As we saw in the first chapter, one of the elements of UP is antiskepticism. Aristotle’s testimony strongly suggests that Plato was, for virtually his entire career, wedded to the view that knowledge (ἐπιστήμη) is possible and that it is not of sensible but rather of ‘separate’ intelligible entities. It seems a straightforward matter to characterize the Skeptics’ position as the contradictory of the claim that knowledge is possible. But in fact many things that both Plato and Skeptics actually say about knowledge should give us pause. First, in Phaedo Socrates claims that “if it is not possible to know anything purely while we are embodied, either nowhere is knowledge possessed or it is only for the dead.”1 It is, I think, a serious mistake to take ‘know purely’ to imply that there is a type of knowledge that is ‘impure.’ This is the case not merely because it would contradict Aristotle’s testimony. More important, if, as Plato himself argues in Republic, knowledge is of Forms, a putative ‘impure’ knowledge of Forms would be something other than ἐπιστήμη. Not only is there not a single word in the entire Platonic corpus to suggest that there is a mode or cognition of Forms other than ἐπιστήμη (or its equivalent νόησις), but the argument in Theaetetus that true belief cannot be knowledge—even if true belief is ‘supplemented’ by some sort of λόγος—is sufficiently broad in its scope that there is no room left over for a mode of cognition of Forms that is not ἐπιστήμη and not belief.2 The words “know purely,” then, should be understood as suggesting that the ne plus ultra of cognition alone is not available to embodied individuals.3
And yet, not too much after this passage we have the Recollection Argument for the immortality of the soul, an argument the principle conclusion of which is that we do in some sense have ἐπιστήμη; for if we did not, we could not make the judgments about the relevant deficiency of instances of Forms in the sensible world. It seems evident that the ἐπιστήμη we perhaps cannot possess while embodied is different in some way from the ἐπιστήμη we must possess if we are to be able to make judgments like “these equal things are deficiently equal.” The relevant distinction is made in Theaetetus. This is the distinction between “possessing” (κεκτήσθαι) knowledge and “having” (τὸ ἔχειν) it.4 This is a distinction between the presence in the knower of that which is knowable (‘possessing’) and the awareness of the presence (‘having’). Deploying this distinction in the argument in Phaedo, we would say that we must ‘possess’ knowledge in order to make judgments about the deficiencies of sensibles, whereas ‘having’ knowledge is definitely problematic for embodied individuals.
If Plato’s antiskepticism regards only the having of knowledge, it remains an open question as to whether he concedes the Skeptics’ claim with regard to the impossibility of possessing knowledge. The account in Republic of the education of the rulers culminating in knowledge via a vision of the Idea of the Good at fifty years of age only slightly mitigates the pessimism of Phaedo.5 At most, this knowledge is possible only for the elite few and then only near the end of their lives.
For many scholars, the temptation to discount Plato’s epistemic rigorism is considerable. One way of doing this is to insist that ἐπιστήμη or a kind of ἐπιστήμη is possible for sensibles as well as for intelligibles.6 Apart from the fact that there is virtually nothing in the texts of Plato to support this interpretation, to make of Plato a proponent of what we might call ‘empirical knowledge’ is, I believe, to ally him with Stoic epistemology in a way that makes the Academic skeptical opposition to that incomprehensible. Another way of doing this is to stipulate that knowledge is (or should be) for Plato something other than a mode of cognition whose sole objects are Forms. So we can just call rational belief (δόξα) ‘knowledge’ if it meets some arbitrarily concocted criterion or even if it does not. Whether the inclination to do this arises from a rejection of Plato’s view on how we possess knowledge or whether it arises from a rejection of the difficulty or impossibility of embodied persons having knowledge is irrelevant. In either case, the view that when in Theaetetus Plato seeks to define ἐπιστήμη he is in fact searching for a stipulative definition of the word ἐπιστήμη or offering an analysis of the concept of ἐπιστήμη as opposed to dialectically constructing a definition of the real thing that ἐπιστήμη is seems to me frankly unbelievable. More to the present point, it makes nonsense of the construction of Platonism within the Academy including its skeptical ‘phase,’ as we will see presently.
With respect to UP, there is a latent tension between the element of antiskepticism and the positive claim that the objects of knowledge are other than sensibles. The fact that the highest mode of cognition is reserved for intelligibles is owing to the instability or relative unintelligibility of the sensible world. Then, an appropriate mode of cognition for the latter is thrown into question, at least insofar as that cognition is supposed to attain truth.7 If the attainment of truth is possible for this mode of cognition, why is it not knowledge or a kind of knowledge? If it is not possible, what is it that the inferior mode does attain? Underlying this question is more than a terminological issue. For the hierarchical metaphysics and epistemology of Platonism is not purchased at the cost of the dismissal of the sensible world as completely unintelligible and of a mode of cognition for it that is conducive, if not to the attainment of embodied knowledge, at least to our psychological and ethical advancement. Plato’s rejection of the extreme nominalism of the Eleatics brought with it a sort of commitment to the rehabilitation of the intelligibility of the sensible world. But this commitment is fraught with difficulties.
