The Epic Rhapsode and His Craft: Homeric Performance in a Diachronic Perspective

  González, José M. 2013. The Epic Rhapsode and His Craft: Homeric Performance in a Diachronic Perspective. Hellenic Studies Series 47. Washington, DC: Center for Hellenic Studies.

14. The Aristotelian tekhnē of hypokrisis

14.1 Technical hypokrisis

We come at last to Rhetoric 1404a12–19, the section where Aristotle makes clear what approach to λέξις and ὑπόκρισις he intends to take in the rest of the treatise:

ἐκείνη μὲν οὖν ὅταν ἔλθῃ ταὐτὸ ποιήσει τῇ ὑποκριτικῇ, ἐγκεχειρήκασι δὲ ἐπ’ ὀλίγον περὶ αὐτῆς εἰπεῖν τινες, οἷον Θρασύμαχος ἐν τοῖς Ἐλέοις· καὶ ἔστι φύσεως τὸ ὑποκριτικὸν εἶναι, καὶ ἀτεχνότερον, περὶ δὲ τὴν λέξιν ἔντεχνον. διὸ καὶ τοῖς τοῦτο δυναμένοις γίγνεται πάλιν ἆθλα, καθάπερ καὶ τοῖς κατὰ τὴν ὑπόκρισιν ῥήτορσιν· οἱ γὰρ γραφόμενοι λόγοι μεῖζον ἰσχύουσι διὰ τὴν λέξιν ἢ διὰ τὴν διάνοιαν.

Now, when it [sc. lexis] comes [into vogue] it will have the same impact as the art of [dramatic] acting, and to a small extent some have endeavored to speak about it, for example Thrasymakhos in his Pities; and to be good at delivery is a natural gift, and [as such] less technical, but [when brought] in connection with lexis [it is] technical. [1] Therefore even those who have this [technical] ability in their turn win prizes, just as also [prizes go to] professional speakers [2] who are such owing to [their natural talent for] delivery; for written speeches are more effective on account of their lexis than on account of their argument. [3]

Fortenbaugh (1985:276) has noted that the Greek of a15–16 (καὶ ἔστι φύσεως … ἔντεχνον) is “irritatingly compressed.” The first difficulty lies in identifying the referent of ἐκείνη: the two obvious choices are λέξις and ὑπόκρισις. [4] Most commentators assume without much discussion that the latter, i.e. ‘delivery’, is exclusively in view. Apparently, the reasons are the parallel with ὑποκριτική and the presumption that, if Thrasymakhos has dealt with ‘it’ in his Eleoi, we should look for a subject that concerns itself with πάθη; and given that these same scholars are commonly of the persuasion that only ὑπόκρισις does so (to the exclusion of λέξις), it follows eo ipso that not style but delivery is meant. I trust I have already succeeded in proving false the impression that Aristotle does not concern himself with ὑπόκρισις in the ensuing chapters of Rhetoric Ⅲ, and that, therefore, the view that λέξις cannot encompass the effective expression of ēthos and pathos is a prejudice contradicted by the evidence. Such considerations, then, do not suffice to remove the ambiguity.Pointing to ἔλθῃ one might advance that, in light of τὰ περὶ τὴν ὑπόκρισιν … ὀψὲ παρῆλθεν of 1403b22–23, ἐκείνη must be ὑπόκρισις. The problem with this view is that the last time we have an explicit reference to this word is at 1403b27, and this, by the demonstrative αὕτη, [5] twenty-one lines before our ἐκείνη; and though ‘delivery’ (in its narrow sense) is still the focus of the following eight lines (b28–35), it is through its concern with φωνή and its properties (hence the περὶ αὐτῶν at 1403b35–36). Perhaps more significantly, at 1403b36 the focus is broadened to τὸ περὶ τὴν λέξιν, which henceforth remains in view as is shown by the syntactic agreement of φορτικόν and ὑπολαμβανόμενον at 1404a1 and the repetition τὸ μὲν οὖν τῆς λέξεως at 1404a8. [6] And λέξις, too, is said to have come up lately (ὀψὲ προῆλθεν 1403b36), which might just as well have triggered the ὅταν ἔλθῃ.

The parallel with acting (ἡ ὑποκριτική) at 1404a13 is more promising, for (explicit attention to) ‘delivery’ is said at 1403b22–23 to have come to τραγική (i.e. to tragic performance) [7] when poets stopped acting out (ὑπεκρίνοντο) their own plays. Thus, one could view ὑποκριτική as tragic ὑπόκρισις. If ἐκείνη referred to delivery, Aristotle would be saying that ‘when delivery comes [to rhetoric], it will have the same effect [there] as acting [did in tragic drama]’; and since ‘acting’ would owe its origin to the application of delivery to drama, we could restate the thought thus: ‘when delivery comes to rhetoric, it will have the same effect there as when delivery came to tragic drama’. [8] This comment is reasonable and apposite and has much to recommend it. But considering the capital importance Aristotle assigns to delivery and its inclusion under style, if, as I argue, the philosopher fully intends to study style on the whole from the perspective of delivery, it would be just as acceptable to refer ἐκείνη to λέξις. [9] Indeed, with the coming of rhetorical [10] λέξις would also come ὑπόκρισις. And surely, for any who hold Aristotle’s perspective, the paramount significance of style coming to rhetoric would lie in the coming of delivery to rhetoric. When one views style and delivery in the intimate connection in which the philosopher places them, it makes no ultimate difference whether ἐκείνη refers to the one or to the other. Only those who have driven a wedge between them make much of the alternative selected. Strictly in terms of meaning and implications, however, I find either equally acceptable, and so perhaps does Aristotle, who apparently did not realize the profound ambiguity he would inflict on future readers by choosing the demonstrative over its referent. Such oversight would be entirely excusable (and a happy one) if in fact λέξις and ὑπόκρισις occupied roughly the same conceptual space in his mind. Nevertheless, having said this much, on contextual grounds I still prefer λέξις as the referent: for from 1403b36 on, Aristotle has addressed himself to style as the more inclusive heading, yet all the while admittedly keeping delivery in mind as its preeminent subdivision. And even though he mentions τὸ ὑποκριτικόν at 1404a15, his interest is in the ‘ability to perform’ insofar as it bears a connection to style, i.e. περὶ τὴν λέξιν (1404a16)—so that the controlling thought is not ὑπόκρισις in its ‘natural state,’ so to say, but in its ‘artistic’ (or ‘artificial’) mode. Therefore, the subject matter does not change at 1404a12 from λέξις to ὑπόκρισις, but remains uniformly the same from 1403b36 on, namely, ‘rhetorical style’. [11]

But why do I speak of rhetorical style? I mean to reflect the all important distinction Aristotle draws at 1404a15–16: whereas τὸ ὑποκριτικόν, ‘an ability to perform’, to the extent that it is a natural endowment (φύσεως) is also ἀτεχνότερον, ‘less within the sphere of art (i.e. of systematic study)’, yet when considered in its connection to style (περὶ δὲ τὴν λέξιν), it is ἔντεχνον, ‘susceptible of artistic (i.e. systematic) treatment’ (or, as Cope 1877:3.8 writes, “when applied to language (declamation) it (the practice of it) may be reduced to an art”). [12] It is impossible to escape the conclusion that Aristotle differentiates between a natural ὑπόκρισις, strictly a matter of talent, that, undisciplined by methodical speech, remains beyond analytical purview, and a ἡ περὶ τὴν λέξιν ὑπόκρισις, where actio is bound with manner of speech (and manner of speech reciprocally with actio), and which on that account is of professional interest and reducible to the systematic treatment of a τέχνη. [13] Where speech is not involved (e.g. pantomime) or else is not of professional concern, the talent itself might safely do without the discipline of training. ἡ περὶ τὴν λέξιν ὑπόκρισις might as well be called, as Cope 1877:8 ad loc. does, “ὑποκριτικὴ [I prefer ὑπόκρισις] κατὰ λέξιν,” and this is the subject matter of chapters 2–12 of the third book of the Rhetoric. Here, at last, it becomes clear precisely how Aristotle relates delivery—that most powerful oratorical device—to the larger field of style; [14] or, reciprocally, the philosopher spells out that a study of style that will serve the ends of the orator must be conducted with a view to delivery. [15] Also clear is what motivates the distinction between artful and artless ὑπόκρισις: after writing that ‘when style-with-delivery comes, it will have the same effect as acting’, the philosopher adds that ‘some have endeavored in a limited way to talk about it’ (my emphasis). Why has the scientific study of this paramount rhetorical matter been prosecuted so inadequately? Why has no one yet succeeded in drawing a full and comprehensive account of it? Aristotle replies: ‘Because the ability to perform can well be a natural talent and, accordingly, less technical’. Ultimately, however, this fact cannot serve to excuse inaction, whatever the history of past failure or the theoretical challenge ahead, [16] and consequently Aristotle sets out to fill the gap and perfect the art of rhetoric with a study of ‘delivery in its connection with style’.

This approach contrasts significantly with the one followed in the Poetics, where delivery, while explicitly owned as part of style, is nevertheless declared to be of interest only to actors (ὑποκριταί) and in consequence rejected as extrinsic to ποιητική. Indeed, at 1456b8–10 we learn that τὰ σχήματα τῆς λέξεως, [21] the ‘forms of utterance’ (Halliwell 1999:97 ad loc.), constitute one type of inquiry into matters connected with λέξις, a study which is for ἡ ὑποκριτική to know as well as for any who have such mastery (ὁ τὴν τοιαύτην ἔχων ἀρχιτεκτονικήν). [22] The ensuing list [23] makes clear that Aristotle has performative settings in view where a skillful employment of the voice (e.g. loudness and intonation) is essential for successful communication. Neither knowledge nor ignorance of these earns ἡ ποιητική any serious blame (ἐπιτίμημα … ὅ τι καὶ ἄξιον σπουδῆς 1456b14–15). Protagoras’ famous criticism of the opening of the Iliad supplies a ready illustration: for why should Homer’s μῆνιν ἄειδε θεά be criticized for allegedly confusing a prayer with a command? The point is that the same form of words can be made into a prayer or a command depending on the tone of voice employed. [24] Therefore this study (θεώρημα) must belong not to poetry but to some other art, the philosopher adds. He is not thereby denying pronuntiatio its rightful place—surely he would have made it the part of any good rhapsode to utter the words with the solemnity and propriety that befits the invocation of a deity—but since the deficiency (if there is one) must be traced to the performer, not the poet, it is therefore possible to dissociate, as he does, the study of ποιητική from considerations of delivery. The degree to which Aristotle, for the purposes of his Poetics, has distanced the poetic texts of epic and tragic drama from their elocutionary performative aspects is nevertheless surprising. [25] In bringing up delivery only to dispose of it as irrelevant, the philosopher surprises the modern reader in a way that is only matched by the striking inclusion under λέξις of the principles of grammar and syntax; [26] for who could think of studying drama today without reference to staging and acting? This move parallels the dismissal of ὄψις at 1450b16–20. [27]

Note now how different the philosopher’s stance is vis-à-vis rhetorical delivery: while he gives the texts of drama and epic priority over performance, [28] so secondary is the written form of an orator’s speech to his public actio (before court or assembly) that Aristotle cannot, in his study of rhetoric, overlook delivery, and the corresponding notion of rhetorical style is quite different from the poetic stylistics that emerges in the Poetics. [29] And yet, even there he owns that the study of delivery is a legitimate θεώρημα of a true aspect of λέξις, however irrelevant to poetics stricto sensu. The difference is readily epitomized by two contrasting definitions. One is found in Poetics 1450b13–14: λέγω δέ … λέξιν εἶναι τὴν διὰ τῆς ὀνομασίας ἑρμηνείαν (‘By lexis I mean expression through wording’); [30] the other, in Rhetoric 1403b16–18: περὶ δὲ τῆς λέξεως ἐχόμενόν ἐστιν εἰπεῖν· οὐ γὰρ ἀπόχρη τὸ ἔχειν ἃ δεῖ λέγειν, ἀλλ’ ἀνάγκη καὶ ταῦτα ὡς δεῖ εἰπεῖν, καὶ συμβάλλεται πολλὰ πρὸς τὸ φανῆναι ποιόν τινα τὸν λόγον (‘To speak about lexis is next; for it is not enough to have [in hand] what one should say, but one must also tell it as he ought to, and this contributes much to the speech coming across as being of a certain quality’; cf. 1404a35–39). This shift in emphasis is further manifest at Poetics 1456a38, where in remarks about the purview of λέξις and διάνοια he assigns pathos to the latter, with only an implicit reference to the corresponding role of lexis understood narrowly not as rhetorical expression but as the argument (λόγος) that falls to inventio. [31] It is true that here the emotions are primarily internal to the play, in and towards dramatic characters (Bywater 1909 ad loc.). More significantly, however, is that Aristotle seems to think of these emotions narrowly as the outworking of the μῦθος (the structure of τὰ πράγματα). Just as they are therefore unrelated to the manner of expression (to the expressive instrumentality of φωνή, gestures, and other elements of delivery), so also is ἦθος similarly divorced from actio: it is rather the articulation of deliberate choice, primarily through argument (the λόγος of διάνοια 1450b8–12) and secondarily through the sequence of events (πράγματα/πράξεις 1450a20–23). [32] Once again, rhetorical style has no inherent bearing on such articulation. Yet, in the ultimate analysis, even for such ἤθη and πάθη is lexis as conceived in the Rhetoric important for plausibility and dramatic illusion vis-à-vis the spectators, all the more so if the performance is to induce in them a catharsis of ‘fear and pity’ (1456b1). [33]

But the little we know about the sophist from the extant sources contradicts such a tendentious reconstruction of his practice. [36] We must, of course, make allowance for the more limited degree of development rhetoric as an art must have received at the hands of this pioneer when compared to Aristotle himself; for I am, after all, arguing that if, according to Aristotle, Thrasymakhos spoke about ἡ κατὰ λέξιν ὑπόκρισις only ἐπ’ ὀλίγον, the philosopher—who observes and, in my opinion, regrets the lack of a τέχνη about delivery—intended to, and in actual fact did, address the gap in the scholarship of his time with his own treatment of λέξις in Rhetoric Ⅲ.1–12. And even though he must have considered the sophist’s discussion deficient enough to warrant his simultaneously holding that ‘no technē has yet been composed’ and that, all the same, Thrasymakhos and others ‘have attempted to speak a little about it’, tradition attests to the sophist’s penning at least one τέχνη, [37] which, as Drerup (1901:227) notes, must have included not only ready-made examples but also accompanying observations of a theoretical nature. And a survey of his contributions quickly reveals that he moves largely within parameters of rhetorical stylistics not unlike those explored by Aristotle in his manual (rhythm, cola, etc.). The reality that emerges from such a survey further underscores the two facts I have endeavored to establish: first, that, far from fencing it off, delivery (in its concern with emotions) was indeed within the scope of Aristotle’s examination of style, just as it received attention (however deficient) by Thrasymakhos; and secondly, that the subjects of rhythm, periodic structure, propriety, and so on could be conceived of and treated under the heading of delivery as a sort of rhetorical stylistics with a view to delivery (ἡ περὶ τὴν λέξιν ὑπόκρισις or ἡ κατὰ ὑπόκρισιν λέξις).

Plutarch, in his Quaestiones convivales 616d mentions Thrasymakhos’ Ὑπερβάλλοντες [sc. λόγοι], which, to judge from the context, must have consisted in arguments built on comparisons (συγκρίσεις). Athenaios X 416a, in turn, writes about the sophist’s προοίμια: it appears, then, that he not only composed examples of closing perorationes but also of opening exordia. These two collections together may correspond to the ἀφορμαὶ ῥητορικαί of the Suidas. As to general qualities of style, Dionysios of Halikarnassos Lysias §6 (p. 14.1–2 Usener) reports that Theophrastos had credited Thrasymakhos with pioneering the middle style [41] (cf. Dēmosthenēs §3 p. 132.14–15 Usener), a manner of expression that ‘compresses the thoughts and expresses them tersely’. [42] In Isaios §20 (p. 122.21–23 Usener) we also learn that Dionysios considered the sophist τῶν … τοὺς ἀκριβεῖς προαιρουμένων λόγους καὶ πρὸς τὴν ἐναγώνιον ἀσκούντων ῥητορικήν, a sentence that Usher (1974:229) ad loc. renders, “[of] those who preferred factual discourses and practical [sic] [43] rhetoric designed for the law-courts” (my emphasis)—an understanding of ἀκρίβεια and ἐναγώνιος that cautions us perhaps not to jump too readily to conclusions as to the meaning of ἀκριβής in Rhetoric Ⅲ.12, [44] where the λέξις γραφική is ἡ ἀκριβεστάτη and the ἀγωνιστική is ἡ ὑποκριτικωτάτη (1413b8–9). A little later in the same section of Dionysios’ Isaios (p. 123.10–14 Usener) we read: Θρασύμαχος δὲ καθαρὸς μὲν καὶ λεπτὸς καὶ δεινὸς εὑρεῖν τε καὶ εἰπεῖν στρογγύλως καὶ περιττῶς, ὃ βούλεται, πᾶς δέ ἐστιν ἐν τοῖς τεχνογραφικοῖς καὶ ἐπιδεικτικοῖς, δικανικοὺς δὲ [ἢ συμβουλευτικοὺς] [45] οὐκ ἀπολέλοιπε λόγους (‘Thrasymakhos is pure, subtle, and terrific both at discovery and speaking at will with terseness or redundancy; but he is entirely into handbooks and epideictic, but has not left forensic [or deliberative] speeches’). Aristotle’s Rhetoric supplies two relevant testimonia. At 1409a2–3 he tells us that orators have used the paean since the time of Thrasymakhos; with this, Cicero appears in broad agreement; [46] and at Rhetoric 1413a6–9 he recounts a brilliant instance of the sophist’s use of similes.

14.2 Hypokrisis and the Use of Writing

Let us now return to Aristotle’s statement about written speeches at Rhetoric 1404a18–19: οἱ γὰρ γραφόμενοι λόγοι μεῖζον ἰσχύουσι διὰ τὴν λέξιν ἢ διὰ τὴν διάνοιαν. The γάρ shows that it is an explanation of the preceding sentence, with its opposition between οἱ τοῦτο δυνάμενοι and οἱ κατὰ τὴν ὑπόκρισιν ῥήτορες. If my understanding of the identity of these two groups is correct, [49] then it follows that those skilled in that delivery which is made ‘artistic’ by its connection to λέξις win prizes because written speeches carry greater force on account of their λέξις than their διάνοια. [50] Doubtless, this need not be the only reason for the success of rhetorically trained speakers. But if the illustration is apposite, we must assume that some orators regularly practiced and trained for their delivery with the aid of written drafts of their own speeches. [51] Accordingly, the philosopher’s argument would run thus: the acting of ‘delivery’ (τὸ ὑποκριτικόν) may be a matter of natural talent and, as such, less technical; but, from the point of view of style, it can be reduced to an art, i.e. systematized, taught, and learned as a discipline; and this is why they win prizes who, through training, have acquired delivery as a skill, even as they do too who have a natural knack for it; for the written speeches that orators use to prepare for and hone their delivery owe their effectiveness less to their argument than to their style—style being precisely what makes delivery susceptible of study. And so we learn that, in Aristotle’s time, the technique of writing was (successfully) used by apprentices of rhetoric and by orators generally but that it remained instrumental and subordinate to delivery. We do not yet have as an ordinary occurrence speeches that are written primarily for a reading public: [52] they are first and foremost scripts of future performances (and only after the fact, secondarily, records of past accomplishment). The circumstances described in the Rhetoric represent, therefore, a midpoint in the evolution of the role of writing as a new technology in the performance culture of ancient Greece: contested in the beginning for the challenge it posed to extemporaneous speaking (and decried by Alkidamas for this reason), it gradually gained broad acceptance among professional practitioners of rhetoric and was likely common in Aristotle’s time. But written speeches were still primarily scripts, ancillary and derivative, mere aids to train for the all-important moment of actual delivery. [53] As Sonkowsky (1959:273) observes, “in the Aristotelian tradition … the techniques of delivery are not merely something that is added in a superficial way after the process of literary composition has been completed, but something that is vitally involved in the very labors of composition anticipating the public presentation.”

We are not surprised that writing should have reached this point by the second half of the fourth century BC. After all, according to the Suidas (s.v. Περικλῆς), roughly a century earlier even Perikles had used scripts to prepare for his public addresses: ‘Public orator and popular leader; who was first to deliver a written speech in court (before him, these were improvised)’. [54] With this note, we might compare Aristotle’s statement, reported by Cicero in his Brutus, that before Corax and Tisias ‘no one was wont to speak according to method or art, but most, however, did with precision and from a written script’ (§46). [55] Though some have emended ‘from a written script’ (de scripto) [56] to ‘precisely ordered’ (descripte), hence putting in doubt Aristotle’s meaning, [57] it is clear, at any rate, that at first the use of scripts was contested and that to it attached a distinct onus of shame that speakers were eager to avoid; so that, if we believe Phaidros, citizens of the greatest power and dignity were ashamed to write speeches and bequeath their writings to posterity for fear that they might be numbered among the sophists. [58] To this sense of shame we might owe that all of Perikles’ scripted speeches perished: [59] perhaps he disposed of them as diligently as he prepared, with their aid, to address the assembly. [60] Whatever the truth about this statesman, there is no doubt in regard to Lysias: for Phaidros introduces him as δεινότατος ὢν τῶν νῦν γράφειν (228a1–2), [61] and Sokrates assumes that his young companion not only bade Lysias repeat his speech but in the end secured possession of the script (τὸ βιβλίον 228b2). Later on, at 258a, Sokrates draws a parallel between ψηφίσματα and λογογραφία, either reckoning, in jest, decrees among the private συγγράματα of those who proposed them, or, more likely, drawing an ironic parallel between the scripts used to prepare for deliberative speeches and the actual language passed by the assembly and inscribed on stone. The picture that emerges presupposes a growing employment of λογογραφία, that is, the writing of speeches with a view to delivery, which made possible lengthy displays (ἐπιδεικνύμενος) of artistic eloquence (σοφία). [62]

Aristotle, too, mentions λογογράφοι at Rhetoric 1388b22, 1408a34, and 1413b13. The first two times he does so in passing, whereas the third mention occurs in chapter 12 of Rhetoric Ⅲ, where the matter of writing resurfaces explicitly, still in the context of the philosopher’s investigation of rhetorical style. Since this chapter closes his examination of the subject, it is reasonable to expect that he should take a synoptic perspective on λέξις and provide the reader with a synthetic summary of the issues involved in ‘style with a view to delivery’. [63] I submit therefore that Rhetoric Ⅲ.12 does not consider particular issues of detail but generic matters broadly applicable to the subject of delivery in its scientific connection to style. This is an important observation for, if ὑπόκρισις and writing reappear explicitly in what constitutes the capstone of Aristotle’s treatment of λέξις, they must be critical elements of rhetorical style, not afterthoughts or secondary considerations. This validates the views I have propounded throughout this and the previous chapter. First, that ὑπόκρισις is not some ethically undesirable feature of the art of oratory to be fenced off and purged from style; rather, it is no more and no less than what the philosopher himself states, namely, ‘that which has the greatest force’, and for this reason it must remain at the heart of the oratorical practice of any successful speaker. And second, that writing in the Rhetoric is not primarily the instrument of speech production for a reading public but a subordinate technology at the service of delivery, and hence studied and approached as such.

What is it, then, that we learn from Rhetoric Ⅲ.12? This chapter starts with what is perhaps the most commonly misinterpreted passage in the entire treatise. I quote it here, for now without a translation, for the convenience of the reader:

δεῖ δὲ μὴ λεληθέναι ὅτι ἄλλη ἑκάστῳ γένει ἁρμόττει λέξις. οὐ γὰρ ἡ αὐτὴ γραφικὴ καὶ ἀγωνιστική, οὐδὲ δημηγορικὴ καὶ δικανική. ἄμφω δὲ ἀνάγκη εἰδέναι· τὸ μὲν γάρ ἐστιν ἑλληνίζειν ἐπίστασθαι, τὸ δὲ μὴ ἀναγκάζεσθαι κατασιωπᾶν, ἄν τι βούληται μεταδοῦναι τοῖς ἄλλοις, ὅπερ πάσχουσιν οἱ μὴ ἐπιστάμενοι γράφειν.

