Dionysius of Halicarnassus On Literary Composition Being the Greek Text of the De Compositione Verborum

CHAPTER XII

Chapter 161,792 wordsPublic domain

HOW TO RENDER COMPOSITION CHARMING

It is not in the nature of all the words in a sentence to affect the ear in the same way, any more than all visible objects produce the same impression on the sense of sight, things tasted on that of taste, or any other set of stimuli upon the sense to which they correspond. No, different sounds affect the ear with many different sensations of sweetness, harshness, roughness, smoothness, and so on. The reason is to be found partly in the many different qualities of the letters which make up speech, and partly in the extremely various forms in which syllables are put together. Now since words have these properties, and since it is impossible to change the fundamental nature of any single one of them, we can only mask the uncouthness which is inseparable from some of them, by means of

[Page 132]

λεῖα μίσγοντα καὶ σκληροῖς μαλακὰ καὶ κακοφώνοις εὔφωνα καὶ δυσεκφόροις εὐπρόφορα καὶ βραχέσι μακρά, καὶ τἆλλα τὸν αὐτὸν τρόπον εὐκαίρως συντιθέντα καὶ μήτ’ ὀλιγοσύλλαβα πολλὰ ἑξῆς λαμβάνοντα (κόπτεται γὰρ ἡ ἀκρόασις) μήτε πολυσύλλαβα πλείω τῶν ἱκανῶν, μήδε δὴ ὁμοιότονα παρ’ 5 ὁμοιοτόνοις μήδ’ ὁμοιόχρονα παρ’ ὁμοιοχρόνοις. χρὴ δὲ καὶ τὰς πτώσεις τῶν ὀνοματικῶν ταχὺ μεταλαμβάνειν (μηκυνόμεναι γὰρ ἔξω τοῦ μετρίου πάνυ προσίστανται ταῖς ἀκοαῖς) καὶ τὴν ὁμοιότητα διαλύειν συνεχῶς ὀνομάτων τε τῶν ἑξῆς τιθεμένων πολλῶν καὶ ῥημάτων καὶ τῶν ἄλλων μερῶν τὸν 10 κόρον φυλαττομένους, σχήμασί τε μὴ ἐπὶ τοῖς αὐτοῖς ἀεὶ μένειν ἀλλὰ θαμινὰ μεταβάλλειν καὶ τρόπους μὴ τοὺς αὐτοὺς ἐπεισφέρειν, ἀλλὰ ποικίλλειν, μηδὲ δὴ ἄρχεσθαι πολλάκις ἀπὸ τῶν αὐτῶν μηδὲ λήγειν εἰς τὰ αὐτὰ ὑπερτείνοντας τὸν ἑκατέρου καιρόν. 15

καὶ μηδεὶς οἰηθῇ με καθάπαξ ταῦτα παραγγέλλειν ὡς ἡδονῆς αἴτια διὰ παντὸς ἐσόμενα ἢ τἀναντία ὀχλήσεως· οὐχ οὕτως ἀνόητός εἰμι· οἶδα γὰρ ἐξ ἀμφοῖν γινομένην πολλάκις ἡδονήν, τοτὲ μὲν ἐκ τῶν ὁμοιογενῶν, τοτὲ δὲ ἐκ τῶν ἀνομοιογενῶν· ἀλλ’ ἐπὶ πάντων οἴομαι δεῖν τὸν καιρὸν ὁρᾶν· οὗτος 20 γὰρ ἡδονῆς καὶ ἀηδίας κράτιστον μέτρον. καιροῦ δὲ οὔτε ῥήτωρ οὐδεὶς οὔτε φιλόσοφος εἰς τόδε χρόνου τέχνην ὥρισεν, οὐδ’ ὅσπερ πρῶτος ἐπεχείρησε περὶ αὐτοῦ γράφειν Γοργίας

