4

Ratio dicendi

What is more uncouth than to speak about speaking (de dicendo dicere), when speaking itself is never other than uncouth, except when it is necessary?

so crassus (1.112), as he reluctantly begins to answer his young friends’ question about whether or not there really is an ars dicendi. Crassus dislikes the self-conscious self-description required for an orator to speak about his own craft; but the phrase de dicendo dicere also reflects the ambiguities inherent in speaking about speech, ambiguities Cicero’s book both describes and enacts. The definition of the various terms that can be used to talk about speech (of whatever sort) is a recurring problem in De oratore: as will be discussed below, the several participants in the conversation offer at times quite different definitions of what an orator is and does. De oratore concerns a range of topics related to speech, and to some extent it explores the language used to talk about speech itself: how does “rhetoric” relate to “oratory”; how does “oratory” relate to “speech”; what is the relationship between eloquence and rhetorical training; how does one become an orator?

This last and most practical question is the starting point of the dialogue: after the participants sat down on their pillows under the Platonic plane tree, Crassus, according to Cotta, began to talk de studio dicendi, praising Cotta and Sulpicius and launching into an encomium of oratory and of the power of speech in general (1.29–30).¹ In a short speech, Crassus manages to attribute a great many capacities to speech itself, and he uses a number of different words to talk about speech. He begins with the broad power of speech (dicendo) to direct and control people’s desires and emotions, something particularly important “in every free people and particularly in states that are calm and at peace” (1.30)—that is to say, in societies that use language rather than force as an instrument of control. He praises both the pleasure and the broad powers given by speech (1.31–32; oratio, used twice), including the ability both to help those in need and to defend oneself against one’s enemies. But then Crassus goes further, praising that one ability, speech, which most distinguishes man from other animals (1.32–34): on the one hand he adduces the pleasures of private conversation (sermo facetus), on the other the power of speech to create cities and civilization. That leads to his concluding encomium to the power of the perfectus orator to give salus to the res publica and an exhortation to Cotta and Sulpicius to continue onward in the pursuit of oratorical success: in id studium in quo estis incumbite.

Crassus’ speech about speech assigns many different capacities and roles to the orator, and he uses a corresponding range of words. The set of terms he uses, moving from dicere to oratio to sermo before returning at the end to the perfectus orator, in part reflects Cicero’s typical desire for variatio and his general tendency in De oratore to avoid using the technical language of the rhetorical schools. At the same time, however, Crassus’ vocabulary seems to mirror his changing definition of his subject: at different moments he sees dicere as the performance of forensic and political oratory within the civic sphere (protecting the afflicted, defending oneself, moving the emotions of the crowd); as all artfully composed utterances, including witty and urbane conversation like that which takes place in De oratore itself; and as the very source and origin of society and law, the basis of all civilization and (in an Isocratean mode) a defining human characteristic and the highest of all human arts.

Few people, one suspects, would consider Crassus’ speech a praise of oratory in any normal understanding of that word: it includes oratorical performance, and it is intended to encourage Cotta and Sulpicius to persevere as orators, but in essence it argues that the successful orator is someone who has complete control of all the arts of speech. Scaevola, in response, is unconvinced, and his own definition of what an orator is and does is much simpler (1.44):²

It’s a big enough thing to be able to guarantee that whatever argument you make in trials appears to be the better and more plausible; that when you speak in public or in the senate your speech is the most persuasive; and finally that you appear to men of wisdom to speak well and to fools to speak the truth.

This definition of the orator as one who speaks effectively in the three major venues for oratory in Rome (in public, in the senate, and in court), corresponds to a small part of Crassus’ definition, but it addresses not the substance of speech—which is essential to Crassus’ definition of orator as conversationalist, lawmaker, and advocate—but its performance and persuasive ability. It seems closely related to the traditional claim of (or accusation against) the sophists, of making the worse argument seem the better. Later in Book 1, Antonius gives a definition that is similar to Scaevola’s (1.213):

I think that he [sc. the orator] is someone who can make use of words that are pleasant to hear and ideas that are likely to be convincing when speaking in court or in public. I call this person an orator, and I want him also to be equipped with a good voice and movement and a certain charm.

Antonius’ analysis of the speaker’s task includes Scaevola’s concern with the audiences for oratory, but it also considers something very important for De oratore, the rhetorical training and techniques through which one learns to be an orator: the use of words pertains to elocutio, convincing ideas to inventio, and voice and movement to actio.³ The reference to charm is not part of the traditional syllabus of the rhetorical schools, but in context it is a nod to Crassus, who had included sermo facetus as an element of the orator’s toolkit in his encomium (1.32), as well as an anticipation of Caesar Strabo’s discourse on humor in Book 2.⁴

All these definitions of the orator, including Crassus’, treat oratory in some ways as performance; only Antonius looks, if only obliquely, at training as well. In the tradition of writing about speech prior to De oratore, however, the focus was on the latter: that is true of Cicero’s own De inventione as well as of the anonymous Rhetorica ad Herennium, both written early in the first century. These are texts not about oratory or the orator, but about rhetoric, the system of education that trains someone to become a proficient orator. Even Crassus, when he reluctantly introduces his version of rhetorical theory in the middle of Book 1, knows what that means, and he cites as the first of the rhetorician’s trita praecepta a definition of the orator very much like that of Scaevola and Antonius, but completely unlike the expansive definition in his own first speech: “The orator’s first responsibility (primum oratoris officium) is to speak in a manner suitable for persuasion” (1.138).⁵ The emphasis on persuasion and proof in all three definitions (ad persuadendum, 1.44, 1.138; probabilior, 1.44; ad probandum, 1.213) matches the words of the rhetoricians’ handbooks, including Cicero’s own (cf. Inv. 1.6); it corresponds to a definition ascribed to the Hellenistic rhetorical theorist Hermagoras of Temnos (F4 Matthes):⁶

Hermagoras said that it was the function of a complete rhetorician (τελείου ῥήτορος) to set out as persuasively as possible whatever civic issue was raised.

But there is a very important difference between Hermagoras’ definition and those of Scaevola, Crassus, and Antonius: the Romans are talking about the orator while Hermagoras is talking about the rhetor. The latter term has an ambiguity that the Latin orator does not: rhetor in Greek refers to both the teacher of rhetoric and the orator, and similarly rhetorike refers to both the study of rhetoric and the performance of oratory itself.

The difference between Greek and Roman ways of talking about speech—the divergent meanings of “rhetoric” and “oratory”—is part of the background to De oratore, and the ambiguity of rhetorike plays a significant role in the argument over the place of rhetoric in the structure of intellectual life, something to be discussed later in this chapter and in more detail in Chapter 6. In De oratore, moreover, rhetoric plays a double role, corresponding to the two meanings of rhetorike: although it is something elementary, a set of rules for teenagers to learn before embarking on a public career, it in fact provides the structure of De oratore itself. Rhetoric is trivial but inescapable: it provides the language through which the characters of De oratore extend its reach far beyond what any teacher of rhetoric ever contemplated.

