APPENDIX A

OBSCURITY IN GREEK

The natural lucidity of the Greek language is sometimes assumed by its modern admirers to extend to all the writings of Greek authors. But the ancients themselves made no such extravagant claims. They might praise Lysias as a model of clearness; but they knew well the difficulties, of subject matter or expression, to be met with not only in Heracleitus[199] or Lycophron, but in masters so great as Pindar, Aeschylus, Thucydides, and the author of that excellent definition which sees in lucidity a fundamental virtue of style—Aristotle himself. Thucydides (to take one writer only out of this group of four) is taxed with obscurity by critics other than Dionysius. Marcellinus, although not otherwise in entire agreement with Dionysius, attributes this particular defect to Thucydides and regards it as deliberate: ἀσαφῶς δὲ λέγων ἐπίτηδες, ἵνα μὴ πᾶσιν εἴη βατὸς μηδὲ εὐτελὴς φαίνηται παντὶ τῷ βουλομένῳ νοούμενος εὐχερῶς, ἀλλὰ τοῖς λίαν σοφοῖς δοκιμαζόμενος παρὰ τούτοις θαυμάζηται ... τὸ δὲ τῆς συνθέσεως τραχύτητος μεστὸν καὶ ἐμβριθὲς καὶ ὑπερβατικόν, ἐνίοτε δὲ ἀσαφές ... ἀσαφὴς τὴν διάνοιαν διὰ τὸ ὑπερβατοῖς χαίρειν (Marcell. Vita Thucyd. §§ 35, 50, 56). An epigram in the Greek Anthology is pitched in the same key:—

ὦ φίλος, εἰ σοφὸς εἶ, λάβε μ’ ἐς χέρας· εἰ δέ γε πάμπαν
νῆϊς ἔφυς Μουσέων, ῥίψον ἃ μὴ νοέεις.
εἰμὶ δέ γ’ οὐ πάντεσσι βατός· παῦροι δ’ ἀγάσαντο
Θουκυδίδην Ὀλόρου, Κεκροπίδην τὸ γένος.

Anth. Pal. ix. 583.

And Cicero, in a more uncompromising way, condemns the Speeches as scarcely intelligible: “ipsae illae contiones ita multas habent obscuras abditasque sententias, vix ut intellegantur; quod est in oratione civili vitium vel maximum” (Cic. Orat. 9. 30).

Obscurity in matter and obscurity in expression are intimately allied. Euripides, in the Frogs, says of Aeschylus that he was obscure in setting forth his plots (ἀσαφὴς γὰρ ἦν ἐν τῇ φράσει τῶν πραγμάτων, Aristoph. Ran. 1122). Dionysius attributes to Lysias, as compared with Thucydides and Demosthenes, a lucidity which embraces matter as well as expression and treats words as the servants of thought: τρίτην ἀρετὴν ἀποφαίνομαι περὶ τὸν ἄνδρα τὴν σαφήνειαν, οὐ μόνον τὴν ἐν τοῖς ὀνόμασιν, ἀλλὰ καὶ τὴν ἐν τοῖς πράγμασιν· ἔστι γάρ τις καὶ πραγματικὴ σαφήνεια οὐ πολλοῖς γνώριμος. τεκμαίρομαι δέ, ὅτι τῆς μὲν Θουκυδίδου λέξεως καὶ Δημοσθένους, οἳ δεινότατοι πράγματα ἐξειπεῖν ἐγένοντο, πολλὰ δυσείκαστά ἐστιν ἡμῖν καὶ ἀσαφῆ καὶ δεόμενα ἐξηγητῶν ... τούτου δὲ αἴτιον, ὅτι οὐ τοῖς ὀνόμασι δουλεύει τὰ πράγματα παρ’ αὐτῷ [sc. Λυσίᾳ], τοῖς δὲ πράγμασιν ἀκολουθεῖ τὰ ὀνόματα (de Lysia, c. 4). So far as the two can be separated, it is with wording rather than with subject matter that the present appendix is concerned.

