Harald Thorsrud. Journal of the History of Philosophy. Volume 40, Issue 1. January 2002.

In an important paper, Couissin argued for what has come to be called the dialectical interpretation of Academic skepticism. On this interpretation, Arcesilaus and Carneades practiced the same, purely dialectical method-they would elicit assent to premises characteristic of their interlocutor’s position and then derive unacceptable consequences exclusively on the basis of those premises. This method allowed them to remain uncommitted to the premises as well as the conclusions of their arguments. They would simply draw logical consequences from their opponents’ own premises for the sake of refutation. This is in keeping with the current consensus that an ancient skeptic is one who has no beliefs, or at least no beliefs of a problematically dogmatic sort. Consequently, the methods of ancient skeptics, including Academic skeptics, are thought to be strictly negative, seeking only the elimination of belief.

So when we come to Cicero’s Philonian version of Academic methodology and find that it explicitly includes provisions for the acquisition of fallible beliefs, it is no surprise that commentators who are sympathetic to the dialectical interpretation tend to view Philo’s position as a degeneration into what Michael Frede has called “dogmatic skepticism” and Gisela Striker, apparently following Couissin, “skeptical Stoicism.” The dogmatic skeptic may assent to some beliefs, not, of course, claiming to know that he knows, but rather with a more modest, fallibilistic attitude. But once we let in such beliefs, according to Frede, there remains only a thin line separating so-called skeptics from dogmatists-“both have views on how things are, one believes that the kind of justification or knowledge which would establish the truth of a view is available and the other that it is not.”

But is this such a thin line after all? Admittedly it is not the gulf separating those who allow themselves no beliefs, the classical skeptics, from those who do, but still it is a crucial distinction. The dogmatist holds his views on “how things are” without ever seriously considering the need to revise them since he thinks of them under the description “beliefs that I know to be true.” He believes that he has more than simply good evidence or plausible arguments, but reasons or proofs that he knows to be conclusive. The Academic (dogmatic skeptic), on the other hand, holds his views on “how things are” with the ever-present prospect of revision since he thinks of them under the description “beliefs that are most likely to be true.” Consequently, the Academic continues to inquire. His target is to determine whether some argument establishes what its proponent claims that it establishes, i.e., whether or not the reasons offered in support are conclusive. The acquisition of fallible beliefs need not hinder this process. Quite the contrary: the aim of acquiring more reasonable beliefs provides a reasonable motivation for engaging in the dialectical practice of refutation, for it is the same skill that enables us to discern the weaknesses as well as the strengths of a position (Luc. 60).

Recently, commentators have objected to the dialectical interpretation or sought to modify it. The central concern here is what to do with Arcesilaus’s and Carneades’ practical criteria, to eulogon (the reasonable) and to pithanon (the plausible) respectively. In both cases, the Academics were apparently responding to the apraxia objection, which claims that suspending judgment on some issue will render one inactive with respect to it (see Plutarch, Adversus Colotes i 12- za-c). For example, if one suspends judgment as to whether or not it is right to pay taxes, she will experience no inclination (the Stoic term is impulse) one way or the other and will end up doing nothing. Not paying taxes in this case should not be considered an action at all. So Arcesilaus and Carneades presented their criteria, at least in part, to show that one may suspend judgment while continuing to live an active life.

But does Arcesilaus himself endorse to eulogon as a criterion? Does Carneades endorse to pithanon? If so, it seems that they were not just revealing inconsistencies, or seeking to relieve their Stoic interlocutors from pretensions to knowledge, but were also offering their own alternative, positive views about how one may live in the absence of knowledge.

Although it is controversial, this interpretation seems to capture the tone of the relevant passages. According to Sextus for example, after Arcesilaus argued against the Stoic criterion of knowledge it was necessary for him to investigate the conduct of life (diexagoge biou) since this requires a criterion (Adversus mathematicos [= M] 7. 158). Sextus does not further explain this “necessity,” but he does say that Arcesilaus provided his criterion, to eulogon, as the key to a happy and wise life, and presumably such a life could be achieved by anyone. Similarly, Sextus reports that after Carneades argued against allof the dogmatic criteria, he too needed a criterion for the conduct of life and the attainment of happiness (M 7.166). Thus it is plausible to attribute these criteria to Arcesilaus and Carneades as positive doctrines that they accepted. On the other hand, if the practical criteria were presented dialectically, why would Arcesilaus and Carneades make such gifts to their adversaries? It would be as if, in Maconi’s words, “Arcesilaus first knocked his opponent to the ground and then gave him a hand up again.” Such generosity would be incompatible with the purely dialectical purpose of refutation.

Further, it is hard to see why the simple possibility of deception should prevent an Academic from accepting a plausible belief. It makes sense for the Pyrrhonist to prefer suspending judgment to risking error insofar as he identifies belief itself as the source of psychological disturbance. But there is no evidence that Arcesilaus or Carneades accepted this ethical diagnosis. It also makes sense for the Stoic sage to withhold belief rather than to risk error since, according to Zeno, the sage will only assent to kataleptic impressions (see n. 30 below), which infallibly guarantee truth. So he will never have to make the choice between withholding belief and risking error as he will never be seriously tempted to believe what is false. But it is clear that neither Arcesilaus nor Carneades believed that the Stoic sage was a real possibility. So, in general, there is no reason to attribute to them the view that it is better to withhold belief than to risk error when faced with these as exclusive alternatives, i.e., that it is always wrong for anyone to believe anything on insufficient, in this case inconclusive, evidence.

This leads, finally, to a serious problem that has not been much discussed. If the dialectical interpretation of Arcesilaus and Carneades is correct, then Cicero seriously misunderstands, or at least misrepresents, his Academic predecessors. Cicero understands the Academic method, as he practices it, as a dialectical inquiry that, by arguing for and against all views, reveals the one that is most likely to be true (Luc. 7). But he also clearly identifies Arcesilaus and Carneades as practitioners of this same method (Tusculan disputations [= Tusc.] 5.11, De natura deorum [= ND] 1.11). In Cicero’s hands, the Academic’s practical criterion is rendered as probabilitas and it is employed both in the conduct of life and in philosophical inquiry (et in agenda vita et in quaerendo ac disserendo, Luc. 32). So if the dialectical interpretation is right, then Cicero has badly misrepresented his predecessors by attributing his Philonian fallibilism to them. His practice would, in that case, be a corruption of the original one.

It is possible that Cicero, under the spell of his teacher Philo, misinterpreted his Academic predecessors, but it is also worth exploring the hypothesis that he did not get them wrong in order to see what kind of a picture emerges. This is especially worth doing since Cicero is our earliest and best informed source for the views of Arcesilaus and Carneades. He is also our most sympathetic source. Since he was entirely uninfluenced by Pyrrhonism, we find none of the polemical attacks on the Academy that are evident in Sextus Empiricus.

A careful consideration of Cicero’s remarks about Carneades and Arcesilaus will reveal, I believe, coherent and interesting portraits that are consistent with Philo’s Academic fallibilism but not with the dialectical interpretation.

I. Arcesilaus

In three separate passages, Cicero identifies the dialectical practice of Arcesilaus with that of Socrates. First, we find that Arcesilaus revived the Socratic method of eliciting opinions from his interlocutors, i.e., those with whom he wished to examine something (disserere); he did this so that he might say what occurred to him in response to their answers. For he “made it a rule that those who wished to hear him should not ask him questions, but state their own views; and when they had done so, he argued against them” (De finibus bonorum et malorum [= Fin.] 2.2). Second, and in a similar vein, Cicero describes the Socratic method (ratio), as revived by Arcesilaus (ND 1.11), as “arguing against, or examining [disserere] all views and openly judging nothing.” Third, at De Oratore [= De Or.] 3.67, we find that Arcesilaus revived the Socratic practice of not stating his own opinion but arguing (disputare) 8 against the opinions put forward by others.

The first of these passages is entirely neutral on the question of whether Arcesilaus held positive philosophical views. The second and third, however, indicate that Arcesilaus did hold positive views. “Openly judging nothing” suggests that Arcesilaus made philosophical judgments, but not openly. If Cicero had meant to say that Arcesilaus simply made no such judgments, he could easily have said so. Next, the phrase “not stating his own opinion” is more naturally read as “not stating the opinions that he holds,” i.e., he has opinions that he does not state, rather than “he does not state his opinions because he does not have any.” Thus it seems likely that Arcesilaus concealed whatever views he did have.

Why did Arcesilaus conceal his views? Cicero never tells us explicitly; however, it is pretty clear why Cicero thought he did. A crucial element in the Academic pursuit of truth is to avoid acquiring beliefs strictly on the basis of authority. Thus, in his authorial voice in the preface to De natura deorum, Cicero admonishes those who wished to know what he himself thought about these things. For, “in disputations it is not so much authority as force of argument that should be sought.” Furthermore, “the authority of those who profess to teach is often a positive hindrance to those who desire to learn; they cease to employ their own judgment and take what they perceive to be the verdict of their chosen master as settling the question” (ND 1.10, see also Luc. 8, 60, Tusc. 5.11, and De divinatione [= Div.] 2.150). Cicero is contemptuous of the appeal to authority made by Pythagoreans and Epicureans. Consequently, it is no surprise that he should think that Arcesilaus would have enough pedagogical sense to conceal his own personal views.