By the dramatic date of De oratore in 91, rhetorical education was well established at Rome, but not without some controversy. In 92, only a year earlier, Crassus and his fellow censor Cn. Domitius Ahenobarbus had issued an edict against the schools of the Latin rhetors, calling them (at least in Crassus’ words in De oratore) “a school of shamelessness.” The exact source of the censors’ disapproval is unclear: according to the wording of the edict, young men were wasting their time in these schools, contrary to mos maiorum; in De oratore, Crassus suggests that Greek teachers of rhetoric at least taught something, while the Latin rhetors taught nothing at all.⁷ Latin rhetorical education clearly continued, but the edict shows that for whatever reason, some people, including Crassus, had found it objectionable. Despite censorial disapproval, however, the two earliest surviving texts of rhetorical theory in Latin were written not much later: Cicero’s own De inventione was certainly written no later than the mid-80’s, and possibly as early as 92 or 91; Rhetorica ad Herennium is later than De inventione, but no later than the 80s.⁸

Formal rhetoric, the rules for composing a speech that were drummed into the heads of teenage future advocates and statesmen, forms the backdrop to the conversation of De oratore. Both De inventione and Rhetorica ad Herennium rely, at one or more removes, on Greek rhetorical theory as it was elaborated in the generations after Aristotle and Theophrastus, who between them had created its basic structure and organization. The most influential Greek ars rhetorica, to which both Cicero and Rhetor are heavily (if indirectly) indebted, was by Hermagoras of Temnos in the middle of the second century; his definitions of rhetoric and of its parts became standard, along with his division of oratorical subjects into those that dealt with general topics (Greek thesis, Latin quaestio) and those that dealt with particular individuals and events (Greek hupothesis, Latin causa).⁹ Hermagoras also articulated the four fundamental issues (Greek stasis, Latin status or constitutio) that served to structure forensic orations.¹⁰ All these questions concern just the first officium of the orator, inventio—the choice of what arguments to deploy.¹¹

There were, of course, a number of anatomies of rhetoric: the old (pre-Aristotelian) method of division into the sections of a speech (introduction, narration, argument, etc.); the Peripatetic structure of the five officia of the orator (inventio, dispositio, elocutio, memoria, actio); the varieties of oratory, also Aristotelian (forensic, deliberative, epideictic); the Hermagorean division into general and particular issues; and the doctrine of status.¹² Cicero in De inventione never got beyond the first of the five officia, although he seems to have planned to write a full treatise on rhetoric.¹³ He first speaks broadly of the divisions of rhetoric (both Hermagorean thesis/hupothesis, which he criticizes, and the older structures) and then plunges into status, before devoting most of Book 1 to an anatomy of invention according to the sections of the speech; in Book 2 he returns to status to analyze the types of argument and topic suitable to the various forms of case. He only briefly discusses deliberative and epideictic oratory at the very end. The structure of Rhetorica ad Herennium is similar, except that Rhetor does not mention status until he discusses the portion of a speech concerning proof, and he devotes far more space than Cicero to the kinds of argument suitable for the different kinds of legal issue; he also devotes more space to deliberative and epideictic oratory.

The fine points of rhetorical theory are not relevant here: for the most part they are not discussed by Crassus and his friends in De oratore either. Cicero himself refers to De inventione as an embarrassing production of his youth (1.5, 1.23); he cites Quintus as thinking that books of rhetoric are agrestes and points to the ieiunitas of writers about rhetoric (2.10). Within the conversation, both Antonius and Crassus apply the adjective perridiculus to rhetoric and its teachers (2.77, 3.75); Crassus speaks of pervolgatas res (3.149, cf. also 2.127) and says that everyone should laugh at anyone who thinks that eloquence is contained rhetorum praeceptis (3.54). Indeed, the words rhetor and rhetorica are invariably pejorative; and when Cicero reports Quintus as saying that one rhetor in the family is enough (2.10, referring to Cicero himself), it is obviously ironic.¹⁴ Criticisms of school rhetoric appear frequently. Crassus, in introducing his first account of the basics of oratory (1.113–15) gives strict limits to the utility of the ars; Antonius (2.64–66) says that ars can only give you a start. Rhetorical instruction is in any case obvious, and either not related to real oratory (as Catulus says, these Greek doctores know nothing about real public speaking, 2.75) or simply derived from it by observation. The most favorable comment, perhaps, is Crasssus’ observation made in the interval of conversation between Antonius’ discussion of inventio and Caesar Strabo’s speech on wit (2.232):

“As if,” said Crassus, “there might be any ars of the things Antonius has been talking about for some time. It is, as he said himself, the observation of the things that work in speeches, and if that could make someone eloquent, then who wouldn’t be eloquent? Who couldn’t learn all this either easily or at least somehow or other? But I think that the effectiveness and usefulness of these rules is not to lead us scientifically to discover what to say, but so that we can be confident that what we achieve by nature, study, and practice is either right or wrong, when we have learned a standard to judge it by.”

The ars rhetorica is derived from studying speakers; its purpose is merely to verify and reassure our natural instincts and experience.¹⁵ As Antonius puts it, “between talent (ingenium) and hard work there’s not much space left for ars” (2.150). Nevertheless, while De oratore insists on the relatively minor importance to the orator of the ars rhetorica, the bulk of the dialogue does in fact take its structure from a traditional Hellenistic ars: Books 2 and 3 deal, respectively, with inventio, dispositio, and memoria and with elocutio and actio. That is not quite the standard order of rhetorical textbooks (memoria belongs after elocutio), but it is definitely the standard syllabus.

Despite its consistent rejection of school rhetoric, De oratore contains at least five artes rhetoricae; before discussing them, it may be useful to give a list:

Location	Size	Speaker
1.17–18	very short	Cicero
1.128	very short	Antonius
1.137–48	short	Crassus
1.78–83	short	Antonius
2.99–3.227 (with interruptions)	very long	Antonius and Crassus

Two of these are very brief and two are slightly longer accounts of standard rhetoric, leading up to the fifth, the large and highly untraditional ars that occupies most of Books 2 and 3. The first very brief and allusive ars comes in the preface, when Cicero is explaining why it is that there are so few orators; his description of the difficulty of the subject includes four of the five standard officia of rhetoric (although not in the standard order, as actio comes before memoria), omitting only dispositio (1.17–18); he also implies other sets of subdivisions, with gestures toward pathos and movere. This is not a full ars, of course, nor is the other highly abbreviated list of rhetorical officia given by Antonius at 1.128, also omitting dispositio.

These two are no more than nods to a familiar organization of rhetorical study. The somewhat longer account of rhetoric given by Crassus in Book 1 is fuller, but in some ways equally allusive. Crassus divides the basic requirements for making a good orator into a traditional set of three categories: natura or ingenium (raw talent and physical attributes), ars, and exercitatio (practice). Of these natura is by far the most important and is given the greatest attention (1.113–33; 147 lines of the Teubner text); exercitatio is given rather less space (1.149–59; 90 lines); and in the middle comes the shortest, ars (1.137–48; 61 lines, including prefatory and concluding remarks).¹⁶

Crassus begins his ars rhetorica by describing what he is about to say as contrita praecepta; he ends by saying once more (1.146) that ars is based on the observation of eloquence, not the other way round. In between he summarizes with bewildering speed the standard fare of rhetorical handbooks: the definition of oratory, the division into quaestio/thesis and causa/hupothesis, the subdivision of causa into the four basic forms of status, and the question of scriptum and sententia all together take fewer than eleven lines of text. Then comes the subdivision of oratory into forensic, deliberative, and epideictic and the loci appropriate to each (8 ½ lines), followed by the five officia (6 lines), the parts of a speech (7 lines), and finally back to elocutio (including a quick glance at the four virtutes of style), actio, and memoria (9 lines) before Crassus says generously that none of these praecepta is useless, but they really only serve as reminders to someone who already knows what he is doing.