One principal cause of obscurity is the anxious search for brevity. Dionysius sees this, especially in regard to Thucydides; and “brevis esse laboro, | obscurus fio” has many an analogue in his critical pages (e.g. ἀσαφὲς γίνεται τὸ βραχύ and διὰ τὸ τάχος τῆς ἀπαγγελίας ἀσαφὴς ἡ λέξις γίνεται, de Thucyd. c. 24 and Ep. ii. ad Amm. c. 2). At the same time, he does not seem to concede enough to the claims of brevity in C.V. 118 1, 2, where it is not simply a question of ‘offending the ear,’ or of ‘spoiling the metre,’ or even of ‘charm.’ The two lines there quoted from Sophocles have something of that πολύνους βραχυλογία which has been justly attributed to Thucydides.[200]

But too many words may be just as fatal to clearness as too few. As Aristotle says (Rhet. iii. 12. 6), lucidity is imperilled when a style is prolix, no less than when it is condensed. A disjointed and rambling diffuseness is condemned by Demetrius (de Eloc. § 192); and Dionysius (Ep. ii. ad Amm. c. 15) remarks that numerous parentheses make the meaning hard to follow (... αἱ μεταξὺ παρεμπτώσεις πολλαὶ γινόμεναι καὶ μόλις ἐπὶ τὸ τέλος ἀφικνούμεναι, δι’ ἃς ἡ φράσις δυσπαρακολούθητος γίνεται).[201]

It is, however, the arrangement of words (even more than their number, large or small) that contributes to lucidity or its opposite. Quintilian (ix. 4. 32) says “amphiboliam quoque fieri vitiosa locatione verborum, nemo est qui nesciat”; and certainly the importance of a right order, in its bearing on clearness, is very great even in the highly inflected languages. Elsewhere (viii. 2. 16) Quintilian gives some good examples of ambiguities to be avoided: “vitanda est in primis ambiguitas, non haec solum, de cuius genere supra dictum est, quae incertum intellectum facit, ut Chremetem audivi percussisse Demean,[202] sed illa quoque, quae, etiamsi turbare non potest sensum, in idem tamen verborum vitium incidit, ut si quis dicat, visum a se hominem librum scribentem. nam etiamsi librum ab homine scribi patet, male tamen composuerit feceritque ambiguum, quantum in ipso fuit.” Quintilian’s ideal is a fine one, but it is not always possible to attain it in Latin or in Greek. The freedom of the classical word-order, so desirable on other grounds, stands in the way here.

Illustrations of a certain degree of ambiguity will be found in some instances of the dependent genitive in Greek, as used especially in Thucydides. Thucydides usually places the dependent genitive before the noun on which it depends.[203] As, however, his rule is not invariable, it cannot be said that in all the following examples (which are designedly of a promiscuous character) the reader is absolved, as Quintilian evidently thinks he should be, from making his conception of the general sense help in determining the grammatical construction:—

(1) καὶ μετὰ τῆς ἥσσονος ἅμα ἐλπίδος ὀλίγων ἡμερῶν ἕνεκα μεγάλου μισθοῦ δόσεως ἐκείνοις ξυναγωνίζεσθαι, Thucyd. i. 143.

(2) εἴ τις ὑπομένοι καὶ μὴ φόβῳ ῥοθίου καὶ νεῶν δεινότητος κατάπλου ὑποχωροίη, iv. 10.

(3) Κερκυραῖοι δὲ μετὰ τῆς ξυμμαχίας τῆς αἰτήσεως καὶ ταῦτα πιστεύοντες ἐχυρὰ ὑμῖν παρέξεσθαι ἀπέστειλαν ἡμᾶς, i. 32.

(4) οἵπερ τῶν ὁλκάδων ἕνεκα τῆς ἐς Σικελίαν κομιδῆς ἀνθώρμουν πρὸς τὰς ἐν Ναυπάκτῳ ναῦς, vii. 34.

(5) ἄπιστα μὲν ἴσως, ὥσπερ καὶ ἄλλοι τινές, δόξω ὑμῖν περὶ τοῦ ἐπίπλου τῆς ἀληθείας λέγειν, vi. 33.