Rhetoric 1413b3–8

Its syntax lends itself all too readily to readings that, not to put too fine a point on it, make nonsense of Aristotle’s argument. Conversely, the most compelling and, I believe, correct interpretation is burdened by a choice of referents for ἄμφω and the bipartition τὸ μὲν … τὸ δέ at 1413b5–6 that is not immediately apparent and only follows from reflecting on the context and the logic of the argument. At the same time, I must emphasize that the correct identification of these referents does not call for a construal of the grammar that is in any way questionable or exceptional. Therefore, even if it does not correspond to the most immediately apparent reading, with the proper perspective Aristotle’s true meaning flows with unexceptional ease. Since the context regards two pairs, the first, the ‘graphic’ and the ‘agonistic’ styles (γραφική and ἀγωνιστική), the second, the ‘demegoric’ and the ‘dicanic’ (δημηγορική and δικανική), a reasonable understanding of the immediately succeeding ἄμφω would seem ‘X + Y’ where ‘X’ is ‘the one,’ and ‘Y’ is ‘the other,’ element of whichever pair is in view. Yet, quite apart from the interpretive difficulties that result from any of the corresponding four possible alternatives, the neuter gender of τὸ μέν and τὸ δέ already hints that perhaps none of them is in view. Aristotle of course could have switched from the feminine-gender adjectives to a neuter bipartition (‘the former [thing] … while the latter [thing]). The mismatch between the feminine λέξις and the neuter verbal predicates (ἐπίστασθαι and μὴ ἀναγκάζεσθαι) might even be enlisted to motivate the switch. But it is unarguable that, had he wanted to, he could have placed beyond doubt that he had in mind one of the four alternatives by writing ἡ μὲν … ἡ δέ instead. If in fact this was his meaning, it is only too ironic that he did not take his own advice at 1407b6–9, the fourth principle of correct Greek usage (ἑλληνίζειν) ascribed to Protagoras, and failed to secure gender agreement with a view to clarity. Aristotle deserves no such criticism, however, if my interpretation below is correct. At any rate, one can infer how hard it is to make sense of the passage from how often most scholars neglect to engage or translate it. This can only be explained as a show of embarrassed silence or as proof that they are oblivious to the interpretive difficulties entailed by the commonplace rendering. On the other hand, full-length translations of the Rhetoric, of which I list the following samples, cannot avoid going on the record:

One should not forget that a different lexis is appropriate for each genus [of rhetoric]. For the written and agonistic [style] are not the same; nor are the demegoric [deliberative] and the dicanic [judicial], and it is necessary to know both. [Debate] consists in knowing how to speak good Greek; [writing] avoids the necessity of silence if one wishes to communicate to others [who are not present], which is the condition of those who do not know how to write.

Kennedy 1991:255

Es preciso no olvidar que a cada género conviene un estilo; no es el mismo el estilo de prosa escrita que el de debate, ni el de hablar ante el pueblo que el forense. Ambos, pues, es necesario que se conozcan: es el uno saber expresarse en griego, el otro no quedar obligado a callar cuando se quiera comunicar algo a los otros, lo cual les pasa a los que no saben escribir.

Tovar 1953:201

Il ne faut pas oublier qu’à chaque genre s’approprie un style différent; car le style des compositions écrites n’est pas le même que celui des débats, et celui des assemblées n’est pas non plus celui des tribunaux. Or, il est nécessaire de connaître les deux: le style des débats suppose la connaissance du grec; grâce à l’autre, on n’est pas forcé de garder le silence si l’on veut communiquer sa pensée aux autres, inconvénient dont souffrent ceux qui ne savent pas écrire.

Dufour and Wartelle 1973:73

It must not be forgotten (lost sight of) that a different kind of language is appropriate to each different kind (of Rhetoric). For the same style is not suitable to written composition (that which is intended to be read) and that which is used in debate (in the contests, the actual struggle, of real life); nor again in (the two divisions of the latter) public and forensic speaking. The orator must be acquainted with both: for the one (debate) implies the knowledge and power of clear expression in pure Greek, and the other freedom from the necessity (lit. the not being obliged to) of suppressing in silence (κατά, keeping down) anything that one may want to communicate to the rest of the world; which is the case with those who have no knowledge (or skill) of writing (i.e. composition).

Cope 1877:3.145 (his emphasis)

Man darf aber nicht vergessen, dass zu jeder Gattung eine andere sprachliche Form passt. Denn die schriftliche Form und die der Kontroverse sind nicht dieselbe, und auch nicht die für die politische Rede und die für die Gerichtsrede geeignete. Beides aber muss man kennen. Das eine nämlich bedeutet, dass man sich auf korrektes Griechisch versteht, das zweite, dass man nicht gezwungen wird zu schweigen, wenn man den anderen etwas mitteilen will, wie es denen ergeht, die nichts vom Schreiben verstehen.

Rapp 2002:1.150

Minime autem opus est nos fallere, aliam singulis generibus congruere locutionem. Non enim eadem est idonea scribendo, et forensibus concertationibus: neque etiam contionalis et iudicialis. Ambo autem necesse est tenere: alterum enim est scire uti Graeco sermone; alterum vero non cogi silere, siquis voluerit participem aliquem facere suar[um] cogitationum. Quo laborant, qui scribere nesciunt.

Vettori 1579:677–678

Of the two pairs of stylistic opposites that open the chapter, scholars make the latter, the δημηγορική and δικανική, an explication (by way of division) of the ἀγωνιστική, the second member of the first and more fundamental opposition around which Aristotle articulates his presentation: there is, on the one hand, a ‘graphic’, and on the other, an ‘agonistic’ style. The agonistic suits the ἀγών, i.e. a contest (in the extended sense) in which various options compete for the allegiance or support of the audience, whether in the law-court or the assembly (calling respectively for the λέξις δικανική and the λέξις δημηγορική). If the relationship commonly assumed between the first pair and the second is right, the graphic style, unlike the agonistic, does not receive a shorthand explication. Instead, after noting that ‘one must know both’ (ἄμφω δὲ ἀνάγκη εἰδέναι 1413b5), the philosopher appends as rationale an enigmatic statement about the possibilities closed to those who ‘know not how to write’: τὸ μὲν γάρ ἐστιν ἑλληνίζειν ἐπίστασθαι, τὸ δὲ μὴ ἀναγκάζεσθαι κατασιωπᾶν ἄν τι βούληται μεταδοῦναι τοῖς ἄλλοις, ὅπερ πάσχουσιν οἱ μὴ ἐπιστάμενοι γράφειν (1413b5–8). This is an observation that has baffled many. Rapp (2002:2.932) sums up the difficulty with characteristic insight: “Nicht sehr erhellend ist die Begründung, die der Abschnitt für die Nützlichkeit der schriftlichen und der kontroversen Form gibt. Wenn die Kontroverse nämlich mit der Fähigkeit in Verbindung gebracht wird, sich auf korrektes Griechisch (ἑλληνίζειν) zu verstehen, dann scheint damit nur eine minimale Bedingung genannt zu sein, und außerdem eine, die auch der schriftlichen Form zugrunde liegt … . Ebenso wenig klar ist, warum die Eigenschaft, nicht zum Schweigen gezwungen zu sein, wenn man etwas mitteilen will, eher die schriftliche als die mündliche Form betrifft.” [

As one readily infers from the translations quoted above, most interpreters take ἄμφω to refer to the graphic and the agonistic styles, i.e. to the first pair, and tie τὸ μὲν … τὸ δέ respectively to either end of its polarity. In particular, most correlate τὸ μέν to ἡ ἀγωνιστικὴ λέξις and τὸ δέ to ἡ γραφικὴ λέξις. [65] Nothing in the grammar of the sentence hinders this approach. As noted above, the problem with it is that its rationale is specious and the resulting interpretation neither suits the context nor yields adequate sense under close examination (see below). The alert reader will not lack signs that all is not good with this consensus reading. There is no reason, for example, why ἄμφω should refer to the first pair and not the second. Only the critic’s sense that the former is the more basic opposition suggests the choice. But what if in fact the second pair does not represent a subdivision of the second element of the first (ἡ ἀγωνιστικὴ λέξις)? What if instead it reprises the selfsame distinction in terms of γένη? Given that the design of Rhetoric Ⅲ.12 is to survey style in terms of what is peculiar to each rhetorical genus, such a reprise would be eminently reasonable. On this assumption, the philosopher would have first set down the polarity that comprises the entire range of rhetorical style (‘graphic::agonistic’) and then he would have illustrated either end with the style of a relevant genus (demegoric and dicanic). If this is what Aristotle did, there would not be any reason to fault him for failing to give comparable due attention to the graphic end of the polarity that engages his attention throughout the chapter. If not impossible, such an oversight at its programmatic opening is at least remarkable. Yet no neglect of either pole attaches to the alternative, which also relieves the apparent ambiguity that burdens ἄμφω, τὸ μέν, and τὸ δέ in the consensus reading: if we posit a reprise it does not matter which pair is selected, for the latter would then take up the εἴδη of the former at the level of γένη. I submit that this is in fact Aristotle’s approach and that only this proposal satisfactorily explains the otherwise mystifying γάρ clause, structured as a τὸ μὲν … τὸ δέ progression [66] that (I will show) hinges, on the one hand, on ὑπόκρισις undergirded by ἑλληνίζειν (basic but non-trivial linguistic competence, including stylistic); and, on the other, on ὑπόκρισις underpinned by the ability to write scripts that help in structuring the discourse and in training for public delivery.

The proposal that the second pair recapitulates the first seems at first precluded by the (to us) natural association of both the demegoric and the dicanic settings with rhetorical ἀγῶνες. But one must not forget that Aristotle is presenting a continuum between two poles and that for him, at least potentially if not in every actual instance, all three γένη—symbuleutic, forensic, and epideictic—share in delivery not only a measure of debate and competition (the ἀγών) but also a measure of written planning. Therefore, as regards their ὑπόκρισις they all fall stylistically somewhere along the continuum between the graphic and the agonistic poles. For the second pair to be able to reprise the first one only needs the dicanic and demegoric styles to be located at clearly distinct points along the continuum. This is indeed the case. In fact, the dicanic style illustrates the graphic in an even more compelling way than the epideictic because the reader could relate to it with greater immediacy: after all, resort to written scripts in the courtroom must have been commonplace by the second half of the fourth century, [67] whereas few would have had (or could be expected to have at some future time) comparable first-hand experience delivering epideictic. Any Athenian resident—every one of Aristotle’s hearers and readers—might have to engage in dicanic oratory sooner or later. To that extent, Aristotle’s preference for the dicanic style to illustrate in relative terms the graphic end of the stylistic range would be straightforward. And it is not as if he depended solely on the cultural context to guide the reader’s interpretation. He both implied and explicitly affirmed the more markedly graphic character of the dicanic style: at 1414a17–18 we read that ‘the epideictic style is the most graphic; for its effect is reading; and the dicanic, the second [most graphic]’. [68] Furthermore, in his analysis of ἀκρίβεια (‘precision’), a property which the graphic style possesses to a relatively greater degree, [69] he clearly contrasts the demegoric style to the dicanic: ‘Now, the demegoric style seems entirely like shadow painting … therefore, precision appears a waste and the worse … ; the judicial plea is a more exact matter’. [70] In what may amount to a striking Gedankenexperiment, [71] he goes so far as to consider a trial with a single judge at which ἀγών is absent (1414a13). Having said all this, I must emphasize that my interpretation of ἄμφω at 1413b5 and of the progression τὸ μὲν … τὸ δέ that follows it does not depend on the reader’s acceptance of my proposal that the second pair at 1413b4–5 recapitulates the first. Still, this proposal makes for a more compelling reading. My argument only requires what most commentators already assume: that ἄμφω refers to the former and more basic opposition ‘graphic::agonistic.’

In my reading of the passage, ἄμφω refers primarily to the ‘graphic::agonistic’ stylistic polarity as a whole and, secondarily, to Aristotle’s illustration of it, i.e. to the demegoric and dicanic styles: the former as the relatively more agonistic and less graphic, the latter as the relatively less agonistic and more graphic. The statement ‘it is necessary to know both’ (ἄμφω δὲ ἀνάγκη εἰδέναι 1413b5) affirms the need to know the full panoply of stylistic resources and strategies, including writing, that make for various registers along the graphic-agonistic continuum. Thus, Aristotle makes mastery of the stylistic polarity the ideal of the well-trained orator who equipped with persuasive delivery successfully addresses audiences in any of the three chief social settings of the classical polis: the deliberative assembly, the courtroom, and the gathering of the people at festivals and other solemn occasions. There is nothing in the context of the passage or the culture of classical Greece that would lead us to expect Aristotle to proffer the ability to communicate with an absent (or distant) audience as motivation for the would-be orator’s need ‘to know both [styles]’. Surely, writing letters to distant relatives and friends hardly qualifies as a rhetorical accomplishment, whether it presupposes personal literacy or not. This can hardly be taken to entail or justify the need to master the graphic style, even if one takes for granted (unjustifiably, in my view) that letter writing inevitably required the interested party to know how to write. If it is to make contextual sense, the ability to write that Aristotle has in view when he mentions ‘any who do not know how to write’ (οἱ μὴ ἐπιστάμενοι γράφειν 1413b8) must serve specifically as adequate motivation for the acquisition of a particular rhetorical skill, not simply for the acquisition of basic literacy. To reduce the philosopher’s meaning to the latter impoverishes his argument, effectively divesting it of rhetorical significance and wresting it from its technical setting—a treatment of rhetorical style. The thought is recast instead as one broadly applicable to the common man without specific interest in oratory as an art. Even on these terms, the alleged meaning is remarkably flaccid, for the need that literacy is supposed to meet—writing private communications to a distant or absent audience—seems marginal and infrequent.

A more promising alternative might seem to promote the written communication from letters to rhetorical speeches. This evades the objections raised above but creates in turn its own share of difficulties. For then only in one historically exceptional case, the orator Isokrates, does the modified interpretation of the τὸ δέ clause [72] makes some sense. One may go so far as to wonder if the reading has not been formulated with him in mind and if without him anyone would have ever thought of understanding the clause thus. For without literacy Isokrates could never have communicated so widely ‘to others’, including distant and absent addressees, what he wished to say. But even in his case the hypothetical consequence of illiteracy, ‘being forced to remain silent’, seems a distinct overreach: he could always deliver his addresses orally to smaller gatherings, as he is actually known to have done on many occasions. Although ordinarily such delivery involved reading (and, therefore, also writing), I doubt that he would have been reduced to silence on the counterfactual assumption that he was illiterate. If someone answers that the consensus interprets Aristotle’s comment not with regard to the bare capacity, and hence the opportunity, to communicate in writing but as regarding the skill to craft such rhetorically accomplished orations as made Isokrates famous, then he is embracing my objection: basic literacy—all that is required by assuming a distant audience—is not the philosopher’s point, but writing as instrumental to rhetorical τέχνη, specifically, to the orator’s scripted composition and delivery of his addresses. At any rate, the case of Isokrates is so extraordinary, and he so poorly exemplifies (or embodies the ideal of) the would-be orator that the Aristotelian treatise envisions, that he cannot possibly serve to explain the τὸ δέ limb of the progression, especially without ever so much as a contextual hint that he is in view there. We must reject as implausible any reading that reduces to so narrow and exceptional a scope what Aristotle surely intended as a general motivation, broadly applicable to his rhetorical pupils, for the need to learn the full ‘graphic::agonistic’ stylistic continuum. Aristotle did not intend to replicate the Isokratean model among his disciples.

What, then, must the reader do with τὸ μὲν … τὸ δέ? First, he should note that stricto sensu the subject of ἐστιν cannot be a style, e.g. ἡ γραφικὴ λέξις or (if he so choose) ἡ δημηγορικὴ λέξις: the agonistic style as a category cannot be strictly equated with any kind of ‘knowing’; nor can the graphic style be equated with the absence of any given compulsion (‘not being forced’). A style is a set of expressive devices and strategies (in Aristotle’s case, with a view to delivery), not the knowledge thereof (or of any other thing). That is the reason why translators do not render ἐστιν by ‘is’ but by ‘suppose’ (Budé), ‘implies’ (Cope), ‘bedeutet’ (Rapp), ‘consists in’ (Kennedy), vel sim. Only familiarity with the consensus reading could have inured us to the inconcinnity of the statement, ‘the agonistic style is to know how to speak good Greek’. To be sure, it is possible to argue oneself into accepting the conceptual sloppiness of this sentence and taking it in its only possible sense, ‘[to know] the agonistic style is to know how to speak good Greek’ (my emphasis). This point may strike my reader as more than a little pedantic, but attending to it actually helps to construe the passage aright. For there is no need to impute such mental and syntactic muddle to Aristotle: the referent of τό in both τὸ μέν and τὸ δέ is the neuter infinitival clause [τὸ] ἄμφω εἰδέναι. One can hardly find fault with the positive equation of skill with skill, of ‘to know’ (εἰδέναι) with ‘to know’ (ἐπίστασθαι); or, conversely, object to the conclusion that if someone possesses a given skill he can evade a corresponding limitation. This amounts to equating ‘to know’ with ‘not to be forced to remain silent’ (μὴ ἀναγκάζεσθαι κατασιωπᾶν), an evasion predicated of ‘any who know how to write’ (οἱ ἐπιστάμενοι γράφειν). To construe both τὸ μέν and τὸ δέ as having and pointing to the same referent, ἄμφω εἰδέναι, entails unproblematic, ordinary syntax. Because they share a single referent I do not call the resulting balanced structure ‘an opposition’ but rather ‘a progression [of thought]’. Another instance of τὸ μὲν … τὸ δέ with a common referent—this one, undisputed—is found in Rhetoric 1356b27–28:

ἐπεὶ γὰρ τὸ πιθανὸν τινὶ πιθανόν ἐστι, καὶ τὸ μὲν εὐθὺς ὑπάρχει δι’ αὑτὸ πιθανὸν καὶ πιστὸν τὸ δὲ τῷ δείκνυσθαι δοκεῖν διὰ τοιούτων, οὐδεμία δὲ τέχνη σκοπεῖ τὸ καθ’ ἕκαστον, οἷον ἡ ἰατρικὴ τί Σωκράτει τὸ ὑγιεινόν ἐστιν ἢ Καλλίᾳ …

Rhetoric 1356b26–30

For since what is persuasive is persuasive to someone, and it is persuasive and credible now immediately in and by itself, now by seeming to be proved through such things [i.e. through means persuasive and credible], and no art considers the particular, e.g. medicine what is healthful for Sokrates or Kallias …’

It is clear that τὸ πιθανόν is picked up both by τὸ μέν and by τὸ δέ. The resulting progression of μὲν … δέ divides into inherent and derivative the reasons for the convincingness and credibility of what persuades. [
73] Because in this particular case these reasons are ordinarily [74] viewed as mutually exclusive, μὲν … δέ may be taken to articulate not a constructive gradation (‘first A; then, building on A, B’) but an alternative (‘either A or B’). Generally speaking, the relationship conveyed by μὲν … δέ depends on the particular context. This use of τὸ μέν and τὸ δέ is closely related to the adverbial one usually rendered as ‘now … now’, ‘partly … partly’, ‘one the one hand … on the other’, ‘both … and’. [75] In the case of Rhetoric 1356b27–28, the only possible latitude in the construal of the grammar is whether one assumes the implicit subject τὸ πιθανόν for ὑπάρχει, with τὸ μέν … τὸ δέ adverbial (‘now … now’) or, more naturally, one grants τό in τὸ μέν and τὸ δέ deictic force as the demonstrative ‘this’ (τό = τὸ πιθανόν). The syntax and meaning of the particles in this case are closely paralleled at Rhetoric 1413b6–8.

What has hitherto led many astray is the assumption, grammatically unexceptional but of impossible meaning, that τὸ μέν pointed to one member of either pair in Rhetoric Ⅲ.12 §1 (usually taken to be ‘the agonistic style’) while τὸ δέ pointed to its complement (‘the graphic style’). This specious construal seemed justified by the emphasis on speaking in the first limb of the progression and the complementary emphasis on writing in the second. But this reading left out of sight the fact that up to that point in his treatise Aristotle had been exclusively concerned—and, in the cultural context even of late fourth-century Greece, could only plausibly be assumed to concern himself—with the oral delivery of rhetorical prose, whether or not composed in advance with the aid of writing. For him to reduce the second limb to basic literacy would have been unthinkable and, in the context of his manual, little short of absurd. [76] Cope (1877:3.145) felt the force of this difficulty: noting that Vettori construed τὸ μὴ ἀναγκάζεσθαι … τοῖς ἄλλοις “of actual writing, that is of letters to absent friends,” he added that this “seems to narrow the meaning of ‘writing’ in such a way as to produce a somewhat ridiculous result. Surely any educated man, whether he be an orator and statesman or not, requires and possesses the knowledge of writing in that sense” (his emphasis). The specious justification for this interpretation dissolves in the light of the rhetorical stylistic continuum, which allows for precisely such apportioning of emphases as a progression that reckons ἑλληνίζειν insufficient for the perfect orator who, rising with confident proficiency to any occasion at hand, aspires for success in all of the social settings that call for public speaking. In the logic of the passage, ‘knowing both’ is explicated first as the linguistic competence of basic, though by no means trivial or contemptible, mastery of Greek expression, and second as the specific rhetorical skill of ὑπόκρισις that the knowledge of writing enables, when the orator prepares scripts the force of whose λέξις allows him to overcome the timidity and fear that might otherwise dissuade him from addressing the audience with the message he wants to make public. The motivation to master the entire stylistic continuum lies in the force of these two examples, a force that would be felt by any Athenian citizen (and, a fortiori, by any would-be orator), who could not fail to appreciate the significance of confident and effective public speaking for his social standing, legal indemnity, and political involvement.

I believe that chiefly in Aristotle’s mind is the deliberative setting of the ekklēsia and the council, where any male citizen who wished, regardless of station, [77] so long as he had not been subject to ἀτιμία, had the right to address the people: ‘The lawgiver expressly made known who should address the assembly and who should not speak before the people. And he does not drive from the platform one who is not from ancestors who have served as generals, nor indeed if he plies some trade to provide for his necessary food, but these he also heartily welcomes and for this reason again and again asks, “Who wants to address the assembly?” (τίς ἀγορεύειν βούλεται;)’. [78] It is the political resonance of this often posed question, τίς ἀγορεύειν βούλεται;—the privilege, duty, and challenge of political initiative—that motivates the philosopher’s ἄν τι βούληται μεταδοῦναι τοῖς ἄλλοις. [79] Although the ideology of the Athenian state encouraged and, ideally, even presupposed the active involvement and initiative of private citizens in the political process, in practice most would have been intimidated and paralyzed by the social and legal stakes of doing so. [80] Hansen (1991:267) observes that ὁ βουλόμενος, the generic Athenian citizen who elected to take the initiative, was not “subject to dokimasia or euthynai like the magistrates; but he could always be made to answer for his initiative, even, in extreme cases, be condemned to death for it” (cf. Demosthenes 10.70). Ober 1989 adds: “The functional difference between the expert politician and the ‘amateur’ … was in part a consequence of the increased risk that political activity brought in its wake. The maze of legal difficulties in which the proposer of a controversial decree might find himself would be enough to scare most citizens away from the bema during intense debate over major issues. The general opprobrium that might follow upon a failed policy decision was also a consideration” (his emphasis). Under the circumstances, it is unsurprising that even this least graphic of all the rhetorical genera should have invited even the experienced ῥήτωρ—and, ordinarily, in the case of the ἰδιώτης probably even required him—to prepare diligently with the aid of a script before communicating his mind to others (τι [διανοίας] μεταδοῦναι τοῖς ἄλλοις). [81] The alternative would have been self-imposed silence, prudent, perhaps, but one that diminished his citizenship by abridging in practice the privilege of ἰσηγορία and παρρησία. [82] A citizen who does not speak up, who flinches from addressing an audience or delivering a speech, lacks the civic voice in law-court or assembly enjoyed by others who need not similarly flinch because their literacy has opened the door to scripted delivery. [83]

For Aristotle, rhetorical training must arguably start and build on ἑλληνίζειν, [84] but it cannot stop there if it hopes to realize the full potential of citizenship in all the relevant public settings. [85] Foundational linguistic competence is not to be despised: it involves a non-trivial mastery of the rules of Greek grammar and the conventions of Greek idiom with a view to rhetorical clarity and suasion. [86] But this necessary foundation looks ahead to the superior control over style and content that writing offers the would-be orator of average talent. The promise of stylistic precision and control over the relationship between form and content—this must be underlined—is not only to be found at the graphic end of the stylistic continuum. Whenever scripts were involved, the degree of precision that could be attained by an average speaker—one not exceptionally gifted by nature—and the control he could exert in delivery over the arrangement and stylistic expression of the material would have been considerably greater than when his address was left to his average improvisatory skills. To the extent that writing was involved at all, even if, in the nature of the case, it served to craft a naturalistically realistic debate style, ‘those who knew how to write’ would enjoy an immeasurable advantage. As often in life, the skills acquired by art would most dramatically improve those who were less exceptionally endowed. In retrospect, then, Aristotle’s statement at 1413b7–8 does not so much concern the graphic end of the stylistic continuum as it illustrates the importance of writing to all styles, including the demegoric; and, to that extent, it also underlines the promise and importance of knowing the graphic style. One may safely assume that the importance of mastering the agonistic style was readily apparent to one and all. Hence, by seizing upon the demegoric genus to illustrate his argument, the philosopher renders it all the more compelling in that he shows the point at issue to be valid even where it is less obviously applicable. [87]

The view that Rhetoric 1413b6–8 is concerned with writing, pure and simple, irrespective of delivery, was built in disregard of the context and traded on a false parallel and several interpretive fantasies. One such fantasy is that μεταδίδωμι pointed to written communication. [88] But ‘to communicate in writing’ is neither its denotation nor even an ordinary connotation. The verb simply puts the accent on the impartation of one’s mind, opinion, or oral argument (διάνοια, δόξα, or λόγος), a nuance that is readily captured by our own ‘to communicate’; or else in certain contexts it conveys the act of granting another the right to speak (lit., giving him a share in speech, λόγος). Now, as it happens, where it is used for ‘to communicate’, it regularly involves a speaker who addresses one or more hearers face-to-face. [89] In these contexts, the instrumentality of writing is decidedly exceptional and chronologically late. A further interpretive fantasy is that τοῖς ἄλλοις meant ‘the rest of the world’ (to use Cope’s translation), a superset that would necessarily contain a preponderance of absent addressees (hence Kennedy’s “others [who are not present]”) and would therefore call for written dissemination. Besides disregarding the context, which cannot reasonably be assumed to refer to a scenario so extraneous to the rhetorical needs and practices of the classical period, [90] to regard the inability to write as ‘the necessity to remain silent’ should have been felt as decidedly odd. A natural reading of κατασιωπᾶν strongly suggests oral sound, not printed writing. In Demosthenes 41.23, Athēnaiōn Politeia 14.2, and Isokrates 8.38 it means, pointedly, ‘not to speak up’, and in the latter two loci the alleged cause is fear. Both the sense and the motive suit the present context very well. Even when it is applied to written communication, as happens with the late Josephus Against Apiōn 2.238, the verbal action it denotes, ‘to speak up’ (the refusal to remain silent), still occurs in the face of an opposing challenge, whether attended by fear or not. (This is also true of Demosthenes 41.23, which concerns the natural tendency to speak up and dispute false and unfair accusations.) The instances surveyed suggest that Aristotle’s words about ‘being forced to remain silent’ were poignant and not without a sting, a sting that is largely absent from the consensus reading.