Everything in this breakneck summary is genuine rhetorical theory, but it is neither adequate nor coherent; it is hard to believe that it is meant entirely seriously. Within two pages, Crassus smashes together several different structures of rhetorical instruction: the question of quaestio and causa comes up twice; the partes come up twice; there is a mixture of instruction on all three branches of oratory with instruction that is primarily aimed at courtroom oratory (the four status and scriptum et sententia). It is not clear how the discussion of the parts of the speech meshes with the officia of oratory. That is in fact a genuine problem in rhetorical theory, and every element in Crassus’ crash course can be paralleled in one rhetorical handbook or another, but the way it is compressed reveals Crassus’ contempt for this kind of instruction. It is no wonder that when Crassus finishes, the young men complain that it went past them much too fast (1.148, 161), and old Scaevola scolds Crassus for not taking their questions seriously (1.164).

The other brief ars is given by Antonius in Book 2 in introducing the extended ars of Books 2 and 3.¹⁷ There, after dismissing contemptuously the Greeks’ ignorant assumption of everyone else’s ignorance (2.77), he allows that their rhetoric is occasionally mildly useful, but that their general theories of speech-making are ludicrous. He then summarizes a (Hemagorean) rhetorical handbook: beginning with the division into quaestio and causa, the rhetoricians subdivide causa into the five officia (here called membra eloquentiae) and then turn to the partes orationis of varying number, assigning particular tasks to particular parts of the speech (2.78–81). Antonius describes this as a tidy system, but one that does not correspond to the needs of the orator in the real world: the speaker’s aims (captatio benevolentiae, or docere, or suadere) need to pervade the whole speech, not just one section of it, and winning a case or an argument is not so tidy a task as schoolteachers make it out to be (2.82–84). He compares the relationship between rhetorical education and real-world oratory to that between a gladiatorial school and real warfare: it teaches you something, but it doesn’t win any battles. While Crassus’ mini-ars in Book 1 is parodic and emphasizes the intellectual emptiness of rhetorical training, Antonius’ short ars is patronizing. He emphasizes how little experience rhetoricians have of the real world and the courtroom, and how impractical and irrelevant their teaching is. That rejection of the standard ars leads directly to his account of inventio, the first of the officia in the huge ars on which Antonius (in Book 2) and Crassus (in Book 3) collaborate. It is a real ars rhetorica, with a great deal of detail—but while it adopts the structure, it totally reshapes the content of any rhetorical handbook that came before it.

Antonius’ long discussion of inventio (2.99–216), in fact, is completely unlike the rhetorical handbooks.¹⁸ Not only does he simplify the discussion of stasis, clarifying it by using major Roman cases as examples, but, much more important, he also reduces the significance of Hermagorean mechanical methods of discovering what arguments to use by paying at least as much attention to ethos and pathos as he does to formal argument. Even within the discussion of argument, moreover, his treatment of the commonplaces (topica) is much closer to Aristotle than to any rhetorical handbook.¹⁹ Cicero may not have read the Rhetoric directly, but his account of invention very clearly draws on the philosophical tradition—as Antonius himself emphasizes.²⁰ What is more, Antonius’ account of invention is supplemented by a long discussion of humor, delivered by Caesar Strabo (2.216–90)—again, something extraneous to the rhetorical tradition.²¹ When Antonius takes over again, he addresses how to organize a speech (dispositio, 2.307–32), an account that is much briefer—as was generally the case in rhetorical handbooks—and is somewhat less different from handbook treatments than is the case with inventio.²² Still, it seems to owe more to Peripatetic rhetoric than to Hermagoras: as in Aristotle, and unlike the handbooks, the discussion of the sections of a speech is subordinated to dispositio rather than inventio. As for the third major section of the ars, elocutio, Crassus in Book 3 turns a discussion of style into a complex exploration of the relationship between rhetoric and philosophy, on the grounds that eloquent language and ornatus must include res as well as verba (3.19–143). Only after that, on Sulpicius’ insistence, does he reluctantly address the questions about diction and figurative language that elocutio traditionally involved (3.148–212).²³

Taken as a whole, the large ars rhetorica of Books 2 and 3 is meant to enhance and supplement the standard fare of the rhetorical schools. It cannot replace the traditional ars—and that is in part why Crassus and Antonius provide their two more traditional mini-artes—because the kinds of elementary instruction the ars contains about the parts of a speech and the officia of the orator are not only useful, but a necessary prerequisite for the more advanced training offered in the rest of Books 2 and 3. Antonius and Crassus have somewhat different ideas about how the ars should be modified: Antonius turns back from the mechanical rhetoric of the Hellenistic age to the far more subtle and psychological framework of Aristotle and the Peripatetics.²⁴ Crassus shares with Antonius a desire to look to philosophical models to improve standard rhetoric but goes well beyond Aristotelian conceptions to suggest that rhetoric has remained much closer to the original shared purpose of philosophy and rhetoric than have the post-Socratic schools, and he goes much further in questioning the relationship between rhetoric and philosophy; the cultural history that forms a major part of Crassus’ speech on elocutio will be discussed more fully later. The approaches to the ars of the two orators also correspond to their approaches to the larger issue of the role and nature of oratory itself, with Antonius favoring the practical orator, Crassus an ideal. That is a topic to which I will return shortly; but before looking more closely at Crassus’ perfectus orator and Antonius’ practical orator, it is worth examining how Antonius and Crassus deal with the most contested element of rhetorical theory, the difficult problem of the rhetorical thesis. Consideration of that will move us closer to the larger question of the relationship between philosophy and rhetoric; and the different answers to that question given by Antonius and Crassus offer a partial explanation of their two very different understandings of what a real orator is and does.

As mentioned above, Hermagorean rhetorical theory made a primary division of the types of oratory into speeches that address a general issue (thesis in Greek, quaestio or quaestio infinita in Cicero’s Latin) and speeches that concern particular events and individuals (hupothesis in Greek, causa or quaestio finita in Latin).²⁵ As Crassus describes the difference in his first swift ars rhetorica, “every speech either concerns an unlimited topic without the mention of specific individuals or dates, or an event tied to specific individuals and dates” (1.138); the same description is echoed at slightly greater length by Antonius (2.41–42) in summarizing the previous day’s conversation for Catulus and Caesar Strabo. These initial accounts of thesis are standard; as the full ars of De oratore Books 2 and 3 develops, however, the division of thesis/hupothesis is addressed quite differently by Antonius in the discussion of inventio and Crassus in the discussion of elocutio.

In the first place, as Antonius complains in discussing the varieties of oratory (2.65), although writers on rhetoric begin by dividing the subject into quaestio and causa, they never, in fact, say anything about how to deal with quaestiones. There are reasons for that, of course. The obvious and practical one is that it is much easier to accommodate the Hermagorean theory of invention, with its division of cases according to the issue to be resolved and its very precise dissection of the varieties of legal argument, to specific cases rather than to moral dilemmas: “Is killing justified?” presents much greater problems of analysis than “Was the killing of Tiberius Gracchus justified?” Rhetorical education was much more useful to budding advocates if it emphasized the kind of thing they needed to be able to do once they graduated. The theoretical issue is more problematic: Hermagoras and his followers seem to have devoted no space at all to the discussion of how to argue a thesis; but if the work of the orator involves only arguments about hupotheseis, then the initial division of the possible subjects of speech into thesis and hupothesis is otiose.²⁶ Does thesis properly belong to the realm of rhetoric at all, or is it, in fact, a part of philosophy?

