Book Search:  


Google full text of our books:


The Sense of the Past:
Essays in the History of Philosophy
Bernard Williams
Edited and with an introduction by Myles Burnyeat

Book Description | Endorsements | Table of Contents

COPYRIGHT NOTICE: Published by Princeton University Press and copyrighted, © 2006, by Princeton University Press. All rights reserved. No part of this book may be reproduced in any form by any electronic or mechanical means (including photocopying, recording, or information storage and retrieval) without permission in writing from the publisher, except for reading and browsing via the World Wide Web. Users are not permitted to mount this file on any network servers. Follow links for Class Use and other Permissions. For more information, send e-mail to

This file is also available in Adobe Acrobat PDF format

Chapter 1



The legacy of Greece to Western philosophy is Western philosophy. Here it is not merely a matter, as in science, of the Greeks having set out on certain paths in which modern developments have left their achievements far behind. Nor is it just a matter, as in the arts, of the Greeks having produced certain forms, and certain works in those forms, which succeeding times would--some more, some very much less--look back to as paradigms of achievement. In philosophy, the Greeks initiated almost all its major fields--metaphysics, logic, the philosophy of language, the theory of knowledge; ethics, political philosophy, and (though to a much more restricted degree) the philosophy of art. Not only did they start these areas of enquiry, but they progressively distinguished what would still be recognized as many of the most basic questions in those areas. In addition, among those who brought about these developments there were two, Plato and Aristotle, who have always, where philosophy has been known and studied in the Western world, been counted as supreme in philosophical genius and breadth of achievement, and whose influence, directly or indirectly, more or less consciously, under widely varying kinds of interpretation, has been a constant presence in the development of the Western philosophical tradition ever since.

Of course philosophy, except at its most scholastic and run down, does not consist of the endless reworking of ancient problems, and the idea that Western philosophy was given almost its entire content by the Greeks is sound only if that content is identified in the most vague and general way--at the level of such questions as 'what is knowledge?' or 'what is time?' or 'does sense-perception tell us about things as they really are?' Philosophical problems are posed not just by earlier philosophy, but by developments in all areas of human life and knowledge; and all aspects of Western history have affected the subject-matter of philosophy--the development of the nation-state as much as the rise and fall of Christianity or the progress of the sciences. Yet even with issues created by such later developments, it is often possible to trace contemporary differences in philosophical view to some general contrast of outlooks which had its first expression in the Greek world.

Granted the size of the Greek achievement in philosophy, and the depth of its influence, it would be quite impossible to attempt anything except a drastically selective account of either. Some very important and influential aspects of Greek philosophy I shall leave out entirely: these include political philosophy (which is the concern of another chapter), and also Greek contributions to the science of logic, which were very important but demand separate, and moderately technical, treatment.1 Moreover, in the matter of influences, I shall not attempt to say anything about what is certainly the most evident and concentratedly important influence of Greek philosophy on subsequent thought, the influence of Aristotle on the thought of the Middle Ages. Aristotle, who was for Thomas Aquinas 'The Philosopher', for Dante il maestro di color che sanno, 'the master of those who know', did much to form, through his various and diverse interpreters, the philosophical, scientific, and cosmological outlook of an entire culture, and the subject of Aristotelianism would inevitably be too much for any essay which wanted to discuss anything else as well. Aristotle's representation in what follows has suffered from his own importance.

After saying something in general about the Greeks and the history of philosophy, and about the special positions of Plato and Aristotle, I shall try to convey some idea of the variety of Greek philosophical interests; but, more particularly, I shall pursue two or three subjects in greater detail than any attempt at a general survey would have allowed, in the belief that no catalogue of persons and doctrines is of much interest in philosophy, and that a feel for what certain thinkers were about can be conveyed only through some enactment of the type of reasons and arguments that weighed with them: of not just what, but how, they thought. In this spirit, if still very sketchily, I shall take up some arguments of Greek philosophers about two groups of questions--on the one hand, about being, appearance, and reality, on the other about knowledge and scepticism. In both, the depth of the Greek achievement is matched by the persistence of similar questions in later philosophy. In another matter, ethical enquiry, I shall lay the emphasis rather more on the contrasts between Greek thought and most modern outlooks, contrasts which seem to me very important to an understanding of our own outlooks and of how problematical they are.

I have said that the Greeks initiated most fields of enquiry in philosophy, and many of its major questions. It may be, by contrast, that there are just two important kinds of speculation in the later history of philosophy which are so radically different in spirit from anything in Greek thought as to escape from this generalization. Greek philosophy was deeply concerned, and particularly at its beginnings, with issues involved in the contrast between monism and pluralism. It is not always easy to capture what was at issue in these discussions: in some of the earlier Greek disputes, the question seems to be whether there is in reality only one thing or more than one thing, but--as we shall see later--it is not easy to make clear what exactly was believed by someone who believed that there was, literally, only one thing. In later philosophy, and already in some Greek philosophy, questions of monism and pluralism are questions rather of whether the world contains one or more than one fundamental or irreducible kind of thing. One sort of monism in this sense which has been known both to the ancient and to the modern world is materialism, the view that everything that exists is material, and that other things, in particular mental experiences, are in some sense reducible to this material basis. Besides dualism, the outlook that accepts that there are both matter and mind, not reducible to one another, philosophy since the Renaissance has also found room for another kind of monism, idealism, the monism of mind, which holds that nothing ultimately exists except minds and their experiences. It is this kind of view, with its numerous variations, descendants, and modifications, which we do not find in the ancient world. Largely speculative though Greek philosophy could be, and interested as it was in many of the same kinds of issues as those which generated idealism, it did not form that particular set of ideas, so important in much modern philosophy, according to which the entire world consists of the contents of mind: as opposed, of course, to the idea of a material world formed and governed by mind, a theistic conception which the Greeks most certainly had.

The other principal element in modern philosophy which is independent of the Greeks is something that first established itself at the beginning of the nineteenth century--that type of philosophical thought (of which Marxism is now the leading example) which places fundamental emphasis on historical categories and on explanation in terms of the historical process. The Greeks had, or rather, gradually developed, a sense of historical time and the place of one's own period in it; and their thought also made use of various structures, more mythological than genuinely tied to any historical time, of the successive ages of mankind, which standardly pictured man as in a state of decline from a golden age (though an opposing view, in terms of progress, is also to be found). Some of the more radical thinkers, moreover, regarded standards of conduct and the value of political arrangements as relative to particular societies, and that conception had an application to societies distant in time. But the Greeks did not evolve any theoretical conception of men's categories of thought being conditioned by the material or social circumstances of their time, nor did they look for systematic explanations of them in terms of history. This type of historical consciousness is indeed not present in all philosophical thought of the present day, but its absence from Greek philosophy is certainly one thing that marks off that philosophy from much modern thought.

It may be that these two, idealism and the historical consciousness, are the only two really substantial respects in which later philosophy is quite removed from Greek philosophy, as opposed to its pursuing what are recognizably the same types of preoccupation as Greek philosophy pursued, but pursuing them, of course, in the context of a vastly changed, extended, and enriched subject-matter compared with that available to the Greeks.

This is not to say that the Greeks possessed our concept of 'philosophy': or, rather, that they possessed any one of the various concepts of philosophy which are used in different philosophical circles in the modern world. Classical Greek applies the word philosophia to a wide range of enquiries; wider certainly than the range of enquiries called 'philosophy' now, which are distinguished from scientific, mathematical, and historical enquiries. But we should bear in mind that it is not only Greek practice that differs from modern practice in this way: for centuries 'philosophy' covered a wide range of enquiries, including those into nature, as is witnessed by the old use of the phrase 'natural philosophy' to mean natural science--The Mathematical Principles of Natural Philosophy is what Newton, at the end of the seventeenth century, called his great work on the foundations of mechanics. It does not follow, however, that these ages did not have some distinction between scientific and what would now be called philosophical enquiries--enquiries which, however they are precisely to be delimited, are concerned with the general presuppositions of knowledge, action, and values, and proceed by way of reflection on our concepts and ideas, not by way of observation and experiment. Earlier ages often did make, in one way or another, distinctions between such enquiries and others--it is merely that until comparatively recently the word 'philosophy' was not reserved to marking them.

It is important to bear this point in mind when dealing with the philosophy of the past, in particular ancient philosophy. It defines, so to speak, two grades of anachronism. The more superficial and fairly harmless grade of anachronism is displayed when we use some contemporary term to identify a class of enquiries which the past writers did themselves separate from other enquiries, though not by quite the same criteria or on the same principles as are suggested by the modern term. An example of this is offered by the branch of philosophy now called 'metaphysics'. This covers a range of very basic philosophical issues, including reality, existence, what it is for things to have qualities, and (in the more abstract and less religious aspects of the matter) God. There is a set of writings devoted to such subjects in the canon of Aristotle's works, and it is called the Metaphysics; and it is indeed from that title that the subject got its name. But the work was probably so called only from its position in the edition of Aristotle's works prepared by Andronicus of Rhodes in the first century B.C.--these treatises were ta meta ta phusika, the books that came 'after the books on nature'. Aristotle's own name for most of these metaphysical enquiries was 'first philosophy'. Nor is it just the name that was different, but so were the principles of classification, both in the rationale given of them and hence in what is included and excluded. Thus Aristotle has an account of his enquiries into 'being in general' which relates the themes of 'first philosophy' in a distinctively Aristotelian way to the rest of knowledge (roughly, he supposed that it was distinguished by having a subject-matter which was much more general than that of other enquiries); and it excludes some enquiries which might now be included in metaphysics, such as a priori reflections on the nature of space and time. These latter Aristotle takes up in the books now called the Physics, which were included among the books 'about nature'; the name Physics itself being misleading, since what their contents mostly resemble is parts of metaphysics, and also what we would now call the philosophy of science, rather than what we now call physics.

These various differences do not stop us identifying Aristotle's enquiries as belonging to various branches of philosophy as we now understand them: this level of anachronism can, with scholarship and a sense of what is philosophically relevant, be handled--as it must be, if we are going to be able to reconstitute from our present point of view something which it would not be too arbitrary to call the history of philosophy. But there is a second and deeper level of anachronism which we touch when we deal with writings to which modern conceptions of what is and what is not philosophy scarcely apply at all. With those writers who did not themselves possess some such distinctions, to insist on claiming them for the history of philosophy as opposed to, say, the history of science, constitutes an unhelpful and distorting form of anachronism. So it is with the earliest of Greek 'philosophers', the earlier Presocratics (a label which as a matter of fact is used not only for thinkers earlier than Socrates, but for some late-fifth-century contemporaries of his as well).

With regard to the earliest of Greek speculative thinkers, Thales, Anaximander, and Anaximenes, who lived in Miletus on the Greek seaboard of Asia Minor in the first seventy years of the sixth century B.C., it is impossible to give in any straightforward modern terms a classification of the kinds of question they were asking. This is not just because virtually nothing remains of their work (Thales, the oldest, in any case wrote nothing) and we have to rely on disputable reports; even if we had all their writings we could not assign them, in modern terms, to philosophy or to science. They are usually represented as asking questions such as 'what is the world made of?', but it is one achievement of intellectual progress that that question now has no determinate meaning; if a child asks it, we do not give him one or many answers to it, but rather lead him to the point where he sees why it should be replaced with a range of different questions. Of course, there is a sense in which modern particle theory is a descendant of enquiries started by the Milesians, but that descent has so modified the questions that it would be wrong to say that there is one unambiguous question to which we give the answer 'electrons, protons, etc.' and Thales (perhaps) gave the answer 'water'.

