falls again in the same syllable the circumflex 
 accent is used, as 6-ra. Unaccented vowels are in a lower pitch of 
 voice. The acute accent is retained before que in all the examples, 
 as metaque. ART 41. 
 
 A small m as in quan-quam^animus is to be entirely neglected. An 
 m- at the end of words, or in m-que, making "position" as jam- 
 nox, certam-que is not to be sounded at all, but is to be made effective 
 by pronouncing the following consonant as if it were double, thus 
 jannox certacque, or by lengthening the preceding vowel before a 
 pause, as, dolo're. ART. 91. 
 
 The numbers following the title of any example, refer to the pages 
 on which that example is cited, and an asterisk indicates the page 
 where special explanations are given. See generally, ARTS. 115 
 
 139- 
 
 Example S. is Italian, not Latin, and illustrates slurred vowels 
 and unwritten consonants assimilated in speech, not quantity. 
 
1 18 QUANTITATIVE EXAMPLES. [A. B. 
 
 A. 
 
 Aeneas' s Introduction to his Account of the Destruction of Troy. 
 See pp. 32. 33. 37. 40. 41. 42. 67. 68. 90*. 
 
 con-ticuere^om-nes, in-ten-tique^6ra tene'ban-t. 
 2 in-de toro pater aene'as- sic or-sus ab al-to : 
 
 m-fan-dum-, regina, jubes- renovare dolo'rem-, ' 
 4 troja'nas ut opes et- lamen-ta'bile reg-nu m 
 
 eruerin-t danai, quaeque^ip-se miser-rima vi'di, 
 6 et- quo'rum- par-s mag-na fin. quis- ta'lia fan-do 
 
 myr-midonum- dolopum-ve w/ aut- du'ri miles ulix-i 
 8 tem-peret a lacrymls ? et- jam- nox- hu'mida coe'lo 
 
 praecipitat-, svaden-tque caden-tia si'dera som-nos-. 
 10 sed-, si tan-tus amor- ca'sus- cognos-cere no's-tros 
 et- breviter- tro'jae sup-re'mu m v _,audrre labo'rem-, 
 12 quan-qua m w animus- nieminis-se w hor-ret- luc-tu'que refugit 
 
 in-cipiam-. frac-ti bel-lo fati's-que repul-si 
 14 duc-to'res- danaum-, tot- .jam- laben-tibus an-nis 
 
 I'n-star- mon-tis equum- divfna pal-ladis ar-te 
 1 6 aedifican-t, sec-ta'que^in-tex-un-t ab-jete cos-tas ; 
 
 / vo'tum- pro reditu simulan-t ; ea fama vagatur. 
 1 8 ^huc delec-ta virum- sor-tfti cor-pora fu'r-ti m 
 
 in-clu'dun-t cae'co later!, penitus-que caver-nas 
 20 ingen-tes .uterum-que^ar-ma'to milite com-plen-t. 
 
 es-t in con-spec-tu tenedos-, notfs-sima fa'ma 
 22 fn-sula, dives opum-, priami dum- reg-na mane'ban-t, 
 
 nunc: tan-turn- sinus et- statio malefida cari'nis ; 
 24 hue se provec-tT deser-to^in li'tore con-dun-t ; 
 nos abiis-se rati, w et- ven-to petiis-se myce'nas. 
 
 VIRG. Aen. ii. I 25. 
 
 ' 
 
 Roman Policy as contrasted with Greek Art. 
 See pp. 32. 33. 34. 40. 41. 68. 94*. 
 
 quo fes-sum- rapitis-, fabii? tu max-imus il-Ie^es, 
 2 unus- qui nobis- cun-ctan-do res-titues- 
 
B. C] QUANTITATIVE EXAMPLES. 119 
 
 ex-cuden-t alii splran-tia mol-lius aeja, 
 4 credo^equidem-, viv5s- ducen-t de mar-more vul-tus, 
 
 5<abun-t causas- melius-, coelfque meaius- 
 6 describen-t radio^et- sur-gen-tia sidera dicen-t; 
 
 tu_ regere^im-perio populos-, romane, memen-to : c 
 
 8 hae tibi^erun-t ar-tes-, pacis-que^im-ponere morem- ; 
 
 par-cere sub-jec-tis, et- debel-lare super-bos. 
 
 VIRG. Aen. vi. 845 853. 
 
 c. 
 