The bestowal of the honorific “founder of skepticism” was probably given to Pyrrho of Elis (c. 360–c. 270) by Aenesidemus, the first-century BCE Academic Skeptic.8 It is not impossible that the supposed founder of Academic skepticism, Arcesilaus (316/5–241/40), knew of Pyrrho, though we have no indication of this. Sextus Empiricus, at any rate, is quite emphatic that, while Arcesilaus shares the true skeptical ‘approach’ (ἀγωγή), his successors, Carneades (214–129/8) and Clitomachus (187/6–110/09), do not.9 The question I wish to pose is what insight if any the skepticism of Arcesilaus and (pace Sextus) Carneades provide for our understanding of Platonism. It would, of course, be a mistake to suppose that the very flexible label ‘Academic’ need indicate anything but the most tenuous tie to Plato. Indeed, a standard scholarly view is that the Academic Skeptics represent a wholesale abandonment of Platonism and that only with Antiochus is Platonism reunited with the Academy. Still, it is worth pursuing the question of whether a skeptical ‘procedure’ or a skeptical approach to knowledge has a Platonic provenance.10
First, it is certain that Academic skepticism is aimed squarely at a Stoic account of knowledge. Here is Sextus Empiricus’s admirably lucid presentation of that account followed by Arcesilaus’s criticism:
For they [the Stoics] hold that three things are linked to each other: knowledge, belief, and, placed between these, grasping. Of these knowledge is sure and stable grasping unalterable by reasoning; belief is weak and false assent; and grasping is what is between these, assent to a graspable presentation. According to the Stoics, a graspable presentation is true and such that there could not be a false one just like it. They say that knowledge is present only in the wise, belief is present only in base men, but that grasping is common to both groups, and that this is the criterion of truth.11
Here is how Arcesilaus apparently responded to this account:
These being the Stoics’ views, Arcesilaus countered them by showing that grasping is in no respect a criterion midway between knowledge and belief. For that which they call grasping and assent to a graspable presentation occurs either in a wise man or in a base man. But if it occurs in a wise man, it is knowledge, and if in a base man, it is belief, and there is nothing else left besides these two but a name.12
The difference between the Stoic account of knowledge that is criticized here and the Platonic account jumps out in the first line. For the Stoics, knowledge, grasping, and belief are “linked” (συζυγοῦσα). The link is provided by the “graspable presentation” that is available to both wise man and fool. This is a mode of cognition that arises primarily from sense perception.13 What differentiates the wise man from his opposite is that the former has “sure and stable grasping unalterable by reasoning,” whereas the latter has weak belief and gives false assent. Thus, there is both knowledge and belief about the same presentations to sense perception. For Plato, however, knowledge and belief are discontinuous; there is no knowledge of that of which there is belief and vice versa. Further, for the Stoics what differentiates the wise and the base men is the false assent that the latter gives to his presentations. This is not equivalent to an assent to a false proposition; false assent is possible to a true proposition. What makes the assent false is that it is unjustified. By contrast, the wise man never gives assent to a proposition that is unjustified. That is why he and he alone has knowledge. So it seems fairly clear that for the Stoic, knowledge is true justified belief or, what amounts to the same thing, a mode of cognition in relation to a true proposition that differs from belief only in that it is “unalterable by reasoning.” Even if, as we have already seen, the passage in Plato’s Meno in which knowledge and belief are joined might be rather implausibly interpreted along these lines, the account in Republic and Theaetetus effectively eliminates this possibility. Knowledge is not for Plato justified belief because knowledge does not have any sorts of propositions as its objects, but rather the Forms themselves.14
The astute criticism of Arcesilaus aims to show that there is in fact no middle ground between a mere belief and knowledge. The putative ‘grasping’ is either knowledge or belief. If it is mere belief it is unjustified, even if the belief is true. If it is knowledge, then a justification is otiose since knowledge is, as defined by the Stoics, unalterable by reasoning. If a so-called justified belief is had by someone, that belief is in fact knowledge; if what one has is not knowledge, it is unjustified. For this reason, the attack on the possibility of knowledge is a fortiori an attack on the possibility of rational or justified belief. A would-be defender of the Stoic account of knowledge is not in a position to concede the impossibility of knowledge as defined by Stoics and instead move to defend rational belief as an almost-as-good alternative. The core skeptical strategy is not to show the impossibility of a rarified form of cognition, thereby leaving something else in its stead. Rather, that strategy eliminates the possibility of rational belief by eliminating the possibility of knowledge. But this conclusion is entirely rooted in the Stoic account that defines knowledge and belief as two forms of the grasping of one and the same proposition. Even a true belief—true because, according to the criterion of truth, the presentation is grasped—is not rational because it is not knowledge.
It is evident that the Stoics’ vulnerability to the Academic critique turns on their insistence that knowledge is “unalterable by reasoning,” that is, that it is in some sense infallible or incorrigible.15 A desire to be faithful to Plato, who also held that knowledge is in some sense infallible (ἀναμάρτητον, ἀψευδής), can hardly be the reason for the Stoics’ claim. It is more likely that the Stoics in fact reasoned that the ne plus ultra of cognitive achievement must be infallible. For the idea that knowledge is fallible (“I know but I may be mistaken”) conflates knowledge with belief. And even if one adds that what one knows must be true, knowledge will not differ from a true belief, which, in turn, does not differ from a lucky guess. But if knowledge is conflated with a lucky guess or an adventitious true belief, then one does not know that one knows. And if this is the case, the knowledge as the supreme achievement of the philosopher becomes something hollow, to say the least. But a true belief differs from a lucky guess only if the former is justified or based on sufficient evidence. Either the putative justification guarantees the truth of the belief, in which case there is knowledge, or else it does not. So suppose a justified true belief is knowledge. It is not enough, however, that there exists some justification for the belief, and that the justification entails that the belief is true. One must also know that the justification or evidence does this. But, then, one must either have justification or evidence for the evidence or else the evidence must be self-evident. That is, one must know that one knows that the evidence entails what it is evidence for.16 On this analysis, one cannot know unless one knows that one knows. That is, knowing is knowing that one knows. So there is no knowing without evidence (including the limiting case of self-evidence) or else one knows that the evidence entails what it is evidence for. But without entailing evidence, there is mere belief, as the Stoics saw, not a justified belief, which would just be knowledge. Stated otherwise, there is only justification when there is knowledge; without justification there is only belief. Worse, the belief is without justification, which is as much as to say that it is irrational. There is literally no more reason to believe a proposition than the opposite.
The Stoics agree that there is a mode of cognition that is better than ungrounded belief. This is knowledge. There is no middle ground between ungrounded belief and knowledge. So the stakes are high in defending against the attack on the possibility of knowledge.17 This defense amounts to showing that it is possible to be in a mental state that is “unalterable by reasoning.” Obviously, one cannot hope to attain such a state merely by cultivating a preternatural stubbornness. Since the state of knowing is one in which the knower assents to the truth of a proposition, a proposition that represents the (graspable) presentation, the knower must maintain that no reasoning could alter the inference that is made from the presentation to the assent to the proposition.18 After all, what makes belief the métier of the fool is that his assent is alterable by reasoning, whether cogent or not.19 Yet the claim that the graspable presentation is, for the knower, such that there could not be a false one like it, must be an inference. For otherwise there would be no difference between the assent of the sage and the assent of the fool to the same graspable presentation. The sage’s mental state is “unalterable by reasoning” because he is convinced that his presentation is sufficient evidence for assent to the proposition representing that presentation.
As Skeptics argued, however, there can be no presentation that is absolutely indistinguishable from one that in fact does not entail the truth of a proposition representing it for the simple reason that the presentation is false:
Among presentations, some are true and some are false. A false presentation is not graspable [percipi non potest]. But every true presentation is such that a false one of the same sort can occur. And where presentations are such that there is no difference between them, it cannot occur that some of them are graspable and some are not. Therefore, no presentation is graspable.20
This is the so-called argument from illusion. Its true provenance is, once again, Stoic epistemology. The force of the argument is usually represented as consisting in the observation that the phenomenology of illusions is indistinguishable from that of veridical perception. So the straight stick in the water looks like a bent stick, which is indistinguishable from what a bent stick really looks like. I do not believe, however, that this is here the central focus. In reply to this argument, the Stoics are said to have added that the graspable presentation must be one that could not come from something that does not exist. So a true presentation is stipulated to be different in some way from a false one. The straight stick that looks bent must presumably be different from the one that is really bent. The point could not be that straight sticks and bent sticks are in reality different; no Skeptic need deny this. The point must be that they are phenomenally different. But this will not do. For the Stoic is committed to an inference from a proposition regarding the way things appear (the graspable presentation) to the truth of a proposition regarding how things really are. It is irrelevant that it is true that things appear a certain way to the knower. Yet it is also question-begging and destructive of the inference to claim that the way things appear is the way they are, for then there is no inference to a proposition about the way they are. In order to create ‘inferential space,’ there must be some criterion according to which the graspable presentation could, but does not necessarily, provide a basis for the truth of the proposition representing this. The argument from illusion is not an argument about whether we are or are not sharp enough to tell illusions from reality; rather, it is an argument that seeks to show that a claim to knowledge, as the Stoics have it, must have a justificatory or evidentiary basis. Such a basis could exist only if the justification or the evidence entails what it is evidence for. But entailment is a logical relation, not a psychological one. That the Stoic sage is more circumspect in his assent than is the fool is beside the point. The problem is that there is no legitimate inference from the way things appear to the way things are even if there is a trivially true inference from the way things epistemically appear, that is, truth-preserving appearance, to the way they are. And this is so because there is no legitimate inference from the way things appear to the truth of a proposition that they are appearing epistemically. If, though, the sage forgoes justification, his claim to know does not differ in weight from the claim of one who merely believes. But if the sage accepts a justificatory burden, in what sense is his mental state “unalterable by reasoning”?
So Arcesilaus urges suspension of judgment regarding all presentations. That is, for any p, no belief that p is the case is more rational than a belief that not-p. In this conclusion, it is not inapposite to compare the sort of argument that Socrates makes in, say, Euthyphro, namely, that if we do not know what the Form of Piety is, we cannot give a λόγος of it, and if we cannot give a λόγος of it, then we are in no position to claim with any justification whatsoever that a deed is pious or not. The only basis for such a claim would be that we could see that the deed bore the hallmark of Piety. But this assumes that we know Piety.21 The point is easily generalizable for any claim about any property in the sensible world whose intelligibility rests ultimately on a separate Form. Plato, like the Academic Skeptic, is not only opposed to the idea that there is such a thing as empirical knowledge, but he is also opposed to the idea that there is such a thing as rational belief unless there is nonempirical knowledge. It is only such knowledge that makes rational belief possible or intelligible. This is the case because only knowledge could provide the justification or the evidence for the belief. In Platonic terms, if Euthyphro is to believe rationally that it is true that prosecuting his father is pious, his justification for this claim is that he knows Piety.
But the rational belief, of course, regards an object that is not the object of knowledge, but only that which participates or imitates the knowable. I am not claiming that Arcesilaus’s attack on the Stoic account of knowledge is made from the vantage point of a Platonic account of knowledge, particularly a Platonic account that assigns the possession of knowledge to a disembodied soul. What I do wish to claim is that the logic of the Skeptic attack on the rationality of belief (the major Skeptic claim), given the impossibility of knowledge as the Stoics conceive of it, is identical to the Platonic attack on the rationality of belief absent at least the possession of knowledge. That Plato believes that knowledge is possible and that he believes that he has proven in the Recollection Argument that without knowledge one cannot make certain judgments about the sensible world that in fact are made leaves open the possibility that Arcesilaus’s attack on Stoic knowledge is compatible with the acceptance of Plato’s claims.22
Sextus reports that Arcesilaus was alive to the criticism that a Skeptic who withheld assent to all propositional claims could not conduct ordinary life without contradicting himself. For in acting, the Skeptic—implicitly or not—must assent to some propositions, namely, those the presumed truth of which makes his behavior rational. Arcesilaus is reported to have held
that he who suspends judgment about everything regulates choices and avoidances and, generally, actions by reasonableness, and proceeding according to this criterion, will act correctly. For happiness arises because of prudence, and prudence resides in correct actions, and a correct action is that which, having been done, has a reasonable defense. Therefore, he who adheres to reasonableness will act correctly and will be happy.23
It was a commonplace even in antiquity that Arcesilaus’s acknowledgment of the possibility of reasonable behavior undercut his principled skepticism. It seems likely that Aenesidemus’s effort to revive a pure, uncompromised ‘Pyrrhonian’ skepticism was aimed precisely at such apparent backsliding.24 We ought to realize, though, that the proposed skeptical criterion of action is not a criterion of truth. A course of action is not rendered less reasonable just because it fails to achieve its goal. That which is reasonable could not be a criterion of truth since ‘the reasonable’ is not used to justify a claim to truth as if it were a kind of evidence. If this is so, in what sense is ‘the reasonable’ a criterion?
In order to answer this question, it will be useful to have before us the famous passage from Plato’s Timaeus in which the principal interlocutor, Timaeus, declares that physical science can only be a “likely story” (εἰκός μῦθος):
Again, these things being so, our world must necessarily be a likeness of something. Now in every matter it is of great moment to start at the right point in accordance with the nature of the subject. Concerning a likeness, then, and its model we must make this distinction: an account is akin to that of which it is an account: an account of that which is stable and secure and discoverable by thought will itself be stable and incontrovertible (and insofar as this is possible and it belongs to an account to be irrefutable and unchangeable one ought not to fall short of this); while an account of that which is made in the image of the other but is only a likeness, will itself be only likely, standing to accounts of the former kind in a proportion: as being is to becoming so is truth to persuasion. If, then, Socrates, in many respects concerning many things—the gods and the generation of the universe—we prove unable to give accounts that are in every way consistent with themselves and exact, you shouldn’t be surprised. If we can provide accounts no less likely than others, we must be content, remembering that I who speak and you my judges are only human, and consequently it is fitting that we should, in these matters, accept the likely story and look for nothing further. (trans. Cornford, slightly altered)25
This passage in a way provides a gloss on the Divided Line in Republic. There we learn that what is believed stands to what is known as the image stands to what is imaged.26 Here, what we learn is that “a likely story” is all that we will attain in regard to the things that become, that is, the objects of physics.
What is particularly interesting about this passage for the current discussion is the use of the word “truth” (ἀλήθεια). There is no doubt that this is the ontological use of the term, that is, a property of being in relation to an intellect.27 At the same time, Plato can qualify the noun δόξα with the adjective ἀληθής, using the term to indicate a semantic property of propositions and derivatively a mode of cognition, δόξα, in relation to these.28 Beliefs can be true, but belief does not attain the truth, we might say, because belief has the realm of becoming, not being, for its objects. The distinction between the ontological and semantic notions of truth does not undercut the dependence of the latter on the former, according to Plato. True belief is possible because of the eternal truth that the objects of true belief are likenesses or images of.29 So the fact that only a ‘likely story’ regarding physics is possible does not mean that there are no true propositions in that story or that they do not differ from false ones. But the only possible justification for the claim that a belief is true is the knowledge of Forms or the existence of ontological truth.30
Returning to Arcesilaus, the claim made by the Skeptic is that there is a criterion of reasonable action, but embracing this presumes no claim to have a justification for any beliefs related to the action.31 What, then, does the ‘reasonable defense’ of the action imply? Unfortunately, Sextus does not record any answer by Arcesilaus to this question. We may speculate, however, that a reasonable defense of an action would consist in or would at least include a statement that happiness was thus obtained. That is, the defense is pragmatic. It is reasonable to act one way or another because in so acting one attains the absence of anxiety or contentment that is, for the Skeptic, the substance of happiness. Since happiness thus conceived is a state of the soul, as opposed to, say, an activity, as it is according to Aristotle, the Skeptic’s ‘reasonable defense’ amounts only to a report on his own subjectivity or ‘what seems to him.’ If this is what Arcesilaus means, he is cleverly exploiting the ambiguity of the meaning of ‘appearances’ (τὰ φαινόμενα) between an epistemic and a nonepistemic sense. To state that “I seem to be content” in the former sense implies a justification consisting in a reason for saying that the appearance is epistemic. In the latter sense, nothing like this is implied. Arcesilaus renounces any interest in whether his appearances are (semantically) true. By contrast, one who took his appearances as epistemic would be open to the same objections that are raised against Protagoras in Theaetetus. Arcesilaus’s skepticism implies a rejection of a dogmatic relativism along with the dogmatism of Stoicism. At the same time, it leaves Plato’s account of knowledge untouched.32
The illustrious successor to Arcesilaus, Carneades, offers a valuable chapter in the history of those struggling to come to terms with Platonic epistemology.33 Sextus tells us that Carneades was no Skeptic, despite any claims to the contrary.34 Sextus reports that Carneades, like Arcesilaus, rejected a criterion of truth. But like Arcesilaus, he offered a criterion
for the conduct of life and for the attainment of happiness . . . introducing the persuasive presentation [πιθανὴν φαντασίαν], and the presentation that is at the same time persuasive, uncontroverted [ἀπερίσπαστον] and thoroughly tested [διεξωδευμένην].35
Apart from Carneades’ substitution of τὸ πιθανόν for Arcesilaus’s τὸ εὔλογον, why does Sextus think that Carneades’ criterion is a betrayal of skepticism? The answer to this question is revealed in Sextus’s observation that Carneades takes the persuasive presentation to be an “apparently true” one (ἡ φαινομένη ἀληθὴς φαντασία).36 As such, he goes beyond Arcesilaus in implicitly claiming to be able to distinguish epistemic and nonepistemic appearances. But practically the whole point of skepticism is the principled rejection of this possibility. That Carneades has abandoned the skeptical stance is evident in his grading of presentations according to whether they are merely persuasive or also uncontroverted, and finally, thoroughly tested. The gradation implies an endpoint or ideal such that when it is reached, the apparently true will coincide with the true.37 In short, the apparently true presentation will be an epistemic appearance. The Skeptic must insist that there cannot be such presentations in principle, which is exactly what Carneades denies with his refined standards of empirical confirmation. For the Skeptic, any and all appearances are nonentailing, that is, there is no entailment from a proposition that represents an appearance to a proposition that represents the truth. Yet this is exactly what Carneades implies when he states that “for the most part” our appearances are “truth revealing” (ἀληθευούσῃ ).38
It is of no use for Carneades to counter the skeptical argument by conceding that the persuasive presentation is not a criterion of knowledge, but only of rational belief. For as we have already seen, rational belief requires that knowledge be at least possible, knowledge the proximity to which makes a belief rational. This proximity has nothing to do with truth and everything to do with evidence or justification. But alas, as Plato himself insists, δόξα has no part in justification.39 A belief may well be true but it is not thereby rendered rational. Nor is its rationality sequentially fortifiable by being arrived at according to the application of Carneadean evidentiary criteria.
As with Arcesilaus, we can only guess at Carneades’ view of Plato’s account of knowledge. His rejection of Stoic epistemology in favor of a calculus of rational belief is not obviously incompatible with the Platonic position if we imagine that our ‘possession’ of knowledge is itself the underlying criterion for the rationalization of belief.40 Such rationalized belief seems not at all unlike what physics is supposed to be in Timaeus. This is admittedly speculative. Yet the Academic argument against the Stoic criterion of knowledge should have resulted in the unqualified rejection of rational belief, as it apparently does for Arcesilaus. Did Carneades think that such radical skepticism was avoidable given that Platonic knowledge is a human endowment?41
One fairly slight piece of evidence in favor of this interpretation is that the Academic Philo of Larissa (158–84 BCE), pupil of Clitomachus, explicitly capitulated on the matter of whether knowledge must be infallible. The little evidence we have for Philo’s view suggests that he summarily dismissed the Stoic criterion of truth as a ‘graspable presentation,’ that is, as one that (a) comes from what is; and (b) is stamped and impressed (in us) in accordance with what is; and (c) in such a way that it could not come from what is not.42 The third clause, said by Sextus to have been added precisely to guard against skeptical arguments, is what Philo attacked. Such an attack is what a Skeptic like Sextus would endorse doing. Yet Philo’s innovation seems to have been to argue that the elimination of (c) does not erase knowledge altogether but only infallible knowledge.43 Thus, an appearance can be epistemic even if it might not have been. The fact that any presentation may be false does not entail that we never attain knowledge from our presentations. The idea of fallible knowledge (“I know but I may be mistaken”) really is a departure from anything authentically Platonic. Insofar as this view is an innovation, we may infer that it was not one that Carneades or Arcesilaus shared. As we will see in the next chapter, it was this innovative deviation from Platonism by Philo that inspired Antiochus of Ascalon to attempt a reinvigoration of the authentic Platonic position.
The Stoics, the principal target of Academic skepticism, adhered to the Platonic claim that knowledge is irreversible by reasoning. The Academics did not challenge this view before Philo of Larissa. The dispute between Arcesilaus and Carneades regarding the existence of a criterion of truth reveals not their rejection of the possibility of knowledge, but their rejection of the possibility of knowledge of the sensible world. This, of course, was the only knowledge that the Stoics thought possible. In their denial of the latter possibility, the Academic Skeptics displayed their Platonic bona fides. It is not until Sextus Empiricus or perhaps his unknown source or perhaps even Aenesidemus that a more comprehensive challenge to the possibility of infallible knowledge tout court was mounted.44 This challenge is aimed at the heart of Platonism far more than anything attributable to the Academy. The reason for this is that Sextus saw that the claim to the possibility of infallible knowledge depended on the existence of an immaterial intellect. If there were no such thing, the putative knowledge of intelligible reality could only be some sort of representation. But just as the Academic Skeptics had no difficulty in showing against the Stoics that representations of sensible reality could not be infallible, so the self-described Pyrrhonians had no difficulty in showing that if intellect is material, then all one could conceivably aspire to would be a representation of immaterial reality, and as such infallibility was impossible.
The proponent of UP cannot forego the possibility of infallible cognition, at least for the Demiurge. The triad of Idea of the Good-Forms-Intellect is, as we have already seen, the core of the positive construct built on the foundation of UP. If the Demiurge cognized only representations of Forms—whatever these might be—he would, among other things, not eternally ‘attain’ the ontological truth that the Idea of the Good provides to the Forms. This state of affairs would import an adventitiousness or, better, haphazardness into the generation of the cosmos that is unthinkable given the comprehensive worldview of Timaeus. More fundamentally, it would place a limitation on the goodness of the first principle of all. That is, the Good could not give ontological truth to the Demiurge, but only to the Forms. As a result, the Demiurge’s desire to make the cosmos like the Forms would be hollow. For the supposed intelligibility imposed on the receptacle would not amount to images of Forms, but only images of the Demiurge’s representations of Forms, thus making the Demiurge more like the earthly craftsman than the maker of heaven and earth.45
As Academic skepticism shows, adherence to the idea that the ne plus ultra of cognition is infallible reveals a new problem about the possibility of rational belief, a problem that the doctrine of ἀνάμνησις is supposed to solve. If we do already know the Forms, then we have the basis at least for a justification for our claims about sensibles. But in order to prove that ἀνάμνησις in the relevant sense exists, one must demonstrate the immortality of the soul. This is the burden of the Recollection Argument in Phaedo. Further, the idea that being a knower is, so to speak, our natural or true state, and that it is the state of the disembodied soul, underlies the ethical dimension of Platonism. Philosophy is a preparation for death, as Socrates says in Phaedo, because the state of an accomplished or successful lover of wisdom is the state to which we aspire, or ought to aspire. This is exactly the case also for the exhortation to assimilate oneself to the divine in Theaetetus.46 This exhortation was to become emblematic of Plato’s ethics for later Platonism. Thus, the immortality and immateriality of the soul and its capacity for the highest form of cognition, that which is possessed by the Demiurge, is a critical part of the Platonic construct. The antiskepticism and the antimaterialism of UP, whatever its precise origin as a challenge to one or another of Plato’s predecessors may have been, was the starting point for the distinctive epistemology at the heart of Platonism.47
The shared conviction of Stoics and Academic Skeptics that knowledge must be infallible also reveals the core problem in the attempts to hive off a ‘nondogmatic’ Socratic philosophy from Platonism. ‘Prudential judgment’ or ‘elenctic knowledge’ or even rational belief, taken to be the main idea in Socrates’ intellectualism, is, as Arcesilaus shows, hollow because there is no justification short of complete and ‘unoverturnable’ justification. Each and every one of Socrates’ claims regarding what he firmly believes or even knows is controvertible. Either Socrates gives no reason for his beliefs—beliefs like “one must absolutely never do wrong”—or his reasons are never conclusive, such as his reasons for thinking that it would be wrong for him to escape from prison. Yet the reasonableness of these claims, if indeed it be such, is restored when we set this putative Socratic intellectualism within the Platonic context. For Plato wants to maintain that recollection or even the midwifery of Socrates in Theaetetus presupposes the possession of knowledge, that is, infallible knowledge of Forms, on the basis of which rational λόγοι about the sensible world, including human life, may be produced.
