Preferred Citation: Dorter, Kenneth. Form and Good in Plato's Eleatic Dialogues: The Parmenides, Theatetus, Sophist, and Statesman. Berkeley:  University of California Press,  c1994 1994.


Form and Good in Plato's Eleatic Dialogues

The Parmenides, Theaetetus, Sophist, and Statesman

Kenneth Dorter

Berkeley · Los Angeles · Oxford
© 1994 The Regents of the University of California

To my mother and to the memory of
my father

Preferred Citation: Dorter, Kenneth. Form and Good in Plato's Eleatic Dialogues: The Parmenides, Theatetus, Sophist, and Statesman. Berkeley:  University of California Press,  c1994 1994.

To my mother and to the memory of
my father


The four dialogues examined here form a natural group with sequential concerns. Since the aim of the present study is to try to understand the group as a whole, I have sacrificed the advantage of greater detail that book-length commentaries would provide, in order to present a more synoptic picture. But although the treatment of individual dialogues will not be as extensively detailed as in book-length studies, I have tried to pay careful attention both to the conceptual arguments and to the dramatic and literary events, and have tried to ensure that the lessening of detail would not mean a lessening of attentiveness.

I call this group of dialogues Eleatic, as a convenient inclusive term, even though the term is only indirectly applicable to the Theaetetus . Unlike the other three dialogues, the Theaetetus is conducted neither by Parmenides nor the Eleatic stranger, and its subject matter is Heracleitean and its dramatic context Megarian (owing to the choice of Eucleides and Terpsion as the introductory speakers). Nevertheless, Parmenides is mentioned at an important juncture as someone whose views ought to be considered as an alternative to the philosophy of becoming that Theaetetus defends without success (180d-181a), and the failure to discuss the Eleatic philosopher is ascribed only to the fact that he is too important to be considered in the available time (183c-184a). He is thus the only alternative indicated to the theories that founder in the Theaetetus , and their collapse may amount to an indirect endorsement of the Eleatic. Nothing is presupposed philosophically by desig-


nating the dialogues in this way. My intention is rather to avoid the presuppositions involved in calling them Plato's "critical" dialogues, as is often done on the assumption that they are partial repudiations of the theory of forms. The latter designation is in any case misleading with regard to the Statesman .

Some of the material has been derived from the previously published studies listed below, and is used with the permission of the editor or publisher. "Justice and Method in Plato's Statesman " (S. Panagiotou, ed., Justice, Law and Philosophy in Classical Athens [Edmonton: Academic Printing and Publishing, 1987] 105-22); "The Theory of Forms and Parmenides I" (J. Anton and A. Preus, eds., Essays in Ancient Greek Philosophy , III: Plato [Albany: SUNY Press, 1989] 183-202); "Diairesis and the Tripartite Soul in the Sophist " (Ancient Philosophy 10 [1990] 41-61); "Levels of Knowledge in the Theaetetus " (Review of Metaphysics 44 [1990] 343-73).

I would like to thank the Social Science and Humanities Research Council of Canada for a one-year research grant, and the University of Guelph for a sabbatical year, both of which helped immeasurably in the preparation of this book. I would also like to thank Thomas Chance, G. R. F. Ferrari, Mitchell Miller, and the late Robert Brumbaugh for their generosity in reading through the work, in whole or part, and for their valuable comments and criticisms; and my students and colleagues, at Guelph and at other departments and conferences where I gave papers on this material, who have often made helpful comments and suggestions.



1. Questions of Chronology


In an earlier study of Plato[1] I tried to come to terms with the nature of his "middle period" philosophy, that is, with the themes for which he is best known, such as the theory of forms, doctrine of recollection, immortality of soul. The present study is the product of an attempt to understand the significance of the set of four dialogues that are usually thought to follow this period, and that pursue philosophy in a way apparently different from that of their predecessors. Whether they were written consecutively is not known, but they are evidently meant to be read as a set: the beginning of the Statesman refers back to the Sophist , the beginning of the Sophist to the Theaetetus , and the Theaetetus to the Parmenides . The latter reference is less explicit than the others. Socrates remarks, "I met the man [Parmenides] when I was quite young and he quite old" (Theaetetus 183e). Since the meeting is unlikely to have actually taken place, the reference would seem to be to the dialogue Parmenides . Many scholars are not convinced by this, and prefer to take the Theaetetus as the earlier of the two, because its aporetic nature more closely resembles Plato's earlier dialogues, while the technical nature of the Parmenides more closely resembles the later ones. Regardless of the order in which the Parmenides and Theaetetus may

[1] Plato's Phaedo: An Interpretation (Toronto: University of Toronto Press, 1982), hereafter PP .


have been written, however, Plato leaves us in no doubt that the Theaetetus is meant to be read as the first member of a trilogy, followed by the Sophist and Statesman : at the beginning of the Statesman (258a) Socrates remarks, "I myself had a discussion with Theaetetus yesterday [the Theaetetus ], and now I have listened to him replying [the Sophist ]." In any case, nothing of substantial importance will depend on the order in which we read the Parmenides and Theaetetus .

We cannot pursue such a study today without becoming part of the forty-year-old debate, as initiated by G. E. L. Owen, about whether Plato radically changed his views at this time—although that question is only incidental to the purpose of these studies. No observations about the development of Plato's philosophy are completely uncontroversial, but some are comparatively so. There is a general though not universal consensus that in the dialogues considered to be early, Plato seems concerned primarily to investigate the nature of the virtues and of other ways of being, especially those that we would call values;[2] that the dialogues considered to belong to his middle period ground this concept of "natures" on the ontological model of the doctrine of separate forms; and that in the "Eleatic" dialogues under consideration here, which are attributed to his last period, this theory of forms itself is brought into question in some sense.

The first part of the Parmenides appears to attack the theory outright, although the significance of this attack is rendered ambiguous by Parmenides' concluding remark that if one does not accept some such theory, "he will not have anything on which to fix his thought . . . and in this way he will utterly destroy the power of discourse" (135b-c). The Theaetetus , for whatever reason, then proceeds to all but ignore the theory of forms in its attempt to define knowledge. This is followed by the Sophist , in which something like a theory of forms is employed, but forms that are very different from those of the middle dialogues. Instead of focusing on values, it focuses on "kinds," which are explicitly said to be value-neutral (227a-b). The Statesman then continues this approach and applies it to political questions. It might seem from this, as many commentators conclude, that Plato has repudiated his "middle period" theory of forms and replaced it with something more Aristotelian. Not only do the "kinds" of the Sophist seem very dose to the

[2] "Value" is a useful term in discussing Platonic philosophy, as long as we do not take it as following from "to value," and therefore as something grounded in individual subjectivity. For Plato it can only mean something inherently good—a source of value rather than a consequence of valuation.


Aristotelian conception of categories, but the particular ones singled out for first mention (being, rest, and motion) are three of the most important categories for Aristotle as well.

There are obstacles, however, to this revisionist interpretation of Plato's Eleatic dialogues as a transformation of his metaphysics into one that is proto-Aristotelian. In no ancient source is there ever any suggestion that Plato changed his views in a radical way. Not only does Aristotle, for example, always write as though Plato consistently defended the theory of forms throughout his life, but he always writes as though Plato's position needs to be continually attacked. Had Plato meanwhile surrendered on the question of the separation of forms, this continued assault would make no sense. Aristotle had only to point to Plato's capitulation as evidence of the untenability of the theory. At the beginning of Metaphysics M.4 Aristotle suggests that the theory of forms was not at first connected with mathematics, but that such a connection was subsequently drawn. If Aristotle was in a position to mention a development so comparatively slight as this, he certainly would have mentioned one that was much greater and very agreeable to him. Neither does Diogenes Laertius, that repository of anecdotes of every stripe, provide the slightest hint of such an occurrence. It is hardly credible that if one of the two greatest thinkers of ancient Greece capitulated to the criticisms of the other, no rumor of that momentous event would have reached those ears. This argument from silence is not the only obstacle to the revisionist view. A second obstacle lies in the internal implications of the Eleatic dialogues themselves. We shall see in the subsequent chapters that these dialogues contain no persuasive evidence that Plato repudiated any significant aspect of the theory of forms, and indeed strong evidence that he retained the theory in all its essentials.

A third obstacle to the revisionist thesis is external to those dialogues. It involves the Timaeus and Philebus , both of which were thought to be later than the Eleatic dialogues, and both of which defend a theory of forms seemingly similar to that of the middle dialogues. Revisionists counter the threats that these dialogues pose to their thesis in complementary ways. They acknowledge the lateness of the Philebus but not the separation of its forms,[3] and they acknowledge the separation of the Timaeus's forms but not the lateness of its date.

[3] See, for example, Roger Shiner, Knowledge and Reality in Plato's Philebus (Assert: van Gorcum, 1974); and Kenneth Sayre, Plato's Late Ontology (Princeton: Princeton University Press, 1983).


There are numerous passages in the Philebus that recapitulate earlier dialogues' assertions about the theory of forms:

First we must consider whether such unities truly exist. Then, since each of them is one and always the same, admitting neither generation nor destruction, how they can nevertheless permanently be this unity. After these, in the things that come to be and that are unlimited, whether we are to assume that it is dispersed in them and has become many, or whether it is whole and apart from itself (which might seem to be most impossible of all), being the same and one, coming to be simultaneously in one and many.

The knowledge of being and reality and what by nature is always completely in itself, I believe that everyone in whom even a little reason is present will consider this to be by far the truest knowledge.

And one knowledge differs from another: the one looking toward things that come to be and are destroyed, the other toward what neither comes to be nor is destroyed but that exists eternally in itself and the same . . . [such as] justice itself. [These two kinds of knowledge are called human and divine, respectively.]

Now, then, the power of the good has hidden itself from us in the nature of the beautiful. For certainly moderation and proportion completely correspond to beauty and virtue.

The repeated disjunction between what is in itself, unchanging, eternal, rational, good, and divine, and what comes to be and passes away, changes, and is characteristic of what is human rather than divine, repeats all the essentials of the middle period theory of forms. Not all the points are made with the same emphasis and detail as in the Phaedo and Republic , but that would hardly be surprising if Plato. had no reason to expect his readers to think that he had abandoned his earlier views. Nevertheless, the lack of detail in his presentation permits advocates of the revisionist thesis to question how conclusively and completely the theory of forms is reaffirmed here.

So the focus of the controversy reverts to the date of the Timaeus — where the affirmation of the ontological difference between forms and individuals is entirely unambiguous—and whether that dialogue can be shown to be later than the Parmenides . The dating of the dialogues has been a perilous enterprise. Earlier accounts, such as Schleiermacher's,[4]

[4] Introductions to the Dialogues of Plato (Cambridge: Deighton, 1836).


that were based on a view of how Plato's thought ought to have developed, obviously have no power as independent evidence for a particular view. The dialogues themselves give us occasional scraps of evidence when one dialogue presupposes prior knowledge of another, as in the case of the Eleatic dialogues, or in the case of the Timaeus , which recalls the Republic . But this does not take us far, and in particular it does not establish whether the Timaeus with its unambiguous affirmation of the separation of the forms is earlier or later than the Parmenides , which questions their separation.

The only other due that has come down from antiquity is Aristotle's statement that the Laws is later than the Republic ,[5] and Diogenes Laertius's remark, "Some say that Philip of Opus transcribed the Laws , which were in wax."[6] The latter is generally interpreted to mean that the Laws was the only dialogue that Plato himself was unable to see through the publishing process before his death, and that it must therefore have been his last. This is a plausible inference but hardly an infallible one, especially in view of the uncertainty of the story itself ("Some say . . .").[7] It is in any case the closest thing that Platonic stylometry has to a Rosetta stone. If one assumed, on the basis of that evidence, that the Laws was the last dialogue, other dialogues that show a strong stylistic affinity with it might be presumed to come from the same period. The search was on for measures of stylistic affinity. Candidates that were found included reply formulas (the responses of the interlocutors— useless, however, in the case of a narrative like the Timaeus ), clausula rhythms (the endings of periods or colons), avoidance of hiatus (following a word ending in a vowel with one beginning in a vowel), and use of hapax legomena (unique appearances of words) or unusual words. But each of these encounters difficulties in measurement. In measuring reply formulas do we take into account the personality of the interlocutor and the nature of the questions being asked? And do we count slight variations as being the same; or formulas imbedded within longer sentences in the same way as isolated formulas? Do we count clausulae only in long periods, which allow more freedom for rhythmic variation, and, if so, how many syllables must the period be in order to get

[5] Politics 2.6.1264 26-27.

[6] Lives of the Eminent Philosophers 3.3.

[7] Jacob Howland gives a careful, scholarly argument to the effect that there is neither reliable evidence on which the chronology of Plato's writings could be established, nor any reason to believe that such a chronology would enhance our understanding of Plato's philosophy ("Rereading Plato: The Problem of Platonic Chronology," Phoenix 45 [1991] 189-214).


counted? Do we count "unobjectionable" (removable by elision, apostrophe, or crasis) hiatus in the same way as "objectionable" hiatus, and, if not, what differences in weighting should be assigned? Is hiatus affected by punctuation, and does it therefore matter which edition is used? In counting unusual words, do we count only the first occurrence or all occurrences, and, if only the first, does the length of the dialogue need to be taken into account at all? In addition to all these decisions, which considerably limit the possibility of scientific objectivity, we must also decide whether to take into account the nature and subject matter of the dialogues. Should we expect to find the same stylistic features in a narrative myth (Timaeus ), an exercise in abstract dialectic (Parmenides —which is so anomalous as to have frequently been dismissed by stylometrists as spurious), or a set of speeches (Symposium ), as in dialogues like the Republic, Theaetetus , or Laws?

In view of all these variables, and of the further complication that several investigators relied on flawed assumptions or on data obtained by faulty counting (whether their own or that in Ast's Lexicon Platonicum ), it is remarkable that in fact a broad consensus emerged from the various investigations. There is considerable agreement, for example, that five other dialogues were written in the same period as the Laws , namely, the Timaeus, Critias, Sophist, Statesman , and Philebus . It is usually assumed that the somewhat abrupt change of style, in several respects, at this point in Plato's career was a result of the impression made on him by Isocrates' treatise on rhetoric.

Using Leonard Brandwood's astute and painstaking survey as a guide,[8] I have summarized in the following chart the results of the investigations as they bear on the question of the relative date or period of the Timaeus and Parmenides . Since the number of periods identified by various investigators varies, all periods other than the first and last have been amalgamated into a "middle" period in order to facilitate comparisons. Where a dialogue is included in the last period, the numbers in parentheses refer to the place assigned within the period, and the number of dialogues assigned to that period, so that "4/6" means "fourth out of six." Of the twenty that I have summarized, Brandwood concluded that the most important are Ritter, Janell, Kaluscha, and Billig (p. iii), while the investigations of Frederking, Kugler, Siebeck,

[8] "The Dating of Plato's Works by the Stylometric Method—A Historical and Critical Survey," 2 vols. (dissertation, University College, London, 1959; Brandwood's study has recently been published as The Chronology of Plato's Dialogues [Cambridge: Cambridge University Press, 1990]). I have omitted only those studies that were inconclusive with regard to both the Parmenides and Timaeus : i.e., P. Droste (1886), E. Walbe (1888), H. Kallenberg (1913), and C. Ritter (1935).


Lutoslawski, Natorp, von Arnim, and Fossum "proved to be of no value for various reasons" (p. 398 n.).





Chief criterion





L. Campbell




unusual words




late (4/6)

P. Blass






[not dated]


late (4/6)

W. Dittenberger






late (1/7)[9]


late (3/7)

A. Frederking









F. Kugler








late (7/7)[11]

M. Schanz




metaphysical terms


[not dated]


late (3/4)

H. Siebeck




reply formulas




[not dated]

C. Ritter








late (4/6)

J. Tiemann






[not dated]


late (1/6)

T. Lina






[not dated]


late (4/6)

G. B. Hussey




terms for




late (4/6)




H. von Arnim




reply formulas




[not dated]

C. Baron









W. Lutoslawsky




500 characteristics




late (4/6)

P. Natorp




1,949 words




late (3/6)

G. Janell









W. Kaluscha




clausula rhythm




late (1/6)

H. von Arnim




reply formulas


late (7/12)


[not dated]

L. Billig




clausula rhythm




late (1/6)

A. Fossum




hapax legomena





[9] But so anomalous with other late dialogues that Dittenberger questioned its authenticity (Brandwood, "Dating" 36).

[10] This result was obtained by Frederking in order to discredit the method, but Brandwood shows that the problem is rather with Frederking's employment of the method.

[11] In the case of the Parmenides different indicators led to different results, from which Kugler tended to doubt the dialogue's authenticity. The seventh "dialogue" added to the set of late dialogues is Republic 8. Brandwood ("Dating" 56) wonders whether Kugler, who was a disciple of Frederking's, was not trying to discredit the method by producing, in some cases, bizarre results (e.g., placing the Crito after Republic 7 and immediately before the Sophist ).

[12] But considered inauthentic by Ritter.


Brandwood concluded that one could separate the Timaeus, Critias, Sophist, Statesman , and Philebus

from the rest to form a final group with the Laws . This was possible because Plato's style in these works differs greatly from that in the preceding works, the change being sudden and abrupt. There would seem to be little doubt that for some reason Plato decided at this point to embrace the precepts of the Isocratean school of rhetoric . . . . Whilst the criterion of hiatus by itself effectively distinguished the six works . . . , that of clausula rhythm produced the order indicated.
     (pp. 399-400)

All the investigations concluded that the Parmenides precedes the Timaeus , but in 1953 G. E. L. Owen published an influential paper in which he argued for the reverse order.[13] He claimed, first, that the clausulae are a better indicator of the sequence of writing than hiatus, in which case the Timaeus does not appear to be among the final three dialogues. Moreover, even though it still appears to be later than the Parmenides , Owen suggested that because of its unusual character Plato may have been led to a different style of writing, which he only later transferred to conversational dialogues (pp. 315-16). What lay behind Owen's challenge to the traditional dating was his conviction that the Parmenides-Theaetetus-Sophist group form a set of "critical" dialogues in which certain features of the middle period's theory of forms (especially their status as paradigms) are repudiated. In that case it would not make sense for Plato to have subsequently written a dialogue like the Timaeus , in which these features are reaffirmed.

Four years later Harold Cherniss replied in detail to Owen's arguments[14] in order to counter their growing influence, and the debate continues today.[15] Brandwood, writing two years after Cherniss, rejected the latter's arguments for placing the Timaeus among the final three

[13] "The Place of the Timaeus in Plato's Dialogues," Classical Quarterly , n.s., 3 (1953) 79-95; reprinted in R. E. Allen, ed., Studies in Plato's Metaphysics (London: Routledge and Kegan Paul, 1965) 313-38. Page references will be to the Allen edition.

[14] "The Relation of the Timaeus to Plato's Later Dialogues," American Journal of Philology 75 (1957) 225-66; reprinted in Allen, ed., SPM 339-78. Page references will be to the Allen edition.

[15] In the past decade, for example, Kenneth Sayre has defended something like Owen's position (PLO appendix B), and William Prior has defended something like Cherniss's position (Unity and Development in Plato's Metaphysics [LaSalle: Open Court, 1985] appendix II). An anomalous position is that of Henry Teloh, who, like Owen and Sayre, maintains the revisionist view, but also accepts Cherniss's arguments for the lateness of the Timaeus , and is consequently uncertain how the two may be reconciled (The Development of Plato's Metaphysics [University Park: Pennsylvania State University Press, 1981] 209-18).


dialogues, but nevertheless regarded it as definitely belonging to the final group of six, and therefore later than the Parmenides . There is no need to discuss here the details of the respective arguments, because the real issue, as Owen showed, is not statistics but philosophy. If the Eleatic dialogues make most sense as a rejection of the earlier theory of forms, then it will be possible to find a loophole in the imprecision of stylometric statistics (or in the initial assumptions) through which one can escape from the late dating of the Timaeus .

This discussion of the problems of dating the Timaeus , in particular, seems to imply my acceptance of the prevailing view about the order in which the other dialogues were written, despite the cautionary note sounded at the beginning. That question, however, will not be an issue here. When it is convenient I shall continue to use the generally accepted sequence as a frame of reference, in spite of its attendant uncertainties, but nothing of substance will be at stake in my so doing. The only sequence that will be relevant to my interpretation is that of the four dialogues examined here (especially the trilogy), and the internal evidence mentioned above gives ample support to that order.

2. Principles of Interpretation

As is often the case with-debates about interpretation, an insuperable obstacle to resolution is the fact that our very perception of the evidence is influenced by our prior beliefs. If we believe that Plato's original formulation of the theory of forms is seriously flawed—an exploratory, unsatisfactory stage on the way to Aristotle's categorial analysis, for example—then we will welcome any evidence that Plato himself came to hold this view, welcome it not only as a confirmation of our own judgment, but also because it would enable us to see Plato in a more favorable light. We would regard this as the most charitable way of interpreting him. The fact that Plato dearly raised serious questions about the theory of forms in one dialogue, and then relegated it to (at best) the background in the next three dialogues, might readily be interpreted in this way. On the other hand, those who have a higher regard for the theory of forms see no reason to expect Plato to change his mind about it, and would accept such a negative interpretation of the Eleatic dialogues only if less drastic interpretations were not possible. The factors that give support to either interpretation are never fully explicit or entirely free from ambiguity (little in Plato is), and so both sides consider themselves supported by the textual evidence.


A second obstacle to complete objectivity in our evaluation of the evidence—which greatly influences the first—lies in our differing conceptions of the nature of philosophy generally, as understood both by Plato and by ourselves. Those who believe that there is a perennial dimension to philosophy, and who think that such timeless concerns were central to Plato's enterprise, will have less reason to expect substantial changes in (as opposed to progressive deepening of) the fundamentals of his thought during the course of his career; they will be inclined to give more weight to the evidence for continuity in his thought than to the evidence for radical discontinuity. For someone, on the other hand, who believes that the primary concern of philosophy is in an area characterized by historical progress, such as descriptive logic, it would not seem condescending to regard Plato's work (like that of Archimedes) as extraordinarily incisive by the standards of his day, but superseded by developments in the intervening millennia—and for that matter continually superseded by his own further progress in the course of his career. Consequently, even though revisionists may approach Plato with considerable respect, the fact that they see him as a pioneer in an enterprise that leads naturally to a more Aristotelian position—and ultimately toward their own analytic philosophy—is likely to cast him as an only partially successful investigator into their own techniques.[16] It is only to be expected, therefore, that he is liable to make mistakes, or to fail to grasp problems, that to us may seem obvious. Thus Richard Robinson suggests that in the Theaetetus Plato "still thought he believed in [the theory of forms], though in his active inquiries he was in fact beyond it, and it functioned as a theory to be criticised instead of as the rock of salvation it had been in his middle period."[17] And Kenneth Sayre, one of the most recent and most prominent defenders of the revisionist view, argues that in the first part of the Parmenides ,[18]

given the arguments as we find them, it would be quite unreasonable to expect these arguments themselves to constitute Plato's reason for wanting to replace this particular theory . . . . [Rather,] he is drawing attention to certain difficulties involved in the theory—but doing so dramatically rather

[16] As Cherniss puts it: "The Analysts of Oxford have succeeded to their own satisfaction in reading the dialogues that they call 'critical' as primitive essays in their own philosophical method. The author of these works, they feel, they could adopt as their worthy precursor, if only he could be absolved of the embarrassing doctrine of ideas that he elaborated in all its metaphysical and epistemological absurdity in the Phaedo , the Symposium , the Republic , and the Phaedrus " ("Relation" 347).

[17] "Forms and Errors in Plato's Theaetetus " (Philosophical Review 59 [1950] 3-30) 19; emphasis in original.

[18] PLO 22, 25.


than dialectically. And the reason Plato chooses this modality of presentation is that the difficulties have to do with aspects of the theory which had remained unintelligible in earlier dialogues, and hence could not even be articulated in the dear manner required for dialectical discussion.

Third, even those who agree about the value of the theory of forms, and about the central concerns of philosophy, may disagree about the degree to which we ought to give Plato the benefit of the doubt when we interpret recalcitrant passages. When we come upon a troublesome section, should we assume, as a (fallible) heuristic principle, that Plato very likely knew what he was doing, and that we therefore need to work harder to see where he is leading us? Or should we assume that he was more likely confused himself, the victim of a rudimentary stage in the development of analytical logic? In short, are the difficulties in the text more likely to arise from a depth and subtlety in Plato's thinking or from a lack of sophistication and self-awareness?

There are obviously substantial correlations among these three sets of predilections. Those who take the theory of forms seriously, as an insightful representation of the primacy of the intelligible and timeless over the corporeal and changing, are likely to believe that the ultimate focus of Plato's philosophy is not susceptible of historical progress, and are likely to read Plato with great respect. Those, on the other hand, who believe that the theory of forms is a perhaps brilliant but nevertheless rudimentary first statement of the principles of categorial analysis are likely to believe that Plato's subject matter is inherently capable of continual refinement, and that Plato is likely to have encountered conceptual difficulties that we are in a much better position to detect and resolve. Nevertheless, it is not always the case that unitarians read Plato in a more charitable way than the revisionists do. No one could approach Plato with more respect than does Myles Burnyeat, for example, in his carefully analytic commentary on the Theaetetus , although Burn-year dearly believes that some kind of revision of the metaphysics of the Republic is taking place.[19] On the other hand, W. G. Runciman has shown by his own example that it is possible to be a unitarian by having a lower regard than most revisionists for Plato's acumen. Runciman believes that Plato did not modify his position in the direction of Aristotelianism precisely because he lacked the sophistication to be able

[19] "We could well read the Theaetetus as a long meditation on that brief passage of the Republic . . . . A meditation of such length and complexity does not leave its topic unchanged. It is no use going back to the Republic now" (The Theaetetus of Plato [Indianapolis: Hackett, 1990] 238-39).


to do so. Thus, Plato did not achieve "in the Sophist a conscious sophistication such as must entail a radical revision . . . of his earlier tenets."[20] And

if only Plato had not been mistaken about the nature of his subject-matter, he would have broken through to an astonishingly sophisticated understanding of the structure and function of language . . . . It is at least tempting to ascribe the prevention directly to Plato's dogged retention of the Theory of Forms, despite the criticisms of his Parmenides and the different method of approach which succeeded them.
     (p. 126)

My own view is that an interpretation in which Plato emerges as one who has important things to teach us about the most fundamental matters of philosophy is preferable to one that is forced to conclude that he does not (except by the example of the limitations of his success). The interpretation to which the present investigations lead will reject the strongly revisionist view that the Eleatic dialogues are a retreat from some of the more distinctive features of the theory of forms, and are a subsequent salvaging of what remains. But it will also reject the diametrically oppposed view, that Plato is not completely serious about the problems he raises in the Eleatic dialogues about the theory of forms, and that the objections are maintained only from a point of view that is not Plato's own. It will argue instead that in the Parmenides Plato is indeed raising serious problems about the middle period theory of forms, but that the problems are not such as can be resolved by a revision of the theory. The problems are symptomatic of serious limitations in the theory, but no other theory is possible that avoids fundamental limitations of one kind or another. On the contrary, the dialogues after the Parmenides show why the theory must be espoused in spite of its limitations. In accordance with what I said above, I entertain no illusion about the possibility of "demonstrating" these claims in a conclusive way. It is possible nevertheless to show that such a reading is completely consistent with what takes place in these dialogues, and that it explains many passages more readily than do other interpretations.

Although in considering these dialogues no one can remain entirely untouched by the polemics that have been swirling around them for several decades, those questions are only incidental to my project here. My main concern is not to prove or disprove any theses about Plato's development, but to learn from Plato as much as I can. I hope that the

[20] Plato's Earlier Epistemology (Cambridge: Cambridge University Press, 1962) 125.


details of the interpretations will have some value in themselves, as attempts to understand the thought of a great thinker, apart from whatever value they may have as evidence for a particular thesis about Plato.

3. Method of Hypothesis and Method of Division

In the dialogues believed to belong to his "middle period," especially the Phaedo and Republic , Plato puts forward the method of hypothesis as an indirect way (

) of discerning the highest truth when we cannot apprehend it directly. According to the Phaedo (100a, 101d-e), the method has two steps: (1) when we posit a hypothesis or theory, we examine its consequences to see whether they lead to harmonious results; (2) then we look for a higher hypothesis until we find one that is satisfactory. According to the Republic (6.511b), this point is not attained until we reach the unhypothetical principle, the Idea of the good.

In the Phaedrus (265c-266b) Plato speaks of another kind of method, which he does not deploy until the Sophist and Statesman : the method of division. According to that method as it is used in the Sophist , we search for the essence of something by beginning with the highest form or essence, then progressively bisecting this into ever smaller species, until we discover what we are looking for as a part in relation to the whole.

What is the relationship between these two methods? Interpreters who approach Plato from the direction of analytic philosophy naturally prefer the second, which, in a manner that anticipates certain features of their own practice, analyzes the nature and structure of class relationships. The method of hypothesis, on the other hand, operates synthetically and speculatively, which may be regarded as a departure from strict logical rigor. It is therefore reasonable to suppose, from the analytic point of view, that Plato himself came to recognize the superiority of the analytic method of division over the synthetic method of hypothesis; and the fact that the method of division was promulgated after that of hypothesis gives some support to this view. The Sophist , which is the first dialogue to make overt use of the method of division, is indeed almost completely analytic. The stranger from Elea, who conducts the investigation, makes no attempt to search for "higher" truths; he simply follows his starting point "down" to its results.

Of course both methods move in both directions. The first part of the method of hypothesis, the deductive moment, goes from higher to


lower; while the second movement, the speculative search for a higher hypothesis, goes in the opposite direction. In the case of the method of division, the division itself is preceded by a collection, in which individuals and species are brought together under a higher genus, so it too proceeds in both directions. There are two important differences between the methods, however. First, in the method of hypothesis the lower is for the sake of the higher: the downward, analytic path leads to the upward, speculative one, for the goal is the unhypothetical principle from which everything else follows. In the method of division, by contrast, the higher is for the sake of the lower: synthetic collection leads to analytic division, for the goal is the definition of a lowest species. Second, because the hypothetical principle is the Idea of the good, the method of hypothesis is ultimately teleological: it seeks the ultimate "why" and is rooted in the ground of value. The method of division, on the other hand, not only refrains from inquiring into the nature of the good, the highest principle, but the Eleatic stranger expressly warns that his method makes no distinction between greater and lesser value:

The method of definitions [

] does not care more or less about sponging than about taking medicine, nor whether one provides us with greater or smaller benefits than the other. It aims at acquiring an understanding of what is akin and what is not akin in all the arts, and, with this intention, it honors all of them equally.

The method of division, then, is very amenable to analytic philosophy, not only because its emphasis is on analysis rather than synthesis, but also because its abstraction from questions of value shares with analytic philosophy a resistance to any attempt to derive values from facts, the ought from the is. In the Statesman , however, Plato once again engages in philosophical speculation about the highest principle of value, and it is unfortunate that this dialogue is almost always ignored by those who study the Theaetetus and Sophist , even though it is the explicit sequel to them. The prevailing tendency to attach great importance to the Theaetetus and Sophist while ignoring their sequel—in other words, to treat what is in fact the beginning and middle of a trilogy as if they were complete and conclusive in themselves—has led to a one-sided perception of the method of division.

The Statesman , as if in accordance with the methodology of division itself, is bisected down the middle. The first half superficially resembles the Sophist with its elaborate bisective divisions, and at 266d the stranger even repeats his earlier injunction against paying attention to


distinctions of value. The second half, however, not only explicitly avoids the technique of bisective division but makes constant reference to value distinctions; in fact the statesman is defined as the one who knows the best thing to do in any particular situation (304a-305d). For some reason, then, in the Statesman the method of division leads to the same kind of axiological ontology that the method of hypothesis leads to in other dialogues.

To understand this tension between the way the method of division is employed in the two dialogues, we must look at the progression of the Eleatic tetralogy as a whole, comprising the Parmenides, Theaetetus, Sophist , and Statesman . In the first of these dialogues Parmenides raises serious questions about the theory of forms, although it is clear that his attack is not intended as an outright refutation since he stipulates in conclusion that unless one does accept some such theory, "he will not have anything on which to fix his thought . . . and in this way he will utterly destroy the power of discourse" (135b-c). Nevertheless Parmenides' arguments show that the theory of forms is problematic because it must constantly fall back on metaphor and analogy: the limitations of human conceptualization make this inevitable. In that case, how can the theory of forms be justified in the face of conceptual aporiae such as are pointed out by Parmenides? The only means available is the method of hypothesis, and in fact the trilogy that follows the Parmenides can be read as an extended application of the method of hypothesis in defense of the theory of forms. Cornford attributed a function of this kind to the Theaetetus , which never brings in the theory of forms at all, when he wrote: "The Forms are excluded in order that we may see how we can get on without them; and the negative conclusion of the whole discussion means that, as Plato had taught ever since the discovery of the Forms, without them there is no knowledge at all."[21] The failure of the Theaetetus's hypothesis, that knowledge is based on the perception of particulars, demonstrates the need for a higher hypothesis on which the residual questions of that dialogue can be answered. The theory of forms, although never appearing explicitly in the Theaetetus , will be visible in the background in a number of ways.

Now, the theory of forms comprises two fundamental claims. First, forms are universals, instances of "sameness" in reality: it is by virtue of the prior reality of a form that a multitude of individuals may be

[21] Plato's Theory of Knowledge (London: Routledge and Kegan Paul, 1935) 28.


identified by the same name.[22] Second, forms are values: they are articulations of the goodness of reality, and therefore the most fundamental forms are the good, beautiful, just, and so forth.[23] What Plato does after the Theaetetus is to reintroduce each of these two aspects sequentially. In the Sophist he shows that by reintroducing the concept of universal kinds in abstraction from the concept of value, we can solve the fundamental aporiae of the Theaetetus .[24] However, the Sophist , although not explicitly aporetic itself, nevertheless ends in a most unsatisfying and unconvincing way. The sophist and philosopher are not distinguished from each other in terms of their values, as had been the case in previous dialogues, but in value-free technical terms: the philosopher is a maker of accurate images, the sophist a maker of inaccurate ones. The distinction is transparently inadequate, for it fails to distinguish love of wisdom from wisdom itself, a distinction that Plato elsewhere treats as fundamental, and so a philosopher becomes identical to a wise person; on the other hand, philosophers who in a moment of ignorance produce an inaccurate image become by definition sophists. On the Sophists definition there is no lover of wisdom—there is either a wise person or a sophist. In the case of sophists, too, the definition fails to distinguish between their products and their intentions, so that sophists who happen to get the facts right about something, that is, to produce an accurate image of it, are philosophers by definition, regardless of the purpose to which they put that image. Sophists cannot be distinguished adequately from philosophers without reference to their goals, that is, to the concept of value.

Accordingly, the next dialogue, the Statesman , responds to these disharmonious consequences with a higher hypothesis; that not only are there universals, but that underlying everything is the good, the "mean." In the light of that concept, the stranger will explicitly dismiss the importance of the distinction between accurate images and inaccurate ones that he had made so much of in the Sophist . Anyone who produces images at all, rather than directly embodying the mean, he calls "the

[22] Cf. Republic 10.596a: "We are in the habit of positing one form with regard to each group of things to which we apply the same name."

[23] Cf. Republic 6.509b: "To the things known [i.e., the forms] not only their know-ability, you must say, is provided by the good, but even their being and essence comes to them from it."

[24] The Theaetetus foundered in an attempt to give an account of the nature of knowledge, stymied in the end by its inability to find a satisfactory' definition of logos, in particular with reference to opinion, doxa. In the Sophist , however, the stranger (who calls his method the method of logoi : 227a) does give a definition of logos, and in relation to doxa (261d-263e).


sophists of sophists" (303b-c). The real difference between the sophistic pretender to wisdom and the genuine possessor of wisdom will be that the latter alone possesses the science of the mean, that is, of particular goodness . Subsequently, in the Philebus and Timaeus the theory of forms will reappear with renewed vigor, and very much in the service of the question of value.

At the same time, the Statesman counters the more radical consequences of the Parmenides ' demonstration of the limitations of metaphorical and analogical thinking, by defending the importance of metaphorical and analogical "paradigms." The underlying problems remain intact to which the criticisms of the Parmenides had pointed, but the limitations of our ability to dispense with them now become more evident.

Plato's belief that ontological thinking can transform us not only intellectually but also morally is supported by a long tradition in both Western and Eastern philosophy. But in the last century and a half this tradition has become eclipsed. Among the most influential movements of the twentieth century there is virtually unanimous agreement that there is no ontological basis for value. On one hand seminal thinkers as diverse as Heidegger, Weber, and Moore have argued (like the Eleatic stranger in the Sophist ) that some kind of nonarbitrary knowing is possible only when one distinguishes beings from values: we can have fundamental knowledge of ontology, facts, or the "is," but not of ethics, values, or the "ought." On the other hand, Nietzscheans and pragmatists, who (like Protagoras in the Theaetetus ) tend to reject this dichotomy altogether, subordinate the ontological dimension to that of value, so that the former has its basis less in the order of things than in human valuation, the will. In both cases the possibility that ontological thinking may lead to moral or spiritual transformation has been eliminated. For those of us who find in this reorientation reason for serious concern, it is important to rethink the insights of someone like Plato, for whom the connection between ontology and moral goodness—the point of union between the divided line and the tripartite soul—is intensely evident.


Chapter One
The Parmenides

1. The Problem of the Parmenides

Plato's intentions are more enigmatic in the Parmenides than in any other dialogue. The first part depicts Socrates, in his youth, haltingly expounding his theory of forms to the aged Parmenides, who proceeds to demolish it with a series of five arguments, only to conclude that "on the other hand if anyone . . . does not admit the existence of forms of things or mark off a form under which each individual thing is classed, he will not have anything on which to fix his thoughts . . . and in this way he will utterly destroy the power of discourse" (135b-c).

For those who consider the theory of forms to be untenable, the dialogue is often welcomed as a sign that Plato himself came to see the error of his ways, and Parmenides' subsequent endorsement of the theory is interpreted as only a limited endorsement that does not extend to all the features of the theory that have come under attack. What remains, they suggest, is a modified theory that no longer postulates the forms as separate from things.

The dramaturgy of the dialogue, however, seems to point in the opposite direction. The Platonic Socrates who here as a youth falls victim to Parmenides' refutations is the same one whom Plato has previously portrayed as continuing to defend the theory in his maturity and even on his deathbed. We are even reminded of the future Socrates of the Republic by the presence of Glaucon, Adeimantus, and someone named


Cephalus (although not the same Cephalus who appears in the Republic ), none of whom serve any other function within the Parmenides itself. Cephalus, to be sure, is the narrator, but since he only repeats Antiphon's narration (127a) it would have been simpler to leave Cephalus (and Glaucon and Adeimantus) out altogether and have Antiphon—or better, Pythodorus, who was the original narrator from whom Antiphon learned it—narrate the story directly. It is as though Plato wanted to underline the gratuitousness of his inclusion of these names. If Plato wanted us to take Parmenides' arguments as decisive, it would be odd for him to make the victim someone whom he has already depicted as subsequently, and in greater maturity, unshaken in his adherence to the theory. It would make far more sense to cast one of Socrates' disciples (Phaedo, for example) as the defeated champion of the theory here (and someone like the Eleatic stranger as its refuter, since Parmenides would no longer be alive). But those who find the theory of forms too speculative for their taste tend to feel the same way about conclusions drawn from Plato's dramaturgy. After all, it is not inconceivable that Plato is simply giving Socrates a "fresh start" here.

The contrary interpretation, according to which Plato considered the theory of forms to be unscathed by these arguments, faces the difficulty of finding a plausible alternative explanation of the arguments, which do, after all, seem to be intended as refutations of the theory. To dismiss Parmenides' arguments as parodies of sophistry or as ironic jokes, as is sometimes done, is too easy. If we can pick and choose which of Plato's arguments to take seriously and which ironically, on as little evidence as this, there will be no escape from arbitrary and capricious interpretation. Moreover, in the dialogues immediately following the Parmenides —the Theaetetus, Sophist , and Statesman —Plato seems to be exploring alternatives to the theory, which suggests that he is indeed dissatisfied with it in some sense. One might reply that Plato, having said in previous dialogues all he had to say about the forms, is now exploring new avenues without renouncing the old; but in the Sophist he does speak as though being a "friend of the forms" is no longer an adequate position (246b ff.).

I shall argue that Plato intended these arguments to raise serious problems about the theory of forms, but that he did not consider the problems to be fatal. We shall see that the mature theory of forms has at its disposal the means to mitigate the devastating conclusions that Parmenides draws, and that Plato even gives us indications of where Parmenides' arguments may be assailed. But we shah also see that the objections cannot be dispelled entirely, so the question of how far Plato


considered the theory to be undermined by them admits of no simple answer. We shall have to take a detailed look at the arguments to see precisely where they lead us.

2. The Theory of Forms

In previous dialogues the forms have served a variety of functions, as causes, universals,[1] ontological paradigms, and referents of knowledge. Some readers have claimed that these functions are incompatible, and that the theory as a whole is incoherent. It is true that the functions are by no means identical, but they are nevertheless compatible and are even related in an order of conceptual entailment,[2] although I would not insist that the order of logical priority necessarily corresponds to the order of their discovery or of their importance in Plato.

1. The theory of forms is always linked to the primordiality of-goodness.[3] The nature of reality is not just a matter of chance or spontaneity, but a consequence of the fact that for Plato being is inseparable from value. The "forms" that reality takes are therefore not random but are consequences of what follows from the nature of goodness: "To the things known [the forms] not only their knowability, you must say, is provided by the good, but even their being and essence comes to them from it" (Republic 6.509b). Accordingly, the most important forms are always those that most obviously manifest value: goodness, beauty, virtue, justice, wisdom, piety, moderation, courage. The others must also be manifestations of goodness insofar as they manifest reality, but the connection is no longer explicit.

2. When Plato speaks of the forms as causes, as he does in the Phaedo , for example (100c ff.), it is dear that he does not mean that they actively produce an effect (efficient causality)—since the forms are not active. They are formal causes, whose power lies not in an activity on the part of the form, but in their capacity to be participated in by something else. Precisely what is meant by this is one of the aporiae of Platonic philosophy, but it is helpful to think of the forms as the possibilities of reality, possibilities of the kinds of things and qualities that can exist in accordance with the nature of reality. The Presocratics

[1] For convenience I use the term "universal" throughout to designate the Platonic "one over many." No post-Platonic connotations are intended.

[2] An earlier and less complete version of the following account was developed in PP 125 -26.

[3] Symposium 205e-206a; Phaedo 99c; Republic 6.509b; Phaedrus 245e; Philebus 13e, 15a; Timaeus 29e.


recognized that the material of which all things are composed must be eternal, and Plato recognized further that the possible forms that this material can take must in some sense be eternal as well. Understood in this way, forms are causes in the purely formal sense that a thing or quality of a certain kind could not exist if such kinds did not follow from the nature of reality. It is a necessary condition for the existence of a thing that such things be possible.

3. Now, if the forms are the possibilities of a kind of thing or quality coming into being, then each form will also be what all instances of that kind have in common, and it will accordingly have the character of a universal. This feature is not always emphasized by Plato (it is not evident in the Phaedo , for example), but it is explicitly mentioned at least in the Symposium (210b), Republic (596a), and Parmenides (132a), and was implicit in the early dialogues, in which Socrates searched for univocal definitions that exhibit the character common to all instances of a certain kind, a quest out of which the theory of forms may have developed.

4. Again, as universals, or the essence of what a number of things or qualities have in common, the forms abstract from everything extraneous and deficient, and so are purely and perfectly that essence. As such they are also paradigms of their kind. This will appear to be an unjustified transition if we think of a universal in purely logical terms such as a class, for a class is not a perfect instance of its members. On such a view the theory of forms will appear incoherent, and it is precisely this imputed incoherence on which the theory of forms is often thought to founder in the Parmenides . For Plato, however, the one that stands over the many and defines it is not a class but an independent ontological reality. As one over many it is universal, and as the bare essence itself it is the pure and therefore perfect quality. Conceived in terms of Plato's conception of reality rather than in terms of ontologies that are foreign to Plato's way of thinking, and based on very different assumptions, the forms' characteristics as universal and paradigmatic are not only compatible but necessarily connected.

5. Finally, as paradigms the forms are characterized by the same quality that particular things possess by participating in them, and are therefore in a certain sense (which will be clarified later) self-referential. Because they possess a determinate character they are in principle knowable—indeed, as paradigms they are the ultimate referents of all our knowledge, as Plato argues in the discussion of recollection in the Phaedo (74a-75a).


I have recalled these features of the theory because they seem to be what guided Plato in his choice of Parmenides' arguments, and perhaps even their order. (1) In the preliminary inquiry Parmenides attacks the valual aspect of the forms by asking whether there can be forms not only of noble things like the just, beautiful, and good, but also of "very ignoble" (

) things like hair, mud, and dirt. (2) The first argument attacks the forms as causes , asserting that the crucial concept of participation, on which the forms' causality is founded, is incoherent. (3) The second attacks them as universals , charging that an infinite regress results from this conception. (4) The fourth attacks them as paradigms , once again pointing to infinite regress, and (5) the fifth attacks them as objects of knowledge , arguing that such a conception leads to a gulf that makes the forms unknowable by us, and ourselves unknowable by the gods. The third argument alone deals with a claim about the forms that has not been part of the theory advanced in previous dialogues—the view that the forms are nothing but concepts—and we shall see that Plato accordingly treats this argument somewhat differently from the others.

Let us go through the arguments now in turn, to see what they prove and what Plato may have thought they prove. After the preliminary inquiry, which is not strictly an argument at all, we shall find that although Socrates answers most of Parmenides' questions with alacrity, he becomes tentative at a crucial step in each argument but the third, using the word

("perhaps") or
("so it seems"). This is no coincidence, for we shall see that in each case the step is a questionable one, and it therefore seems likely that the anomalous tentativeness of Socrates' replies is meant to call this to our attention. The third argument is the only one in which no such hesitation takes place, and it is also the only one whose target—conceptualism—was not part of the previously developed theory. The refutation of this un-Platonic conceptualism is, I believe, the only refutation that Plato fully endorses, and therefore the only one in which Socrates' agreement is never tentative.

3. Preliminary Inquiry (128e-130e)

The main part of the conversation begins when Socrates proposes to solve a paradox of Zeno's by means of the theory of forms. Only the conclusion of the paradox is given: "If beings are many, it is necessary for them to be simultaneously similar and dissimilar, which is impossible" (127e). Perhaps the argument itself (which Plato may have assumed his readers would be familiar with) was to the effect that we are


similar because we are all human beings (or share some other attribute), but dissimilar because we are distinct individuals.[4] Whatever the argument may have been, Socrates replies to it by speaking of similarity and dissimilarity as separate forms, whose very separation from the things that participate in them resolves the paradox: there is no contradiction in saying that I participate in similarity and dissimilarity, rather than being both similar and dissimilar (129a). The same goes for comparable paradoxes about the one and many, or rest and motion (129b-e).

Under Parmenides' questioning, Socrates says that there are forms themselves-by-themselves (

) also of the just, the beautiful, and the good (130b); that he is at an impasse (
) about whether there are separate (
) forms such as human being, fire, or water (130c); and that he thinks there is no separate form of hair, mud, and dirt, but they are just as we see them—be is troubled by the thought that what is true of one should perhaps be true of all, but is afraid this will make him fall into a pit of nonsense (130c-d). Parmenides replies: "You are still young, Socrates, and philosophy has not yet taken hold of you in the way that, in my opinion, it eventually will, at which time you will not despise any of these. But now you still consider people's opinions because of your age" (131e). There are a number of conclusions we can draw from this passage. First, Socrates' immediate acceptance of value-laden forms like justice, beauty, and goodness, together with his outright rejection of the existence of forms for "very ignoble and base" (
, 130c) things like hair, mud, and dirt, shows the importance to him of the forms' role as the bearers of value, a role that makes them hard to reconcile with what is trivial or unclean. The mature Socrates' description of them as offspring of the good (Republic 6.509b) is testimony to the fact that reality is shaped by the teleology of goodness,[5] and so the fundamental possibilities of reality will also be manifestations, however indirect, of the nature of the good.

Not only is there a connection between reality and teleological goodness, but there is also a connection between philosophy's attempt to apprehend that reality, and moral goodness. Such a connection has already been drawn in detail in the Republic 's doctrine of the tripartite

[4] Such an interpretation is suggested by Socrates' extrapolation of the argument to the "one and the many" at 129c. Also see R. E. Allen, Plato's Parmenides (Oxford: Blackwell, 1983) 78.

[5] At least as far as is permitted by the nature of material necessity. Cf. Timaeus 47e-48a.


soul, and the present section contains a graphic reminder of that doctrine. Zeno remarks that he wrote his polemic when he was young, in a spirit of youthful combativeness (

) rather than that of an older man's love of honor (
, 128e). These alternatives represent youthful and mature versions, respectively, of the spirited part of the tripartite soul, the irascible part. It is interesting, then, that when Socrates treats Parmenides and especially Zeno so disrespectfully[6] —so insensitively to their love of honor—that Pythodorus expected them to be angry with him, they only smile in admiration of Socrates' cleverness (130a). We are shown that Zeno's youthful spiritedness did not in his maturity develop into an equally aggressive "older man's love of honor," but is rather replaced by a noncombative appreciation of rationality. Antiphon, on the other hand, undergoes an opposite transformation: "When he was a youth he studied [Pythodorus's account of the conversation] with great care; though now he devotes most of his time to horses . . .. When we asked him to repeat the conversation, he was at first unwilling—for he said it was a lot of trouble— but then he did so" (126c-127a). Thus he seems to have gone from a youth motivated by rationality to an adult motivated primarily by pleasure. There is the suggestion in all this that philosophy, if one perseveres at it, can transform us for the better—but not if one allows oneself to become lazy and self-indulgent. That suggestion will be amplified in the dramatic background of the Theaetetus . For now, however, we may observe the presence here of the categories of the tripartite soul (pleasure, competitiveness, reason), and therefore the implicit reminder of the moral dimension of philosophy, which does not come under active consideration in the dialogue.

Second, value is not the only factor that determines for Socrates what forms there are, for he introduced the theory in terms of the forms of "same and not-same, plurality and one, and rest and motion" (129d-e), which have no obvious connection with values. The valual aspect of the forms was introduced by Parmenides' questions; prior to that the characteristic that Socrates had emphasized was the forms' separation from corporeal things. The quality of the form as separate (

) from materiality, that is, as subsisting itself-by-itself (
), is, together with value, the fundamental concern of the present passage. Very likely this is why Socrates balks at acknowledging the existence of forms

[6] This is rightly taken by Allen to be a way of emphasizing Socrates' immaturity (PP 67). Allen also refers to Euthydemus 273a, where Ctesippus is described as "insolent due to his youth" (n. 38).


for natural kinds, like human beings, fire, and water. These examples are not trivial or lacking in nobility, like hair, mud, and dirt, so his uncertainty about the existence of separate forms for the former three cannot be on account of an "unworthiness." They may be unconnected with value in an obvious way, but the same was true of sameness, not-sameness, plurality, oneness, rest, and motion, which Socrates himself had offered as his first instances of forms. The reason that it is difficult to think of human beings, fire, and water as separate from their corporeal manifestations must lie elsewhere. I suspect it lies in the fact that human beings, fires, and water actually exist in the physical world; whereas sameness, not-sameness, plurality, oneness, rest, motion, justice, goodness, and beauty never exist corporeally as such,[7] but only as attributes of physical things. There is no risk of identifying these latter forms with corporeal instances, because there are no corporeal instances of them as such; corporeal entities are (in Aristotelian terms) substances, not attributes. But since there are corporeal instances of substantial forms like human being, fire, and water, it would be harder to establish that their essential being lies outside themselves, somehow separate from individual human beings, fires, and waters.[8] Because Aristotle rejected the separation of forms from individual things, the forms of natural kinds—substantial forms or species—are for him the most important, and attributes have only a subordinate importance. For "Socrates," the reverse is true.

Third, Socrates nevertheless needs to recognize that in order to be consistent he must accept that what is true in some cases is true in all. Why should some things, but not others, owe their nature to formal causality, and how can Socrates divide the one from the other without being arbitrary? Hence his indecision about natural kinds. If the theory of forms recognizes that there is such a factor as formal causality opera-five in the world, then it must generalize the causal efficacy of forms to all cases. The forms' character as necessary causes of phenomena within

[7] That is why the "method of paradigms" is so important in the Statesman (277d ff.).

[8] Allen's interpretation here, although conceived in very different terms, ultimately points in the same direction: "Such Ideas [as Man, Fire and Water] are substantival; their distinctive feature is that, unlike those Socrates has just accepted, they have no opposite (Phaedo 104a-b, 105a, Republic VII 524d-525a). Since Socrates introduced the theory of Ideas to explain why qualification by opposites does not imply the identity of opposites, it may seem an open question, so far as the Parmenides is concerned, whether Ideas are required where opposition is not involved (as is true of Republic VII 523c ff.)" (PP 107). The connection between Allen's interpretation and my own may be expressed in Aristotelian terms: it is much more difficult to regard a substrate as separable from matter than so to regard the contraries by which it may be characterized.


the physical world must be insisted upon as rigorously as their separation from the physical world.

Fourth, as the end of the passage shows, Socrates' failure to answer successfully the questions posed by Parmenides is meant to illustrate his immaturity. He is too concerned with what others will think of him, and not yet fully committed to philosophy. In other words, like the young Zeno, he is still too much under the influence of the spirited part of his nature and its concern with repute, and not sufficiently dedicated to the rational. The defense of the theory of forms has not been put into the most capable of hands. Later, when Socrates has shown himself unable to cope with Parmenides' first four challenges to the theory of forms, Parmenides introduces the fifth and greatest challenge by saying that only "someone with much experience" (

) would be able to disprove the contention that even if the forms exist they will be unknowable (133b-c). The reference to the need for "much experience" is an allusion to the youth and inexperience of Socrates that we already see in this passage. Parmenides goes on to make that clear, and to suggest how Socrates might attain greater sophistication (135c-136a). We must therefore keep in mind that what is true of this preliminary inquiry is meant to prefigure the subsequent arguments—that Socrates' failure to answer the later challenges should, as here, be taken as a reflection on his youth at least as much as on the limitations of the theory of forms. If Plato balieved that the "greatest difficulty" could be met by a thinker of sufficient ability and experience, he would presumably believe the same to be true of the other, lesser arguments.[9] And if Plato were seriously presenting us with a refutation of the theory of forms or one of its major components, it does not seem likely that he would impugn the competence of its defender. This is not to deny that he regards the challenges to the theory of forms as serious and important.

4. First Argument: Participation (130e-131e)

The first argument focuses on the forms' function as causes, and con-dudes that the relation of participation, the mechanism of this causality, is incoherent. The argument turns on a dilemma: either (a ) the form will be whole and entire in each of the things that participate in it, or

[9] Cf. Mitchell Miller, Plato's Parmenides: The Conversion of the Soul (Princeton: Princeton University Press, 1986), 60 n. 24 (p. 219).


(b ) it will be partitioned among them. If a , "then being one and the same it will be whole in many separate individuals at once, and would thus itself be separate from itself" (131b). If b , then it will no longer be possible to maintain that the form is "one" (131c), a unity, as Socrates' theory had assumed (131a). To drive home the point, Parmenides adds three more specific reductios : the parts of bigness will no longer be big since they are small relative to the whole, the parts of equality will no longer be equal to equality since they are smaller than equality as a whole, and smallness will be larger than its parts since it embraces them all (131d-e).[10]

Clearly a is closer than b to representing the theory as put forward in previous dialogues (although, as we shall see, it can be misleading even to think of the forms as wholes since they are simple and incomposite, as at Phaedo 78c f.). In fact, by treating forms as if they were spatially extended and capable of addition, subtraction, and division, b exemplifies the confusion between formal and material causality that Socrates warned against in the Phaedo (96d f., 101a f.). Accordingly a is the horn Socrates tries to grapple with. In reply to Parmenides' conclusion at a , Socrates says: "It would not [be separate from itself] if it were like day, which is one and the same, is in many places at once, and yet is not itself separated from itself; so each of the forms too would be one and the same in all its participants at once."[11] Parmenides replies, however: "You make one and the same to be in many places at once, just as if you spread a sail over many people and then said it was one and the whole of it was over many. Isn't something like that what you mean to say?" "Perhaps" (

), is Socrates' hesitant reply (131b-c).

Socrates may well have misgivings about the substitution Parmenides

[11] This example, according to Proclus, derives from the discourse of Zeno (Commentary on Plato's Parmenides [Princeton: Princeton University Press, 1987] 229).


makes here. Commentators sometimes take Parmenides' point to be that, if Socrates compares participation to the presence of daylight to different individuals, we can still reply that the light that falls on one individual is a different part of the light than that which falls on another—like Parmenides' example of a sail.[12] But Socrates did not compare forms to light but to day (

), and this is a different matter entirely, for when I say that today is present to all of us, this does not mean that different parts of the day are present to each of us.[13] To be sure, it has been pointed out since antiquity that a day can be divided, in the sense that the sun reaches some places later than others. But this would not be true within the same city, for instance (or along any longitude), so Socrates' example is valid. Compresent individuals are separated from one another spatially, not temporally, but a unit of time (like a day) is divisible only temporally and not spatially, so it cannot be divided among us. (Light, on the other hand, is spatial, like a sail, and so can be partitioned among spatially discrete individuals.) By substituting a spatial metaphor for Socrates' temporal one Parmenides misrepresents Socrates' argument.

Although the youthful Socrates is not quite quick enough to put his finger on what went wrong, his dubious "perhaps" suggests that Plato made the switch deliberately and sought to arouse our suspicion and make us pursue the matter further. It might be thought that Plato portrays Socrates as tentative here and in the other arguments only to show that Socrates senses defeat and would like to be able to find another answer. Certainly that is part of it, but there are three reasons why I doubt that it is the whole explanation. First, Socrates hesitates only at one crucial step in each argument, whereas if the hesitation were simply a matter of dramatic verisimilitude we should expect his hesitations to be more extensive. Second, each step where he hesitates is in fact a questionable rather than overwhelming one, which is the reverse of what we should expect on the other hypothesis. Third, Socrates is not made to hesitate at all in the third argument although he is again refuted, and this is what we should expect if the hesitations are flags of fallacious steps, for this argument (against conceptualism) is the one

[12] Both Allen (PP 116-17) and Miller (PP 49 n. 22 [209]) interpret Socrates in this way, and accordingly conclude that the retort is meant to show Socrates' lack of sophistication.


argument with which Plato is presumably in sympathy, and it does not in fact seem to involve any deliberate fallacies.[14]

The first argument shows, in any case, not only that the simultaneous presence of the forms to many individuals is unintelligible on the basis of our experience of spatial things, but also that it can be made intelligible on analogy with temporal units. The concept of participation has not therefore been shown to be incoherent, but neither has it been entirely vindicated. A weak interpretation of Socrates' point is that since it is possible for some things, like days, to be simultaneously present to many individuals without being divided, then such a situation is possible in general and may therefore be possible also for forms. Clearly this does not show that (or how) forms can be undividedly present to many things, but only that it is not prima facie impossible. Even giving a strong interpretation to his point does not help much. We can interpret it as an elliptical argument by analogy, assuming we are meant to understand that there is a certain resemblance between forms and temporal units in that both are aspatial. Thus, if a temporal unit can be undividedly present to many individuals because it is aspatial, then so can a form since it too is aspatial. This is only an analogy, however, because the way that formal causes may be present to multiple individuals would hardly be the same as the way this is true of temporal units; and because it is only an analogy, neither we nor Parmenides are required to accept the inference from the observed similarity (aspatiality) to the imputed one (undivided multipresence). That is why we resort to analogies only when we are not able to give an account of something in itself,[15] and so the need here to fall back upon (at best) analogy is an admission of a limitation in the explanatory power of the theory of forms.

The limitation is not merely contingent, a function of the Socrates' lack of sophistication,[16] for even in the Phaedo , where the concept of

[15] Cf. Republic 506c-d, Phaedrus 246a.

[16] Miller argues otherwise. He, too, believes that "in his portrayal of Socrates, Plato is serving notice that an indirect, analogical understanding of the doctrine of forms is not enough" (PP 65). But he takes this as a reflection on the as-yet-inadequate level of conceptualization on the part of young Socrates, rather than as a reflection on the limitations of the mature theory (56 n. 39 [217]). Young Socrates is not yet sufficiently adept at abstract conceptualization to understand the nature of the forms in themselves; he can only conceive of them in corporeal terms and therefore only in terms of simile and metaphor. The second part of the dialogue, Miller argues, is designed to address this deficiency by training him in abstract conceptualization. Thus the Parmenides functions as a kind of stepping-stone toward the Republic . "The elder Socrates' philosophical command of the theory of forms in the Republic is the goal, albeit retrospectively posed, toward which the conceptual labor prescribed in the Parmenides is a needed means . . . . The youthful Socrates' task is to appropriate fully, by critical and conceptual inquiry, what was given to Glaucon and Adimantus more in the manner of authoritative pronouncement and in the modes of image and analogy" (20). But since even the Socrates of the Republic con- stantly uses "image and analogy" in speaking of the forms, what reason is there to sup- pose that Plato believed a purely conceptual account to be possible? I agree with Miller that the Socratic wisdom of the Republic is more like the goal at which the Parmenides aims, than a rejected ideal from which it flees; but I do not think we are given any reason to expect that wisdom to be able to purge its concepts of metaphor and analogy.


participation receives the most extensive treatment, it is never characterized other than metaphorically. Socrates is uncertain there whether to describe the relationship between form and individual as presence (

), communion (
), or in some other way (100d)— previously he had described it as a striving (
) of things to emulate the form (74d), and as the participation (
) of the thing in the form (e.g., 100c). All these are metaphors or analogies, and they are never translated into a purely conceptual account. In the present account, therefore, the theory has not been refuted, but one of its limitations has been made apparent.

5. Second Argument: Universals (132a-b)

The second argument is no longer concerned with the causal relation by which forms are somehow present to particulars, but with the aspect of forms as universals:[17]

"I think you believe that each form is one for some such reason as this," said Parmenides; "when many things appear to you to be big, perhaps there seems to you, as you look at them all, to be some one Idea that is the same in all of them, and from this you conclude that the big is one."

"What you say is true," he said.

"What about the big itself and the other big things; if you look at them

[17] Allen interprets the whole series of arguments as a single extended argument (e.g., PP 95, 168), but the connections he makes between them sometimes seem strained. In the present case he writes, "The Paradox of Divisibility [i.e., the present argument] assumes the second disjunct of the Dilemma of Participation: if there is participation, there is participation in parts of Ideas" (134). This seems unlikely for two reasons. First, there is nothing in the second argument to suggest that the members of the infinite regress are related as parts and wholes; even if there were, this would follow as the conclusion rather than being an assumption. Second, Allen acknowledges that "the implication of the [first argument] is that both disjuncts of the Dilemma of Participation are false" (130), in which case it would be odd to assume the truth of one of them as the basis for the next argument.


all in the same way with your soul, will not a single big appear again by which all these appear to be big?"

"So it seems" [


"Then another form of bigness will appear, beyond bigness itself and the things that participate in it; and after all these yet another, by which they will all be big; and each of your forms will no longer be one, but their number will be unlimited."

Best known by its post-Platonic name of the Third Man argument, this second argument, sometimes in conjunction with the fourth, has received far more attention than all the others combined. In an enormously influential article Gregory Vlastos argued that the regress arises from an incompatibility between the self-referential aspect of the forms (bigness is itself big) and the principle of separation or "nonidentity" (bigness is distinct from what is big).[18] These two principles are incompatible, however, only if the sense in which a form has its own quality is exactly the same as the sense in which corporeal things have it, and this is not in fact the case.

Let us begin at the point where Socrates again displays a certain hesitation, replying only "so it seems" when Parmenides suggests that if he looked at the big itself and the other big things in the same way, there would have to be another big to account for what they have in common. Here again there are good grounds for hesitation. Parmenides' instructions are to look at the form and things in the same way with our soul. "With our soul," because forms cannot be perceived with our eyes but only with the mind, that is, with reason. An integral part of the theory of forms has always been the insistence that the body perceives only individuals, and the mind (reason) only forms: rational knowledge is not sense perception but "recollection." The incommensurability between the two kinds of perceiving is why, for example, not only are the prisoners in the cave (who are at the level of sense perception) unable to see the forms outside the cave when they look, but the philosophers outside (who perceive the forms with their soul) are no longer able to see the individualities within. Consequently, while it is evident that the

[18] "The Third Man Argument in the Parmenides," Philosophical Review 63 (1954) 319-49. Cf. Wilfrid Sellars, "Vlastos and the Third Man," Philosophical Review 64 (1955) 405-37; Peter Geach, "The Third Man Again," Philosophical Review 65 (1956) 72-82; and Vlastos, "Postscript to the Third Man: A Reply to Mr. Geach," Philosophical Review 65 (1956) 83-94. All but the Sellars are reprinted in Allen, ed., SPM. The debate to which these articles have given rise continues to this day, but I am inclined to agree with Robert Turnbull ("The Third Man Argument and the Text of Parmenides ," in John Anton and Anthony Preus, eds., Essays in Ancient Greek Philosophy 3, Plato [Albany: SUNY Press, 1989] 203-25) that "most of the Vlastos-inspired controversy is irrelevant to the interpretation of that text and its context" (203).


form of bigness is to be perceived with our soul rather than our eyes, how are we supposed to perceive big things (individuals) "in the same way with your soul ," since individuals are perceived not by reason but by the senses? The term "soul," or psyche, may be broad enough to include both sense perception and reason, as it does in Aristotle, but one is hardly using the soul "in the same way" in both cases. What Parmenides suggests is simply inconceivable in terms of the mature Socrates' theory of forms.

We need not review the epistemology of the mature theory of forms, however, to see that this is so. The present argument is often discussed in terms of the post-Platonic example of the form of "man" rather than Plato's example of the form of "big," but this is unfortunate because Plato's example seems strategically chosen to illustrate something that is more obscure in the other. We can see from Plato's example that the form and the things cannot be looked at "in the same way" because "big" cannot be predicated univocally of both. Predicated of things , "big" means occupying more space than normal, and it clearly cannot mean this when predicated of the form since forms are aspatial (as we were reminded by the implications of the first argument). Plato could scarcely have chosen an example better calculated to show the impossibility of univocal predication, that is, of looking at the two "in the same way."

In fact this very example was used in the Phaedo to illustrate the incommensurability between forms and things. Big things can be both big and small at the same time, as Simmias is big in comparison with Socrates and small in comparison with Phaedo, whereas bigness itself can never be both big and small (102b-d; cf. 103b). It was on similar grounds that Socrates had earlier argued that it is impossible to derive the concept of equality itself from our sensible experience of equal things (74b-75b). Socrates had made a similar point earlier in the Parmenides , using an example that will become prominent in the second part of the dialogue, namely, that of the "one." He says that it is easy to show that the same thing is both one and many—since a corporeal thing is, on one hand, a collection of parts and therefore many, and, on the other hand, a discrete individual and therefore one—but that the one itself can never be many or the many itself one. Forms, in other words, display an absolute nature, while things display only an equivocal nature, and one cannot look at them in the same way. It is problematic how any predicate within the corporeal world can be applied univocally to an incorporeal reality.


On the other hand, neither can the predication simply be equivocal, as the fifth argument will show, for then the form and thing would bear no essential relation to each other and the theory of forms would be pointless. The only alternative is to take the dual predication analogically (as medieval theology does in explaining how predicates derived from the corporeal world can be applied to God) and say that bigness itself is not big in the sense of taking up space, but in some sense that is the intelligible analogue of taking up space. The reply to Parmenides' second argument, then, like the first, must fall back upon analogical reasoning. In the first argument it was a three-term analogy (forms are to things as temporal units are to things) such as Plato had used in the "affinity" argument of the Phaedo (78b ff.), and here a four-term analogy (bigness itself is to big as the intelligible is to the corporeal) such as he used in the Divided Line.

Once again the apparent refutation of the theory of forms contains within itself a reminder of how the objection can be met—in terms already familiar to us from the middle dialogues—but once again the implicit defense can do no more than partially vindicate the theory. In the first argument the analogy gave us no way of-understanding intrinsically precisely how a form is wholly present to many things, and here we find that we have no understanding of what bigness itself (or any form) is intrinsically, but only analogically.

6. Third Argument: Concepts (132b-d)

The next argument focuses on the forms' function as concepts. The youthful Socrates sees that the key to answering Parmenides' second argument is to show that the forms are fundamentally different from things, so that the two cannot be taken together in the same way. However he does not yet formulate this difference in the sophisticated manner of the Phaedo and Republic , but suggests instead that "each of these forms may be only a thought, in which case it would not be appropriate for it to exist anywhere other than in our souls; for in that way each would be one and would no longer suffer the consequences that were just mentioned." Since this is the only feature of young Socrates' defense of the theory of forms that was not part of the mature Socrates' conception of the forms in other dialogues, it is not dear why Plato introduces it here at all—only to depict it as completely rejected. It may be that because of Plato's opposition to materialism, and his elevation of the mind over the senses as the touchstone of reality, he was seen as an ally by some of those who (like the followers of Protagoras and Gorgias)


believe reality to be entirely subjective.[19] In that case this central one of the five arguments may be meant to signal his opposition to that extreme interpretation of his antimaterialism. The argument discussed in the next paragraph, in any case, seems to prefigure the refutation that Socrates undertakes against Protagoras in the Theaetetus .

In a series of steps Parmenides establishes that if the form is a thought, the thought must be of an independently real universal: (1) a thought must be of something, (2) namely, of something that is, and (3) it must be common to all its instances (132b-c). The point seems to be that conceptualism implies its negation, realism, because thoughts are of independently real existents, in which case conceptualism is incoherent. But why must the referents of thoughts be real independently of the mind? Although this premise may seem question-begging, it may instead have been taken from the historical Parmenides' claim that "it is the same thing that can be thought and that can be" (fr. 3), and "not without being . . . will you find thinking" (fr. 8.35-36).[20] We can interpret these fragments in a Platonic, rationalistic way, to mean that reason (thinking,

) is the test of what is real. Against the view of those in the cave, who define reality in terms of what is visible, Plato (following Parmenides and anticipating Hegel) argues that the real should be defined in terms of the rational and that the rational consequently implies the real.[21] On such a principle one may reject a conceptualism that

[19] Klaus Oehler may be right in suggesting that Plato is the ancient thinker who comes closest to modern idealism: "Der wesentliche Bezug dieses Denkens auf Öffent-lichkeit ist auch der letzte Grund, warum dieses Denken sich hie in der Weise spaterer Reflexion auf sich selbst zurückgezogen und ein im strengen Sinne philosophisches Selbstbewub tsein konstituiert hat, um von sich aus, aus seinem Inneren, die Welt 'systematisch' zu entwerfen, obwohl es gerade Platon, war, der dieser Moglichkeit von allen antiken Denkern, Augustinus ausgenommen, am nächsten gekommen ist" (Die Lehre vom noetischen und dianoetischen Denken bei Platon und Aristoteles [Munich: Beck, 1962] 105).

[21] Hans Rochol ("The Dialogue Parmenides : An Insoluble Enigma in Platonism?" International Philosophical Quarterly 11 [1971] 496-520) goes so far as to suggest that "this theory, concluding from the existence of thinking to the existence of an objectively existing object of thinking, is at the same time the gist and basic theory of Platonism" (506-7).


treats rationality as independent of reality, and that claims that the being of concepts may be confined to the mind.

Parmenides follows this with a further point, that if forms are thoughts, and things are what they are by participating in forms, then "either everything is made of thoughts, and all things think, or despite being thoughts, they are without thought." If I have interpreted this second, rather obscure part of the argument correctly, it does not seem to have much force, since the second alternative, that things may be constituted by thoughts without themselves thinking, is not absurd. However, the second horn would no doubt have seemed more paradoxical in the days before Berkeley and Kant than it does to us, Protagoras notwithstanding. In any case, the argument as a whole is the only one in which Socrates displays no hesitation (see above, n. 14).

Socrates had introduced this argument to circumvent the consequences of the Third Man argument, recognizing that the key to blocking the infinite regress is to deny that forms and things can be looked at by the soul in the same way. If the forms are concepts and cannot "exist anywhere other than in our souls," then clearly the soul cannot look at material things in the same way as forms. But this tactic turns out to be too extreme, for it would preclude any ontological relationship at all between forms and beings (a difficulty that arises on other grounds in the fifth argument). Parmenides' refutation makes this point, but only in a negative way: it rejects Socrates' inappropriate conception but without offering an alternative.

The middle dialogues give us a three-level ontology, comprising corporeal matter, thinking soul, and eternal forms. Whereas modern transcendental philosophy seeks to make form a function of mind, Plato's middle dialogues keep soul rigidly distinct from the forms. Knowing is not reflexive but is the soul's apprehension of something other than itself, and the first part of the present argument reaffirms this separation.[22] Whether this argument arose in response to misunderstandings of the theory of forms or conceptualist challenges to it is impossible to determine. Conceivably it is meant to make dear the difference between, on one hand, the Platonic position that truth is to be found in the mind as separated from the body and the physical world generally (e.g., Phaedo 65d-66a); and, on the other, the Protagorean position that truth is subjective , with no ascertainable objective referent.

[22] Hans-Georg Gadamer suspects that in the second part of the dialogue there may be hints even of a distinction between psyche and nous, which paves the way for the Plotinian position ("Plato's Parmenides and Its Influence," Dionysius [1983] 3-16).


The argument functions in another way as well. The first two arguments explored the purely ontological aspects of the forms, as causes and universals, while the last two introduce the epistemological dimension, first implicitly in the fourth argument's concept of paradigm, then explicitly in the fifth argument's focus on the problem of knowledge. The third argument is the transition between these pairs, and in its refutation of extreme conceptualism the relationship between being and thinking is briefly indicated.

7. Fourth Argument: Paradigms (132d-133a)

The way to reconcile the ontological function of forms with their epistemological relation to concepts is not, as Socrates now sees, by identifying forms with concepts but by thinking of them as paradigms that serve both as the ontological basis of individual things and as the referent of our concepts.[23] He proceeds, accordingly, to put forward a new interpretation of forms now as paradigms, although without explicit reference to their epistemological aspect.

"The most likely view appears to me to be this, that these forms exist in nature as paradigms and other things resemble them and are like them; and their participation in the forms turns out to be nothing but an assimilation to them."

"Then if anything," Parmenides said, "resembles the form, is it possible for that form not to be like the thing that resembles it, insofar as the thing has been made to resemble it; or is there any way that the like can be unlike the like?"

"There is none."

"And is there not a great necessity that the like participate in one and the same form as its like?"


"And won't that, by participation in which like things are made like, won't it be the form itself?"


"Then it is not possible for anything be like the form or for the form to be like anything else; otherwise beyond the form another form will always appear, and, if that is like anything, still another, and without end a new form will always arise continuously, if the form is like what participates in it."

"What you say is very true."

[23] As Allen (PP 163) and Miller (PP 56) note, the paradigm-copy model of participation would also offer Socrates a way out of the earlier dilemma of participation, since a model is neither divided among nor present in its copies.


"Then it is not by likeness that other things partake in the forms, but it is necessary m seek some other means by which they partake."

"So it seems" [


Here Parmenides does not merely ask Socrates to "look at the form and things in the same way," as he did in the second argument, but offers an argument why it ought to be possible to do so—the reciprocality of the relation of likeness. Socrates, who has already conceded the infinite regress in the second argument, now assents unhesitatingly to each of the preliminary steps, but he hesitates this time at the conclusion that Parmenides draws from them. From the point of view of the middle dialogues it is not surprising that Socrates hesitates. Parmenides' conclusion is that the infinite regress resulted from Socrates' claim that the forms are paradigms and therefore resemble the particulars in some way. However the regress does not follow specifically from this claim, but from Parmenides' treatment of the concept of resemblance as a symmetrical relation. In the middle dialogues the resemblance between forms and particulars is represented by the mature Socrates not as a symmetrical one but as one between unequals. In the Phaedo , for instance, using the example of equality, he says that particulars strive to be like the form but fall short (74d ff.), and he makes a similar point in the Republic in terms of the examples of the beautiful, the just, and the pious (479a). The relationship is thus an uneven one in the sense that the forms have an absolute perfection that the things participating in them can never achieve. Forms and things are ultimately incommensurable and cannot properly be set into a reciprocal relationship in the way that Parmenides proposes. It makes sense to say that equal sticks strive to be like equality itself, but not to say that equality itself strives to be like equal sticks. Indeed, the fifth argument will trade on this very type of incommensurability. And since the forms are self-referential in any case, there is no need to posit further forms to account for their nature; they are the ultimate referent both of themselves and of particulars.

On the other hand, if the resemblance between forms and things is ultimately incommensurable, how can it be a resemblance at all? If the forms are the standards against which particulars are measured, how can the latter be incommensurate with their measure? This is a fundamental aporia in the theory of forms, which was noted as early as the Phaedo , where Socrates, in passing, calls our attention to the difficulty of deciding whether or not things can be said to be "like" forms (74c-d).[24] But


although Plato seems to have been aware of the problem from the time he first elaborated the theory of forms, he gives no evidence of believing that it could definitively be solved. As in the first and second arguments, the "solution" seems to be that the resemblance between forms and the things they participate in is not a univocal one (and therefore does not give rise to an infinite regress) but an analogical one: the beauty of the beautiful itself is to the beauty of beautiful things as the absolute is to the relative, not different in degree—and certainly not symmetrical—but different in dimension.

Once again there is nothing in the fourth argument that the mature Socrates of the middle dialogues could not and did not answer, but once again the answers must ultimately fall back on analogy and thus lack the transparency to which reason aspires.

8. Fifth Argument: Separation (133a-134e)

There is an important introduction to the fifth argument, to which commentators do not always pay enough attention.

"Do you see, then, Socrates, how great the impasse is, if someone maintains that forms are marked off as entities themselves-by-themselves [


"Yes, indeed."

"You may be sure," he said, "that you do not yet, so to speak, grasp how great the impasse is if you maintain that each form is one and is something always marked off apart [

] from particular things [

"How is that?" said he.

"There are many factors," he said, "but the greatest is this: if anyone should say that it is not even fitting for the forms to be known if they are such as we say they must be, no one would have any way to show the speaker that he was wrong, unless the disputant were a man of much experience and not without natural ability, and were willing to follow a long and complicated proof; otherwise he who insists that they are unknowable would be unconvinced."

What is remarkable here is first of all Parmenides' claim that it would be an error to believe that if the forms are separate (

, 133a9, b2) they are unknowable, even though it would not be easy to demonstrate the error. For this is to concede in advance that the fifth argument, which makes precisely this claim, is in principle answerable although only with great difficulty. Moreover, because of the relationship of this argument to the previous ones, Parmen-


ides' words suggest that those too should be answerable, if only in an elusive manner. Remarkable too are Parmenides' words (in the last speech above), "if [the forms] are such as we say they must be," which imply that he shares Socrates' view. The context suggests that Parmenides is using the first person plural earnestly, and not as a patronizing form of the second person singular. If the previous arguments were to have been regarded as effective refutations, he could hardly include himself as one of the theory's supporters. And since the reference is specifically to the assertion that the forms are independent and distinct from things, Parmenides is here professing to share Socrates' advocacy of the theory of separate forms.

That the historical Parmenides should be portrayed as sympathetic to the theory of forms is perhaps not surprising in light of the terms in which the theory is stated above: "each form is one and is something always marked off apart from particular things." This statement could be taken equally as a depiction of Parmenides' own criterion of reality, except that the forms are multiple. Plato always speaks of Parmenides with respect, and it is widely agreed that he sees the theory of forms as an extension of the Parmenidean view of reality. Accordingly he may well have felt that Parmenides would have accepted the theory of forms. But that this dramatis persona Parmenides, who is leading an attack upon the theory of forms, should speak here as an advocate of that theory seems quite implausible unless his arguments are to be regarded not as hostile refutations but as the "friendly" criticisms of one who is sympathetic to the theory but not blind to its limitations. An awareness of its limitations would entail a rejection of the theory only if one were able to replace it with one that is free of such limitations, and Plato does not seem to have considered such an alternative possible.

The fifth argument is the longest of the five, and the one that Parmenides considers the most important (133b). In it the epistemological implications of the forms' character as paradigms become explicit for the first time in the dialogue. The argument may be summarized in the following steps.

1. "Whoever claims that the essence of each thing is itself-by-itself [

. . .
] would agree, first, that none of them are in us" (133c).

2. "Those ideas that are what they are in relation to one another [

] have their essence in relation to one another," that is, in relation to other forms rather than in relation to concrete things.


"And the things among us that have the same names as those, will likewise stand in relation to each other and not to the forms." For example, if we are a master or slave we are master or slave of another person, not of a form, and the forms of mastery and slavery are similarly relative to each other, not to particular persons (133c-134a).

3. Thus "that which is knowledge itself would be knowledge of that which is truth itself," that is, of the forms; whereas "the knowledge that exists among us would be knowledge of the truth that exists among us" (134a-b).

4. Now, since "the forms themselves, as you agree [in step 1], we do not have, and they cannot exist among us"; and since "what each of these kinds themselves is, is known by the form itself of knowledge" [cf. step 3]; "then none of the forms is known by us, since we do not participate in knowledge itself."

"It seems not [

]" (134b-c).

5. Moreover, since the forms are far more perfect than the particular things among us, it is appropriate that god's knowledge will be of them, in which case he will be unable to know our world, since "the forms there are not able to have any relationship to the things among us, nor the things among us to those there" (134c-d).

6. And in this case god will no more be able to know the things among us than we can know the forms, nor will the gods be able to be our masters since their commerce is with the form of mastery, which does not stand in relation to us at all (step 2; 134d-e).

The first three steps (to which the young Socrates assents without hesitation) could easily be accepted by the Socrates of the middle dialogues since they insist only on the separation of forms from things, and on the fact that we are not capable of absolute knowledge or wisdom, both of which claims are common enough in the theory of forms (e.g., Phaedo 66b-e). The fourth step, however, draws a more radical consequence from our alienation from absolute knowledge—that is, that we can have no knowledge of the forms—and it is here alone that Socrates hesitates. If we look at what has happened, we can see why his agreement is uncertain ("It seems not"), although he cannot quite identify what is wrong. Parmenides has argued from the fact that the terms constituting certain relations, such as mastery and slavery, stand in relationship only to other terms at their own level (forms to forms, particulars to particulars), to the conclusion that no relationship can exist between


forms and particulars, including the relation of knowing. Knowledge itself knows truth itself and knows those entities (

), the forms; and knowledge among us knows the truth and the entities (
) that exist among us (134a-b). From the point of view of the middle dialogues, thus far the claims are not unjustified. Knowledge itself (wisdom) is indeed knowledge of the forms and is indeed beyond us, as we noted in the preceding paragraph, and our knowledge is of the truth and being of the things of our world. What is missing, however, is the further claim of the middle dialogues that the truth and being of the things of this world necessarily refer to the truth and being of the forms; the forms are the truth of this world.[25] In order to establish such a premise at this point, the young Socrates would have to argue that not all relations are of the kind that Parmenides describes, that some relations do in fact bridge the gap between the world of forms and things. The middle dialogues give us at least three such relations: participation, recollection, and eros.

Beginning with the Meno , the mature Socrates bridges the gap between our corporeal realm and the realm of the forms by the concepts of participation and recollection: since corporeal things participate in the forms, we can recognize or recollect the forms from their limited presence in corporeal things. This does not quite enable us to know them as they are in themselves ("knowledge itself" or wisdom is not available to us), but it enables us to know something of the forms. Here, however, Plato makes the young Socrates respond to Parmenides' argument as if he had not yet formulated the bridging concept of recollection; and Parmenides has previously discouraged Socrates' attempts to formulate the concept of participation (perhaps Socrates' present hesitation indicates a suspicion that some such formulation can be found). It is ironic then that Parmenides should use the very term participation here, where the concept itself is conspicuously omitted: "Then none of the forms is known by us since we do not participate [

] in knowledge itself" (134b)—a phrase that is repeated in step 5 (134c10).

In step 2 was an even more pregnant reminder of the doctrine:

Those ideas that are what they are in relation to one another have their essence in relation to one another, but not in relation to the things among us—likenesses or whatever rise we call them; and we who participate in them take our names from them. And the things among us that have the same names as those will likewise stand in relation only to one another and not to the forms.

[25] Cf. Phaedrus 247d, 249d.


Socrates forthrightly asks, "What do you mean?" and Parmenides responds with the master-slave example. But although that example illustrates his point, Parmenides had just provided us with two that do not: participation and naming. We and all things in our world participate in the forms, and the predicates that apply to us are the names of the forms in which we participate, such as animal, just, beautiful, big, and even master or slave. If particular things are thus related to the forms by participation and naming, then our knowledge of these things is ipso facto a knowledge, however imperfect, of the forms, as the doctrine of recollection maintained.

This point is connected also with the lesson of the third argument: that thinking refers to a being that is not reducible to thought. The activities of thinking and knowing stand in relation to the forms without collapsing the distinction between thinking and being (forms). Like eros in Socrates' speech in the Symposium , thinking and knowing are functions of soul, and they therefore mediate between the corporeal and the intelligible or divine without belonging exclusively to either. The fifth argument arrives at its paradoxical conclusion by ignoring the mediating nature of knowledge and trying to account for knowledge entirely in terms of the opposite poles of the divine and the corporeal.

But although the mature theory of forms has the resources to withstand this final assault too, the latter was no more pointless than were the other critiques. We saw in the first argument that the relation of participation is conceptually aporetic, marking one of the boundaries where conceptual clarity begins to break down and give way to metaphor and analogy. There it resulted in a limitation as to what could be said about the role of participation in explaining the ontological relationship between forms as causes and the particulars that depend upon them. In this fifth argument it results in a similar limitation about the epistemological relationship between forms as paradigms and the particulars that enable us to know them. The doctrine of recollection may show that we are led from particulars to knowledge of the forms, and that Parmenides' dichotomy is therefore a false one, but recollection is no less metaphoric than participation. Parmenides' attack shows that it is far from obvious how the realms of particularity and of form can stand in relation to each other, and how therefore we are to conceive of the relations of participation and recollection that make knowledge possible. More than that, in emphasizing the deft, rather than the bridges, between human thinking and the nature of reality in itself, the fifth argument illustrates why all philosophical theorizing must ultimately fall short of complete adequacy to reality.


The argument concludes with a remark similar to the one that preceded it:

"And yet, Socrates, the forms contain these difficulties and still very many more besides them, if these ideas of things exist and we mark off [

] each form as something in itself. Therefore he who hears such assertions is at an impasse and contends that the forms do not exist, or, even if they do exist, it is very necessary that they are unknowable by human nature. And he thinks that there is something in what he says, and, as we just said, he is amazingly hard to convince. Only a man of very great natural ability will be able to understand that there is a certain genus and essence, itself-by-itself [
], for each thing, and only a still more amazing man will be able to discover all these things and teach someone else to evaluate them properly."

"I agree with you, Parmenides," said Socrates, "for what you say is very much after my own mind."

"But yet," said Parmenides, "if anyone, in view of these and other such difficulties, will not permit the existence of forms of things or mark off [

] a single form in each case, he will not have anything on which to fix his thoughts, as long as he does not permit the idea of each thing to be always the same, and in this way he will utterly destroy the power of discourse. You seem to me to have been well aware of this."

"What you say is true," he said.

In the face of this seemingly unambiguous conclusion in favor of the theory of forms, why would anyone maintain that Plato regarded the preceding arguments as fatal objections to it? The usual response is that Plato still believes that some theory of forms is necessary, but that he is abandoning one of the most characteristic features of his previous theory, the separation of forms from things.[26] But to maintain this is to focus on the second main paragraph above to the exclusion of the first, where Parmenides maintains that "a man of very great natural ability will be able to understand that there is a certain genus and essence, itself-by-itself [

], for each thing." It is hard to imagine that
can mean anything other than existing distinctly from sensible things,[27] especially since at 130b8

[26] Sayre, for example, writes, "What Parmenides obviously has omitted from this last-minute defense of the theory is any mention of the Forms as entirely separate entities . . . . What Plato has Parmenides saying is that certain aspects of this theory Socrates has been maintaining, here and in earlier dialogues, indeed are essential for intelligible discourse—namely those aspects providing Forms that are definite and always the same—but that the notion that these Forms exist in total separation from the things that participate in them not only is inessential but leads to endless difficulties" (PLO 22).


by-itself") is used interchangeably with

("separate") at b3, b5, c1, and d1.

The same conclusion is implied by the passage with which the final argument began, where Parmenides said that someone "of much experience and not without natural ability, and . . . willing to follow a long and complicated proof" would understand why it would be wrong to "say that it is not even fitting for the forms to be known if they are such as we say they must be," namely, "marked off as entities themselves-by-themselves [

] . . . apart [
] from particular things [
]." Indeed, in view of this use of
, we may say that
("mark off") in the present passage also means "separate." The face that Plato does not here use the word
("separate") is not significant when he uses synonymous terms such as
. The present passage, therefore, affords no evidence that Plato at this point did anything but reaffirm the theory of forms in all its essentials.

On the basis of this examination there seems no reason to conclude that Plato intended these arguments to be a refutation or recantation of the theory of forms, and several reasons to believe that he did not:

1. The arguments are easily answered on the basis of features of the theory that were prominent in the middle dialogues and that could plausibly be omitted here only by portraying Socrates as being in the early stages of developing the theory.

2. By having Socrates hesitate precisely at the point where such factors can be brought into play to repel Parmenides' attack, Plato seems to hint at the inconclusive nature of the arguments.

3. Not only the answers but the problems themselves were anticipated in the middle dialogues: the problem of giving nonmetaphorical accounts of participation and recollection, and the ambiguity of resemblance and predication with regard to forms and things, are clearly present in the Phaedo , where the complete theory was first introduced. Plato evidently recognized these problems from the beginning but felt that the theory was not vitiated by them.


4. The final argument begins and ends with speeches that show Parmenides to be Socrates' ally in the theory, a devil's advocate rather than nemesis, which would hardly be possible if the arguments were intended as serious refutations. The motive for exhibiting and even exaggerating these problems may have been partly to remind us that the theory of forms cannot be regarded as a dogma or perfected doctrine but only as a valuable (perhaps indispensable) although inexact way of interpreting the world.

In Plato's letters it is said that philosophical truth cannot ultimately be put into words, and must instead be nurtured indirectly in its recipient:

There is no composition of mine concerning these things, nor will there ever be. For it cannot be expressed in speech like other kinds of knowledge, but after a long attendance upon the matter itself, and communion with it, then suddenly[28] —as a blaze is kindled from a leaping spark—it is born in the soul and at once becomes self-nourishing.
     (7.341c-342a; cf. 2.314a-c)

The authenticity of the letters has never been established conclusively, but on this point the letters are supported by the mythic, ironic, doubt-sowing character of the dialogues themselves, in which all words are uttered through mouths other than that of the author. The Phaedrus in particular gives direct support to this sentiment:

He who thinks that he has bequeathed in his writings some expertise [

], and he who receives them thinking that anything dear and firm will be in them, would be very simpleminded . . . if he thought that written words are more than reminders, to the one who already knows, of that about which they are written . . . . You might think that they speak with a certain wisdom, but if you question what is said, because you want to learn, they always say the same one thing.

I think a much finer way of being serious comes about when someone, using the art of dialogue, takes an appropriate soul, and plants and sows it with knowledgeable words that are able to help both themselves and their planter; words that are not barren, but have within them seeds from which other words grow in other characters, which are ever able to make them immortal, and making their possessor happy to the greatest extent that is humanly possible.


Accordingly, the first part of the Parmenides may readily be regarded as Plato's way of showing that his own theory of forms (which is prominent in the Phaedrus ) is no exception to the limitations on the possibility of philosophical disclosure that the latter dialogue stipulates. The theory of forms, too, can at best be an insemination, never a dogma.

9. Transition to the Hypotheses About the One (134e-137c)

Parmenides had said that "only a man of very great natural ability will be able to understand that there is a certain genus and essence, itself-by-itself, for each thing, and only a still more amazing man will be able to discover all these -things and teach someone else to evaluate them properly" (135a-b). He now concludes, "What then will you do about philosophy? Where can you turn if these things are unknown?" (135c). Socrates is uncertain "at present," and Parmenides attributes this—as he had Socrates' reservations about forms of dirt, mud, and hair—to his philosophical immaturity. Socrates has tried to define beauty, justice, good, and all the other ideas before he has been trained in the kind of dialectic that seems to most people to be useless loquacity, but is in fact the means of capturing truth.[29] Socrates has been right to confine his attention to intelligible forms rather than sensible things, that is, to pay attention to likeness and unlikeness. "But it is necessary to do this as well as that: in the case of each hypothesis not only must you examine what follows if what is hypothesized exists, but also if it does not exist—if you wish to be more fully trained" (135e-136a). If we showed only what followed from the hypothesis that something exists, then any problems we encountered might convince us prematurely to reject our hypothesis; whereas if we also considered what followed from denying it, we might find the consequences more objectionable than those of accepting it. Thus the Theaetetus will show that the problems consequent on denying the existence of the forms are even more problematic than those that follow from affirming their existence. And in the second part of the Parmenides we will find that the problems of affirming the existence of the One will not be as severe as those of denying its existence.

With this, we are led into the second part of the dialogue, which is

[29] Zeno later seconds this: "The many are ignorant that without this passage and wandering through all things it is impossible for the mind to happen upon truth" (136d-e).


even more enigmatic than the first. Proclus (pp. 29-34) distinguishes four different ways of interpreting even the overall intent of the arguments, regardless of any disputes about their content: (1) "Aporetic," although not in our sense of the word as "inconclusive," but rather in the sense of elenctic or refutation. (2) "Gymnastic," in the sense of logical exercises intended to develop our powers of abstract reasoning. (3) "Ontological," in the sense of a study of the nature of the One. (4) "Ontological" in the sense of "ontic," that is, a study of the beings that derive from the One. We may add at least two more approaches that Proclus does not mention. (5) "Jocular": A. E. Taylor writes that "the Parmenides is an elaborate jeu d'esprit . . . . Its purpose is to 'have some fun' with Monists who regard the sensible as illusion, and very little more."[30] (6) "Logical," in the sense of working out the implications of the concept "one." With some exceptions, this latter has been the dominant interpretation at least since Cornford.[31]

Even those who agree, either that the arguments are positive demonstrations or negative refutations, do not agree on what they demonstrate or refute. The "aporetic" or elenctic view often held that the arguments are refutations of Zeno, but more recently it has been argued that they are refutations of the historical Parmenides (Cornford, Fried-länder), or even of Plato's own earlier theory of forms, whether on account of their unity[32] or their separateness.[33] In the more positive in-

[30] Plato: The Man and His Work (Atlantic Highlands: Humanities Press, 1952) 350. This approach, too, goes back to antiquity, according to Robert Brumbaugh (Plato on the One: The Hypotheses in the Parmenides [New Haven: Yale University Press, 1961] 5). W. F. R. Hardie makes an appropriate response to it: "It seems to me clear that the passage of transition . . . is calculated to lead us to expect that the second part of the dialogue will throw some real light on the difficulties which have been raised in the first . . . . The reference in the Theaetetus (183E) to the 'noble depth' of what was uttered by 'Parmenides' on this occasion seems to me to tell strongly against the view that the hypotheses of the second part are nothing but an obscure joke or a long and tedious parody" (quoted by F. M. Cornford, Plato and Parmenides [London: Routledge and Kegan Paul, 1939] 114).

[31] "It is a natural inference that a main purpose of the whole exercise must be to point out that even the apparently simplest terms, such as 'One' and 'being', which will appear at the threshold of any metaphysical discussion, are dangerously ambiguous" (Cornford, PP 110). Recent exceptions to the logical interpretation include Paul Fried-lander, Plato , vol. 3 (Princeton: Princeton University Press, 1969 [orig. 1960]); Egil Wyller, "Plato's Parmenides : Another Interpretation" (Review of Metaphysics 15 [1962] 621-40), which summarizes his book Platons Dialog Parmenides in seinem Zusammenhang mit Symposion und Politeia (Oslo, 1960); Brumbaugh, PO ; and Miller, PP .

[32] Maximilian Beck writes: "It is demonstrated indirectly in this dialogue, namely, by carrying the contrary thesis to absurdity, that ideas are neither unifies nor multitudes, neither wholes nor parts, have neither spatial nor temporal extension; ideas are beyond any quantitative category; they are purely qualitative identities" ("Plato's Problem in the Parmenides," Journal of the History of Ideas 8 [1947] 232-36 at 234).

[33] Thus Sayre. He argues that the four hypotheses under "if the One is" should be paired with their analogues under "if the One is not." In that case, he says, if we put together hypothesis I ("if Unity exists, then this Unity with reference to itself has no characters") with hypothesis VI ("if Unity does not exist, then this Unity with reference to itself has no characters"), we get the conclusion that a unity that refers only to itself (i.e., a separate form, according to Sayre) has no characters unconditionally (i.e., whether or not it exists) (PLO 44-46). In other words, the very concept of separate forms is empty. What Sayre does not notice, however, is that by the same logic we can arrive at the opposite conclusion by pairing hypotheses II and V. Hypothesis II is that "if Unity exists, then this Unity with reference to others has all (contradictory) characters," and hypothesis V is that "if Unity does not exist, then this Unity with reference to others has all (contradictory) characters." On Sayre's principles, therefore, the two together mean that a unity that refers to other things (a nonseparate form) is self-contradictory unconditionally (i.e., whether it exists or not). So Sayre's conclusion, that Plato is here repudiating the concept of separate forms in favor of forms that are not separate, is nullified. We shall find that as long as we concentrate only on the stated conclusions of the hypotheses, and do not penetrate to the implicit distinctions within the details of the arguments, we will not be able to reduce the appearance, of a mutual canceling-out of all the hypotheses, into a positive conclusion.


terpretations, candidates for what is being demonstrated include the nature and relationship of forms and things (Brumbaugh, Miller), the implications of the original "participation" refutation in Part 1 (Allen), and "a nonhierarchical linguistic ground intrinsic to all expression of thought" (Sternfeld and Zyskind).[34]

Whether "the One" refers to Parmenides' monism, as some of the above views hold, or to some Platonic principle, is not easy to demonstrate. Both views are defensible. The latter seems more plausible, however, for the first part of the dialogue deals with difficulties implicit in the theory of forms, and we would expect there to be a thematic continuity between the two parts. The forms are "one over many," and therefore Ones. Their unity was threatened in the first part by at least three of the arguments: the first, in which the forms divide into their individual manifestations, like so many pieces of a sail; and the second and fourth, in which the forms multiply in infinite regress. It would not be surprising, therefore, to find the second half of the dialogue devoted to these questions, whereas there seems no reason suddenly to introduce Parmenides' (or Zeno's) principles in order to refute them—except insofar as they may coincide with principles underlying the theory of forms as well. Not only is each form a One in itself, but all together are unified by their common origin in the good (Republic 6.509b). It may be, then, that the following discussions apply at one level to each of the forms, and at another level to the Platonic One, the good.[35]

[34] Robert Sternfeld and Harold Zyskind, Meaning, Relation, and Existence in Plato's Parmenides: The Logic of Relational Realism (New York: Lang, 1987) 90.

[35] The latter interpretation is often referred to as the Neoplatonic interpretation because the Neoplatonists saw their own doctrines in it. But not everyone who considers Plato to believe in an overarching transcendent principle is a Neoplatonist. Proclus's commentary on the Parmenides , which is the paradigmatic Neoplatonic interpretation, is dogmatic and question-begging to a degree that would deter most commentators from accepting the same appellation.


In the first part of the dialogue we found a due to interpretation in the fact that Socrates hesitated at a single crucial juncture of each argument but the one against conceptualism. Will a similar approach help us to understand what is going on in the second half? This possibility has had at least one defender,[36] but the evidence seems to be against it. Aristoteles hesitates some fifty-three times during the eight extended arguments, and there is no conspicuous pattern. suggested by the juncture at which the hesitations take place. The following chart shows how often Aristoteles hesitates in each hypothesis, and where he hesitates first (subsequent hesitations are often extensions of the initial one).





Initial Hesitation





138e6: What neither has parts nor is a
whole cannot come to be.





143a3: The One that is will be unlim-
ited in quantity.





156e2: The instant is in between
change and rest.





157d7: Something cannot be part of a
multitude that includes it.





159c4: The One and the others will
never be in the same thing.





160e1: There is otherness in the One.





163d8: The nonexistent One neither
perishes nor comes to be.









If the hesitations were signals of flaws in the reasoning, the most flawed arguments, relative to their length, would be hypotheses II (if the One is, then everything will be inconsistently true of it in relation to the

[36] Brumbaugh (PO 14) believes that Aristoteles' hesitations are significant, but he makes no sustained attempt to show this. Cf. Allen's reply to Brumbaugh (PP 308 n. 78).


others) and V (if the One is, nothing can be said of the others in relation to the One). And the only ones free of flaws would be VII (if the One is not, everything will be inconsistently true of the others in relation to themselves) and VIII (if the one is not, nothing can be said of the others in relation to the One). But there seems no reason why Plato would want us to find the former two least persuasive and the latter two most persuasive. Nor does the initial hesitation seem always to coincide with a crucial step of the argument. This suggests that, unlike in Part 1, there is nothing to be gained by looking for keys to the refutation of Parmenides' arguments in the replies of his respondent.

Young Socrates was portrayed as astute and impressive, and it was his own theory that was under discussion. One would expect his moments of hesitation to be revealing about the status of the argument. Young Aristoteles, on the other hand, was chosen to replace him in the second part because Parmenides—professing himself old, and tired, and dreading the labors before him—wanted the youngest, least demanding, respondent (137a-b). The fact that he is introduced as "the one who become one of the Thirty [tyrants]" (127d) does not seem calculated to encourage our confidence in his perspicuity. Moreover, since the theory under discussion has nothing otherwise to do with him, his hesitations do not seem to be more than casual responses.

To illustrate his recommendation that we consider what follows both from the truth and falsity of a hypothesis, Parmenides uses Zeno's hypothesis "if the many exist": (I) if the many are, we must ask what must happen to the many themselves in relation to themselves and in relation to the One, and what must happen to the One in relation to itself and to the many; and (II) if the many are not, what happens to the One and the many, both in regard to themselves and each other (136a). Schematically, this may be represented as shown in Figure 1.

The eight hypotheses that are subsequently considered, however, are posed not in terms of the "many" but in terms of the "One."[37] A schema that began from the One and followed the same order as the preceding example would look like the following (Parmenides subsequently substitutes the term "the others" for "the many"). Comparing it with the actual order of the hypotheses (the last line) we can see that the first four hypotheses follow this schema exactly, but the last four do

[37] Strictly speaking there are only two hypotheses: if the One is, and if the One is not. The number eight is reached by specifying each of these two hypotheses in four different ways. It has become conventional, however, to refer to each of the specifications as a hypothesis, and I shall observe that convention here.



Figure 1


Figure 2

not (see Fig. 2). Because the previous hypotheses were based on the opposition between the many and the One, it seems likely that "the others" is meant as a synonym for "the many." In most of its appearances this seems to be the case, and in one place the equivalence is explicit (165e), but "the others" is denotatively less determinate, which will give the argument greater flexibility in places. (See below, note 44.)

A full treatment of the problem ought to repeat the eight hypotheses, once under the hypothesis that the others exist, and again under the hypothesis that they do not; or at least the eight hypotheses about the many should be added to the eight about the One. Moreover, Parmenides also says that he must do the same with likeness, unlikeness, motion, rest, genesis and destruction, and being and not-being. In other words one must consider each thing under both hypotheses, both in relation to itself and to every other thing that you wish, "if you are to


train yourself to see through to the truth in a completely masterful way" (136b-c). The eight hypotheses dealt with here are therefore arbitrarily limited, and any attempt to see them as representing an eight-part complete classification of the relationship between the One and the many would be questionable.

Even prior to Proclus there had been some question as to whether the hypotheses are to be regarded as eight or nine in number. The appendix to the first two hypotheses begins with the words:

Let us discuss it again for the third time. If the One is as we described it, being both one and many and neither one nor many, and participating in time: will it not be necessary, on one hand, that since the One is, it will at one time participate in being, and on the other hand, since it is not, it will also at some time not participate in being?

The reference to "the third time" has convinced some readers, from antiquity to the present, that the appendix is a separate hypothesis, and that the entire argument should be construed as comprising nine sub-arguments rather than eight.[38] In that case, just as the third argument would reconcile the tension between the first two, the sixth might reconcile that between the fourth and fifth, and the ninth that between the seventh and eighth. The scheme might even be elaborate enough for the last group of three to reconcile a tension between the first two groups. Such an interpretation would accord well with Plato's proclivity for triadic groups, and even so intricate an architecture as this would be compatible with the care with which he wrote. But I cannot see any convincing way of so interpreting the present arguments. Moreover, even apart from the fact that the initial schema works by bifurcation and therefore implies eight rather than nine possibilities, the hypothesis with which the appendix begins is, "if the One is as we described it," which is a different order of hypothesis from the others ("if the One is" [i.e., "exists"], "if the One is not").

The basis for Plato's choice of the concepts dealt with within each hypothesis, and for the order in which they are presented, has been a matter of considerable speculation. Cornford makes a convincing case for ascribing it to the treatise of Zeno,[39] but it is not a question that will

[38] Proclus pp. 352, 402. The most recent defenders of this schema are Sternfeld and Zyskind, chap. 3.

[39] "From the fragments and from references in Plato and Proclus we obtain the following list of contraries which appeared in the various sections. Some conjectural items are added in italics [Cornford's sources have been omitted from this summary]: One and Many. Divisible and Indivisible. Finite and Infinite in number. At Rest and in Motion. In itself and in another . Same and Different. Like and Unlike. In contact and Not in contact . Large and Small. Equal and Unequal. This list corresponds pretty closely with the series of contraries in the Hypotheses of the Parmenides " (PP 57).


concern us here. More important is the question of what the arguments are intended to demonstrate.

The final contusion of the argument is that "whether the One is or is not, both it and the others, both in relation to themselves and each other, both are and are not in every way, and both appear and do not appear to be so" (166c). This leaves us with two alternatives: either the object of the discussion is to produce complete skepticism with regard to the powers of philosophy (i.e., misology), or the paradoxes are intended to motivate us to discern hidden distinctions that render the paradoxes apparent rather than real. Clearly the second of these two is what we would expect of Plato, especially in light of Parmenides' stated aim of developing Socrates' power of conceptual reasoning (135c-d); and an examination of the eight hypotheses will bear this out. In fact the seventh hypothesis introduces the distinction between false appearance and truth. In the following synopses of the hypotheses I have tried to reduce some very complicated arguments to what I take to be their basic structure. But whatever success I may have had in tracing that structure accurately has had to be purchased through the sacrifice of considerable richness of detail.

10. If the One Is (137c-160b)

Hypothesis I. If the One Is, What Follows for It in Relation to Itself: Nothing Can be Said of It (137c-142A)

The argument may be divided into two distinct sets of deductions, each of which follows from an initial admission: (a) that the One cannot have parts, and (b) that we cannot predicate of the One any attribute whose meaning is not identical with "one." (a) Since the One cannot be many, it cannot be a whole, which would imply multiple parts (137c-d); since it has no parts, it cannot have shape (it is indefinite,

), or therefore place, or therefore motion or rest. (b) It cannot even be the same as itself because, since "same" is not identical with "one," multiplicity would result (139b-e); if it cannot be called "same" it cannot be called "like" or "equal" and therefore cannot be spoken of in terms of age or time, and if it cannot be said to exist at some time, it cannot be said


even to exist (141e). Accordingly there is no name, logos, knowledge, or perception of it.

All the admissions that were deduced from a could have been deduced from b alone. None of the terms, "whole," "shape," "place," "motion," or "rest," means the same thing as the term "one," and therefore—according to b 's principle that nonidentical attributes imply a lack of unity—to predicate any Of them of the One is to introduce multiplicity and destroy its oneness. But Parmenides does not generalize that principle, and therefore makes use of two premises instead of just premise b . Even though Plato might have dispensed with principle a on grounds of logical economy, it is likely that he wanted to bring it in because of the important role that the whole-part relationship will have here (and subsequently in the Theaetetus ).

Hypothesis II. If the One Is, What Follows for It In Relation to the Others: Everything Is (Inconsistently) True of It (142A-155E)

This argument may be divided into four main sections, which are sequential rather than (as in Hypothesis I) parallel.

a. If the One is, it must participate in and therefore be different from being (

). In that case the One and being (existence) must be parts of "the existent One" (
) (142b-d). Thus the existent One is a whole of parts, and since each part is again an existent One, the whole is divided ad infinitum (
) and contains unlimited (
) number.

b. Since the One is divided, it cannot be a whole; it can only be parts limited by an extrinsic whole (144d). (This contradicts 137c, which states that to be a whole it must be divided into parts.)[40] Consequently the One is both one and many, whole and parts, limited and unlimited; and since it is both whole and parts, it is both in itself and in another.

c. From this it is fallaciously deduced (by equivocation) that the One must be not only at rest but also in motion, since "in another" means "changing." (But in fact "in another" was previously used in a

[40] There is a similar tension in the Theaetetus . At 204a-b Theaetetus distinguished between "wholes," which are not reducible to their parts, and "sums," which are so reducible. Socrates refutes this at 205a simply by begging the question against Theaetetus and assuming that a whole too is simply the sum of its parts.


very different sense, that of spatial inclusion.) And it must be not only the same as itself but other than itself as well, because it is "in itself" and the relation "in" can hold only between different things. (Another equivocation.) The One must also be both other than the others (by definition) and the same as the others (because it is other than itself: 145e-147b). Because it thus participates in sameness and otherness it must be both like and unlike itself and the others. It also follows that the One both touches and does not touch itself and the others (again because of the "in" relationship), and is both equal and unequal to itself and the others (147c-131e).

d. Finally, since the One is, and "is" implies present time, the One participates in time and grows older. But if it is becoming older than itself, then the latter is becoming relatively younger, so it is both older and younger than itself. However, the age difference between two existing things can never increase or decrease, so it must be the same age as itself. Consequently "the One both is and becomes both older and younger than both itself and the others, and neither is nor becomes either older or younger than either itself or the others" (153c). Accordingly, there is knowledge, opinion, and perception, as well as name, logos, and so on, of it.

These first two hypotheses obviously are designed to appear to contradict each other to the greatest possible extent. The challenge to the reader is to dissolve the contradictions as far as possible by drawing out suppressed distinctions, as part of the training referred to at 133c-d. Most obviously, some of the internal contradictions of II disappear if we distinguish a purely conceptual sense of "in itself" and "in another" from the spatial sense given to them in c . But the other ambiguities are less straightforward.

In hypothesis I there are no indefensible inferences. There are only inconsistencies with hypothesis II. Even the conclusion—that nothing can be said of the One, not even that it exists—makes sense if one pushes the historical Parmenides' own position to the extreme. At a purely literal level Parmenides' conclusion here may appear incompatible with the historical Parmenides' poem, which insists above all else that "it is." But the poem also avers that "it is one." The present argument shows that if we attempt to think the oneness of the world we must abstract from all determinations. As soon as we even conceive of a duality between oneness and the multiplicity of determinations, the oneness that we attempt to conceive collapses.


As for the exclusion of the One from the determination of existence in particular, it is important to notice how this conclusion is arrived at. The historical Parmenides had argued that "it" is not temporal: one cannot speak of it in the past tense or future tense, but only in terms of the timeless now (fr. 8.5). These claims were echoed in the course of the first hypothesis's argument, and it was then concluded that what does not exist in time cannot exist at all. Accordingly the argument shows only that the One cannot exist in the sense of having temporal existence. It does not occur to Aristoteles to wonder whether there might be some mode of being that is nontemporal. But Parmenides and Plato certainly thought that there was. The conclusion of this first hypothesis, that the One does not participate in being (

), echoes the Republic's doctrine that the good is beyond being (

In both cases there is at work a distinction between "being" in the sense that is applied to temporal reality, and in some other sense that is related to the first only analogically. We saw that this "ontological difference" was pointed to in Part 1 by the Third Man argument: "large" cannot mean the same thing with reference to forms as it does with reference to spatial beings. The same is true of "existence." The being of forms is of a different order from that of physical things, and if we define existence with reference to the latter (as does hypothesis I), the former cannot be said to "exist." This conclusion follows whether we take the One to refer to the overarching principle of the good, or to the multiplicity of individual forms. Insofar as a form is a one-over-many with regard to physical things, it cannot be said to exist in their sense. And insofar as the good is the one-over-many with regard to the mul-


tiple forms, neither can it be said to exist in their sense. The three are related as individuality, form, and telos. In Plato, as in Aristotle, the formal cause and the final cause never completely coincide.

The inconsistency of hypothesis I with hypothesis II arises because the latter begins with a premise in flagrant contradiction to the "transcendent" position of hypothesis I. Whereas the first hypothesis ended by arguing that the One does not even participate in existence (

. . .
, 141e), the second hypothesis starts with the claim that if the One is, it must participate in existence (
. . .
, 142b). Once this is granted it becomes a wedge to split the One progressively into an unlimited many, until we end up with a position that is the very reverse of Parmenides'—not unity but indefinite divisibility. Perhaps Plato has in mind the philosophy of Heracleitus, which he contrasts with that of Parmenides in the next dialogue, the Theaetetus . Or perhaps, as the frequent reiteration of the word
("unlimited") suggests, he has Anaximander in mind. But the primary candidate may be Anaxagoras, with his doctrine of infinite divisibility. The dialogue begins with the request by disciples of Anaxagoras, who have traveled from Clazomenae, to hear the account of Socrates' meeting with Parmenides.[42] Whether Plato had one of these philosophers in particular in mind, or the tradition as a whole that emphasized the unlimited aspect of reality, what is important is the contrast between this view and the view put forward in the first hypothesis: the contrast between unity and dissolution. Although historical antecedents of this contrast may be postulated in terms of specific historical figures, its primary importance for Plato is in terms of his own contrast, as reported by Aristotle, between the One and the indefinite dyad, indefinite plurality.[43]

The first hypothesis conceives of the One in relation to itself, while the second conceives of it in relation to the unlimited many. The One of the first hypothesis is what is beyond being, like the Idea of the good, while that of the second hypothesis participates in being. It is the difference between a first principle conceived as independent of space and time (i.e., of being qua spatiotemporal existence), and one conceived as spatial and temporal. Hypotheses I and II do not ultimately contradict each other, because their conclusions are predicated of different subjects—of transcendent being in one case and of immanent being in the

[42] Proclus attests that "the visiting nature philosophers [were] the followers of the teaching of Anaxagoras" (p. 25). Brumbaugh (PO , and "The Purpose of Plato's Parmenides" [Ancient Philosophy 1 (1980) 39-48]) and Miller (PP 26-28) stress the importance of this for the context of the dialogue as a whole.

[43] Aristotle, Metaphysics A.6.987 25-988 1.


other. This explains the discrepancies between the two hypotheses on the nature of "whole" and "unlimited" (

). According to hypothesis I, which considers the One only in relation to itself, the One cannot be a whole, because a whole implies parts (137c-d). According to hypothesis II, which considers the One in relation to the others, the One cannot be a whole precisely because it is partitioned, and wholeness implies being undivided (144d). Thus the whole is in one sense something that transcends its parts, and in another sense something that reduces to its parts. Once again the distinction reflects the difference between the Parmenidean and pluralistic viewpoints. The same ambiguity in the conception of whole and part will reappear in the Theaetetus (203e-205e), directly in the shadow of the distinction between the Heracleitean and Parmenidean worldviews (180d-184a).

Throughout the eight hypotheses, the ambiguity between whole and parts is reflected generally in an ambiguous use of the term "the One." In the first hypothesis the One implies a supervenient unity. In the second it refers not to a transcendent principle but to the aggregate of material reality. In the third it will refer both to the latter (qua whole) and to the unity of each part of the whole.[44] The ambivalences among these three senses of the One (transcendent unity, totality, individuality of each part) recur throughout the eight hypotheses and reflect precisely the moments of the whole-part dialectic: wholeness is the principle of unification, but since the entity itself is an aggregate of parts, the wholeness in some sense transcends the physical entity as such; on the other hand, since this wholeness is of the physical entity, the entity has an inherent integrity and so a wholeness that is immanent ; but this implies parts and therefore conceptual dissolution of the wholeness of the entity into component unities, which are themselves wholes.[45]

[44] The ambiguity of the term "others" depends on that of the term "One." The following chart shows how the two terms are used in the various hypotheses:










immanent, infinitely divisible


infinitely divisible [unities?]




mediating between unlimited and form




without characteristics








without characteristics


nonexistent (unity)


apparent, not real, characteristics


nonexistent (unity)



[45] I suspect that these are the same three senses that Proclus has in mind when he writes, "'The One' can be used in three senses. We have the One that transcends all beings, and that which is present together with all beings, which also, with the One, produces all the orders of beings, and thirdly we have that which is inferior to Being and which is, as it were, 'swallowed down' by it" (p. 381).


In the same sense that the One is both a whole and parts, it is both limit (whole) and unlimited (parts: 144e-145a). This is implicit also in the ambiguous way that "unlimited" (

) has been used in these two hypotheses. In hypothesis I it means "formless," that is, beyond spatiality (137d), whereas in hypothesis II it means coextensive with spatiality and therefore "unlimitedly divisible" (143a). "Unlimited" (beyond spatiality) in the first hypothesis corresponds to "limit" (beyond divisibility) in the second hypothesis, that is, the supervenient whole. The general contrast between the two anticipates not only the distinction between limit and unlimited in the Philebus (23c-25b), but also parallel distinctions in other dialogues, such as between the formalists and the materialists in the Sophist (246a-c), between relative measure and the mean in the Statesman (283e-285c), and between reason and chaos in the Timaeus (29a-30b). It also recalls such earlier accounts as the Symposium's opposition between the divine and the human (202c-d) and the Phaedo's conflict between the materialistic hypothesis (96a-e) and the hypothesis of forms (100b-e). None of the other dialogues, however, presents the opposition as starkly as does the Parmenides . In each of the other dialogues the antithesis is resolved by means of a synthesis of the two,[46] but here any such synthesis must be conjectural. The key to the possibility of such a synthesis in the Parmenides lies in the appendix to the first two hypotheses.

This appendix, which combines both moments of the previous antinomy, concludes that the One sometimes participates in being and sometimes does not. But since it cannot "sometimes have and sometimes not have the same thing unless it receives it at some time and again loses it," it must go through successive stages of generation and destruction, changing from the One to the many and back again (155e-156b). The conclusion loses its paradoxical character if we can interpret these successive stages as changes of aspect or relation rather than changes of state. Such an interpretation is suggested by the previous antinomy, since the first hypothesis considers the One in relation to itself and the second considers it in relation to the others. The paradox was generated by obscuring the difference between these two aspects: reality appears

[46] Philebus 25b-27b, Sophist 246c-249d, Timaeus 47e-48a, Symposium 202d-203a, Phaedo 105b-c. Since the Statesman is there making use of the method of division, the synthesis is the unity of the two within the genus of measure.


as both one and many, depending on whether one looks to its unified character or its diversity.

Parmenides himself provides the foundation for resolving the paradox in this way. When the One alternates between unity and multiplicity it alternates between motion and rest, and when it does so "it is necessary for it not to be in any time." The point of change between motion and rest must itself neither be the one nor the other, "and there is no time in which something can neither be in motion nor at rest at once" (156c). Parmenides introduces the concept of the instant (

) as this limiting case of time in which change can be explained (156c-e), and goes on to suggest that the changes of the One between existence and nonexistence, and between being one and many, and so forth, are to be explained in the same way (156e-157b). The fact that these changes are said to take place not in time but at the limiting case of time, at the boundary of time and nontime, provides a basis for regarding the switch between unity and multiplicity as nontemporal also in the way suggested above, that is, as a change of aspect rather than a change of state. If the change between unity and plurality occurs at the boundary between time and nontime, that change is evidently nothing but the contiguous sides of the boundary itself.

The concept of the instant provides us with the mediation between the Parmenidean and pluralistic antinomies inasmuch as it posits a point of contact between the timeless and the temporal, between the supervenient unifying principle of reality and physical reality itself. It is in this intersection, too, that we must conceive participation and recollection to occur, and accordingly the concept of the instant provides an elusive but nonmetaphorical substitute for the metaphors whose limitations were exposed by the arguments of Part 1.

Hypothesis III. If the One Is, What Follows for the Others In Relation to Themselves: Everything Is (Inconsistently) True of Them (157B-159B)

Here again, as with hypothesis II, the inconsistencies will turn out to be only apparent. Since the others are other than one they cannot be unities and therefore must consist of parts. The parts are parts not of a many (

) or a sum (
) but of a whole (
), because a many or a sum has no character as a totality but is simply each part.[47] But as

[47] This is perhaps the most obscure passage in the dialogue. A literal, if not always intelligible, translation may be rendered: "If something is a part of a many, in which it itself is, it will be a part of itself, which is impossible, and of each one of the others, if indeed it is a part of all. For if it is not a part of one of them, it will be a part of the others besides this one, and thus it will not be a part of each one, and not being a part of each it will be a part of none of the many. But it is impossible for something that is a part of none of these many, of which it is not a part of any, to be any other kind of part at all" (157c-d).

The argument seems to be that if we call something a part of a many, this is different from calling it a member of the many, or a part of one of the members. To be a part of the many means to be a part of the many distributively , i.e., to be a part of each member. So in the case of something that is a part of a many in which it is also a member, then it must be a part of itself, which is impossible. Plato's point would then be that if we speak of reality in terms of wholes and parts, this should not be assimilated to the model of a sum (or a many) and its members.


parts of a whole they participate in one Idea (

, 157e), and each part also participates in the One since it is a unity. But before they participated in the One they participated in a form (
) by virtue of which each of them is unlimited in quantity (
).[48] Thus in relation to the One they participate in limit, but in their own nature they are unlimited (158c-d). Because of this duality they are also both like and unlike, the same and not the same, in motion and at rest, and so on (158e-159b).

We can see that this hypothesis develops the synthesis, only hinted at in the appendix, of the opposition established in the first two. Here is made explicit the opposition of limit and unlimited as connected with the One and the infinitely divisible many, and the ambiguity of the totality as a supervenient unity (whole) or mere aggregation (sum).[49] The relationship between the many and their supervenient character (whether as unified or unlimitedly divisible) is expressed in terms of participation in forms; the metaphor is reintroduced after the abstract formulation of the appendix. It is not therefore Plato's intention to do away with metaphor entirely, which would hardly be possible, but to make us aware of our reliance on it and the way that this reliance limits the transparency of our concepts. The Statesman , in fact, will defend the importance of metaphoric and analogic "paradigms."

Hypothesis IV. If the One Is, What Follows for the Others In Relation to the One: Nothing Can be Said of Them (l59B-160B)

Since the One and the others are jointly exhaustive, there can be no third entity to mediate between them, so they must be absolutely sepa-

[48] This disjunction between specific form and unity can only be ex hypothesi for Plato.

[49] Cf. Theaetetus 203e-205e.


rate, and the others cannot participate in the One. Since the others cannot in any sense be one, neither can they be many, since each would then be one part. Nor can we speak of them in terms of likeness or unlikeness, because then they would participate in two forms (

, 159e;
, 160a), which is impossible if they cannot participate even in one. Since they are unable to participate in anything, they cannot be characterized in any way at all.

If the third hypothesis provided a positive mediation between the antitheses of the first two, the fourth provides its negative counterpart by showing that unless participation in forms is possible, we will have the result that Parmenides had warned of in the first part of the dialogue: "If anyone . . . does not admit the existence of forms of things or mark off a form under which each individual thing is classed, he will not have anything on which to fix his thoughts . . . and in this way he will utterly destroy the power of discourse" (135b-c). The present argument shows that some such postulation must be accepted if meaningful speech is to be at all possible. It thus anticipates the Theaetetus's pervasive argument that a world of flux without stable forms is unintelligible and would make speech unintelligible.[50] The premise on which this fourth hypothesis rests, that since the One and the others are jointly exhaustive there can be no tertium quid through which they can stand in relation to one another, had already been answered in the appendix in terms of the concept of the instant as the tertium quid between motion and rest, and between all such other oppositions. It had also been answered in the middle dialogues in terms of the relation of participation, by which the unity of the form and the multiplicity of corporeality are mediated. That very term appears here eight times in the half-page between 159d and 160b.

If Parmenides had asked whether anything could exist besides the One and the many , "the relationship between them" would have suggested itself as a possible answer. But because he had substituted for the term "many" the more general term "others,"[51] the question is put in a way that precludes that answer and exaggerates the paradox. His phrase "the things that are other than the One" (159b) must by definition include even the relations between the One and what is other than the One, and so arbitrarily excludes any answer.

Combining the first four hypotheses, Parmenides concludes that "if One exists, the One is all things and nothing at all in relation both to

[50] Cf. esp. 182c-183b.

[51] At the end of hypothesis VII "the many" is used as a synonym for "the others" (165e).


itself and to all others" (160b). We have seen, however, that although on the surface these conclusions appear to cancel one another out, reducing their mutual antecedent to absurdity, the details of the arguments reveal a consistent underlying teaching. They show how the theory of forms mediates the antagonism between the Parmenidean and pluralistic worldviews and overcomes the one-sidedness of each.

11. If the One Is Not (160b- 166c)

The second group of four hypotheses is odd in two ways. First, three of the hypotheses are absurd in their very formulation. If the One does not exist, it makes no sense to ask how the One is related to itself or to the others, or how the others are related to it. Second, it seems odd that Plato's Parmenides would even entertain the hypothesis that the One is not, since the historical Parmenides vehemently denied that it is possible to say or even think that "it is not."[52] Perhaps, then, the second four hypotheses are seriously meant to be what the first four only appeared to be: a reductio ad absurdum of the antecedent. In that case they would amount to an indirect demonstration of the One by showing the incoherence of its denial. This group, unlike the previous one, does not contain an implicit escape from incoherence. The arguments may further function as a way of preventing the first four from being taken as a reductio , for there is no point in rejecting the hypothesis that the One exists, on the grounds that it leads to absurdity, if the rejection of that hypothesis leads to absurdity as well. Even if this group functions as a reductio , however, the care that Plato has put into the arguments suggests that there is something to be learned from their content as well as their form.

Hypothesis V. If the One Is Not, What Follows for the One In Relation to the Others: Everything Is (Inconsistently,) True of It (160B-163B)

To say that the One is not, it must be that we can distinguish the One from other things, and we must therefore have knowledge of it. It must therefore partake of signifiers such as "that," "some," "this," and so on, and relations such as likeness and unlikeness, equality and inequality, and so forth, and we will be able to say what is true of it. "But if we say what is true, it is dear that what we speak of must exist" (161e). Not only does what-is-not thus participate in being, but what-is participates in not-being, insofar as its own nonbeing is not , and so being and

[52] Cf. fragments 2, 6, 7. Cornford, however, takes these hypotheses to be Plato's controversion of the historical Parmenides' claim (PP 220, 240).


not-being participate in themselves and in each other (162a-b). Since it participates in opposites it must change, but since it is neither in space nor capable of becoming other than itself it must be at rest. Thus it must both change and not change, both come into being and perish, and neither come into being nor perish.

This seems to be an indirect confirmation, by reductio , of Parmenides' warning against trying to speak about nonbeing. To speak of non-being is to treat it as a being, which leads to absurdity. In the Sophist a disciple of Parmenides will show that it is possible in one sense to speak of nonbeing, but only in a predicational sense, not (as here) in an existential sense. This indirect consequence of the present argument becomes the direct consequence of the next one.

Hypothesis VI. If the One Is Not, What Follows for the One In Relation to Itself: Nothing Can Be Said of It (163B--164A)

Since the One is not, it cannot participate in anything, and therefore none of the above qualities (nor any others) pertain to it.

Hypothesis VII. If the One Is Not, What Follows for the Others In Relation To Themselves: Everything Is (Inconsistently) True of Them (164B-165E)

The others must exist if we can speak of them, but they cannot be "other" than the One, for ex hypothesi there is no One, so they must be "other" than each other. They can only be so in general (

) rather than qua individuals, since individuals imply unity, which does not exist. A given mass of "others" is unlimited (
) in quantity and infinitely divisible, and is therefore both one (in appearance) and many (in reality). Since they combine apparent unity with multiplicity, they will appear to possess number, largeness, smallness, equality, and limit. Moreover, depending on whether one looks to their appearance or their reality, they will appear as both unlimited and limited, one and many, like and unlike, same and different, touching and separate, in motion and at rest, coming to be and perishing, and the rest.

The contradictions are of course only apparent, since they are between what the many appear to be (as we imagine them now) and what they really are. The hypothesis reverses the Platonic worldview: the inner reality beyond. appearances is no longer the unity and form of things, but their diversity and formlessness. To be precise, given the indefinite and individual nature of these "others," Parmenides is talking


about what would later be called prime matter, or what the Timaeus calls chaos, that is, the material basis of existence hypothetically denuded of all form. Thus, as with the preceding pairs of hypotheses, the consequences of this first member of the fourth pair only apparently contradict those of its companion. Here we are made to realize that without oneness reality would reduce to unformed matter. The next hypothesis shows that nothing can be conceived or said about reality so conceived. As Aristotle was subsequently to conceive it, matter is only potentially for existence, which cannot become actual existence in the absence of form (unity).

Hypothesis VIII. If the One Is Not, What Follows for the Others In Relation To the One: Nothing Can Be Said of Them (165E-166C)

Since "many" is a plurality of ones, then if there is no such thing as "one," the others not only cannot be one, but cannot be many either. Nor do they even appear to be one or many, because what is cannot be conceived of (

) in terms of what is not. If they cannot be conceived in relation to the One, they cannot be conceived as like or unlike, same or different, touching or separate, or anything else. It follows that "if the One is not, nothing is."

12. Conclusion

To put the results of our examination of the second part of the dialogue as succinctly as. possible, the distinctions implicit in the first four hypotheses have led to the conclusion that unless there is something by which the limiting One and the unlimited many can stand in relation to each other, meaningful speech and thought will be impossible. The next four illustrate this point negatively by showing what happens if we try to conceive of the many otherwise than in relation to the One. Hypotheses VII and VIII show negatively and positively that reality without unity reduces to something like pure potentiality. Such a reality may appear (to our imagination) as a kind of existence (VII), but it cannot in fact exist without unity (VIII). In that case any attempt even to conceive or speak of it is deluded, for it will appear to be both incoherent in relation to others (V) and empty in relation to itself (VI).

The Republic's remark that the forms are grounded in the good, both in terms of their knowability and their existence (509b), shows


that all things in the world of becoming, which participate directly in the "specific" oneness of a form (being), also participate indirectly in an "absolute" One, the good, which is beyond being. The fundamental problem for the theory of forms is to give an account of the nature of these two relationships.[53] The first part of the Parmenides casts doubt on our ability to give a nonmetaphorical account of the nature of participation by things in forms (i.e., by "becoming" in "being"), while concluding, however, that unless we nevertheless affirm the existence of forms and participation we will not be able to account for the possibility of thought and discourse (135b-c). The second part of the dialogue has now shown, in its dizzying way, that a similar conclusion follows from a consideration of the relationship between our world of becoming and the One that is beyond being. Once we enter into the "gymnastic" dimension of these arguments and respond to their challenge to draw crucial distinctions, we can distinguish between the arguments that reflect genuine paradoxes, and those that are only formulated in a paradoxical manner but can readily be resolved by means of the distinctions with which Plato supplies us. This has led us to conclude that although the relationship between the temporal (becoming) and the timeless (the One) can be hinted at in terms of the concept of the "instant," a fully elaborated account remains recalcitrant, and the third hypothesis is forced to return to the metaphor of participation. But just as the first part of the dialogue concluded by saying that, despite the problems of conceptualizing the relation of participation, to attempt to do away with that relationship would destroy the possibility of thought and discourse; here too the fourth hypothesis shows the impossibility of dispensing with the relation of participation. And the final four hypotheses show that it is impossible to avoid the problems of the first four simply by doing away with the concept of the One (whether applied to the specific forms or to the good), for in that case the resulting paradoxes become wholly intolerable.

We now turn to the Theaetetus , which can, in fact, readily be interpreted as an exploration of the claim that, if someone does not recognize the existence of forms, "he will not have anything on which to fix his thoughts, as long as he does not permit the idea of each thing to be always the same, and in this way he will utterly destroy the power of logos."

[53] Cf. PP chap. 8 §5.


Chapter Two
The Theaetetus

1. Knowledge and Virtue (142a-153e)

Near the beginning of the dialogue Socrates asks Theaetetus whether knowledge and wisdom are the same thing (145e). Theaetetus answers in the affirmative, and Socrates responds noncommittally: "Now it is this very thing that perplexes me, and I am not sufficiently capable of grasping by myself what knowledge is." There matters are allowed to rest. After Republic 4's analysis of human wisdom as knowledge together with self-mastery (the subordination of appetite and competitiveness to reason in the tripartite soul, 442c), this uncritical identification of wisdom with knowledge is provocative, and leads us to wonder whether the aporetic ending of the dialogue is in any way connected with this oversimplified beginning. Plato's readers would hardly have forgotten the doctrine of the tripartite soul so quickly, and there seems to be a deliberate reminder of it in the names of the initial speakers, Eucleides ("Renown") and Terpsion ("Delight"), which correspond to the two lower motivations of the soul, "love of honor" and "love of pleasure."

Later, in the long, central digression that recalls the Republic and Phaedo ,[1] where Socrates speaks not as a midwife but from his own


conviction, he says that we must seek to escape the evils of the corporeal world, and that "to escape means assimilation to god, and assimilation means to become just and pious, with wisdom" (176b). The word for wisdom here is phronêsis (sophia is used in the earlier passage and in the Republic ), but the point is the same, that the highest rational attainment goes beyond intellectual knowledge alone (which may be in the service of our irrational passions) and involves the subordination of "corporeal evils" to the "divine." Accordingly, Socrates goes on to say that "knowledge of this [how to assimilate ourselves to god] is true wisdom [sophia this time] and virtue" (176c).

There are more direct reminders of the Republic doctrine. Before Socrates meets Theaetetus, Theodorus describes him as being quick to learn, gentle, and courageous, and remarks that this is a combination which otherwise

I would not have supposed to exist, nor do I see it. Rather, those who are as sharp as he is, and quick and with retentive memories, are also for the most part quick-tempered, . . . manic rather than courageous. Those on the other hand who are more sedate are also somewhat sluggish when they come up against their studies, and are forgetful.

We can restate this passage, which recalls the qualities sought for in the guardian class of the Republic (2.375b-376c), in terms of the categories of that dialogue: intelligent people are almost always dominated by their spirited nature, and those who are not spirited tend to be lazy or sluggish, that is, more interested in comfort than in effort. Most people are thus dominated by love of honor or love of pleasure, and it is only Theaetetus's nature that makes Theodorus realize that it is possible to be intellectual without being dominated by spiritedness: that is, that love of reason is distinct from love of honor, and that there are three types of persons rather than two (this classification will become important again at the end of the trilogy, in the contusion of the Statesman ).

The lazy, forgetful type was exemplified by the two characters with whom the dialogue opens.[2] When Terpsion asks Eucleides if he can repeat the conversation that Theaetetus had with Socrates, he replies,

[2] Although their names reflect the opposition between appetite and spiritedness, their behavior seems in both cases to be motivated by a variety of appetite. The excessively spirited type, on the other hand, is mentioned only indirectly, in Socrates' reference to the irrationally angry response of people whose views he has shown to be false (151c-d).


"No, by Zeus! Not by heart" (142e). But Socrates has recited it for him verbatim and refreshed his memory every time Eucleides went to Athens, until by now Eucleides has almost all of it written down (143a). Not only is Eucleides' memory not impressive, but he has gone about this task in a lazy, piecemeal fashion. Terpsion too is unimpressive for his memory or energy: he has "always intended" to ask Eucleides to read the account of the conversation between Socrates and Theaetetus, but has not actually done so until now—whether out of forgetfulness or lazy procrastination, he does not say. He wants to hear it now, however, because he is tired and needs to rest (143a). Eucleides, too, is tired and would like a rest, so he decides to have his slave read the conversation to them while they are resting (143b). He also mentions that he put the conversation in the form of direct discourse rather than narrative because it would be too much trouble (

) to put in Socrates' narrative asides, such as, "And I said," between all the speeches (143c).

This combination of poor memory and lazy lack of spirit becomes immediately evident within the conversation in the character of Theodorus, who cannot remember who Theaetetus' father is (although Socrates, who has never met him, knows his father and native city as soon as he sees Theaetetus),[3] and who fearfully resists every attempt to draw him into the discussion.[4]

The significance of memory, which Plato foreshadows by these dramaturgic means, will emerge in due course; but some preliminary remarks may be made about courage (which is here opposed both to laziness and to fearfulness). Throughout the ensuing discussion the need for courage and boldness is continually emphasized.[5] A clue to the reason for this may be found in a passage of the Meno . There, after introducing the doctrine of recollection, Socrates concludes that this refutes Meno's paradox by showing that learning is possible

if one is courageous and does not desist from seeking; for seeking and learning are the whole of recollection. One must not be convinced by that conten-

[3] Ronald Polansky suggests another interpretation of this: "Theodorus, by contrast [with Socrates], is so unconcerned with the fathers of the youths he instructs that he forgets the name of Theaetetus' parent. His mathematical instruction demands little political caution or involvement with the personal background of the youth" (Philosophy and Knowledge: A Commentary on Plato's Theaetetus [Lewisburg: Bucknell University Press, 1992] 40).

[4] 146b, 162a-b, 165a-b, 169a-b, 177c, 183c-d.

[5] E.g., 148d, 151d, 151d-e, 157d, 166a-b, 177d, 187b. Cf. 153b. See also Sophist 261b. "Courage" in this sense, however, does not necessarily imply all the attributes ascribed to it as one of the cardinal virtues in Republic 4 (442c).


tious doctrine [Meno's paradox]; that doctrine will make us lazy and is pleasant for soft people to hear. This one, however, makes people energetic and searching.
     (81d, emphasis added)

The Theaetetus , in fact, recalls the Meno at almost every turn. For example: (1) at 146c-d Theaetetus is rebuked for giving a list of examples in answer to the question, "What is knowledge?" as Meno had been for "What is virtue?" (72a). (2) At 148c Socrates offers a definitional example of day as earth mixed with water, as he had offered Meno the example of shape defined as what always accompanies color (75b). (3) The Meno took as its model the knowledge that the square root of an area of eight is not expressible as a whole number but may be expressed as a diagonal (82b-85b); the Theaetetus proceeds to take as an example of knowledge the distinction between areas whose roots are expressible as whole numbers (squares) and those whose are not (oblongs; 147d-148b).[6] (4) Socrates' remark that some people think he is a most strange person who reduces others to an impasse (149a) precisely echoes Meno's complaint at 79e-80b. (5) The Theaetetus (187b ff.), like the Meno (97a ff.), discusses knowledge by comparison with true opinion. (6) The Theaetetus more than once (198c-d; cf. 196d-e, 209e) alludes to Meno's paradox (80d). (7) After Anytus warns Socrates that he may find himself in serious trouble for critical remarks he has just made (94e), the Meno ends with Socrates saying, "Convince your friend Anytus of these things of which you are now convinced, so that he may become more calm. If you convince him you may also benefit the Athenians," a pointed anticipation of Socrates' trial and execution; the Theaetetus ends with Socrates going off to answer the indictment of Meletus, of which Anytus was coauthor. (8) F. M. Cornford, who notes several of the above as well, suggests also that Socrates' midwifery corresponds to the Meno's doctrine of recollection.[7] His suggestion is often rejected on the grounds that the answers elicited by midwifery are frequently wrong,[8] but we might say the same about the answers elicited by Socrates from Meno's slave (82e, 83e).

[6] Malcolm Brown points out that "already in antiquity this point of parallelism was made by the anonymous commentator on Theaetetus (eds. H. Dials and W. Schubart, Berliner Klassikertexte ii, 1905): 'the [side of the] .two-foot square is also incommensurable . . . but he left it out, they say, because it is in the Meno " (" Theaetetus : Knowledge as Continued Learning," Journal of the History of Philosophy 7 [1969] 360 n. 4).

[7] PTK 27-28. As a ninth parallel we may add Runciman's observation (p. 7) that the conclusion of the Theaetetus is only apparently aporetic, like that of the Meno .

[8] E.g., John McDowell, Plato: Theaetetus (Oxford: Clarendon Press, 1973) 117.


Nevertheless midwifery, unlike recollection, cannot be considered a necessary condition for attaining knowledge, so the correspondence is only approximate.

Whether or not Plato deliberately plays out the Theaetetus against the backdrop of the Meno's doctrine of recollection, the Theaetetus reaffirms the latter's claim that, whatever theoretical knowledge may be, it is not something easily acquired. It requires the courage to persist amid difficulties and frustrations, and the boldness to pursue hypotheses that may fly in the face of common sense. According to the digression in the middle of the dialogue, what is called for ultimately is nothing less than the courage to change our way of life. This is not the kind of knowledge that Theaetetus will take as his model, however. Urged on by Socrates' promise to act as the midwife of his conceptions (a metaphor that will function in a number of ways in the dialogue), Theaetetus brings forth the idea that knowledge is sense perception.

2. The Heracleitean-Protagorean Problematic (151e-160e)

That Plato should assign to a mathematician the role of defining knowledge as sense perception is not surprising when we consider that for the Greeks mathematics centered on geometry, whose proofs were illustrated by diagrams. The Meno , however, reminded us that what one learns only by looking at the diagrams is not knowledge at all. Socrates says there about the slave: "At present these opinions, having just been stirred up in him, are like a dream. If, however, one were to ask him the same things many times and in many ways, you know that finally he would have knowledge of them that is no less accurate than anyone's" (85c-d). The slave's opinions will not be transformed into knowledge until he frees himself from dependence on particular diagrams or formulations.

In the Theaetetus Socrates pursues the opposite path, moving within the realm of sense perception rather than abstracting from it. To begin with, Theaetetus's model of surds and roots bears only a superficial resemblance to the diagram of the Meno . Whereas Socrates used that diagram as a means of discovery, Theaetetus uses his model only as a means of classifying what is already known. It is a preliminary application of the method of collection and division ("we tried to collect them [the roots] into a unity," 147d; "we divided all number into two," 147e). Moreover, when Theaetetus uses this mathematical example as


an instance of knowledge, Socrates does not proceed to make a connection between the nature of mathematics and the nature of knowledge generally, and to use this as an impetus to lead his partner in the direction of the intelligible, as in previous dialogues. Socrates instead pushes him in the contrary direction, to the most phenomenalistic way of conceiving knowledge. The world is just as it seems to each observer. Plato begins the dialogue with the most elementary conception of knowledge, that is, the lowest grade of information, mere sense experience. From this he will generate under the pressure of criticism progressively more complete models, in accordance with the method of hypothesis.[9]

By beginning in this way Plato is able to respond to the attack on stability launched by Heracleitus, who insisted that conceptual distinctions are always arbitrary, regardless of whether they refer to values such as beauty, or factual demarcations such as up and down, day and night, or alive and dead. The world of thought, like the world of beings, is pure becoming or flux, and conceptual knowledge is therefore delusory. The next generation took the next step and wondered how, if Heracleitus is right, it is possible for him to say so without inconsistency. Accordingly, his disciple Cratylus rejected his teacher's claim that we cannot step into the same river twice, for we cannot step into it even once. And to make his point with greater consistency Cratylus abandoned speech altogether and limited himself to pointing with his finger.[10]

It is against this background that the Theaetetus takes place, a dialogue explicitly concerned with the Heracleitean foundations of fifth-century sophistry. Today's "postmodernists" have advanced worldviews that are parallel in some ways to that of the Ephesians. Like their Presocratic counterparts, they attempt to break down the perceived structures of experience into negativity and flow—the problematic of the Theaetetus is of interest today not only for historical reasons. It is no accident that the reaction of the analytic-minded mathematician Theodorus to the school of Heracleitus is evocative of the reaction of contemporary analysts and traditionalists to Deconstructionists:

It is no more possible, Socrates, to discuss these doctrines [with their adherents] . . . than with maniacs. For they are, in accordance with their trea-

[9] For this interpretation of the method of hypothesis, see my PP 127-38. Kenneth Sayre points out that "the Theaetetus unquestionably is Plato's most ambitious and sustained attempt to apply the method of hypothesis in matters of philosophic argumentation" (Plato's Analytic Method [Chicago: University of Chicago Press, 1969] 232).

[10] Aristotle, MetaphysicsG .5.1010 11-13.


tises, completely in motion; and as for keeping to an argument or a question and calmly answering and asking in turn, there is less than nothing of that in them . . . . If you ask one of them something, he pulls an enigmatic little phrase out of his quiver and shoots it off. And if you try to get an account of what be said, you will be hit anew by another turn of phrase. You will never reach any conclusion with any of them; nor, indeed, do they themselves with one another; but they take very good care to let nothing be stable, either in an argument or in their own souls.

Socrates' rejoinder, that the Heracleiteans probably exaggerate these qualifies to Theodorus because of his hostility, reminds us that Plato himself shows considerable respect for Heracleitus's doctrines, and in the Timaeus describes the cosmos as partly grounded in chaos. The degree of his endorsement of Heracleitean destructuring is obscured by his concern about the propriety of disseminating such views even if they are true. Half a page after Theodorus's remarks, Socrates says,

Have we not heard from the ancients, who concealed it from the many by means of poetry, that the origin of all things, Oceanus and Tethys, are flowing streams, and that nothing stands still? And also from the modems who, because they are wiser, reveal these things openly so that even the cobblers may hear them and learn their wisdom and cease from their foolish belief that some things stand still while others are in motion, and, once they have learned that all things are in motion, may honor these teachers?

In view of his belief that we do ordinary people no favor by convincing them that stability is an illusion, we must expect that whatever affinity Plato has for the views of Heracleitus will not be straightforwardly acknowledged. Nevertheless, these doctrines are taken very seriously in the Theaetetus . We need to consider how receptive Plato is to the objections against natural stability and to what extent his own philosophy of form justifies itself against the considerations that lead to the destabilizing of what appears to be stable. In the Parmenides Plato threw the theory of forms into uncertainty. And even though Parmenides reaffirmed the need for forms if thinking and discourse are to be possible, the forms are missing from the Theaetetus , at least on the surface. We would expect them to make an appearance when Socrates discusses the problem of quantitative relativity: it seems unproblematic to assert that six dice are more than four by a half and less than twelve by a half (154c), but, he says (in a dear echo of the method of hypothesis), we need "to observe our thoughts in relation to themselves, whichever ones we think, to see whether for us they are consonant with one another or


not at all" (154e; cf. Phaedo 100a, 101d). In the present case, three such beliefs produce tension with the statement about the relative size of numbers:

Nothing can ever become more or less, either in size or number, as long as it is equal to itself.

And second, that to which nothing is added and from which nothing is subtracted, is neither increased nor decreased, but is always equal.

Third, that something previously was not, but later is, without becoming, is impossible.

"These three admissions fight with themselves in our minds when we talk about the dice," or when we say that if Theaetetus grows taller, then Socrates goes from being taller to being shorter without changing size (155b-c). Their "fight with themselves" presumably consists of the fact that each of them seems clearly true when taken just by itself, but clearly false when applied to the relative largeness and smallness Of numbers (the dice) and sizes (Socrates and Theaetetus).[11] It is important to keep in mind that the fight must be a tension within each statement, rather than a tension among them, for the examples of the dice and Theaetetus falsify either all three of the admissions together or none at all. (It is misleading therefore to translate

as "contend with one another, " which implies that if we got rid of two of them the remaining one would be unproblematic.)

In the Phaedo (100e-103a) such problems are resolved by means of the theory of forms: relations like larger and smaller are not corporeal properties of individuals. They are therefore not subject to the three admissions mentioned above, which apply only to nonrelational subjects. They are conceived instead as relational essences, which are distinct from any corporeal individual, but which may be participated in by individuals in certain circumstances. Accordingly, we would not say, in violation of the first principle ("admission"), that Socrates, while remaining equal to himself, has gone from being taller to being shorter, but only that in one comparison he participates in the relation "taller" and in another "shorter." Nor would we say, in violation of the second principle, that Socrates has decreased without anything having been

[11] Polansky 94 puts it somewhat differently, by describing the tension as resulting from the possibility of interpreting each statement either as referring to something in itself, or as referring to it relative to something rise. Thus in each case the subject is unchanged in itself but may be different relative to something else.


subtracted from him, but only that he participates in a different relation because the size of the other referent (Theaetetus) has changed. Nor again, in violation of the third principle, that Socrates has gone from not-short to short without a process of becoming, but only that he participates in one relational form rather than the other because of a becoming that attached to the other referent.

Unlike the Phaedo , the Theaetetus makes no mention of the theory of forms and offers no solution. Theaetetus himself is left in a state of perplexity by the puzzles, and Socrates remarks, "This feeling—wonder—very much pertains to philosophy. For there is no other beginning of philosophy than this, and it seems that the one who said that Iris is the child of Wonder did not genealogize badly" (155d). This metaphor of "parent and child" pervades the Theaetetus . It was implicit at the beginning of this passage as well. If Socrates only wanted to illustrate the simple "paradox" that six dice could be both more (than a smaller quantity) and less (than a greater one) without changing, why did he needlessly complicate the example by making the larger and smaller quantities, not five and seven as we would expect, but four and twelve—the extremes of which six is the harmonic mean? He even goes to the trouble of pointing out, for no apparent reason, that six is not only more than four and less than twelve, but more than four by a half and less than twelve by a half (154c). The only purpose this would seem to serve is to make us think of six as a kind of product or "offspring" of four and twelve, as the mean that unifies them.[12] The parent-child relation is in fact the dominant leitmotiv of the dialogue. The Theaetetus contains at least six explicit references to parentage, and at least seven references to the relation for which parental procreation is a metaphor, that is, the explanation of something as a product of the intercourse of two prior elements.

The explicit references begin (1) when Socrates, after being told about Theaetetus by Theodorus, immediately asks who Theaetetus's father is (144c). (2) Later he speaks of his own mother, Phaenarete (149a)—the only time in any dialogue that he does so.[13] (3) He then goes on to compare the formulating of opinions to giving birth (151, 157c-d), and (4) subsequently refers to the deceased Protagoras's theory as an orphan (164e). In between were (5) the reference to Wonder as

[12] Cf. Rosemary Desjardins, The Rational Enterprise: Logos in Plato's Theaetetus (Albany: SUNY Press, 1990) 186.

[13] Thomas Chance reminds me that Socrates mentions his father, Sophroniscus, at Euthydemus 297e.


the father of philosophy (155d), and (6) the forthcoming discussion of perception as the "twin offspring" of objective and subjective becoming (136a-157c).

The same phenomenon—the explanation of an existent as the product of two progenitors—is operative without the parentage metaphor in (1) Socrates' definition of day as the mixture of earth and water (147c), which is to serve as a model for Theaetetus in his search for a definition of knowledge. (2) In response to Socrates' example of clay, Theaetetus recounts his and his young friend Socrates' idea of classifying all numbers into those that are the product of two equal roots (squares) and those that are the product of two unequal roots (oblongs). (3) Socrates himself is presented as a mixture of Theaetetus's looks (143e) and young Socrates' name (147d; cf. Statesman 257d). (4) In the present passage we have seen that relations like bigger and smaller can be explained only as the product of two referents, not as the property of one. (5)The analysis of syllables at 203 shows that they are normally the product of mixing vowels and consonants. (6) At 209d Socrates refers to the skytalê "a staff about which a strip of leather was rolled, on which dispatches were so written that when unrolled they were illegible until rolled again upon another staff of the same size and shape" (Fowler). It too is therefore a model of intelligibility based on the intercourse between two elements. (7) The dialogue as a whole, that is, the account of the conversation between Socrates and Theaetetus, is a product of the joint efforts of Eucleides and Socrates (143a).

The significance of all this emphasis on parentage will be considered later on. At this point, after the implicit demonstration that relations must be a product of (at least) two terms, and the reference to Wonder and Iris as father and child, we are given a biparental model of sense perception. If none of the uninitiated is listening—by whom Socrates means coarse materialists who deny the existence of anything nonperceptible, including change—Socrates will introduce Theaetetus to the mysteries of much cleverer people. These are evidently the Heracleiteans. No criticism is offered of this doctrine, and the presumption seems to be that it is a view that Plato accepts,[14] but Socrates is noncommittal when Theaetetus tries to find out whether he subscribes to this theory (157c).

[14] Not all scholars would agree. See, for example, Terence Irwin, "Plato's Heracleiteanism" (Philosophical Quarterly 27 [1977] 1-13); David Bostock, Plato's Theaetetus (Oxford: Clarendon Press, 1988) 153. The question remains continuously in view during Burnyeat's discussion of the first part of the dialogue (TP 7-65).


Within Heracleitean flux changing things may be described as gradual processes or "slow motions," some of which are capable of acting upon or being acted upon by others, in such a way that perception results (156a). Perception is accordingly like the offspring of two parents. The progeny is always twins. When the slow motion that is a gradually changing object comes within range of the slow motion that is a gradually changing eye, they produce the twins, perception and the perceived thing—for example, the perception of whiteness and the representation of a white object (156d). These progeny of the slow motions—that is, of the gradual motions of changing things—are quicker because they move from place to place: from their mutual birthplace between the eye and the object, the perception moves to the eye, and the perceived object moves to its perceived location. Accordingly, "nothing is one , itself by itself, but it always comes to be for someone" (157a). Any perceived thing is only phenomenal, in something like Kant's sense: it is the product of the intercourse between a thing in itself and our organs of perception. The physical world is therefore only a construction; the world in itself is pure flux or motion. The world of discrete and self-identical things is "objective" only in the sense that the phenomenal world is so for Kant: it is the world that is "given" in normal experience. The doctrine applies not only to individual things but also to "universals." "It is necessary to speak in this way both with regard to individuals and about multitudes collected together [

]. It is to such collections that they apply the terms 'human' and 'stone,' and every animal and form" (157b-c). A little later Socrates includes "good" and "beautiful" (157d).

One of the reasons that Plato is usually held to subscribe to the flux theory of perception is that it fits in with his view of the physical world as "becoming" rather than "being." But all the evidence of the previous dialogues indicates that this further extension of the theory, by which "multitudes collected together," or universals, are relativized in the same way as sensibles, is one to which he does not subscribe. On the Heracleitean hypothesis, however—which is being explored here—the natural interpretation of universals is that they are artificial constructs abstracted (not "recollected") from particular experiences.

It might seem, Socrates points out, that we can dispute this doctrine by pointing to the fact that in dreams, madness, and other illnesses, perceptions of reality contradict those of normal, waking perceivers, and are objectively false (157e-158a). If we recognize that some opinions are false, then we must be able to recognize a standard of correct-


ness, in which case, pace Protagoras, not everyone is the measure of truth. This objection proves to be without substance, for according to the theory our judgments do not have the same referents as those of anyone else—sick people and dreamers included—and therefore do not contradict one another and cannot be considered false. If wine that everyone judges to be sweet is judged by me to be bitter because I am sick, there is no contradiction. When I say "this wine" I am referring not to the "wine in itself" but to one of the twin offsprings of both the wine and my organs of perception. This offspring is numerically different from the offspring that anyone else intends by the phrase "this wine," and that is the partial offspring of their organs of perception. The doctrine is thus compatible with the principle of noncontradiction[15] and not falsifiable on any obvious grounds. Accordingly, Socrates proceeds to explore more subtle problems to which the hypothesis leads.

3. Perception and Understanding (160e-168c)

Socrates' immediate concern will not be the theoretical model of perception underlying Theaetetus's definition of knowledge as equivalent to perception, but the consequences of that equivalence. If the hypothetical equivalence is discredited through its consequences, then the model from which it follows cannot be accepted in its entirety either. Socrates interprets the equivalence of knowledge and perception as a denial of the possibility that knowledge can be falsified. When Protagoras says that a person is the measure of all things, this means that there is nothing outside our individual perceptions by which they might be rendered false. Socrates launches an initial attack of this conception in four stages.

1 (161b-163a). Protagoras might as well have said that the measure of all things is not a person but a pig or baboon. Then he could laugh at us for thinking him as wise as a god when in fact he is no wiser than a tadpole. In that case it would make no sense for anyone to pay to be his student, or to practice Socratic midwifery or dialectic, since truth is already to be found in mere perception (161c-e).

Neither Theodorus nor Theaetetus can find anything wrong with this

[15] "What is in every way different from something else cannot in any way have the same capacity as the other" (158e).


refutation, but Socrates points out that it is an example of demagoguery (162d), and that Protagoras would accuse them of accepting appeals to mere likelihood (162e). We are not to be deterred by the seeming absurdity of saying that a pig or baboon is the measure of all things. On the basis of the foregoing theory of perception, which asserted that the object of perception is always relative to the perceiver, it is plausible and even necessary to conclude that pigs and baboons are the measures of all things (i.e., all that they perceive).

But a residue of Socrates' objection survives this reply. His second point still stands. On Protagoras's account it makes no sense to consider one person to be wiser than another or for one person to presume to teach or criticize another. In fact, however, Protagoras charged for teaching, and Socrates' midwifery and dialectics were considered valuable by his students. Protagoras's claims are not invalidated by this point only because he is really talking about a different level of knowledge. The objection speaks of interpretive knowledge—understanding— rather than perceptual information. But although this implicit distinction does not refute Protagoras at present, it will eventually become a wedge with which to dislodge Protagorean relativism. Accordingly, while the argument seems at first inconsequential, on closer inspection it implies a distinction between two levels of knowledge, a distinction that will turn out to be important. The implications of the next three objections will progressively specify what is involved in the implicit distinction between perception and understanding, and they will do so in terms of concepts that seem to recall the doctrine of recollection: recognition, memory, and intelligibility.

2 (163a-c) . What about hearing a foreign language, Socrates asks, or even seeing written words in our own language when we cannot read? How can it be maintained that perception is knowledge when we perceive these sounds and symbols but do not understand them? Theaetetus replies:

We shall say that we know about them just what we see and hear. In the one case we both see and know the shape and color, and in the other case we hear and at the same time know the higher and lower sounds. However, those things that the grammarian and the interpreter teach about them, we neither perceive by sight or hearing, nor know.

Socrates praises this answer but adds, "I had better not disagree with you about this, so that you will grow."


Coming after the last objection, the basis on which Socrates might have disagreed is not hard to discern. Once again the two levels of knowledge are visible, sensory information and understanding, but here the latter is made evident rather than merely implicit. At the first level— sounds and shapes—everyone's knowledge is coextensive with the information supplied by their senses, and none is any better than any other. But at the second level it is undeniable that some people (especially grammarians) recognize the meaning of these phenomena better than other people, and this second kind of knowledge is not coextensive with sense perception. Once again a portion of Socrates' objection remains untouched by the reply. There is an interpretive, recognitive, as well as a sensory kind of knowledge, and the former is not reducible to the latter.

3 (163c-165b) . On the hypothesis that knowledge is perception, if we see something we must know it. But if we dose our eyes, then, even if we still remember the object, we must be said no longer to know it. This would be a "monstrous" conclusion (163c-164b).

No reply is made to this objection, but Socrates remarks that their conclusion was derived from a contentious rather than philosophical style of argument (164c) and that if Protagoras were here he would have much to say in reply (164e). The fallacy of the argument may be expressed as a collapsing of the distinction between memory of knowledge and memory as knowledge. I can look at a book and say that I perceive and therefore know that a book is on my desk. I can then dose my eyes and say that I remember perceiving and knowing that a book was on my desk. But it does not follow that I still know that there is a book on my desk. There is no contradiction or monstrous conclusion. Nevertheless memory is a kind of knowledge, although of a different order than perception. And as Aristotle mentions at the beginning of the Metaphysics (A.1.980b 29-981a 1), a plurality of memories constitutes experience, which is an important kind of knowing different from sense perception. Memory is in fact a necessary condition for interpretive knowledge, and thus a precondition for the distinction implied by the previous examples.

4 (165b-e) . The final objection in this series, introduced indeed as "the most formidable," is that if we look at something with one eye dosed, then we both see and do not see, and accordingly both know and do not know the object. This is said to reduce Theaetetus's hypothe-


sis, that perceiving is knowing, to absurdity (165d). Theaetetus is dearly not persuaded by the argument, but lacks the weapons with which to fight it. With some encouragement he might have hit upon the distinctions made in the definition of contradiction in the Republic ,[16] which stipulates that a genuine contradiction must refer to the same time, the same part of the subject, and the same object (in fact, the converse of that definition appeared above at 158e: see n. 15). In the present case we are speaking of different parts (eyes) of the subject, and so there is no contradiction. Socrates, however, does not give Theaetetus encouragement, but the reverse.

What is the point of this "most formidable" but most transparently fallacious argument? Might the two different eyes, one open and one dosed, be meant metaphorically? The previous three objections have reminded us of the difference between perceptual and interpretive knowledge, and the difference between perception and the precondition for interpretive knowledge, memory. In other dialogues, especially the Meno —the dialogue most often alluded to in the Theaetetus —memory was used as a metaphor for a latent component of knowledge, furnished not by the senses but by the mind itself, the analogue of Aristotle's "active intellect." Activated by perception, this latency may be "recollected," making possible judgments of attribution ("this is beautiful") and understanding ("justice is the harmony of the tripartite soul"). In the dialogues after the Meno , recollection is pictured as an intellectual "seeing" of the forms.[17] And shortly hereafter in the Theaetetus we will be told that there are two kinds of seeing and two kinds of failure to see.[18] The philosopher sees what lies "above" although he may be blind to what lies at his feet or in front of his eyes (174a-c), while others see what is at their feet and before their eyes but cannot see the whole nor what is "above" them (174e-175d, 176e). The "higher" realm of the philosopher is that of divinity and goodness, while the other is that of the mortal and evil (176a).

Are these two kinds of seeing prefigured in the argument about the open and closed eye, an argument that is announced as "the most formidable" but that is a joke if taken literally? Socrates goes on to say that similar problems would arise if someone were to ask whether we can know the same thing sharply and dimly, close by but not at a dis-

[16] Republic 4.436b ff.

[17] E.g., Phaedrus 247c ff., Republic 7.516a-b, Parmenides 132a.

[18] Cf. Republic 7.517d-518b; also Aristotle's distinction between what is most dear to us and what is most dear in itself (Metaphysicsa .1.993 9-11, Z.3.1029 3-12).


tance, intensely and quietly (165d). According to the doctrine of recollection, one might say that sensibles are perceived sharply, dose by, and intensely, while intelligibles are perceived dimly, at a distance, and quietly. Perhaps there is some prefiguration of this fundamental distinction in the opening words of the dialogue: "Just now, Terpsion, or long ago . . . ?" In a dialogue devoted to discovering the sources of knowledge, it would not be beyond Plato's dramaturgy to give these words a double meaning. On the literal level they ask when Terpsion arrived from the country, but they are also appropriate to the fundamental epistemological alternatives of empiricism and rationalism: on the empirical model knowledge begins just now when we perceive something; on the rational model what we perceive now calls into play something acquired long ago. We need not try to decide whether Plato intended these connections or not. The only important question will be whether the doctrine of the direct apprehension of forms may in fact be brought usefully to bear on the problems of the Theaetetus . That question, which is much debated in the literature, will be answered affirmatively in the course of this study, and it may be that the present passage is meant to anticipate that answer.

The distinction between perception and understanding becomes all but explicit in what follows. Socrates, speaking in the persona of Protagoras, defends Protagoras's theory against the preceding refutations by means of a distinction between knowledge and wisdom (a distinction that Theaetetus had collapsed at the beginning of the dialogue). He reaffirms that each of us is the measure of all things because "what is" cannot mean anything other than what appears to a perceiver. But although in this sense everyone is equally knowledgeable, wisdom may be distinguished from such knowledge as the ability, "when bad things appear and are for someone, to implement a change and make good things appear and be to him" (166d). Thus, understanding (wisdom) exists in addition to perceptual knowledge, but it is of a pragmatic rather than factual nature. It does not tell us what exists but only what is desirable and how to achieve it. It is in this sense that doctors, educators, and sophists are wiser than ordinary people. They replace the worse with the better, but not the false with the true. There is no such thing as falsity, "because it is impossible to think [

] what is not" (167a).

"Protagoras" doses with a Socratic appeal for fairness and seriousness in argument so that "your partners will blame themselves for their confusion and aporia, rather than you, and they will follow and love


you, and hate and flee from themselves to philosophy in order that, by becoming different, they may be liberated from their former selves" (168a). The sentiment is obviously Socratic rather than Protagorean, and points up the difference between them, which will soon be elaborated in Socrates' digression. For Protagoras wisdom means the ability to eliminate unpleasant perceptions in favor of pleasant ones; for Socrates it means overcoming one kind of life in favor of another.[19]

4. Understanding and Value: Beginning (168c-172b)

The first set of four objections was aimed at the Protagorean definition of knowledge as perception. Insofar as the arguments succeeded in forcing a distinction between perceptual and interpretive knowledge (understanding), that definition has been seriously compromised. The next set, also of four objections, will focus on interpretive knowledge alone. Here Socrates finally succeeds in pressuring Theodorus to replace Theaetetus as his partner.

1 (170a-171d) . The first refutation of this series is the famous palintrope or self-refutation argument: "Shall we say that people always believe[20] truly, or sometimes truly and sometimes falsely? In both cases it follows that they don't always believe what is true but both [what is true and what is false]" (170c). The conclusion obviously follows from the second alternative, of which it is a restatement. The subsequent argument is designed to show that the conclusion must follow from the first alternative as well.

The argument may initially be simplified as follows. The minor premise is that people generally disagree with Protagoras's claim that each person is the only judge of what is true for him. They think that different people have different degrees of wisdom about different things, and that wisdom is true thought and that ignorance is false opinion (170c; cf. 170a-b). The major premise is that Protagoras claims that what people believe is true (171c). The conclusion follows, that Protagoras must concede the general opinion to be true, that not everything we

[19] This radical conception is already present in several of the early dialogues. Thus Martha Nussbaum points out that in the Protagoras , "Socrates offers us, in the guise of empirical description, a radical proposal for the transformation of our lives" (The Fragility of Goodness [Cambridge: Cambridge University Press, 1986] 117).


believe is true. Since this contradicts his own position the latter must be false.

The actual course of the argument is more complicated because of Protagoras's insistence that truth is always relative to some believer. An opinion is not true simply, but true for someone. Accordingly, the way the argument puts it is that Protagoras's theory may be true for him but false for tens of thousands of others (170e). Moreover, if the theory were right, then if no one believed it, it would ex hypothesi be false for everyone and therefore false. And if no one believed it but Protagoras, then:

First, by as much as those who believe it outnumber those who do not, it is that much more not true than true. . .. Second comes a most elegant point: he accepts that the tenet of those who believe in opposition to him about his own tenet—in that they believe it is false—must somehow be true, since he agrees that what anyone believes really is.

The validity of this argument has been much debated.[21] It is sometimes felt that the reasoning depends on an illicit transition from "true for someone" to "true" simply: Protagoras would accept that his theory is not true for most people, but it would still be true for him and no contradiction would arise. Such a defense, although technically valid, would be disingenuous. Protagoras wants to persuade us that his theory is true for everyone, otherwise his arguing for it, publishing it, and teaching it would be inexplicable. It would be damaging for Protagoras to be forced to admit that his theory is true only for himself (and perhaps a few others), but false for everyone else. Moreover, having admitted that, it would be difficult for him to deny that the theory is false in general .[22]

What this argument demands of Protagoras is that he acknowledge that at the level of interpretation or understanding not all judgments are equally valid. He was willing to acknowledge that at this level we can distinguish opinions that are pragmatically superior from those that are pragmatically inferior, but not opinions that are true from those that

[21] Cf. Sayre, PAM 88-92; Myles Burnyeat, "Protagoras and Self-Refutation in Plato's Theaetetus (Philosophical Review 85 [1976] 172-95); Jay Newman, "The Recoil Argument" (Apeiron 16 [1982] 47-52); and Bostock 92-95.

[22] If Protagoras were to rise from the ground up to his neck, Socrates says, he would accuse me of talking nonsense before sinking back and running off (171d). This is sometimes taken to be an admission by Plato that the argument is flawed, but, as Burnyeat points out ("Protagoras" 191), the fact that Protagoras would run away after repudiating the conclusion suggests that his reasons would not be good ones.


are false. The present argument makes the point that, on the contrary, Protagoras does regard his interpretations as truer than those of non-relativists, and that unless he acknowledges that his perceptual relativism ceases to be relativistic at the level of interpretation or theory, he cannot help but undermine his entire position. He must concede that no one "is the measure of any single thing that he does not understand [

]" (171c). Perception may be relativized, but understanding may not.

2 (171d-172b) . The next refutation is interrupted by Socrates' digression. Socrates begins the argument by recapitulating the claim made earlier in Protagoras's defense: although sensible qualities are just as they appear to each of us, one person may be wiser (i.e., "more effective") than another in pragmatic pursuits such as medicine (171e). The same dichotomy now appears in the larger context of the state. According to the theory, values are relative to the state, as sensa are to the individual:

In political affairs, with regard to what is noble and shameful, just and unjust, pious and not: however each state legislates these in accordance with its opinions, that is how they in truth are for it. And in these matters no one is wiser—neither one individual than another, nor one state than another. . .. None of these has by nature an essence [

] of its own, but rather the common opinion becomes true when it is believed, and for as long a time as it is believed.
(172a, b)

But here, too, Socrates replies, when it comes to what is advantageous or disadvantageous to the state, Protagoras would not deny that one adviser differs from another, and one state from another, with respect to truth. He would not dare to say that whatever a state believes to be in its advantage necessarily is so (172a-b). The escape from relativism, made possible by the distinguishing of understanding from perception, extends to value as well. Not everyone understands equally well what is good, and if in many cases there is no way of adjudicating disputes about what is beneficial, there are at least some occasions, especially in politics, when someone may proven right or wrong.

5. Socrates' Digression (172c-177b)

It is significant that the digression begins just at the point where values are ascribed to convention rather than nature, for one of the functions of the digression is to repudiate the claim that justice and piety are arbitrary values without essence in nature. Rather, they are precisely the


natural essences that the philosopher strives to know (175c, 176b). It is acquisition of this type of knowledge that requires the courage spoken of in the dialogue's opening passages. Knowledge of this kind would be different both from the perceptual and interpretive knowledge of the corporeal world that were distinguished above. The latter two correspond to the lowest levels of the Divided Line, eikasia and pistis. Eikasia , as portrayed in the Cave, is the uncritical awareness and memory of passing perceptions, and pistis , which is by contrast the highest awareness of the corporeal world, is therefore our interpretation of the former experiences.[23] The kind of knowledge referred to in the digression, on the other hand, would correspond to the Republic's category of noêsis . The remaining kind of knowledge (according to the Divided Line), dianoia —the drawing-out of the implications of our initial postulates— has been illustrated throughout the dialogue by the deductive aspect of the method of hypothesis, and will be illustrated more generally in the "aviary" modal of knowledge.

The digression is reminiscent of the middle books of the Republic (especially the Divided Line and the Allegory of the Cave) because of its placement as well as because of its content: like the central books of the Republic , it occurs in the very center of the dialogue and breaks into the beginning of the discussion of a political question, which is subsequently resumed as if the digression had never taken place.[24] The substance of the digression, again like the middle books of the Republic , deals with the difference between a life devoted to corporeal, mortal values, and one devoted to intelligible, divine values. It is a sustained comparison of the life of the philosopher with the life of those who devote themselves to law and politics. Where the denizens of the Cave were described in the Republic as prisoners, here politicians and orators are described as slaves, inasmuch as their success depends on their adherence to arbitrary rules and on the approval of their audience (172d-172e). As in the Republic , the philosopher who enters their world will appear to be ignorant and unable to see what is in front of him (174b), but "when he drags the other upward . . . then the situation is reversed" (175c, d) and the political person will then betray his own inability to see (174e-175a, 176e). There are also important differences between this account and that of the Republic , especially that the pres-

[24] Cp. Republic 449a-b with 543c-544a.


ent account focuses on political life in particular, rather than on the ordinary generality of people depicted in the Cave. In that respect the digression is a contrast between knowledge pursued for its own sake and knowledge pursued for the sake of social rewards and honors (a theme that was present in the Cave but in a subordinate way: 516c-d), that is, a contrast between reason as an end in itself and reason as a servant of spiritedness.[25] But here, as in the Republic , the fundamental contrast is between the corporeal world and the intelligible world. For the philosopher, "in reality only his body occupies and dwells in the city, while his mind, considering all such things as of little or no importance, is contemptuous of them" (173e).

The contrast is drawn to such an extreme degree that even Socrates may not qualify as a philosopher, for the pure type represented here has no knowledge of where the marketplace is, or the law courts, nor does he know anything about parties with flute girls. In fact—in further contrast to Socrates—he is so oblivious to these things, and to himself (174b), that "all these things he doesn't even know that he doesn't know" (173e). Ordinary people think they know things when they do not; Socrates has risen so far above their vanity that he knows that he is ignorant; but the true philosopher has risen so far above even this self-consciousness that he no longer knows even that he is ignorant. Socrates does not fit the description of the philosopher here, for it is a conception, like that in the Phaedo , or in the Republic's Islands of the Blessed (519c), of the most extreme transcendence of the corporeal realm imaginable, in favor of the intelligible. The transcendence is not for the sake of abstract intellectual knowledge, but for the sake of goodness. As in the Phaedo , the corporeal world of mortals is the abode of evil, while among the gods exists goodness alone (176a). The goal of life, therefore, is to become as godlike as possible, which means "to become just and pious together with wisdom" (176b). For,

Two patterns, my friend, are set up in reality, one divine and most blessed, the other godless and most miserable. Unjust people do not see that this holds true, and because of their foolishness and complete lack of understanding they are unaware that they become more like one of them, due to their unjust behavior, and less like the other. For this they pay the penalty of living the life that is an image of what they resemble.

[25] According to the Republic , this is one of the two fundamental species of injustice in particular and vice in general, the other being the subservience of reason m appetite (442b-443e).


In view of Plato's continuing interest in politics and social justice, however, we may assume that the reason why Socrates' own character fails to conform to the conception of the philosopher here is not because he has not yet reached that stage, but rather because, like the philosopher who returns to the Cave from the Islands of the Blessed, his experience of transcendence enables him to return to the corporeal world, in a transformed way.

6. Understanding and Value: Conclusion (177c-187a)

2 (continued, 177c-179d) . After the digression Socrates reiterates the beginning of the previous argument: people like Protagoras may claim that justice is only a matter of what is legislated by the state, but no one would say that whatever a state thinks is good or "advantageous to itself really is so" (177d). Whether it is so or not can be determined only in the future, and Protagoras can hardly maintain that each of us is the measure of what is going to happen. Rather, the ability to make predictions is what sets experts apart from ordinary people in such matters as medicine, food, and music, and what sets Protagoras apart in matters of law (178a-e). It follows that some of us are wiser than others, and that it is they who are the measure, not ordinary people (179b).[26] "Protagoras" had already agreed that some people are wiser than others in that they are able to replace worse sensations with better ones, but he denied that this had anything to do with truth or falsity (166d). Socrates here counters that denial by pointing out that the ability to make such replacements successfully is the ability to predict what will happen, and predictions are indeed qualifiable as true or false. At this point Socrates makes fully explicit the difference between the (infallible) perceptual and (fallible) interpretive levels of knowledge that has been implicit throughout the earlier discussions:

There are also many other ways to establish that not every opinion of everyone is true. However, with regard to the passing impressions from which our sense perceptions and the corresponding opinions come to be, it is harder to


confirm that they are not true. . .. It may be that they are unassailable, and that those who say they are fully dear and instances of knowledge are perhaps saying what is really so.

I have been referring to understanding as "interpretive knowledge," but it is not dear what makes such interpretation and understanding possible. According to the flux theory, even such concepts as human, stone, and "every animal and form," including the good and the beautiful, arise in the same way as perceptions (157b-d). In that case the concepts, by which we understand our perceptions, must be said to arise out of those perceptions themselves. Those who have superior understanding of goodness can be said only to have extracted or generalized instances of them more effectively.

There are two problems with this account. First, if nonrelativistic knowledge is impossible at the perceptual level, how can relativity be overcome at the level of understanding merely by abstraction? Second, there would be a circularity not unlike that of Meno's paradox (which is often recalled in this dialogue): in order to extract concepts from the flux of experience we must be in possession of the interpretive principles that already presuppose those concepts. The Platonic answer to both these problems has been that interpretive knowledge has its source not only in the senses' relation to the corporeal world, but also in reason's relation to the intelligible world. In the remaining two arguments of this section Socrates will reaffirm this answer both negatively and positively: negatively, by showing that if we try to account for knowledge only in terms of the flux model, we will be reduced to silence (like Cratylus); positively, by showing that at least some of the concepts by which we interpret experience have their source in something other than sensory experience.

3 (180c-184a) . If everything is in flux, then all that exists is the transitory impressions of the perceiver (152d)—this was the point of the theory of perception developed earlier (156a-157c) and repeated here (182a-b). There can be no other knowledge than this. The flux theory is that everything is in motion, not only in the sense of movement in space, but also in the sense of alteration (181c-e). It follows that the sensuous qualities that we perceive are changing at the very moment we perceive them, and the act of perception itself is always changing into nonperception. Moreover, since perception is knowledge, knowledge too ultimately collapses into an identity with nonknowledge: each is


continuously changing into the other (182c-e). Consequently, "if everything is in motion, every answer about anything one is asked will be equally right," and language itself will break down (183a-b).

Parmenides had told Socrates that

if anyone does not admit the existence of forms of things, or mark off a form under which each individual thing is classed, he will not have anything on which to fix his thoughts, as long as he does not admit that the Idea of each thing is always the same, and in this way he will utterly destroy the power of discourse.
     (Parmenides 135b-c)

Is the present argument meant to remind us of this warning, and thereby of the theory of forms? (Socrates' remark that "I met him when I was quite young and he quite old" [183e] seems clearly meant to remind us of that dialogue.) If so, it would explain the puzzling fact that Parmenides is mentioned immediately before and after the present passage but to no obvious purpose. Beforehand Socrates says that he "nearly forgot that others declare the opposite" of the flux theory. These others are Melissus and Parmenides, whose views Socrates proposes to examine after they examine the proponents of flux (180d-181a). Afterwards, Theaetetus reminds Socrates of this next task, but Socrates declines to pursue it, on the grounds that they could not do justice to Parmenides' views except at great length (183c-184a).

4 (184b-187a) . If this was an indirect reminder of the theory of forms, the next section is a direct reminder of it. Socrates raises the question whether there is some one form within us (which we might call the soul) with which we perceive together whatever each of the senses perceives only separately—something that perceives sounds and sights, and the like, each of which alone is proper to a specific sense (l84d-e). The test is whether there is anything we can think about that involves more than one sense. If there is, this common factor cannot be reduced to what the individual senses give us and must somehow be provided by or through the mind or soul itself.

In fact there are several such common qualities. We can think that the objects of seeing and hearing both are , and that each is different from the other and the same as itself (185a). Socrates thus generates three of the five "greatest kinds" of the Sophist , existence, sameness, and difference (the other two, motion and rest, are already evident in the Heracleitean model and its rejection by the Eleatics). Socrates further establishes that the mind will discern that each of these objects is


one , and both together are two , and that it can also ask whether they are similar or dissimilar (185b). To these qualities Theaetetus adds odd and even (185d), two of the traditional Platonic forms. And Socrates, remarking that Theaetetus has shown himself to be not ugly, as Theodorus had claimed, but beautiful and good, proceeds to add to the list the forms beautiful and ugly , and good and bad (186a), which had earlier been assimilated to the flux model of perception (157d). What all these qualities have in common is that the soul somehow perceives them through itself rather than through the body's sensory faculties (185e).

Now, conceptions about being (

) and value (
) can be attained, if at all, only through a long and difficult education, and truth and knowledge are inaccessible unless we can discern being. Accordingly, knowledge can be found only through the qualities that the soul finds by itself, rather than those that it received from the bodily senses. Knowledge cannot therefore be the same as sense perception (186c-e).

This argument is put to the epistemological purpose of refuting Theaetetus's claim that knowledge is perception. But it also has consequences for the ontological foundation on which, according to Socrates, Theaetetus's position rests—that all is flux. Clearly, the intended inference is that all these forms, which are not themselves in flux, are real and imply some kind of stability within Heracleitean flux.

Are these qualities in fact the Platonic forms?[27] Like the forms of the Republic they are apprehended only as the result of a long and difficult education,[28] but whether they may be regarded as "separate" forms or not cannot be answered on the strength of this passage. At the very least they correspond to the forms' aspect as "universals," although even this is not entirely explicit. Unlike the characterization of forms in Republic 10, they are not said to be posited for "each multiplicity to which we give the same name" (596a). Instead, we have a plurality of senses (sight, hearing, etc.) to which we can apply the same intepretive categories. But it comes to the same thing. To say that we both see and hear

[27] Cf. Bostock 98-99: "He does make it quite dear in this digression that he has not stopped believing in the forms, for the philosopher's wider vision is explicitly credited to his concern with the forms of justice, injustice, happiness, and so on (175c, and possibly 176e). The same conclusion is also dear from later passages in our dialogue. But what is not very dear is how exactly Plato now conceives of the forms, or indeed whether he has a definite view on this question at all . . . . All that one can say with complete confidence is that they are still regarded as imperceptible entities (185c-e, 195d-196a)."

[28] In the Republic six stages are specified: (1) arithmetic, 525a-b; (2) plane geometry, 526c; (3) solid geometry, 528b; (4) astronomy, 528c; (5) harmony, 530d; (6) dialectics, 531d. Theaetetus has studied the first five of these with Theodorus (145c-d).


something beautiful is to say that "beautiful" is not a unique name, but rather one that can be applied to a plurality of sensory experiences; that is, it is a universal. The present passage affirms the need for universals, and further affirms that these universals are not reducible to sensory information, "but rather the soul, itself by itself, discerns what is common to all" (185d-e).

Let us consider what is implied ontologically by each example, and to what extent each may indeed be considered necessarily given in our experience.

"Being" is the first example because it has the greatest generality, but for that same reason it indicates nothing more than the bare need for such concepts.[29] The example of "different" and "same" does, however, imply a differentiation within reality, and is moreover a contrast that is almost universally employed: even for modern neo-Heracleiteans like Nietzsche and Derrida there are identifiable selves as well as differentiation (however uncompletable). Cratylus might perhaps reject such a concept, in view of his insistence that we cannot step into the same river even once. Yet insofar as he continued to point his finger he continued to differentiate, and to imply the relative integrity of what he indicated.

The examples of "one" and "two" suggest that we are intrinsically capable of distinguishing unity from multiplicity, and thus that we are by nature capable of discriminating between part and whole. This follows from the previous dichotomy, for, as Socrates shows, if we can distinguish one thing from another, we can distinguish between the unity of the pair and the particularity of its members. If Cratylus, pointing, distinguishes one thing from another, he distinguishes them also as parts within a complex.

The next pair, the similar and dissimilar, has more radical implications. If we discern similarity and dissimilarity, then we discern common features; and if there are common features, then something like the existence of forms is indicated. It is not necessary to assign an ontological status to these forms, as Plato or Aristotle did—they may so far be interpreted merely epistemologically, as in Descartes and Kant, or linguistically as in contemporary thought—but there is no reason in the present passage to suppose that Plato's Views are any different here from those expressed in the Phaedo and Republic . In any case, we may at least grant to Plato that with our mind's eye we see resemblances,

[29] For a different view, see Jason Xenakis, "Essence, Being and Fact in Plato: An Analysis of One of Theaetetus 'Koina'" (Kant-Studien 49 [1957-58] 167-81).


common properties, universals, in the world. There is then at least a prima facie case for an underlying ontological structure.

With the concepts of odd and even, Theaetetus explicitly introduces mathematics, which was only implicit in the earlier example of one and two. One of Plato's favorite arguments for the ontological significance of a priori forms of knowledge is the efficacy of mathematics. If the principles of mathematics are known to us by nature, as the Meno argues, and if they also turn out to be the principles by which external nature operates, as astronomy and Pythagorean science suggest, then we are by nature attuned to understanding the structure of reality. This would suggest not only that reality is comprehensible for practical purposes, but also that it is inherently rational, and even that it has value, since the rational is the good.

The rational, value-laden aspect of the forms is made explicit by the last two examples, the beautiful and the good, and with them the entire theory of forms is present by implication. This is clearly the direction to which Plato points us in the attempt to answer the extreme Heracleiteans.

7. The Parentage of Knowledge

If the Theaetetus is hinting that the Heracleitean problematic can be answered by means of the theory of forms, such an answer seems open to the following objection. In sense perception, according to the preceding model, the thing qua perceived is not the same as the thing in itself (the affective slow motion). The offspring is not the same as the parent. If the same model is to be used now for knowledge, it would follow that the act of knowing also requires two parents—the perceived object and our interpretive concepts—and that it too produces twins, knowledge and the thing qua known. The latter is not identical either with the thing qua perceived or with the thing as it is in itself. Knowledge would then be possible only in an equivocal sense. There are three levels: (1) the thing in itself, that is, the objective slow motion; (2) the perception of that thing, which is not identical with it, but is one of its twin offspring; and now (3) knowledge of that offspring, which is a new twin offspring. The thing qua known, therefore, is the grandchild of the thing in itself; and knowledge, rather than bringing us closer to reality than perception did, takes us a step farther away. In this respect Kant and Heracleitus are ultimately allies. But Plato would not disagree either. In fact such considerations form the basis of his claim that knowledge of


the empirical world is not possible. However, the forms, which are the objects Of true knowledge for Plato, cannot be radically discontinuous from the physical world, or else the doctrine of recollection would make no sense. How then would Plato be able to find formal structure in a destructured reality of pure flux—assuming that he accepts that view of the physical world?

His argument that flux cannot be the whole story is the counterpart of the Parmenides ' claim, quoted above, that unless there are forms all discourse becomes impossible. Here the argument is that if there is only flux, then the referents of words are constantly changing, and words will have no stable meanings and will be ultimately self-contradictory and incoherent (181b-183b). Even if Protagoras were to reply that words need only refer to perceptual, not noumenal, reality, Socrates would press the point that, in a world of pure flux, words cannot refer to anything individuated, whether objective or subjective. Socrates' previous remark to Theodorus (who is the respondent here) that "they are no friends of yours" (180b) suggests that mathematics too gives us reason to reject the extreme form of Heracleiteanism.

Since the subjective, relativistic aspect of the perceptual object must be passed along to its offspring, the object of empirical knowledge, such knowledge does not reach as far as physical reality itself, which is pure becoming. That is why Socrates had said that we must search for knowledge "in whatever name the soul has when, itself by itself, it is occupied with what is" (157a). In other words, if it is impossible to have knowledge of the offspring (the empirical world), or of one of its parents (the perceived quality), it is nevertheless possible in the case of the other parent, the forms. The forms must be known, no longer in their children but in themselves. Like light (one of Plato's favorite metaphors), their presence is detected initially by their illumination of empirical objects, but they can be known in themselves only if we abstract from the objects illuminated.

The importance of this way of looking at the problems of knowledge and perception explains Plato's persistent use of the parent-child theme as a dramatic leitmotiv in the early part of the dialogue—it appears in at least thirteen direct or indirect instances (see above, section 2). In one of those instances, Socrates turned out to be a mixture of Theaetetus's looks and young Socrates' name. Our looks or appearance reflect our changing nature, the slow motion of our flux. Our name, on the other hand, reflects our constant identity; and when we turn from individuals to common natures, the name reflects the eternal form. Theaetetus turns


out to be the representative of appearance both in terms of what he shares with Socrates and in terms of the way he attempts to define knowledge.[30] In fact his attempt to give an account of knowledge will fail because he sees knowledge only in terms of one of its parents, and is blind to the other.[31]

8. Five Models of Knowledge (187a-201c)

The above discussions show, Socrates says, "that we should not seek [knowledge] in perception at all, but in whatever the name is, when the soul, itself by itself, is engaged with what is real" (187a). Theaetetus's response to this is minimal: he revises his definition of knowledge to "true opinion" (

). But Socrates, in return, wonders how an opinion could ever be otherwise (187b-d). In an attempt to answer this he develops five models of knowledge, which, like the earlier refutations, may be construed as progressively more adequate hypotheses.

1 (188a-c) . The first model is the simplest, deliberately abstracting from learning and forgetting, and concentrating only on knowing and not knowing. On this model false opinion can mean only that we think that (1) something we know is either (a) something else that we know or (b) something that we do not know; or else that (2) something we do not know is either (a) something else that we do not know or (b) something that we know. All these are interpreted as judgments of identity, as if we said, (1a) "Socrates, whom I know, is Theaetetus, whom I also know," (1b) "Socrates, whom I know, is someone whom I do not know," (2a) "Socrates, whom I don't know, is Theaetetus, whom I also don't know," or (2b) "Socrates, whom I don't know, is Theaetetus, whom I know." Consequently they are dismissed as implausible accounts. Nevertheless, in cases of mistaken identity any one of these kinds of judgments might arise. In bad light or from a distance (1a) I

[30] Plato is perhaps conscious, too, of the irony that Theaetetus died of "flux," dysentery (142a). Although this may be historically factual, there was no need to mention it.

[31] Recall the digression's distinction between two kinds of seeing and two kinds of failure to see (174a-c, 175d, 176c, 176e, and perhaps 165b-c).

There have also been literal references to the parent-child relationship, in Socrates' remarks about his mother, and his inquiry about the identity of Theaetetus's father. The result is that Theaetetus is identified in terms of his father, and Socrates in terms of his mother. If this is not merely a coincidence (it is, however, the only time in the dialogues that Phaenarete is mentioned, and perhaps the only time that Socrates inquires after someone's paternity), perhaps it is a hint that there is something one-sided (but jointly complementary) about both Theaetetus's emphasis on perceptual material, and Socrates' "sterile" (as he describes himself) rational criticisms of the progeny of others.


might mistake Socrates for his younger look-alike, Theaetetus; or (1b) I might fail to recognize him and think he is someone I do not know; or, (2a) never having seen either Leukippus or Democritus, I might hear the former lecture and think he is the latter; or, (2b) if there were some resemblance between them, I might see Leukippus and think at first that it was Socrates. But this model abstracts from sense perception as well as from learning and forgetting. It is concerned only with the logical relationship between the concepts of something known and something not known. It is meant to be inadequate, as a way of forcing us into more sophisticated formulations. The paradox on which it is based is elegantly formulated by Burnyeat: "A necessary condition for mistaking X for Y is also a sufficient condition for not mistaking X for Y . The necessary condition is that one know X and Y . But this, it is claimed, is a sufficient condition for knowing that X is not Y ."[32]

2 (188c-189b) . The second model is ontological rather than epistemological. It substitutes "being" for "knowing," so that to have a false opinion means to believe "what is not" about something (188d). Here the "is" of judging is interpreted as existential rather than identificatory, but the revision does not resolve the difficulty. Earlier Protagoras had insisted that there is no such thing as falsity "because it is impossible to think what is not" (167a). And here too Socrates concludes that "thinking what is not" means "thinking nothing," which means "not thinking" (189a). Once again important distinctions have been suppressed, in particular the distinction between the two senses of "not being" that we will encounter in the Sophist : "nonexistence" and "difference." Presumably this is done to enable us to perceive the inadequacy of the most simplistic models of explanation, and thus make us better able to appreciate the need for the progressively increasing complexity that will follow.

3 (189b-190e) . In the next model Socrates combines the first two. Now false opinion is "other-believing" (

), which means that we "always have an opinion about something that is, but of one thing instead of another" (189b-c). The first clause of that description is existential ("opinion about something that is") like the previous model, the second identificatory ("one thing instead of another") like the first one. Theaetetus approves of this model, "for when someone thinks beautiful instead of ugly, or ugly instead of beautiful, then, most truly, his opinion is false" (159c). After rebuking him for the oxymoronic

[32] TP 77.


phrase "truly his opinion is false,"[33] Socrates demurs, saying that we would never say that "the beautiful is ugly" or "the unjust is just" or "the odd is even" (190b).

Socrates has perverted Theaetetus's meaning. Clearly, Theaetetus was thinking of predication: my opinion is false if I believe that a beautiful thing is ugly (something unfamiliar may seem ugly to me at first, but beautiful on further acquaintance). But Socrates misrepresents the judgment as one of identity, as in the first modal. Plato gives with one hand and takes back with the other. He has Theaetetus remind us that the function of judgment may be predication (which would go a long way toward solving the present aporia), but he then has Socrates suppress the concept.[34] In fact this is the closest that the Theaetetus ever comes to exploring predication, even though it is dear from other dialogues (both those considered to be earlier and those considered to be later) that that is where the models for true and false opinion must be sought.

4 The Wax Block (191c-196d) . At this point learning and memory are added to the model (191c-d), after having been expressly excluded since the beginning. Learning is compared to the impression made by a shape in a block of wax, and memory is the retention of that shape. Socrates now goes through an odd, selectively exhaustive, list of types of judgment, in order to discover cases where false opinion is possible

[34] Since other dialogues explain predication in terms of the participation of things in forms, the present passage may be intended as an indirect reminder of the theory of forms, in keeping with Cornford's suggestion that "the Forms are excluded [from the Theaetetus ] in order that we may see how we can get on without them; and the negative conclusion of the whole discussion means that, as Plato had taught ever since the discovery of the Forms, without them there is no knowledge at all" (PTK 28). We can connect this interpretation with "Parmenides'" remark that "in the case of each hypothesis not only must you examine what follows if what is hypothesized exists, but also if it does not exist" (135e-136a). The Theaetetus may be taken as such an exercise with respect to the theory of forms. Cornford interprets the Theaetetus's many allusions to earlier dialogues, in which the theories of forms and recollection were presented, as hints that those doctrines should be brought to bear on the present discussions. McDowell, on the contrary, thinks that such allusions "can be read as an implicit criticism of the Theory of Forms and the Theory of Recollection" (p. 219), and that "it is hard to see how Plato could have supposed, as Cornford's thesis would imply, that a restatement of the Theory of Forms would solve all these problems at a stroke" (p. 258)—a view that he shares with Glenn Morrow ("Plato and the Mathematicians: An Interpretation of Socrates' Dream in the Theaetetus [201e-206c]," Philosophical Review , 79 [1970] 309-33 at 312) and Bostock 241-43. Others, such as Sayre (PAM 58 n. 2, 135) and jürgen Sprute ("Über den Erkenntnisbegriff in Platons Theaitet," Phronesis 13 [1968] 47-67, esp. 52, 67), are closer to Cornford's position. I shall argue that the theories of forms and recollection can, in facts largely overcome the aporiae of the Theaetetus .


(192a-c).[35] It is not possible to make false judgments of the following kinds. (K = something Known, R = something Remembered, P = something Perceived, C = something whose imprint we place in Correspondence with what we are perceiving [

]. Letters within square brackets indicate terms that seem to be presupposed but are not mentioned. The subscripts indicate cases where the text explicitly speaks of mistaking one such thing [x] for a different one [y ] [although that condition is presumably implicit in the other cases as well]. Thus 1a translates as, "One thing that we know and remember but do not perceive is something else that we know and remember but do not perceive.")


a. (KR~P)x




b. K




c. ~Kx




d. ~K




a. Px




b. P




c. ~Px




d. ~P




a. (KPC)x




b. (KPC)x


Ky [~P]


c. (KPC)x


Py [~K]


a. (~K~P)x




b. (~K~P)x


~K, [P]


c. (~K~P)x


~Py [K]


"It remains in the following cases, if indeed anywhere, that [false
judgment] will come about" (192c-d):


a. Kx




b. Kx




c. (KP)x




"False judgment remains in the following case(s)":


a. (KP~C)x


(KP~C)y (193b-c, 194a)


b. KP


K~P~C (193d)

According to 5, false judgment occurs when something we know (and possibly but not necessarily perceive) is mistaken for something

[35] Jacob Klein makes the intriguing observation that just as "Theodorus . . . distinguished fourteen oblong rectangles from the three equilateral ones; . . . Socrates also distinguishes fourteen cases in which false opinion is precluded from the three cases which admit it" (Plato's Trilogy [Chicago: University of Chicago Press, 1977] 128).


else that we perceive (and possibly but not necessarily know). The mistake can happen only because we have failed to put the wax imprint of the thing we know (x ) together with the present perception (y ) well enough to see that they do not fit.[36] This becomes explicit in 6, where false judgment is said to occur when something we know and perceive (but whose imprint we may not place in correspondence with the perception) is mistaken for something else that we know (and possibly but not necessarily perceive) and whose imprint we do not place in correspondence with the perception. In other words, in the case of at least one of the things that we are confusing with each other, we fail to compare adequately our knowledge (imprint) of a thing with our perception of it. The general conclusion that follows from all this (although it is not clear why Socrates chooses just the examples that he does) is that false judgment occurs because we sometimes fail to compare present perceptions with past imprints properly.

The wax model is thus successful in accounting for error in at least some cases of sense perception, but Socrates now proceeds to show that it fails when we attempt to apply it to intelligible rather than visible things; for in that case we can no longer speak of error as having to do with the fitting-together of knowledge with perception. For example, when we mentally add five and seven, and think the answer is eleven, we then think that eleven, which we know, is twelve, which we also know. But this would mean that we think that one thing that we know (but do not perceive) is another that we know (and do not perceive)—a scenario that has been declared impossible according to the previous survey of permutations (1a), so the model on which that survey was based must be discarded (196a-b).

Now that this hypothesis has been discredited, a fifth one is proposed, but before we turn to that I would like to raise a question about the elaborate classification that we have just reviewed: What happens to the category of memory (R)? Although memory is presented as the distinctive feature of the wax model (191d, 194d-e), it is mentioned in almost none of the cases specified.[37] It is present on both sides of 1a, then on only one side of 1b, and then not at all in the rest of the classification. Moreover, when 1a, 1b, and 1c are restated between steps 5

[36] In the case of 5a and 5c, where we not only perceive but also know y , we must also have failed to fit the perception of y with its own imprint.

[37] Accordingly Burnyeat (TP 97) and Polansky (188) do not even mention it in their schematic representations. All of Socrates' examples are represented in terms of P and K alone.


and 6, R is left out altogether even though P is now specified more explicitly (193a-b). It is left out again when Socrates reduces all the examples to a general statement: "I could never have false opinions about you and Theodorus either when I know both of you or when I know one but not the other; and the same applies to perceiving, if you follow me" (193b). And it is left out of the two summaries as well: at 194a-b Socrates says that false opinion turned out to be impossible about things that we do not know and have never perceived, but possible about things that we both know and perceive; at 195c-d he summarizes their findings as, "false opinion exists neither in the relation of perceptions to one another nor in thoughts but in the fitting-together of perception with thought." Thus, in case 6a we were told that the reason for the mismatch between the two objects is that the perception is indistinct (193b-c), but nothing was said about the obvious possibility that one's memory of one of them might be indistinct, although the possibility is explicitly built into the model (194e).

The reason that memory disappears from consideration seems to be that it is conceived in a way that equates it with active knowledge. This can be seen from the fact that, in the three places where it does appear, its truth value is identical with that of knowledge. On this model, to know is to remember and to remember is to know. Memory remains implicitly present in terms of C (which was introduced after R was dropped), for the ability to match the imprint to the perception implies having a correct memory of our former perception, but to substitute C for R is to leave out what is distinctive about memory, that is, the fact that it may become partially but not wholly lost, that it may exist in a state of latency. That will be remedied by the "having/possessing" distinction made in the next model, the aviary. If all this is meant to make us aware of the inadequacy of the wax-block model of memory, the elaborateness of the device would seem to be an indication of the importance of memory to the dialogue's concerns.

5 The Aviary (196d-200c) . The fact that memory can be latent rather than actualized is illustrated at the very beginning of the aviary model. Socrates asks Theaetetus, "Have you heard what people now say that knowing is?" and Theaetetus replies, "Perhaps, but I do not remember at present" (197a). Appropriately, Socrates goes on to distinguish "having" knowledge, which implies awareness, from "possessing" it, which does not. When we learn something, we possess it; but if like Theaetetus we cannot recall it, then we cannot be said to have it at


that moment. It is as if our mind were an aviary containing "all kinds of birds, some in flocks apart from the others, others in small groups, and some alone flying randomly through them all" (197d).

It is far from dear what these last details refer to. Plato may have in mind the sort of progression that Aristotle writes of at the beginning of the Metaphysics : from individual perceptions (random individuals), to the experience that results from a multiplicity of similar perceptions (small groups), to the science that discerns the principle common to all such perceptions (distinct flocks). Alternatively, the picture may be a reference to the method of collection (which was alluded to at 147d). The single birds may be individual knowledges that have not yet been related to others, as when we do not yet see that our knowledge of Socrates and our knowledge of Theaetetus belong together within a knowledge of the species of human beings. The small groups may represent various knowledges of species that have not yet been discerned as embraced within a genus, as when we recognize the species of human beings but do not yet see its relationship to other animals. And the flocks may represent our knowledge of genera. On either explanation the aviary appears to illustrate the progression of knowledge from individual perceptions to universal kinds.[38]

The aviary is empty at our birth, and the knowledges that we acquire through learning are birds that we capture for the aviary. When we first catch one and imprison it, we may be said to "possess" it, but we do not actually "have" it until we catch hold of it again (197c-e), that is, make use of it. The modal has the advantage over the wax block that it can account for knowledge that is latent rather than actual. But it has the disadvantage that it is no longer possible to match knowledge against perception—the birds do not seem to refer to anything outside the aviary. This does not seem at first to be a disadvantage, however, for Socrates' examples are no longer concerned with perceptual knowledge but only with mathematics. It is as if we have now moved beyond pistis to dianoia on the Divided Line. But the model cannot be assimilated to the doctrine of recollection, because it posits a mind empty at birth and filled entirely by empirical means. In fact the suggestion that we learn mathematics by having it handed over from teacher to student (198a-b) flies in the face of the Meno .

[38] Polansky, citing Lewis Campbell (The Theaetetus of Plato [Oxford: Clarendon Press, 1883] 199 n. 11), offers a different suggestion: "The small groups contain knowledge of species and genera, which embrace these individuals. The solitary birds flying through all are the common things, such as being, like, good, and so on, referred to earlier (185a ff)" (197 and n. 40).


By distinguishing between possessing knowledge (latently) and having it (actively) the aviary model enables us to avoid the paradoxical conclusion that false opinion is simply not knowing what one knows (199c). We can now say that it may be not having what one possesses. But two other difficulties arise. If false opinion is the mistaking of one bird for another—grasping the knowledge of eleven, for example, when we ought to be grasping that of twelve—then we make a mistake precisely by grasping a knowledge , which is a very strange conclusion. Thus,

first, for a person having knowledge of something, to be ignorant of this very thing, not through his ignorance but through his knowledge; second, to have the opinion that this is something else and something else is this; how can it not be very absurd for the soul, when knowledge has come to it, to know nothing and be ignorant of everything?

Theaetetus suggests circumventing the problem by supposing that the aviary contains ignorances as well as knowledges (199e), but Socrates replies that in that case the problem that the aviary was meant to solve—"How can we mistake one thing for another?"—reappears within it. We must ask how we can mistake an ignorance for a knowledge, and any attempt to answer the question would involve either an aporia or an infinite regress (200a-c). But in a sense Theaetetus is right, and we do have ignorance within the aviary: that was precisely the point of distinguishing between the (latent) possession and (active) having of knowledge. When we cannot grasp a knowledge that we possess, we are at that moment not in a state of knowing. Possessing, as distinguished from having, is a mixture of knowledge and ignorance. Let us consider the model more closely. Socrates says that mistaking eleven for twelve would be like mistaking a pigeon for a dove (199b). The analogy becomes clearer when later, in a different context, Socrates says that we have the same number in mind "when we say one, two, three, four, five, six; or twice three; or three times two; or four plus two; or three plus two plus one" (204b-c). This means, if we apply it to the present case, that our knowledge of eleven must include 5 + 6 and 4 + 7, while our knowledge of twelve would include 5 + 7, which sufficiently resembles the other two that it is readily mistaken for them, as a pigeon is for a dove.[39] When we first make such a mistake, we place into our

[39] Also see R. Hackforth, "The Aviary Theory in the Theaetetus " (Classical Quarterly 32 [1938] 27-29) 28. The fact that we can know a number (especially a large one) without knowing all the various operations by which it may be derived ("one, two, three, four, five, six; or twice three; or three times two; or four plus two; or three plus two plus one"), shows that a number is a whole that is not merely the sum of its parts. This will be of some importance later on.


aviary an "ignorance," that 7 +5 = 11, which we may continue to find there.

How then would we answer Socrates' question as to how we can think that something we know is something we do not know? On the wax model such false opinions were explained as a mismatching of perception to knowledge (because, for example, of the former's indistinctness). But the aviary model cannot provide such an explanation, because the birds, unlike the wax impressions, do not refer to anything beyond themselves. It often happens in Platonic dialectic that if two hypotheses are rejected a third is proposed that combines the positive features of each while avoiding their weaknesses, as number 3 in this section combined numbers 1 and 2.[40] No sixth model is proposed here to follow the wax and aviary hypotheses, but if we try to imagine what such a model would have to be like, it would be one that combined the "recognition" factor of the wax modal with the "latency" factor of the aviary model.[41] In view of the emphasis on memory both in the last two sections and in the dramatic byplay of the opening of the dialogue, and in view of the frequent allusions to the Meno , it may be significant that the doctrine of recollection does in fact incorporate both the features of latency and recognition.

That we are meant to come away with something positive from these discussions is suggested by the fact that Socrates could have refuted the "knowledge is true opinion" definition at the very outset if he had chosen to. After the aviary model is dismissed, Theaetetus reiterates this definition as still the best he can devise, and Socrates replies that true opinion cannot be the same as knowledge, because jurors can be persuaded to have a true opinion about something they have not witnessed, whereas only eyewitnesses have knowledge (201b-c).[42] Since Socrates

[40] Desjardins has documented in detail how this kind of dialectic forms the back-bone of the Theaetetus , as well as figuring strongly in other dialogues. She schematizes it as a kind of destructive dilemma, in which the unacceptable consequences of an original disjunction imply the need to replace that disjunction with a third possibility, in which the original opposition is reconciled—by means of what she terms the "complex emergence of new entities." See her RE .

[41] John Ackrill points out that "at the very beginning (191d5), Socrates says: 'whatever we want to remember, of the things we see or hear or think of, we imprint on the block.' Nothing is made of this last case within the wax block section. But at the transition to the aviary it is dearly implied that items thought of and imprinted on the block are (or include) abstract or universal ideas Thus the account of misidentification in terms of the misconnecting of two items . . . can be widened to cover misdescription and misclassification" ("Plato on False Belief: Theaetetus 187-200," The Monist 50 [1966] 383-402 at 394). Hence Plato prepares us in advance for the fact that the wax modal can be made to converge with the aviary model.

[42] As Burnyear points out (TP 124), a few lines earlier Socrates seems to say that it might have been possible for the advocates to give the jurors such knowledge through their oratory, if time were not so limited (201a-b). The tension between these passages is an interesting one, but it does not affect the present point. I shall comment on that tension later on.


does not offer this simple but crushing refutation at the outset, but chooses first to develop the abortive models in detail, it is worth trying to see the value of their implications.

Socrates' remark that we can only know what we have seen (201b) is reminiscent (at a different level) of the claim that lies at the basis of the doctrine of recollection. As the Meno puts it, knowledge is possible because in some sense we have already "seen" reality.[43] We can extend the aviary model in this direction. In some sense we have latent knowledge of reality a priori , and we can add this kind of bird to the latent memories of a posteriori knowledge with which Socrates stocked the aviary. Because the former is only latent we cannot always grasp it, just as we cannot always grasp the correct bird in the original aviary. When we perceive something (whether with the senses or the mind), it reminds us of one of these birds; and if we can grasp the correct bird, we then have knowledge of the thing perceived. But because many of the birds resemble one another, and because they are (as latent) indistinct,[44] we can mismatch a perception with a latent knowledge. This model (even in cases that do not require an a priori factor) avoids the paradoxes of the other one because when we make a mistake we are not in active possession of knowledge. Active knowledge arises only from the correct match (i.e., recognition) between latent knowledge and perception.

Even the revised aviary model, however, would leave us with the problem of how we can tell when we are matching correctly. How can we distinguish in practice between the true matching of 5 + 7 = 12 and the false matching of 5 + 7 = 11, or, more recalcitrantly, between "Justice is an arbitrary convention" and "Justice is a natural value"? If the wax block provided us with a correspondence model of truth, the aviary, which makes no reference to anything outside us, implicitly provides us with something more like a coherence model. Five and seven do not "correspond" to the twelve-bird in the same sense that our remembered knowledge of Theaetetus corresponds to the Theaetetus whom we now perceive; rather, the proposition 5 + 7 = 12 is internally coherent.

In the case of the arithmetical examples that Socrates gives through-

[43] 81c; cf. Phaedrus 249e-250a.

[44] Cf. the discussion of 165d, p. 83 above.


out the aviary discussion, we can see how the truth of the propositions might be tested by counting,[45] but no method is given to us by which we might test the truth of our opinions generally . That seems to be the function of the concept of logos , which will now be introduced, and which plays a similar role in the Meno (98a). If the wax and aviary models are related to each other as the correspondence and coherence models of truth, then in the Theaetetus correspondence leads to coherence, which in turn leads to the need for some kind of methodology or logos. A central problem of the correspondence model of truth has always been that of validation: How can we tell whether our thoughts correspond to something outside them, since we cannot get outside our thoughts to compare the two? Accordingly, proponents of a correspondence theory of truth tend in general to make use of a coherence model in order to validate correspondence. Since no other validation is possible, if we can assume that reality is logically coherent, the degree of coherence Within our thoughts will be a good test of the degree to which they correspond with reality. This was certainly Plato's view, and together with the correspondence model implicit in the theories of recollection and purification, a coherentism is implicit in the methods of hypothesis and division.[46] Thus in the Meno Socrates defends the dianoetic (coherence) approach to noetic (evidentness and correspondence) knowledge by saying, "Since all nature is akin, and since the soul has learned all things ["prenatally," but has since "forgotten" them], then if it recollects even one thing—which people call learning it—nothing prevents it from discovering all other things if one is courageous and does not stop searching" (81d).

We saw earlier (n. 42) that Socrates gives two apparently conflicting accounts, in close proximity with each other, of how empirical knowledge may be acquired. At 201a he says,

[45] Cf. Burnyeat: "What corresponds in the Aviary to the Wax Block's fitting of an imprint to a perception is counting . Counting is the attempt to identify, not the items themselves, abstract or concrete, which have number, but their number" (TP 109).

[46] The correspondence theory is possible, moreover, only on the assumption that something is evident to us independently of our thoughts. Accordingly, if correspondence leads to coherentism as a method of validating its success, it leads to evidentness or "illumination" as a justification of its possibility (cf. Phaedo 99e, Republic 507d- 509c, Phaedrus 248d). This triune conception of truth may be found in Aristotle as well: not only is he the first explicitly to formulate the correspondence theory, but he is also the first to work out in detail the logic of coherence as a tool or "organon" of achieving correspondence, and he also makes use of the image of evidentness or illumination (e.g.,' Metaphysics a . 1.993 9-11). It is often said that Descartes and his successors replaced the correspondence theory with the coherence theory, but it would be more accurate to say that they disengaged the coherence theory from the correspondence theory.


Or do you think there are any teachers so clever that, in cases of robbery or other violence where there were no witnesses, they are able to teach, within the short time allowed by the water clock, the truth about what happened?

But at 201b-c he goes on to say,

When jurors are justly persuaded about matters that can be known only by an eyewitness and not otherwise, if then they decide these things from hearsay, acquiring a true opinion, don't they decide without knowledge, being rightly persuaded if they passed judgment well?

The first statement appears to accept that, if more time were available, teaching could take the place of witnessing; the second insists that that is not possible. We could escape from the tension by interpreting the first remark in the sense of "especially within the short time allowed by the water clock," which would remove the tension but would not explain why Socrates bothers to mention the clock at all, a reference that only confuses the issue. Rather than trying in this way to eliminate the tension, between the direct knowing of the witness and the indirect knowing of those to whom the facts are demonstrated by argument, we will do better to notice that this tension, which is operative with respect to knowledge of the physical world, drops away at the level of noetic knowledge. As the Meno's slave demonstration showed, at this level rational argument may indeed bring us to the point where we are capable of seeing for ourselves, where persuasion and witnessing, coherence and correspondence, coincide (85c-d).[47]

9. The Logos of Knowledge (201c-210d)

In response to Socrates' counterexample about juries, Theaetetus suddenly recalls something that he had forgotten (at 148e he said that he had never heard a definition of knowledge): he once heard someone say that knowledge is true opinion with a logos (201c). In an inadvertent illustration of the phenomenon of recollection, he does not think that he can explicate this claim himself, but thinks that he could follow someone else who did. Socrates replies, '"Listen, then, to a dream in


exchange for a dream, for I seemed to hear it from certain people" (201d-e).

The puzzling description of this theory as a dream has given rise to several explanations.[48] It is reminiscent of the use of that term in a previously cited passage of the Meno : "At present these opinions, having just been stirred up in him, are like a dream. If, however, one were to ask him the same things many times and in many ways, you know that finally he would have knowledge of them that is no less accurate than anyone's" (85c-d). Perhaps the use of the term "dream" here in the Theaetetus is meant to suggest that the theory that follows is one that we should be able to recognize as true, but only indistinctly—as Meno's slave recognized the truth of the mathematical demonstration. It is a not-yet-adequate "recollection" of the nature of knowledge. But because of its lack of distinctness, Theaetetus, who like the slave can follow it but not exhibit it himself,[49] will never successfully formulate it in this dialogue.

Earlier, Socrates described the soul's thinking as "nothing other than a talk [

] with itself, in which it asks itself questions and answers them, and affirms and denies" (189e-190a). According to Meno's paradox, if we need to ask the questions in the first place, how can we answer them by ourselves? Conversely, if we can answer them, why did we need to ask them? On the theory of recollection this kind of dialogue


is possible because we know the answers latently but not overtly, and our self-questioning is designed to bring the "dream" into clearer focus. The same is true of Socrates' maieutic questioning. Socrates says that he is like a midwife in that he is sterile (with regard to wisdom), and that he has always been so (150b). But he said earlier that although midwives must be past childbearing, they must also have previously given birth, "because human nature is too weak to acquire an art concerning things with which it is not experienced" (149b-c). It does not seem possible, then, that Socrates was always sterile—or, at least, if he can be a midwife to wisdom, he must have something like a "memory" or "recollection" of wisdom, even if he never actually had wisdom.

With Socrates as his midwife, Theaetetus produces three versions of the dream theory, the third of which itself has three divisions. The theory is "that the primary elements from which we and everything else are composed have no logos." They can only be named; we cannot even say that an element is or is not, or we would be adding being or not-being to it. But we can give a logos of composite things by naming the elements of which they are composed (201e-202c). Theaetetus recognizes this as the theory he has heard. The paradigms that this theory has in mind, Socrates says, are "the elemental letters and composite syllables of writing [

means both "elements" and "letters"]. Or do you think that the one who said the things we have mentioned was looking somewhere else?" (202e). Theaetetus answers in the negative, but others have not always been so sure.[50]

1 (203a-d) . Socrates first points out that it makes no sense to say that the syllable can be known only on the basis of its elements, its component letters. For since the letters are elements, they are not reducible to a further logos; but since knowledge ex hypothesi requires a logos, the letters must be unknowable. In that case they can hardly confer knowability Upon the syllable. For example, if knowing the first syllable of "Socrates" means that we can give an account (

) of it by analyzing it into the letters S and o , we will be unable to know these

[50] E.g., Winifred Hicken ("Knowledge and Forms in Plato's Theaetetus," Journal of Hellenic Studies 77 [1957] 48-53), Morrow 328, Friedländer 3.186. It is generally recognized that certain aspects of the dream theory are reminiscent of the theory of forms. Whether the refutation of the dream theory is thereby also a refutation of the theory of forms is, however, a matter of dispute. Rorty 235 and Kunion Watanabe ("The Theaete-tus on Letters and Knowledge," Phronesis 32 [1987] 143-65, esp. 163) think that it does refute it, while Hicken 50-51 and McDowell 243-44 argue that it does not.


elements themselves because they are not further analyzable into something more elemental still. It is thus problematic how a complex can be known in terms of elements that are themselves not known.

2 (203e-206c) . Socrates then distinguishes a "whole" (

) from a "sum" (

SOCRATES: Perhaps we should have proposed, not that the syllable is its elements, but that from these a single form [

] arises, which itself has a single Idea [
] of its own, different from the elements.

THEAETETUS: Absolutely. And perhaps it will even be better this way than the other.

SOCRATES: Let it be then as we are now saying, the syllable is a single Idea [

] arising from the several conjoined elements, and it is the same in writings and in all other things.

THEAETETUS: Absolutely.

But Socrates quickly cuts off this route of escape:

SOCRATES: Isn't it the case that there must not be parts of it?


SOCRATES: Because if something has parts, the whole is necessarily the sum of parts. Or do you also say that the whole that arises from the parts must be some single form [

] that is different from the sum of the parts? (203e-204a)

This new hypothesis is attacked with a dilemma, the first horn of which immediately collapses the new distinction between whole and sum. "A whole is . . . that from which nothing is missing, and that from which something is missing is neither a whole nor a sum, which together become the same for the same reason" (205a). Socrates' objection begs the question by assuming that no account of a "whole" can be given that would satisfy the original stipulation that it is "without parts" and "different from the parts."[51]

Can such an account be given? If a whole is without parts, how can we speak of it in terms of "the parts" at all? It is this oddity that makes Socrates' refutation plausible. The answer would seem to lie in estab-

[51] See above, Chap. 1 S10, Hypothesis lib. Also see McDowell 243-44: "At Parmenides 157c4-e2 Plato sets out an argument, exploiting the same principles as the above reductio ad absurdum , in order m show that what a part is a part of, i.e. a whole, is not an entity designated by the standard use of the expression 'all the parts.' . . . Plato deliberately, and pointedly, uses against the dream theory a premise which he knows to be false."


lishing that a whole is correlated to a sum, so that one can speak of the parts of the sum in relation to the whole, but not as parts of the whole. There are various examples that can illustrate such a relation.

a. A species can be thought of as a whole, and in a sense we can think of its members as parts of the species; but they are not parts in the strict sense, because the species retains its integrity when individual members cease to exist or come into being. Correlated with the species is the totality of its present members, which stands to the species as a sum to a whole. The totality is affected by addition and subtraction of parts, but the species is not.
b. In the case of a syllable, since historically speech preceded writing, the whole sound was given first—as a unity that was only subsequently analyzed into letters by later grammarians. Consequently, in different languages the same sound is frequently represented by different collections of elements. The long o, in the first syllable of "Socrates," can be regarded as a single sound, as represented by the omega in Greek; or as a combination of two sounds, a short o (o ) followed by a long u , as represented by "ou" in the Wade-Giles and pinyin systems of transcribing Chinese; or, for that matter, "oh" in English, which does not have separate vowels corresponding to the Greek omega (long o ) and omicron (short o ), and therefore adds the h to indicate that the sound must be lengthened. Even to the extent that a syllable can be analyzed unproblematically into elements, the sum of the elements is not the same as the syllable unless they are properly united with regard to sequence, relative duration, emphasis, and so on. Thus, although in one sense we can speak of the syllable as the sum of its constituent letters, in another sense it is a pregiven whole to which we must look, as a paradigm, in order to put the right elements together in the right way. Wholeness implies an organization of the parts, whereas a sum is a simple aggregation. In terms of Socrates' later example of the parts of a wagon: when a wagon has been dismantled, or wrongly put together, the parts constitute a mere sum. The sum does not imply wholeness until the parts have been put together in a uniquely correct way.
c. According to the theory of forms, as well, the essential character of a thing precedes it as a paradigm, rather than following upon it as a consequence.[52] The importance of the sum/whole distinction to the


theory of forms was already visible in the first argument of the Parmenides, the argument from participation (Chapter 1, section 4). An analogous distinction was visible in relation to that dialogue's treatment of the One, in the distinctions implicit in the differences between the first and second hypotheses, and explicit in the third (Chapter 1, section 10).
d. In the Phaedo Socrates rejects Simmias's conception of the soul as an epiphenomenon of bodily parts (a sum). The soul is rather what makes possible the organization of the bodily elements into a unity in the first place.

How the two present hypotheses (that a syllable is its elements, and that a syllable is a whole that is not identical with its parts) may be integrated to achieve the model of a whole that is related to its parts but is not identical with them will become evident in the third hypothesis, below.[53]

The second horn of the dilemma is that if the syllable is a whole that

[53] The problem as presented here is related in part to Gail Fine's interpretation of the dream theory ("Knowledge and Logos in the Theaetetus," Philosophical Review 88 [1979] 366-97). She argues that although one cannot know the whole by enumerating the parts, "understanding any system consists in understanding how its elements are interrelated." But since it is also the case that "one does not understand a discipline's elements until one understands the system to which they belong," it follows that "accounts proceed in a circular fashion" (p. 356). The circularity is not, she feels, "an unfortunate problem. Rather, it is one of Plato's significant contributions to epistemology to have seen that we do not possess bits of knowledge in isolated, fragmented segments" (p. 396). In another sense, of course, Plato does regard the circularity as an unfortunate problem—i.e., insofar as it gives rise to Meno's paradox. The only way out is that the whole must in some sense be given independently of the parts. The metaphor of recollection of forms would be one way of representing such givenness. Alexander Nehamas writes, "The interrelation model of knowledge was actually first located in the dialogue by May Yoh . . . . Yoh, however, proceeds to connect this promising model, quite gratuitously, with the theory of Forms" ("Epistêmê and Logos in Plato's Later Thought," in Anton and Preus, eds., 288 n. 32). Yoh's account differs from the present one in that she does not connect the forms with the wholes, or with the interrelation of the elements, but with the elements themselves ("On the Third Attempted Definition of Knowledge, Theaetetus 201c-210b," Dialogue 14 [1975] 420-42, esp. 430). In this respect I would not disagree with Nehamas that the connection is gratuitous. But it is not gratuitous to bring in the forms when they are conceived not as elements but as principles of unity.


is not composed Of parts, then it is as irreducible as the letters, and equally unknowable (by logos): "The syllable falls into the same form [

] [as the elements] if it has no parts and is a single Idea [
]" (205d).

Socrates adds a more general objection to the dream theory. Our experience in the learning of writing and music has taught us that it is easier to know the elements than the composites, which is the opposite of what the theory claims.[54] He adds that this can be demonstrated in other ways as well (206c). In view of the way that the language in the above quotations irresistibly reminds us of the theory of forms, it would not be surprising if the primacy of knowing "uniform" forms over multiform individuals may be what is meant .[55]

3a (206c-e) . Socrates leaves aside the sum/part/whole question and turns to the question of what is meant by logos. The first hypothesis is that it means the mirroring in words of one's opinions. The hypothesis is dismissed because logos in this sense is natural to all normal people, so nothing would be gained by adding "with logos" to the definition of knowledge as right opinion. The image 'with which the hypothesis was presented is nevertheless a striking one: "the making dear of one's own thought [

] through the voice with verbs and nouns, as in a mirror or water, imaging the opinion in the stream through the mouth" (206d). Here words are viewed from the other side, as primary elements of meaning rather than as derivative complexes of elemental letters. Nouns and verbs are the regularities that make it possible to give meaningful form to the otherwise undifferentiated stream of voice through the mouth. In the same way the forms always were for Plato the regularities that made it possible for a meaningful world to arise out of the undifferentiated stream of Heracleitean flux. Here the analogy between nouns and verbs, on one hand, and forms, on the other, is never made explicit, but in the Sophist nouns and verbs will function as the signs of individual forms (259e-264b). There the Eleatic stranger will echo Socrates, saying that speech (
) manifests our thought (
) in "a stream proceeding from the soul through the mouth with voice" (263e);

[54] 206a-b. Burnyeat notes that "his silence about musical structure has the same provocative intent as his silence about the order of letters in a syllable. In music it is even more obvious than in spelling that the enumeration of elements is not enough. It is actually impossible to enumerate musical elements without reference, implicit or explicit, to structure" (TP 210 n. 94).


but he adds that speech is possible only "because of the interweaving of forms with one another" (259e).

3b (207a-208b) . The second hypothesis is that logos means an account of something in terms of its elements, such as listing the parts of a wagon. This is refuted by the observation that one may be able to enumerate elements without having knowledge in the normal sense. Someone might say, for example, that the first syllable of "Theaetetus" is spelled "The" (Q + e ), but incorrectly think that the first syllable of "Theodoros" is spelled "Te" (T + e ). In this case he does not know how to spell the syllable, but gets it right in the first case by right opinion. Therefore, on this understanding of logos, we can satisfy the definition without having knowledge (207d-208b).

Although this hypothesis does not further the investigation directly, the examples used have implications that further it indirectly. In the Theaetetus Socrates refutes the "dream" theory in terms of the way we learn to read. Syllables are wholes, and all knowledge is of a whole in terms of its parts, such as the syllable in terms of its letters; but on that definition of knowledge the ultimate elements or letters will be unknowable since they have no parts (e.g., S and o in the first syllable of "Socrates"), and knowledge of the syllable will be indistinguishable from right opinion (e.g., "The" in the names "Theaetetus" and "Theodorus"). But when we think about this example of reading, we find that Socrates has inverted it. We noted above that historically it was speech that preceded writing, rather than the reverse, and it is even more obvious that as individuals we learn to read only after we learn speech. Therefore we learn letters in terms of syllables, and syllables in terms of words, instead of the other way around. In fact the Statesman will make this point explicitly. In a passage that recalls Socrates' example here of the child who can spell the syllable "The" in "Theaetetus" but cannot spell the same syllable in "Theodorus," the Eleatic stranger says that children learn letters by seeing at first how they constitute the simplest syllables, and then using this knowledge as a paradigm for recognizing the same letters in more difficult syllables. The letters are then known in terms of their sameness with and difference from the letters in other syllables (277e-278c). Thus we come to know parts by noticing the similarities and differences among wholes, and seeing how these are reflected in the parts. Knowledge of the whole precedes that of the parts.

This point follows even more dearly from the example of knowing the wagon by listing its parts. Socrates adds that it would not, on this


definition, count as knowing the wagon if we could name the wheels, axle, body, rails, and yoke, but not the "hundred pieces of wood" from which they are built (207a); nor as knowing the name "Theaetetus" if one could list the syllables but not the letters (207b). The reference to knowing the name by knowing the syllables reminds us that the present discussion of whole and parts had altogether abstracted from the name or word as a whole, and asked only about the relationship between letter and syllable. But the meaning of the syllable comes from two directions: from the letters, which furnish its materials, and from the word itself—the nouns and verbs mentioned at 206d—which gives the syllables their purpose and meaning. Similarly, the basic parts of the wagon can be explicated either in terms of the hundred pieces of wood from which they are constructed, or in terms of the unity of the wagon, which is their reason for being. The hundred pieces of wood are not a wagon until they are properly unified.[56]

Implicit in the previous discussion was a conception of a "whole" that is not reducible to its parts. Implicit in this one is the conception of a unifying form that can explain the parts of a sum from above instead of from below. The sixth of the fifteen aporiae that Aristotle raised in Book B of the Metaphysics is,[57]

whether it is the genera that should be taken as elements and principles, or rather the primary constituents of a thing . . . . To judge from these arguments, then, the principles of things would not be the genera; but if we know each thing by its definition, and the genera are the principles or starting-points of definitions, the genera must also be the principles of definable things . . . . And some also of those who say unity or being, or the great and the small, are elements of things, seem to treat them as genera.

Essentially the same question underlies the present discussion of the Theaetetus . In response to this, Plato, like Aristotle, turns to a consideration of the nature of definition—a course subsequently pursued more intensively in the Sophist . There the "primary elements" will not be the minutest particulars but the most universal genera. The Sophist accordingly furthers the inconclusive inquiry of the Theaetetus in at least three ways. (i) The Dream paradox, that complexes are knowable in terms of their demerits but elements are ultimately unknowable, is, if not re-

[56] This has not gone unnoticed. Cf. E. J. O'Toole, "Forms and Knowledge in the Theaetetus " (Philosophical Studies 19 [1970]) 115; Rorty 237; McDowell 245; Rosen, "Socrates' Dream" 185-86; Burnyeat, "Socrates and the Jury: Paradoxes in Plato's Distinction between Knowledge and True Belief" (Supplementary Volume LIV of the Aris-totelian Society [1980] 173-92) 187-88, and TP 193-94; and Polansky 220.

[57] B.3.9958-21- 11; Ross's translation.


futed, at least undercut. The reciprocity of collection and division means that the whole and part are somehow implicit in each other: through collection (which appears only rarely in the Sophist ) we understand the whole in terms of the parts, but through division we understand the part in terms of the whole. (ii) Even though the whole may be known in terms of its parts, this knowledge is not reducible to its parts (as in 2, above), for we can know what "production" or "acquisition" is without knowing every kind of production or acquisition. (iii) Because parts can be known in terms of the whole, the logos of a thing need not be understood as the additive sum of the constitutive elements of the thing. Rather, the correct "method of logoi" (Sophist 227a) is to determine the essence of something by dividing the pregiven whole into derivative species. The method of division always takes precedence over the method of collection.[58]

The way that the Sophist goes beyond the Theaetetus can also be seen in its handling of the same illustrations. We have already seen that in the first model of "logos" (3a) the regularity inherent in the paradigm of "nouns and verbs" implied the combinability of universal forms, and that this conception becomes explicit in the Sophist . The same is true of the present paradigm of "letters." Letters in their own way represent regularities that limit the otherwise unlimited stream of vocal sound (cf. Philebus 17b), and as such provide illustrations of the stability and combinability of forms. The Eleatic stranger makes use of this paradigm too, in the Sophist , to illustrate the properties of forms: "Since some forms will combine and others not, they are in virtually the same condition as written letters; for some of these do not fit each other, but others do" (252e-253a). The Statesman takes the comparison farther still: not only do letters provide an image of the stability and selective combinability of the forms, but our ability to read language becomes a paradigm of our ability to understand incorporeal reality. The fact that we can transfer our reading proficiency from easy syllables to difficult ones, because of the regularity of letters, is a paradigm of our ability to discover in corporeal paradigms the regularities that enable us to "read" the highest things, "the long and difficult syllables of actuality [

]" (278c-d). Because the highest forms are of qualities (good-

[58] See Michel Fattal ("Le Sophiste: Logos de la synthèse ou logos de la division?" in Michel Narcy, ed., Etudes sur le Sophiste de Platon [Bibliopolis: C.N.R., Centro di Studio del Pensiero Antico, 1991) 156 and n. 17. Fattal suggests a connection between the priority that Plato assigns to division over collection, and Plato's preference for Parmenides over Heracleitus: the concept of logos, Fattal argues, is assemblative in Heracleitus but articulatory in Parmenides.


ness, beauty, justice, and even sameness and difference) rather than species (human being, bird, pig) there are no obvious visible examples of them: "Of the greatest and most valuable [

] things there is no image made clear for human beings, by the exhibition of which someone wishing to satisfy the soul of the inquirer can, by applying it to one of his senses, sufficiently satisfy it . . . . For the incorporeal, being finest and greatest, can be dearly exhibited only by reason and not by anything else" (285e-286a). However, we can use the visible world as a paradigm of the intelligible in an indirect way. Learning to read, by using the combinations of letters in easy syllables as paradigms of those in difficult ones, is offered as a paradigm of this kind of paradigm (277d).

3c (208c-210b) . The final hypothesis is what hoipolloi would say—that logos is the ability to name the sign by which one thing is distinguished from everything else, that is, the definition. But this hypothesis too must fail, because we must already know the difference between one thing and another in order to have an opinion about it in the first place, and so nothing new is gained by the addition (209a f.). This definition of knowledge will be either absurd or circular, depending on whether the logos about a thing's distinctness is itself regarded as opinion or knowledge. If it is an opinion, then we are told to add an opinion (logos) about something to the right opinion we already have about it; and "to command us to acquire the very things that we have, so that we may learn the things that we already believe to be so, greatly resembles someone completely in the dark" (209e). If, on the other hand, it is knowledge, then knowledge is defined in terms of itself, and the definition is circular (210a).

In the evaluation of this hypothesis the only example considered is the definition of "Theaetetus" (209a-c). Consequently, definition is conceived only in terms of an individual thing rather than a form or kind. But as a preliminary model Socrates had defined the sun as "the brightest of the heavenly bodies that revolve around the earth" (208d). This example is ambiguous. Although the sun is, like Theaetetus, an individual, it is a unique individual of its kind and therefore, like universals, admits of a definition by species and differentia. It is not made dear here whether a definition of an individual within a many-membered infima species is possible at all (209b-c); it is at least much easier to define a universal (208d). Nevertheless, even if Theaetetus cannot be defined, he can be known in the sense that enabled Theodorus and Soc-


rates to recognize him at the beginning. This knowing is not dependent on a single specifiable mark, but on his overall "look," which makes him easily distinguishable from Socrates, despite the similarity of individual features or "elements."[59] Even if a definition of Theaetetus is not possible, the example shows in an analogous way the importance, for knowledge, of the unifying "form." Nor should we forget that the other example, the sun, tends to be associated in Plato with the theory of forms.[60]

The definition of the sun is a counterinstance to the present argument's conclusion that definition is not possible. How are we able to define the sun in spite of Socrates' reductio? The reductio is a restatement of Meno's paradox, but there is no question in this case of recollection, since we are asking about a visible object. Nevertheless the answer is analogous: the explicit definition can be sought and recognized because we already know it implicitly. We have all the information necessary to conclude that the sun is the brightest heavenly object, even if the unification of the information into that description has never explicitly occurred to us. If we ask in turn how definitions are possible of the common properties that Socrates speaks of at 208d (unity, goodness, etc.), we will be led to an analogous conception of latency, this time regarding nonempirical knowledge. Such a conception is to be found in the theory of recollection. In the case of latent empirical knowledge the data are given, but the principle of unification is not; in the theory of recollection's conception of latent a priori knowledge the unity itself is given, but only in an elusive way.

10. Knowledge and Wisdom

We have seen that the Theaetetus's examination of knowledge goes through a progression of several different kinds of knowledge, a progression that reflects in a general way that of the Divided Line. It passes from perceptual (

) to interpretive (
) to mathematical (
) knowledge, before ending in aporia after a discussion that con-

[59] See Mitchell Miller, "Unity and Logos: A Reading of Theaetetus 201c-210a" (paper presented at the Society for Ancient Greek Philosophy, Dec. 27, 1989) 16. Also Burnyeat, TP 230.

[60] E.g., Republic 507a-517c, Phaedo 99d. Cf. Sayre PAM 135. The sun is referred to not only here in the very last argument of the dialogue, but also in the very first argument (153c-d, the Iliad's "golden rope"), and implicitly in the central digression (174a), where Thales is represented as the paradigmatic philosopher, who is concerned with the heavenly bodies (of which the sun is here defined as the brightest).


stantly evokes (but never invokes) the theory of forms and doctrine of recollection. The next step would be to return to the suggestions made by Socrates in the digression (and previously suggested by the dramatic byplay at the beginning of the dialogue), but never incorporated into the dialogue proper: in particular, the suggestion that the pursuit of wisdom is not ultimately satisfied even by adequate definitions, but eventually entails a change from one kind of life to another, like the "turning around of the soul" in the Allegory of the Cave (Republic 7.518c f.).

To the modem ear it sounds strange that our way of life should have anything to do with our intellectual ability to know things. There seems no obvious reason why thoroughgoing hedonists who pursue philosophy as a profession because they are clever and can make money at it should not be able to have a purely intellectual grasp of the nature of things without reforming their values and way of life. Even Plato's own doctrine of the tripartite soul seems to countenance this view, for it makes clear that even if we know the truth about things, we may not have the self-mastery necessary to act on it in opposition to the demands of appetite and ambition. Thus we may have knowledge without being good. Although this is true of ordinary knowledge, at the highest level of knowledge it is no longer true (accordingly, Socrates repeatedly insists on the inadequacy of his definitions of the virtues and the good in the Republic ).[61] Here, what we know and what we are coincide. The consummation of the Divided Line coincides with the consummation of the tripartite soul. This is the doctrine of "purification," which Plato advances in the language of the mysteries.[62] The highest, "moral" forms can be adequately grasped only to the extent that we are capable of experiencing moral truth within ourselves, and we will be capable of this only to the extent that we are free of attachment to the pleasures of appetite and ambition.

Perhaps the dearest illustration of this is to be had in Aristotle's Nicomachean Ethics , where we find that no adequate conceptual definition of goodness is possible, although we can define it merely formally as "that at which all things aim" or the "mean between the extremes of excess and deficiency." Only good individuals themselves can give content to these formulas; only they infallibly recognize goodness in concrete situations. The wisdom of the good person is not propositional,

[61] 4.435c-d, 6.504b, 6.509c.

[62] E.g., Symposium 210a-d, Phaedo 82d-83e, Republic 7.519a-b.


but is closer to what we earlier called "understanding."[63] This is Plato's point as well, that knowledge of the highest things requires an inner recognition that is inseparable from our devotion to those things.[64] And like Plato, Aristotle affirms that the difference between philosophy and sophistry is a difference in the kind of life that one chooses (

, Metaphysics F.2.1004622-25).

We have seen numerous reminders of the theory of forms throughout the Theaetetus , and have suggested that the aporetic nature of this dialogue may be a consequence of the overt absence of the forms. Nevertheless, it is best not to short-circuit Plato's enterprise in these dialogues by too hastily supplying "solutions" from other dialogues. There are certainly, to be sure, numerous reminders within the Theaetetus of those earlier ways of resolving some of the present questions; but nevertheless the Sophist and Statesman are clearly intended to be sequels to the Theaetetus , and it would be premature to interpose the earlier doctrines in their original form before we have followed the trilogy's own line of inquiry to its conclusion. In our discussion of 3b (above) we saw that the Sophist furthers the inquiry of the Theaetetus by showing that the part may be known in terms of the whole, as well as the other way around; that knowledge of the whole is not reducible to knowledge of the parts; and that a logos (definition) can succeed by beginning with the whole and dividing it progressively into species, rather than trying to proceed by aggregation from unknowable elements. Subsequently, in our discussion of 3c we saw that the Theaetetus's project of definition failed because, as is the case throughout the dialogue, the focus was on (unknowable) individuals rather than genera or kinds. This too will be remedied in the Sophist .

[63] cf. Burnyear: "Much of what Plato says about knowledge and its relation to true belief falls into place if we read him, not as misdescribing the concept which philosophers now analyze in terms of justified true belief [thus Runciman 183 n. 22], but as elaborating a richer concept of knowledge tantamount to understanding" ("Socrates" 186). Also see his TP 217, where he adds, "I have an ally in Bishop Berkeley, who conceived himself to be following Plato when he wrote, 'We know a thing when we understand it.'"

[64] Cf. Pascal: "Whereas in speaking of human things we say they must be known before they can be loved, the saints on the contrary say in speaking of divine things that they must be loved in order to be known" ("On Geometrical Demonstration," in Pascal The Provincial Letters, Pensées, Scientific Treaties , vol. 33 of Great Books of the Western World [Chicago: Encyclopaedia Brittanica, 1952] 440).


Chapter Three
The Sophist

1. The Sophist and Its Predecessors (216a-217e)

In the middle dialogues such as the Phaedo and Republic , Plato defines reality with reference to the criterion of rationality. Reason apprehends what is universal and unchanging, but not what is particular and in flux. The senses apprehend what is particular and in flux, but not what is universal and unchanging. Since reason is a more trustworthy guide to truth than are the changeable and deceptive senses, true reality is to be identified with "being" (the universal and unchanging) rather than "becoming" (the particular and fluid). This is the dichotomy represented later in the Sophist by the gods (friends of the forms) and giants (materialists), respectively. The former maintain against the materialists that "through the body we have intercourse with becoming by means of the senses, and by means of reason through the soul we have intercourse with real being, which always remains the same in the same respects, whereas becoming is different at different times" (248a). The leader of this dialogue is not Socrates but an unnamed stranger from Elea, who apparently is proposing to give up this dichotomy by neutralizing the difference between the gods and giants—in which case he would destroy the theory of forms in one of its most fundamental features. Consequently it is more important in the case of the Sophist than with most other dialogues to consider its standpoint in relation to that of its predecessors. There are in fact notable differences between the way soph-


istry—the defining focus of the present dialogue—is portrayed here and in the Socratic dialogues.

In the Republic , where the dichotomy between the divine and the corporeal is most thoroughly worked out, there is no thematic discussion of the nature of the sophist, but through the characterization of Thrasymachus and the views he champions in Book 1 the sophist appears as the paradigm of the unjust man, who subordinates his rational faculty to one or both of his lower faculties, appetite and spiritedness (which in him take the extreme forms of greed and ambition).[1] The entire teaching of the Republic , especially that of the tripartite soul, arises out of the challenge posed by these views in their restatement by Glaucon and Adeimantus in Book 2. The connection between sophistry and injustice will seem arbitrary and even question-begging to us if we think of sophistry as skepticism about objective knowledge, which eventuates in techniques for winning arguments. Skepticism in itself is neither just nor unjust, and the techniques of argument can be employed for just as well as unjust purposes; the Republic itself insists on this double-edged nature of techne (333e-334b). For Plato, however, sophistry is not simply equivalent to techniques for winning arguments. Otherwise, Socrates would have been for him, as for Aristophanes, a kind of sophist. The key to Plato's equation of sophistry with injustice lies in the sophist's boast to enable the weaker argument (

) to defeat the stronger. One whose art aims at such a goal cannot be motivated by reason (
); the motivation for wishing to defeat the strongest reasons must lie outside reason. According to the doctrine of the tripartite soul, it must therefore lie in a desire for personal gain, either in the form of the satisfaction of appetite or in the furthering of ambition. The power of sophistry enables one to defeat reason with its own weapons, and therefore to put reason in the service of appetite or spiritedness. It is the public equivalent of "rationalization," where the conclusion is dictated in advance by our self-interested desires (appetitive or spirited) and reason is directed to justify this with the appearance of rationality. Sophistry necessarily involves the governance of reason by the irrational parts of our nature, and therefore (in accordance with the Republic's definition) is intrinsically unjust.

Accordingly, in the Gorgias Socrates says that there seems to be no use for Gorgias's techne except as a means to unjust ends (481b). There,

[1] I have discussed this at length in "Socrates' Refutation of Thrasymachus and Treatment of Virtue" (Philosophy and Rhetoric 7 [1974] 25-46).


too, injustice means the defeat of reason by appetite.[2] Gorgias's techne is called rhetoric (449a), however, which Socrates distinguishes from sophistry: rhetoric is related to the sphere of justice (

), and sophistry to that of lawmaking (
).[3] Nevertheless, since the two differ only in their sphere of employment, their basic nature must be similar.

This conclusion is supported by the Protagoras , whose fide character is a sophist, both by general repute (311e) and by his own avowal (317b). Protagoras, however, is less candid than is Callicles or even Polus in the Gorgias , and he cites the popular view that an unjust person ought never to admit that he is such (323b). He does reluctantly admit that self-control (

), which is a virtue, does not necessarily imply justice (333d). But this does not commit him to an approval of injustice, since, unlike Socrates, he is willing to consider the different virtues as independent of one another. Nevertheless the fact that he was the first to charge for teaching virtue (349a) would be damning enough in the eyes of Socrates, who says in the Gorgias that it is shameful to charge money for teaching justice and virtue. If one succeeds in teaching it, the student, being just, will of his own volition compensate the teacher who has benefited him. If the student does not do so, the teacher has failed to make him just and deserves no compensation (521d-e). Accordingly, in the Protagoras Socrates sees the sophist not as a teacher or thinker but as a moneymaker, more concerned with selling his wares than with whether they are any good (313c-e). He is governed by the values of the appetite rather than reason and is, therefore, essentially unjust.

Coming to the Sophist from the Protagoras, Gorgias , and Republic , we would expect that here too the sophist will be identified in terms of his fundamental injustice, his use of reason not as an intrinsic good but as an instrument of gain, whether for the sake of power or the satisfaction of appetites. It comes as a surprise, therefore, to discover that at the end of the dialogue the sophist is defined in merely technical terms, as (1) the maker of semblances rather than likenesses (266d-e), (2) op-

[3] Gorgias 464b-c. In the same place rhetoric and sophistry are, in turn, related to each other as medicine is to gymnastics. The Sophist (227e-229a) will use medicine and gymnastics as analogues of, respectively, chastisement of evil souls (resulting in justice) and teaching of ignorant souls (resulting in knowledge). So in both dialogues justice appears as analogous to medicine in distinction from gymnastics.


erating on the basis of opinion rather than investigation into his subject (267b-e), and (3) making his claims with irony (268a).[4] There is no explicit reference to the tripartite soul, and nothing is said about the subordination of reason to ambition or hedonism. Indeed, sophistry is made an aspect of production rather than acquisitiveness (265a).[5] The absence of any reference to appetite and ambition is the more perplexing in view of the fact that these categories return as central at the end of the Sophist's sequel, the Statesman , where they define the two fundamental types of citizen, the moderate and the courageous.

A second question, which will turn out to be related to the previous one, is why Plato begins the dialogue with several apparently inconsequential preliminary captures of the sophist. For the purposes of demonstrating the method of division, the Angler division would have sufficed. The key to both these questions lies in the original one, the relationship of the Sophist to the doctrine of the tripartite soul in the Republic . Although that doctrine is never explicitly mentioned in the Sophist , there are numerous indirect references to it in the preliminary divisions of the dialogue, which depict the sophist in terms of moneymaking and conquest. Following the Angler division, there are a total of six "sophist" divisions (according to the enumeration at 231d-e): four main divisions plus suggestions for two additional versions of the second.

2. Preliminary Divisions (218a-232b)

The fundamental dichotomy from which all the subsequent divisions proceed is between the art of production and the art of acquisition, and

[4] Moreover, as Guthrie 5.158 points out, in the final dichotomy the sophist is identified with the maker of contradictions rather than speeches, the reverse of what Plato's usual characterization of sophists (in the Gorgias, Protagoras , and Republic 1, for example) would lead us to expect.


it is remarkable that although the subsequent final division will locate the sophist under the form of production (265a), all but the last of the preliminary divisions find the sophist under the form of acquisition. Moreover, the latter, which alone tries to locate him within the productive rather than acquisitive arts, fails to discover the sophist at all (230e-231a), and precisely for that reason.

The stranger summarizes the results of these preliminary divisions at 231d-e. In the first division the sophist showed himself as "the paid hunter of rich young men, . . . second, as a kind of merchant [

, traveling businessman] of things learned in the soul, . . . third, as a retailer [
] of these same things." Theaetetus interrupts to say, "and fourth, as a vendor [
] of things learned that he himself produced." The stranger's response, "You have remembered correctly," is ironic, for what Theaetetus calls the fourth version had originally been part of the third: "Buying some, and making by himself other things learned, and selling them, he makes his living from this" (224d). It was not assigned a separate number, and Theaetetus had gone on to call the next one number four (225e). That is now called the fifth appearance: the sophist as "an athlete in contests of words, who has appropriated the eristic art" (231d-e). The sixth is a purifier of the soul (231e).

The precise number of divisions is thus overtly ambiguous. In fact there are only four fundamentally different preliminary divisions of the sophist, and they systematically reflect the Republic's account of the sophist based on the tripartite soul. In order to avoid the numerical ambiguity, and to collapse the variations of merchant-retailer-vendor into one, I shall designate the four as A [= I], B [= II/III/IV], C [= V], and D [= VI].

The demonstration division of the Angler (218e-221c), which provides the model for the others, begins with the words, "Shall we consider him as having an art or as someone without an art but with some other power?" (219a). Implicitly, then, the division begins with the form of power (

—into which the stranger later collapses the two-world distinction) divided into the species of art (
) and nonart (see Figure 3). The summary, apart from its lack of reference to the preliminary form of "power," is complete and accurate, as are the final two— which makes more intriguing the fact that the next three are not.[6]

The Angler division is followed by division A of the sophist (221c-223b), which recapitulates steps 5 and 6 of the Angler division, before

[6] The inaccuracies are noted also by Klein (PT ), Rosen (PS ), and Benardete.



Figure 3



Figure 4

taking its own departure from 6a (see Figure 4). Although the summary is more explicit than the division in some places, it also drops one (and only one) of the division's steps, the distinction between persuasion and violence in step 8, and thereby the distinction between reason and spiritedness. We shall return to this point later.

The third subdivision of the Angler (exchange/conquest) corresponds


to the Republic's distinction between the commercial and spirited alternatives to reason. Accordingly, the fact that the first division of the sophist begins on the side of conquest suggests that we are here looking at him under his spirited aspect. But this is somewhat moderated by the next subdivision, step 5, which casts him in the role of a hunter rather than a fighter. Since a hunter aims at providing himself with sustenance, hunting makes reference to the appetitive as well as to the spirited side of our nature.[7] The stranger's description of this sophist as "the paid hunter of rich young men" (231d) combines both the attributes of commerce and aggression. Where this first division incorporates the spirited and appetitive elements jointly in its starting point, the starting points of divisions B and C will present each one separately.

Division B follows immediately, beginning under the heading of exchange, that is, subdivision 3a under Angler (see Figure 5). Here again the summary omits precisely one of the original steps, namely, step 5 (there is also a difference in the wording of step 8), the distinction between selling one's own products and those produced by someone else. The collapsing of that distinction implies that a sophist may be a merchant of his own products, which completes the symmetry: merchant of someone else's products (division II), retailer of his own products (division III), and retailer of someone else's products (division IV). Moreover, it implies that what matters is not whether he is productive, but that he is mercenary. In this appearance the sophist is guided exclusively by appetitive motives. He is out to make money by selling his knowledge of virtue.

The implicit symmetry mentioned above is made explicit when the stranger suggests that the sophist need not travel (224d), and might therefore be a retailer (6a) rather than a merchant (6b), and (224e) that he might make his own products (5a) rather than selling someone else's (5b). If these alternatives had been pursued, they would have yielded three more divisions (all variants on the appetitive aspect of the sophist). But they are not pursued, and Theaetetus, as we saw, counts this in one place as one more division (225e) and in another place as two (231d). He might equally have counted it as three (as the discrepancy between the division and the summary implied above), or as none, since all are minor variations on the preceding one. The introduction of so many variations emphasizes that the crucial point is the selling, not the

[7] In the Republic , hunters enter the city (373b) only when the formerly healthy, basic polis becomes feverish (372e) and begins to change from an agrarian (appetitive) society to a warrior (spirited) one (373d-e).



Figure 5

further distinctions under 5 and 6. The point common to them all is the restriction of sophistry to commerce, which is reflected later in the stranger's description of this group as showing the sophist to be a merchant or retailer (231d-e).

After this comes division C, which begins with species 4a under Angler (fighting), although the summary begins with division 2b (acquisition). The italics indicate where the initial steps are recalled from the Angler division in accordance with (although in more detail than) the summary (see Figure 6). Once again, a single step is omitted from the summary, namely, the distinction in step 3 between exchange and conquest.

Here it is the spirited element that is isolated as the starting point. "Fighting" rather than "hunting" leaves no doubt that only the spirited



Figure 6

element is initially considered. The spirited nature of this aspect of the sophist is confirmed by the stranger's summary of the present division, which reveals the sophist to be "an athlete in contests of words, who has appropriated the eristic art" (231d-e). Nevertheless the final distinction is between moneywasting and moneymaking, which suggests that even at his most spirited the sophist is never entirely free of mercenary motives—a point that is emphasized by the summary's implicit collapsing of the exchange/conquest distinction, and its addition of the prior category of "acquisition." By contrast, in the case of the specifically appetitive sophist of division B, no spirited factors entered at all.

The Republic's third element, reason, is present in all three divisions,


but always in the role of a means to an end, rather than an end in itself or point of departure. In division A it appears in the form of persuasion (8b), which the sophist as hunter uses in lieu of violence to catch his prey. Its subordinate role is emphasized even more in the summary, which omits the distinction in step 8 between persuasion and violence. In division B it appears among the works that serve the soul rather than the body (7b), particularly those concerned with virtue (9b), which are sold by the sophist as retailer or as merchant. And in division C it appears as the verbal rather than physical weapons that are employed by the sophist as fighter (6b). In each case reason is subordinated to appetite or spiritedness, and the three cases correspond to the two types of injustice, and their combination, which are developed in the Republic and illustrated in the person of Thrasymachus.

Division D, which follows, is the only one in the dialogue (including the final one) that does not follow from the Angler division, and the only one that follows from a preliminary "collection," in accordance with the way the method is described in the Phaedrus (265c-266b).[8] The stranger surveys various types of tasks performed by servants (226b-c), the common denominator of which is that they all involve an act of division (

) and are accordingly given the name of discrimination (
). The division proceeds as shown in Figure 7.

Here we do not find the sophist (or any discrepancy in the summary). What we find is a species about which the stranger says, "I'm afraid to say [these are] sophists . . . lest we accord to them too great an honor"; they resemble sophists, but only as the dog resembles a wolf (231a). He calls this species instead "noble sophistry" (231b). The "noble sophist" is generally recognized to be a reference to Socrates (cf. 230a-d), and the general form from which the division begins, "discrimination" or "division" (

), shares its latter name with the philosophical method of the stranger.[9] The stranger and Theaetetus do not find the sophist precisely because they begin with a purely rational form and make no reference to appetite or spiritedness. This division demonstrates negatively what the others demonstrated positively, a necessary connection between sophistry and the lower parts of the tripartite soul. Where we find the sophist we are within the realm of one or both of the

[8] Socrates' opening remark in the Sophist , that the stranger may be a god in disguise (216a), is perhaps meant to remind us of the Phaedrus , where someone who is able to perform collections and divisions properly is compared to a god (266b).

[9] The belief that ignorance is never voluntary, is held not only by both the noble sophist and Socrates, but by the stranger as well (228c).



Figure 7

irrational parts of the soul, and when we confine ourselves to the soul's purely rational operations, we do not find the sophist. If we think in terms of appetite (division B), or spiritedness (division C), or both together (division A), we will find the sophist, but if we confine ourselves to rational discrimination alone (division D), we will not.

The most direct reference to the tripartite soul occurs in division D, step 3, when the stranger draws a distinction between chastisement, which corrects vice, and teaching, which corrects ignorance (227d-e). The difference between vice and ignorance is that ignorance, like deformity in the body, is a kind of missing of the mark, whereas vice, like disease in the body, is a kind of disharmoniousness.[10] The latter is in

[10] The distinction turns out to be less absolute than it seems at first. At 230c the stranger compares one aspect of the instruction that corrects the soul's ignorance to the practice of physicians. This draws an implicit connection between ignorance and disease, whereas previously ignorance was connected with deformity, and vice with disease. The implication is that even though ignorance is no longer conceived as a necessary condition for vice (as it was by Socrates), it is still conceived as a sufficient condition. Someone with the wrong ideas about goodness will not be virtuous, even though the right ideas are no guarantee of virtue either. For a perceptive and carefully annotated discussion of this passage, see Paul Gooch, "'Vice Is Ignorance': The Interpretation of Sophist 226A-231B," Phoenix 25 (1971) 124-33.


fact the definition of vice (injustice) given in Republic 4 in terms of the tripartite soul—both as disharmony (444b) and disease (444c-e). Moreover, during the discussion of vice the stranger remarks, "In the souls of people who are in a worthless condition, do we not see opinions opposed to appetites [

], anger (
) to pleasures, reason (
) to pains, and all of these opposed to one another?" (228b). Here we see not only echoes of the oppositions used in Republic 4 to establish the three parts of the soul (437b-441c), but the very terms by which the three parts are denominated. Thus we might say that division D provides us with, among other things, the schema by which to recognize the pattern (the doctrine of the tripartite soul) underlying the previous ones.[11]

Even the distinction between ignorance and vice already implied the tripartite soul. The Socrates of the early dialogues had equated virtue with knowledge, and thereby vice with ignorance. But these equations could not explain what Aristotle later called moral weakness (

), in which our knowledge of what is right cannot overcome the opposing pull of our passions. The tripartite division of the soul enabled Plato to introduce the concept of self-mastery (Republic 4.430e-431a), that is, control of anger and appetite by means of reason, which supplied the deficiency in the Socratic account and at the same time denied the simple equation of virtue with knowledge and vice with ignorance. If our soul is not harmonized by self-control, then knowing the good will not ensure doing the good. The distinction here between vice and ignorance therefore implies the concept of self-control and, to readers of the Republic , further implies the tripartite soul.

[11] All this means that the opening divisions have more substantial importance than is generally recognized. What importance they are usually thought to have is limited to their exhibition of diversity. For Taylor 381 they are meant to show that the sophist is the one who "can masquerade in all these guises" and is accordingly "a pitter of discourse against discourse, a contradiction-monger." For Cornford (PTK 187), "the first six Divisions actually, though not formally, serve the purpose of a Collection preliminary to the seventh," a view shared by Sayre (PAM 154-55), except that for him "the first five Sophists serve as a collection for the seventh, and . . . the definition of 'Sophist VI' illustrates the way in which a common property is to be sought." Other writers such as Friedländer, Rosen, and Benardete take the details of the opening divisions more seriously and find much of interest to say about them, although they do not see the divisions as embodying an underlying pattern.


The fact that the Sophist does retain—even if only implicitly—the doctrine of the tripartite soul makes more interesting the question why Plato now abandons his former practice of using this model in order to distinguish philosophy from sophistry. All the more so, because the distinction that he draws between them at the end of the dialogue, without that model, proves to be unsatisfactory.

3. Beginning of the Final Division (232a-236c)

The stranger now undertakes the search for the sophist, for which the preceding divisions were preparations. From the very beginning this final division lacks the rigor of its predecessors. Several of the steps remain entirely implicit. Moreover, when the division is resumed after a long digression, there will be some significant departures from the present model. Because of the first of these factors I shall summarize the steps before diagramming them, and because of the second I shall use Roman numbers here, reserving the Arabic numbers for the final form of the division.

i. In another reminder of the Republic , the sophist is now described as someone who can produce everything through a single skill (233a f.), as the poet had been described in Republic 10 (596c f.).[12] The stranger asks, "[What] if someone would claim that, by a single art, he knew how, not to speak or dispute about, but to produce and do all things whatever?" (233d). This is an odd beginning, for after thus contrasting "speaking about" and "producing" as two different arts concerning all things, the present division looks for the sophist under the species of a producer rather than a talker, which is not what we would have expected.

ii. The stranger continues, "And furthermore, having quickly produced [images of] them all, he sells them for very little money" (234a). As with step 4 of division B, the sophist sells his products rather than giving them away.

iii. The fact that he charges little for "all things" means that he must be playing (rather than serious; 234a).

iv. This species of play is a form of imitation (234b).

v. Specifically, it is imitation in words rather than pictures, resulting in spoken images (

) (234b-c).

[12] Compare especially Sophist 234b with Republic 598b-c.


vi. Images may further be divided into likenesses (

) and semblances (
; 235d-236b). The difference between them is that a likeness represents the proportions of the original accurately, while a semblance compensates for the distorting effect of the audience's perspective. A nonverbal example of semblance is a tall statue whose upper portions are exaggerated in size to compensate for the fact that they are farther from the viewer and would otherwise appear too small (235d-236b).[13] Thus, in an important sense, the semblance is not the original. The question of how a semblance cannot be what it seems to be leads into the long central discussion of the dialogue.

The preceding steps may be schematized as shown in Figure 8.

In all three "successful" preliminary divisions, the sophist was sought among the acquisitive rather than productive arts (division D, which was unsuccessful insofar as it caught only a noble cousin to the sophist, did not make use of the productive/acquisitive distinction). Here for the first time he will be sought among the arts of production. And where the first four divisions classified him under verbal (or pedagogical) skills, here he is classed as producing things rather than talking about them.[14] Why then is he portrayed here as productive rather than, as before, acquisitive?

There would have been no difficulty in pursuing the present course once again under the heading of acquisitiveness. The model for doing so was already provided in division B, which begins with acquisition instead of production (Angler 2), by exchange rather than by conquest (Angler 3), and by selling rather than giving (Step 4). At that point there is a distinction between selling one's own products (5a) and those other than one's own (5b), which leads (according to Theaetetus's enumeration at 231d) to the fourth and third appearances of the sophist, respectively. Accordingly, the final division might have begun with 5a, in which the sophist would be seen as a vendor of his own products, and the division might then have proceeded to specify the precise nature of the products, as it does here. The difference would be that all these divisions, including his art of production, would fall within the form of acquisitive arts. In fact there is nothing here against so interpreting the present starting point (i.e., as proceeding from 5a of division B rather

[13] Cf. the Phaedrus 259e-260d, where Socrates criticizes rhetoricians for being more concerned with accommodating themselves to their audience's beliefs, whether true or false, than with accommodating themselves to the truth.

[14] I.e., in step i. The verbal component is, however, readmitted in step v as a species of production: images in words are distinguished from images in pictures.



Figure 8

than 2a of the Angler division), except that no explicit encouragement is given to such a move. The only reference to the sophist's acquisitiveness (and therefore the only common ground with the Republic's conception of sophistry) is to the fact that the sophist sells his products rather than giving them away, and even here the money seems relatively unimportant, since he sells them "for very little money." Curiously, when this beginning is recalled at the end of the dialogue, even the slight reference to acquisitiveness is eliminated, as well as any ambiguity about whether the starting point might be located within division B, and therefore within the form of acquisitive rather than productive art.

4. Likenesses and Semblances (236a-242c)

The division founders on the problem of how to say that something "is not." If something is an inaccurate image, then in a significant sense it "is not" the original. But Parmenides argued (and his arguments in-


spired sophists like Protagoras) that it is impossible to say or think that something is not. When we attempt to do so we must conceive the nothing (what is not) as if it were something, and so we fall into contradiction. If this aporia cannot be resolved, the present attempt to define sophistry in terms of false images will fail (236a-239c).

The stranger now appears to shift his ground. It seemed at first that the problem lay only with the falsity of semblances. But now a less obvious problem arises that tacitly broadens the problematic to include not only semblances (inaccurate images) but likenesses (accurate images) as well. Without calling attention to the shift, the stranger now asks, not for a definition of semblances in particular, but for one of images in general (239d). The ensuing discussion will consistently put its questions in terms of images in general rather than semblances in particular (239d-240c, 241e, 264c). Most surprisingly, in one place the problematic is formulated not only not in terms of false images— semblances—but not even in terms of images in general, and rather in terms of true images, likenesses. The stranger says, "Without really being, then, it really is what we call a likeness?" (240b). Shorty thereafter he explicitly collapses the distinction between accurate and inaccurate images, and puts all images on the same basis for the purposes of the present discussion: he speaks of "false statements or opinions—whether images, likenesses, imitations, or semblances" (241e).

Why has the stranger broadened his attack to include all images, whether accurate or inaccurate? A semblance "is not," in the sense that its proportions are not those of the original. As long as we confine our example (as the stranger does) to the relationship between a statue and its model, the difference between semblance and likeness is clean But the present case is about semblances and likenesses in words (

). What does it mean to speak of accurate or inaccurate representation in words of the true proportions of a thing? We have already had at least four different images of the sophist, at least three of them somewhat persuasive. A fifth is on the way, which will be more rigorous in some ways but less persuasive in others. Throughout the dialogues, including the Theaetetus , Plato shows himself well aware of the difficulty, and perhaps impossibility, of giving exceptionless definitions of philosophical concepts, that is, perfectly accurate images in words. Does this mean that when he has given us imperfect verbal models, as Socrates often admits to doing, he is producing semblances rather than images, and practicing sophistry rather than philosophy? Unlike the sophists he does not deliberately distort his models in accordance with his audience's point of view. The distortion is involuntary and unavoidable.


But the question of whether the distortion is voluntary or not is irrelevant: if the image is a distortion, it cannot be a likeness. The intention of the producer tells us what was aimed at, but not what was produced. Only the nature of the product tells us that.

To put the matter more radically, in a sense any image is by nature a distortion of the original. At the very least the material from which it is made is different from that of the original.[15] This is true to a relatively small degree if we compare a well-sculpted and well-painted statue with the original model, and it approaches a vanishing point if we compare a painting of a painting or a sculpture of a sculpture with the original painting or sculpture; but it becomes clearly evident when we try to compare justice itself, for example, with words about justice. The medium of words is so different from that of forms that the very concept of an accurate image is seriously problematic. Since physical things are images of forms, as the Divided Line avers, then they necessarily fall short, if only because their mode of existence is so different. The Phaedo and the Parmenides ,[16] too, allude to the ambiguity of whether things can be said to be similar to the forms in which they participate, but the clearest statement of the problem is in the Cratylus :

SOCRATES : Would there be these two such things as Cratylus and a likeness [

] of Cratylus, if one of the gods not only copied your color and shape, like painters, but also made all the things within the same as within you, and bestowed the same softness and warmth, and put in them movement, soul, and intelligence like yours. In a word, everything that you have, he would place another instance of beside you. Would then this be Cratylus and a likeness of Cratylus, or two Cratyluses?

CRATYLUS : They seem to me, Socrates, to be two Cratyluses.

SOCRATES : Do you see then, my friend, that we must seek a different kind of correctness of likenesses, and of the things we were just talking about [names], and not require that if something subtracts

[15] Rosen puts the point more strongly: "An (accurate) image of a given original 'is and is not' the original, and in the same respect, not in two different respects. An accurate copy of a certain look is the same look. And yet, precisely as the copy (not, in other words, as a distinct material realization of the copy look), it is not and cannot be the same as the original. In sum: if it is the same, it is not a copy. If it is a copy, it is not the same" (PS 191). I am not sure, however, that one can speak of a copy as "not . . . a distinct material realization of the copy look." It does not seem possible to distinguish a copy and original as separate things unless an individuating material substratum is built into the modal.

Gilles Deleuze is highly sensitive to this ambiguity/instability in his treatment of the Sophist in chap. 1 of Difference and Repetition.

[16] Phaedo 74c-d (see PP 60-62) and the second and fourth arguments of the Parmenides (132a-b, 132d-133a).


or adds something it is no longer a likeness? Or do you not perceive how far likenesses are from having the same qualities as those things of which they are likenesses? (432b d)

If no image can perfectly represent the original, we can understand why Plato extended the problem of falsity from overtly false images (semblances) to images in general, expressly including overtly true ones (likenesses). Accordingly, if the following discussion is to be successful, it must explain not only how sophistry is possible but also how philosophy is possible. We will see that this is not adequately achieved in the Sophist ; and when we get to the Statesman we will find that the stranger, without distinguishing between likenesses and semblances, relegates all images and imitations of the unformulatable "science of the mean" to the same inferior status, and calls those who produce the images "sophists of sophists," regardless of whether the images they produce are the best possible (the regime of a constitutional monarch) or the worst (that of a tyrant; 303b-c).

Moreover, if, according to the Divided Line, physical things are related to forms as images to originals, the ontological status of physical things also comes into question at this point. In fact, the stranger will later suggest that connection himself when he divides the form of divine production into "entities themselves" and "images" (265e-266a)—a dear reminder of the two-world ontology of the Divided Line. What makes it even more emphatic is that there was no need to divide the form of divine production at all, since the sophist is pursued within the collateral form of human production.

There is no need to wait until that point in the dialogue for a reminder of the two-world ontology. As soon as the problem of images is defined, the stranger asks Theaetetus for a definition of "image," and Theaetetus, as he had done with Socrates in the Theaetetus , responds with a list of examples. One way in which this is suggestive is that the list uses examples reminiscent of those used to explicate eikasia and pistis in the Divided Line and the Cave: "images in water and mirrors, and, further, in paintings and statues and all other such things" (239d). Even more suggestive is the stranger's reply. He does not simply say, as Socrates did in similar circumstances, that what he wants is not a list but the quality that enables all items on the list to be called by the same name. Instead he prefaces that request with the comment that a sophist would reply by acting as though he had no eyes and had never seen these things; he would want to know "what follows from words alone" (239e-240a). By putting it in this way the stranger draws a distinction


between the visible world and the intelligible world, and insists that philosophy (for which "the sophist" here stands ironic proxy) is concerned only with the intelligible world.

Theaetetus's ensuing attempt to give a proper definition of "image" complements the problematic with which the section started. Previously the stranger emphasized the falsity of images. Now Theaetetus conversely emphasizes the truth of nonimages. An image is "another such thing made in the likeness of the true one [

]" (240a). But the original problem now reappears in this obverted form. As the stranger points out, if "true" means what really is (
), then because the image does not have true being, we cannot say that it really is, even though it "really is" an image (240b). Theaetetus acknowledges this, using language that anticipates the language by which the problem will be solved: "Not-being does appear to have become twisted together with being, by some such weaving [
], and it is very strange" (240c). The only way out of this aporia, the stranger says, is to distinguish among senses of not-being, so that there will be a sense in which it is meaningful to say that not-being can be. Some commentators take this passage as evidence that Plato saw himself as going beyond Parmenides, who said that we can neither say nor think that "it is not." They point to the Eleatic stranger's use of the word "parricide" in relation to his teacher, "father Parmenides." But the stranger's words are, "Do not think that I am becoming a kind of parricide" (241d). If anything, these words suggest that Plato does not regard these distinctions as fatal to Parmenides' position.

5. Being and Not-Being (242c-251a)

"When I was younger," says the stranger, "whenever anyone spoke of what now brings us to an impasse—'not-being'—I thought I understood it exactly. But now you see what an impasse we are in with regard to it . . . . Then perhaps we have admitted this same condition into our soul no less with regard to 'being'" (243b-c). He proceeds to establish this by showing that (a) neither the pluralists nor the monists have given an adequate account of being, and (b) neither those who equate being with corporeal matter nor those who equate it with intelligible form have done so either. His refutations of these positions will be along the general lines of the Phaedo's method of hypothesis, as will his replacement of them by a higher hypothesis. Accordingly, the method of division cannot be regarded as a replacement of the method of hypothesis,


as is often claimed. They complement rather than compete with each other.

A. The One and the Many (243d-245d)

In a passage that recalls the Phaedo's complaint that previous philosophers relied on metaphor and physical explanations to a degree that blurred the distinction between philosophy and mythology (99c), the stranger mocks the anthropomorphic explanations of his predecessors who, as if they were telling stories to children, speak of their principles as fighting wars with each other, or falling in love, marrying, and having children (242c-243a). Because their fundamental principles were never clarified and subjected to rigorous scrutiny, the stranger proposes to undertake such scrutiny now, especially with respect to the concepts "being" and "not-being," which appeared so opaque in the previous section.

The stranger focuses first on dualism, which he takes as paradigmatic for pluralism generally. What is the relationship between the dualists' fundamental principles and "being"? Being must be either an additional principle, or identical with one of the original two principles (e.g., the hot and the cold), or with both of them in combination. If it is an additional principle, then there are three principles, and the dualists were wrong about the number of principles. If it is identical with one of them, then the other cannot be said equally to be. If it is identical with the combination, then there really is only one fundamental principle, and the dualists become monists (243d-244a).

Next, using arguments familiar to us from the Parmenides , the stranger goes on to show that monism is no more able than dualism to give an adequate account of being.

1. If being is the same as the One, we end up with two names for the same. Are the two names identical with the One, or different from it? (a) If the name is different from the. One, there will be two beings rather than one. (b) If the name is identical with the One, then the One will be a name, either a name of nothing (since there is nothing besides the One) or a name of a name (i.e., of itself; 244b-d).

2. Is the One the same as the whole (i.e., the world)? (a) If it is the same, then, since a whole has parts, the One will be multiple. (b) If the one being is not the same as the whole, then being will lack something of being. (c) If we deny that the whole exists, then neither coming into


being nor quantity can be said to be (244e-245d). "And so myriad other questions, each involving limitless aporiae, will appear to anyone who says that being is either two kinds or only one" (245d).

B. Form and Matter (245e - 249c)

The stranger then considers the other dichotomy mentioned above, that between the champions of matter and the champions of form, who are assimilated to champions of motion and champions of rest. This opposition between the "friends of the forms" and the materialists is described as gigantomachia , a war against the giants. The reference is to the mythical war in which the giants tried to drag the gods out of heaven. The giants, literally, the "earthborn" (248c), clearly refer to the materialists, while the friends of the forms, like the gods, derive their weapons from the invisible (immaterial) realm above (246b). In Hesiod, the gods defeat the giants. Whether that will be true here in the long run remains to be seen; in the short term there will be no victor.

For the materialists, "being and corporeality are identical." For the others, "true being is certain intelligible and incorporeal forms," while corporeality is not being but "a certain becoming that is in motion" (246a-c). The materialists, says the stranger, are too dogmatic and intolerant to participate in civilized argument, so we need to suppose them to be better than they really are, if any rapprochement is to be possible (246c; repeated at 247c). We may take this as an admission that the following argument would not in fact prove its conclusion to the materialists' satisfaction, although it does so to the stranger's (and perhaps Plato's) satisfaction. The argument is as follows:

1. The existence of living animals implies the existence of souls (246e).

2. The fact that souls can be just or unjust, wise or unwise, implies the possession and presence of these qualities (247a).

3. Such qualities exist without being visible or corporeal (a-b).

4. Therefore the materialists' exclusive materialism collapses (c).

We can see why the materialists would have to be reformed to accept this argument. Step I would not bother them since they conceive souls in material terms, as Theaetetus points out (247b). But the implication in step 2, and assertion in step 3, that the virtues (and perhaps the vices) have an existence that is distinct from that of bodies, would certainly be


unacceptable to them (as indeed it will be to Aristotle). The argument is similar to the second of the three refutations in the Phaedo of epiphenomenalism (the theory that the soul is a harmonia of corporeal elements), which similarly appeals to the reality of the virtues (92e-94b). That argument was not irresistible,[17] nor is this one, as the stranger's need for "reformed" adversaries acknowledges. Nevertheless it perhaps articulates the stranger's own reason for rejecting materialism. It is significant that that reason is based on the reality of virtue, rather than (as we might expect) the reality of universal kinds, for we shall find that the stranger's method of division cannot accommodate distinctions among "values," whether moral or otherwise.

The reformed giants have now acknowledged two kinds of being— corporeality and incorporeal virtue—and so some definition of being must be found to embrace both. The stranger suggests the definition of being as "power" (

), the power either to affect or be affected by something else. The definition is only provisional, however: "Perhaps later something else will occur both to us and to them" (247d-e).

The stranger next seeks to extend this definition to the position of the friends of the forms. They acknowledge that there is a difference between becoming and being. We commune (

) with becoming through the body by means of perception, and with being through the soul by means of reason. The unity of these two, he suggests, like the unity of the two realms of the reformed materialists, lies in the concept of power (248a-b). However, the friends of the forms will object that being is neither active nor passive, and therefore not a kind of power (248c). Power implies change, and for the friends of the forms being is unchangeable. The stranger counters this by insisting that if the forms are known they must be acted upon by our minds, and in this sense they are not at rest but undergo an alteration (248e). This is sometimes seen as a radical departure from earlier conceptions of the forms. It is, to be sure, a different way of talking about them, but it is not a different way of conceiving them. One feature of the earlier presentation of the theory of forms was that what is unchanging may nevertheless be a cause of change—in a sense comparable to Aristotle's "formal" cause rather than "efficient" cause.[18] We may in this sense speak of forms as active, insofar as they are responsible for an attribute that has come into being in a particular case (for example, the beauty of someone who becomes beautiful). And we may speak of them as passive insofar as they are

[17] See PP 103-6; cf. 108-14.

[18] Cf. Phaedo 99a-100e.


objects of knowledge. But in both cases we are speaking only analogically, as we always do when we speak of forms. In the crucial sense forms are not active: they do not produce alteration in other things by means of any kind of alteration in themselves. Nor in the crucial sense are they passive: in being "acted upon" they are not in any sense altered.

What this discussion does illustrate is that we conceive of forms in relation to the realm of change. Forms and changing things are not two radically distinct worlds, or we would be vulnerable to the fifth argument of the Parmenides . Forms are the timeless aspect of changing things, the being of becoming. But in that case they must somehow presuppose the realm of change and motion. This is the direction that the stranger now takes, by means of an argument analogous to the one he used against the materialists:

STRANGER : What then, by Zeus? In truth shall we be so easily persuaded that motion and life and soul and wisdom are not present to perfect reality, and that it neither lives nor thinks, but that august and holy, without mind, it is immovably fixed?

THEAETETUS : That would certainly be a terrifying statement, stranger, to agree with . . . .

STRANGER : But then are we to say that it has mind and life and soul, and yet stands absolutely immovable although it is ensouled?

THEAETETUS : All these things seem to me to be unreasonable. (249a-b)

So there must be motion. But there must also be what is not in motion. The stranger had previously forced the reformed materialists to concede the existence of incorporeal virtues but had said nothing about whether or not these are in motion. To make the latter point, he now focuses not on virtue but on knowledge. On one hand, because knowledge comes and goes in the mind, the existence of mind forced the friends of the forms to concede that reality includes motion. On the other hand, because mind cannot exist without "what is in the same respects and in the same way and in relation to the same thing," and the latter cannot exist without rest (249b-c), the existence of mind also forces the reformed materialists to concede that reality includes rest.

Although in a purely formal way, then, it is the category of power that reconciles the materialists and the friends of the forms, in a concrete way it is the existence of mind that reconciles them. There is thus a double reconciliation: a formal one (the definition of being as "power") and a substantial one (each party's recognition of the reality of its opponent's realm). When the stranger said that the positions of both the materialists and the friends of the forms can be regarded as claiming


that being is a kind of power, he did not thereby show that there was any common ground between the two positions except in words only (the concession that the virtues were incorporeal was not strictly necessary to that reconciliation). He merely effected a collection of the two positions within a common class.[19] What the two camps meant by power might still be very different. But if he can now show that they share a belief in the existence of mind, which requires the reality of both motion and rest, then a reconciliation of substance may be possible.

It is not dear, however, on what grounds the champions of rest are required to accept the "perfect reality" (

) of motile mind, or why the champions of motion must agree that mind requires the unvarying existence described as "what is in the same respects [
] and in the same way [
] and in relation to the same thing [

To take the latter question first: it may seem as though the formula is a reference to the forms, but it is in fact simply an echo of the principle of noncontradiction formulated in Republic 4: "It is dear that the same thing [

] will not do or suffer opposites in the same respect [
] and in relation to the same thing [
] simultaneously [
]" (436b). What the materialists are being asked to concede is that the object of knowledge must have a stable self-identity. But why must the self-identity be stable rather than evolving? In the Republic's formula that kind of stability was not required. The word "simultaneously" meant that we are concerned with the nature of an object only at a given instant; whether it changes from one instant to the next is irrelevant. Here, however, there is no temporal qualification, and the implication is that what is self-identical must be unchanging, that is, at rest. The link between the possibility of knowledge and the need for stable self-identifies has already been noted in the Theaetetus , where Socrates argues that if everything is always in flux and becoming its

[19] Sayre writes that since tangibility and unchangeability are not opposites, "'the battle of the Gods and Giants' is not a conflict between two groups who follow opposite branches in an exhaustive dichotomous subdivision of a common Kind" (PAM 166). However, although they are not opposites in their intention, they imply each other's contradictory, and so are opposites in their extension. In other words, what is tangible is changeable and thus opposed to the unchangeable, and what is unchangeable is intangible and thus opposed to the tangible. Cf. Phaedo 80b: "Soul is most similar to the divine and immortal and intelligible and uniform and indissoluble and what is always related to itself in the same way about the same things [sc . "unchangeable"]; and body [sc . "what is tangible"] is most similar to the human and mortal and multiform and unintelligible and dissoluble and what is never related to itself in the same way about the same things." I suspect that Plato chose nonantonyms to emphasize that the contrast is between dusters of concepts rather than one specific opposition.


opposite, then knowledge must itself always be becoming nonknowledge (182c-e). He inferred that "if everything is in motion, every answer about anything one is asked will be equally right," and language itself will break down (183a-b). Even if the stranger's formula is not meant initially as a reference to the forms, therefore, it must ultimately imply the forms nevertheless. Only the forms fulfill the condition of being at rest in the requisite way. Nevertheless we cannot suppose that the unregenerate materialists would be convinced by this, since the Heracleitean and Protagorean schools, at least, would not be troubled by the relativity of knowledge. But if the materialists were "reformed" enough to accept the independent reality of immaterial virtues, presumably it would not be too much for them to accept the necessity for absolutely stationary objects of mind. The stranger does not claim to have refuted materialism; he has only put forward a position that "more reasonable" materialists ought to accept.

The other question raised above has, of course, more serious implications for "Platonism." If the champions of rest are required to include within the concept of "perfect reality" not only the static forms but also mind, motion, life, and soul, how seriously is the classical theory of forms compromised? There have been approximately four ways of answering this.

(1) The phrase "friends of the forms" is conceivably not a reference to Plato's middle period theory of forms at all, but to some other theory.[20] Such an approach goes back at least as far as Proclus, who construes the phrase as a reference to "the wise men of Italy," that is, the Pythagoreans.[21] But Plato can hardly have been unaware that, in the absence of any other identification, such a phrase would naturally call to mind his own earlier writings. (2) On the assumption that the reference is to proponents of the earlier theory of forms, it may be that this episode reflects a radical revision of Plato's earlier thinking on the subject.[22] (3) We may, on the other hand, accept the assumption that the reference is to the earlier theory, but deny that any change in that theory is indicated. This would be possible if "life," "soul," "motion," and "reason" referred not to these qualities as phenomena but only as

[20] E.g., John Burnet (Greek Philosophy: Thales to Plato [London: MacMillan, 1914]) 280, Taylor 385, Runciman 76. W. D. Ross further cites Schleiermacher, Zeller, Bonitz, Stallbaum, and Campbell (Plato's Theory of Ideas [Oxford: Clarendon Press, 1951] 105 nn. 4-6).

[21] Proclus, p. 106.

[22] E.g., Burner 349; Sayre., PAM 165-66 n. 25; Guthrie 5.141.


forms—that is, the form of life, the form of soul, and so on.[23] However, such an interpretation goes against the whole sense of the passage, which portrays the friends of the forms as making a concession and extending their concept of reality.[24] (4) It may be that a change in the earlier theory of forms is indeed indicated, as the second position claimed, but the change may not affect any of the central doctrines of the theory.[25]

For the reasons given in the preceding paragraph I believe that the evidence is against the first and third interpretations, which are advanced by more recent friends of the forms in the hope of rescuing the earlier theory. Does it need to be rescued? Are the changes required by the stranger of a kind that are crucial to the integrity of the earlier theory? That theory posited a dichotomy between what is apprehended as real by reason and what is perceived as real by the senses: the former is universal and unchanging; the latter, particular and in flux. Since reason is more trustworthy than the senses, therefore, the true reality must be "being" (the universal and unchanging) rather than "becoming" (the particular and fluid). These are the alternatives that are now championed, respectively, by the friends of the forms and the materialists, so if the stranger's reconciliation of the two camps implies the collapse of this dichotomy, then the classical theory of forms would indeed have been modified in one of its most fundamental features.

But matters are not so simple. The primary polarity within the theory of forms was between the intelligible realm of being and the visible realm of becoming. Where do life, soul, and mind belong in this dichotomy? They are not forms, nor are they entirely at rest. Nevertheless, they are known by reason and not by the senses; which means, according to the earlier theory of forms, that they are species of being rather than becoming.[26] Throughout the Phaedo there is a contrast be-

[23] Harold Cherniss, "The Relation of the Timaeus to Plato's Later Dialogues" (1957), in Allen, ed., SPM 352.

[24] See Guthrie's (5.144) reply to Cherniss.

[25] See, e.g., Cornford, PTK 246; Ross 111; Friedländer 3.268-70.

[26] Recall 248a: "Through the body we commune with becoming by means of the senses, and by means of reason through the soul we commune with real being- which always remains the same in the same respects, whereas becoming is different at different times." Guthrie 5.144 observes that "the idea that Plato should ever have allowed the sensible world to cross the bridge between Becoming and Being is contradicted by every other dialogue early or late Here however he says that Being includes not only the Forms Rest and Motion, but also 'whatever is unmoved and whatever is moved', which gives colour to the view . . . that it includes all or some things in the physical world." This latter point will be addressed below.


tween being, characterized as intelligible, unchanging, and eternal, and becoming, characterized as corporeal, changing, and mortal. Even there, in the locus classicus of the theory of forms, soul (and by extension reason and life) is assimilated to being rather than becoming: we are told that soul is eternal because it resembles the invisible and unchangeable, which is eternal—rather than resembling the corporeal and changeable, which is mortal (79b-80b). Accordingly, in the Phaedo and other dialogues of this period, soul, life, and reason quite dearly do not belong to the realm of becoming; they do not quite belong to the intelligible realm either, because they are active. They really act as a kind of intermediary between the two realms,[27] enabling the visible realm to know the intelligible realm (by means of individual souls) and enabling the intelligible realm to be present in the visible (by means of the world soul). Although they are intermediate between being and becoming, they do not constitute a completely independent third type, distinguished equally from both, but are closely assimilated to the realm of being. To move from the Phaedo's position that life, soul, and reason have a kind of quasi being, to the position of the Sophist that they may be included in a broader conception of being, is not a fundamental change. They may be said to have being because they are (as the Phaedo argued) eternal. Even though they may be characterized by "becoming" insofar as they are active, nevertheless they do not change in the way corporeal things change. They are not undergoing generation and destruction. Moreover, because they are not corporeal, they are perceived by reason rather than the senses. The rigorous distinction between the intelligible world and the visible world seems to be as firmly drawn as ever, perhaps even more firmly drawn, since the middle is now more completely assimilated to one of the poles.

Socrates concludes with the words,

For the philosopher who most honors these things, there is, it seems, every necessity on these grounds neither to accept the account of those who say that the universe [

] is at rest, whether as one or many forms, nor should he listen at all to the account of those who set being [
] completely in motion. Rather he must say, following the children's prayer, that "however many things are unmoved and moved," being [
] and the universe [
] are both together.

Here again the two-world distinction is maintained rather than collapsed, because rest and motion are predicated of different subjects: the

[27] Cf. the concept of the daimonic in the Symposium (202d-203a).


philosopher who honors these things is to reject those who say that the universe (literally, the "sum" or "all") is at rest and those who say that being is in motion. It follows then that the (corporeal) universe is in motion but that being is at rest. The final sentence retains this linkage between the universe and motion, and being and rest: "unmoved and moved, being and the universe." The philosopher who honors these things embraces both poles within the more general class of "being" (i.e., "power"), but without collapsing the distinction between them. The reason that there is so much disagreement and confusion among commentators about what is being said in this section is that the stranger uses the same word, "being," both generically (as "power") and specifically (as "unchanging," or the counterpart to "becoming"). The definition of "passive power" (being affected) has just been made broad enough to include "being known"; and now the definition of "being" is broadened to include instances of active power.[28] The position of the friends of the forms (who understand "being" as unchangeability or rest) is collected together with the position of the materialists (who understand "being" as corporeality or motion) both formally and substantively, as I remarked earlier. Formally, they have been collected into the genus "power." Substantively, they have been combined into an extended conception of reality that comprises both "being" and "becoming," with soul (as the principle of both mind and motion) as the middle term that bridges the extremes. But the change from the earlier theory of forms is only terminological. In the middle dialogues Socrates had reserved "being" for what is unchanging and "becoming" for what is changing. He was concerned to distinguish rather than combine them, and so he offered no name for the whole. The stranger chooses instead to call the whole by the name of its primary species, but does not in other respects depart from the earlier ontology.[29] "Being" is now said to include "becoming," but this genus is no longer what the friends of

[28] Thus Charles Kahn takes this sense of "power" to be an instance of the "locative-existential" use of "to be," which "means something like 'to be effectively there, as a physical presence'" (The Verb 'Be' in Ancient Greek [Dordrecht: Reidel, 1973] 376 n. 6).

[29] This is essentially similar, I believe, to what Rosen means by saying that "There are two distinct ontologies at work in the Sophist.. . The doctrine of forms, so to speak, is the Stranger's version of a precise speech about being. The doctrine of the divine ousia is his version of imprecise speech" (PS 223; emphasis in original). Rosen goes on to say that "the Stranger . . . does nothing to reconcile these two ontologies" (ibid.). If by this Rosen means that the stranger does not collapse the distinction between them, then I would agree. But the stranger does reconcile the two in another sense, i.e., by comprehending the "precise" one, together with its counterpart (materialism), within the "imprecise" one, both formally and substantively. The greater logical generality of the imprecise one does not, however, confer upon it ontological primacy.


the forms meant by "being"; hence their principle has not been repudiated but has only been subsumed under a more inclusive class. The principle of the friends of the forms continues to retain its integrity, and continues to be referred to as "being" or "rest." As Socrates will say in the Philebus (12e-13a), the fact that things may be classed within the same species does not prevent them from being fundamentally opposed to each other in nature.

The preceding discussion has suggested that the problem of being can be solved by defining it as "power." But the stranger remarks that this solution may be undermined in the same way that dualism was undermined earlier. There the problem was in the relationship of being to the opposed principles of the dualism. Here we again have the opposed principles of motion and rest, and consequently the problem of coordinating them with being. If being is common to both rest and motion, it must be different from either one. But in that ease it would be neither at rest nor in motion, which is impossible. As the stranger predicted at the beginning of this section, the concept of being has collapsed into aporia just as had the concept of not-being. He now suggests that the discovery of the true nature of either one of them may entail the discovery of that of the other (250e-251a). This symmetry between them will turn out to be significant.

6. Combining of the Forms (251a-259e)

The aporia about being was how it is possible to call it by more than one name—rest and motion, for example. This difficulty may be no more puzzling, the stranger suggests, than the fact that we attribute to a person colors, shapes, sizes, vices, and virtues. Thus, "we say that a person is not only a person but also good, and endless other things" (251b). Only the young, and those who learn late in life, would say that what is one cannot be many in this way, and "would not allow us to call a person good, but only the good good, and a person a person" (251b-c). The question, then, is whether such things as "being" and "rest" can combine as do "person" and "good"—more precisely, whether all such things combine, or only some of them, or none at all (251d).

This may be the first time that Plato thematically explores the combinability of forms, but that does not mean—as is often inferred—that in the middle period the forms are conceived as absolutely discrete and incapable of combination. On the contrary, there would have been no point in Socrates' insistence that opposite forms cannot combine, as he


insists in the Phaedo (102d-105e), unless it were assumed that non-opposite forms are capable of combining. In fact it is clear from the same passage that in some sense the form of three combines with that of oddness, the form of fire with that of heat, and the form of snow with that of cold. Similarly, in the Republic it is dear that not only do the just and beautiful combine with the form of the good (505a-506a), but so in some sense do all other forms (508e-509b). The Sophist may be the first dialogue to explore the combinability of forms in detail, for it is the first one that examines the formal basis of language (and not just names, as in the Cratylus ), but in so doing it expands on what Plato has said elsewhere, rather than contradicting it.

The stranger addresses his question by pointing out that if none of these things combine, then motion and rest will not share in being and will not be. In fact, if different things cannot combine, then atomism is ruled out as well, since the atoms will not be able to join together (252b—although presumably the atomists would reject this assimilation of physical combination to formal interaction). More important, the "late learners" can thus be seen to contradict themselves, for they cannot express themselves without employing all kinds of verbal attribution, and this makes a mockery of their claims. On the other hand, if all forms combined, then so would opposites, and we would be left with contradictions, such as that motion is at rest, and rest in motion (252c-d). The remaining possibility, that some forms combine and some do not, is likened to the way letters combine to form syllables, or the way high and low sounds combine to form music. It is for the art of dialectics-the free person's art, philosophy—to know which forms combine and which do not, just as that of grammar (

) knows which letters combine, and that of music knows which sounds combine (253a-b).

In music there are no specific elements (notes) that account for the ability or inability of the notes to combine. That ability has a formal rather than a material explanation. Sequences of sounds that manifest beauty or harmonia may be said to combine, and those that do not manifest it do not combine. The example thus suggests the possibility that the question of combinability may ultimately rest with forms of value , not merely forms of kind. In the other example, grammar, the combination of letters into syllables is made possible by vowels (although nothing is said of the combination of letters or syllables into meaningful words). It is possible to see the three examples as forming a progression. In music there are specific elements neither for combining


nor for separating the other elements. In grammar there are elements that enable combination, but no elements that account for uncombinability. In dialectics there will be both.

The question, then, is whether there are some forms that make possible the combining of forms, and others that are responsible for their separation (252e-253c). There follows one of the most puzzling passages to be found anywhere in Plato:

STRANGER : To divide according to kinds, without thinking the same form to be different or a different one the same, shall we not say that this belongs to the science of dialectic?

THEAETETUS : Yes, we shall.

STRANGER : Will not he who is able to do this sufficiently perceive (1) one Idea extended everywhere through many, each one of which lies apart; (2) many Ideas different from one another, embraced from without by one; (3) and again one through many wholes brought together into unity; (4) and many forms apart from each and separate? (253d)

If only the first two were given, they would naturally seem to be examples of the two forms we are seeking: the first, that which combines ("one idea extended through many"); the second, that which separates ("different . . . embraced from without"). But the third and fourth categories apparently repeat this dichotomy with slight variations. The term "wholes," in 3, seems to be equivalent to "forms" in 4, in which case it refers not to individuals but to species: a species is already a combination of all the lower species into which it can be further divided, and therefore a whole of parts. In that case, "wholes" in 3 and "forms" in 4 seem to mean the same thing as "Ideas" in I and 2, and it is hard to see what the difference is between the first pair and the second.

The passage is so laconic that any interpretation runs the risk of being arbitrary. According to Julius Stenzel, writing in 1917, the most common interpretation was

that in (2) Plato is describing the inclusion of subordinate classes within a higher one, and in (4) contrary classes, mutually exclusive to each other, such as Rest and Motion. But Maier rightly draws attention to the point that (2) and (4) partly coincide: "Sheer disparity between them," he says, "is unthinkable." He is entirely in the right.

Stenzel's own interpretation is:[30]

Where he is concerned with the Division of wholes or unities, he says

. From the point of view of sense, the important thing is not that the

[30] Plato's Method of Dialectic (New York: Russell and Russell, 1964 [orig. 1940]) 97, 100-101.


Division extends through "all," but that it passes through wholes or unities. (2) Conversely, where the process of Collection is described, as it was in the preceding phrase, the essential thing is to include all the kinds (

) under certain higher ones . . . . On our view, then,
[wholes] takes the place of the plural of
[one] . . . . (3) and (4) are closely related, and describe the process of obtaining the definition,
, of a single form,

Cornford, basing his own interpretation on that of Stenzel, argued that the first pair refers to the genus and species in the act of collection, and the second pair refers to them in the act of division.[31] This has been the prevalent interpretation in our century. It has the virtue of bringing the passage into a clear relationship with the overall theme of division in the Sophist ; although against this must be balanced the fact that there is no discussion of collection in the Sophist . The interpretation has the further disadvantage of requiring that the pairs represent movement in opposite directions—the first pair upward and the second pair downward—and the text gives no clear encouragement to such a reading. Not surprisingly, in spite of the general prevalence of Cornford's interpretation, a daunting variety of alternatives to it have been proposed. I shall list (without comment) four more of them to give some idea of the diversity of possible interpretations.

a. G. E. M. Anscombe: "(i) Species-individual [individual forms], (ii) genus-species [unifying forms such as "one," "whole," "being"], (iii) transcendentals-other forms, (iv) perfectly specific forms considered in themselves [such as the series of natural numbers]."[32]

b. Alfonso Gomez-Lobo: "Soph . 253d1-e2 does not describe Division, it anticipates the comparison of Being and Not-Being with other Forms which will ultimately provide Plato's answer to the dilemma of Parmenides."[33]

c. Jacob Klein: "Let us understand what the Stranger says, by means of examples. Example a : 'mammal' extends every way through 'lion,' 'camel,' 'dog.' . . . Example b : 'Being' embraces, 'from the outside,' Change and Rest . . . . Example c : 'animal is assembled into unity through 'mammal,' 'fish,' and 'bird.' Example d : 'justice,' 'cloud,' and 'fish' are entirely apart and separate."[34]

[31] PTK 267. He has been followed by Sayre (PAM 178-79), Guthrie 5.131, and Teloh 190, among others.

[32] "The New Theory of Forms," The Monist 50 (1966) 403-20 at 419-20.

[33] "Plato's Description of Dialection in the Sophist 253d1-e2," Phronesis 22 (1977) 29-47 at 47.

[34] PT 52.


d. Seth Bernadete: "Every idea of type A1 [ = 1] is manifest as one idea among many other ideai , all of which are comprehended by another idea of type A2 [ = 2]. And every idea of type B1 [ = 3] is manifest in manifold ideal of type B2 [ = 4], which are not comprehended by another idea ."[35]

Faced with such a bewildering variety of readings, the most prudent course might be to suspend judgment.[36] Nevertheless, I prefer to offer a different interpretation, recognizing at the same time that there is insufficient evidence for maintaining it (or any other) against its rivals— especially Cornford's—with much confidence. It seems to me that the simplest explanation would be to take the passage as describing a pair of divisions. Thus I ("one Idea extended through many, each one of which lies apart") refers to the starting point of division, and 2 ("many Ideas different from one another, embraced from without by one") to the results of that division. The opening words of 3 ("and again , one through many wholes brought together into unity") may be intended to suggest that we do not stop after one division but continue further to subdivide the preceding species. In step 3 we recognize that each resultant species is a whole with further subordinate parts brought together within it. Finally, in 4 ("many forms apart from each and separate") we distinguish the subspecies within the previous species. The point of listing four steps instead of only two would thus be to illustrate that the combining and differentiating functions that underlie diaeresis operate at more than one level.

On that interpretation, the present passage would be the counterpart of an equally obscure passage of the Theaetetus , where the aviary is said to contain "all kinds of birds, some in flocks apart from the others, others in small groups, and some alone flying randomly through them all" (197d). Although the Theaetetus makes no explicit mention of the methods of division and collection, they had already been adumbrated in the Phaedrus , and the Theaetetus passage may be intended as an illustration of collection, an illustration that is now comprehended in the Sophist by an illustration of division. On that interpretation, the single birds represent specific or individual cognitions that have not yet been brought into relation with others. The small groups represent col-

[35] 2.145.

[36] Rosen forthrightly says, "I prefer not to read my speculations into the text, because I see no secure basis for a detailed positive account of what is here meant by 'dialectic'" (PS 261).


lections taken only to the first level—the first species or genus. And the flocks represent the possibility of bringing the results of our first collections into progressively more inclusive ones.

However that may be, the question at hand concerns the combinability and divisibility of the forms. We have already seen that rest and motion do not combine with each other because they are opposites, but that both combine with being (250a-d). This means that the three are distinct: all are different from each other and each the same as itself (254d). What then is the status of "same" and "different"? Are they additional forms, or are they somehow reducible to the other three?

STRANGER : But surely motion and rest are neither different nor the same.


STRANGER : Because whatever we would call motion and rest together cannot be either of those two.


STRANGER : Because motion will be at rest and rest will be in motion; for with regard to both of them, whichever one becomes the different would force the one that is different to change to the opposite of its nature, since it would participate in its opposite.


STRANGER : Both surely participate in the same and the different.


STRANGER : Then let us not say that motion is the same or the different, or that rest is. (255a-b)

In other words, if motion were equated with the different, then rest, which is different from motion and therefore participates in the different, would participate in motion (which ex hypothesi is the same as the different). Thus rest would be in motion. On the other hand, if we identified rest with the different, then, by the same reasoning, motion would be at rest. Same and different cannot, therefore, be reduced to motion and rest.

Nor can they be reduced to "being." If "same" were equivalent to "being," then (1) "motion and rest both are " would mean (2) "motion and rest are both the same." "Then it is impossible," the stranger concludes, "for the same and being to be one." Theaetetus replies only, "Virtually so" (

; 255b-c). His reply to the stranger's next question-whether they shall therefore consider the same to be a fourth form—is much more emphatic: "Absolutely" (
). But still we may wonder whether there is something to be learned from his hesitation in the preceding exchange. Does sentence 2 really imply a para-


dox if interpreted in the same terms as sentence 1? Sentence 2 is in fact ambiguous. It may mean " . . . the same as themselves" (as it does half a page later, 256a), in which case there is no paradox. Or it may mean (as the stranger takes it to mean) " . . . the same as each other." In fact it is only in the first (innocuous) sense that sentence 2 follows from sentence 1. Sentence 1, in saying "both are " does not mean that both are each other ; it is existential rather than attributive, and therefore intransitive rather than transitive. But the stranger encourages Theaetetus to interpret the second (putatively isomorphic) sentence as being transitive . The two sentences must, however, be interpreted in parallel ways for the inference to be cogent. If sentence I speaks of same and different not in relation to each other but only in relation to themselves, then this is the only legitimate way to understand sentence 2: motion and rest are the same as themselves . With this argument, unlike the preceding one, the reductio fails. But it need not have failed. The stranger could have used an argument similar to the preceding one. He could have argued that if we equated one of the pair same-different with "being," then, if the other member of the pair is , it becomes its own opposite. That is, if "same" were equivalent to "being," then when we said that the different exists, we would be saying that the different is the same; and similarly if different were equivalent to being, we would be forced to conclude that the same is different. Why did the stranger abandon this successful kind of reductio in favor of a questionable one? Theaetetus's initial hesitation suggests that Plato may not have been unaware of its weakness. By reflecting on Theaetetus's hesitation, we saw that "being" and "the same" are in one way very dose in meaning. We will shortly find that one sense of "nonbeing" is "the different," so it should not surprise us to find that one important sense of "being" is "the same" (

).[37] In the Battle against the Giants section, we saw

[37] Rosen observes that to "replace 'sameness' by 'identity' and explain that, in turn, as one sense of 'is' . . . destroys the distinction between the two forms being and sameness " (PS 272). Nevertheless, the fact that "difference" is a sense of "nonbeing" does encourage such an explanation. It may well be that the distinction between "being" and "sameness" is only one of quasi genus and species. (Quasi genus because "being" cannot be a true genus: the stranger's dialectic against the One [244b-245d] implies that if being were a true genus, the differentia, which must come from outside the genus of "being," would by definition be "nonexistent.") In fact we can regard all four of the other "greatest kinds" as species of "being": "sameness," "motion," "rest," and even "not being" qua "difference." "Rest" and "motion" are species of "being" if one thinks of them in terms of the positions of the friends of the forms and the materialists, respectively: "being as opposed to becoming" (i.e., "rest") and "becoming" (i.e., "motion") are species of the strangers "being as power." Again, since "not being" qua "difference" exists as negativity, it too is a way or species of "being."


that the friends of the forms are never asked to give up their belief that reality is unchanging. They are asked to accept the reality of what is in motion, but not the reality of what is in a state of essential becoming. Life, soul, and reason are essentially connected with motion, but, unlike visible things, they never become different from themselves. A living body becomes a nonliving body, solid stone becomes sand, fire extinguishes, water evaporates; but life itself never becomes nonlife, soul never becomes nonsoul, reason never becomes unreason. In the highest sense, to be is to be selfsame or "in oneself" (

). The stranger will in fact use this phrase two speeches later. Perhaps, then, this questionable reductio is meant to remind us of the continuous parallelism between the dialogue's treatment of being and not-being. The themes that we have been led to consider in reflecting on it will, in any case, become important subsequently.