The problem of thesis is addressed by Antonius several times. His introductory discussion of the scope of rhetoric begins with the standard division of thesis/hupothesis (2.41–42), but he then points out that hupothesis is only really relevant to a few rhetorical contexts, “either to pleading cases or to giving advice” (43). In other words, standard rhetorical theory limits itself to forensic and deliberative oratory and has nothing to say about the third Aristotelian division, speeches of praise and blame (i.e., epideictic) (43–47), or, for that matter, about a whole range of other forms of speech: giving evidence (48), making reports or giving instructions (49), obiurgatio and consolatio (50) and, at considerable length, historiography (51–64); along the way, Antonius simply dismisses technical philosophical writing and poetry as unintelligible or not worth the effort (61). After this list of gaps in rhetorical theory, Antonius returns to the division between thesis and hupothesis (65) and seems to propose a modification of standard theory. On the one hand, thesis (here infinita quaestio) seems to include all topics that are not limited by person or event, and thus requires the ability to talk about the size of the sun and the shape of the earth as well as mathematics and music (66); that is clearly not something Antonius wants to require of the orator. On the other hand, the orator must talk about broad ethical and political issues, albeit within limits: “so we should take up that area too, but in such a way that it is limited to a moderate scope” (67). And, finally, the orator should be able to speak about civic matters not the way philosophers do, separatim, but in the way lawgivers and the founders of states do, “in a simple and grand style, without chains of arguments and without a dry crowd of words” (68). None of this, however, needs separate instruction: once you learn the basics of forensic and deliberative speech, the rest is easy (69–70).

Antonius’ first criticism of the Hermagorean division is minor: it is theoretically wrong in two ways, first because it does not cover all the areas of rhetorical speech (including history as well as other forms of oratory), and second because the orator does need to speak about broader issues than fall under the narrow category of hupothesis. On the other hand, Antonius recognizes that not all broad topics are suitable for rhetorical instruction: the orator needs to be able to talk about ethical and political issues, but not in the same way that a philosopher does. Antonius’ discussion thus leaves intact what seems to have been one of the standard approaches to the problem: namely, that thesis, speaking on general topics, was the province of philosophy. The fact that he uses the magnitude of the sun as his first example of what an orator should not have to speak about is indicative: it was a standard thesis in philosophical education.²⁷ Antonius’ problem here concerns the main issue in all the discussions of thesis in De oratore: what belongs to rhetoric, and what belongs to philosophy?²⁸

The separation of rhetorical and philosophical thesis that Antonius is suggesting in this passage is not his last word on the subject. Once he moves from the general discussion of the purview of rhetoric and oratory to the specific topic of inventio, the issue of thesis surfaces once more. Thus, in discussing how to discover the fundamental issue of a case (in the technical language of Greek rhetoric, the krinomenon), Antonius goes through the basic situations of forensic oratory, the three most important staseis: whether or not something happened (factumne sit); its characteristics (quale sit); and how to define it (quod nomen habeat) (2.132).²⁹ But for Antonius, these basic courtroom arguments hinge not on the particular facts of a case but on the more general questions that apply to those facts (2.133):

This is where you have to see how big a mistake those teachers make, the men to whom we send our children—not that this is particularly relevant to oratory, but so that you can see what sort of dumb and rough people those men, who think they are learned, really are. In setting up the categories of speeches they distinguish two types of case, and identify one as the kind in which there is a universal question not involving people or times, and the other as defined by specific people and times. They just don’t understand that all disputes are taken back to a basic, underlying general question!

Antonius elaborates on this at greater length: all causae are based on quaestiones (ex genere quaestionis pendere causas, 2.139).

That is, in fact, a much stronger argument against the Hermagorean division of thesis and hupothesis than Antonius’ earlier objections.³⁰ There, the problem was simply that the division was too crude and did not recognize that orators often have to talk about broad issues of ethics or civic life. Here, the point is that thesis is in fact the essential core of all legal arguments (except for questions of fact, the status coniecturalis, but he does not discuss that).³¹ Hence to make such a strong distinction between the general and the specific, the broad and the particular, is not merely useless, but wrong: what moves a jury is not a claim about individuals and specific circumstances, but a claim that a particular action is part of a larger pattern of moral or civic behavior. But although Antonius’ argument touches on the crucial question of the division between rhetoric and philosophy, his objection to the Hermagorean division remains firmly within the realm of rhetoric and oratory: Antonius criticizes schoolroom rhetoricians because they do not understand how forensic argument actually works, but he does not address the larger and more important issue of the place of thesis in the broad organization of the disciplines of language.

That, on the other hand, is just what Crassus does when it is his turn to attack the rhetorical theory of thesis in Book 3.³² In the first place, he argues, the most admirable kind of eloquence is to emphasize a subject’s importance through elaboration: amplificare rem ornando (3.104). One of the best ways to do that is through the use of praise and blame, particularly through the use of loci communes (105–6), some of which can only be used in one direction—attacks on moral turpitude or deprecatio—while others can be used either favorably or unfavorably (107). But these loci, Crassus says, are now the province of philosophers, specifically of Academics and Peripatetics; and although they used to be the shared possession of all speakers, they now must be borrowed from those who have taken them from us (108):

Since we have been driven from our possession [sc. of this wider area of argument] and left in a tiny little property that is itself subject to dispute, and since although we act as patrons for others we have been unable to hold on to and protect what is ours, we must borrow what we need from the people who have broken into our ancestral home—and that is truly disgraceful.

From the idea of commonplaces (loci) and the need to borrow them from philosophers, Crassus slides unobtrusively into the question of thesis: the philosophical schools have divided speech on public topics (civilem orationem) into these two categories of general and specific questions; rhetoricians have followed this division in their teaching, but they are clearly losing the battle with philosophers, and while the rhetoricians barely cling to their claim to own the theseis, some philosophers (Philo, the head of the Academy) are now laying claim even to the hupotheseis.³³

Like Antonius, Crassus objects to the Hermagorean division and to the rhetoricians’ abandonment of thesis, but his solution steps outside the rhetorical framework accepted by Antonius. Throughout the first part of his discussion of elocutio, Crassus is attempting to explain and explore the relationship between rhetoric and philosophy and hoping to reclaim rhetoric as the universal discipline of intelligent speech. In accordance with that, he here, as elsewhere in this speech, claims an original unity of the two as represented in early sapientes like Solon and Lycurgus. In the past, such people were called politici philosophi, taking their title from the whole city, whereas nowadays they are named only from a tiny space, such as the Peripatos or the Academy (3.109);³⁴ in other words, the separation of philosophy from wisdom and speech has led to a parochialism in philosophy as well as in rhetoric. As far as the theseis are concerned, Crassus not only rejects the rhetoricians’ abandonment of thesis out of incompetence, he also proposes that just as the theory of status constructs a system of arguments for dealing with particular cases, so too can there be a parallel system of general arguments for dealing with theseis; and he proceeds to construct one. In other words, rather than accept, as Antonius reluctantly does, a damaging and artificial division between the two great disciplines of speech, Crassus creates a rhetorical structure for philosophical argument based on thesis. Instead of permitting philosophers to claim ownership of rhetoric, he provides a way for orators to reclaim the ownership of ideas. Antonius’ two arguments about thesis together with Crassus’ argument form a sequence.³⁵ Antonius starts from the minimalist position that rhetoric should borrow just a little from the philosophers to talk about general ethical or civic ideas; he then progresses to the stronger stance that in fact almost all legal issues turn not on the particulars of a case, but on the underlying thesis, and thus that thesis is a proper part of rhetoric. Crassus takes that one large step further, by arguing that it is just as possible to construct a system of thesis-arguments within rhetoric, one that sets up an analysis of abstract ethical-political issues that, as it should, subsumes hupothesis within it. From a system of arguments about cases, he constructs a system of arguments about concepts.