(6) τά τε τῆς ἀντιμιμήσεως αὐτῶν τῆς παρασκευῆς ἡμῶν τῷ μὲν ἡμετέρῳ τρόπῳ ξυνήθη τέ ἐστι κτλ., vii. 67.

(7) τοὺς γὰρ ἂν ψιλοὺς τοὺς σφῶν καὶ τὸν ὄχλον τῶν Συρακοσίων τοὺς ἱππέας πολλοὺς ὄντας, σφίσι δ’ οὐ παρόντων ἱππέων, βλάπτειν ἂν μεγάλα, vi. 64.

(8) καὶ τοῦ Κλέωνος καίπερ μανιώδης οὖσα ἡ ὑπόσχεσις ἀπέβη, iv. 39.

(9) καὶ τριήρης τῇ αὐτῇ ἡμέρᾳ ἁλίσκεται τῶν Ἀθηναίων ὑπὸ τῶν Συρακοσίων ἐφορμοῦσα τῷ λιμένι, vii. 3.[204]

Similarly in other authors: e.g. καὶ δὴ καὶ τότε τοῦ Θρασυμάχου τὴν ἀπόρρησιν οὐκ ἀπεδέξατο, Plato Rep. ii. 357 A (where, however, the meaning may be “would not accept from Thrasymachus his withdrawal”); and ὣς φάτο, τῷ δ’ ἄρα πατρὸς ὑφ’ ἵμερον ὦρσε γόοιο, Hom. Il. xxiv. 507; and

τούτων ἐγὼ οὐκ ἔμελλον, ἀνδρὸς οὐδενὸς
φρόνημα δείσασ’, ἐν θεοῖσι τὴν δίκην
δώσειν.

Soph. Antig. 458-60.[205]

If in some of these instances the order is not absolutely unambiguous, still less is it so in other and more miscellaneous extracts about to be given. The writer of artistic prose, as of poetry, has to satisfy claims which are often hard to reconcile: those of clearness, of emphasis, and of euphony.[206] The result may often be a more or less unconscious compromise in which one of the elements prospers at the expense of the others. Euphony, to take that element alone, is expected to please the ear in many different ways—by the avoidance of harsh letters (found singly or in combination), of short syllables in close succession, of monotony in word-terminations, of monotony in every shape and form. Obscurity may well ensue, especially in a literature which does not aid the eye by means of punctuation, capital letters (to denote proper names or the beginning of a sentence), italic type, or division into paragraphs and chapters. To set against these deficiencies, there was the help provided by the reciter or the skilled anagnostes; and it is often interesting to speculate how, by a slight pause or modulation of the voice, a practised reader would be able to remove a seeming ambiguity. In poetry, again, metre would often be an aid to clear delivery, though its exigencies might on the other hand have led to some ambiguities in the actual writing. No careful modern student of a highly-wrought speech, like the Crown of Demosthenes, can have failed to be arrested momentarily, here and there, by some slight ambiguity which, as far as he can judge, might have been removed by an equally slight change in the word-order; and he gains much in the appreciation of Demosthenes if he is thus led to consider what are the subtle laws of rhythm and melody to which an absolutely unimpeachable lucidity has (in however small a degree) given way. He will certainly be led to the conclusion that, in Greek, good order is by no means the simple thing it may seem when achieved, but rather is the highly complex result of the play of many forces. The following examples, drawn from various authors in poetry and in prose, may be found suggestive. They are of set purpose presented without any attempt at sequence or classification, except that a considerable number of extracts from the de Corona are grouped together:—

(1) καί μοι τὸν υἱόν, εἰ μεμάθηκε τὸν λόγον
ἐκεῖνον, εἴφ’, ὃν ἀρτίως εἰσήγαγες.

Aristoph. Nub. 1148.

(2) ἀλλά μιν αὖτις ἀναρπάξασα θύελλα
πόντον ἐπ’ ἰχθυόεντα φέρεν βαρέα στενάχοντα.

