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.

13. Rhapsodic hypokrisis and Aristotelian lexis

13.1 Why Aristotle on Ὑπόκρισις Matters

‘The matter of hypokrisis’ (lit., ‘what is connected with hypokrisis’) is here said to have come ‘late’ to rhapsodic delivery. This observation seems to mark a development in the manner of performing Homeric poetry—for, whatever else ῥαψῳδία denotes, it must at least denote this—and hence the report has a prima facie claim to our attention in a study of the epic rhapsode and his craft. The coordinate mention of τραγική makes clear that what has come late to rhapsody is not the exercise of ‘delivery’: τὰ περὶ τὴν ὑπόκρισιν came late to tragic drama too, which could not exist without on-stage delivery (surely, in the context of drama, ὑπόκρισις must be what the ὑποκριτής does). The two statements that follow make clear that the philosopher is speaking about the study and formal instruction of ‘delivery’: ‘For initially the poets themselves used to act their tragic plays’. We must remember that early plays featured only one actor. [5] Thus, if a talented poet himself acted his plays, he had no reason critically to reflect upon and write about the principles that made for successful delivery. No one needed to learn them from him. Doubtless, Aristotle assumes the poet’s natural gift not only for composing but also delivering his lines effectively. [6] A second statement helps to clarify the meaning of τὰ περὶ τὴν ὑπόκρισιν: δῆλον οὖν ὅτι καὶ περὶ τὴν ῥητορικήν ἐστι τὸ τοιοῦτον ὥσπερ καὶ περὶ τὴν ποιητικήν, ὅπερ ἕτεροί τινες ἐπραγματεύθησαν καὶ Γλαύκων ὁ Τήϊος (1403b24–27). There is ‘that sort of thing’ in relation to rhetoric which also exists with reference to poetics; [7] which Glaukon of Teos and some others have taken in hand (in connection with poetics). Evidently, in view is studying, teaching, and writing about poetics. And οὔπω δ’ ἐπικεχείρηται (1403b21–22) points in the same direction: ‘not yet taken in hand’ does not address the attempt to practice delivery, but its analysis and instruction by scholars. [8] What could the philosopher mean when he writes that ὑπόκρισις has come to rhapsody late? Pointing to the paring of ῥαψῳδία and τραγική, someone might conclude that what Aristotle has in mind is the theatricality of rhapsodic recitation: an exaggerated stage presence, with overdone histrionic vocal intonation, gestures, attire, perhaps even too ‘mimetic’ an impersonation of Homeric characters when reciting their speeches. The thought of the passage would be construed thus: at length, under the influence of tragic performance, someone started to reflect upon dramatic technique—on what made for effective delivery on stage—and brought in turn such reflection, in writing and teaching, to bear on rhapsody. It is true that as early as the fourth century BC (if not before) tragic drama exerted such an influence upon rhetoric and rhapsody, with the corresponding overemphasis on histrionic delivery. But restricting Aristotle’s meaning to this alone would drastically impoverish his thought, not only in regard to what he says about rhetoric but also, more germanely, what he says by implication about rhapsody. My argument is simple to state, though it takes the following two chapters to make it convincingly: the philosopher’s view of ὑπόκρισις is not simply the superficial one of dress, gestures, and emotive vocal delivery, which a critic of the encroachment of acting on the practice of orators and rhapsodes might decry. Vocal delivery, to be sure, is involved (for he says that ὑπόκρισις is in the voice, 1403b27); and πάθη are certainly in view, for his focus is on how to use the voice ‘for each pathos’ (πρὸς ἕκαστον πάθος 1403b27–28). But his intent is to study ὑπόκρισις as a most powerful means to the end of persuading an audience. ‘Delivery’, then, is an essential aspect of the orator’s task, a crucial element of rhetoric, and it must be diligently considered from the point of view of the civic psychology of emotions that the Rhetoric undertakes to investigate.

The inquiry that follows seeks to address two principal objections. The first, that Aristotle in fact does not think of delivery in the terms just outlined: as an essential part of the oratorical task that needs to be understood, embraced, and practiced by one who desires to succeed as an orator. Rather, critics claim, Aristotle thinks delivery inherently an ethically objectionable matter that is best set aside and disposed of; accordingly, his goal in Rhetoric Ⅲ.1–12 is not ὑπόκρισις, ‘delivery’, but λέξις, ‘style’—and a notion of style, at that, purged from the moral stain of ties to delivery. Although this stance is variously advocated, it finds particular expression in a view of φαντασία at 1404a11 that contemptuously glosses it as ‘mere appearance’, ‘outward show’, vel sim. A second chief objection, built upon the first, is that the statement about writing at 1404a18–19 has nothing to do with delivery but solely, if anything, with style. By stating his case thus, the critic has implicitly divorced style and delivery, has made them separate in Aristotle’s thought and treatment, and has mapped his views on style onto our modern own—views, in our case, that are inextricably linked to the written word and hence connote matters that can be satisfactorily captured by what is on the page, with only a derivative reference (if any) to vocal utterance or performance.

These views are so insidious—if only, because they so readily fit the mold of our own thinking and reflect a longstanding scholarly consensus—that it takes much effort to undo them. To that end, I undertake a comprehensive rereading of Rhetoric Ⅲ.1–12 that places the philosopher’s thought back in its own historical context, without distorting anachronisms, however appealing or natural to us. [10] To anticipate my results: in Aristotle’s thought λέξις and ὑπόκρισις cannot be divorced, nor are they, in fact, divorced in his actual treatment. Delivery (as an element, or even the characteristic shape of, rhetorical style) is in view all throughout these twelve chapters; φαντασία is not ‘outward show’, but the soul’s [re]presentational device that mediates between sense perception and man’s critical faculties; and hence, the writing that is in view in 1404a18–19 is an element of the orator’s delivery: a technique only recently introduced and used with a view to delivery. It is in that light that we must in particular read chapter 12 of Rhetoric Ⅲ. Only after establishing that this is Aristotle’s thought regarding oratorical delivery may one bring the investigation to bear upon the rhapsode and his training and practice (see above, chapter 9). For if the γραφόμενοι λόγοι of 1404a18 had really been a matter of style tied to the written word and not to delivery, one could legitimately challenge its applicability to rhapsodic performance, which, after all, is concerned not with the written word but with vocal utterance.

13.2 Relationship between Λέξις and Ὑπόκρισις

Rhetoric 1403b18–22 marks the transition to the new subject:

The ‘natural order’ (κατὰ φύσιν) of the inquiry had led to what was ‘first by nature’ (ὅπερ πέφυκε πρῶτον), τὰ πράγματα, specifically, ‘whence it gains τὸ πιθανόν’—this, clearly, a reference to the study of πίστεις in Books I and Ⅱ. The headings that follow under δεύτερον and τρίτον are not a recapitulation of the basic outline of the Rhetoric (πίστεις, λέξις, and τάξις), as is the opinion of those who equate τὸ ταῦτα τῇ λέξει διαθέσθαι with πῶς χρὴ τάξαι τὰ μέρη τοῦ λόγου. For the verb διατίθεσθαι does not mean here ‘to arrange’ (it is therefore not a synonym of τάξαι); indeed, if arrangement were in view, with ‘style’ playing the organizing principle, we would expect κατὰ λέξιν (or similar) instead of the instrumental τῇ λέξει. [
15] Rather, the DGE s.v. διατίθημι B.Ⅱ.2 correctly cites Rhetoric 1403b20 under ‘explicar, exponer, narrar’, placing the emphasis not on the structure but on the character of the speech in view. [16] Thus, taking at face value the statement of intention at 1403b15 περὶ δὲ τῆς λέξεως ἐχόμενόν ἐστιν εἰπεῖν, we should read the δεύτερον and τρίτον that follow the reference to the foregoing material (the πρῶτον) as a twofold conceptual division of λέξις—that second great head in the overall outline of the treatise as we know it today. [17]

It is at this point in the text of Rhetoric Ⅲ.1 that scholars start sensing difficulties with its terminology. For some render ἐπεὶ καί ‘since even’ at 1403b36, as if the philosopher’s treatment of ὑπόκρισις marked a detour, opening a parenthesis on ‘delivery’ before the matter of λέξις (properly considered) is finally taken up in section eight. [22] So, for example, Kennedy 1991 renders the passage as follows: “An Art concerned with [the delivery of oratory] has not yet been composed, since even consideration of lexis was late in developing.” [23] When it comes to 1404a8, however, he translates τὸ μὲν οὖν τῆς λέξεως as “the subject of expression,” adding in a footnote, tellingly: “Lexis, here apparently including delivery.” In other words, whereas he had thus far sought to separate λέξις and ὑπόκρισις, now at last he must bring them together. Kennedy is forced to make inconsistent exegetical choices—lexis now excluding, now including, ‘delivery’—because he wishes to restrict the characterization of φορτικόν at 1404a1 to ὑπόκρισις (hence also the “since even” that suggests a detour). But now, at 1404a8–9, Aristotle speaks of the ‘small necessary place’ of λέξις in all teaching, a regrettable consequence (as the μὲν οὖν makes clear [24] ) of the corruption of the audience (διὰ τὴν τοῦ ἀκροατοῦ μοχθηρίαν). If φορτικόν applies only to ὑπόκρισις, and μοχθηρία in the hearer calls for λέξις (in a limited way, as a concession to weakness), then it follows that now λέξις must include ὑπόκρισις. But another option (defended here) opens before the interpreter: that Aristotle’s view of λέξις, from the vantage point of an ideal ethics, is more radically negative than commentators usually allow, and that φορτικόν applies not only to ὑπόκρισις but to λέξις as a whole.

Those who accept the suggestion of a detour usually blame the undeniable awkwardness of the resulting thematic outline on an unfinished (or unpolished) redaction, either by Aristotle himself or by later members of the Academy who allegedly merged two originally independent works into a single treatise. [25] But the suggestion of a detour or parenthesis cannot be sustained. As argued above, following upon his clear statement of intention, Aristotle sets out to discuss λέξις, of which he reckons ὑπόκρισις a subordinate division. From our modern point of view, this choice is by no means self-evident: why should ‘oral delivery’ (if we accept this as a valid tentative rendering of the term) be considered part of ‘style’? It is quite possible to think of ‘style’ as including only what can be immediately conveyed by the written word, without embracing matters more narrowly connected with oral delivery. This is the way most interpreters think about the items covered in chapters 3–11: [26] word choice suitable to prose, metaphors, similes, frigidity, grammatical correctness, conciseness, appropriateness, prose rhythm (for loud reading), periodic style, elegance, expressiveness, use of proverbs, and hyperbole. This, we find eminently possible to do because of the predominance in our conceptual universe of the written word (and, hence, of written discourse) over performance. [27] We think of the word, first and foremost, as reified, on the page (so to speak), and ‘style’, therefore, as subsuming what belongs to the literary study of written texts—with voice and its qualities logically falling under some other head, ‘delivery’, ‘acting’, or similar. My argument here resists this proclivity by highlighting the discontinuity between Aristotle and later ancient scholars in the way they articulated the relation between λέξις and ὑπόκρισις. Although I do not fully subscribe his argument and conclusions, at least in this regard I agree with Graff 2000: “[W]hile it is true that style (λέξις, elocutio) would become a regular component in the major rhetorical systems after Aristotle, its status is less clear in works prior to the Rhetoric and also in the Aristotelian formulation of the art” (4–5). The discontinuity, however, is not simply one of content, of topics covered. It arises primarily from the emphasis the philosopher placed on the oral and aural dimensions of style and from his corresponding focus on delivery as its controlling τέλος. But as writing acquired greater pedagogical and cultural prominence, it did not take long before λέξις came to mean what we ourselves generally understand by style: the diction and composition of the written word, a set of formal qualities at times even better suited for reading than hearing. This development is already clear in Demetrios’ On Style: [28] he shows no interest in the management of the voice, and references to ὑπόκρισις are restricted to stage acting [29] or to short, passing observations about the γραφική and ὑποκριτικὴ λέξις, [30] comments that turn on the use of σύνδεσμοι and are largely derivative of Aristotle (Rhetoric Ⅲ.12). Typical of his attitude is On Style §195, where he abruptly cuts off his analysis of the scope for acting in a particular scene of Euripides’ Iōn with the words: ἀλλ’ οὐ περὶ ὑποκρίσεως ἡμῖν τὰ νῦν ὁ λόγος. [31]

This is not Aristotle’s view: though the destabilization of the oral culture of ancient Greece was already under way in his own time (and, to go no further, chapter 12 of Rhetoric Ⅲ may well serve as witness), ὀνόματα, ‘words’, are for him still primarily mimetic and a matter of the φωνή: ‘The poets, as is natural, were the first to set [this] in motion; for words are instances of imitation, and the voice also, the most mimetic of all our constituent parts, was [ready at hand]’. [32] This statement, the cause of much dismay for some scholars, as one writer notes, [33] has nothing to do with the Platonic theory of words in the Kratylos, but is rather an admission of the degree to which, even as late as Aristotle, ποιητική and ῥητορική (‘literature’ and ‘oratory’, to use our terms), not to mention ῥαψῳδική and ὑποκριτική, were all preeminently oral professions, and scholarly study of these would not have easily dissociated their ‘strictly oral’ dimensions (e.g. intonation) from their (from our cultural perspective) ‘more literary’ qualities (say, the use of tropoi or grammatical correctness). A careful reading of Rhetoric Ⅲ.3–12 will uncover many instances where the oral dimension of delivery is clearly in view, [34] well beyond what a study of style strictly bound to the written word would lead the scholar to expect. But Aristotle’s perspective is particularly clear in chapters 1 and 12, where (as I will presently argue) ὑπόκρισις, though first apparently limited to the voice and its properties, quickly expands its purview, and soon comes to stand more generally for λέξις—while, reciprocally, λέξις stands for a style that is tailored to, and controlled by, ὑπόκρισις. This move, which has confused many a scholar, should not surprise us: though we might have wished for greater terminological clarity throughout, the philosopher’s usage opens to us a window into his thought, still primarily controlled by, the oral dimensions of rhetoric, which is sooner embodied by the performance of the speaker before his audience than by the written text of his oration; this allows ‘delivery’ to stand for the ‘stylistic shape’ of the performance, including, yes, the use of metaphors, similes, elegance, conciseness, grammatical correctness, and so on.

13.3 Ὑπόκρισις, Not a Detour

That Aristotle actually dealt with matters that fall strictly under the narrow definition of ὑπόκρισις (i.e. those connected with the management of the voice) should be considered a priori plausible. After all, at least on two different occasions he states that no one had attempted a treatise about them: at 1403b21–22 he notes that τὰ περὶ τὴν ὑπόκρισιν had not yet been taken in hand (οὔπω δ’ ἐπικεχείρηται); [35] and at 1403b35–36 he adds: οὔπω δὲ σύγκειται τέχνη περὶ αὐτῶν, where, as mentioned above (§13.2), the αὐτῶν most likely points to the three qualities of the voice (1403b30–31). There is no existing manual (τέχνη), the philosopher says, ‘since the [broader] field of λέξις itself has come up late’ (ἐπεὶ καὶ τὸ περὶ τὴν λέξιν ὀψὲ προῆλθεν 1403b36). Here ὀψὲ προῆλθεν parallels ὀψὲ παρῆλθεν at 1403b23, this latter noting the introduction of ὑπόκρισις into the dramatic and rhapsodic arts. There is a third statement at 1404a12–15, in my opinion to be listed along with the other two; though I will consider it in detail below (see §14.1), I may now advance some of my conclusions here for the sake of clarity in my presentation. The text runs as follows:

ἐκείνη μὲν οὖν ὅταν ἔλθῃ ταὐτὸ ποιήσει τῇ ὑποκριτικῇ, ἐγκεχειρήκασι δὲ ἐπ’ ὀλίγον περὶ αὐτῆς εἰπεῖν τινες, οἷον Θρασύμαχος ἐν τοῖς Ἐλέοις.

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.

The ἐκείνη, as I will argue below, takes λέξις as its immediate antecedent, but does so not to the exclusion of ὑπόκρισις, which is subsumed under ‘style’ as the larger heading: [
36] this conceptual overlap alone accounts for the grammar (which calls for λέξις) and the context (which calls for ὑπόκρισις)—a fact of great significance from which it follows that where ‘style’ is mentioned ‘oral delivery’, its primary subdivision, is preeminently in view. ‘When it comes’, [37] ὅταν ἔλθῃ, clearly picks up on ὀψὲ παρῆλθεν and ὀψὲ προῆλθεν; and ἐγκεχειρήκασι δὲ ἐπ’ ὀλίγον answers to and qualifies οὔπω δ’ ἐπικεχείρηται (1403b21–22). Should there be any need further to establish the tie between this last and the first two passages, note the statement about ἆθλα at 1404a17, with the pointed πάλιν that precedes it, which sends us forward to the καθάπερ clause that follows it and back to the similar comment at 1403b32–34. Aristotle’s focus is on the lack of an appropriate scholarly treatment of ὑπόκρισις and, more broadly, λέξις. In light of this, is it plausible to think that he, too, would have failed to cover ὑπόκρισις in his Rhetoric? After pointing out the gap (twice, if not thrice), would he also have failed utterly to fill it? Scholars by and large seem to think so, and Kennedy (1991:219n8) may serve as illustration; for, translating 1404a12–13 “when delivery comes to be considered,” he adds the footnote: “As it apparently was by Aristotle’s student Theophrastus.” If he is right, how are we to justify the philosopher’s failure? One might perhaps adduce that ὑπόκρισις is not that important after all: but Aristotle calls it ὃ δύναμιν μὲν ἔχει μεγίστην (1403b21). Or one may claim that its study is undesirable, for it is only likely to corrupt the audience. But, to this view and its underlying presumptions (which I examine below and find wanting), Aristotle’s own considered judgment may suffice: ‘Nevertheless, the subject of lexis has some small necessary place in every demonstrative discipline’. [38] But, one may yet protest, what the philosopher does study is λέξις, not ὑπόκρισις, and therefore he does, after all, fill the gap. The problem with this view, however, is that it drives too wide a wedge between ὑπόκρισις and λέξις, wider than the philosopher’s thought and words will allow; and that it fails to account for the two statements that are unequivocally about ὑπόκρισις in its narrow concern with φωνή: thus a glaring failure to address what Aristotle himself owns as being of the greatest moment would stand—and that, without either any explicit admission of what must then be considered by all an intentional oversight or a rationale for this a priori unexpected course of action.

Another option must be considered: that Aristotle did, in fact, treat ὑπόκρισις in chapters 3–12, though in a manner that has so failed to meet the expectations of modern scholars as to lead them to believe that he did not do so. A full demonstration must wait until I consider 1404a12–19 in detail below; but I can now anticipate my conviction that most scholars have read Aristotle with anachronistic expectations, tacitly presupposing that only the presence of material on delivery such as is attested in later works would justify the claim that the philosopher had indeed written, if not a whole independent treatise on it, a least a section of his Rhetoric on delivery. Thus it will be helpful, at this point, to survey the chapters in question and see if there are any sections where voice and its properties—under the subheadings of loudness, harmony, and rhythm—come explicitly into play. As I do, we must remember at all times that the primary application of ὑπόκρισις is πῶς αὐτῇ [i.e. φωνῇ] δεῖ χρῆσθαι πρὸς ἕκαστον πάθος. I am not thereby conceding that only passages that explicitly discuss the voice and the emotions would qualify as a study of ὑπόκρισις, for my main contention is that, for Aristotle, all of λέξις is intimately bound with delivery; but identifying such sections would help to make the point that the philosopher, far from excluding delivery as socially noxious and unworthy of his attention, is clearly engaged with what he has already granted is of the greatest consequence for the practicing orator. [39]

13.4 Ὑπόκρισις, Not Just in Rhetoric Ⅲ.1

When it comes to propriety of style, we are easily misled into thinking of it primarily (or even exclusively) in terms of word choice. [43] Such a view would not be problematic if we reckoned as part of that choice its effect on delivery, i.e. its ensuing sound shape: intonation, loudness, rhythm, the melodic line of the resulting phrase, etc. But for the average modern literate sensibility, word choice is mostly a matter of lexical semantics, and propriety connotes a register suitable to the topic and the character or social standing of the ‘notional speaker’ (often merely the author of the text, whose speech is usually read in silence)—without giving thought to any performative dimensions. Aristotle once again confutes such assumptions; for though he speaks of propriety as ‘contracting’ or ‘augmenting’ the tone—and illustrates his meaning a fortiori from poetry by censuring slaves and youth who use fine language—when it comes to a real-life example, he mentions ‘the voice of Theodoros’ (ἡ Θεοδώρου φωνή, my emphasis): ‘for his [voice] seemed to belong to the one speaking, those [of the other actors] to someone else’ (ἡ μὲν γὰρ τοῦ λέγοντος ἔοικεν εἶναι, αἱ δ’ ἀλλότριαι 1404b23–24). A concentrated focus on the voice is all too easy to understand, considering that the actor was masked, and hence could not change his facial expression to enhance his acting. [44] He could of course gesticulate and move about the stage, but the lion’s share of his dramatic art would necessarily fall to the voice. It is perhaps Aristotle’s conceptual dependence upon the stage as he develops his own ideas about the orator’s λέξις and ὑπόκρισις that explains why he largely failed to include in his study the use of the face, hands, and any appropriate scope for gestures. [45] But the fact remains that, in illustrating proper word choice, as he holds up the goal of a ‘natural art’ that hides its artifice (1404b18–19), he makes clear that his concern is with its impact on the voice, i.e. on the orator in performance.

Another place where φωνή and its properties are clearly in view is chapter 7. We must remember, as I pointed out above (see §13.4), that delivery is primarily concerned with how one should employ the voice in regard to every πάθος. Thus, when we read at 1408a10–11 that ‘style will possess propriety if it is expressive of πάθος and ἦθος and proportional to its subject matter’, we should immediately think of the voice in performance, how its loudness, harmony, and rhythm are to be deployed to achieve these goals. [46] Hence, to speak αὐτοκαβδάλως or σεμνῶς shall be a matter of word choice in the extended sense discussed above: not only with attention to the semantic register involved, but also to intonation, loudness, and any other quality that serves to project a persuasive persona or communicate the requisite feeling. We shall not be surprised that it is λέξις that aims to be παθητική if in Aristotle’s mind, as argued above (and further, below), this term readily subsumes ὑπόκρισις; nor shall we wonder that the list at 1408a16–19 would affirm the propriety of the ‘angry style’ when dealing with insolence, of the ‘indignant’ and ‘reticent’ when handling impious and shameful matters, of speaking with admiration of what is worthy of praise, and humbly of pitiable things. Some readers might indeed miss here the more explicit hints offered by the Rhetorica ad Herennium in its section on pronuntiatio [47] and conclude, with its author, that nemo de ea re diligenter scripsit; or else judge the philosopher deficient in comparison with Cicero’s De oratore Ⅲ §§213–227. But it would be an error to let ourselves be guided by anachronistic notions of what a proper account of delivery should look like and fail to see that, in his own way (admittedly compressed by contrast), Aristotle offers guidance as to the proper voice one must use successfully to convey such πάθη as arise in the presence of insolence, impiety, shame, pity, etc. Apparently, he was satisfied to indicate the connection between a cause and its corresponding emotional response, leaving it to the student to consider how volume, melodic line, timbre, and so on should combine to express anger. [48] The outcome is that ‘the one listening always experiences a like feeling with the one speaking emotionally, even if he says nothing. Therefore many overwhelm their hearers by raising a clamor’. [49] Note the performative setting: the goal is a community of feeling between the hearer and the speaker. [50] Rapp (2002:2.862) calls this the “musikalisch-sympathetische Wirkung der emotionalen Rede,” and he thinks that this outcome of the παθητικὴ λέξις has nothing to do with the artistic (“kunstgemäßen”) emotional arousal considered in Rhetoric Ⅱ.1–11. He is mistaken: for Aristotle makes clear that an inference takes place, an ill-founded one (παραλογίζεται 1408a20), to be sure, [51] but a mental reckoning nonetheless, which on the basis of personal experience (ἐπὶ τοῖς τοιούτοις a21) weighs the truthfulness of the claim that purports to give rise to the emotion displayed. Rapp is wrong in thinking that the mind is irrationally bewitched and thereupon does away with the qualifications set forth in Rhetoric Ⅱ—e.g. those of chapter 8 concerning pity, where we read that “on the whole, [a person feels pity] when his state of mind is such that he remembers things like this happening to himself or his own or expects them to happen to himself or his own”; [52] and that “since sufferings are pitiable when they appear near at hand … necessarily they are more pitiable who contribute to the effect by gestures and cries and display of feelings and generally in their acting [hypokrisis].” [53] Nevertheless, this scholar helpfully draws attention to Politics Ⅷ.5, where μέλη, ‘melodies’, are called μιμήματα τῶν ἠθῶν, an assertion justified by the distinct moods with which the hearers are affected in listening to the several musical modes. The passage makes clear that pathē are in view, [54] and it is followed by a similar statement about the effects of ῥυθμοί (Politics 1340b7–10). The balance is to underscore the potential inherent in the voice’s ἁρμονία and ῥυθμός to communicate πάθη—precisely the stated aim of ὑπόκρισις.

