1.3.1 Socratic Wisdom and Paradoxes

A natural way to signal epistemic neutrality is to acknowledge one’s ignorance. Consider, for example, Protagoras’ religious agnosticism when he claims about the gods, ‘I cannot know (ouk echō eidenai) whether they exist or whether they do not’.Footnote ²⁸ Similarly, when someone asks me which candidate will win the next election, I reply that I don’t know. In my case, it means that I have not made up my mind about it. I could be interested in knowing, aware of the options and recent polls, and still fail to form a belief. Maybe public preferences are shifting, and it is difficult to predict who will come on top. Protagoras’ case might be a more committed agnosticism, though. He could have considered the matter and concluded that he would never be able to form a belief about the subject or refuse to inquire because he has good reasons to think he will not succeed.

Some philosophers distinguish between opinion and knowledge. In such cases, someone could claim no knowledge of a subject yet hold all sorts of opinions about it. Take Xenophanes, who suggests that in religious matters, the fallback position is opinion, not suspension.Footnote ²⁹ However, Protagoras seems to mean that he not only holds no knowledge but also no opinions about the existence of the gods. Similarly, when I acknowledge my ignorance about who will win the election, what I mean is that I have no knowledge or opinion about it, despite inquiring and being interested in the question.

Expressing one’s ignorance is not the most precise way to convey a neutral stance or posture. As we saw, mere lack of belief is not suspension (Section 1.1). But in many communicative contexts, if someone discloses ignorance, it is often safe to infer that what they mean is suspension of belief.Footnote ³⁰ But what would we need to be more explicit and precise? As I showed, some modern philosophers understand suspension as higher-order beliefs or knowledge of ignorance, plus some additional conditions. To use Reference MasnyMasny’s (2020, 5024) definition again, maybe I could say I believe that this year I neither believe nor disbelieve that Barcelona will win the Champions League, I neither believe nor disbelieve it, but I intend to judge it before the end of the season. We seem to find examples of neutrality expressed in analogous ways in ancient philosophy.

Socrates (469–399 BCE) is famously known for saying that ‘he knew nothing except the fact that he knew nothing’ (DL 2.32). On its own, this claim sounds like disclosure of a general, almost global, neutral stance. Such an interpretation might go as far back as Arcesilaus, the first sceptical Head of the Academy. Or at least to Cicero (106–43 BCE), when he reports that ‘Arcesilaus used to deny that anything could be known, not even the residual claim Socrates had allowed himself, i.e., the knowledge that he didn’t know anything’ (Acad. 1.45).Footnote ³¹

Socrates’ famous claim, however, is nowhere to be found in Plato or any of the other early sources. The closest we get is Plato’s Apology 21d. When Socrates discovers that those who appear wise were not, he thought to himself:

In this passage, the scope of Socratic ignorance, the force of the claim and its formulation differ from the claim found in Diogenes Laertius and Cicero. For Plato, the extent of Socrates’ ignorance is not absolute but restricted to what is worthwhile (kalon kagathon; lit. ‘beautiful and good’). Socrates might have some other types of knowledge, and he does not entirely rule out the possibility of having some worthwhile knowledge. Moreover, Plato’s Socrates does not claim to have knowledge of his ignorance. He only claims that he does not think to know what he ignores (although he considers himself wiser than others because of that).Footnote ³³

The difference between these Socratic claims is not trivial. But both cases seem to convey not mere lack of knowledge but something akin to suspension of belief.Footnote ³⁴ Given that Socrates was engaged in a philosophical inquiry into the topics he claims to be ignorant of, his position falls closer to higher-order content conceptions of suspension. We might be tempted to construe Plato’s Socrates as saying that he does not believe (nor disbelieve) that he knows what he ignores and does not know what he ignores but intends to judge it.

If generalised, this type of higher-order content formulation risks becoming paradoxical. For example, when I say I know that I know nothing. If it is true, then I do know something, which makes what I said false. But the claim cannot be both true and false. Someone might worry that these paradoxical formulations are self-defeating nonsense. But as mentioned in the previous section, Sextus Empiricus, at least, embraces them and says that the sceptics can use them precisely because these expressions are self-cancelling, which protects the sceptic against any accusation of affirming or denying them, even if using them alienates the sceptic’s interlocutors.

A formulation like the one Cicero attributes to Arcesilaus (nothing can be known) could be accused of the same problem. However, since it is not a knowledge claim, Arcesilaus avoids paradox.Footnote ³⁵ Cicero attributes a similar but more extreme version of this statement to Metrodorus of Chios (fl. 4th century BCE), who ‘claimed at the beginning of his book On Nature: I declare that we don’t know whether we know anything or nothing, not even whether we know that or not, or altogether whether anything is the case or not’ (Cic. Acad. 2.73).Footnote ³⁶ Metrodorus’ formulation also seems designed to be as general as possible while avoiding contradiction or paradox (more on this in Section 3). But is the implication mere ignorance or suspension?Footnote ³⁷

1.3.2 Aphasia and Aporia

We find the term aphasia already in Euripides (c. 480–c. 406 BCE), meaning speechlessness caused by surprise or fear (Hel. 549 and Heracl. 515). In Plato, we find it only twice. Both instances report a certain incapacity to respond to an argument. The first is in Philebus. After listening to Socrates, Protarchus says: ‘this argument has left me absolutely speechless (aphasian pantapasi) for the moment’ (21d4). The second is in the Laws (636e), where Megillus confesses that he and his companions cannot say anything (aphasia) to respond to the Athenian’s arguments. Although using other terms or phrases, we find confessions like this in many other dialogues. Like the previous ones, most happen when someone cannot refute or contest an argument.Footnote ³⁸ Sometimes, aphasia occurs after aporia.