But despite Cicero’s adaptation of the dialogue form, he was less successful, and in some cases even less interested, in concealing his own views. He was similarly less successful in concealing what he took to be the positive views of Arcesilaus. In particular, we find two distinct, though related, philosophical views attributed to Arcesilaus: one regarding (a) the apparent unknowability of things and suspension of judgment (epoche), and the other regarding (b) the infallibility of the sage.

I.I Acad. 44-6

Arcesilaus’s motivation in attacking Zeno was his recognition of the obscurity of things, i.e., that “the senses are limited, the mind feeble, the course of life brief and as Democritus put it, truth is submerged in the depths, everything is governed by opinion and custom, nothing of the truth remains, all things are constantly wrapped in darkness.” As Cicero puts it, Arcesilaus used to believe that all things lie concealed in the depths (omnia latere censebat in occulto, Acad. 45). That this was indeed Arcesilaus’s view and not merely an admission he made for purely dialectical purposes is revealed at De Or 3.67. In a brief discussion of the history of Plato’s Academy, Cicero remarks that Arcesilaus appropriated from the various books of Plato and other Socratic dialogues the view that there is nothing that can be apprehended with certainty either by the senses or the mind. So it is clear enough that Arcesilaus accepted this rather pessimistic conception of our epistemic abilities.

Because of this pervasive obscurity, Arcesilaus held that it is not possible to know anything. Whereas Socrates claimed to know at least that he knew nothing, Arcesilaus would not even allow himself that much. Nevertheless, it is clear that Arcesilaus believed that knowledge is not possible, for he drew practical conclusions on the basis of this claim. What is more, in these passages, Cicero gives no indication that Arcesilaus drew these conclusions exclusively on behalf of his Stoic interlocutors. On the contrary, Cicero presents these Arcesilaan conclusions as perfectly general and applicable to all.

All this skeptical talk of knowledge does not necessarily commit Arcesilaus to the view that truth does not exist, i.e., that some beliefs are not in fact true. According to Cicero, Democritus did deny the very existence of truth, but on this point, speaking as an Academic, he dissents (Luc. 73). As far as he is concerned, there may very well be truth, submerged or concealed in the depths. So although he would admit that it is possible that some of his beliefs are true, he does not claim to have any knowledge. That is, he does not claim to know which of his beliefs, if any, are in fact true. I take it this is what he means by claiming that he does not abolish truth altogether, but discerns some things as true and some as false (Luc. 111). Some things seem true to him (ND 1.12), yet he is keenly aware that for most views the arguments for and against are well matched. Thus he is not convinced that any of the arguments he relies on to support what seems true to him are conclusive.

Indeed, it was most likely this same lack of conviction that led Arcesilaus to his views regarding the possibility of knowledge and the obscurity of things. It would hardly be fitting for a supposedly open-minded inquirer who claimed to be reviving a Socratic form of inquiry to start with the view that knowledge is impossible. The natural order of things is first to go in search of the truth and then later to come to appreciate the difficulties surrounding the acquisition of knowledge. The crucial interpretative issue here is how to understand Arcesilaus’s appreciation of these difficulties, and particularly, what he took to be the practical consequences of our epistemic limitations.

Whenever he was presented with a position, Arcesilaus argued against it, and as a result he led most of those who heard him to adopt his own view about the impossibility of knowledge and the obscurity of things. So, when equally weighty reasons were discovered in favor of each of the opposed views, it was easier to withhold judgment from either side (Acad. 45). Maconi (op. cit. 1988, 234) cites this passage (along with Sextus’s Pyrroneion Hypotyposeon I.232) in support of the claim that the suspension of judgment was Arcesilaus’s real aim in arguing against all views. But even though this was clearly Sextus’s view, it is unlikely that it was Cicero’s, for it would strain to the breaking point Arcesilaus’s connection with Socrates. Yet Cicero insists on this connection in the three passages cited above (Fin. 2.2, ND 1.11, De Or. 3.67) as well as the passage under consideration here. In each case, Cicero identifies Socrates as the inspiration for Arcesilaus’s methodological innovations within the Academy. If we are to give due weight to this debt, we cannot follow Maconi in claiming epoche as the real aim of Arcesilaus’s dialectical method.

How then did epoche figure into the Socratically inspired method of Arcesilaus that Cicero reports? What epoche does, I will argue, is not to prevent us from having beliefs, or even to prevent us from having false beliefs; what it does is to prevent the premature conclusion to inquiry. A conclusion is premature when one comes to believe that he knows what in fact he does not.