The focus on delivery is unchanged in the section of Rhetoric Ⅲ.7 that follows, where Aristotle considers λέξις as expressive of ἦθος. We may best understand his definition of ἠθικὴ λέξις, namely ἡ ἐκ τῶν σημείων δεῖξις (1408a26), if we heed Labarrière’s observation that the philosopher regularly uses σημεῖα as the units of meaning of φωνή, i.e. of articulate sound, whether man’s or the animals’ (in contrast to σύμβολα, which he restricts to λόγος). [55] And indeed, the fitting style (ἡ ἁρμόττουσα [λέξις] 1408a26) involves not only the actual words selected but also the manner of their utterance. [56] Rapp (2002:2.864) himself points out that there is no ready opportunity for deception by the speaker here (the grounds on which, he claims, Aristotle contemns delivery)—and by doing so, he undermines his own view that the doctrine of this chapter equips the orator with such artifice as will allow him to trick the audience into false inferences. The point, rather, is that only by speaking in character can the orator clothe his subject with persuasion. Rapp may be too quick, however, to assume that the listed categories cannot be blurred; for the orator who boasts of manliness, say, should not display stylistic traits that stereotype or prejudice declare typical of the female sex; nor should a defendant who pleads the naivete of youth clothe his appeal in sophisticated delivery. [57] The goal is successful μίμησις, and the voice—we well know—is the most ‘mimetic’ of our constituent parts. One further item proves the full involvement of voice and face in delivery: if the words uttered are harsh, to avoid the appearance of artificiality and speciousness, one should not also use a harsh voice and countenance. [58]

Chapter 8 of Rhetoric Ⅲ is yet another section where delivery is clearly in view and the canon of naturalism controls the outcome. Prose is to strike a middle course between an ἔμμετρον (‘metrical’) and an ἄρρυθμον σχῆμα (‘arrhythmic shape’): μέτρον would make the λόγος into a ποίημα and this is to be avoided; but the lack of ἀριθμός renders the speech unformed, without boundaries, and therefore unpleasant and beyond our grasp. Aristotle has in view certain metrical shapes for cola, chosen to meet given canons of propriety detailed in this section, which avoid the sort of predictable recurrence characteristic of poetry; [59] these, he calls ‘rhythm’: ‘Rhythm is the quantity of the pattern of lexis, whose sections are the meters; therefore speech must have rhythm, but not meter’. [60] But ‘rhythm’ is, of course, the third of the basic properties of voice (cf. 1403b30–32). This proves that we are once again dealing with ὑπόκρισις in its narrow sense of ‘oral delivery’. The expressed need for σεμνότης in this context—used to disqualify the trochaic meter as ‘too much like the cordax’ (1408b36)—should alert us to the connection between propriety and the voice (intensity, intonation, and rhythm), a connection I have tried to highlight above (see §13.4) as I do again here. The recommendation of iambic as the λέξις of the people (οἱ πολλοί) does not focus on its abstract metrical qualities, but on its status as a performative commonplace: ‘therefore, of all meters people in conversation utter (φθέγγονται λέγοντες) iambics most’ (1408b34–35).

I must still consider ἁρμονία, listed as the second property of voice. There is no agreement about its meaning at 1403b31: we know it concerns τόνοι, whether the pitch is ὀξύς, βαρύς, or μέσος. Some think this refers to pitch accent at the word level, [61] but the textual evidence renders this interpretation too narrow. Cope (1877:3.5) seems to leave his options open, for he translates “accents (or tones of voice)” (his emphasis), and cites as comparanda on “the modulation of the voice in the expression of various emotions” Cicero’s De oratore Ⅲ §§215–219. In Cope 1867:380 he is nearer the rhetorical meaning when he parallels it with the Latin apta compositio. Of course, in its primary musical sense, it stands for “the orderly succession of certain sounds, determined by definite intervals, which appeals to an instinctive sense or taste in the human mind … and constitutes ‘tune’ or ‘melody’” (Cope 1867:380, his emphasis). The difficulty resides in transferring this concept from a musical to an oratorical context. This semantic move is attested elsewhere in Aristotle, specifically, in the Poetics. Indeed, at its first appearance (Poetics 1447a22) it refers to ‘melody’, [62] which explains why μέλος takes its place at a later point (1447b25). [63] But, significantly, at Poetics 1449a28, ἁρμονία designates a quality of prose: “For the iambic trimeter, more than any other metre, has the rhythm of speech: an indication of this is that we speak many trimeters in conversation with one another, but hexameters only rarely and when diverging from the colloquial register (ἐκβαίνοντες τῆς λεκτικῆς ἁρμονίας).” [64] Lucas (1968:86) tersely remarks ad loc.: “‘ἁρμονίας’ is said to refer to the pitch of the voice used by the Greeks in conversation (cf. R. 1403b31). We should have expected rather a reference to rhythm.” Though I would take ‘pitch of the voice’ in the extended sense of ‘melodic contour’ or ‘intonation’ of an utterance, which was arguably the actual effect of the pitch accents at the higher level of the sentence, [65] Lucas acutely senses the intimate relationship in prose between ἁρμονία and ῥυθμός, a relationship in evidence in chapter 8 of Rhetoric Ⅲ. [66]

The vulgate reading at Rhetoric 1408b32–33 is: τῶν δὲ ῥυθμῶν ὁ μὲν ἡρῷος σεμνὸς καὶ λεκτικὸς ς ἁρμονίας δεόμενος, where ‘ς’ is the usual abbreviation for καί. Both textual criticism and sense commend Bekker’s and Spengel’s text (which Cope adopts): τῶν δὲ ῥυθμῶν ὁ μὲν ἡρῷος σεμνὸς καὶ λεκτικῆς ἁρμονίας δεόμενος. [67] Kennedy (1991:238), like Kassel, follows Vettori (1548:519) instead, who argued for the insertion of οὐ before λεκτικός, [68] and translates: “Of rhythms, the heroic [dactylic hexameter] is solemn and not conversational and needs musical intonation.” To explain his translation of ἁρμονίας δεόμενος he adds the note: “That is, it is chanted. In Aristotle’s time rhapsodes no longer used a lyre.” The problem with this view, of course, is that one could hardly be expected to understand ἁρμονία in this unusual technical sense (i.e. ‘melodic contour of non-instrumental chanting’) without further explicit textual support, as its use for ‘tunings’ or ‘attunements’ [69] is by far the commonest technical one, of which the λεκτικὴ ἁρμονία is an attested extension. But by itself, ἁρμονίας δεόμενος would almost certainly be understood in its generic sense as ‘lacking harmony’, or, if in its technical sense, as ‘lacking [instrumental] melody’ or perhaps even ‘needing [instrumental] melody’. ‘Lacking [instrumental] melody’ will not suit the contrast; ‘needing [instrumental] melody’ will not do either, for Kennedy’s translation calls for non-instrumental chanting. As to ‘lacking harmony’, Cope (1877:3.86) rightly says that “[it] is absurd in itself, and contradictory to the evidence of our ears, and all ancient authority.” Of Vettori’s emendation [70] the same writer comments that “[it] leaves ἁρμονίας δεόμενος to explain itself as it best may” (Cope 1877:3.87). But for the substitution of ἀλλά for καί, Roemer’s text [71] is near identical to Spengel’s and superior to Vettori’s. Paradoxically, the meaning I advocate implies Kennedy’s, for lacking ‘the melody of common speech’ suggests that ‘the melody of heroic verse’ (with ‘melody’ in its extended sense) contains non-conversational cadences such as reduced melody would produce. [72] Ross 1959 emends more heavily, [73] but his meaning is much the same in the event. [74]

Thus we have in Rhetoric Ⅲ.8 ἁρμονία and ῥυθμός linked together as related properties of the voice in delivery: the latter denotes the measured delivery of cola for rhythmical effect, which facilitates the hearers’ grasp of the orator’s meaning; the former denotes the melodic contour of the speech, [75] correlated with the orator’s rhythmic utterance, shaping the speech to express the necessary ēthos and pathos. This interpretation finds support in Aristoxenos, who explicitly recognized a kind of prose melody: λέγεται γὰρ δὴ καὶ λογῶδές τι μέλος, τὸ συγκείμενον ἐκ τῶν προσῳδιῶν τῶν ἐν τοῖς ὀνόμασιν· φυσικὸν γὰρ τὸ ἐπιτείνειν καὶ ἀνιέναι ἐν τῷ διαλέγεσθαι. [76] This same author also observed that ‘there are three things that can be made rhythmic: utterance (λέξις), [77] melody (μέλος), and bodily motion; thus utterance will divide time by means of its own parts, e.g. letters, syllables, words, and all such things’. [78] A fragment of Theophrastos in Plutarch’s Quaestiones convivales I.5.2 (623a) further reinforces the analogy between the melody and rhythm of music and the melody and rhythm of speech: [Θεόφραστος] ἀρχὰς μουσικῆς τρεῖς εἶναι λέγει, λύπην ἡδονὴν ἐνθουσιασμὸν, ὡς ἑκάστου τούτων παρατρέποντος ἐκ τοῦ συνήθους καὶ ἐγκλίνοντος τὴν φωνήν (fr. 90 Wimmer). [79] In De oratore Ⅲ §§173ff. Cicero follows Aristotle in connecting numerus (~ῥυθμός) and modus (~ἁρμονία). His usage is not uniform throughout; e.g. at §171, when he first introduces them, he apparently designates rhythm by modus quidam and harmonia by forma; [80] but once he employs the more technical numerus, he makes clear that modus stands for harmonia:

In this case, of course, the vocal melody will depend on notions of intonation and rhythm that are more familiar to us than the ancient Greek modulation, of which tonal accents were so important a component. One can speak similarly of the famous chapters 39–41 of [Longinos] Peri hypsous, written at a time when accentual stress had replaced pitch; and yet these sections preserve a recollection of the Aristotelian connection between ἁρμονία and ῥυθμοί in the orator’s choice and arrangement of words (σύνθεσις). [

This survey, I trust, has shown that, far from excluding ὑπόκρισις, Aristotle has more than once explicitly made voice and its properties the focus of his study in chapters 2–11 of Rhetoric Ⅲ. Considering the notorious thematic compression characteristic of this work, his coverage, in substance and significance, is such as to warrant the view that he does indeed endeavor to instruct the orator in proper delivery, if only we suppress anachronistic prejudices of what an adequate, comprehensive treatment of pronuntiatio by the philosopher should look like. My survey completed, I return to the peculiar Aristotelian relation between ὑπόκρισις and λέξις.

13.5 Semantic Development of Ὑπόκρισις and Λέξις

The term ὑπόκρισις, first employed at 1403b22 to refer to that subordinate division of λέξις that wields the greatest influence and is preeminently concerned with voice and its attendant properties, becomes at 1404a12–19, from the point of view of terminology, synonymous with λέξις itself, as I explain below. [84] This is perhaps not to be wondered at, for by exchanging the terms Aristotle merely uses the most important facet of λέξις to refer metonymically to it. The corresponding semantic development takes place in a passage prompted by the ethically equivocal status of λέξις, which becomes a necessary expedient because of rhetoric’s concern with δόξα:

Rhetoric 1404a1–12

This passage contains a remarkable admission regarding the entire rhetorical enterprise, one that belies the attempt of some to restrict to ὑπόκρισις the unflattering δοκεῖ φορτικὸν εἶναι of 1403b36–1404a1, which has instead τὸ περὶ τὴν λέξιν as its proper subject. [
91] On the contrary, μέγα δύναται (1404a7), as καθάπερ εἴρηται underscores, merely restates ὃ δύναμιν μὲν ἔχει μεγίστην (1403b21), and the correspondence between these two statements presupposes a clear, if qualified, equivalence between τὰ περὶ τὴν ὑπόκρισιν at 1403b22 and τὸ περὶ τὴν λέξιν as the implicit subject carried over from 1403b36 into this section. It is impossible, therefore, to divorce the philosopher’s notions of λέξις and ὑπόκρισις. Accordingly, even Cope 1877 ad loc. must make the following admission: “[N]ot only ὑποκριτική, but the whole of Rhetoric, is directed πρὸς δόξαν. So that φορτικόν here must stand, as it often does, for the vulgarity which is shewn in unphilosophical habits of mind … and, as applied to a study or art, may signify popular, showy, unsubstantial.”

If Aristotle in this section considers λέξις an expedient, strictly speaking it is such only from the point of view of an ideal society. We are not faced here with an avoidable course of action, and therefore an unprincipled choice out of personal ease, selfish gain, or self-promotion. This is not to deny, of course, that a given speaker may harbor ethically dubious motivations; but the philosopher’s point is that even an orator with the purest of intentions must have recourse to oral delivery, because no polis can boast of citizens who will embrace the bare facts of an issue without having to overcome the potential obstacles of misunderstanding and prejudice. The triumph of the ἔξω τοῦ ἀποδεῖξαι [92] —arguably necessary, but, in strict justice, superfluous [93] —is perhaps best exemplified by the abuse to which it is open, e.g. by the base appeal to emotions, a practice, however, not exclusively the province of delivery (strictly considered), [94] but arguably most readily illustrated by the use of the φωνή, namely, πῶς αὐτῇ δεῖ χρῆσθαι πρὸς ἕκαστον πάθος (1403b27–28). [95] Thus, ὑπόκρισις understood strictly as ‘delivery’ becomes the preeminent exhibit in the trial against the potentially unethical facets of λέξις, and the connection made exempli gratia with the upper hand the ὑποκριταί have over ποιηταί (a parallel extended next to the political arena) follows all too naturally (1403b32–35). But even if ὑπόκρισις is ‘exhibit A’ (so to speak), Aristotle wishes to apply the conclusion to λέξις as a whole, as the syntactic agreement of φορτικόν with τὸ περὶ τὴν λέξιν makes clear (1403b36–1404a1), notwithstanding the efforts of many to construe the grammar otherwise. [96] From that point onward, the text is primarily engaged with λέξις (principally in its neuter form, τὸ περὶ τὴν λέξιν [97] or, at 1404a8, τὸ τῆς λέξεως) and calls on ὑπόκρισις only subordinately, to make the wider point concerning λέξις by mutual implication. [98]

According to Aristotle (1404a8–10) attention to style has a small but necessary place in every διδασκαλία, for it makes a difference to clarity (πρὸς τὸ δηλῶσαι), yet not so much (τοσοῦτον)—i.e. its importance should not be overstated—but all this is φαντασία directed towards the hearer. [99] πρὸς τὸ δηλῶσαι must not be overly restricted to intellectual clarity [100] (as Rapp 2002:2.815 implicitly does by referring to “Gedanken” and “die gedankliche Anordnung”): Aristotle has in view such a presentation before the hearer as communicates persuasively to him in full the expressive aims of the speaker, not only conveying clearly his logos (in the limited sense of 1356a3–4) but also projecting a distinct ēthos and by its lucidity affecting the audience with the intended pathē. [101] Thus, at 1404b1–3, where ὁ λόγος is characterized as a kind of sign (σημεῖον γάρ τι ὁ λόγος ὤν) that fails to achieve its proper end unless it make its meaning clear (ἐὰν μὴ δηλοῖ), the philosopher implicitly calls to mind the entire communication process, with the complete circle of its constituent parts (cf. 1408a25–26). [102] It is true that here ‘clarity’ is ascribed to ‘proper words’ (τὰ κύρια ὀνόματα) and stylistic propriety to all the others mentioned in the Poetics 1457b1–3: loan word, metaphor, ornament, neologism, lengthening, contraction, and modification (Halliwell’s terms in his Loeb Classical Library translation). [103] If this appears, on the surface, to restrict clarity to a mere subset of the whole range of stylistic devices (word order among them; cf. 1407b21–25 and 1410a19–22), this is but a distorted impression that fails to account for what the philosopher says or implies elsewhere. So, for example, metaphors, just classified among the τἆλλα ὀνόματα that contribute to propriety of style, are nevertheless ranked at 1404b31–32 with τὰ κύρια and τὰ οἰκεῖα as alone being serviceable (χρήσιμα) to prose style, so that, if one composes (with these) well, “there will be an unfamiliar quality and [the composing may] escape notice and will be clear (σαφηνιεῖ 1404b36–37).” [104] In other words: we must grant a wider role to form, including ornamentation generally, in bringing about the requisite perspicuity; [105] and since form is connected with sense perception, this implies that φαντασία is instrumental to clarity and must be part and parcel of effective communication. The οὐ μέντοι τοσοῦτον at 1404a10–11 (‘yet not so much’) finds its conceptual (negative) correlative in the ἀλλά clause. The one who wishes to claim too broad a role for λέξις, Aristotle says, should consider that it works on the hearer through φαντασία: this fact will help him the better to give it its due measure, neither under- nor overestimating its import. Happily, the philosopher does not leave us to feel our way blindly to the understanding of φαντασία, but has already given us a working definition at 1370a28–30: [106] it is said to be ‘a kind of weak perception’ (αἴσθησις), connected not only with sense perception but also with the mental faculties of memory and hope. Nevertheless, it appears to be the universal assumption of the translators of this passage that by φαντασία Aristotle means ‘external show’, ‘mere appearance’, ‘fancy’. [107] The reasons for this unanimity seem obvious: the appeals to μοχθηρία and φορτικόν (1403b35; 1404a1, a8) arguably call for a negative judgment; the concern with λέξις, Aristotle admits, far from desirable in itself, is but a necessary concession (1404a2–3); ideally, what is not strict demonstration should be superfluous (1404a6–7); the often assumed close tie between δηλῶσαι and conceptual clarity seems designed to render the bulk of stylistic analysis external show, superficial fancy.

13.5.1 Φαντασία, ‘mere fancy’?

Nevertheless, there are good reasons to resist this interpretive consensus. Though perhaps understandable, it is but the fruit of our own prejudices against rhetoric and issues from a live suspicion of what this art portends for the pursuit of truth. Such mistrust has a distinguished ancient pedigree: the mention of μοχθηρία and the use of φορτικόν show that Aristotle is not impervious to it. All the same, I claim that the philosopher’s attitude is far less negative (and, thus, more balanced) than the one that prevails among us, and that if we translate φαντασία as ‘mere appearance’ vel sim. we displace the center of gravity of his corrective and drive it to an unintended extreme. Let us revisit the reasons adduced for the consensus view. One may well regret the reality of μοχθηρία (who would not decry the weaknesses that play into the hand of oratorical abuse seeking to tread the ‘facts’ underfoot?), yet it is but a consequence of the ethical cast inherent in any political [108] process. [109] This is clear from the contrast drawn between rhetoric and geometry, echoed again at 1417a18–21, this time by the opposition between μαθηματικοὶ λόγοι, [110] which have no ἤθη because they lack moral purpose, and the Σωκρατικοί, which do: it is the very nature of ῥητορική—the fact that it addresses the need of the political assembly (deliberative, forensic, or epideictic) to build consensus and manage dissent—that makes not only logos, but also ēthos and pathos its necessary ingredients, and for this reason it holds both the promise of effective democratic governance [111] and the danger of the unprincipled exploitation of man’s prejudice and vanity. [112]

Now, as to the phrase καὶ δοκεῖ φορτικὸν εἶναι, we are supposed to appraise it in light of the ensuing qualification, καλῶς ὑπολαμβανόμενον. Cope 1877 ad loc. renders it ‘and rightly so considered’, adding Vettori’s alternative ‘when considered aright’, which he nevertheless rejects because the former alone “is the more natural interpretation of ὑπολαμβάνειν; which will not in fact bear the meaning assigned to it by Victorius ‘Si vere iudicare volumus’” (his emphasis). If Cope is right, at least he cannot claim a large following, [113] nor is it clear why his is the ‘natural’ translation and on what account Vettori’s meaning (which glosses his translation recte ponderatum) [114] is not allowable: for ὑπολαμβάνειν, as LSJ s.v. Ⅲ.1 states, can mean ‘to take up a notion, assume, suppose’, and hence ‘understand a thing to be so’ or ‘conceive of something in a certain way’. [115] The suggestion, then, is that the label φορτικόν for τὸ περὶ τὴν λέξιν holds in a restricted, yet basic, sense; and it does not take much effort to discover that here, as before, the logic behind this dim view of style is again the interaction between the unprincipled orator and his uncultivated audience, whose weaknesses he finds all too easy to exploit for selfish ends. As noted above (§13.5.1 n. 112), the adjective has already made its appearance at 1395b1–2, but it is the treatment in chapter 26 of the Poetics that brings out most clearly its rationale. There, as the philosopher debates whether epic or tragedy is the superior μίμησις, we learn that the label ‘vulgar’ does not so much inhere in the subject matter or practice at hand as it is intimately dependent on the target audience and the consequent interaction between performer and public: ἡ ἧττον φορτικὴ βελτίων, τοιαύτη δ’ ἡ πρὸς βελτίους θεατάς ἐστιν ἀεί (1461b27–29). In view is the practice of actors and αὐληταί who feel the need to add beyond what is proper or called for, on the assumption that the public will not otherwise notice and understand the performance. Epic, on the other hand, addresses itself πρὸς θεατὰς ἐπιεικεῖς who need not such σχήματα, tragedy πρὸς φαύλους (1462a2–4). But, he adds, the fault is not really poiētikē’s, but should be laid at the door of hypokritikē (1462a5–6). [116]

So far, then, we have seen that μοχθηρία and φορτικόν share a common root concern, one that does not inhere so much in λέξις (and, by implication, ὑπόκρισις) as in the potentially corrupt interplay between orator and audience. On the other hand, if used in a principled way, style and delivery can play a significant role in advancing the principles of justice; for they help rhetoric fulfill its social promise, enabling truth and right to assert their natural superiority. [117] This consideration, by itself, should caution us against rendering φαντασία by so utterly dismissive a gloss as ‘mere fancy’. Several additional reasons concur with this judgment. After all, the entire rhetorical enterprise is branded as πρὸς δόξαν, and interpretive consistency would demand that a φαντασία which is no more than ‘outward show’ tarnish with its stain the very art of oratory. Cope 1877 ad loc. appears to realize this when, quoting Eudemian Ethics I.4.2, λέγω δὲ φορτικὰς μὲν [τὰς τέχνας] τὰς πρὸς δόξαν πραγματευομένας μόνον, he adds: “This I suppose must be meant of arts that have nothing solid and substantial about them, but aim at mere outside show, ostentatious and hollow, πρὸς δόξαν contrasted with πρὸς ἀλήθειαν”; and though parenthetically he glosses πρὸς δόξαν as “directed to τὸ δοκεῖν, mere outward show, not τὸ εἶναι,” he softens the outcome in translation by rendering the offending sentence: “But since the entire study and business of Rhetoric is directed to mere opinion, is unscientific.” In fact, one might even question his comparandum, for Rackham’s Loeb Classical Library translation of Eudemian Ethics 1215a29–30 is not ‘ostentatious’ or ‘hollow’, but the far more neutral “pursued only for reputation.” (Reputation and truth need not be at odds.) Since the semantic range of δόξα is broad enough indeed to allow for ‘show’, ‘ostentation’ (should the context call for it), it might be inadvisable to look for guidance in the Eudemian Ethics, especially when the statement needing clarification contains a sweeping characterization of oratory as a τέχνη and the Rhetoric itself does not fail to provide us with parallels that, I believe, make the present one clear.

I start, then, with the passage adduced by Cope above, when he opposes πρὸς ἀλήθειαν to πρὸς δόξαν to justify his “ostentatious and hollow.” This opposition does, in fact, occur at 1365b1, in the seventh chapter of Rhetoric I, where Aristotle considers greater and smaller in connection with the potential disagreement between opposing parties over the degree of significance of a matter that is the object of debate. The abstract notions greater and smaller are illustrated with particular oppositions, such as often versus seldom, proper versus acquired, or ends versus means. The polarity that now occupies us (‘what has respect to truth is greater than what has respect to doxa’) is just one of these, and the philosopher offers the following definition:

ὅρος δὲ τοῦ πρὸς δόξαν, ὃ λανθάνειν μέλλων οὐκ ἂν ἕλοιτο. διὸ καὶ τὸ εὖ πάσχειν τοῦ εὖ ποιεῖν δόξειεν ἂν αἱρετώτερον εἶναι· τὸ μὲν γὰρ κἂν λανθάνῃ αἱρήσεται, ποιεῖν δ’ εὖ λανθάνων οὐ δοκεῖ ἂν ἑλέσθαι. καὶ ὅσα εἶναι μᾶλλον ἢ δοκεῖν βούλονται· πρὸς ἀλήθειαν γὰρ μᾶλλον. διὸ καὶ τὴν δικαιοσύνην φασὶ μικρὸν εἶναι, ὅτι δοκεῖν ἢ εἶναι αἱρετώτερον· τὸ δὲ ὑγιαίνειν οὔ.