At its core, aporia is a distinctive way of being puzzled or perplexed due to an inquiry into a difficult philosophical question. It is an intellectual impasse in which we perceive no way out or lack the resources to answer a question we have been discussing. Speechlessness or incapacity to articulate thoughts is a frequent effect.Footnote ³⁹ This philosophical use of the term aporia goes back to the everyday use of the word aporos, which means ‘impassable’, ‘pathless’, ‘without means or resources’, and ‘at a loss’Footnote ⁴⁰ Heraclitus (fl. c. 500 BCE) uses it in this way when he says, ‘If one does not expect the unexpected one will not find it (exeurein), for it cannot be searched out (anexereunēton) nor arrived at (aporon)’.Footnote ⁴¹

We can also trace back aporia to the reception of Xenophanes (c.570–c.478 BCE) and Zeno of Elea (c.495–c.430 BCE). In one report, we read, ‘Xenophanes was in aporia about all things and held as his only dogmatic view that all things are one and that this is god, who is limited, rational, and changeless’ (Ps-Gal. Hist. phil. 7=DK21A35).Footnote ⁴² The contrast between aporia and holding dogmatic views suggests that the former entails suspension of belief.Footnote ⁴³

Aristotle, in turn, reports that ‘Zeno was puzzled (ēporei) because if place is something, it will be in something’ (Ph. 210b22-25).Footnote ⁴⁴ We get the following explanation of the puzzle in Simplicius: ‘For every existing thing ought to be somewhere. But if place is an existing thing, where could it be? Surely in another place, and that in turn in another and so on indefinitely’ (In Phys. 563.17–20, Eudemus fr. 78). Therefore, we are invited to infer that place does not exist. But how does that lead to aporia instead of, for example, a reductio ad absurdum? It causes aporia if someone, for example, is neither completely ready to reject the argument nor give up the belief that place must exist. The normative recommendation might be to suspend belief, but the discomfort of such a fallback position propels the inquiry forward. Maybe it does not even require us to suspend our belief about the existence of place, but only about whether we can explain why that is the case.

The obvious place to study the use of aporia is Plato’s dialogues. Plato offers different, seemingly inconsistent accounts of aporia. He often describes it as a passive state of mind. Meno, for example, famously describes it as having touched a torpedo fish and getting one’s mind and tongue numb (Meno 80a-b). In Laches, Socrates describes aporia as being storm-tossed by an argument (194c). But other passages suggest a more active picture. In the Meno, Socrates claims to be more perplexed than anyone else about what virtue is, but his perplexity makes him keep the enquiry going.Footnote ⁴⁵ Later, he stresses that aporia fosters a desire to find the correct answer to the question under discussion (Meno 84b-c).Footnote ⁴⁶ This suggests that Socrates conceives aporia as an active cognitive process consisting of puzzling over a question to find the answer. But then, which is it? Is it a passive perplexity or a restless mulling over a question? The answer is both.

The Platonic images offer, on the one hand, a descriptive illustration and, on the other, a normative account of aporia. Secondary interlocutors like Meno and Laches provide descriptive illustrations: a window into how some people react to aporia. They show a spectrum of possible reactions to perplexity caused by argument. Some people get upset or frustrated or assume their interlocutor has harmed them. For example, after failing to answer what courage is, Laches acknowledges aphasia and frustration: ‘I am really getting annoyed at being unable to express what I think in this fashion’ (Lach. 194a8-b1). But Laches’s aphasia does not entail suspension: ‘I still think I know what courage is, but I can’t understand how it has escaped me just now, so that I can’t pin it down in words and say what it is’ (Lach. 194b1-4). Unlike Plato’s Socrates (and the Sceptics), Laches disconnects his puzzlement and speechlessness from his claim of knowledge. Thus, he claims to know even if he cannot articulate it in words. If knowledge is not connected to explanation, Laches’s position is rational. But if we accept, with Plato’s Socrates, that knowledge is necessarily linked to explanation, then Laches’s position is irrational (see Section 2.2).

In contrast, Plato’s Socrates provides a normative account. He explains what we ought to do when we find ourselves in aporia. In other words, he is interested in the epistemically responsible thing to do. When someone is receptive to aporia, it marks cognitive progress in the investigation. It lets you become aware of your ignorance and mistakes, inviting you to withhold or withdraw judgement and keep looking.Footnote ⁴⁷ In other words, if in aporia, you should suspend belief. These beneficial features of aporia also help distinguish it from the puzzlement derived from eristic argumentation. The difference, however, is not so much in the puzzling itself but, as Jan Reference Szaif, Politis and KaramanolisSzaif (2018, 41) remarks, in ‘the ethos underlying the refutational practice’. Eristic argumentation aims to entertain, impress, and defeat interlocutors.Footnote ⁴⁸ Instead, aporetic philosophical elenchus seeks to find the truth and acquire knowledge, or at least avoid erroneously claiming to know what one does not know. Could the difference be in the persistent intention to judge the issue?

Aporia is an important concept in philosophy after Plato, too. We find it in Arcesilaus, Aenesidemus, Aristotle, Plutarch (c.45–120 CE), Sextus Empiricus, Alexander of Aphrodisias (fl. late 2nd and early 3rd centuries CE), and other later thinkers.Footnote ⁴⁹ Here I cannot discuss the whole evolution of the term, but I shall return to some of these authors in later sections. For now, it suffices to say that philosophers used aporia as a philosophical tool and assumed that the problems that cause this type of puzzlement must be resolved to make up our minds about the questions under discussion.

1.3.3 Lack of Opinion in Pyrrho

Pyrrho (c.360–270 BCE) wrote nothing and what we know from him comes from different sources. His thought appears influenced by DemocritusFootnote ⁵⁰ and Anaxarchus (c.380–320 BCE).Footnote ⁵¹ Some people think that he had contact with Indian philosophy.Footnote ⁵² In one of the surviving reports, we get the following argument from his disciple, Timon (c.320–230 BCE):

The interpretation of this passage is the epicentre of heated debate. One of the sources of disagreement is how to read the adjectives adiaphora, astathmēta, and anepikrita. Are they characterising reality or our epistemic limitations to access it? We could read the Greek terms in both ways. Depending on how we take them, the first claim [a] is either metaphysical, epistemological or both.Footnote ⁵⁴ Many also take issue with the second sentence [b]. Reference StopperStopper (1983, 293), for example, complains that the inference from [a] to [b] looks ‘zany’, and it should go the other way around. He suggests, following Reference ZellerZeller (1909, 501), amending the text from dia touto (‘for this reason’) to dia to (‘on the grounds that’). In this way, the unreliability of our sensations and opinions would explain that the things are adiaphora, astathmēta, and anepikrita, not vice versa. This change would place Pyrrho closer to what later sceptics argue. However, his suspension would be derived from the nature of our epistemic faculties, not from equipollence.Footnote ⁵⁵ According to this, we can reconstruct the main argument in at least two ways:

I will not adjudicate here which interpretation is correct. I only want to highlight that the normative recommendation is a suspension of belief in both cases. And the neutral stance results from voluntarily withholding or retracting belief. That is why when we perceive something or ponder an opinion, we should remind ourselves that it no more is than is not, or it both is and is not, or it neither is nor is not.Footnote ⁵⁷ Given the metaphysical indeterminacy of sensible things, our cognitive limitations, or a combination of both, the rational thing to do, Pyrrho suggests, is to remain adoxastos and aklinēs. So, we ought to get rid of all our beliefs.Footnote ⁵⁸ Moreover, notice that suspension follows from the argument and that aphasia follows suspension, not vice versa, as in Sextus Empiricus. The evidence is insufficient, but it is tempting to construe both versions of Pyrrho’s argument as higher-order accounts akin to a settled agnosticism instead of an active and inquiring suspension.