Innes (2007:154–155) is the most recent proponent of the audience of absent addressees who can only be reached in writing. Her strategy is to quote three passages from Isokrates where the orator uses οἱ ἄλλοι to deliberate rhetorical effect. In the first, Isokrates 11.44, he wants to establish a notional opposition between the addressee of his work, for whom he alleges keen pedagogical concern, and ‘the rest’, an interloping generality of potential hearers whose approbation sophists eagerly seek in display performances (ἐπιδείξεις) that are self-seeking and indifferent to the true well-being of their audiences. Isokrates’ point is that sophistic epideictic may well have a large following and reach a numerous audience but it does not have the interest of its hearers at heart. Isokrates, in contrast, may not reach as far and may not be as popular and influential, but his goal is not thereby frustrated for it is not display and showiness: he genuinely cares for his addressee. Thus, Isokrates Bousiris 44 opposes ἐπίδειξιν τοῖς ἄλλοις ποιούμενος to ὑποδεῖξαί σοι βουλόμενος: the contrast is not between present and absent addressees but between a self-serving ‘display’ (ἐπίδειξιν) and a pedagogically informed ‘indicating’ (ὑποδεῖξαι). [91] The orator does not imply that the interloping οἱ ἄλλοι are any more absent that the addressed σύ. Their existence is mooted only to be rhetorically excluded. Innes’s second example, Isokrates 2.7, not only fails to support her point but even nicely parallels what I argue is Aristotle’s meaning in Rhetoric 1413b7: Isokrates contrasts the composers of prose or poetry (οἱ συντιθέντες), who at the design stage hold their works in their minds (ἐν ταῖς διανοίαις), with ‘the rest’ (τοῖς ἄλλοις), i.e. the audiences, who acquire their knowledge of these works only in performance (ἐπιδειχθέντα). Far from establishing a contrast between the orator’s speech and epideictic displays intended for an absent worldwide readership, here he calls To Nikoklēs a protreptic ἐπιχείρημα that serves both king and subjects (Isokrates 2.8). Finally, Isokrates 9.74 builds on the well-known topos of the winged word that carries an individual’s name far and wide, a topos best known from Theognis’ celebrated promise to Kyrnos: ‘I have given you wings with which you will fly, soaring easily, over the boundless sea and all the land. You will be present at every dinner and feast, lying on the lips of many’ (Theognis 237–240). [92] Isokrates contrasts immobile statues with λόγοι that ‘can be carried out to Greece (ἐξενεχθῆναι … εἰς τὴν Ἑλλάδα) and, once distributed among (διαδοθέντας) the leisured gatherings of enlightened men, are welcome by them’. [93] Although ἐκφέρω and διαδίδωμι may be used of publication, they neither presuppose nor require strictly written dissemination. [94] In this passage Isokrates is studiously ambiguous and does not explicitly claim that his λόγοι are written speeches anymore than Theognis’; on the contrary, we are explicitly told that they are mediated by performance in the leisured gatherings (ἐν ταῖς διατριβαῖς) of enlightened men. Although he acknowledges his writing soon thereafter at 9.76, such recognition is tempered by the paring of λέγειν καὶ γράφειν at 9.80. [95] Here, at any rate, we are far away from a public of absent, isolated readers, and Isokrates’ point surely turns, at least in part, on the fact that his speech, ostensibly a gift for Nikokles (a speculum principum), has carried Evagoras’ reputation all the way from Cyprus to Athens where (such is the conceit) it will make his intellectual temper known by embodying it ‘in words that are spoken’ (τὰς διανοίας τὰς ἐν τοῖς λεγομένοις ἐνούσας 9.75). [96] The public that is notionally absent from Cyprus is actually present with Isokrates, whose Athenocentric vantage point readily equates Athens with Hellas. [97]

To return to τοῖς ἄλλοις at Rhetoric 1413b7, the implicit contrast is clear: there is, on the one hand, the individual who wants to impart his mind; on the other, those who are apprised of it. This is the ordinary pragmatic context of the communication process. The passage gives no intimation that the receiver is distant from the emitter and hence can only be reached through writing. It is easy to demonstrate that the sense of οἱ ἄλλοι is simply ‘the rest’, not ‘the rest [of the world]’ (much less a ‘rest’ that is absent): in Rhetoric 1355a19 Aristotle uses it to contrast himself with the rest of those who have written technically about ῥητορική; at 1362a7, the opposition is between the rest of the brothers who are shameful and the one who is noble; at 1363b3, ‘the rest’ are all the ⟨φιλο⟩τοιοῦτοι other than the explicitly mentioned φιλόνικοι, φιλότιμοι, and φιλοχρήματοι; at 1369a16, those not already mentioned who act in accordance with their habitual characters; at 1388b7, all not included under the adjacent relative οὕς; at 1408a36, οἱ ἄλλοι are all but ὁ ἀκούων. Extending this short survey to cases other than the nominative yields the same outcome: ‘the rest’ is contextually defined and ordinarily does not mean ‘the rest [of the world]’. There is therefore no warrant for the view advanced by the consensus other than the presupposition that writing to a distant readership is entailed by the passage. To this presupposition, a suitable meaning of τοῖς ἄλλοις is made to correspond. This sense does not flow from the statement. It is read into it.

But some believe to have found the necessary justification for this reading in Alkidamas On the Sophists §31. This section comes in the context of the sophist’s justifying his paradoxical use of writing for the broadside against oratorical writing. He makes clear that he does not entirely dismiss the ability to write (τὴν γραφικὴν δύναμιν) but merely considers it inferior to performing extempore (§30). He resorts to writing, moreover, not with an attitude (οὐκ ἐπὶ τούτῳ μέγιστον φρονῶν) but only to show those who pride themselves on this skill that with little effort he succeeds in blotting out and demolishing their arguments for it. Then comes the following:

On the Sophists §31

There is nothing in the passage that suggests strictly written dissemination of Alkidamas’ writings. εἰς τοὺς ὄχλους ἐκφερομένων regards delivery before the crowds, with ὄχλους hinting at the sophist’s contempt for the multitudes who, spoiled by their frequent hearing of speeches prepared with the aid of writing, can no longer appreciate the finer and superior skill of extempore speech-making. This negative connotation of ὄχλοι also seems present in Gorgias’ Helen 13, which envisions crowds delighted and persuaded by speeches that are artfully written but spoken without truth; and in Palamēdēs 33, where a mob (ὄχλος) that can be swayed by piteous wailing, prayers, and the entreating of friends is contrasted with those who are and appear as the foremost of the Greeks and are persuaded by one who exhibits the truth with manifest justice. So it is also with Isokrates, who in To Nikoklēs 49–50 compares the crowds (ὄχλοι) that others seek to please, rather than admonish and counsel, with Nikokles, who receives instead his paraenetic address; and who in Nikoklēs 21 contrasts advisers skillful at addressing the crowds (ὄχλοι) with advisers who know how to deal with the issues. [
103] Finally, Aristotle’s own use of the word in Rhetoric 1395b28–29 also seems decidedly negative: ‘For this is why the uneducated are more persuasive with the crowds than the educated, just as the poets say that the uneducated are more accomplished in speaking before a crowd’. [104] Mariß (2002:291) remarks on the undertone of ‘regret’ in Alkidamas’ acceptance of writing for delivery before crowded audiences: “[Alkidamas sagt] daß gemeinhin ἐπιδείξεις veranstaltet würden und daß er sich wegen dieses allgemeinen Trends dem Schreiben zuwende.” To the hint of contempt he expresses with τοὺς ὄχλους, add the fact that in §11 ἀκρόασις refers to passive and politically inactive hearers under a tyranny (Mariß 2002:295). The entire context of §31, then, is one of oral delivery, whether the display is improvised or scripted. Hence the mention of ‘[vocal] recitations’ (ἀκροάσεις), of an audience spoiled by ‘listening’ (ἀκροᾶσθαι) to the speeches of others, and of forming an opinion of Alkidamas on ‘hearing’ (ἀκούοντες) him speak extempore. Nothing other than the bare fact of the sophist’s resort to writing and the failure of scholars to realize that, just as in the rest of the broadside, scripted delivery is in view supports the implausible reading that Alkidamas is now referring to the written dissemination of his speeches through the book market. If so, why should he refer to the alleged readership as ὄχλοι? A gathered throng was the desired ideal of the display sophist (Isokrates Letter 1.6): what justification does the modern scholar have to turn it into a potential multitude of individual solo readers? Hudson-Williams (1949:65) had it right: “Alcidamas, a contemporary of Isocrates, says that one reason which prompted him to compose an artistic written λόγος instead of ‘improvising’ was the prospect of giving recitals to crowded audiences … . Epideictic λόγοι were intended to be read [or, better, performed] before an audience, as a play was intended to be exhibited in the theatre.” [105]

In my view, then, the graphic and agonistic styles are not two mutually exclusive subjects with no conceptual overlap; and so, it is not the case that knowledge of one translates into one thing (namely, speaking correct Greek), while knowledge of the other, into another (namely, not having to keep silent etc.). ‘Graphic’ and ‘agonistic’ are both connected with delivery and denote stylistic emphases that point to one or another of two ends of a continuum. [106] Ignorance of either emphasis would handicap the orator’s delivery and compromise his ability to persuade his audience. Since both the graphic and the agonistic styles presuppose the use of written scripts in honing and rehearsing one’s delivery, ‘those who do not know how to write’ describe individuals whose rhetorical training is wanting: laymen who approach their speech without the benefit of schooling and formal training in the discipline of oratory. At best, such people can be expected to speak good Greek but they are not up to the task of addressing the rest of their fellow citizens, whether in the law-court or the assembly. My reading, then, makes ‘one must know both’ equivalent to ‘one must be trained in scientific delivery’: this involves knowing the rules of style that make delivery an art in the first place; and these rules include suiting the style to the kind of oratory in view, that is, choosing the stylistic emphases that make the moment of truth—the performance before the audience—successful.

Thus, I translate: ‘One must not overlook that a different style suits each genus of rhetoric. For the graphic and agonistic are not the same, nor the demegoric and dicanic. And one must know both: for knowing both is, on the one hand, to know how to speak correct Greek, and also, on the other, not to have to keep silent if one wishes to share something with the others—which [is what] those who do not know how to write experience’. Aristotle envisages a graded rhetorical skill: on the one hand, the foundation and ἀρχὴ τῆς λέξεως (1407a19), which was possibly shared by many an ordinary citizen [107] and reached to linguistic competence; on the other, a superior ability that was the province of the rhetorically trained, [108] who did not flinch from taking the initiative in political affairs, proposing and defending their own motions, and who were unafraid to prepare (and argue, as they must) their own cases in forensic settings. As observed above, classical Athens had many citizens who, though present in the assembly, were largely passive and easily manipulated by the naturally gifted or better trained ῥήτορες who made a career of their political activities. Though, as Ober (1989:106) notes, ὁ πολιτευόμενος vied with ῥήτωρ as a label for the professional political orator (in some ways, a label not far from our own ‘politician’), the free citizen’s ideal of civic life was to be fully and actively involved in the affairs of the polis. It is to that ideal that Aristotle addresses himself in enjoining rhetorical training—particularly, mastery of the use of written scripts as an aid to delivery—on those who wish to escape the status of mere silent spectators at their political assemblies. The degree of proficiency requisite for discharging the political responsibilities of the free citizen, in Aristotle’s view, is reliably available only to the one who has mastered the skill of using written scripts to train for the moment of delivery. Ober nicely captures the contrast at 1413b5–8: “Skill in public address was sine qua non for the politician. This meant not only skill at putting words together but also in putting them across” (1989:113).

The inattentive reader of the Rhetoric might think that the graphic style alone calls for writing, a belief that contradicts what I have advocated above. And the comment at 1414a17–18 might seem to lend credence to this view: ἡ μὲν οὖν ἐπιδεικτικὴ λέξις γραφικωτάτη· τὸ γὰρ ἔργον αὐτῆς ἀνάγνωσις. But the γάρ clause ‘for its peculiar work is reading’ is not primarily a statement of design, as if the philosopher thought epideictic speeches composed first and foremost for a reading public (something that may have been true of Isokrates but certainly not so of other practitioners of this genus). Rapp, who thinks that the philosopher fails to explain why the epideictic genus is the most graphic, suggests that his analysis may have been influenced by the circulation in writing of Isokrates’ speeches, which Aristotle took as heuristic models. Be that as it may, this scholar is right to observe: “Allerdings scheint Aristoteles die Schriftlichkeit nicht durchgehend als ein Merkmal der vorführenden Rede angesehen zu haben, denn in Kapitel I 9 ist davon nicht die Rede, und auch in Ⅲ 14, 1415b28f., spricht Aristoteles explizit von einem ‘Hörer’ der vorführenden Rede” (Rapp 2002:2.938). By asserting that the peculiar work of epideictic is reading, the philosopher seems to be restating the relatively lesser dependence of the epideictic style on the effective use of the voice, especially of a loud voice (1414a15–17). The employment of the common devices of style (κοινῇ 1414a28) characteristic of epideictic, e.g. how it utilizes connective particles and repetition, is contrary in tendency to the corresponding use by the more agonistic styles. This is the proper way to understand τὰ ὑποκριτικά at 1413b17: the term denotes not ‘speeches suited for delivery’ (as some suppose) but general stylistic devices and strategies (e.g. asyndeton), [109] viewed—in good Aristotelian fashion, as we should expect from my argument in this and the last chapter—as instrumental to delivery. [110] Thus, to connective particles as stylistic devices (subsumed, like other such resources, under τὰ ὑποκριτικά) correspond two delivery strategies of opposite tendency: the agonistic downplays their use and prefers relatively more frequent asyndeta, entrusting their characteristic function instead to the phrasing of the speaker’s voice; the graphic, on the contrary, makes liberal use of connectives, and for this very reason depends less for its clarity and persuasiveness on a marked vocal intonation. By reason of their respective peculiar styles (ἰδίᾳ 1414a29), [111] the script of an epideictic speech survives much better the absence of vocal performance than the script of a symbuleutic or a forensic speech: the clarity and persuasiveness of the former (always inextricably tied to each other) are relatively less affected than those of the latter two. Nonetheless, there is still something logically odd about asserting that the work of the epideictic style is reading, for stricto sensu ‘reading’ cannot be the ‘work’ of any style (what a given style does or busies itself with) in the same way, for example, that ‘to observe the cause is the work of an art’ (τέχνης ἔργον 1354a11), ‘to see the existing means of persuasion in each case’ is the business of rhetoric (1355b10–11), or ‘to know what constitution is beneficial on the basis of past history’ and ‘to know what constitutions the others have’ is the business of the art of politics (1360a37–38). Unless style is reconceived as a ‘doing’, it can neither be considered ‘an employment’ (χρῆσις) nor can anything be said to be its product (Eudemian Ethics 1219a13–18). [112] One cannot even appeal to the general definition of ‘the ἔργον of something’ as ‘its τέλος’ (Eudemian Ethics 1219a8), for it would run counter to Aristotle’s argument to think of any style (not even the most graphic of them all) as attaining its fulfillment in anything other than delivery. This is the approach adopted by Kennedy, who translates, “its objective is to be read” (1991:257). The best solution seems to me to assume that Aristotle has so deeply internalized the view of λέξις as subservient to an action—the actio that is ὑπόκρισις—that he allows for a conceptual overlap between the manner of speaking (1403b17) and the action of speaking before an audience. λέξις then becomes ‘a manner of λέγειν’, and the delivery of epideictic, as the genus most dependent on scripts because its precision calls for more extensive use of writing, naturally approaches, when it does not simply consist in, the reading of its scripts. One might even argue that the ποικιλία of style and thought characteristic of epideictic called for a more precise memorization of its script and made its commitment to memory a more critical prerequisite to delivery. To that extent, its delivery would naturally match its public reading. And, since appropriate delivery may rightly be called the telos of any given style, ἀνάγνωσις, a delivery peculiarly suitable for epideictic, can be pronounced ‘the ergon of the epideictic style’. [113]

Aristotle had remarked in 1413b14–19 that the speeches of the graphic poets, when subject to the Gedankenexperiment of delivery in an agonistic context, [114] appear thin or meager, [115] whereas those of actual orators [116] feel amateurish when read. [117] Therefore, dramatic speeches pervaded by emphases proper to the graphic style, [118] having only marginally, if at all, been written for actual on-stage delivery, would suffer little loss of impact if they were not produced and could only be experienced by the interested reader. Note that, if my understanding of the passage is correct, to illustrate the graphic style Aristotle has selected not epideictic addresses but speeches by graphic/anagnostic poets. I believe that he owes this choice not simply to his prior resort to poets, actors, and plays to illustrate the ethical and emotional styles (although this probably contributed to it). His aim would have been not to wedge apart epideictic from the other two species of rhetoric, since, after all, he viewed all three as inherently performative genera whose scripts, apart from delivery, could only imperfectly represent concrete, real-life speech-making. One must remember that ordinarily epideictic speeches were delivered in their own characteristic settings (note the mention of a ‘listener’ at 1415b29). [119] One is therefore not surprised to learn that Isokrates complained about the loss of persuasiveness experienced by his written discourses, not only because popular prejudice took for granted that written speeches were composed for self-seeking display and profit but also because ‘when a discourse is deprived of the prestige of the speaker, the sound of his voice, the variations that happen in rhetorical delivery, and, besides, of the critical timeliness and earnestness that attaches to the action, and … [when] someone reads it out unpersuasively and without investing it with character, but rather as one tallying up an account—in these circumstances it is natural, I think, that it should seem trivial to its hearers’ (Isokrates 5.26–27). [120] If even Isokrates’ epideictic writings came across as weak and ineffective when read without the requisite rhetorical vocalization, agonistic speeches would suffer comparatively greater loss and feel jejune if deprived of their intended ὑπόκρισις. It is clear, then, that Aristotle has in view the technology of writing throughout the entire chapter. This is beyond question for the graphic style, whereas with respect to the agonistic Rhetoric 1413b15–19 contemplates holding illustrative scripts of it in one’s hands and reading them aloud, [121] thus depriving the style of proper delivery (ἀφῃρημένης τῆς ὑποκρίσεως). [122] Aristotle’s analysis suggests that writing had been adopted wholesale by successful rhetorical practitioners, and he correspondingly enjoins its mastery on the rhetorical apprentice.

It is now time to inquire what the philosopher means when he assigns ἀκρίβεια to the λέξις γραφική and contrasts it with ὑπόκρισις: ἀλλ’ ὅπου μάλιστα ὑπόκρισις, ἐνταῦθα ἥκιστα ἀκρίβεια ἔνι (1414a15–16). [123] We must notice first that the terms τὸ ἀκριβές and ὑπόκρισις are subject to degrees: this is clear enough from the superlatives at 1413b8–9 (ἔστι δὲ λέξις γραφικὴ μὲν ἡ ἀκριβεστάτη, ἀγωνιστικὴ δὲ ἡ ὑποκριτικωτάτη) and the comparative at 1414a10–11 (ἡ δὲ δίκη ἀκριβέστερον); and it is not surprising, since the inverse relation in which 1414a15–17 puts ἀκρίβεια and ὑπόκρισις [124] suggests that these are two tendencies that point to the extremes of a stylistic continuum. In other words, it is not a matter of a style being ἀκριβής, ‘precise’, and thus not ὑποκριτική, ‘histrionic’ (i.e. ‘not bearing relation to delivery’), [125] or vice versa. What the philosopher is underlining through this contrast is that not all the elements and expressive strategies of style (τὰ ὑποκριτικά), however much they can be put at the service of delivery according to the occasion, exhibit the same intimate relationship and immediate connection to voice in its expressive use ‘for each pathos’ (1403b27–28). We readily understand, for example, that prose rhythm depends on the voice to a greater extent than the employment of metaphors; and though even instances of the latter, insofar as instrumental to delivery, are carefully designed and chosen to evoke apposite φαντάσματα, [126] when restricted to silent reading (to take an extreme, anachronistic case) metaphors retain much more of their natural effectiveness (their ἔργον; cf. 1413b18) than prose rhythm does. And however much the intonation that should attend the delivery of a script may be imagined by a reader and recreated in his mind’s ear, clearly this is an essentially ‘histrionic’ feature (to use Aristotle’s terminology) that, although perhaps susceptible of ‘(over)precise’ articulation in the context of the graphic, [127] finds a natural home in the agonistic style. The illustration supplied further clarifies what is in view: actors go after such effects in dramas, i.e. after what gives the greatest scope to their ὑπόκρισις (which is eminently concerned with giving expression to ēthos and pathos), whereas poets seek such characters. [128] Authors whose works are towards the graphic end of the stylistic continuum are the ἀναγνωστικοί, e.g. Khairemon, who is ‘precise as a logographer’ (1413b13).