 Lament for Marcellus. 
 Se<Tpp. 32. 33. 41. 71*. 97*. 
 
 o nate^in-gen-te%; luc-tum- ne quaere tuoru m , 
 2 os-ten-den-t ter-ris hun-c tan-tum- fata, neque^ultra 
 
 es-se sinen-t. nimium- vobls- romana propago . 
 4 visa poten-s, super!, propria^haec- si dona fuis-sen-t. 
 
 quan-tos il-le virum- mag-nam- mavor-tis ad ur-bem- 
 6 cam-pus aget- gemitus- ; vel- quae, tiberlne, videbis- 
 
 funera, quum- tumulum- praeter-labere recen-tem- ! 
 8 nee puer iliaca quis-quam- de gen-te latinos 
 
 in- tan-tum- spe tol-let avos-, nee- romula quon-da m 
 10 ul-lo se tan-tum- tel-lus jac-tabit alum-no. 
 
 heu pietas, heu pris-ca fides, in-vic-taque bel-lo 
 12 dex-tera ! non il-li se quis-qu k a m ^im-pune tulis-set 
 
 ob-vius armato, seu quum- pedes iret in hos-tem- 
 1 4 seu spuman-tis equi foderet- cal-caribus ar-mos. 
 
 heu, miseran-de puer-, si qua fata w as-pera rum-pas 
 1 6 til mar-cel-lus eris-. manibus- date lilia plenis- ; 
 
 pur-pureos- spar-gam- flores, animam-que nepotis 
 1 8 his sal-te m w ac-cumulem- donis, et- fun-gar inanT 
 
 txumere. VIRG. Aen. vi. 868 8 
 
120 QUANTITATIVE EXAMPLES. [D. E. 
 
 D. 
 
 No One Contented with his Lot. 
 See pp. 17. 32. 33. 40. 68. 99*. 
 
 qul fit-, maecenas, ut- nem5, quam- sibi sor-tem- 
 2 seii rati5 dederit-, sen for-s ob-jecerit, il-la 
 
 con-ten-tus- vivat-, laudet- diver-sa sequen-tes ? 
 4 "5 for-tunatT mer-catores ! " gravis an-nis 
 
 miles ait-, mul-t5 jam- frac-tus- mem-bra labore. 
 6 con-tra mer-cator-, navim- jac-tan-tibus austris : 
 
 "militia^es-t potior-. quid enim-? con-cur-ritur ; h5rae 
 8 momen-to^aut cita mor-s venit, aut vic-toria laeta." 
 
 ag-ricolam- laudat- juris- legum-que peritus-, 
 10 sub- gal-li cantum- con-sul-tor ubi^os-tia pul-sat. 
 
 il-le, datis- vadibus- qul rure^ex-trac-tus in ur-be m ^es-t, 
 12 sSlos- felices- viven-tes- clam at in ur-be. 
 
 cetera de genere^hoc, adeo sun-t mul-ta, loquacem- 
 14 delas-sare valen-t fabium-. ne te morer, audi, 
 
 quo rem- deducam-. si quis- deus, "en ego," dicat. 
 1 6 "jam- faciam- quid- vol-tis ; eris- tu, qul modo miles-, 
 
 mer-cator-; tu, con-sul-tus- modo, r\is-ticus ; hm-c vos, 
 1 8 vos hin-c mutatis- dis-cedite par-tibus. - eja ! 
 
 quid- statis-?" nolin-t. at-qui licet esrse beatis-. 
 20 quid- causae^est, merito quin il-lis Jupiter; am-bas 
 
 iratus- buc-cas in-flet- ? neque se fore pos-tliac 
 22 tarn- facilem- dicat-, vptis ut praebeat aure ? 
 
 HOR,, Sat. i. I 22 
 \ 
 
 E. 
 
 Dedication of the Odes to Maecenas* 
 See pp. 73. 101*. 
 
 I 
 
 maecenas, atavis edite regibus, 
 Z. 6 et- praesidiu m w et- dul-ce decus- meum-, 
 
E. F. G.] QUANTITATIVE EXAMPLES. 21 
 
 sun-t quos- cur-riculo pulvere m w olym-picum- 
 4 col-legis-se juvat-, metaque fer-vidis 
 
 evitata rotis- pal-maque n5bilis- 
 6 ter-rarum- dominos evehit ad- decs-. 
 
 me doc-taru m ^hederae praemia frorYtium- 
 8 dis- mis-cen-t superis- ; me gelidum- nemus 
 
 nym-pharumque leves- cum- satyris- choii 
 10 secer-nun-t populd, si neque tifjias 
 
 euter-pe cohibet-, nee- polyhym-nia .^-^JL 
 12 les-boum- refugit- ten-dere bar-biton. 
 
 quod- si me lyricis- vatibus in-seris, 
 14 sub-llmi- feriam- sidera ver-tice. 
 