1.3.4 The Debate between Stoics and Academic Sceptics

This debate goes back to the Platonic Academy under the headship of Polemo (c.350–267 BCE) and Crates (died c.268–264 BCE). Polemo taught both Zeno of Citium (c.334–262 BCE),Footnote ⁵⁹ the founder of Stoicism, and Arcesilaus, who became the head of the Academy after Crates and began what is known as the New or Sceptical Academy, which focused on arguing dialectically, cross-examining other philosophical traditions, without revealing their own views.Footnote ⁶⁰ The evidence suggests that the interchange between Zeno and Arcesilaus was civil and focused on the characterisation of Stoic wisdom and the Stoic sage, specifically, on what the sage can know and the practical implications of answering that question.Footnote ⁶¹ Later, the debate became a complex and lengthy interchange between the Stoics and Academics. I will only mention some aspects relevant to our topic.Footnote ⁶²

According to the Stoics, our souls receive different types of impressions or presentations (phantasiai) from the world around us, which affect us by revealing themselves and their causes (DL 7.49–51; Aëtius 4.12.1–5). We can evaluate these impressions to either accept or reject them. For example, when I open my bedroom window, I receive the impression that it is daytime. I usually accept this impression as true and form the belief that it is in fact daytime. The Stoics distinguished, however, between different types of assents to impressions. Sometimes, they thought, assent is firm and secure; other times, however, it is given to false or non-cognitive impressions, hastily or in a weaker form than required. In these latter cases, they think we have no knowledge but only opinions.Footnote ⁶³

The Stoics held that only the sage has true knowledge, virtue, tranquillity, and happiness. Moreover, they argued sages are infallible, and their assent to the impressions would always be stable, unimpeded, and perfectly free. And given the way they conceive opinions, the sage must be unopinionated. To guarantee this rather demanding image of the sage, the Stoics held that a sage would only assent to what they call ‘cognitive impressions’ (phantasiai katalēptikai). These were special kinds of impressions, but the translation might be misleading. The Greek word katalēpsis means ‘grasping’, but one of Cicero’s ways of referring to it in Latin was cognitio, where we get the Standard English translation (Acad. 2.17).Footnote ⁶⁴ However, its difference with other impressions is not that they have to do with cognition as we understand the term now, but that they were obviously true, precise, cannot be or become false, and are the basis for knowledge.Footnote ⁶⁵ The way they put it was to say that cognitive impressions arise from ‘what is’, formed in such a way that they could not have come from ‘what is not’, and are so patently true that are their own guarantee; they do not require further justification beyond themselves and, thus, can be the foundation for our knowledge.Footnote ⁶⁶

The Academics granted (perhaps only for the sake of the argument) that the sages should not assent hastily but challenged the existence of cognitive impressions. Then, they argued that the Stoic sage would have to suspend judgement on all matters:

For the Academics, if the Stoics cannot show that there are cognitive impressions, then their normative fallback position should be epochē. In other words, they should have no opinions and make no claims of knowledge.

In this dialectical context, Cicero reports: ‘Arcesilaus used to deny that anything could be known […] He thought that everything was hidden so deeply and that nothing could be discerned or understood. For these reasons, he thought we shouldn’t assert or affirm anything, or approve it with assent: we should always curb our rashness and restrain ourselves from any slip’ (Cic. Acad. post. 1.45; LS68A1, part).Footnote ⁶⁸ It is not easy to know whether Arcesilaus was committed to these views. But, at least in some sources, he seems to hold to an epistemological theory in which suspension is the normative outcome of the arguments (I return to this point in Section 3.2.1).

The Stoics responded in various ways to the Academics, but here I will mention only one reply, illustrated in the aforementioned report about the Stoic Sphaerus (c.285 BC–c.210 BCE):Footnote ⁶⁹

The king presents Sphaerus with an indiscernibility challenge analogous to the Stoic objection using a realistic wax pomegranate. But how are we supposed to understand the story? Scholars have offered two readings. One is to think that Sphaerus concedes he assented to a reasonable impression instead of a cognitive one, as the last sentence suggests. That would mean that Sphaerus had an opinion: he assented to something that could turn out false. But this does not necessarily mean the king has refuted Sphaerus. Sphaerus could still answer that the king mistakenly considered him a sage when he never claimed to be one.Footnote ⁷¹ Later Stoics are quick to acknowledge they are not sages, so it is possible that Sphaerus would take a similar approach.Footnote ⁷²A different reading assumes that Sphaerus considers himself a sage but argues that he assented to the cognitive impression ‘it is reasonable that this is a pomegranate’, which was true. In this case, the king’s challenge fails because it has not shown that the Stoic sage has any opinion.Footnote ⁷³ Maybe a combination of these two readings is correct: Sphaerus did not consider himself a sage, and he assented to a cognitive impression.

The debate over the existence of cognitive impressions and the possibility of universal epochē was kept alive by subsequent generations of Stoics and Academics. But by the time the Academy was under the headship of Philo of Larissa (159/8–84/3 BCE), the Academics had started to argue for more restricted forms of epochē. Perhaps they were still arguing only dialectically. However, Aenesidemus, who may have studied with Philo, famously described the debate of his time as ‘Stoics fighting with Stoics’.Footnote ⁷⁴

I will come back to this debate later (Sections 3.2.1 and 3.2.2), but here I want to highlight that we can see two different ways of conceiving suspension of belief. On the one hand, the Stoic sage arrives at suspension of all beliefs without suspending all their judgements. This is possible because, for the Stoics, opinion is a weak assent to what is false or non-cognitive,Footnote ⁷⁵ which the sage avoids by only assenting to cognitive impressions and thus acquiring infallible knowledge, not mere beliefs. On the other hand, the Academic sceptics remain unpersuaded of the existence of cognitive impressions and claims of knowledge, arguing for universal epochē, even if they are not endorsing the argument or allow weak, pragmatic, or residual commitments to some claims, as I shall show in Sections 3 and 4.