The various approaches to thesis taken by Antonius and Crassus travel a long way from the jejune Hermagorean manuals of rhetoric, but that is exactly what De oratore does. Over the course of Books 2 and 3, the structure and purpose of traditional rhetoric are placed in a larger perspective that reduces their importance. It is all very well to insist, as the handbooks do, that inventio—the gathering of material to use in a speech—is limited to the discovery and construction of appropriate arguments; Antonius instead emphasizes the importance of having an effect on the jury, whatever the arguments may be. It is all very well to say that the forensic or deliberative orator is concerned only with particular actions and specific individuals; but how can a jury understand the significance of this case without attending to the larger questions of ethics or policy it raises? And why should philosophers claim that only they can understand and explain those larger issues, when, at least according to Crassus, it is the task of the orator to construct arguments about those issues? We begin by thinking of rhetoric as a self-contained system of rules for training orators, but we are gradually disabused of such a simplistic notion: not only is rhetoric itself an insufficient preparation for public speaking of any kind, but there is a vast range of ideas and subjects that are subsumed in the training of the true orator. That, at least, is Crassus’ view of the matter; on the other hand, it is important to recognize that, just as Antonius and Crassus take rather different approaches to the enlargement of rhetoric and the employment of theseis, so too they differ about what is to be expected of the orator. This chapter began with the contrast between Crassus’ definition of the orator in his opening speech and the far more limited role assigned to him by Scaevola and Antonius. And, having looked at the different ways Crassus and Antonius reconstruct rhetoric itself, we can now bring into clearer focus the orator himself, the real subject of De oratore.

Crassus ended his encomium of oratory by describing the perfectus orator (1.34):

It is my settled belief that the moderation and wisdom of the complete orator (perfecti oratoris) preserve not only his own honor but the well-being of a great many individual citizens and of the commonwealth as a whole.

The orator’s civic role is paramount here: he speaks in public on behalf of others and on behalf of the res publica—as well as being able to preserve his own dignitas. In responding to Crassus, Scaevola (1.35) emphasizes less these civic functions than the characteristics pertaining to sociability: “that even apart from the forum, the rostrum, the courts, and the senate you stated that the orator is complete (oratorem . . . esse perfectum) in every type of humane speech.” Scaevola’s version of Crassus’ orator perfectus is defined in terms of his ability to use language in any social context (omni genere sermonis) as well as in the standard venues (the civil courts, a public audience, the criminal quaestiones, and the senate) for public speech.³⁶

In answering Scaevola in turn, Crassus uses a slightly more elaborate version of the phrase and offers quite a different definition: “the fully complete orator (oratorem plenum atque perfectum) is a person who can speak resourcefully and inventively on every subject” (1.59). This description says nothing about the orator’s civic function emphasized earlier; here, he displays a combination of rhetorical skill (copiose varieque dicere) and universal knowledge, a combination repeated a little further on (1.71–73), when Crassus emphasizes the requirement that the perfectus orator must possess both oratorical ability and a vast range of learning.³⁷ As the dialogue progresses, Crassus occasionally draws attention to certain specific requirements for the perfectus orator: he must know the civil law (1.197); he must combine either Peripatetic or Academic skill at disputation with the experience and technique of someone trained in rhetoric (3.80);³⁸ and he must have knowledge of philosophy (3.145).

Although the exact wording and the exact requirements vary somewhat, Crassus has a fairly consistent image of the ideal orator whose shaping is the object of the entire discussion. When, after much prodding, he reluctantly begins to describe the training and attributes of this true orator, he starts with natura (1.113): there is no point in training anyone who does not naturally have the right physique and talent for oratory. Some orators, like the revolting Varius (1.117), may be talented enough to overcome natural disabilities, but that is irrelevant to the construction of an ideal (1.118):

We’re seeking the orator, and so we must shape in our conversation (fingendus est nobis oratione nostra) an orator with all faults removed and praiseworthy in every respect (omni laude cumulatus).

This seems to be Cicero’s own opinion as well, since Crassus’ words here echo Cicero’s in the preface (1.20):

My own opinion is that nobody will be an orator praiseworthy in every respect (omni laude cumulatus) if he has not achieved knowledge of all the great subjects and areas of knowledge.

The repetitions of omnis in both definitions are emphatic: the true orator must deserve praise in all respects, he must be free of all flaws, he must be expert in all branches of knowledge.³⁹ Whether or not this is a practicable ideal (to be discussed below), what is most striking is the repeated stress on the content of oratory, not rhetorical training and expertise: what the true orator does not know, in effect, isn’t knowledge.

The perfectus orator belongs to Crassus: of the fourteen occurrences of this or very closely related phrases in De oratore, Crassus speaks ten. But even in Crassus’ usage, perfectus orator can mean two quite different things. On the one hand, it is “the ideal orator”: a speaker who is truly perfect, an unattainable and possibly Platonic model against which the striving orator can measure himself, the complete master of all skills and all knowledge. That is the perfectus orator of Cicero’s claims in 1.20 and Crassus’ in 1.71, a level of achievement that can only be emulated, but not equaled—and one that Cicero did not attempt fully to describe until Orator, written in 46.⁴⁰ On the other hand, it is, as I have generally translated it, “the complete orator”—a speaker with all the skills and knowledge necessary for successful oratory, a goal that is difficult to reach but attainable with ability and effort.⁴¹ This lesser goal of completeness is one that Crassus sometimes seems to accept, as in his initial account of the perfectus orator at 1.34 and later on at 1.129–30, when he speaks of the absolutio and perfectio of his imagined orator as being comparable to the level of attainment of a Roscius in acting, or at 3.71, where Demosthenes is taken as a model for emulation, exemplifying “the glorious and outstanding image and beauty of the complete orator (oratoris perfecti).”

Roscius and Demosthenes are real people; they represent an attainable level of expertise, but at times Crassus is perfectly well aware of the unreality of the ideal perfectus orator. Thus at 1.71, he speaks of him as eum quem fingo, and at 1.118 he programmatically demands that the object of their discussion be an ideal, not a flawed reality, again using fingendus; at 2.123 Antonius echoes the same language, speaking of hunc oratorem quem nunc fingo.⁴² The use of fingere is significant and is, as Ingo Gildenhard has shown, the equivalent of Plato’s use of πλάττειν for the construction of his imaginary Kallipolis; the comparison is apt, as the search for the perfectus orator in De oratore is just about as realistic as the construction of the Kallipolis of the Republic.⁴³ In the world of De oratore, as will be seen later, comparison to Plato is scarcely a compliment, and even if Crassus appears to endorse such an ideal, the dialogue as a whole does not: we are not looking for an oratory of the imagination, but (as even Crassus seems to recognize from time to time, as in 1.34) an oratory that works in Rome.⁴⁴

If even Crassus at times seems to waver between “ideal” and “complete” versions of the perfectus orator, it is worth looking at how others use the phrase. Cicero himself never uses it in De oratore in his own voice; Scaevola uses it once (1.35) in criticism of Crassus’ ideal; and Antonius uses it three times. Antonius’ comments on the problem of perfection allow us to see that (as with the setting of the dialogue itself) while ideal visions are wonderful things, they are (more or less by definition) unreal. In his first use of the phrase, Antonius describes such a figure as extraordinarily rare in comparison to other kinds of expert because of the many difficult demands placed on him (1.128):

But in an orator we have to look for the sharpness of dialecticians, the pithy sayings of philosophers, the vocabulary (more or less) of poets, the memory of lawyers, the voice of tragedians, the physical gestures of almost the best of actors. That’s why there is nothing less common in the human race than the perfectus orator.