 HOR. OOd. i. i. I 6. 2936. 
 
 F. 
 
 Leuconoe Recommended not to Peer into the Future. 
 See pp. 67. 73. 104.* 
 
 tu ne quaesieris-, sclre nefas- ! quern- mihi, quern- tibi 
 2 flnem- dl dederin-t, leuconoe ; nee- babylonios- 
 
 ten-taris- numeros. ut- melius-, quic-quid erit-, patl ! 
 4 seu pliires hiemes-, seu tribuit- Jupiter ul-timam-, 
 
 quae nunc op-positis debilitat- pumicibus- mare 
 6 tyr-rhenum-: sapias-, vma liques, et- spatio brevi 
 
 spem- Ion-gam- reseces-. dum- loquimur-, fiigerit in-vida 
 8 aetas- ; car-pe diem-, quam- minimum- credula pos-tero. 
 
 HOR. OOd. i. ii. 
 
 G. 
 
 While Ponticus writes Heroics, Propertius keeps to Erotics. 
 See pp. 74. 105*. 
 
 dum- tibi cadmeae dicun-tur-, pon-tice, thebae, 
 2 ar-inaque frater-nae tris-tia militiae. 
 
QUANTITATIVE EXAMPLES. [G. H. I. 
 
 at-que, w ita sim- felix-, pnmo con-ten-dis Homero, 
 
 sin-t modo fata tuis- mol-lia car-mlaibus- ! 
 nos, ut- con-svemus-, nos-tros agitamus amores, 
 
 at-que^aliquid- duram- quaerimus in- dominam- ; 
 nee- tan-tu m ^in-genio, quan-tum- servire dolorl 
 
 cogor, et aetatis- tem-pora diira queri. 
 
 PROP. Eleg. 
 
 H. 
 
 Penelope writes to Ulixes. 
 See pp. 74. 105*. 
 
 han-c tua penelope len-to tibi mit-tit, ulix-e, 
 
 ml- mihi res-cribas, at-tamen ip-se veni. 
 troja jacet- cer-te, danals in-visa puel-lis- : 
 
 vix- priamus- tan-ti totaque troja fuit. 
 o utinam- tun-c quum- lacedaemona clas-se petebat 
 
 ob-rutus m-sams es-set adul-ter aquis- ! 
 non ego deser-to jacuis-sem- frigida lec-to, 
 
 nee- quererer- tard5s Tre relic-ta dies ; 
 nee- mihi, quaeren-tl spatiosam- fal-lere noc-tem-, 
 
 las-saret viduas pen-dula tela manus. 
 
 Ov. Her. I, 
 
 I. 
 
 Hymn to Phoebus and Diana. 
 See pp. 68. 74. 106*. 
 
 phoebe syl-vanim-que poten-s diana, 
 2 lucidum- coeli decus, 5 colen-di 
 
 sem-per et- cul-ti, date quae precamur- 
 4 tern-pore sac-ro 
 
I. K. L.] QUANTITATIVE EXAMPLES. 123 
 
 quo sybil-lmi monuere ver-sus 
 6 virgines- lec-tas puero's-que cas-tos- 
 
 dis, quibus- sep-tem- placue're col-les, 
 8 dicere car-men. 
 
 haec- jovem- sen-tire deo's-que cun-ctos 
 10 spem- bonatn- cer-tam-que domum- repor-to, 
 
 doc-tus tt- phoebi chorus et- dianae, 
 12 dicere laudes. 
 
 PJoR. Carm. Saec. I 8. 73 76. 
 
 K. 
 
 The "Just" Man. 
 See pp. 25. 74. 75. 106*. 
 
 jus-tu m ^ / et tenacem- propositi virum- 
 2 non civiu m w ar-dor- prava juben-tium-, 
 
 non vul-tus m-stan-tis- tyran-m 
 4 men-te quatit- solida ; neque^auster-, 
 
 dux- in-quieti tur-bidus had-riae, 
 6 nee- ful-minan-tis- magna jovis- manus-, 
 si frac-tus il-labatur or-bis 
 im-pavidum- ferien-t rumae. 
 
 HOR. OOd. iii. 3, i 8. 
 
 L. 
 
