THE INCONSISTENCY CHARGE IN CICERO'S DE FINIBUS 1–2

Abstract

Cicero often challenged Epicureanism on the grounds of inconsistency. Cicero personifies the charge through his character Torquatus, who defends Epicureanism in De finibus 1–2. Cicero highlights the discrepancies among Torquatus’ beliefs and between them and his behaviour. Torquatus holds that the senses incontestably verify the tenets of Epicureanism, and that logic is superfluous. Yet he is sensitive to the fact that Epicurus’ teachings are not intuitive and require a fair amount of logical argumentation in its defence. Therefore, he defends his school against Cicero's criticisms. But by engaging in a defence of the system, Torquatus has already spoken against his commitment to the obviousness of Epicureanism and his disavowal of logic.

Cicero's presentation of Epicureanism is polemical.¹ This fact is especially striking given that his Academic commitments generally motivate a more even-handed presentation of opposing systems.² The Tusculan Disputations suggest various motives for this hostility: a commitment to hedonism is a temptation to behave poorly (Tusc. 2.16), Epicurus’ wise man acts ludicrously (Tusc. 2.17) and the system appears inconsistent (Tusc. 5.31).³ In De finibus 1–2, the last charge receives special attention.⁴ Torquatus, the dialogue's defender of Epicureanism, displays a sceptical attitude towards logic and complete trust in sense perception. The two go hand in hand—there is little need to debate the phenomena if they are so readily observable. Consistent with this position, he insists on expressing his beliefs by oratio perpetua, and not in a cross-examination.⁵ One should not complicate one's perceptions by captious debate when one can just express what one observes. Given these commitments, Cicero has no need to cross-examine Torquatus. He only needs to show that the phenomena are not obvious or easy to interpret.⁶ That demonstrated, Cicero would seriously undermine popular Epicurean epistemological claims. It would, on the one hand, (1) justify the need for logic to adjudicate complications in perception, and (2) call into question Epicurean hedonism since it, like their epistemology, is justified by the obvious use of sense perception. In other words, if Cicero shows that complicated sensory data require some sort of arbitration, a blow is already struck against Epicureanism. In this paper, I will comment on various moments in De finibus 1–2 that speak to these points, with a special emphasis on Cicero's initial arguments against Epicureanism (1.17–26) and Torquatus’ response to them (1.29–42).

The Epicureans were not always clear about the relationship between thought and speech, which is perhaps why Cicero marks both at the beginning of the dialogue. Torquatus accuses Cicero of disdaining Epicurus because he does not employ the ‘ornaments of oratory’ of Plato, Aristotle and Theophrastus. That is the only possible explanation for why ‘the things that [Epicurus] perceived (senserit) do not seem true to you’ (1.14). The charge cleverly takes for granted the Epicurean worldview. Pleasure is the best guide of human behaviour, and Torquatus assumes that Cicero's estimation of Epicurus’ philosophy relates to the (lack of) pleasure he derives from his writings. The charge is also plausible because Cicero elsewhere seems to suggest that one's speech is a general measure of the quality of one's thought (Tusc. 5.47).⁷ But exceptions occur, and in the prologue (1.4–5) Cicero cites one.

cum Sophocles uel optime scripserit Electram, tamen male conuersam Atilii mihi legendam putem, de quo Lucilius: ‘ferreum scriptorem’, uerum, opinor, scriptorem tamen, ut legendus sit. rudem enim esse omnino in nostris poetis aut inertissimae segnitiae est aut fastidii delicatissimi.

I think that people should read Atilius’ badly translated version of the Electra, even if Sophocles’ version was best. Licinius said that Atilius was an ‘iron author’, but still, he was an author, and as such people should read him. Not to know our own poets at all is a sign of either extreme laziness or extreme pickiness.

However much Cicero may value style, he makes it clear that he can look past it (si non habeat, non admodum flagitem, 1.15). Since Cicero is willing to read and even praise Atilius, he might extend the same favour to Epicurus.⁸

Yet in Tusculan Disputations 2.7, Cicero claims that he does not read the works of philosophers who write poorly, by which he must mean the Epicureans. How can Cicero defend the rough Latin poets, but refuse to read the writings of the Epicureans? He seems to offer a clue further down in the prologue of De finibus 1: some people look down on Latin letters ‘because they come upon certain unadorned and uncouth productions, translated from bad Greek into worse Latin’ (1.8). Cicero agrees with these critics (ego assentior). We can suppose that for Cicero the quality of the translations of such books and the fact that ‘no one moderately educated is ignorant of what those from that school say and feel (sentiant)’ (Tusc. 2.7) may exempt one from having to read them.⁹ Epicureanism is not difficult to learn (Fin. 1.27), and so there is no need to wade through their dreadful prose to understand the system. The result is that Cicero does not reject Epicureanism because of aesthetic dissonance while reading its texts, as Torquatus may like to believe.

We can now return to Torquatus’ accusation in 1.14. He claims that Epicurus ‘sensed’ (senserit) the teachings of his own system. He cannot believe that Epicurus’ perceptions ‘do not seem true’ (uideantur) to Cicero. How does Cicero fail to perceive what Epicurus has perceived? And for that matter, which perceptions does Torquatus think Cicero ought to recognize as true? We have two clues. The first is that the theory which Epicureans thought was most obvious was the avoidance of pain and the seeking of pleasure (1.71), something available even to infants and animals. The second is that Torquatus had prefaced his question to Cicero with the claim that Epicurus uniquely taught people how to live ‘well and happily’, framing the question within their ethical doctrine. So, the source of Torquatus’ confusion (even if it is just rhetorical) is how Cicero does not agree with Epicurean hedonism—if indeed he actually disagrees with Epicurus (something of which Torquatus can ‘scarcely be persuaded’) and is not merely confused. He presents the argument effectively: by conceding that his master writes poorly, he makes it appear that Cicero is receiving conflicting sensations of (1) the truth of hedonism and (2) the harshness of Epicurus’ style, and has fallen into error by exaggerating the importance of the second perception. Epicureans were aware of such possibilities. Indeed, the ‘cradle argument’ seems to present the opposite scenario, in so far as adults’ ‘hedonistic intuition [is] often obscured by an artificially imposed value system’.¹⁰ But in both cases, there is some disconnect between pleasure and belief, with a claim of resulting error—a difficulty for the system.¹¹ In any case, Cicero parries Torquatus’ accusation: ‘His speech does not offend me’, he says, on the grounds that Epicurus speaks as he wishes and plainly enough to be understood (1.15). Epicurus may not have the literary skill of Plato, Aristotle and Theophrastus, but there is a difference between rhetorical ornament and clear writing.

Cicero responds to Torquatus with a three-part critique (1.17–25). In the first place, he attacks Epicurean natural science (φυσική). ‘It is not normal for a natural scientist to believe that anything is a minimum, which [Epicurus] never would have thought, if he had preferred to learn geometry from his friend Polyaenus rather than make Polyaenus ignorant of it’ (1.20). Cicero's joke is somewhat obscure. In the Academica (2.106), Cicero explains that Polyaenus had been a gifted mathematician, until Epicurus convinced him of the falsity of geometry. Epicurus posited the existence of minima, an entity beyond which one cannot imagine anything smaller.¹² Although minima may solve certain problems (like Zeno's divisibility paradox), they introduce their own. For example, they make all sides in geometrical figures commensurable, since the minima exclude the possible of irrational numbers. Yet basic geometrical theorems, like the Pythagorean, demonstrate the existence of irrational numbers. So, one must reject geometry to believe in minima, which is apparently how Polyaenus evaded the difficulty.¹³ The second problem with Epicurean natural science is astronomical. Cicero reproaches the school for believing that the sun was only a foot in diameter. This was something that not even Democritus believed, an innovation proposed in order to make the sun ‘as big as it seems’ (quantus uidetur 1.20). No fragment of Epicurus adopts the figure of a single foot, though it certainly was part of the tradition to believe in a small sun.¹⁴ It fits in with their insistence on the truth of sensory perception. The point here, and in the previous example of Epicurean geometry, is not to disprove the existence of minima or the idea of a one-foot sun. Rather, perceiving the absurd consequences of the Epicurean theory of perception is already a sort of refutation.¹⁵

Cicero then addresses the Epicurean rejection of logic (λογική) (1.22). Unfortunately, the section is lacunose, though what we do have is telling, and Cicero thematizes the Epicurean rejection of logic throughout his conversation with Torquatus. Their rejection of logic, and consequently of debate, was perhaps an especially sore point for Cicero. In the prologue to the second book of Tusculan Disputations, Cicero attributes the success of Greek philosophy to ‘contentions and disputations’, even as he praises the members of his own school for their willingness to ‘confute others … and be confuted without resentment’ (Tusc. 2.4–5). In contrast, Epicureanism had no place for the elements of logic, such as definitions, division, reasoning (ratio), resolving fallacies and ambiguities, because ‘Epicurus placed the judgement of all things in the senses’ (1.22). Epicureans did not appeal to logic to resolve conflicting data, whether sensory or linguistic, because (1) they did not believe sensory data could err and (2) they believed that, ideally, the meaning of words should be so obvious as to make debate unnecessary. As for (1), Epicureans would argue that the only way to adjudicate sensory data is to acquire more of them. That does not mean using new data to refute prior data, but allowing them to stand next to each other to fill out the picture. Accumulated sensations confirm or refute our opinions and eliminate what is obscure.¹⁶ Now, (2) is a corollary to this idea. Perception provides one with an obvious understanding of how things are, but one also ought to describe how things are in an obvious way. Otherwise, one's clear vision of reality will be obscured by dialectical hair-splitting (cf. 2.17). The problem here is most people's undeniably obscure use of language. To resolve this, Epicurus might suggest ‘seeing the first meaning of every utterance’, such that each term ‘needs no demonstration’ (Diog. Laert. 10.38). Making the distinction between sensation and language, he hoped that they would ultimately be commensurate.¹⁷ To put it another way, he believed natural science to be such as to compensate for logic, since ‘by this science [φυσική], the force of words and the nature of speech and the reasoning out (ratio) of consequences and contradictions can be perceived’ (1.63, cf. Diog. Laert. 10.31). There is no need to search for definitions, let alone to syllogize, if by mere observation of the material world one acquires knowledge.

The Epicurean claim that their philosophical commitments are so obvious as to be sense-perceptible, and hence that debate would only obscure them, is already a dialectical strategy. We might consider Torquatus’ later claims that: ‘(1) After one's senses having been taken away, nothing remains. (2) It is nature that must judge whether something is in accord with or contrary to nature’ (1.30). This rhetorical first inference is not proven, and in fact is complicated by Torquatus’ concessions in 1.31, treated below. The second claim does not follow from the first, unless we grant that Torquatus is using ‘nature’ to mean ‘the senses’. And the point remains: what if the data are unclear? What if the senses provide contradictory information? Such questions were apparently a source of contention within the school. Torquatus mentions certain Epicureans who thought that unreflective hedonism was insufficient (negent satis esse quid bonum sit aut quid malum sensu iudicari), and so made an application of their ‘mind and reason’ (animo etiam ac ratione) (1.31). As for what this might entail, Torquatus immediately clarifies: ‘an almost natural and inborn notion in our minds, so that we may sense (sentiamus) that one thing is to be sought, and another avoided’. He seems to refer to the Epicurean theory of preconceptions (προλήψεις), by which one has knowledge of a concept that one can use to identify that concept in the world.¹⁸ If so, it is curious that Torquatus claims that only some Epicureans think along these lines, although this may simply reflect the divergence between ‘popular’ and ‘educated’ Epicureanism (cf. 1.25).¹⁹ Epicurus’ less reflective disciples might interpret his declarations about the priority of sense perception in an overly literal manner, while others would know that some medium is necessary for interpreting the world.²⁰

There was yet another group of Epicureans who, in the face of so many attacks from other philosophical schools, thought that Epicureans should ‘not confide excessively in the cause, but that [hedonism] must be argued for and carefully examined and disputed with reasons carefully sought’ (1.31). Torquatus declares his sympathies with this group (ego assentior), and his position may serve to justify his long defence of the system.²¹ Torquatus is theoretically committed to the obviousness of his system, but in practice acknowledges that his philosophy like any other wants argumentation. He does not explain how such a defence is compatible with Epicurean epistemology.²² And in fact, perhaps it is most reasonable to assume that Cicero wants us to understand that it is not.²³ Cicero likely wishes for us to sympathize with Torquatus’ measured, almost sceptical appraisal of his own system (non … nimium nos causae confidere), and recognition of the need to argue. Perhaps the only way to remain a philosophically serious Epicurean is to be ‘better than one's principles’ with regard to trust in sense perception.²⁴ For Cicero the author, this is quite a refutation. We are to understand that it is easier to argue for the system if one leaves behind the points that become difficult to explain—but such a stance does not speak highly of the original system.

Cicero then moves on to ethics, in which the Epicureans made the most investment. His first argument is historical, more rhetorical than demonstrative. ‘Nature has made us for greater things [than pleasure and pain]’ (1.23) is the claim, which he defends with examples from the lives of Torquatus’ ancestors.²⁵ Surely his family's founder, the original Torquatus, was not pursuing pleasure in fighting the Gaul or executing his son? Or the other Titus Manlius Torquatus, who banished his own son? And who could deny that Regulus was a happy man, though he died by torture? Admittedly, the examples are by no means clinchers.²⁶ That is not to say that they are ineffective. Cicero does not need to demonstrate that Epicureanism is false, but only show that there are plausible arguments against it. Such an outcome would strongly point to logic as a means to resolve the uncertainty—but as we know, Epicureans reject logic. The result is that if neither the truth or falsehood of Epicureanism is evident, then the Garden already stands refuted, since the school claims that its position is evident.

Cicero then considers the popularity of Epicureanism among the masses (1.25):

et quod quaeritur saepe, cur tam multi sint Epicurei, sunt aliae quoque causae, sed multitudinem haec maxime allicit, quod ita putant dici ab illo, recta et honesta quae sint, ea facere ipsa per se laetitiam, id est uoluptatem. homines optimi non intellegunt totam rationem euerti, si ita res se habeat. nam si concederetur, etiamsi ad corpus nihil referatur, ista sua sponte et per se esse iucunda, per se esset et uirtus et cognitio rerum, quod minime ille uult, expetenda.

There are many reasons why so many people are Epicureans (a question often asked), but what most attracts the crowds is that they think that [Epicurus] says that to do right and honourable things is a delight (that is, a pleasure) in itself. The best sort of people do not understand that the whole system is overthrown, if that is the way things are. For if it were conceded that these things are pleasant in themselves (even if nothing were referred to the body), virtue and knowledge of things would be sought for their own sake, which is what he least wants.

In context, Cicero has just considered the fact that many people (like the Torquati of old) undergo pains and labours for the sake of some good, and tells Torquatus that no one (not even Epicurus and Metrodorus) claim that those toils undertaken in the pursuit of virtuous ends could be called pleasant in themselves. To defend the position, Cicero adduces an apparent explanation of why so many people are Epicureans, one which is odd for many reasons. The idea that the multitude became Epicureans after realizing that they could acquire pleasure in their right and honourable deeds is questionable. It is far more believable—and rhetorically punchier—to assume that people become Epicureans to justify the pleasure-seeking they want to or would have engaged in anyway.²⁷ Yet even if we assume Cicero's claim to be true, the inference Cicero draws from it is questionable, namely that if virtue and knowledge are pleasurable, then they would be sought for their own sake. The masses would act virtuously because they want the consequent pleasure, and hence pleasure would still be the end. Torquatus could add that many people eat for the pleasure of the experience, and not merely for nutrition. The argument here is weak.

But then, many of Cicero's arguments in this section have been weak. Torquatus himself has little difficulty in explaining that his ancestors behaved the way they did to avoid greater pain or obtain more pleasure (1.34–5), and that people undertake difficult tasks for some greater advantage (1.32). So, the question arises: why launch such implausible arguments, especially in the area in which Epicureans took the most pride? One possibility is that his arguments are purposely weak. Cicero the narrator may signal as much: ‘I said these things more to draw him out (prouocarem) than for me to speak’ (1.26). In other words, Cicero is inviting Torquatus to engage in discussion, and has made it easy for Torquatus to do so. Cicero goes on to say (1.27):

fieri, inquam, Triari, nullo pacto potest, ut non dicas, quid non probes eius, a quo dissentias. quid enim me prohiberet Epicureum esse, si probarem, quae ille diceret? cum praesertim illa perdiscere ludus esset. quam ob rem dissentientium inter se reprehensiones non sunt uituperandae.

When one disagrees with another, Triarius, it is not possible to keep silent about the point in question. What would stop me from being an Epicurean, if I agreed with what he said? Especially since what he said is quite easy to learn. For that reason, we should not disparage attempts to refute those with whom we disagree.

It is not just that Cicero disagrees with Torquatus. Cicero claims to perceive correctly what Epicurus teaches because it is ‘quite easy’ (ludus) to master. But since they disagree, they should do what ordinary people do: talk it out and resolve the disagreement by argumentation. Torquatus appears to be of the same mind (1.28), once more setting himself apart from orthodox Epicureanism. He does not realize that to respond to Cicero is to engage in the art that Epicurus completely lacked (disserendi artem nullam habuit), as Triarius had just reminded him (1.26).²⁸ So, for Torquatus to work through Cicero's arguments and refute them is already a mark against himself. Yet he cannot easily ignore these arguments, as Triarius has already expressed his sympathy with them. We might take this to be a subtle refutation on Cicero the author's part.

Torquatus chooses to defend himself. Cicero had attacked him for his natural philosophy, logic and ethics. Torquatus demurs on the first front, even as he reasserts his faith in Epicurus’ physics (1.28). He does not formally respond to the criticisms of logic, only offering a brief rejoinder in his defence of hedonism. We might infer that by debating at all he has already bent his school's rules on the matter, even if he does belong to a group of Epicureans who see value in it. Yet at least he will not engage in questions and answers, but only give a speech (oratio perpetua, 1.29). To engage in dialectic may be going too far, as ‘there [is] no need for reasoning or debate (ratione neque disputatione) … these things are felt, like feeling that fire is hot or snow is white, and honey is sweet. There is no need to confirm any of these things with elaborate reasoning, it is enough only to be reminded’ (1.30). This reluctance to engage in questions and answers may refer, if we assume that the Platonic tradition is in the background, to a certain critique.²⁹ Sophists and rhetores preferred to discourse in a certain type of long speech-making, and were often reluctant to engage in the dialectic typical of Socrates and his friends. The latter is riskier, as those who engage in dialectic open themselves up to refutation even as they attempt to refute others (cf. Grg. 458a).³⁰

Torquatus then mentions some methodological points: ‘first of all, I will proceed in the way that would please the very founder of our school: I will establish what it is we're investigating and what it is like. Not that I think you don't know, but so that the speech may proceed in a logical way (ratione et uia)’ (1.29). Cicero the interlocutor says something quite similar later in De finibus, employing the formula [ratio] + [uia]. As the formula is somewhat uncommon, we may feel invited to read the two passages together.³¹ Here is the second in full (2.3–4):

Every speech that involves a certain method and reasoning (quae uia quadam et ratione habetur), must first make a praescriptio (praescribere) when investigating … so that, among those debating, there is agreement as to the topic of debate. This was set down in Plato's Phaedrus, a rule of which Epicurus approved, and he perceived that it ought to happen in every debate. But he did not see the corollary to this commitment. For he says that he dislikes making definitions, without which it is occasionally not possible for those who are debating to come to an agreement as to what is under discussion.

So: (1) Epicurus agreed with Plato's claim that disputants ought to agree as to what is under dispute; (2) they cannot agree about what is under dispute without definitions; (3) Epicurus did not believe in formulating definitions; therefore, (4) Epicurus cannot debate and remain consistent to his principles. The hinge is (2). For an Epicurean to maintain a semblance of consistency, he must explain how interlocutors can speak to each other without hammering out definitions. For Cicero, the only way to proceed seems to be that of establishing a praescriptio. This was ‘an agreed-upon verbal account … which served to restrict the scope of the action being brought before the court’.³² But what is the nature of this account? Apparently, Cicero thought the praescriptio was an apt metaphor for whatever is implied by the ‘rule laid down in the Phaedrus’. In the relevant passage, Socrates says: ‘as if they did know [the essence of each thing], they do not reach an agreement in the beginning of their investigation, and advancing they pay the likely penalty: they do not agree with themselves or each other … [so] as to what love is and what power it has, [let us] set down a definition (ὅρος) by agreement (ὁμολογία)’ (Pl. Phdr. 237c–d). Prior to an investigation, one must reach an agreement (ὁμολογία) as to the definition of what is under discussion.³³ After reaching an agreement, one ‘advances’ by testing claims against the agreement, by the standard of ‘what is probable’. So, even if one does not yet know the definition of something, one can at least agree upon a working definition with which to investigate further. In the present context, the definition to be agreed upon is that of ‘pleasure’, and they must decide whether it is to be defined kinetically or statically (2.6–7).³⁴ One thing is sure, this ‘rule laid down in the Phaedrus’ is inescapably dialectical. Only with great imagination can the Phaedrus passage be reconciled to the Epicurean manner of discovering definitions. Torquatus may arrange his speech in a manner consistent with his ideal of ratione et uia, but ultimately he cannot do so and claim to disdain logic.³⁵

One of Torquatus’ key commitments is establishing that Epicurean teaching mirrors what ordinary people observe. The greatest proof for this commitment is the universal pursuit of pleasure, even from infancy, and even among animals. This was held to be so obvious as to make the system ‘brighter and clearer than the sun’ (1.71). Torquatus had argued for this position earlier in the dialogue with an anecdote (1.39). Once, while he was in Athens, his father pointed out to him a statue of Chrysippus with his hand extended. He told Torquatus that the statue signified a syllogism: ‘“Does your hand, in its current disposition, have any desire?” “No”. “But if pleasure were the good, it would have a desire”. “I believe so”. “Therefore pleasure is not the good”.’ Torquatus considers this a cheap shot. The joke only works, he explains, if pleasure is ‘that which as it were tickles the senses, so to say, and flow to them with delight and flow into them’. But Epicureans prioritize the absence of pain (katastematic pleasure) over feeling an active (kinetic) pleasure.³⁶ So, even if Chrysippus’ hand is not feeling any kinetic pleasure, it is apparently free from pain, and so already participating in the highest good of (katastematic) pleasure.

Cicero would note that most people do not use the word ‘pleasure’ in the specialized Epicurean senses. Yet maintaining the two types of pleasures was important, as it distinguished Epicureans from the Cyrenaics (1.39), whose hedonism extended merely to kinetic pleasures. Epicurus escapes the criticism of the Cyrenaics by using the word ‘pleasure’ in a specialized, scientific sense when need arises (2.16, 39), but was also committed to the position that no erudition was required to understand his teachings.³⁷ All this is quite crafty: they prove the system by leaving the term ‘pleasure’ ambiguous, and distinguish it when criticism arises, shuffling out kinetic pleasure when it becomes inconvenient.³⁸ Of all Cicero's allegations of Epicurean inconsistency, this one seems to be the most difficult.³⁹

After Torquatus finishes his speech, he and Triarius look to Cicero for a response. Cicero asks them not to consider him a philosopher who will deliver a lecture (schola)—he aspires to imitate Socrates, not Gorgias, on the question of long speeches (2.1):

ego autem arbitror, quamquam admodum delectatus sum eius oratione perpetua, tamen commodius, cum in rebus singulis insistas et intellegas quid quisque concedat, quid abnuat, ex rebus concessis concludi quod uelis et ad exitum perueniri. (2.3)

Although I was very much delighted with his long speech, I personally think it is more apt for one to consider each thing individually and understand what each person concedes, what he rejects, and what you wish to conclude from the concessions and to arrive at the conclusion.

Cicero's pleasure in Torquatus’ speech is not necessarily complimentary. In the Gorgias, to which Cicero had alluded not long before, Socrates claims that rhetoric ‘aims at pleasure (ἡδύς) apart from the best. I do not say that it is an art but a knack, because it has no account of the nature of anything to which it applies itself, such that it is unable to provide a cause of each thing’ (465a). Though Cicero may not be consciously referring to the Gorgias in the present passage, his delight in Torquatus’ speech but doubts regarding its logical force still invite us to question whether Torquatus can provide an adequate account of what is under discussion, namely pleasure. And indeed, when Cicero attempts to draw from Torquatus an account of what pleasure is, he balks: ‘Who, I ask, could be ignorant of what pleasure is? Or who could wish for a definition of it so as to understand it better?’ (2.5).

Torquatus objects to ‘dialectical quibbling’ and presses Cicero to offer his response in the form of a speech (2.17). We might wonder: would Cicero concede defeat by playing by Torquatus’ rules? By offering a long oratio, he would seem to reinforce Torquatus’ distrust of dialectic. But for Cicero there is such a thing as philosophical rhetoric, and with that understood he can accommodate Torquatus’ request. This sort of rhetoric is akin to dialectic (2.17). The words that Cicero applies to the latter can just as well apply to philosophical rhetoric: it ‘contains the whole science both of seeing what is in each matter and judging what sort of thing each thing is, and of disputing in a rational way’ (2.18). It is not the form of either rhetoric or dialectic that matters so much as its method. With that understood, for Torquatus to reject Cicero's questions and answers but assent to his ‘philosophical rhetoric’ does not make much sense, as both are predicated on the same principles.

In this speech, Cicero thematizes the inconsistency of the Epicureans. At one moment (tum) they define pleasure kinetically, at another (tum) they define it as katastematic (2.75). A good Epicurean would be willing to die for his friends, even if that seems to contradict Epicureanism (2.80); indeed, the contradiction between life and teaching, when life is nobler, refutes teaching (2.81). Epicurus claims that pleasure cannot be divorced from friendship, but also says beautiful remarks in favour of friendship that do not ‘accord with his reasoning and opinion’ (2.84). He said inconstantissime that while pleasure of a limited duration is sufficient for happiness, pain is worse the longer it lasts (2.87–8). Epicureans praise the behaviour of the virtuous, but cannot account for their apparent disinterest in pleasure (2.67). On his deathbed Epicurus pronounces himself happy (2.97), even though all his senses were in agony. Like Torquatus, Cicero portrays Epicurus as better than his self-contradictory principles: ‘Nothing in this famous letter of Epicurus is congruent and fitting with respect to his teachings … he refutes himself, his writings are refuted by his own goodness and habits’ (2.99). The Epicurean system, despite its pretensions to simplicity and obviousness, for which it rejects the niceties of logic, ultimately does include serious theoretical problems (friendship, katastematic vs kinetic pleasures, the role of virtue, etc.).⁴⁰ It may be possible to defend and explain these things, but the system disavows the logical tools necessary to synthesize the contradictions.

In De finibus Cicero presents Epicureanism as self-contradictory. The extent to which Cicero's Epicureanism corresponds to the historical views of Epicurus is an open question and has received ample treatment elsewhere. At a minimum, I do think that Cicero has presented a vision of Epicureanism that was both common and plausible to his contemporaries, even if an Epicurean could clarify and nuance the debates of De finibus 1–2. But perhaps the more important result of Cicero's work is to indicate problematic questions within Epicureanism. Every system and theory have features that defenders find difficult to explain. The difference is that Epicureans employed a rhetoric of obviousness that makes even the indication of such difficulties especially acute.