Now follows a series of examples that are designed to show the features peculiar to either end of the stylistic continuum. Aristotle’s Gedankenexperiment consists in taking a speech out of the context to which its style is suited, and using the resulting deficiencies to highlight the corresponding strengths when it is left in its natural rhetorical environment. Hence, compared to those crafted for competitive settings, graphic speeches seem thin; [129] whereas those delivered by professional public speakers (the implication being that they must have received high praise) [130] feel amateurish when reduced to mere scripts in the hands of the lay reader. In selecting speeches by professional public speakers, Aristotle has made his illustration as pointed as possible and rendered the ἰδιωτικοί (1413b16) sharper by contrast. As explained above, τὰ ὑποκριτικά denotes the sum total of rhetorical stylistic elements and strategies. These look to delivery and its corresponding use of the voice in whatever measure is judged appropriate. Therefore, the more critically any one of them depends for its effectiveness on delivery (on vocal articulation and the expression of character and emotion), absent ὑπόκρισις and reduced to its mark on the page the more it looks silly and lacking in persuasiveness. We are offered two examples: asyndeta and the frequent repetition of the same thing, two stylistic elements relatively more suitable for the histrionic end of the spectrum and, conversely, proportionally less suited for the graphic end. [131] Hence, rhetors put them to good use in debate contexts that call for the agonistic style. Asyndeton, in particular, requires changes in ēthos and tonos (1413b30–31). These illustrations allow us to fill in the meaning of ἀκρίβεια: the precision in avoiding asyndeta lies in making the script, by itself—without giving voice to it by way of delivery—a fuller record of the meaning and attitude of the author; the connecting particles would attempt to capture something of the missing ēthos and tonos that voice would readily impart. The more ‘precise’ script, when brought to life by delivery, would feature a lesser contribution of vocal phrasing to the clarity and persuasiveness of its speech. The style of the resulting communication would be comparatively less histrionic and proportionally more graphic. Concerning the effect achieved by frequent repetition in performance, a graphic style that had infrequent recourse to it would have to state more explicitly, perhaps in the form of a direct proposition adequately expressed by suitable stylistic devices, the particular impression effected by the repetition. Hence, if in delivery the speaker accentuates another’s evil intent by piling up κλέψας, ἐξαπατήσας, and προδοῦναι ἐπιχειρήσας, the equivalent graphic statement may be reduced to the metaphoric and explicit, ‘he was bent on harming you,’ possibly aided, for effect, by other elements of style. The ἀκρίβεια, then, consists in a more complete propositional expression shaped by such stylistic elements as survive better a diminished role for ὑπόκρισις—the vocal expression of ēthos and pathos—all carefully targeted for the intended audience with a view to clarity and persuasiveness. When ὑπόκρισις is present and exploited to the full—i.e. when character is conveyed and emotion is elicited by the speaker’s full and compelling use of the volume, harmonia, and rhythm of his voice—the more prolix nature of the graphic style turns, comparatively speaking, into a stylistic burden that hinders the effective expression of ēthos and pathos of which a resourceful use of voice is capable.

Now, there is a significant corollary to the Aristotelian notion of stylistic ‘precision’: insofar as it puts less emphasis and is less dependent on what is the essence of delivery, i.e. the use of the voice for the expression of character and the evocation of emotion, it is more appropriate to situations that may be allowed to lean for their persuasiveness more decisively on the subject matter of the orator’s discourse (τὰ πράγματα). In so doing, it achieves a different balance between the possible rhetorical πίστεις. And since style/delivery often holds the potential to provoke the suspicion of the hearer, should one find himself in a situation where this suspicion is likely to arise, adopting a less histrionic style might be desirable. This, at least in part, is the situation envisaged at 1414a11–14, where a lawsuit is in the hands of one judge only.

But before I get to this, first in order of presentation is the demegoric style, which the philosopher introduces as ‘altogether like skiagraphia’. [132] Although the nature of σκιαγραφία as a painting technique is disputed, it is clear from Aristotle’s and, especially, Plato’s use of it that it corresponded to a representational technique that was intelligible only from a distance. [133] Because of its illusionistic character, σκιαγραφία was used by Plato as a metaphor for illusion and deception: a case in point is Republic 602d2–3, where Sokrates notes that it ‘falls nothing short of witchcraft’. [134] One of the central features of σκιαγραφία is a kind of mixing, [135] and for this reason Plato uses it metaphorically to criticize the fallacious commingling of what cannot (or should not) coexist as a mixture. Thus, in the Republic 583b, in the course of discussing the pleasures of three kinds of people—the φρόνιμος, the φιλότιμος, and the φιλοκερδής—we are told that ‘excepting the man of intelligence, the pleasure of the rest is neither altogether true nor pure (καθαρά), but is a kind of shadow painting (ἐσκιαγραφημένη)’ (583b3–5). The word καθαρός reappears at 586a6, where Sokrates says that those who have no experience of φρόνησις and ἀρετή have never tasted pleasure that is stable and pure (καθαρᾶς ἡδονῆς); to which he adds the question: ‘Are not the pleasures they live with of necessity mixed with pains, phantoms of true pleasure and shadow painted (ἐσκιαγραφημέναις), so colored by juxtaposition that either kind seems intense … ?’ (586b7–c1). We see that the opposition ‘pure’ vs. ‘mixed’ is central to the use to which Plato puts σκιαγραφία. So it does not come as a surprise that at Rhetoric 1414a13–14 Aristotle, in making his own argument, should also bring it forth while contrasting the demegoric style with a judgment that is pure (καθαρὰ ἡ κρίσις).

The demegoric style is declared to be entirely like skiagraphia: the greater the crowd, the more distant the view. At first one might consider taking θέα in its literal sense: on average, a large audience necessarily separates speaker from listener so that even a clear hearing of what is being said becomes difficult. [136] But this cannot have been the philosopher’s point: for, in this respect, the forensic setting would have hardly provided the orator with a much better public environment, as juries sometimes numbered in the hundreds. And if this were its meaning, the comment would invert the implication of shadow painting, i.e. the growing unintelligibility of what is viewed as the spectator gets close. If the style of the assembly speaker were skiagraphic not by design but as a consequence of the limitations of his setting—the distance between him and his audience that prevents them from hearing him clearly—then, at least in principle, one could remedy the deficiency without a change in the manner of speaking simply by having them approach each other. This suggests (the possibility of) increasingly detailed intelligibility in proportion as the intervening distance is abridged. But this is precisely the opposite of what the metaphor of a stylistic σκιαγραφία implies. The literal understanding of θέα also faces the objection that, if true, the philosopher would have paralleled ‘viewing’ with hearing, since according to him voice, not sight, is central to delivery. [137] To this objection, one might reply that the metaphor dictates its own terms and is responsible for the lack of a precise fit. But there is a better approach: Aristotle’s observation is not a veiled criticism of a style that should, but due to its setting cannot, be more precise. Rather, σκιαγραφία is a compelling metaphor that serves to specify the degree of precision peculiar to the demegoric style. It also helps the philosopher to explain why this style’s skiagraphic lack of precision actually suits the rhetorical context for which it has been formulated. Aristotle’s endorsement of this stylistic design is signaled by the use of χείρω at 1414a10 in connection with τὰ ἀκριβῆ (see below). This confirms his belief in its appropriateness for the deliberative performance setting. We must therefore take θέα in an extended sense: the statement concerns ‘a distant [rhetorical] view’, a notion that hints at, and is the consequence of, the various and potentially divergent views of each listener in the assembly. By assimilating the delivery style, of which the speaker is the efficient cause, to a σκιαγραφία, the text commends a perspectival inversion with the hearers as viewers and the speech as the object they observe: each member of the audience has his own ‘viewing’, θέα, and all these θέαι are ‘distant’. But distant from what? I submit that they are potentially distant from each other and from the speaker’s own θέα and that they represent mutually divergent viewpoints. πορρώτερον ἡ θέα (1414a9) regards the expectation that public opinion in the deliberative assembly will grow increasingly variegated with crowd size. Divergence of views in the democratic decision-making process is the challenge that calls for rhetorical consensus building. One may equate each θέα with each citizen’s δόξα, the individual opinion that the demegoric speaker is attempting to influence and draw closer to his own. The assembly setting requires his appeal to be as broad as possible, with its breadth corresponding to the size of his target. His approach should seek to circumscribe the complexity of the issue under discussion to the possibilities of a mixed deliberating body. Deliberative matters are inherently more difficult to discuss, because they must cope with uncertain outcomes and future circumstances [138] and because they cannot be treated with the intellectual precision of a philosophical argument: [139] not only are the abilities present in a mixed crowd widely variable [140] but the orator cannot possibly furnish his delivery with carefully crafted φαντάσματα that will seem persuasive to each and every one of his listeners. [141] Hence, there are limitations of precision regarding ‘persuasion through the subject matter’ and ‘persuasion through the hearer’, [142] limitations that must translate into a corresponding style that is artfully imprecise. Imprecise as to subject matter: with enough definition to capture the basic features of a particular issue, yet vague enough to do so in bold strokes that avoid philosophical niceties incomprehensible to men of average ability. Imprecise as to delivery: with such stylistic expression of ēthos and pathos as will have broad appeal to the promiscuous crowd that attends to the hearing, sidestepping private prejudice and building on common sentiment. [143] Hence, it is apparent that both in σκιαγραφία and in the demegoric style τὰ ἀκριβῆ are not only περίεργα but positively χείρω. [144] Paradoxically perhaps but with striking faithfulness to the guiding metaphor, the perspectival distance actually becomes instrumental to the clarity and persuasiveness of the demegoric style. Rhetorical style and rhetorical setting coexist in perfect symbiosis. [145]

Moving on to the forensic setting, one is struck by the fact that Aristotle does not compare ἡ δημηγορική with ἡ δικανική (the emendations of some editors notwithstanding) but with ἡ δική. [146] This proves that he is concerned with the sort of style that will suit not only the hearer peculiar to either setting (the κριτής can either be an ἐκκλησιαστής or a δικαστής) [147] but also, as I have argued above, with the corresponding subject matter: in the case of symbuleutic rhetoric, τὸ συμφέρον καὶ βλαβερόν (Rhetoric 1358b22). One can readily understand the philosopher’s point: justice, insofar as it surveys the past and calls for a determination upon it, is less uncertain than the issues of policy that dominate deliberative assemblies. If no further considerations attended the analysis, this alone would justify the implication that the dicanic style, called to suit an inherently more precise subject matter, should reflect this precision in its own precision: with metaphors that have narrow targets, with a comparatively greater abundance of conjunctions and connectives that draw out more explicitly the logic of the argument, featuring φαντάσματα that are more acutely crafted as to their impact on ēthos and pathos, etc. Greater precision per se, it must be emphasized, calls for neither more nor fewer stylistic devices and strategies (although it may call for more or less frequent resort to a given device or strategy). One must eschew the notion, often presupposed by scholars, that ἀκρίβεια entails ‘more of style’ (it may well require restrictions, as when it calls for concision). [148] Since Aristotle’s notion of λέξις is circumscribed by ὑπόκρισις, the rhetorical practice it subserves, the correct view is that a precise delivery calls for a precise style and vice versa. The inverse proportion in which the philosopher puts ἀκρίβεια and ὑπόκρισις (1414a15–16) must be understood not with regard to ὑπόκρισις as a rhetorical practice—which, as such, cannot be quantified but may be qualified (in particular, it can be more or less precise according as its stylistic means are more or less exact)—but in relation to vocal expression (loudness, harmonia, and rhythm) as the material cause of the homonymous technical practice. Because the degree to which character and emotion are involved as rhetorical πίστεις—I say ‘the degree to which,’ not ‘the fact that,’ because they always are—is in turn proportional to the greater or lesser use of the voice, ἀκρίβεια is also inversely proportional to the stylistic prominence of ēthos and pathos.

If, as noted above, the dicanic style should match its relatively more precise subject matter with a correspondingly more precise style, this is so a fortiori (ἔτι δὲ μᾶλλον) when justice is in the hand of one judge alone. [149] Then there is no debate (ἀγών) between, nor even a gathering of, δικασταί who hold a variety of opinions; [150] then the κρίσις is ‘pure’ (καθαρά) because the skiagraphic features of deliberative settings, with their multiplicity of views (translated by the painting metaphor into distance), are absent when there is but a jury of one. ‘For [then the trial] is least in rhetorical devices’: [151] it is not that rhetoric per se and, a fortiori, delivery are inconsequential, but that, focused on one single individual, the scope is the narrowest possible and, for success, the orator must fine-tune every aspect of his speech (including style) to his one-member audience and to the subject matter at hand. [152] Here, Aristotle is being true to life, for we know from experience that a style (including, say, lavish praise) that may be acceptable in addressing a large audience may meet with suspicion if focused on one individual. When one man alone is the target of persuasion, the danger is great that he will resist the thrust of the speaker’s rhetoric; for being self-conscious of, and sensitive to, the fact that the presentation is crafted for him, this invites him to second-guess what is said and whether the appeal to him to embrace a particular opinion is legitimate or not. [153] Consequently, the orator who would persuade one man must narrow the artifice of his delivery so as not to raise his natural suspicion; just as he ought to bear in mind his one listener as he tries to craft stylistic φαντάσματα that hold the greatest promise of persuasion. The use of εὐσύνοπτον (1414a12) is particularly apposite, picking up again on the metaphor of a σκιαγραφία: demegoric settings involve distant views that blur the precise correspondence that might otherwise obtain between the subject matter of the speech and the stylistic design of the speaker; conversely, trial by a single judge allows the orator to see more readily what in his stylistic rhetorical kit (ἐν ῥητορικοῖς) is οἰκεῖον and ἀλλότριον to the subject matter. [154] The usual reading, ‘[what is] proper and foreign to the subject matter’ is acceptable (οἰκεῖον may govern the genitive or the dative). But this determines neither the sphere from which τὸ οἰκεῖον and τὸ ἀλλότριον are taken; nor the standard by which, and the point of view according to which, whatever belongs to either adjective is pronounced ‘proper’ or ‘foreign’ (as the case might be). I join the consensus in understanding the general sphere to be the presentation of the speaker, i.e. his rhetorical practice, including style. But I depart from the consensus in my choice of viewpoint and standard: it is not the opinion of the single κριτής about the closeness of the fit between the speaker’s rhetorical practice and his subject matter. Instead, it is the speaker’s own judgment about the appropriateness of his ῥητορικά—given the context, chiefly apprehended under the head of style—to his one-man audience. Because for Aristotle stylistic design regards the fit in delivery between the subject matter and the audience (how to say what must be said with a view to clarity and persuasion), the speaker’s judgment of pertinence and impertinence becomes one of closeness of stylistic fit between the subject matter (which ‘clarity’ regards) and audience (which stylistic persuasion, i.e. persuasion through ēthos and pathos, regards). This fit is what the philosopher calls ἀκρίβεια.

It is perhaps helpful to underline again that style plays an instrumental function to delivery and, as the rhetorical practice that regards ‘how one must speak’ (ὡς δεῖ εἰπεῖν 1403b17), it mediates the translation of subject matter (ἃ δεῖ λέγειν 1403b16) into actual vocal performance before an audience. Because style subserves delivery and the end of delivery is suasion with the use of the voice through character and emotion, proper use of style must consider the fit between the devices and strategies that shape the manner of speaking and what makes for persuasion (including the sine qua non of clarity) for the particular kind of audience addressed in a given setting (1356b26–34). This aspect of style, i.e. how closely it fits the target audience, most immediately regards its ψυχαγωγία. This ψυχαγωγία is circumscribed by the occasion: it not only looks to the individual hearer but also to the pragmatics of the communication. What is useful and clear for a given individual in the context of an ἐπίδειξις may be counterproductive and obfuscating to him in the more agonistic symbuleutic address. But this is not all. For style looks not only to target audience and performance occasion but also to subject matter, and with respect to the latter there is a corresponding fit between the manner of expression and the subject at hand. Stylistic ‘precision’, ἀκρίβεια, trades on both ends of this dual perspective, i.e. on how closely style fits both its subject matter and its target audience at a given occasion. Any failure on either end of this stylistic communicative mediation incurs a loss of persuasiveness. A speaker may owe this failure either to excessive or to defective precision: in an agonistic context, where the audience has a hard time following the logic in the heat of the moment, a style whose precision corresponds to an expansive argument compromises clarity and hence persuasiveness. One might say that argumentative fit—where the argument is detailed and expansive—compromises audience fit. In such contexts, compelling speakers adopt stylistic practices that put greater emphasis on character and emotion through the use of the voice, and the stylistic devices that subserve this greater prominence of vocal delivery are such as to commend a congruent simplification and selectivity of argumentative πράγματα. Given style’s dual mediation between subject matter and situated audience, a speaker may place a comparatively greater stylistic emphasis on persuasion through the facts themselves or on persuasion through character and emotion. [155] The former relies less on the voice, the latter, more. The former presupposes an audience with a higher capacity to follow oral argument or a context that allows for its more reposed and effective intellection; the latter relies instead on, and seeks to exploit, the hearer’s settled opinions and attitudes, his preexisting interests and commitments (emotional and intellectual), his underlying understanding and experience of the world—culture and society—in which he lives. Insofar as these constitute a more basic and reliable ‘common denominator’ for the audience, the agonistic style addresses itself more to these latter considerations and refers to them for guidance. On the other hand, typically (and perhaps ideally) the graphic style has a more generic audience in view, one that is more equanimous in the absence of passion and polemic. This allows this style to place a relatively greater emphasis on the closeness with which it fits the subject matter. In fact, the epideictic setting largely allowed speakers to regard, and even to construct, more uniform ideal audiences with reference to generic cultural and social stereotypes (patriotism, Panhellenism, φιλοτιμία, etc.). Because from the start they could select and shape their subject matter with regard to these generic stereotypes, epideictic made fewer and less stringent demands on the persuasive appeal through character and emotion effected by style in delivery. This afforded epideictic a potential for ἀκρίβεια greater than that enjoyed by the other two rhetorical genera. The inherently dual perspective of style as a rhetorical practice is responsible for Aristotle’s shifts in emphasis between form and content, shifts which often lead modern scholars to think that he has confused two different senses of ἀκρίβεια: stylistic precision and precision of theme or argument. [156] This twofold emphasis is not the outcome of confused thinking. It is a faithful reflection of the Aristotelian conception of style and of what this conception in turn entails for stylistic precision, i.e. for the fit it effects between the ‘what to say’ and the ‘how to say it’ with a view to persuading a given situated audience in the vocal performance of delivery. But, nota bene, even in its regard for subject matter ‘style’ continues to look to persuasion through ēthos and pathos by the speaker’s use of his voice in delivery. In other words, although ‘style’ as a sort of ‘formal cause’ takes the obligatory and inevitable account of ‘τὰ πράγματα’ as the rhetorical ‘material cause,’ it is ‘persuasion’ (the ‘final cause’) through ‘vocal delivery’ (the ‘efficient cause’) that drives Aristotle’s analysis. The relation of stylistic precision to target audience is a particular manifestation of what Aristotle considered in Posterior Analytics 87a31–b7 in a wider context: ‘A science is more precise than another … if it depends on fewer items and the other on an additional posit’ (ἀκριβεστέρα δ’ ἐπιστήμη ἐπιστήμης … ἡ ἐξ ἐλαττόνων τῆς ἐκ προσθέσεως). [157] The opposite of an ‘additional posit’ (πρόσθεσις) is ‘abstraction’ (ἀφαίρεσις), and it is precisely through abstraction that epideictic constructs its public, whereas the agonistic genres circumscribe their subject matter and deliver it with rhetorical stylistic brushstrokes that will appeal to and sway a majority of their diverse audiences. [158]

The alternative understanding of dicanic stylistic precision proposed above is to be preferred to the consensus on contextual grounds. Aristotle’s statements on the matter are meant to illustrate the greater precision of the dicanic style that is suitable for a trial by one judge (relative to the dicanic style employed before a numerous audience). For this reason, we expect him to explain why such a trial promotes greater precision in the speaker’s stylistic design, not why a single judge is more capable of discerning when the speaker talks off subject (ἔξω τοῦ πράγματος), [159] whether he be assumed to draw on extraneous subject matter, on emotional appeals that bear no relation to, or even invert, the claims of truth, or to adopt yet some other diverting strategy. This view is to be rejected not only because there are no grounds for the claim that one judge, merely because he is one, is better able than a plurality to detect inartistic rhetorical devices that offend truth and right [160] but also because it suggests, if it does not entail, a displacement of the argument’s center of gravity away from stylistic precision (ἡ ἀκρίβεια τῆς λέξεως) and towards πίστεις through αὐτὰ τὰ πράγματα (1403b19). Although one might still evade the second objection by insisting on viewing τὸ οἰκεῖον τοῦ πράγματος in relation to style (the one judge more readily sees what in the stylistic practice of the speaker properly belongs to the subject matter), the only way to make this productive of greater accuracy of style is to translate the judge’s ability into the speaker’s self-policing of his stylistic practice on this account. But it is implausible to assume that the judge’s temperament, capacity, and whatever else makes εὐσύνοπτον predicable of him will ordinarily translate into such temperament, capacity, etc. in the speaker as will enable him, under the judge’s careful watch, to modify his stylistic practice accordingly. A similar flaw does not attach to my reading, which offers the objective fact of the single-member audience and, hence, the absence of a multiplicity of divergent views that would otherwise blur the stylistic practice of the speaker (with a view to persuasion) as the reason for the closer fit that λέξις effects in delivery between τὰ πράγματα and the single hearer. [161] τὸ οἰκεῖον τοῦ πράγματος καὶ τὸ ἀλλότριον is not ‘what in the speaker’s delivery is on or off subject’ but ‘what pertains (οἰκεῖον) or does not pertain (ἀλλότριον) to his situated delivery of the subject matter before the single judge’. [162] This comment recalls Aristotle’s observation in Rhetoric I.1.10: because deliberative rhetoric regards more general (i.e. public) interest, it lends itself less to speaking off subject than forensic. [163] In the assembly the κριτής judges about affairs that are pertinent to him (περὶ οἰκείων), whereas in a trial the decision is about the affairs of others (περὶ τῶν ἀλλοτρίων γὰρ ἡ κρίσις). The dicast is therefore liable to look to his own interests (πρὸς τὸ αὑτῶν) and to listen with partiality; and, after being won over by the appeal of the litigants, to give himself to them rather than render true judgment. [164] To sum up: the perspectival horizon of εὐσύνοπτον is best construed not as that of the one judge but the litigants’: since he is only one, the identity and range of τὸ οἰκεῖον τοῦ πράγματος is correspondingly narrow and well defined. τὸ οἰκεῖον τοῦ πράγματος regards the manner in which the facts of the case (τὸ πράγμα) should be translated at the time of delivery with a view to clarity and persuasion into audience-directed stylistic practices. In contrast to the perspectival narrowness of the one-judge trial, before a crowd of dicasts τὸ οἰκεῖον and τὸ ἀλλότριον, translated into style for effective delivery, would be compounded and multiplied by a variegated patchwork of views, interests, and indifference. The speech, both style and subject matter in the intimacy of their mutual rhetorical connection, would be blurred into an oratorical skiagraphia; then it would be advisable for the litigants, in crafting their speeches as to form and content, to tailor them broadly with a view to moving the entire jury towards the desirable δόξα. In such a situation, the deployment of rhetoric’s full resources is all the more critical, and success depends to a larger extent on the level of the speaker’s training in the oratorical art. In the opposite case, when directed to a single judge, the scope of rhetoric is the least.


[ back ] 1. Alternatively, with ἔστι as ‘it is possible’ governing the complementary εἶναι (Bonitz 1870:220 col. 2 lines 6–8) and τό not as the article of the infinitival clause τὸ ὑποκριτικὸν εἶναι but, more narrowly, as forming the abstract τὸ ὑποκριτικόν (cf., for example, Poetics 1453b1–2): ‘and it is possible for the ability to perform to be a natural talent and [as such], less technical, but in connection with lexis [it is possible for it to be] technical’. With the grammar thus construed φύσεως is coordinated with ἀτεχνότερον (φύσεως … καὶ ἀτεχνότερον) and one should not print a comma before the conjunction καί. I actually prefer this translation, which provides a more compelling fit for the context and more naturally accounts for the infinitive εἶναι: if Aristotle had merely intended to say that ‘a capacity for acting is natural’, why not simply write καὶ ἔστι φύσεως τὸ ὑποκριτικόν? At any rate, since I do not want to give the impression that the validity of my argument depends on my preferred translation, I have followed the more common ‘and for someone to be hypokritikos is of nature’.

[ back ] 2. Although not office holders per se, Athenian ῥήτορες were readily identified in the fourth century BC as a well-defined group with an important institutional role. To use a familiar label (with the proper adjustments; cf. Hansen 1983a), they were ‘politicians’ who consistently engaged the democratic process as public speakers. Hence one may fairly call their rhetorical activities ‘professional,’ whether they owed their oratorical skills to technical training or to untrained talent. Rhetoric Ⅱ.11 1388b15–18 conveniently offers Aristotle’s own estimate of ῥήτορες: ‘Those are emulated who possess these and similar things. And these things are the ones [already] mentioned, viz. courage, wisdom, and office (ἀρχή; cf. 1388b5); for rulers (ἄρχοντες) can do good to many, [I mean] generals, rhētores, and all those who can do such things’. For ῥήτορες as ‘politicians’ and hence ‘professional’ public speakers, see Perlman 1963:354–355, Hansen 1981 esp. 368–370, Hansen 1983a esp. 46–48, Hansen 1983b, Hansen 1987, Ober 1989:104–127, Schiappa 1999:68, and Worthington 2007. Epigraphically, ῥήτωρ is first attested in IG I3 46a.25, dated by the editors to ca. 445 BC (cf. Meiggs and Lewis 1969:128–133 no. 49).