 Tke Usurer's Anticipated Country Pleasures. 
 See pp. 42. 68. 77*. 107*. 
 
 beatus il-le, qui procul- neg5-tiis 
 2 ut- pris-ca gen-s mor-talium-, 
 
 pater-na riira bubus ex-er-cet suis- 
 4 solutus om-m foenere ; 
 
 neqiie^ex-citatur- clas-sico miles- truci, 
 6 ' nequejhor-ret Iratum- mare, 
 
124 QUANTITATIVE EXAMPLES. [L. M. 
 
 forum-que vitat et- super-ba civium- 
 8 potentiorura- limina. 
 
 libet- jacere modo sub an-tiqua^ilice, 
 10 modojm- tenaci gramme. 
 
 at- quum- tonan-tis an-nus hiber-nus- jovis 
 12 im-bres nive's-que com-parat, 
 
 aut trudit acres hin-c et hin-c multa cane 
 14 apros in ob-stan-tes plagas ; 
 
 aut amite lev! rara ten-dit- retia, 
 1 6 tur-dis edacibus- dol5s ; 
 
 pavidum-que lepore m ^et ad-venam- laqueo gruem-, 
 1 8 jucun-da cap- tat- praemia. 
 
 haec ubi locutus- foenerator al-fius, 
 20 jam-jam- futurus- rus-ticus, 
 
 om-nem- red ! "git Tdibus- pecuniam- ; 
 22 quaerit- calen-dis- p5nere. 
 
 HOR. Ep. 2, i 8. 23 24. 2936. 6770. 
 
 M. 
 
 Natural Melody as shewn by Latin Accent. 
 See pp. 8 note, 72. 79. 87. 108*. 109*. 
 
 es-t aute m in- dicen-do etiam- quidam- can-tus ob-scurior. 
 2 in-quo il-lud etiam- notan-dum- mihi videtur ad- studium- 
 
 per-sequen-dae svavitatis in- v5cibus. ip-sa enim- natiira, 
 4 quasi modularetur hominu m 6ratione m , in om-m ver-bo posuit 
 
 acutam- vocem-, nee una plus-, nee a pos-trema syl-laba citra 
 6 ter-tiam- : quo magis-naturam- duce m ad aurium- volup-tatem- 
 
 sequatur in-dus-tria. ac-v5cis quidem- bonitas op-tan-da est. 
 8 non est eni m in- nobis-, sed-trac-tatio at-que iisus in- nobis. 
 
 er-go il-le prin-cep-s variabit et- mutabit ; om-nes sonorum-, 
 10 tu m in-ten-den-s, turn- remit-ten-s, per-sequetur- gradiis. Cic. 
 
 OOrator. 579. 
 
N. O. QUANTITATIVE EXAMPLES. 125 
 
 N. 
 
 Natural Feeling of Uneducated People for Long and Short 
 Quantities, and High and Low Pitch. 
 
 See pp. 2 note, 9 note, 79. 109*. 
 
 in- ver-su quidem- theatra tota ex-claman-t si fuit lina syl- 
 2 laba brevior aut lon-gior. nee ver5 mul-titiido pedes- novice 
 
 nee ul-los- numeros- tenet : nee il-lud- 5 quod of-fen-dit, aut 
 4 cur, aut in- quo oMen-dat, in-tel-ligit : et- tajnen . /pm-nium- ^ 
 
 lon-gitudinu m et- brevitatu m in- soms-, slcut acutarum- gravium- 
 6 que vocum- judiciu m ip-sa natura in auribus- nostris- col- 
 
 locavit. Cic. OOrdtor. 173. 
 
 O. 
 
 Quantitative Rhythm in Prose, and the Resemblance of Greek Lyrics 
 and Latin Comic Senaril to Prose Rhythms. 
 
 See pp. 7. 9. 20. 79. 109*. 
 
 es-se er-go in oratione nurnerum- quen-dam- non es-t dif- 
 2 ficile cog-nos-cere. judicat enim- sen-sus : in- quo imquu m 
 
 est, quod ac-cidit-, non cog-noscere, si, cur id ac-cidat-, re- 
 4 perire nequeamus. neque eni m ip-se ver-sus- rati5ne es-t cog- 
 
 nitus-, 'sed- natura at-que sen-su, quern- dimen-sa ratio docuit-, 
 6 quid ac-ciderit. ita notati5 naturae et animad-versio peperit 
 
 ar-tem-. sed in- ver-sibus- res es-t aper-tior- : quan-qua m etia m 
 8 a modls- quibus-dam- can-tu remote, soluta es-se videatur oratio. 
 
 max-ime'que in op-timo quoque eorum- poetarum-, qui \vptKol 
 10 a graecis- nominan-tur- : quos cum- can tu spoliaveris-, nuda 
 
 paene remanet oratio. quorum- similia sun-t quaeda m etia m 
 12 apud- nos-tros, quae, nisi cum- tibicen ac-ces-sit, oration! sun-t 
 
 solutae simil-lima. at- c5micorum- senarii prop-ter- similitu- 
 14 dinem- ser-m5nis sic saepe sun-t ab-jec-ti ut- non-nun-quam- 
 
 vix in his numerus et- ver-sus in-tel-ligi possit. quo es-t ad 
 1 6 in- venien-dum- dif-ficilior in oratione numerus-, qua in- 
 
 ver-sibus. Cic. Q Orator. 1834. 
 