First it should be noted that Cicero tells us that it is when equally weighty reasons are discovered on each side of an issue that it would be easier to withhold judgment. He does not indicate that Arcesilaus was committed in advance to reaching this end, nor does he imply that Arcesilaus always did, in fact, find such equally weighty reasons, although he was surely more often successful than not. In order to maintain the Socratic flavor of Arcesilaus’s method, we would do better to assimilate the epoche that Arcesilaus induces in himself and his interlocutors to the state of being in aporia. Many of Plato’s Socratic dialogues end in aporia, not because Socrates sought to reduce his interlocutors to perplexity, but because they were, all together, unable to reveal the truth of the matter under discussion. Aporia, like epoche, is the unfortunate consequence of a failed attempt to discern the truth. It is not an aim that is intentionally sought. Unfortunate as such failures may be, they do provide support, but not conclusive support, for Arcesilaus’s view that knowledge is impossible.

Because of all this, Arcesilaus believed that

no one should openly affirm or positively assert [neque profiteri neque adfirmare] anything, nor should he approve by assenting, and he should restrain and hold back from every slip his rashness [temeritas], which is revealed when he approves of something that is false, or has not been understood [incognita]. For there is nothing more shameful than for assent and approval to rush ahead of knowledge and comprehension. (Acad. 45)

This passage is problematic. It sounds as if Arcesilaus is proposing a general ban on any belief that falls short of certainty. If so then his own views, reported just a few lines earlier, represent a clear violation of this ban. It is important to stress that Arcesilaus seems to be offering practical advice: we must restrain our rashness in matters of belief. But what is it that characterizes this epistemic rashness? If it is rash merely to have any philosophical beliefs, then Arcesilaus would be a poor student of his own teaching.

Alternatively, we may fit this passage into the Socratic picture I have been developing. Now Arcesilaus cannot claim to know when he, or anyone else, has approved of something false. If he were able to do this he would, presumably, know when he had approved of something true, and there would then be no reason to deny the possibility of knowledge. What Arcesilaus can do, and in a thoroughly Socratic manner, is to reveal when someone has approved of something that she is not able to adequately defend because she has mistakenly taken her argument to be conclusive. And this would indeed be rash and shameful if it resulted in the belief that she knows what in fact she does not. So the dogmatic rashness that Arcesilaus reveals occurs when we approve of something that has not been adequately, i.e., conclusively, established as ifit had been so established.

This is how Cicero seems to understand the rashness (teme7itas) of assent. In his introduction to De divinatione, he remarks that because of the exhaustive arguments of Carneades, he is inquiring into the proper judgment to make regarding divination. That is, in his typical Academic fashion, he is setting out to discover the most probable view. And he undertakes the project by examining the arguments for and against the reliability of reading divine signs. “For in all matters, rashness [temeritas] and error in assenting is shameful” and he is afraid that he might assent rashly either to a falsehood or something that has not been sufficiently understood (non satis cognita, Div. 1.7). He is not proposing that it is shameful to believe something that turns out false-again, this is the inescapable risk that every non-sage faces in believing anything. There is no guarantee that Cicero’s examination of the arguments for and against the reliability of divination will lead him to the truth. But after his careful consideration of these arguments his tentative approval that one of the views is most likely to be true is no longer rash. So we should not read satin cognita in this passage as a characteristic of Stoic episteme. In other words, Cicero is not saying that he is afraid that he might assent to something that fails to meet the Stoic criterion for knowledge, but rather something that has not been sufficiently understood in a much less rigorous sense. Similarly, the rashness that is revealed in the Acad. 45 passage cited above need not be the result of assenting to something that fails to meet the Stoic criterion. The incognita in that passage may serve the same purpose as [non] satin cognita in Div. 1.7: a standard of inspection or argument, less rigorous than the Stoics,’ that enables the Academic to avoid assenting rashly.

In the spirit of this less rigorous criterion, Cicero remarks, “just as I judge it to be most honorable to discern the truth, so it is most shameful to approve of falsehoods as truths” (pro veris probare falsa turpissimum est, Luc. 66). What is most shameful is not to approve of a falsehood as something that is probably true, but rather to approve of it as something that is definitely true. If Cicero had considered the mere holding of fallible beliefs to be shameful, it is very unlikely that he would go on to admit later in this passage (Luc. 66) to being a great opinion holder. The same holds for Arcesilaus: Cicero would not have intentionally portrayed his venerable Academic predecessor as shameful by his own (Arcesilaan) lights.

1.2 Luc. 66-7 and 76-8

Arcesilaus also thought that the sage was infallible. He “believed, in agreement with Zeno, that the greatest strength of the sage was to avoid being deceived, for nothing is further removed from our conception of the dignity of the sage than error, frivolity or rashness” (Luc. 66). The idea that the sage will hold no opinions seemed true to Arcesilaus and both honorable and fitting for a sage (Luc. 77). Cicero does not tell us how Arcesilaus arrived at this conception, but he does show us how he made use of it as a premise in an argument aimed at Zeno.