[ back ] 3. This translation draws on my analysis above (see in particular §13.3) and immediately below. To interpret this difficult passage correctly one must observe the following guidelines. Aristotle is primarily concerned here with exploring the technical side of the capacity for delivery (this follows from correctly identifying the referent of ἐκείνη). Therefore, in ‘διὸ καὶ … πάλιν … , καθάπερ … ’, structured as ‘therefore also A in turn, just as B’, a comparison that explains A in terms of a better known B—such is the logic of comparisons (cf. K-G Ⅱ.2.490 §579)—one would expect A to regard technical delivery (the new consideration thitherto only to a small extent the subject of scholarly reflection) and B, non-technical talent (the obvious, broadly accepted fact). That B refers to a non-technical talent for delivery as the better known member of the comparison (and hence the one following καθάπερ) should not be controversial: whatever competitions (formal or informal) Aristotle has in mind when he mentions ἆθλα, it should be obvious that by and large no technical training in delivery would have assisted the majority of the competitors. Otherwise, Aristotle could not claim that delivery had only received limited and trifling technical attention before him. This means that the ῥήτορες who are defined as κατὰ ὑπόκρισιν must be professional speakers who have become such in consequence of their natural talent. This interpretive approach is supported by the use of τοῦτο, which more plausibly refers to the closer clause, περὶ δὲ τὴν λέξιν ἔντεχνον, than to the more distant, ἔστι φύσεως τὸ ὑποκριτικὸν εἶναι, καὶ ἀτεχνότερον. Aristotle could have made the opposite choice clear with the demonstrative ἐκεῖνο. A final consideration: the γάρ sentence that closes the quotation most naturally applies to the comparison as a whole (‘therefore also A in turn, just as B’) rather than to its καθάπερ clause alone. οἱ γραφόμενοι λόγοι, entailing as it does the technical notion of lexis, cannot plausibly apply to a natural talent for delivery; hence, if one takes the statement in its most natural sense, i.e. as explaining the comparison as a whole and, by implication, the explicandum διὸ καὶ τοῖς τοῦτο δυναμένοις γίγνεται πάλιν ἆθλα, the γάρ clause further supports the view that τοῦτο points to τὸ ἔντεχνον ὑποκριτικόν (=τὸ περὶ τὴν λέξιν ὑποκριτικόν or ἡ περὶ τὴν λέξιν ὑπόκρισις), and not to the ἡ ἄτεχνος ὑπόκρισις regarded in οἱ κατὰ τὴν [ἄτεχνον] ὑπόκρισιν ῥήτορες.

[ back ] 4. See also above, §13.3. Another option, but less likely, is τέχνη, which appears at 1403b35 in the sense of ‘treatise’ (note the σύγκειται). But here it would not be acceptable in that sense, for ἐκείνη is picked up by περὶ αὐτῆς, and ἐγκεχειρήκασι δὲ ἐπ’ ὀλίγον περὶ αὐτῆς εἰπεῖν τινες clearly calls for a subject, not a treatise. τέχνη would be possible, however, in the sense of ‘art’ or ‘principles that guide a professional practice’, which by itself would fail to resolve whether λέξις or ὑπόκρισις or, as I argue, a combination of both is in view.

[ back ] 5. On αὕτη versus αὐτή see above, §13.2 n. 19.

[ back ] 6. The syntactic agreement by itself would not be decisive, as a superficial substitution of ὑπόκρισις by τοῦτο, without any change of the notional subject—a common occurrence—might be in view. But τὸ μὲν οὖν τῆς λέξεως is clearly intended to summarize and drive home the implications of the preceding passage, i.e. 1403b36–1404a8, and this suffices to show that the broader τὸ περὶ τὴν λέξιν, a topic neither limited to nor exclusive of ὑπόκρισις, is its subject matter.

[ back ] 7. τραγική, as its coordination with ῥαψῳδία shows, does not in the first instance refer to tragica ars but to actio tragoediarum. Nevertheless, 1403b24–26 assumes a natural connection between dramatic ὑπόκρισις and tragic ποιητική, and from it infers that a similar relation obtains for ἡ ῥητορική. The parallel might be closely drawn: so long as oratory remained chiefly the province of untrained, well-endowed public speakers, little explicit attention was paid to delivery; once numerous citizens of middling talent had need of it and, on this account, sought some form of training, rhetorical instruction began to focus on the rules that made for successful delivery.

[ back ] 8. ‘When delivery comes to rhetoric’ would mean ‘when the study of delivery comes into vogue’ and orators apply their technical knowledge of it regularly and effectively.

[ back ] 9. There is no difficulty, of course, in having a feminine demonstrative for what has been spoken of thus far periphrastically in the neuter as τὸ περὶ τὴν λέξιν or τὸ τῆς λέξεως, for the notional subject matter all along is the feminine λέξις. Cf. the use of feminine adjectives at 1403b29–30, notionally in agreement with φωνή, even though the preceding clause speaks of τοῖς τόνοις. Or the use of αὕτη for ὑπόκρισις at 1403b27, which until then has been referred to by a neuter plural (τὰ περὶ τὴν ὑπόκρισιν 1403b22) or the neuter singular (τὸ τοιοῦτον … · ὅπερ 1403b25–26). One might also wonder why ἐκείνη is preferred to αὕτη. The reason is not, as one might think, to direct the reader’s attention away from λέξις to the further removed ὑπόκρισις; the distance felt is rather that between the demonstrative and λέξις, which appears four lines earlier and is separated from ἐκείνη by an explanation, a qualification (with the feminine φαντασία), and an example, none of which contains or refers to ‘style’ at the syntactic level.

[ back ] 10. On the use of ‘rhetorical’ to qualify ‘style’, see immediately below, §14.1.

[ back ] 11. Rapp 2002:2.817–818 writes: “Tatsächlich werden im vorliegenden Abschnitt die Belange der sprachlichen Form (λέξις) mit denen des mündlichen (und vom Schauspieler verlangten) Vortrags verglichen. Daher liegt nichts näher, als unter ‘ὑπόκρισις’ dieselbe Kunst oder Fähigkeit zu verstehen, die schon in 1403b22 damit bezeichnet wurde, und ‘jene’ auf die sprachliche Form (λέξις) zu beziehen; letzteres ist schon deswegen angeraten, weil bereits im vorausgegangenen Abschnitt 1404a8–12 von der Beachtung der sprachlichen Form und nicht von der Vortragskunst die Rede war.” Although I cannot agree with him that τῇ ὑποκριτικῇ in 1404a13 (which Rapp rewords as ‘ὑπόκρισις’) regards oratorical delivery, he is doubtless right to observe that the context points to lexis as the referent of ἐκείνη.

[ back ] 12. Schloemann 2000: “Hier ist der Begriff [sc. ὑπόκρισις] eng mit dem der λέξις verbunden und hat eine viel stärkere ‘technische’ Konnotation: er ist auf künstlerische Gestaltung bezogen” (211).

[ back ] 13. The ‘coming of ἐκείνη’ would then signal the arrival of ‘scientific delivery,’ whether we designate it by ‘delivery connected with style’ or ‘style with a view to (or according to) delivery’. For overviews of rhetorical actio see Wöhrle 1990 and Steinbrink 1992.

[ back ] 14. This interpretation is further supported by the lines that follow (1404a18–26), for once again style and delivery are mentioned in close succession. At 1404a18–19 written speeches are said to owe their effect more to style (λέξις) than to argument (διάνοια); after which Aristotle immediately launches into a historical survey in which the mimetic quality of φωνή plays a fundamental role and is credited with the origin of the arts of rhapsodic performance, acting (ὑποκριτική), and others—only to return immediately thereafter once more to poets as pioneers of λέξις.

[ back ] 15. For a study of the intersection of rhetoric and poetics in Aristotle’s treatment of λέξις, see Ricoeur 1996. There, I find much I would qualify or disagree with, but I subscribe in substance the central insight that the Poetics and the Rhetoric structure the field of style in ways that are distinctive and reflect their various aims. Thus, for example, observing that in the former Aristotle rejects the analysis of λέξις under “modes of speech” (i.e. illocutionary forms of speech), he writes: “Hardly has this line of analysis been alluded to when it is interrupted by the remark: ‘Let us pass over this, then, as appertaining to another art …’ (1456b19). This other ‘art’ can only be rhetoric” (328). Ricoeur thinks that Aristotle applies himself to “clos[ing] the gap, if not fill[ing] a void” between the theory of λέξις in poetry and a “truly rhetorical theory of lexis” (343). I must, however, part company with him when, following the consensus view, he asserts that, as a how, ὑπόκρισις is further removed than λέξις from the what of a speech (343): perhaps so from our perspective, but only true of Aristotle in a limited sense, for ultimately the entire art of rhetoric is oriented towards effective and successful delivery. Valid logical distinctions cannot obscure necessary practical connections; yes, “it is through our lexis that we teach” (344), but in fifth- and fourth-century Greece it is preeminently in oral delivery that the act of teaching takes place. For this reason I reject his interpretation of, and rationale for, Rhetoric 1404a18–19: “To the degree that style is the external manifestation of discourse, it tends to separate the concern to ‘please’ from that of ‘arguing.’ It is doubtless because writing constitutes a second degree of exteriorization that the separation is particularly dangerous in this case” (344).

[ back ] 16. Cf. similar pronouncements regarding metaphors in Poetics 1459a6–7 and Rhetoric 1405a9–10 (with the note in Cope 1877:21–22 ad loc.); and yet no one would deny that much effort is expended in teaching their proper use in accordance with the principle that art improves nature. Cf. Newman 1998:76n20, Newman 2005:46n20, and Burkett 2011:91–92.

[ back ] 17. Namely, Roberts (in Ross 1924), Jebb (in Sandys 1909), and Rapp 2002. On the other side of the divide we find the old Latin translation (apud Spengel 1867), Cope 1877, Freese 1926, Tovar 1953, Dufour and Wartelle 1973, and Kennedy 1991 (cf. Arnhart 1981:165).

[ back ] 18. Or τὸ δὲ περὶ τὴν λέξιν ἔντεχνον, which is nowhere attested. And while one might wonder whether τό may have fallen out, the placement of δέ, between περί and τήν, rules this out (cf., for example, Aristotle De generatione animalium 715a14 and Historia animalium 491b21, 493a28, 511a27, 597a32; there are no instances of τὸ περὶ δέ). In point of fact, neither does the context recommend adding it, for the text makes good sense as is.

[ back ] 19. See above, §14.1 n. 3. Interpreters have failed to grasp the distinction Aristotle is drawing here. Had he merely wanted to say “just as [prizes go] to orators on the basis of their delivery” (Kennedy’s translation, which is representative of the rest), he would have written instead καθάπερ καὶ τοῖς ῥήτορσιν κατὰ τὴν ὑπόκρισιν. In the attributive position κατὰ τὴν ὑπόκρισιν defines these orators (it could be otherwise had Aristotle used a participle, e.g. λέγουσιν, which the prepositional clause could then modify adverbially); and if Aristotle meant to ascribe to them superior delivery, adding this qualifier was a very odd way indeed of doing so. In fact, it is hard to see how κατά by itself could bear this meaning, while the slightly modified τοῖς κατὰ τὴν ὑπόκρισιν διαφέρουσιν ῥήτορσιν would have made this plain. Once we accept the implications of the syntax and reflect on the context, it becomes clear that in this expression ὑπόκρισις must stand for τὸ ὑποκριτικόν as the natural talent that is ἄτεχνον and bears no relation to λέξις.

[ back ] 20. Fortenbaugh (1985:287n29) interprets the contrast incorrectly. He is right, I think, to reject the translation in Sonkowsky (1959:261) that opposes “actors with histrionic ability” to “orators who excel in delivery” (this would invert the terms of the comparison, which should rather have had the καθάπερ go not with ῥήτορες but with οἱ τοῦτο δυνάμενοι, i.e. the alleged actors); but thinking that this chapter is an introduction to style, which he divorces from delivery, he concludes that the philosopher is only trying to show that “style, too, conveys power and produces prizes.” On this reading, the γάρ sentence that follows would make the point by noting that written texts owe their power more to style than to thought (for my own notions about this, see below, §14.2). This interpretation would be allowable if Aristotle had written ἡ δὲ λέξις ἔντεχνος at 1404a16; as it stands, the text will not permit τοῦτο to refer to style and will therefore not support an alleged literary view of style as embodied by written compositions. Fortenbaugh is otherwise correct in contrasting τὸ [φύσει] ὑποκριτικόν, ‘a natural gift for delivery’ (“orators gifted in delivery”) with ‘style’—not just any concept of style, but ἡ κατὰ λέξιν ὑπόκρισις, the choice that Cope rewords “ὑποκριτικὴ κατὰ λέξιν” and Fortenbaugh dismisses as “second best but … at least consistent with Aristotle’s focus on style.” Why it is “second best” remains unexplained.

[ back ] 21. The expression recalls Rhetoric 1408b21 and b28–29, where it is used specifically of rhythm. Though in the Rhetoric its semantic range is much narrower than the one we encounter in the Poetics, both occurrences share the performative emphasis.

[ back ] 22. Poetics 1456b10–11. Presumably the ‘architect’ of acting displays a technical knowledge of his field (rationale, principles, and application), as opposed to actors who, having learned their trade by trial and error, might rely for their guidance on experience. This probably relates to the observation at Rhetoric 1403b22–24.

[ back ] 23. οἷον τί ἐντολὴ καὶ τί εὐχὴ καὶ διήγησις καὶ ἀπειλὴ καὶ ἐρώτησις καὶ ἀπόκρισις καὶ εἴ τι ἄλλο τοιοῦτον (1456b11–13). Cf. 1457a21–22.

[ back ] 24. This is not do deny Twining’s obvious comment that “[of] Homer’s pronunciation or action … Protagoras knew nothing” (1812:2.251, his emphasis). All that is implied in the sophist’s censure is an incorrect use of the imperative, of which there is an echo in the scholia of the Venetus A ad loc. (Erbse 1969–1988:1.5.41–42 §d): ὅτι κατὰ τὴν ποιητικὴν ἤτοι ἄδειαν ἢ συνήθειαν λαμβάνει τὰ προστακτικὰ ἀντὶ εὐκτικῶν. It is wrong, however, to interpret these performative settings strictly as σχήματα διανοίας (e.g. in Quintilian Institutio oratoria 6.3.70) and suppress in turn the ‘manner of expression’ (but cf. Twining 1812:2.253). Note Lucas’s humane observation: “The point raised here probably arose from [Protagoras’] perception that what we call the imperative mood was associated with the tone of command … . Those who discover new principles often try to apply them too rigidly” (Lucas 1968:198). On the notion and history of σχήματα διανοίας see Lausberg 1998:271–273 §§600–603, 335–410 §§755–910.

[ back ] 25. Paradoxically, what makes such distance possible is Aristotle’s belief in the inherent expressive power of διάνοια, i.e. of the argument articulated by λέξις/λόγος (1456a36–37; cf. 1462a11–13) that allows a reader to recreate the effect of performance (or at least to ‘replay’ the events of the plot, 1453b1–6) in his mind’s eye. Note that the philosopher readily makes even the articulation of emotions the purview of λόγος (1456b5–8; see immediately below for an important qualification), although whatever de-emphasis on delivery this entails is partially counterbalanced by his pairing ἀνάγνωσις and τὰ ἔργα at 1462a17–18 (I adopt Madius’s preferred reading over against the vulgate ἀναγνωρίσει: “ideo legendum est ἀναγνώσει, ut codices aliqui calamo exarati Venetiis habent” [Madius and Lombardus 1550:297]). To be sure, the distance between λόγος in performance and διάνοια as the verbal expression of the argument of a play abstracted from actio (including pronuntiatio and rhetorical elocutio) only makes sense where the text under discussion is notionally, if not actually, independent of the performer, i.e. viewed as a script that can be interpreted with a varying degree of competence. Then one may rightly question how successfully the performer has brought out the expressive potential of the text or acted out what its author intended. This argues for a significant degree of fixation already by Aristotle’s time of a standard Athenian version of the Iliad, though it need not rule out a comparatively small degree of mouvance. For one, the philosopher’s focus is not on epic but on drama, whose authorship and textual status are far more straightforward than those of the Homeric poems; and yet we know that even the text of Athenian tragedies experienced so-called ‘textual interpolations,’ which, some scholars will admit, are of histrionic origin (cf., for example, Page 1934; see also above, §13.5 n. 42). Moreover, the example selected by Aristotle, Iliad 1.1, is likely to have been among the earliest passages to attain fixity; the rhapsode’s irrelevance (narrowly conceived) to the textual status of this line (in a poem, besides, that is notionally fixed and uniformly ascribed to the man ‘Homer’), is thus, on the point of analogy, a good parallel to the corresponding ‘irrelevance’ of acting to drama. For the concept of ‘notional fixity’ and why it does not rule out a measure of mouvance, see §7.1 above.

[ back ] 26. Bywater (1909:260) notes on Poetics 20: “This whole chapter has been condemned by Ritter and others as an interpolation; and it must be admitted that, according to our notions of the divisions of knowledge, the matter in it belongs to grammar and philology rather than to an Art of Poetry.”

[ back ] 27. ἡ δὲ ὄψις ψυχαγωγικὸν μέν, ἀτεχνότατον δὲ καὶ ἥκιστα οἰκεῖον τῆς ποιητικῆς· ἡ γὰρ τῆς τραγῳδίας δύναμις καὶ ἄνευ ἀγῶνος καὶ ὑποκριτῶν ἔστιν, ἔτι δὲ κυριωτέρα περὶ τὴν ἀπεργασίαν τῶν ὄψεων ἡ τοῦ σκευοποιοῦ τέχνη τῆς τῶν ποιητῶν ἐστιν (‘Theatrical spectacle moves the heart but it falls quite outside, and is least pertinent to, the art of poetry; for the effect of tragedy is present even without public contest and actors, and, besides, the production of visual effects belongs more to the art of the equipment and props maker than to the art of the poets’). Just as τὸ [φύσει] ὑποκριτικόν is ἄτεχνον, so also ‘spectacle’ is ἀτεχνότατον, not in absolute terms (for otherwise how could Aristotle in this very passage refer to ἡ τοῦ σκευοποιοῦ τέχνη?) but with respect to ἡ ποιητικὴ τέχνη (so, correctly, Susemihl 1874:101, 235; Gudeman 1934:190; and Halliwell 1986:340 §d). In contrast to ὑπόκρισις, however, which can be readily associated with λέξις—an association of φωνή with λόγος—and thus becomes the object of systematic study, the symbolic grammar of dramatic spectacle fails to benefit from a similar analytic strategy. The absence of a natural correspondence between verbal and visual language makes the crucial difference. Cf. Poetics 1453b1–9 and 1462a11–12.

[ back ] 28. Not necessarily as written texts (though also available thus), but as oral literature that can be actualized on demand by performance. It is doubtful that rhetorical addresses were ever similarly considered.

[ back ] 29. Addressing this distinction, Halliwell (1993:51) remarks apropos Rhetoric Ⅲ: “[T]he subject of these chapters is rhetorical lexis, not lexis in general, a point emphasized at 3.1.10, 04a 36–9.” See further above, §13.2 n. 31 and §13.5 n. 100.

[ back ] 30. To this definition of λέξις, Aristotle adds the further observation, ὃ καὶ ἐπὶ τῶν ἐμμέτρων καὶ ἐπὶ τῶν λόγων ἔχει τὴν αὐτὴν δύναμιν (‘which has the same potential in both verse and prose’, Poetics 1450b14–15). In its light, his earlier statement at 1449b33–34 is but a particular application of ἡ διὰ τῆς ὀνομασίας ἑρμηνεία to metrical expression (cf. Dupont-Roc and Lallot 1980:209–210). Note the divergent point of view of Rhetoric 1404a28: ἑτέρα λόγου καὶ ποιήσεως λέξις ἐστίν (‘The lexis of prose is different from that of poetry’). If the λέξις of the Poetics denotes ‘expression’ (equivalent in this sense to λόγος broadly understood as ‘speech’ in communication, as observed by Lucas 1968:195 ad 1456a37), in the Rhetoric the focus is more narrowly on the form of such expression, in particular on how it must be shaped for persuasive delivery before an audience.

[ back ] 31. See Rhetoric 1403a36–b1 with Grimaldi 1988:369 ad loc. Cf. also Poetics 1450b4–5: τοῦτο [i.e. ἡ διάνοια] δέ ἐστιν τὸ λέγειν δύνασθαι τὰ ἐνόντα καὶ τὰ ἁρμόττοντα. Lucas (1968:195) relates τοῦ λόγου in Poetics 1456a37 to λέξις as ‘expression’. But Aristotle not only conceptualizes rhetorical lexis and poetic lexis divergently; even the concept of λέξις that the Poetics explores in the following section—and tellingly fails to tie explicitly to pathos—diverges from the λόγος that is strictly instrumental to διάνοια. A distinction is presupposed at 1456a33–34, where λοιπόν divides into λέξις and διάνοια. And so it is that, after dismissing διάνοια and referring the interested reader to his Rhetoric, the philosopher then takes up λέξις at 1456b8 from the point of view of its σχήματα.

[ back ] 32. Cf. Aristotle’s comment at 1450a24–26 about ἀήθεις τραγῳδίαι (‘tragedies deficient in character’; cf. Lucas 1968:103 ad loc.).

[ back ] 33. There is, as scholars note, a parallel in the passage of the Rhetoric that starts at 1356a1, but with this important distinction: that, as befits an introduction, Aristotle there uses logos in the most general sense of ‘discourse, speech’ (cf. Grimaldi 1980 ad loc.) and makes no attempt to apportion ēthos or pathos to one or another component of the art of oratory. Therefore Kennedy’s remark, “Aristotle is not thinking of style and delivery but of the thought and contents” (1991:38n41), is unnecessary and a little misleading. That the philosopher does not exclude style any more than he does thought is clear from the comment that follows, namely, that persuasion through character happens whenever the speech is spoken in such a way as to make the speaker worthy of trust.

[ back ] 34. For the relevant testimonia and fragments, see DK 85 and Radermacher 1951:B.Ⅸ. For studies of his contribution to the development of prose and rhetoric, see Blass 1887:244–258, Schwartz 1892, Drerup 1901:225–251, and Kennedy 1963:68–70.

[ back ] 35. See below, §14.2.

[ back ] 36. Quintilian’s passing comment (Institutio oratoria 3.3.4) that Thrasymakhos had ascribed actio to natura rather than ars is not as definitive in its implications as might seem at first. For on the basis of Rhetoric 1404a15–16 the same can be said, grosso modo, about Aristotle himself. And yet we have seen how the philosopher, with this broad statement as a foil, introduces a technical notion of actio in its particular connection to elocutio. All the same, the logic of his thought has often escaped the modern scholar and we cannot be sure that Thrasymakhos’ words, even if faithfully reported by Quintilian, were not accompanied by such instruction as would have made clear, explicitly or implicitly, that in certain respects actio was, after all, susceptible of systematic analysis. Quintilian may have missed this, may have depended on a source that was not sufficiently accurate or detailed, or may have chosen only to report the general statement without any further elaboration. His remark, then, may do nothing more than place the sophist all the more firmly as a forerunner of Aristotle in his approach to ὑπόκρισις.

[ back ] 37. The Suidas s.v. Θρασύμαχος states: ἔγραψε συμβουλευτικούς, τέχνην ῥητορικήν, παίγνια, ἀφορμὰς ῥητορικάς (2.726 Θ no. 462 Adler). The scholiast to Aristophanes Birds 880, however, mentions a μεγάλη τέχνη: is it the same as the τέχνη ῥητορική of the Suidas? Cf. Plato Phaidros 261c2 and 266c3.

[ back ] 38. τῶν γε μὴν οἰκτρογόων ἐπὶ γῆρας καὶ πενίαν ἑλκομένων λόγων κεκρατηκέναι τέχνῃ μοι φαίνεται τὸ τοῦ Χαλκηδονίου σθένος, ὀργίσαι τε αὖ πολλοὺς ἅμα δεινὸς ἁνὴρ γέγονεν, καὶ πάλιν ὠργισμένοις ἐπᾴδων κηλεῖν, ὡς ἔφη· διαβάλλειν τε καὶ ἀπολύσασθαι διαβολὰς ὁθενδὴ κράτιστος. τὸ δὲ δὴ τέλος τῶν λόγων κοινῇ πᾶσιν ἔοικε συνδεδογμένον εἶναι, ᾧ τινες μὲν ἐπάνοδον, ἄλλοι δ’ ἄλλο τίθενται ὄνομα (267c7–d4).

[ back ] 39. Radermacher (1951:71) comments perceptively on the rhythmic structure Plato imparts to the words τῶν γε μὴν οἰκτρογόων ἐπὶ γῆρας καὶ πενίαν ἑλκομένων λόγων (Phaidros 267c7–8): “Ceterum notabis puros numeros … ex quo conicias in ἐλέων exemplis admodum numerosam fuisse sophistae orationem.”