126 QUANTITATIVE EXAMPLES. [P. Q. 
 
 . 
 P. 
 
 Oratorical Effect of the Rhythmic Collocation of Words in Prose. 
 See pp. 9. 18. 26. 65. 79. 109*. 
 
 .... per-soliitas." dlchore-us-. nihil eni m ad- re 1 ", 
 2 ex-trema il-la, lon-ga sit an brevis. [quia pos-trema syl-laba 
 
 brevis an- longa sit, ne in-ver-su quidem- refert. 217] dein- 
 4 de, "patris- die-turn- sapien-s, temeritas filii com-probavit. " 
 
 hoc dichoreo tan-tus- clamor- concionis ex-citatus es-t, ut acl- 
 6 mirabile es-set. quaerS, non-ne id- numerus ef-fecerit-? ver- 
 
 boru m or-dine m im-muta. fac- sic-, "com-probavit filii terae- 
 8 ritas"; jam- nihil erit, et-si "temeritas" ex tribus- brevibns 
 
 et- lon-ga es-t : que m Aris-toteles ut op-timum- probat, a quo 
 10 dis-sen-tio. at eadem- ver-ba, eadem- sen-ten-tia, animo 
 
 is-tuc- satis es-t, auribus- non- satis. ClC. OOrdtor. 214 5. 
 
 Q. 
 
 On Running Words together. Slurred Vowels^ Open and Gating 
 Voivels, and the Poetic Use of Open Vowels. 
 
 See pp. 16 note, 19. 36. 42. 53. 55. 79. no*. 
 
 ut in- Iegen-d5 oculus-, sic animus in- di-cen-d5 prospiciet- 
 2 quid- sequatur, ne ex-tremorum- ver-borum- cu m m-sequen- 
 
 tibus- prlmis- con-cur-sus, aut hiul-cas- voces ef-ficiat aut as- 
 4 peras-. quam-vls enim- svaves- graves-que sen-ten-tiae, tamen- 
 
 si in-con-ditls- ver-bis ef-ferun-tur, of-fen-den-t aures ; quaru m 
 6 es-t judicium- super-bis-simum-. quod- quidem- Latma lin-gua 
 
 sic ob-ser-vat-, nemo ut- tarn- rus-ticus sit-, quin- vocales- 
 8 nolit- con-jun-gere sed- grace! viderin-t ; nobis-, ne 
 
 si cupiamus- quidem-, dis-trahere voces- con-ceditur. in-dican-t 
 
Q.] QUANTITATIVE EXAMPLES. 127 
 
 10 orationes il-lae ip-sae hor-ridulae catonis : in-dican-t om-nes- 
 
 poetae, praeter e5s-, qui ut- ver-sum- faceren-t, saepe hiaban-t : 
 12 ut en-nius semel- ; "sclpio in-vic-te," et- quidem- nos: "hoc- 
 
 motu radian-tis etesiae in- vada pon-ti." h5c idem- n5s-tri 
 14 saepius- non tulis-sen-t, quod- graecl laudare etiam- solen-t. 
 
 Cic. Orator. 150. 152. (See also 153162.) 
 
 See hiatuses in Ramsay's Prosody, p. 115, where he quotes 
 
 from Virgil, the hiatus being indicated by a dash : 
 
 Amplrion Dircaeus in Actaeo Aracyntho. 
 
 EC. ii. 24. 
 addam cerea pruna honos erit huic quoque porno. 
 
 EC. ii. 53. 
 et siicus pecorl et lac subducitur agnls. 
 
 EC. iii. 6. 
 et, longum, formose vale, vale in quit, Tola. 
 
 (the e in the second vale shortened,) EC. iii. 79. 
 stant et juniper! et castaneae hirsutae. 
 
 EC. vii. 53. 
 credimus, an qui amant, ipsi sibi somnia fingunt. 
 
 (the i in qut shortened,) EC. viii. 108. 
 et vera incessu patuit dea ille ubi matrem. 
 
 Aen. i. 405. 
 
 insulae Tonio in magno, quas dlra celaen5. 
 