We can say something--and we shall touch on this later--about the features of these speculations which make them more like rational enquiries than were the religious and mythological cosmologies of the East, which may have influenced them. And this is in fact a more important and interesting question than any about their classification as 'philosophy', something which in the case of these earliest thinkers is largely an empty issue.


The involvement of Greek philosophy in the Western philosophical tradition is not measured merely by the fact that ancient philosophy originated so many fields of enquiry which continue to the present day. It emerges also in the fact that in each age philosophers have looked back to ancient philosophy--overwhelmingly, of course, to Plato and Aristotle--in order to give authority to their own work, or to contrast it, or by reinterpretation of the classical philosophers to come to understand them, and themselves, in different ways. The Greek philosophers have been not just the fathers, but the companions, of Western philosophy. Different motives for this concern have predominated in different ages: the aim of legitimating one's own opinions was more prominent in the Middle Ages and the Renaissance (which, contrary to popular belief, did not so much lose the need for intellectual authority, as choose different authorities), while the aim of historical understanding and self-understanding is more important in the present day. But from whatever motive, these relations to the Greek past are a particularly important expression of that involvement in its own history which is characteristic of philosophy and not of the sciences.

It has been a characteristic also of literature, though the nature of the involvement in that case is very different. It has been suggested2 that our conceptions of Western literature have room for the notions both of a 'relative classic'--a work which endures and has influence and stands at least for a period of time as an exemplar--and of the 'absolute classic', above all the Aeneid, which defines for ever the high 'classical' style. Adapting these notions to philosophy, we might say that the classical philosophers Plato and Aristotle are classics in the sense that it has been impossible, at least up to now, for philosophy not to want to make some living sense of these writers and relate its positions to theirs, if only by showing why they have to be rejected: this is a status which they have shared, in the last 200 years, only with Kant. But they might be said also to define a classical style of philosophy--meaning by that a philosophical, not a literary, style. They are both associated with a grand, imperial, synoptic style of philosophy; though beyond that very general description, they have been acknowledged from ancient times to define two different styles, Plato being associated with speculative ambitions for philosophy, seeking to establish that another world of intellectual objects, the Forms, accessible to reason and not to the senses, was ultimately real, while Aristotle renounced these extravagant other-worldly hypotheses in favour of a more down-to-earth, classificatory, and analytical spirit, more respectful of the ordinary opinions of men--but defining a grand style for all that, since the systematic impulse was directed to producing one unified, ordered, and hierarchical world-picture.

Oppositions of the Platonic and the Aristotelian spirits have been a commonplace. In our own century, Yeats wrote, in Among School Children:

Plato thought nature but a spume that plays
Upon a ghostly paradigm of things;
Solider Aristotle played the taws
Upon the bottom of a king of kings . . .

Most famously, the received contrast is expressed in Raphael's fresco in the Vatican called The School of Athens, which displays the two central figures of Plato and Aristotle, the one with his hand turned towards heaven, the other downwards towards earth. In this connection one must remember the mystical elements which were associated with Platonic thought: not altogether falsely, so far as some of Plato's own writings are concerned, but very heavily selected for and modified by the Neo-Platonist tradition. It is connected with this image of Plato that for a period in the early Middle Ages only the Timaeus (in Latin translation) was known, an untypical dialogue in which a theistic cosmogony is advanced.

Looked at more than superficially, the famed contrast is a very complex and ambiguous matter. The spirit of Plato has sometimes been associated with the religious impulse as such; but equally, and in fact more importantly, where the framework of thought is already religious, an expanded Aristotelianism has represented an ordered and stable understanding of the world in relation to God, while Platonism has been taken to represent variously humanism, magic, or individual rational speculation.

The old picture by which the Middle Ages built on Aristotle, but the Renaissance got its inspiration from Plato, has been much qualified by modern scholarship, but it retains enough truth,3 and more than one important Renaissance thinker agreed with the words of Petrarch, that Plato 'in that group came closest to the goal that may be reached by those whom heaven favours'. Much of this Platonic influence flowed into humane studies and the betterment of the soul, rather than the study of nature; and where the study of nature is pursued in the Renaissance, outside the continuing traditions of Aristotelian science, there is deep uncertainty and disagreement about what kinds of procedure or lore may prove effective in uncoding the messages hidden in phenomena. It is rather later, and with a vision much closer to modern conceptions of mathematical physics, that Galileo expressed what is still a Platonic influence in saying:

(Natural) philosophy is written in that vast book which stands forever open before our eyes, I mean the universe; but it cannot be read until we have learnt the language and become familiar with the characters in which it is written. It is written in mathematical language, and the letters are triangles, circles, and other geometrical figures, without which means it is humanly impossible to understand a single word. (Il Saggiatore, Question 6)

From this point on, the business of decipherment could be more readily detached from notions of an arcane mystery, which were present in the Renaissance, as they were originally in the early Pythagorean sects which influenced Plato; it could become the public task of critical scientific discussion.

Thus in one context Platonism may represent a mystical or cabbalistic interest, against which Aristotelianism stands for a cautious, observational approach, concerned to stick to the phenomena; in another, while a Platonic influence encourages rational enquiry into nature, Aristotelianism can be seen (as it was by Descartes, despite his occasional dissimulations) as an obscurantist attachment to mysterious essences and muddled vitalistic analogies. An opposition of the Platonic and Aristotelian spirits is indeed something real, which can be traced through very complex paths in the history of Western thought; but it defines not so much any one contrast, as rather a structure within which a large number of contrasts have in the course of that history found their place.

Various as these contrasts have been, what can be said is that the majority of them have been associated with interpretations of these philosophers' views and, in many cases, with what have been believed to be their systems. Under these various interpretations, they have still been seen as authors of large world-views, as classical system-builders. Modern scholarship, encouraged by a philosophical scepticism about system-building, has tended to reduce the extent to which these philosophers are seen as expressing systems. In both cases, their works are now more clearly seen as the product of development over time, with corresponding changes of outlook; while discussions which in the past were taken to be fundamentally expository can be seen to be more provisional, exploratory, and question-raising than was supposed. If this point of view is accepted, does it mean that the importance of Plato and Aristotle, as more than a purely historical recognition, will for the first time be radically reduced? Perhaps not: the power and depth of their particular arguments may come to be what command admiration and interest rather than the breadth and ambition of their systems. Yet it would be superficial to rest too easily on this idea. The interest that these two philosophers have always commanded in the past has been generated not merely by admiration for their undoubted acuity, insight, and imagination, but, very often, by a belief that they had vast and unitary systematic ambitions, of a kind which we now have rather less reason to ascribe to them.

Apart from these issues of how the work of Plato and of Aristotle is to be interpreted, there are in any case other, more general influences likely to affect their traditional standing. Those features of twentieth-century culture which have weakened the hold of the classic, and of the idea that past works can have any authority over the taste of the present, apply in some degree to philosophy. Past geniuses of philosophy, as of the arts, look different under the influence of our idea, deeply felt and largely correct, that twentieth-century experience is drastically unprecedented. Again, in more technical areas of contemporary philosophy, there have been developments from which some of it has attained the research pattern of a science, and in any such area its interest in any of its past, let alone its Greek past, becomes necessarily more external and ultimately anecdotal. For both these reasons, the role of an absolute classic in philosophy, the role which Plato and Aristotle have peculiarly played, is one that quite conceivably may lose its importance. The question here is not whether philosophy might cease to be of interest--there is more than one dispiriting kind of reason why that might prove to be so--but whether, granted philosophy retains its interest, Plato and Aristotle might not do so, and might become finally historical objects, monumental paradigms of ancient styles. It is not impossible, but if it were to happen at all, there is one reason why it is less likely to happen to Plato than to his great companion: the fact that Plato's work includes as a vivid and independent presence the ambiguous figure of Socrates, whose aspect as ironical critic of organized philosophy can be turned also against the Platonic philosophies, which at other points he is presented as expounding.


The pre-eminent status of Plato and Aristotle is both the cause and the effect of their work being quite exceptionally well preserved: though in the case of both, and particularly of Aristotle, there was some luck involved. Of Plato's works, we have everything that he is known to have published. Of Aristotle, we do not have his dialogues (for which he was most admired in antiquity), but we do have a large body of treatises which contain material prepared by him or in some cases by close associates or students.

Work later than Aristotle will not in general be touched on here except for some discussion of ancient scepticism; but we should not forget the large influence exercised on Western thought by the later schools, particularly the Stoics and Epicureans, quite apart from those influences on Christianity which are discussed in another chapter. The Presocratics will be of closer concern to us. They are known to us through fragments of their writings, and in many respects the situation is as described in the chapter on Greek science, that we have to rely on summaries and accounts by later writers, who may be remote in time, or stupid, or--as in the case of Aristotle, who was neither--have their own axe to grind. There does remain one very considerable and nearly continuous fragment of Presocratic writing, asubstantial amount of the poem of Parmenides (who was born probably c. 515 B.C.): we owe this entirely to the Neoplatonist scholar Simplicius, who, in the commentary on Aristotle's Physics which he wrote in the sixth century A.D., copied out long extracts from the poem on the ground that Parmenides' book had at that time become very rare. Thanks to Simplicius, we have enough to reconstruct a continuous argument (which we shall turn to in the next section).

By contrast, another challenging figure, Heracleitus, who was almost certainly rather earlier than Parmenides and perhaps born c. 540 B.C., is known only from an assemblage of brief disconnected fragments, conflicting and obviously puzzled reports, and a number of unreliable anecdotes illustrating an original, pessimistic, and contemptuous personality. In his case, it is not clear what has been lost, either in terms of works, or, indeed, in possibility of understanding: it seems anyway that he wrote in the form of brief and dense epigrams, and he was famed already in antiquity for his obscurity. Plotinus said of him (Enneads, 4. 8), 'He seems to speak in similes, careless of making his meaning clear, perhaps because in his view we ought to seek within ourselves, as he himself had successfully sought.' The idea of searching within oneself was in Heracleitus, as it was in Socrates: 'I searched myself', Heracleitus said (fr. 101).4 But he was almost certainly not careless of making himself clear: rather, his conception of truth was of something that essentially could not be expressed in a direct, discursive way. He probably thought of philosophical speech as he said of 'the king whose oracle is at Delphi: he does not say, and he does not conceal--he gives a sign' (fr. 93). In this, Socrates vitally differed from him.

Heracleitus' views, so far as they can be discovered, centre on the necessity to the cosmos of constant change and 'warfare' between opposing principles, though these are held in some kind of reciprocal relation and balance. They elicited, at some remove, the respect of Lenin, but it was undoubtedly Nietzsche's admiration that came closer to him--and not only because of his greater sympathy for Heracleitus' contempt for the masses. Heracleitus has seldom had followers, but his deliberate ambiguities and startling images (as in fr. 52: 'time is a child at play, playing draughts: the kingdom is a child's') have contributed to the deep resonances which he has occasionally evoked in later philosophy, most recently in Heidegger.

Of the later Presocratics, again no complete work survives; the most numerous fragments are of two contrasted writers. One is Democritus (roughly a contemporary of Socrates, being born c. 470 B.C.), who was concerned both with ethical questions and with the explanation of natural phenomena; he is most famous as one of the first theorists of physical atomism. The other is the riddling figure of Empedocles, who came from Acragas in Sicily, and was falsely represented by tradition as having died by throwing himself into Etna. He wrote not later than 450 B.C. two poems which later came to be called On Nature and Purifications. As these titles suggest, naturalistic elements, an interest in physical explanation, coexisted in his thought with a religious strain, and it is still disputed how they were combined, and whether his interest in nature was subordinate to magical concerns rather than the product of curiosity and free enquiry. Curiosity did to some extent motivate the Milesian thinkers, and free enquiry was consciously practised by Democritus and others of similar temper, such as the ingenious thinker Anaxagoras (born c. 500 B.C.), who is said to have been prosecuted by the Athenians for holding an impiously naturalistic view of the heavenly bodies.

