Plato and the Norms of Thought

Abstract

This paper argues for the presence in Plato’s work of a conception of thinking central to which is what I call the Transparency View. According to this view, in order for a subject to think of a given object, the subject must represent that object just as it is, without inaccuracy or distortion. I examine the ways in which this conception influences Plato’s epistemology and metaphysics and explore some ramifications for contemporary views about mental content.

1. Introduction

I would like to discuss some aspects of Plato’s conception of thinking, and in particular to investigate what he supposes is required for a subject to be thinking of a given object. The topic falls, in modern parlance, under the rubric of intentionality,¹ though to avoid prejudice to either Platonic or modern conceptions I shall largely steer clear of that term.

I shall argue that Plato upholds a demanding view of what such thinking requires, which I label the ‘transparency view’. In brief, that stands for the following necessary condition: if a thought is to be of a given object, that object must be represented by the subject just as it is, complete and accurate, without omission or distortion. To give an intuitive sense of what this means: if in attempting to think of Socrates, who has (let us say) blue eyes, I represent someone with green eyes, or eyes of no determinate colour, then I am not thinking of Socrates; if, in attempting to think of a pig, pigs having curly tails, I represent a creature with a straight tail, or with no tail, or I simply fail to represent it as having a tail, then I am not thinking of a pig.

The case for this reading proceeds as follows. I begin, in section two below, with the account of thinking as internal speech given at Theaetetus 189–90. Section three outlines some contemporary views, by way of motivating Plato’s conception. Sections four to six offer a detailed analysis of Theaetetus 208–9, where we find the most direct statement of what I am calling the transparency view.

There are other key Platonic texts on the relation between mind and world that reference to the transparency view may help illuminate. I discuss from this perspective the puzzle about knowledge and false judgement at Theaetetus 188 (Sect. 7); the Meno’s paradox of enquiry and theory of recollection (Sect. 8); the theory of recollection in the Phaedo (Sects 9–10); and Forms versus particulars as objects of cognition in the Republic and elsewhere (Sects 11–14).

In section fifteen, I conclude with a brief assessment of what claim the transparency view ought to have on us. By arguing that, and how, the view features in Plato’s work, my aim is to uncover a neglected but important aspect of his discussions of the mental. I hope thereby to clarify the view as a philosophical thesis and encourage reflection on both its contrast with, and challenge to, certain contemporary views about mental content.

2. Plato on thinking

Our starting-point is the famous account of thinking (dianoeisthai) at Theaetetus 189e–190a, where Socrates describes it as speech (logos) which the soul has with itself.² Thinking is portrayed as a rather careful process. The soul ‘questions and answers itself’ on the topics under consideration before making a judgement (doxa), the latter also characterized as silent speech.

Now the context in which this account of thinking occurs is Socrates’ discussion of the difficulties of explaining how false judgement is possible. The account is embedded in, and apparently intended to serve, the argument that it is impossible to think of a thing (if one is to be thinking of it at all) as anything other than the thing that it is, which causes the so-called ‘other-judging’ theory of false judgement to run into trouble. I cannot falsely judge yonder cow to be a horse, given that I am ‘grasping both [cow and horse] with my soul’ (190c6-7), since I need to be grasping a cow for it to be the case that my mistake is about a cow, and a horse for the mistake to be that of identifying the former as a horse. Without both items in mind, Socrates and Theaetetus agree, the possibility of mistake does not arise within a theory that seeks to explain falsehood in terms of judging that one thing is another (190d8–11). But if making a judgement is silent speech, then the mistake would involve the absurdity of my formulating the statement ‘That cow is a horse’ (190c2–3).

Why should that be? What presupposition underlies this model such that the account of falsehood as other-judging runs aground? The portrayal of thinking as internal speech is attractive enough for Plato to have it reiterated by a different protagonist (the Eleatic Stranger) at Sophist 263e–264a, and it would be incautious to assume that he repudiates the inference that Socrates draws from it in the Theaetetus, that the only way to grasp or have a thought of a cow or a horse is in effect to think of them with just those qualities that cows or horses distinctively have.³ For he evidently supposes that to think of them is already to have identified each as it is; so if mistaking cow for horse is to mean reporting to oneself that the one is the other, the absurdity is generated by the items purportedly being identified having already been picked out as distinct. To get to think of a given object is by this measure a test, and one that it is not hard to fail. Merely representing a horse or a cow as a large domesticated quadruped will not count, for example. If it did, there would be no puzzle about how one could be confused with the other.

My use of ‘transparent’ to signify the representation of an object just as it is stands in deliberate contrast with the sense that term has when featuring in modern semantic distinctions between opaque and transparent, or de dicto and de re.⁴ In asking why Plato should hold that a subject’s thinking that that cow is a horse is equivalent to the subject’s stating ‘that cow is a horse’, one is asking why the thought should be regarded, in the modern sense, as opaque rather than transparent, the latter holding of the statement, about a cow, that ‘that is a horse’, which is not in any obvious way absurd.⁵ But if to think or speak of a cow is already to represent a cow with just those features that cows distinctively have, transparently in my sense, and so too for thinking of a horse, then the idea of thinking of the cow in question as a horse is indeed tantamount to stating ‘that cow is a horse’.

Perhaps this suggests that the transparency view (hereafter TV) is obviously wrong, and that a distinction along the lines of the modern contrast between the transparent and the opaque, unavailable to Plato (as we may agree that it was), can be adduced to solve the problem. But reflection on the discussion of other-judging indicates a rather different point: that the semantic distinction between transparent and opaque is not basic, but presupposes some notion, weaker than TV, of what the requirements are for thinking or speaking of a given object. If so, commentators who suppose that an application of the distinction will straightforwardly dissolve Socrates’ critique of other-judging have stopped at the wrong place.⁶ Plato, rightly it seems to me, is focused on the more fundamental question of what it is for something to be an object of one’s thought or speech.⁷ Whether the semantic distinction between opaque and transparent can be usefully deployed depends on how one answers that prior question.

In the Sophist, the reiterated account of thinking comes in the wake of that work’s advances in the understanding of falsehood. Though the relation between the Theaetetus and Sophist (the latter written as sequel to the former) is not a topic to discuss in detail here,⁸ one may see the Sophist as taking itself to have succeeded in neutralising the second of the Theaetetus’ puzzles about, or failures to explain, falsehood (falsehood as not-being). But it is silent on the latter’s more psychologically oriented puzzles (knowing/not-knowing and other-judging).

The Sophist offers to show, via its analysis of not-being, that ‘Theaetetus flies’, for example, is not empty of content, and so is a genuine, truth-evaluable thought. It therefore removes one important block to a successful analysis of false judgement and in so doing offers support for the model of thinking as internal speech. It hopes to have demonstrated that ‘Theaetetus flies’ should not count as a statement about nothing merely because it states what is not. Indeed one can hardly overplay the Sophist’s achievement in distinguishing ‘what a statement is about and what gets said about it in the statement’ (Frede 1992, p. 416).

Yet the Sophist has no pretension to move on from drawing that vital distinction to laying down the conditions for what makes ‘Theaetetus flies’ or indeed ‘Theaetetus sits’ a statement, or a thought, about Theaetetus.⁹ That topic is allotted to the Theaetetus itself. The Sophist’s achievement is logical rather than psychological. Whatever problems may be left over from the Theaetetus in the psychological domain are not within the Sophist’s provenance to address. (This is in part signalled by the presence of the Eleatic Stranger as protagonist rather than Socrates.) As has been well remarked, in the Sophist ‘there is no explicit substitute for the Tht. view of what it is to think about an object’ (White 1976, p. 193, n. 39).¹⁰ That view, encapsulated (so I believe) by TV, stands as the mature expression in Plato’s work of the relation between thought and object. I shall have more to say in section four about its presence in the Theaetetus.

3. Thinking, ancient and modern

On something like a modern or ‘folk’ view of the matter, I take it that there will not be a great difference between attempting to think of a given object and succeeding. Thus, in contrast to what I consider to be Plato’s conception, though one may talk of success, there is not, generally, a notion of failure to compare it with. If you ask me to try to think of an elephant, for example, and I affirm sincerely that I am thinking of an elephant, then that is generally taken as sufficient for it to be the case that I am thinking of an elephant. I might of course not be familiar with the term ‘elephant’ — but then it makes little sense to say that I was even trying to think of one.

One may likewise doubt that the thought could so much as be attempted if I had only a limited familiarity with the term, perhaps merely aware that it referred to some animal or other. Alternatively, I might have some peculiar mental block whereby I am quite familiar with the term but am simply unable, say, to form an image of an elephant in my mind. Perhaps here at any rate one could be said to be trying to do so. That this would indeed be an odd situation shows, I think, that there is little gap on this picture between the notion of trying to think of a given object and that of succeeding. In standard cases, it will be regarded as sufficient for my getting to think of a thing that I simply put my mind to thinking of it.