[ back ] 40. Just as Aristotle does in Rhetoric Ⅲ.19, though chapter 14.7 makes clear that such devices are also profitably used in prooimia.

[ back ] 41. Or a middle style, if Grube 1952 is right (cf. O’Sullivan 1992:114), though his insistence upon rendering λέξις narrowly as ‘diction’ (word choice stricto sensu) seems misguided, as his own comments on pp. 264–266 hint. Cf. also Innes 1985:260–262.

[ back ] 42. ἡ συστρέφουσα τὰ νοήματα καὶ στρογγύλως ἐκφέρουσα λέξις (§6 p. 14.9–10 Usener). Grube (1952:260) translates: “A manner of writing (or speaking) which gathers up its ideas compactly and brings them out in bold relief.”

[ back ] 43. Presumably reading ἀσκοῦντας with F1, although he too prints ἀσκούντων.

[ back ] 44. See below, §14.2.

[ back ] 45. E. Schwartz delevit: not only is it contradicted by the testimony of the Suidas but, as Blass (1887:250) notes, by established facts about the sophist, including the extant fragments. It is true that as a non-Athenian he would not have had the right to address the assembly, but neither are all speeches of a symbuleutic character necessarily for oral delivery or at least for delivery by the author himself. Thus, the reported Ὑπὲρ Λαρισαίων (DK 85 B2), whose single extant line indeed reads like a political address, may have belonged to a political pamphlet or it may have been written for a Thessalian embassy (so Blass 1887:250). And the Περὶ πολιτείας (DK 85 B1) is so generic an exemplum, as Radermacher (1951:B.Ⅸ.10) notes, that it may well have been extracted from a rhetorical manual. (Cf. Drerup 1901:227, who thinks it must have been a political pamphlet. For Dionysios [Dēmosthenēs §3 p. 132.18 Usener] it is a παράδειγμα from one of his δημηγορικοὶ λόγοι.) In any case, it is surprising that one who arguably took great interest in the rhetorical impact of emotions and whose style was arguably most fitting to the oratory of the law-court should not have left any forensic speeches (as Dionysios Isaios §20 observes).

[ back ] 46. Orator 175: “nam neminem in eo genere scientius versatum Isocrate confitendum est, sed princeps inveniundi fuit Thrasymachus, cuius omnia nimis etiam exstant scripta numerose.”

[ back ] 47. He means Dēmosthenēs §3, although on p. 68 he erroneously refers to “Isaeus 3.”

[ back ] 48. Cf. Grube 1952, who disputes Thrasymakhos’ alleged part in the development of the periodic style. But he is countered by Innes (1985:262), who suggests that Theophrastos may have recognized the sophist as the pioneer of “the mean between the unshaped and the overperiodic sentence.”

[ back ] 49. See above, §14.1 with n. 3.

[ back ] 50. This is an affirmation of the stylistic potential that writing holds for delivery; it does not of course assert that any and all written speeches in fact exhibited the force of their style at the expense of their argument. Some have sought to weaken the implication that οἱ γραφόμενοι λόγοι are introduced to make a point about the practice of performing orators. But, ceteris paribus, given its context one must surely grant not only the plausibility but even the likelihood that this implication is deliberate and true to fact. After all, it is while discussing the role of ὑπόκρισις qua art in its connection with λέξις that Aristotle observes that speakers who master this aspect of ῥητορική win prizes. Hence follows that the competitive success asserted of orators competent in ἡ περὶ τὴν λέξιν ὑπόκρισις must be the very fact that the ensuing γάρ clause seeks to explain when it draws attention to the ‘force’ of written speeches on account of their λέξις. Some translate οἱ γραφόμενοι λόγοι as ‘speeches intended for reading’, thereby dismissing any connection with performance other than perhaps an adventitious one. But no prizes were ever awarded to written speeches qua written, and this view effectively (and inadmissibly) severs the γάρ clause from the context it is supposed to illustrate. A sounder proposal, if in my opinion still too limiting, is that the statement chiefly regards epideictic showpieces (cf. Rhetoric 1414b24–25), whose authors would have hoped for a broad readership and even crafted them with reading in mind. But even under this proposal prize winning would still require their public oral performance (including perhaps reading them in public). This hypothetical scenario would only represent an extreme case of orators’ dependence on scripts for their delivery. Nonetheless, I believe that a broader reference than ‘epideictic’ is commended both by the context, itself a general survey, and by the present participle γραφόμενοι (on which see below). Sonkowsky (1959:261) is even more definitive and asserts that “the suggestion … that the epideiktikon genos is meant cannot be correct unless Aristotle is inconsistent,” because “[e]pideictic in the lexis graphikē is the most accurately written of the three kinds of discourse and depends the least upon delivery, the counterpart of style, for its effectiveness.” As a signal case of ‘epideictic’ (broadly understood) one might think of Isokrates. But although Aristotle knew some of his speeches and occasionally used them to illustrate his argument, these were not delivered by their author in a public setting such as would encourage the notion that they adequately illustrate the meaning, much less that they are the intended referent, of οἱ γραφόμενοι λόγοι. Isokrates famously declared himself neither well-endowed nor courageous enough to deal with the crowd and the speaker’s platform (Isokrates 5.81, 12.9–10, and Letter 8.7); and in Isokrates 5.7 he even contrasts the (arguably private or semi-public) oral delivery of a speech with its ‘publication’ (διαδοθέντος τοῦ λόγου; cf. Hudson-Williams 1949). Our current knowledge of Greek rhetoric suggests that there were no formal rhetorical contests with prizes during the classical period. Isokrates explicitly says so of his own time in Isokrates 4.1–3, a passage that equates δωρεαί, τιμή, and ἆθλον. Ancient reports point to Gorgias’ delivering addresses at Olympia and Delphi; the setting of Lysias 33 is Olympia (Lysias 33.2 ascribes to Herakles the convening of a twofold assembly featuring a physical ἀγών and an intellectual ἐπίδειξις [φιλοτιμίαν glosses ἀγῶνα and the Aldine δέ should be rejected]; the word choice underlines the different status of the contest and the display). Isokrates 5.12–13, 15.147, and Letter 1.6 confirm that ἐπιδείξεις were regularly delivered at πανηγύρεις and further suggest the absence of prizes (cf. Carey 2007:238). But if ἆθλα therefore stands not for placing first in a contest but metaphorically for success in the competitive settings of the courtroom (where plaintiff and defendant contended; cf. Plato Euthydēmos 305b6–9), the assembly (where competing policies were advocated), and the festival (where orators hoped that their own showpieces would receive a more enthusiastic hearing than their rivals’), then we are emphatically not dealing with ‘speeches intended for reading’, except perhaps in the very restricted sense of ‘intended for reading aloud’ where, exceptionally (cf. Alkidamas On the Sophists §18), delivery permitted an overt show of dependence on a script. Alkidamas’ On the Sophists further supports connecting οἱ γραφόμενοι λόγοι with performance. It proves that written speeches were viewed by traditionalists in the rhetorical trade as unwelcome parvenus not to Athenian bookshops but to public delivery. It is surely significant that Aristotle follows up his comment about written speeches with further observations on the role of voice and the rise of rhapsodic and dramatic performance as τέχναι; and that he adds the opinion of the uneducated that those whose public delivery style bears heavy poetic color ‘speak best’ (διαλέγεσθαι κάλλιστα). In sum, nothing suggests that οἱ γραφόμενοι λόγοι are ‘speeches intended for reading’, while there is much to commend the view that they are precisely what Aristotle calls them, ‘written speeches’. One final point: note that the philosopher calls them neither οἱ γραπτοὶ λόγοι (cf. Alkidamas On the Sophists §§1, 14, 18, 25, 31–33) nor οἱ γεγραμμένοι λόγοι (cf. Plato Phaidros 277e5; Aristotle Rhetoric 1407b11–12, Sophistical Refutations 166b3–4, Topics 105b12–13, and the many mentions of γεγραμμένοι νόμοι in the Rhetoric), both, alternatives that would have emphasized the speeches as written products. Overlooking for a moment all the other arguments that preclude the meaning ‘speeches intended for reading’, if nothing else these alternatives per se would more plausibly hint at the circulation of written texts apart from performance. By using instead γραφόμενοι, which as a present-tense participle cannot in this context be particular (Smyth §§2052, 2728a), Aristotle shows that he has in view not an actually existing class of speeches but a contingency (‘speeches that are [being] written’ meaning ‘if/whenever speeches are written’) that at the very least marginally draws attention to process and practice. Such practice must in turn be regarded against the context of technical ὑπόκρισις. Cf. Alkidamas’ mention of οἱ εἰς τὰ δικαστήρια τοὺς λόγους γράφοντες (On the Sophists §13). For Aristotelian parallels, cf. De interpretatione 16a3–4 (τὰ ἐν τῇ φωνῇ, ‘what one speaks’, versus τὰ γραφόμενα, ‘what one writes’); and Historia animalium 515a34–35 (ὥσπερ ἐν τοῖς γραφομένοις κανάβοις) vis-à-vis De generatione animalium 743a2 (καθάπερ οἱ τοὺς κανάβους γράφοντες ἐν τοῖς τοίχοις).

[ back ] 51. And perhaps transcripts of others, not their own, whose past success commended them as models to imitate. Cf. Plutarch Dēmosthenēs §8.2.

[ back ] 52. Isokrates’ case is truly exceptional and resulted from his peculiar incapacitating infirmity. Compare §14.2 n. 50 immediately above with Rhetoric 1414a16–17. See also Harris 1989:85–86 (where “the author of the Rhetorica ad Alexandrum” at 86 should be corrected to ‘Aristotle’). Nota bene that the contested (for Harris, “highly dubious”) Aristotelian report preserved by Dionysios of Halikarnassos (cf. his note 99) refers to Aristotle’s, not Isokrates’, times.

[ back ] 53. Arguably, this point of view underlies an otherwise puzzling fact: the examples Aristotle selects to illustrate his exposition in the Rhetoric generally fail to demonstrate first-hand, detailed acquaintance with actual forensic and deliberative speeches (Trevett 1996a). Precisely because their scripts (and transcripts) were considered ancillary and derivative they must have enjoyed, if at all, rather limited circulation (pace Dionysios of Halikarnassos Isokratēs 18; see n. 52 immediately above) and must have been hard to acquire. And ordinarily—there are exceptions—the philosopher may have considered them too ephemeral to admit of cogent and intelligible contextualization, or to suit a treatise that to escape premature obsolescence had better avoid potentially obscure historical singularities (cf. Trevett 1996a:377–378). By reason of their greater popularity, reach, and expected longevity, various poetic works were naturally (if paradoxically) a better source of stylistic exempla (cf. Graff 2005, esp. 322–334). The sole significant exception was epideictic rhetoric (Trevett 1996a:375–376), especially Isokrates’, whose written speeches, one must assume, were considered of such polish and were so readily available to, and familiar among, those interested enough in oratory to consult Aristotle on the subject that the philosopher thought them suitable for illustration. The relatively greater difficulty of procuring oneself with actual scripts of forensic and deliberative oratory, and even an estimate of their finish and excellence that perhaps compared unfavorably with those of epideictic and poetry, do not preclude Aristotle’s conviction and acknowledgment, expressed in Rhetoric 1404a16–19, that orators in all performance settings, not just where epideictic was concerned, were the better and more successful for their mastery of technical delivery, and that written scripts played a central role in reaching the desired proficiency.

[ back ] 54. ῥήτωρ καὶ δημαγωγός· ὅστις πρῶτος γραπτὸν λόγον ἐν δικαστηρίῳ εἶπε, τῶν πρὸ αὐτοῦ σχεδιαζόντων (4.100 Π no. 1180 Adler). On this, Blass (1887:35n5) remarks: “[D]iese letzte Notiz geht ganz entschieden auf Aristoteles oder Theophrast zurück.”

[ back ] 55. “Nam antea neminem solitum via nec arte, sed accurate tamen et de scripto plerosque dicere” (Cicero Brutus §46). Although the vulgate descripto has been plausibly emended to descripte (or even discripte) and plerosque seems arguably an overstatement, one cannot entirely rule out ‘from a script’ as the true reading, especially if the report of Perikles’ practice preserved in the Suidas is accurate and goes back to peripatetic sources. Note that this assertion in the Brutus is immediately followed by “scriptasque fuisse et paratas a Protagora rerum illustrium disputationes” (my emphasis). Cf. Blass 1887:27n1, 34–39. Even if Quintilian was right to reject the attribution of certain extant writings to the famous Athenian (Institutio oratoria 3.1.12), and even if these same writings are the ones that Cicero (Brutus §27, De oratore Ⅱ §93) apparently accepted as genuine—which of these two critics is one to prefer?—it need not follow that Cicero inferred the use of scripts from the falsely attributed texts. I am not convinced, at any rate, by the thoroughgoing skepticism of Hudson-Williams 1951 and Trevett 1996b:433–436. Plutarch’s observation in Periklēs §8.7 (perhaps confirmed by Plato Phaidros 257d–e) that Perikles bequeathed no writings to posterity hardly proves that he never used scripts. In fact, that the Suidas should have claimed him as the first to deliver a written speech not before the assembly (as one might otherwise have expected) but in a courtroom argues for the veracity of the report; and the scrupulousness implicit in the practice agrees with the portrait of a man who was περὶ τὸν λόγον εὐλαβής (Periklēs §8.6). I am, however, persuaded (and Alkidamas and Rhetoric Ⅲ.12 support the view) that ceteris paribus deliberative speeches were less likely to depend, and in fact less frequently depended, on a script for their delivery; and that, when they did, the scripts were probably less elaborate and detailed both as to form and content.

[ back ] 56. For this acceptation of de, see the OLD s.v. 7c: “indicating personal or other source of information.” E.g. de scripto sententia dicta (Cicero Pro Sestio 129); or de tabulis publicis recitat (Cicero Pro Flacco 40).

[ back ] 57. See n. 55 immediately above.

[ back ] 58. καὶ σύνοισθά που καὶ αὐτὸς ὅτι οἱ μέγιστον δυνάμενοί τε καὶ σεμνότατοι ἐν ταῖς πόλεσιν αἰσχύνονται λόγους τε γράφειν καὶ καταλείπειν συγγράμματα ἑαυτῶν, δόξαν φοβούμενοι τοῦ ἔπειτα χρόνου, μὴ σοφισταὶ καλῶνται (Plato Phaidros 257d4–8: ‘And you yourself know, I suppose, that those of the greatest power and dignity in our cities are ashamed to write speeches and leave behind writings of their own, fearing the opinion of future times lest they be called sophists’). Cf. Phaidros 257c6, where λογογράφος is considered a term of opprobrium.

[ back ] 59. Plutarch Periklēs 8.7 (156c10–11): ἔγγραφον μὲν οὖν οὐδὲν ἀπολέλοιπε πλὴν τῶν ψηφισμάτων.

[ back ] 60. Cf. Blass 1887:36.

[ back ] 61. To which Stallbaum 1857b ad loc. adds: “per totum enim librum Lysias non ut orator, sed tanquam orationum scriptor carpitur.”

[ back ] 62. Phaidros 258a6–8: ἔπειτα λέγει δὴ μετὰ τοῦτο, ἐπιδεικνύμενος τοῖς ἐπαινέταις τὴν ἑαυτοῦ σοφίαν, ἐνίοτε πάνυ μακρὸν ποιησάμενος σύγγραμμα (‘Then he speaks after this, displaying his own wisdom to his fans, having composed at times a very long writing’). For λογογραφία taken broadly as ‘composition of speeches’ see Yunis 2011:172 ad 257e1–2 and Innes 2007:154 ad Rhetoric 1388b21–22. See also above, §9.6.

[ back ] 63. I agree with O’Sullivan (1996:127) contra Graff (2001:22 and 37) when he observes in connection with Rheroric Ⅲ.12 that “[t]he written and the spoken, of such importance in the fifth century, continue to be the framework [for the analysis of rhetorical prose] as the world moves into an increasingly bookish Hellenistic age.” Graff, on the other hand, believes that O’Sullivan overestimates the significance of the written/spoken division and maintains that “[t]he twelfth chapter [of the Rhetoric] and the stylistic distinction it contains have the character of an afterthought” (22). That Aristotle should have planned and executed the composition of his treatise so haphazardly strikes me as psychologically implausible and un-Aristotelian. Such a notion is in fact belied by the deliberate manner in which the philosopher moves from one thematic head to the next. I do not mean to suggest, however, that all transitions are flawlessly executed. Aristotle was arguably comfortable with a measure of redundancy that many think awkward (e.g. at Rhetoric 1403a34–b8). The expectation that Rhetoric Ⅲ.1–12 should exhibit unity of design and expository coherence seems warranted even under the assumption that it was composed as an independent work. Chapter 12 starts with a transition that is not only smooth but also of the type expected for a closing summary that ties the threads of the argument together: ‘But one must not forget that a different style suits each genus [of rhetoric]’. Having surveyed λέξις without reference to, and insofar as common to all, genera, Aristotle now attempts to close with the stylistic emphases peculiar to each genus. At 1414a18–29 he notes explicitly that this has been the design of the entire section (chapters 1–12): ‘To subdivide lexis further, [to state] that it must be pleasant and elevated, is superfluous. For why [state this] rather than [that it should be] temperate and liberal and whatever other virtue of character there is? Obviously what has been discussed will bring about that it is pleasant, if indeed the virtue of lexis has been rightly defined … . So then, I am done speaking about lexis, from the point of view both of what is common about all [genera] and of what is peculiar to each genus’. Clearly the ‘common point of view’ (κοινῇ a28) regards the earlier chapters whereas the ‘peculiar’ (ἰδίᾳ a29), the present summation. As it is, if with Rapp (2002:2.948) one takes the presence of transitions and summations as proof that Rhetoric Ⅲ.12 is merely an appendix, then one will be denying Aristotle the expository strategies that would precisely allow him to integrate his earlier exposition of common elements of style with his summation of the subject from the vantage point of genera. According to Rapp, the view that chapter 12 is integral to the philosopher’s conception of rhetorical style remains implausible in the absence of explicit advance notice of its content earlier in the treatise. Apparently, only prior warning would adequately prepare the reader for a conclusion to ‘style’ that makes application to the three main social settings of rhetorical practice. Had Aristotle actually included such notice in chapter 1, I suspect that many would think it interpolated into an outline that did not include the corresponding material. On the internal coherence of Rhetoric Ⅲ.1–12 and its integration (and, more generally, the integration of Book 3) with the rest of the treatise, cf. Grimaldi 1972, Kennedy 1991:299–305, Graff 2000:12–17, Innes 2007:151n2, and Burkett 2011:xi–xxxv.

[ back ] 64. A different attempt to construe the passage is to invert the reference of τὸ μὲν … τὸ δέ, with the former pointing to the graphic, the latter to the agonistic, style. This is Thurot’s strategy, who replaces by μόνον the μὴ of οἱ μὴ ἐπιστάμενοι γράφειν. To this, Roemer 1898 ad loc. writes: “sed tum verba τὸ μὲν ἑλληνίζειν ἐπίστασθαι ad λέξιν γραφικὴν referenda essent, quod fieri non potest, contra verba ἄν τι βούληται μεταδοῦναι τοῖς ἄλλοις latiore sensu intellegenda significare videntur, quod Bonitzius cum omnibus interpretibus declarat ‘tradere per libros scriptos’.”

[ back ] 65. Greek has no set rule by which τὸ μέν consistently means ‘the former’ and τὸ δέ ‘the latter’ or vice versa. The choice must be guided by the context (LSJ s.vv. ὁ, ἡ, τό A.Ⅵ.1; cf. K-G Ⅱ.2.264 §527.3a Anm. 1). As early as Β 52 οἱ μέν is found to point to ‘the latter’ and τοὶ δ[έ] to ‘the former’.

[ back ] 66. I do not call it ‘opposition’ for reasons detailed on §14.2 below.

[ back ] 67. Cf. Rapp 2002:2.931.

[ back ] 68. ἡ μὲν οὖν ἐπιδεικτικὴ λέξις γραφικωτάτη· τὸ γὰρ ἔργον αὐτῆς ἀνάγνωσις· δευτέρα δὲ ἡ δικανική.

[ back ] 69. ἔστι δὲ λέξις γραφικὴ μὲν ἡ ἀκριβεστάτη (1413b8–9).

[ back ] 70. ἡ μὲν οὖν δημηγορικὴ λέξις καὶ παντελῶς ἔοικε τῇ σκιαγραφίᾳ … διὸ τὰ ἀκριβῆ περίεργα καὶ χείρω φαίνεται … · ἡ δὲ δίκη ἀκριβέστερον (1414a7–11).

[ back ] 71. Rapp 2002:2.937.ⅳ.

[ back ] 72. I will make clear below that, even on these terms, this consensus reading is impossible. I accept it here simply for the sake of argument.

[ back ] 73. Grimaldi’s translation, “and since one thing is suasive and credible in and by itself, another by the fact that it seems to be proved by statements that are such” (1980:52 ad b28), although equivalent to Aristotle’s meaning, is not syntactically accurate. The subject of ἐστι is τὸ πιθανόν, an abstract concept that I might somewhat inelegantly render as ‘the persuasive’. This concept embraces any and every particular instance of what persuades. Since the ensuing observation is predicated of this abstract notion, stricto sensu it must be translated ‘the persuasive is either X or Y’ or, better, ‘the persuasive is now X, now Y’. It may well be the case that ordinarily no instance of ‘the persuasive’ will be both inherently suasive at times and derivatively suasive at others (or even suasive partly in and by itself, partly derivatively), although this possibility cannot be absolutely excluded. But insofar as some instances are inherently suasive/credible and some derivatively suasive/credible, the larger notion of ‘the persuasive’ to which all belong can and must be pronounced ‘now inherently persuasive, now derivatively persuasive’. Had Aristotle written instead, ἐπεὶ γὰρ τὰ πιθανὰ τινὶ πιθανά ἐστι, one could then take Grimaldi’s translation as Aristotle’s meaning stricto sensu. For another instance of the abstract τὸ πιθανόν see 1355b15–16.

[ back ] 74. Nothing prevents the inherently and immediately plausible from being susceptible of inferential proof too.

[ back ] 75. K-G Ⅱ.2.264 §527.3a: “τὸ μέν … τὸ δέ, τὰ μέν … τὰ δέ und τοῦτο μέν … τοῦτο δέ … , teils … teils, einerseits … andererseits, sowohl … als auch.” See also LSJ s.vv. ὁ, ἡ, τό A.Ⅷ.4. An adverbial construal, with the elided subject understood from the previous clause, is satisfactory too: ‘For [to know both] is both (τὸ μέν) to know how to speak good Greek and (τὸ δέ) not to be forced …’.

[ back ] 76. Why introduce here this strange notion of publishing books for distant audiences? Why not simply say, ‘knowing the graphic style makes possible persuasive speaking when precision is required’? If Aristotle’s point was to provide a compelling example of the orator’s need to master the written style, he should have chosen something that was ordinarily within the scope and interest of the citizen orator-to-be. Such would arguably be addressing the assembly and courtroom, a compelling prospect that I claim must be what Aristotle has in mind here.

[ back ] 77. Plato Prōtagoras 319c8–d7.

[ back ] 78. ὁ νομοθέτης διαρρήδην ἀπέδειξεν οὓς χρὴ δημηγορεῖν καὶ οὓς οὐ δεῖ λέγειν ἐν τῷ δήμῳ. καὶ οὐκ ἀπελαύνει ἀπὸ τοῦ βήματος, εἴ τις μὴ προγόνων ἐστὶν ἐστρατηγηκότων [υἱός], οὐδέ γε εἰ τέχνην τινὰ ἐργάζεται ἐπικουρῶν τῇ ἀναγκαίᾳ τροφῇ, ἀλλὰ τούτους καὶ μάλιστα ἀσπάζεται, καὶ διὰ τοῦτο πολλάκις ἐπερωτᾷ ‘τίς ἀγορεύειν βούλεται;’ (Aiskhines 1.27). Also reflecting this right and duty are the many references to ὁ βουλόμενος in the democratic process ([Demosthenes] 13.11 and Demosthenes 24.63; SEG 26.72, line 34; Aiskhines 1.23, 2.65; Andokides 1.23; Athēnaiōn Politeia 29.2).

[ back ] 79. Cf. Demosthenes 18.170, 191; Aristophanes Akharnians 45, Women at the Thesmophoria 379, Assemblywomen 130; and Alkidamas On the Sophists §11. On political initiative in Athens see Hansen 1981, Hansen 1991:266–268, and Ober 1989:108–112.

