Ancient Skepticism (Part 3)

3. Academic Skepticism

3.1 Arcesilaus

With Arcesilaus (316/5–241/0 BCE) and his role as leader of the Academy (266/268 BCE), Plato’s Academy turns skeptical. Arcesilaus does not refer to himself as a skeptic—this nomenclature is a later designation. However, Arcesilaus stands at the beginning of a re-orientation in the history of Platonically inspired philosophy. He rediscovers Socrates the examiner. Socrates’ commitment to investigation, to the testing and exploring of one’s own and others’ beliefs, and his passion for weeding out falsehoods, are the starting-points of his Academic skepticism (Cicero, Acad. 2.74, 1.46). Throughout the history of this skeptical school, these traits, and the corresponding commitment to a life guided by reason, remain alive (Cooper 2004b, Vogt 2013). When, as we shall see below, Arcesilaus defends a skeptical life without belief, this is because, as he thinks, reason itself, if properly and faithfully followed, leads us to live that way. To Arcesilaus, the skeptical life is a life lived following reason, a life based on reason—just as the competing Stoic and Epicurean lives are alleged by their proponents to be. Arcesilaus engages with the epistemologies of these contemporaries of his. In particular, the Academics call into question that there is a criterion of truth, as both Epicureans and Stoics, beginning in the generation before Arcesilaus, claim there is.


Like Socrates, Arcesilaus did not write anything. His views must be unearthed from Sextus’ comparisons between Pyrrhonian and Academic skepticism, from Cicero’s discussions in the Academica, and from a range of shorter (and sometimes hostile) reports. Major themes in Arcesilaus’ philosophy are (i) his dialectical method, (ii) discussion of whether there is a criterion of truth, and (iii) his defense of the skeptic’s ability to act.

(i) Method. Arcesilaus embraces what scholars call a dialectical method (Couissin 1929 [1983], for a reading that calls the dialectical interpretation in question, cf. Perin 2013). This method is inspired by Socrates. It proceeds by asking one’s real or imaginary interlocutor what they think about a given question, then plunging into an examination of their views, employing their premises. Can they explain their position without running into inconsistencies, and without having to accept implications that they want to resist? As a consequence of this method, it sometimes appears as if a skeptic, while examining someone’s view and its consequences, makes a positive claim: “so, such-and-such is not so-and-so.” However, within a dialectical exchange, this should be read as “according to your premises, such-and-such follows.” This method remains a key ingredient of Greek skepticism. While the different skeptical schools develop variants of the dialectical method, skeptical argument is often characterized by the fact that skeptics think of themselves as engaging with “dogmatic” interlocutors. (In the skeptical tradition, as articulated for example by Sextus Empiricus (see section 4.4.), “dogmatists” are philosophers who put forward, and defend, positive answers to philosophical questions about reality, knowledge, ethical values, etc. They need not do so dogmatically or rigidly or without consideration of alternatives in order to count, in skeptical terms, as “dogmatists.”)

(ii) The Criterion of Truth. Zeno, the founder of Stoicism, and roughly 20–30 years Arcesilaus’ senior, was for a time a student at the Academy. He was still in the Academy when he formulated key Stoic doctrines. Like Arcesilaus, he claims Socratic ancestry. Zeno is inspired by some of the same ideas that inspire the skeptics. In particular, he engages with the Socratic idea that knowledge is integral to virtue. Contrary to Arcesilaus, Zeno aims to give accounts of knowledge and virtue, and holds them up as ideals that human nature permits us to achieve. For him, knowledge is very difficult to attain, but ultimately within the reach of human beings. From the point of view of Arcesilaus, Zeno’s claim to Socratic heritage is almost offensive: Zeno seems to be too optimistic about our cognitive powers to be following Socrates (Frede 1983). Scholars have traditionally envisaged an exchange of arguments between Zeno and Arcesilaus, where each modified his views in the light of the other’s criticism. However, Zeno most likely formulated his views between 300–275, and Arcesilaus argued against him c. 275 to 240, when Zeno (who died c. 263) was probably already retired (Brittain 2006, xiii; Alesse 2000, 115 f.; Long 2006, ch. 5).