2.1.1 Balance of Evidence

Many philosophers suppose or argue that we ought to suspend our belief when there is a balance of evidence on the question under investigation. I understand ‘balance of evidence’ in a broad sense, including cases where one is confronted with two equally good arguments or comparable empirical observations, and also when there are two equally bad arguments, or the arguments or questions are muddled in such a way that they cause utter confusion. Similarly, insufficient, or total lack of evidence could balance the scales in a way that leave suspension of belief as the only rational alternative.

Ancient philosophers often used a balance of evidence to argue for suspension of belief. Consider Aenesidemus’ ten modes. Each of them points to different contexts in which appearances conflict. The second mode ‘concerns the various natures of men and their idiosyncrasies. […] one man has a passion for medicine, another for farming, and another for commerce. And the same things harm some men and benefit others. Hence one must suspend judgment’ (DL 9.80–81).Footnote ⁸³ If someone asks what the good life is, Aenesidemus will point at various pieces of conflicting evidence. If the outcome is a balance of evidence, or, as he calls it, an equipollence (isostheneia; lit. ‘of equal strength’) where competing possibilities are equally persuasive, the only rational answer will be suspending our beliefs.Footnote ⁸⁴ Reference Annas and BarnesAnnas and Barnes (1985, 24–25) explain the general structure of these arguments like this:

This type of argument has a long tradition. Academic sceptics used them too.Footnote ⁸⁵ They oppose sense perceptions, conceptions, arguments, ideologies, and cultural practices, among other things. But the opposition of perceptions and arguments goes back at least as far as Protagoras.Footnote ⁸⁶ He claims that ‘it is possible to argue every position pro and con with equal plausibility –including the very question whether every position can be argued pro and con’ (Sen. Ep. 88.43).Footnote ⁸⁷ This thesis, however, does not lead him to suspend his beliefs. Instead, he maintains his famous ‘man measure’ doctrine: ‘Of all things the measure is man, of the things that are that they are, of the things that are not that they are not’ (Sext. Emp. Math. 7.60, Pl. Tht. 152b2-3=DK80B1).Footnote ⁸⁸ Plato, Aristotle, and Sextus took this to mean that all opinions and judgements are as they appear to each individual, so that each one of us is the ultimate authority on one’s own judgements. This was complemented by Protagoras’ claim that all opinions are true (for each believer) and that contradicting anyone is impossible.Footnote ⁸⁹ Since this interpretation faces some objections, some scholars have interpreted Protagoras not as a relativist but as a pluralist, a perspectivist or a pragmatist,Footnote ⁹⁰ but my point here remains. For him, the opposition of arguments does not necessarily lead to suspension of belief.

Democritus also opposes perceptions to cast doubt on their reliability:Footnote ⁹¹

Democritus’ argument concludes with a disjunctive: either there is no truth in appearances, or it is unclear to us (he uses the term adēlon, which means ‘unclear’, ‘unknown’, or ‘obscure’).Footnote ⁹³ When I perceive that the apple in front of me is red, either that has no truth value,Footnote ⁹⁴ or, if it has, it is unclear to me whether it is true or false, given that others perceive the same object differently. Either way, it seems that I should not rush my judgement. Could Democritus’ position defend a neutral stance regarding our sense perceptions? It seems so, although it is difficult to be sure.

According to Sextus, Democritus considered the senses a ‘bastard’ or ‘obscure’ (skotios) kind of cognition. Although he is clear that they cannot be a source of knowledge only of opinion, he also recognises that knowledge begins with the senses. In other texts, Democritus claims that perceptual properties are relative and dependent on agreement: ‘By convention colour, by convention sweet, by convention bitter, but in reality atoms and void.’Footnote ⁹⁵ So, it is unclear if he thinks we should dismiss or withhold from forming sensory judgements or simply acknowledge their relative and conventional status.Footnote ⁹⁶ Moreover, in other places, Aristotle attributes to Democritus the thesis that truth lies in appearances,Footnote ⁹⁷ which some find difficult to reconcile with sense perception scepticism. Despite these tensions, it might be possible to fit together Democritus’ claims. If he thought that suspension of belief follows when we consider our sense perceptions alone, that would be compatible with thinking that they are a necessary requirement for knowledge or that they contribute to knowledge when supplied with other sources of information.Footnote ⁹⁸

In Plato’s Protagoras (360d8–362a4) we find another interesting case. After conversing for a while, Socrates tells Protagoras that the discussion has turned against them. As if it had a voice on its own, it mocks them both for changing sides. As a result, matters stand now ‘terribly confused’ (tarattomena deinōs; 361c3). Socrates is willing to continue investigating, but Protagoras uses the conundrum to exit the conversation, promising to revisit the topic at another time. Socrates’ fallback position seems to be suspension of belief.Footnote ⁹⁹

Another possible case of suspension due to unclarity is the Stoic response to the sorites paradox. The paradox is generated by asking to specify a clear boundary between vague terms like bald, few and many: ‘Chrysippus thinks that when one is asked to specify gradually whether, for example, three things are few or many one should come to rest (hēsuchazein, as they put it) a little before one reaches “many”’ (Cic. Acad. 2.93=SVF 2.277). According to Cicero, the recommendation cannot simply be to stop answering but must also include withholding assent.Footnote ¹⁰⁰

As mentioned earlier, Academic sceptics also used balance of evidence to advocate for suspension of belief. Arcesilaus would ask his interlocutors to state and defend their opinions while he would argue against them (Cic. Fin. 2.2). This is how Cicero describes Arcesilaus’ dialectical method: ‘by arguing against everyone’s opinions he drew most people away from their own, so that when reasons of equal weight were found on opposite sides on the same subject, the easiest course was to withhold assent from either side’ (Acad. 1.45).Footnote ¹⁰¹ Later, Arcesilaus’ successor as head of the Academy, Carneades (214–129/8 BCE), would become famous for arguing both sides of a case or question.Footnote ¹⁰² According to some sources, he was sent by Athens as an ambassador to Rome, where he gave a series of lectures on justice, scandalously arguing in its favour one day and against it the next morning.Footnote ¹⁰³

These cases use a balance of evidence either to reach suspension of belief or to conclude that perceptual knowledge is impossible. But for Aristotle, recognising that there is a balance of evidence in certain topics is just the beginning of the philosophical inquiry. In Topics 1.11 (104b3-5), he describes dialectical problems as a point of contention where people ‘either have no opinion, or the public think the opposite of the wise, or the wise think the opposite of the public, or each of these groups has opposed opinions within itself’.Footnote ¹⁰⁴ He later explains that there are also dialectical problems of which ‘there are contrary deductions (for there is a puzzle whether it is so or not, because there are persuasive arguments about both sides), as well as those about which, because they are vast, we have no arguments’ (Top. 104b12-15).