1. Phd. 66E4–6: εἰ γὰρ μὴ οἷόν τε μετὰ τοῦ σώματος μηδὲν καθαρῶς γνῶναι, δυοῖν θάτερον, ἢ οὐδαμοῦ ἔστιν κτήσασθαι τὸ ἐἰδέναι ἢ τελευτήσασιν. Cf. D7–E2, 65E1–4, 67A2–6. In the context of the passage, it is clear that the words used here for ‘knowledge’ (γνῶναι, ἐἰδέναι) are being used for the ‘highest’ form of cognition, knowledge or wisdom (σοφία).
2. See Gerson 2009, 44–55. The mode of cognition that Plato calls διάνοια (‘understanding’ or ‘thought’) in Republic is clearly not the required tertium quid. It is not a kind of ἐπιστήμη. Cf. 533C7–E2.
3. Cicero, De nat. deo. 1.11, refers to Socrates as the originator of the idea of refraining from judgment and Arcesilaus as having revived it (repetita). This implicitly drives a wedge between Socrates and Plato, but only if we identify Socrates not with the historical figure, but, rather arbitrarily, with the figure in the ‘aporetic’ dialogues alone. Shields (1994, 343–45) argues that Arcesilaus, “by extending certain Socratic practices” in what Shields assumes to be the early dialogues, can be understood not to have completely misunderstood Plato. I do not agree that Arcesilaus’s possible Platonic bona fides requires accepting an arbitrarily selected group of dialogues (‘early’) as containing a distinctly skeptical position.
4. See Tht. 197B–D. Aristotle employs basically the same distinction, using his invented term ἐνέργεια. He makes a distinction between first and second ‘actuality’ in cognition.
5. See Rep. 540Aff. It is clear from the text of the analogy of the Divided Line (511B–D, esp. C8–D2) that it is only by seeing the Idea of the Good that the philosopher is able to have knowledge of Forms.
6. See Fine 2003, where this position is developed in many papers, especially in one devoted to knowledge and belief in Plato’s Republic. See also Burnyeat 1990.
7. See Tht. 186C9–10: Οὗ δὲ ἀληθείας τἰς ἀτυχήσει, ποτὲ τούτου ἐπιστήμων ἔσται; (If someone cannot hit upon the truth of something, will he then have knowledge of it?) It does not follow logically from this claim that if one does hit upon the truth, then one has knowledge. For it seems that one may hit upon the truth adventitiously, as it were, without knowing it. But to assume that this is what Plato meant to allow is to gainsay his hierarchical metaphysics such that truth—whether in the intelligible or sensible realms—is one thing. As I will argue, this view constitutes a misreading of Plato’s Platonism because it undermines the possibility of a consistent positive construct on the basis of UP.
8. Two of Aenesidemus’s works are Pyrrhonian Discourses and Outline Introduction to Pyrrhonian Matters. See Photius, Bib. 212. But see D.L. 9.70, where the medical doctor Theodosius is said in his book Skeptical Chapter to deny that Pyrrhonism should be identified as skepticism for the fittingly skeptical reason that we cannot know what it was that Pyrrho taught.
9. See Sextus, PH 1.220–35.
10. See Krämer 1972, 14–107, who finds a number of elements in the later works of Plato and in the works of the members of the Old Academy apt for skeptical interpretation. See Sedley 1996, 98, on the anonymous commentary on Plato’s Theaetetus that argues that the members of the New Academy “virtually all endorsed dogmatic Platonism.” Sedley (84) concurs with Tarrant in dating this commentary to the late first century BCE.
11. Sextus, M. 7.151.1–153.1: τρία γὰρ εἶναί φασιν ἐκεῖνοι τὰ συζυγοῦντα ἀλλήλοις, ἐπιστήμην καἰ δόξαν καἰ τὴν ἐν μεθορίῳ τούτων τεταγμένην κατάληψιν, ὧν ἐπιστήμην μὲν εἶναι τὴν ἀσφαλῆ καἰ βεβαίαν καἰ ἀμετάθετον ὑπὸ λόγου κατάληψιν, δόξαν δὲ τὴν ἀσθενῆ καἰ ψευδῆ συγκατάθεσιν, κατάληψιν δὲ τὴν μεταξὐ τούτων, ἥτις ἐστἰ καταληπτικῆς φαντασίας συγκατάθεσις· καταληπτικὴ δὲ φαντασία κατὰ τούτους ἐτύγχανεν ἡ ἀληθὴς καἰ τοιαύτη οἵα οὐκ ἄν γέ νοιτο ψευδής. ὧν τὴν <μὲν> ἐπιστήμην ἐν μόνοις ὑφίστασθαι λέγουσι τοῖς σοφοῖς, τὴν δὲ δόξαν ἐν μόνοις τοῖς φαύλοις, τὴν δὲ κατάληψιν κοινὴν ἀμφοτέρων εἶναι, καἰ ταύτην κριτήριον ἀληθείας καθεστάναι. Cf. Stobaeus, 2.73.19 = SVF 1.68–9; Cicero, Acad. 1.41 = SVF 1.60. The term μεθόριος does not indicate a third possibility (either knowledge or belief or something between these) but rather the ‘boundary’ between knowledge and belief, that is, something that they both share. See Gerson 2009, 100–111, for an account of Stoic epistemology.
12. Sextus, M. 7.153.1–154.1: ταῦτα δὴ λεγόντων τῶν ἀπό τῆς Στοᾶς ὁ ’Aρκεσίλαος ἀντικαθίστατο, δεικνὐς ὅτι οὐδέν ἐστι μεταξὐ ἐπιστήμης καἰ δόξης κριτήριον ἡ κατάληψις. αὅτη γὰρ ἥν φασι κατάληψιν καἰ καταληπτικῆς φαντασίας συγκατάθεσιν, ἤτοι ἐν σοφῷ ἤ ἐν φαύλῳ γίνεται. ἀλλ’ ἐάν τε ἐν σοφῷ γένηται, ἐπιστήμη ἐστίν, ἐάν τε ἐν φαύλῳ, δόξα, καἰ οὐδὲν ἄλλο παρὰ ταῦτα ἤ μόνον ὄνομα μετείληπται.