2 εὐπρόφορα] εὔφορα F 3 συντεθέντα F 4 πολλὰ ... (5) πολυσύλλαβα om. P. 7 μηκυνόμενά τε γὰρ F: μηκυνόμεναί τε γὰρ M 8 προίστανται F 9 τε τῶν Us.: τέ τινων F, E: τινῶν PMV 11 φυλασσομένους EF: φυλαττόμενον s || ἐπὶ FE: om. PMV || ἀεὶ μένειν EF: διαμένειν PMV 14 ὑπερτείνοντας Us.: ὑπερτείνοντα libri 17 τἀναντία FE: τοὐναντίον PMV 19 ὁμοιογενῶν EM: ὁμοίων γενῶν F: ἀνομοίων PV || ἀνομοιογενῶν EFM: ὁμογενῶν PV 22 τόδε χρόνου FMV: τὸ λέγειν P 23 πρῶτον P

2. Compare the scholia of Maximus Planudes on the π. ἰδ. of Hermogenes: τοῦτο γάρ φησι καὶ Διονύσιος, ὅτι δεῖ μιγνύειν βραχέσι μακρὰ καὶ πολυσυλλάβοις ὀλιγοσύλλαβα, τοῦτο γὰρ ἡδέως διατίθησι τὴν ἀκοήν (Walz _Rhett. Gr._ v. 520).

12. Cp. Anonymi scholia on Hermog. π. ἰδ. (Walz vii. 1049), διὰ τοῦτο κάλλους ἴδιον ὁ ῥυθμός, εἴτε βέβηκεν εἴτε μή· ἐπειδὴ κατὰ Διονύσιον ἡδύνει τὴν ἀκοὴν καὶ ποικίλλει, καὶ μὴ ἄρχεσθαι ἀπὸ τῶν αὐτῶν, μηδὲ λήγειν εἰς αὐτά, ἀλλὰ τὸ ἐξ ἁπάντων καλῶν ῥυθμῶν, τουτέστι ποδῶν, συγκεῖσθαι τὸν λόγον· ἀνάγκη γὰρ αὐτὸν οὕτω καλὸν εἶναι· τάττει δὲ τὸν σπονδεῖον μετ’ αὐτῶν.

14. =ὑπερτείνοντας ... καιρόν=: lit. ‘exceeding due measure in either case.’ On the whole, Usener is perhaps right in reading the plural here and in l. 11; clearness, and variety of termination, recommend the change. But (1) all MSS. have ὑπερτείνοντα, (2) the singular has been used in ll. 1, 3, 4 _supra_, and so might as well be maintained to the end, while φυλαττομένους (instead of φυλαττόμενον) might arise from the initial σ of σχήμασι.

20. =τὸν καιρὸν ὁρᾶν=, ‘to have an eye to (or observe) the rules of good taste,’ is a natural and appropriate expression. The use of θηρατός in =134= 3 is no argument for reading θηρᾶν here, but rather tells against the anticipation of so pronounced a metaphor. Moreover, the _middle_ voice is found in this sense in _de Demosth._ c. 40 τὴν εὐφωνίαν θηρωμένη καὶ τὴν εὐμέλειαν. With ὁρᾶν cp. _de Demosth._ c. 49 ἄλλως τε καὶ τοῦ καιροῦ τὰ μέτρα ὁρῶν and _de Thucyd._ c. 1 τῆς προαιρέσεως οὐχ ἅπαντα κατὰ τὸν ἀκριβέστατον λογισμὸν ὁρώσης (where θηρώσης is given in Usener-Radermacher’s text).

21. Quintil. xi. 1. 1 “parata, sicut superiore libro continetur, facultate scribendi cogitandique et ex tempore etiam, cum res poscet, orandi, proxima est cura, ut dicamus apte; quam virtutem quartam elocutionis Cicero demonstrat, quaeque est meo quidem iudicio maxime necessaria. nam cum sit ornatus orationis varius et multiplex conveniatque alius alii: nisi fuerit accommodatus rebus atque personis, non modo non illustrabit eam sed etiam destruet et vim rerum in contrarium vertet.”

22. =τόδε χρόνου=: Usener reads τόδε γε (without χρόνου), in view of P’s τὸ λέγειν. But τόδε γε is unusual in this sense, whereas ἔτι καὶ εἰς τόδε χρόνου is found in _Antiqq. Rom._ i. 16. Cp. i. 38 _ibid._ καὶ παρὰ Κελτοῖς εἰς τόδε χρόνου γίνεται: also i. 61, 68, iii. 31, vi. 13.