The core of the dispute between Arcesilaus and the early Stoics concerns the question of whether there is a criterion of truth. The notion of a criterion is introduced into Hellenistic discussions by Epicurus, who speaks about the kanôn (literally measuring stick) and kritêrion. For Epicurus, a criterion is that evident thing, viz., the content of a sense-perception, against which claims about the non-evident are tested. For example, physics advances claims about non-evident things, such as atoms and void. These are not accessible to the senses, and accordingly, do not count as evident. Perception rules out various physical theories. For example, a physical or metaphysical theory according to which there is no movement can be dismissed because it is in disagreement with the evident.

Zeno argues that a certain kind of impression—namely a cognitive impression (phantasia katalêptikê)—is the criterion of truth (cf. Shogry forthcoming-b). Zeno’s conception of cognition (katalêpsis, literally grasping, apprehension), which figures in the notion of a cognitive impression, tries to resolve a basic epistemological problem. Belief-formation aims at the truth; there is a norm inherent in the practice of believing that one should only believe truths. It is not transparent to us, however, which of our beliefs, or claims to the effect that such-and-such is true, aim successfully at the truth. Zeno argues that some impressions are cognitive. That is, on his account, they are stamped and reproduced from something that is, exactly as it is; so they grasp it as it really is. Hence, being so stamped and reproductive, they reveal through themselves that they are cognitive. For example, when I look at my computer screen while typing this, I may very well have a cognitive impression that this is my computer screen. When I look up and out of the window, I have an impression of a friend walking across campus that is probably non-cognitive. This impression might be true. But since I see her from such a distance, it is pretty surely not cognitive. That is, not all true impressions are cognitive, but all cognitive impressions are true. We should only assent to cognitive impressions, and so in forming our beliefs, hold only those certified beliefs (cf. Brittain 2014 on the Stoic view that assent to cognitive impressions is not compelled). All other beliefs—based on assents to non-cognitive impressions—violate the norm for believing that you should only believe truths, even if what you do believe may be true.

Arcesilaus calls into question whether there are impressions of this kind. His main point seems to be that there could be an impression that is phenomenologically indistinguishable from cognitive impressions, but nevertheless misrepresents the matters it gives an impression about. To use an example that may derive from Carneades (see section 3.2), there is no impression of a given egg such that no impression of any other egg could be phenomenologically indistinguishable from it. In response to this, the Stoics defiantly added a clause to their definition, given above, of the cognitive impression: “and of such a kind as could not arise from what is not” (Long and Sedley (1987) [= LS] 40; DL 7.46, 54; Cicero Acad. 1.40–1, 2.77–8; SE M 7.247–52). Absent a criterion of truth, Arcesilaus’ skeptic suspends judgment about everything (PH 1.232). Reason itself, Arcesilaus thinks, demands such suspension.