 (the ae in ir.sulae shortened, but o in wnio slurred, which word, 
 but for the necessity of verse, "ut versum faceret," would have been 
 ionio, with the first short and second long,) 
 
 Aen. iii. 211. 
 
 nomen et arma locum servant ; te amice nequivi. 
 (the e in te shortened,) Aen. vi. 507. 
 
 Ramsay c,uotes also many other passages from Catullus, Virgil, 
 Horace (OOd. i, 28, 24; ii, 20, 13; Ep. 5, loo, not in Sat., 
 except the passage with num, supra p. 65, where he adds the con- 
 jecture " coctove num adest,") Propertius, Ovid, among Augustan 
 writers ; shewing that the practice was not uncommon, and justifying 
 Cicero's remark, (Q. n.) 
 
128 QUANTITATIVE EXAMPLES, [R. 
 
 . 
 
 Sk. 
 
 R. 
 
 The Story of Gracchus 's Piper and its Theory. 
 See pp. 20. 72. 79. in*. 
 
 quid ad aures- nos-stras et aotionis- svavitatem-, quid es-t 
 
 2 vicis-si-tudine, et varietate, et com-mutatione ap-tius? itaque 
 
 idem- grac-chus-, (quod- potes audire, catule, ex licino clien-te 
 
 4 tuo, Iit-terat5 homini, quern- ser-vum- sibi il-le habuit ad 
 
 manum-), cu m ebuf-neola solitus est habere fistula qui staret 
 
 6 oc-cul-te pos-t ip-sum- cum- con-tionaretur-, perilu m hominem-, 
 
 qui in-flaret- celeriter eum- so-num-, quo iMu m aut remis-su m 
 
 8 excitaret aut a con-ten-ti5ne revocaret. 
 
 audivi, meher-cule, in-quit catulus, et saepe su m ad-miratus 
 10 hominis- cum- diligen-tiam- tu m etiam- doc-trina m et scien-tia m . 
 ego vero, in-quit- cras-sus, ac doleo quide m il-los viros in 
 12 earn- fraude m in republica es-se delap-sos .... 
 
 mit-te, ob-secro, in-quit-, cras-se, Julius, ser-mone m is-tu m , 
 i^ et te ad- grac-ch! fis-tulam- refer : cujus ego non-clum- plane 
 
 ratione m in-tel-ligo. 
 
 1 6 in om-nl voce, in-quit cras-sus, es-t quid-dam- medium-, sed- 
 
 suum- cuique voci. hinc gradati m ascen-dere voce m utile et 
 
 1 8 suave es-t. na m a prin-cipi5" clamare, agres-te qmd-da m es-t, 
 
 et il-lud, ide m ad- nr-man,-du m est vocem- salutare. dein-de 
 
 20 es-t quid-dam- con-ten-tionis ex-tremum-, quod- tamen m-ferius 
 
 es-t qua acutis-simus- clamor-, quo te fis-tuja pr5gredi non 
 
 22 sinet, et- tamen ab ip-sa con-ten-tione revoc&bit. es-t item- 
 
 con-tra quid-da m in remis-sione gravis-simum-, quoque tan- 
 
 24 quam- sonorum- gradibus-. descen-ditur. ha.ec varietas, et hic- 
 
 per om-nes- sonos- vocis cur-sus, et- se tuebkur, et ac f -ti6ni af- 
 
 26 feret- svavitatem-. sed fis-tulat5rem- domi relin-qifetis, sen- 
 
 su 1 ^ hujus con-sue-tudinis v5bis-cu m ad- forum- deferetis. 
 
 ClC. De OOrdtore. lib. iii. 225228. 
 
S.] ITALIAN EXAMPLE, 129 
 
 Opening of "Jerusalem Delivered" an Example of Slurred Voweh 
 and Unwritten Assimilated Consonants. 
 
 See pp. 37. 52. 54. 55*, 113, 
 
 -canto Parmi pietose w , e'l capitano 
 2 -che'l gran sepolcro libero -di cristo. 
 
 -molto^egli^oprd -col senno w , e -con la mano; 
 4 -molto soffri -nel glorioso w acquisto ; 
 
 e w invan 1'inferno^a -lui s'oppose^, e^invano 
 6 -s'armo -d'asia^e -di libia w il popol misto ; 
 
 -che^il -ciel gli die -favore w , e -sotto w ai santii 
 8 segni ridusse^i suoi compagni^erranti, 
 
 o musa tu, che di caduchi^allori 
 10 non circondi la fronte w in elicona, 
 
 ma su nel cielo^infra^i beati co-ri 
 12 hai di stelle^immortali^aurea corona; 
 
 lu spira^al petto mio celesti^ardori, 
 14 tu rischiara w il mio canto^, e tu perdona 
 
 se^intesso fregi^al ver, s'adorno^in parte 
 1 6 d'altri diletti che de' tuoi le carte. 
 