This is connected with the similarly modern idea that the subject has the authority to determine that she is thinking of the given object.¹¹ The subject, that is, has the authority to determine that the attempt to think of the object succeeded. The matter will be unaffected by the fact that, for example, what I might conjure up when asked to think of an elephant is an image of a blurry grey lump with little or nothing in the way of determinate distinguishing features. So long as I the subject am, as it were, satisfied that it is an elephant that I am thinking of, then it is an elephant that I am thinking of.

Certainly there are cases where it seems we may want to suspend the rule that if a subject sincerely affirms he is thinking of an object, then he is thinking of that object. For example, if I affirm sincerely that I am thinking of Barcelona but in fact am conjuring up an image which, when I describe it, reveals characteristics that seem actually to belong to Madrid, we would perhaps want to say that if it is any city I am thinking of, it is Madrid, not Barcelona. Yet it still seems more plausible to say, not that I scored a failing grade in my attempt to think of Barcelona, such that my lack of accurate representation of it is the grounds for denial that I was thinking of it, but that I was either mislabelling Madrid as Barcelona or misremembering as Barcelona the city, Madrid, that I was thinking of.

This sort of case, then, does not imply folk recognition of norms of accuracy to be applied to the content of a thought in order to determine whether it is a thought of the given object. After all, as in the elephant case, a subject’s conjuring up a vague image of Barcelona that did not even mark it out distinctively as that city would not in itself be any bar to the truth of the proposition that the subject was thinking of Barcelona, so long as she affirmed that she were.¹²

I have moved rather blithely, in this section and earlier, between particulars such as Barcelona or Socrates as objects of thought and items such as ‘an elephant’ or ‘a pig’, which do not (or not uncontentiously) refer to particular entities in the world. Fictional or mythical objects such as Pegasus would also fall under that category. Thoughts involving such items are, one might say, thoughts with contents without their being thoughts about objects. So perhaps one should exclude these kinds of case from the ambit of thoughts about objects, and restrict the scope of our discussion of the conditions of such thoughts accordingly.¹³

On the other hand, it seems somewhat arbitrary to release this sort of case from considerations of accuracy altogether. One may disagree about whether thinking of ‘an elephant’ is to be regarded as a case of thinking of an object; I err on the side of generosity with the scope of ‘object’ so as not to prejudge the discussion in favour of one ontology rather than another. In any case, intuitively there are more or less accurate ways to be thinking of ‘an elephant’ as there are more or less accurate ways to be to be thinking of Jumbo the (particular) elephant. Perhaps the applicability of standards of representational accuracy might itself serve as a criterion of objecthood. For a proponent of TV, there is correspondingly no reason to exclude such cases from its strictures, even if what counts as a representation worthy of being a thought of the object in question will presumably differ between the different types of case: with Jumbo, it will be getting all his features right, with ‘an elephant’ those features characteristic of elephants as a species. So too with the case of a mythical object such as Pegasus: one had better be representing a horse with wings rather than (say) a pig with propellers if one is to be thinking of that mythical creature.

There is, of course, a well-known contemporary debate about mental (intentional) states, according to the ‘externalist’ side of which, whose paternity is generally agreed to be shared by Putnam and Burge,¹⁴ the content of such states is determined in part by facts independent of the subject, for example social facts or facts about the environment. There will then prima facie be a good deal more than the authority of the individual subject that determines what the subject’s thought is about.

One of the most vigorous aspects of this debate has been whether externalism threatens some familiar conceptions of authoritative self-knowledge; and if it does, whether that is a problem for externalism or for those conceptions. I shall not attempt to trace the contours of the debate here.¹⁵ But it does seem uncontroversial to say that externalism is not intended to deny that I can have thoughts about a certain item even if my representing of it is flawed. Thus, for an individual who has some false beliefs about, say, arthritis (as in Burge’s famous thought-experiment), those beliefs can be about arthritis, because the subject belongs to a linguistic community which has that concept rather than one which corresponds fully to the subject’s beliefs.

Indeed a worry that has been raised in attributing instead the concept of ‘tharthritis’ to the subject (one that would correspond to her beliefs) is that the existence of false belief is then ruled out¹⁶ — what the subject will have is not false beliefs about arthritis but true beliefs about tharthritis. Here we are close to the sort of consideration that lies behind some of Plato’s own problems with false belief. If one holds that a subject does not possess the concept of arthritis unless she captures that condition accurately, putative false beliefs about the condition will disable the claim that arthritis was the concept possessed.

In practice, externalists generally share with their opponents a presumption that, although expert knowledge has a crucial role to play in individuating content, one is not required to possess an expert’s knowledge of a thing oneself in order to have thoughts about that thing. This relaxation in turn gets much of its purchase from the idea that language is fundamentally a shared, public phenomenon, such that individual speakers can, as it were, borrow their cognitive credentials from other members of the community in which they participate (one may contrast here Plato’s notion of thought as internal speech).¹⁷

Thus, only an expert metallurgist can distinguish (say) aluminium from molybdenum, but according to Putnam’s division-of-linguistic-labour hypothesis, this will not prevent a speaker who lacks such knowledge from referring to (or by extension having thoughts about) aluminium.¹⁸ Rather, the existence within the community of such knowledge is what enables the reference to be secured. Likewise, on a variant of that well-known planet Twin Earth, where we are to imagine that the relative scarcity of those metals is the reverse of what it is on Earth, it is regarded as common ground between externalist and critic that the ordinary inhabitant incapable of distinguishing the two ‘succeeds in talking about molybdenum, since this is how those in the know on Twin Earth would take his words’ (Crane 1991, p. 12).

Putnam himself puts the point thus: ‘I can refer to gold … perfectly well without being able to distinguish gold from non-gold because there are others in the community — experts — upon whom I can rely’ (Putnam 1995, p. xvi). The question presses, though: once we grant that the subject is talking or thinking about x, not in virtue of her own shaky grasp, but because of the availability of the expert conception, have we not conceded that only the expert is really talking or thinking about x? When an Olympic victor from a small country declares that her victory was a victory for all her compatriots, a hard-nosed response would be that, actually, it was a victory for her. Non-expert members of a linguistic community may bask in reflected glory; but cognitive success, like its Olympic counterpart, belongs to those who have attained the standards necessary to achieve it.¹⁹

4. Thinking in the Theaetetus

Plato, no sentimentalist about such matters, does not offer reliance on vicarious expertise as a means of granting the non-expert subject successful thought. At its strictest, his conception of what it is to be thinking of an object is that one must, oneself, be representing it just as it is. I have termed this the Transparency View (or TV). To revisit our earlier example: if Socrates has blue eyes, and what I represent in my thought is someone with green eyes, or eyes of no determinate colour, then I am not thinking of Socrates.

Why not? To get more closely to grips with this idea, let us examine the passage in Plato where TV is most clearly stated, Theaetetus 209b–c. Socrates is mounting a critique of the third and final version of the proposal that knowledge is true judgement with an account, ‘account’ (logos) now construed as the expression of a thing’s distinguishing mark. Socrates argues, against this version, that Theaetetus will not even be an object of one’s judgement unless one has such a mark in mind, and for this one would need to grasp Theaetetus’ distinctive snub-nosedness ‘and in the same manner the other things from which you [Theaetetus] are composed’ (kai talla houtō ex hōn ei su, c8–9).

In my example, I suggested that one violates TV if one represents someone with blue eyes as having green eyes. That remains true, but now it transpires that if one thinks, correctly enough, of Theaetetus as snub-nosed, let alone as simply having a nose, one would no more be thinking of him than of Socrates, who is also snub-nosed (209b2–c3). To be thinking of Theaetetus, such that he can be an object of one’s judgement, one must get at least his distinctive snubness in mind (c5–8). As Theodorus had helpfully pointed out at the start of the dialogue, Theaetetus’ snub nose is less pronounced than Socrates’ (143e8–144a1). Yet even with this on board, Socrates goes on to say that for Theaetetus to be an object of one’s judgement one must have in mind his other features in like manner too. Each one of his parts or ‘limbs’ (cf. hen hekaston tōn melōn, 209b5–6) must be represented in a way that distinguishes them, as far as may be the case, from those of others. Unless I have in mind Theaetetus in all his particularity, I do not have Theaetetus in mind. That is the stance that lies at the heart of TV.