(iii) Action. If skeptics suspend judgment, argues their dogmatic opponent, they are not able to act. Stoic philosophy conceives of three movements of the mind: impression, assent, and impulse (Plutarch, Col. 1122a-d). All three figure in action. The agent assents to the impression that A is to be done; their assent is an impulse for the action A; if there is no external impediment, the impulse sets off the action (Inwood 1985). It is a cornerstone of Stoic philosophy that there can be no action without assent, and so without the belief that the action done is to be done. The Stoics aim to avoid the kind of determinism according to which actions are not ‘up to’ the agent (Bobzien 1998); for them, impressions, but not assents, are caused by external things. In response to the Apraxia Charge, Arcesilaus seems to have argued that the skeptic can act without having assented (Plutarch, Col. 1122A-d), and so without believing that the action done is to be done. However, this is not his complete response. From the point of view of the Stoics, skeptical action, if performed without the relevant kind of assent (that is, assent that it is up to the agent to give, and that is a rational acceptance of the impression), is like the action of a non-rational animal, or like the automatic movement of plants when they grow and flourish. Arcesilaus is robbing people of their minds (Cicero Acad. 2.37–9; Obdrzalek 2013). But Arcesilaus need not and does not go so far as to compare human agents with non-rational agents. As human beings, skeptics have rational impressions. They perceive the world conceptually, and think about it. Arcesilaus does not suggest that skeptical action is causally set off by impressions, or in the way, whatever that is, that animal actions are set off. This would be a problematic proposal, for it would disregard that the skeptic has a human mind. Given the complexity of human thought, the skeptic is likely to have several, and often competing impressions. If all impressions triggered impulses, the skeptic would be inactive due to a kind of paralysis. The second component of Arcesilaus’ reply, thus, is that the skeptic, in acting without assenting, adheres to the reasonable (eulogon) (SE M 7.158; 7.150; Striker 2010). That is, Arcesilaus aims to explain skeptic action as rational agency (Cooper 2004b). Arcesilaus disputes the dogmatic claim that some impressions can be identified as true, and the related claim that one can only act on the belief that some impression is true. But he does not argue that there are no differences between impressions which agents could take into account. His agents are rational: they think about their options, and go with what looks, in one way or another, more plausible.

Arcesilaus defends skeptical action also against Epicurean critics (Plutarch, Col. 1122A-d), again by showing on the basis of the Epicureans’ own premises that skeptical action is possible. Can the sceptic explain why, when leaving a room, they go through the door rather than running into the wall? Arcesilaus seems to have exploited the Epicurean view that, while all sense-perception is true, belief can introduce falsehood. Like the Epicurean, the skeptic can keep apart the perception and a view formed on its basis. By not assenting to the perception, thus adding belief (“here is the door”), the skeptic guards against the source of falsehood, namely belief. But a skeptic has perception of the door available to them, which is enough for not running into walls.

3.2 Carneades

Like Arcesilaus, Carneades (214–129/8 BCE) refrains from writing and philosophizes in a Socratic spirit. Famously, on an embassy from Athens to Rome in 156/5 BCE, Carneades argues for justice one day, and against justice the next. His aim is not to overthrow justice. Carneades wants to show that the supporters of justice—including Plato and Aristotle—do not have the successful arguments they think they have to show what justice is and what it requires (Lactantius, Epitome 55.8, LS 68M). Like Arcesilaus, Carneades (i) engages with Stoic epistemology. His account of skeptical action includes (ii) a detailed proposal regarding the criterion. As part of his less radical skepticism, Carneades seems (iii) to allow for a certain kind of assent, and perhaps for belief.

(i) The Stoic-Academic Debate. Chrysippus, the third major Stoic (after Zeno and Cleanthes), and his student Diogenes of Babylon, revise Zeno’s epistemology, defending it against Arcesilaus’ arguments (Brittain 2006, xiii). In response to their arguments, Carneades continues the exchange with the Stoics that Arcesilaus began (SE M 7.402–10). His first move addresses the link between mental states and action. People in states of madness, he argues, act just as easily and naturally on their impressions as other people, even those who act on cognitive impressions (if there are any). From the point of view of exhibited behavior, there does not seem to be a difference: any and all impressions, even those the Stoics think clearly arise from something that is not, are in all respects relevant to action completely on a par. Cognitive impressions, if there are any, have no superiority.