There is one further group among the predecessors and contemporaries of Socrates, whom we should mention at this point: the so-called 'Sophists', whose interests were neither cosmological nor religious, but more practically orientated, largely towards the training of pupils in techniques for political and forensic success, a training for which they received money. These activities earned them an extremely bad reputation from Plato, whose attitude to them, expressed in all modes from the glittering mockery of the Protagoras to the contempt and disgust of the Gorgias and Republic, has not only left the Sophists in low esteem, but has helped to make the word 'sophist' useless for any historical purpose. This is particularly because Plato tended to conflate four different charges against them: that their teaching had a practical rather than a purely theoretical bent; that they took money; that they produced bad arguments, designed to puzzle and impress rather than to get at the truth; and that they advanced cynical, sceptical, amoral, and generally undesirable opinions.

It is not at all easy to disentangle these elements, nor to establish how far the Sophists, or some of them, had what we would now identify as genuine philosophical interests. They were prone to confuse, as has been well said,5 the force of reason and the power of the spoken word, two things which Socrates' method of question and answer gave a way of taking apart. Gorgias of Leontini, a celebrated stylistic innovator who influenced one of the greatest geniuses among Greek writers, the historian Thucydides, was a teacher of rhetoric whose excursion into metaphysics, a lost work called On What Is Not, may well, to judge from later summaries of it, have been parodistic. But more serious claims can be made for Protagoras of Abdera (born c. 490 B.C.), who commanded enough interest and respect from Plato for him to construct in his Theaetetus a sensitive elaboration of a relativistic theory of knowledge starting explicitly from a Protagorean basis. It may be true, as a recent writer has said,6 that 'he dominated the intellectual life of his time without being a truly original thinker', setting rather an intellectual tone, sceptical and irreverent; but it is possible that he articulated more searching and systematic thoughts about knowledge and society than this implies. It would be interesting to know more of his work than we do, both to learn about the radical strain in fifth-century thought, and to form a more detailed idea of developments, which certainly occurred, in the theory of knowledge and the philosophy of language before Plato. It would be interesting, too, as a matter of sheer curiosity, to know how he continued his book On the Gods, of which we have only the discouraging first sentence: 'About the gods I cannot know, whether they exist or not, nor what kind of beings they might be; there are many obstacles to knowledge, both the obscurity of the subject and the shortness of man's life.'


Greek philosophy started at the edges of the Greek world: on the offshore islands and the western seaboard of Asia Minor--Ionia--and to the far west, in the Greek colonies of Southern Italy and Sicily. The latter were not in any case independent of Ionian influence. Many received new colonists from Ionia after it was annexed by the Persians in the sixth century, and, in particular, the city of Elea in Southern Italy, famous for the philosophy of Parmenides and his pupil Zeno (thus called the 'Eleatics'), was founded by the citizens of Phocaea, a city in Ionia, who had emigrated in large numbers.

The question has been much and inconclusively discussed, of why systematic cosmological thought, embodying an element of rational criticism, should have arisen in Ionia at this time. The great empires of the East had acquired a good deal of empirical information about measurement, positional astronomy, and such matters, while the Babylonian tradition embodied considerable sophistication in mathematical computation, though with little impulse, it seems, to discover an a priori order in the mathematical subject-matter. These various techniques, moreover, coexisted with pictures of the origin and structure of the universe which were straightforwardly mythological. Knowledge of these beliefs, transmitted through the Persian empire, may have played a role in the formation of Ionian cosmology, but, if so, they were essentially modified in a more critical and less mythological direction. The relatively autonomous political life of the small Greek cities perhaps played a part in the growth of critical and reflective thought, as contrasted with those 'Asiatic vague immensities', in Yeats's phrase, of the great empires.

Open speculative enquiry was a necessary condition for the development of Greek philosophy, but it would be a mistake to think that everything which eventually fed into it was equally an example of that openness. The Milesians and Eleatics formed 'schools' only in the sense that these thinkers were connected by ties of intellectual influence and teaching; the Pythagorean school, on the other hand, which was founded in Croton in South Italy towards the end of the sixth century, was more like a religious brotherhood or secret society. The history of this school and its founder is wrapped in obscurity and legend, though we know that Pythagoras himself was another who emigrated from Ionia, having been born and having gained a reputation in Samos. The Pythagorean school played an important, if much disputed, role in the development of mathematics, though it is doubtful to what extent those studies figured in its earlier years. Its life was devoted, more certainly, to an ascetic religious discipline centring on concepts of the purification of the soul and reincarnation: ideas and practices perhaps influenced by shamanistic beliefs which would have reached Greece through the Thracians and Scythians.

Pythagorean ideas were to play an important part in the development of the idea of a rational, immaterial soul, separate from the body, an idea which was much developed by Plato; and which passed from him through Augustine to become the basis of Descartes' dualism--though it lost, in the context of seventeenth-century mechanical science, a basic feature which it shared with all Greek ideas of 'soul', namely the conception that it was the presence of soul which gave living things their life. (Descartes marked the difference when he said something which Pythagoreans, Plato, Aristotle, would all equally have been unable to understand: 'it is not that the body dies because the soul leaves it--the soul leaves it because the body has died.') Whatever exactly the early Pythagoreans did and believed, they did it in secret, and the concept of being initiated into a mystery applies in their case better than that of making an intervention into an open rational debate. Their existence, contemporary with the later Milesians, reminds us also of something else: that from its beginnings two motives were brought to Western philosophy which have been active alongside one another ever since, the desire for salvation and the desire to find out how things work.

We have already suggested that while the question of how far Milesian thought was philosophical is an unhelpful one, the question of how far its enquiries were rational may be a better one. There are of course various criteria of rationality, but certainly one very important expression of it is to be found in reflection, guided by general principles, on what questions require an answer. There is a very striking example of such thought in a famous argument of Anaximander, who worked in the first half of the sixth century and, according to an ancient writer, was 'the first of the Greeks, to our knowledge, who was bold enough to publish a book on nature'.7 The argument relates to a question which bothered other Presocratics: what keeps the earth in its place? Others were to appeal to material supports of various kinds: but Anaximander argued that the earth was symmetrically placed in the centre of the universe, and thus needed no support. This argument represents an early application of a purely rational principle, the Principle of Sufficient Reason. If the earth were to move in one direction rather than another, there would have to be a reason for this, in the form of some relevant asymmetry or difference: so, if there is no such asymmetry, the earth will not move in any direction rather than another, i.e. will stay where it is. This impressive argument brings out clearly how the application of rational principle, even if it is to basically primitive cosmological materials, marks out such thought from mythological picture-making.

A different exercise of rationality, however, and a much more purely abstract one is represented by the extraordinary work of Parmenides. Parmenides expressed his philosophy in verse, a choice less eccentric than it would be now, but still a choice (the Milesians wrote in prose). The effort to express abstract and logical considerations in epic hexameters gives an intense but also strained effect, and his style was poorly viewed in antiquity. His aim throughout seems to be to achieve as much clarity as possible, and the syntactical obscurities that remain are the unintended results of the language being drastically bent to his unprecedented subject-matter. His ambiguities are thus of a very different kind from the revelatory puns of Heracleitus. Even the very little we have of Heracleitus (compared with 154 lines of Parmenides) shows that he was the more controlled and sophisticated writer; but Parmenides was attempting something quite different from him or anyone else before,8 which was to determine the basic nature of reality entirely by argument from premisses self-evident to reflection--just one premiss, in fact, though Parmenides (fr. 5) says that it makes no difference where one starts. Whatever exactly we say about the Milesians, in this undertaking we can certainly recognize the first example of pure metaphysical reasoning: it remains one of the most ambitious.

Parmenides' poem represents a goddess as revealing to him the true way of enquiry. What she gives as the key to the true way is this: 'it is, and it cannot be that it is not.' We must not try to think 'it is not': for 'you could not know what is not (that is not practicable), nor speak of it (fr. 2). The same thing is there to be thought, and to be (fr. 3). What is there to be spoken of and thought, must be; for it is there to be, but nothing is not (fr. 6).'

We will leave for the moment the question of what is meant by 'it' in 'it is' and 'it is not' (in the Greek the verb 'is' stands by itself). Parmenides' first conclusion is that there is no coherent or possible enquiry into what is not, or which uses the thought 'it is not'; this is because 'the same thing is there to be thought, and to be', and what is not, nothing, is not available, so to speak, for thought. Parmenides' ultimate backing for this radical claim is hard to recapture with total precision, and is still the subject of controversy. Some believe that the basic argument (as given in fr. 6, 1-2, the last sentence quoted above) is this: with regard to what can be thought and spoken of, it is true (at least) that it could be--and this might be conceded even by those of us who suppose that some things which can be thought and spoken of do not, as a matter of fact, exist (unicorns, for instance). But now consider: of nothing, it is not true that it could be. So a thing which can be thought and spoken of cannot be identical with nothing. But then it must be something; and so, contrary to what you first thought, must actually be.

This is at least a clear fallacy. But in the strange phrase translated as 'is there to be', Parmenides has a more primitive conception than this version captures, a notion of language and thought having a content only because they touch or are in contact with what is--the touching and seeing models of thought and meaning operate more directly on Parmenides' ideas than is quite brought out by the excursion through what could be. But however exactly we are to reconstruct Parmenides' rejection of the thought 'it is not', his rejection of it is clear and total, and he proceeds to deduce from that rejection, in order, a series of surprising consequences. What is can have no beginning or end; if it had, then, before or after, it would not be, and that is excluded. He adds to this proof of 'its' having no beginning, another based on an elegant use of the Principle of Sufficient Reason: 'what necessity would force it, sooner or later, to come to be, if it started from nothing?' (fr. 8, 9-10).

'It neither was nor will be, since it is altogether now': here Parmenides gives the first expression to an idea of eternity. His conception is not of something outside time altogether, as some later conceptions of eternity have it, something to which no temporal notions apply at all. It is now. But, equally, it is not merely indefinitely old--it has no past, and no future. Its time, such as it is, is represented as a perpetual present. 'It' is uniform, unchanging, has no divisions, is the same under any aspect--for to deny any of this would involve thinking that there was some place, or some time, or some respect, with regard to which it was not, and this, once more, is excluded.

Above all, there is only one of it. For 'there is and will be nothing besides what is' (fr. 8, 36-7)--anything else would have to be something which was not; and 'what is', itself, cannot consist of two distinguishable things or be divided, 'since it all, equally, is; it is not more or less in any way . . . so it is all continuous, for what is sticks close to what is' (fr. 8, 22-5). Once the uniqueness of 'it' is seen to be a conclusion of Parmenides' argument, and not (as some earlier scholars supposed) a premiss, the question of what 'it' is lapses. It is just that thing, whatever it is, that we are thinking and speaking of, when we succeed in thinking and speaking of something--and Parmenides certainly supposes that we can think and speak of something, though very evidently it is not what, in our everyday error, we take ourselves to be thinking and speaking of.

The philosophical legacy of this remarkable argument is very extensive and various. The concept of eternal, unchanging, and uncreated being is one which Plato was to use in characterizing his Forms; his debt to Parmenides was explicit and acknowledged, though he had to differ from him, as he gravely concedes in the Sophist, by admitting into reality also principles of change. He differed from him already, however, about the world of unchanging being: Plato had held, from the earliest introduction of the Forms, that there were many of them, which could be intellectually distinguished. How this could be, however, is something he did not take up until in that same late dialogue, the Sophist, he directly faced the challenge of Parmenides' proof and sought to meet it by systematically distinguishing different senses of 'is not'.