[Page 133]

mingling and fusion and juxtaposition,—by mingling smooth with rough, soft with hard, cacophonous with melodious, easy to pronounce with hard to pronounce, long with short; and generally by happy combinations of the same kind. Many words of few syllables must not be used in succession (for this jars upon the ear), nor an excessive number of polysyllabic words; and we must avoid the monotony of setting side by side words similarly accented or agreeing in their quantities. We must quickly vary the cases of substantives (since, if continued unduly, they greatly offend the ear); and in order to guard against satiety, we must constantly break up the effect of sameness entailed by placing many nouns, or verbs, or other parts of speech, in close succession. We must not always adhere to the same figures, but change them frequently; we must not re-introduce the same metaphors, but vary them; we must not exceed due measure by beginning or ending with the same words too often.

Still, let no one think that I am proclaiming these as universal rules—that I suppose keeping them will always produce pleasure, or breaking them always produce annoyance. I am not so foolish. I know that pleasure often arises from both sources—from similarity at one time, from dissimilarity at another. In every case we must, I think, keep in view good taste, for this is the best criterion of charm and its opposite. But about good taste no rhetorician or philosopher has, so far, produced a definite treatise. The man who first undertook to write on the subject, Gorgias of Leontini, achieved nothing

[Page 134]

ὁ Λεοντῖνος οὐδὲν ὅ τι καὶ λόγου ἄξιον ἔγραψεν· οὐδ’ ἔχει φύσιν τὸ πρᾶγμα εἰς καθολικὴν καὶ ἔντεχνόν τινα περίληψιν πεσεῖν, οὐδ’ ὅλως ἐπιστήμῃ θηρατός ἐστιν ὁ καιρὸς ἀλλὰ δόξῃ. ταύτην δ’ οἱ μὲν ἐπὶ πολλῶν καὶ πολλάκις γυμνάσαντες ἄμεινον τῶν ἄλλων εὑρίσκουσιν αὐτόν, οἱ δ’ 5 ἀγύμναστον ἀφέντες σπανιώτερον καὶ ὥσπερ ἀπὸ τύχης.

ἵνα δὲ καὶ περὶ τῶν ἄλλων εἴπω, ταῦτ’ οἴομαι χρῆναι φυλάττειν ἐν τῇ συνθέσει τὸν μέλλοντα διαθήσειν τὴν ἀκοὴν ἡδέως· ἢ τὰ εὐμελῆ καὶ εὔρυθμα καὶ εὔφωνα ὀνόματα, ὑφ’ ὧν γλυκαίνεταί τε καὶ ἐκμαλάττεται καὶ τὸ ὅλον οἰκείως 10 διατίθεται ἡ αἴσθησις, ταῦτα ἀλλήλοις συναρμόττειν, ἢ τὰ μὴ τοιαύτην ἔχοντα φύσιν ἐγκαταπλέκειν τε καὶ συνυφαίνειν τοῖς δυναμένοις αὐτὴν γοητεύειν, ὥστε ὑπὸ τῆς ἐκείνων χάριτος ἐπισκοτεῖσθαι τὴν τούτων ἀηδίαν· οἷόν τι ποιοῦσιν οἱ φρόνιμοι στρατηλάται κατὰ τὰς συντάξεις τῶν στρατευμάτων· 15 καὶ γὰρ ἐκεῖνοι ἐπικρύπτουσι τοῖς ἰσχυροῖς τὰ ἀσθενῆ, καὶ γίνεται αὐτοῖς οὐδὲν τῆς δυνάμεως ἄχρηστον. διαναπαύειν δὲ τὴν ταυτότητά φημι δεῖν μεταβολὰς εὐκαίρους εἰσφέροντα· καὶ γὰρ ἡ μεταβολὴ παντὸς ἔργου χρῆμα ἡδύ. τελευταῖον δὲ ὃ δὴ καὶ πάντων κράτιστον, οἰκείαν ἀποδιδόναι τοῖς 20 ὑποκειμένοις καὶ πρέπουσαν ἁρμονίαν. δυσωπεῖσθαι δ’ οὐδὲν οἴομαι δεῖν οὔτε ὄνομα οὔτε ῥῆμα, ὅ τι καὶ τέτριπται, μὴ σὺν αἰσχύνῃ λέγεσθαι μέλλον· οὐδὲν γὰρ οὕτω ταπεινὸν ἢ ῥυπαρὸν ἢ ἄλλην τινὰ δυσχέρειαν ἔχον ἔσεσθαί φημι λόγου μόριον, ᾧ σημαίνεταί τι σῶμα ἢ πρᾶγμα, ὃ μηδεμίαν ἕξει 25 χώραν ἐπιτηδείαν ἐν λόγοις. παρακελεύομαι δὲ τῇ συνθέσει