 TASSO. Ger. Lib, i. I 16, 
 
INDEX OF AUTHORS CITED. 
 
 THE numbers in parentheses refer to the passage in the works of the author 
 cited. The other numbers give the page in this tract. An added n refers to a 
 note. 
 
 Academy, on Aspiration, 35 n. CICERO (continued) 
 
 BARTHOLOMEW, title, 2 n. on the natural feeling of unedu- 
 
 BLAIR, title, i n. cated people for long and short 
 
 (P 95)> refers to Lachmann on Lu- quantities and high and low pitch 
 cretius, 44 n. (ibid. 173), App. N, p. 125 ; and 
 
 (P- 99)> n transitional muffled in, how to read the same, 108, 109. 
 
 46 n. on quantitative rhythm in prose, 
 
 BONAPARTE, Prince Louis Lucien, his and on the resemblance of Greek 
 
 rule for "energetic" Italian conso- lyrics and Latin comic senarii, to 
 
 nants, 54 n. prose rhythms (ibid. 183 4) ; App. 
 
 CANNING, his Needy Knifegrinder, O, p. 125 ; and how to read the 
 
 75 n. same, 109. 
 
 CATULLUS, on inserted h (Odes Ixxxii. feet to be measured by syllables 
 
 or Ixxxiv), 35 . _ not intervals (ibid. 194), 25 n. 
 
 has written pure iambics (iv. xxvii.), on regulation of prose rhythm by 
 76. quantity (ibid. 194 5), 80. 
 
 CICERO : on the oratorical effect of rhythmic 
 
 allows simple feet to be of more than collocation of words in prose (ibid. 
 three syllables, 23. 2145), App. P, p. 126 ; how to 
 
 to Terentia and Tulliola, 88. read the same, 109, 1 10. 
 
 on natural melody as shewn by Latin on the nature of Gracchus's Ora- 
 accent (OOrdtor, 57 9). App. tory (De OOrdtore, book iii. 214), 
 M, p. 124 ; how to read the same, 112. 
 
 108, 109. on the story of Gracchus's piper 
 
 quotes Lucllius (ibid. 149) 42. and its theory (ibid. 225 228), 
 
 on running words together, slurred App. R, p. 128 ; and how to read the 
 vowels, open and gaping vowels, and same, in. 
 
 poetic use of open vowels (ibid. CORSSEN, title, i n. 
 
 150, 152), App. Q., p. 126; how to final m omitted in inscriptions (I. 
 
 read the same, no, in. 267 271), 47, 48. 
 
 on various effects of running words on ejus as eijus (I. 301 3), 66 n. 
 together, omitted consonants, &c. on Betonung, or Accentuation (II. 
 (ibid. i 53 162), 'referred to but not 794 892), 9 n. 
 
 cited, in App. Q, p. 127. on distinction of words by differ- 
 
 on cun nobls (ibid. 154), 55 n. ence of accent (II. 808), 79 n. 
 
 vowels in com, in long before fol- on accentual hexameters in third 
 lowing f, s, but short otherwise (ibid. century A.D. (II. 942), 84 n. 
 
 159), 17 n. CURWEN, title, 12 n. 
 
 used to omit h, m, in pttkros, &c. on rhythm, 12 n ; 
 (ibid. 1 60), 35 n. on setting pattern, 14 n. 
 
 there must be no verses in prose DANIEL, Thesaurus Hymnologicus, 39. 
 writing according to Aristotle {ibid. DIEZ, on final nasals, 52 n. 
 
 170173), 9 n. DINDORF, his edition of Euripides, 28 n. 
 
INDEX OF A UTHORS CITED. 
 
 DIONYSIUS of Halicarnassus ; identity 
 of accent and alteration of musical 
 pitch (jrepi avvtito-ewg bvo/jLardav, ch. 
 xi.), 27 11. 
 
 DONATUS, on mussitare, 58 n. 
 
 ELLIS (ROBINSON), on the present state 
 of the pronunciation of Latin in 
 English schools and colleges, p. v. 
 
 EURIPIDES, (Orestes 1403), 28. n. 
 
 FAVARGER, French syllabication, 19 n. 
 
 HALDEMAN, title, i n. 
 
 on nasalisation of final w, (Lat. 
 pron. art. 105), 62 n. 
 