That same attempt was also to provide the solution, as Plato hoped, to another problem which directly related to Parmenides' argument, the problem of falsehood. To think, surely, is to think something--to think nothing is not to think at all. So what is the 'something' that is thought by one who thinks falsely? Thought or speech which is false cannot be nonsensical: what relation to reality is possessed by speech which has a meaning, but is not true? This problem Plato made a powerfully original attempt to solve, in the course of which he developed a distinction essential to these issues, that between a name and a statement. What is in many ways the same set of questions has recurred in increasingly sophisticated forms to the present day; and Wittgenstein's Tractatus, a metaphysical work comparable in both boldness and abstractness to Parmenides', takes its start from a question which implies the converse of Parmenides' principle: 'how can we say what is not?'

Other strains of Parmenidean influence come from his denials of plurality and change. His pupil Zeno invented a series of famous paradoxes which apparently deduce contradictions from the suppositions that there is plurality, or that motion is possible; paradoxes such as that of the Arrow, which (in its shortest form) runs thus: an arrow in flight occupies at each instant a space which is just its own length; but any body which at any time occupies just such a space is, with regard to that time, at rest; so the arrow is at rest at each instant; so it is at rest at every instant, that is to say, it does not move. These paradoxes gave rise to a complex debate which belongs as much to the history of mathematics as to philosophy, from which there emerged eventually the concepts of the continuum and of a limit. But even after mathematical techniques had been established to characterize the phenomena which Zeno thought could not be coherently characterized, there have remained philosophical problems about the application of mathematics to physical space and time in which some of Zeno's arguments have still played a role; while the method which he invented, of generating from a set of assumptions an infinite regress (or progress)--a method which can be used either destructively, or constructively to determine some infinite set of items--has remained an essential resource of analytical thought.

Apart from difficulties for common sense from the Eleatic arguments, there were particular problems for the most advanced form of theoretical pluralism, atomism, which held that the world consisted of atoms moving in empty space--for how was totally empty space to be conceptualized to avoid the Eleatic argument (which impressed others, such as Anaxagoras) that it would have to be nothing, and hence could not exist? It seems that the first of the Atomists, Leucippus (born near the beginning of the fifth century) asserted the existence of the void against the Eleatics by saying that void is not being, yet is--a formulation which seems too much like a contradiction. Aristotle's treatment of this subject in the Physics represents a great advance in the conceptualization of empty space, and although he himself does not accept a void, he does not include the Eleatic type of argument among the several bad arguments he uses against it. It is all the more remarkable that Descartes, in the seventeenth century, when he denied a vacuum on the basis of his own physics (which involved a close assimilation of matter and space), was able to use a startlingly Eleatic type of argument: 'If it is asked what would be the case if God removed all the matter from a vessel and let nothing else take the place of what had been removed, then the answer must be, that the sides of the vessel would be contiguous. For if there is nothing between two bodies, they must be next to each other' (Principles of Philosophy, ii. 18).


Parmenides' poem had a further part, mostly lost, in which the goddess expounded a pluralistic cosmological theory; which, however, she was committed to regarding as nonsense, and probably advanced only as a sophisticated example of the kind of thing she had warned against at the beginning of the poem, the way of ignorant mortals, who 'drift along, deaf and blind, amazed, in confused throngs: they think that to be and not to be are the same, and not the same' (fr. 6, 6-9)--that is to say, they think, confusedly, that it is possible for what is here and now, not to be at other times and places. There has been much discussion of what relation Parmenides supposed the opinions of men to bear to reality as he explained it. But if that discussion tries to rest anything on what it would be consistent for Parmenides to hold, it must recognize from the beginning the important fact that there is nothing which Parmenides could consistently hold on this subject. For the opinions of men certainly change, and are different from one another: so if everything is (literally) one and (literally) nothing changes, there are no such opinions.

This point applies just as much to the true thought of the instructed philosopher. Some interpreters have claimed that Parmenides believed being and thought to be one, that nothing existed except thought (Parmenides would thus be something like an idealist, in the sense in which it was claimed earlier that no ancient philosopher was an idealist). This view is based partly on highly resistible interpretations of two ambiguous lines (fr. 3; fr. 8, 34), but also on the argument that since Parmenides thought everything was one, and agreed that there was thought, he must have supposed, not being stupid, that thought was the one thing there was. But this type of argument ignores the obliquities of the metaphysical imagination. One might as well argue that since Parmenides thought everything was one, and conceded (since he refers to himself more than once) that he existed, he must have supposed that he was the only thing there was. It is clear9 that Plato regarded himself, more than a century later, as forcing Parmenides to face the question of the existence of thought as part of reality. Let us call Parmenides' one thing 'It'. Then one of Plato's points was that Parmenides agreed that there was at least a name of It; but if there is only one thing, then It must be that name; and since It is a name, then Its name must be the name of a name; so Parmenides' theory comes out as the view that there is only one thing in reality, a name which is the name of a name.

This mildly jocular argument contains in fact both a narrower and a wider point. Since naming is, by both Parmenides and Plato, closely connected with thinking, it raises the question of thought being part of reality, a question which Plato goes on to pursue. But it raises also the general issue of what it is to take a thesis like Parmenides' seriously. Is it, for instance, to take it literally? An Eleatic might reply that of course it was never meant to be taken in the literal way in which Plato's argument takes it; but then the question can be pressed, as it was repeatedly pressed against metaphysical arguments by G. E. Moore in the present century, of how it is to be taken. Moore himself was burdened by a prejudice that to take something seriously was to take it literally; we do not have to agree with that, in order justifiably to demand some directions from the speculative metaphysician about how to take him seriously. One guide about how to take him seriously is provided by the direction of his arguments: but in Parmenides' case, this gets us no further on, since his argument either proves nothing at all, or proves just that literal absurdity which Plato objected to.

Parmenides had a theory so simple and radical that, taken literally, it leaves no room even for what he regarded as correct thought. With regard to other, false, ideas, the deluded beliefs of men, and indeed the pluralistic world itself as it seems, he and his followers were disposed to relegate these to the category of 'appearance'.10 This contrast between appearance and reality can be aligned, as it is by Parmenides (fr. 7), with a contrast between sense-perception and reason: sense-perception is deluded by mere appearance, it is the power of reason that grasps reality. But such a distinction, whatever else may be said about it, does not solve the problem that we have been pressing on Parmenides. For even if men are deluded by the senses, and appearances conceal rather than reveal reality, at least it is true that there are appearances, and any full account of what actually exists must include the actual existence of (misleading) appearances. As the English twentieth-century metaphysician F. H. Bradley insisted, appearance must itself be part of reality.

The point was seen, once more, by Plato in that late dialogue the Sophist to which we have already referred. But it was a truth which Plato himself had to learn to take as seriously as it needs to be taken. In his middle-period dialogues, above all the Republic, he had offered a picture of knowledge and reality which was itself open to this criticism, or at least was deeply ambiguous on the issue. On the one hand, there was the world of Forms, immaterial and unchanging objects of purely intellectual knowledge, which were supposed, in that simple and ambitious theory, to solve a lot of problems at once: to explain, for instance, what mathematical truths are truths about (for evidently they are not about such things as the inaccurate geometrical figures one sees on blackboards), and, at the same time, to be what give general terms a meaning. Over against these, were the objects of sense-perception and everyday belief, the things of the natural world which are mistaken for reality by the 'lovers of sense-experience', who are contrasted with the philosophers, the lovers of truth.

In the Republic, the distinction between these worlds is hammered home by a series of dichotomies: in the model of the Divided Line, which separates the realm of Forms from that of matter, and assigns reason to the one, the senses to the other; and in an image which has haunted European thought, which represents the philosopher's education as a journey into the sunlight from a cave, in which ordinary men, prisoners of their prejudices, raptly watch a flickering procession of shadow images. This distinction, and the ordering of value that goes with it, Plato sometimes represents as one between 'being' and 'becoming', where 'becoming', we are told, is constituted by some unsatisfactory and unstable combination of being and not being. Plato was to abandon these formulations, though certainly not his belief in eternal intellectual objects.

Interpreters have not found it easy to capture exactly what Plato meant when, in the Republic and other dialogues of his middle period, he claimed 'real being' for the Forms, and denied it to the everyday objects of sense-perception. There is, in fact, more than one level of difficulty. There is the very general philosophical problem, which we have just touched on in the confrontation of Parmenides and Moore, of giving a sense to metaphysical assertions which deny the reality of some large and evident dimension of experience. Problems of that general sort are still with us. But there is also an historical problem, of understanding those particular metaphysical formulations which belong to a time before the development of any systematic logical theory, and which we are particularly likely to misrepresent in the light of later conceptions. Beyond that again, there is a very specific historical problem of understanding Plato, who seems himself to have become dissatisfied with some of these formulations and to have become, in his later work, a critic of his earlier self. If Plato became dissatisfied with these formulations, there is really not much reason to suppose that they ever had some fully determinate sense which we could now recover: it is rather that Plato is one of those who helped to put us into a position from which these formulations may be seen to have no fully determinate sense at all.

His later dissatisfaction with what he had said in the Republic lay in some part in technical issues about the idea of being: certainly he came to a clearer understanding of that notion, and also to a more patient and analytical conception of the kind of philosophical enquiry that such an understanding demanded. Related developments away from a simple Republic image occurred, as we shall see, in Plato's conception of knowledge.

It may be also that, more broadly, he became less governed by images of the rational mind being clogged or imprisoned by the empirical world. Those images themselves, it must be said, always stood in an uneasy relationship to another kind of picture which at the same time he offered of the material world, equally unfavourable to it, but in a contrary direction--that it was evanescent, flimsy, only appearance. The world of matter had to be ultimately powerless but at the same time destructively powerful, two conflicting aspects which stand to one another as the shadows of the Cave stand to the fetters which bind its prisoners. Such tensions express something very real in Plato's outlook (notably, his own deep ambivalence towards political power, and towards art), but their theoretical costs, for so ambitious a theory, are high, and Plato seems to have become aware of them.

The Republic theory, however, refuses to go away; it is perhaps Plato's most famous doctrine, and besides its appearances in history and literature as 'the Platonic philosophy', it itself, or at least its terminology, has recurred in many forms. Its tensions themselves help to explain how it keeps a hold on the philosophical imagination; and here one factor to be mentioned--which particularly relates to the associated Platonic doctrine of love, expressed in the Symposium--is that there is a constant and vivid contrast, in these middle-period works, between Plato's world-denying theories and his literary presentation of them. The resonance of his images and the imaginative power of his style, the most beautiful ever devised for the expression of abstract thought, implicitly affirm the reality of the world of senses even when the content denies it.

A more general point is that it is only philosophers and historians of philosophy who worry much about what is entailed by a theory such as that of the Republic when it is taken strictly. Others--artists, scientists--get what they need out of it, and if Plato's theory is taken broadly enough, much more can be got from it than is strictly in it. That includes the rationalist spirit so important to the seventeenth-century scientific revolution, which we have already referred to in the person of Galileo, the spirit which sought an underlying mathematical structure under the flux of appearances. It is clear how this can be thought to be in the spirit of Plato's Republic; it is clear also how it contradicts what is actually said there, since Plato quite explicitly says that there is no hope at all of giving a scientific account of the material world. His message is, and quite clearly, not that physics should be mathematical, but that one should give up physics and pursue mathematics. If philosophers are going to be influential, it is as well that they should be misunderstood.