1 οὐδὲν F: οὐδ’ MV: om. P || καὶ F: om. PMV 5 αὐτόν FM: om. PV 6 ἀγύμναστον F, γρ M: ἀνάσκητον PM^{1}V || σπανιωτέρ(αν) P, MV 9 ἢ EFM: om. PV 10 ἐκμαλάττεται F: μαλάττεται PMV 15 συντάξεις FM: τάξ[ει]ς cum litura P, V 16 ἐπικρύπτουσι EF: συγκρύπτουσιν P, MV 17 ἄχρηστον FE: μέρος ἄχρηστον PMV 20 κράτιστον EF: ἐστὶ κράτιστον PMV 21 καὶ πρέπουσαν om. F 22 δεῖν om. F || ὅτι καὶ τέτριπται EF: ὅτ’ (οὔτ’ V) ἐπιτέτραπται PMV 23 μέλλον EF: om. PMV 24 ῥυπαρὸν EF: ῥυπαρὸν ἢ μιαρὸν PV: μιαρὸν M || ἔχον om. F 26 δὲ EF: δὲ ἐν PMV

1. For οὐδ’ ὅτι (as read by Schaefer) Dobree suggested a number of alternatives,—οἶδ’ (= οἶδα), οὐδὲν, οὐδ’ ὁτιοῦν.

7. The passage that begins here is, itself, a good example of rhythmical and melodious writing.

10. =τὸ ὅλον=: cp. Long. p. 207, s.v. σύνολον.

15. The description in _Iliad_ iv. 297-300 may be in Dionysius’ mind. Cp. Cic. _Brut._ 36. 139 “omnia veniebant Antonio in mentem; eaque suo quaeque loco, ubi plurimum proficere et valere possent, ut ab imperatore equites pedites levis armatura, sic ab illo in maxime opportunis orationis partibus collocabantur”; Xen. _Cyrop._ vii. 5. 5 ἀναπτυχθείσης δ’ οὕτω τῆς φάλαγγος ἀνάγκη τοὺς πρώτους ἀρίστους εἶναι καὶ τοὺς τελευταίους, ἐν μέσῳ δὲ τοὺς κακίστους τετάχθαι.

19. Cp. Dionys. Hal. _Ep. ad Cn. Pompeium_ c. 3 ὡς ἡδὺ χρῆμα ἐν ἱστορίας γραφῇ μεταβολὴ καὶ ποικίλον: Aristot. _Eth._ vii. 1154 b μεταβολὴ δὲ πάντων γλυκύ, κατὰ τὸν ποιητήν: Eurip. _Orest._ 234 μεταβολὴ πάντων γλυκύ. Dionysius’ whole-hearted faith in the virtues of μεταβολή (considered in its widest bearings) rests on a basis of permanent truth. If we open Shakespeare at random, we can see how the verbal forms (‘remember,’ ‘bequeathed,’ ‘sayest,’ ‘charged,’ ‘begins’) are varied in the opening sentence of _As You Like It_; and this though our language is almost wholly analytical. And the words that fall from Lear in his madness (_King Lear_ iv. 6) are full of the most moving μεταβολαί, as well as of the most pathetic variations from τὸ εὐμελὲς to τὸ ἐμμελές.