Antonius here echoes, somewhat sarcastically, Cicero’s own explanation in the preface (1.16–17) of why there are so few great orators: the list of demands is a combination of Crassus’ requirements for immense learning with the regular officia described by the rhetoricians. Antonius is also fleshing out the aphorism he earlier cited from his own libellus about oratory, that he had seen many diserti, but no eloquentes.⁴⁵

At the end of Book 1, although he does not bother repeating the phrase, Antonius attacks Crassus’ exacting view of the orator, substituting the much simpler, if still demanding, picture of a person who has mastered the techniques of rhetoric and knows enough to make use of other kinds of expert and expertise, someone who functions in a Roman setting and who is judged by results, not by an impossible abstract standard of knowledge.⁴⁶ At the beginning of Book 2, however, as Antonius begins his long account of inventio, he seems to reverse his opinion and embrace the ideal. Antonius himself says as much, and offers an explanation of his change of opinion: in Book 1, he says, he opposed Crassus in the hope of luring his disciples Cotta and Sulpicius away from him; but now, in the presence of the seriously learned brothers Catulus and Caesar Strabo, he will say what he really thinks (2.40), and he is in essential agreement with Crassus.

That, at any rate, is how Antonius’ self-presentation at the opening of Book 2 is generally interpreted: that his rejection of the elevated portrait of the orator in Book 1 was merely for the sake of argument, while his praise of Crassus’ perfectus orator in Book 2 reflects his genuine belief—or at the very least shows that he has decided for present purposes to endorse Crassus’ ideal. Some caution, however, is necessary, not only because (as has been argued in Chapter 3 and will be again in Chapters 5 and 6) Antonius and Crassus are presented as very different types of orator throughout the dialogue, and not only because Antonius, like Laelius in De re publica, is consistently presented as un-idealistic, practical, and somewhat cynical, but also because what he says here about the perfectus orator is remarkably different from Crassus’ depiction.⁴⁷

Antonius begins his presentation with a mock proclamation (2.28–29):

“Hear ye, hear ye!” he said. “You are about to hear a man straight from the lecture room, educated by a teacher and in Greek literature. . . . But since this whole thing, whatever it is, whether it’s an artificial creation or simply the desire to speak, can be nothing at all without a big mouth (os),⁴⁸ I shall teach you, o my pupils, what I have not learned myself, namely what my opinion is about every sort of speech.”

Laughter follows, and what follows the laughter is Antonius’ demonstration that oratory cannot be an ars, because an ars concerns things that are known, while oratory is just the opposite (30): “For we speak before those who are ignorant and we speak things about which we are ourselves ignorant.”⁴⁹ Antonius’ swift and blunt rejection of the idea that oratory is an ars is, of course, completely undercut when he eventually begins his detailed and scholarly discussion of inventio. It is an opening that must be read as pure irony, on Cicero’s part as well as on Antonius’: the preface to Book 2, only a few pages earlier, had attempted to refute the standard opinion that Antonius had no formal education in rhetoric or Greek culture by arguing that he suppressed that knowledge in public because he wished to be thought uneducated in such matters. Here Antonius reflects both opinions that Cicero offers: on the one hand he claims to be thoroughly educated; on the other he claims to have had no education at all. The internal audience recognizes this as a joke; the external audience is supposed to recognize that the fictional Antonius is playing off the reputation of the real Antonius.⁵⁰

From that, and after some discussion of whether it makes any difference whether rhetoric is an ars or not (2.31–32), Antonius turns to a highly rhetorical encomium of Crassus’ perfectus orator the opening of which deserves to be quoted at some length (33–34):

But more about me later. Now I’m setting out something of which I have convinced myself, namely that even if it isn’t an ars, still there is nothing grander than a perfectus orator. I’m not talking about the practical use of speech, which rules over every peaceful and free state (in omni pacata et libera civitate); there is such delight in the very capacity of speech, that the human ear and mind can perceive nothing more pleasant. [34] What song (cantus) can be found that is sweeter than a well-tuned speech? What poem (carmen) is neater in the artful rounding-off of its words? What actor is more pleasant in his imitation of truth than the orator in its presentation? What is more subtle than a crowd of pithy epigrams? What is more admirable than material illuminated by the gleam of words? What is more replete than a speech stuffed with every sort of material?

This whole passage responds to Crassus’ own encomium of oratory: Crassus had praised the importance of oratory in every pacata civitas (1.30), and Antonius adopts the phrase, but immediately dismisses that aspect of the subject. The perfection of Antonius’ perfectus orator has nothing to do with the social value and context of oratory; he is concerned purely with its aesthetic aspects. Six rhetorical questions follow, of approximately the same length: the first three devoted to the pleasures of the ear, the second three to the pleasures of the mind; in the first half there are comparisons with the performances of singers, poets, and actors, and in the second half Antonius speaks about sententiae, verba, and res.⁵¹ It is worth setting out the Latin text of the passage quoted more fully above (34):

Qui enim cantus moderata oratione dulcior inveniri potest?

Quod carmen artificiosa verborum conclusione aptius?

Qui actor imitanda quam orator suscipienda veritate iucundior?

Quid autem subtilius quam crebrae acutaeque sententiae?

Quid admirabilius quam res splendore illustrata verborum?

Quid plenius quam omni genere rerum cumulata oratio?

The whole set displays impressive stylistic control and the kind of virtuoso skill at ornatus that a good orator should possess: six different comparative adjectives, with the first three coming late in their clauses and the last three at the beginning, and in the whole set of admirable qualities of oratory only three words are repeated (other than quid): res, verba, and oratio. There are some peculiarities, however: in the first pair of questions, for instance, cantus must mean song, while carmen can refer to any rhythmical utterance ranging from song to poem to legal formula. But even if cantus and carmen do not refer to the same form of utterance,⁵² the sentences in which they appear seem to say much the same thing: that speech, like cantus, has rhythm, and that speech, like carmen, has rhythm.⁵³ The middle pair of questions stress artificiality: the orator as performer and the decoration of a speech with frequent sententiae. The third and final pair emphasize the relationship between res and verba, but they do so in a banal and redundant fashion: subject matter has to be expressed in words, and words need a subject to talk about.

This glorification of the perfectus orator has virtually nothing in common with the perfectus orator envisaged by Crassus. Crassus’ orator is a master of all knowledge; Antonius’ orator is a brilliant and precious stylist, who seems to aim at aural effect more than sense, and certainly more than substance. He is repetitive in thought without being repetitive in language, and Antonius’ emphasis on the relationship of res and verba is a trivialization of the argument about res and verba that Crassus makes throughout the dialogue, just as the emphasis on style rather than substance or civic context is a trivialization of everything that Crassus believes about the perfectus orator.

The rest of Antonius’ breathless and verbose praise of oratory continues in this vein: the set of six rhetorical questions quoted above, introduced by various forms of quis and qui, is followed by a set of three statements introduced by huius, eiusdem, and eadem; that in turn is followed by another set of six rhetorical questions or phrases introduced by quis (2.35), followed shortly by another sequence of three clauses, si qua est ars alia . . . aut si quisquam dicitur nisi orator . . . aut si via ulla nisi ab hac una arte traditur (2.36). This last tricolon of si quis offers a hint at what Antonius is doing: in his praise of the glories of early Roman law (1.193), Crassus had offered three reasons in three parallel sentences with anaphora, each beginning sive quem; Antonius here may be parodying what is in fact one of Crassus’ most outrageous arguments, which Antonius had (rightly) ridiculed at the end of Book 1.⁵⁴ Finally, what is perhaps even more indicative of Antonius’ ironic approach to Crassus’ high-flown ideals comes when he actually begins to talk about rhetoric: he abruptly dismisses the beauties emphasized in his encomium and says that instead we now need to put the orator in his proper place, in foro atque in oculis civium (2.41), something his praise of the perfectus orator had completely ignored and which itself ignores the fact that an ideal orator can only be imagined—and not in the forum.