Finally, the remaining possibility is eliminated. The different cannot be equated with being, because some things have being in themselves (

), and nothing can be different in itself, but only in relation to something else (255c-d). Thus we have not three but five forms or kinds—being, rest, motion, sameness, difference—which blend with each other in various ways, and which the stranger designates as the "greatest kinds."[38]

In discussing the sense in which these forms do and do not combine with each other, the stranger implicitly distinguishes between the "is" of sameness and the "is" of connection. This distinction is more familiar to us as the distinction between identity and predication. However, if we speak in this way it is important to bear in mind that the grammatical relations of identity and predication are only analogues of ontological relationships in the stranger's account.[39]

[39] Cf. Paul Seligman, Being and Not-Being: An Introduction to Plato's Sophist (The Hague: Nijhoff, 1974) 72. We should also keep in mind Kahn's important observation that the distinctions among senses of "to be" do not imply an ultimate lack of unity in the concept of "being" generally: "The verb has a number of distinct uses or meanings that are all systematically related to one fundamental use" (VBAG 401). Also see his article "The Greek Verb 'to Be' and the Concept of Being," Foundations of Language 2 (1966) 245-65.


We have just seen that in one sense "to be" means to be the same. In that sense, when we say x is F we can only mean that x is the same as F. This was the only sense admitted by the "late learners"; but the stranger has now expanded the meaning of "is" from that restrictive sense in such a way as to include predication. That distinction is brought to bear now in a series of conclusions drawn from the foregoing discussion:

1. Motion is not in any sense rest It is neither the same as rest (255e), nor does it participate in rest (256b).

2. Motion both is and is not the same. It is not the same in the sense that it is different from (not identical with) "the same," but it is the same insofar as it is the same as itself (it combines with sameness with respect to itself; 256a-b). Keeping in mind the caveat on the preceding page, we can make the distinction clearer by saying that it is not identical with sameness, but sameness may be predicated of it.

3. Motion is different from the different, so in one sense (identity) it is not "different," but in another sense it is different (it combines with difference with respect to other things) (256b-c). As With the previous case, we might say here that motion is not different if we are using the "is" of identity, but is different if we are using the "is" of predication.

4. Even though motion is , in the sense that it combines with being (256a), it nevertheless is not , insofar as it is different from being (250b, 256d).

In the same manner as 4, we may say that each of the four greatest kinds other than "being" both is and is not. Accordingly, "'being' is many, while 'not-being' is unlimited [

] in quantity" (256e). Just as there is only one way to be right, and an unlimited number of ways to be wrong, the number of things that a thing is not, is incomparably greater than the number of things that it is.

The stranger's implicit distinction of a connective sense of "being" is now transferred to "not-being." "When we say 'not-being,' it seems, we do not speak of the opposite of being, but only of what is different" (257b). In other words, what is not-F need not mean what does not


exist but only what is different from F. The different is entirely relational. It is analogous to knowledge, which is divided into various technai and epistêmai in accordance with its objects. In a similar way the different is particularized and defined by the object from which it distinguishes itself (257c-258c). Of the three examples that are given— beautiful/not beautiful, large/not large, and just/not just—the first and third are once again examples of value. Even large (

) has a quasi-valual import, especially in the present context of the greatest (
)—that is, most important—kinds.

This conclusion, that not-being exists qua difference, formally contradicts Parmenides' prohibition against saying or thinking that not-being exists (258d). But it does not contradict the spirit of that prohibition, because "we long ago said goodbye to any talk about an opposite of being, whether it exists or not, and whether it can be spoken of or is absolutely incapable of being spoken" (258e). To say that not-being exists qua difference is not to say that it exists qua the opposite of existence .

We saw that the stranger has maintained a parallel between the nature of being and the nature of not-being. The present discussion has distinguished two senses of being and two senses of not-being. (a) The two senses of "being" are "is the same as" (even difference is "the same as" itself ) and "combines with" (even not-being, i.e., difference, "is" a kind of being). The primary sense has been combinability, since identity has been distinguished as the additional kind, sameness; but the latter sense has hovered in the background as the legacy of the late learners. (b) The two senses of "not-being" are "difference" and "nonexistence." The primary sense has been "difference," but the sense of "nonexistence" has been present in the background as the legacy of the sophist, who, as the stranger earlier remarked, has eluded capture by hiding himself in apparently unintelligible not-being (239c). Are the two pairs parallel? In fact they are not, as we can see from the following chart:













Sameness and difference. are counterparts, as are being and not-being, but combinability and nonexistence are not. The counterpart of combinability is not nonexistence but uncombinability, and the counterpart


of nonexistence is not combinability but existence. The lack of symmetry is disconcerting because at 253c the stranger spoke of the importance of discovering those forms that make combination possible and those that make separation possible. "Being" qua combinability certainly answers to the first of these, but we have not been shown the second. Commentators often assume that the form responsible for uncombinability is "difference,"[40] but to say that forms are different is not to say that they do not combine. "Rest" and "being" are different, but they combine. Otherwise combinability would be synonymous with sameness, which is just what the late learners claimed, and what the stranger is at pains to deny.

The example that we are given of uncombinability is "motion" and "rest" (252d). This suggests that the form—or at least one form— responsible for uncombinability is "opposition," a conclusion that accords with the Phaedo (102d-103c) and that is suggested by the stranger's frequent use of this word. We might go on to make "opposition" a species of "difference," as "difference" is a species of "nonbeing," and in that case it would be true to say that "difference" (in one of its species) is responsible for the inability of some forms to combine. But that goes beyond anything that is explicit in the stranger's presentation. It would seem more natural to make "opposition" another species altogether. Thus, one species of "nonbeing" would be "difference," that is, simple diversity, while another would be "opposition."

This would give us "opposition" as the counterpart to "combinability," but we still have no counterpart for the third species of nonbeing, "nonexistence." For this we need to recall the stranger's earlier collection of the senses of "being," proposed by the materialists and by the friends of the forms, into the form "power." "Power" appears to be the counterpart to "nonexistence." This would explain, too, why the stranger treats "power" as the most ultimate form, from which all but one (division D) of his divisions begin. The equation of existence with power, like that of uncombinability with opposition, accords with Plato's earlier dialogues. When Cebes, for example, asks Socrates to show that the soul exists after death, he asks for proof that it has power and intelligence (Phaedo 70b). In that context "power" seems to mean something like "existence." Again, in the Republic the exalted status of the good is emphasized by saying that it is "beyond

in dignity and power "


(6.509b). In that case, we can complete the chart above' and vindicate the stranger's claim of parallelism between being and nonbeing:













We can see from this that Plato has not abandoned his earlier, metaphysical, conception of being in favor of a purely logical or grammatical one, as is often argued. The metaphysical concerns are relegated to the background of the Sophist , for reasons that have already been suggested and that I shall return to later, but we are reminded of them nevertheless, however briefly and however indirectly.[41] The fact that the discussion of being as "power" is not explicitly followed up in the present passage does not mean that we are entitled to disregard it when we consider the implications of this section as a whole, as commentators sometimes assume.[42] When we draw inferences from the stranger's deliberate parallelism between the problematic of being and the problematic of nonbeing, it is important that we take into consideration all aspects of that parallelism. It is true that the stranger focuses primarily on the concept of not-being as relational ("difference"), but he makes it quite dear that the relational sense is not the only sense of "not-being." "Not-being" can also mean "nonexistence" (

, 237b). Consequently "being" also means "existence," or "power."[43]

[42] Thus Ross concludes that "the supposition that knowing is an action and being known a passivity, which is only one of several suggestions put forward in 248d4-7, is simply dropped . . . . It is silently dropped when the consequence of accepting it has been pointed out" (pp. 110-11). Ross is far from being the only one to believe that the previous discussion can be safely ignored—although usually the question is not even addressed.


It is worth noting that the two concepts that are combined in the above passage from the Republic , dignity and power, are both used by the stranger as well, but only one is endorsed by him. He appears to be as indifferent to the concept of dignity as he is partial to that of power. His method, he tells us, abstracts from all questions of whether something is more or less dignified,[44] or whether it is more or less ridiculous (227b). Curiously, however, despite his apparent indifference to distinctions of value, when the stranger now recapitulates his dissatisfaction with the late learners, he does so precisely in terms of the language of values: the ability to ignore the late learners' quibbles and make the necessary distinctions is "both difficult and beautiful" (

; 259c). On the other hand: "My good man [
], the attempt to separate everything from everything rise not only is not melodious [
] but indeed belongs to someone completely unmusical [
] and unphilosophical," because it obliterates the possibility of logos (259d-260a). His choice of words reminds us that differences of kind (those that admit blending, those that do not) are sometimes only fully intelligible when expressed as differences of better and worse.

7. Logos And the Forms (260a-264c)

The Theaetetus , which made no explicit use of forms, or kinds, or any universal concepts, foundered in an attempt to exhibit the nature of logos. Here such an exhibition turns out to depend on the discussion of the forms just completed. The purpose of forcing the late learners "to allow different things to combine with each other . . . was to show that discourse [

] is for us one of the kinds of being" (260a)—for discourse (which combines subjects with predicates) would be impossible without combination. If discourse is to be meaningful, however, it must also be possible for some statements to be false. So the question to be considered is whether the form "not-being" combines with the forms "opinion" and "discourse." If it does, then the possibility of false statement, and therefore of discourse, will be vindicated. Moreover, since images, likenesses, and semblances were conceived—in relation to the sophist and philosopher—as possibilities of "false" discourse (i.e., of saying what is not), the paradoxes surrounding them would dissolve as well (260c).

The point that was previously made about letters and about the greatest kinds is now extended to names: some are capable of com-


bining, and some are not. The test is whether or not the resultant sequence is a meaningful one, and it will not be so unless nouns are combined with verbs, as in the minimal sentence "Man understands" (261a-262c).[45]

Every sentence must be about something and must have a quality (262e).[46] In the case of the two sentences "Theaetetus sits" and "Theaetetus, with whom I am now conversing, flies," the subject is the same but the truth quality is different. The first states the facts (

), while the other states something different from the facts—it states what is not as if it were. In such cases "there really and truly comes about a false logos" (263a-d). It has often been observed that the second example seems intended to combine contingent falsity with necessary falsity. It is contingently false because, since we are told that Theaetetus happens to be sitting, we know that he cannot be doing anything that is incompatible with sitting, such as flying.[47] It is necessarily false because Theaetetus is a human being, and being human is incompatible with being able to fly. The ability to fly requires wings, and to be winged belongs (oddly, to be sure) only to species within the class of water animals (220a-b). Human beings, on the other hand, are a species of land animal (222b-c).[48]

[46] The fact that in the following examples this quality will be truth or falsity does not mean that Plato failed to recognize that some sentences are neither true nor false. Questions and commands may be said to have a quality also, but of a different kind; there is no point in discussing them here, since the entire discussion is subservient to the question of whether false logos is possible. Runciman is therefore precipitous in speaking of Plato's "apparent belief that every combination of noun and verb must be a statement and as such must be either true or false" (p. 109).

[47] Commentators have wondered whether "sitting in an airplane" is an interesting refutation of the stranger's claim. Cf. Sayre: "It does not go without saying, as Xenakis avows, 'that "sitting" means here "sitting on the ground " not, say, "in an airplane"'"(PAM 208 n. 85; emphases in originals). Certainly the stranger intends sitting and flying to be mutually exclusive (otherwise the conjunction of the two sentences loses much of its force). Such an intention is entirely defensible, for sitting is normally construed as a passive state (being supported), while flying is an active state. Strictly speaking, when we are in an airplane, it is not we who fly but the airplane; and when the airplane is not flying we say that it "sits" on the ground. Actively flying, as distinct from being carried by a vehicle that is flying, is incompatible with sitting. It is parallel to the earlier opposition between motion and rest.


How do these examples demonstrate that meaningful discourse is possible because of the combining of forms? In the second case the answer is clear enough. Flying cannot combine with sitting (or with being human), and so "Theaetetus flies" is contingently (and necessarily) false. But the first sentence appears to contain only one form, "sits," and a proper name, "Theaetetus." Why did the stranger choose this as an example of a true sentence, instead of his earlier example, "Man understands," in which a combination of forms may be discerned? It is tempting to try to minimize the tension between the example (an individual participating in a form) and the thesis it is meant to illustrate (discourse as rooted in the combinability of forms) by taking only one of the two literally and "making allowances" for the other.[49] But it would be better not to have to make allowances at all.

Assuming, for the moment, that Plato both stated and illustrated his thesis competently, we need to ask how "Theaetetus sits" can involve a combination of forms. "Theaetetus flies" may be false either because "flying" does not combine with "sitting," or because it does not combine with "being human"; but in the case of a true sentence matters are not so simple. It is important to bear in mind, however, that the combination of forms is meant to explain initially not how sentences come to be true or false, but how they come to be meaningful (259e). To be meaningful they must be capable of truth and falsity, but to explain how a sentence is able to be true or false is not necessarily to explain whether it is in fact true or false. We saw earlier that meaningfulness requires both that some things combine and that some things be unable to combine. If nothing combined, then predication would be impossible, as the late learners contend; but if everything combined, then opposites could be predicated of each other, which would destroy meaning (251d-252e). If, now, we take "sitting" and "flying" to be in opposition to

[49] Cornford gives the example precedence over the thesis when he writes, "All discourse depends on the 'weaving together of Forms' . . . . We can also make statements about individual things. But it is true that every such statement must contain at least one Form" (PTK 300). The last sentence twists the first one beyond recognition. A single form cannot illustrate a "weaving together of forms." Ross, on the other hand, gives the thesis precedence over the example. He states that the two present sentences ("Theaetetus sits," "Theaetetus flies") "do not illustrate Plato's thesis," and dearly regards this as a lapse on Plato's part (p. 115). See Guthrie 5.161-62 for a discussion of some other approaches to this issue. John Ackrill, "S YMpL OKH EIDW N" (in Vlastos, ed., Plato 1.201-9), gives a perceptive analysis of this passage.


each other, then the examples do indeed illustrate thee stranger's thesis. Their opposition can be expressed by interpreting "flying" as overcoming the pull of gravity (or, in classical terms, overcoming the downward tendency of the earth in us) and "sitting" as a suspension of struggle against the pull of gravity (cf. n. 47, above). Therefore, if one of these can be truly predicated of Theaetetus, the other will be excluded. Understood in this way, "Theaetetus sits" is meaningful because some forms combine and some do not. It is meaningful because its truth entails the falsity of statements whose predicates are forms that do not combine with "sitting," such as "flying." Presumably the stranger chose "flying," rather than "standing," "walking," or "lying down," to illustrate his thesis because it implies necessary falsity as well as contingent falsity. But in other respects any of these other examples would have done as well.

The status of logos or discourse has thus been legitimated: it has been shown to be meaningful and thus to combine with being. Since thinking (

) is nothing but the soul's silent discourse (
) with itself, it too has being (263e). Corresponding to affirmation and denial in discourse is "opinion" in thinking; therefore opinion has being as well. And when opinion is brought about by means of sense perception, the result is "semblance" (
, 264a). Here "semblance" is used synechdochally to stand for the whole class of images—whether true (likenesses) or false (semblances)—as "likeness" had been used previously (240b). Now that we have seen how semblance can after all have being, we are ready to resume the hunt for the sophist that had led us to that class earlier.

8. Resumption Of the Final Division (264c-268d)

The final division, as it is now recalled and completed, takes the form shown in Figure 9.

No longer is the initial distinction an opposition between speaking about and producing all things, as it was prior to the digression. Now, as in the preliminary divisions, it is between acquisitive and productive arts. It is no longer even conceivable that the division might fall within the form of acquisitiveness. In addition, the reference to selling has been removed, so all question of selfish motivation is now eliminated. Why should Plato have taken so much trouble to establish the sophist as acquisitive in all the preliminary divisions, only to end up looking for him in the form of productive arts? The stranger explains:



Figure 9


STRANGER : Did we not begin by dividing the productive and acquisitive arts from each other?


STRANGER : And under acquisitive art the sophist showed himself to us in the arts of hunting, competition, commerce, and other such forms.

THEAETETUS : Absolutely.

STRANGER: But now, since the mimetic art has encompassed him, it is dear that we must first of all divide the productive art itself in two. (265a)

This is not sufficient grounds for removing the sophist from the acquisitive form, since division B (the commercial sophist) showed us that the productive art can reemerge within the acquisitive arts (step 5a) when the production is ultimately for the sake of gain, as it was for the sophists. If the sophist makes his products not because he sees such creativity as an intrinsic good, but rather because they are instrumental to profit,[50] then he belongs in the acquisitive class. He becomes an example of the sophist as a merchant who sells his own products (224d-e). There is thus no immediately clear reason why the stranger shifts the inquiry out of the sector within which all the preliminary divisions led us to believe the sophist is to be found. It is not a question we would need to worry about if the ensuing division led to a satisfactory definition of the sophist; but it doés not. The sophist is defined as (1) the maker of semblances rather than likenesses (step 4, 266d-e), (2) operating on the basis of opinion rather than knowledge of his subject (step 6, 267b-e), and (3) making his claims with irony (step 7, 268a). This definition gives rise to a number of serious questions.

Since the chief burden of the dialogue is to distinguish the sophist from the philosopher, how successfully can we differentiate them in terms of the above definition? If we look to the first of those three distinctions, the difference will be that the models (images) used by the philosopher will always be accurate (likenesses), and those Of the sophist inaccurate (semblances). But the difference between philosophy and sophistry is not that the former is always true and the latter always false. In that case a philosopher whose conceptions were not perfectly accurate would be nothing but a sophist; and a sophist who happened to give an accurate description of some state of affairs, when it suited his

[50] In the Republic Socrates shows that Thrasymachus's unjust person is unjust not by virtue of his knowledge of the principles of ruling, but because he uses that knowledge in the service of moneymaking (345c-346e).


purpose, would be a philosopher, however manipulative his intentions might be. The stranger's point would be more defensible if he meant that for the purposes of a sophist semblances are adequate , whereas for the purposes of a philosopher likenesses are necessary. But to introduce the notion of purpose, whether explicitly or implicitly, is to show that what really distinguishes the philosopher and the sophist are their values and goals, to which their "products" (likenesses or semblances) are only a means. In this way too the present definition misses what is essential. We shall return to this point.

If we try the second distinction, between knowledge and opinion, there will be an analogous distortion. The philosopher always knows what he is talking about, while the sophist has only an opinion. In that case, once again, there could be no such thing as a mistaken or even partially ignorant philosopher; he would be a sophist by definition. The Socrates of the aporetic dialogues, who frequently professes his ignorance (as he does again in the Theaetetus ) and even the Socrates of the Republic , who claims to have only opinion rather than knowledge of the good (506b-e), would be a sophist by definition. And any sophist who happened to have attained knowledge about his subject matter would be a philosopher, regardless of the use to which he put that knowledge. Neither of these criteria, whether taken singly or jointly, provides an adequate account of the essential difference between philosophy and sophistry.

The third distinction, in terms of irony, sounds more promising. To say that the sophist is ironic may seem to suggest that he is insincere and disguises his true intentions, in which case he might be using argument as a means to mercenary or political ends. But it turns out that this is not what the stranger means at all, for those who are characterized by irony are said to "have a great suspicion and fear that they are ignorant of the things that they give themselves the appearance of knowing in front of others." Their irony is thus a kind of modesty, remarkably like Socratic irony. The species with which they are contrasted, the counterpart of irony, is the simplicity (

) of people who are "simpleminded [
], believing that they know things about which they only have opinion" (268a). The distinction, then, is between those who fear that they really know less than they seem to, and those who seem to themselves to know more than they really do, that is, between modesty and conceit. By definition, neither of them has knowledge, since they are the two species of "opinion" (step 6b). In both cases the appearance of knowledge falls short of the reality, and the only issue be-


tween them is whether they suspect and worry about this shortcoming or not. It is rather surprising that Plato places the sophists in the class of those who do have this suspicion and worry. For throughout the dialogues Plato portrays the sophists not only as makers of semblances, but as pleased with their calling, and as denying any ultimate distinction between knowledge and opinion—or, by extension, between likeness and semblance.[51] They are never depicted as suspecting, much less fearing, that they have only opinions when they ought to have knowledge. By endowing them with what is, in effect, Socratic irony (which in this case may be represented by the Socratic awareness of "knowing only that one does not know"), Plato prevents the present dichotomy, like the other two, from distinguishing between the sophist and Socrates, or any philosophers who are aware of their own limitations.[52]

However we interpret this species of irony, it must ultimately be the home of the philosopher as well as the sophist. At 268b the stranger asks whether the type of person that is isolated at step 8b is the wise man (

) or the sophist (
). Theaetetus replies that he must be the sophist, "since we posited him as not knowing" (step 6). So in the final analysis what the dialogue distinguishes sophistry from is not philosophy (
) after all, but rather knowledge and wisdom (
).[53] The Symposium insists that philosophy, "love of wisdom," must be distinguished from wisdom itself; as the seeking of wisdom it can only be in between wisdom and ignorance (203e-204b). This distinction is implicit throughout the dialogues,[54] and Plato has given us no reason to believe that it ought to be collapsed here. Given the stark opposition between knowledge and opinion in step 6, not only must we put the wise man on the side of knowledge, as Theaetetus does, but to distinguish between wisdom and philosophy we must put the philosopher on the side of opinion. Given the present alternatives, we cannot keep the philosopher distinct from both the wise man on one side and the sophist on the other. We must either collapse the distinction be-

[51] Cf. 264a, where semblance is the product of opinion.

[54] Only in the Republic is it even hypothesized that philosophers might finally reach their goal of wisdom—and that hypothesis is always subservient to the attempt to envision the (perhaps unrealizable) perfectly just city, which by definition would have to have wise rulers.


tween the wise man and the philosopher in step 6, or collapse the distinction between the philosopher and the sophist in step 7. The only alternative would be to relegate the philosopher to those who are "simpleminded, believing that they know things about which they only have opinion."

The goal of the dialogue was to distinguish the sophist from the philosopher and the statesman (cf. 217a), but the final division has not succeeded in doing so. The definition of the sophist is too broad, because sometimes philosophers, too, (a) are forced to rely on opinion, (b) are then destined to produce semblances, and (c) because of this they will sometimes feel "a great suspicion and fear that they are ignorant of the things that they give themselves the appearance of knowing in front of others." The definition is also too narrow, because it is conceivable that sometimes sophists will (a) have knowledge of their subject, and (b) create a likeness of it; and (c) because of this, when they do produce semblances (which is most of the time) they will sometimes think they "know things about which they only have opinion." At most, the definition provides an initial basis for a distinction between the sophist and the philosopher, to the extent that it implies that for the sophist semblances are adequate, whereas the philosopher must strive for likenesses. But we can understand this difference in their "instruments" only if we understand the difference in their values and goals: only their end can explain their means. As I previously mentioned in section 4, in the Statesman (303b-c) the stranger will collapse the distinction between likeness and semblance, and describe as "sophists of sophists" all those who produce images of the uncodifiable "science of the mean." Since this does not distinguish between images as semblances and images as likenesses, even the philosopher, as defined above in terms of likeness making, would fit that description and be classed with the sophists. We will have to wait for the Statesman to hear the stranger's final word on the distinctions among sophistry, philosophy, and wisdom.

Although the sophist is never successfully distinguished from the philosopher, he is distinguished at least in a perfunctory way from the statesman. Theaetetus agrees that the person located under step 8a (making speeches) is a demagogue rather than a statesman, and that the sophist differs from the demagogue by employing the art of contradiction rather than extended speeches (268b-c). The distinction between the sophist and statesman is therefore only indirect, since they are not distinguished from each other but mutually from the demagogue; and indeed the distinction between the statesman and the demagogue is only


nominal, since Theaetetus does not explain in what way they differ from each other. The philosopher , on the other hand, is never even mentioned in the final division. The failure to differentiate the sophist not only from the statesman and the wise man, but also from the philosopher, is a serious shortcoming, and not only for reasons of definition. The difficulty of distinguishing between the philosopher and the sophist was a major factor in Socrates' conviction and execution (which was alluded to at the end of the Theaetetus , hence implicitly at the beginning of the Sophist , 216a); and this difficulty has been kept before our eyes throughout the dialogue. For example, (1) the philosopher, like the sophist in this dialogue, shows himself in a multitude of forms—one of which is that of the sophist (217c). (2) Both the philosopher (218d) and the sophist (221d) are depicted as hunters. (3) When the sophist (who makes inaccurate images, or semblances) is said to hide in an impenetrable place (236d, 238c), we find that the philosopher (who makes accurate images, or likenesses) is there with him. As we saw, the impenetrability is due to the nature of images in general (239d-240c, 241e, 264c), not of semblances in particular, and in one place they are referred to only as likenesses (240b). (4) The sophist resembles the Socratic philosopher as the wolf resembles the dog (231a). (5) Like the philosopher (cf. Theaetetus 146c-d), the sophist rejects strings of examples in place of definition (239d-240a).

However, not only does the final division never completely define the sophist in a wily that makes explicit how he differs from the philosopher; it is the only division in the dialogue in which we cannot even implicitly locate the distinction between the sophist and the Socratic philosopher. In division A's depiction of the sophist as a hunter of youths, step 10 distinguished the lover who hunts youths in order to give them something, from the person who hunts them in order to take something from them. If we divided 10a ("as a gift") along the same lines as 10b ("for remuneration"), the Socratic philosopher would be visible as one who teaches virtue as a gift rather than for financial reward. In division B the sophist is once more identified as a teacher of virtue (9b) for the sake of marketing (4b) rather than as a gift (4a). The Socratic philosopher could therefore be found by means of a parallel division within 4a, that is, as one who teaches virtue freely. In division C (step 9) the sophist is a verbal warrior engaged in eristic:

STRANGER : That which, within an art, argues both about justice itself and injustice and about all the others generally, are we not accustomed to call this eristic?


THEAETETUS : Of course.

STRANGER : Of eristic, one kind actually wastes money, and the other makes money.

THEAETETUS : Absolutely.

STRANGER : I suppose that the one that, through the pleasure of this pursuit, causes someone to neglect his affairs, and the style of which gives no pleasure to most of his listeners, is, in my opinion, called nothing other than loquacity.

THEAETETUS : That's more or less what it is said to be.

STRANGER : The opposite of this, which makes money from private eristic [is the sophist]. (225c-e)

As is often observed, the first of these subdivisions sounds very much like a satiric description of Socrates,[55] so here again the differentia between sophistry and Socratic philosophy would be the subordination of the former to personal gain. Division D distinguishes the two negatively, by showing that if we confine our attention to reason without reference to the acquisitive parts of the soul, only the philosopher and not the sophist will be found.

Not only is the sophist excluded from division D, but for another reason so is the Eleatic stranger's method, and noticing the reason for this will help us understand why, in the course of the employment of that method, the dialogue moves gradually from its Republic- like beginnings to its sterile conclusion.

It is worth noticing, in passing, that even if the final division is not successful in answering the original question of this dialogue (that of the fundamental difference between the sophist and the philosopher), it is successful in offering a plausible answer to the Theaetetus's question about the nature of knowledge—although it does so only implicitly. Since the difference between likeness and semblance is that the first is an accurate image of reality while the second is only specious, it is possible to define knowledge as the ability to produce likenesses of reality.[56] Those who have knowledge can infallibly distinguish between a likeness and a semblance, whereas those who are guided only by opinion cannot. The final division is also successful in giving a plausible answer to

[56] This does not mean that those with knowledge will always want to create a likeness, but only that they are able to do so.


the related question on which the Theaetetus finally foundered: the nature of epistemic logos. Such a logos would be a likeness in words (cf. subdivision vi), and the Sophist abounds with examples of this kind. I do not mean to suggest by this that the aporiae of the Theaetetus have thus been definitively resolved; I do not think that Plato regarded such questions as fully answerable. Consider, for example, the question of whether logos makes knowledge possible or knowledge makes logos possible—a question that haunts the dialogues from the Meno to the Theaetetus , and is in between laid to rest only provisionally by metaphors like "recollection." If logos makes knowledge possible, then we must be able to formulate a logos of something before we know that thing, which seems impossible. If, on the other hand, knowledge makes logos possible, then we must have knowledge before we can formulate a logos of that thing, in which case philosophical inquiry no longer seems possible. But even though the Sophist provides no definitive resolutions of such perplexities, and provides us with no model of knowledge that confronts the problem addressed by the doctrine of recollection (as the Statesman will do in the "method of paradigms," 277d-278e), our understanding of what is involved in the questions is raised to a higher level by virtue of the dialogue's focus on kinds rather than individuals.

9. Diaeresis and Value

The form from which division D began was the art of division (

) or discrimination (
; 226c). So far the stranger's own method is expressly included, since it is called a method of division (
). But the first subdivision is that, "in the discriminations just mentioned, there was one that separated better from worse, and another like from like . . . . I do not know the name of the latter, but the name of the kind of discrimination that keeps the better and throws away the worse . . . [is] purification" (226d). The former (purification) is the path that they pursue, and on which they later discover Socratic philosophy, but the latter (dividing like from like) is the form within which the stranger's method belongs. As he describes it a few lines later, in the course of this same division,

the method of definitions [

] does not care more or less about sponging than about taking medicine, nor whether one provides us with greater or smaller benefits than the other. It aims at acquiring an understanding of what is akin and what is not akin in all the arts, and, with this intention, it honors


all of them equally. Moreover, in view of their similarities it does not consider one of them more ridiculous than another.

Unlike the Socratic pursuit of philosophy, then, the stranger's method does not discriminate better from worse. It is value-free. That is why the stranger—once he has aligned himself with the pursuit of sorting like from like rather than the purificatory separation of better from worse—no longer has the means to distinguish the sophist from the philosopher.[57] In fact, at the end of division D the stranger does make an informal distinction between the sophist (whom the division did not find) and the Socratic philosopher (who was found instead), and the distinction was precisely in terms of value: "I am afraid to call them sophists . . . lest we confer upon them too much honor [

]" (231a). The philosopher is a "noble" sophist (231b). The stranger's method does not, however, lend itself to exploring this difference of value.

In principle, Plato might still have had the stranger discriminate between the sophist and the Socratic philosopher by distinguishing the giver of gifts from the vendor of wares as two different types, neither of which is posited as better than the other. The fact that he does not do so after leading us, by means of the preliminary divisions, to expect

[57] Stanley Rosen is, as far as I know, the first commentator to appreciate the implications of this passage. He suggests that the strangers "insistence upon the apolitical nature of his method . . . [is] falsified by his inability to keep sharply distinct the criteria of like/unlike and better/worse" (PS 326). The reason is that "Plato was experimenting with a technical resolution to the problems of non-being and falsehood and at the same time was indicating . . . that a technical resolution to these problems is not feasible. In other words, Plato is both presenting a plausible philosophical accusation against Socrates, by means of the persona of the Stranger, and himself refuting that accusation . . . . [T]he Statesman is the Strangers recantation" (p. 308). I agree with Rosen about the importance of this tension, but it is not necessary to regard Socrates and the stranger as serious adversaries whose enmity can be reconciled only through recantation. I argued in the previous chapter that in the Theaetetus Socrates leaves the forms in the background in order to show how the inquiry fails without them. In the Sophist the stranger brings back the categorial aspect of the forms, but leaves the valual aspect in the background to show the shortcomings of the inquiry without it. When he reintroduces it in the Statesman , therefore, this can be construed as the completion of the project of the Sophist , rather than a recantation of it. It is the positive counterpart of the Sophist's indirect demonstration of the importance of value. The Theaetetus-Sophist-Statesman trilogy is a cumulative exploration of the elements implicit in the theory of forms, in terms of both the absence and the presence of those elements individually.

It might be objected that when the strange/rejects value distinctions in the present passage he is doing no more than his "master" Parmenides had done in the eponymous dialogue, when he rebuked Socrates for doubting that there are forms of lowly things: "You are still young . . . and philosophy has not yet taken hold of you as, in my opinion, it eventually will; at which time you will not despise any of these" (Parmenides 130e). In that passage, however, the question was not whether some forms are more important than others, but whether some kinds of things were too lowly to have forms at all.


some such differentiation may be a way of suggesting that the distinction cannot fully be appreciated without reference to better and worse. It is not merely a question of whether philosophy is "better" than sophistry, although that is at least implicit. What distinguishes the philosopher and the sophist is just their differing beliefs about what is good. For the philosopher what is good is the love of wisdom, and the sharing of it with others; for the sophist what is good is personal gain—whether in terms of honors or riches—and skill in reasoning is good only as a means to this end. According to every other dialogue in which the question arises, the difference between the two is that, in the philosopher the better (part of our nature) rules the worse; in the sophist the worse rules the better. As long as we try to distinguish them without reference to the idea of value—the good—but only by means of the products that they generate in the course of these pursuits, we will be no better than those whom Socrates ridicules in the Phaedo for trying to understand reasons in terms of their material results rather than in terms of their underlying purposiveness. Such people would say that the reason Socrates sits in jail instead of running away is the mechanical arrangement of his limbs in a sitting position; whereas the true reason is his belief that it is better not to run away (98c-99a). The difference between the philosopher and the sophist is not that the philosopher always creates accurate images and the sophist inaccurate ones (for a sophist does not stop being a sophist when he happens to give an accurate description of something, nor does a philosopher cease to be a philosopher when his descriptions are imperfect), but that the philosopher always desires to create accurate images, while the sophist desires to create persuasive ones whether or not they happen to be accurate; and so the former strives for likenesses while the latter is satisfied with semblances. Only if we understand the differing values from which these different desires spring, will the distinction between likeness and semblance tell us anything about the real difference between the sophist and the philosopher. As Aristotle remarks, in a passage cited in the previous chapter, what distinguishes the philosopher from the sophist is a difference in the purposes of their lives.[58]

In the Parmenides Plato explored the difficulties that result from adherence to the theory of forms. But his character Parmenides also insisted that "in the case of each hypothesis you must examine what follows not only if what is hypothesized exists, but also if it does not exist "


(135e-136a). In the trilogy that follows he explores the difficulties that result from hypothesizing the nonexistence of the central features of the forms. The Theaetetus was a negative demonstration of the importance of the forms, which shows that if the epistemological aspect of the forms is absent, it will not be possible to give an account of knowledge. Subsequently, something like this epistemological aspect of the forms is reintroduced in the Sophist in terms of the stranger's conception of "kind" (generally

), which leads to an account of the difference between knowledge (likeness of reality) and error (semblances of reality). But by confining itself to the categorial aspect of the forms without their valual aspect, the Sophist proves unable to distinguish knowledge from wisdom—or, therefore, sophistry from philosophy. Wisdom, according to the Republic , is knowledge combined with self-mastery, and the latter factor is what the sophist lacks. But self-mastery implies an ability to discriminate the better from the worse (the proper master from the subordinate), and this is just what the stranger's "devalued" forms cannot accomplish. It is in the Statesman that the further lack will be supplied, as the stranger reintroduces the concept of value in terms of his doctrine of the mean.[59] It is a serious mistake to read the Sophist (as is almost always done)[60] without bearing in mind that it is only the second installment in a trilogy. Only in the light of the Statesman is the landscape of the Sophist clearly visible. The Statesman , together with its successors, the Philebus , the Laws , and perhaps the Timaeus , is centrally concerned with value—the very thing from which the stranger's method, as employed here, abstracts.

With this in mind let us recall the stranger's substitution of "power" as the criterion of reality or existence, in place of Plato's earlier criterion of "rationality." It is striking that in that very passage where the stranger gives his value-free speech, two examples reminded us of the value-committed speech of the friends of the forms. To secure the materialists' agreement that some things are both immaterial and real, the stranger used the example of virtues (247a-b). And to illustrate the sophists' technique of failing to distinguish between the "is" of identification and the "is" of attribution, he compared "Man is good" with "Man is man" and "Good is good" (251a-b). He further reminded us that the phi-

[60] Klein (PT ), Rosen (PS ), and Benardete are exceptions.


losopher can never be free of valuation, for "the philosopher especially honors"[61] knowledge, wisdom, and reason (249c).

There, in the middle of the dialogue, we were reminded of the limitations of the stranger's value-free method. The importance of value was also anticipated at the beginning of the dialogue, with Socrates' opening speech about the gods' concern with whether or not we are virtuous (216b). Now, at the end of the dialogue, we are once again reminded of the primacy of virtue.[62] In step 6 of the final division, the stranger divides imitation into that based on knowledge and that based on opinion. He uses the following example:

What about the pattern of justice and of virtue taken as a whole? Aren't there many who, although they are ignorant of it, still have a kind of opinion, and who try hard to be eager to make this, which they believe to be virtue, appear to be in them, by imitating it in their behavior and speech as much as possible?

I believe that it is against this background that the substitution of "power" for "reason" is to be understood. The middle dialogues were concerned with distinguishing the better from the worse; but the stranger is interested here only in distinguishing similarities and differences. On that basis he is able to ask a question that Plato has never before asked: What do being and becoming have in common? For the friends of the forms, the task was to distinguish being from becoming. The stranger, on the other hand, wants to collect them together into a common form, and he uses the form with which he began the Angler division—that of power, implicitly understood as existence. But at the same time the repeated references to virtue and other values remind us of what has had to be sacrificed in order to proceed in this way. Another such reminder may be found in steps 2 and 3 of the final division, where the stranger divides the form of divine production into "entities themselves" and "images" (265e-266a). The division appears to be a deliberate reminder of the two-world ontology of the Republic's Divided Line. The resemblance is intensified by the fact that in step 3 the stranger divides not only the right side of the division (human production), but also the left side (divine production). "Since there are two," he says, "cut each of them again into two" (265e). There is certainly no need to cut both, because the sophist is pursued only on the right side, and nothing is

[62] Cf. Friedländer 3.279.


gained by dividing the left. This is in fact the only place where the stranger symmetrically divides both the left- and right-hand forms—a "superfluity" that is repeated in the summary. Since there was no methodological basis for the symmetrical division, and indeed every precedent has been against it, the reason for it must lie elsewhere.

Not only is the procedure of step 3 without precedent, but step 2, on which it is based, is invalid. A properly exhaustive division should be, first, between divine and mortal production, with mortal production then subdivided into human and animal. Even this latter division may require intermediate steps.[63] The most obvious explanation for making the present kind of division despite the above objections is that, thus cut, the divisions resemble those of the Divided Line. The four resultant kinds are divine production, images of divine production, human production, and images of human production. Taken together, they present a vivid echo of the Divided Line seen in terms of the Allegory of the Cave (509d-517c): (1) the divine forms (represented in the Cave allegory by natural objects, especially celestial bodies), (2) images of the divine forms (represented by shadows and reflections of natural objects), (3) the physical world (represented by human-made artifacts), and (4) images of the physical world (represented by shadows of the artifacts).

The implications of such a reminder in the context of the Sophist should not be underestimated. Two in particular are worth pointing out. The Sophist has treated the realms of being and becoming as parallel divisions. within the form of "power." Even here, in the passage just discussed, the divine and the human are treated as parallel. In the Republic , however, the entire visible realm is treated as an image of the intelligible realm (532c), and so human production would be an image of divine production (cf. Republic 596a-597b). The importance of this is that an original model has an ontological priority over its imitated image, as the stranger himself had earlier pointed out (240a-b). But the stranger abstracts from all such ontological priority when he collapses being and becoming into collateral species of "power," and when he collapses the divine and human into collateral species of "production." The reference to the Divided Line and the Cave reminds us that the stranger's principle of abstracting from value distinctions such as noble and base, or higher and lower, is not without its questionable side.

The other implication leads in a similar direction. Any reminder of the Line-Cave section suggests the possibility of a comparison between

[63] cf. Statesman 262a ff.


the strangers unifying concept of "power" and Socrates' unifying concept of "the good." There are three important differences between these two concepts. First, the good is a first principle, not an inclusive class. Second, it is a source of value, not only of existence. Third, the good is directly relevant only to the level of being, not to that of becoming—it is the ground of the being of forms, not of the existence of particular things and events. These differences lead us to notice the same deficiencies in the stranger's ontology that we observed in the previous paragraph. In theory, one might turn this argument around and suggest that the reference is not meant to make us criticize the Sophist in the light of the Republic , but to criticize the Republic in light of the Sophist . The reminder of the Republic might be a way of calling attention to the position that is being superseded . There are two reasons why this is unlikely. The first is that in his dialogues after the Sophist , Plato consistently returns to an ontology more like that of the Republic , in which being is given ontological priority over becoming—the intelligible world over the visible. The second reason is that the recurring reminders of the importance of value, which we have observed at work throughout the Sophist , would be inexplicable in a dialogue whose fundamental orientation was toward an overcoming of the connection between ontology and value.

The stranger's provisional overcoming of that connection, in the service of showing how all existence can be neutrally classified in terms of sameness and difference, is an important philosophical point. It represents the isolation of one of the necessary conditions for the theory of forms, namely, the principle of "sameness." But the fact that it operates as a bridge between the ontology of the friends of the forms and that of the materialists does not mean that it is a dialectical synthesis of and therefore improvement over each. It is rather the kind of synthesis that belongs to a common denominator, and it stands midway between them.

The first three of the five "greatest kinds"—being, rest, and motion—correspond to this highest kind of quasi genus (see above, n. 37) and its two primary aspects. The highest kind is being as power or existence (

). The aspect "rest" corresponds to the friends of the forms' concept of "being," as the realm of changeless forms (which must here be designated as "rest" because its previous name, "being" [
], is now reserved for the more inclusive form). And the aspect "motion" corresponds to becoming, formless flux.[64] The other two great-

[64] I cannot, therefore, agree with Cornford that there is no special significance to the use of motion and rest here, and that "the only fact about them that is relevant is that they are contrary and incompatible" (PTK 277-78). The present interpretation also gives significance to 243d-245d, where the stranger shows that if we take. as an ontologically first principle, either a dyad of opposites or the One itself, absurdity will result. The passage in question may be seen as embracing both the rejected one-sided positions dialectically: the first principles are both a One (existence or power) and a dyad (rest and motion). (Cf. Klein's documentation of the leitmotiv of "both": PT 60 and passim .) A precisely analogous dialectical ascent occurs in the Phaedo to illustrate the upward stage of the method of hypothesis: from the "wise" (100c) but unreliable hypotheses of the materialism (96a-97a), to the "safe" but "simpleminded" and "ignorant" version of the theory of forms (100d), to their eventual synthesis in a more complex formulation of the theory of forms—forms as bearers of material properties. The latter is still "safe," like the previous formulation, but now "sophisticated" instead of "ignorant" (105b-c), because it includes the physical causality of the "wise" materialists.


est kinds, sameness and difference, are the necessary conditions for division of a form into its aspects (254d-e). The aspects are collected into the common form by virtue of their sameness and distinguished within the form by virtue of their difference. Such a classification is obviously of great philosophical interest; but the considerations discussed above show that, while classifications such as these can enrich and extend the valuational inquiry of the friends of the forms, they can not replace or supersede it.

The stranger refers to the antagonism between the friends of the forms and the materialists as a gigantomachia or "war against the giants" (246a), a reference to the battle between the giants and the gods. "Giants" explicitly refers to the materialists (246a, 248c), while the friends of the forms, whose weapons come from the invisible (immaterial) realm above (246b), are implicitly identified with the gods. The metaphor suggests that the stranger appreciates that there is a difference not only of kind but also of value between the two philosophies. It may be that he is on the side of the "gods" after all, but is pushing Theaetetus and us to make the connections ourselves. This possibility gives heightened significance to Socrates' opening words in the dialogue: "Did you not notice, Theodorus, that you have brought not a stranger but some god?"


Chapter Four
The Statesman

1. Statesman As Shepherd (257a-267c)

After the fanfare and multiple demonstrations with which the stranger introduced the method of symmetrical division in the Sophist , it is surprising to find that that is the only dialogue in which Plato makes use of it in its pure form. The same Eleatic stranger conducts the inquiry of the Statesman , but halfway through the dialogue the divisions cease to be symmetrical and bisective, and the stranger increasingly resorts to myth, analogy, and metaphor—sespecially the metaphor of weaving—in a way that resembles Socratic inquiry much more than it resembles the diaeretic method employed in the Sophist .[1] In fact that method is employed in a dubious way from the beginning of the dialogue. The first division may be schematized as in Figure 10.

The details of this diaeresis show a remarkable decline from the rigor and precision of the divisions in the Sophist . Three points in particular are worth commenting on.

1. In the shorter way the distinction between leathered and non-leathered is made twice: the differentia of step 7, "leathered" (or

[1] The shift is preceded by a discussion of the importance of paradigms (277d-279a). We may think of that discussion as a deliberate tempering of the Parmenides ' critique of analogical and metaphorical reasoning. The discussion acknowledges that paradigms are only an indirect way of presenting what we would prefer to present directly, but that they are sometimes indispensable.



Figure 10



), is repeated in step 9 (
).[2] This is an obvious violation of the method. Once "leathered" is excluded by step 7, it can no longer be implicit in the species to be divided in step 9. Since this manifest anomaly involves the species of birds, it calls our attention to a more serious problem about that species, the fact that the stranger's divisions, both here and in the Sophist , destroy the species of birds. In both dialogues there is a distinction between land animals and water animals. The Sophist locates birds within the genus "water animal" (220b) while the Statesman locates them within the genus "land animal" (264e). Thus we can locate water birds within one genus and land birds within another, but there is no genus that divides into the species of birds per se . We can isolate classes that are too general (animals) or too specific (water birds, land birds) but not the precise species of birds itself. The problem could easily have been avoided by making "feathered" into a prior differentia, that is, by dividing animals into feathered and nonfeathered, and nonfeathered into swimmers and walkers (birds could then have been located by appropriate distinctions within "feathered"). In fact this very point is suggested by the summaries, both here and at 276a, which omit step 6 and thereby separate the species "feathered" prior to any distinction between water and land.

All this would be of little consequence if the method of division were presented merely as a convenient method of classification, but it is offered rather as a method of arriving at real essences, by discovering which species belong by nature within which genera. The stranger speaks of distinguishing one kind of thing from everything else in terms of its "single form" (

, 258c5), and later says,

When someone at first perceives the common character of many things, he should not turn away until he sees in it all the distinctions, as many as lie in the forms [

]. When on the other hand he at first perceives all sorts of differences in a great many things, he must not become discouraged and stop until he has put those that are akin into a single similarity, enclosed in the genus to which they really belong [

The problems we have seen, however, (and will continue to see) should make us wonder whether the method of bisective division is being recommended as wholeheartedly as at first it seemed.


2. The diaeresis begins with the same flagrant violation of methodology with which it ended. Just as step 9 of the shorter way used a differentia already excluded by step 7, step 2 uses one already excluded by step 1. Step I distinguishes practical from intellectual arts on the grounds that the former eventuate in products (

, 258e). But when the latter (which by definition now exclude the producing of products) are divided into critical and directive, the distinction is that the directive art is for the sake of producing (
, 261a-b) and necessarily has products (
, 261b). Here too a characteristic supposedly excluded by the previous division reappears and needs to be excluded again in the subsequent one (even though the stranger could have made his point without inconsistency, as we shall see). This is reinforced by an oddly enfeebled conclusion that the stranger draws: not that the art of the king is within the class of theoretical rather than practical arts, as the stranger has just claimed to show, but only that his art is "more akin to" the theoretical than the practical (259c-d)—a weak identification that is hardly adequate to the requirements of definition by division (although it will be echoed in the "weaver" division), but that faithfully reflects the lack of rigor in the first two steps. All this seems too confusing even for the stranger. Here he locates kingship within the directive sciences, which are distinguished from the critical; but later he will say that kingship was differentiated from other sciences as being both directive and critical (292b).[3] More than methodological carelessness is involved here. After the stranger distinguished between practical and intellectual arts on the basis of whether or not they involve products, he remarked that "any king can, with his hands or his whole body, do little to hold his rule, compared with what he can do with his entire soul and its force" (259c). This distinction between what can be accomplished by the body and what by the soul, which was mentioned only in passing, could have been used to avoid the problem that we just noticed. Suppose that step 1 distinguished the intellectual sciences from the practical not by claiming that the latter have products and the former do not, but that the latter have corporeal products (resulting from the activities of the body) while the former have intelligible products (resulting from the activities of the soul). In that case, the intellectual sciences would not have been defined in such a way as to preclude them from making products, and there would be no contradiction when step 2 uses this as the differentia to

[3] Cf. Benardete 3.126.


distinguish the directive intellectual sciences from the critical. Step 1 would have rejected only corporeal products while step 2 affirmed intellectual products. Thus the incoherence is egregious and the stranger himself has shown us how he could easily have made these divisions without tangling his distinctions. Why does he refer to the corporeal/ intellectual distinction but not use it—especially when it would have saved him from incoherence? This issue will reappear in an intensified way within the next passage discussed.

3. After step 5, where herds are distinguished from individuals, Socrates' namesake, young Socrates, proposes that the next division be between human herds and animal herds, and is rebuked by the stranger for impatiently cutting off a small part (herds of humans) from the whole (herds) rather than gradually narrowing the field by progressive symmetrical divisions (262a-b). However, the stranger has evidently not dosed the door on shortcuts altogether, for he soon offers young Socrates the choice of a longer or shorter route by which to track down the statesman. The longer is methodologically more correct, he repeats, but "it is possible to go by whichever of these ways we wish" (265a). Young Socrates understandably would like to see both, so they take first the longer, then the shorter way, and arrive at two apparently quite different conceptions of what the statesman is. Both conceptions start from the assumption that he is a herdsman of human beings, but the way a human being is conceived is very different. According to the more famous shorter way, the lowest genus to which humanity belongs is the genus of two-footed animals, which includes only the two species "humans" and "birds." The difference between the two is that humans have no feathers, so our essence is "featherless biped." According to the longer way, the lowest genus to which humans belong is the genus of tame, hornless, noninterbreeding herd animals, which includes only the two species "humans" and "pigs."[4] The difference between the two is


that humans have two legs and pigs have four, so our essence is "two-footed, tame, hornless, noninterbreeding herd animal."

There are at least two problems here. First, the concept "human" is radically different in each case, with "two-footed" acting as the genus in one and the differentia in the other. Second, neither definition is very impressive: according to one, humans are like birds with the feathers missing; according to the other we are like pigs with two legs missing. And yet this method is supposed to lead us to a thing's very essence (285a-b). How does young Socrates react to all this? "The argument was beautifully done," he replies (267a). Perhaps, then, the problems are meant to reflect on young Socrates' lack of conceptual discipline and rigor, rather than on the method itself, for he has also ignored the stranger's recommendation to take the long way rather than the short (262a-c, 265a).[5]

On the other hand, perhaps he ignored that recommendation because of its conditional character: "It is finest to distinguish what is sought from everything rise immediately, if that correctly reflects how the things really are . . . [But] it is safer to make one's cuts by going down the middle, and one would more likely hit upon the boundaries between the forms" (262b). Conceptual rigor and safety are serious concerns in Plato, as he previously had shown in the Phaedo (101d-e, 105b-c). But if a different approach is called the finest or most beautiful one (

), this is certainly a still more impressive recommendation. Even the safer way is not entirely secure: it is only "more likely" to hit upon the boundaries. Accordingly, while the longer way would be safer, it would not be unreasonable for young Socrates to hope that the stranger has found a valid shortcut, which would, after all, be the "finest" solution. In fact the shorter way does, in effect, "distinguish what is sought from everything else immediately." Since leathered creatures were already ruled out by step 7, all the stranger needs to do is distinguish the resultant species of (featherless) walking animals into two-legged and four-legged (there was no need to exclude feathers again in

[5] Cf. Mitchell Miller, The Philosopher in Plato's Statesman (The Hague: Nijhoff, 1980) 21-29.


step 9, as we saw). Strictly speaking, then, the definition that the shorter way leads to is not "featherless biped" but "two-footed walker." The subsequent reintroduction of the already discarded category of "feathers" can only be regarded as a kind of joke since both divisions begin from the same point, if the shorter way still needed to exclude feathers so would the longer way. The stranger's intention, in this calculated attempt to make the shorter way look less respectable than it really is, is perhaps to reinforce his effort to discourage Theaetetus from looking for shortcuts before he knows how to do so properly. Had Theaetetus been able to demonstrate at this point the ability to see through the stranger's subterfuge, it would have been a different story.

The shorter way turns out to be more convincing than the longer one not only in its one-step definition ("two-footed walker"), but also in its procedure. Since the genus from which the longer and shorter ways began was "walking," it is more natural to use "number of feet" as the differentia, rather than "presence of horns." Nor do the differentiae "horned/hornless" and "interbreeding/noninterbreeding" tell us anything significant about human nature. Even though they are technically more correct because they proceed by more symmetrical cuts, they are artificially devised and are more in the nature of distractions from, rather than concentrations upon, the subject to be defined. (As Socrates points out in the Philebus [17a], it is possible to go from the one to the many too slowly as well as too quickly.) It is perhaps an implied criticism of their artificiality that, when this division is later recalled, the order of these two unilluminating differentiae is reversed (276a)." It is far more natural to define human beings simply as two-footed animals than as two-footed noninterbreeding hornless ones.[6] Consequently it is the definition of the shorter way that serves as the model for Aristotle, who commonly defines "human being" as the "two-footed animal" (and describes us as "by nature a political [cf. "herd"] animal": Politics 1.2.1253a 2-3).

If we think of the shorter way in terms of the definition to which it really leads, rather than the redundancy that the stranger tacks onto the end in order to belittle it, then there is a very pointed difference between the results of the two methods. Both take "two-footed" to be the final

[6] Moreover, as Harvey Scodel points out, the longer way "is somewhat perverse in its implication, for it construes human nature . . . as a surd, while that of its four-footed counterparts is rational" (Diaeresis and Myth in Plato's Statesman [Göttingen: Vandenhoeck and Ruprecht, 1987] 65). The reference is to the stranger's description of human nature in terms of the square root of 2 at 266b.


differentia of human beings, but the shorter way uses this to distinguish human beings from all walking animals, whereas the longer way uses it to distinguish us only from pigs. Which of these reflects more accurately the distinctive nature of humanity? It cannot be denied that by dividing walking animals immediately into two-footed and four-looted we produce a seemingly unbalanced division—humans on one side, all four-footed animals on the other—whereas by the longer way the divisions are more evenly balanced at each step. But on the other hand the longer way ignores the distinctive nature of human beings. A rigid insistence on balanced divisions may conform to the principle of relative measure, but violate that of the mean (to anticipate a distinction that the stranger will make later on). It may be that human beings are so distinctive that it makes more sense to speak of humans as one side and all "other animals" as the other, as young Socrates was inclined to do.[7] The stranger did, it is true, rebuke young Socrates for doing something of the sort when he simply divided "animals" into "humans" and "beasts" (262a if.); however, the problem with young Socrates' immediate division of animals into human and nonhuman is not that it is asymmetrical, but that it gives no indication of what he takes to be the essential nature of humanity. Rather than identifying the determinative differentiation of the species "humanity," young Socrates simply makes the species its own differentia, and so the definition, in its impatience to arrive at the infima species , bypasses the all-important step of discovering what the real differentia of humanity is. Because it dispenses with an identification of the essence, the distinguishing feature of "human being," it fails to give us the very thing that the process of division is supposed to accomplish.

In a later remark the stranger shows that it does in fact make sense to set human beings apart from all other animals as long as we can discern the true differentia: "human beings, who are different and more divine [

] animals, govern the lower species of them" (271e). Later still we are told that the human soul comprises a divine part as well as an animal one (309c), which once again distinguishes us immediately from all other animals.[8] Although these distinctions make no reference to the number of our legs, it is not unusual to regard our two-

[7] J. B. Skemp points out that "Aristotle argues very thoroughly against any attempt to reach any of the infimae species of the animal world by a process of division by dichotomy [De Partibus animalium , I, 2-4; 642b,5 644b,20]" (Plato's Statesman [London: Routledge and Kegan Paul, 1952] 70).

[8] Scodel 57, too, notes that "Plato himself continues to invoke the distinction between man and beast despite the Stranger's stricture (cf. Phil . 16a and Laws 653e)."


legged posture, which sets our head heavenward, as the outward sign of our distinctive nature:

The name anthrôpos ["human being"] signifies this: that, on one hand, the other animals do not examine, reason about, or look up at what they see; but a human being, as soon as it has seen—and this means viewed—then it looks up and reasons about what it has viewed. Therefore alone of the animals the human being is rightly named anthrôpos , since it looks up [anathrôn ] at what it has viewed [opôpe ].
(Cratylus 399c)

For of all animals man alone stands erect, in accordance with his god-like nature and substance. For it is the function of the god-like to think and to be wise; and no easy task were this under the burden of a heavy body, pressing down from above and obstructing by its weight the motions of the intellect and of the general sense.
(Aristotle, Parts of Animals 4.10.686a 27-32, rev. Oxford translation)

In challenging young Socrates' division of animals into human and not human, the stranger said that any other intelligent animal, as for example the crane is thought to be, might make an analogous division: for example, cranes and noncranes (263d). The stranger's primary reason for speaking of other intelligent animals here is that only an intelligent animal could make such a distinction; but there is a secondary consequence as well. If there are other intelligent animals, we are prevented from using intelligence or rationality as the obvious differentia by which to distinguish humans from other animals. The crane may be rational too. However, as we soon learn, only humanity is godlike. So there really is a natural differentia by which we might have been distinguished from all other animals immediately. The implication is that what distinguishes us from other animals is not a technical distinction but a difference of value, "divinity." The crucial difference between ourselves and other animals lies not merely in the number of legs, or even the presence of intelligence, but in the divinity of our nature, that is, in the nature of our soul . The longer way, in insisting on equal divisions, may often have to bypass the "fitting" division, as in this case. When it does so, it succeeds according to relative measure but fails according to the mean.

The key to the two preceding problems lies in the difficulty, for the method of division, of taking account of nonvisible marks such as the distinction between body and soul, or between one kind of soul and another. This difficulty stems from the method's inability to recognize


differences among levels of being that imply differences of value. Such differences are alluded to even though they are never addressed. The stranger said that "the science of the king is never one that supervises soulless things, like architecture, but it is nobler [

] in that it always exercises its power among living beings and in relation to just these" (261c-d). Later, when he draws the distinction between "more divine" and "lower" souls (271e), it is clearer still that any adequate account would have to make distinctions of value. Since the first part of the dialogue defined statesmanship as the nourishing of human beings the way shepherds take care of their flock, we might have expected the concept of value to enter into the notion of "nourishing." However, the stranger never inquires into the meaning of "nourishing," but only into the meaning of "human being." Later he calls attention to this as a failure, saying that their investigation did not specify in sufficient detail how the statesman rules (275a). The myth will redress this reticence about nourishment—as will the final part of the dialogue (310a ff.)— explicitly discussing the nourishment of the human herd, both corporeal and otherwise (271e-272c).

In the Sophist the stranger had emphasized that the method abstracts from considerations of value (227a-b), and this turned out to be why the dialogue's eventual distinction between the sophist and the philosopher is ultimately unsatisfactory. Here, too, at the end of the longer way, the stranger defends the juxtaposition of pigs and people in the ultimate genus by repeating the prohibition against paying attention to differences of value (266d). But the present dialogue, unlike the Sophist , will violate that principle with increasing frequency throughout its course,[9] as it replaces the longer way with the shorter. It even begins with an explicit recognition of the importance of value:

SOCRATES : I owe you a great debt of gratitude, Theodorus, for my acquaintance with Theaetetus and also for that with the stranger.

THEODORUS : Soon, Socrates, you will be three times as indebted, when they have worked out the statesman and philosopher for you.

SOCRATES : Indeed? Shall we say that this, my dear Theodorus, is what we heard from our great calculator and geometrician? . . . You are placing equal value on each of these three, who are farther apart in honor than your art of proportions can express.

[9] Cf. Michel Fattal: In the Statesman , "le dialecticien semble par là contrevenir à la règle qu'il s'est pourtant fixée dans le Sophiste en 227a-b et dans le Politique lui-même en 266d, celle qui proscrit l'usage des 'valeurs' comme critère de division" ("La diairesis dans le Politique de Platon" [paper presented at the Third Symposium Platonicum, "Plato's Politicus, " Bristol, August 25-30, 1992] 13).


THEODORUS : By our god Ammon, Socrates, you have rebuked me well and justly and with presence of mind. (257a-b)

The second and third of the problems we noted regarding the opening division of this dialogue may be indirect confirmations of Socrates' warning: that an attempt to account for distinctively human activities without reference to value is doomed to confronting externals rather than essentials. And in view of Plato's association of the good with the image of the sun, there may be an oblique reference to value as well in Theodorus's oath by Ammon, who, in his characteristic of Ammon-Ra, was the Egyptian sun god.