However, Aristotle thinks that we ought not to inquire into every problem, but only those which cause us to be at a loss (aporeō), and we can solve by argument (Top. 105a3-9).Footnote ¹⁰⁵ Aristotle does not say we should suspend belief when we find a balance of evidence. But he thinks that if we are genuinely puzzled, we should examine all sides of the question to make up our minds about it.

In Nicomachean Ethics 9.8, for example, when discussing ‘whether a person should love himself or someone else most of all’ (1168a28-29),Footnote ¹⁰⁶ Aristotle writes: ‘It is quite natural that there is a puzzle (aporeitai) about which view we should follow since both are plausible; presumably, then, we should separate arguments like this from one another and determine how far and in what way those on each side are true’ (1168b10-12).Footnote ¹⁰⁷ These are two views about how to lead one’s life, and without determining which one to follow, we are either stuck, like Buridan’s Ass, or else behaving epistemically irresponsibly.

In Metaphysics Β, Aristotle incorporates the aporiaiFootnote ¹⁰⁸ and the examination of the competing arguments to answer them as a required first part of the inquiry that leaves us in a better position to make a judgement:

Notice that Aristotle’s first step is to raise the puzzles, even if he presupposes that we can solve them with the appropriate analysis. But for a puzzle to be such, all sides must appear convincing and generate an initial impasse.Footnote ¹¹⁰ Moreover, the puzzles might reveal the objective complexity of an issue under investigation.

Aristotle recognises that being aware of a puzzle or genuinely puzzled about an issue might not be enough to move the inquiry forward. So, he adds a conditional: ‘for those who wish to be puzzle-free it is useful to go through the puzzles well’. Otherwise, one might opt for committed neutrality or disregard the puzzle.Footnote ¹¹¹ But Aristotle sides with those who wish to be puzzle-free. At the same time, he recognises that without a balance of evidence there is no puzzle; we would simply select the side with the stronger argument or evidence.

Plotinus employs a similar method, setting up aporiai and trying to solve them (Enn. I.1). But he explicitly acknowledges that some aporiai might not have a solution. He argues that in cases like the aporiai about the soul, even if we do not find a solution, we might benefit from the enquiry by finding out what in this area does not admit of solution:

Dillon and Blumenthal (2015, 166) comment on this passage, ‘It is notable, and typical of Plotinus’ approach to philosophy, that he should recognize that the honest admission that some problems may not admit of solution is a valid conclusion of philosophical investigation.’ This suggests that Plotinus is open to the possibility of agnosticism in some topics: a committed neutral stance that results from the negative results of an inquiry. We can understand this knowledge as follows:

Indeed, this does not necessarily close the door to further investigation and changing one’s mind about the question. But it explains one’s claim of knowledge and sets a higher threshold for engaging in further inquiry.

Suspension of belief also follows when there is little or no evidence, which we can also see as a balance of (lack of) support on both sides of the question. In this case, there is no support, or the evidence is insufficient to decide which side is better. On the question whether the number of stars is odd or even, we have nothing at all to help us choose between one or the other.Footnote ¹¹³ And as for little evidence, consider again Protagoras’ religious agnosticism. The full fragment reads as follows: ‘Concerning the gods, I cannot know whether they exist or whether they do not, or what form they have; for there are many obstacles to knowing, including the obscurity (adēlotēs) of the question and the brevity of human life.’Footnote ¹¹⁴ One way of reading the passage is this: nothing people say about the gods counts as sufficient evidence for or against the existence and nature of the gods, or at least not at first glance. But adjudicating all the views would take too long for a human being to sort out, so Protagoras prefers to suspend his belief and remain neutral.Footnote ¹¹⁵

In Plato’s Statesman, we find another passage where the lack of historical records or information about a topic leads the main character, the Eleatic Stranger, to suspension of belief. The context is the cosmological myth, where the Eleatic Stranger narrates that before current humans, there was another race guided by minor gods and living a life without toils. The question is whether they had a more blessed life than us. The Eleatic Stranger thinks humans are not well equipped to answer this question:

The Eleatic Stranger’s proposal to leave the question to one side results from not having adequate sources. Notice that there is no aporia, no genuine philosophical puzzle. The question would be easily judged if they had access to relevant factual evidence or reliable witnesses. Plato uses the Greek verb aphiēmi (Pol. 272d2), which means ‘to give up’, ‘let go’ and ‘put away’. The rational recommendation is to stop discussing the topic because both options seem equally reasonable in the current context.

2.1.2 Methodological Suspension and Hypothetical Reasoning

Suspension of belief can be part of a philosophical method in multiple ways. In some cases, reaching a genuine suspension due to the refutation of our ideas might be crucial for inquiring about the truth of a given topic. As I showed, part of the aporetic tradition uses suspension in this way (Section 1.3.2). In this section, however, I will discuss another way in which suspension of belief may be used as part of a method of inquiry. That is to avail of a sort of ‘pretend’ and temporal suspension to assess the consequences of an argument or a hypothetical scenario. This suspension of belief is not an involuntary stance or the normative conclusion derived from a balance of evidence. You are not required to suspend your beliefs, but you are invited and allowed to do it (i.e., it is epistemically permissible). The reason for suspension is the hope that it might help us find insight or the truth about a topic.