Rhetoric 1365b1–8

A definition of pros doxan [is] what one would not choose if he is likely to escape notice. Therefore also ‘faring well’ would appear preferable to ‘doing good’; for a man will prefer the former even if he escapes notice, but seemingly he would not choose to do good unnoticed. And everything they want to be more than to seem [would appear preferable]; for [all these things are] more pros alētheian. Therefore they say that even justice is a small thing, because seeming [just] is preferable to being [just]; not so with being healthy.

Aristotle could not be clearer: that has respect to doxa which one would not choose when likely to escape others’ notice. The focus is clearly on appearances in a social context—what I might call ‘social pretense’ (in the spatial sense of praetendo) if ‘pretense’ did not carry such negative connotations. The concern is arguably for one’s reputation in society, for affecting the views that others have of us, for creating a social standing or managing our neighbors’ attitudes towards us. To use Aristotle’s own word as I believe he conceives of it, this would be a matter of φαντασία, of how we appear, of being aware of and trying to control the impact that one’s behavior has on the way we come across, of giving expression to a particular ēthos in a manner, if not so technically proficient as that of a professional orator, yet not entirely different in kind from it. [
118] Such φαντασία could, of course, involve empty show and ostentation, but it need not do so; and it has a legitimate claim on our interest as social beings. The examples point this out well: doing good versus faring well, or even justice versus health. No sham is involved in the act of conferring benefits on others: the suggestion is not that it looks as if someone is doing good while he is not; the point is simply that the benefaction is done with an eye on the profit that accrues to one’s reputation with his neighbors. As to the latter opposition (i.e. justice versus health), no one would insist on a necessary connection between doing justice and mere outward show or ostentation. [119] But, clearly, whether a person or action is just is subject to debate and opinion in ways that the physical condition of a man—healthy or diseased—is not. It is in this sense, because it is open to judgment and pretension, that justice is said to be of small value. The overriding concern, therefore, is with reputation, with social appearances (one might say ēthos), and this is precisely what πρὸς δόξαν is intended to convey. [120]

Another reason to reject the extreme translation of φαντασία is the claim at 1404a8–9 that ‘the matter of λέξις has some small [but] necessary part in every διδασκαλία’. Since arguably by this assertion Aristotle seeks to justify the place of λέξις in his study of rhetoric, interpreters have tacitly assumed that διδασκαλία, while no doubt embracing other subjects, must ultimately have ἡ ῥητορικὴ τέχνη in view. The logic of the comment would be as follows: ‘style’ plays a small but necessary part in every scientific and technical domain, for it is part and parcel of making one’s meaning clear; but, if arguendo its status in the art of rhetoric is momentarily left open, in subjects like geometry it obviously contributes, if at all, largely dispensable ornamentation; ergo—one readily infers from such clearer cases—it must also lend rhetoric ‘mere outward show’. Presumably, then, the statement aims to vindicate the orator’s unhappy need to resort to λέξις in his practice. But this logic faces difficulties: with the words ‘in every διδασκαλία’ is not the philosopher claiming that style plays a small but necessary role in teaching any given subject (here, a τέχνη) and not, as tacitly assumed, in practicing it? [121] In other words, if the art of rhetoric is really in view, does not the argument then regard its teacher rather than the professional orator? In fact, it is hard to see how διδασκαλία could stand without comment for τέχνη. Perhaps we are to take ‘instruction’ metonymically for the knowledge imparted, [122] but this would still place a focus on the process of teaching that mystifies the common interpretation. One might have expected Aristotle simply to affirm instead that ‘style plays a small but necessary part in [the practice of] every technē.’ The accompanying comment, διὸ οὐδεὶς οὕτω γεωμετρεῖν διδάσκει, fails to dissipate the puzzle; for οὕτω could be construed with διδάσκει: ‘no one follows this method in teaching geometry’, thus making style the province of teaching; or with γεωμετρεῖν: [123] ‘no one teaches geometers to follow this method’, a translation found nowhere, yet implied by the common interpretive framework [124] according to which style is unnecessary for the geometer. [125] Yet, arguably, something like the common interpretation seems to be required by the context, for, after all, we are dealing with style as a component of rhetorical practice, not with a meta-linguistic assessment of style’s contribution to rhetorical instruction. [126]

A solution to the puzzle, however, lies at hand in the demonstrative nature of oratorical practice. [127] Aristotle might have in mind the centrality of rhetorical ‘proof’ to trial and deliberative assembly (and, in smaller measure, to epideictic speeches). [128] And, in this restricted sense, the speaker can be said to ‘teach’ his audience the relevant facts—just as the geometer can be said to prove a particular theorem. For the former, however, the demonstration is πρὸς τὸν ἀκροατήν, whereas the latter does it, so to speak, πρὸς αὑτόν. Not that the geometrical proof cannot be directed at others; but that, by the nature of its reasoning, it is either correct or incorrect; and, if the former, it is so for one and all—the geometer as much as anyone else. Not so in the case of a rhetorical demonstration, which can be compelling to one, yet fail to convince another. This, I think, is the philosopher’s point and why he speaks of διδασκαλία rather than τέχνη. [129] If so, then, φαντασία cannot be ‘mere outward show’ in connection with superfluous ornamentation. Rather, it concerns the demonstrative force requisite to the rhetorical project. This is also why the philosopher writes, ἀλλ’ ἅπαντα φαντασία ταῦτ’ ἐστὶ καὶ πρὸς τὸν ἀκροατήν, where following Kassel I do not, like others, punctuate with a comma before καί, which I take as adverbial, not conjunctive; [130] for φαντασία is present even in the study of geometry, but it is primarily πρὸς αὑτόν, not πρὸς τὸν ἀκροατήν as in the case of rhetoric or, more broadly, wherever teaching takes place and the instructive and persuasive aspects of any τέχνη are called into action. [131]

13.6 Φαντασία in the Rhetoric

A final reason to oppose ‘mere show’, ‘ostentation’ for φαντασία lies in the use Aristotle makes of this term. [132] The initial suggestion that in Rhetoric 1404a11 φαντασία meant ‘pomp, ostentation’ goes back to Freudenthal’s “Prunk” (1863:17–18) and his attempt to explain the restriction ‘if we are not speaking metaphorically’ in De anima 428a1–4:

Nussbaum (1978:254), the one scholar after Cope principally responsible for popularizing Freudenthal’s idea, [
134] realizes that it renders Aristotle’s comment in De anima utterly trivial: “[H]e seems to be saying, ‘Assuming when we say phantasia we mean the faculty in virtue of which we are appeared to in such-and-such a way, and are not using the transferred sense according to which it means (mere) show, then it can be said that in virtue of phantasia we tell truth or falsehood—whereas to say, “in virtue of ostentatiousness we tell truth or falsehood” would be silly.’ It will be objected that this is a trivial point. But for Aristotle it is never trivial to recognize all the senses of a word …” (her emphasis). This last observation notwithstanding, the point is trivial, and it is hard to believe that Aristotle would have felt the need to preclude such a misunderstanding—not to mention that, as argued below (§13.6), this meaning was simply not commonly available until the much later time of Polybios. [135] Now, this is not the place to conduct a survey of the intricate and extensive scholarly debate on Aristotle’s concept of φαντασία. [136] Some have even questioned whether he held a single, consistent view of its meaning throughout his works. [137] But if one can detect some measure of disjunction between (and at times even within) his various writings, this takes place against a background of overall conceptual coherence. It is, at any rate, clear that the meaning alleged by Cope, and widely accepted by other translators, cannot be paralleled in any other passage of Aristotle. Thus, Wedin (1988:68) notes: “Following Freudenthal, [Nussbaum] remarks that φαντασία can mean ‘(mere) show, pomp, ostentatiousness’ and argues that this is the metaphorical sense meant in [De anima] 428a1–4. The remark on the point of usage is acceptable, but that 428a2 counts as a case in point is, I submit, mistaken. An initial reservation is that only one passage in Aristotle can be marshaled in support of the Freudenthal reading, namely Rhetorica 1404a11” (my emphasis).

But there are some discordant voices. Indeed, while not calling for a gloss so extreme as ‘showy, ostentatious’, similarly tending towards superficiality is Halliwell’s proposal that the term is not used in its “psychological sense, but [taken] to mean merely ‘appearance’, as at Sophistic Refutations 4.165b25” (Halliwell 1993:59n16). [138] My problem with this comment is that his perhaps otherwise unobjectionable “merely ‘appearance’”—where ‘appearance’ can be neutral enough simply to denote ‘what appears to the thinking (or sensing) subject’—seems to connote ‘mere appearance’, which in turn is glossed by ‘show’; thus we move quickly from ‘appearance’ to the objectionable ‘mere ostentation or show’, the very meaning that cannot be substantiated from any Aristotelian passage—unless, of course, one chooses to call ‘mere show’ any appearance that happens to be false (as φαντασία can certainly be). The locus adduced in support is a case in point: ἔστι δὲ τὰ μὲν παρὰ τὴν λέξιν ἐμποιοῦντα τὴν φαντασίαν ἓξ τὸν ἀριθμόν· ταῦτα δ’ ἐστὶν ὁμωνυμία, ἀμφιβολία, σύνθεσις, διαίρεσις, προσῳδία, σχῆμα λέξεως (165b24–27). [139] Doubtless here the ‘appearance’ in question is false: false logic is the focus of the treatise, a point its opening reiterates by referring to οἱ φαινόμενοι ἔλεγχοι (164a20–21) and by distinguishing between real syllogisms and those that seem [true] but are not. [140] There is nothing here of ‘mere ostentation’, for the superficiality of false reasoning is quite another from the ‘sensual show’ alleged at Rhetoric 1404a11. Sophistical Refutations is not, after all, an ethical treatise that looks into the motivations of deceitful sophists in order to condemn them for their ostentatiousness. Furthermore, I fail to see why this instance of φαντασία (or the one at 168b19) should not have “its psychological sense.” [141] For Aristotle himself draws the parallel between the inexperienced, who reasons and refutes falsely, and ‘those who view things from a distance’ (164b27), a formulation strongly reminiscent of the passage in De anima (428b17–22) where, discussing why φαντασία can be false, Aristotle distinguishes between perception of τὰ ἴδια and perception of the αἰσθητά to which these ἴδια belong: [142] “As to the whiteness of an object, sense is never mistaken, but it may be mistaken as to whether the white object is this thing or something else.” [143] This comment must in turn be read against 430b29–30: ‘But just as sight perception of a proper object (τὸ ἴδιον) is [always] true, while [our perception] whether the white thing is or is not a man is not always true, so it is with immaterial objects’. Thus, when Aristotle in Sophistical Refutations mentions the error of ‘those who view things from a distance’, De anima leads me to believe that most likely he has in mind the misleading φάντασμα that results from viewing an object from too far a distance. I might add that the sources of false φαντασία in Sophistical Refutations 165b26–27, “the ambiguity of a term, the ambiguity of a proposition, the possibility of wrong disjunction, the possibility of wrong conjunction, the possibility of wrong accentuation, and similarity of termination” (Poste 1866 ad loc.), have their grounds on aural or visual φαντασία. [144] Such errors come from the application of νοῦς to what one hears or reads (the φαντάσματα), e.g. in syntactic or semantic parsing, where learning (and hence memory) and deductive logic are involved. This would seem to me to fall squarely under the psychological sense of φαντασία. [145]

Another discordant voice is Fortenbaugh (2002:96–100), who, opposing the analysis of human emotions as “phantasia apart from belief,” takes an approach similar to Halliwell’s: he denies any ‘scientific’ intent in Aristotle’s account of φαντασία in the Rhetoric and cites in support Sophistical Refutations 164a20–24. Accordingly, any references to “appearance” would not pertain to the “biological faculty of phantasia” (96): recognizing that courts and assemblies make decisions on the grounds of probabilities, not certain knowledge, the philosopher “is careful to speak of what appears to be the case,” thus “calling attention to the fact that human emotions are caused by beliefs, which may or may not be true” (97). But I fail to see why such a stance would prevent φαντασία from being the psychological faculty more fully (and precisely) discussed in De anima. Of course it is possible to use φαίνεσθαι without implying any particular psychological framework: the word, by itself, will not settle whether its register is technical or colloquial, nor, if technical, the degree of precision invoked. But Fortenbaugh elides the fact that not only the verb but also the noun, φαντασία, is used, for which it is harder to argue a colloquial meaning devoid of any technical import. That Aristotle associates φαντασία and αἴσθησις at Rhetoric 1370a28–30, in fact, seems to militate against a strictly colloquial register for the verb and its noun, even in the context of the oratorical treatise. I would argue, moreover, that, considering the philosopher’s undeniable interest in the epistemological role of φαντασία, one should assume, ceteris paribus, that in the rhetorical context of truth-seeking and decision-making φαίνεσθαι is more likely to bear a degree of technical precision than to be strictly colloquial, ‘unscientific,’ and devoid of psychological overtones. It seems to me that Fortenbaugh’s rationale for a cleavage between De anima and the Rhetoric (see, for example, Fortenbaugh 2002:100) is the distinction drawn in De anima Ⅲ.3 between φαντασία and δόξα, which allows a degree of psychological detachment to a subject who denies his φαντασία conviction (the πίστις that accompanies δόξα), resisting its implications (as when, looking at a picture, we know—however horrifying the depiction—that we need not flee from it as if from imminent danger). But it is precisely this effect of attachment to or detachment from a particular ‘view’ that opponents at law try to induce among the jury. Once rhetorical persuasion is achieved, the ‘picture’ carries the conviction of truth and action follows. Naturally, Aristotle need not raise in the Rhetoric aspects of φαντασία that follow from non-rational animals’ possession of this faculty, since the purview of oratory is strictly logos-endowed man. [146]

Now, it is not only the case that the works of Aristotle fail to produce a single instance for which ‘ostentatiousness’ correctly translates φαντασία: in fact such a meaning finds no parallel in the literature before Aristotle’s time. The closest approach is the use of the verb φαντάζειν (not the noun φαντασία) in Herodotos 7.10ε, [147] where, it is true, ‘to show oneself’ might be deemed less compelling than ‘to make an arrogant display of oneself’. But we are simply not entitled to read our view of what is contextually compelling into the received lexical meaning of a term—a meaning, besides, that in this case quite adequately suits the context—unless we can support the corresponding modification of, or extension to, the lexicon by additional instances that clearly require it. Here it is not so. Artabanos’ comment hinges on the contrast between ὑπερέχοντα ζῷα, ‘prominent animals’, and σμικρά, ‘small’: god strikes the former with his thunderbolts and does not suffer them to show themselves—a hyperbole allowed for its exemplary value. Some might draw the inference that ‘prominent animals’ are “creatures of greatness,” wont “to display their pride” (to quote A. D. Godley’s Loeb Classical Library translation): in their eyes, this may color φαντάζεσθαι, whose proper meaning is ‘to appear’, with a note of arrogance. But if so, such semantic coloring is strictly contextual and this instance by itself should not receive its own lemma in the lexicon (LSJ s.v. Ⅱ.2). To be sure, such hypothetical contextual color would resemble the late use of φαντασία for ‘outward show’ and ‘ostentation’—for which there are, however, no examples before Aristotle, and which, in my view, the alleged passage of the Rhetoric does not illustrate either. Indeed, even if arguendo one should grant the validity of the acceptation of the LSJ ‘to make a show’, one cannot argue convincingly from what would be, under those circumstances, a single, exceptional meaning of the verb φαντάζειν in Herodotos [148] that its corresponding noun, φαντασία, also in that same exceptional sense, was conceptually available to Aristotle; and, furthermore, that he actually used φαντασία with that anomalous meaning once and only once in his entire oeuvre, departing in so doing from other senses frequently attested elsewhere in his works. I find this unpersuasive, especially when the number of such instances of φαντασία exceed one hundred, giving more than ample scope for potential parallels—parallels that fail to materialize. Are we really to believe on the basis of a precedent so slim that, at best, only one Herodotean passage can be adduced for it that Aristotle departed from the received meaning of the noun and his own usage elsewhere, attested by many passages? Admitting such a departure would only be justified under rather stringent contextual constraints, i.e. only if the local context should categorically demand it. I have already listed several reasons that, in fact, suggest that no such necessity obtains. My case will be further strengthened if I can show that the philosopher’s use of φαντασία elsewhere also suits Rhetoric Ⅲ.1: this I shall do presently. [149] Aristotle’s interest, our critics notwithstanding, is with the psychology of perception and the ways in which φαντασία subserves sense ‘perception’ (αἴσθησις) and ‘supposition’ (ὑπόληψις): ‘For phantasia is different from both perception and thought; it does not happen without perception, nor supposition without it’ (De anima 427b14–16). [150] If rhetoric studies the means for persuasive speech, surely its practical goal is to persuade or dissuade the audience; hence, in its polis setting and at the level of civic action, rhetoric becomes a legitimate object of interest in the study of ‘animal motion’, though here the ζῷον in question is one that possesses φωνή and λόγος and is eminently πολιτικόν. It is in this context that φαντασία and φαντάσματα—their role in voluntary movement—enter into the considerations of the orator and the teacher of oratory.

But before considering the Rhetoric’s own engagement with φαντασία, we must survey statements elsewhere that illuminate the conceptual background of this word. De anima 433b27–30 offers a convenient starting point: ‘Thus, then, in general terms, as already stated, the animal is capable of moving itself just insofar as it is appetitive; and it cannot be appetitive without phantasia. Now all phantasia is rational [i.e. connected with reason] or perceptual [i.e. connected with perception]. Of the latter the other animals also have a share’. [151] Here Aristotle notes that motion must be traced to desire, desire that is not without φαντασία; and that φαντασία can be categorized by its connection with λόγος or perception. [152] Such λογιστικὴ φαντασία recalls a later section of the same work, which speaks of ἡ βουλευτικὴ [φαντασία] ἐν τοῖς λογιστικοῖς (434a7), ‘the deliberative φαντασία in rational beings’: [153] using reason a man decides whether he will do this or that (pursuing the greater good), and proves that he is not moving ἀορίστως (434a4), like the lowest animals, but measuring by a single standard; thus it follows that he is able, from many φαντάσματα, to fashion one course of action. [154] The operative word here is βουλευτική, which highlights the corresponding role assigned to φαντασία in deliberating a course of action. [155] There is only a small distance from this to a corresponding symbuleutic, i.e. social, dimension, as the orator artfully crafts φαντάσματα [156] that will prompt his hearers to ‘do this or that’. This argument precludes the facile criticism that rhetoric merely addresses itself to no more than the passions of the audience, in a manipulative attempt to elicit behavior that is as irrational [157] as it is beneficial to the speaker (just as ‘desire’ is said at times to overpower βούλησις, De anima 434a12–13). No; the text tells us plainly that choosing a plan of action is the work of λογισμός, and, in so doing, promotes φαντασία (with its ethical and emotional components and the λέξις that expresses them) to a cardinal tool of the rhetorical task.

For Aristotle, then, φαντασία—the soul’s [re]presentational [158] device—mediates between sense perception and the critical faculties [159] that, apprehending the object as desirable or undesirable, move one towards or away from it. But αἴσθησις is not the only cause of φαντάσματα; hope and memory too are associated with φαντασία, [160] and more generally λόγος and νόησις: τὰ μὲν γὰρ ὀργανικὰ μέρη παρασκευάζει ἐπιτηδείως τὰ πάθη, ἡ δ’ ὄρεξις τὰ πάθη, τὴν δ’ ὄρεξιν ἡ φαντασία· αὕτη δὲ γίνεται ἢ διὰ νοήσεως ἢ δι’ αἰσθήσεως (De motu animalium 702a17–19). [161] The causal chain is: ‘thought’ or ‘sense perception’ → φαντασία → ‘desire’ → ‘bodily affections’ → motion. [162] As the translation indicates, in the previous passage πάθη stands for bodily changes (chillings and heatings), [163] not for the psychic affections that attend φαντασία and that are studied in Rhetoric Ⅱ (see below, §13.6). These latter πάθη would not follow, but precede, desire; in other words, they would constitute motivations for judgments or actions. [164] Perception and thought are ‘critical’ faculties the soul uses to judge: ἡ ψυχὴ κατὰ δύο ὥρισται δυνάμεις ἡ τῶν ζῴων, τῷ τε κριτικῷ, ὃ διανοίας ἔργον ἐστὶ καὶ αἰσθήσεως … (De anima 432a15–16); [165] and if φαντασία according to Aristotle depends on perception, and thought, in turn, cannot happen without it, [166] it should not surprise us to read the following: “Now we see that the movers of the animal are reasoning and phantasia and choice and wish and appetite. And all of these can be reduced to thought and desire. For both phantasia and sense-perception hold the same place as thought, since all are concerned with making distinctions [lit., ‘for all are involved in judging’ (κριτικὰ γὰρ πάντα)]” (De motu animalium 700b17–21). [167] It can hardly be accidental that, in arguing for the usefulness of rhetoric, Aristotle should censure κρίσεις that happen μὴ κατὰ τὸ προσῆκον (Rhetoric 1355a22–23), implying that “the agent responsible for permitting the bad judgments is a rhetoric which does not achieve its perfection as rhetoric, and thus fails to realize its usefulness” (Grimaldi 1980:27). Though ‘judgments’ here are the decisions of the courts, these are but the social expression (at the civic level) of the individual’s proper use of his own faculties of judgment. [168]

This line of reasoning is of a piece with Aristotle’s division of the soul into two parts: one that possesses λόγος, the other ἄλογον (Nicomachean Ethics 1102a27–28). The former he divides further into the ἐπιστημονικόν, which studies things whose principles (ἀρχαί) cannot be otherwise, and the λογιστικόν, which makes calculating (λογίζεσθαι) and deliberating (βουλεύεσθαι) its task: [169] ‘For to deliberate and to calculate are the same thing, and no one deliberates about what cannot be otherwise’. [170] Drawing our attention to the appearance of τὸ ἐπιστημονικόν at De anima 434a16, Labarrière (1984:30) plausibly argues that the subdivision of the soul that has λόγος is in view throughout this section of De anima, where the philosopher discusses the λογιστικὴ φαντασία. The effect of this terminology is to underline the involvement of λόγος—ratio and oratio—as νοῦς makes use of the corresponding φαντάσματα in forming a ὑπόληψις. [171] Like Plato before him (but with greater conceptual clarity), [172] Aristotle placed desire in the sphere of λόγος, and therefore made it, in some measure, the object of persuasion and rational appeal. We find this clearly stated at Rhetoric 1370a18–27: [173] ἐπιθυμία (which he defines as ‘a desire for what is pleasant’) can be ἄλογος or μετὰ λόγου; [174] the former kind (e.g. hunger, thirst, or sleep) does not come ἐκ τοῦ ὑπολαμβάνειν, ‘from forming an opinion’; of the latter kind, he writes: ‘[I call] “attended by logos” all those things people desire because they have been persuaded; for often, after hearing and being persuaded (ἀκούσαντες καὶ πεισθέντες), they long to see and own’. [175] Thus, the craving in this case comes from persuasion (whence, by implication, one forms ὑπολήψεις)—in particular, from aural persuasion, ‘when they have heard and have been persuaded’. I do not think that ἀκούσαντες and πεισθέντες are conceptually independent, as if the former merely referred to learning about something by word of mouth, while the latter conveys the exercise of reason that results in conviction. Rather, I believe that Aristotle has selected the common scenario of rhetorical persuasion to illustrate the division (only too apposite a choice given the subject matter of his treatise), and that the participles, conceptually subordinated, might be translated thus: ‘for they often long to see and own when they are convinced by hearing [an oral argument]’. This serves well to remind us that, for Aristotle, the rhetorical endeavor is preeminently of an oral (and hence aural) nature.

Given its focus on persuasible desires, it is hardly surprising to find the first occurrence of φαντασία at this point in the Rhetoric (at 1370a28), where it is featured in an argument about the very desires that are open to persuasion and should therefore be of concern to the orator:

Rhetoric 1370a27–32

Since the experience of pleasure inheres in perceiving a certain affection and since phantasia is a sort of weak perception and since some kind of phantasia of what a person remembers or hopes will always attend his remembering and hoping—if this is the case, it is clear that pleasures accompany those who are remembering and hoping, since indeed perception [is involved here] too.