As I have suggested above, the three dialogues of the trilogy may in this way be regarded as an extended application of the method of hypothesis. The Theaetetus , by hypothesizing that knowledge is a species of perception or doxa, foundered in a sea of unacceptable consequences. The Sophist proposed a "higher" hypothesis, that besides the individuals that comprise the world of perception and doxa, there are also universal kinds. This higher hypothesis resolved some of the aporiae of the Theaetetus (regarding the nature of knowledge and of epistemic logos), but led to unacceptable consequences of its own—an inability to find the true differentia between the sophist and the philosopher. Finally the Statesman will present us with a higher hypothesis still, that of the mean, which implies value and the good. On the basis of this hypothesis the residual problems of the Sophist (e.g., the essential difference between philosophy and sophistry) will be more convincingly resolved. The doctrine of the mean will not be introduced explicitly until 283d if., but it has been present (in different ways) by indirection in Socrates' opening remarks about correct proportion and in the stranger's problematic opening diaeresis.[10]

2. Myth of Cosmic Reversal (267c-274d)

The division is not yet complete, for merchants, farmers, grain workers, physical trainers, and physicians would all claim that they, rather than the statesman, are the "shepherds" of humanity. Since shepherds perform all such duties and more, the claims of these rivals must be taken seriously. Instead of simply dividing further to distinguish these rivals from the shepherd, the stranger proposes to start entirely over, taking


an altogether different road (268d). He relates a myth, according to which the universe's rotation alternates directions. When the god is directly in contact with it, the cosmos rotates together with him. At this time all life springs fully mature from the earth and ages in reverse, growing younger and younger until it disappears back into the earth, from which it is eventually reborn.[11] Food is always available without effort, the weather is always gentle, and all animals live in harmony, each species under the governance of a subordinate deity. But nothing corporeal can continue indefinitely without change, so eventually, when each soul has fulfilled its allotted cycle of rebirths, the god lets go. This results in an earthquakelike shock that causes great destruction. The cosmos is a living and intelligent being, however, and does not remain at rest. Because of its blessed nature (269d) and its memory of governance by the god (273a-b), it seeks to emulate its former state, but its corporeality limits both the divinity of its nature (269d) and its ability to remember its former state (273b), so its emulation must be deficient. Consequently it rotates in the opposite direction, which is why birth, aging, and death are reversed in our experience. Moreover, when the god lets go, the subordinate gods let go of their own charges (272e), and the cosmos reverts to the harshness and injustice intrinsic to its corporeal nature (273b-c). Enmity arises among the animals, and hardship brought on by scarcity of food and inclemency of weather. It was as a compensation for this that Prometheus gave us fire, and Hephaestus and Athena gave us the arts (274b-c). Over time the cosmos continues to forget its divine legacy, and as it becomes more forgetful the proportion of disorder to goodness increases, until the cosmos and all within it are in danger of destruction. Then the ordering god, concerned that it might founder in confusion and break up in the sea of boundless dissimilarity,[12] returns to his place, orders it, and, correcting it, makes it immortal and ageless again (273b-e). The cosmos is thus immortal only so long as the god periodically intervenes. Otherwise it will perish en-tropically in complete undifferentiation.

The myth is similar in many ways to that of the Timaeus .[13] Here, as well as in the Timaeus , the irrationality of the cosmos is due to its cor-

[11] The souls must somehow age in the earth, for they each undergo a number of rebirths before the cycle reverses (272e).

[12] One way of looking at the method of division is that it aims to provide boundaries within the sea of dissimilarity.

[13] Also see Luc Brisson, "Interprétation du mythe du Politique " 5-6, and T. M. Robinson, "Forms, Demiurge and World-Soul in the Politicus " (both papers presented at the Third Symposium Platonicurn, "Plato's Politicus ," Bristol, August 25-30, 1992).


poreal nature, which "partook in great disorder before entering into the present cosmos" (273b); and here too it is partly overcome by a memory of the divine, a memory that is obscured by the corporeal nature. Moreover, just as the demiurge of the Timaeus is distinguished from the Olympian gods, here too the divine "helmsman of the universe" (272e) is not one of the Olympians, although this is not at first apparent. At one point young Socrates asks, "But the life that you say existed during Kronos's reign, was it in those revolutions or these? For it is dear that the reversal of the stars and sun comes about in both periods" (271c). The stranger replies, "You have followed the story well. But the life you asked about, when everything came about by itself for people, does not belong to the presently established period, but this too was of the previous one" (271c-d). Is the stranger merely ironic when he praises young Socrates' grasp of the story? Is young Socrates really asking something as foolish as whether the self-generation of all things from the earth is happening now? In fact the question was a legitimate one (and the praise probably sincere), but it is answered elliptically. At the beginning of the myth the stranger mentioned the story of Zeus's reversing the direction of rotation of the sun and the heavens. Young Socrates has heard of it, he says. The stranger establishes that Socrates has also heard the story of the reign of Kronos, and the story of those who were born from the earth (269a-b). It appears that young Socrates now wonders whether the reign of Kronos and the time of the earthborn coincide, for it evidently occurs to him that if Zeus was the god who reversed the rotation (as in the traditional account of the Atreus story), then both directions of rotation would occur in Zeus's reign, and the same would probably have been true of Kronos's reign. As he puts it, "it is dear that the reversal of the stars and sun comes about in both periods" (271c). So there should be a period of the earthborn in both Zeus's and Kronos's reign, and it is perceptive of young Socrates to wonder which of these two is the fabled one. But the stranger, after his initial praise, answers as if the whole of Kronos's reign is the golden age, and the whole of Zeus's reign the present age. The reversed age, he says, "was that of the people of Kronos's time, but the present one is said to be at the time of Zeus" (272b). In that case it could not have been Zeus who reversed the course of the heavens, but rather some more fundamental god who is in charge during both periods (although in control only in one of them). This is supported by the fact that the ruling god never leaves the scene, and therefore could not have been replaced by a different god. Rather than debarking from


the ship of the universe, he only leaves the helm and moves to a place of lookout (272e), from which he later returns to the helm (273e).[14]

The peculiar form of the question and answer at 271c-272b thus points to a tacit emendation in the story. The god who is celebrated in the myth is beyond Kronos and Zeus, who are by implication relegated to the position of subordinate gods. The reign of Kronos corresponds to the period when the divine helmsman takes charge of the tiller, and the reign of Zeus corresponds to the period when the helmsman has let go. These distinctions are never explicit; with Socrates' trial in the offing, perhaps the stranger is cautious about introducing "other, new gods" in Athens.

Despite the similarities between the myth here and in the Timaeus , in the latter dialogue the divine and corporeal seem to be united in a perpetual polarity, whereas here they are united sequentially. In view of the political character of the dialogue, perhaps the sequential nature of the myth is meant to assimilate the myth to the nature of political history, for it is clear from the Republic that Plato sees political states as displaying a quasi-cyclical sequence: a rise from the agrarian to the warrior state and thence to an intelligent "civilization," which will progressively decline into appetitiveness (

): from "aristocracy" to oligarchy to democracy, anarchy, and tyranny. Here, too, when the god lets go of the cosmos it reverts to "its innate appetitiveness [
]" (272e). The myth combines the cosmic point of view of the Timaeus with the political point of view of the Republic . Another function of the sequential relationship of the poles may be, as G. R. F. Ferrari has suggested, that the distinction between ruled-by-god and ruled-after-god prefigures the dialogue's later distinction between rule by the statesman and rule by law.[15] This suggestion is strengthened by the fact that the statesman (297e-299c), like the god, is compared to the captain of a ship.

Conspicuous throughout is the theme of value. At the beginning, when the stranger refers to the sign of the gods' favor that was bestowed upon Atreus (in his rivalry with his brother, Thyestes, over the succession to the rule of Mycenae), young Socrates thinks he means the golden lamb, whereas he means instead the subsequent reversal of the heavenly directions (268e-269a). Young Socrates' expectation is understandable since they had just been talking about shepherds and flocks, but more

[14] Cf. Brisson 7-8.

[15] "Myth and Conservatism in Plato's Politicus, " paper presented at the Third Symposium Platonicurn, "Plato's Politicus, " Bristol, August 25-30, 1992.


than dramatic naturalism is involved. The golden lamb was a token that the gods considered Atreus more fit to rule than Thyestes. The golden attribute signifies, therefore, the difference between better and worse. Finding the statesman, we are later told (303d-e), is like refining gold to separate it from less precious metals (as here it is said that the statesman must be made to stand apart and pure: 268b). It will turn out not to be possible to do this without appealing to differences of value.

Similarly, one of the most important questions raised in the myth is in which of the two periods people are happier. The answer is that if those in the previous period used their special advantages for the pursuit of philosophy and wisdom, then they must have been ten thousand times happier than we; but if they used them only for idle pleasures— eating, drinking, and storytelling—then the reverse is the case (272b-d). Not only is it taken as beyond dispute that the value of wisdom is immeasurably greater than that of pleasure, but it is also dear that the stranger no longer treats differences of value as irrelevant to philosophical inquiry. The complete superiority—the incomparably greater value—of the cosmos's divine component over its corporeal one is as strongly emphasized in the Statesman myth as it is anywhere in Plato. "The corporeal element in the cosmos's composition . . . partook of great disorder before its entry into the present cosmos," the stranger says. "For from its composer it acquired all beautiful things, but from its previous state comes everything that is harsh and unjust in the heavens" (273b-c). From this passage in particular, and the myth in general, it is quite clear that Plato has not abandoned his "two-world" view.[16] The distinction between the divine and the corporeal—and the polarity between them as the matrix of our world—is as unambiguous here as it ever was in the earlier dialogues. The passage may be a "myth," but it is a myth that represents a two-world conception of reality. It would be unaccountable if in this very trilogy he were abandoning that conception.

3. Paradigm of Weaving (274e-283b)

The stranger remarks that the myth has shown an error in their original diaeresis. "In one way the error was relatively small, but in another way it was of very noble [

] proportions and much larger and

[16] For detailed discussions of this question, see Richard Mohr, "Statesman 284c-d," Phronesis 22 (1977) 232-34; idem , "Disorderly Motion in Plato's Statesman," Phoenix 35 (1981) 199-215.


greater than the other" (274e). It treated the present human statesman as if he were the shepherd-god of the divine cycle (this was the more serious aspect of the error), and did not specify in sufficient detail how he rules (the less serious aspect). The difference between the divine shepherd and human king is that the latter shares approximately the same nurture and education as his charges, whereas the former (like shepherds generally) does not. The former is in fact greater (

) than the latter (275b-c). The difference between the divine and the human, like that between the divine and corporeal (and between the two aspects of the error), is primarily a difference in value.

To remedy their mistake it is necessary, first, to think of the statesman not as a herdsman, since he is not different in kind from his charges, but yet as someone who has something in common with herdsmen. The stranger collects these activities together into the form of "caring for" (

, 275e). In the preceding diaeresis the production of living things (4b) was called nourishing or feeding (
: e.g., 261d), and so the stranger proposes that they replace "nourishing" with "care" (
) and continue the division as before (276a-d). We may notice that the method of hypothesis is at work here again, in the rejection of the original conception of the statesman as a kind of shepherd, in favor of a more adequate or "higher" conception. In fact the concept of nourishing will later be distinguished from, rather than subsumed within, statesmanship (288e-289a), and the statesman will be explicated on the hypothesis that he is like a weaver rather than like a shepherd.

They made another great error in addition to this, he continues, by not making further divisions in their final definition. They ought to have divided the shepherd of human beings into a divine and a human shepherd, so that the statesman could be distinguished from the god. And they ought to have distinguished the statesman's rule into rule by force and rule by consent, so that the king could be distinguished from the tyrant (276c-e).[17] But if the essential difference between kings and tyrants is said to be that the former are obeyed voluntarily and the latter involuntarily, here again the method of division has led us not to the true distinguishing essence of the thing, but only to a superficial characteristic. The writer of the Republic , for whom the tyrant was the paradigmatic unjust man, can hardly have believed that the distinguishing feature of the tyrant is simply the reluctance of his subjects, or that

[17] It is odd that this is called an additional error. The two elements seem to correspond to the major and minor aspects of the original error.


hoi polloi will necessarily obey even just laws voluntarily. In fact the stranger will later repudiate this definition explicitly (296b).

Even apart from this the stranger is not satisfied, for the account is not only too long, but still unfinished. "It is difficult, you divine man, to show without paradigms any of the greater things. For each of us knows everything almost as if in a dream, and then is ignorant as if he has awakened" (277d). As with the doctrine of recollection (which Socrates, too, compares to a dreamlike state: Meno 85c-d), we have an implicit knowledge of all things, which we cannot discern when we try to bring it into explicit focus. How can we find the statesman unless we already know what he is, and why do we need to find him if we do know what he is? Again as with the doctrine of recollection, paradigms make possible the transition from merely implicit knowledge to explicit knowledge.

"I need to give a paradigm, my blessed one, of paradigm," the stranger says. Children who can correctly read short and easy syllables become confused when the same letters are used to make longer ones. So we can use the short syllables as paradigms of how the letters sound, from which the child will be able to read the longer ones. Thus "a paradigm comes about whenever something that is rightly believed to be the same in some second distinct thing is compared with the first, so that the two together result in one true opinion" (277d-278c). The application of the epithets "divine" (

) and "blessed" (
) to young Socrates may also remind us that the mundane can serve as a reminder or paradigm of the divine. Implicit as well in terms like "blessed" and "divine" is the connotation of value. After all, the method of paradigms is said to be especially important in dealing with the greatest and most valuable (
) things (285e).

As a paradigm of something that has "the same activity as statecraft" (279a), the stranger gives a diaeresis of the art of weaving (see Figure 11).[18] For a division that is meant to be "paradigmatic," it is, like the previous one, remarkably lax. The first two subdivisions are both nonnormal, in opposite ways. The Sophists divisions began by distinguishing production from acquisition, but this one, on the contrary, begins by conflating them: "All things whatever that we make or acquire are for the sake either of doing something or of protecting ourselves against suffering something" (279c). Whereas step 1 thus fails to divide the starting point

[18] Charles Griswold suggests an interesting connection between the choice, of the weaving of woolen garments as a paradigm of statesmanship, and the inclement conditions of the age of Zeus: "Politikê Epistêmê in Plato's Statesman, " in Anton and Preus, eds., 141-67, esp. 152, 165 n. 22.



Figure 11


as it should have done, step 2, on the contrary, performs a double division instead of a single one for no apparent reason: not only 2b (barriers) but also 2a (protective drugs) is divided in half. More remarkable than either of these lapses is the summary, which is by far the most inaccurate of all, misrepresenting the original order at almost every step. The original step 4 is missing altogether (or else it is assimilated into step 5 in a very diffuse way: cf. 280d), and only in one case is the sequence between two steps—5 and 6—accurately repeated. Not even the first and last steps are correctly recalled. The most serious deficiency is the conclusion. After having defined clothesmaking, the stranger says, "weaving, to the extent that the greatest part of it deals with the making of clothing, does not differ except in name from this art of clothes-making" (280a). But if only part of weaving (even if the greatest part) is concerned with the making of clothing, then the definition is too narrow, and avowedly so. The false note was struck with step 6 (which, perhaps significantly, is treated as the ultimate step by the summary), when the stranger divided carpets from wrappings and seeks the weaver in the latter class, even though carpets too can be woven. Without this step the stranger would not have had to weaken his conclusion with the unscientific qualification "for the most part."[19]

Not only is the definition too narrow, it also turns out to be too broad: not only are not all weavers clothesmakers, but not all clothes-makers are weavers, the stranger reminds us. They are so only for the most part:

STRGER : The one who works at the first stage of clothesmaking appears to do the opposite of weaving.


STRANGER : The work of weaving is a kind of combining.


STRANGER : But the other is the separating of what is joined or matted together.


STRANGER : The work of the carder's art. (280e-291a)

Not only carders but also fullers and menders consider themselves involved in the making of clothing, although they will admit that weaving is "the greatest part" of clothesmaking (251b).

[19] Recall the stranger's equally unscientific conclusion after the second step of the opening diaeresis, that the art of the king is only "more akin to" the theoretical than the practical (259c-d).


Even those who make the instruments for these arts are part of the art of clothesmaking, but these might be classified as contributory causes rather than true causes (281d-e). The distinction between contributory and true causes is reminiscent of the Phaedo's distinction between the true cause and "that without which the cause could not be a cause" (98b-99c ff.). The true cause is teleological, while the contributing causes are the material conditions that allow the purpose to be fulfilled. That is the case here as well. It may seem surprising that the stranger begins by dividing "doing" from "protecting against," and looking for weaving under the latter. We can find it just as easily under the former, so it is odd that the distinction is made at all. It is evidently made in order to illustrate that an activity is best defined by its purpose, we might even say its "value." The reason that the sophist was never convincingly differentiated from the philosopher in the Sophist was that their most important difference, the difference in their goals, was never taken into account. They were treated as makers of products, and the purpose for which the products were made was ignored. Here, on the contrary, the purpose is insisted on right from the beginning of the division, and distinguished from mere making. Similarly, it is only because their ultimate purposes coincide that instrument makers and carders can be considered part of the art of clothesmaking. In a technical sense, the maker of the loom is not a clothing maker, nor is the person who cards fleece into woolen strands. But the stranger is now no longer concerned with technical definitions as in the Sophist , but with teleological, purposive, value-laden ones. In fact we might even distinguish the weaver from these others by calling weaving the most beautiful and greatest (

) of them, he says. Although there would be some truth (
) in this—a possibility that he might not have entertained earlier—it would not yet be a clear and complete distinction (281c-d). Accordingly, he expands the previous division, as shown in Figure 12.

Here again we find remarkable failures in the rigor and even usefulness of the division, signaled once again by the stranger's own words. There is a comparatively minor oddity in that the left-hand side of the division has to be taken a step farther than the right-hand side before the latter becomes fully intelligible, but the serious problem emerges in step 13. There the stranger says: "Of woolworking there are two divisions, and each of them is by nature a part of two arts" (252b). We would have expected him to say that each has two arts as its parts. To say that a species is part of two genera means that it is not unified, not



Figure 12

wholly subsumed within either genus (as was prefigured earlier by the split species of "birds," part of which was subsumed under "water animals" and part under "land animals")—in other words, that the division is incorrect, not made at the natural joints between genera. The stranger shows this in his otherwise unnecessary division of the left-hand species, "dividing." For both combing and carding turn out to be partly in one and partly in the other genus. This is explicit in the case of combing, about which the stranger says, "half of combing . . . belongs to the art of dividing" (252b; the other half belongs to the art of composing). We would expect this to be true of carding as well, since he had said that both divisions are parts of two arts, and that is in fact the case. Carding includes not only raking the fleece into a web to straighten it and remove the impurities[20] ("dividing") by rubbing it be-


tween fiat surfaces set with teeth, but also the subsequent condensing ("composing") of the web into a sliver, which can then be twisted and spun. So carding too, as the stranger remarked, is part of both arts. The reason that combing as well as carding is in both genera is that the weaver uses the comb partly to separate the strands and partly to combine the warp and woof. But since it is in the act of weaving that the comb is used, the ambiguity in combing attaches to weaving itself and undermines the definition. Carding for the most part separates, and weaving for the most part combines, but that is hardly an adequate way to distinguish them, and it certainly is not in keeping with the precision demanded by the method of division.

At this point the method of division by bisection, the "longer way," which has been used in an increasingly ineffectual way throughout the dialogue, becomes completely abandoned. The stranger will continue to make divisions, but rather than dividing by halves, he will divide immediately into the ultimate species—the very thing that he had warned young Socrates against at 262b. The reason he will give for this is that here division by bisection simply is not possible, but we shall see that that is no more true here than it was previously. A more convincing reason emerges from the ensuing discussion of the two different kinds of measure.

4. Relative Measure, the Mean, and Diaeresis (283b-287b)

"Why ever didn't we straightaway answer that weaving is the combining of the warp and woof, instead of going around in a circle and distinguishing very many things pointlessly?" the stranger asks (283b). On the basis of the foregoing discussion we can reply that one reason for the roundabout approach is that it locates weaving within a framework of purpose and value: weaving is for the sake of production. To define it merely in terms of its mechanical activity is to ignore what it is good for, and therefore its "true" cause or reason for being. Later the stranger will claim that "it might plausibly be said that whatever exists, among all that is, is an instrument or contributing cause of at least one thing" (287d). Everything, therefore, points toward something as its purpose. Looms are for the sake of woven cloth, woven cloth for the sake of warmth, warmth for the sake of health, health for the sake of life, and life (as the myth suggests) for the sake of embodying the divine. Only what is beyond being, the good itself, does not point beyond itself. The ultimate meaning of each thing is its value.


Although young Socrates does not feel that all this Was Unnecessarily lengthy or a waste of time, the stranger worries that he may come to think so in the future, and seeks to preclude this by explaining two different ways of measuring length.

STRANGER : One is with respect to the shared bigness or smallness of things toward one another. The other is with respect to the necessary essence of coming into being [


YOUNG SOCRATES : What do you mean?

STRANGER : Doesn't it seem to you that, in the nature of it, we must say that the greater is greater than nothing other than the less, and, again, the less is less than the greater and nothing else?

YOUNG SOCRATES : It certainly does.

STRANGER : But what about this? With regard to what exceeds or what is exceeded by the nature of the mean, whether in words or actions, must we not also say that it really exists? And that in this lies the chief difference between those of us who are bad and those who are good?

YOUNG SOCRATES : Evidently. (283d-e)

The "mean" (

) does not refer here to a mathematical mean (average), any more than it does in Aristotle's Nicomachean Ethics .[21] It refers to the correct degree between two indeterminate extremes such as too long and too short. If we wonder, then, whether a discussion, like the previous one, is too lengthy, we may answer either by comparing its length with that of other discussions, or by comparing it with an independent standard of moderation. Similarly, if we wonder whether individual people are good or bad, we may compare their qualities with those of some other person, or we may compare them with an independent standard of the mean. For Plato, as later for Aristotle, we can speak of good and bad, excess and deficiency, only in relation to such a mean. The concept of value has now been introduced in a fully explicit way.

None of the arts—including that of statesmanship—would be possible without such a measure, for any art necessarily presupposes a standard of goodness at which to aim (284a-b). This is in part what the stranger meant by calling the mean "the necessary essence of coming


into being." Not only human creation but, as the myth showed, all coming to be is ultimately dependent on an absolute or "divine" standard. The stranger goes on, however, to make the dependence between the mean and the arts reciprocal. "If this exists, they exist, and if they exist, this too exists. But if one of them does not exist, the other never will either" (284d). The mean cannot therefore be equated with the Republic's Idea of the good, whose existence has absolute priority and is unhypothetical. The mean is not goodness absolutely, but the measure of goodness in words and actions, hence inseparable from them. The confusion that we witnessed at the beginning of the dialogue, as to whether statesmanship is a practical or theoretical science, now seems to be a deliberate adumbration of the nature of the science of the mean, in which praxis and theôria are inextricably linked. In the earlier passage we saw that in step 1 statesmanship was intellectual rather than practical because it does not result in products (258d-e), while in step 2 it was directive rather than critical because it does result in products (260b). Now we see that it is indeed both, because they cannot be separated. The greatest science, the science of the mean, is double-sided: what we know and what we do are ultimately inseparable. The Socratic equation between knowledge and virtue reappears here in a more subtle form. We may assume that this is also the reason that acquisition and production, which were fundamentally distinguished in the Sophist , were recombined at the beginning of the weaver division (279c): knowing may be regarded as a kind of acquisition, and doing as a kind of production. The rule-free flexibility of the science of the mean corresponds to Aristotle's concept of practical wisdom (pbronêsis ),[22] that is, the capacity for discerning in practical situations the mean that embodies what is right or good. Plato is here treating of the individual application of the principles that were discussed only in a general way in the Republic .[23]


The distinction between relative measure and the mean has an important consequence for the method of division as it has been used here. In principle the method ought to be concerned with the mean, rather than with relative measure:

When one first sees what is common among many things, one should not retire until one sees within it all the differences, however many lie in forms. And again, on the other hand, when all kinds of dissimilarities are seen in a plethora of things, one should not be liable to get discouraged or stop until one has compassed all the related things within a single similarity and enclosed them in the essence of some genus.

In other words, the mean between the extremes of seeing everything either as an identical unity or as completely unrelated consists of discovering the same and the different as it really is. Nothing is said about dividing into halves. When the stranger had earlier recommended the longer way of bisection into equal halves, over the shorter way of immediately separating off the species as they really are, he in effect recommended relative measure over the mean.

Let us reconsider the episode of the longer and shorter ways, where the stranger, for the only other time in the dialogue, gives us alternative procedures. In a previously quoted passage he had said, "It is finest to distinguish what is sought from everything else immediately, if that cor-


rectly reflects how the things really are . . . [But] it is safer to make one's cuts by going down the middle, and one would more likely hit upon the boundaries between the forms" (262b). The dichotomy is precisely parallel with that between relative measure and the mean. "Going down the middle" means comparing the two species to make sure that they are of approximately equal size, that is, dividing according to relative measure. The true mean, however, would be the precise boundary that one is seeking. Since the mean is connected with values such as beauty (

), it is finest (
) to find the proper measure (mean) of a thing immediately, the one that "correctly reflects how the things really are."

But in this earlier passage the stranger recommends relative measure over the mean, because of its greater "safety." The tension between the earlier endorsement of the longer way, which corresponds to relative measure, and the later endorsement of the nonrelative mean illuminates both the particular peculiarity of the episode of the two ways, and the general peculiarity of the fate of the method of bisective division, which is first employed with increasing ineffectuality, and then entirely abandoned for no obvious reason. The episode of the two ways now seems to have been an intimation of the fact that the safe, bisective approach to division involves only relative measure and can never be fully adequate. And the progressive but unacknowledged failure of the divisions that follow shows how a mechanical application of the formal rules of division can lead to unsatisfactory results. As a confirmation of this, in the remainder of the dialogue the stranger will always use the shorter rather than the longer way. He will distinguish forms immediately into their infimae species instead of proceeding indirectly by the relative measure of making artificial distinctions designed to produce symmetrical halves. We shall see that this is not because such artificial longer ways cannot be devised. It must be, then, that the shift is meant to recommend the shorter way as ultimately preferable. The rigid but safe relative measure of the longer way turns out to be a pedagogical steppingstone to a subtler method of division. And the stranger's distinction between the two measures turns out to be the principle of division that bifurcates the dialogue itself into its two halves—in more ways than one. Not only does it occur directly in the middle of the dialogue, and function as the boundary that separates the stranger's use of the longer way in the first half from the shorter way in the second, but more important, it provides the differentia by which the difference between the longer and shorter ways can be identified.


That the longer way, the method of symmetrical bisection, should eventually be superseded should not entirely surprise us, for in the same passage where the stranger distinguished the longer and shorter ways, he said, with regard to his explanation of the method of division, "I must try to speak even more dearly, in cognizance of your nature, Socrates. In the present circumstances it is not possible to make it clearly evident" (262c). What is it that the stranger cannot yet explain to young Socrates because of his nature? The method of division, as the stranger has so far employed it, is evidently some kind of simplification. We will consider in the next chapter what it may be that the stranger is holding back here, but perhaps a due to the way that he accommodates the method to young Socrates' nature is to be found in the stranger's previously mentioned reference to mathematics at the beginning of the dialogue: "Shall we say that this, my dear Theodorus, is what we heard from our great calculator and geometrician? . . . You are placing equal value on each of these three, who are farther apart in honor than your art of proportions can express" (257a-b). Consider the reference to mathematics when the stranger divides the mean from relative measure:

It is dear that we should divide the art of measurement by cutting it into two, in accordance with what has been said. One section includes all those arts that measure number, length, depth, breadth, or thickness in relation to their opposites. The other includes all those that measure in accordance with what is due, fitting, timely, required, and everything rise that dwells at the mean and away from the extremes.

Thus, mathematics, like the longer way, is a species of relative measure. To consider the significance of this, let us recall the elder Socrates' first meeting with Theaetetus, young Socrates' classmate. Socrates asks whether Theaetetus has been learning geometry, astronomy, harmonics, and arithmetic from Theodorus, and Theaetetus replies that he has (145c-d). These were, as we saw, the first five of the six disciplines that, in the Republic , Socrates had said would turn our soul around from its imprisonment in the Cave and lead it to an apprehension of the good. But those five were to be no more than a preparation for the sixth, dialectic (Republic 531d). This, however, has not been part of their training. Not only is there no mention of it in Socrates' initial conversation with Theaetetus, but less than a page later Theodorus turns down Socrates' invitation to discuss the nature of knowledge with him, because "I am not accustomed to dialectic" (146b).

Theaetetus and young Socrates are to be thought of as talented mathe-


maticians who lack any acquaintance with dialectical thought. They are, in the language of the Republic , well versed in dianoia but not yet at the level of noesis.[24] Accordingly, it is appropriate for the stranger to address them at the level of dianoia, by means of a method that, like mathematics, follows postulates to their consequences but not to their origins. This is how mathematics and dianoia in general were described in the Republic (510c), and it applies as well to the longer method of division. The starting point of the division is never called into question as part of the method: one simply postulates it and proceeds to make symmetrical derivations from it. It seems that the stranger is providing Theaetetus and young Socrates (and that Plato is providing us ) with a multistage training in dialectic. As he says, "Why did we set ourselves the search for the statesman? Was it for the sake of this subject itself, or rather for the sake of becoming better dialecticians about all subjects?" "It is for the sake of all of them," young Socrates replies (285d).

The first stage of their training in dianoia picks up where mathematics leaves off. There is still the procedure of reasoning downward from a given starting point, but that reasoning now shifts from reasoning about quantity to reasoning about quality. It is perhaps meant to alert us to this that when Socrates uses the term "quality" (

) in the Theaetetus , he does not expect Theaetetus to know what he means.[25] And perhaps it is to enhance this shift from quantity to quality that Plato has both Socrates and the stranger use letters and syllables as examples of rational thought, whereas in previous dialogues he used mathematical examples. At this stage one learns to think in terms of "same" and "different," by making use of the relative measure of "equal halves" to arrive at contradictory categories. But the conclusions at which one arrives do not always turn out to be consistent or convincing. A standard of correctness needs to be added, then, to the simple concept of relative differentiation. This is the mean. But if we are indeed capable of discerning the mean, then the elaborate artificiality of the relativistic divisions of the longer way are unnecessary and even detrimental—as all the divisions of this dialogue have demonstrated. The longer way functions as a stepping-stone to the shorter, a step for which young

[24] Miller observes: "by presenting young Socrates as immersed in mathematics and not yet introduced to philosophy, [Plato] seems to address his students at just the point in their education where they must first turn from the one to the other" (PPS 9).

[25] 182a (it may also be that this term is first coined here, and that Socrates needs to explain it on that account). The contrast between quantity and quality would perhaps not look quite so simple to the Greeks, since for them the paradigm of mathematics was geometry, in which quantity is to a certain extent assimilated to quality (shape).


Socrates was not ready earlier. At this stage value enters into the picture once again, for correctness and truth are valued above error and falsity. Thus the method of division, which at the earlier stage had abstracted from value, is now itself justified precisely on the basis of its value: "By far the greatest and primary consideration is to honor [

] the method itself of being able to divide according to forms" (286d).

After this point, where Plato distinguishes the mean from relative measure, he will never again make use of the longer way, the relative measure of bisective division. Nor will any of his characters, including the Eleatic stranger, reiterate the stranger's ban on making distinctions based on value. On the contrary, just as the nature of the mean (unlike that of relative measure) is inseparable from that of value, the primacy of value will permeate the dialogue from now on. The stranger, as we have seen, defines the mean in terms of value : "good and beautiful" (284b), "due, fitting, timely, required" (284e).

To put the matter somewhat differently, there are two conceptions lacking in mathematical dianoia that need to be supplied if a transition to noesis is to be possible. One is the ability to reduce the indefinite multiplicity of the world to a synoptic order, by perceiving the world in terms of connected forms or kinds or essences rather than individuals (cf. Sophist 253d).[26] The other is the ability to discern the inner necessity of the existence of such essences, the fact that, as Socrates puts it in the Republic , they all spring from the nature of the good. Thus what is required is that the mathematicians learn to think qualitatively rather than only quantitatively, and eventually learn the teleological mode of thought implied by the myth and presupposed by the concept of the mean.

5. Division Without Bisection (287b-293e)

When the stranger moves from the "safe," longer way of the first half of the dialogue to the "finest" (262b), shorter way of the second half, he makes it seem that he is forced to do so by the nature of the subject rather than by his choice of methodology, but this claim is disingenuous. It is a stratagem complementary to the one he employed in the first half in regard to the superiority of the shorter way. Then he tacked a redundant step onto the end of the shorter way's definition to make the ap-


proach seem absurd ("featherless biped" instead of "two-looted animal"); now he claims that he is proceeding immediately to the final species only because these genera do not admit bisection. Both stratagems are attempts to minimize the temptations, to impatient natures, of the shorter way; but both pretexts fall away under scrutiny.

The stranger says: "The king, then, has been separated from most of his companions, or rather from all who have to do with herds. There remain, we say, the arts that have to do with the city itself, arts of both contributory causes and causes, which we must first separate from one another" (287b). So the following division picks up where the original one left off. We are now looking at the class of contributory causes and causes, of the episteme, intellectual, directive, and originary, of the production or nurture of two-footed, herd animals.

"Let us then divide them, like sacrificial animals, by their limbs, since we cannot cut them into two. For it is always necessary to cut into the closest number possible" (287c). He thus suggests that he is dividing them by the limbs only because division by bisection does not happen to be possible in these cases. But in the Phaedrus Socrates had introduced the method of division as "the ability to divide according to forms, at the natural joints; and not to attempt to hack off a part, in the manner of a bad butcher" (265e). The object of the method of division is not, then, to divide down the middle, but to divide correctly into the forms. So the reason that the stranger wants to divide at the natural joints is because this is what the method of division aims to do—and not, as he makes it seem, because symmetrical bisection, although inherently superior, happens not to be possible in this case. He never explains why bisection will not work here, but says only: "The reason, I think, will be no less evident as we proceed." We shall test this claim, and try to discover the reason that the stranger refers to, by attempting as far as possible to do what he says cannot be done, that is, to derive all the remaining specifications in the dialogue by means of bisective division. A reason why this cannot be done with complete success will, in fact, become evident, but it will be a reason that applies to all subjects, not only to this one in particular. As long as we are content with the previous standards of definition, we will be able to arrive at the remaining species by progressive bisection as easily as in all the previous cases.

In fact the genus itself is already bifurcated—"contributory causes and causes"—and the stranger tacitly divides it along those lines before he distinguishes the first species, instruments: "Whatever arts make any


instrument whether large or small for the city, all these must be classed as contributory causes" (287d). Is the genus, then, from which the further divisions are made "both contributory and primary causes," as the stranger said at 287b, or only "contributory causes," as he now says? Actually, it appears to be neither, for the stranger proceeds to the second species with the words, "All the same, let's say this about another kind of possession in the city" (287d-e), and from then on all these species will be classed as kinds of possession. So the stranger appears to have made a further silent bisection, this time of the class of contributory causes into those that produce possessions and those that contribute in some other way. It will later appear that this other class is that of servants.

The stranger lists a total of seven kinds of possessions that are contributory causes of statesmanship and need to be distinguished from the art of the statesman. These are instruments, receptacles, supports, defenses, playthings, raw materials, and nourishment. Compared with the divisions of the Sophist , an extensive list like this hardly seems like a division at all. But given the stranger's distinction between the longer way and the shorter way, we can see that he is now employing the shorter way, the immediate articulation of the genus into its ultimate species, without intermediate bisections. Had the stranger wished, however, he could easily have derived them by a bisective division like the one shown in Figure 13. In the summary (289a-b), the stranger mentions that raw materials (called there the "firstborn form,"

) should have been put first, which confirms something like the sequence in the diagram.

This supplied series of bisections is as rigorous as those of the Sophist , and more so than those of the Statesman ; so why did Plato not devise one similar or better? The reason cannot be that this subject in particular lends itself more naturally to division into seven, for even without the constraint of bifurcation the stranger mentions that certain classes have been left out and can only be included by force:

Whatever we have left out, if we have forgotten anything not very important, can be fit into one of these. Thus with the class [

] of coins, seals, and every other kind of engraved dies. These do not constitute among themselves a large genus with a common name, but some can be made m fit under "playthings," and others under "instruments," although the amalgamation is very forced. With regard to the possession of tame animals, except slaves, the previously partitioned art of herd nurturing will show itself to include them all.



Figure 13

The admission that some things have been left out and should be forced, not very naturally, into the classes mentioned is quite surprising: since he proceeded immediately to the ultimate species without intermediate divisions, all the stranger needed to do was add more species.

Since the present diaeresis was introduced as a continuation of the one with which the dialogue began, it is no problem to assign the species of tame animals to one of the earlier classifications as the stranger recommends (presumably they would be placed under step 6b, "land animals," as they were in the Sophist's division A, 222b-d), but it is surprising that he mentions them at all, since they are not produced by an art, a "contributing cause." It may be that, like slaves, they are mentioned even though they are excluded from this classification, because they too are possessions, even if not products of contributory arts. But in that case it sounds as though "possessions" is the genus, and "contributory arts" the differentia, rather than the other way around. The lines of derivation are becoming tangled. However, this does not mean that the subject itself does not admit of progressive bisection, but only that (as with the weaver diaeresis) the method has not been employed as rigorously as it was in the Sophist .


The stranger's next remark is ambiguous as well: "The class of slaves and servants in general remains, among whom somewhere, I prophesy, will become evident those who dispute with the king about this very weaving [i.e., the art of statesmanship]" (289c). Does this class "remain" because all other species of possessions have now been separated and it is the remaining species of possessions, or because it remains outside the class of possessions altogether? One would expect the former, because it would seem pointless for the stranger to have taken the trouble to divide possessions into seven species if he were going to look for his quarry elsewhere; but, on the other hand, slaves are no more "produced by contributory arts" than are tame animals, and the other members of the class of servants are not possessions at all. Rather, they belong in the class that I earlier suggested is the counterpart to that of "contributory arts of production" within the genus of contributory arts generally, that is, the class of servants. Presumably that is why the stranger, although he had in the previous speech mentioned slaves in the context of possessions, now speaks of them in the context of "servants in general."

Again without bisection, the stranger immediately proceeds to list (i.e., divide according to the shorter way) five kinds of servant: slaves, distributors, public servants, diviners, and priests. This time, not only is the list capable of being derived by progressive bisection, but the very order in which it is given is what would most naturally follow from such bisection (see Figure 14, which also includes the implicit preliminary bisections that we have already noted).

Another ambiguity attaches to this representation, however. Are we right to put the class of servants under contributory as distinct from primary causes (B), or does it proceed from the wider class, that of causes generally (A)? If it belongs under A, then why did the stranger bother to distinguish the two kinds of cause and classify possessions as a species of contributory cause rather than of cause generally? But if it belongs under B, then all the subsequent distinctions are going to be within the class- of contributory rather than primary causes, in which case we may never be able to find the statesman at all; for the stranger seems to suggest that the statesman is the primary cause of the city, as the weaver is the primary cause of weaving (287c-d). In that case the stranger might as well say now that the true differentia of the statesman is that he is the primary cause of the city whereas the others are contributory causes. It will turn out, however, that the class of servants includes rulers, such as kings, and so apparently contains primary causes



Figure 14

as well as contributory ones. It seems that the class of servants belongs under A after all.

As with the weaver division, and to a lesser extent the shepherd division, the lines of division are badly tangled. However, that is not because the subject itself resists bifurcation more than any other subject. Plato could have made these distinctions as cleanly as any of the previous ones, but there was always an artificiality about that cleanliness, a pretense of definitive rigor, which is now dropped. The tangled lines of derivation illustrate an important problem in the longer way. All things are related to one another in a multitude of ways—there can be no definitive conceptual map of reality—and yet the longer way arbitrarily insists on a single, thoroughly determined path of derivation. The present tangles arise because the stranger abandons the artificial tidiness of the Sophist's divisions and exhibits the multifariousness of relationships


by introducing more than one line of derivation at a time: tame animals can be classified under animals or under possessions; slaves are species both of possessions and of servants; and the statesman will turn out to be in one way a master but in another way a servant, in one way a primary cause but in another way a contributing cause (cf. 287d: "it might plausibly be said that whatever exists, among all that is, is an instrument or contributing cause of at least one other thing"). The difficulty in finding the statesman through just one line of derivation was prefigured in the shepherd division, where birds are found under two-looted animals, land animals, or (in the Sophist ) water animals; and in the weaver division, where not only were both combing and carding parts of two (contradictory) arts, but weaving itself turned out to be a species of clothesmaking that was only partly concerned with making clothing.

Now that they have separated these off, however, another very large crowd becomes visible:

Many of the men resemble lions and centaurs and other such creatures. Whereas very many others resemble satyrs and other weak and cunning beasts. And they quickly exchange their form [

] and power into one another.

These turn out to be "the greatest impostors of all the sophists," and in order to distinguish them from the statesman, the stranger distinguishes different types of government (291c-292a), for it will turn out that the impostors are the monarch, aristocrat, democrat, oligarch, and tyrant. The types are distinguished initially by whether the ruler is one, few, or many. The stranger then bisects the first two of these, monarchy and oligarchy, although he does so not on his own initiative. He merely reports that "people today" employ distinctions such as enforced/voluntary, poor/rich, and lawful/unlawful. The result is five types of government: monarchy and tyranny, aristocracy and oligarchy, and democracy (which is called by the same name in both cases).[27] It is unclear why the stranger confuses the issue by including the distinctions enforced/voluntary and rich/poor, since the only important differentia

[27] Benardete 3.138 notes an interesting parallel: "The number of common names for the kinds of regimes . . . is equal to the number of disguises in which the sophist showed up in the stranger's divisions in the Sophist . The parallel extends even farther. Not only is the one name of a regime (kingship) the only possible name for the correct regime, just as Socrates' cathartics is the one sophistry noble in descent, but the double form of democracy, which hides under a single name, is exactly parallel to Theaetetus' mistaking the double form of the stranger's third division for the third and fourth." The conclusion that Benardete draws from all this seems somewhat anticlimactic, however: "Young Socrates learns that the city is the natural locus of all sophistry."


turns out to be lawful/unlawful, and the other two pairs are not coextensive either with lawful/unlawful or with each other. King and tyrant, aristocrat and oligarch, all are likely to be rich, and their subjects may be ruled equally involuntarily. The stranger's only stated reason for adding the other differentiae appears to be that at the time people made distinctions on the basis of such criteria (291e), and bringing the criteria into the discussion enables him to dismiss them. In fact he will go on to dismiss all three "fashionable" criteria as irrelevant to the definition of the statesman (298a-299b), but the lawful/unlawful distinction turns out at least to be relevant to distinguishing other kinds of rulers (301a-c). It is worth noting, however, that the other two criteria, force and wealth, correspond to the lower levels of the tripartite soul. They therefore represent ways that the dictates of reason can be subverted.

Since statesmanship was agreed to be a kind of science, "the boundary between the statesman and his rivals is not 'few' or 'many,' nor 'voluntary' or 'involuntary,' nor 'poor' or 'rich,' but some science" (292c). And indeed to the extent that we attain to this science the symmetrical technical divisions according to the longer way (few/many, voluntary/involuntary, poor/rich) become inadequate and irrelevant. The only criterion that really matters is that of goodness:

As long as they employ science and justice and preserve it [the city], changing it from worse to better as far as possible, this must, at that time and by those measures, be for us the only right regime. All the others, we say, must be said to be not genuine nor really existent [in their own right], but imitating this one. Those that we speak of as having good laws imitate it more nobly, and the others more basely.

So instead of the five forms of government that "people today" distinguish by relative measure, there is really only one, determined by the mean (301a). As far as its imitations are concerned, distinctions of value are now anything but irrelevant to the stranger's method of definition. That was already implicit in earlier passages, but now that he has replaced the value-free method of division by relative measure (equal halves) with the value-grounding mean, in the above passage it becomes in a sense the only relevant distinction.

6. Distinguishing the Statesman (293e-305e)

The stranger had said that if someone possesses the science of the mean, the distinctions between few and many, voluntary and involuntary, rich


and poor are irrelevant (292c). A more controversial claim, that it is also irrelevant whether such a ruler rules with or without law, is with caution introduced more gradually. It is not mentioned at all in the passage referred to above, and at 293b it is only hinted at with the words, "in accordance with writings or without writings." Not until 293c does the stranger specifically mention the irrelevance of law, at which point young Socrates responds by saying only, "Fine" (

). The implications do not sink in until the stranger's next speech, in which he says that, as long as it is for the good (
) and in accordance with science and justice, the statesman may undertake such actions as killing or exiling citizens (293d-e). Only at this point does young Socrates reply, "The rest of what you said, stranger, seems moderate [
]. But the statement that it is necessary to rule even without laws sounds rather difficult" (293e). Even at this late point the stranger feels that young Socrates is rushing him: "You anticipate me a little," he says.

However outrageous the claim may seem, it follows necessarily from the doctrine of the mean, for the mean is not something that can be formulated into writings. Just as the flexibility and responsiveness of oral teaching makes it in principle superior to written teaching (Phaedrus 275d-277a), and the greater pliancy of the shorter way makes it ultimately preferable to the rule-bound longer way, so too is the science of ruling (like Aristotle's phronêsis ) superior to any set of laws:

A law would never be able, by comprehending accurately what is best and most just for everyone at once, to enjoin what is best. For the dissimilarities among human beings and actions, and the fact that nothing is ever, so to speak, at rest in human affairs, do not allow any art to declare a simple rule in any case regarding all people and for all time.

In this respect the statesman is like a physician or ship's captain, and the stranger shows at length the absurdities that would result from having these two professions supervised by the rich or the many, or from subordinating them to strict written laws of practice. Young Socrates is persuaded: "It is clear that all the arts would be completely destroyed" (299e).

The stranger does, however, address young Socrates' concerns about rule without law, by distinguishing between two kinds of lawlessness. Although rule by law is inferior to rule by science, it is superior to the abrogation of law for the sake of profit or individual favor. One who overrides the laws without doing so on the basis of science does so


on the basis of ignorance and desire (

, 301c). As in Republic Book 9, this embodiment of desire is the tyrant. On the basis of the above distinction, the stranger divides all the unscientific forms of government first into three—rule of one, few, or many—and then each of these into two: with law or without law. It seems at first that he is reverting to the technical, value-free kind of distinctions characteristic of the Sophist , but then he proceeds to rank them in the order of their relative goodness (302b-303b): "Monarchy, then, when harnessed within good writings [
], which we call laws, is the best of all six. But without law it is difficult and the most oppressive to live with" (302e). Nothing is said about monarchy with bad laws. The reason for this appears to be that only "good writings" are called laws. As attempts to imitate science, laws are good to the extent that they succeed in imitating it; and to the extent that they fail to imitate it, they cannot be called laws at all.

When a government is lawful, the stronger and more effective it is, the better; and when it is lawless, the reverse is true. The strongest government is monarchy, followed by oligarchy and democracy. Accordingly, when governments are lawful, monarchy will be best, oligarchy next, and democracy third; but when they are lawless, democracy will be best, oligarchy next, and monarchy (tyranny) last. This is the order of goodness of the six normal constitutions, but the statesman himself is a god as compared with all of them. Since the other six merely "preside over the greatest images, they themselves are such. And, being the greatest imitators and impostors, they turn out to be the sophists of sophists" (303b-c). In this way, the stranger concludes, we can distinguish the statesman from the centaurs and satyrs that were confused with him earlier (at 291a-b). Since satyrs were there called weak and cunning, and since they were traditionally associated with the appetites, they may perhaps be meant to indicate the lawless species, which follow desires . Centaurs, grouped in the earlier passage with animals like lions, are what the Republic classes as spirited , and function there as upholders of the law. So perhaps we may take these terms to refer to the lawless and law-abiding species, respectively.

More important, the stranger has now succeeded in doing what he failed to do adequately in the Sophist : he has provided a convincing criterion for distinguishing between the genuine knower and the sophist. It is no longer a question of what kind of images (

) they produce in their words—whether accurate (likenesses) or inaccurate (semblances). We saw that in the Sophist many of the problems initially ascribed to


the concept of "semblance" ended up being spoken of in terms of "likeness" and "image" generally, although the stranger never called attention to this in a thematic way. Any image is always at a remove from reality itself. Accordingly, here in the Statesman the stranger no longer castigates the sophists because they produce inaccurate images, but because they produce images at all , whether accurate or inaccurate: "Because they preside over the greatest images [

], they themselves are such. And, being the greatest imitators and impostors, they turn out to be the sophists of sophists" (303b-c). They do not fall short because their images are inferior. On the contrary, they preside over the "greatest images" and are the "greatest imitators"; even the monarch who rules "with good writings" is not exempted from the stranger's attack. The problem is that they provide only an image, an imitation, of the principle of goodness; and images, because of their inflexibility, necessarily petrify and distort the unformulaic nature of goodness.[28] The statesman, by contrast, who possesses the science of the mean, embodies goodness within himself and does not need to distort it in inflexible formulations. As we saw in section 10 of the chapter on the Theaetetus , wisdom is not only a matter of intellectual knowledge (i.e., images of reality), but a way of being: at the highest level, what we know and what we are coincide, in precisely the same sense that Aristotle would later argue that the good in particular cases can be known only by a good person. Moral forms can be grasped only as far as we can experience moral truth within ourselves, and this means freeing ourselves from attachment to the pleasures of appetite and ambition. Precisely this is impossible for the sophist, whose enterprise is inseparable from these passions (as the preliminary divisions of the Sophist reminded us). The mark of the science of statesmanship, then, which distinguishes the statesman from impostors, is the same as the mark of any true science or techne: it is the ability to discern the mean, the best course.


It is no accident that the weaver division lumps together the acquisitive/productive distinction (279c) that was so central to the Sophist , for sophists are no longer to be understood only in terms of acquisitiveness, as in the Republic and earlier dialogues, nor in terms of their products, as in the Sophist itself. They are rather those who imitate the one who knows, but without the science of the mean, by which the latter knows what is good. It is not that the imitations (images) produced by sophists happen to be merely semblances rather than likenesses, but that, lacking the science of the mean, it is impossible that their products be otherwise. Just as the Sophist completed the project of the Theaetetus by introducing the conception of universal kinds, the Statesman completes that of the Sophist by introducing that of goodness.

Once again the stranger sees a group from which the statesman still needs to be separated. The task is compared to the refining of gold (303d-e), the separation of more precious from less precious metals, what is of greater value from what is of lesser value. The baser impurities have already been removed, so that everything that remains (types of political episteme) is valuable (

). These are the arts of the general and the judge, and the royal rhetoric that is in the service of justice (303e-304a). "By means of music we must attempt to make [the statesman] clear," the stranger says (304a). When it is a matter of learning music, or the handicrafts in general, there is always another science by which we can decide whether we ought to learn these or not, and that therefore governs and rules over them.[29] Thus the science that decides whether to use rhetoric or force takes precedence over the science of rhetoric, and the one that decides whether to make war or peace supervenes over the science of making war. These evaluational roles belong to statesmanship, which, in a different way, supervenes over jurisprudence as well, for the judge is bound by the statesman's laws (304b-305d). The highest science is thus distinguished from the technical sciences by its status as a science of valuation.

In the Republic (332e-333d) Socrates argues that justice does not in itself do anything; at best it governs something else that is an activity. It we insist on specifying a function for justice, it will turn out to be something absurd, like the guarding of money when the money itself is not in use. Here, too, the stranger has separated the statesman's science of justice from any specific kind of action. It is the art that decides when

[29] Why did the stranger single out music in particular? Is it a reference to music's role as a sensuous symbol of the divine?


and how the others are to be used: "The art that is really kingly should not itself act, but should rule over those that are able to act, recognizing the right time and wrong time to begin and urge on the greatest things in the cities. And the others should do what it orders" (306c-d). Statesmanship thus weaves together the other arts (306e). Just as the weaver oversees all the arts that are contributory to weaving and exist for its sake, the statesman oversees all these contributory arts of service (except the "imitators"), for they all exist ultimately for the sake of the polis. The polis is their purpose, their reason for being, and therefore the ultimate referent of their value.

At this point the statesman has been fully distinguished from his rivals. Here, as earlier, we can see that although the stranger pretended that bisective division was no longer possible, it could well have been applied if he had chosen to do so. Figure 15 (which incorporates the earlier bisective classification of "servants") both illustrates how this might have been done and gives an overview of the statesman in relation to his various rivals.

The statesman, it seems, is ultimately found within the form of "servant," although this is never stated precisely. The stranger says only: "But I don't think I was seeing a dream when I said that somewhere in this realm would appear those who most of all dispute about the claim to statesmanship. However it would seem to be extremely strange to seek them in some subservient [

] part" (290b-c). It is not dear whether "those who most of all dispute about the claim to statesmanship" includes the statesman himself or only his rivals; but since the statesman is distinguished from his rivals only within the last genera, he must be present in the same class to begin with. A little later the stranger says that the priest's and diviner's arts "both are somehow parts of the art of service [
] . . . . So finally we seem to me to have caught on to a trail that leads where we are going" (290d). The traditional king will be found in this region as well (290e-291a ff.); and if he is a servant, then so must be the statesman. That is not surprising, since the ruler is in the service of the state (the only true masters, as the myth showed, are divine: cf. 274e-275a). The Republic makes it clear that any practitioner of an art or science serves what the science is about (1.342c-d).

Whether the class of servants is, in turn, to be found within the class of causes in general ("contributory causes and causes") or only contributory causes was, as we have seen, ambiguous. The ambiguity is appropriate, for although on one hand it is clear that the statesman and



Figure 15.


those whom he supervises are related as primary cause to contributing causes, on the other hand everything was said to be a contributing cause in some sense (287d). According to the Phaedo the only true cause is the purpose, the good, to which every other kind of cause is merely instrumental (98b-99c). In the same way the statesman, although in one sense the author of his instructions, is in another sense an acolyte of the mean, and his activity is the transmission of the good into the city.

This final diaeresis is problematic in another and more fundamental way as well. By distinguishing statesmanship as a science that supervises rather than acts directly, the dialogue has in fact come full circle back to the opening steps of the first diaeresis, where the science of statesmanship was differentiated from other sciences as being intellectual rather than practical (258d-e), directive rather than critical (260b), and originary rather than transmissive (258d-261a). Some of the same language is even used. Statesmanship was distinguished from the productive or practical arts (

, 258d-e) in step I of the shepherd diaeresis, and is now distinguished from arts like those of the general, rhetorician, and judge in the same terms (
, 305d)—even though such arts should not be able to appear within the region that remained after step 1. In step 2, where the statesman was classified as directive, the word used (
, 260b) is a form of the term for "to order" (
); here another form of the same word is used when the stranger says that the practical arts must follow the orders (
, 305d) of the statesman.

How is it possible that a division that begins as an explicit continuation of the shepherd division (287b), ends up by subsuming within itself the opening steps of its predecessor and thus giving birth to its parent? This phenomenon, of earlier differentia reappearing within later species that ought to exclude them, was the besetting sin of the beginning and end of the shepherd division (step 2, and step 9 of the shorter way), and its multiple appearance there now seems to have been a foreshadowing of its grander manifestation in the whole sweep of the dialogue: not merely as a redundant step (as in the shepherd division) but as a great circle. Earlier the stranger had spoken of "going around in a circle" (

, 283b) with reference to the weaver division. The weaver division was not in fact circular, and the stranger's phrase was nothing more than a hyperbolic metaphor for taking a roundabout route. But it can be taken quite literally as a foreshadowing of the fate of the statesman diaeresis, of which the weaver diaeresis was said to be a paradigm.


This circle is the most extensive manifestation of the tangles that we have witnessed throughout the dialogue, and that bear witness to Plato's belief that no definitive mapping of reality is possible. The multiplicity of ways that things are related to one another means that any such map will be characterized by arbitrariness or inconsistency, or both.[30] The longer way of diaeresis attempts to produce just such a map, but that technique can never be final, however much it can be useful. Even in the Sophist we saw that the sophist can be defined in many ways—ways that lead to very different conceptions of his nature—depending on our starting point and how we divide up the territory. We should not forget that the Phaedrus , where the method of the longer way is first introduced (265e-266b), is also the dialogue that most strongly warns of the impossibility of adequately capturing reality in writings (275c, 276d). Rather than there being a single hierarchy of precedence and subordination, of presupposition and entailment, the overarching circle of the Statesman illustrates that any place that we begin can lead us to any other place, once we know what we are doing.

Plato was never one-sided: he always recognized the importance of methodological rigor, but he also always recognized its limitations. Methodical rigor is important for leading us out of the Cave, up from eikasia , the lowest level of the Divided Line, through pistis , and to mastery of deductive thinking, dianoia. But when it is time to pass from dianoia to noesis we must pass from the rigidity of deductive rigor to something more flexible and fluid. Just as the rule of the statesman supersedes the rule of law because laws are too rigid to apply in a nonarbitrary way to the realities of life, so too the fluidity of the shorter way of thinking, guided by the mean, supersedes the rigidity of the longer way, guided by the artificial rigor of relative measure. The relationships within reality are too intricate to be captured in a formal way. In one way or another the limitations of formal and systematic theorizing, such as were exposed in the first part of the Parmenides , will always reemerge.

7. The Nature of Virtue (305e-311c)

"That a part [

] of virtue is opposed to a form [
] of virtue is very easily attacked by those who dispute about words, according to the

[30] Cf. Joseph Novak, "Plato and the Irrationals" (Apeiron Vol. XVI [1982] pp. 71-85 and Vol. XVII [1983] pp. 14-27): "Plato is using the operations with irrationals as a model for establishing divisions which are to terminate in definitions . . . [J]ust as the exact numerical value of an irrational quantity and the nonterminating algorithm, so a Form can never be definitively articulated on account of enormously complex relationships revealed by a division . . . [A]t 266A the Stranger asks Theaetetus and young Socrates to make a division . . . 'by the diagonal and again by the diagonal of the diagonal' . . . This reference to division by irrationals, beyond the obvious pun, probably makes reference to the way irrationals function as a model in division" (21-2).


opinion of the many," the stranger says.[31] Nevertheless, he continues, it is dear that courage and moderation are often, in a certain way (

), fiercely opposed to each other (306a-b). The qualifying phrase "in a certain way" is important; but it tends to be overlooked by those who see in this passage a divergence from the doctrine of the unity (or at least compatibility) of the virtues, found in the Phaedo and Republic . In fact there is no divergence except nominally (i.e., for "those who dispute about words"). The doctrine continues to be maintained in the Laws (963c-965d). In the Phaedo Socrates distinguished between the nature of courage and moderation when they exist in their ordinary sense and when they follow from philosophical purification (68c-e). Only in the latter sense are they truly virtuous and unified. And in Book 4 of the Republic they are compatible only when they are defined in a very sophisticated way. Someone is courageous

when his spirited part preserves, through both pains and pleasures, the instructions of reason about both what is to be feared and what is not . . . . And is he not moderate by the friendship and consonance of these same parts [the rational, the spirited, and the appetitive] when the ruler and its two subjects agree that the rational part should rule and they do not rebel against it?

This conception of virtue, like the more specialized one in the Phaedo , is meant to reflect a philosophically purified condition, one in which reason rules. It is not intended to deny that in normal parlance what we mean by courage is fearlessness and what we mean by moderation is restraint, nor that these two are normally in opposition. On the contrary, the Republic too affirms that "a gentle nature is opposed to a spirited one" (375c).[32] It is with these normal senses that the discussion in the Statesman begins, and the discussion will, like its predecessors, point to a refined form of these virtues in which they are compatible

[32] Cf. Guthrie 5.191.


with each other. That they are not meant to be ultimately incompatible was already dear at the beginning of the trilogy, when Theodorus recounted his surprise at finding them united in Theaetetus—something that he had not expected to be possible (Theaetetus 144a-b).

Even in their normal senses courage and moderation are virtues, although not unqualifiedly so. Courage is not the same as recklessness, nor is moderation the same as apathy. They are virtues because they are good, even in their unpurified state. Statesmanship, like any other art, can create good products only by using good materials. Just as the other arts discard bad materials and use only the good ones, statesmanship discards "those characters that lack the power to share in courage and moderation, and whatever other qualities tend toward virtue" (308e). The others, which "tend toward godlessness, violence, and injustice, carried away by the force of a bad nature, it removes by punishment with death, exile, and the greatest dishonors [

] . . . . And again those who wallow in ignorance and extreme abasement it yokes to the genus of slavery" (309a). However things may stand with the art of diaeresis, it certainly could not be said of the art of statesmanship that "it honors [
] all equally" (Sophist 227b; cf. Statesman 266d).

Even though courage and moderation in their normal sense are good, and therefore virtues, they are not completely so. In the strict sense they can be said only to "tend toward virtue." The moderate type "lacks drive and a certain sharp and active quickness," and may even be simpleminded; while the courageous type "is lacking in justice and caution" and "inclines towards brutality" (309e, 311a-b). The statesman's job will be to weave these two natures together as the literal weaver does the warp and woof (and in a way analogous to the philosopher-king's weaving together of the appetitive and spirited natures), and thus remove their initial incompatibility.[33] Since the human soul comprises both a divine and an animal part, there will be both a divine bond and a human one (309c). By the divine bond the stranger means "that when the really existent true opinion, together with constancy, about the beautiful, the just, the good, and their opposites comes to be in souls, it comes to be as divine [

] in a godly [
] race" (309c). When

[33] Benardete 3.112 suggests that "since the warp, or vertical threads, which is kept distinct on the loom, stands to the woof, or horizontal threads, which must be packed tightly together, as courage to moderation, the stranger seems to imply that moderation must be adapted to courage and not vice versa."