Socrates actually speaks of distinguishing Theaetetus’ snubness, and so on, from other examples that he has ‘seen’ (heōraka, 209c7). But while this acknowledges practicalities (Why care if he cannot tell Theaetetus from someone he has never met?) the qualification does not, I think, weaken the requirement for a thing’s being an object of one’s judgement. Socrates may, after all, have acquaintances who (like Socrates and Theaetetus themselves) look like each other, perhaps twins who are harder to tell apart from one another than from anyone else, known or unknown. That may be why the immediately surrounding context lacks the qualification, Socrates speaking simply of differentiating one thing from ‘the rest’ (tōn allōn, 209a7, d5–6, 8, 9–10); and why it was also laid down that Theaetetus needs to be distinguished from the ‘furthest Mysian’ (209b7–8), not a person Socrates is likely to encounter, indeed not a concrete individual at all. So the idea seems to be that in order to get a thing in mind one must represent it with sufficient precision to distinguish it from all other things whatsoever; and this is how the notion of a distinguishing mark was introduced, as that by which one thing differs from ‘everything’ (tōn hapantōn, 208c8).²⁰

The language of the passage suggests there is at least one extra element involved in getting a thing in mind. Its terminology recalls, no doubt deliberately, the Wax Tablet model that was presented earlier in the dialogue, with the idea of an object’s features being ‘stamped’ on the subject (cf. ensēmēnamenē, 209c8 with e.g. 191d8). It seems natural to read this as meaning that the cause of the stamping must be the object itself. So there is a causal condition for getting x in mind as well as a qualitative one²¹: I must represent the qualities of x accurately, this condition indicated by the shift from nose to snub nose and then to Theaetetus’ distinctive snub nose (and other features); and the representation must have been brought about by x itself.²²

Thus we have a pair of necessary conditions: an object will not feature in my judgement ‘before’ (prin, 209c6) the subject represents the object’s particular features, as stamped by the object on the subject’s mind. These conditions are apparently also jointly sufficient: when Theaetetus’ distinctive snubness (along with his other features) is so stamped, it will cause the subject to form true judgements about him (poiēsei ortha doxazein peri sou, c10).²³

There is reason to think that for Plato causal and qualitative conditions are tightly connected, which may explain why the language of stamping is resuscitated at the point where TV is introduced. Given a standard Platonic assumption that what is F cannot be the cause of what is not-F (e.g. Phaedo 101a–b, cf. Theaetetus 199d), it will be hard to maintain that a certain representation can be caused by Theaetetus if it misidentifies his features (e.g. as pale when he is dark, blue-eyed when he is green-eyed, and so on). Accurate representation of x will be a necessary condition for x being the cause of that representation; the causal condition delivers the qualitative one. Any inaccuracy would not have x but some other factor, presumably in the environment or the soul of the subject, as its cause. In terms of the analogy, wax that, say, by being too soft leaves a distorted impression of Theaetetus, is to be regarded as causally responsible for that outcome, and not Theaetetus, even if the latter was a necessary condition for the impression’s having been made; whereas in the case of accurate representation the wax has, as it were, simply allowed the object to exercise its causal powers.

5. Thought and belief

There is a causal condition also operative in the dialogue’s famous ‘jury’ passage, which had argued that knowledge is different from true judgement or belief. The courtroom lawyer’s persuasive speech is said, in a phrase similar to that later found at 209c10, to cause the jury to judge the truth (cf. doxazein poiountes, 201a9; doxasai … poiēsai, b6), whereas it is only an eye-witness who can have knowledge. The jury, that is, lack a necessary condition for knowledge that the eye-witness possesses, which given the causal vocabulary and the later echo is plausibly construed as being that knowledge must be caused by its proper object, in this instance the events in question.

The echo of the jury in the exposition of 209c indicates at the same time a key difference between the two passages. By 209c, the stamping of a thing’s features on one’s mind is needed for one to have so much as true belief about that thing. The jury passage implicitly takes this as a requirement only for knowledge. A distinction is thereby suggested between getting a thing in mind or thinking of it, signified by the stamping, and having true beliefs about it. The latter is made dependent on the former at 209c but one might wonder how necessary the connection is. Maybe the jury is unable not just to know but even to think of the events in question, since those events did not play the appropriate causal role — they never, so to speak, entered the jury’s head. Still, the jury is, thanks to the lawyer’s speech, well enough able to form true beliefs about them.

Socrates acknowledges the force of the inference from ‘I do not have x in mind’ to ‘I cannot have beliefs (true or false) about x’; indeed this inference is partly responsible for his trouble in accounting for false belief in the Theaetetus. But an objector might reiterate that having beliefs about x is independent of one’s being able to think of x. So even if I cannot think of the Battle of Trafalgar, since I did not witness it, I can surely have beliefs, true and false, about it. It seems therefore that one could set tough conditions for thinking of a thing, such as those Socrates offers at 209c, without regarding that as having any potentially troubling implications for one’s ability to have beliefs about that thing.

But the objector needs to offer an account of what makes it the case that my beliefs about Trafalgar are beliefs about that battle; and a natural, perhaps inevitable, move is to say that my beliefs only succeed in being about that battle if they succeed in picking that battle out. Hence to have beliefs about it I need to have that battle in mind. Then everything rests on what the conditions are for getting the battle in mind. What the resonance between 209c and the jury passage suggests is that without a specification of these conditions it was premature to grant the jury even beliefs about the events in question.

Thus it seems to me a mistake to claim that the requirement set out at 209c, that one needs to distinguish a thing from others in order to form true beliefs or judgements about it, is only supposed to apply to ‘a quite narrow subclass of judgements’, namely judgements of identity (Burnyeat 1990, p. 227, n. 112). The passage certainly focuses on issues of identification, but that is not because it restricts itself to identity judgements. Rather, it uses a premiss about individuation in thought as the basis for a conclusion about judgement generally.

Socrates says that if I think (dianoēthō, 209c1) of one who has a snub nose and protruding eyes, let alone just a nose and eyes, then it will be no more Theaetetus than Socrates whom I will judge (doxasō, c2), and this is part of a tightly structured argument that begins with Socrates’ scepticism at 209b2–3 about whether Theaetetus can be an object of judgement in virtue of features he shares with others, a scepticism justified by reference to the inadequacy of using such features to identify him in thought (dianooumenon, b3–4, dianoia b6, dianoeisthai, b7). Socrates goes on to infer at b10–c3 that on this basis Theaetetus will not be an object of one’s judgement, and concludes at c5–10 that he will be so only if his particular features are stamped. Conflating thought with judgement, as Socrates is careful not to do here, makes it easier to suppose that identity judgements are all he is interested in.²⁴ His point, rather, is that failure to identify an object in one’s thought, by registering its distinctive features, prevents it from being an object of one’s judgement.

If the range of judgements covered by our Theaetetus passage should be interpreted broadly, so too should its range of objects. Although it focuses on individual particulars such as Theaetetus, there is no reason to confine its requirements for being an object of judgement to these and exclude types or kinds. Its initial proposal had stated simply that knowledge results from the addition of a distinguishing feature to true judgement ‘about any of the things that are’ (208e3). In favour of excluding kinds from Socrates’ subsequent critique, it might be argued that the grasp of, say, an animal kind’s distinguishing features ‘is the special province of the expert in biology’ (Lee 2008, p. 431). But that would seem to show not that one is licensed to waive the requirement to pick out distinguishing features when it comes to making judgements about kinds, but that having one kind rather than another as an object of one’s judgement may require expertise. The reasoning of the passage stands or falls irrespective of ontological category. If I cannot make judgements about Theaetetus unless I can tell him apart from Socrates, there are no obvious grounds for ruling that I can make judgements about crocodiles without being able to distinguish them from alligators.

6. Questions of scope

The notion of stamping does, however, suggest that only intrinsic features of the item in question, and among these only its external features, need be represented for TV to be satisfied. This seems a reasonable constraint when it comes to thinking of perceptible objects, though it should be noted that the Wax Tablet model does not restrict the source of impressions to what is perceptible (191d6–7, 195a5–6), only that falsehood is to be explained by mismatch of perception to impression. Indeed Theodorus had earlier picked out Theaetetus as possessing a unique combination of intellectual and moral gifts (144a–b), suggesting, at least in the case of persons, an alternative route to individuation.

As to the intrinsic, its scope is murky, and in the Platonic scheme some features that one might regard as relational may be included.²⁵ Thus, a thing’s being taller than another is on occasion considered to be analysable as that thing having tallness compared to the shortness in the other (Phaedo 101a, 102b–c). On this view, an object’s being taller than another means that object possessing tallness, which may then be available for stamping when the object acts on my mind — and of course one may quite naturally behold a single individual and say ‘she is tall’, the relational underpinning notwithstanding.

The Theaetetus offers some subtle and pertinent observations about relational features and individuation. The opening pages of the dialogue hint, in rapid succession, at two promising-looking ways of picking out an object, which can be seen to be problematic once the discussion of 209c is on the table. At 144c1 Theodorus, having sung the praises of the gifted young man who will shortly be revealed as Theaetetus, tells Socrates that he is the middle one of the group approaching them,²⁶ a fact that prompts Socrates to recognize the youngster and identify him as the son of Euphronius of Sunium (c5), Theodorus then supplying the additional information that his name is Theaetetus (d1). As 209c makes clear, individuation means nothing if it does not give the means to reidentify — Socrates speaks there of being able to pick out Theaetetus again ‘if I meet [him] tomorrow’ (c9). Neither previous spatiotemporal location nor knowledge of origins suffices for that, while even at 144c it is evidently Theaetetus’ intrinsic features that enable Socrates to recognize him as the youth he knows as Euphronius’ son.

Because of its presence as an explicit thesis, my focus thus far in exploring the transparency view has been the Theaetetus. I am not, to be sure, wedded to the notion that TV (or any other doctrine) is something that Plato is unequivocally committed to even when it is presented as an explicit thesis. Plato fascinates because he never directly tells us what he thinks. So there is room to wonder, for example, whether in certain parts of the Theaetetus TV is intended to play the role of a problematic assumption that the reader should see must be discarded. But I hope to show that the norm that TV expresses — that of representing a thing just as it is — exerts a major influence on Plato’s accounts of cognition throughout his work, and serves in various key passages as a condition either of knowledge or of thought more generally. That, I suggest, gives us reason to take it seriously, or at least to assume that Plato did. If it is a view he finds problematic, that may be because the conditions for thinking about the world that TV implies are extremely demanding — which is not to say he believes they are dispensable, or that we should.

One might in fact care to admit that TV has a certain intuitive force, despite what I have described as the modern view, whereby shoddy representation of x can still count as a thought of x. One should at least have qualms about the idea that, for example, representing a man with green eyes should count as thinking of Socrates. On the assumption that Socrates has blue eyes, then ipso facto, in representing a man with green eyes I am not representing him,²⁷ though it might, depending on how well his other features are represented, count as a near miss in my attempt to do so.

A sceptical reader might none the less wonder whether there is any substantive difference between speaking of a near miss in an attempted thought of Socrates and speaking of a thought of Socrates albeit a somewhat inaccurate one. Perhaps what is at stake is only a matter of terminology without deeper roots. Yet the sceptic is likely to continue to show a marked preference for the latter form of expression over the former, and one reason for this is that the latter offers reassurance that our minds hook up with the world’s objects on a regular basis, albeit with varying degrees of precision and perhaps always imperfectly, whilst the former signals that these objects may be cognitively out of reach. Thinking of them becomes ‘the untouched goal, not the daily achievement’, as Rudebusch (1985, p. 536) aptly puts it. If that is a point about terminology, it seems that terminology can disclose a more sizeable issue.

In the remainder of this paper I intend to take some further steps towards evaluating TV as an account of thinking by exploring how it might help shed light on various Platonic discussions concerned with the question of how thought and its objects are related. An examination of TV’s influence on these discussions will, I hope, provide resources with which to assess its credentials as a philosophical thesis, in particular regarding a theme that will be important in much of what follows, and which I have claimed marks out Platonic from some contemporary approaches, namely the nature of the connection between cognition and expertise.

7. Thought and falsehood

To see, first, how TV might be doing further work in the Theaetetus itself, recall that a significant portion of the dialogue (187d–200d) is occupied with the question of how it is that we can go wrong in our thinking — how false belief or judgement is possible. At 188a–c Socrates argues that whether I know two things, x and y, or not, I can never mistake one for the other, given that if I know a thing, I cannot think it is anything else, and if I do not know it, I cannot so much as have it as an object of my thought in the first place such that I could then go on to make a mistake about it.

One may protest that Socrates presents a false dichotomy — surely I can know Socrates and still make mistakes about him; I can know Socrates without having everything about him right.²⁸ But it follows from TV that if I know Socrates then I do have all his features right: getting something about him wrong would mean that it was indeed not him I have in mind, and therefore (a fortiori) not him that I know.²⁹ And then the dilemma is real, for either I get everything about Socrates right, in which case it follows straightaway that I cannot represent him as anything other than who he is, or I fail to get everything about him right, in which case it is not him I have as an object of thought such that I can then make a mistake about him. Socrates says that if I do not know either of two things then I will not be able to ‘take into thought’ (eis tēn dianoian labein, 188b9) that the one is the other. Grant, then, that I do know the two things. What makes this problematic for falsehood is precisely the idea that getting them in one’s thought means representing each just as it is.³⁰

It might be argued that the condition that to know (let alone think of) a thing one needs to have everything about it right is unnecessarily strong, and that to generate the puzzle one need only assume that to know a thing requires that one be able to tell it apart from other things, a ‘much weaker and more plausible’ hypothesis (Sedley 2004, p. 121). Whilst I agree that a liability to confuse x with other things indeed plausibly undercuts a claim to know x, it may be that, in order to satisfy the condition that I not misidentify, the first (apparently stronger) condition needs to be satisfied too, at least where intrinsic properties are concerned.³¹ Imagine three objects with two properties each, one round and red, the second round and blue, the third square and red. To tell the first apart from the others I need to have in mind both — that is, all — its properties.

Generalizing, it might be the case that each of an object’s properties is a unique distinguisher of that object from one or other different object, so that to avoid mistaking the object for anything else I do need to know everything about it. Theaetetus’ snub nose may be less pronounced than Socrates’, but as pronounced as that of some third person, so a different characteristic of Theaetetus will be needed to distinguish Theaetetus from that person. Perhaps there is someone in the world with the same features as Theaetetus except blue eyes instead of green; then I had better make sure that I represent someone with green eyes if I am to be thinking of Theaetetus. Completeness of representation may be unnecessary to individuate an object, but it may not be. So even if, in many cases of plentiful coarse-grained differences between objects, such comprehensiveness is not needed for successful individuation, Socrates is prudent if one of his motives for declaring that all of Theaetetus’ parts be accurately represented is to insure against any possibility that the supposedly weaker condition fails to be met. This gives him a principle of full applicability, and nicely explains why Socrates generalizes from Theaetetus’ snubness to all of his features at the culmination of his reflections at 209b–c on what it takes to differentiate one object from others.³²

The generalization does not explicitly say that all of Theaetetus’ features need to be represented with precision. But in stating at 209c8 that ‘the other’ (talla) features of his be represented in the same fashion as his snub nose, Socrates seems to mean more than that his point is relevant not especially to Theaetetus’ nose but to whatever feature may serve to discriminate him in a given case.³³ The listing of Theaetetus’ various features — nose and eyes and mouth and so on — at 209b5 is given as the first stage in the attempt to think of him. It functions as an insufficient but presumably not an unnecessary step (otherwise why include it?), signifying that comprehensiveness of representation is part of the conception here. Indeed in many cases there is something intuitively unsatisfying about the idea that I am thinking of an object just by picking out a single, even uniquely individuating, feature of that object. Does thinking of Theaetetus’ particular snub nose really equate to thinking of Theaetetus? Theaetetus is a man, after all, not a nose; and it seems to me that we need to have in mind a man (not just a nose) to be thinking of him.

What, though, of clones or exact copies? Here it seems that even maximally comprehensive representation will fail to deliver successful individuation. True, the causal condition might break the tie, making it the case that one qualitatively identical item is the object of my thought rather than another; but that will not help the subject tell one apart from the other. The question about duplicates might be dismissed as an anachronism from a Platonic perspective, though it was certainly a live issue for the Stoics, whose staunch response was to deny that any two objects could be qualitatively indistinguishable.³⁴ But in fact, in Plato’s Cratylus (432b–c) the participants agree that if all Cratylus’ qualities were exactly reproduced we would have two Cratyluses,³⁵ which suggests that picking out either as Cratylus would count as correct. If so, and allowing for the metaphysical oddity of treating two numerically distinct objects as the same individual, TV will remain the guarantor of successful individuation even in what might seem to be a limiting case.

8. The objects of enquiry

Let us now consider a better known Platonic passage, but one which has obvious structural affinities with the puzzle about knowing and not-knowing in the Theaetetus,³⁶ namely the paradox of enquiry at Meno 80d–e, often dubbed ‘Meno’s paradox’. This paradox too, it is tempting to claim, betrays a false dichotomy between total knowledge and blank ignorance.³⁷ It states, in Socrates’ reformulation (80e2–5), that if I know a thing, there is no need to seek it; if I do not know it, then I cannot seek it, because I do not know what it is I am looking for. Then it may be said among other things that the second limb of the dilemma will hold only if I know nothing about my object of enquiry, an objection positively invited by Meno’s original formulation which talks of the would-be enquirer being unable to search when not knowing the object ‘at all’ (to parapan, 80d6, cf. b4), this in turn picking up Socrates’ claim at the start of the dialogue not to know ‘at all’ what virtue is (71b3). Maybe if I know something about the object, or just have some true beliefs about it, I have something to go on and enquiry can proceed.

But what does this response amount to? The fact that the ‘at all’ qualification is dropped in Socrates’ restatement of the paradox should give pause to those whose diagnosis places weight on it.³⁸ The phrase’s prominence, however, makes it doubtful that Socrates is simply treating not knowing and not knowing at all as blandly interchangeable. Part of his purpose may be to preview his solution by hinting that subjects know the object of their enquiry ‘in a sense’ (Moravcsik 1971, p. 57), as the recollection theory may suggest. But Socrates’ switch also implies that blank ignorance is not a necessary condition for the second half of the dilemma to bite.

To see this, take a case removed in specifics but I think close in spirit to what is animating the paradox. Imagine that an African elephant and an Indian elephant have escaped from my local zoo and that I am tasked to search for the former. Let us say that I know lots of things about African elephants except those few characteristics that distinguish them from the Indian variety. I step outside my door, and lo and behold there are the two escapees in my spacious driveway! (Let us grant, what is by no means unproblematic, that I can take the pair to be the escapees.) Question: can I, even in this hugely advantageous position, search for the African elephant on the basis of what I know?

The answer is surely not. All I can do is stare dumbly from beast to beast, with no principled way of taking things further. Moreover, in this situation, it seems we might well deny that I have knowledge of the African elephant, given that I cannot pick it out from a choice of two, other than by guesswork. It is but a small step from there to the thesis that I am unable to think of an African elephant either, since at best I can think of an object that is indifferently African or Indian. Yet by hypothesis I had a plentiful stock of information about the African, far from the situation of rank ignorance that Meno’s formulation might have suggested was the crucial barrier to the carrying out of a search.

The conclusion looms that in the absence of identificatory knowledge there is no basis, beyond guesswork, for enquiry. The same would hold if Jumbo and Jimbo, two African elephants, were the escapees and my task was to search for Jimbo without my knowing his distinguishing features. In general, search is thwarted whenever the information I have about the object of my search fails to track these features through the population in which the search is conducted. Given that differences between things may be few and slight there is nothing to prevent this occurring in the absence of a full grasp of the object’s features.

Now Socrates evidently holds that Meno’s paradox can be defeated, via the theory of recollection, but in this his sympathies lie with the second limb of the dilemma. He holds, that is, that one cannot seek x without information sufficient to identify x. The solution proceeds by rejecting the inference (one that at least in the case of abstract objects appears plausible) that if I already have such information x is no longer an object that needs to be sought. It appeals to the idea of information stored within oneself, but needing to be recovered via a process of recollection. This framework both leaves room for and enables genuine enquiry: consider how possession of, but inability to recall, the name of the person one met at yesterday’s party can allow the subject to identify suggested names as further from, closer to — or identical with — the actual one.

One might none the less object to my description of the elephant example as being in the spirit of the paradox. It may at any rate seem implausible that Socrates should suppose that one discovers, say, what an African elephant is through recollection as one might, according to him, discover what virtue is. Yet the Meno is studiously unrestrictive as to the scope of recollection. The soul has seen ‘all things’ (panta chrēmata, 81c6–7) in its previous existences here and in Hades, while the paradox itself is stated in quite general terms as to object. Socrates’ talk at 85e1–3 of the slave eventually being able to recollect the whole of geometry ‘and all the other branches of learning (mathēmata)’ indicates no doubt a focus of interest on ‘a systematically arranged body of knowledge’ (Dimas 1996, p. 1, n. 1), but not a restriction of scope. Rather, systematicity is given the widest possible scope by Socrates, who speculates that ‘all nature is akin and the soul has learned all things (hapanta)’ in its prior incarnations (81d1). However one judges recollection as a solution, this generality does not look accidental. It indicates, I think rightly, that there is no object of enquiry to which Meno’s challenge does not in principle apply.

9. Recollection and the objects of thought

More explicitly than the Meno, the Recollection Argument (RA) of the Phaedo concerns itself with the conditions for thinking of an object. A good deal of scholarly discussion has centred on exactly what sort of cognition the argument is intended to account for. Is it the humdrum grasp that any regular human has of concepts such as equality or justice? Or is it the higher-level grasp striven for by experts — philosophers, as Plato would have it — manifested in the ability to give accounts or definitions of such concepts?³⁹

These alternatives throw up some interesting parallels with what I have described as differences between modern and Platonic views of thinking. Since Plato considers that items such as justice are real objects, standardly termed (by scholars more than by Plato himself) Forms, the question boils down to how stringent are the conditions that he thinks need to hold in order for it to be true that a subject is getting in mind a given object. Can an ordinary person, with only a rough and ready conception of (say) justice, be regarded as getting, however sketchily, the Form of Justice in mind by acquiring or exercising that conception? Or is it only the philosopher, in attaining a precise grasp of the concept, who does so?

Certainly, TV would suggest the latter. But that is to get ahead of ourselves. In order to explore how stringent a conception of recollection is operative with regard to Forms, it might be helpful to consider how stringent a conception Plato is working with in the everyday sense, where an individual such as Simmias might be its object. After all, the RA takes pains to establish the conditions for ordinary recollection first, before going on to argue that they can be applied to our cognition of concepts such as equality. So those who wish to take a stand on this latter phase ought to pay close attention to what the first phase tells us about the conditions for recollection in general. An examination of this phase of the argument will require a certain amount of textual analysis, since there has been something not far short of a scholarly consensus about Socrates’s basic approach here that I take to be radically mistaken. Showing this will enable us to get clearer on the view of thinking that the RA espouses, and thus see how stringent a conception of thought is operative here even with regard to our cognition of ordinary objects.

I want to focus on one crux in particular: Socrates’ remark that, whenever one recollects from ‘likes’ (e.g. Simmias from his picture), it is necessary to get in mind (ennoein) whether or not the object that prompts the recollection ‘falls short’, with respect to likeness, of the object whose recollection it prompted (74a5–7). Why, it may be asked, should it be thought necessary (anankaion, a6), in looking at a picture of Simmias that causes me to think of Simmias, that I form a view about its quality as a likeness of Simmias? One can see why the modal claim has been dubbed ‘highly questionable’ (Scott 1995, p. 63, n. 12). My suggestion is that it might be fruitful to explore a connection between this requirement and the requirement on thinking imposed by TV. A small clue that some link may be intended comes at Theaetetus 209c9, where Socrates says that it is only Theaetetus’ distinctive features being stamped that will ‘remind’ him (me … anamnēsei) of Theaetetus when he next encounters him. This tough condition for being reminded perhaps serves as a reminder of another passage more centrally devoted to recollection and provides a hint about how to interpret it.

To motivate my reading of the RA, let me first consider a different approach to tackling the crux that has won a good deal of favour with scholars. Take Socrates to mean it is necessary for the subject to get in mind not whether but that the image falls short of the recollected object, and read this as implying that to get Simmias in mind from his picture I must recognize that the picture is indeed just a picture of Simmias, not Simmias himself; for otherwise there is no question of my moving beyond the picture to get to think of another thing, Simmias.⁴⁰

This initially attractive response, call it the ‘closed’ interpretation, founders on the text. Socrates plainly takes as open the question to be asked about the image’s falling short. He says it is necessary that the subject get in mind whether the picture (or whatever it may be) falls short in some way or not (eite ti elleipei … eite mē, 74a6–7) of the recollected object. Why include the second disjunct as a possibility if what the recollecting subject has to get in mind is the first?⁴¹ Socrates clearly envisages that one might get in mind that a picture does not fall short of its original. So he must be claiming, as his words in any case naturally suggest, that the subject has to get in mind whether, and if so to what extent, the image fails to be a good likeness of the original.

That this is Socrates’ meaning is brought out when he speaks of this step as something the subject ‘undergoes in addition’ (prospaschein, 74a6) to the recollection. If his point were that the subject needs to recognize the likeness as just a likeness, this would be in order that the subject be in a position to recollect, whereas his language indicates that the getting in mind of whether the likeness falls short or not happens in the wake of recollection, not as a means to it. Forming a view as to the quality of the likeness is indeed the sort of thing that may happen once likeness has reminded one of original, giving a smooth fit with Socrates’ choice of words.

Now Cratylus 432b–c, as we saw earlier, asserts that something which reproduced all the qualities of an object would be not an image, but another of that object — ‘two Cratyluses’ (c5–6), for example. That passage goes on to say that an image must therefore ‘fall short’ of its original, and so is often cited in support of the closed interpretation of the Phaedo. But as we have just seen, the Phaedo’s wording allows that the subject has the possibility of noticing that the image does not fall short, yet this is not supposed to mean he is not recollecting — on the contrary, the disjunction of falling or not falling short is introduced as applicable ‘whenever someone recollects’ original from image (74a5). So one should resist the suggestion made by Harte (2006, p. 40, n.14) that the ‘or not’ disjunct is meant to apply to cases in which a would-be reminding object fails, because not recognized as merely a likeness of something else, to trigger recollection. Socrates speaks of the subject having to notice whether the likeness does or does not fall short ‘of that which he recollected’ (hou anemnēsthē, 74a7), the past tense confirming unequivocally that the scenario envisaged is either way one in which recollection has taken place. It follows that replication along Cratylus lines cannot be what is meant in the RA by not falling short.

The difficulty, then, as indicated earlier, is why Socrates thinks that a subject must, in being reminded of Simmias by his picture, get in mind how poor or good a likeness it is.⁴² Certainly, that may happen; and perhaps in order to be caused at all to think of Simmias by his picture, it must not be so poor a likeness that it cannot be recognized as such (the opposite problem from replication); but again, Socrates insists that whether and to what extent it falls short is an open question. So there must be more to it than one’s having (perhaps implicitly) to recognize that it is a good enough likeness for recollection to occur. Socrates’ wording suggests, rather, one who forms a view as to the picture’s quality as a likeness, good or bad — why does he think that this rather refined judgement is inevitable whenever a picture of x has put one in mind of x?