This famous passage, with its αἴσθησίς τις ἀσθενής, does not, of course, collapse the conceptual complexity of φαντασία surveyed above into a facile equation between it and perception. The enclitic τις here, as often, signals a simplification; it warns us of an approximation that, while suitable to the context and argument at hand, yet lacks the philosophical sophistication and accuracy that might be necessary and present elsewhere. [

The connection between φαντασία and αἴσθησις is helped by the semantic range of αἰσθάνομαι, which—Aristotle teaches in De anima Ⅱ.6—covers not only αἰσθητά perceived in themselves, but also what might be called ‘incidental objects’, perceived κατὰ συμβεβηκός. The former are proper, ἴδια, to one of the ordinary senses (e.g. seeing white or tasting sweetness); or common, τὰ κοινά, to all or some (e.g. size or number). The incidental, he illustrates as follows: “An object of perception is spoken of as incidental, e.g. if the white thing were the son of Diares; for you perceive this incidentally, since this which you perceive is incidental to the white thing. Hence too you are not affected by the object of perception as such” (418a20–24). [178] It is clear that perception of incidental objects is of a higher order, in that it calls for processing and integrating with memory an array of data (proper and common). [179] And yet at one level—and certainly in popular parlance—we can still say that we ‘see the son of Diares’ just as we might as well say that we ‘recognize’ him: seeing considers it from the perspective of the senses, recognizing, from that of the mind. It is precisely this double-sidedness of perception that gives rise to the concept of an αἰσθητικὴ φαντασία, and the discourse on φαντασία seeks to understand the individual roles of soul and body and their mutual interplay when such ‘perception-as-realization’ takes place. Only in this sense is it right to say that we feel an emotion: not as a bare affection, as a bee may be said to be angry—for emotion is not a sensory datum impinging on the senses—but by the awareness of a relevant fact through a [re]presentational φάντασμα that is, in turn, attended by pleasure or pain in their various forms. [180] This explains the words ἡ δὲ φαντασία ἐστὶν αἴσθησίς τις ἀσθενής: if αἴσθησις is attended by pleasure or pain, [181] and memories and hopes—which involve φαντάσματα of the things one remembers or hopes for—are painful or pleasurable, then, on the point of analogy, φαντασία is a sort of weak αἴσθησις. [182]

The other instances of φαντασία in the Rhetoric follow the established norm; as Grimaldi (1980:256) notes, through “imagination, the presentative faculty,” φαντάσματα affect our appetitive system. Indeed, at 1370b33 victory is said to give rise to a φαντασία of superiority (ὑπεροχή), ‘which all strongly or mildly desire’; at 1371a9 honor and a good reputation are reckoned among the most pleasurable possessions, because on their account each has the φαντασία ‘that he is such as a worthy man (ὁ σπουδαῖος), all the more when they say so who he thinks speak truly’; at 1371a19, elaborating on why being loved is pleasant, the reason offered is that ‘there too [the one loved enjoys] a φαντασία that being good is his nature, [183] [a thing] which all who perceive it desire’; at 1378b9 the φαντασία attends ‘dwelling upon the thought of revenge’; [184] and at 1382a21, 1383a17, and 1384a23 the word enters into the respective definitions of fear and courage (which involve a future expectation—the ἐλπίς noted at 1370a29—of evil or safety) and shame (‘a φαντασία of loss of honor’). In none of these instances is there an emphasis on the unreality of what is ‘[re]presented’ to the mind; [185] the focus is rather on the cognitive process that consists in entertaining a given notion or idea, a process that brings pleasure or pain to the one who engages in it. The point of view taken is δόξα, [186] often that of a third-party, and there is a keen interest in how our notions are affected by those around us. In one case we even read explicitly that the φάντασμα is what others say about us; and that the persuasion operative with the φαντασία is enhanced in proportion as the speakers are trustworthy—an argument that points the way to the ‘scientific’ (ἔντεχνον) study of reasoned emotional appeal, an appeal that, insofar as it is based on λόγος, has a legitimate place in the art of rhetoric. [187]

De anima 433a10–22 calls φαντασία a kind of thinking (νόησίς τις): like δόξα, it is a critical faculty; [188] but, significantly, in contrast to δόξα, it does not carry πίστις (De anima 428a20–24)—arguably the ultimate goal of rhetoric. This fundamental distinction is one that we must bear in mind as we think of the oratorical task: by affecting their φαντασία, the very [re]presentational system of the διανοητικὴ ψυχή, the speaker shapes the perception of his audience. Such molding influence works at all the available cognitive levels (including the πίστεις inherent in the αὐτὰ τὰ πράγματα, just as even the geometer must use φαντάσματα), but it pertains especially to pathos, as the orator endeavors to secure a desirable outcome or avoid a disagreeable one. Even a speaker’s successful self-presentation, which turns on his ability to communicate proper ēthos, can be reinterpreted as φαντασία towards the hearers with a view to creating such πάθη as will move them to act to the speaker’s advantage. [189] And λέξις, of course, studies which use of λόγος—including, we must always remember, φωνή and its attributes—best conveys the speaker’s meaning, best achieves his goals, inducing such φαντάσματα as will support his case. Thus, he will adopt the ‘style of an angry man’ when there is insolence; the style of man ‘disgusted’ and ‘cautious even to speak’, with regard to impious and shameful matters; he will speak with admiration of what is worthy of praise, with lowliness of pitiable things (Rhetoric 1408a16–19). All these, being matters of style and delivery, will imply an appropriate register for the orator’s voice—a careful choice of intonation, loudness, and prose rhythm—and a fitting countenance that does not raise the suspicion of speciousness (1408b4–7). These visual and aural stimuli translate into φαντάσματα that render style effective in expressing ēthos and pathos (1408a10–11). The resulting community of feeling between hearer and speaker creates plausibility and persuasion: ‘The proper lexis also makes the matter credible: the mind’s reasoning that one speaks truthfully is ill-founded, because when faced with such things people react in this way and consequently think that the circumstances are thus—even if [190] they are not as the speaker [represents them]—and the hearer’s feelings track those of the emoting speaker, even if he is [really] saying nothing’ (1408a19–24). Thus conceived, rhetorical performance, deliberate in its λέξις and ὑπόκρισις, seeks to bring the audience under the spell of the orator: through carefully targeted sensory φαντάσματα—to which voice, gestures, and all the resources of style are instrumental—he attempts to place before his listeners a particular [re]presentation of the facts; this [re]presentation, successfully brought before his mind’s eye, becomes the hearer’s own λογιστικὴ φαντασία (the sensory φαντάσματα now having noetic status). The orator’s design and hope is that the audience will feel the force of this φαντασία—that πίστις will attend the thinking it subserves—and that his hearers will embrace his [re]presentation with their own δόξα. [191] Thus, φαντασία not only denotes the soul’s [re]presentational faculty: more loosely, it may also be used interchangeably with φαντάσματα or even stand for the oratorical technique that manipulates them for rhetorical ends. This technique, Aristotle tells us, works best when, drawing upon his own φαντασία, the orator turns its internal [re]presentations into sensory stimuli of adequate λέξις and σχήματα expressive of his own emotions. Then is his pathos best able to stir the equivalent feelings in his listeners and to excite in them a corresponding φαντασία that becomes instrumental to the art of persuasion. [192]


[ back ] 1. See above, Introduction.

[ back ] 2. See above, Introduction.

[ back ] 3. Ordinarily, I will use the line numbers of Kassel’s edition of the Rhetoric (Kassel 1976). Since these do not always agree precisely with the numbering of the more readily available OCT by Ross, the reader who consults the latter should know occasionally to adjust the count slightly.

[ back ] 4. Rhetoric 1403b21–24.

[ back ] 5. Cf. Aristotle Poetics 1449a15–17.

[ back ] 6. Cf. his assertion at Rhetoric 1404a15.

[ back ] 7. ποιητική presumably includes τραγική and ῥαψῳδία.

[ back ] 8. This understanding receives further support from the related ἐγκεχειρήκασι of 1404a13–15.

[ back ] 9. See above, §9.5.

[ back ] 10. For a recent commentary on this book, see now Burkett 2011. He too joins the consensus: “Unlike lexis, Aristotle views delivery as largely a matter of natural talent …” (26).

[ back ] 11. At Rhetoric 1403a36 the philosopher also refers to πίστεις with the expression τὰ περὶ τὴν διάνοιαν.

[ back ] 12. Though ‘style’ is its usual translation, λέξις denotes broadly how thought is expressed in words. Thus, depending on the context it may be rendered as ‘language’, ‘word choice’, ‘expression’, vel sim. (Kennedy 1991:216). I will argue that oral expression is most often in view.

[ back ] 13. Rhetoric 1403a36–b2 and 1403b7–8.

[ back ] 14. Kennedy’s 1991 translation ad loc. (Kennedy 1991:217–218).

[ back ] 15. Cf. Burkett 2011:25.

[ back ] 16. Though related to it, this acceptation is not the same as the (later common) ‘to set forth’ (LSJ s.v. B.6), a meaning illustrated by διατίθεσθαι with such objects as λόγους, ἐπαίνους, or δημηγορίαν. The DGE collects these instances s.v. B.Ⅱ.1, a division that differs from B.Ⅱ.2 in that it emphasizes the publication of the discourse. Thus, I agree with Cope (1877:3.3) when he dismisses the meaning in publicum proponere, in medium proferre, choices that would make τὸ ταῦτα τῇ λέξει διαθέσθαι near identical with the third head of Aristotle’s list, ὑπόκρισις. Kennedy 1991 ad loc. translates, “how to compose this in language”; Dufour and Wartelle 1973, less literally, “la valeur que leur prête le style”; while Cope helpfully writes that “διαθέσθαι denotes … the investing of the speech with a certain character, putting it in a certain state, by the use of language … . It does not mean here distribution, ordering, arrangement, which is not the special office of the graces and properties of language or style” (1877:3.3).

[ back ] 17. Cope (1877:3.3) glosses τούτων in 1403b21 as “of such things as these, the divisions of Rhetoric.” If he means the threefold division of rhetoric into pisteis, lexis, and taxis, this cannot be right, for then τὸ ταῦτα τῇ λέξει διαθέσθαι would have to stand for ‘arrangement’ (a reading refuted immediately above). It would also follow that for Aristotle τὰ περὶ τὴν ὑπόκρισιν would adequately substitute for an explicit mention of λέξις (a corollary that should discomfort those who seek to divorce these terms). The alternative is to refer τούτων to the discrete series of topics that suggest themselves in their natural order (κατὰ φύσιν 1403b18) to the scholar of rhetorical art. On this view, the philosopher might have gone on to taxis in his enumeration, but he stops at hypokrisis in order to fulfill his promise to consider lexis next (1403b15).

[ back ] 18. Cf. the use of κυριώτατος at Rhetoric 1355a7, 1356a13, and 1358b17.

[ back ] 19. With Cope and the anonymous commentator (Anonymi in artem rhetoricam commentarium, p. 159 ad loc., ed. Rabe), I read αὕτη at 1403b27. As Cope observes, “αὐτή seems to have no meaning here” and it is regularly translated as if it read αὕτη.

[ back ] 20. The feminine gender is used throughout, which makes clear that φωνή, ‘voice’, is still conceptually in view under the plural τόνοι.

[ back ] 21. οὔπω δὲ σύγκειται τέχνη περὶ αὐτῶν, ἐπεὶ καὶ τὸ περὶ τὴν λέξιν ὀψὲ προῆλθεν· καὶ δοκεῖ φορτικὸν εἶναι, καλῶς ὑπολαμβανόμενον (1403b35–1404a1).

[ back ] 22. The logic would be a fortiori: lexis appeared late, hypokrisis the more so. Kennedy 1991:217 writes: “A third beginning is then supplied in section 3 [of Rhetoric Ⅲ.1], followed by some remarks on delivery, and the actual discussion of lexis does not begin until section 8” (my emphasis).

[ back ] 23. Despite Kennedy’s acknowledgment that “lexis … refers to the ‘way of saying’ something” (1991:216, my emphasis), his perspective as a modern scholar naturally gravitates towards a reading of the Rhetoric that makes the written word its primary focus; this frame of reference in turn leads him to introduce hard distinctions—in this case, between ὑπόκρισις and λέξις—where the relation between the terms is more nuanced. Freese’s translation ad 1403b35–36 avoids the impression of a dichotomy: “But no treatise has yet been composed on delivery, since the matter of style itself only lately came into notice” (1926:347). καί (as the “itself” hints) may well be emphatic, but if so the comparison does not intend a contrast between ‘delivery’ and ‘style’ that sets the former apart from the latter. It would signal an argument a maiore ad minus, one from the part to the whole: τὸ περὶ τὴν λέξιν denotes the sphere of all matters related to λέξις; if only lately did the general sphere of ‘stylistic matters’ itself receive scholarly attention, the correspondingly late focus on ‘delivery’—a subordinate component of Aristotelian λέξις—is only to be expected. A translation like Kennedy’s that renders ἐπεὶ καί ‘since even’ does not necessarily exclude this meaning but it tends to obscure it. Contrast this with Ross 1924, where Roberts translates ad loc., “[I]ndeed, even the study of language made no progress till late in the day,” but in a footnote clarifies: “From this and other indications it would seem that Aristotle regards delivery as a subordinate part of λέξις, ‘expression’.” And, sensitive to the ‘oral’ overtones of the terminology, he adds: “The classification of ὑπόκρισις under λέξις is helped by the relation of the latter to λέγειν. λέξις is ‘a mode of speaking’.” Cope 1877:6, in his note to §5, offers an acceptable translation, but unduly restricts the φορτικόν to ἡ ὑποκριτική (obscuring its clear, and immediately preceding, neuter referent, viz. τὸ περὶ τὴν λέξιν; see further below, §14.1 n. 6): “But no art has been as yet composed of it; for in fact it was not till late that that of composition made any advance.”

[ back ] 24. Reviewing the uses of μὲν οὖν, Denniston (1954:470–481) divides them into three main categories: 1) retrospective and transitional οὖν with prospective μέν; 2) οὖν emphasizing a prospective μέν; 3) οὖν emphasizing an adversative or affirmative μέν. This passage falls under the first category (the repeated mention of the ‘hearer’ at 1404a8 and a12 makes this clear): with transitional and inferential force Aristotle retakes the immediately preceding μέγα δύναται in order to restate his point (διαφέρει γάρ τι) with an additional qualification (οὐ μέντοι τοσοῦτον), deriving from it as well a practical consequence for the study of rhetoric (ἔχει τι μικρὸν ἀναγκαῖον ἐν πάσῃ διδασκαλίᾳ). The ὅμως of a9 might seem at first to hinder my view that this section presents a restatement and an inferential summary (which would call for overall agreement with, and perhaps expansion of, the foregoing, but not for an adversative); but, just as the μὲν οὖν, this ὅμως also looks back, responding to the tension between the philosopher’s censure of the ἡ τοῦ ἀκροατοῦ μοχθηρία that gives λέξις its force, and his implicit repudiation of any device that does not strictly answer to the bare facts—this latter, an imperative of justice that only an ideal society, devoid of the corrupting allure of style, might realize. This, moreover, is not the only ὅμως in the passage: an ἀλλ’ ὅμως in the previous sentence (1404a7) makes explicit the tension between the status of λέξις as περίεργα ἔξω τοῦ ἀποδεῖξαι (a6–7) and its influence (μέγα δύναται).

[ back ] 25. I am not denying the possibility (which has much to commend it) that the present-day Book Ⅲ of the Rhetoric may originally have been an independent work: this would account for the admittedly awkward double transition from Rhetoric Ⅱ to Ⅲ: ἐπεὶ δὲ δὴ τρία ἔστιν ἃ δεῖ πραγματευθῆναι περὶ τὸν λόγον at 1403a34–35 and ἐπειδὴ τρία ἐστὶν ἃ δεῖ πραγματευθῆναι περὶ τὸν λόγον at 1403b6–7. For a possible reconstruction of the historical development of the treatise, see Kennedy 1991:217 and 299–309 (his appendix Ⅱ).

[ back ] 26. I make clear below (see §13.4) that I disagree with this narrow construction of the contents of these chapters.

[ back ] 27. Note, in this connection, the helpful distinction between primary and secondary rhetoric in Kennedy 1999:2–3.

[ back ] 28. Until recently scholars dated most, if not all, of the treatise Peri hermēneias (or De elocutione; I shall refer to it by the English On Style) to no later than the second century BC, with a few excepting only what they thought were minor editorial touches by a later hand. During the last decade, however, several studies have advocated a later, Roman-imperial, date (Roberts 1902:64 had already advocated long ago a date in the range I BC–I AD.) See Innes 1999:312–321 and Marini 2007:1–20.

[ back ] 29. On Style §§58, 195.

[ back ] 30. On Style §§193–194, 226, 271.

[ back ] 31. Aristotle himself planted the seeds of this development: his divergent analyses of λέξις in the Poetics and the Rhetoric implied that poetry (or, more broadly, literature) and oratory called not merely for two different styles (cf. Rhetoric 1404a27–28), but for entirely different ways of conceptualizing style (see below, §13.5 n. 100 and p. 588). To understand how natural it was for the study of λέξις to shift its focus from oratory to literature, we need only remember the philosopher’s dismissal of ὄψις as ‘least integral to poetics’ because the δύναμις of tragedy did not depend on actors or on competitive performance (ἀγών) (Poetics 1450b16–20); and recall that he thought reading unassisted by the κίνησις of performance quite able to reveal the qualities of tragedy (Poetics 1462a11). Letteraturizzazione is the Italian term used to denote this move away from oratory—in Kennedy’s words, “the tendency of rhetoric to shift focus from persuasion to narration, from civic to personal contexts, and from speech to literature, including poetry” (1999:3). It is difficult to say whether this development was already under way with Theophrastos. His writing an independent treatise on delivery (cf. Diogenes Laertios 5.48) suggests that he held to a greater autonomy of λέξις and ὑπόκρισις than Aristotle did. A passage of Athanasius (Fortenbaugh et al. 1992:558 fr. 712) shows that to voice, he added the study gestures (κίνησις τοῦ σώματος), and that he was more successful than his predecessor in integrating voice and gesture with an analysis of human emotions. (Yet facial expressions are not entirely ignored by the Rhetoric: cf. 1386a31–33 [see below, §13.4] and 1408b5–7 [see below, §13.4].) But fragments 687 and 691 (Fortenbaugh) show an emphasis on the senses (sound and sight) that is remarkably similar to Aristotle’s own (cf. Rhetoric 1405b6–8, b17–19). Thus, it is hard for me to tell how great a departure Theophrastos’ treatment really represents. Cf. Fortenbaugh 1985.

[ back ] 32. ἤρξαντο μὲν οὖν κινῆσαι τὸ πρῶτον, ὥσπερ πέφυκεν, οἱ ποιηταί· τὰ γὰρ ὀνόματα μιμήματα ἐστίν, ὑπῆρξε δὲ καὶ ἡ φωνὴ πάντων μιμητικώτατον τῶν μορίων ἡμῖν· (1404a19–22).

[ back ] 33. Cf. Rapp 2002:2.819–820.

[ back ] 34. Note, for example, how at 1407b11–12, where Aristotle turns his attention explicitly to the written text (τὸ γεγραμμένον), he is chiefly concerned that it be εὐανάγνωστον, ‘easy to read aloud’, and εὔφραστον, which LSJ s.v. renders as ‘easy to make intelligible’, an acceptable gloss that has, nevertheless, an undeniable speaking semantic component and is therefore translated as ‘easy to speak’ (Kennedy), ‘easy to utter’ (Freese Loeb Classical Library), or ‘leicht auszusprechen’ (Rapp).

[ back ] 35. By itself, this might mean that his own treatise had not yet addressed it, but the second occurrence makes the sense clear, namely, that no scholar had yet given it sustained attention. Of course, had he meant the former, this would imply that he himself would presently take it up.

[ back ] 36. Thus, I agree with Cope when he glosses ἐκείνη as “that (the art which applies ὑποκριτική to Rhetoric)” (1877:3.8, his emphasis). This art, a subdivision of ῥητορική, is precisely lexis.

[ back ] 37. For which we must doubtless supply something like ‘into use’ or ‘into vogue.’ Aristotle must have intended his own treatment in Rhetoric Ⅲ to contribute to such bringing it into use.

[ back ] 38. τὸ μὲν οὖν τῆς λέξεως ὅμως ἔχει τι μικρὸν ἀναγκαῖον ἐν πάσῃ διδασκαλίᾳ (1404a8–9). According to Aristotle, τὰ περὶ τὴν ὑπόκρισιν had only come to the dramatic art of late because the poets themselves had at first acted out their own tragedies. Hence there was no need to instruct others in delivery—the assumption being, perhaps, that poets were ‘natural’ actors (cf. 1404a15). Once some distance intervened between the poet and the performer, the conceptual space necessary for abstract study arose. But, as regards rhetoric, Aristotle lived during a time when teachers of the art were many and in great demand; so, it is only reasonable to expect that his rhetorical manual would have made sure to impart the all essential instruction in oral delivery.

[ back ] 39. Cf. Plutarch Dēmosthenēs §7.5–6.

[ back ] 40. See above, §13.2.

[ back ] 41. Rhetoric 1405b17–19.

[ back ] 42. This is what is involved in ‘placing the thing before the eyes’ (τὸ ποιεῖν τὸ πρᾶγμα πρὸ ὀμμάτων 1405b12–13), a matter treated at greater length in Rhetoric Ⅲ.10–11 (cf. also Poetics 17). For an exploration of φαντασία in connection with rhetoric, see below, §13.6. One must remember that, although it takes its name from the preeminent sense, sight, φαντασία does not exclude the other senses (De anima 429a1–4; cf. Busche 2003:34). It is helpful to cite here Kennedy’s insightful comment about Aristotle’s distinctive approach to the study of metaphors: “The account [of ornament] is a subtle one and seeks to penetrate to an understanding of the psychological effect of a metaphor” (1963:107). To this, he adds: “[The knowledge communicated by metaphors], like rhythm and like the sense of grammatical completion, produces a feeling of pleasure and satisfaction and is thus a characteristic of good style. Happiness is as much the object of Aristotle’s theory of style as of his ethics” (112). Kennedy is acknowledging that, even when it comes to metaphor—what many would consider the most traditional of stylistic topics—the philosopher is eminently concerned with the emotions, the very target of delivery. And his concern, as we have seen by his statement at 1405b17–19, centers on the sensory qualities of metaphors.

[ back ] 43. Thus, Kennedy (1991:220) titles Rhetoric Ⅲ.2 “The Aretē, or Virtue, of Good Prose Style; Word Choice and Metaphors.”

[ back ] 44. His eyes, though visible through the mask, would have been too small for his audience to see.

[ back ] 45. Stage gestures were exaggerated and, in his view, unbecomingly crude (cf. Poetics 26 and below, §13.5.1). The same attitude may underlie his vehement criticism of Kleon’s manner in addressing the assembly (cf. Athēnaiōn Politeia 28.2).

[ back ] 46. As Rapp (2002:2.861) remarks, one should not construe that style be ἀνάλογον to its subject matter as a third requirement unrelated to its ability effectively to convey ēthos and pathos: “Natürlich liegt die Angemessenheit nicht schon dann vor, wenn die sprachliche Form ‘emotional’ und ‘charaktervoll’ ist; also wird man den letzten Teil des Definiens ‘dem zugrunde liegenden Gegenstand entsprechend’ auf die Merkmale ‘emotional’ und ‘charaktervoll’ zurückbeziehen müssen, so dass die Behauptung wäre: Die sprachliche Form muss dem jeweiligen Gegenstand entsprechend emotional oder charaktervoll sein.”

[ back ] 47. Book Ⅲ, chapters 11–15 §§19–27.

[ back ] 48. One might reasonably argue that, since we all have experienced (in ourselves or others) the full range of emotions under consideration, such explicit directions would have been superfluous. There is, besides, a marked tendency in the Rhetoric towards compression of treatment, so that material not strictly necessary is omitted or a subject already mentioned is not repeated, even when it calls for development from a new vantage point. Cf. Striker 1996:289 and 300n9 and Brinton 1988:208 on the scope of Aristotle’s study of πάθη.