13. There can also be presentations from incorporeal and nonevident corporeals like god. See D.L. 7.49–52 = SVF 2.52 (in part).
14. Scott (1995, 215) claims that in Meno, although Socrates argues for recollection, there is no commitment to the theory of Forms “or to the other assumptions that support the argument in the Phaedo.” I take this claim to be contradicted by Aristotle’s testimony.
15. Aristotle, Top. Ε 2, 130b15–16; Ε 4, 133b28ff.; Ε 4, 134b16–17; Ζ 8, 146b1–2, gives “ὑπόληψις incontrovertible by λόγος” as a definition of knowledge. It is not clear whether this was the accepted Academic definition or not. But in any case, it was certainly open to the Stoics to reject it in favor of something more modest had they wished to do so.
16. Cf. Cicero, Acad. 2.145 = SVF 1.66. In this passage, Cicero quotes the Stoics as maintaining that the one who grasps (but only believes) is like one who ‘grasps’ something with the fist, whereas the one who knows is like one who grasps that fist with the other fist. This seems to me best interpreted as a claim that knowing entails and is entailed by knowing that one knows.
17. The Stoics generally conceded that there were no living sages, that is, no one who actually had knowledge. Although the existence of such a sage would show that knowledge is possible, the actual nonexistence of one would not entail that knowledge is not possible.
18. On assent being to a proposition, see Stobaeus 2.88.4 = SVF 3.171: “Propositions are the objects of acts of assent.” Perhaps assent to the proposition is extensionally equivalent to assent to the graspable presentation.
19. Cf. Tim. 51E4, where true belief is distinguished from knowledge (here, νοῦς) by the fact that the former is overturnable by persuasion (μεταπειστόν) whereas the latter is not. The fact that the Stoics adhere to a Platonic distinction in the face of Academic criticism is not without significance. One might have guessed that an embrace of materialism would have led Stoics to abandon the claim that knowledge is infallible precisely because infallibility would be applicable to cognition of immaterial entities only by immaterial entities.
20. Cicero, Acad. 2.40; cf. Sextus, M. 7.402–10.
21. So Eu. 11A6–B1. Cf. Men. 71B1–8, where knowledge of the Form of Virtue is necessary both for knowledge of its properties—like teachability—and for ‘knowing’ whether something instantiates it.
22. It is difficult to know quite what to make of the consistent ancient evidence that Arcesilaus regarded himself as a disciple of Plato. See D.L. 4.33; Sextus, PH 1.232; Plutarch, Adv. Col. 1121F–1122A. The difficulty is principally whether Arcesilaus’s supposed adherence to Plato refers to an aporetic or dialectical method (cf. Cicero, Acad. 1.46) or to substantive doctrines. It seems not implausible that Arcesilaus would be correct in claiming a sort of fidelity to Plato insofar as he held that there is no ἐπιστήμη of the sensible world. That there might be more to his fidelity to Plato than this is suggested by Cicero, Acad. 2.60, which may or may not be the passage Augustine is thinking of at C. Acad. 3.43. Gigon (1944, 1972) stresses the Stoics as primary Academic targets in support of his argument endorsing the texts that claim that the Academic Skeptics did not reject Plato’s “esoteric teaching.” Glucker (1978, 296–306), criticizing Gigon, identifies this supposed “esoteric teaching” with the theory of Forms, and rightly complains that this theory is hardly concealed in the dialogues. But the esotericism perhaps has more to do with the positive construct of UP, including the important addition that Plato could well be interpreted to have held that at least occurrent ἐπιστήμη is not available to embodied individuals.
23. Sextus, M. 7.158.5–12: ὅτι ὁ περἰ πάντων ἐπέχων κανονιεῖ τὰς αἱρέσεις καἰ φυγὰς καἰ κοινῶς τὰς πράξεις τῷ εὐλόγῳ, κατὰ τοῦτό τε προερχόμενος τὸ κριτήριον κατορθώσει· τὴν μὲν γὰρ εὐδαιμονίαν περιγίνεσθαι διὰ τῆς φρονήσεως, τὴν δὲ φρόνησιν κεῖσθαι ἐν τοῖς κατορθώμασιν, τὸ δὲ κατόρθωμα εἶναι ὅπερ πραχθὲν εὔλογον ἔχει τὴν ἀπολογίαν. ὁ προσέ χων οὗν τῷ εὐλόγῳ κατορθώσει καἰ εὐδαιμονήσει. The “reasonable defense” is an obvious reaction to the Stoic doctrine of κατορθώματα, which, supposedly based on knowledge, have such a defense.
24. See Burnyeat 1980, 27–31.
25. Tim. 29B1–D3: τούτων δὲ ὑπαρχόντων αὖ πᾶσα ἀνὰγκη τόνδε τὸν κόσμον εἰκόνα τινός εἶναι. μὲ γιστον δὴ παντὸς ἄρξασθαι κατὰ φύσιν ἀρχήν. ὧδε οὖν περί τε εἰκόνος καἰ περἰ τοῦ παραδείγματος αὐτῆς διοριστεόν, ὡς ἄρα τοὐς λόγους, ὧνπέρ ἐισιν ἐξηγηταί, τούτων αὐτῶν καἰ συγγενεῖς ὄντας· τοῦ μὲν οὖν μονίμου καἰ βεβαiόυ καἰ μετὰ νοῦ καταφανοῦς μονίμους καἰ ἀμεταπτώτους—καθ’ ὅσον ἷόν τε καἰ ἀνελέγκτοις προσήκει λόγοις εἶναι καἰ ἀνικήτοις, τούτου δεῖ μηδὲν ἐλλείπειν—τοὐς δὲ τοῦ πρός μὲν ἐκεῖνο ἀπεικασθέντος, ὄντος δὲ εἰκόνος εἰκότας ἀνὰ λόγον τε ἐκείνων ὄντας· ὅτιπερ πρός γένεσιν οὐσία, τοῦτο πρός πίστιν ἀλήθεια. ἐὰν οὖν, ὦ Σώκρατες, πολλὰ πολλῶν πέρι, θεῶν καἰ τῆς τοῦ παντὸς γενέσεως, μὴ δυνατοἰ γιγνώμεθα πάντῃ πάντως αὐτοὐς έαυτοῖς ὁμολογουμένους λόγους καἰ ἀπηκριβωμένους ἀποδοῦναι, μὴ θαυμάσῃς· ἀλλ’ ἐὰν ἄρα μηδενός ἧττον παρεχώμεθα ἐικότας, ἀγαπᾶν χρή, μεμνημένους ὡς ὁ λέγων ἐγὼ ὑμεῖς τε οἱ κριταἰ φύσιν ἀνθρωπίνην ἔχομεν, ὥστε περἰ τούτων τὸν ἐἰκότα μῦθον ἀποδεχομένους πρέπει τούτου μηδὲν ἔτι πέρα ζητεῖν.
