[2] Reference may also be made to pages 27-29, 33, 34, 40-55, 74-85, 92-95, 98 ff., 122-127, 134-137, 154-167, 184-193, 200-207, 236-241, 264-281. Especially to be noticed is that warm praise of simplicity (pp. 76-85, 134-137) which should suffice to prove that Dionysius is not a ‘rhetorician’ in any invidious sense.

[3] See Glossary, s.v. s???es??.

[4] de Isocrate c. 2, d???e?e? ??? ? d?????a p??????? t? ???? t?? ???e??, ?a? t?? ????? ?e?peta? t? ???????? ... ???eta? d? ? f?s?? t??? ???as?? ?pes?a? t?? ?????, ?? t? ???e? t? ???ata.

[5] The Greek word (?ef??a?a, capita) corresponding to ‘chapters’ occurs several times in the C.V. (see Glossary, s.v.); and one (pe?????) of the words corresponding to ‘paragraph’ is found in the de Thucyd. c. 25. The paramount importance and dignity of the p?a?at???? t?p?? is indicated in the C.V. 66 9-15, and in the de Demosth. c. 58 fin.

[6] Quintilian (Inst. Or. ix. 4. 23) applies the term naturalis ordo to such collocations as viros ac feminas, diem ac noctem, ortum et occasum. But even here the order, though perhaps natural, is certainly not necessary.

[7] A good example of the severance of ?????? from its article by an adjectival phrase will be found in the C. V. itself, 222 22: ??f??? ??? ?f???? s???pteta? t? ? t? t ?a? d?a???e? ????????? d??as?? ? eta?? t?? te p??s???????? t?? “pa?da?da???” ?a? t?? s??a???f?? t?? s??apt????? a?t? ??????. The convenience of this articular bracket is obvious.

[8] Cp. ??????? ... p???a?????t??, Hom. Il. ii. 459-63.

[9] Attention is called to the elaborate word-order by Mr. P. N. Ure in his edition of this portion of Thucydides. The extent to which prepositions can be parted from cases, in post-Homeric as well as in Homeric Greek, is worth notice as a somewhat different illustration of the freedom of Greek order. See, for example, the remarks in Liddell and Scott’s Lexicon on the position of e??.

[10] In Caesar B.G. ii. 25 more than a hundred words come between the subject, Caesar and the main verb processit.

[11] e.g. ‘A quarrel had arisen between a big and a little boy about a big and a little coat.’

[12] A good illustration of the freedom of order possible (at any rate theoretically) in Greek, even within the limits of verse, is supplied in a letter from Richard Porson to Andrew Dalzel: “There is a passage of Sophocles three times quoted by Plutarch, and always in a different order, but so as in the three variations to remain a senarian. Now the fragment consists of five words, and the sense is this: ‘(The physicians) wash away bitter bile with bitter drugs [p?????? p????? ??????s? fa?????? ?????].’ The five words, you know, will admit of one hundred and twenty permutations, and what is extremely odd, these words will admit twenty transpositions [which Porson proceeds to indicate], and still constitute a trimeter iambic.”—Luard’s Correspondence of Richard Porson pp. 91, 92.

[13] Horace Ars Poetica 40,

cui lecta potenter erit res,
nec facundia deseret hunc nec lucidus ordo.

Can the obscure potenter here be a Latin translation of some such technical term (found by Horace or Neoptolemus in the Greek writers on literary criticism) as d??at?? or de???? or p??a????

[14] Demetrius, for example, evidently expects to find more lucidity in the plain style (the ?s???? ?a?a?t??) of a Lysias than in the elevated style (e?a??p?ep?? ?a?a?t??) of a Thucydides: see the summary in Demetrius on Style pp. 33, 34. And a principal reason for this is that the former keeps more closely than the latter to the normal order of words in Greek (de Eloc. §§ 191 ff.). For Herodotus as compared with Thucydides cp. de Imit. ii. 3. 1 t?? saf??e?a? d? ??af?s?t?? ???d?t? t? ?at????a d?d?ta? (quoted in the editor’s Dionysius of Halicarnassus: the Three Literary Letters p. 173).