[ back ] 49. … συνομοιοπαθεῖ ὁ ἀκούων ἀεὶ τῷ παθητικῶς λέγοντι, κἂν μηθὲν λέγῃ. διὸ πολλοὶ καταπλήττουσι τοὺς ἀκροατὰς θορυβοῦντες (1408a23–25). It is only natural that a section concerned with φωνή should mention causing an uproar as a strategy sometimes used to confound the audience. Just as the volume (μέγεθος) of one’s voice can be modulated in accordance with the art of rhetoric without detriment to the truth, it can also serve the purposes of less principled orators. Thus, I disagree with Rapp (2002:2.863), who takes θορυβοῦντες and the concessive statements that accompany it (on which see note 50 immediately below) as paradigmatic of Aristotle’s view of delivery. He thus reasserts his opinion that the philosopher has entirely excluded ὑπόκρισις from his treatment of style (as pernicious to the interests of justice), and he denies that this section develops the art of oratorical delivery.

[ back ] 50. The concessive clauses εἰ καὶ μὴ οὕτως ἔχει, ὡς ὁ λέγων (1408a22–23) and κἂν μηθὲν λέγῃ (a24) do not condone, much less enjoin, deception or encourage emotional appeals ἔξω τοῦ πράγματος. They only state the obvious: that rhetoric, as any other social endeavor, is open to the manipulation of deceit, all the more so inasmuch as its potential impact is great. On the valence of ἔξω τοῦ πράγματος, see Grimaldi 1980 ad 1354a15–16. On this point Halliwell (1994:212) writes: “Here the formulation (assisted by a certain latitude in the established terminology of speaking ‘on/outside the subject,’ peri/exō tou pragmatos) somewhat elides two things that might properly be distinguished: first, an emphasis on the importance of criteria of judicial relevance; second, a general deprecation of (distortingly) emotional appeals in rhetoric (cf. 3.1.5, 1404a5–8).” I would only modify his statement to say ‘a general deprecation of (potentially distorting) emotional appeals as ideally superfluous.’

[ back ] 51. See below, §13.6 n. 190.

[ back ] 52. καὶ ὅλως δὴ ὅταν ἔχῃ οὕτως ὥστ’ ἀναμνησθῆναι τοιαῦτα συμβεβηκότα ἢ αὑτῷ ἢ τῶν αὑτοῦ, ἢ ἐλπίσαι γενέσθαι ἢ ἑαυτῷ ἢ τῶν αὑτοῦ (1386a1–3).

[ back ] 53. ἀνάγκη τοὺς συναπεργαζομένους σχήμασι καὶ φωναῖς καὶ αἰσθήσει καὶ ὅλως ἐν ὑποκρίσει ἐλεεινοτέρους εἶναι (1386a31–33). Both this and the quotation immediately preceding (with its supplements) are from Kennedy 1991:153–154 ad loc.; ἐν ὑποκρίσει should have been translated ‘by their delivery’. At 1386a32 Kassel’s codex F reads ἐσθῆτι for αἰσθήσει, which Spengel corrected to ἐσθῆσι.

[ back ] 54. The words used are ὀδυρτικωτέρως, συνεστηκότως, μαλακωτέρως, μέσως, καθεστηκότως, and ἐνθουσιαστικούς (Politics 1340a42–b5; cf. b10–12). For the relationship between ἦθος and πάθος in this context, see Susemihl and Hicks 1894:622–624.

[ back ] 55. Labarrière 1984:34–40. σημεῖον, as the unmarked term, may also apply to λόγος.

[ back ] 56. οὐ γὰρ ταὐτὰ οὐδ’ ὡσαύτως ἂν ἀγροῖκος ἂν καὶ πεπαιδευμένος εἴπειεν (1408a31–32).

[ back ] 57. The material in this passage and the approach Aristotle takes is best understood if we suppose his ideas to have developed under the influence of stage acting, where the need for a successful characterization, e.g. of females by males or the old by youths, would be real and acute. Cf., for example, Plato’s Republic 395d5–e3. Significant too is the fourth-century phenomenon of λογογραφία, which further separated composition from delivery; the professional speech writer had to place himself in his client’s shoes and make him say nothing inconsistent with his origin, social status, occupation, etc. See above, §9.5 n. 76.

[ back ] 58. δοκεῖ γὰρ ἀληθὲς εἶναι, ἐπεὶ οὐ λανθάνει γε ὃ ποιεῖ τὸν λέγοντα. ἔτι τοῖς ἀνάλογον μὴ πᾶσιν ἅμα χρήσασθαι· οὕτω γὰρ κλέπτεται [ὁ ἀκροατής]. λέγω δὲ οἷον ἐὰν τὰ ὀνόματα σκληρὰ ᾖ, μὴ καὶ τῇ φωνῇ καὶ τῷ προσώπῳ [καὶ τοῖς] ἁρμόττουσιν (1408b3–7). The language of ‘escaping notice’, ‘cheating the hearer’, etc. is not so much an admission of the potential for deceitfulness inherent in naturalistic art as it is a statement of the paradox essential to its success: for then ars adeo latet arte sua. Cf. Rhetoric Ⅲ.2.4–5 and 8.1. Fortenbaugh (1996:161) calls to mind an interesting analogy. In Politics 1310a2–12, Aristotle counsels that demagogues in a democracy should seem to speak (δοκεῖν λέγειν) on behalf of the wealthy; whereas in an oligarchy, oligarchs should exhibit (ὑποκρίνεσθαι) a correspondingly favorable attitude towards the δῆμος. Fortenbaugh comments: “Those words may suggest feigned concern, and undoubtedly there are moments to be disingenuous” (161), but “[t]he verb hypokrinesthai … need not imply a feigned exhibition” (161n34).

[ back ] 59. ῥυθμὸν δὲ [δεῖ ἔχειν τὸν λόγον] μὴ ἀκριβῶς· τοῦτο δὲ ἔσται ἐὰν μέχρι του ᾖ (1408b31–32, supplementing from the preceding sentence).

[ back ] 60. ὁ δὲ τοῦ σχήματος τῆς λέξεως ἀριθμὸς ῥυθμός ἐστιν, οὗ καὶ τὰ μέτρα τμητά. διὸ ῥυθμὸν δεῖ ἔχειν τὸν λόγον, μέτρον δὲ μή (1408b28–30).

[ back ] 61. Kennedy 1991:218 and n. 5; cf. Burkett 2011:28–29.

[ back ] 62. Lucas 1968 ad loc. writes: “I translate it by ‘melody’, though this does not exclude the notion of rhythm.” That it entails musical tune is clear from 1447a23, which only includes instruments.

[ back ] 63. The triad ῥυθμός, λόγος, and ἁρμονία at 1447a23 corresponds to ῥυθμός, μέλος, and μέτρον at 1447b25. This can be easily explained if we note that μέλος already implies words, and that μέτρον corresponds to metrical λόγος (Lucas 1968:61). ‘Meters’, as Poetics 1448b21–22 shows, are ‘categories’ (μόρια) of rhythms (to use Halliwell’s Loeb Classical Library translation).

[ back ] 64. Translation by Halliwell 1999. The entire passage runs thus: μάλιστα γὰρ λεκτικὸν τῶν μέτρων τὸ ἰαμβεῖόν ἐστιν· σημεῖον δὲ τούτου, πλεῖστα γὰρ ἰαμβεῖα λέγομεν ἐν τῇ διαλέκτῳ τῇ πρὸς ἀλλήλους, ἑξάμετρα δὲ ὀλιγάκις καὶ ἐκβαίνοντες τῆς λεκτικῆς ἁρμονίας (1449a24–28).

[ back ] 65. Cf. Probert 2006:54–55, who in turn directs the reader to Devine and Stephens 1994:376–497.

[ back ] 66. Apparently, such correlation was true even in their technical musical sense; cf. West 1992:178 and 181.

[ back ] 67. See Spengel 1867:2.386–389.

[ back ] 68. The apparatus critici in Roemer 1898 and Kassel 1976 ad loc. suggest that Vettori printed καὶ οὐ λεκτικός. Strictly speaking, however, although commended for its sense (“… ipse aliquando suspicatus sum desiderari particulam negandi”), neither in Vettori 1548:517–519 nor in Vettori 1579:615–617 was the emendation adopted. In either edition the Greek text reads: τῶν δὲ ῥυθμῶν, ὁ μὲν ἡρῶος, σεμνὸς καὶ λεκτικὸς καὶ ἁρμονίας δεόμενος; which he translates: “e numeris autem, herous grandis est, dignitatisque plenus et orationi aptus quique requirit harmoniam.” On the assumption that the text was sound, he noted: “valet λεκτικὸς hoc in loco, sonorus ac grandiloquus. quod sane mirum videtur, cum omnis locutio λέξις a Graecis vocetur.”

[ back ] 69. Often called ‘modes’ by modern scholars, harmoniai represent ancient scales (i.e. distinctive series of intervals assembled as scales). In this technical sense ἁρμονία does not imply words; it might denote the music of an instrument (ψιλὴ ἁρμονία) or the voice’s singing (with or without words). μέλος, on the other hand, referred primarily to the melody of singing and in this sense involved words. (Because it was the singing of poetry and it was usually accompanied by instruments, μέλος also stood for a poetic composition with musical accompaniment, especially lyric poetry.) By conceptually abstracting the music from the singing, μέλος was also at times used for ἁρμονία (West 1992:177–178). Clearly, one would be more likely to construe the statement ‘the heroic meter calls for ἁρμονία’ as implying that it calls for musical accompaniment than that it calls for chanting. Quite apart from the intrinsic unlikelihood of the textual reading (given the unquestionable application of ἁρμονία at 1403b31 to the voice of the orator, which rules out chanting; and the incontrovertible use of λεκτικὴ ἁρμονία in the Poetics 1449a28, which renders the restoration of the same expression in the present passage distinctly plausible), had this been Aristotle’s meaning one would have expected τοῦ ᾄδειν δεόμενος vel sim. One may still argue that, for an audience familiar with the performance practice of epic (assuming ex hypothesi non-instrumental chanting, i.e. a reduced melody not far from speech tones; cf. West 1981:114), the meaning of ἁρμονίας δεόμενος would have been transparent: I agree. This would justify Vettori’s reading and Kennedy’s translation. But, even then, if ἁρμονία can be used of the melodic contour of non-instrumental chanting, given its application in the Poetics to the melodic contour of prose, we would still have chapter 8 of Rhetoric Ⅲ concerned with ἁρμονία as one of the three components of φωνή that affect the orator’s delivery.

[ back ] 70. This he quotes not as οὐ λεκτικός (as Roemer 1898 ad loc.) but as οὐ λογικός (Cope 1877:3.87). Cf. Demetrios On Style §42.

[ back ] 71. τῶν δὲ ῥυθμῶν ὁ μὲν ἡρῷος σεμνὸς ἀλλὰ λεκτικῆς ἁρμονίας δεόμενος (Roemer 1898:189).

[ back ] 72. Demetrios On Style §42 uses the term ἠχώδης. See above, §10.2.1.

[ back ] 73. τῶν δὲ ῥυθμῶν ὁ μὲν ἡρῷος σεμνῆς ἀλλ’ οὐ λεκτικῆς ἁρμονίας δεόμενος.

[ back ] 74. His contribution is to oppose a σεμνὴ ἁρμονία to the attested λεκτικὴ ἁρμονία, a contrast in itself plausible; and to motivate the αὐτή in the following clause: ‘whereas heroic verse calls for an elevated—not a colloquial—intonation, iambic by itself [i.e. without any peculiar intonation] is the λέξις of the common people’.

[ back ] 75. I do not agree with Cope (1877:3.86), who, although using ‘harmony’ in his translation “[to leave] open whether we are to understand by ἁρμονία ‘harmony’ in its ordinary musical sense,” nevertheless calls this “a somewhat non-natural interpretation.” The sense he favors is “an adaptation or fitting of parts into an organized whole” (i.e. ‘structure’ or ‘style of composition’). On the impact of Greek accents upon the melodic shape of the utterance, see Allen 1973:230–234, West 1981:114–115, and §13.4 n. 65 above.

[ back ] 76. Elementa harmonica 23.13–16 (da Rios). Barker translates: “For there is indeed said to be a kind of melody which belongs to speech, that constituted by the tone-patterns that occur in words, since tension and relaxation belong naturally to speech” (1984–1989:2.138).

[ back ] 77. That Aristoxenos here means ‘utterance’ (the action of λέγειν) and not ‘style’ or any other more technical meaning is clear from its use in Elementa rhythmica 17.15–23 (Pighi): ὥσπερ γὰρ τὸ σῶμα πλείους ἰδέας λαμβάνει σχημάτων, ἐὰν αὐτοῦ τὰ μέρη τεθῇ διαφερόντως, ἤτοι πάντα ἤ τινα αὐτῶν, οὕτω καὶ τῶν ῥυθμιζομένων ἕκαστον πλείους λαμβάνει μορφάς, οὐ κατὰ τὴν αὑτοῦ φύσιν, ἀλλὰ κατὰ τὴν τοῦ ῥυθμοῦ. ἡ γὰρ αὐτὴ λέξις εἰς χρόνους τεθεῖσα διαφέροντας ἀλλήλων λαμβάνει τινὰς διαφορὰς τοιαύτας, αἵ εἰσιν ἴσαι αὐταῖς τῆς τοῦ ῥυθμοῦ φύσεως διαφοραῖς. ὁ αὐτὸς δὲ λόγος καὶ ἐπὶ τοῦ μέλους καὶ εἴ τι ἄλλο πέφυκε ῥυθμίζεσθαι τῷ τοιούτῳ ῥυθμῷ ὅς ἐστιν ἐκ χρόνων συνεστηκώς. Barker (1984–1989:2.185) translates the key sentence as follows: “For the same utterance [lexis], when disposed into durations that differ from one another, takes on differences of a sort that are equal to the differences in the nature of the rhythm themselves.” Of manifest significance is the aural effect of the words in performance, a significance best captured by the gloss ‘utterance’, which notionally subsumes any particular rhythmic realization. We find a similar use of λέξις at Aristotle’s Sophistical Refutations 165b23–24 and Rhetoric 1401a1, where παρὰ τὴν λέξιν is variously translated ‘verbal’, ‘of diction’, ‘depends on language’, ‘beruhen auf dem sprachlichen Ausdruck’, vel sim. Cf. Halliwell (1993:53–54), who identifies three different ways in which Aristotle uses λέξις; though I am not entirely comfortable with the manner in which he apportions the semantic range and with the illustrations adduced in support, he does to my mind correctly identify the two ends of the spectrum: “ordinary speech,” roughly equivalent to my ‘utterance’, and “style or expressiveness,” an evaluative term that, among other things, focuses on register, genre, and tone. (I would transfer to his category C some of his examples for B that, by his own admission, display a “semantic emphasis.”)

[ back ] 78. Elementa rhythmica 19.15–18 (Pighi): ἔστι δὲ τὰ ῥυθμιζόμενα τρία· λέξις, μέλος, κίνησις σωματική. ὥστε διαιρήσει τὸν χρόνον ἡ μὲν λέξις τοῖς αὑτῆς μέρεσιν, οἷον γράμμασι καὶ συλλαβαῖς καὶ ῥήμασι καὶ πᾶσι τοῖς τοιούτοις. This division of λέξις is indebted to Aristotle’s treatment in the Poetics 20. For Aristoxenos’ Elementa rhythmica see now Marchetti 2009. The passage immediately above is found in Marchetti 2009:37 §9, with translation at 67 and commentary at 116–119. Marchetti’s rendering of λέξις as ‘text’ is disappointingly flat and I cannot endorse it. He considers more broadly its Aristoxenian meaning at 92–99 (“a group of words chosen by the poet but which may be composed into different verses in different word order” 98). Although Barker (1984–1989:2.186) ad loc. prints the standard gloss ‘diction’, as noted in the footnote immediately above at §4 (=17.19 Pighi) he renders it more pointedly as ‘utterance’.

[ back ] 79. The text suffers from various uncertainties which do not, however, affect the point for which I quote it. Hubert in his Teubner edition prints it thus: λέγει δὲ μουσικῆς ἀρχὰς τρεῖς εἶναι, λύπην, ἡδονήν, ἐνθουσιασμόν, ὡς ἑκάστου τῶν … αὐτῶν παρατρέ⟨ποντος⟩ ἐκ τοῦ συνήθους ⟨καὶ παρ-⟩εγκλίνοντος τὴν φωνήν. Fortenbaugh et al. (1992:572–575 fr. 719A) adopt Bernardakis emendation τῶν ⟨παθῶν⟩ τούτων for τῶν … αὐτῶν and translate: “[A]nd he says that there are three sources of music, pain, pleasure and inspiration; for each of these emotions turns the voice aside and deflects it from its usual (inflection)” (575).

[ back ] 80. May and Wisse (2001:276) ad loc. translate them as ‘cadence’ and ‘form’. At §173 Cicero repeats the terms: modus etiam et forma verborum.

[ back ] 81. May and Wisse 2001:278 ad loc. Wilkins 1892 ad loc. comments: “[H]ere [modus] covers variations both in duration and in pitch, i.e. what we call ‘tune’.” He further translates vocis moderatio ‘the modulation of the voice’, and verborum conclusio ‘the periodic arrangement of the words’. I might also mention Hermogenes Peri ideōn I, where rhythm is defined thus: … καὶ τὸ ἐξ ἀμφοῖν τούτοιν [sc. συνθέσεών τε καὶ ἀναπαύσεων] συνιστάμενον ὁ ῥυθμός· ἡ γὰρ ποιὰ σύνθεσις τῶν τοῦ λόγου μερῶν καὶ τὸ ὡδί πως ἀναπεπαῦσθαι τὸν λόγον ἀλλὰ μὴ ὡδὶ ποιεῖ τὸ τοιόνδε ἀλλὰ μὴ τοιόνδε εἶναι τὸν ῥυθμόν (p. 218.22–26 Rabe).

[ back ] 82. See especially §39.

[ back ] 83. See below, §14.1.

[ back ] 84. See below, §14.1.

[ back ] 85. Here as elsewhere I quote Kassel 1976 ad loc. Ross’s Greek text, despite differences in detail, ultimately bears the same meaning. He emends the vulgate οὐκ ὀρθῶς at 1404a2–3 to οὐχ ὡς ὀρθῶς; but an explicit ὡς is not required and, as Spengel (1867:2.357) notes, should be understood here from the ἀλλ’ ὡς that follows. Kassel emends ζητεῖν to ζητεῖ at 1404a4; and preserves πλείω at 1404a4 and ὡς at 1404a5 over against the πλέον and ὥστε suggested by the old Latin translation (and adopted by Ross). Yet other than emphasis there is hardly a difference between ‘justice is to seek nothing more (μηδὲν πλέον) than’ and ‘justice does not at all (μηδὲν) seek more (πλείω) than’. I prefer the transmitted infinitive ζητεῖν to Kassel’s indicative (‘justice is not at all to seek more than’), since ζητεῖ would call for οὐδέν. As to ἢ ὥστε vis-à-vis ἤ ὡς, the choice is semantically indifferent (cf. Smyth §2007).

[ back ] 86. The immediate referent in agreement with the neuters ἔχοντος and ἀναγκαίου is τὸ περὶ τὴν λέξιν at 1403b36. Alternatively, the referent might be ὑπόκρισις (from αὕτη/αὐτή at 1403b27) or τὰ περὶ τὴν ὑπόκρισιν (from 1403b22). The translation would then be: ‘not as being a right, but a necessary, thing’. These more distant potential referents, however, are far less likely than the proximate τὸ περὶ τὴν λέξιν, whose case is further buttressed by its reappearance at 1404a8 in the form τὸ τῆς λέξεως. See further below, §14.1 (with n. 6). My grammatical and contextual analysis, here and below, demonstrates (contra Burkett 2011:36 and others) that one should not think of τὸ τῆς λέξεως as a new subject (or digression) introduced at 1404a8—a notion that, if true, would call for an earlier contrasting referent, presumably ‘delivery’ (ὑπόκρισις) viewed as distinct from ‘style’ (λέξις). The text controverts this reading: i) by unambiguously subsuming τὸ τῆς λέξεως under φαντασία ‘towards the hearer’ (πρὸς τὸν ἀκροατήν 1404a11–12, my emphasis), thus evincing that Aristotle has aural effects in mind (i.e. ‘delivery’); and ⅱ) by stating that its ‘coming into vogue’ will accomplish the same as ἡ ὑποκριτική (1404a13), plainly echoing 1403b22–26 where ὑπόκρισις is undeniably in view (and not some notion of λέξις allegedly distinct from ‘delivery’).

[ back ] 87. This ὅμως concedes that μέγα δύναται seemingly controverts τἆλλα … περίεργά ἐστιν.

[ back ] 88. This μὲν οὖν is transitional (see Denniston 1954:470–473), with retrospective οὖν inferentially regarding μέγα δύναται and prospective μέν anticipating the adversative οὐ μέντοι in 1404a10 (which substitutes for οὐ δέ; cf. Denniston 1954:404 §2.i).

[ back ] 89. This ὅμως acknowledges that the need for lexis in every demonstrative discipline seemingly controverts that it owes its efficacy to the baseness of the hearer.

[ back ] 90. This translation draws on the discussion that follows.

[ back ] 91. For my reasons to identify τὸ περὶ τὴν λέξιν as the subject, see immediately above (n. 86) and below (§13.5); see, further, §14.1 (with n. 6).

[ back ] 92. On the related ἔξω τοῦ πράγματος see above, §13.4 n. 50. τὸ ἀποδεῖξαι here refers narrowly to rational demonstration (i.e. the enthymeme derived from πράγμα or λόγος); for ἀπόδειξις and ἀποδεικνύναι in the Rhetoric see Grimaldi 1972:139–141.

[ back ] 93. Too much has been made of 1404a2–7, which states that ‘justice (τὸ δίκαιον) consists in not at all seeking more in connection with one’s argument (or ‘speech’, περὶ τὸν λόγον) than that one should cause neither pain nor pleasure’ (hinting at pathos, cf. 1378a20–22); and that ‘it is just to contend (ἀγωνίζεσθαι) only with the facts’ (αὐτοῖς τοῖς πράγμασιν). Accordingly, some infer that departing from this ideal would not be merely undesirable, but even unjust; and they think that οὐκ ὀρθῶς ἔχοντος at 1404a2–3 corroborates this view. But the comparison aims not at discriminating justice from injustice, but at evoking an ideal vision of what strict justice calls for. Although ἀγωνίζεσθαι can be used of an epideictic competition (or even a political debate), I think here Aristotle is primarily thinking of the court setting, which leads him naturally to state this ideal in terms of ‘justice’. Justice, strictly speaking, must always bind defendant and plaintiff to the facts of the case: the facts are their one and only necessary point of reference; strictly speaking, everything else must be judged superfluous. This does not mean, however, that the contribution of ēthos and pathos is unjust; simply that justice does not require it. This interpretation is conclusively proved by the ὥστε clause, which declares whatever else (τἆλλα) falls outside the realm of demonstration not unjust, but ‘superfluous’, περίεργα. Therefore, in ‘not because it is right’ (οὐκ ὀρθῶς) ‘right’ does not mean ‘morally right’ (with ‘unjust’ or ‘morally wrong’ as its opposite), but ‘correct’, as in ‘the correct choice’, the choice demanded and strictly justified by the circumstances (its opposite, then, being ‘erroneous’ or ‘incorrect’). This is similar to its use in the expression ὀρθῶς λέγειν, ‘strictly speaking’ (cf. LSJ s.v. ὀρθός Ⅲ.2), where accuracy and inaccuracy, correctness and error, are opposed, not justice and injustice. Jebb’s translation conveys the meaning well: “… we must give our attention to this subject, considered as necessary, not as desirable in itself; for, strictly speaking, our sole aim in our language should be to give neither pain nor pleasure; our facts ought to be our sole weapons.” Cf. also Dufour and Wartelle 1973 ad loc. On ὀρθός, Irwin (1985:391) writes: “‘Orthos’ indicates success in pursuing an end or correctness in picking it (1144a20), as opposed to error … . It is not confined to moral rightness: nor is any special moral sense of the term required.”

[ back ] 94. Is not striking imagery itself (to mention only one aspect of the wider field of style) quite able forcefully to affect the emotions of the hearers? Rapp (2002:2.814 §3.1) suggests that Aristotle’s goal is to move the ethically objectionable ὑπόκρισις off the center of his study, in order to focus instead on those aspects of style that do not so crassly take advantage of the μοχθηρία of the hearers: “Der mündliche Vortrag wird als eine Sache der Begabung vom Zentrum der folgenden Überlegungen ausgeschlossen. Die sich tatsächlich anschließende Behandlung der sprachlichen Form konzentriert sich dagegen auf die Erleichterung des Verstehenprozesses.” But here, realizing keenly that this, too, looks to affect the hearer in ways that go beyond a strict appeal to the facts themselves, he tellingly adds: “… was zwar auch wirkungsbezogen ist, gleichzeitig jedoch in den Dienst der Sache gestellt werden kann” (my emphasis). With this admission, the case against ὑπόκρισις contracts to whether Rapp is right in asserting—I submit that he is not—that Aristotle thinks of ‘delivery’ strictly as a matter of talent and not susceptible of technical treatment.