26. Rep. 510A9. Here δόξα encompasses both πίστις and ἐἰκασία. But see Tim. 52A5, where δόξα (with αἴσθησις) seems equivalent to πίστις in 29C3.
27. Cf. Rep. 508E1, where it is the Idea of the Good that gives ‘truth’ to the Forms.
28. See Tim. 37B9, 51D6; Johansen 2004, 48–68.
29. Cf. Parm. 135B5–C2, where Parmenides warns Socrates that if the problems with Forms cannot be resolved, thus relegating them to the realm of the impossible, then the ability to have intelligent conversation (τοῦ διαλέγεσθαι δύναμιν) will be completely destroyed. Kahn (1996, 297) takes the ability to refer to philosophical dialectic, not to ordinary conversation. But if for Plato dialectic is about Forms, the consequence of giving up Forms is, of course, that the philosophical study of them will be impossible. It seems implausible that Parmenides is saying what is trivially true; rather, he is best taken to be making the extremely portentous and controversial claim that without Forms ordinary conversation turns into nonsensical noise.
30. See Tim. 28A2, 28C1, 52A7, where δόξα is acquired μετ’ ἀἰσθήσεως. There is no suggestion here that sense perception provides the justification for, as opposed to the source of, belief. Indeed, at 28A3, sense perception is characterized as ἄλογος. How could this ever be the justification for anything?
31. Sextus, M. 7.150, says that Arcesilaus did not define a criterion “in the proper sense” (προηγουμένως). I take it that this is intended to imply that a criterion of action is not a criterion in the proper sense because it is not a criterion of truth.
32. See Glucker 1978, 296–306, for some acute observations about the tantalizingly ambiguous evidence regarding the existence of ‘an esoteric Platonism’ within the skeptical Academy.
33. Carneades was actually about a hundred years younger than Arcesilaus. We are informed that between the Academic headship of Arcesilaus and that of Carneades there were a number of heads of lesser distinction including Lacydes, Telecles, Evander, and Hegesinus. Virtually nothing is known about these philosophers, including whether or not they embraced a form of skepticism.
34. See Sextus, PH 1.226–32, who, for this reason, distinguished the ‘Middle Academy’ of Arcesilaus from the ‘New Academy’ of Carneades.
35. M. 7.166.
36. M. 7.174.
37. See Tarrant 1985, 13–21, whose ingenious thesis it is that Carneades was arguing for the coincidence of the seeming true with the true for universal judgments, not particular ones. These universal judgments required the possession of ‘common notions’ or concepts that, on Tarrant’s view, were a sort of simulacrum for Platonic Forms.
38. Sextus, M. 7.175. It may be objected on Carneades’ behalf that the presentation that is ‘truth revealing’ is not a proposition and so is not supposed to entail another proposition representing the truth. But the claim that the apparently bent stick really is bent is first a claim about how things appear to someone. Given this, there must be some inference to the claim that things are as they appear. The only way to avoid the inferential claim is to identify the appearance with reality, which is presumably what a consistent Protagoras would have done. But, of course, that amounts to a form of dogmatism, too, because it is a claim about how things really are.
39. See Tim. 51D5–E4, where δόξα—true δόξα—is ἄλογον.
40. That Carneades was primarily focused on rebutting Stoic epistemology seems clear from his well-known quip (quoted by Diogenes Laertius, 4.62), “If it was not for Chrysippus, I would not be.” And yet Sextus, M. 7.159, says that Carneades aimed his arguments against the criterion of truth of all of his predecessors, including, it would seem, Plato.
41. Cf. Cicero, TD 1.57, for the claim (certainly not Stoic, but perhaps intended as Academic) that our embodied conceptualization would not be possible if we did not possess previous disembodied cognition of Forms. Also, see Karamanolis 2006, 48, n. 11, who cites Acad. 1.13, 2.11–12, 18, and Sextus, PH 1.235, as suggesting that the Academics did not “denounce” knowledge in the non-Stoic sense.
42. See Sextus, M. 7.248.1–4: καταληπτικὴ δέ ἐστιν ἡ ἀπὸ ὑπάρχοντος καἰ κατ’ αὐτὸ τὸ ὑπάρχον ἐναπομεμαγμένη καἰ ἐναπεσφραγισμένη, ὁποία οὐκ ἄν γένοιτο ἀπό μὴ ὑπάρχοντος. Cf. D.L. 7.46; Cicero, Acad. 2.18.
43. See Barnes 1989, 71–74; Brittain 2001, chap. 3.
44. See Gerson 2009, 129–33.
45. The debilitating result of having only representations of Forms in the intellect and not the Forms themselves only intensifies if Forms are Numbers and their representations are only images of these.
46. See Tht. 176A8–B3: διὸ καἰ πειρᾶσθαι χρὴ ἐνθένδε ἐκεῖσε φεύγειν ὅτι τάχιστα. φυγὴ δὲ ὁμοίωσις θεῷ κατὰ τὸ δυνατόν· ὁμοίωσις δὲ δίκαιον καἰ ὅσιον μετὰ φρονήσεως γενέσθαι. (For this reason, it is necessary to try to flee from here to there as quickly as possible. This flight is the assimilation to the divine as much as possible. And this assimilation is to become just and pious with accompanying wisdom.)
47. See Tim. 52C–E, where antimaterialism follows from the distinction of knowledge and true belief, since knowledge must be exclusively of immaterial entities.