[15] e?a????t?? ??? t???? e?s?? ???? p??ta t? T????d?d?? s?a?e??, ?a? ??d’ ??t?? ????? ?????se?? ??aat???? ???a, de Thucyd. c. 51.

[16] ?? ??? ????a???? ?????p??? ??d’ ?p?d?f????? ? ?e???t???a?? ??d? t??? ?????? ?? ? et?s??? ?????? ??e??e???? ta?ta? ?atas?e???es?a? t?? ??af??, ???’ ??d??s? d?? t?? ????????? a???t?? ?p? ??t?????? te ?a? f???s?f?a? ???????s??, ??? ??d?? fa??seta? t??t?? ?????, de Thucyd. c. 50. A comprehensive condemnation of ?s?fe?a is found in the same essay, c. 52: ? p??ta ??a?????? t? ?a?? ?a? s??t?? pa?????sa ta?? ??eta?? ?s?fe?a.

[17] See, further, the Appendix headed “Obscurity in Greek.”

[18] In the same way, Dionysius must surely feel the loss both of clearness and of emphasis involved in transferring ? ??? ??p?? (112 1 and 4) from the middle to the end of the sentence. ????? and p???? may cover these cardinal points: “no clearness no charm,” he might well say,—“no emphatic order no full expression of feeling.”

[19] Cp. Demetrius on Style p. 278 (Glossary, s.v. ?fas??).

[20] Cp. Lewis Campbell in the Classical Review iv. 301, and Goodell in the paper named on p. 33 infra. In the matter of emphasis, Greek sentences are usually constructed on a diminuendo, English sentences on a crescendo principle. The English of ? ’fe??e??? ????? te ?a? ????? ?a (Soph. Antig. 281) is, as Jebb gives it, “lest thou be found at once an old man and foolish.” As fuller examples, in prose and verse, Mr. L. H. G. Greenwood suggests the Phaedrus 230 B, C (?? t?? ??a? ... Fa?d?e) and the Rhesus 78-85, 119-130.

[21] The views of Quintilian and Demetrius with regard to rhythm are applicable also to emphasis: Quintil. ix. 4. 67 “nam ut initia clausulaeque plurimum momenti habent, quotiens incipit sensus aut desinit: sic in mediis quoque sunt quidam conatus, iique leviter insistunt. currentium pes, etiamsi non moratur, tamen vestigium facit”; Demetrius (de Eloc. § 39) p??te? ???? ?d??? t?? te p??t?? ????e??e? ?a? t?? ?st?t??, ?a? ?p? t??t?? ?????e?a, ?p? d? t?? eta?? ??att?? ?spe? ?????pt????? ? ??afa????????.

[22] The initial emphasis is here reinforced by ?? and d?: elsewhere by the chiastic arrangement, as in (10).

[23] Compare the occasional postponement of a relative pronoun with the same object: e.g. Thucyd. i. 77 ???es?a? ??? ??? ?? ???, d????es?a? ??d?? p??sd???ta?.

[24] Our poets can, and do, imitate the emphatic position of a word placed at the beginning of a line with a stop immediately following (as ???’ in Hom. Il. i. 52, ??pt’ in Odyss. ix. 290, and haesit in Virg. Aen. xi. 803):—

And over them triumphant Death his dart
Shook, but delayed to strike.

Milton Paradise Lost xi. 491.

Or (still nearer to the ‘me, me, adsum,’ of Virgil):—

Me, though just right, and the fixed laws of Heaven,
Did first create your leader—next, free choice,
With what besides in council or in fight
Hath been achieved of merit—yet this loss,
Thus far at least recovered, hath much more
Established in a safe, unenvied throne,
Yielded with full consent.

Milton Paradise Lost ii. 18-24.