[ back ] 80. Plato Prōtagoras 319c1–7, although hardly an objective source, may be safely assumed to paint a realistic picture of the humiliation inflicted on a speaker by his fellow citizens. Cf. also Demosthenes 19.113 (and 19.14, one of many instances in which a speaker entreats the assembly not to interrupt with their clamoring).

[ back ] 81. My suggestion assumes, of course, that the citizen in question had decided in advance to take the initiative in the debate and could therefore make a script of his planned motion; or else that the scheduled debate regarded a subject sufficiently well known and defined to allow a script readied in advance of the meeting, with minimal adjustment at the time of delivery, substantially to represent the individual’s contribution to the discussion. For the view that these meetings had agendas and that one could seek to get an item added to them, see MacDowell 1975:64.B2. Obviously, Alkidamas’ scornful sketch in On the Sophists §11 cannot be taken to preclude scripted delivery before the people; for he ridicules the thought of a ῥήτωρ hastening to his writing tablet not only upon hearing the herald asking ‘Which citizen wants to speak?’ but also after the water clock in the courtroom is running. Yet no one, I think, will dispute the currency of scripted forensic delivery, often facilitated by the professional λογογράφος. Cf. Mariß (2002:169–170), especially her comment at 170 that “die Existenz eines schriftlichen λόγος δημηγορικός … wird suggeriert,” a suggestion however that she dismisses as objectively untenable and mooted by Alkidamas only to serve his rhetorical polemic. Be that as it may, we have the tantalizing notice in Thoukydides 8.68 that Antiphon was the one man best able to help before court and assembly whoever consulted with him in a matter. Pace Hudson-Williams (1951:69–70), the parallel between the forensic and the deliberative settings is telling. Once again, we must beware not to conflate the separate questions whether deliberative speeches were ‘published’ and whether speakers prepared at all with the aid of writing for their addresses before council and assembly. Such preparation could have been partial, chiefly comprising model proems, perorations, commonplaces, topics, aphorisms and the like—all elements that could be quickly adjusted to the need of the moment. Whatever the practice in the late fifth and early fourth centuries, all that I need for my present purposes is that the use of scripts for demegoric addresses should have been unexceptional enough by Aristotle’s time to motivate the philosopher’s recourse to them when he makes the case for writing as needful for rhetorical proficiency and uninhibited political engagement. I believe it is of some significance to my proposal that, starting soon after the year 403/402 BC, in accordance with what MacDowell (1975:65) calls the “Old Legislation Law,” anyone wishing to propose a new law had to submit a written proposal in advance (Demosthenes 20.89–99). This requirement, while eventually relaxed by the “New Legislation Law” (Demosthenes 20.91, with MacDowell 1975:65–66), remained in force for the “Review Law” (Demosthenes 24.20–23, with MacDowell 1975:66–68, esp. 67.D4).

[ back ] 82. Cf. Demosthenes 18.308 and 22.30. It is precisely the combination of voluntary speaking (ἄν τι βούληται) and self-imposed silence (ἀναγκάζεσθαι κατασιωπᾶν) that convinces me that the forensic setting is not primarily in view. There, one would more naturally be said ‘to have to speak’ rather than ‘to want to speak’; and neither plaintiff nor defendant could afford to remain silent even if he did not enjoy the aid of a script, whether drafted in person or supplied by the λογογράφος. On ἰσηγορία see Lewis 1971 and Hülsewiesche 2002.

[ back ] 83. Americans are used to a highly abstract concept of ‘speech’ that is deemed, at times, to include even the way money is spent. Hence the polemic whether placing restrictions on expenditures during political campaigns constitutes an unlawful abridgment of the constitutional right to free speech. But surely for Aristotle ‘to keep silent’ must have denoted sooner one who does not speak up than one who does not publish a written communication. On κατασιωπᾶν, see further below.

[ back ] 84. Hence Rapp’s remark (2002:2.932) that “[mit ἑλληνίζειν] scheint … nur eine minimale Bedingung genannt zu sein, und außerdem eine, die auch der schriftlichen Form zugrunde liegt … .”

[ back ] 85. The consensus reading implies that ἑλληνίζειν constitutes the essence of the ‘agonistic’ style and, furthermore, that both the demegoric and the dicanic are subsumed under it. This necessarily entails a corresponding reduction of the dicanic to basic oral competence, a reduction that is hard to accept for a setting that often depended emphatically on scripted delivery.

[ back ] 86. Cf. Kennedy 1991:231. It consists of five elements: first, the correct use of connectives; second, the use of specific names and steering clear of circumlocutions; third, the avoidance amphibolies; fourth, keeping the genders distinct; and, fifth, observing number.

[ back ] 87. The fact that Aristotle is considering the necessity and importance of writing even for the demegoric style, the least precise and most agonistic of all styles, shows how misleading it is to equate ‘agonistic’ with ‘oral’ and ‘graphic’ with ‘written.’ This, which Aristotle studiously avoids, Rapp (2002:2.931) does when he writes: “Insofern der Gegenbegriff zur Form der Kontroverse der des schriftlichen Stils ist, mag man dazu neigen, die Kontroverse ganz mit dem mündlichen Stil gleichzusetzen; das entspricht der Argumentation des Kapitels auch im Großen und Ganzen, an einer Stelle jedoch, scheint eine mündliche Rede berücksichtigt zu sein, die nicht kontrovers ist (vgl. 1414a12–14)” (his emphasis). This last observation displays the failure of the schematism.

[ back ] 88. Bonitz 1870:459 s.v. μεταδιδόναι: “tradere per libros scriptos.” Cf. §14.2 n. 64 above.

[ back ] 89. Here follows a list of such contexts: Euripides Iphigeneia among the Taurians 1030 ‘tell me what you think’ (δόξης μετάδος, ὡς κἀγὼ μάθω), Hiketides 56–58 ‘share now your thought with me’ (μετά νυν | δὸς ἐμοὶ σᾶς διανοίας, μετάδος δ’), Orestēs 153 ‘tell us/give us an account’ (λόγου μετάδος); Plato Phaidōn 63c9 ‘or will you share [your thought] with us?’ (ἢ κἂν ἡμῖν [τῆς διανοίας] μεταδοίης;), Kratylos 426b6–8 ‘these [notions], then, I will share with you if you want me to; and you, if you can find anywhere something better, try also to share it with me’ (τούτων οὖν σοι μεταδώσω, ἂν βούλῃ· σὺ δ’ ἄν τι ἔχῃς βέλτιόν ποθεν λαβεῖν, πειρᾶσθαι καὶ ἐμοὶ μεταδιδόναι), Lysis 204a2–3 ‘we spend our time mostly in conversation, in which we would gladly let you share’ (ἡ δὲ διατριβὴ τὰ πολλὰ ἐν λόγοις, ὧν ἡδέως ἄν σοι μεταδιδοῖμεν); Aiskhines 2.52 ‘allow to speak’ (ἀλλ’ οὔτ’ ἂν ὑμῖν ὁ ῥήτωρ οὗτος οὔτ’ ἂν ἐμοὶ λόγου μεταδοίη); Polybios 29.27.4 ‘consult with his friends about the circumstances’ (μεταδοῦναι τοῖς φίλοις ὑπὲρ τῶν προσπεπτωκότων), 38.8.1 ‘when Golosses communicated the conversation to the general’ (τοῦ δὲ Γολόσσου μεταδόντος τῷ στρατηγῷ περὶ τῶν εἰρημένων); Dionysios of Halikarnassos Roman Antiquities 6.82.3 ‘for we will no longer allow you to speak’ (οὐ γὰρ ἂν ἔτι μεταδοίημεν ὑμῖν λόγου), 10.9.3 ‘admitting no outsider either to their counsel or conversation’ (μεταδιδόντες οὐθενὶ τῶν ἔξωθεν οὔτε βουλεύματος οὔτε λόγου), 10.13.7 ‘and not even allowing the tribunes to speak also’ (καὶ οὐδὲ λόγου τοῖς δημάρχοις ἔτι μεταδόντες), 10.40.2 ‘but preventing those who wished to oppose the measure from speaking’ (τὸ δὲ μὴ μεταδιδόναι λόγου τοῖς ἀντιλέξαι βουλομένοις); Plutarch Quaestiones convivales 697d7–8 ‘but because he participates and shares in conversation’ (ἀλλ’ ὅτι λόγου μεταλαμβάνει καὶ μεταδίδωσιν); [Plutarch] Moralia 102b3–4 ‘I thought it good to share with you these consolatory words to relieve your grief’ (καλῶς ἔχειν ὑπέλαβον τῶν παραμυθητικῶν σοι μεταδοῦναι λόγων πρὸς ἄνεσιν τῆς λύπης); Dio Chrysostomus 37.1 ‘I shared my oratory with the people and your officers’ (τῶν λόγων μετέδωκα τῷ δήμῳ καὶ τοῖς τέλεσι τοῖς ὑμετέροις), 38.4 ‘why are you giving yourself the right to speak when we have not given it to you?’ (τί δὲ σεαυτῷ λόγου μεταδίδως, οὗ σοὶ μὴ μετέδομεν ἡμεῖς;), 38.5 ‘to have allowed a friend to speak who is willing to do so in vain’ (ἀνδρὶ φίλῳ λόγου μεταδοῦναι βουλομένῳ μάτην εἰπεῖν), 56.11 ‘if although being a king he allowed the others to speak’ (εἰ βασιλεὺς ὢν μετεδίδου λόγου τοῖς ἄλλοις); Josephus Bellum Judaicum 2.142 ‘he swears not to share their doctrines with anyone’ (ὄμνυσιν μηδενὶ μὲν μεταδοῦναι τῶν δογμάτων); 1 Thessalonians 2.8 ‘to share with you not only the gospel’ (μεταδοῦναι ὑμῖν οὐ μόνον τὸ εὐαγγέλιον); Akhilleus Tatios 4.7.8 ‘let her come before my eyes and converse with me’ (εἰς ὀφθαλμοὺς ἡκέτω τοὺς ἐμοὺς καὶ λόγων μεταδότω); Athenaios 1.2a ‘will you then agree to share with us too the fine talk you had over the cups … ?’ (ἆρ’ οὖν ἐθελήσεις καὶ ἡμῖν τῶν καλῶν ἐπικυλικίων λόγων μεταδοῦναι … ;), 14.643e–f ‘all the ones I remember I will also share with you … . But those names of flat cakes we copied we will also share with you’ (ὅσων μέμνημαι τούτων σοι καὶ μεταδώσω … . ἡμεῖς δὲ ἃ μετεγράψαμεν ὀνόματα πλακούντων τούτων σοι καὶ μεταδώσομεν); cf. Eusebios Ecclesiastical History 2.1.8; Cassius Dio 5.18.8; Philostratos Vitae sophistarum 2.18.599.9–10.

[ back ] 90. What fourth-century Athenian, successful in court and assembly, would have ever thought that his ability to communicate with others depended on his disseminating written documents?

[ back ] 91. Livingstone (2001:185) writes: “Isocrates regularly insists that his works are not intended for ἐπίδειξις, public performance and self-display … . Isocrates’ antithesis between (mere) ἐπίδειξις and practical, useful rhetoric parallels the Platonic Socrates’ contrast between ἐπίδειξις and dialectic. It … reappears in places where, as here, the practical objective is to give useful advice … [I]n Busiris … it is all the more necessary to remind his readers that this is not ἐπίδειξις for its own sake, but for a serious purpose.” On ὑποδεῖξαί σοι he adds: “[T]he verb ὑποδεικνύναι suggests a more ‘restrained’ presentation than ἐπιδεικνύναι … [It] means ‘to illustrate’, to show in brief … [and] can be used of getting a message across without stating it publicly … . In the present passage the idea of a concealed communication … is reinforced by the antithesis τοῖς ἄλλοις/σοι. Compare especially Ep. I 6, where Isocrates uses the fact that he is writing to Dionysius personally, rather than speaking at a public festival, as evidence that his aim is serious and practical, not ‘epideictic’” (185–186). There is therefore nothing here about ‘the world at large’ or ‘absent addressees’ who cannot be reached but by the medium of writing. (In fact, in Letter 1 Dionysios is the absent addressee, while the contrasting πλεῖστοι of §6 enjoy face-to-face delivery!) The point is narrowly rhetorical and regards grounding the orator’s claim about the character of his address in the identity and scope of his alleged target audience.

[ back ] 92. I take the translation from Gerber’s Loeb Classical Library edition. σοὶ μὲν ἐγὼ πτέρ’ ἔδωκα, σὺν οἷσ’ ἐπ’ ἀπείρονα πόντον | πωτήσηι, κατὰ γῆν πᾶσαν ἀειρόμενος ‖ ῥηϊδίως· θοίνηις δὲ καὶ εἰλαπίνηισι παρέσσηι | ἐν πάσαις πολλῶν κείμενος ἐν στόμασιν. Race (1987:149–150) explores related Pindaric precedents and echoes.

[ back ] 93. τοὺς δὲ λόγους ἐξενεχθῆναί θ’ οἷόν τ’ ἐστιν εἰς τὴν Ἑλλάδα καὶ διαδοθέντας ἐν ταῖς τῶν εὖ φρονούντων διατριβαῖς ἀγαπᾶσθαι. I take this sentence to be the reason why Innes mentions Isokrates 9.74 even though it does not feature οἱ ἄλλοι. τοῖς ἄλλοις ἅπασιν does occur at the end of the section, but in a context that bears no relation to her argument.

[ back ] 94. Hudson-Williams 1949.

[ back ] 95. Cf. Hägg 2012:33; pp. 30–41 provide a convenient introduction to the oration.

[ back ] 96. Oral delivery is further suggested by ἐν εἰδόσιν ποιούμενοι τοὺς λόγους (9.5); and τοὺς δ’ ἄλλους ἐθιστέον ἀκούειν περὶ ὧν καὶ λέγειν δίκαιόν ἐστιν (9.7). Isokrates inscribes his συγγράφειν (9.8) in the notional context of performance before an audience of hearers (cf. 9.4).

[ back ] 97. It is well known that Evagoras was an honorary citizen of Athens (cf. IG I3 113, Isokrates 9.54, and Demosthenes 12.10), so that Isokrates’ encomium, however thin in actual biographical detail and historical accuracy, had every hope of being warmly received by his fellow citizens. Braund (1998:54) suggests that “Isocrates’ oration was clearly immediately influential. Its effect can be seen in Xenophon’s Agesilaus” (but cf. Hägg 2012:41). We also know that Athens honored Evagoras ca. 393 BC with a decree from which three small fragments have survived. In connection with it Lewis and Stroud (1979:187) remark that “the enthusiastic welcome which Konon received as a returning hero was to a large extent shared by his Cyprian patron.”

[ back ] 98. Although τοὺς γραπ- before -τούς could have fallen out, I for one do not agree with the editors that the text must be emended. Alkidamas’ complaint is precisely that the use of writing is so widespread that ‘the speeches of the rest’, i.e. of all others who make oratory their regular practice, are scripted and therefore do not fail to spoil their audiences. It is precisely the hyperbolic τῶν ἄλλων τοὺς λόγους that explains why Alkidamas has reluctantly given in to the practice and does not even consider that any, perhaps even a majority, of his potential new hearers may have been exposed to, much less might well appreciate, extempore speaking.

[ back ] 99. For ἐκφέρω as ‘oral delivery’ see Hudson-Williams 1949:65–66. Cf. Mariß 2002: “[ἐκφέρειν] charakterisiert eher die mündliche Epideixis” (291).

[ back ] 100. My translation seeks to capture what Mariß (2002:294) rightly observes: “Vielmehr wird die präpositionale Wendung ἐξ ἐκείνου τοῦ τρόπου durch den ὅταν-Satz erklärt und konkretisiert: ‘fordern wir auf, auf jene Weise eine Probe von uns zu nehmen, wenn wir jeweils (da das Publikum das richtige ist) über jeden vorgelegten Gegenstand εὐκαίρως καὶ μουσικῶς sprechen können’.” Because the temporal contingency of ὅταν hinges on the presence of the right kind of public (those who often meet Alkidamas), insofar as restricted to them this contingency turns factual and the ὅταν clause is equivalent to an explanation (‘because then’). In other words, τοῖς πολλάκις ἡμῖν ἐντυγχάνουσιν notionally supplies a temporal reference point, a τότε, which ὅταν answers (Smyth §2240; cf. On the Sophists §13).

[ back ] 101. Mariß (2002:295) notes that here Alkidamas uses μὲν … δέ first to present, and then to expand on, a particular point of chronology: διὰ χρόνου μέν, ‘after a time’, is thus reprised by μηδεπώποτε δὲ πρότερον, ‘never yet before’. These do not regard two different groups but one and the same set of potential hearers who finally come to hear Alkidamas perform. διὰ χρόνου draws attention to the fact that they have long had the opportunity to hear him yet have never before availed themselves of it; it also hints at Alkidamas’ dissatisfied impatience with them.

[ back ] 102. Mariß (2002:295) correctly renders ἀκροάσεις “Rezitation,” “Lesung,” and suggests recourse to τοῖς μὲν γὰρ πολλάκις ἡμῖν ἐντυγχάνουσιν for clarification. She also adds that “Schriftliches liegt jedenfalls zugrunde.”

[ back ] 103. Cf. Isokrates 5.81 and 15.192. An only slightly less negative assessment seems implicated by Isokrates 2.16 and 12.263. In Isokrates 4.150, in a context that does not regard public speaking, the majority of the Persians are called ‘a mob without discipline’ (ὄχλος ἄτακτος). The connotation of the term is neutral at 4.96, 6.64, 6.73, 6.78, 18.9, and perhaps at 11.26. But none of these regard a rhetorical context and are therefore without relevance. With a different acceptation, ὄχλος as ‘mass’ (i.e. ‘great number’) occurs at 12.192 and 15.310; at 12.211 and 15.320 it denotes ‘trouble’.

[ back ] 104. τοῦτο γὰρ αἴτιον καὶ τοῦ πιθανωτέρους εἶναι τοὺς ἀπαιδεύτους τῶν πεπαιδευμένων ἐν τοῖς ὄχλοις, ὥσπερ φασὶν οἱ ποιηταὶ τοὺς ἀπαιδεύτους παρ’ ὄχλῳ μουσικωτέρους λέγειν (Rhetoric 1395b26–29). The other instance of ὄχλος in the Rhetoric is at 1414a9 (see below); it too seems to exhibit a somewhat negative color. Cf. Mariß 2002:292.

[ back ] 105. An alternative way of reading the τὸ μὲν … τὸ δέ clauses in Rhetoric 1413b5–6 takes ἄμφω as gathering together the graphic and the agonistic styles into a larger notion which the philosopher then considers from a different point of view. To illustrate this reading, suppose I should say: ‘a student of Latin must learn both grammar and vocabulary; for one thing is to know paradigms and another to be able to read Cicero.’ It would be incorrect to infer that the following two equations hold true: ‘to know grammar’ = ‘to know paradigms’ and ‘to know vocabulary’ = ‘to be able to read Cicero’. (Knowing paradigms would be a subset of knowing grammar, while knowing vocabulary simply would not correspond to being able to read Cicero.) My point would be, rather, that merely knowing paradigms bespeaks a limited linguistic ability, while a proficient grasp of both grammar and vocabulary confers a working knowledge of the language. Note that neither grammar nor vocabulary by itself would suffice. As an alternative to the reading I have defended above, one could argue that something similar happens here: just as the knowledge of both styles, graphic and agonistic, would equip the orator successfully to address an audience, failure to master either would prevent him from doing so. To paraphrase Aristotle’s point: ‘For one thing is to know how to speak correct Greek; another to be able successfully to communicate (and, presumably, to persuade) an audience, avoiding the need to keep silent because one lacks the requisite skills.’ When he uses τὸ μὲν … τὸ δέ, the philosopher often specifies what each points to by making them follow a genitive (e.g. Categories 4b20, Nicomachean Ethics 1109a33, Metaphysics 1072b3, Politics 1334b20, or De respiratione 471a8) or by having ἐν + dat. indicate their sphere of reference (e.g. On Generation and Corruption 317a23); at least once, τὸ μὲν … τὸ δέ merely stand in apposition (Metaphysics 1065a21). What has misled interpreters of Rhetoric 1413b5–8 is the (otherwise natural) assumption that the referent of μὲν … δέ is ἄμφω; and that since ἄμφω (to put it schematically) is made of the explicitly mentioned ‘X’ and ‘Y’, τὸ μέν must equal ‘X’, τὸ δέ, ‘Y’ (or vice versa). It would then be as if the philosopher had written: ἄμφω δὲ ἀνάγκη εἰδέναι· ὧν τὸ μὲν … τὸ δέ. But no ὧν is implied and τὸ μέν may be rendered as ‘one thing’, while τὸ δέ, as ‘another thing’. Aristotle could have written τὸ μὲν … ἄλλο δέ or τὸ μὲν … ἕτερον δέ: but perhaps he did not because this might have conveyed too sharp a contrast between speaking correct Greek and being able to speak to an audience in public address. (Note, for example, the ἓν μὲν … ἕτερον δὲ … ἄλλο δέ in Politics 1291b18–19, where the various categories do not overlap.) After all, we take for granted that proficient public speakers will be linguistically competent. One clear instance of τὸ μὲν … τὸ δέ which does not pick up on anything explicit that immediately precedes them is Sophistical Refutations 181b9–10: ‘If one thing is good and another evil (εἰ τὸ μέν ἐστιν ἀγαθὸν τὸ δὲ κακόν), it is true to call them good and evil’ etc. τὸ μέν and τὸ δέ do not respectively point to anything in the immediate context: they must be translated simply ‘one thing’ and ‘another thing’. So also, according to this alternative reading, in Rhetoric 1413b5–8: ‘One thing is to speak good Greek; another, not to be forced to keep silent’ etc. The general sphere of reference, were it to be made specific, would have to be expressed somewhat along the lines of τὸν γὰρ καὶ ἐξετάζειν καὶ ὑπέχειν λόγον καὶ ἀπολογεῖσθαι καὶ κατηγορεῖν ἐγχειροῦντα with τὸ μέν … τὸ δέ (cf. Rhetoric 1354a5–6). A paraphrase might run thus: τὸ μὲν γάρ ἐστιν τὸν καὶ ἐξετάζειν καὶ ὑπέχειν λόγον καὶ ἀπολογεῖσθαι καὶ κατηγορεῖν ἐγχειροῦντα ἑλληνίζειν ἐπίστασθαι, τὸ δὲ πιθανῶς καὶ σαφῶς μεταδοῦναι τοῖς ἄλλοις ἐπίστασθαι (or πιθανῶς καὶ σαφῶς μεταδοῦναι τοῖς ἄλλοις οἷόν τε εἶναι). The alternative construction ἔστιν μὲν … ἔστιν δέ would have made a poorer choice; for it would only make marginal sense to write: ‘it is possible, on the one hand, to know how to speak correct Greek; and possible, on the other, not to have to keep silent’ etc. ‘Not to have to keep silent’ may well be a possible scenario, but, strictly speaking, an inability rather than a possibility. I should make clear that I consider this alternative interpretation inferior to the one I have propounded above, although effectively its meaning is much the same. I only offer it to my reader here to underline the fact that there is very little to recommend joining the consensus on this passage of the Rhetoric.

[ back ] 106. A fact already recognized by Sonkowsky 1959: “The styles are a pair of ideal classifications; [Aristotle] recognizes a relative scale between them, along which, presumably, the two could be mixed together in varying degrees” (260–261). So also O’Sullivan 1992: “The two styles to be considered are, of course, the ‘written’ and the ‘unwritten’, but I use the quotation marks advisedly. For they are, in a sense, purely nominal designations, indicating tendency rather than method of presentation” (43). And, more recently, Schloemann 2000: “[Aristoteles] stellt Lesen … und öffentliche Rhetorik gegeneinander. Es ist dadurch nicht gesagt, daß die öffentlichen Reden auf der Ebene der Produktion mündlich sind … . Auch Reden im agonistischen Stil können schriftlich konzipiert sein” (209).

[ back ] 107. Cf. Rhetoric 1354a6–7.

[ back ] 108. Cf. Rhetoric 1354a7–11.

[ back ] 109. Cf. the use of τὰ ῥητορικά at 1375a8 and 1414a12 with Grimaldi 1980:314.