[ back ] 95. Cf. De interpretatione I 16a3–4: ἔστι μὲν οὖν τὰ ἐν τῇ φωνῇ τῶν ἐν τῇ ψυχῇ παθημάτων σύμβολα.

[ back ] 96. On this see immediately above, §13.5 n. 86, and further below, §14.1 n. 6.

[ back ] 97. As noted above, §13.5 n. 86, this must be the noun phrase implicitly in agreement at 1404a3 with the participle ἔχοντος and the adjective ἀναγκαίου.

[ back ] 98. Here again my understanding diverges from Rapp’s, much the most sensitive and careful of recent readers of the Rhetoric, whose translation and commentary of Rhetoric Ⅲ is the most important extended treatment of that book to appear since Cope 1877. Rapp (2002:2.812 §1) understands the uncertainty facing the reader as to the respective boundaries of ὑπόκρισις and λέξις: “[D]ie vorigen Abschnitte [erweckten] den Eindruck … als handle es sich beim mündlichen Vortrag und der sprachlichen Form um zwei klar voneinander getrennte Bereiche.” But 1403b35–36, he adds insightfully, hints “dass das vorliegende Kapitel von einem Diskussionsstand ausgeht, auf dem sprachliche Form und mündlicher Vortrag eng miteinander verknüpft sind, bzw. letzterer einen Teilbereich der sprachlichen Form ausmacht.” I find his suggestion misguided, however, that the conceptual development of the argument lies in so delimiting the purview of style that a discussion of delivery, its subordinate component, can be excluded. At 2.814 §3.1, after recalling that it is “durchaus unklar, inwieweit vom mündlichen Vortrag und inwieweit von der sprachlichen Form … die Rede ist,” he offers two possible explanations to καὶ δοκεῖ φορτικὸν εἶναι (1403b36–1404a1). The first, which I believe correct, flows readily from the text (as he himself admits): “Gemeint ist tatsächlich die sprachliche Form, denn zu Beginn des Abschnitts wird ausdrücklich das genannt, ‘was die sprachliche Form betrifft’. Dann würde auch für die Behandlung der sprachlichen Form gelten, dass sie nur notwendig, aber nicht richtig ist.” The second posits as aim a limited concept of style, one that has been “bereinigt,” ‘cleaned up’ (so to say), from which delivery has been excluded, and that focuses on clarity of meaning. But the alleged goal (which I here dispute) of an unimpeachable λέξις that makes clarity its sole ambition proves elusive: for why then would Aristotle add that ‘no one teaches geometry thus’ (1404a12)? If in fact λέξις (in its alleged restricted sense) pursues clarity (διαφέρει γάρ τι πρὸς τὸ δηλῶσαι 1404a9–10), why would it not also apply to geometry and every other philosophical or scientific inquiry? But if geometry can achieve clarity without recourse to style, why should not the orator who renounces an appeal to man’s baser inclinations not simply dispense with style altogether? The conclusion seems inescapable: if with a view to its effect upon the hearer (“aus Gründen der Wirkung”) the orator is willing to compromise intelligibility (as Rapp claims), and, therefore, clarity becomes a stylistic trait—with style teaching us to negotiate the corresponding trade-off—how can λέξις be said to have escaped ethical indictment? Newman 2005 is another recent work that, while taking Aristotelian lexis seriously as more than rhetorical ornament, still views it narrowly through the lens of ‘intellectual clarity’ without reference to ‘delivery’—as if that were all that the philosopher meant when he asserted, ὡρίσθω λέξεως ἀρετὴ σαφῆ εἶναι (Rhetoric 1404b1–2; cf. Poetics 1458a18). In rhetorical theory and practice ‘expressive clarity’ cannot—and for Aristotle emphatically did not—exclude full-blown consideration of ὑπόκρισις.

[ back ] 99. I take καὶ πρὸς τὸν ἀκροατήν as epexegetic (cf. Smyth §2869a), i.e. as adding by way of clarification: ‘I mean [φαντασία] towards the hearer’.

[ back ] 100. That is, it must not be overly restricted to such clarity of thought and argument, including a perspicuous prose style, as would characterize a well-written, readily intelligible book. In the Rhetoric the pragmatics of oral delivery is of controlling importance for Aristotelian λέξις. Not so for his divergent conceptualization in the Poetics, where he defines λέξις as ἡ διὰ τῆς ὀνομασίας ἑρμηνεία (1450b13–14). This allows him at 1456b10 to sever ‘poetics’ from ἡ ὑποκριτική—here, understood as ‘[dramatic] delivery’ and encompassing the knowledge of forms of utterance (τὰ σχήματα τῆς λέξεως)—as different kinds of investigation (Poetics 1456b8–19). On Aristotle’s divergent analyses of poetic and rhetorical lexis, see further above, §13.2 n. 31, and below, §14.1. Cf. Morpurgo Tagliabue 1967:201–207, Halliwell 1986:344–349, and Halliwell 1993. Newman’s emphasis on ‘clarity’ and her controlling focus on ‘metaphor’ prevents her from registering clearly the differences between the philosopher’s respective notions of poetic and rhetorical style.

[ back ] 101. Rapp (2002:2.816) uses ‘Gedanken’ to gloss διάνοια, which at 1403a36 seems to encompass ēthos and pathos insofar as addressed in Books I and Ⅱ of the Rhetoric, where they are covered from the point of view of εὕρησις (cf. Grimaldi 1988:369 and Poetics 1456a33–b7). But his discussion suggests that with ‘Gedanken’ he regards conceptual perspicuity only. Cf. Rapp 2002:2.829 §3 and his clear statement at 2.804: “Der Gedanke ist das, was dasselbe bleibt, wenn sich die sprachliche Form allein ändert.”

[ back ] 102. Here I follow Ross, who adopts an emendation in Richards 1915:111 that is clearly superior to the paradosis: it not only accounts for the otherwise troublesome ὡς but makes for a more coherent explanation of the importance of clarity. By justifying the need for perspicuity on the grounds of the logos otherwise missing its target, the transmitted text offers a circular argument. Kassel’s edition prints Vahlen’s ὥστ’ for the ὡς of cod. Parisinus 1741 (Vahlen 1867:262), a fine alternative that Kennedy 1991 and Rapp 2002 prefer. Both Ross and Kassel read τι for ὅτι. Their texts make largely the same point; either: ‘for logos, since it is a kind of sign, will not perform its proper function unless …’ (Ross); or: ‘for logos [is] a kind of sign, so that it will not perform its proper function unless …’. Here logos is best understood as ‘articulate speech’ (cf. Poetics 1457a23–24 and De interpretatione 16b26–27) and as implying the speaker’s oral argument.

[ back ] 103. If we do not usually think of a ‘metaphor’ as a ‘word’ (ὄνομα), Aristotle does: ἅπαν δὲ ὄνομά ἐστιν ἢ κύριον ἢ γλῶττα ἢ μεταφορὰ ἢ κόσμος ἢ πεποιημένον ἢ ἐπεκτεταμένον ἢ ὑφῃρημένον ἢ ἐξηλλαγμένον. Cf. Newman 2005:29; and Lucas 1968: “μεταφορά means both the transference of a word from its usual reference and the word so transferred” (204). Burkett (2011:79–80) condemns as an “error” the notion of ‘metaphor’ as “one word” (his emphasis).

[ back ] 104. Kennedy 1991:223 ad loc., slightly modified (my emphasis).

[ back ] 105. It may be well further to drive this point home with the aid of another passage. There, quite apart from strict conceptual clarity, we learn that, when it comes to ēthos (and the same might be said of the designs of pathos on the affections of the audience) it is essential to communicate it clearly if it is to gain its proper end: ‘[T]he [forensic] narrative must indicate character (ἠθικήν); and it shall be so if we know what makes for ēthos. One way, indeed, is to make the motivation clear (τὸ προαίρεσιν δηλοῦν); [then] the ēthos, of a certain kind, by the motivation being such; and the motivation, of a given sort, by its end’ (1417a16–19). [In translating προαίρεσις by ‘motivation’ (which should be understood as deliberate purpose), I follow Nicomachean Ethics 1139a31–33: πράξεως μὲν οὖν ἀρχὴ προαίρεσις—ὅθεν ἡ κίνησις ἀλλ’ οὐχ οὗ ἕνεκα—προαιρέσεως δὲ ὄρεξις καὶ λόγος ὁ ἕνεκά τινος.] Though this comment belongs to the section on τάξις and no particular stylistic devices to secure a strong ethical cast are mentioned, it is clear from the observations at 1404b18–25 that style, particularly the stylistic register selected, will further or hinder this goal insofar as it lends conviction to, or detracts from, the portrayal of the man who is (so goes the claim) motivated thus, and acts in accordance with the alleged purposes: one must seem to speak in his own voice or else he will fail to persuade his audience. This explains why τὸ πρέπον, ‘propriety’—here meant as stylistic propriety—immediately follows clarity as the λέξεως ἀρετή. Cf. Halliwell 1993:57n12, where he makes a distinction between the “referential” and “expressive” uses of δηλοῦν in the Rhetoric, and lists some examples.

[ back ] 106. ἐπεὶ … , ἡ δὲ φαντασία ἐστὶν αἴσθησίς τις ἀσθενής, καὶ [κἀεὶ?] ἐν τῷ μεμνημένῳ καὶ τῷ ἐλπίζοντι ἀκολουθοῖ ἂν φαντασία τις οὗ μέμνηται ἢ ἐλπίζει, … (‘Since … , and [since] phantasia is a kind of weak perception, and in the one who recalls or hopes some sort of phantasia of what he is recalling or hoping for will [always?] go with [his remembering or hoping], …’). This passage is notoriously difficult and textually uncertain. Here I print Kassel’s text and, with Cope, assume that the protasis εἰ δὲ τοῦτο at 1370a30 resumptively sums up the ἐπεί clauses and the optative ἀκολουθοῖ ἄν (‘—if this [is so] …’). See Grimaldi 1980:251–252 and cf. Rhetoric 1362a29–31 on ἀκολουθέω (the optative does not so much make the attendance of phantasia a likely contingency as it rhetorically softens the statement the more gently to invite and elicit the assent of the reader to it). The bibliography on Aristotelian phantasia is massive. A partial list might include Freudenthal 1863, Cosenza 1968, Nussbaum 1978:221–269, Schofield 1992 [1978], Armisen-Marchetti 1979, Armisen-Marchetti 1980, Watson 1982, Ward 1984, Labarrière 1984, Rispoli 1985, Wedin 1988, Watson 1988, Fattori and Bianchi 1988, Frede 1992, Turnbull 1994, Sisko 1995, Caston 1996, Everson 1997, Labarrière 1997, Lefebvre 1997a and 1997b, Busche 1997, Manieri 1998, Lefebvre 1999, Formigari et al. 1999, Birondo 2001, Lefebvre 2001–2002, Barnouw 2002:49–147, Labarrière 2002 and 2003, Lefebvre 2003, Busche 2003, Veloso 2004, Krow 2005, O’Gorman 2005, Linguiti 2006, Stevens 2006, González 2006, Yurdin 2008, Riu 2009, and Dow 2009. For more on 1370a28–30 see below, §13.6.

[ back ] 107. Cope 1877 ad loc. renders it ‘fancy’, adding that φαντασία is “the mental presentation, a mere copy, without reality” (with a reference to his note on Rhetoric I.11.6); Kennedy 1991 and Freese 1926 prefer ‘outward show’; Jebb, more cautious, retains the more traditional ‘imagination’ (in Sandys 1909 ad loc.); Roberts, too, chooses ‘fanciful’ (in Ross 1924); and Rapp 2002, ‘reiner Anschein’.

[ back ] 108. Here I employ ‘political’ in that basic sense most readily illustrated by Aristotle’s famous dictum that ὁ ἄνθρωπος φύσει πολιτικὸν ζῷον (Politics 1253a2–3).

[ back ] 109. I need only cite 1356a25–27, where, from the tripartite division of rhetoric into logos, ēthos, and pathos, Aristotle concludes: ὥστε συμβαίνει τὴν ῥητορικὴν οἷον παραφυές τι τῆς διαλεκτικῆς εἶναι καὶ τῆς περὶ τὰ ἤθη πραγματείας, ἣν δίκαιόν ἐστι προσαγορεύειν πολιτικήν (‘so that it happens that rhetoric is, as it were, a kind of offshoot of dialectic and of the study of ethics which one may fairly label politics’, my emphasis). Cf. Nicomachean Ethics 1094a27–b3. For the negative perspective, cf. Plato Gorgias 463d1–2.

[ back ] 110. Aristotle considers geometry a part of mathematics, as can be gathered from many passages in his works (cf. Heath 1949:1–16). In his Posterior Analytics 77b26–33, for example, after stating that ‘in mathematics (ἐν δὲ τοῖς μαθήμασιν) fallacy does not happen similarly’, he illustrates this assertion with reference to the circle: ἆρα πᾶς κύκλος σχῆμα; ἂν δὲ γράψῃ, δῆλον. τί δέ; τὰ ἔπη κύκλος; φανερὸν ὅτι οὐκ ἔστιν (‘Is every circle a figure? If one draws it, clearly yes. What then? Is the epic [cycle] a circle? Clearly not’). Another clear statement can be found in Metaphysics Ε 1026a25–27, where in a parenthetical comment the philosopher remarks: “even the mathematical sciences (ἐν ταῖς μαθηματικαῖς) differ in this respect—geometry and astronomy deal with a particular kind of entity, whereas universal mathematics applies to all kinds alike” (Tredennick’s Loeb Classical Library translation); this shows that mathematics includes geometry as a subdivision (see Ross’s note in his commentary ad loc.). In Proklos’ Commentary on the First Book of Euklid’s Elements 38.4–12 (Friedlein) we learn that Geminos had described geometry as a μέρος of μαθηματική, a view that goes back, in turn, to Peripatetic scholarship (specifically, to Eudemos of Rhodes), thus confirming Aristotle’s own classification (cf. F. Hultsch in RE Ⅶ 1 s.v. “geometria,” col. 1210).

[ back ] 111. Hence the transmitted reading πολιτειῶν at 1403b35, which Kassel prints and Lossau 1971 justifiedly defends against the suspicions of Spengel (1867:2.357), whose conjecture was accepted by Ross in his OCT.

[ back ] 112. One need only remember that φορτικότης had made its appearance at 1395b1–2, in the chapter on maxims, long before the matter of style was broached. There Aristotle reflected on the pleasure a hearer experiences when an orator hits upon opinions he already holds. Cope 1877 ad loc.: “The φορτικότης here ascribed to vulgar audiences is much the same as the μοχθηρία τῶν ἀκροατῶν Ⅲ 1.5, the vices or defects, which oblige the orator to have recourse to τἆλλα ἔξω τοῦ ἀποδεῖξαι in order to convince them, because they are unable to appreciate logic alone.” Nota bene 1395b13–14, where, as might be expected, we learn that the ethical character of the discourse is chiefly in view. Cf. also 1415b4–8: δεῖ δὲ μὴ λανθάνειν ὅτι πάντα ἔξω τοῦ λόγου τὰ τοιαῦτα· πρὸς φαῦλον γὰρ ἀκροατὴν καὶ τὰ ἔξω τοῦ πράγματος ἀκούοντα, ἐπεὶ ἂν μὴ τοιοῦτος ᾖ, οὐθὲν δεῖ προοιμίου, ἀλλ’ ἢ ὅσον τὸ πρᾶγμα εἰπεῖν κεφαλαιωδῶς, ἵνα ἔχῃ ὥσπερ σῶμα κεφαλήν (‘One must not forget that all such things fall outside the argument; they regard a hearer who is vulgar and listens to what is extrinsic to the matter, since, if he were not such, a proem is not necessary other than to headline the matter, so that it may have a head just as [it has] a body’).

[ back ] 113. Freese writes “and rightly considered it is thought vulgar”; Kennedy, “delivery seems a vulgar matter when rightly understood”; Jebb, “and, properly viewed, the subject is thought vulgar”; Dufour and Wartelle, “il semble d’ailleurs que ce soit là un art grossier à en juger sainement”; Rapp, “auch scheint es, richtig betrachtet, ungebührlich zu sein”; Tovar, “parece que es asunto fútil, bien considerado.”

[ back ] 114. Vettori 1579:544.

[ back ] 115. Forming an opinion entails weighing the facts (‘ponderare’) and making a judgment (‘iudicare’). These entailments of ὑπολαμβάνω are easily borne out by a survey of the verb elsewhere in the Rhetoric, where it is frequently used.

[ back ] 116. One other passage may be mentioned. In Politics Ⅷ.6 (1340b20ff.) Aristotle takes up the question whether the education of freeborn youth πρὸς ἀρετήν should include μουσική, specifically, learning to sing and play an instrument. The inquiry—1340b34–35 makes clear—addressing itself to those who claim that it is a menial occupation (βάναυσον εἶναι τὴν ἐπιμέλειαν), suggests that μουσική is beneficial, so long as the degree to which it is practiced is carefully regulated, and melodies, rhythms, and instruments vetted for propriety (1340b42–1341a3). Later on, at 1341b8–18, before turning his attention to ‘harmonies’ and ‘rhythms’, Aristotle again makes passing mention of the general disapproval bestowed on ‘professional education’ (ἡ τεχνικὴ παιδεία); he goes on, in a long parenthesis, to explain the opprobrium as follows: τεχνικὴν δὲ τίθεμεν τὴν πρὸς τοὺς ἀγῶνας· ἐν ταύτῃ γὰρ ὁ πράττων οὐ τῆς αὑτοῦ μεταχειρίζεται χάριν ἀρετῆς, ἀλλὰ τῆς τῶν ἀκουόντων ἡδονῆς, καὶ ταύτης φορτικῆς, διόπερ οὐ τῶν ἐλευθέρων κρίνομεν εἶναι τὴν ἐργασίαν, ἀλλὰ θητικωτέραν· καὶ βαναύσους δὴ συμβαίνει γίγνεσθαι· πονηρὸς γὰρ ὁ σκοπὸς πρὸς ὃν ποιοῦνται τὸ τέλος· ὁ γὰρ θεατὴς φορτικὸς ὢν μεταβάλλειν εἴωθε τὴν μουσικήν, ὥστε καὶ τοὺς τεχνίτας τοὺς πρὸς αὐτὸν μελετῶντας αὐτούς τε ποιούς τινας ποιεῖ καὶ τὰ σώματα διὰ τὰς κινήσεις (‘We reckon that education technical which regards contests; for in it the performer does not practice for the sake of his own virtue but for the pleasure of the audience, and a vulgar one at that. Precisely for this reason we do not deem this practice proper for free [citizens] but more typical of paid laborers. And the result is that they grow illiberal, for the target they aim at is base; for the hearer, being vulgar, is wont to change mousikē [for the worse], whence also it actually follows that the professionals themselves who ply their trade with him in view he makes of some such kind [i.e. illiberal], as well as their bodies on account of their motions’). Note that, once again, what qualifies a practice as φορτική (and, in this case, earns it the label θητικωτέρα) is the focus on the hearers, specifically, on their pleasure, which is compared unfavorably with personal ἀρετή, so that ‘professional’ practice is rendered illiberal: a θεατὴς φορτικός corrupts the τεχνῖται who ply their trade with him in view. Cf. Politics 1342a18–21, Plato’s Laws Ⅱ 659b–c, and especially Plato’s Gorgias 512e5–513c2, with its insistence on the necessary conformity of the orator (as μιμητής) to his audience and their πολιτεία if he desires to wield influence in the city (μέγα δύνασθαι ἐν τῇ πόλει 513a3–4).

[ back ] 117. Cf. 1355a20–24, a pivotal passage whose text is the object of much disagreement but whose implications are surprisingly tolerant of the various opinions. See further below, §13.6.

[ back ] 118. Though here φαντασία and φαίνεσθαι do not themselves occur, one finds several instances of their semantic opposite in this context, viz. λανθάνω.

[ back ] 119. As scholars have suggested, the choice of justice as an illustration might hint at an ongoing polemic with some of the more outrageous sophists, who may have publicly owned appearing just preferable to actually being so. (The φασί would then have specific subjects in view, which the reader in turn would be expected to identify. Thus, for example, in Plato’s Republic Ⅱ 362e4–363a5 Adeimantos notes that parents commend justice for the good repute that accrues from it and the benefits that attend on such public esteem [cf. 365b4–7, 366d7–e5, and 367b6–c1].) This would go some way towards explaining the statement, ὅτι δοκεῖν ἢ εἶναι αἱρετώτερον. It would, indeed, be surprising if this apparently sweeping and rather pessimistic judgment represented the view of the common man. Or are we to believe that most Athenians really thought the appearance of justice more desirable than its reality?

[ back ] 120. Translators agree, rendering πρὸς δόξαν at 1404a2 ‘to influence opinion’ (Freese), ‘with opinion’ (Kennedy), ‘auf die Meinung abzielt’ (Rapp), ‘ne s’attache qu’à l’opinion’ (Dufour and Wartelle). Jebb’s ‘aims at appearance’ and Tovar’s ‘apariencia’ approach Cope, but show greater restraint as neither carries the negative connotations of ‘show’ and ‘ostentation’. Indeed, both can be argued to imply ‘opinion’ and hence are, in my view, acceptable equivalents. Of the other occurrences of δόξα in the Rhetoric, those at 1360b22, 1362b20, 1367a16, 1368b23, 1371a16, 1388a2, 1388a6, and 1404a25 clearly (1397b26 probably) carry the meaning ‘reputation’ (cf. also 1372b21–23); at 1381b19–20 τὰ πρὸς δόξαν (as 1381b30–31 makes clear) refers to anything that affects the opinion the public has of us (hence our ‘reputation’), and is opposed to τὰ πρὸς ἀλήθειαν, ‘what is actually true’, for there may, of course, be a gap between one’s reputation and one’s true character; at 1384a23–29, too, the argument hinges on ‘loss of reputation’ (ἀδοξία), an opinion (δόξα) men heed only on account of those who hold it (οἱ δοξάζοντες); ‘opinion’ is the proper rendering too at 1369a22, 1377b18, 1378b10, 1391b24, 1395b3, 1403a32, and 1412a27 (here best translated ‘expectation’), and perhaps at 1384b23 (though I rather incline, with Kennedy, to ‘reputation’, construing καταφρονοῦσι with τῆς δόξης and not with τοῦ ἀληθεύειν). (The well-known idiomatic παρὰ δόξαν needs no discussion.) The only other relevant section is Rhetoric Ⅱ.1, whose principal thrust is the importance of showing oneself to be, and rendering the kritēs, of a certain type (1377b24, the former pertains to ēthos, the latter to pathos—ποιόν τινα … κατασκευάζειν is construed apo koinou with αὑτόν and τὸν κριτήν, though its sense varies with either term). The point is constructing a convincing persona, articulating and sustaining a particular view of oneself, managing one’s reputation with his hearers, how one comes across: hence the recurrence of φαίνεσθαι (at 1377b26, 29, 31; 1378a5, 17). Thus, there is no parallel to be found in the Rhetoric for Cope’s translation of πρὸς δόξαν, “aim[ing] at mere outside show, ostentatious and hollow” (Cope 1877:3.7); and even his alternative, “is directed to mere opinion, is unscientific,” although it contains the otherwise acceptable ‘opinion’, distorted as it is by ‘mere’ and ‘unscientific’, fails to offer Aristotle’s meaning. The philosopher makes clear (see 1355b35–1356a4) that ēthos and pathos are ἔντεχνοι, that they are (so to speak) ‘scientific’. Cf. Grimaldi 1980:38 ad 1355b35 on ἔντεχνοι (he uses Roemer’s lineation) and 349–356.

[ back ] 121. This question bears tangentially on the debate reviewed by Rapp (2002:2.92–94) ad 1355a1325–26, subsection 1, whether Aristotle designed his theory of demonstrative science to guide and formalize scientific research or was instead exclusively concerned with the teaching of knowledge already acquired.