[25] Here t??t??? is emphasized by ?a? as well as by its position well in front of the verb which governs it, while ?s??? depends for its emphasis on its position alone. ‘But even these hidden piles did divers (entering the water) saw off—for pay.’ Compare the analysis which Quintilian (ix. 4. 29) gives of Cicero’s “ut tibi necesse esset in conspectu populi Romani vomere postridie.”

[26] For the rhetorical and metrical effect Sandys (ad loc.) compares Milton Paradise Lost vi. 912, “Firm they might have stood, " Yet fell.”

[27] In this sentence the orator would probably pause slightly before ?e??a???, and thus (1) emphasize it; (2) separate it from d?d?. Other means (illustrated by various examples in this Introduction) of throwing a word into relief are: the interposition of a number of unemphatic words, the use of particles such as ?? and d?, the placing of emphatic words in contrasted pairs near together or remote from one another.

[28] The order here (1) avoids the juxtaposition of too many accusative-terminations; (2) provides a conclusion which satisfies ear and mind alike.

[29] The position of t?’ here may be compared with that of ???? in Eurip. Med. 1045 ??? pa?da? ?? ?a?a? ???? (‘for they are mine’). In English, too, both the end and the beginning may be emphatic: e.g. “silver and gold have I none.”

[30] Quoted by Dionysius (C.V. c. 3), though without any special reference to the point of emphasis.

[31] Quoted by T. D. Goodell School Grammar of Attic Greek p. 296. ?e?? seems to owe some at least of its emphasis to its late insertion. If placed immediately after ????sae?, it would, surely, lose a little in weight. Goodell does right to include some treatment of the question of Greek word-order in a Grammar intended primarily for use in schools. It should be pointed out even to beginners that so simple a sentence as ?? d’ ????a??? ?????sa? t??? ?a?eda??????? can be arranged in half-a-dozen ways, each with its own separate shade of meaning. Compare the remarks of W. H. D. Rouse with regard to the teaching of Latin: “It is possible by question and answer to make clear from the first the essential structure of an inflected language, as depending for emphasis on the order of words; and this lies at the root of style. Thus a simple sentence may give matter for several questions. Take Caesar Labienum laudat. I may ask, Quem laudat Caesar? Answer: Labienum laudat Caesar. Question: Quid facit Caesar? Answer: Laudat Labienum Caesar. If all the texts read are treated in this way, the pupils become used to correct accidence, syntax, and order, and learn the elements of style” (Classical Review xxi. 130; cp. also W. H. S. Jones The Teaching of Latin p. 33). An instructive contrast might be drawn, with reference to the context in either case, between Romanus sum civis in Livy ii. 12, and Civis Romanus sum in Cicero Verr. II. v. 65, 66.

[32] With “verbi transgressio” cp. “verborum concinna transgressio” in Cic. de Orat. iii. 54. 207.

[33] A modern reader might be disposed to see an example of emphasis in the illustrative passage which “Longinus” here quotes from Herodotus vi. 11. In hyperbata the Treatise on the Sublime itself greatly abounds, being much influenced (in this as in other ways) by Plato. For examples of hyperbaton in Plato see Riddell’s edition of the Apology, pp. 228 ff. Among modern English writers, Matthew Arnold had a curious and perhaps half-humorous trick of securing emphasis by a “bold and hazardous” hyperbaton (cp. de Sublim. xxii. 4), which keeps back the verb till the end of the sentence: e.g. “And a good deal of ignorance about these there certainly, among English public men, is”; “the grand thing in teaching is to have faith that some aptitudes for this every one has”; “one thing that Protestants have, and that the Catholics think they have a right, where they are in great numbers, to have too, this thing to the Prussian Catholics Prussia has given.” Such oddities are, in English, usually of a playful and undress character: e.g. “it was really a party that one might feel proud of having been asked to; at least I might, and did, very” (Life and Letters of Sir Richard Claverhouse Jebb p. 93; cp. J. D. Duff’s remarks, on the same page, with regard to the literary adequacy of the following English translation of a pathetic sentence in one of Demosthenes’ greatest speeches: “this woman in the first instance merely quietly to drink and eat dessert they tried to force, I should suppose”).

[34] The immediately preceding sentence in Quintilian is “venio nunc ad ornatum, in quo sine dubio plus quam in ceteris dicendi partibus sibi indulget orator.” This may be compared with Dionysius’ view that it is the accessory arts (such as the heightening of style) that best reveal the orator’s power: ?? ?? ???sta d??d???? ? t?? ??t???? ???eta? d??a?? (de Thucyd. c. 23). In this attitude there is always some danger (unless, like Dionysius himself, a writer has a saving belief in the virtue of simplicity) of falling into that vice of Écrire trop bien, which, according to M. Anatole France, is the worst of all literary vices.

[35] If we were to say that in a Greek sentence there are two kinds of arrangement, viz. (1) grammatical arrangement which aims at clearness, and (2) rhetorical arrangement which aims at (a) emphasis, and () euphony; then it must be admitted that Dionysius’ real subject is (2) ()

[36] The lines quoted from Homer in c. 16 are particularly telling.

[37] C.V. 244 23. Perhaps ‘spontaneous’ or ‘subconscious’ would be a better translation than ‘instinctive.’ Dionysius certainly does not intend to exclude training.

[38] The judgment of the ear appears to be indicated by the words t?? p???? etap?pt??t?? ???t????? at the end of c. 24.

[39] Cp. C.V. c. 6.

[40] Cic. ad Att. xiv. 20. Dionysius Halic. Ant. Rom. i. 1 ?p?e???? ??? ?pa?te? ??????s?? e????a? e??a? t?? ???st?? ????? t??? ??????. Buffon Discours de rÉception À l’AcadÉmie, 1753: “le style est l’homme mÊme.” Cp. Plato Rep. iii. 400 D t? d’ ? t??p?? t?? ???e??, ?? d’ ???, ?a? ? ?????; ?? t? t?? ????? ??e? ?peta?;

[41] Cp. p. 24 supra. The desire to avoid monotony of termination would seem to be the main explanation of such collocations as ?? t??? ?????? e???es?a? p??a???e???s? t??? t?? f???? fe????s? t?? d??a? and t? a?t? ????ta? ??? t??t? [Antiphon v.]. Additional emphasis, too, falls on t??? ?????? and t? a?t?, as on s?t???a? ?sfa?? in Demosthenes’ peroration.

[42] In describing the smooth or elegant style of composition (as practised by Isocrates and his followers, including Theopompus), Dionysius notes, as one of its characteristics, the avoidance of hiatus. This avoidance is to be noticed in the recently discovered Hellenica; and without basing any positive conclusion on the fact, Grenfell and Hunt point out that the author usually avoids hiatus “even at the cost of producing an unnatural order of words, e.g. ?p?????? ?se?? ?sa? t??? ?a?eda??????? and ??e?, ? ??d?e?, ?f?, p???ta?, ?p? t??? t????????” (Oxyrhynchus Papyri v. 124).

[43] e.g. the greater tendency in Latin to place the principal verb at the end of the sentence. Cp. Quintil. ix. 4. 26 “verbo sensum cludere, multo, si compositio patiatur, optimum est. in verbis enim sermonis vis est. si id asperum erit, cedet haec ratio numeris, ut fit apud summos Graecos Latinosque oratores frequentissime. sine dubio erit omne quod non cludet, hyperbaton, et ipsum hoc inter tropos vel figuras, quae sunt virtutes, receptum est.” In Latin the words et? d? ta?ta ?? p???? ?ste??? ????a ?p?st? ?p’ ????a??? would naturally run “haud multum postea Euboea ab Atheniensibus defecit” (J. P. Postgate Sermo Latinus p. 7).

[44] On the other side, the classical writers not seldom yield to the temptation to write long and rambling sentences, whereas the best English authors are stimulated by the very absence of inflexions to arrange their thoughts with great care and clearness within the sentence and the paragraph. By these and other means English prose becomes, in the hands of a great master, an instrument of surpassing force and beauty. As there are differences in word-order between Greek and Latin, so are there among the modern analytical languages, though (in a comparison) it may be legitimate to group those languages together. An order regarded as natural (i.e. customary) in one modern language will not be so regarded in another. Further, a language like German (though it is often unable to follow the Greek order without ambiguity: cp. Lessing’s Laocoon c. 18) possesses a greater number of inflexions than English or French. Welsh, too, has certain syntactical features which enable it often to reproduce the Greek order more faithfully than English can do. For example: in St. John’s Gospel xvii. 9 where the Greek has ?? pe?? t?? ??s?? ???t?, ???? pe?? ?? d?d???? ??, ?t? s?? e?s??, the Welsh version gives Nid dros y byd yr wyf yn gweddio, ond dros y rhai a roddaist i mi; canys eiddot ti ydynt. And Plato Apol. c. 33 ?a? ??? ta?ta p???te, d??a?a pep????? ??? ?s?a? ?f’ ???, a?t?? te ?a? ?? ??e??: Welsh, Ac os hyn a wnewch, yr hyn sydd gyfiawn fyddaf fi wedi ei dderbyn oddiar eich llaw, myfi a’m meibion. [These Welsh instances are given on p. 38 of the present editor’s chapter on the Teaching of Greek, in F. Spencer’s Aims and Practice of Teaching.] In Appendix II. at the end of this volume will be found a few idiomatic modern renderings (in English, French, and German) from Greek prose originals.

[45] LemaÎtre Les Contemporains i. 205.

[46] Boileau L’Art poÉtique i. 133.

[47] Edinburgh edition of Stevenson’s works, iii. 236-61 (Miscellanies). “It is a singularly suggestive inquiry into a subject which has always been considered too vague and difficult for analysis, at any rate since the days of the classical writers on rhetoric, whom Stevenson had never read” (Graham Balfour’s Life of Robert Louis Stevenson ii. 11). S. H. Butcher (Harvard Lectures pp. 242, 243) regards the essay as “a pretty precise modern parallel to the speculations of Dionysius,” and quotes some passages in proof. The following is an example of such points of contact. Stevenson: “Each phrase in literature is built of sounds, as each phrase in music consists of notes. One sound suggests, echoes, demands and harmonizes with another; and the art of rightly using these concordances is the final art in literature.” Dionysius (C.V. c. 16): ?ste p???? ?????? ?a??? ?? e??a? ????? ?? ? ?a?? ?st?? ???ata, ?a??? d? ????t?? s???a?? te ?a? ???ata ?a?? a?t?a e??a?, ?de??? te d???e?t?? ?? t?? ?d????t?? t?? ????? ???es?a?. Compare p. 40 infra as to the music of sounds; and see Demetrius on Style p. 43, as to Stevenson and other English writers on style.

[48] Compare especially the speeches in Il. ix., and the warm eulogies they have drawn from Quintilian (x. 1. 47; cp. x. 1. 27, with reference to Theophrastus) and from many others since his time. Dionysius’ versification of Demosthenes, and prosification of Simonides, in c. 25 and c. 26, may not seem altogether happy, but one or two points should be remembered in his favour. He does not recognize merely mechanical conceptions of literature: such as are implied in the Latin-derived words prose and verse, or in literature itself. He would probably have agreed with Aristotle that “Homer and Empedocles have nothing in common but the metre, so that it would be right to call the one poet, the other physicist rather than poet” (Aristot. Poet. i. 9, S. H. Butcher). He might probably have also maintained that, in essentials, Theognis is less of a poet than Plato. And in modern times, if he had known them, he might have called attention to the rhymed rhetoric which often passed as poetry in eighteenth-century England, and have asked whether the elevation of thought and the measured cadences of Demosthenes did not entitle him to a higher poetic rank than that.

[49] Of Thucydides: p???t?? t??p?? ??e???s????? (de Thucyd. c. 24). Of Plato: ?s?et? ??? t?? ?d?a? ?pe????a??a? ?a? ???a ??et? a?t? t? d????a??, ? ??? ?? ?d?s??? ??? ???e?? ?????? ??. t??t? d? pa?e?? ????e?, ?? ??? ?????, t?afe?? ?? ?? t??? S???at????? d?a?????? ?s???t?t??? ??s? ?a? ????est?t???, ?? e??a? d’ ?? a?t??? ???? t?? G?????? ?a? T????d?d?? ?atas?e??? ??as?e?? (Ep. ad Cn. Pomp. c. 2; de Demosth. c. 6. See further in Demetrius on Style p. 14, n. 1).

[50] It will be noticed that the only question here is about differences of form. But it is one of Dionysius’ great merits to have proclaimed so clearly the leading part which beauty of form (not simply verse, but expression generally) plays in all high poetry. Aristotle was by no means insensible to this essential element, but he is apt to dwell more fully (though we must remember the fragmentary condition of the Poetics) on the associations of p???t?? than on those of ???d??. It is in connexion with prose rather than with poetry, that it seems necessary to lay most stress upon the intellectual and logical elements involved, and to pay heed not only to the nature of the subject matter itself but to the sustained argument in which it is presented. Reason in prose and emotion in poetry: these are perhaps the two leading elements, if any distinction of the kind is to be attempted.

[51] Aristot. Rhet. iii. 1. 9; 8. 1 and 3; 2. 1. Cp. Cic. Orat. 56. 187 “perspicuum est igitur numeris astrictam orationem esse debere, carere versibus; sed ei numeri poËticine sint an ex alio genere quodam deinceps est videndum”; 57. 195 “ego autem sentio omnes in oratione esse quasi permixtos et confusos pedes; nec enim effugere possemus animadversionem, si semper eisdem uteremur, quia nec numerosa esse, ut poËma, neque extra numerum, ut sermo vulgi, esse debet oratio: alterum nimis est vinctum, ut de industria factum appareat, alterum nimis dissolutum, ut pervagatum ac vulgare videatur.” Also ibid. 51. 172; 57. 194-196; 58. 198; 68. 227. Cicero’s correct attitude is the more noticeable that he is commonly supposed to have been swayed by Asiatic rather than by Attic influences.

[52] C.V. c. 25 ????? d? t?? ???st?t????? a?t???a?, ?t? ??a??a??? ?st?? ?pe???a??es?a? t??a? t? pe?? ???e? ??????, e? ????? t? p???t???? ?pa???se?? a?t? ??????, ?? t?? pe??a? t?? a?t?? ???seta?.

[53] The modern custom is to view with some suspicion these inversions when found in prose composition, though in German prose they are common enough. It would be interesting to take two such sentences of the New Testament as e???? ? ??te?? ?fes??? (Acts xix. 28, 34) and ?pese?, ?pese? ?a???? ? e???? (Apoc. xiv. 8), and see how they have been rendered into various modern languages by translators generally (both in authorised and unauthorised versions). It would probably be found that the French language here has been true to what Dionysius would call its ????e?de?a, or essentially prose character. In English the justification of the inversion would be the emotional nature of the original passages, which may be held to raise them to the same plane as poetry. [It would, on the other hand, be not good but bad journalism to write, “Uproarious were the proceedings at yesterday’s meeting of the Grand Committee.”] For the effect of word-order in English verse see an extract from Coleridge’s Biographia Literaria in the notes, p. 79 infra. Coleridge was fond of offering, as a rough definition of poetry, “the best words in the best order.”

[54] See the notes on c. 25; particularly that on 256 11.

[55] The words “How art thou” are, it will be noticed, differently divided in these two lines with a kind of Dionysian freedom.

[56] Ruskin continually, and Carlyle often (e.g. Sartor Resartus bk. iii. c. 8), provides examples of iambic rhythm. So George Eliot Mill on the Floss bk. vii.: “living through again, in one supreme moment, the days when they had clasped their little hands in love, and roamed the daisied fields together.” And Blackmore, in Lorna Doone c. 3: “The sullen hills were flanked with light, and the valleys chined with shadow, and all the sombrous moors between awoke in furrowed anger.” [Blackmore sometimes falls also into the hexameter rhythm, as in the same chapter: “And suddenly a strong red light, cast by the cloud-weight " downwards, " spread like " fingers " over the " moorland, "" opened the " alleys of " darkness, and " hung on the " steel of the " riders.”]

[57] Cicero’s conception of the requirements of rhythmical prose (as compared with those of verbal fidelity) is curiously illustrated by the way in which he is supposed to have recast the letter sent by Lentulus to Catiline. Sallust Cat. 44 “quis sim ex eo quem ad te misi cognosces: fac cogites in quanta calamitate sis et memineris te virum esse: consideres quid tuae rationes postulent: auxilium petas ab omnibus etiam ab infimis.” Cicero Cat. iii. 12 “quis sim scies ex eo quem ad te misi: cura ut vir sis et cogita quem in locum sis progressus: vide ecquid tibi iam sit necesse et cura ut omnium tibi auxilia adiungas, etiam infimorum.” Cp. A. C. Clark (reviewing Zielinski) Classical Review xix. 172.

[58] Cp. C.V. 176 20 ?? ??? ?pe?a??eta? ????? ??de?? ?? t?? ??t??? ???e??, ?spe? ?? t?? ??t???. With regard to the occasional presence in prose of metrical or quasi-metrical lines, the likely explanation seems often to be one which Dionysius does not favour (p???? ??? a?t?s?ed???e? ?t?a ? f?s??, 256 19), rather than one which recognizes ?t?a ?a? ?????? t??a? ???atateta?????? ?d???? (254 3).

[59] D. B. Monro Modes of Ancient Greek Music p. 118.

[60] From the essay (already mentioned) on Style in Literature.

[61] de Demosth. c. 22.

[62] So that, in 126 15, t?? ???? t???? = ‘the high pitch’ = ‘the acute accent.’

[63] W. H. D. Rouse’s edition of Matthew Arnold on translating Homer Introd. p. 7.

[64] A. J. Ellis and F. Blass (in the publications mentioned later).

[65] Arnold and Conway Restored Pronunciation of Greek and Latin pp. iv. 3, 7, 20-26. Cp. also the pamphlet on the Pronunciation of Greek issued by the Classical Association in 1908 (pp. 348-51 infra). In the Contemporary Review of March 1897 the history of Greek pronunciation in England is ably sketched by J. Gennadius.

[66] Even the pronunciation of the poet’s name has changed with the lapse of centuries; and the spelling Shakspere is preferred by some authorities not only because it has excellent manuscript authority, but because it may serve to remind us that “he and his fellows pronounced his name Shahk-spare, with the a of father in Shahk, and with the French e (our a) in spare” (Furnivall).

[67] Quintil. i. 10. 17 “siquidem Archytas atque Aristoxenus etiam subiectam grammaticen musicae putaverunt,” etc.

[68] C.V. 68 7-11, ... t?? pe?? t?? s????se?? t?? ????t?? p?a?ate?a? ??????? ?? ?p? ???? ?????sa?, ?s?? t?? ???a??? ??t?????? ? d?a?e?t???? s?????a?a? t???a?, ??de?? d’ ?????? ??d’ ?p?????t?? ???? t?? pa???t?? ??e???as????, ?? ??? pe???a?.

[69] Some reference to Quintilian’s own apparent indebtedness to the de Imitatione of Dionysius will be found in Demetrius on Style p. 25.

[70] de Sublim. xxxix. 1. In the editor’s article on the “Literary Circle of Dionysius of Halicarnassus” (Classical Review xiv. 439-42), an endeavour is made to view the literary life of Dionysius in relation to its Roman surroundings.

[71] The more recent writers on rhetoric (?? ???? te??????f??, de Isaeo c. 14) would not greatly appeal to Dionysius.