Consider, for instance, thought experiments. When we use them, we may suspend temporarily and deliberately one or more of our beliefs. When I ponder the consequences of owning a ring that can make people invisible, I suspend my belief that such a magical ring is impossible.Footnote ¹¹⁷ These exercises depend on our capacity to assess counterfactual scenarios as factual and take fictional narratives as schemes with real implications. We pretend not to hold some beliefs to see where that can lead us. Similarly, when we argue and ask someone to assume a hypothesis, we require them to put their beliefs into ‘brackets’ for the sake of the argument.Footnote ¹¹⁸ In these cases, suspension arises from our willingness to consider a hypothetical scenario rationally permitted under certain circumstances and rules and with specific goals in sight.Footnote ¹¹⁹

Rene Descartes’ (1596–1650 CE) hyperbolic doubt is perhaps the most famous and explicit example of this type of methodological suspension. It consists of intentionally withholding assent to all beliefs in things that are not absolutely certain.Footnote ¹²⁰ This suspension is only an intellectual exercise, a steppingstone to finding a foundation for all knowledge. And it is only temporary and ‘pretend’. One can follow the method without genuinely doubting one’s beliefs, and Descartes never assumes he has offered sufficient reasons to withhold his beliefs. Descartes was well acquainted with ancient and modern revivals of scepticism, and his Meditations respond to the challenge they represent.Footnote ¹²¹ The question here, however, is whether ancient philosophers used suspension of belief in a similar way.

Although most ancient philosophers did not explicitly refer to suspension of belief when discussing hypothetical reasoning, they certainly assume that imagination and thought allows us to suspend our beliefs to follow these types of argumentations. Moreover, many of the arguments Descartes uses in his Meditations have clear antecedents in ancient philosophy. So, from a retrospective point of view, we might say that ancient philosophers used reasoning that implies this pretend methodological suspension, even if they do not conceive it in exactly that way. In what follows, I offer some examples of ancient approaches to hypothetical reasoning.

First, we have examples of what we call thought experiments.Footnote ¹²² Many imply a methodological suspension of belief in so far as they set up a hypothetical scenario.Footnote ¹²³ Consider the following fragments from Xenophanes’ poetry:

Xenophanes’ thought experiment requires us to suspend our beliefs about the actual abilities of oxen, horses, and lions. We must put those beliefs into ‘brackets’ to imagine the animals drawing their gods. Even though Xenophanes does not explicitly require us to suspend our beliefs, the analogy would not work if one completely refused to entertain the counterfactual scenario and relate it back to the real world. If someone replied to Xenophanes, ‘What does it matter? These animals cannot draw!’ we would rightly think the person missed the fragments’ point.

We could say the same of other ancient thought experiments. For example, the Pythagorean Archytas of Tarentum (first half of the 4th century BCE) invites us to consider what would happen to someone who, standing at the outermost edge of the universe, extends his hand or his staff. One must disregard any beliefs concerning the practical impossibility of reaching this place to follow the argument.Footnote ¹²⁶ Similarly, when Academic sceptics invite us to imagine what would happen if god presented us with false impressions indistinguishable from true ones, one cannot complain that god would never do that. The hypothetical scenario consists of putting into brackets our beliefs about god and imagining what would happen if he behaved this way.Footnote ¹²⁷

This methodological suspension seems presupposed in other forms of hypothetical reasoning, even if we do not see them as thought experiments.Footnote ¹²⁸ Take, for example, the passage in Plato’s Republic 5, when Socrates is compelled to argue that the sharing of women and children is possible and beneficial:

Socrates must persuade Glaucon and his friends of two theses. The natural order would be to prove the possibility of sharing women and children first and argue for its benefit to society later. But, maybe for rhetorical reasons, Socrates wants to argue for them in reverse order.Footnote ¹³⁰ For that, his interlocutors must be willing temporarily to put into parentheses their current beliefs about the possibility of sharing women and children. Notice that accepting an assumption does not necessarily involve entertaining a counterfactual scenario. We might assume what turns out to be a fact, but we take it as an assumption because we lack sufficient evidence or proof. We suspend our actual beliefs about the matter in question to see where the argument leads.

Moreover, when we accept an assumption, we must for the sake of argument be willing to suspend other beliefs incompatible with it. If we assume that learning the history of philosophy is not a waste of time, and I tell you I spent an evening reading Plato, you cannot reply that I have wasted my time.Footnote ¹³¹ You must suspend your belief for the sake of argument.

Similarly, when we investigate a question and posit a hypothesis, the status of our hypothesis implies that we have not made up our minds about it. This means that, at least in some cases, our default position during the investigation is a genuine suspension of belief. But other times, we might believe the hypothesis is correct even if we lack definitive reasons to accept it. Instead of discarding it, we treat it as if it were true, entertaining a methodological suspension of incompatible beliefs (for hypotheses as criterion for action, see Section 4.1).

Consider the passage in Plato’s Meno where Socrates clarifies what he means by ‘hypothesis’. With regard to virtue, he confesses that ‘since we don’t know (ouk ismen) either what it is or what sort of thing it is, let us first make a hypothesis about it’ (Meno 87b3-4). This acknowledgement of ignorance is not absolute nor prevents them from having opinions about the matter. Socrates believes that virtue relates to the soul and rapidly agrees with Meno on a series of claims that they think are obvious. Then they decide to hypothesise that virtue is a sort of knowledge (87c), and from there, they deduce that if that were true, it follows that virtue is teachable (89c). But Socrates later reveals some reasons to doubt this hypothesis is true. And if those reasons hold up, it will turn out that virtue is not something one can teach (89d–90c). So, we could say that Socrates’ willingness to examine both hypotheses required him to bracket some of his beliefs.Footnote ¹³²

3.2.1 Arcesilaus

The scope of Arcesilaus’ suspension of belief is difficult to establish. He left no writing. His philosophical method involved engaging in Socratic cross-examination and arguing for both sides of a question without revealing his views.Footnote ¹⁴² This suggests we cannot ascribe him any claim preserved in his arguments, and at the same time, we cannot rule out that he might have had some views he revealed to close associates (Sext. Emp. PH 1.234; cf. Cic. Acad. 2.60).

However, almost all sources report not only that he did not reveal his views but that he suspended all judgements. Sextus Empiricus, for example, writes that Arcesilaus ‘is not found making assertions about the reality or unreality of anything, nor does he prefer one thing to another in point of convincingness or lack of convincingness, but he suspends judgement about everything’ (PH 1.232).Footnote ¹⁴³ This description puts Arcesilaus close to Pyrrhonian scepticism. However, Sextus quickly adds that, unlike the Pyrrhonists, Arcesilaus held that epochē was good and assent bad and that some sources say, ‘he appeared to be a Pyrrhonist but in truth was a Dogmatist’ (PH 1.233–234).