In a second argument, Carneades points to objects that are similar to one another: can the wise person discern any two eggs, two grains of sand, and so on? The Stoics have multiple replies. It is conceivable that, in some contexts of action, the wise person assents to what is reasonable (eulogon) (DL 1.177), without having a cognitive impression of how things are. Or, if faced with the task to identify grains of sand while lacking a cognitive impression, the wise person can suspend judgment. However, the wise will train themselves so as to be able to perceive minute differences (Cicero, Acad. 2.57), where it might be important to do so. This point is backed up by Stoic physics: no two items in the universe are identical, and their differences are in principle perceptible. Carneades replies that even if no two things were exactly alike (consistent with his general line of argument, he does not take a stance on such questions), a very close similarity could appear to exist for all perceivers (Cicero, Acad. 2.83–5); that is, the impressions of two items, though in fact these items might differ from each other, could be indistinguishable (Shogry forthcoming-a and forthcoming-b). Discussion continues with a move on the part of the Stoics: they add to their definition of the cognitive impression “one that has no impediment.” Sometimes an impression is—as it were, by itself—cognitive, but is unconvincing due to external circumstances (SE M 7.253). It is a difficult question whether this addition harms the Stoics more than it helps them. If the initial conception of a cognitive impression hangs on the idea that something about its phenomenological nature, or something internal to the impression, marks it as cognitive, the Stoics give up on a crucial assumption if they grant that sometimes there are “impediments.” If, however, cognitive impressions are differentiated by a causal feature (the way they are caused by the ‘imprinter’ which causes the ‘imprint’), the further addition might help (Frede 1983, Nawar 2014), since the impediment might need to be removed before the causal connection could be confirmed (on the Stoic-Academic exchange, cf. Hankinson 2003).

(ii) Carneades’ Criterion. Even though Carneades further pursues a discussion begun by Arcesilaus, he does not simply continue within the framework of Arcesilaus’ skepticism. The distinctiveness of his position is best seen in the context of his criterion: the persuasive (pithanon). The notion of the persuasive can be understood in two distinctively different ways. Persuasiveness might be a causal feature, so that a persuasive impression sets off a physiological process of being moved in a certain way. But there may also be a rational kind of persuasiveness. Carneades construes persuasiveness in rational terms. For him, the persuasive is the convincing, or perhaps even the plausible.

Carneades develops a three-stage criterion: (1) In matters of importance, skeptics adhere to the persuasive. (2) In matters of greater importance, they adhere to the persuasive and undiverted. A persuasive impression is undiverted if there is no tension between it and its surrounding impressions. (3) In matters that contribute to happiness, skeptics adhere to persuasive, undiverted, and thoroughly explored impressions. A persuasive impression is undiverted and thoroughly explored when it and the surrounding impressions are closely examined without its persuasiveness being diminished (SE M 7.166–84). Consider an example. A skeptic looks in a dark room for a rope. Before they pick up what appears to them to be a rope, they look closely and poke it with a stick. Coiled objects can be ropes, but they can also be snakes. The persuasive impression that this is a rope must be examined before the skeptic adheres to it (M 7.187).

The three-stage criterion is put forward in the context of action. However, Sextus describes Carneades’ criterion as a criterion of truth, not a criterion of action (M 7.173). Carneades might take himself to offer more than a practical criterion. His discussions of the persuasive come close to a general epistemological theory (Couissin 1929 [1983], Striker 1980, Bett 1989 and 1990, Allen 1994 and 2004 [2006], Brittain 2001). Cicero renders the Greek pithanon as probabile (and sometimes as veri simile), which modern editors sometimes translate in terms of what is probable or likely to be true. Some scholars think that Carneades is an early thinker about likelihoods, and argue that he develops a fallibilist epistemology (Obdrzalek 2004).

(iii) Assent and Belief. Does adherence to persuasive impressions involve belief? Carneades coins a term for the kind of adherence he has been describing: approval (Cicero, Acad. 2.99). He distinguishes it from assent in the sense of Stoic and other dogmatic theories, which establishes a belief that something is in actual fact true; but he nevertheless describes it as a kind of assent (Cicero, Acad. 2.104). Carneades’ disciples disagree on whether approval is any kind of genuine assent. That is, they disagree on whether, in approval, one forms a belief. Philo and Metrodorus think that Carneades allows for some kind of belief, close to or identical with belief as the Stoics understand it. Clitomachus disagrees, and Cicero follows Clitomachus (Acad. 2.78, see also 2.59, 2.67). Scholars continue to debate these issues, and the basic problem remains unchanged. It is not clear whether there is a plausible notion of belief according to which belief falls short of ‘holding to be true’ (or according to which, though some kind of ‘holding true’ is involved, the relevant affirmation-as-true is weaker than in beliefs as Stoics and other dogmatist epistemologists conceive of them). In any event, affirmations as true, at least of the full and flat-out (“in actual fact”) sort the Stoics think of, are precisely what the skeptic does not make.