[ back ] 110. In other words, it is an error to think of τὰ ὑποκριτικά as the stylistic devices and strategies of the agonistic style, as if the graphic featured analogous τὰ ἀναγνωστικά (it does not). (Graff 2001:21 goes even further and mistakenly identifies the agonistic style with τὰ ὑποκριτικά.) Since all genera are performed and all styles are conceived with a view to delivery, τὰ ὑποκριτικά denotes the entire collection of stylistic devices and strategies that an orator can choose from in order to suit the subject and setting of his address. The clarity and persuasiveness that various stylistic devices and strategies are capable of in performance are unequally devitalized when deprived of delivery. So also is the subset of them peculiar to a given style denaturalized when employed unchanged in the wrong setting (the agonistic, in the setting proper to the graphic, and the graphic, in the setting proper to the agonistic).

[ back ] 111. That is, by reason of the expressive strategies characteristic of each genus, of the devices to which each of them typically resorts as well as in what manner, with what frequency, etc. it does so.

[ back ] 112. Bonitz 1870:285 s.v. ἔργον: “hoc discrimen in usu voc[is] ἔργον ita est conspicuum, ut modo operae et actionis notionem complectatur, modo ipsum opus effectum significet … .”

[ back ] 113. Graff (2001:21) is therefore wrong to assume that “by ‘reading’ Aristotle here … means a reading aloud, but presumably one lacking the dramatic vocal and gestural accompaniments of full hupokrisis.” Unless, of course, by “full hupokrisis” he means the delivery peculiar to the agonistic style. If so, this would then be a trivial observation, for no one would expect Aristotle to recommend an agonistic delivery for a graphic style. He is also wrong to wedge apart the goal of delivery from those stylistic qualities that make for easy reading: “Aristotle’s account here [sc. at Rhetoric Ⅲ.8.6] emphasizes the benefits to the reader. That Aristotle is thinking here not of the delivery of a speech, but rather its reading (aloud but conceivably to oneself) is unambiguous. No fourth-century Greek would deliver (ὑποκρίνεσθαι) a speech from a piece of papyrus before the assembly or even in court” (Graff 2000:198–199, his emphasis). The issue whether a speaker’s explicit show of dependence on a script was culturally acceptable or not in a particular social setting is a red herring. Quite apart from the undeniable benefits that a proper style would provide to the reader of lightly punctuated scriptio continua, it should be obvious that a script that does not make for a fluent and intelligible reading aloud, when adequately internalized for delivery, would neither make for a fluent and intelligible vocal performance. Hence Aristotle’s explicit equation at 1407b11–12 between written material that is εὐανάγνωστον (‘easy to read [aloud]’) and εὔφραστον (‘easy to give oral expression to’). Note that this last comment occurs in the section on ἑλληνίζειν, a subject that, for Aristotle, indisputably concerns the linguistic competence of oral expression. Cf. Graff 2001:28.

[ back ] 114. Alternatively, one might conjecture that, at least for some of these speeches, their own dramatic context was agonistic and that, if judged accordingly, they appeared meager. If so, at any rate, fans of Khairemon’s works cannot have felt this fact as a serious blemish.

[ back ] 115. Without Kassel’s emendation, οἱ μέν at b14 must refer to the poets (cf. οἱ ἀναγνωστικοί at b12); τῶν γραφικῶν (b15) in turn must either refer to the poets too (‘and side by side, those of the graphic [class of poets] …’) or else, rather more likely, to their dramas (τῶν γραφικῶν [δραμάτων], i.e. ‘those poets who compose graphic dramas’; cf. b11). With Kassel’s καὶ παραβαλλόμενοι ⟨οἱ λόγοι⟩ οἱ μὲν τῶν γραφικῶν (as with Ross’s similar καὶ παραβαλλόμενοι οἱ μὲν τῶν γραφέων ⟨λόγοι⟩), the comparison regards the speeches, not the poets; and τῶν γραφικῶν presumably refers to the ‘graphic poets’ (γραφικοί standing for ἀναγνωστικοί). I believe that ultimately none of these alternative texts differs much in meaning from any other. (I suggest a different emendation below, §14.2 n. 129.) I am not troubled by Aristotle’s substituting the poets for what stricto sensu must be the speeches in their plays and calling them ‘thin’ in debate, although I can appreciate the inconcinnity between οἱ μὲν τῶν γραφικῶν and οἱ δὲ τῶν ῥητόρων and understand why the editors would wish to emend it away.

[ back ] 116. If one accepts ἢ τῶν λεχθέντων at 1413b16 (see the footnote immediately below), here ῥήτορες, following its common political use, must narrowly denote assembly and forensic speakers.

[ back ] 117. It is hard to know what to do with ἢ τῶν λεχθέντων at 1413b16, which Kassel brackets. I am tempted to accept it on the grounds that οἱ [λόγοι] τῶν ῥητόρων ἢ τῶν λεχθέντων nicely comprises the three rhetorical genera, which Aristotle may well be comparing in toto to the speeches found in those poets who are suitable for reading. The point would be that the layman who takes in hand a script devised for delivery before an audience cannot with an ordinary (non-dramatic) reading compensate for the missing ὑπόκρισις. Aristotle may have added ἢ τῶν λεχθέντων from a feeling that, by only writing οἱ δὲ τῶν ῥητόρων, he would appear to leave out epideictic speeches. The addition, as the superset of actually delivered speeches (‘the speeches of orators or [those] from the ones spoken [that are compared]’), would not only include οἱ τῶν ῥητόρων but also incorporate epideictic. At any rate, if one follows Kassel in excising the words, it is still possible to interpret τῶν ῥητόρων as denoting all professional public speakers comprehensively. Their speeches would still comprise all three of the standard rhetorical genera.

[ back ] 118. Aristotle subsumes all these emphases under the umbrella of ἀκρίβεια (see below).

[ back ] 119. If my reader prefers the view that the anagnostic poets and their speeches straightforwardly represent the epideictic genus, then he need only excise ἢ τῶν λεχθέντων at 1413b16 and restrict τῶν ῥητόρων at b15–16 to symbuleutic and forensic speakers. Ultimately, the alternative preferred is of no great consequence to my overall argument.

[ back ] 120. ἐπειδὰν γὰρ ὁ λόγος ἀποστερηθῇ τῆς τε δόξης τῆς τοῦ λέγοντος καὶ τῆς φωνῆς καὶ τῶν μεταβολῶν τῶν ἐν ταῖς ῥητορείαις γιγνομένων, ἔτι δὲ καὶ τῶν καιρῶν καὶ τῆς σπουδῆς τῆς περὶ τὴν πρᾶξιν, καὶ … ἀναγιγνώσκῃ δέ τις αὐτὸν ἀπιθάνως καὶ μηδὲν ἦθος ἐνσημαινόμενος ἀλλ’ ὥσπερ ἀπαριθμῶν, εἰκότως, οἶμαι, φαῦλος εἶναι δοκεῖ τοῖς ἀκούουσιν.

[ back ] 121. This observation is true regardless of how one construes the relation of the epideictic genus to the anagnostic poets and their speeches.

[ back ] 122. Sonkowsky (1959:261) observes: “The technical term lexis graphikē must not be confused with graphomenos logos, which means ‘a speech that is written down.’ Speeches in the lexis agōnistikē can also be graphomenoi.”

[ back ] 123. For a concise classification of ἀκρίβεια according to its various uses, see Grant 1885:452 ad Nicomachean Ethics 1.7.18 1098a26–33 (cf. 1094b11–1095a13). Kurz 1970 offers a foundational study of the concept down to Aristotle’s time. For ‘precision’ in the stylistic context, see O’Sullivan 1992 passim (esp. 42–62); Graff 2000:53, 101–108, 171–177, and 195–198; Graff 2001; and Innes 2007:161–164.

[ back ] 124. ἀλλ’ ὅπου μάλιστα ὑποκρίσεως, ἐνταῦθα ἥκιστα ἀκρίβεια ἔνι. τοῦτο δέ, ὅπου φωνῆς, καὶ μάλιστα ὅπου μεγάλης.

[ back ] 125. Since the analysis concerns how to suit style to the various genera dicendi (cf. 1413b3–5), ὑποκριτικὴ λέξις is best rendered as ‘suited to the portrayal of ēthos and pathos that is the peculiar function of delivery’ (note ἠθική and παθητική at 1413b10). For the sake of convenience, I will use ‘histrionic’ as shorthand for the fuller, more accurate translation.

[ back ] 126. See above, §13.4 and n. 42.

[ back ] 127. Cf. Alkidamas On the Sophists §16: ὅταν γάρ τις ἐθισθῇ κατὰ μικρὸν ἐξεργάζεσθαι τοὺς λόγους καὶ μετ’ ἀκριβείας καὶ ῥυθμοῦ τὰ ῥήματα συντιθέναι, καὶ βραδείᾳ τῇ τῆς διανοίας κινήσει χρώμενος ἐπιτελεῖν τὴν ἑρμηνείαν … (‘For whenever one is accustomed to work out his speeches in detail and to put together the words with precision and rhythm, and to perfect his expression by using his mental process sluggishly …’).

[ back ] 128. Though certainty is impossible, I think that τοὺς τοιούτους are not the actors but ‘those [dramatic personae who are] such’: not all the characters of a play give the same scope to histrionic displays. In any case, whatever the identity of the elided nouns, τὰ τοιαῦτα doubtless means τὰ ἠθικὰ καὶ παθητικά and οἱ τοιοῦτοι, οἱ ἠθικοὶ καὶ παθητικοί.

[ back ] 129. If οἱ μέν stands for οἱ λόγοι, I believe that the text is best emended to καὶ παραβαλλόμενοι οἱ μὲν τῶν γραφικῶν ⟨τοῖς⟩ ἐν τοῖς ἀγῶσι στενοὶ φαίνονται (i.e. by construing the comparison as one between οἱ τῶν γραφικῶν and ⟨οἱ⟩ ἐν τοῖς ἀγῶσι). It is easy to see how the first τοῖς might have fallen out: ‘On the one hand, those [speeches that are] of the graphic ones appear thin when compared to those [used] in agōnes’.

[ back ] 130. Hence the glosses εὖ λεχθέντων in F (speakers are praised); and εὖ μὲν λεχθέντες (speeches are well delivered) in the anonymous commentary (Anonymi in artem rhetoricam commentarium, p. 221 ad loc., ed. Rabe).

[ back ] 131. Strictly speaking, it is not right to affirm that the graphic style discourages the use of frequent repetition. It would be more accurate to say that, in those social performance settings for which a relatively more graphic style is suitable, recourse to frequent repetition detracts from the effectiveness (clarity and persuasiveness) of the delivery. In such contexts a speaker would rightly disapprove of its use, and this is what is meant by ‘frequent repetition of the same in the graphic style is rightly rejected as inadequate’ (1413b19–20). As bundles of expressive devices and strategies with a view to clarity and persuasiveness formulated for performance in given social settings, styles are similar to genres and call for a comparable fit between cultural occasion (with its characteristic subject matter) and the manner of communication that obtains between a performer and his audience.

[ back ] 132. On σκιαγραφία, sometimes translated ‘shadow painting’, see Keuls 1975 [1997:107–144]; Pemberton 1976; Keuls 1978:72–87 (and the index s.v.); Rouveret 1989:24–26 and 50–59; Koch 2000:137–153; and Rouveret 2006.

[ back ] 133. Insofar as it depended for its effect on viewing at a distance, we may think of it as a sort of ‘classical impressionism.’ Besides the passage of the Rhetoric that now concerns us, Aristotle refers to it in Metaphysics 1024b23 and Protreptikos fr. 104 (Düring). Some say it is also in view in De sensu 439b20–23 and 440a29–30. Plato makes more frequent mention of it, and a list of loci can be found in Keuls 1978:78–79.

[ back ] 134. For a review of this and other passages where Plato touches on painting, see Demand 1975. Cf. Aristotle Metaphysics 1024b23, where σκιαγραφία is classed with dreams (ἐνύπνια).

[ back ] 135. Disputed matters are whether it involved several colors or just hues of one and whether the mixing was by superposition or juxtaposition. Cf. Pemberton 1976.

[ back ] 136. Cf. Johnstone 1996a.

[ back ] 137. No advantage is gained by understanding θέα literally as ‘point of view’ (i.e. not ‘opinion’ but ‘place from which one looks out’). One must extend its meaning into the realm of discourse (see immediately below).

[ back ] 138. Cf. Rhetoric I.3, esp. 1358b2–5 and b13–17.

[ back ] 139. Cf. Rhetoric 1369b31–32: δεῖ δὲ νομίζειν ἱκανοὺς εἶναι τοὺς ὅρους ἐὰν ὦσι περὶ ἑκάστου μήτε ἀσαφεῖς μήτε ἀκριβεῖς (‘One should consider definitions to suffice when they are neither unclear nor precise about each matter’). Here, ἀκριβεῖς (‘precise, exact’) seems as undesirable as ἀσαφεῖς (‘unclear’). This has troubled some, who have therefore tried to soften it by assuming an excess (i.e. not just ‘precise’ but ‘excessively precise’). Kennedy even translates it as ‘inexact’, inverting its sense, although in an accompanying footnote he explains that by ‘inexact’ he really means ‘overly technical’(!). But perhaps it is best understood in contrast to ἱκανούς: ‘sufficiency’ implies a judgment of adequacy relative to a need, while Aristotle measures ‘exactness’ with reference to an objective standard. If so, eo ipso ‘a definition that suffices’ cannot stricto sensu be objectively ‘accurate’. If one wonders how precision renders a definition inadequate, the answer must be that the philosopher judges it counterproductive, exhibiting a prolixity that inhibits, rather than facilitates, discussion: ‘But one must … not look for precision in the same way in everything, but in each case in accordance with the underlying matter, and as much as belongs to the method … . One should proceed in the same way in other cases too, so that the side issues do not overwhelm the main ones’ (Nicomachean Ethics 1098a26–33). Thus, the two adjectives joined by repeated μήτε at Rhetoric 1369b32 are not like each other: obscurity is undesirable; precision, if ideally desirable, is impractical and in the real world even counterproductive. If we avoid both the undesirable and the impractical, we end up with something that is adequate, sufficient. Cf. Nicomachean Ethics 1094b11–14: “Now we must be satisfied (ἱκανῶς) with the statement of our science, if its distinctness be in proportion to the nature of the subject matter. For exactness (τὸ ἀκριβές) is not to be expected equally in all reasonings, any more than in all the productions of art” (translation by Grant 1885:427).

[ back ] 140. Aristotle often evinces an unflattering attitude towards the promiscuous multitudes. In a crowd one is bound to find some individuals who cannot be persuaded even with the most exact knowledge, and on this account the speaker must resort to what is commonly held (Rhetoric 1355a24–29); the audience is presumed incapable of reasoning from a distant point and of arriving through many stages at a comprehensive view (1357a3–4); the judge is assumed to be a simple man (1357a11–12); maxims are of great help because hearers are vulgar (1395a32–b2); the uninstructed are more persuasive with the crowd than the educated (1395b26–28); a proem’s raison d’être is to address remarks outside the argument to the base hearer (1415b5–9); the weakness of the audience makes lengthy interrogation inadvisable (1419a18).

[ back ] 141. Cf. Rhetoric 1356b26–28.

[ back ] 142. Cf. Rhetoric 1356a1–4.

[ back ] 143. I think it is in that sense that Rapp (2002:2.940) writes ad Rhetoric 1414a8–9: “Das ist eine der wenigen Stellen, an denen Aristoteles andeutet, dass die besonderen Bedingungen der öffentlichen Rede auf Seiten der Zuhörer nicht nur aus den intellektuellen Unzulänglichkeiten der einzelnen Zuhörer … herrühr[en], sondern auch aus den besonderen massenpsychologischen Umständen öffentlicher Reden.”

[ back ] 144. When the peculiar nature of the imprecision that σκιαγραφία metaphorically represents is not understood, interpreters are not able to explain why precision would be not just wasted effort but, in actual fact, worse.

[ back ] 145. If one should ask why Aristotle does not also assimilate the dicanic style to a σκιαγραφία, the reason must be that it is susceptible of comparatively greater precision. This is so, first, because its subject matter deals with the factual past and not the uncertain future (1358b13–17; cf. 1354a26–28) and is therefore inherently more precise. And, second, because the judge’s own welfare is not immediately implicated by the judicial proceedings and, for this reason, his viewpoints, convictions, and proclivities are more immediately relevant to, and more likely to be engaged by, the stylistic shape of the argument. The unprincipled orator may seek to exploit the more precise stylistic fit of the delivery to the audience by speaking off subject. Precisely this scenario is in view at 1354b31–1355a1: yet one must not take this passage as Aristotle’s considered opinion about the proper working of judicial rhetoric; nor infer from it what rhetorical practice, including style, he would advocate. If his disheartening assessment of a jury’s propensity for unscrupulous behavior (οὐ κρίνουσιν 1355a1) might be taken to imply that the speaker should adopt a singularly emotional delivery, this per se does not make the dicanic style less precise than the demegoric. One should also note that Aristotle himself ascribes to both the symbuleutic and the forensic settings audiences whose judgment is obscured by love, hate, and personal interest (cf. 1354b8–11). At any rate, the converse of Aristotle’s dim view of actual forensic rhetorical practice is the promise of securing a more effective hearing before an audience by a more precise stylistic fit. A final thought: the adverb παντελῶς hints perhaps that after all the metaphor of ‘shadow painting’ is also applicable to the dicanic style, only in a less ‘absolute’ sense.

[ back ] 146. Cf. Rhetoric 1358b10–11. Rightly, Kassel ad loc. contra Ross, Roemer, and others.

[ back ] 147. Rhetoric 1358b3–5.

[ back ] 148. Just as a more emotional style—a style, therefore, toward the agonistic end of the stylistic spectrum (cf. 1413b9–10)—may entail more of a given device (at 1408b10–12, for example, the use of compound words, more epithets, and unfamiliar terms). On ἀκρίβεια and stylistic concision see O’Sullivan 1992:44.

[ back ] 149. ἔτι δὲ μᾶλλον ἡ ⟨ἐν⟩ ἑνὶ κριτῇ (1414a11).

[ back ] 150. Aristotle cannot mean that there is no ἀγών between opponents at law, for that controversy remains, whether there be one judge or many. That members of the jury did not in any formal way debate each other and that they voted immediately upon the close of the arguments is immaterial. Quite apart from the fact that individual jury members could, and probably did, speak to those seated nearby, their yelling approvingly or disapprovingly (as the case may be) as well as shouting out their reactions and engaging in other forms of disturbance—practices that speakers often complained about—should more than suffice to motivate Aristotle’s view that forensic debate (the ἀγών of trial) was not to be restricted to the opposing arguments of defendant and plaintiff. Either party must have keenly felt the divergent pull of competing interests espoused, and convictions held, by diverse members of their audience. It would have been in their best interest to seek to gratify, and to avoid antagonizing, any jury member. The ensuing pull of potentially contradictory impulses must have been felt an essential element of the agonistic setting. This pull would be absent in a lawsuit before a single judge.

[ back ] 151. I follow Kassel’s text and understand ἡ δίκη (for ἡ κρίσις) as the implicit subject of ἐστιν with adverbial ἐλάχιστον: ἐλάχιστον γάρ ἐστιν [ἡ δίκη, i.e. ἡ κρίσις] ἐν ῥητορικοῖς. ῥητορικά are ‘rhetorical devices and strategies’, broadly understood (cf. 1375a8). For the superlative neuter singular ἐλάχιστον effectively used as an adverb, cf. Aristotle Historia animalium 494b15–16 and Meteorologica 343b21–22. To judge from his translation, Kennedy also understands the syntax thus. An alternative construal with ἐλάχιστον as the subject largely yields the same meaning: ‘For [then] there is least [sc. scope or efficacy vel sim.] in rhetorical devices’.

[ back ] 152. Innes (2007:162–163) suggests that the historical background to Aristotle’s ‘trial by one judge’ is the process of arbitration (see her references ad loc. and Todd 1993:123–125, 128–129, also with bibliography). This is certainly possible, for although the term for ‘arbitrator’ was διαιτητής (not δικαστής) it not only seems reasonable to subsume him under the more general figure of a κριτής, and the arbitration procedure, under δίκη (so in Athēnaiōn Politeia 53.2), but Aristotle himself in De caelo 279b10–12 writes that ‘those who intend to judge the truth adequately must be arbitrators and not legal adversaries’ (καὶ γὰρ δεῖ διαιτητὰς ἀλλ’ οὐκ ἀντιδίκους εἶναι τοὺς μέλλοντας τἀληθὲς κρίνειν ἱκανῶς). This shows that the terminology was flexible enough to subsume under the respective labels of ‘judge’ and ‘trial’ the figure of the ‘arbitrator’ and the procedure of ‘arbitration’. Add to this the fact that Athens featured forty magistrates (οἱ δικασταὶ οἱ κατὰ δήμους Athēnaiōn Politeia 48.5; cf. Demosthenes 24.112) who could render summary judgment in cases that involved small sums (Athēnaiōn Politeia 53.1; cf. Todd 1993:129). These too could be added to the historical background, if one be needed, behind the philosopher’s one-judge δίκη. I do not think, however, that Innes is correct to infer the absence of ἀγών from ‘arbitration’. The very fact that it was subsumed under ἡ δίκη argues otherwise. At any rate, it is also possible, with Rapp (2002:2.937), to consider the absence of ἀγών simply a Gedankenexperiment that allows Aristotle to demonstrate what is peculiar to the forensic setting vis-à-vis the demegoric. This ideal setting shows all the more clearly why the dicanic style is relatively more precise.

[ back ] 153. Cf. in this connection Innes 2007:162–163, who draws attention to Quintilian Institutio oratoria 11.1.44, a passage that “ridicules the idea that the orator might speak overwhelmed by emotion ‘in minor matters and minor cases … speaking seated before an arbitrator’” (her ellipsis).

[ back ] 154. Even though in my reading εὐσύνοπτον at 1414a12 regards not the judge but the speaker, the mere fact that it is used at all marks the conceptual distance traveled from Rhetoric 1357a3–4 and hints at the truth of Rapp’s surmise that here Aristotle has in mind “ein einziger (vernünftiger; vgl. 1354a34–b1) Richter” (Rapp 2002:2.937).

[ back ] 155. I do not mean to revalidate the dichotomy between ὑπόκρισις and λέξις that I have labored to refute in this book; nor to vindicate a hard dualism of form and substance. My point is only one of relative emphasis. For Aristotle, persuasion through character and emotion can, and ideally does, regard the interests of truth and corresponds to the facts of the case.

[ back ] 156. Cf. Rapp 2002:2.939 and Innes 2007:162 with n. 41.

[ back ] 157. Cf. Kurz 1970:128 and Barnes 1994:190.

[ back ] 158. Cf. Metaphysics 982a25–28 and 1078a9–14.

[ back ] 159. The consensus tacitly equates τὸ ἀλλότριον [τοῦ πράγματος] to τὰ ἔξω τοῦ πράγματος.

[ back ] 160. To vindicate the claim one has to assume that the judge, because he is one, is eo ipso more proficient, perceptive, skillful, etc., an assumption for which there is no motivation other than the need to rescue the otherwise flawed reading.

[ back ] 161. This approach meets Innes’s desideratum (2007:162) that Rhetoric 1414a10–11 not move from stylistic precision to precision of argument. And because it does not equate τὸ ἀλλότριον [τοῦ πράγματος] with τὰ ἔξω τοῦ πράγματος it does not appear inconsistent with (or contradictory to) Rhetoric I.1 (1354b27–1355a1), an appearance that Rapp (2002:2.936) probes: in this passage, Aristotle regards speaking off subject less serviceable in assembly than in the law-court because members of the former have personal interest in the matter under discussion while ordinarily jurymen do not. If the exclusion of τὰ ἔξω τοῦ πράγματος is the exclusion of the ἀλλότριον [τοῦ πράγματος], and the latter exclusion makes rhetorical style more exact, then one expects the demegoric style to be more precise than the dicanic, because ex hypothesi it will have less frequent recourse to (the, for the setting, less serviceable) speaking off subject. This inference seems confirmed by 1355a2–3: ‘There [sc. in the assembly] the judges themselves sufficiently look to this [sc. precluding the speaking off subject]’. But the facts are the inverse of this expectation. Note that the use of οἰκείων at 1354b30 and ἀλλοτρίων at 1354b33 regards a standard (of judgment) and a point of view identical to the ones I claim for the statement in 1414a12–13: they have in view what pertains or does not pertain to the κριτής—only, here the sphere is the subject matter, not the speaker’s stylistic practices.

[ back ] 162. I.e. εὐσύνοπτον [sc. τῷ λέγοντι vel sim.] Cf. Politics 1323b6–7.

[ back ] 163. On the text of 1354b29, see Grimaldi 1980:17.

[ back ] 164. This proves that the absence of agōn at 1414a13 cannot refer to the judge’s freedom from prejudice on the grounds of his lack of personal involvement in the matter at issue.