The key, I think, lies in recognizing that it is a refined judgement he is talking about.⁴³ Consider the sort of person who would, as we might say, be bound to notice whether an image failed accurately to capture its original or not. It seems to me that, in one way or another, such a person would have a kind of expertise. Take a painting of an African elephant. Presumably a zoologist with a special interest in elephants would be especially well placed to pronounce on whether it misrepresented, say, the proportions of the ears, or the shape of the trunk. The better one knew one’s elephants, the more, it seems plausible to say, one could not help but notice whether a depiction misrepresented features of the original. Someone who had only a rough and ready conception of an elephant would not be struck by imperfections in the representation as our expert zoologist would be.

Socrates himself acknowledges the role of expertise in a similar case. At Theaetetus 144e–145a he insists, albeit in a jocular tone, that even in the mundane case of judging whether, as Theodorus reports, Socrates and Theaetetus look alike, a certain sort of expertise is appropriate — that of a painter, he says, rather than a mathematician such as Theodorus. Perhaps rightly, Socrates implies that a painter whose business was to depict animal form might be a better candidate for expert knowledge of external physiognomy than a zoologist whose concern may be primarily with the animal’s internal structure. But it is expertise of one sort or another that is called for.

Of course, if the elephant in our painting were depicted in a flagrantly sloppy manner — the ears, say, being grotesquely out of proportion — then even a regular viewer might be bound to notice. But that is merely to confirm that there is a correlation between expertise and the necessity of being struck by misrepresentation. The less knowledgeable the viewer, the more flagrant the misrepresentation needs to be to command attention. The greater the expertise, the more difficult for even relatively subtle misrepresentations to escape notice. In the case of someone — perhaps a merely hypothetical figure — who had a comprehensive knowledge of elephants, one might well say that in the normal course of events misrepresentations could not fail to jump out at such a person, for whom the original is brought, in all its glorious detail, to mind.

One might be prepared to concede that a complete expert is bound to notice whether or not a representation falls short of its original. But Socrates in the RA does not put things in terms of expertise. Indeed the puzzle is what makes him apparently think that whoever gets in mind an object from its representation must notice whether or not the latter misrepresents the object. The gap can be bridged quite straightforwardly, however, if we assume that for Socrates, here as in the Theaetetus, the ‘getting in mind’ of an object requires the subject to grasp its features just as they are. If so, we have a ready explanation of why he is able to insist that in getting an object in mind from a representation of it I must notice whether or not the representation falls short. For if, in order to get the object in mind, I must get in mind its features just as they are, it seems far from implausible to say that I am bound to notice how, if at all, the painting in front of me fails to measure up to the object that I am thinking of with such transparency. Sensitivity to the quality of the likeness is indeed necessary on this view of what it takes to have the original in mind.

10. Representation and the objects of thought

If one presses hard enough, it may seem that what one will turn out to be doing on this reading, no less than on the one I have rejected, is to notice not whether, but that, the likeness falls short; for one may be sceptical about whether any painting, say, could be a faithful enough likeness to survive scrutiny on these terms. Imperfection, however slight, will be registered by one who has the appropriately lucid recall. In that case my reading will fail on the very point that was supposed to motivate it.

Elsewhere, however, Socrates offers a view of what makes a good likeness that is tough but not absurd. At Cratylus 431c the painter who depicts ‘all the appropriate colours and shapes’ of the original is said to produce a good likeness, whereas one who adds, subtracts, or presents a feature of disproportionate size fails to do so. If that is Socrates’ view, then we may take him to regard it as possible that a painter produce a likeness that is not liable to be found wanting. On these terms it remains a genuine alternative that likeness not fall short of original.⁴⁴

Does Socrates’ statement, at 431c11–13, that a likeness that does fall short by these criteria may yet be a likeness, albeit a bad one, mean that he must similarly regard a thought that does not, by these criteria, represent the features of its object, as not necessarily excluded from being a thought of that object? That would present a difficulty for my reading of the RA, given that its open question seems to presuppose that likenesses may be bad as well as good. Notice, however, that in the Cratylus the analogy fails to persuade Cratylus, who at 431e–432a is still insisting that a name that adds to or subtracts anything from its nominatum fails to be its name. That is why Socrates has to point out that if this were taken literally, names would be duplicates rather than images of their nominata, and thus could no longer name them. We would get, say, two Cratyluses rather than Cratylus and a representation thereof (432a–d).

Thence Socrates extracts Cratylus’ concession that a name with an inappropriate letter can still do the business of naming (432d–433a); but that of course does not follow.⁴⁵ A name that does not literally duplicate its nominatum — being merely composed of letters — can still be composed of all the right letters (however that is to be determined; not an issue I shall pursue here). Socrates’ inference is equivalent, with reference to his analogy with painting, to the conclusion that a painting which counts as a likeness by not being a duplicate of its original may therefore just as well count as a likeness even if it, say, misrepresents the colour or shape of the original.⁴⁶

Behind this dialectic of duplication versus misrepresentation lies a crucial point: an accurate representation does not have to reproduce the qualities of what it represents. A painting can accurately represent a spherical object without being a spherical painting, and so on. A representation may capture its original just as it is without being a duplicate and so no longer a representation at all. On Socrates’ telling, there will, no doubt, be a marked tendency for an expert assessor to find that any likeness does fall short of its original, given the tough requirements for a likeness not to fall short that he sets out in the Cratylus. But that serves simply to highlight the unity of his account of recollection in the Phaedo, since those who aim to recollect Forms pronounce with vigour that the perceptible particulars that trigger their recollection do indeed fall short of them, just as we may imagine there would be fault to be found with most paintings in relation to their originals when judged by Socrates’ standards.

As far as recollection of ordinary objects is concerned, one may yet feel entitled to ask whether, given the transparency requirement for thinking of an object, the getting in mind of these objects can take place at all on the interpretation I am offering, other than perhaps by our hypothetical complete expert. But recall here the parallel question from which this investigation of the RA began: whether, in the succeeding portion of the argument, concerning Forms, those who recollect include anyone possessing a rough and ready grasp of the relevant concept, or whether the only ones who do so are those elusive experts who can give a proper account of it, which seems to mean at the dramatic date of the dialogue either no one or at most Socrates.⁴⁷

If, as I have suggested, the requirement that thinking of an object be transparent lurks behind the discussion of ordinary recollection, this would not, in context, be odd; it would simply lend support to that side of the debate that sees the discussion of recollection of Forms as restricted to the precise grasp that philosophers aim at. The relation between the two parts of the RA is then smoother for advocates of that reading than for its opponents, and smoother than if its advocates were to rest content with the idea that the first part offers a more relaxed view of recollection than the second. Instead of moving from one notion of recollection, whereby getting in mind a rough and ready idea of the object in question can count as success, to a quite different one where only those with a precise grasp of the object are regarded as getting it in mind, we would be dealing with the same stringent notion in both cases. Thus we remain true to Socrates’ appeal to the recollection of ordinary objects as explicatory of the recollection of Forms.

11. The limits of cognition

My concern is not to present a full-scale analysis of the RA that would decisively favour one side of the debate over the other. But all parties can perhaps agree that the requirements for getting a thing in mind should be consistent throughout the argument, on pain of ignoring one of its basic structural features. In this regard it seems to me that those who have sympathy (as I do) with a reading that restricts a grasp of Forms to philosophers should be careful not to look the wrong way round at the question of what it takes to grasp a Form. Thus recollection of ordinary objects is straightforward, the story might go, while recollection of Forms is hard, and that is why one might remain untroubled by a divergence between a relaxed conception of ordinary recollection and a more stringent one for Forms.

Is that true to Plato’s texts? Instead of surmising that it is easy to get, say, Simmias in mind, but rather tough to think of a Form, it might better fit what Plato actually says about Forms and particulars if we consider the possibility that only Forms can properly speaking be grasped. After all, the Parmenides, in the wake of its critique of Forms, is ready to declare that without them one will have ‘nowhere to turn one’s thought’ (135b8). Less dramatically, Forms are, as I shall put it, cognitively privileged items in Plato’s ontology. Whatever the highest type of cognition happens to be in a given context, Forms are routinely singled out as its proprietary object. To state the matter succinctly, if roughly: only Forms can be known.⁴⁸

That this is a Platonic view has sometimes been denied. The touchstone for debate has been the middle books of the Republic, in particular the argument concerning knowledge and belief at the end of Book V. Yet for all the careful analysis that has been devoted to upholding this denial,⁴⁹ it is hard to read these books without the sense that what they express is a conviction that to be an object of Φ is a privilege reserved for Forms, where Φ stands for the highest type of cognition, whether this be knowledge as in the dyadic structure of the Book V argument, or understanding (noēsis or nous) in the more elaborate scheme of the Line analogy of Book VI.

Now it is true that those who have reached the Republic’s cognitive pinnacle are said to ‘know (gnōsesthe) what each of the images are, and of what they are images, through having seen the truths concerning beautiful and just and good things’ (VII 520c4–5).⁵⁰ But what cognitive achievement does this passage describe? Primarily that of distinguishing image from original: picking out an image as a certain image (knowing ‘what it is’), and identifying the original that it is an image of, say Beauty itself in the case of a beautiful vase.⁵¹ What is evidently being referred to is the philosopher’s avoidance of the mistake made by the lovers of sights and sounds (on whom see further Sect. 13 below), who, by contrast with philosophers, fail to distinguish images of beauty, and so forth, from their originals. That it is the philosophers’ grasp of Forms that enables images to be thus distinguished is not in doubt; but it does not follow that their grasp of images is the same as their grasp of Forms. So the passage does not count against the thesis that for Plato only Forms can be Φ-ed.⁵²

Earlier on, in introducing the topic of the Good, Socrates did say, at VI 506a, that Guardians will need to know in what way just and beautiful things are good, and that no one will know (gnōsesthai) them adequately otherwise. His purpose here is evidently to show why knowledge of beauty, justice, and so forth, is lacking without knowledge of goodness; it is a point about the relation between goodness and other attributes, not between particulars and Forms. There is no commitment here to a thesis that particulars can be cognized in just the same way that Forms can.

If that is right, one feature of Plato’s position seems to me to have been paid insufficient regard. If, as might be assumed, Forms are the really difficult objects to grasp, why is it that, of all possibilities, they are privileged as knowable? Should not Plato’s position be the other way round? If it is easy enough to get in mind equal sticks, or a beautiful vase, but hard to get in mind Equality or Beauty itself, should it not be that perceptible particulars are regarded as knowable, with Forms, given the difficulty of fulfilling the philosopher’s quest, being regarded as primarily objects of some less exalted type of cognition? Since even Socrates does not pretend to knowledge of Forms, why not say that the best, or at least the standard thing, one can do with Forms is have (possibly false) beliefs about them? By contrast, should not Socrates feel confident in declaring that, along with everyone else, he has knowledge of heaps of ordinary things, so that they would stand as the principal candidates to be objects of knowledge?

Something has clearly gone wrong here. The problem with perceptible particulars is not that they are too simple, but that they are too complex, to be grasped. Let us then hypothesize that Plato may be working with the idea that, if not cognition in general, then at least the highest form of cognition will be that which gets the features of its objects just right. If so, then it will evidently be impossible, or as near as makes no difference, to have perceptible particulars as objects of that form of cognition. No one is going to be capable of getting their features right, since they will be too complex to capture just as they are.

The Theaetetus offers some suggestive remarks to this effect. It regards a type such as a wagon as complex, but of sufficient determinacy to be captured in thought — grasp of the ‘one hundred timbers’ (in Socrates’ citation from Hesiod at 207a3–4) that are its elements is rejected not as too strong but as too weak to count as knowledge, and by implication such a grasp is considered perfectly possible to attain. Perhaps a human being qua type could also be regarded as consisting of a determinate number of parts, such as an anatomist might be capable of enumerating. But what of perceptible tokens of a type considered as individuals? To pick out Theaetetus one cannot just say ‘this is a man with a nose, eyes, mouth’ and so on for each of his parts (209b4–6). One must, as we have seen, go on to identify his snub nose, and that as distinct from other snub noses, and the same with his other parts.

In principle this process may have to go well beyond a grasp of the gross functional parts, assuming they are determinate in number, if individuation is to succeed. The detail that might need to be captured has no obvious limit. It is not that such detail is required every time the object is mentioned;⁵³ both listing of elements and articulation of distinguishing features are things the subject must merely be ‘capable’ of (dunaton, 206e7; echein … eipein, 208c7–8) in order to grasp the relevant object. Rather, in the case of particulars, fulfilment of even the modal condition becomes problematic, and this not simply because the unending level of detail that may obtain threatens to elude representational capacities of any sort. Representation is treated, in addition, as involving an ability to articulate its content, a point that is prefigured in the identification of thinking with silent speech,⁵⁴ and reinforced by the way representation of Theaetetus is portrayed in terms of the enumeration of Theaetetus’ features.

Plato’s tendency to regard representation as carrying with it the ability to articulate is not confined to the Theaetetus,⁵⁵ even as his view of what capacities have representational power varies. The Theaetetus denies that power to sense-perception once the latter is distinguished from knowledge at 184–6⁵⁶, retaining it for thought (including thus far thought about perceptible objects), which in turn is treated as a kind of speech.⁵⁷ In Republic VII the representational character of sense-perception is intact, and conveyed by having the senses ‘report’ (paraggelei, 524a2) or ‘state’ (legei, a7) that such-and-such is the case. An even closer dependence of representation on articulation is evident in the analysis of thinking in the Philebus, where the job of the painter in our soul is merely to produce images of the words that the scribe in our soul has written (39b6–7).

If conceptualization is a weaker notion than articulation (the latter implying the former but not the other way round), Plato would have less sympathy still with the idea that representational content can outstrip our ability to conceptualize what is represented.⁵⁸ And a motive for his stance readily suggests itself, illustrated for example in the way that the Line analogy of Republic VI regards clarity and truth as increasing when one ascends from lower cognitive power to higher:⁵⁹ a less sophisticated cognitive capacity should not be able to capture the world more accurately than a more sophisticated one. If not all representational content can be articulated, that principle will be breached. And certainly it seems as if, for example, the precise shape of Theaetetus’ snub nose that distinguishes it from others is unlikely to be decipherable by articulation,⁶⁰ whether or not that same fineness of grain defeats other forms of cognitive capacity too.

12. Knowledge and complexity

Given the potentially indeterminate complexity of Theaetetus’ individuating features, wherein he contrasts with a type such as a wagon, successful representation of him is rendered problematic, and here there is a parallel with the central Platonic contrast between Forms and particulars. The complexity of the latter is described in a number of ways. Sometimes, as at Theaetetus 209b–c, it is treated statically — each perceptible particular just has a lot of different bits to it.⁶¹ The Republic implies that any perceptible particular is ‘indeterminately’ multiple (apeira to plēthos, VII 525a5–6), terminology that recurs at Philebus 14e3–4.⁶² Sometimes the complexity is treated more dynamically — particulars are, as Forms are not, subject to change.⁶³ So even if I could accurately grasp a particular at one moment, it is always changing over time, so that in subsequently calling it to mind I will inevitably fail to represent it accurately.

Hence the Cratylus can assert that knowledge is not possible of that which ‘at the very instant the knower-to-be approaches would become different and of a different quality’ (439e7–440a2).⁶⁴ The mutability of perceptible particulars in this regard is nowhere more vividly described than in the Symposium, as it recounts the ways in which every living being is in flux (207d–208a). But the non-living are also prone to the vagaries of change: the Phaedo refers to ‘humans, horses and cloaks’ as never being in the same state (78d10–e4), whilst Forms are, by contrast, not subject to any change (78d4–7). These passages do not claim that perceptible particulars are always changing in every respect; rather, they maintain the eminently plausible thesis that, in some respect or other, such objects are always changing. Forms, in sum, are simple and unchanging, particulars are complex and unstable.

In the argument of Republic V this is conveyed by the allotting of knowledge to ‘what is’ and belief to ‘what is and is not’. Much illuminating work has been done in analysing the sense or senses in which the verb ‘to be’ might be functioning here. But perhaps that has resulted in slightly too much of a shift in emphasis. After all, the key term that demarcates the territory of belief from that of knowledge is not ‘is’, but ‘not’. However one parses ‘is’, the conjunction of ‘is’ and ‘is not’ is plausibly intended to capture in general form the notion of change. For change, one might suppose, is just a thing’s switching from being in some state or condition to not being in it (whether the switch be synchronic or diachronic). The debate over the correct reading of ‘is’ should not obscure the point that the basic contrast at work in distinguishing the domains of knowledge and belief is that between stability and change.

If one takes seriously the emphasis on simplicity and stability as the key features that differentiate Forms from particulars, it is not hard to see why Forms would be cognitively privileged. Given that Plato’s cognitive rankings are keyed to such notions as clarity and truth,⁶⁵ it seems plausible that only Forms could be grasped clearly and without falsehood — could be represented, that is, just as they are. Only Forms are of a sufficiently stable and simple nature to enable a clear and truthful representation of them to be made. Note in this regard how Socrates emphasizes in the argument of Republic V that knowledge is not just of what is, but of what is ‘as it is’ (hōs esti 477b11, hōs echei 478a7).⁶⁶ One may infer that there is no knowing a perceptible particular ‘as it is’ because it is too complex and unstable to be captured, or even to be, just as it is.