[ back ] 122. So the statement ‘listen to my instruction,’ while referring to the contents of the teaching, draws attention to the teacher’s active training of his pupil. The DGE s.v. διδασκαλία I.3 documents its late, reified acceptation, ‘doctrine’, in Christian literature. But Aristotle, aligning himself squarely with the classical usage, arguably evokes the process of teaching. Thus, in Rhetoric 1355a25–26, with διδασκαλίας γάρ ἐστιν ὁ κατὰ τὴν ἐπιστήμην λόγος Aristotle declares that following (superlatively precise 1355a24) knowledge in one’s speech is ‘[characteristic] of didaskalia’ (or ‘proper to’ vel sim.). In view here is not a reified, written speech, but the orator’s speaking: note the participle λέγοντας at 1355a25. I cannot accept Kassel’s unnecessary emendation of the vulgate διδασκαλίας to the nominative case, which unduly recasts the relationship between the terms of the sentence as a straightforward identity. Neither is Rapp (2002:2.92) right to equate the (semantically broader) διδασκαλία here with the (narrower) διδασκαλικοὶ [λόγοι] in Sophistical Refutations 165b1–3. Not every διδασκαλία need be a λόγος guided by ἐπιστήμη; τέχνη must be allowed a place too, as one learns from the occurrence in Rhetoric 1355b29 of διδασκαλική in connection with τέχνη. More on this passage immediately below.

[ back ] 123. Cf. DGE s.v. διδάσκω I.1.d and Plato Menōn 85e1.

[ back ] 124. So, for example, although Halliwell (1993:55) glosses didaskalia as “systematic instruction,” he also calls ‘geometry’ (not, nota bene, the teaching of geometry) “[Aristotle’s] example of didaskalia”; at 56, after asserting that “[t]he Rhetoric defines didaskalia as ‘discourse grounded in knowledge’,” he adds that “the reference here is certainly to systematic bodies or methods … of knowledge,” thereby not necessarily implying, but certainly suggesting, speech reified as a ‘body of knowledge’ rather than didactic discourse and the active process of instruction (notice that he calls ‘philosophy’ itself ‘[a] knowledgeable didaskalia’). Admittedly, professional engagement with abstract subjects like geometry may often in practice approach teaching in manner.

[ back ] 125. For example, in Plato Theaitētos 162e γεωμετρέω denotes the practice of geometry as discourse on, and argument about, geometry.

[ back ] 126. After stating unequivocally that style has some small necessary part in every διδασκαλία, it would be striking if Aristotle would contradict himself by asserting that geometric instruction—arguably a διδασκαλία—does not, after all, call for style in any measure, however small. But, strictly speaking, this is not the nature of the claim, and the διό may simply reflect common (if inadvisable) practice: style is rarely (if ever) involved in teaching (and applying) geometry, not because the discipline has no place for it (it does, though small); but because, insofar as it offers the smallest imaginable scope for style, its neglect in this case hardly occasions any harm. In other words, established practice in connection with geometry illustrates the principle at issue by taking it to the limit. Cf. Grimaldi (1980:36) ad 1355b29 apropos διδασκαλικὴ καὶ πειστική: “A[ristotle] acknowledges that the use of language in all the disciplines is always something more than notional and rational. The word greatly extends the area of ‘persuasive speech’ and recognizes that almost all discourse with another inevitably seeks to win acceptance for itself from the other. … A.’s examples of geometry and arithmetic and his inclusion of ἐπιστῆμαι (55b31) are even more interesting in view of the fact that no kind of persuasion is ordinarily identified with scientific discourse.” Cf. Rihll 1999:12–13, 22. Aristotle Sophistical Refutations 165b3 (δεῖ γὰρ πιστεύειν τὸν μανθάνοντα) furnishes a rare acknowledgment of the need for persuasion in scientific discourse. Rapp (2002:2.94) reads this statement through the lens of Topics 159a28–30, which he interprets as stating the need for a learner to trust his teacher even before he can grasp that/why he is right (“denn niemand versucht, Falsches zu lehren” 159a29–30). But I believe that he has misunderstood the passage in the Topics and that he is wrong to use it to displace a learner’s ‘trust’ away from ‘commonly accepted, reputable opinions’ (τὰ δοκοῦντα in the sense of ἔνδοξα) and onto his teacher’s intent not to deceive him. τῷ … μανθάνοντι θετέον (159a29) should be rendered ‘a pupil must posit’ (i.e. grant) not ‘a pupil must state’; τὰ δοκοῦντα (159a29) are not uniquely the learner’s, but generally embraced, notions (cf. 100b4–6); and, as often (cf. Denniston 1954:64–65), the second γάρ clause does not qualify the first but shares its reference (the contrast drawn at 159a26–28). In fact, Topics 159a3–14 sets the stage by detailing explicitly the conditions under which a learner should or should not posit (i.e. grant) axioms and propositions harder to argue than the original thesis. Clearly, the analysis does not turn on the trustworthiness of the teacher but strictly on the context (learning vs. training) and the character of the claims involved (whether convincing, well-known, and so on). Note in particular the occurrence of πιστεύειν in connection with πιστά (not διδάσκαλος or ὁ διδάσκων). This is not to deny, of course, that positing or not particular claims affects the student-teacher dynamics (159a13–14). Topics 141a30 should suffice to settle that both teaching and learning entail demonstration.

[ back ] 127. Rhetoric 1355b26–32 owns that all τέχναι and ἐπιστῆμαι have a demonstrative facet. See Grimaldi’s comment ad loc., quoted in the footnote immediately above.

[ back ] 128. Notice the related use in Poetics 1456b5, which opposes φαίνεσθαι ἄνευ διδασκαλίας to ἐν τῷ λόγῳ ὑπὸ τοῦ λέγοντος παρασκευάζεσθαι καὶ παρὰ τὸν λόγον γίγνεσθαι (cf. Lucas 1968:196 ad loc.). Note also Aristotle Sophistical Refutations 165b1–3, where διδασκαλικοὶ [λόγοι] are defined as arguments that ‘deduce from the principles proper to each subject and not from the opinions offered by the answerer’ (διδασκαλικοὶ μὲν οἱ ἐκ τῶν οἰκείων ἀρχῶν ἑκάστου μαθήματος καὶ οὐκ ἐκ τῶν τοῦ ἀποκρινομένου δοξῶν συλλογιζόμενοι); the sentence that precedes this one makes clear that Aristotle’s taxonomy concerns ‘arguments used in conversing’ (τῶν ἐν τῷ διαλέγεσθαι λόγων τέτταρα γένη, my emphasis), a perspective that brings out the discursive and demonstrative character of such arguments (cf. Sophistical Refutations 171a31–32, a38–b2). A few lines later (165b9) the philosopher even substitutes ἀποδεικτικῶν for διδασκαλικῶν, bringing out explicitly the tie between formal instruction and demonstration, a tie that, as just noted, given the importance of ‘proof’ (ἀπόδειξις) for oratory, would readily suggest the use of διδασκαλία at Rhetoric 1404a9. Plato too associates διδασκαλία with rhetoric. In Gorgias 453d7–454e2, both rhetoric and διδασκαλικαὶ τέχναι are said to work conviction (they are πειθοῦς δημιουργοί 453e4–5). For Sokrates, however, the parallel between them breaks down when it comes to truth-value: whereas πίστις can be false, μάθησις (the goal of διδάσκειν) cannot.

[ back ] 129. Attention to the need for persuasion in scientific discourse seems also at issue in Rhetoric 1355a24–29, a passage to which I have already drawn attention above. There we learn that not even fortified with the sharpest knowledge (ἐπιστήμη) would a speaker be able to convince one and all: ἔτι δὲ πρὸς ἐνίους οὐδ’ εἰ τὴν ἀκριβεστάτην ἔχοιμεν ἐπιστήμην, ῥᾴδιον ἀπ’ ἐκείνης πεῖσαι λέγοντας· διδασκαλίας γάρ ἐστιν ὁ κατὰ τὴν ἐπιστήμην λόγος, τοῦτο δὲ ἀδύνατον, ἀλλ’ ἀνάγκη διὰ τῶν κοινῶν ποιεῖσθαι τὰς πίστεις καὶ τοὺς λόγους, ὥσπερ καὶ ἐν τοῖς Τοπικοῖς ἐλέγομεν περὶ τῆς πρὸς τοὺς πολλοὺς ἐντεύξεως (‘And yet, in dealing with certain people, not even if we should have the most exact knowledge would we find it easy by drawing upon it to convince them in conversation; for discourse that conforms to systematic knowledge is characteristic of didaskalia, and this is impossible [with such people], but rather one must fashion his proofs and [oral?] arguments utilizing what is commonly held, as we said in the Topics concerning conversation with the crowd’). Although in this passage the philosopher appears categorically to deprecate didaskalia in the context of rhetoric, in point of fact his criticism is restricted to ‘instruction’ (i.e. demonstrative discourse) that adheres to the canons of exacting knowledge (what we might call ‘formal instruction’). He does not intend to deny rhetoric an instructive or demonstrative force but to make room for an art of persuasion that does not regard as its own domain the collection of (demonstrative and persuasive) facts peculiar to any particular body of knowledge. Aristotle aims to preserve the distinction between a system of knowledge whose purview is bounded by its peculiar subject matter and oratory, whose search for τὸ πιθανόν embraces all others only in their common relation to λόγοι. Indeed, an argument strictly based on ἐπιστήμη would be characteristic of formal instruction, but rhetoric is a combination of ‘analytical knowledge’ (τῆς ἀναλυτικῆς ἐπιστήμης 1359b10) and ‘that [political] science concerned with ethics’ (τῆς περὶ τὰ ἤθη [πολιτικῆς] 1359b10). The philosopher affirms in 1359b8–16 that one who failed to furnish (as he ought) rhetoric or dialectic as the δυνάμεις (1359b13; cf. 55b26) that they are and instead remade them into ἐπιστῆμαι (59b13) would obscure and change the character of rhetoric from an ἐπιστήμη of speech (59b16) to one of certain underlying subjects (59b15–16, e.g. medicine, geometry, arithmetic, or some other 55b29–31). It seems, then, that Aristotle sometimes loosely applies to rhetoric the label ἐπιστήμη (it is ‘an epistēmē of discourse’ as well as a combination of analytical and ethical epistēmai; cf. Grimaldi 1980:93 ad 1359b16), sometimes disowns it (discourse based on epistēmē amounts to formal instruction that is impossible with the crowd). In sum, where διδασκαλία is involved the teacher brings out the clear and necessary consequences of a subject’s own peculiar principles, and he secures the necessary assent of the learner. This happens characteristically in the formal instruction of an ἐπιστήμη (say, geometry), given a pupil of the requisite ability; but to adhere strictly to this manner of presentation with most members of an ordinary audience is impracticable, certainly so where considerations of ēthos and pathos play a role (cf. 1417a19): hence one must resort to commonly held notions (τὰ κοινά) for one’s proofs (πίστεις) and arguments (λόγοι). Cf. Rapp 2002:2.92–96 and Grimaldi 1980:28–29.

[ back ] 130. My argument, of course, does not turn merely on a matter of punctuation—whether a comma should precede the καί. But understanding φαντασία as ‘ostentation’ has recently been helped by a tendency to weaken ever so slightly the immediacy of its connection in this sentence to the hearer: ‘all this is mere show, and [all of it is directed] towards the hearer’. But if, as I claim, Aristotle has in view the orator’s crafting of φαντάσματα for the hearing (and viewing) of his audience, one might expect him to say ἀλλ’ ἅπαντα φαντασία ταῦτ’ ἐστὶ πρὸς τὸν ἀκροατήν. And he could further emphasize the target of this φαντασία by placing an adverbial καί before πρός. This, I submit, is precisely the case; and if such is the point of the phrase, the editor should not punctuate with a pause after ἐστί. Omitting the comma by itself does not, of course, prejudice the interpretation and would be the safest editorial choice: hence neither Bekker, nor Spengel, Roemer, or Kassel, not even Cope and Sandys feature the comma; Ross and Dufour and Wartelle do. Yet a translation, when there is one, often makes the structure of the editor’s thought clear.

[ back ] 131. Once again I draw attention to Rhetoric 1355b29 and to Grimaldi’s comment above, §13.5.1 n. 126.

[ back ] 132. Cf. González 2006, which was based on an earlier version of this material. Less narrowly concerned with the Rhetoric, O’Gorman 2005 helpfully explores the psychology of lexis, although he places to my mind excessive emphasis on the visual (as opposed to the aural and cognitive) dimensions of φαντασία without comparable regard for the pragmatics of delivery. Nevertheless, his qualification of lexis as “phantasmatic” (26) is a step in the right direction, even if its connection to delivery remains muted (only mentioned in passing at 16, 23, 26) other than to say that style sometimes subsumes it.

[ back ] 133. I have quoted Ross’s text (cf. Ross 1961:286 ad 428a3–4), even though there is much disagreement whether the last clause is in fact a question. Note his insertion of ἆρα and his punctuation, for which he has no ms. support. Some scholars, obviously uncomfortable with φαντασία as a faculty of judgment, recast the statement as a query (to be answered later in the negative) or, like Bywater (who added ζητῶμεν εἰ), as a summary proposal for the examination that follows. (Cf. Watson 1982:106n10 and Wedin 1988:47–48 with nn. 29–30.) But contrary to the assumption of many, the list at 428a4–5 (τοιαῦται δ’ εἰσὶν αἴσθησις, δόξα, ἐπιστήμη, νοῦς) need not be exhaustive (otherwise, we would expect ταῦται, not τοιαῦται, at De anima 428a4), and therefore φαντασία need not be, as claimed, ruled out as a faculty of judgment. [Bekker and those who depend on him (like Biehl and Hicks) print ταῦτα as the reading of LWX; but Förster (1912:108) ad loc., who independently perused the witnesses and whom Ross follows, corrects Bekker and instead reports “τοιαῦται” for LW and “τοιαῦ (sic)” for X.] As Wedin (1988:47n30) himself acknowledges, we do not need a question here to establish that, in the event, for Aristotle φαντασία turns out to be something less than a full faculty; he might well have described it as a δύναμις that enables us to arrive at truth or falsity, while simultaneously asserting its subordinate role to the standard (truly independent) faculties. According to this view, φαντασία would then be “a system of internal [re]presentations that enables a person to have desires, beliefs, and thoughts about objects and situations in the world” (Wedin 1988:22).

[ back ] 134. Cf. Cope 1877 ad Rhetoric I.11.6.

[ back ] 135. For a better solution to what Aristotle means by κατὰ μεταφοράν, see Wedin 1988:69–70.

[ back ] 136. A debate, however, that has overlooked almost entirely Rhetoric Ⅲ.1. For helpful overviews see Rees 1971, Schofield 1992 [1978], Watson 1988, and Wedin 1988. An expanded bibliography may be found above, §13.5 n. 106.

[ back ] 137. Cf. Frede 1992:279–282.

[ back ] 138. Halliwell 1993:60n19 offers a mild corrective to this statement.

[ back ] 139. ‘The things that produce the appearance [of refutation] in consequence of language are six in number; these are: homonymy, ambiguity, combination, division, prosody, and form of expression’. An alternative (freer) translation of the first clause that better conveys the neutrality of τὴν φαντασίαν [τοῦ ἐλέγχου] as the functional equivalent of τὸν φαινόμενον ἔλεγχον (164a20–21; cf. 169b38) might be: ‘The linguistic phenomena that make a fallacy look like a refutation are six in number’. φαντασία, the (psychological) movement triggered by the perception of the phenomena listed, should be carefully distinguished from φάντασμα, the attendant ‘image’ or ‘impression’. Thus, more accurately: ‘The things that cause [a refutation] to appear [to someone] …’ (or ‘the appearing [of a refutation]’).

[ back ] 140. οἱ μὲν εἰσὶ συλλογισμοί, οἱ δ’ οὐκ ὄντες δοκοῦσι (164a23–24). Contra Poste, one should read the καί that precedes τῶν φαινομένων μὲν ἐλέγχων as epexegetic, as do Pickard-Cambridge (‘i.e.’), Forster (‘that is’), and Dorion (‘c’est-à-dire’). See Schreiber 2003:192n1 and 173–178 (Appendix 1).

[ back ] 141. LSJ s.v. 1.b places 165b25 under ‘less scientifically, appearance’, still not its fourth division ‘parade, ostentation’.

[ back ] 142. On ἴδια see below, §13.6.

[ back ] 143. Translation by Hicks 1907 ad 428b21–22.

[ back ] 144. Concerning non-visual φαντάσματα, see Frede 1992:285.

[ back ] 145. This crucial observation precludes a ‘metaphorical’ meaning in Simplicius’ sense (In Aristotelis De anima 428a1–2), that is, the meaning that obtains ὅτε ἐπὶ τοῦ φαινομένου τῇ φαντασίᾳ χρώμεθα καὶ ἐπὶ αἰσθήσεως καὶ ἐπὶ δόξης ‘when we use φαντασία for τὸ φαινόμενον (what appears to be the case) in both perception and belief’ (Hayduck 1882:208, lines 7–8). In other words, since it involves φαντάσματα and αἴσθησις, ‘appearing’ in a non-psychological, transferred sense is not in view (cf. Wedin 1988:69). The verb φαίνεσθαι by itself (apart from its context) does not decide whether the ‘appearing’ is or is not psychological (i.e. whether it involves perception and ‘images’ or simply denotes ‘seeming [to be the case]’); according to Simplicius, Aristotle was making clear his concern with ‘psychological φαντασία.’

[ back ] 146. Cf. the helpful analysis in Rapp 2002:2.575 and 2.621 (ad Rhetoric 1382a21).

[ back ] 147. Warning Xerxes about the peril of rash arrogance, Artabanos remarks: ὁρᾷς τὰ ὑπερέχοντα ζῷα ὡς κεραυνοῖ ὁ θεὸς οὐδὲ ἐᾷ φαντάζεσθαι, τὰ δὲ σμικρὰ οὐδέν μιν κνίζει. How and Wells 1912 ad loc. write: “[P]roperly show oneself (ⅳ.124.2; ⅶ.15.2); here se ostentare, ‘make a show of oneself’,” noting, moreover, its agreement with Polybios’ usage. Powell 1938 s.v. φαντάζομαι 2 glosses it as ‘[to] be conceited’.

[ back ] 148. I have not been able to turn up any other incontrovertible use of the verb in this peculiar sense before Aristotle. As to the noun φαντασία, the only other text cited by LSJ s.v. 4 that, in the opinion of some, might lend credence to the modern consensus on Rhetoric 1404a11 is [Hippokrates’] Decorum 7, which indeed belongs under ‘ostentation’ (cf. Jouanna 1999:75–111). But according to a recent detailed linguistic study by García Valdés 1992:304, “with a fair degree of certainty” the earliest possible date for this treatise is the second century AD.

[ back ] 149. Quite apart from the uniqueness of Herodotos 7.10ε and the reasons just offered for a reading of φαντάζεσθαι sensu stricto, it is pertinent to remember that verbs and nouns do not always possess the same semantic range. In this connection, Martínez Hernández (1997:196) criticizes precisely the failure to keep lexical fields and spheres separate when verbs and nouns are mixed, forgetting “[el] carácter paradigmático del campo, que exige establecer las oposiciones sobre las que se estructura entre lexemas de una sola clase” (cf. Martínez Hernández 1997:76–77 and Coseriu and Geckeler 1981:56–59). Geckeler (1971:218) offers a clear statement of this principle: “Wenn also das Wortfeld in struktureller Hinsicht als ‘ein lexikalisches Paradigma’ … charakterisiert ist, so kann es—das ist ein Wesenszug des Begriffes ‘Paradigma’—notwendigerweise nur eine Wortart umfassen. In einem einzigen Paradigma können nicht z. B. Substantive, Verben, Adjektive und Adverbien nebeneinander funktionieren, d. h. in direkter Opposition zueinander stehen” (his emphasis). To illustrate this point, one need only consider ποιέω, ποίησις, and ποιητής, which have areas of overlap and disjunction; or, to use an English example, the word ‘trip’: its noun, widely used for ‘short journey’, now rarely (if ever) for ‘nimble step’ or ‘stumble’; its verb in turn hardly ever for ‘to make a trip or excursion’. There are also adjectives without corresponding verbs (‘pregnant’ is very common; ‘pregnate’ rare and obsolete, ‘impregnate’ taking its place), verbs without their nouns (‘cleave’, ‘to part’, owns ‘cleft’; but ‘cleave’, ‘to adhere’, lacks a noun), etc. On such gaps, see Geckeler 1976:158–160 and Gutiérrez Ordóñez 1989:104–105. Lexical semantics is too complex to allow for unexamined extrapolations. Not even when, in their full diachronic sweep, a verb and its noun are attested with one and the same meaning should their synchrony be assumed. On this caution cf. Coseriu 1981:109–113. Cf., further, Schofield 1992:251n11.

[ back ] 150. φαντασία γὰρ ἕτερον καὶ αἰσθήσεως καὶ διανοίας, αὕτη τε οὐ γίγνεται ἄνευ αἰσθήσεως, καὶ ἄνευ ταύτης οὐκ ἔστιν ὑπόληψις. Cf. De anima 427b27–28.

[ back ] 151. Translation by Hicks 1907 ad loc., modified. ὅλως μὲν οὖν, ὥσπερ εἴρηται, ᾗ ὀρεκτικὸν τὸ ζῷον, ταύτῃ αὑτοῦ κινητικόν· ὀρεκτικὸν δὲ οὐκ ἄνευ φαντασίας· φαντασία δὲ πᾶσα ἢ λογιστικὴ ἢ αἰσθητική. ταύτης μὲν οὖν καὶ τὰ ἄλλα ζῷα μετέχει.

[ back ] 152. Though ἢ … ἤ are disjunctive, the alternatives need not be mutually exclusive: they may simply offer two complementary ways of viewing any φαντασία, ways that depend on the point of view chosen. That is to say, even the λογιστική may, on further consideration, turn out to be connected with αἴσθησις in a manner still to be determined. That this is in fact so becomes clear at De anima 432a3–10, where we learn that “apart from sensible magnitudes (παρὰ τὰ μεγέθη τὰ αἰσθητά) there is nothing, as it would seem, independently existent” (Hicks 1907 ad loc.); therefore, ‘it is by the perceptual forms (ἐν τοῖς εἴδεσι τοῖς αἰσθητοῖς) that the objects of thought (τὰ νοητά) exist’ (my translation, taking Wedin into account). Wedin (1988:114) interprets the point of the passage as follows: “[O]ne cannot think of things that are of a kind not to have existed or, even more liberally, … one cannot think of things whose salient parts are of a kind not to have existed.” Cf. also Labarrière (1984:47n32), where “la φαντασία produite par la νόησις” should be modified to ‘φαντασία involved in νόησις’ vel sim. That νόησις cannot, in fact, produce φαντασία is explained below, §13.6 n. 161.

[ back ] 153. ἡ μὲν οὖν αἰσθητικὴ φαντασία, ὥσπερ εἴρηται, καὶ ἐν τοῖς ἄλλοις ζῴοις ὑπάρχει, ἡ δὲ βουλευτικὴ ἐν τοῖς λογιστικοῖς (πότερον γὰρ πράξει τόδε ἢ τόδε, λογισμοῦ ἤδη ἐστὶν ἔργον· καὶ ἀνάγκη ἑνὶ μετρεῖν· τὸ μεῖζον γὰρ διώκει· ὥστε δύναται ἓν ἐκ πλειόνων φαντασμάτων ποιεῖν) (De anima 434a5–10). ‘Perceptual phantasia, as noted, also exists in the other animals, but the deliberative [only] in the rational (for whether one will do this or that already is a task for reasoning; and it is necessarily the case that one measures by one [standard], for one pursues what is greater; it follows that one is actually able from more images to make one’.