Paul Friedländer writes: "Is not the art of weaving as practiced by the true statesman a copy of the world 'bound together' by the law of proportion (Timaeus 31b et seq .), the 'living garment of the deity'?" (3.290-91). Cf. his n. 24 (3.528) on the theme of the "'world fabric' . . . in the beliefs of different peoples."


this is present, virtue is consummated, and the courageous type "is made genre and is most willing to share in the just things," and "the orderly nature . . . is made really moderate [

], and wise" (309d-e). It is dear then that the kind of moderation that is opposed to courage is not "really moderate," and the present analysis does not contradict in spirit those of earlier dialogues. As in the Republic , virtue in the fullest sense is possible only in conjunction with truth about the good and beautiful.

The human bond will be intermarriage—an ironic conclusion to a dialogue that began by defining humanity as noninterbreeding. Instead of marriages being arranged as they now are for the sake of wealth or power (310b)—that is, the lower levels of the tripartite soul—they will be arranged for the sake of virtue. By interbreeding the courageous and moderate types, the statesman will hope to produce fully virtuous offspring in which both qualities are combined. The trilogy ends as it began, with the extraordinary nature exemplified by Theaetetus. It is not impossible that the Theaetetus's dramatic emphasis on parents and offspring is a foreshadowing of this—but only in part, since that function would be unintelligible before the trilogy was complete.

The two types would not normally marry—are naturally noninterbreeding—since like is attracted to like; but their mutual repulsion can be overcome by means of the divine bond. If both have the same opinion about the beautiful and the good, then they will no longer be incompatible. Thus the statesman weaves them together by "co-opinions, that is [

], honors, dishonors, opinions, and the giving in marriage of pledges to one another" (310e). This bonding of the opposite tempers within the higher yoke of the state is a political version of the epistemic method of hypothesis. Each type, by itself, relies on a thesis whose consequences are in some sense disharmonious. The view that raw courage is good cannot be reconciled with the fact that it leads to brutality, injustice, and even madness (310d). And the view that simple moderation is good cannot be reconciled with its tendency to lead to passivity, simple-mindedness, and even to sloth and mutilation (310e). The one-sidedness of each hypothesis is overcome by a higher—truer—hypothesis or opinion about the good, according to which each of these appears only as a partial consequence. Here, as in the Phaedo and Republic , the journey can come to an end only with a vision of true goodness. The method of division and the method of hypothesis converge in the form of the good: the latter by the overcoming of incompleteness, the former by the shorter way of dividing by the mean rather than the average.


Chapter Five
An Overview

1. The Trilogy and the Problem of Value

After the critique of the Parmenides , the theory of forms has been re-established in its essential features by means of an implicit use of the method of hypothesis throughout the Eleatic trilogy. The method operated, on one hand, within each dialogue: in the Theaetetus , Theaetetus's hypotheses about the nature of knowledge were progressively superseded by means of the hypothetical method's procedure of testing consequences and seeking a higher or more inclusive starting point; in the Sophist the same was true of the hypotheses of the monists and pluralists, and of the friends of the forms and the materialists; in the Statesman the method of hypothesis led to the replacement of the statesman-as-shepherd hypothesis with the statesman-as-weaver hypothesis, and the mutually exclusive hypotheses about virtue, of the moderate and the courageous citizens, were replaced by one that reconciled their polarity. On the other hand, the movement among the dialogues, too, was characterized by the dynamic of the method of hypothesis: the Theaetetus foundered in an attempt to provide an account of knowledge without reference to anything but sensory particularity; the Sophist resolved this difficulty by means of its theory of value-free universal "kinds," but without being able to distinguish philosophy in its truest sense convincingly from the imposture of sophistry; and the Statesman accomplished the latter by means of the reintroduction of value, especially in the form


of the "mean." Thus the method of hypothesis continues to function, as it had in the Phaedo , as an indirect approach (

) in the progress toward an understanding of the nature of the good.

In the Theaetetus at its most radical, we were plunged into the "sea of dissimilarity,"[1] the sensible world's welter of particularity, where not only are universal kinds absent in favor of nominalistic individuality, but even individuals collapse into rivers of ceaseless change. All that exists are motions, not even individuals, and even less universals. It is a world familiar to us also from the second part of the Parmenides , where unity collapses into the fathomless multiplicity of the apeiron . Rescue is possible only if "the same" can be found amid "the different." Accordingly, the importance of the philosopher Parmenides, the apostle of the Same, is stressed in the Theaetetus despite the absence of any discussion of his work. The positive discernment of the Same amid the Different is the thesis of the Sophist . Indeed, the difficulty for the Sophist is no longer how to account for sameness, as in the Theaetetus , but how to account for difference, "nonbeing." The discernment of sameness is the role of collection, which first affirms the integrity of individual instances, and then affirms the objective reality of the qualities common to diverse individuals. Questions that could not be answered without hypothesizing the reality of the Same can now be answered as a consequence of that hypothesis, even though the Sophist makes no explicit reference back to the aporiae of the Theaetetus .

Today we might oppose the Sophist's line of argument by appealing to the familiar distinctions between metaphysical and epistemological, or linguistic, form, that is, between objectivity and subjectivity. However true it may be that we are constrained to organize our experience into sortal categories, this fact may tell us something not about the working of reality in itself, but only about that of our oven mind, or perhaps not even that but only about the way language works. And there are some who challenge even the necessity for language to work in this way, who emphasize as much as possible the way sameness is permeated by difference even within linguistic structures. Against such challenges no definitive defense is possible. Plato may point out that without universal referents neither logos nor knowledge is fully possible, but Kant and his successors (like Protagoras but unlike Theaetetus) willingly accept the unattainability of "objective" (noumenal) knowledge. Knowledge today is readily acknowledged to be like a con-

[1] Cf. Statesman 285b.


struction, or even like a conversation, but not like a mirror. Logos itself is not an image of something beyond itself, but more like a pharmakon , a drug.

If such objections cannot be refuted, however, they can be deprived of the appearance of primacy, once it is pointed out that the objector assumes no less than the defender. Plato, it is true, assumes that our urge toward conceptualization reflects the structure underlying reality itself. This is, in fact, a deliberate hypothesis (Phaedo 100b), for he is well aware of the limitations of demonstration. The attempt to overcome the hypothetical nature of thinking is a rewarding but uncompletable task. Moreover, as the Statesman emphasized, when we try to reflect on the highest, most elusive things, analogical thinking (paradigms) becomes indispensible, and so philosophical speech carries with it an unavoidable metaphorical aspect, the limitations and dangers of which were vividly displayed in the Parmenides , and reaffirmed in the stranger's condemnation of all image making as a form (however unavoidable) of sophistry (Statesman 303c).

But if Plato cannot avoid making certain crucial assumptions about the correspondence of the forms of language or thought to those of reality, it is no less an assumption when Kant or his successors insist that our epistemological or linguistic imperatives do not correspond to the constitutive principles of reality. The hypothesis that Kant calls his "Copernican revolution" is defended only by contrast with the empiricist hypothesis, and holds no advantages over the realism of the classical philosophies.[2] The same is true of the more radical post-Kantian positions. The success of modern science has allowed it to dictate the criteria of knowledge to a degree that almost no thinkers have been able to escape. Even those who least accept the goals of modem science find themselves allowing it to define the terms of what is possible. Descartes (who does accept the goals of modem science) captures this spirit in his resolve to count as false whatever cannot dearly and distinctly be shown to be true. It follows from this that if our "innate ideas" cannot be proven to be true of independent reality, they must be regarded as false to it and merely subjective. It eventually follows from this principle that if language cannot be proven to mirror the world in itself, then it must be regarded as mere conversation, inscription, and potion.

[2] See my "First Philosophy: Metaphysics or Epistemology?" Dialogue 11 (1972) 1-22.


Even Nietzsche's anti-Cartesian rejection of the concept of "truth" rests on this Cartesian foundation: truth is impossible because certitude is impossible.

However, these argumenta ad ignorationem were explicitly rejected by Plato and Aristotle.[3] Those thinkers were not unaware that their own positions contained indemonstrable assumptions, but they were more aware than many of their successors that the alternative positions were no less hypothetical. If Plato's position in part reflects a choice that he made and can never justify with mathematical rigor—a choice about what difficulties he was willing to live with in order to resolve others—the same is no less true of his rivals, ancient and modem. Such a choice is not merely a matter of "aesthetic" preference. The questions of whether there are in fact structural forms by which we organize our experience, and whether such forms are to be regarded as revelatory of the nature of Being or only of the nature of thinking, might be regarded by some as combining undecidability with inconsequentiality, and therefore academic in the worst sense of the word. How we answer these questions would then rest on nothing more stable than the aesthetic taste of each individual. If, however, the forms are for Plato noumenally constitutive, and for Kant only phenomenally constitutive, and for others merely historical constructions, these divergences are not self-contained hypotheses that are therefore inconsequential. They fundamentally affect our attitude toward the world. Plato's thesis is a powerful advocacy for the priority—the higher value —of the intelligible over the sensible, and for a commitment to increased spirituality and selflessness. A nominalist position, on the other hand, which takes only individuals to be real, leads more naturally although not inevitably toward egoism. Alternatives such as spirituality and materialism, altruism and egoism, are reflections and consequences of our beliefs about the nature of goodness, and its presence or absence in the nature of reality. As such, they have important implications for the nature and basis of morality. They are anything but inconsequential or academic. They are aesthetic only in the sense that the beautiful is the symbol of the good.

Those who are inclined to dismiss the Platonic forms as—at best— regulative principles of the mind rather than structurally constitutive principles of reality may well make the same charge against the purported reality of the good, especially when it itself is called an Idea. But

[3] For example, in the Theaetetus (184e-156e) and De Anima (3.4.429 18f.).


in the case of the good it is especially clear that more is at stake than choices based on nothing further than personal taste. The entire fabric of morality is at stake.

Pragmatists, and others who deny any objective truth to the concept of goodness, argue that it is meaningless to speak of the real (noumenal) nature of things. We can know reality only insofar as it enters into our consciousness, and whether the reality that we experience cognitively corresponds to some mind-independent reality is irrelevant. If, as Peirce argued, the meaning of a theory is the practical difference that it makes, then the correspondence theory of truth is meaningless: there is no way to see whether objective reality corresponds to our representations of it, so it is irrelevant to our behavior, and consequently without meaning. As long as we confine our attention to epistemology, this is a plausible argument. When we turn to the question of value, however, the matter is different. Many things may be decided by fiat or convention, but not value itself, as Socrates points out in reply to Protagoras's pragmatic argument in the Theaetetus . No one would say that whatever a state thinks is good or "advantageous to itself, really is so" (177d). Even for Nietzsche, the sustained attempt to regard value as a product of human creativity exists in perpetual tension with an "objective" preference for natural value (will to power) over reactive value (asceticism). The tension is not a weakness in Nietzsche but the mark of an intellectual penetration and honesty that refuses to dismiss the experiential evidence of a nonarbitrary factor in values, however awkward that factor may be for other features of his perspectivism. He could easily have achieved greater consistency by embracing a thoroughgoing personal relativism, but in order to do so he would have had to be a more shallow or less honest thinker. The message of the digression that grows out of the above passage of the Theaetetus is that the standard of goodness, in the light of which alone such judgments can validly be made, is the ultimate goal of philosophy; and that it is something we need to discover rather than invent. Not until the doctrine of the mean is developed in the Statesman is the theme of the digression further elaborated.

The value of the doctrine of categorial forms can ultimately be established only in relation to our understanding of goodness. This is true not only in the tautological sense that value implies a conception of goodness, but also in the significant sense that the moral implications of a way of interpreting the world are the most important criterion for choosing among objectively undecidable alternative interpretations—a point that even Kant recognized in the second critique. Hence the hy-


pothesis that the Same has objective reality, as well as epistemological force, is not treated as self-justifying; its justification points beyond itself. The assumption of the reality of the Same justifies our tendency to classify entities, but the very notion of better or worse classifications already implies the need for a standard of measure, and merely relative measure does not suffice. The doctrine of the Same can find ultimate justification only if its ultimate (and not merely regulative) value can be established.

The doctrine of the Same lies behind every kind of episteme and techne, built as they are on universal principles. Moreover, as the Eleatic stranger pointed out, these involve not only universalization but aspirations toward goodness; without the mean they would be inconceivable. There would be no possibility of distinguishing skill (techne) from incompetence, for example, unless our judgment were governed by some sense of what is right and good. Even in the form of modem science episteme makes use of the value-laden heuristic principle that reality should be explained in the most rational possible way. However much Ockham's razor militates against teleology in one sense, in another it presupposes it by encouraging rationally unified models over random and chaotic ones. In this sense even the Timaeus can be considered very much in keeping with the spirit of Ockham's razor.

An underlying message of the Timaeus is that, if the universe is to be conceivable as rational and good, it can be so only at the level of inherent structure, that is, universal pattern, not at the level of individuality. The kinds of things that exist can be justified, but individual events cannot, since the realm of individuality is at the mercy of the errant cause, chance.[4] Accordingly, the necessity of the reality of kinds follows for Plato from the postulation of a teleological justification of Being. This means that the justification of the Sophist's postulation (against the Theaetetus ) of the reality of kinds is ultimately predicated on the conception of goodness inherent in the Statesman's doctrine of the mean. The Sophist points beyond itself for its completion, toward the States man and beyond it to the idea of the good that is only implicit in the latter.

An argument analogous to this cosmological reasoning appears at the microcosmic level. Moral goodness implies the overcoming of ego-

[4] Many religious and philosophical traditions believe that even at the level of individuals a beneficent fate overrides mere contingency, but Plato does not seem to have subscribed to such a view—unless one interprets in this way the doctrine of individual "guardian daimons" in the Phaedo myth.


centric desires in favor of intentions that are equally concerned with what is good for others. To achieve this attitude means to aim at something like the standpoint of the whole, and for Plato our capacity for such a standpoint is "reason." Justice, and virtue generally, is the subordination of the egocentric desires for corporeal pleasure (appetite) and power (spiritedness) to reason. Whereas appetite and spiritedness operate with primary reference to the individuating body, reason operates with primary reference to the intelligible realm of universal and holistic form. The Sophist's project, of teaching us to think in terms of the sameness of universal kinds instead of the differentiation of corporeal individuality characteristic of the Theaetetus , can thus be justified on moral grounds as well as epistemological ones: learning to see the world in terms of the primacy of kinds over individuality is a way of overcoming the standpoint of personal bias and self-interest. The mean, by which the statesman evaluates proposed action, is a standard that abstracts from every merely relative measure, and can be discerned only by one who has achieved the disinterestedness of a universal point of view (although the application of the mean is always to particulars). The dialectical methods of the Eleatic stranger enable us to approach progressively the state of mind that makes such thinking possible.

Here again, if we ask whether it matters which hypothesis we adopt— the Platonic doctrine of forms or the Protagorean doctrine of individuality—we can point not only to the former's superiority in accounting for the nature of knowledge, but, more importantly, to its moral superiority. Its alleged superiority in accounting for knowledge, taken by itself, might easily be dismissed by Protagoras as nothing more than a self-aggrandizing presumption, since for him "knowledge" is simply an ignorant or pretentious term for "opinion." If Protagoras can establish that we can function in the world by means of opinion alone, Plato's epistemological argument will carry less weight, for worldviews might be conceived merely as convenient conceptual schemes, any number of which may have the same pragmatic serviceability. They may be regarded as distinctions without difference, free-floating metaphysical speculation. But many of these worldviews do nothing to promote an attitude toward the world that transforms us morally, and several of them promote attitudes that actively discourage morality. In such cases the distinctions imply real and important differences.

It is usual today to deny any connection between Being and goodness, fact and value, is and ought. Heidegger, for example, dismisses any such connection in the strongest terms in his Letter on Humanism ,


and on this issue all major contemporary philosophical traditions stand together. It is largely for this reason that contemporary moral thought is often accused of displaying an impoverished groundlessness. Only from pragmatists would Plato receive any conceivable support for his attempt to ground the theory of forms in the nature of the good—an ironic turn, since it was precisely to counter the pragmatism of Protagoras that the move toward formal kinds and ultimate teleology came about in the trilogy. In fact only the second, microcosmic, argument would be acceptable on pragmatic grounds: if belief in the primacy of the intelligible and universal, over the particular, fosters greater morality, then there are pragmatic reasons for accepting it. But Protagoras and contemporary pragmatists can never in fact accept it, for it would lead to the elimination of their first principle, the priority of the practical over the theoretical. For Plato, on the contrary, the theoretical is primary, and for practical as well as theoretical reasons, since it is the stronger ground of morality. Just as the Epicureans began with the life of the body but end up with the life of the mind, a Platonic pragmatism begins with praxis but is led to theoria.

2. Beyond the Statesman

The Statesman is not the end, but only the end of the beginning. It is possible to ascend still farther, we were reminded: young Socrates has been shown only a limited version of the stranger's method (262c). Would the fuller explanation reveal to us not only a more complete portrait of the statesman, but also the nature of the philosopher? The statesman's mean was only a principle of praxis; it could no more exist without the arts than they could without it (284d). The philosopher, by contrast, would be concerned with a more ultimate principle still, the unhypothetical, independent source of value, which is prior to any practical application.

At the beginning of the Sophist it is implied that, after investigating the nature of the sophist and statesman, the Eleatic stranger will continue by investigating that of the philosopher (216d-217b). This expectation is reasserted by Theodorus at the beginning of the Statesman (257a-c), but half a page later the elder Socrates suggests a symmetrical pattern instead: he himself has already conducted an investigation with Theaetetus as his partner; the stranger had then done so as well; now the stranger is to conduct one with young Socrates as his partner, so, Socrates suggests, young Socrates "will reply to me afterwards" (258a).


Now it seems that not the stranger but Socrates will lead the investigation into the nature of the philosopher. Perhaps this is another indication of the way that the stranger's approach begins to converge with the Socratic approach to philosophy during the course of the Statesman . Perhaps, on the other hand, the ambiguity about who will conduct the investigation is meant to suggest that the status of the projected dialogue is at least uncertain. After all, the second (314a-c) and seventh (341c-342a) letters tell us that such a project cannot be accomplished, and should not even be attempted except perhaps with the utmost caution. No philosopher was more aware than Plato of the self-overcoming dialectic of philosophic speech. The Parmenides shows how this is true of structuralism, and the Theaetetus shows how it is true of nonstructuralism. Accordingly, Plato's dialogues always have the character of withholding just what they seem to offer. Plato's own reticence, Socrates' irony, the interlocutors' reservations, and various literary devices continually combine to deny definitiveness to whatever is put forward. And yet things are put forward, for any attempt to evade the dialectical tension of philosophy by refraining from assertion only results in the no less problematic tensions of skepticism. There is an ultimate dialectic for Plato between aporia and natural value—ultimate in the sense that it must be preserved rather than overcome. To overcome the tension in favor of aporia results in misology. To overcome it in the direction of fully determinate values results in dogmatism. The difference between Plato and Protagoras is not that Protagoras alone recognizes the limits of objectivity, but that Plato alone recognizes both poles of the ineluctable tension of philosophy.

The investigation of the Philosopher was perhaps never intended to be written down in a dialogue, but only presented as the implicit goal of the Sophist and Statesman investigations. It is a goal that is necessarily left to readers to pursue further on their own. Nevertheless, there are hints that help us to find a path on which to begin. Consider the stranger's remark that the demonstration of the method of division in the Sophist and Statesman is not fully adequate. What would it be like in its fully realized form? "In the present circumstances," he said, "it is not possible to make it clearly evident" (262c). What does he mean by this? It seems likely that the limiting "circumstances" (

) refers to the limits of young Socrates' present abilities. As we saw, the application of the method seems tailored to one who has mastered the first five dianoetic disciplines but not dialectic. The method, as utilized in the Statesman , was not impressive for its rigor—


particular divisions were often arbitrary (sometimes avowedly so: 289b) and even contradictory—and was hardly displayed in way that would justify Socrates' claim in the Phaedrus that only a godlike man could perform it.

It was to overcome conceptual limitations like those of young Socrates that the stranger introduced the method of paradigms. His language was reminiscent of the Republic's "soul writ large," and more generally reminiscent of the theory of recollection:[5]

It is difficult, you divine man [

], to mark out any of the greater things without making use of paradigms. For it is as if each of us knows all that he knows as in a dream and then, like someone waking up, knows nothing of it at all.

We can see a paradigm of this technique in the way children can be taught to read:

Take them first to those cases in which their opinions were correct, and then take them and set them in front of the ones they do not recognize, and by comparing the two show that they are the same and have the same nature in both cases.

The same kind of problem occurs in philosophy, and a similar technique may be used to overcome it:

To some things there are by nature visible likenesses that are easy to understand, and that are not hard to show to people when, in response to a request for an explanation of one of them, someone wishes to indicate it easily, without trouble, and without a verbal explanation. To the greatest and most valuable [

] things, however, there corresponds no image made clear for human perception, which someone who wishes to satisfy the mind of an inquirer can display to one of his senses and succeed in satisfying them. Accordingly we must train ourselves to be able to give and to understand a rational account of each thing. For incorporeal things, which are the most beautiful and greatest, can be exhibited deafly through reason only and through nothing else, and it is for their sake that all this has been said. But practicing is in every case easier on lesser things rather than on the greater ones.

On the basis of this, the stranger undertook an elaborate division of the art of weaving, as a paradigm to help with the search for the states-

[5] As Skemp 76 points out, "It would seem that here [265b, 277c, 285e] in this doctrine of visible likenesses we have the late form of the earlier doctrine of Recollection. The visible likeness helps the mind to comprehend the Form of which it is a likeness."


man. It turned out that just as the weaver combines the warp and woof of the two kinds of woolen strands, the statesman weaves together the warp and woof of the two kinds of subjects. The weaver is a paradigm for the statesman because what the weaver does is visible in a sense that what the statesman does is not. In the case of the weaver we can easily see whether he understands the mean that leads to excellence in his craft, and we can see whether he is successfully weaving the warp and woof; we cannot see these things in the case of the statesman. More to the point, we can clearly see what the weaver's intentions and goals are, but not those of the statesman. The paradigm was well chosen, for despite the difference in their visibility we can appreciate the parallel between the two endeavors, and thus understand the statesman by analogy with the weaver.

Was this the whole point of the exercise, however? Why was the paradigm developed by means of an elaborate bisective division—at fifteen steps by far the longest in either of the two diaeretic dialogues— when the technique of bisective division was never to be used again, either in the remainder of the Statesman or in any other dialogue? Why did the stranger not simply say that weaving, insofar as it is a combination of warp and woof, will give us an image by which to picture the statesman's activity? Weaving is not only the subject of the diaeresis, and a paradigm for the art of the statesman, but is also a paradigm of the method of diaeresis itself: it is an intimate concatenation of combining and separating, as the diaeretic method is an intimate concatenation of collecting and dividing. Perhaps not only the extrinsic subject of this division but the very technique of division itself is intended as a paradigm of something that cannot easily be made evident to us. If the stranger has not been able to show us the nature of division in its highest sense, perhaps what he has shown us here is a paradigm of a higher kind of division.

The mean toward which the statesman looks is not the same as the Idea of the good, for the mean can exist only if there are arts, while the good is prior to all such specification. The statesman, in discerning the mean, must do so in the light of the good, but he need not perceive the good directly. We must assume that this is to be what ultimately distinguishes him from the philosopher. According. to the Republic , once the dialectic of noesis has led the philosopher to an apprehension of the Idea of the good, then, "hanging onto the things that hang on this beginning, it thus descends again to a conclusion, making


use of nothing perceived by the senses, but moving from forms themselves, through forms, to them, it concludes in forms" (511b-c). It is plausible that this downward movement of noesis, which proceeds from the good to the subsequent forms, is the consummation of the method of division. Or rather, that the method as portrayed in the Statesman (and Sophist ) functions as a concrete analogue or paradigm of this true method. It performs, at the prenoetic level of Theaetetus and young Socrates, the kind of derivation of kinds that can be carried out adequately only at the level of one who has attained the wisdom necessary to the true philosopher.

If that is the intention, then the stranger's method of division operates on several levels. (1) It functions as a useful tool for classification and definition. (2) It introduces the bridging concept of quality (in abstraction from value) into the erstwhile exclusively quantitative dianoia practiced by young Socrates and Theaetetus. (3) It illustrates the upward path of the method of hypothesis or dialectic by the abandonment of the shepherd modal in favor of the weaver model of the statesman (in which the concept of value makes its appearance in terms of the mean). (4) In accordance with the method of analogical paradigms, it uses this avowedly inadequate version of the method of division to intimate to us the consummate but as yet unattainable employment of this method, the downward path of dialectic from the good.

Because the exhibition of the method was, by admission, greatly simplified, it would be a mistake to try to assimilate the Republic passage very closely to the models we have been given in these two dialogues. On such a model we might, for example, have bisected the good into intelligible and corporeal species, and the latter into active (acting) and passive (perceiving), and thus derive by division the three major forms after the good, namely, wisdom, virtue, and beauty. But these would be merely formal divisions that tell us nothing of the content of the terms. The results would be nearly as extrinsic to the essential nature of the subjects as was the definition of "human being" in terms of the absence of feathers or legs. The actual derivation of the other forms from the Idea of the good must presuppose a detailed insight into the nature of the good, such as not even the Socrates of the Republic claims to be capable of. And the movement in the Statesman from bisective division (relative measure) to immediate division into numerous species suggests that a mechanical bilateral division of the good, as illustrated above, would never be satisfactory. The true relationship of the good to the


subordinate species of reality would need to be much more subtle and nuanced.

Of the dialogues believed to be later than the Statesman , the Philebus is the only one in which the method of division is employed even in a minor role. In one place, for example, "lack of self-knowledge" is divided into three species: with respect to wealth, with respect to beauty, and with respect to virtue; and the latter is divided into the twin species of weakness and strength (48d-49c). In another place a division is begun that might almost have come from the Sophist , but it does not last long. Knowledge is divided into knowledge of production on the one hand, and knowledge of education and nurture on the other. Productive arts are then divided into the inexact, like music, and the exact, like building. The most exact of the arts, arithmetic, is next divided into reckoning with unequal, concrete units (e.g., two armies, two oxen), and with equal, abstract units (two, four). Socrates observes that the division can be carried at least one step further, but breaks it off because it has already served its purpose, which is to show that there are pure and impure species of knowledge, just as there were pure and impure species of pleasure (55d-56e).

Neither of these exercises in the method of division is very ambitious, nor is there any other example that would justify the high expectations that the Phaedrus, Sophist , and Statesman led us to have. In another way, however, the concept of division does play a serious role in the Philebus . Socrates says:

And the ancients, who were superior to us and lived closer to the gods, handed down this saying, that all things that are ever said to be come from one and many, innately possessing limit and unlimitedhess in themselves. Since this is how things are ordered, we must always posit one Idea for each thing, and seek it—for we shall find it in the thing. Then, if we grasp it, after one we must look for two, if that is how many there are; but if not, then three or some other number. And again the same with each of those ones until we see not only that the original one is one and many and unlimited, but also how many it is. We should not apply the Idea of the unlimited to the multiplicity before we discern the whole number of it between the unlimited and the one.

This is evidently a reference to collection and division: by collection we discover the class in which the object is to be sought, and then we divide that class until we locate the object in the infimae species. Since we are instructed to repeat the process with each of the ones we discern, the method will eventually lead to the parent of all Ideas, the Idea of the


good (cf. Republic 6.509b). Ultimately, then, the goal is to be able to see in retrospect the lines of division that lead from the good to all other forms, as had been adumbrated in the Republic . Accordingly, later in the dialogue we are told to posit—as a further division beyond limit (form), the unlimited ("materiality"), and their combination—the cause of their union (23c-d), and that can only be the good (64c f.). As Socrates introduces this unifying conception, Protarchus asks, "What do you mean, my good man?" (

, 23d).

The method is not employed in the Philebus in anything like a conclusive way: the dialogue ends with Protarchus's enjoinder to Socrates not to leave yet because they have not yet finished (we are not told what remains to be done). It nevertheless has a crucial function in the search for the nature of the good. By means of the method of hypothesis, Philebus's hypothesis that the good life is the life of pleasure, and Socrates' hypothesis that the good life is the life of wisdom, have given way to the higher hypothesis that the good life is what combines wisdom with pleasure. Since this means that the nature of the good life may now be conceived as a product of wisdom (limit) and pleasure (unlimited), the method of collection and division, as described above, can be brought into play in order to seek the unifying cause of this mixture. Socrates concludes: "Then if we cannot catch the good with one Idea, let us capture it with three: beauty, proportion, and truth. Let us say that we may ascribe the causality to these, taken as one, more correctly than to the components of the mixture. And through this the mixture itself has come to be good" (64e-65a). Thus we have found the Idea of the good, by virtue of which the mixture of wisdom and pleasure constitutes the good life.

We have not, however, grasped it as it is in itself, but only in terms of three aspects. It is tempting to think of this as the first division of the Idea of the good, a division into the species of beauty, proportion, and truth. The first we might further divide into natural and artistic beauty, and thus be led to the realms of eros and art; by means of the second, to the realm of mathematics, and by means of the third to the realm of philosophy. But what of morality and politics? Just prior to this, Socrates had said, "Now, then, the power of the good has hidden itself from us in the nature of the beautiful. For certainly moderation and proportion completely correspond to beauty and virtue" (64e). Virtue, however, is left out of the subsequent triune description of the good. It turns out, then, that the division of the good into three parts is hardly definitive. Not only have we not grasped the good as it is in itself, but even


the three species in terms of which we have conceived it are to some extent arbitrary and incomplete. Nevertheless we begin to have a clearer idea of how the methods of division and of hypothesis may, both individually and in combination, give us guidance and direction in our asymptotic attempts to grasp the oneness of reality, and in particular to grasp the necessitating ground that alone can complete our understanding and endow the whole with value. When Plato does attempt to give more than a hint of content to the Idea of the good, he eschews any formal methodology whatever, and resorts to the ambiguous world of myth (although even in the myth of the Timaeus there is an emphasis on the forms of Same and Different, which lie at the basis of the method of collection and division).[6]

It was Hegel who ultimately attempted a rigorous systematization of the whole and of its necessitating principle. Indeed, the dialectical method by which he did so bears a strong resemblance to the methods of both hypothesis and division. It resembles the former in its progressive resolution, through higher theses, of the conflicts that result from the incompleteness of lower ones. And it resembles the latter in its preservation of the lower theses, so that lower and higher are related to one another through lines of derivation and subsumption—although the way in which the lower are present in the higher is not the same as the way species (even the quasi species of the Sophist and Statesman ) are present in the genus. The rigor that Hegel claimed for his unification, however, is often belied by arbitrary steps within the dialectic, and by questionable assumptions at the base of it. Plato's avowed mythopoietic presentation reflects a more realistic assessment of the possibility—or rather, impossibility—of consummating our aspiration toward wisdom with an episteme of wisdom.

Dramatically, the Timaeus is presented as a sequel to the Republic . This is obviously fitting because of the centrality of the creative power of the good in the latter dialogue. But the Timaeus is also the natural sequel to the investigations of the Philebus , especially if we regard the Timaeus 's mythic presentation as an indication that the Philebus has taken us as far as is possible through conceptual methodology alone. Whether chronologically it was in fact a sequel to the Philebus is impossible to ascertain, but ultimately the historical date of the Timaeus makes little difference. Regardless of when the Timaeus was written, it remains the case that only under the cover of mythology did Plato ever

[6] 35a, 36c, 39a, 40a-b, 43d.


feel comfortable enough to attempt to give content to his conception of the good. His famous lost lecture on the good was said to have been full of mathematics; but whether the mathematics were any less qualified by whimsy and aporia than those of the Timaeus must remain a matter of speculation. We cannot dismiss the fact that he never published it, and in this respect the warnings of the second and seventh letters ring true.


Bibliography of Works Cited

Ackrill, John. "Plato on False Belief: Theaetetus 187-200." The Monist 50 (1966) 353-402.

———. "Sg MPL OKH EIDW N." In Vlastos, ed., Plato (q.v.) 201-9.

Allen, R. E. Plato's Parmenides . Oxford: Blackwell, 1983.

———, ed. Studies in Plato's Metaphysics . London: Routledge and Kegan Paul, 1965.

Anscombe, G. E. M. "The New Theory of Forms." The Monist 50 (1966) 403-20.

Anton, John, and Anthony Preus, eds. Essays in Ancient Greek Philosophy , vol. 3, Plato . Albany: SUNY Press, 1989.

Beck, Maximilian. "Plato's Problem in the Parmenides ." Journal of the History of Ideas 8 (1947) 232-36.

Benardete, Seth. The Being of the Beautiful . Chicago: University of Chicago Press, 1984, 1986.

Bostock, David. Plato's Theaetetus . Oxford: Clarendon Press, 1988.

Brandwood, Leonard. The Chronology of Plato's Dialogues . Cambridge: Cambridge University Press, 1990.

———. "The Dating of Plato's Works by the Stylometric Method—A Historical and Critical Survey." 2 Vols. Dissertation, University College, London. 1959.

Brisson, Luc. "Interprétation du mythe du Politique ." Paper presented at the Third Symposium Platonicum, "Plato's Politicus ," Bristol, August 25-30, 1992.

Brown, Malcolm. "Theaetetus : Knowledge as Continued Learning." Journal of the History of Philosophy 7 (1969) 359-79.

Brumbaugh, Robert. Plato on the One: The Hypotheses in the Parmenides . New Haven: Yale University Press, 1961.


——— "The Purpose of Plato's Parmenides ." Ancient Philosophy 1 (1980) 39-48.

Burner, John·Greek Philosophy: Thales to Plato . London: MacMillan, 1914.

Burnyeat, Myles. "The Material and Sources of Plato's Dream." Phronesis 15 (1970) 101-22.

———. "Protagoras and Self-Refutation in Plato's Theaetetus ," Philosophical Review 85 (1976) 172-95.

———. "Socrates and the Jury: Paradoxes in Plato's Distinction between Knowledge and True Belief." Supplementary Volume LIV of the Aristotelian Society (1980) 173-92.

———. The Theaetetus of Plato . Indianapolis: Hackett, 1990.

Campbell, Lewis. The Theaetetus of Plato . Oxford: Clarendon Press, 1883.

Chance, Thomas. Plato's Euthydemus . Berkeley and Los Angeles: University of California Press, 1992.

Cherniss, Harold. "The Philosophical Economy of the Theory of Ideas." In Allen, ed., Studies (q.v.) 1-13. [Originally published in 1936.]

———. "The Relation of the Timaeus to Plato's Later Dialogues." American Journal of Philology 75 (1957) 225-66. [Reprinted in Allen, ed., Studies (q.v.) 339-78.]

Cornford, F. M. Plato and Parmenides . London: Routledge and Kegan Paul, 1939.

———. Plato's Theory of Knowledge . London: Routledge and Kegan Paul, 1935.

Deleuze, Gilles·Difference and Repetition . Translated by Constantine Boundas. Unpublished manuscript.

Desjardins, Rosemary. "The Horns of Dilemma: Dreaming and Waking Vision in the Theaetetus ." Ancient Philosophy 1 (1951) 109-26.

———. The Rational Enterprise: Logos in Plato's Theaetetus . Albany: SUNY Press, 1990.

Diels, Hermann. Die Fragmente der Vorsokratiker . Revised by Walther Kranz. 12th ed. Zurich: Weidmann, 1966.

Dorter, Kenneth. "First Philosophy: Metaphysics or Epistemology?" Dialogue 11 (1972) 1-22.

———. "Justice and Method in Plato's Statesman ." In S. Panagiotou, ed., Justice, Law and Philosophy in Classical Athens 105-22. Edmonton: Academic Printing and Publishing, 1987.

———. "Nussbaum on Transcendence in Plato and Aristotle." Dialogue 32 (1993) 105-15.

———. Plato's Phaedo: An Interpretation . Toronto: University of Toronto Press, 1982.

———. "Socrates' Refutation of Thrasymachus and Treatment of Virtue." Philosophy and Rhetoric 7 (1974) 25-46.

Fattal, Michel. "La diairesis dans le Politique de Platon." Paper presented at the Third Symposium Platonicum, "Plato's Politicus ," Bristol, August 25-30, 1992.

———. "Le Sophiste: Logos de la synthèse ou logos de la division?" In Narcy, ed. (q.v.), 145-63.


Ferrari, G. R. F. "Myth and Conservatism in Plato's Politicus ." Paper presented at the Third Symposium Platonicum, "Plato's Politicus ," Bristol, August 25-30, 1992.

Fine, Gail. "Knowledge and Logos in the Theaetetus ." Philosophical Review 88 (1979) 366-97.

Fowler, H. N. Plato -Statesman. Loeb Classical Library. Cambridge, Massachusetts: Harvard University Press, 1925.

Frede, Michael. Prädikation und Existenzaussage: Platons Gebrauch yon ". . . ist " und ". . . ist nicht " im Sophistes . Hypomnemeta 18. Göttingen: Vandenhoeck and Ruprecht, 1967.

Friedländer, Paul. Plato . Vol. 3. Translated by Hans Meyerhoff. Princeton: Princeton University Press, 1969.

Gadamer, Hans-Georg. The Idea of the Good in Platonic-Aristotelian Philosophy . Translated by P. Christopher Smith. New Haven: Yale University Press, 1986.

———. "Plato's Parmenides and Its Influence." Dionysius 7 (1983) 3-16.

Gallop, David. Parmenides of Elea: Fragments . Toronto: University of Toronto Press, 1984.

Geach, Peter. "The Third Man Again." Philosophical Review 65 (1956) 72-82. [Reprinted in Allen, ed., Studies (q.v.) 265-77.]

Gomez-Lobo, Alfonso· "Plato's Description of Dialection in the Sophist 253d1-e2." Phronesis 22 (1977) 29-47.

Gooch, Paul. "'Vice Is Ignorance': The Interpretation of Sophist 226A-231B." Phoenix 25 (1971) 124-33.

Griswold, Charles. "Politikê Epistêmê in Plato's Statesman ." In Anton and Preus, eds. (q.v.), 147-67.

Guthrie, W. K. C. A History of Greek Philosophy , vol. 5. Cambridge: Cambridge University Press, 1978.

Hackforth, R. "The Aviary Theory in the Theaetetus ." Classical Quarterly 32 (1938) 27-29.

Harrison, Joan. "Plato's Prologue: Theaetetus 142a-143c." Tulane Studies in Philosophy 27 (1978) 103-23.

Hicken, Winifred. "Knowledge and Forms in Plato's Theaetetus ." Journal of Hellenic Studies 77 (1957) 48-53.

Howland, Jacob. "Rereading Plato: The Problem of Platonic Chronology." Phoenix 45 (1991) 189-214.

Irwin, Terence. "Plato's Heracleiteanism." Philosophical Quarterly 27 (1977) 1-13.

Kahn, Charles H. "The Greek Verb 'to Be' and the Concept of Being." Foundations of Language 2 (1966) 245-65.

———. The Verb 'Be' in Ancient Greek . Foundations of Language, Supplementary Series, vol. 16. Dordrecht: Reidel, 1973.

Klein, Jacob. A Commentary on Plato's Meno . Chapel Hill: University of North Carolina Press, 1965.

———. Plato's Trilogy . Chicago: University of Chicago Press, 1977.

McDowell, John. Plato: Theaetetus . Oxford: Clarendon Press, 1973.

Miller, Mitchell, Jr. The Philosopher in Plato's Statesman . The Hague: Nijhoff, 1980.


———. Plato's Parmenides: The Conversion of the Soul . Princeton: Princeton University Press, 1986.

———. "Unity and Logos: A Reading of Theaetetus 201c-210a." Paper presented at the Society for Ancient Greek Philosophy, Dec. 27, 1989.

Mohr, Richard. "Disorderly Motion in Plato's Statesman ." Phoenix 35 (1981) 199-215.

———. "Statesman 284c-d ." Phronesis 22 (1977) 232-34.

Morrow, Glenn. "Plato and the Mathematicians: An Interpretation of Socrates' Dream in the Theaetetus (201e-206c)." Philosophical Review 79 (1970) 309-33.

Narcy, Michel, ed. Etudes sur le Sophiste de Platon . Bibliopolis: C.N.R., Centro di Studio del Pensiero Antico, 1991.

Nehamas, Alexander. "Epistêmê and Logos in Plato's Later Thought." In Anton and Preus, eds., (q.v.), 267-92.

Newman, Jay. "The Recoil Argument." Apeiron 16 (1982) 47-52.

Novak, Joseph. "Plato and the Irrationals." Apeiron 16 (1982) 71-85 and 17 (1983) 14-27.

Nussbaum, Martha. The Fragility of Goodness . Cambridge: Cambridge University Press, 1986.

Oehler, Klaus. Die Lehre yore noetischen und dianoetischen Denken bei Platon und Aristoteles . Zetemata 29. Munich: Beck, 1962.

O'Toole, E. J. "Forms and Knowledge in the Theaetetus ." Philosophical Studies 19 (1970) 102-18.

Owen, G. E. L. "The Place of the Timaeus in Plato's Dialogues." Classical Quarterly , n.s., 3 (1953) 79-95. [Reprinted in Allen, ed., Studies (q.v.) 313-38.]

———. "Plato on Not-Being." In Vlastos, ed., Plato (q.v.) 223-67.

Panagiotou, Spiro. "The Consequences of the Divisibility of Forms in Plato's Parmenides ." Phoenix 36 (1982) 45-52.

Pascal, Blaise. "On Geometrical Demonstration." Translated by Richard Scofield. In Pascal: The Provincial Letters, Pensèes, Scientific Treatises , vol. 33 of Great Books of the Western World, 430-36. Chicago: Encyclopaedia Brittanica, 1952.

Polansky, Ronald. Philosophy and Knowledge: A Commentary on Plato's Theaetetus . Lewisburg: Bucknell University Press, 1992.

Prior, William. Unity and Development in Plato's Metaphysics . LaSalle: Open Court, 1985.

Proclus. Commentary on Plato's Parmenides . Translated by Glenn Morrow and John Dillon. Princeton: Princeton University Press, 1987.

Robinson, Richard. "Forms and Error in Plato's Theaetetus ." Philosophical Review 59 (1950) 3-30.

Robinson, T. M. "Forms, Demiurge and World-Soul in the Politicus ." Paper presented at the Third Symposium Platonicum, "Plato's Politicus ," Bristol, August 25-30, 1992.

Rochol, Hans. "The Dialogue Parmenides : An Insoluble Enigma in Platonism?" International Philosophical Quarterly 11 (1971) 496-520.

Rorty, Amelia· "A Speculative Note on Some Dramatic Elements in the Theaetetus ." Phronesis 17 (1972) 227-38.


Rosen, Stanley. Plato's Sophist . New Haven: Yale University Press, 1983.

———. "Socrates' Dream." Theoria 42 (1976) 161-88.

———. "Sophrosyne and Selbstbewusstsein ." Review of Metaphysics 26 (1973) 617-42.

Ross, W. D. Plato's Theory of Ideas . Oxford: Clarendon Press, 1951.

Runciman, W. G. Plato's Earlier Epistemology . Cambridge: Cambridge University Press, 1962.

Sayre, Kenneth. Plato's Analytic Method . Chicago: University of Chicago Press, 1969.

———. Plato's Late Ontology . Princeton: Princeton University Press, 1983.

Schleiermacher, Friedrich. Introductions to the Dialogues of Plato . Cambridge: J. & J. Deighton, 1836.

Scodel, Harvey. Diaeresis and Myth in Plato's Statesman . Göttingen: Vandenhoeck and Ruprecht, 1987.

Seligman, Paul. Being and Not-Being: An Introduction to Plato's Sophist . The Hague: Nijhoff, 1974.

Sellars, Wilfrid. "Vlastos and the Third Man." Philosophical Review 64 (1955) 405-37.

Shiner, Roger. Knowledge and Reality in Plato's Philebus . Assen: van Gorcum, 1974.

Shorey, Paul. "A Lost Platonic Joke (Politicus 266 D)." Classical Philology 12 (1917) 308.

Skemp, J. B. Plato's Statesman . London: Routledge and Kegan Paul, 1952.

Sprute, Jürgen. "Über den Erkenntnisbegriff in Platons Theaitet ," Phronesis 13 (1968) 47-67.

Sternfeld, Robert, and Harold Zyskind. Meaning, Relation, and Existence in Plato's Parmenides: The Logic of Relational Realism . New York: Lang, 1987.

Stenzel, Julius. Plato's Method of Dialectic· Translated and edited by D. J. Allan. New York: Russell and Russell, 1964. [Originally published in 1940.]

———. Studien zur Entwicklung der platonischen Dialektik von Sokrates zu Aristoteles . 3d ed. Darmstadt: Wissenschaftliche Buchgesellschaft, 1961. [2d ed. Leipzig, 1931.]

Taylor, A. E. Plato: The Man and His Work . 6th ed. Atlantic Highlands: Humanities Press, 1952.

Teloh, Henry. The Development of Plato's Metaphysics . University Park: Pennsylvania State University Press, 1981.

Tordesillas, Alonso. "Le point culminant de la métrétique." Paper presented at the Third Symposium Platonicum, "Plato's Politicus ," Bristol, August 25-30, 1992.

Trevaskis, J. R. "The

and the Vowel Analogy of Plato, Sophist 253." Phronesis 11 (1966) 99-116.

Turnbull, Robert. "The Third Man Argument and the Text of Parmenides ." In Anton and Preus, eds. (q.v.), 203-25.

Vlastos, Gregory. "Postscript to the Third Man: A Reply to Mr. Geach." Philosophical Review 65 (1956) 83-94. [Reprinted in Allen, ed., Studies (q.v.) 279-91.]


———. "The Third Man Argument in the Parmenides ." Philosophical Review 63 (1954) 319-49. [Reprinted in Allen, ed., Studies (q.v.) 231-63.]

———, ed. Plato , vol. 1. Garden City: Doubleday-Anchor, 1971.

Watanabe, Kunio. "The Theaetetus on Letters and Knowledge." Phronesis 32 (1987) 143-65.

Wyller, Egil. Platons Dialog Parmenides in seinem Zusammenhang mit Symposion und Politeia . Oslo: 1960.

———. "Plato's Parmenides : Another Interpretation." Review of Metaphysics 15 (1962) 621-40.

Xenakis, Jason· "Essence, Being and Fact in Plato: An Analysis of One of Theaetetus 'Koina.'" Kant-Studien 49 (1957-58) 167-81.

Yoh, May. "On the Third Attempted Definition of Knowledge, Theaetetus 201c-210b." Dialogue 14 (1975) 420-42.


Index of Subjects


Anaxagoras, 58

Anaximander, 58

Apeiron. See Unlimited

Aristophanes, 122

Aristotle, 3 , 26 , 58 , 66 , 82 , 93 , 106 n46, 133 , 143 , 204 , 219

De Anima



A.1.980b27-981b10: 87 n23, 102

A.1.980b29-981a1: 81

A.6.987b25-988a1: 58 n43

a .1.993b9-11: 82 n18, 106 n46


G .2.1004b22-25: 120 , 175

G .5.1010a11-13:73 n10

Z.3.1029b3-12:82 n18


Nicomachean Ethics , 119 , 205 n23

2.6.1106a26-b28: 203 , 203 n21

Parts of Animals




1.2.1253a2 -3


Aviary model of knowledge, 101 -7


Becoming. See Flux

Being, 54 -64, 91 , 93 , 140 -62, 176 -80, 231 -35

Belief. See Opinion

Berkeley, 120 n63

Blessed, Islands of the, 88 -89


Cause. See Contributory causes; Purpose

Cave, Allegory of, 32 , 35 , 87 -9, 119 , 139 , 178 , 207 , 224

Circle, argument of Statesman as, 223 -24

Collection, 102 , 131 , 145 , 154 -55, 160 , 180 , 196 , 229 . See also Division, method of

Concepts, forms as, 34 -37

Contributory causes, 200 , 202 , 210 -15, 221 -23

Courage, 21 , 25 , 69 -72, 87 , 225 -27

Cratylus, 73 , 90 , 93


Definition, 14 , 16 n24, 22 , 71 , 77 , 112 n52, 115 -17, 119 , 137 , 143 , 210

Derrida, 93

Descartes, 93 , 106 n46, 230

Dialectic, 47 , 92 n28, 104 , 104 n40, 151 -165, 207 -9, 236 , 239 -42. See also Division, method of; Hypothesis, method of

Difference (dissimilarity, other), 53 -4n39, 54 -67 passim , 75 n11, 91 , 93 , 97 , 99 , 110 , 114 , 117 -18, 121 , 136 -41, 150 -62, 179 -80, 183 , 205 , 208 , 229 , 242


Diogenes Laertius, 5

Divided Line, 17 , 34 , 87 , 102 , 115 -19, 138 , 139 , 177 -78, 208 , 219 n28, 224

Divine, 4 , 41 -43, 60 , 69 , 82 , 87 -88, 120 n64, 122 , 131 n8, 142 , 178 -80, 188 -98, 202 , 204 , 218 , 220 n29, 221 , 226 -27, 231 , 240

Division, method of, 13 -17, 72 , 106 , 116 , 124 -36, 140 -41, 143 , 153 -55, 172 -80, 182 -90, 192 n12, 196 -224, 227 , 236 -42

Doxa: See Opinion

Dream, 72 , 78 -79, 107 -9, 197 , 221 , 237 ; ("Socrates' Dream")


Epicureans, 235


Falsity, 51 , 54 , 75 -86, 96 -105, 136 -40, 162 -65, 209

Flux, 73 -74, 77 -79, 90 -95, 113 , 116 , 121 , 147 , 179 , 229

Forms, theory of, 15 -16, 21 -47, 90 -96, 98 n34, 109 n50, 111 -120, 121 , 139 , 142 -80, 232 -35, 238 -243

Friends of the forms, 142 -50, 177 -80


God(s). See Divine

Good, form of the, 16 n23, 21 & n3, 24 -26 passim , 49 , 57 , 57 n41, 66 -67, 160 -61, 179 , 202 , 204 , 207 , 209 , 227 , 231 -35, 238 -43


Hegel, 35 , 242

Heidegger, 17 , 234

Heracleitus, 58 -59, 73 -74, 77 -79, 91 -94, 113 , 146

Hermeneutics. See Language; Writing

Hypothesis, method of, 13 -17, 73 , 73 n9, 74 -75, 106 , 140 -41, 196 , 227 -31, 241 -42


Iliad, The , 118 n60

Image. See Similarity

Indefinite. See Unlimited

Interpretation. See Interpretive knowledge; Language; Writing

Interpretive knowledge (understanding); 79 -86, 89 -95, 120

Irrational(s), 71 , 187 n6, 192 -93, 224 n30

Isocrates, 6 , 8


Judgement. See Opinion

Justice, 4 , 21 , 24 -6 passim , 69 , 86 , 88 n25, 89 , 107 , 122 -23, 131 , 133 , 142 , 151 , 157 n50, 177 , 192 , 195 -97, 204 -205n23, 216 -20, 226 , 234


Kant, 78 , 93 -94, 229 -32


Language, 90 -91, 137 -40, 146 , 150 -65, 229 -30. See also Writing

Law, 87 , 89 , 123 , 217 -20, 224

Laziness, 25 , 69 -71

Letters and syllables, 109 -20, 150 -51, 197 , 208 , 237 . See also Nouns and verbs

Likeness. See Similarity

Logos, 16 n24, 106 -20, 122 , 162 -65, 173 , 219 n28, 229 -30


Materialism, 34 -35, 60 , 77 , 142 -50, 231

Mathematics, 3 , 70 n3, 72 -73, 92 -95, 102 -6, 108 , 118 , 207 , 224 n30, 240 -41, 243

Mean, 16 -17, 89 n26, 139 , 188 -89, 191 , 203 -219, 223 -24, 227 -35 passim , 238

Measure, 89 , 89 n26, 203 , 207 . See also Mean and Relative measure

Melissus, 91

Memory, 70 -72, 81 -83, 98 -109, 125 , 192

Meno's Paradox, 70 -71, 90 , 108 , 112 n53, 115 , 173

Midwifery, Socrates', 71 -72, 79 -80, 109

Moderation (self-control), 4 , 21 , 123 , 124 , 225 -27

Moore, 17

Morality, 231 -35, 241

Music, 89 , 113 , 113 n54, 151 -52, 220 , 220 n29, 240


Nietzsche, 17 , 93 , 231 -32

Noncontradiction, principle of, 79 , 79 n15, 82 , 145

Not-being (nonbeing), 64 -67, 136 -41, 150 , 156 , 158 -62, 229

Nouns and verbs, 162 -65. See also Letters and syllables


Ockham's Razor, 233

Opinion, 16 n24, 71 , 72 , 76 , 78 , 84 -86, 89 , 96 -120, 162 , 165 , 172 , 234

Other. See Difference


Paradigms, 22 -23, 26 , 37 -39, 43 , 116 -17, 173 , 181 n1, 197 , 230 , 237 -38

Parentage, 76 -77, 94 -96


Parmenides, 35 , 40 , 48 -49, 56 -59, 64 , 64 n52, 65 , 91 , 136 -37, 140 , 159 , 229

Participation, 42 -43, 45

Pascal, 120 n64

Peirce, 232

Power, 125 , 143 -50, 160 -62, 176 -80

Pragmatism, 83 -6 passim , 232 -35

Proclus, 48 , 49 -50n35, 53 , 53 n38, 53 n39, 58 n42, 59 -60n45, 146

Protagoras, 79 -86, 89 -95, 97 , 137 , 146 , 229 , 234 , 236

Purification, 106 , 119 , 173 -75, 195 , 225

Purpose, 16 , 24 , 115 , 122 , 168 , 175 , 200 , 202 , 209 , 221 , 223 , 233 -35


Rational. See Reason

Reading. See Letters and syllables

Reality. See Being

Reason, 7 , 25 -26, 32 -33, 35 -36, 60 , 69 , 83 , 88 , 90 , 94 , 107 , 117 , 121 , 122 , 131 -32, 143 -50, 176 -80, 231 , 233 , 237

recollection, 21 , 27 -31, 42 -43, 45 , 70 -72, 80 -83, 95 , 95 n34, 102 , 197 , 104 -9, 118 -19, 173 , 237 , 237 n5

Relative measure, 188 -89, 203 -16, 224 , 233 -34, 239


Self-mastery, 68 , 119 , 133 , 176

Semblance. See Similarity

Similarity (image, likeness, semblance, etc.), 23 -24, 30 , 37 -39, 88 , 91 , 93 -94, 114 , 116 -17, 134 -41, 155 , 172 -76, 179 -80, 192 , 208 , 218 -19, 229 -35 passim , 237 -39, 242

Sophists, 16 -17, 73 -95, 121 -40, 159 , 215 , 218 , 228 , 230

Soul, 32 -33, 36 , 91 , 111 -12n52, 112 , 142 , 144 -50, 157 , 184 , 189 -90

Soul, tripartite, 17 , 24 -25, 68 -71, 88 , 85 n25, 119 , 122 -34, 216 , 219 , 225 , 227 , 234

Syllables. See Letters and syllables


Teleology. See Purpose

Truth, coherence, correspondence, and evidentness models of, 105 -7, 231 -32


Understanding. See Interpretive knowledge

Universals, 22 -23, 31 -34, 92 -94, 102 , 104 n41, 115 -16, 117 -18, 121 , 220 , 228 -35 passim

Unlimited, 54 -66, 158 , 192 , 229 , 240 -41


Value, 2 , 2 n2, 16 -17, 21 , 23 -25, 73 , 84 -94, 117 , 119 , 131 , 143 , 151 , 159 , 162 , 168 , 170 , 173 -80, 189 -91, 194 -97, 200 , 202 -3, 206 , 209 , 216 -28 passim , 231 -35

Verbs, 113 . See also Nouns and verbs

Virtue, 234 , 240 -41. See also Courage, Justice, Moderation, Morality, Wisdom


Wax block model of knowledge, 98 -101, 102 , 104

Weber, 17

Whole-part relation, 55 , 55 n40, 93 , 103 n39, 110 -20, 141 -42, 152 -55

Wisdom, 16 -17, 21 , 41 -42, 46 , 68 -69, 74 , 83 -94, 119 -20, 219 , 235 , 241

Writing, 9 -13, 46 -47, 214 , 217 , 224 , 224 n30, 236 -43. See also Language


Zeno, 23


Index of Platonic Dialogues and Passages Cited


Cratylus : 151

432b-d: 138 -39

399c: 189



273a: 25 n6

297e: 76 n13



449a: 123

464b-c: 123 n3

467a: 123 n2

481b: 122

491e: 123 n2

517c-518a: 123 n2

521d-e: 123


Laws : 5 -6, 176

653e: 188 n8

963c-965d: 225

Letters : 243

2.314a-c: 46 , 219 , 236

7.341b-d: 219 n28

7.341c-342a: 46 , 236


Meno : 42 , 70 -72, 82 , 94 , 102

81c: 105

81d: 70 -71, 106

85c-d: 72 , 108 , 197

98a: 106



date of: 4 -9

Phaedo : 22 , 45 , 68 , 88 , 93 , 95 n33, 121 , 209 n26, 229

65d-66a: 36

66b-e: 41

68c-e: 225

70b: 160

74a-75b: 22 , 31 , 33 , 38 , 78 , 138 n16

78b: 34

78c: 28

79b-80b: 147 -48

80b: 113 n55, 145 n19

82d-83e: 119 n62

85e-88b: 112

92e-94b: 143

96a-e: 28 , 60

96a-105c: 180 n64

98b-99c: 175 , 200 , 223

99a-100e: 143 n18, 219

99c: 21 n3, 141

99d: 118

99e: 106 n46

100a: 13 , 21 n3, 74 -75

100b-e: 31 , 60 , 229

100e-103a: 75

101a: 28

101d-e: 13 , 21 n3, 74 -75, 186

102b-d: 33


Phaedo (continued )

102b-d: 33

102d-105e: 151 , 160

103b: 33

105b-c: 60 n46, 186


245e: 21 n3

246a: 13 n15

247c ff: 82 n17

247d: 42 n25

248d: 106 n46

249d: 42 n25

249e-250a: 105

259e-260d: 135 n13

265c-266b: 13 , 131 , 131 n8, 210 , 224

275c-d: 46 , 224

275d-277a: 217 , 219 n28, 224

276e-277a: 46

Philebus : 3 -4, 17 , 176 , 240 -42

12e-13a: 150

13e: 21 n3

15a: 21 n3

16a: 188 n8

16c-d: 240

17a: 187

17b: 116

23c-25b: 60 , 241

25b-27b: 60 n46

48d-49c: 240

55d-56e: 240

64c: 241

64e-65a: 241

Protagoras : 84 n19

311e: 123

313c-e: 123

317b: 123

323b: 123

333d: 123

349a: 123


Republic : 68 -69, 93 , 121 , 169 n54, 177 -79, 194 , 197 , 204 , 204 n23, 218

332e-333d: 220

333e-334b: 122 , 205 n23

342c-d: 221

345c-346e: 167 n50

352d-354a: 205 n23

357a-367e: 122

372d: 185

372e-373e: 128 n7

375b-76c: 69 , 225

430e-431a: 133

435c-d: 119 n61

436b: 82 n16, 145

437b-441c: 133

442b-443e: 68 , 70 n5, 88 n25, 225

444b-e: 133

449a-b: 87 n24

479a: 38

504b: 119 n61

505a-506a: 151

506b-e: 30 n15, 168

507a-517c: 118 , 178

507d-509c: 106 n46

508e-509b: 151

509b: 16 n23, 21 , 21 n3, 24 , 49 , 57 , 57 n41, 66 -67, 160 -61, 241

509c: 119 n61

510c: 208

511b-c: 13 , 238 -39

516a-b: 82

516c-d: 87 n23, 88

517d-518b: 82 n18

518c-521c: 119

519a-b: 119 n62

519c: 88

525a-531d: 92 n28, 207

532c: 178

543c-544a: 87 n24

596a: 16 , 22 , 92

596a-597b: 178

596c: 134

598b-c: 134 n12


Symposium 43

202c-d: 60

202d-203a: 60 n46, 148 n27

203e-204b: 169

205e-206a: 21 n3

210a-d: 22 , 119 n62


Timaeus 17 , 66 , 74 , 176 , 192 -94, 233

29a-30b: 21 n3, 60

35a-43d passim : 242

47e-48a: 24 n5, 60 n46

date of: 3 -9, 242 -43



G&S Typesetters


10/13 Sabon




Braun Brumfield, Inc.


Braun Brumfield, Inc.

Preferred Citation: Dorter, Kenneth. Form and Good in Plato's Eleatic Dialogues: The Parmenides, Theatetus, Sophist, and Statesman. Berkeley:  University of California Press,  c1994 1994.