 HORACE : 
 
 slurs duwi) turn, num, cum, quum, 
 sum, nam^ quatn, rem, 64. 
 
 dedication of his odes to Maecenas, 
 (OOdae, i. i, i 6, 29 36), App. E. 
 p. 120 ; and how to read the same, 
 102 104. 
 
 Leuconoe recommended not to peer 
 into the future (OOdae, i. n), App. 
 F. p. 121 ; how to read the same, 
 104 105. 
 
 (OOdae, 3, i, 5), read for final m, 69. 
 
 the "just" man {OOdae, iii. 3. 
 i 8), App. K, p. 123 ; how to read 
 the same, 106, 107. 
 
 the usurer's anticipated country 
 pleasures \Epodl, 2, i 8. 23 24. 
 29 36. 6770), App. L, p. 123 ; 
 how to read the same, 107 108, 
 
 beginning and end of hymn to 
 Phoebus and Diana (Carmen Saecii- 
 Idre, i 8. 73 76), App. I. p. 122 ; 
 how to read the same, 106. 
 
 Carmen SaecttZdre, phonetic tran- 
 scription of English reading, sup- 
 posed to be made by Maecenas redi- 
 vlvus, 74 n. 
 
 no one contented with his lot, (Sat. 
 i. i 22), App. D. p. 1 20 ; and how 
 to read the same, 99101. 
 
 dum slurred and positional in one 
 line (Sat. i. 5, 13), 43. 
 
 does he use mtm short before a 
 vowel (Sat. ii. 2, 28) ? 65, 127. 
 
 uses agit- long before a vowel (Sat. 
 ii. 3, 260), 65. 
 
 use//z (A. P. 65), 16 n. 
 
 his remark on eyes and ears (A. P. 
 180), does not apply to reading and 
 speaking, 45 n. 
 
 on .the introduction of spondees into 
 iambics (A. P. 251262, 77 n. 
 
 cannot scan Plautus(^l. P. 270 
 T 4), 75- . 
 
 JOBERT, title, 19 n. 
 
 on French syllabication, 19 n. 
 
 KENNEDY, title, i n. 
 KEY, title, 30 n. 
 
 denies acute accent before enclitic, 
 30 n. 
 
 his views of accent and of Quin- 
 tilian's non-accented prepositions and 
 relatives, 32 n. 
 
 notes that later Latin grammarians 
 speak of the laws of accent in the 
 past tense, 32 n. 
 
 LACHMANN, on coopertus, &c. (on 
 
 Lucretius, p. 136), 44 n. 
 LUCILIUS, slurred compostaeut, tes- 
 
 serulaeomnes, 42. 
 
 LUCRETIUS, his use of open um, 53. 
 MEYER, P., quotes old French on 
 
 school Latin, 32 n. 
 
 on French nasality, 62 n. 
 MOZART, extracts from his Nozze, 38. 
 MUNRO, title, i n. 
 
 admits acute accent before enclitic, 
 30 n. 
 
 OVID, commencement of Penelope's 
 epistle to Ulysses, (Her. i, i 10), 
 App. H. p. 122, and how to read 
 the same, 105-106. 
 
 PALMER and MUNRO, title, i n. 
 
 PHAEDRUS, his prologue, 78. 
 
 PRISCIAN on obscure, open, and mid- 
 dle m, 58 n. 
 
 on distinguishing words by differ- 
 ence of accent, 79 n. 
 
 PROPERTIUS, while Ponticus writes 
 heroics, keeps to erotics (Eleg. 7, 
 i 8). App. G., p. 121 : how to read 
 the same, 105. 
 
 PROUT (FATHER), his Latin rhymes 
 on eggs and bacon, 86. 
 
 QUINTILIAN, from his Instittitio 
 OOrdtona. 
 
 on Cicero's aijo maija. (i, 4, n) 
 66 n. 
 
 quantity of syllables changed poeti- 
 cally (i, 5, 1 8), 17 n. 
 
 variability of h (i, 5, 19 21), 36 n. 
 
 prepositions and relatives are with- 
 out accent when they are added on 
 to following word (i, 5, 27), 30 , 
 32 n. 
 
 the position of the accent alters 
 with the quantity of the syllable 
 before, mute and liquid (i, 5, 28), 
 17 n. 
 
 rule for the place of the accent (i, 
 
 5> 3- 3 1 ) 3 1 n -' 
 
 on marking length of vowels (i, 
 7, i 3), 16 n. 
 
 ad and at, cum and ^wum (i, 7, 5), 
 
 44 5 
 
Emm 
 
VStf: ilM 
 
 P