[ back ] 154. With Wedin (1988:82–83), I take ‘deliberative phantasia’ to mean ‘imagination connected with deliberation’, just as the λογιστική is that connected with λόγος, without prejudicing the question whether such ‘imagination’ always preexists reflection or else can also be forged by the deliberative process. There is no clear proof here of φαντασία as functionally complete, and Wedin’s proposal (1988:45–63) may stand. This conclusion holds even if we understand the ἕν of ἓν ἐκ πλειόνων φαντασμάτων as ἓν [φάντασμα], for deliberation would still have the active role. Functional incompleteness only requires that φαντασία not be regarded as the subject of δύναται … ποιεῖν, an interpretive point that seems entirely warranted (πράξει, διώκει, and δύναται share a common subject which must therefore be ‘rational animals’ vel sim.). That ἕν does not stand for ἓν [φάντασμα] but for ‘one plan, thought, or course of action’ plausibly follows from Wedin’s observation that combining in one’s head elements from different images into a single collage seems a rather odd way of conceiving the ordinary psychology of deliberation. The involvement of several images need not be “reduced to mentally pasting together images” (83, his emphasis), even if one regards this procedure as a rare possibility. Not germane to this question is De anima 434a10–11: καὶ αἴτιον τοῦτο τοῦ δόξαν μὴ δοκεῖν ἔχειν, ὅτι τὴν ἐκ συλλογισμοῦ οὐκ ἔχει, αὕτη δὲ ἐκείνην. Its correct interpretation hinges on the last clause (‘but the latter [has] the former’), which Ross declared “clearly unmeaning” but in fact serves to rule out ‘phantasia’ and δόξα as the respective referents of αὕτη and ἐκείνην; for after 427b14–26 and 428a18–26 the mere notion that φαντασία may have δόξα is out of the question. Such a claim, moreover, seems hardly apropos in its context (hence Ross’s comment and emendation). Since αὕτη cannot stand for φαντασία, neither can it be the elided noun in τὴν ἐκ συλλογισμοῦ: one must therefore read τὴν ἐκ συλλογισμοῦ [δόξαν] (with Wedin 1988:147n60 and Modrak 2001:252, contra Hicks 1907:567 and Nussbaum 1978:264n66, among many). The passage addresses the view that (at least some) animals have an attenuated form of belief (cf. Hamlyn 1968:153; Irwin 1980:127 with n. 24; Wedin 1988:147n60). If so, the question arises: does it entail phantasia? In other words, the context aims to elucidate the manner in which animals move and the role, if any, that phantasia plays in such motion. Animals that lack logos, Aristotle notes, do not have deliberative phantasia. His proof issues from the endoxical view that animals lack δόξα, at least the strong δόξα of inference (which precludes their deliberating; cf. De memoria 453a14: τὸ βουλεύεσθαι συλλογισμός τίς ἐστιν); and because ‘[inferential] doxa has deliberative phantasia’ he concludes that animals do not have deliberative phantasia. ‘Therefore’ (διό) ‘the appetency [of such animals] does not have the deliberative faculty (τὸ βουλευτικόν)’. Thus interpreted, the assertions are contextually compelling and agree with Aristotle’s earlier statements in the treatise. Cf. Bywater 1888:65–67.

[ back ] 155. Cf. De anima 431b6–8: ὁτὲ δὲ τοῖς ἐν τῇ ψυχῇ φαντάσμασιν ἢ νοήμασιν, ὥσπερ ὁρῶν, λογίζεται καὶ βουλεύεται τὰ μέλλοντα πρὸς τὰ παρόντα. Clearly, these φαντάσματα are λογιστικά and βουλευτικά.

[ back ] 156. Stricto sensu, he does not craft, but provokes, φαντάσματα: what he crafts are the visual and aural stimuli that become the object of his audience’s perception and, hence, move their φαντασία (i.e. he ultimately aims at particular φαντάσματα).

[ back ] 157. Or less than rational.

[ back ] 158. Though my analysis does not depend for its validity on it, I am attracted to the view of φαντασία in Wedin 1988 as functionally incomplete (cf. De anima 429a1–2) and co-occurring with actual exercises of the functionally complete faculties (see his chapter 2 for an explication). Thus I also follow his use of brackets for ‘[re]presentation’, which is intended, he notes, “to alert the reader to the fact that I am not foisting on Aristotle the view that we do not actually perceive objects but only make inferences to them from Hume-like images” (Wedin 1988:17n27).

[ back ] 159. τῇ δὲ διανοητικῇ ψυχῇ τὰ φαντάσματα οἷον αἰσθήματα ὑπάρχει (De anima 431a14–15). Cf. Wedin 1988:110–113.

[ back ] 160. Rhetoric 1370a28–30. On this passage and its context see above, §13.5 n. 106, and below, §13.6.

[ back ] 161. Nussbaum 1978 ad loc. translates: “For the affections suitably prepare the organic parts, desire the affections, and phantasia the desire; and phantasia comes about either through thought or through sense-perceptions.” Since here διά governs νοήσεως, Wedin (1988:57n42) is doubtless right to insist that γίνεται διά + gen. must not be construed with “acquisitional or genetic” force. Posterior Analytics Ⅱ.19 makes clear that νόησις cannot occasion φαντασία in the same generative sense in which perception produces it (428b10–11); hence, ‘thought’ cannot be said to engender φαντασία (cf. Wedin 1988:42–43, 74–75). Wedin therefore translates αὕτη δὲ γίνεται ἢ διὰ νοήσεως ἢ δι’ αἰσθήσεως: “imagination comes about because of … thinking … or perceiving” (57n42, my emphasis). The sense of διά + gen., then, is the very one of qualified co-occurrence expressed at De anima 429a2 by ὑπό + gen. (where ὑπό retains some of its locatival force; cf. Wedin 1988:52). Although both might be fairly characterized as ‘causes’, ‘engendering’ φαντασία (which for Aristotle ‘thought’ cannot do) is very different from ‘arousing’ φαντασία (which ‘thought’ arguably does). Cf. Modrak 2001:249–256.

[ back ] 162. Wedin 1988:57 explains: “If, then, imagination is not a full faculty, it is surely involved in the actual use of such faculties. Perception, desire, and thought require it and so do memory and even dreams. Imagination is required because images are required not as the object toward which the faculty is directed but as a means by which a faculty accomplishes [its task].”

[ back ] 163. Cf. Nussbaum 1978:154n19.

[ back ] 164. On πάθη influencing judgment as opposed to action, see Striker 1996:292–293.

[ back ] 165. ‘The soul of animals is defined with reference to two faculties: both by its capacity to discriminate, which is the task of thought and perception …’.

[ back ] 166. διὸ οὐδέποτε νοεῖ ἄνευ φαντάσματος ἡ ψυχή (De anima 431a16–17).

[ back ] 167. Translation by Nussbaum 1978 ad loc. ὁρῶμεν δὲ τὰ κινοῦντα τὸ ζῷον διάνοιαν καὶ φαντασίαν καὶ προαίρεσιν καὶ βούλησιν καὶ ἐπιθυμίαν. ταῦτα δὲ πάντα ἀνάγεται εἰς νοῦν καὶ ὄρεξιν. καὶ γὰρ ἡ φαντασία καὶ ἡ αἴσθησις τὴν αὐτὴν τῷ νῷ χώραν ἔχουσιν· κριτικὰ γὰρ πάντα. Citing a study by John Cooper, she also notes that “there is no need to interpret [κρίνειν] as implying that any kind of explicit or reflective judgment is taking place—and in particular … it need not be associated with ‘explicit verbal performance or the disposition to such—as indeed we can readily infer from [Aristotle’s] ascription of κρίνειν to animals’” (Nussbaum 1978:334). The point is well taken. But a restriction necessary in the case of animals without logos must not disallow the otherwise legitimate implications of the philosopher’s statement for the social world of the polis. Therefore, I think it is right to consider his analysis in the context of κρίσις that involves discursive thought and decision making.

[ back ] 168. Cf. Rhetoric 1377b21–22, which plainly states that ἕνεκα κρίσεώς ἐστιν ἡ ῥητορική.

[ back ] 169. Nicomachean Ethics 1139a3–8.

[ back ] 170. τὸ γὰρ βουλεύεσθαι καὶ λογίζεσθαι ταὐτόν, οὐδεὶς δὲ βουλεύεται περὶ τῶν μὴ ἐνδεχομένων ἄλλως ἔχειν (Nicomachean Ethics 1139a12–14).

[ back ] 171. Cf. Wedin 1988:145–156. One must be careful to regard ‘images’ (φαντάσματα) not as objects of thought but as what they really are, viz. vehicles for thought. As Wedin (1988:116) pointedly observes, “[t]o contemplate with an image is not to contemplate an image.”

[ back ] 172. Cf. Fortenbaugh 2002:23–44.

[ back ] 173. Kassel brackets these lines as a likely Aristotelian addition (Kassel 1976:xix and ad loc.). However one chooses to explain the contradiction noted below (n. 174) that leads him (with Spengel 1867:1.159) to place the passage under suspicion, ultimately what matters is that we realize the underlying consistency of Aristotle’s classification of causes for human action, irrespective of his terminological inconsistency.

[ back ] 174. The terminology employed here contradicts 1369a1–4, where βούλησις (described as ‘a desire for what is good’) is assigned to λογιστικὴ ὄρεξις but ὀργή and ἐπιθυμία to the ἄλογος ὄρεξις. (My point is indifferent to the various textual uncertainties of the passage; I follow Ross’s text.) The contradiction, however, is superficial, for as Grimaldi (1980:231) explains the same conceptual schema, detailed in Nicomachean Ethics 1102a26–1103a10, underlies both passages. Indeed, at Nicomachean Ethics 1102b13–14 Aristotle mentions a subdivision of the soul’s ἄλογον μόριον that ‘somehow shares in λόγος’ ([ἄλλη τις φύσις] μετέχουσα μέντοι πῃ λόγου), for it responds differently in the continent and incontinent man. And a few lines later, at 1102b29–31, he structures the opposition as one between the φυτικόν, which ‘in no wise shares in λόγος’, and the ἐπιθυμητικόν (and in general the ὀρεκτικόν), which ‘somehow does share [in λόγος], in that it hearkens to it and obeys it’ (τὸ δ’ ἐπιθυμητικὸν καὶ ὅλως ὀρεκτικὸν μετέχει πως, ᾗ κατήκοόν ἐστιν αὐτοῦ καὶ πειθαρχικόν).

[ back ] 175. Reading πολλά adverbially, and the verbs and participles absolutely: μετὰ λόγου δὲ ὅσα ἐκ τοῦ πεισθῆναι ἐπιθυμοῦσιν· πολλὰ γὰρ καὶ θεάσασθαι καὶ κτήσασθαι ἐπιθυμοῦσιν ἀκούσαντες καὶ πεισθέντες (1370a25–27).

[ back ] 176. In printing κἀεὶ ἐν I follow Roemer, who adopts Susemihl’s emendation. Alternatively, the κἄν of ΘΠΓΣ (using Ross’s sigla) or β (in Kassel’s stemma) may actually be κἀν = καὶ ἐν (so Kassel, after Arnet), where the ἀεί required by the syllogism, though not explicit, is nevertheless understood. In any case, it is clear that εἰ δὲ τοῦτο summarizes the three protases, and that δέ must therefore be resumptive. Cf. Grimaldi 1980:251 and see also above, §13.5 n. 106.

[ back ] 177. Cf. Watson 1982:103n6 and Wedin 1988:89. I am not hereby endorsing the widespread view of the Rhetoric as a treatise lacking in exactitude. For a survey of the literature for and against this view see Fortenbaugh 1974:222n4 and 223n5. Cf. also Striker 1996:286–288.

[ back ] 178. Hamlyn 1968 ad loc. The Greek runs as follows: κατὰ συμβεβηκὸς δὲ λέγεται αἰσθητόν, οἷον εἰ τὸ λευκὸν εἴη Διάρους υἱός· κατὰ συμβεβηκὸς γὰρ τούτου αἰσθάνεται, ὅτι τῷ λευκῷ συμβέβηκε τοῦτο, οὗ αἰσθάνεται· διὸ καὶ οὐδὲν πάσχει ᾗ τοιοῦτον ὑπὸ τοῦ αἰσθητοῦ. Note how πάσχειν and αἰσθητόν are connected, just as αἰσθάνεσθαι and πάθος.

[ back ] 179. Cf. Wedin 1988:94–95: “[P]erceiving Socrates qua Socrates, perhaps even qua man, involves something more than just proper perception. For one thing it involves the identity of the object, not simply feature identification. So it is likely that belief, thought, and even language may be needed as well” (95).

[ back ] 180. πάθος covers anything that comes about through πάσχειν (cf. DELG s.v. πάσχω). At De anima 403a7, ὀργίζεσθαι, θαρρεῖν, ἐπιθυμεῖν, ὅλως αἰσθάνεσθαι are listed as πάθη τῆς ψυχῆς (for sweetness, cold, heat, etc. as πάθη, see Categories 9b2–9). At 403a17–18 the list is: θυμός, πραότης, φόβος, ἔλεος, θάρσος, ἔτι χαρὰ καὶ τὸ φιλεῖν τε καὶ μισεῖν; all these happen ‘with the body’, the philosopher adds, ‘because together with these the body feels something’ (ἅμα γὰρ τούτοις πάσχει τι τὸ σῶμα 403a18–19). The concept of πάθος extends further to states of mind and body such as sleep (De insomniis 462a3–4: ἐὰν μὲν αἰσθάνηται ὅτι καθεύδει, καὶ τοῦ πάθους ἐν ᾧ ἡ αἴσθησις τοῦ ὑπνωτικοῦ), to μνήμη, ‘memory’ (De memoria 449b4–6), and even to φαντάσματα (De memoria 450a10–11: καὶ τὸ φάντασμα τῆς κοινῆς αἰσθήσεως πάθος ἐστίν [bracketed by Ross, after Freudenthal]).

[ back ] 181. In ἐστὶν τὸ ἥδεσθαι ἐν τῷ αἰσθάνεσθαί τινος πάθους (1370a27–28) πάθος stands broadly for τὸ πάσχειν τι (see above, n. 180). It might thus be rendered by ‘affection’—not in the narrow sense of ‘emotion’ (though emotions are certainly included) but the more general one of ‘the state of being influenced or acted upon’. The connection of αἴσθησις with πάθη is natural, since sense organs must be affected by the corresponding stimuli for perception to occur (De anima 424b25–26). But just as every sense has its corresponding pleasure (Nicomachean Ethics 1174b20–21), so also the emotions (πάθη as animi perturbationes) are accompanied by pleasure and pain: λέγω δὲ πάθη μὲν ἐπιθυμίαν ὀργὴν φόβον θάρσος φθόνον χαρὰν φιλίαν μῖσος πόθον ζῆλον ἔλεον, ὅλως οἷς ἕπεται ἡδονὴ ἢ λύπη (Nicomachean Ethics 1105b21–23).

[ back ] 182. Cf. Wedin 1988:89. Of course, Aristotle’s reasoning does not deduce, but assumes as a premise, what I have presented as the conclusion; his goal is to argue that memory and hope are, in fact, attended by pleasure and pain: εἰ δὲ τοῦτο, δῆλον ὅτι καὶ ἡδοναὶ ἅμα μεμνημένοις καὶ ἐλπίζουσιν, ἐπείπερ καὶ αἴσθησις (Rhetoric 1370a30–32).

[ back ] 183. I read the (notionally) underlying construction as ὑπάρχει αὐτῷ τὸ ἀγαθὸν εἶναι. Once this turns into an articular infinitive qualifying φαντασία, the τό, not strictly required, might be dropped (so cod. Parisinus 1741) to avoid the close succession of τοῦ + inf. and τὸ + inf. (Without the notionally underlying presence of τό, one would have expected ὑπάρχει αὐτῷ ἀγαθῷ εἶναι.)

[ back ] 184. διότι διατρίβουσιν ἐν τῷ τιμωρεῖσθαι τῇ διανοίᾳ (1378b7–8).

[ back ] 185. Concerning 1384a23, Kennedy 1991:146n56 notes: “As usual, this means a mental ‘visualization’ of the effects, not (as the English word may imply) a false conclusion.”

[ back ] 186. Labarrière 1984 argues for an understanding of λόγος that puts the accent on oratio rather than ratio. While animals do not have reason, he says, they are not entirely deprived of rationality: τὸ αἰσθητικόν—which animals obviously have—cannot be easily classified as either ἄλογον or λόγον ἔχον (De anima 432a30–31); besides, many (if not most) animals have φαντασία and μνήμη of particulars (Nicomachean Ethics 1147b5); and, as a survey readily shows (cf. Labarrière 1984:34–40), Aristotle assigns to some a kind of μάθησις and διδασκαλία, and to birds, ἑρμηνεία ἀλλήλοις and even a διάλεκτος. Moreover, following De generatione animalium 786b23–25 and De anima 420b29–33, Labarrière links φωνή—a σημεῖον of pain and pleasure (Politics 1253a10–11)—to αἰσθητικὴ φαντασία. The implication is that the opposition between αἰσθητική and λογιστικὴ φαντασία reflects the one between φωνή and λόγος; and hence, the λόγος in question is not so much ratio as oratio. Whether one agrees with him or not (cf. Wedin 1988:146–152), one can hardly deny that, at the very least, wherever ratio is involved, persuasion is active as “l’espace intersubjectif et dialogal de l’opinion et de la délibération” (32), a space that turns readily into the public sphere of δόξα where the community engages in dialog and persuasion.

[ back ] 187. Cf. Brinton 1988.

[ back ] 188. Even if not a full, i.e. independent, faculty. On its critical nature, see above, §13.6.

[ back ] 189. That such overlap between ēthos and pathos exists is generally denied by Fortenbaugh, who in a series of works (see Fortenbaugh 1996 passim, esp. 147, with bibliography in nn. 3 and 4) has argued that persuasion through character is not intended to arouse emotion in the audience and does not compromise the objectivity of the juror. (His only concession is that at Rhetoric Ⅱ.1 eunoia is thought of as an emotion, not of the audience but of the speaker; but cf. Rhetoric 1415a35–36, where it is paired with ὀργίσαι.) By insisting on the distinction Aristotle draws between εὔνοια and φιλία in Nicomachean Ethics 1166b30–35, he seems to imply that should the audience, for example, see the orator manifest intense φιλία towards them (or their city), they would be inclined to mistrust him; or else that the speaker should seek to restrain such feelings for fear of warping his own or his hearers’ judgment (Fortenbaugh 1996:164). Neither implication squares with intuition or actual practice. (Of course, if such protestations were overdone, they may look inauthentic and fail to convince, but this is entirely another matter.) And even if the distinction of Nicomachean Ethics holds, many an Athenian orator must have boasted strong affection for his city and convinced her of the honesty of his claims, or else the parody in Aristophanes Knights would ring hollow (cf., for example, 732, 773, and 1339–1355). Far better, then, to acknowledge with Carey (1994:35) that, “[i]n practice, ethos and pathos are closely connected, for one of the effects of ethos, as well as inducing a degree of trust, is also to produce a feeling of goodwill in the audience toward the speaker.” (Cf. Carey 1994:39 and 43, and Russell 1990:205–206 and 212.) In other words: the orator succeeds when his εὔνοια and φιλία are reciprocated by his audience.

[ back ] 190. Here εἰ καί, like κἄν at a24, must mean ‘even if’ (cf. Smyth §2378), for the point is not to assert the necessary disjunction between the facts and the pathos of the λέξις but to underline the effectiveness of the style even if such disjunction obtains. Hence, the mind’s reasoning is ill-founded but not necessarily false. In Aristotelian usage, the verb παραλογίζεσθαι (1408a20) and its noun παραλογισμός—nothing suggests semantic variance between them—may be applied to inferences that, although ill-founded, hit upon the truth. Hence it need not entail, though in context it well may imply, that the conclusion drawn is not true to fact. Two considerations establish beyond question this (otherwise reasonable) semantic claim. The first is Aristotle’s equating παραλογισμός and ὁ φαινόμενος ἔλεγχος/συλλογισμός in the Sophistical Refutations (see Sophistical Refutations 164a20–21 with §13.6 n. 140 above; and Sophistical Refutations 169b18–20, 170a9–10 with Schreiber 2003:174–175 and 229n3); the second, his explicit mention of a sophistic reasoning that is ‘seemingly deductive’ (φαινόμενος συλλογιστικός) on subjects which only dialectic is fitted for testing ‘even if its conclusion is true (for it deceives about the cause)’ (κἂν ἀληθὲς τὸ συμπέρασμα ᾖ (τοῦ γὰρ διὰ τί ἀπατητικός ἐστι) 171b10–11). This precise scenario is in view at Rhetoric 1408a19–24.

[ back ] 191. At 1408b12 Aristotle refers to this embrace with the word συγγνώμη, which some erroneously think an expression of condescending toleration and translate as ‘it is excusable’ (e.g. Kennedy ad loc.: “for it is excusable that an angry person calls a wrong ‘heaven-high’”). Here συγγνώμη is best taken instead etymologically as ‘con-sensus’, i.e. ‘concurrence in view or judgment’, a sort of ‘critical fellow-feeling’ from which the acceptations ‘forbearance’, ‘excuse’, and ‘pardon’ derive. In view is the audience seeing something from the speaker’s vantage point once he has emotionally attuned them to his message: ‘Compound words … suit one who speaks with emotion; for there is commonality of mind with an angry speaker who calls an evil “heaven-high” … and whenever he has already gripped his audience and has caused them to grow deeply stirred with praises, reproaches, anger, or friendship … . For they utter such things when they are deeply stirred and also in consequence approve [of him/this], obviously because they are similarly affected’ (1408b12–18: τὰ δὲ ὀνόματα τὰ διπλᾶ … ἁρμόττει λέγοντι παθητικῶς· συγγνώμη γὰρ ὀργιζομένῳ κακὸν φάναι οὐρανόμηκες … καὶ ὅταν ἔχῃ ἤδη τοὺς ἀκροατὰς καὶ ποιήσῃ ἐνθουσιάσαι ἢ ἐπαίνοις ἢ ψόγοις ἢ ὀργῇ ἢ φιλίᾳ … . φθέγγονταί τε γὰρ τὰ τοιαῦτα ἐνθουσιάζοντες, ὥστε καὶ ἀποδέχονται δῆλον ὅτι ὁμοίως ἔχοντες. Note that I do not punctuate with a period before καὶ ὅταν but read καὶ [συγγνώμη] ὅταν; for the construction τε γὰρ … ὥστε καί see Bonitz 1867:676–677 contra Denniston 1954:536). The relevant sense of συγγνώμη, which regards the hearers’ embrace of the speaker’s [re]presentation with their own δόξα, is best defined by Aristotle himself at Nicomachean Ethics 1143a19–24: ἡ δὲ καλουμένη γνώμη, καθ’ ἣν συγγνώμονας καὶ ἔχειν φαμὲν γνώμην, ἡ τοῦ ἐπιεικοῦς ἐστὶ κρίσις ὀρθή. σημεῖον δέ· τὸν γὰρ ἐπιεικῆ μάλιστά φαμεν εἶναι συγγνωμονικόν, καὶ ἐπιεικὲς τὸ ἔχειν περὶ ἔνια συγγνώμην. ἡ δὲ συγγνώμη γνώμη ἐστὶ κριτικὴ τοῦ ἐπιεικοῦς ὀρθή· ὀρθὴ δ’ ἡ τοῦ ἀληθοῦς (‘The so-called “judgment” according to which we say that people “exhibit commonality of mind” and speak of “being of a mind [to]” is a right discrimination of what is reasonable. The proof: we say that the reasonable man most of all “exhibits a sympathy of judgment” and that it is reasonable to exhibit commonality of mind regarding some things. Commonality of mind is a right critical judgment of what is reasonable; and correct judgment is a judgment about the truth’). On this passage cf. Rowe and Broadie 2002:185, which renders συγγνώμονας as “having a shared sense” and τὸ ἔχειν συγγνώμην as “to be sympathetic.”

[ back ] 192. This is elaborated in connection with the poet’s art in Poetics 17 1455a22–32. The meaning of λέξις in τῇ λέξει συναπεργάζεσθαι (a22–23) goes beyond mere ‘verbal expression’ to include the rhetorical qualities of propriety (the presence of τὸ πρέπον and the absence of τὰ ὑπεναντία 1455a25–26) and vividness (ὅτι μάλιστα πρὸ ὀμμάτων τιθέμενον 1455a23). On this passage see Meijering 1987:14–17.

[ back ] 193. For the Greek, see above, §13.4 n. 53. This passage of the Rhetoric is similar in sentiment to the beginning of Poetics 17, on which see immediately above, n. 192. Cf. Dupont-Roc and Lallot 1980:281–284.

[ back ] 194. Rhetoric 1386a33–34.

[ back ] 195. Even in the case of φαντάσματα strictly subservient to the ‘bare facts’ (such as might be employed by the geometer), one must always remember that ‘forms’ (universals), not φαντάσματα, are the objects of thought: τὰ μὲν οὖν εἴδη τὸ νοητικὸν ἐν τοῖς φαντάσμασι νοεῖ (De anima 431b2; cf. 429a27–29 and 432a2–3). Hence, Wedin (1988:91) notes: “The contemplating geometer may have an image ‘before his eye,’ but he does not extract information from it by inspection or any other procedure. Information is forthcoming in virtue of his doing something, namely, proving theorems. And although this requires images, the images merely subserve the thinking. They are not what the thinking is about” (his emphasis).

[ back ] 196. For πάθη as rhetorical πίστεις see Solmsen 1938 and Conley 1982.

[ back ] 197. By implication, this may be said to apply even to ῥητορική as a whole because its ultimate orientation is πρὸς δόξαν, i.e. its goal is to shape opinion.