Hom. Odyss. xxiii. 316.[207]

(3) ἠδ’ ὡς εἰς Ἀίδεω δόμον ἤλυθεν εὐρώεντα,
ψυχῇ χρησόμενος Θηβαίου Τειρεσίαο,
νηῒ πολυκλήιδι.

id. ib. xxiii. 322.[207]

(4) ὅτι Ἱππίας μὲν πρεσβύτατος ὢν ἦρχε τῶν Πεισιστράτου υἱέων.

Thucyd. i. 20.

Here τῶν Πεισιστράτου υἱέων depends on πρεσβύτατος ὤν, not on ἦρχε.

(5) κράτιστα τοίνυν τῶν παρόντων ἐστὶ νῷν
θεῶν ἰόντε προσπεσεῖν του πρὸς βρέτας.

Aristoph. Eq. 30, 31.

Here the actor would pause slightly after νῷν, at the end of the metrical line.

(6) τοῦτ’ οὖν ἔβλαψα τί δράσας;

id. Ran. 1064.

Careful delivery would make it quite plain that the meaning is: τί οὖν ἔβλαψα, δράσας τοῦτο;

(7) σαφῶς γὰρ ἄν, εἰ πείθοιμι ὑμᾶς καὶ τῷ δεῖσθαι βιαζοίμην ὀμωμοκότας, θεοὺς ἂν διδάσκοιμι μὴ ἡγεῖσθαι ὑμᾶς εἶναι.

Plato Apol. c. 24.

(8) καὶ ἐς τύχας πρὸς πολλῷ δυνατωτέρους ἀγωνιζόμενοι καταστῆναι.

Thucyd. i. 69.

(9) οὐδ’ ἐκλογίσασθαι πώποτε πρὸς οἵους ὑμῖν Ἀθηναίους ὄντας καὶ ὅσον ὑμῶν καὶ ὡς πᾶν διαφέροντας ὁ ἀγὼν ἔσται.

id. i. 70.

ὑμῖν is probably to be connected with ὁ ἀγὼν ἔσται. Its present position has the effect of marking the contrast between ὑμῖν and Ἀθηναίους, and further of breaking the monotony of the accusative-endings οἵους Ἀθηναίους ὄντας. It should, however, be remembered that in a highly inflected language like Greek a noun may stand in a vague general case relation (genitive, dative, or accusative) to the whole sentence in a way that is impossible in an uninflected language. This may be so here, and in some of the other passages quoted.

(10) ῥηθήσεται δὲ οὐ παραιτήσεως μᾶλλον ἕνεκα ἢ μαρτυρίου καὶ δηλώσεως πρὸς οἵαν ὑμῖν πόλιν μὴ εὖ βουλευομένοις ὁ ἀγὼν καταστήσεται.

id. i. 73.

Similarly ὑμῖν (‘you will find,’ etc.) is to be taken with ὁ ἀγὼν καταστήσεται. It is contrasted with πόλιν and paves the way for βουλευομένοις.

(11) ἔνθ’ ὅ γε τοὺς ἐλεεινὰ κατήσθεε τετριγῶτας·
μήτηρ δ’ ἀμφεποτᾶτο ὀδυρομένη φίλα τέκνα.

Hom. Il. ii. 314-15.

Connect ἐλεεινὰ τετριγῶτας, and ἀμφεποτᾶτο φίλα τέκνα.

(12) ὡς οὖν δεινὰ πέλωρα θεῶν εἰσῆλθ’ ἑκατόμβας.

id. ib. ii. 321.

Connect θεῶν ἑκατόμβας.

(13) καίτοι σ’ ἐγὼ ’τίμησα τοῖς φρονοῦσιν εὖ.

Soph. Antig. 904.

εὖ with ἐτίμησα. The line occurs in the suspected portion of the Antigone. But, so far as this particular point is concerned, cp. the order of μόνος in—

τὰ κοινὰ χαίρων οὐ δίκαια δρᾷ μόνος.

Eurip. Ion 358.

(14) τίνος δ’ Ἀτρεῖδαι τοῦδ’ ἄγαν οὕτω χρόνῳ
τοσῷδ’ ἐπεστέφοντο πράγματος χάριν,
ὅν γ’ εἶχον ἤδη χρόνιον ἐκβεβληκότες;

Soph. Philoct. 598.

Here strict lucidity is sacrificed to emphasis. τίνος must be joined with πράγματος (not with τοῦδε).

(15) στέμματ’ ἔχων ἐν χερσὶν ἑκηβόλου Ἀπόλλωνος
χρυσέῳ ἀνὰ σκήπτρῳ.

Hom. Il. i. 14.

(16) περὶ τούτων δ’ ὄντος τουτουὶ τοῦ ἀγῶνος, ἀξιῶ καὶ δέομαι πάντων ὁμοίως ὑμῶν ἀκοῦσαί μου περὶ τῶν κατηγορημένων ἀπολογουμένου δικαίως, ὥσπερ οἱ νόμοι κελεύουσιν, οὓς ὁ τιθεὶς ἐξ ἀρχῆς Σόλων κτλ.

Demosth. de Cor. § 6.

δικαίως qualifies ἀκοῦσαι: cp. the position of γενναίως in de Cor. § 97 (quoted in Introduction p. 24 supra). The present order is not only emphatic, but also serves to connect δικαίως closely with ὥσπερ κτλ., and thus to a certain extent actually to avoid ambiguity.

(17) σκέψασθ’ ὦ ἄνδρες Ἀθηναῖοι καὶ θεωρήσατε ὅσῳ καὶ ἀληθέστερον καὶ ἀνθρωπινώτερον ἐγὼ περὶ τῆς τύχης τούτου διαλεχθήσομαι.

Demosth. de Cor. § 252.

(18) τὸ μὲν τοίνυν προελέσθαι τὰ κάλλιστα καὶ τὸ τῶν οἰηθέντων Ἑλλήνων, εἰ πρόοιντο ἡμᾶς, ἐν εὐδαιμονίᾳ διάξειν, αὐτῶν ἄμεινον πράττειν τῆς ἀγαθῆς τύχης τῆς πόλεως εἶναι τίθημι.

id. ib. § 254.

(19) τοῦ μὲν οὖν γράψαι πράττοντα καὶ λέγοντα τὰ βέλτιστά με τῷ δήμῳ διατελεῖν καὶ πρόθυμον εἶναι ποιεῖν ὅ τι ἂν δύνωμαι ἀγαθόν, καὶ ἐπαινεῖν ἐπὶ τούτοις, ἐν τοῖς πεπολιτευμένοις τὴν κρίσιν εἶναι νομίζω.

id. ib. § 56.

(20) οὐ γὰρ ἂν ἥψατ’ αὐτῶν | παρόντων ἡμῶν, κτλ.

id. ib. § 30.

The vertical stroke, here and elsewhere, may serve to indicate the possibility of a slight pause in utterance, and Aristotle’s remarks on the obscurity of Heracleitus may be recalled: τὰ γὰρ Ἡρακλείτου διαστίξαι (‘to punctuate’) ἔργον διὰ τὸ ἄδηλον εἶναι ποτέρῳ πρόσκειται, τῷ ὕστερον ἢ τῷ πρότερον, οἷον ἐν τῇ ἀρχῇ αὐτοῦ τοῦ συγγράμματος· φησὶ γὰρ “τοῦ λόγου τοῦδ’ ἐόντος ἀεὶ ἀξύνετοι ἄνθρωποι γίγνονται”· ἄδηλον γὰρ τὸ ἀεί, πρὸς ὁποτέρῳ ‹δεῖ› διαστίξαι.

Aristot. Rhet. iii. 5.

(21) λοιπὸν τοίνυν ἦν καὶ ἀναγκαῖον ἅμα | πᾶσιν οἷς ἐκεῖνος ἔπραττ’ ἀδικῶν ὑμᾶς ἐναντιοῦσθαι δικαίως.

Demosth. de Cor. § 69.

(22) ταῦτα τοίνυν εἰδὼς Αἰσχίνης οὐδὲν ἧττον ἐμοῦ | πομπεύειν ἀντὶ τοῦ κατηγορεῖν εἵλετο.

id. ib. § 124.