Not all philosophical thought that regards reality as different from appearances need be as drastically dismissive of appearances as Parmenides, or the Republic taken strictly. It may rather encourage, like the Republic taken loosely, some rationalistic, perhaps scientific method for uncovering the reality from the appearances. One important difference between these attitudes is that the rationalist programme which finds an intellectual order under appearances may also find it to be, to some extent, systematically related to appearances, so that discovery of the hidden order can lead to control of what happens even as it appears: all control of the environment which is grounded in physical theory is of this character.

The fact that the view of the Republic was not really of this kind presented a serious difficulty to Plato. He was not, of course, interested in physical technology; but he was concerned with social technology, and the dream of the Republic is that philosophers, who have seen the truth about reality, would return to the Cave and, after their intellectual sight had adjusted to the darkness of empirical life, would be able to order things better than those who had never left. But despite some hopeful references to the paradigm that they carry in their memory, he does not provide enough to bridge the disjunction between the two worlds, and, as we shall see rather later, the theory of knowledge which he offers is unable to help in the basic task of political education.

The search for a coherent theory of how scientific knowledge might be possible was a preoccupation of some late-fifth-century thinkers. Anaxagoras had said (fr. 21A) that appearances were 'a glimpse of the hidden', and for this he was praised by Democritus, who evidently struggled with these questions. Democritus took the point that while sense-perception could be misleading, and thought had in some sense to get behind appearances, nevertheless it was only with the help of other perceptions that this could be done: 'colour, sweetness, bitterness, these are matters of convention', he interestingly said (fr. 125), 'and what there is in truth are atoms and the void'; but he represented the senses as replying, 'Poor mind, are you going to overthrow us when you take your beliefs from us? If you throw us, you fall over.'

In trying to resolve the epistemological problems of atomism (and we do not know how far he got), Democritus was facing not only the Eleatics, who thought that they knew something incompatible with his atomism, but also a range of Sophists who thought that they had arguments against anyone's knowing anything at all--or at least, anything of a theoretical, general, or scientific character. The inconclusive speculations of the earlier Presocratics, and in particular the mind-numbing conclusions of Eleatic logic, served to encourage attitudes of scepticism.

A general sense that certainty, at least on any large or speculative issue, is impossible, is itself an early phenomenon. But the Sophists, or some of them, pursued a more aggressive line against philosophical theory of any kind and the use of dialectic to support it; wishing in this to advance their own claims to teach something useful, in the form of rhetoric and the all-important power to persuade in the courts and the political assembly--activities in which, as they agreed with their critics, scrupulous logical demonstration was not at a premium. The arguments used in these attacks on the possibility of knowledge seem now a mixture of almost childish muddles or tricks, and penetrating insights into real difficulties; a few arguments embody both at once, as some of those recorded in Plato's Euthydemus, or found in a rather rough and ready compilation of dialectical material called the Dissoi Logoi or 'Double Arguments', which is generally taken to date from this period. We do not know how far Protagoras himself developed the positions which Plato ascribes to him, offering a relativized view of truth and knowledge, by which what seems to each man is true for him: but we do know that Democritus used against him, and may have invented, a form of argument which was to be very important in the later history of scepticism and the theory of knowledge. This form of argument is called the peritrope or 'reversal', and consists in applying a philosopher's criterion of knowledge, truth, or meaningfulness to his own statements--in this case, asking Protagoras whether his own thesis is supposed to be (non-relatively) true.

Some of the material which survives from these early excursions into scepticism seems naïve--naïve, that is, not just by some arbitrary standard of later logical theory, but by the contemporary standards of insight set by Herodotus or Thucydides or, differently, Sophocles--adult persons, compared (it seems) with clever children. The point is not about the individual psychological fact, of the talents or maturity of Sophists compared with those of historians or tragedians; the question is about the social fact, that these arguments were capable of genuinely impressing and bewildering the Sophists' contemporaries. The basic question is, as Nietzsche unforgettably said about Socrates, how did they get away with it? Here it is important to remember the gap that always exists between intelligent practice and the theoretical reflective understanding of that practice; and, more particularly, how utterly puzzling the theory of reasoning must have seemed at this point. On the one hand, there existed already startling intellectual achievements in mathematics, and some systematic thought about such subjects as medicine made sense; while, even more evidently, the practice of argument in everyday life could be seen to rest on some assumptions about the connexions between proof and truth--at the very least it was possible to show a person through dialectical refutation that he was contradicting himself and must be wrong somewhere. But at the same time, the Eleatic arguments--which were in fact deep and powerful--led to impossible results; it seemed that one could prove anything. Many other invalid arguments, neither deep nor powerful, could not be decisively shown up because no systematic vocabulary of logical criticism yet existed. The fundamental achievements of Plato and Aristotle in setting logic and the philosophy of language on their feet can conceal from us how random and unstructured reflective logical thought was before the fourth century.

Plato and Aristotle sought foundations for philosophical and (in Aristotle's case at least) scientific enquiry which would resist scepticism. Aristotle's theory of knowledge is complex, and no general account of it will be attempted here. It judiciously combined appeal to some intuitively or self-evidently known principles, with an important role for sense-experience. It also made a very characteristic appeal to the consensus of informed and thoughtful persons: Aristotle champions a programme which applies equally to metaphysics, ethics, and science, of considering and seeking to reconcile the views of the best authorities, and when he says that one's theory should accord with ta phainomena, 'the appearances', he includes in that not only data of observation, and what competent speakers would be disposed to say, but also, at least presumptively, existing well-entrenched theoretical opinions. The weight of proof, for Aristotle, is against those who would try to unseat such a consensus. Even granted that the strength of the presumption is not necessarily very strong, so that Aristotle can throw it over with some ease if he thinks he has a strong argument; granted, too, the element of preselection that Aristotle exercises in what is to count as a worthwhile opinion; nevertheless, the fact that he can hope to find any soil in which to ground such a method shows how far things have travelled by his time from the age of the Sophists.

For Aristotle, the advance of knowledge is a collective and on-going enterprise, to which earlier thinkers, unless too exotic, primitive, or capricious, can be seen as contributors. That idea exists powerfully today in the conception of a scientific community, whose practitioners are recruited through an apprenticeship in experimental and observational techniques, and again there is a presumption in favour of expert consensus. But in a world where there were few experimental techniques, the question of who was to be counted as part of the informed consensus was interpreted differently, and there was a strong pull towards intellectual activity coming to be seen (as it scarcely was by Aristotle, but was by many later) as the scholastic undertaking of harmonizing the contents of authoritative books.

This methodological respect for an informed consensus provides a contrast between Aristotle's outlook and Plato's, something which emerges in particularly stark terms with regard to ethics. But there is another set of beliefs about knowledge which they share, and which has been of the greatest importance for the history of philosophy: beliefs which represent knowledge as, in more than one way, quite special, and in particular very different from mere belief or opinion, even true opinion. One idea of this kind is that real knowledge, as opposed to random true belief, should form a system, should be theoretically organized in a way which itself corresponds revealingly to the structure of the subject-matter. This idea relates most directly to an ideal body of scientific knowledge, an ideal which Plato (in relation to philosophy and mathematics) did much to form, and Aristotle carried much further. It can be seen, however, also as a condition on what it is for a particular person to know anything. It represents a person's thought as real knowledge only insofar as that thought approximates to the system--the knower is the savant, one in whom some part of the ideal body of theoretical knowledge is realized.

This requirement leaves out, needless to say, a good deal of what in everyday acceptance would count as knowledge. This divergence is increased when there is added a further idea, that organized theoretical knowledge can be had only of an unchanging subject-matter, that contingent and particular and changeable matters of fact are no subject for science. Taken together, these ideas yield the conclusion that no person's thought can strictly and properly be said to be knowledge unless it relates to a necessary and unchanging subject-matter. This conclusion--and there are other routes to it besides this--exerted a notable fascination on both Plato and Aristotle, and has since recurred in philosophy more than once.

In Plato's thought, a development on this subject can very clearly be followed. In the Meno, a dialogue which marks a boundary between the early and the middle period of his work, his views are in a rich and unstable solution. Faced with a sophistic puzzle about how it is possible to learn anything at all, he introduces for the first time the doctrine of anamnesis or 'recollection', which represents the process of learning as the recovery of opinions already in the soul but forgotten. In the dialogue, this process (or, more strictly, its earlier steps) is illustrated by a scene in which Socrates elicits from a slave-boy, by questioning, assent to a geometrical truth of which the boy had no conscious idea before. A great deal could be said about this famous doctrine, and the Pythagorean ideas of pre-existence, reincarnation, and immortality which Plato attached to it, sketchily in the Meno, but more extensively in the Phaedo. The present point, however, concerns only one feature of it: that as an account of learning, it could not really look appropriate to anything except a necessary or a priori subject-matter, such as mathematics. There is indeed something which is striking and demands explanation in the fact that one can elicit from a pupil, by argument, mathematical conclusions which have never occurred to him before; but no amount of Socratic questioning could elicit from anyone a set of particular facts of geography or history which he had not already, in the mundane sense, learned. The reader of the Meno, however, finds that Socrates seems to hold also all of the following: that knowledge can be acquired only by such 'recollection'; that there is a distinction between knowledge and mere true belief; and that this last distinction can be applied not only to mathematics, but also to contingent matters--we can distinguish between a man who knows the way to Larissa and a man who merely has true beliefs about it. If we accept the obvious fact that 'recollection' does not apply to such matters (and it is not entirely clear whether the Meno accepts that point or not), these claims produce an inconsistency.

However it may be with the Meno, there is no such inconsistency in the Republic, where Plato makes it clear that for him the distinction between knowledge and belief is a difference of subject-matter: they relate to those two ontological worlds represented by the Divided Line. This neat co-ordination, however, leads to absurd conclusions, compounded by the fact that at this stage Plato has no adequate theory of error. The consequence that there is no empirical knowledge presents a problem, of which we have already seen the outline in discussing the Cave, of how the philosophers' knowledge can play any constructive role in this world at all; for to apply knowledge to this world requires propositions which are about this world, and if no such proposition can ever be more than believed, then it is incurably obscure how the philosopher kings' knowledge can, with regard to the empirical world (which is where, reluctantly, they rule), make them better off than others. Not only can there be no empirical knowledge--equally there can, strictly speaking, be no mathematical or other a priori belief, and the situation of apprentice or lucky mathematicians (let alone mistaken ones), which had been discussed in the Meno, becomes indescribable. Plato has, indeed, got a place in his classification for something roughly analogous to a priori belief, but that, interestingly, concerns not so much individual knowers or believers, as the status of a whole subject, the partially axiomatized mathematics of his day, which he believed to lack foundations.

In his later work Plato went back to the view that knowledge and belief could relate to the same subject-matter, and he may very well have accepted that there was empirical knowledge. The Republic represents the high-water mark for him of a theory of knowledge controlled by the categories of subject-matter, by the ideal of a body of a priori knowledge, rather than by questions about what has to be true of someone who knows something (as contrasted, for instance, with someone who merely believes that same thing). This emphasis in the Republic deeply defeats Plato's own purposes. Plato's anxious question, to which he repeatedly came back, and to which the Republic was supposed to give the great answer, was how moral knowledge could be institutionalized and effective in society, as opposed either to the rhetoric of the Sophists, or to the unreasoned and hence vulnerable perceptions of conservative tradition. Knowledge had to be present in society in the form of persons who knew, and who commanded an effective theory of education. Real knowledge, and the ability to impart it--or rather elicit it--went together.

This idea helped in the understanding of the life of Socrates, for it served to join something which Socrates admitted, that he had no knowledge, with something that had to be admitted about him, that his influence did not necessarily make his friends better: it was a fact, which contributed to Socrates' condemnation, that among his associates were such men as the brilliant deserter, Alcibiades, and Critias, prominent among the Thirty Tyrants. Plato's theory of effective moral education was meant to complete the work, and the apology, of Socrates. The Republic's account of knowledge seems at first to yield just such a theory; but in fact it totally fails to do so. It says quite a lot about what it is for a body of propositions to be knowledge, and something about what it is for a person to acquire such knowledge, but it says ultimately nothing about the cognitive difference that that process is supposed to make to a person's handling of matters in the everyday world which, by ontological necessity, lie outside that body of knowledge altogether.

There is another way in which knowledge can seem to make quite special demands. This arises from considering the standards which should govern personal or individual knowledge; whereas the last line of thought was more concerned with the question of what constitutes an impersonal body of scientific theory. More intimate to the concept of knowledge itself, it was equally started in Greek reflection, and has played an even more prominent part in subsequent theory of knowledge. This is the idea that knowledge implies certainty; that an individual cannot be said to know a thing unless he is certain of it, where that implies not only that he feels utterly sure of it, but that--in some sense which it has been a repeated undertaking of philosophy to try to make clear--he could not, granted the evidence he has, be wrong.

This is not, as some modern philosophers have implied, a merely arbitrary condition on knowledge. It is a quite natural suggestion to arise from reflection on knowledge; by more than one route, perhaps, but one could be the following. Obviously, there is a distinction between knowing a thing, and being right about it by luck--even ordinary speech, which is lax about ascriptions of knowledge, distinguishes between knowing and guessing correctly (even where the guesser actually believes his guess). But now consider the condition of a man who believes on ample evidence that a given thing is true, but whose evidence is such that he might still be wrong. Then even if he is not wrong, that seems to be, relative to his state of mind, ultimately luck. Here we can take the case of two men, each of whom has, on two different occasions, exactly the same kind and amount of evidential basis for his belief in a certain kind of fact; but, as it happens, one is right and the other is wrong. There is real pressure to say that the one who, luckily, was right, did not really know, and a natural English phrase marks this exactly, when it is said of him that for all he knew he might have been wrong. By this kind of argument, it can be plausibly claimed that so long as one's evidence falls short in any way of conclusive certainty, one does not, even if one is right, really know.

It is just possible that this powerfully influential line of argument was sketched out near the beginning of Greek philosophy, by the poet Xeno-phanes of Colophon (born about the middle of the sixth century), who wrote lines which can be translated (fr. 34):

No man has discerned certain truth, nor will there be any who knows about the gods and all the other things I say: for even if by chance he says what is totally correct, yet he himself does not know it; appearance (or opinion) holds over all.

Plato in the Meno apparently refers to this as expressing the sceptical view that knowledge is unattainable because you would not know when you had attained it--which is another version of the demand for certainty. But Plato may have been wrong about Xenophanes' meaning; the sense is much disputed, but it is most probable11 that he speaks only of a distinction, itself very important to Greek thought, between what one has seen for oneself or established at first hand, and what can only be the subject of inference, such as questions about the gods. But besides some good reasons for so taking it, one bad one has been advanced:12 that on the view of the lines as expressing a general sceptical point, there is no way in which the second sentence could stand as a reason for the first--it would rather have to be a consequence. On the contrary, the second sentence might express a subtle and powerful reason for the first--'noone knows about these things, because if he did know, it would have to be more than luck that he was right, which it cannot be.' The trouble about this as an interpretation of Xenophanes is not that it is too weak an argument, but rather that it is, by a century or so, too sophisticated.

But what Xenophanes probably did not say was eventually said. The requirement of strong certainty having been deduced from the concept of knowledge, a variety of thinkers took the step of claiming that strong certainty, and hence knowledge, were not to be had. Plato attempted to answer such a sceptical conclusion, while sharing the premiss that knowledge demanded strong certainty. But the negative view recurred, and it is interesting that it was, much later, members of the school that Plato founded, the Academy, who made some of the more interesting contributions to the rather episodic intellectual movement which is called Scepticism.

Our knowledge of ancient Scepticism comes in good part from the writings of an undistinguished medical writer of the second century A.D. called Sextus Empiricus. Sextus himself belongs not to the Academic school of scepticism, but to that called 'Pyrrhonian', after Pyrrho of Elis (c. 360-275 B.C.); Pyrrho himself is a shadowy figure, whose views came to Sextus as reported and amplified by his pupil Timon and other writers. Later Pyrrhonism inherited from the Academic Arcesilaus the technique of laying alongside any set of evidences or supposedly convincing argument another with contrary effect, in order to induce total suspension of assent--an attitude which was expressed in a phrase which already had an earlier history in philosophy: ou mallon 'no more this than that'. The aim of this technique was practical, to achieve that state of mind which more than one ancient school made its aim, ataraxia, quietude of mind or freedom from disturbance.

The Pyrrhonists were careful to withhold assent even from the claim that there was no knowledge; they recognized that, expressed dogmatically, it would be open to the peritrope or charge of self-refutation, and this very reflection helped them to get rid of that dogma along with others. The sceptical proposition, they said, was like the purge which 'does not merely eliminate the humours from the body, but expels itself along with them'.13 Correspondingly, the slogan 'no more . . .' was to be taken, not as a theoretical statement or the right answer to a theoretical question, but as an element in a practice which leads to the same state as having the right answer would lead to, if there were such a thing as having the right answer. Ataraxia followed, for the Pyrrhonists, not on answering fundamental questions, but on being induced to give up asking them. They illustrated the point with a story of the painter Apelles, who, despairing of being able to paint a horse's foam, flung his sponge at the canvas, which produced the effect of a horse's foam.

The later Pyrrhonists criticized the Academic school, of which the outstanding figure was Carneades (c. 213-129 B.C.), for being less prudent in withholding assent, and accused them of dogmatism, for asserting definitely that there was no knowledge. It is clear that Carneades worked very directly on the conception of knowledge as entailing certainty. His target, and the focus of his problems, was set for him by the theory of knowledge advanced by the Stoic school, which had been founded c. 305 B.C. by a gruff eccentric, Zeno of Citium. The theory had been developed in the late third century by a figure important in the history of logic, Chrysippus. (A line of verse said of him that if he had not existed, neither would the Stoic school, and equally elegantly Carneades added, 'if Chrysippus had not existed, neither would I.') The Stoics' theory of knowledge cannot be discussed here, but it is notable for pursuing quite directly the requirements which follow from the argument set out earlier against 'luck': needing, as they believed, some certain criterion of truth, they had recourse to a supposedly self-validating state of mind, one which would eliminate the possibility that what was assented to could be false. They introduced the concept of a 'kataleptic impression'--a form of conviction which was supposedly both subjectively indubitable and objectively unerring. It was this that Carneades attacked, by trying to show that no impression which had the first of these characteristics could be guaranteed to have the second. This was the first enactment of a dispute which was to become central to much modern philosophy, above all through Descartes' appeal, in his notion of a 'clear and distinct perception', to what is, in effect, a kataleptic intellectual impression.

The views of ancient Sceptics are not altogether easy to reconstruct from the accounts, rambling and sometimes inconsistent, offered by second- or third-rate thinkers such as Sextus or Cicero. To some, and varying, degrees they were actually sceptics, denying the possibility of knowledge or indeed of truth, or Pyrrhonianly withholding assent even from these denials. But at the same time there were strains, particularly in Carneades, of what would in modern philosophy rather be called empiricism or positivism, which ascribes certainty only to statements about impressions of sense or subjective appearances, and emphasizes verifiability, the probabilistic character of all empirical inference, and the heuristic uselessness of deduction ( J. S. Mill's criticism of syllogistic inference as circular was anticipated by ancient Scepticism). It may be that to Greek thinkers the two strains of scepticism and of radical empiricism seemed more closely associated than they do in modern philosophy, where radical empiricism has sometimes been invoked (as by Berkeley) precisely against scepticism. But to Greek thought the distinction between appearance and reality was so basic, and knowledge so associated with reality, that knowledge which was merely of subjective appearances perhaps did not count as genuine knowledge at all. This is a large subject,14 but if this line of argument is correct, it illustrates once more a point made before, that subjective idealism was not a view which occurred to the Greeks.

What is certain is that both the empiricist and the more purely sceptical strains in ancient Scepticism were to be of great importance later. Sextus Empiricus was destined to be one of the most influential of Greek philosophical writers. The translation into Latin and printing of his works (1562, 1569) coincided with an intellectual crisis precipitated by the Reformation about the criterion of religious faith, and it has been shown15 how sceptical arguments from Sextus became important instruments in subsequent controversies. The weapons of scepticism were used both against, and in defence of, traditional religious faith. One style of defence was expressed by Montaigne, who emphasized the inability of man to reach knowledge, and his pretensions in trying to do so; among the innumerable considerations assembled to support this outlook are the arguments of ancient scepticism. A fideistic, unfanatical attachment to traditional religious belief emerges as the basis of the life of ataraxia. As he winningly puts it in his celebrated Apologie de Raymond Sebond: 'La peste de l'homme, c'est l'opinion de sçavoir. Voilà pourquoy l'ignorance nous est tant recommandée par nostre religion comme pièce pro-pre à la créance et à l'obeïssance.'

In sharp contrast is the attitude of Descartes, whose use of the armoury of sceptical devices in his Method of Doubt was designed to be pre-emptive, and to enable him to arrive at certainties which, as he put it, 'the most extravagant hypotheses of the sceptics could not overthrow'. Descartes goes through doubt, not to give up philosophy, but to establish it. Finding certainties, as he supposes, first about himself as a rational soul, then about God, then about the structure of the physical world, he attempts a project which is, in effect, to reverse the relation of Carneades to the Stoics: he advances beyond doubt to a new form of kataleptic impression, and it is significant that among his first and basic certainties are those about subjective states of mind which, we have suggested, neither the Stoics nor the Sceptics regarded very highly as truths about reality. But both Descartes' conception of certainty, and still more some of the propositions which he regarded as certain and which were essential to his system, lacked kataleptic effect on his critics, and Descartes' fundamental achievement, contrary to his hopes, was to help to radicalize doubt, not to eliminate it.

When Montaigne said that Christianity should be taken on faith, because all arguments defeat one another, he almost certainly meant what he said; when Hume and Bayle, in the eighteenth century, spoke in similar terms, they did not. By that time, Pyrrhonian ataraxia was to be found not in Christianity, but in as little enthusiasm as possible for any religious issue. In cultivating that, as much against militant atheists as against zealots of the Church, Hume was a genuinely Pyrrhonian thinker, as also in his conservative social views; and besides the standard sceptical material which he used, a basic element in his epistemology, the theory of 'natural belief', can be found crudely prefigured in Sextus.

Hume and the ancient Pyrrhonians had something else in common. For all of them, the rejection of philosophy was the eventual rejection of philosophy, and ataraxia a state of mind achieved by working at sceptical considerations and then letting natural belief have its sway, so that one ends up living calmly by the customs of one's society (or, rather, by some critical liberalization of them). These thinkers would not have been impressed by the suggestion that it might have been simpler never to have started reflecting at all; or if they express envy for those innocent of reflection, this attitude is formed and expressed at a level of self-consciousness which does not invite the reader to take it simply as it stands. Some, and notably the ancients, believed that people who had never embarked on any reflection did not in general experience ataraxia, but were rent by passions and prejudice; but even those who were less sure of that would not have favoured an educational or psychological regime which produced the benefits of a passionless rationality by entirely unphilosophical means. Scepticism remained an intellectual posture, and for all these thinkers, the Pyrrhonian outlook was both a minority state and (what is not quite the same thing) an achievement. The Pyrrhonist had, in relation to the rest of society, the role of a sage: a very quiet one.

This is one of several reasons why this posture is no longer possible. There is in modern society no serious role of a sage, as opposed to those of the expert, the commentator, or the entertainer. There is also no serious point of view, or at least none which can be publicly sustained, by which wars, calamities, and social upheaval can be quite so distantly regarded as Scepticism suggested they should be. Again, outlooks shaped by Romanticism and by modern psychological theory demand a deeper view of the emotions than Pyrrhonism had, and a more sceptical view of ataraxia itself. These points are well, if negatively illustrated by the thought of Bertrand Russell, whose philosophical stance in the theory of knowledge was, broadly, that of a twentieth-century Humean, but who notably failed to reconcile his social and moral concerns with his theoretical scepticism about ethics, or the strength of his feelings with his understanding of the mind. A book about Russell was called The Passionate Sceptic; while there could still be some outlook to which that phrase applied, it is notable, and a significant comment on Russell's own difficulties, that in the terms of ancient or even Humean Pyrrhonism, it is a contradiction in terms.


'The discussion is not about any chance question,' Socrates says towards the end of Book I of the Republic (352D), 'but about what way one should live.' The discussion was with the sophist Thrasymachus, who had claimed that it was only ever a second-best situation in which a man had reason to act in accordance with the requirements of dikaiosune--'justice' as we necessarily translate it, though in the Republic it covers a wide ground, and relates to all aspects of being concerned for others' interests as well as one's own. One often does have, according to Thrasymachus, a reason, as things are, for acting in this way, but this is only because one's power is limited--typically, by the greater power of another; one whose power was not so limited would have no such reason, and would be a lunatic if he put others' interests before his own. This view Socrates sets out to refute. Discontented with what he offers against Thrasymachus, and confronted rather later with a more sophisticated version of this kind of thesis, he is represented by Plato as spending the rest of the Republic in giving the ultimate answer to it.

Although the speaker is Socrates, and although the question of what exactly in the Platonic Socrates was Socratic is still unanswered,16 there would be much agreement that the Republic's answer was Platonic, but the problem was Socratic. It was a problem raised by Sophistic scepticism, a form of it more genuinely alarming than scepticism about cosmological speculation or logic, for in this case there existed recognizable and possibly attractive alternatives to the considerations displaced by sceptical criticism.

The nub of the sceptical attack was that there was no inherent reason for anyone to promote or respect anyone else's interests, and that the belief that there was such a reason was the product of various kinds of illusion: in particular, it stemmed from an innocent failure to see that the rules and requirements on people's conduct which were found in different societies obtained only 'by convention', a concept which for the Sophistic critics meant that such rules were social products, about which it could be asked whose interest they served. There were, on the other hand, perfectly good 'natural' motives to self-interested conduct, and this was well illustrated by the behaviour of agents where there was no such framework of convention, notably by the behaviour of one city-state to another--a set of considerations brilliantly and grimly represented in the famous 'Melian dialogue' in Book 5 of Thucydides' History.

Part of the problem was set by this kind of use of the concepts of 'nature' (phusis) and 'convention' (nomos),17 and the attendant question of what kind of life it was 'naturally' rational to live; together with the suggestion that it was 'naturally' rational to pursue self-interest, the ideally satisfying forms of life being represented, in some of the more uninhibited expositions, in terms of sheer gangsterism. Thrasymachus offers this kind of picture; at this level what is in question is not only an entirely egoistic conception of practical rationality, but also a very simple schedule of egoistic satisfactions, in terms of power, wealth, and sex.18 This set of considerations just in itself yields the materials of fear and envy, rather than any on-going structure of social relations, and indeed Thrasymachus' view, reduced totally to these elements, turns out to be even descriptively quite inadequate for any account of society.

However, this picture was superimposed on, and derived some appeal from, something different: a picture of a certain kind of social morality, which does offer some impersonal criteria of who is to be admired and respected, but finds them particularly in certain kinds of competitive success and inherited position--an aristocratic or feudal morality. It was from the context of such a social morality that the fifth and fourth centuries inherited the concept of arete, 'personal excellence' (the standard translation of this term as 'virtue' is only sometimes appropriate, and can be drastically misleading). This term carried with it certain associations which Plato, and probably Socrates, made strong efforts to detach from it: in particular, the notion of being well thought of and spoken of, cutting a good figure. Here a vital term is kalos, 'fine', 'noble', 'splendid', a word more strongly aesthetic than agathos, 'good', and an important term of commendation, but bearing with it implications of how one is regarded; as its opposite, aischros, 'base' or 'shameful', carries implications of being despised or shunned.

The deeds that made one admired if one was a Homeric hero were typically but not exclusively individual feats of arms, and one's arete was displayed in such. One could be shamed and lose repute not only by failing in such feats, but by being mistreated--such things led to the anger of Achilles and the suicide of Ajax. What happened to one mattered for one's esteem as well as what one did, and among things one did, competitive success ranked high: all this, of course, among those who themselves ranked high, for women and members of lower orders had other aretai and kinds of repute. In this area, there are two importantly different points, which discussion of this subject has often confused. One is that, for such a morality, shame is a predominant notion, and a leading motive the fear of disgrace, ridicule, and the loss of prestige. A different point is that excellence is displayed in competitive and self-assertive exploits. While socially and psychologically these two things often go together, they are independent of each other: in particular, the occasion of shame and disgrace may be a failure to act in some expected self-sacrificing or co-operative manner. The confusion of these two things is encouraged by measuring Greek attitudes by the standard of a Christian, and more particularly of a Protestant, outlook. That outlook associates morality simultaneously with benevolence, self-denial, and inner-directedness or guilt (shame before God or oneself). It sees the development of moral thought to this point as progress, and it tends to run together a number of different ideas which have been discarded--or at least rendered less reputable--by that progress.

The ideas of arete, shame and reputation, were of course much older notions than the self-interest conceptions of the Sophists and the simply reductive social theory that went with those. Insofar as these Sophistic speakers (and, still more, conventional persons influenced by them, such as Meno) appeal to notions of arete, and offer for impersonal admiration the ideal of a man of power, they are in fact expressing ethical conceptions which have an aristocratic structure in itself old-fashioned by the end of the fifth century; but these conceptions have been given a new, opportunistic, content, and detached from the base in traditional society which had originally made them part of a working social morality.

This structure of ideas is thus more old-fashioned than another theory presented in the Republic, the theory offered by Glaucon and Adeimantus in Book 2. This represents the conventions of justice not as a device of the strong to exploit the weak (which was Thrasymachus' formulation), but as a contractual device of the weak to protect themselves against the strong. This theory, only sketched in the Republic, is the prototype of many which view public norms as the solution to a problem which would now be expressed in the language of games-theory. It reaches, in fact, outside the most characteristic terms of Greek ethical theory, concerned as that was with arete. In two important respects, it resembles modern Utilitarian and contractual theories. First, the notion of a rule or practice is more fundamental in this theory than notions of character or personal excellence. Second, the desires which are served by the institutions of justice and, generally, the practices of morality are in the first instance self-interested desires: morality is represented as a device for promoting egoistic satisfactions which could in principle occur without it, but which are as a matter of fact unlikely to do so because of everyone's weak position in an amoral state of nature.

This instrumental or contractual view of morality was rejected by Socrates, Plato, and Aristotle. It is in many ways different from the crude Thrasymachean outlook--indeed, in expression it is its opposite. Yet for Plato it shared a basic fault with that outlook: morality was represented by both as an instrument for the satisfaction of non-moral, selfish desires which existed naturally in independence of morality. This was not just a moralizing prejudice on Plato's part, a desire for the moral motivations to appear more dignified. Still less was it the expression of an idea, later insisted on by Kant, that there can be no reason for moral conduct at all, except that it is one's duty--that the very nature of morality requires it to consist in a completely autonomous demand which cannot be rationalized or explained by anything else. The point for Plato was precisely that there had to be a reason for moral conduct, but that no theory of the instrumental kind could provide it. A theory of morality, in his view, had to answer Sophistic scepticism by showing that it was rational for each person to want to be just, whatever his circumstances. The contractual theory failed in this respect: if one were powerful and intelligent and luckily-enough placed, it would cease to be rational for one to conform to the conventional requirements of morality. This is readily admitted by Glaucon and Adeimantus in the dialogue; indeed, they basically agree with the Platonic Socrates in viewing the contractual theory, not as an answer to Sophistic scepticism about morality, but rather as a more sophisticated expression of that scepticism.

The contractual solution was particularly weak because it was unstable relative to a superior agent, one more intelligent, resourceful, and persuasive than the average. It was above all for that kind of agent that Plato thought that the sceptical demand had to be met, and the objectives of morality and justice shown to be rational. The life of Alcibiades had been scepticism in action, and the answer had to apply to a man of his superior powers. Here the first feature, too, of the contractual theory had to be rejected, the view that notions of character came second to the notions of a desirable or useful practice. The demand to show to each man that justice was rational for him meant that the answer had to be grounded first in an account of what sort of person it was rational for him to be. If anything outside the soul (as Socrates and Plato said) or outside the self (as we might put it) is what primarily has moral value--some rule, for instance, or institution--then we are left with a possible contingency, that there could be a man whose deepest needs and the state of whose soul were such that it would not be rational for him to act in accordance with that rule or institution; and so long as that contingency remains possible, the task that Socrates and Plato set themselves will not have been carried out.

It has been said by Kantian critics that Platonic morality is egoistic, in a sense incompatible with the real character of morality. This misses the point. It is formally egoistic, in the sense that it supposes that it has to show that each man has good reason to act morally, and that the good reason has to appeal to him in terms of something about himself, how and what he will be if he is a man of that sort of character. But it is not egoistic in the sense of trying to show that morality serves some set of individual satisfactions which are well defined antecedently to it. The aim was not, given already an account of the self and its satisfactions, to show how morality (luckily) fitted them; it was to give an account of the self into which morality fitted.

For Plato, as also for Aristotle, it was a trivial truth that if it is rational for one to pursue a certain course of life or to be a certain sort of person, then those things must make for a satisfactory state of oneself called eudaimonia--a term which can only be translated as 'happiness'. But not everyone now will regard it as a triviality, or even as true, that it is only rational to do what in the end makes for one's own happiness. Moreover, many people who do agree that that is true will not in fact be agreeing with the same thing as Plato and Aristotle meant. These facts are due not only to imperfections of that translation, but also to changes in views of life--changes which themselves have no doubt affected our understanding of the term 'happiness'. A proper charting of the complex relations of these words would involve a whole history of Western ethical thought. What is certain is that eudaimonia did not necessarily imply the maximization of pleasure; and when Plato, supposedly having shown in the Republic that justice is the proper state of the soul, goes on to argue that the life of the just man is also a large number of times more pleasant than that of the unjust, this is meant to be an entirely additional consideration. It is in this respect much like Kant's assurance that virtue will be rewarded in an after-life, coming as that does after his insistence that it must be regarded as its own reward (a manoeuvre which Schopenhauer disobligingly compared to slipping a tip to a head-waiter who pretends to be above such things). The state of eudaimonia should be interpreted as that of living as a man best could, and when one finds some Greek thinkers suggesting that one can attain eudaimonia although one is the victim of torture, the linguistic strain that is undoubtedly set up expresses not just a semantic difficulty, but, under that, the substantial difficulty of supposing that being tortured is compatible with living as one best could.

The Platonic aim, then, can be seen as this, to give a picture of the self such that if one properly understands what one is, one will see that a life of justice is not external to the self, but an objective which it must be rational for one to pursue. That is the sense of Socrates' question with which we started, about the way 'one should live': the 'should' is formally that of egoistic rationality, but the task is to reach the right understanding of the ego.

Both Socrates and Plato gave that account in terms of reason and knowledge. Plato saw the fullest expression of these powers in the form of systematic theoretical understanding, something which led to the consequence that the philosopher was the happiest and most fully developed of human beings; it led also to the Utopian political system of the Republic. Socrates himself certainly never developed the latter ideas (though the view, popularized by Popper,19 that Socrates himself was politically a democrat who was betrayed by the authoritarian Plato has no historical basis). The idea, however, that the real self, which is fully expressed in the life of justice, is the self of the discursive intellect, is only a development of Socratic conceptions. It may be that Socrates laid more weight than Plato on 'knowing how to act', and less on knowledge expressed in systematic theory, but certainly the notion that knowledge had to be reflective and rational was already there. An 'interest in definitions', as well as a concern with ethical questions, is what Aristotle plausibly tells us can be ascribed to the historical Socrates, and the interest in definitions with regard to ethical matters certainly took the form of trying to reach a reflective and articulate understanding of the criteria of virtuous action, which would make good practice more rationally lucid and self-critical.

If the essence of virtuous action lay in rational knowledge exercised by the soul, then there could be no separate motives represented by the various virtues, as conventionally distinguished: justice, self-control, courage, and the rest. All of them could only be expressions, in different spheres or aspects of conduct, of the same basic rational motivation. When Socrates taught the 'unity of the virtues' under rational prudence or intelligence, he did not mean that there were no ways of distinguishing one virtue from another. He meant rather that they were not basically different motivations: they were the same power of the soul, under different manifestations. Since, further, rationality must be displayed in balancing one kind of demand against another, and an exaggeration of, say, 'courageous' behaviour would not in fact be an expression of real rational understanding of what was required of one, it followed that it would not be an expression of the one underlying power of reason, and hence not of any virtue at all. So the unity of the virtues implied, as might be expected, that one does not properly display any virtue unless one displays all of them.

Virtue is the pursuit of one's interest, construed as a rational agent--the proper interest, as Socrates put it, of the soul, and this was probably already taken by Socrates in a way which implied that the interests of the soul were a separate matter from those of the body, an implication which Plato's drastically dualistic theory of soul and body was to pursue further in the direction of asceticism. Virtous action is a matter of the calculation of what truly matters most to one, and what matters most to one is what matters most for one's soul: these are the demands of the virtuous life, of courage, honour, justice. Hence if one does not act in accordance with those demands, one acts to defeat what matters most to one; no man can consciously act in such a way; so wrong action must involve a failure of knowledge and understanding, and be something which one could not possibly have chosen with open eyes. So all error is involuntary, and 'no-one willingly errs', as Socrates put it: a conclusion still discussed under the name of 'the Socratic paradox'.

The paradox raises in fact two different questions. The first is whether a person can voluntarily do one of two things, while fully and consciously holding that he has stronger reason to do the other. The second question is whether a person must, if clear-headed, admit that he always has stronger reason to do acts of justice, honour, and so forth, rather than acts of mean temporal self-interest. Most would now find it hard to give a simply Socratic answer to the second question, supported as that is by the ascetically dualistic view of the self. One difficulty that such a view inevitably raises, and which Plato himself treats uneasily, is the marked contrast between the spiritual view of one's own interests which is needed by the account of morality's motivations, and the less spiritual view of other people's interests which is needed by its subject-matter. Socrates thought that the good man cannot be harmed, for the only thing that could touch him would be something that could touch, not his body, but the good state of his soul, and that is inviolable. But--apart from other and perhaps deeper weaknesses of that picture--we must ask why, if bodily hurt is no real harm, bodily hurt is what virtue so strongly requires one not to inflict on others?

To the first of those two questions, however, the one in terms purely of conscious action and rationality, some philosophers would still give the Socratic answer. To those of us whose actions seem often very divergently related to what we take to be our reasons, that answer will still seem a remarkable paradox. It should be offered, if at all, not as a demure tautology about action and reason, but rather as conveying an ideal (a highly problematical one) of a state in which action becomes wholly transparent to the agent. That is still, itself, very much a Socratic ideal.

It is surprising how many elements in Socratic-Platonic morality are still to be found in the complex and very interesting ethical theory of Aristotle, different though it is in certain central respects. It is different, most importantly of all, because not all the weight is put on intellectual excellence and pure rationality. Aristotle distinguishes between 'intellectual excellences' and 'excellences of character', and emphasizes the importance to the latter of the correct formation of desire and motivation through training. Without correct upbringing nothing can be done: the hopes for the regenerative powers of philosophy itself which are implicit in the Socratic stance have gone, as has the sense of any combative scepticism against which morality has to be defended. A more settled order is in question. Aristotle, moreover, did not believe in a soul wholly separate from the body, and that denial goes with a rejection of Socratic asceticism, and with more worldly possibilities for eudaimonia. The old link of arete and public approval, which Plato sought to cut altogether, cautiously reasserts itself in the Aristotelian account, though his theory of the motivation of the virtues is much more sophisticated than anything that had been achieved at an earlier time, or indeed by Plato himself.

Yet, granted these differences,Aristotle still ends by regarding the life of theoretical reason as the highest form of human life, a conclusion which does not follow as directly, or even as coherently, from his pre-misses as it did from Plato's. He preserves also something like the Socratic paradox about action and reason. He even preserves, in effect, the Socratic conclusion about the unity of the virtues, since he thinks that one cannot genuinely have any one excellence of character without the presence of phronesis, 'practical reason' (itself one of the intellectual excellences), but if one has phronesis, then one must have all excellences of character. This emphasis on the rational integration of character, as also on the integration of a good life over time, its retrospective rational shapeliness, is indeed a central feature of Aristotle's outlook. In the matter of the ultimate unity of virtuous traits of character, certainly this is one issue on which the Greek view seems far from ours: nothing is more commonplace to us than that particular virtues not only coexist with, but carry with them, typical faults. But this is one of the many differences with the Greeks where the contrast itself points to an illuminating area of discussion: what divergences in the understanding of human nature underlie these different conceptions of a rationally desirable life.

It is worth bringing together several features of Greek ethical thought which mark it off in many ways from current concerns and from the moral inheritance of the Christian world. It has, and needs, no God: though references to God or gods occur in these writers, they play no important role. It takes as central and primary questions of character, and of how moral considerations are grounded in human nature: it asks what life it is rational for the individual to live. It makes no use of a blank categorical moral imperative. In fact--though we have used the word 'moral' quite often for the sake of convenience--this system of ideas basically lacks the concept of morality altogether, in the sense of a class of reasons or demands which are vitally different from other kinds of reason or demand. The sharp line that Kantianism, in particular, draws between the 'moral' and the 'non-moral' is very partially paralleled by another sharp line, Plato's line between soul and body; but the parallelism is far from total, the distinctions are drawn on quite different principles, and the discussion of the merits and failings of each will be a quite different sort of discussion. Relatedly, there is not a rift between a world of public 'moral rules' and one of private personal ideals: the questions of how one's relations to others are to be regulated, both in the context of society at large and more privately, are not detached from questions about the kind of life it is worth living, and of what is worth having or caring for.

In all these respects the ethical thought of the Greeks was not only different from most modern thought, particularly modern thought influenced by Christianity, but was also in much better shape. There are of course respects in which its outlook could not be recaptured now, and some in which we could not want to recapture it. Some of its thoughts express a certain integration of life which perhaps existed for a short while in the city-state, but which, as Hegel emphasized, would have to be recovered, if at all, only in some totally changed form. Other features of its perceptions, its substantive attitudes to slavery, for instance, and to the role of women, we must hope will never be recovered at all.

At a more theoretical level, it is important that Greek ethical thought rested on an objective teleology of human nature, believing that there were facts about man and his place in the world which determined, in a way discoverable to reason, that he was meant to lead a co-operative and ordered life. Some version of this belief has been held by most ethical outlooks subsequently; we are perhaps more conscious now of having to do without it than anyone has been since some fifth-century Sophists first doubted it. But when all that has been said, it is true that Greek ethical thought, in many of its basic structures and, above all, in its inability to separate questions of how one should relate to others and to society from questions of what life it is worth one's leading and of what one basically wants, represents one of the very few sets of ideas which can help now to put moral thought into honest touch with reality.

In these last remarks I have mentioned 'Greek ethical thought', and that principally refers, of course, to the philosophical ideas of Socrates, Plato, and Aristotle which, very sketchily, I have discussed. But there is a question which I should like to raise in closing, which reaches behind them, and behind some other aspects of Greek philosophy which have been touched on in this chapter.

I have mentioned already Socrates' saying, that the good man cannot be harmed: it expresses an ideal of rational self-sufficiency, of freedom from the damage of contingency. There is an analogy, not merely superficial, between this type of assertion of rationality, and that cognitive demand for the elimination of luck, which appeared in the discussion of knowledge, certainty, and scepticism. The ideal of self-control, always high among Greek aspirations, turned into the aim that, in both cognition and action, what is of highest value, what matters most, should be entirely under the self's control. In later schools, this theme reappeared in various forms: in the Cynic exaggeration of Socratism, that virtue was sufficient for eudaimonia and that the good man really could be happy on the rack; in that hope for a state of ataraxia which the Sceptics were not alone in cultivating. Aristotle expressly discussed the question of how far eudaimonia, the ultimately desirable state, could be subject to risk, and replied that to a small but ineliminable degree it had to be. This represented, however, not so much any large or perilous aspiration, as rather the entirely sensible thought that it is unreasonable to leave out of account the apparatus of social life within which men live and express themselves, and which is subject to fortune. Very notably, a dimension of life which to us is one of the most significant precisely because of its reaching outside the defended self, friendship, is discussed by Aristotle in a way which now seems bizarre in its determination to reconcile the need for friendship with the aim of self-sufficiency.

A deeper sense of exposure to fortune is expressed elsewhere in Greek literature, above all in tragedy. There the repeated references to the insecurity of happiness get their force from the fact that the characters are displayed as having responsibilities, or pride, or obsessions, or needs, on a scale which lays them open to disaster in corresponding measure, and that they encounter those disasters in full consciousness. A sense of such significances, that what is great is fragile and that what is necessary may be destructive, which is present in the literature of the fifth century and earlier, has disappeared from the ethics of the philosophers, and perhaps altogether from their minds. Nietzsche found Socrates to blame for this, with his excessive distrust of what cannot be discursively explained, his faith in the 'fathomability' of nature, and his 'Alexandrian cheerfulness'.20 Those remarks belong, in fact, to the first period of Nietzsche's long and ambivalent relations to the figure of Socrates, and it was a period in which Nietzsche thought that the 'metaphysical solace' of tragedy could be understood only through a fundamentally aesthetic attitude to life, an attitude which we have even greater reason to reject than Nietzsche eventually had. But however much he or we may qualify his account of Greek tragedy and Greek thought, what he pointed to is truly there: Greek philosophy, in its sustained pursuit of rational self-sufficiency, does turn its back on kinds of human experience and human necessity of which Greek literature itself offers the purest, if not the richest, expression.

If there are features of the ethical experience of the Greek world which can not only make sense to us now, but make better sense than many things we find nearer to hand, they are not all to be found in its philosophy. Granted the range, the power, the imagination and inventiveness of the Greek foundation of Western philosophy, it is yet more striking that we can take seriously, as we should, Nietzsche's remark: 'Among the greatest characteristics of the Hellenes is their inability to turn the best into reflection.'21

Return to Book Description

File created: 8/7/2007

Questions and comments to:
Princeton University Press

New Book E-mails
New In Print
PUP Blog
Princeton APPS
Sample Chapters
Princeton Legacy Library
Exam/Desk Copy
Recent Awards
Princeton Shorts
Freshman Reading
PUP Europe
About Us
Contact Us
PUP Home

Bookmark and Share