(A1) If the sage ever assents to anything, he will sometimes hold an opinion [as long as (A1,1) he has no criterion guaranteeing that what he assents to is true].

(A2) But he will never hold an opinion [because that would make him vulnerable to error and the sage never errs].

(A3) Therefore, the sage will never assent to anything [i.e., he will suspend judgment on everything]. Cicero also reports (Luc. 76-7) a crucially important argument of Arcesilaus’s in support of (Ai.i). This argument held the field for the two centuries separating Arcesilaus’s Academy from Cicero’s. Zeno thought that the sage would never hold mere opinions because he would assent only to propositions that he knew to be true in accordance with his infallible criterion, the kataleptic impression. But Arcesilaus pressed him on this point, paving the way for future Academic attacks on the Stoics. To summarize the attack: for any sense-impression S, received by some observer A, of some existing object O, which is a precise representation of O, we can imagine circumstances in which there is another sense impression S’, which comes either (i) from something other than 0, or (ii) from something non-existent, and which is such that S’ is indistinguishable from S to A. The first possibility (i) is illustrated by cases of indistinguishable twins, eggs, statues, or imprints in wax made by the same ring (Luc. 84-7). The second possibility (ii) is illustrated by the illusions of dreams and madness (Luc. 88-91). On the strength of these examples, Cicero, following Arcesilaus, concludes that we may, in principle, be deceived about any sense impression, and consequently that the Stoic account of empirical knowledge fails.

It does not follow from this argument that there is no infallible criterion of judgment. What does follow is that the Stoics have not succeeded in showing that their kataleptic, sensory impression is such a criterion. The outcome of the exchange between Zeno and Arcesilaus is aporia: neither of them know whether or not knowledge is even possible, despite Zeno’s willingness to teach that it is. Cicero reveals in these passages an Arcesilaus who leads Zeno, dialectically, to unacceptable conclusions on the basis of Zeno’s own views, but it is equally clear that Arcesilaus himself accepts some of the views along the way: in particular, the apparent unknowability of things (as in [a] above) due to the lack of an infallible criterion of judgment, and the conclusion to the above syllogism, (A3), the infallibility of the sage (b).

2. Carneades

As Couissin noted (op. cit. 1929/ 1983, 42), “Carneades is above all the successor to Arcesilaus: in eadem ratione permansit (“he kept to the same standpoint,” Acad. 1.45).” Cicero frequently portrays Carneades as advancing arguments not because he accepts the conclusions but for the sake of examination (causa disserendi, Fin. 5.20, Luc. 131, 139). But what is the point of such an examination? If the dialectical interpretation is right, then Carneades would have sought exclusively to refute his interlocutor’s views.

Lactantius, relying on a portion of Cicero’s De Republica 3 that we no longer have, reports that when Carneades was sent as an ambassador to Rome he argued one day in favor of some Aristotelian and Platonic conception (s) of justice and the next day he argued against his previous conclusions. He did so “not because he thought justice ought to be disparaged, but to show that its defenders had no certain or firm arguments about it.” It is likely that Carneades’ audience contained some who accepted the arguments and conclusions of his first day’s discourse. Thus it seems, on the second day, he sought to relieve them of views that had not been adequately established. We can understand this in terms of seeking to relieve others from the deception of thinking they know what in fact they do not know. If this were Carneades’ goal, it would explain why he sought to show that the proponents of either side had no certain or firm arguments in their favor.

We find a similar use of dialectic at ND 3.43-4. Carneades employed a version of the sorites against Stoic theology in the following way. “If gods exist are the nymphs also goddesses? If the nymphs, the Pans and Satyrs also are gods.. . ” By making such small steps Carneades threatens the Stoics with an absurdly large number of deities, some of which, like the streams of the under-world, clearly should not be counted as such. The further Carneades proceeds the more trivial and clearly ungodlike the examples become. But if these latter examples are not gods neither are the former. To draw a line separating the former from the latter would be arbitrary.

The point, as Cicero presents it, is that Carneades advanced this argument not to do away with the gods, but in order to demonstrate that the Stoics had settled nothing regarding theology. In other words, he did not seek primarily to show the Stoics that their theological views were false, but rather that their arguments did not establish what their authors thought they established.

Carneades also argued against the theological view that the whole world is ruled and governed by divine intelligence, and that many natural things were constructed by the immortal gods expressly for human use. But he argued (disserere) against this view to such an extent that he excited in people a desire for the discovery (or investigation) of the truth (AD 11-4). In keeping with these other examples, it is likely that Cicero thought that Carneades did not himself accept the conclusion to this argument either, but rather that he set it forth causa disserendi. So what these examples indicate is that Cicero thought Carneades employed his negative dialectic to relieve others from reaching conclusions prematurely and thinking they know what in fact they do not, so that they may more effectively seek truth.

But is it correct to attribute any positive methodology for acquiring beliefs to Carneades?

2.1 The Dialectical vs. the Fallibilist Carneades: Clitomachus of Philo?

One of the central issues in Cicero’s account of Carneades is whether Philo understood him correctly as a modest skeptic with a fallibilist epistemology or Clitomachus (as usually interpreted) got him right, portraying him as a pure dialectician, committed to none of his premises or conclusions and drawing these only from his interlocutors. Cicero refers to this difference of opinion regarding Carneades in the Lucullus. The view in question is that the sage will not assent to anything, but nevertheless, he might still form opinions in the absence of apprehension, the perceptual grasping of an impression about which one cannot be mistaken. This view is said by Philo and Metrodorus to have been accepted by Carneades (a Carneade dicitur probatum); however, Cicero puts more trust in Clitomachus’s interpretation that Carneades defended this view dialectically, i.e., that he didn’t approve of it himself but rather advanced it for the sake of argument (Luc. 78).

Generally, these interpretive options are taken as incompatible: dialectician or fallibilist, but not both. Burnyeat argues that Carneades actually developed two distinct dialectical responses to the Stoic apraxia objection: “One way is to act without assent, merely following appearances [Clitomachus/Cicero]; the other is to assent responsibly, with a clear understanding that your opinion is only an opinion and may turn out to be wrong [Philo/Metrodorus].” These are clearly incompatible options insofar as the former involves a passive responding to, or being moved by, appearances and the latter requires some active engagement with the phenomena in order to arrive at one’s best estimate of the truth.

But if the Philonian and Clitomachean options are incompatible, then Cicero is extraordinarily muddled in the Lucullus, for he endorses Clitomachus’s view of Carneades as well as Philo’s general, fallibilist view of Academic methodology. Not only does he register his agreement with Clitomachus’s assessment at Luc. 78 and 108, he also quotes, apparently verbatim and with approval, from Clitomachus’s books on Carneades. Cicero’s description of the tota Carneadis sententia that stretches from Luc. 98 to 104 is taken from volume one of Clitomachus’s four volumes on the withholding of assent (Luc. 98) and from Clitomachus’s volume written for the satirist and member of the Scipionic Circle, Gaius Lucilius (Luc. 102). What is more, Cicero promises to quote from the latter pretty much as it was written since he is familiar with the system or doctrine [institutio et quasi discipline] contained in that book, and since this doctrine is the very subject they are dealing with in the Lucullus. Clearly, Cicero thinks Clitomachus got Carneades right, at least in the passages he relies on as the source for Carneades’ views.

It is unlikely, however, that Cicero wholeheartedly supported the Clitomachean version of Carneades since he approves of a remark that Carneades customarily made: “Clitomachus repeated the substance of his teaching, but Charmadas reproduced the style also.” Perhaps what Carneades meant was that Clitomachus accurately recorded his battery of dialectical arguments-we know from Luc. 16 that Clitomachus was the most industrious of Carneades’ students and wrote numerous books-but that he did not master the rhetorical skills necessary to accomplish Carneades’ goals. The point is that merely having a catalogue of arguments pro and contra does not reveal how these arguments are to be used or what they are supposed to accomplish. What Carneades literally says in this passage is: “Clitomachus said the same things, but Charmadas also said [them] in the same way.”

As for Philo, his mixture of fallibilism and skeptical, open-ended inquiry is attested to on page after page of Cicero’s philosophical dialogues. In the Lucullus (7-9, 66), Cicero clearly endorses Philonian fallibilism-the whole purpose of Academic inquiry is to seek for and discover the most probable view, the one most likely to be true. But again, if the dialectical interpretation of Clitomachus’s view of Carneades is right, then Cicero’s general endorsement of Philonian fallibilism is at odds with his particular endorsement of Clitomachus’s view of Carneades.

There is a further, related difficulty. Even if we ignore the Philonian framework in the Lucullus, we cannot do so in the revised version of the Academic books (see n. 7 above). Cicero remarks in a letter to Varro (Fam. 9.8) that he has taken the part of Philo in his Academic books. Since this letter is to Varro, Cicero certainly meant that he had taken the part of Philo in his revised version in which Varro stars as the Antiochean interlocutor. In the earlier version, Cicero assigns the role of Antiochean to Lucullus. Now, in both the original and revised versions, Cicero defends the New Academic position of Arcesilaus and Carneades. But in the earlier version he does so by relying heavily on Clitomachus’s accounts of Carneades, and in the later version he does so explicitly as a Philonian. Thus, if the dialectical, Clitomachean interpretation is incompatible with the fallibilist, Philonian interpretation, we have another serious problem. Why would Cicero complete his defense of the Clitomachean, dialectical Academy and then within a month begin revising this account by relying on the Philonian, fallibilist interpretation? The reason Cicero decided to revise his Academic books had to do with whom he thought most deserved the compliment of being cast as his Antiochean foil-there is no mention of his changing his mind about the proper interpretation of Carneades or the Academy (Att. xiii.12, and 13). The more natural interpretation is that Cicero defended the New Academy against Antiochus in both editions and that he felt comfortable using both Philo and Clitomachus to accomplish this end because they are consistent and mutually supportive sources.

Thus it seems that we must find a way to reconcile the apparently incompatible Clitomachean and Philonian interpretations, at least when it comes to Carneades’ stand on withholding assent.

2.2 Carneades on Withholding Assent

There is another point on which Cicero agrees with Clitomachus, but here there is no mention of a disagreement with Philo and Metrodorus: “I believe Clitomachus when he writes of the nearly Herculean labor endured by Carneades when he cast assent out of our minds, like a wild and savage beast, that is mere opinion and thoughtlessness” (Luc. 108). The question to ask here is, what exactly did Carneades cast out?

Michael Frede’s view is that there are two distinct, and again, incompatible, answers to this question, Clitomachus’s and Catulus’s. Frede does not explicitly say that he takes Catulus’s view to be Philo’s, but this seems to be the case as he considers it part of what he calls the dogmatic skeptical view of the Academy that confers some epistemic status (e.g., verisimilitude) on the impressions accepted, as opposed to the classical skeptical view of Clitomachus that does not confer any such status, but rather allows the skeptic to acquiesce in impressions without actively accepting them as true or even like the truth. Along these lines, Frede remarks,

We find Clitomachus making a distinction of two kinds of assent, obviously trying to give an interpretation of the distinction which will not commit Carneades to the view that it is wise to have mere opinions (Luc. 104). And it seems clear from Catulus’ remarks [Luc. 148] that the opposing party [Philo’s view prior to his innovations] similarly made a distinction of two kinds of assent, but exactly in such a way that Carneades would be committed to the view that the wise man will have opinions. (op. cit. 1984, 267)

Although Frede’s view of the incompatibility between Philo and Clitomachus is different from Burnyeat’s, I believe it encounters the same difficulty-it forces us to read the Lucullus as a defense of two incompatible views of the Academy. Frede remarks that it is clear that Cicero “does not think that the view Catulus expresses [the Philonian view] is the general view of the Academy” (ibid.). In other words, in a dialogue in which Cicero takes the part of Philo in defending Academic philosophy, Cicero disagrees with Philo’s general view of the Academy. Surely this cannot be right.

Richard Bett does, I think, get Carneades’ view on withholding assent right. He argues that the sort of assent Carneades cast out is the conscious acceptance of an impression as true. Originally, the Stoics claimed that action is impossible without a preceding mental event that they called assent. Whether or not we are aware of it, according to the Stoics, an assent always precedes an action. If Carneades accepted this account of assent he would have to agree that withholding assent does make life unlivable. Instead, what he proposed, according to Bett, was to provide for a different sort of mental event or attitude, approval, which is somewhere between a taking-to-be-true and the complete indifference that might seem to be the alternative. Thus, Bett remarks,

Assent and approval are distinguished not by degrees of certainty, nor by the presence or absence of deliberation [nor by claiming that one is passive, the other active], but simply by whether or not one takes the impressions in question to be true. To assent to an impression is to take a stand on the truth of the impression. Approval, by contrast, involves no commitment, one way or the other, as to whether the impression approved is really true or false. (op. cit. 1990.10)

This distinction does successfully show that a life without assent is possible. Furthermore, though Bett does not make this point, it allows for a Socratically examined life, insofar as it allows for self-conscious and rational selection among one’s impressions. Also, this distinction shows how Carneades can consistently approve of his own distinction, and perhaps even other philosophical views. Although Bett does not discuss the tangled details regarding Philo and Clitomachus, I believe his account of Carneades, which derives primarily from the Clitomachean material, is perfectly acceptable to a Philonian.

2.3 A Unified, Ciceronian View of Carneades

We can now examine some of the details of a resolution of the alleged incompatibility between Philo’s and Clitomachus’s views of Carneades. The passage in which this incompatibility is found, with some of the surrounding context, is as follows.