13. Knowledge and identification

This approach to the cognitive privileging of Forms is borne out if one considers a further aspect of the argument of Republic V. The fundamental difference between the lovers of sights and sounds (who are confined to belief) and philosophers (who can attain knowledge) is not, or not just, that the former restrict their gaze to perceptible things whereas the latter look to Forms. It is that the former mistake the one for the other — they identify beauty with perceptible beautiful things, and so on — whereas philosophers recognize that beauty is different from perceptible beautiful things (476c–d). Thus what distinguishes belief from knowledge in this context is that those with belief are, as those with knowledge are not, liable to get things wrong, in the sense of mistaking one thing for another.⁶⁷

In misidentifying the beautiful perceptibles as beauty, do the lovers of sights and sounds thereby get in mind, however sketchily, the Form of Beauty? Socrates insists that their thought is unable to see ‘the nature of the beautiful itself’ (476b6–7), which might seem to leave room for thinking of the beautiful itself in some way that does not involve the grasp of its nature. Yet the Form just is the nature of the beautiful, pure and simple, since Socrates immediately follows up by remarking on how rare are those capable of getting to ‘the beautiful itself’ and seeing it ‘by itself’ (b9–10), standard locutions for the Form. The phrase ‘by itself’ picks up the point that each Form is one despite the guise of plurality taken on when in commerce with perceptible things (a5–8). Thus the lovers, who cannot see beyond the plurality of beauty, fail to see the Form. I therefore doubt that Socrates envisages a thought that could be of the Form, but not of the nature of the beautiful itself. If one fails to grasp it accurately, the Form is not an object of one’s thought.

The idea this implies, of intellectual advance not as the gradual increase of one’s cognition of an object, but as progression towards the stage at which one does cognize it — that is, grasp it just as it is — characterizes the two great ‘ascent’ passages in Plato, in relation to the Form of Beauty in the Symposium, and the Form of Good in the Republic. In the Symposium, at the climax of the ascent, the lover is said to ‘suddenly (exaiphnēs) discern something wondrously beautiful in nature’ (210e4–5), the ‘suddenly’ suggesting that this is not an object that was gradually hoving into view, but one that is seen for the first time at journey’s end.⁶⁸ The Republic equally emphatically states that the Form of Good is ‘the final thing to be seen, and with difficulty’ (VII 517b8–9),⁶⁹ a formulation that would be ill-fitting if cognition of the Form were conceived of as coming in degrees.⁷⁰

This is not to imply that Plato thinks we cognize Forms by some faculty of intellectual intuition in addition to, or instead of, by reasoning.⁷¹ My claim is rather that, if reasoning is the method,⁷² it is not until the reasoning is complete that one is regarded as cognizing the Form. That seems to me the most satisfactory way to explain why the Platonic texts that concern the cognition of Forms regularly emphasize both an arduous process of intellectual ascent and a moment of vision at the climax. The arduous process is wholly responsible for one’s cognition of the Form; but only at the close of the process is it the Form, and not some more or less inaccurate approximation thereof, that one has in mind. The language of vision serves simply but appropriately to indicate that.⁷³

A similar notion of intellectual progress is to be found in certain aspects of the celebrated portrayal of Socrates as midwife of ideas in the Theaetetus. Here the theme of a subject’s making progress, under Socrates’ care, is explicit (150d5, 151a4–5, cf. 210c1–2) but it sits alongside a strict dualism of true and false with regard to the conceptions that a subject may bring forth. Socrates will test to see ‘whether the young man’s thought is producing a semblance and a falsehood, or what is viable and true’ (150c2–3). It soon emerges that a conception of knowledge that is incorrect does not count as a genuine conception at all, but a phantom or ‘wind-egg’ (151e6), as turns out to be the case with Theaetetus’ own various attempts to say what knowledge is (210b8–9). We are surely not to conclude that Theaetetus has made no progress, under Socrates’ tutelage, towards producing a conception of knowledge. But equally clearly, in terms of the procreative imagery, he has not yet produced one.

This view of cognition may seem to pose a threat to the possibility of genuine dialogue,⁷⁴ considered vital in Plato for the achievement of intellectual advance, whatever imagery it may be couched in. Such dialogue might be intrapersonal, as in Socrates’ description of thinking as internal question and answer, or interpersonal, as manifested throughout Plato’s work. In the case of, say, a debate concerning the nature of justice, if the parties do not have an accurate conception of justice, then one would be committed to saying that they are not really debating (or thinking) about justice at all; and if one party, but not the other, has an accurate conception, they would, unwittingly, be talking at cross-purposes.⁷⁵

Plato, however, is untroubled by the idea that people may take themselves to be thinking and talking about things that they are not in fact thinking or talking about. He has, for example, no brief for the notion that subjects have authoritative access to the contents of their own thoughts.⁷⁶ He is also prepared explicitly to acknowledge that the parties to a debate may be talking past each other (for anyone familiar with the full canon of Socratic encounters this will occasion little surprise). Socrates worries, at Theaetetus 196d–e, that he and Theaetetus may not be ‘understanding anything from one another’ (e4) in discussing knowledge without their knowing what it is. But Socrates is equally adamant that the discussion must go on (196e–197a).⁷⁷ What is important is the possibility of debate making progress, and ultimately of getting things right. And here TV presents no particular obstacle. Intellectual progress will be a matter not of the parties getting more accurately in mind an object of enquiry that they already have partially or imperfectly in mind, but of their moving closer to, in the ideal case succeeding in, getting the object in mind. And this, I have suggested, is exactly the picture of progress that the language of the ascent passages evokes.⁷⁸

14. Knowledge and thought

Let us consider, then, who in general would not be liable to the kind of mistake committed by the lovers of sights and sounds, of identifying one thing as another. The answer looks to be a kind of expert. Take the case of a forger who makes a copy of a famous painting. It seems plausible that an expert specializing in the works of the painter concerned would in general be the best candidate for distinguishing forgery from original. But if we make this more precise, and ask who would be liable to notice some inaccuracy, however slight, in the forgery, it would surely be someone who had an exact grasp of the features of the original — a complete expert, as we earlier labelled such an individual.⁷⁹

But this then raises the question: what sort of objects could be grasped in all their detail? Particular perceptible objects, paintings or otherwise, given their complexity, would seem to be ruled out. No kind of expert (however deserving of the title) could be immune from mistake. So in that sense there can be no knowledge of the perceptible realm — any claim to it can be compromised by one’s liability to misrepresent. An object such as a Form, however, would have the simplicity to be grasped as it is. This is not to deny that it remains very hard to successfully represent a Form — to give a correct account of it. But at least it is only very hard and not, as seems to be the case with particulars, impossible.

Thus there turns out to be a close connection between, on the one hand, the idea that only those with the knowledge of the complete expert will have the capacity to avoid misidentification and, on the other, the question of what sort of object could be grasped in such a way that it could not be mistaken for some other.⁸⁰ Plato’s answer to this latter question, as I read him, is that an object with the character of a Form can be grasped in that way, but perceptible particulars cannot, and so it is philosophers, with their proprietary interest in Forms, who operate in a sphere in which misidentification can be avoided.

The requirement that to know a given object one must be able to distinguish it from other objects therefore excludes, as far as the Republic is concerned, perceptible particulars from the ambit of knowledge, though it still allows that they can be objects of belief. It is less clear that they are regarded as objects of thought in that work. The state of belief is said to characterize the lovers’ ‘thought’ (dianoia) at Republic V 476d4–5; but the Line analogy of Book VI separates thought from belief, and confines it to the intelligible realm. And one might certainly contend (recall the Battle of Trafalgar example in Sect. 5 above) that it is possible to have beliefs about an object that one is unable to think of. I argued in section 5 that the Theaetetus casts doubt on the possibility; the Republic seems to leave it open.

What is closed off, once the Line is in place, is the idea that perceptible particulars function as objects of thought. I do not believe, from a Platonic perspective, that this is supposed to undermine the transparency view or reduce it to absurdity. On the contrary, I take it that one of Plato’s main motivations for the introduction of Forms is that without them there might be no objects of thought (recall the quotation from the Parmenides in Sect. 11 above). It may be that the Theaetetus, in its unfolding of the transparency view, itself points towards the need for Forms in this regard, though its ontological commitment is not overtly so narrow. For example, a perceptible type, such as a wagon, may serve as an object of thought in so far as it has a determinate set of characteristics. Both Socrates’ apparently exhaustive listing of a wagon’s functional parts and his reference to its more fine-grained ‘one hundred timbers’ suggests that such types would fit the bill. So too, in a more complex way, would natural kinds. The transparency view may have sombre implications for perceptible particulars as objects of thought; but it is not a view that entails any kind of cognitive nihilism.

15. Conclusion

In the light of these considerations, how seriously should we take the transparency view? By way of sketching a reply, let me cite an extract from an influential modern work:

It is a tribute to the education of philosophers that they have held this thesis [that in referring to an object a subject must believe she has picked it out under some uniquely identifying description] for such a long time. In fact, most people, when they think of Cicero, just think of a famous Roman orator, without any pretension to think either that there was only one famous Roman orator or that one must know something else about Cicero to have a referent for the name. (Kripke 1972, p. 81; his emphasis)

Kripke’s chief topic here is reference (that highly refined product of philosophers’ education), not thought.⁸¹ But irrespective of what most people believe themselves to be doing, I submit that thinking generically of a famous Roman orator is no more sufficient for thinking of Cicero than thinking generically of a car is sufficient for thinking of a Rolls-Royce. Such sloppiness deserves to provoke a degree of Socratic dissatisfaction, whatever problems TV may carry with it.⁸² We ought therefore, as partisans of Socrates or even of common sense, to be wary of abandoning TV’s rigorous principles in favour of a cheerful appeal to looser conceptions of what it takes to think of an object.⁸³

References