Antonius’ final mention of the perfectus orator is equally dismissive (2.298):

Remember that I’m not talking about some divinity (de . . . divinitate quadam) of a perfectus orator, but about the middling level of my own training and custom.

Divinitas is distancing: the perfectus orator is a wonderful thing in his place, but his place is not in this world: in the real world, successful oratory is the product of hard work, perspiration rather than inspiration. It is no accident that at the end of Antonius’ long discourse, Catulus notes that he had wondered about the source of Antonius’ divinitas in the courtroom (2.362), and now knows that it comes from hard work and study.

De oratore contains more than one version of what an orator is, and that is scarcely surprising: it is, after all, a dialogue. Although May and Wisse translate the title as On the Ideal Orator, it should be remembered that Cicero chose not to put perfectus in the title. On the whole range of subjects connected to oratory, Crassus and Antonius differ considerably—in their organization of rhetoric, in their treatment of the problem of thesis, in their larger understanding of the relationship of both rhetoric and oratory to other disciplines of language, notably philosophy, in their views of what the orator should know and how he should use it, and in what they believe the orator should be and should do. The next chapter will look at some of the same issues as they are framed in somewhat different terms in Book 1, looking in particular at the stories it tells about the encounters between Romans and Greeks and between rhetoric and philosophy.

1. Crassus’ fairly traditional praise of eloquence owes much to Isocrates in particular, although other philosophical sources have been proposed. For thorough analysis and discussion see LPW 1:102–7, and on the broad cultural ideals of oratory which Cicero shares with Isocrates, see Smethurst 1953: 275–80 and LPW 1:64–65. Antonius’ similar encomium (2.33–38), has a very different purpose; see further below in this chapter. On the opening speeches of Crassus and Scaevola, see also Barwick 1963: 17–20.

2. LPW 1:131–35 provide a convenient collection of ten definitions of oratory in De oratore and give a concise history (with bibliography) of such definitions, of which there are a great many; the most important ancient collection is found in Quintilian 2.15.

3. On the standard elements of rhetorical training and technique, see further below in this chapter.

4. Note also that Cicero himself had referred to the orator’s lepos quidam facetiaeque (1.17); as discussed below, Scaevola’s response to Crassus melds what Crassus had said about the orator and what Cicero himself said in the preface.

5. Antonius later (1.260) quotes this definition as if it were Crassus’ own; see below, Chapter 6.

6. On Hermagoras, see further below in this chapter.

7. De oratore 3.93–94; the edict is quoted by Suetonius DGR 25.2 and Gellius, NA 15.11.2 and it is also mentioned by Tacitus, Dial. 35.1. For a succinct and clear discussion, see Kaster 1995: 273–75. Gruen 1990: 179–91 is far too trusting of the evidence of De oratore, and there is not enough evidence to be dogmatic about the intention of the edict.

8. Cicero and Rhetor overlap considerably and clearly share a common Latin source, either a written text or the classroom of a particular teacher. The reference at Inv. 1.80 and Ad Her. 2.33 to the trial of Fulvius Flaccus in 113 gives a terminus post quem for the shared source (whether written or oral) of the two works; see above, Chapter 3 n. 36. Corbeill 2002 is a good introduction to De inventione and Ad Herennium; I follow his dating here. See also Kennedy 1972: 103–39. Putting De inventione in 92/91 may assume too great a degree of precocity in Cicero: it is a juvenile work, but perhaps not the work of a fifteen-year-old.

9. For Hermagoras’ definition of the ideal orator (or rhetorician), see above. Many details about the reconstruction of Hermagoras’ work are uncertain, and little is known about the wider range of rhetorical studies in the second century. When I speak of Hermagorean rhetoric, I mean the general structure of rhetorical theory associated with him, but it should be recognized that many questions about his work remain. The fundamental discussion of Hermagoras is Matthes 1958, and the standard edition of his fragments is Matthes 1962. Woerther 2012 adds useful discussion and bibliography but does not significantly alter the picture.

10. The four basic Hermagorean staseis are (to use the standard Latin names) coniectura (did he do it?); definitio (he did it, but it’s not what you think it is); qualitas (there are mitigating circumstances); and translatio (this is not the proper forum for this case). Only the first three are relevant to Roman forensic conditions; they are described but not named by Crassus at De or. 1.139 as (1) factumne sit; (2) quale sit aut etiam quo nomine uocetur; and (3) rectene factum esse uideatur. The four staseis are paralleled by the four quaestiones legales that concern the interpretation of documents; the only one that matters in De oratore is scriptum et sententia/voluntas (literal meaning and intention), which Crassus mentions in his summary of rhetoric and which is central to the causa Curiana. On the theory of stasis see the clear summary of Kennedy 1994: 97–101 and the full treatment of Calboli Montefusco 1986; on its place in De oratore, see LPW 3: 26–32. For a worthwhile treatment of the Aristotelian origins of Hermagorean stasis, see Marsh 2012.

11. The five traditional divisions of rhetorical instruction are sometimes referred to as partes and sometimes as officia. I use the latter term to avoid confusion with the partes (prooemium, etc.) of the speech itself.

12. For a lucid (and still valuable) account of the various rhetorical traditions known to Cicero, see Solmsen 1968: 2.178–215 (originally published 1941). For a brief and clear introduction to Cicero’s use of Aristotelian and Peripatetic rhetorical theory in De oratore, see LPW 1: 61–64.

13. In the last paragraph (2.178) of Inv. Cicero announces his intention of dealing with the rest of rhetoric in further books.

14. For rhetoric as pejorative, see also 1.52, 55, 86, 87, 133; 2.58; 3.75, 122. LPW rightly delete rhetoricum at 3.80; see their notes ad loc. and on 1.187 ratione dicendi. A parallel case of the derogatory use of a Greek title in Latin is that of philosophus; see Hine 2016. For the various criticisms of standard rhetoric in De or. see also Barwick 1963: 71–73.

15. Similarly Antonius 2.32–33.

16. On Crassus’ summary of the ars rhetorica, see LPW 1: 231–33 and Merklin 1987: 151–53.

17. See Merklin 1987: 153–55.

18. There is a clear and succinct summary of Antonius’ version of inventio and its difference from standard (Hermagorean) treatments at LPW 3: 20–24; see also Merklin 1987: 155–58. It should be noted that Cicero avoids the word inventio, just as he tries to avoid all the technical terms of rhetoric.

19. See the useful summary of LPW 3: 102–3.

20. Dillon 2016 provides a succinct and sensible discussion of the evidence as to whether Cicero knew Aristotle’s work directly and which works he knew. There are significant differences (above all, in the meaning of ethos) between Antonius and Aristotle, but that does not mean either that Cicero read Aristotle and misunderstood him or that he read someone else who had modified Aristotle for dumb Romans. What is most important is that Antonius is made to emphasize his use of the philosophical topica derived from Aristotle rather than the rhetorical topica as found, for instance, in Ad Herennium. On the general Aristotelian background to Antonius’ discussion of inventio, see Solmsen 1968: 2.197–98; Wisse 1989; Fantham 2004: 164–77. On the importance of ethos and pathos in Antonius’ account and its relationship to Aristotle’s Rhetoric, see (in addition to Wisse 1989) LPW 3: 125–33. On thesis and topica, see further below.