Another approach to Carneades’ stance toward belief is to ask whether he might invoke Platonic considerations. Consider that Socrates, when asked in the Republic what he thinks the good is, refuses to reply because he thinks beliefs without knowledge are shameful (Rp. 506c). In response to this, his interlocutors point out that there is a difference between putting forward one’s beliefs as if one knew them to be true, and putting them forward with the proviso that they are merely beliefs (Vogt 2012a, ch. 2). The shamefulness of mere belief might disappear through this proviso. A passage in Cicero’s Academica suggests that Carneades invokes this thought. According to Carneades, the wise person can hold beliefs if they fully understand them to be beliefs (2.148). Along similar lines, it has been suggested that Carneades might conceive of a hypothetical mode of believing (Striker 1980 [1996, 112]), perhaps engaging with a move in Plato’s Meno. Investigation cannot get off the ground if we do not, in some sense, begin with our beliefs about the matter under investigation. But how can we do so without endorsing our beliefs, not knowing whether our views are true? Plato’s answer at this point is: by hypothesizing our beliefs. Today we would insist that hypotheses are not beliefs. However, it is conceivable that Carneades argued along these lines, and that the details of his vocabulary got lost or confused in doxography.

3.3 Later Academic Skepticism

Carneades was an enigma to his students and immediate successors. Clitomachus (head of the Academy from 127 to 110 BCE) seems to have attempted the impossible: to adhere closely to Carneades’ philosophy, even though he never understood what Carneades truly meant (Levy 2010). The cornerstone of his adherence lies in the view that Carneades argues for suspension of judgment and against beliefs understood as Stoics understand them. Philo of Larissa, another student of Carneades, interprets his teacher as allowing for tentative beliefs in the skeptic’s life. With Philo, the skeptical era of Plato’s Academy comes to an end. Philo’s philosophy seems to divide into two phases. In Athens, and as head of the Academy, he stays relatively close to Carneades. Moving to Rome later in his career, he develops a markedly different position. He argues only narrowly against the Stoic criterion and their conception of cognition. One can apprehend things and so come to know them—one just cannot apprehend them in the way in which the Stoics construe cognition (PH 1.235). The fact that there is no apprehension in the sense of the Stoics does not mean that there is no knowledge (Acad. 2.14). This move shifts the discussion in several important ways. First, Philo can be interpreted as a kind of externalist: one can know something without knowing that one knows it. Absent Stoic cognitive impressions, we are not able to identify which instances of ‘holding-true’ qualify as knowledge; but we nevertheless have some knowledge (Hankinson 2010). Second, this proposal is a step toward modern skepticism, which is not concerned with criteria of truth, but with knowledge.

Cicero’s skeptical philosophy in his own philosophical writings is again distinctively different. In line with his notions of what is probable (probabile) or likely to be true (veri simile), Cicero often examines a range of philosophical positions, aiming to find out which of them is most rationally defensible. He thinks it is better for us to adopt a view that is likely to be true, rather than remain unconvinced by either side (Thorsrud 2009, 84–101). Cicero is of the greatest importance for the transition between ancient and early modern skepticism. As in other fields of philosophy, Cicero’s influence is partly the influence of the translator. In transposing philosophical ideas into the language of a different culture, the ideas change. Cicero sometimes speaks of doubting, dubitari (e.g., Acad. 2.27, 106; however, he often sticks with the earlier language of assent and suspension). But doubt has no place in Greek skepticism (see section 1).

(End of Part 3)


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

Blog at

Up ↑

%d bloggers like this: