Introduction
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.
(15b)
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.
(58a)
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.]
(61d-62a)
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)
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.). | |||||||||
Investigator | Date | Chief criterion | Parmenides | Timaeus | |||||
L. Campbell | 1867 | unusual words | middle | late (4/6) | |||||
P. Blass | 1874 | hiatus | [not dated] | late (4/6) | |||||
W. Dittenberger | 1881 | particles | late (1/7)[9] | late (3/7) | |||||
A. Frederking | 1882 | particles | early[10] | late | |||||
F. Kugler | 1886 | particles | uncertain | late (7/7)[11] | |||||
M. Schanz | 1886 | metaphysical terms | [not dated] | late (3/4) | |||||
H. Siebeck | 1888 | reply formulas | late | [not dated] | |||||
C. Ritter | 1888 | expressions | middle[12] | late (4/6) | |||||
J. Tiemann | 1889 | expressions | [not dated] | late (1/6) | |||||
T. Lina | 1889 | prepositions | [not dated] | late (4/6) | |||||
G. B. Hussey | 1889 | terms for | uncertain | late (4/6) | |||||
"aforementioned" | |||||||||
H. von Arnim | 1896 | reply formulas | middle | [not dated] | |||||
C. Baron | 1897 |  | middle | late | |||||
W. Lutoslawsky | 1897 | 500 characteristics | middle | late (4/6) | |||||
P. Natorp | 1899 | 1,949 words | middle | late (3/6) | |||||
G. Janell | 1901 | hiatus | middle | late | |||||
W. Kaluscha | 1904 | clausula rhythm | middle | late (1/6) | |||||
H. von Arnim | 1912 | reply formulas | late (7/12) | [not dated] | |||||
L. Billig | 1920 | clausula rhythm | middle | late (1/6) | |||||
A. Fossum | 1931 | hapax legomena | early | late | |||||
[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 (
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.
(227a-b)
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.