Other reports ascribe him an even stronger commitment to universal epochē and other doctrines (again, it is difficult to say if these are reports of his views or assessments that depend on further inferences).Footnote ¹⁴⁴ Cicero, for instance, reports one of his arguments against Zeno as follows:

Here Arcesilaus’ epochē is not the residual state derived from relentless philosophical cross-examination but the conclusion of an argument he is said to have accepted. However, notice that this reasoning does not commit him to epochē unless he also considered himself a sage, which seems unlikely since the notion of sagehood operating here is distinctively Stoic.

Cicero offers a different account in Acad. 1.44–5. In this passage, Arcesilaus’ epochē derives from inapprehensibility (akatalēpsia), the doctrine that nothing can be discerned or understood, and not from an argument with Stoic premises.Footnote ¹⁴⁷ We can reconstruct it as follows:

Instead of a refutation of someone else’s doctrines, this seems to be an argument for scepticism. The first premise is the least controversial. Most people would agree on our sensory and cognitive limitations and the brevity of human life. But the second premise makes a bold claim. Although it is difficult to deny that, in principle, everything is up for debate, why would that mean everything is hidden in darkness? The claim assumes a balance of evidence on every question. Otherwise, one could reply that not everybody’s opinion is worth the same. Maybe the premise depended on witnessing Arcesilaus demolish everybody else’s position and argue both sides of every question.Footnote ¹⁵⁰ If so, the aforementioned would be regarded as Arcesilaus’ closing argument, which concludes with the normative recommendation that one should suspend all judgements.

The different accounts raise two questions. First, to what extent was Arcesilaus committed to universal epochē? And second, what would the scope of such suspension be? If Arcesilaus assented to universal epochē, his scepticism would be self-defeating or paradoxical (he would form the belief that he should not form any belief). Although this result would be welcome by Pyrrhonian sceptics like Sextus, most scholars offer competing interpretations to avoid such a conclusion for Arcesilaus.Footnote ¹⁵¹

Some defend a ‘dialectical reading’ and argue that Arcesilaus was never committed to universal epochē. His views, if any, remain undisclosed. This is an attractive solution but, unfortunately, runs against almost all sources that explicitly attribute universal epochē to Arcesilaus.

Others argue that epochē is restricted to rational belief, which leaves weaker forms of commitment open to Arcesilaus. We can identify at least three options:

• Psychological reactions (i.e., postures, impulses, and desires): involuntary non-rational non-beliefs that are not normative stipulations and can be expressed as commitments that neither are beliefs nor are based on reasons.

• Rational hypothesis: voluntarily rational stances that fall short of belief (i.e., rational non-beliefs) and can be expressed as commitments that are not beliefs but are based on reasons.

• Residual beliefs: voluntary non-rational commitments which are beliefs that are not based on reasons.Footnote ¹⁵²

The first option sees Arcesilaus’ commitment to epochē as an involuntary feeling or impulse. The second understands it as a rationally grounded supposition that falls short of belief. Finally, the last sees epochē as a residual belief from Arcesilaus’ philosophical method, rationally unwarranted but somehow unavoidable. These three options, however, imply attributing to Arcesilaus a ‘type restriction’ to his suspension of belief: rational beliefs. But rational belief here means mental content we can assent to or reject consciously, which suggests that, under these interpretations, Arcesilaus’ epochē requires us to consider the belief in question directly, and leaves unclear whether it prohibits indirect suspension or not.

3.2.2 Carneades

Carneades’ exact position concerning the scope of epochē is also contested.Footnote ¹⁵³ His erudition and talent as a dialectician and controversialist were even greater than Arcesilaus (DL 4.62). But unfortunately, he also wrote nothing. All we know from him comes from his student Clitomachus, whose reports were passed on by Philo of Larissa (159/8–84/3 BCE) to Cicero. But things are still more complicated. Clitomachus and Philo disagreed on how to interpret Carneades, and some later sources offer misleading amalgamations of these two interpretations.Footnote ¹⁵⁴

According to Clitomachus, Carneades argued for inapprehensibility and universal epochē, just like Arcesilaus, but introducing two types of impressions and two types of assent. However, in other passages, Clitomachus depicts Carneades more dialectically, admitting that when listening to him, he could never tell which views he adhered to (Cic. Acad. 2.139).Footnote ¹⁵⁵ Since Carneades’ innovations in this interpretation address the practical implications of suspension, I will discuss them in Section 4. For now, I will focus only on how Philo of Larissa interpreted Carneades.

Philo thought that, unlike Arcesilaus, Carneades accepted that it is sometimes rational to hold opinions. This meant that even the sage could err, but by taking these assents as opinions and nothing more, the sage understands that nothing is apprehensible (Cic. Acad. 2.59).Footnote ¹⁵⁶ Carneades, according to this interpretation, agrees with the Stoics that assent is necessary for action but disagrees with the existence of cognitive impressions. The result is a commitment to an epochē with a more restricted scope than that of Arcesilaus and the Clitomachean interpretation earlier. A summary of Philo’s reading, or something close to it, is preserved by Eusebius:

The last sentence seems to merge elements from the Clitomachean reading, but the main argument seems to follow Philo’s interpretation:

1. i. Everything is inapprehensible (as Arcesilaus has argued; see [3] earlier).

2. ii. But not everything is unclear.

3. iii. Thus, there is a difference between the inapprehensible and the unclear.

4. iv. Humans can only suspend judgement over an issue if and only if it is unclear.

5. v. Therefore, humans cannot suspend judgement over everything (against [4] earlier).Footnote ¹⁵⁸

Scholars typically understand inapprehensibility here as the denial that there are Stoic cognitive impressions that allow knowledge. In other words, all impressions are non-cognitive (noncataleptic).Footnote ¹⁵⁹ If so, the argument implies that sometimes we have a clear impression of things we cannot claim to know for sure. In Philo’s interpretation, there is no distinction between types of assent. Therefore, it is sometimes rational to assent (strongly) to clear yet inapprehensible impressions and form opinions. But since the sage knows the clarity or persuasiveness of the impressions does not guarantee their truth value, the assent is accompanied by the understanding that there is room for being mistaken. Although this epochē is more restricted than the previous versions, it still contrasts with the Stoic commitment to the infallibility of the sage.Footnote ¹⁶⁰

3.2.3 Aenesidemus

A contemporary of Cicero, Aenesidemus criticised Academics who made strong assertions and restricted the scope of their epochē. According to some sources, he defected from the Academy, and took inspiration from Pyrrho to promote his version of scepticism, which later had a huge influence, especially on Sextus Empiricus. Although we only preserve a few fragments and summaries from his works, we attribute to him the ten modes of epochē, the eight modes against causation, and the revival of the Pyrrhonian sceptical tradition.