[Page 135]

worth mentioning. The nature of the subject, indeed, is not such that it can fall under any comprehensive and systematic treatment, nor can good taste in general be apprehended by science, but only by personal judgment. Those who have continually trained this latter faculty in many connexions are more successful than others in attaining good taste, while those who leave it untrained are rarely successful, and only by a sort of lucky stroke.

To proceed. I think the following rules should be observed in composition by a writer who looks to please the ear. Either he should link to one another melodious, rhythmical, euphonious words, by which the sense of hearing is touched with a feeling of sweetness and softness,—those which, to put it broadly, come home to it most; or he should intertwine and interweave those which have no such natural effect with those that can so bewitch the ear that the unattractiveness of the one set is overshadowed by the grace of the other. We may compare the practice of good tacticians when marshalling their armies: they mask the weak portions by means of the strong, and so no part of their force proves useless. In the same way I maintain we ought to relieve monotony by the tasteful introduction of variety, since variety is an element of pleasure in everything we do. And last, and certainly most important of all, the setting which is assigned to the subject matter must be appropriate and becoming to it. And, in my opinion, we ought not to feel shy of using any noun or verb, however hackneyed, unless it carries with it some shameful association; for I venture to assert that no part of speech which signifies a person or a thing will prove so mean, squalid, or otherwise offensive as to have no fitting place in discourse. My advice is that, trusting to the

[Page 136]

πιστεύοντας ἀνδρείως πάνυ καὶ τεθαρρηκότως αὐτὰ ἐκφέρειν Ὁμήρῳ τε παραδείγματι χρωμένους, παρ’ ᾧ καὶ τὰ εὐτελέστατα κεῖται τῶν ὀνομάτων, καὶ Δημοσθένει καὶ Ἡροδότῳ καὶ τοῖς ἄλλοις, ὧν ὀλίγῳ ὕστερον μνησθήσομαι καθ’ ὅ τι ἂν ἁρμόττῃ περὶ ἑκάστου. ταῦτά μοι περὶ τῆς 5 ἡδείας εἰρήσθω συνθέσεως, ὀλίγα μὲν ὑπὲρ πολλῶν θεωρημάτων, ἱκανὰ δὲ ὡς κεφάλαια εἶναι.

XIII

εἶἑν. καλὴ δ’ ἁρμονία πῶς γένοιτ’ ἂν εἴ τις ἔροιτό με καὶ ἐκ ποίων θεωρημάτων, οὐκ ἄλλως πως μὰ Δία φαίην ἂν οὐδ’ ἐξ ἄλλων τινῶν ἢ ἐξ ὧνπερ ἡ ἡδεῖα· τὰ γὰρ αὐτὰ 10 ποιητικὰ ἀμφοῖν, μέλος εὐγενές, ῥυθμὸς ἀξιωματικός, μεταβολὴ μεγαλοπρεπής, τὸ πᾶσι τούτοις παρακολουθοῦν πρέπον. ὥσπερ γὰρ ἡδεῖά τις γίνεται λέξις, οὕτω καὶ γενναία τις ἑτέρα, καὶ ῥυθμὸς ὥσπερ γλαφυρός τις, οὕτω καὶ σεμνός τις ἕτερος, καὶ τὸ μεταβάλλειν ὥσπερ χάριν ἔχει, οὕτω καὶ 15 πίνον· τὸ δὲ δὴ πρέπον εἰ μὴ τοῦ καλοῦ πλεῖστον ἕξει μέρος, σχολῇ γ’ ἂν ἄλλου τινός. ἐξ ἁπάντων δή φημι τούτων ἐπιτηδεύεσθαι δεῖν τὸ καλὸν ἐν ἁρμονίᾳ λέξεως ἐξ ὧνπερ καὶ τὸ ἡδύ. αἰτία δὲ κἀνταῦθα ἥ τε τῶν γραμμάτων φύσις καὶ ἡ τῶν συλλαβῶν δύναμις, ἐξ ὧν πλέκεται τὰ ὀνόματα· 20 ὑπὲρ ὧν καιρὸς ἂν εἴη λέγειν, ὥσπερ ὑπεσχόμην.

XIV

ἀρχαὶ μὲν οὖν εἰσι τῆς ἀνθρωπίνης φωνῆς καὶ ἐνάρθρου

2 χρωμένους EFMV: χρ(ω)μεν(ος) P 4 ὀλίγον F: sed cf. =154= 7 7 εἶναι· εἶἑν sic P, FM: εἶεν V 8 με καὶ F: ἢ PMV 9 μὰ PMV: νὴ F 10 οὐδ’] οὐκ PV || ἡ F: om. PMV 13 οὕτω καὶ PMV: οὕτω F 14 ἑτέρα PMV: ἄρα F || σεμνός τις F: σεμνὸς PMV 15 ἔχει P: ἔχει (ἔχειν V) τινὰ FMV 16 πινόν (θ suprascripto) P: πιθανόν V: τὸ πῖνον M: πόνον F 18 δεῖν] δὴ F 20 ὀνόματα PE: ὀνόματα ταῦτα FMV 22 φωνῆς καὶ ἐνάρθρου REF: καὶ ἐνάρθρου φωνῆς αἱ PMVs

6. =ὑπέρ= = περί: l. 21 _infra_, =96= 2, etc. Reiske’s ἀπό is attractive; but does ὀλίγα really = ὀλίγα θεωρήματα?

8. =εἶἑν= = “So!” The breathing on the last syllable (as given by the best manuscripts, here and in other authors) helps to distinguish this word from the third pers. plur. optat. of εἰμί.

9. In a negative sentence, =μὰ Δία= is to be preferred to νὴ Δία.

13. =λέξις=: μέλος (cp. l. 11 _supra_) is here in question. Hence Usener suggests μέλισις. Perhaps λέξις (‘the words,’ ‘the libretto’) is here felt to include the music,—‘a passage set to music’: cp. =124= 22 καὶ γὰρ ἐν ταύτῃ καὶ μέλος ἔχουσιν αἱ λέξεις (‘the words’) καὶ ῥυθμὸν καὶ μεταβολὴν καὶ πρέπον, and contrast =126= 20-1.

16. =πίνον=, ‘mellowness,’ ‘ripeness’ (see Gloss.). The readings of FPMV seem all to point in this direction. πόνον (F’s reading) might possibly mean either ‘involve trouble’ (to the author) or ‘suggest painstaking’ (to the reader). Usener conjectures τόνον.

22. Chapter xiv., which in some respects is the most interesting in the treatise, might easily be ridiculed by one of those scoffers whom Dionysius elsewhere (=252= 17) mentions with aversion. In _Le Bourgeois Gentilhomme_ (ii. 4) there is much that could serve for a parody of the _C. V._—the Maître de Philosophie with his “Sans la science, la vie est presque une image de la mort” (_nam sine doctrina vita est quasi mortis imago_), his “tout ce qui n’est point prose est vers; et tout ce qui n’est point vers est prose,” and (particularly) his remarks on _l’orthographie_: “Pour bien suivre votre pensée et traiter cette matière en philosophe, il faut commencer selon l’ordre des choses, par une exacte connaissance de la nature des lettres, et de la différente manière de les prononcer toutes. Et là-dessus j’ai à vous dire que les lettres sont divisées en voyelles, ainsi dites voyelles parce qu’elles expriment les voix; et en consonnes, ainsi appelées consonnes parce qu’elles sonnent avec les voyelles, et ne font que marquer les diverses articulations des voix.” These remarks include descriptions (many of which are taken almost verbatim from De Cordemoy’s _Discours physique de la parole_, published in 1668) of the mode in which various letters are formed, and (incidentally) M. Jourdain’s exclamation, “A, E, I, I, I, I. Cela est vrai. Vive la science!”

[Page 137]

effect of the composition, we should bring out such expressions with a bold and manly confidence, following the example of Homer, in whom the most commonplace words are found, and of Demosthenes and Herodotus and others, whom I will mention a little later so far as is suitable in each case. I think I have now spoken at sufficient length on charm of style. My treatment has been but a brief survey of a wide field, but will furnish the main heads of the study.