Arcesilaus showed that there is no impression that is true and of such a sort that it is not possible that there be another of the same sort that is false. This is the one dispute that has persisted to the present day. For the point that the sage will assent to nothing did not pertain to this controversy; for indeed, he was allowed to apprehend nothing and nevertheless hold an opinion-which was said to have been approved by Carneades, but for my part, trusting Clitomachus more than Philo and Metrodorus, I believe this was argued [dialectically] by him rather than approved. But let us drop this … if I show that apprehension is not possible, you must concede that the sage will never assent. (Luc. 77-78)

The reason the sage’s non-assent does not pertain to the controversy, Cicero thinks, is that the primary issue is whether there are any impressions about which we cannot be mistaken. In other words, before we can determine whether or not the sage assents, we must first establish whether or not katalepsis is possible, that is whether there is anything worthy of a sage’s assent. But in the Lucullus, katalepsis is neither shown to be possible nor shown to be impossible. What we are left with is the claim that the Stoics have not successfully established the foundation of their epistemology-they have not shown that there are impressions about which we cannot be mistaken.

Similarly, Carneades held the view that it is probable, but not certain, that katalepsis is not possible (Luc. 110) Consequently, since he had not conclusively settled the issue regarding the possibility of knowledge, he should not have been willing to positively assert either that the sage will assent or that he will not-i.e., Clitomachus will have got him right. This seems to be the case as we find Carneades associated with both the assertion (Luc. 59, 67) and the denial (Luc. 78) of the proposition that the sage assents; for he would have argued for both positions for the sake of examining the issue. Moreover, if an interlocutor were to say that the sage sometimes forms an opinion (and thus, assents), Cicero would refrain from combating him, especially as even Carneades does not vehemently combat this position (Luc. 112). All of this suggests that Carneades suspended judgment on the issue of whether the sage will ever assent and hold mere opinions.

This is the most reasonable position for a non-sage to hold. For how can we know what a sage will or will not do prior to becoming one, or, short of that, prior to being able to identify one? This modest attitude toward one’s ethical ideal does not preclude speculation as to what the sage will do, though it may preclude positive endorsement. Indeed, such speculation may be a necessary part of the struggle to achieve the ideal. It is in this spirit that Cicero undertakes the main topic of investigation in the Lucullus, the sage (Luc. 66).

Returning to the troublesome Luc. 78 passage: Philo’s mistake, from Cicero’s perspective, is to attribute a definite view regarding the sage to Carneades. In order for Carneades to approve of the view that the sage will hold opinions, he must also believe that he has settled the prior issue of whether or not katalepsis has been shown to be possible, which he would not have done. This is not to say that Philo misunderstood Carneades’ method as a whole. Indeed, Cicero does not consider Philo’s mistake terribly important; instead, he dismisses it to get on to more important matters. Although we do not know whether the sage will assent and hold opinions, we non-sages may continue to do so responsibly in the pursuit of truth. On this point, if I am right, Clitomachus, Philo, and Cicero’s Carneades were all in agreement.

3. Conclusion

There are two sorts of reasons in favor of discounting Cicero’s view of his predecessors: (i) he intentionally misrepresents them for the sake of a prestigious historical continuity stretching back to Plato and Socrates, or (ii) he unknowingly misrepresents them because he has been misled by his teacher Philo, or other contemporary Academics, who may have been motivated by (i).

(i) If Cicero had intentionally engaged in historical revisionism then he also would have taken pains not to be obvious about it. And indeed, there is not the slightest hint of intentional revisionism in Cicero’s remarks regarding Arcesilaus and Carneades. But it seems that in the absence of any solid evidence we should take these passages at face value, as statements of Cicero’s own view of the dialectical practices of his great predecessors, the model for his own phiosophical method.

(ii) If we thus deny (i), and still wish to discount the historical accuracy of Cicero’s testimony, then we will need to attribute to him fundamental errors in his understanding of the history of the Academy. In this case Cicero will be guilty not of dissimulation but rather ignorance. Yet this is a hard line to toe, for Cicero is arguably our most extensive, most sympathetic and best-informed source for the philosophy of Arcesilaus and Carneades. His credentials are excellent: he was a student of the Academy from his youth and had the opportunity to learn from philosophers who themselves had been students of Carneades. These facts do not, of course, guarantee that Cicero’s view of his predecessors is accurate. However, they do provide good reason to think that he knew what he was talking about. It remains possible that our best source is himself quite poorly informed, but I believe it makes better sense to proceed on the assumption that this is not the case. I also believe that a more thorough treatment of the evidence regarding Arcesilaus and Carneades would bear out the conclusions presented here. In particular, Cicero was correct to claim that he practiced a philosophical method that extended from Arcesilaus and Carneades down to the time of Philo’s innovation.