The sources, however, offer incompatible pictures. On the one hand, they report that Pyrrhonists like him are aporetic, entertaining doubts about every thesis and free from all dogma, determining nothing, not even that they determine nothing or that one thesis is more convincing than unconvincing.Footnote ¹⁶¹ On the other hand, descriptions of his position sometimes suggest that he endorsed some type of relativism or describe him as a full-on negative dogmatist, that is, someone committed to the claim that no statement of a certain type is true or no statement of a certain type is knowledgeable.Footnote ¹⁶² This has given rise to two families of competing interpretations.

According to some, Aenesidemus’ brand of scepticism and scope of epochē is self-cancelling, topic neutral, and universal with respect to content available to the Pyrrhonian inquiry. A revisionist interpretation proposes, in contrast, that Aenesidemus’ arguments do not aim at epochē at all but at aporia and the resulting negative conclusions of the type ‘x is not by nature F’ and allowing relativised claims like ‘x is F for y, at time t, or in circumstance C’.Footnote ¹⁶³ This second interpretation, however, has been criticised. One of the recurring objections is that the evidence does not force it and instead is incompatible with Aenesidemus’ attitude towards the Academics. Moreover, it risks rendering Aenesidemus’ scepticism incoherent.Footnote ¹⁶⁴

3.2.4 Agrippa’s Five Modes

Agrippa’s five modes – the mode of disagreement, regress, relativity, hypothesis, and reciprocity – are among the most powerful resources in the Pyrrhonian sceptical toolbox. Sextus Empiricus uses them extensively but simply attributes them to ‘recent sceptics’ (PH 1.164). However, Diogenes Laertius (9.89) attributes them to an Agrippa (fl. first cent. CE?) from which we know nothing else.

Like those attributed to Aenesidemus, these modes are argumentative devices used by the Pyrrhonian sceptic to guide dogmatic people towards epochē.Footnote ¹⁶⁵ Modern philosophers still debate the merits of the radical epistemological challenge raised by a subset of these modes – the mode of regress, hypothesis, and reciprocity – known in the literature as Agrippa’s trilemma.Footnote ¹⁶⁶ So, it is natural to wonder about the scope of epochē that these modes promise to produce. However, things are not so simple.

We preserve two descriptions of these modes. One in Diogenes Laertius (9.88–89) and the other in Sextus Empiricus (PH 1.164–169). However, there are some significant differences between these reports. The one found in Diogenes Laertius does not mention epochē, gives no indication that the modes can be used together, and describes them as arguments for negative dogmatism.Footnote ¹⁶⁷ In contrast, the version presented by Sextus is explicitly dialectical, aims at epochē, and offers some indication of how the modes might work together. For this reason, here I will only refer to this latter version. Sextus’ text reads as follows:

Sextus adds that every object of investigation can be referred to the five modes (PH 1.169). In other words, the sceptic can contrapose these modes to any dogmatic assertion.Footnote ¹⁷¹ Note, for instance, that the mode of disagreement, which Sextus always puts first,Footnote ¹⁷² applies to both philosophical and everyday matters. The reach of the mode of relativity is also extremely broad; it includes anything that appears to us as a possible object of judgement. Finally, the other three modes target different attempts at justifying one’s beliefs or claims to knowledge.

Given the meagre indications on how the Pyrrhonian sceptic is supposed to use them, scholars have offered different reconstructions, sometimes leaving aside the mode of relativity, disagreement or both.Footnote ¹⁷³ But one way to understand their structure and use is to conceive them as a list of questions or challenges the Pyrrhonian sceptics pose to their dogmatic interlocutors or themselves whenever they feel tempted to make up their minds about something. Although Sextus does not always use them all in the same passage and often applies them in different orders or with modes from other lists, some structural things remain constant (although that only illustrates Sextus’ sceptical practice and has no normative value).Footnote ¹⁷⁴ For example, all the modes can lead to other modes or to epochē, disagreement always comes first, and nothing prevents the sceptic from reapplying the modes over and over, even if this means going into higher orders in the conversation (v.g., a disagreement about a disagreement and so on). An illustration of how a Pyrrhonian sceptic might use the five modes looks like the following interchange:

As this illustration shows, when used together, the scope of application of the five modes has no thematic restriction; they are topic neutral (PH 1.165).Footnote ¹⁷⁵ Moreover, by bouncing from one to the other, the Pyrrhonian sceptic can loop the conversation, never allowing any justification for our beliefs.Footnote ¹⁷⁶

However, it is unclear whether the five modes really aim at a global or universal epochē. First, someone might think these modes are implicitly committed to a normative principle of rationality. However, if they are all dialectical strategies applied by following some simple recipe, there is no need for the sceptic to believe any such principle. The dogmatist would have those commitments, not the sceptic who is only asking the questions and following impressions. But surely, one might reply, the Pyrrhonian sceptics follow certain logical rules and understand when to respond with one mode rather than the other. Could not we accuse them of believing in such things? Not really. Consider that it is possible to program machines to do all sorts of actions, including following instructions in a logical order, without beliefs of any kind.Footnote ¹⁷⁷ This includes a chatbot programmed to respond to our knowledge claims with Agrippa’s modes. Moreover, the Pyrrhonian sceptics could respond by applying the mode of disagreement and then using the other four modes to our accusation.Footnote ¹⁷⁸ Thus, it seems that using the five modes does not imply or commit the sceptic to any belief. But could they deliver universal epochē?

Even if anything can be an object of epochē, it does not follow that the sceptic can suspend all their beliefs without any residue or that they suspend them all at once. The five modes can certainly challenge any attempt to establish a foundation or framework for knowledge and belief formation. But it does not seem that their use can deliver a general conclusion or guarantee no residue (that would require some dogmatic assumptions!). Just as the sceptic can loop the conversation and filibuster our attempts at demonstration, nothing in the five modes stop us from keep trying to offer an acceptable justification without committing ourselves to the idea that the explanation has to go ad infinitum (we might suspend judgement over that), or that the fallback position must always be suspension of judgement and not, for example, belief or disbelief with the proviso that it is open to revision or that it appeals to the best explanation. Moreover, the five modes cannot target irrationally held beliefs. Nor is it useful against the Protagorean relativism that worries Plato and Aristotle, where every opinion is true for the person having it and disagreement is impossible.Footnote ¹⁷⁹