21. Caesar Strabo’s speech has been discussed a great deal, particularly as to its sources; see Rabbie’s introduction in LPW 3: 172–212, and for a rather different approach, Dugan 2005: 125–33. In introducing his account of humor (2.235) Strabo emphasizes the analytic quality of his ars with all its subdivisions and categories. That is almost certainly a parody, and the whole ars of jokes seems to be Cicero’s own construction, probably making some use of (lost) Greek treatises on humor but largely his own; it is itself a joke on the dry and boring analytic tradition of rhetoric.

22. Although there are differences: see LPW 4: 13–23 for detailed analysis.

23. On Crassus’ speech on elocutio, see further below, Chapter 7.

24. Indeed, the whole ars of Books 2 and 3 adheres more closely to the Peripatetic organization of rhetoric than any other extant text; cf. Solmsen 1968: 2.192.

25. The related problems of thesis and the topica (loci communes) are complex, as is the relationship of both to the doctrine of status in the composition of forensic speeches. Thesis and topica are (as will be discussed below) addressed more than once in De oratore, by both Antonius and Crassus. Others have discussed these aspects of rhetorical theory in detail (and I am probably oversimplifying); see, for instance, Calboli Montefusco 1991; Reinhardt 2003: 3–35, 346–50. Michel 2003: 201–34 is a broad treatment, rightly emphasizing the differences between Crassus’ and Antonius’ approaches. For further discussion (focusing on Crassus’ discussion of the issues at 3.107–19) see below.

26. Cicero criticizes Hermagoras for precisely this at Inv. 1.7–8.

27. Note that the young Cicero himself (Inv. 1.8) excluded such topics from rhetorical education: “quae sit mundi forma?” “quae sit solis magnitudo?” quas quaestiones procul ab oratoris officio remotas facile omnes intellegere existimamus.

28. By the time Antonius talks about this problem in Book 2 it has already been discussed in less technical terms several times in Book 1, when the Roman speakers report their meetings with Greek philosophers and rhetoricians. See below, Chapter 5.

29. See n. 10 above.

30. Cf. LPW 3: 76 on the relationship between this passage and the earlier treatment at 2.65.

31. There was a theory of rhetoric that opposed Hermagoras’ hierarchy of thesis and hupothesis; it is associated with his younger contemporary Athenaeus, who is said to have spoken of general topics as pars causae (Quintilian 3.5.5), in other words, inverting causa and quaestio, in the way that Antonius here seems to do. On this argument and on pars causae, see Barwick 1963: 61–63; Athenaeus’ role is questioned by LPW 5: 46.

32. For a fuller discussion of Crassus’ speech on elocutio, see below, Chapter 7. Crassus’ argument in this passage is extraordinarily difficult; it is marred by a textual problem at the beginning of 3.110, where rhetores or something similar must be supplied for the argument to make any sense at all (on this, see also Barwick 1963: 51), and Crassus’ anatomy of the varieties of argument in 3.111–18 would be better represented by a chart (as is helpfully supplied by both Brittain 2001: 343 and LPW 5: 39). I am more than usually indebted to LPW for my understanding of this passage; see particularly Wisse’s explanation of Crassus’ argument and of thesis in LPW 5: 38–57, with bibliography. See also n. 25 above.

33. For Philo’s rhetoric, see Brittain 2001: 296–342, relying too heavily on De oratore for reconstructing the rhetorical theories of Charmadas and Philo. For criticism, see LPW 5: 52–53, 74 (on 3.110).

34. On the unusual phrase politici philosophi, see LPW ad loc.

35. A set of “waves,” to use LPW’s Platonic metaphor for the way in which Cicero sets out the stages of complex arguments. Wave theory is most clearly applicable to the presentation of the orator’s need for philosophical knowledge in Book 3; see LPW 5: 1 for a clear summary.

36. Scaevola is drawing on Crassus’ discussion of sermo at 1.32 and changes the meaning of humanitas (from the contrast between humans and animals to the elegant sociability of gentlemen); see LPW ad loc. Barwick 1963: 28–31 takes Crassus’ mention of sermo to reflect a strict theoretical division between contentio (a word which, as Barwick notes, is not used here), referring to senatorial and forensic oratory, and sermo, referring to all other kinds of speech, and traces it back to Hellenistic divisions between thesis and hupothesis. Crassus, like Cicero, is not that rigid.

37. A list of the necessary studies for the ideal orator is found in Barwick 1963: 10–13.

38. There are textual problems in this sentence, but it is clear that the two things that Crassus demands are philosophy and rhetoric, in some fashion.

39. LPW ad loc. describe this as the maximalist definition (restated in slightly different language at 1.64).

40. On the Platonic ideal of the orator, cf. Or. 1–19, with the commentary of Kaster 2020. Like many scholars, LPW believe (cf. 4: 88–91 for a summary of their position) that Cicero in De oratore has a single and unified concept of the perfectus orator, different aspects of which are emphasized in different passages (and to some extent by different speakers). I view the different articulations of what the orator is, particularly the differences between Crassus and Antonius, as genuine disagreements, and I take Cicero’s Academic skepticism as fundamental to the dialogue: he presents more than one view, and however much he may wish for the existence of the ideal, he recognizes its practical unattainability. See further below, Chapter 5.

41. This use of “complete” is classified as archaic by OED, §5 (“Of persons: Fully equipped or endowed; perfect, accomplished, consummate; esp. in reference to a particular art or pursuit, as a complete actor, horseman, merchant”) but it is less burdened by metaphysics than “ideal,” on which see below. On the various phrases C. uses to describe his ideal orator, and on possible Greek antecedents, see Barwick 1963: 7–8.

42. Of these passages, the most important is clearly 1.118, quoted earlier. On Cicero’s use of fingere, see also Dugan 2005: 79 n. 11.

43. LPW 1: 219 (on 1.118) attempt to distinguish fingere as used for Crassus’ portrait of the orator from fingere as used to describe Plato’s Kallipolis, but see Gildenhard 2013a: 231–37 on the Platonic idealism of Crassus’ vision of the orator and its contrast with Antonius’ quite different understanding.

44. For a somewhat different approach to the same issue, see also Fox 2007: 128–31.

45. On this, see further below, Chapter 5.

46. On the paired speeches of Crassus and Antonius at the end of Book 1, see below, Chapter 6.

47. LPW 2:221–22 refer to Antonius’ “conversion” (Bekehrung) and minimize the differences both between Antonius’ and Crassus’ conceptions of the perfectus orator and between Antonius’ encomiastic prologue and the substance of his discussion of inventio.

48. On the double meaning of os here (literal mouth or brazenness), see LPW ad loc.

49. On the long-debated question of whether rhetoric is an ars, see LPW 1: 190–94 and particularly Barnes 1986; see further below, Chapter 5.

50. LPW 2:222 rightly recognize that Antonius’ reversal would not be possible without the preceding preface to Book 2.

51. So LPW on 2.34. The six questions comprise respectively 23, 20, 24, 18, 19, and 20 syllables; that is not isocolon, but it comes close.

52. As they do, for instance, at Mil. 80.

53. It is perhaps indicative that LPW note on the first question “Moderatus bezieht sich auf den Rhythmus (numerus)” and on the second “Hiermit wird die rhythmische Periode bezeichnet.” They are not identical, but they are remarkably close, and the distinction (between clausular rhythm and periodic balance) has no connection at all to the difference, such as it is, between cantus and carmen.

54. On the speeches of Crassus and Antonius at the end of Book 1, see below, Chapter 6.