(23) συνέβαινε δ’ αὐτῷ | τῷ πολέμῳ κρατοῦντι, κτλ.

id. ib. § 146.

(24) τότε τοίνυν κατ’ ἐκεῖνον τὸν καιρὸν ὁ Παιανεὺς ἐγὼ Βάτταλος Οἰνομάου τοῦ Κοθωκίδου σοῦ | πλείονος ἄξιος ὢν ἐφάνην τῇ πατρίδι.

id. ib. § 180.

(25) εἰ γὰρ ὡς οὐ τὰ βέλτιστα ἐμοῦ πολιτευσαμένου | τουδὶ καταψηφιεῖσθε, ἡμαρτηκέναι δόξετε, οὐ τῇ τῆς τύχης ἀγνωμοσύνῃ τὰ συμβάντα παθεῖν.

id. ib. § 207.

(26) οὐκ ἂν οἷα σὺ νῦν ἔλεγες, τοιαῦτα κατηγόρει, παραδείγματα πλάττων | καὶ ῥήματα καὶ σχήματα μιμούμενος κτλ.

id. ib. § 232.

(27) σὺ τοίνυν ταῦτ’ ἀφεὶς ἐμὲ τὸν παρὰ τουτοισὶ πεπολιτευμένον αἰτιᾷ, καὶ ταῦτ’ | εἰδὼς ὅτι, καὶ εἰ μὴ τὸ ὅλον, μέρος γ’ ἐπιβάλλει τῆς βλασφημίας ἅπασι, καὶ μάλιστα σοί.

id. ib. § 272.

Here may be added, from R. Y. Tyrrell’s edition of Eurip. Bacchae p. 36, an interesting note suggested by the distance which parts μόσχων from ἀγελαῖα βοσκήματα in Bacch. 678: “The Greek writers are not nearly so sensitive about the order of words as we are. Surely we have something at least as strange in the order of words in 684 where ἐλάτης certainly depends on φόβην not on νῶτα. See Comm. on 860 for more curious inversions of the natural order; and compare in Soph. Oed. R. 1251 χὤπως μὲν ἐκ τῶνδ’ οὐκέτ’ οἶδ’ ἀπόλλυται; O.C. 1427 τίς δὲ τολμήσει κλύων | τὰ τοῦδ’ ἕπεσθαι τἀνδρός; Perhaps the best instance in Greek of a violent hyperbaton is Ar. Thesm. 811 οὐδ’ ἂν κλέψασα γυνὴ ζεύγει κατὰ πεντήκοντα τάλαντα | ἐς πόλιν ἔλθοι τῶν δημοσίων ‘nor would a lady ride in her chariot to the town after pilfering the public exchequer to the tune of 50 talents.’” Probably the Greek authors, in such instances, were not blind to the liberties they were taking with the natural and lucid order of words; but they trusted to delivery’s artful aid. And about the order adopted in the passage quoted from the Thesmophoriazusae there seems to be a touch of intentional comedy.

It is worth notice, in connexion with Thucydides and word-order, that the Vatican manuscript B, which is at its best from vi. 92 to the end of viii., frequently exhibits an order of words which is peculiar to it and may point to a reviser’s deliberate effort after greater lucidity. In reference to the text presented by the newly discovered Commentary on Thucydides ii., Grenfell and Hunt (Oxyrhynchus Papyri vi. p. 113) say: “As usual, the text of the papyrus is of an eclectic character and does not consistently agree with either family [of the MSS. of Thucydides]; but it supports the ABEFM group seven times against only four agreements with the other [viz. CG]. Several new readings occur of which we append a list.”

With regard to the 27 passages quoted above from various authors it may be remarked in general that, while in some of them there are real obscurities, in others the ambiguity is purely grammatical. And it might almost be laid down as a principle of Greek language that grammatical rules may be freely neglected where the neglect of them does not make the meaning seriously ambiguous, and is desirable in order to secure emphasis, euphony, or some similar object.