THE YALE REVIEW (1959)

THE SUBLIME AND THE BEAUTIFUL REVISITED

By IRIS MURDOCH

What I have to say does not fall into the domain of literary criticism, as it is understood nowadays. My subject lies on the borders of literature and philosophy, but it is important to insist that I am not a critic. My remarks will he at times more personal and throughout more abstract than those of a critic; and I would like to say at the start that although what follows may sound like a manifesto and may imply a dogmatic tone of voice, I am not all that sure that what I say is right.

I want to connect a literary problem with a more general political and moral problem. Roughly, I think one may be enlightened by connecting the question: Is the Liberal-democratic theory of personality an adequate one? with the question: What is characteristic of the greatest literary works of art? The latter question could also he put in the form: What, chiefly, makes Tolstoy the greatest of novelists and Shakespeare the greatest of writers? I shall be concerned here mainly with the novel, and I approach the problem as a novelist concerned with the creation of character. Wherein does the reality of a person reside and in what way can one, or should one, display that reality? More precisely, I want to treat recent changes in the portrayal of character in novels as symptoms of some more general change of consciousness. It has for some time now been the fashion to say that we are in a morass, and to attempt to get out of the morass by attacking Romanticism; and I am going to do this too. The word "Romantic" is best defined by what it is opposed to, and I shall he opposing it to different things from some to which it is usually op-posed. But I hope it will become tolerably clear what I mean by it.

I shall pursue my line of thought first in the region of philosophy, and come later, having erected a sort of philosophical structure, to discuss literature; and I shall start by considering certain ideas of Kant. Kant is the father of all modern forms of the problem of freedom,
248
and also incidentally the father of most modern theories of art. To him, even more than to Rousseau or to Hegel, we may impute both the initial strength and the later weakness of the Liberal theory of personality, which is to such an extent also the Romantic theory of personality. I shall argue later that it is desirable to purge the Liberal theory of many of its Romantic elements. I am using the word "Liberal" here of course in its traditional historical sense, the sense in which the philosophy of John Stuart Mill is a Liberal philosophy.

Kant's moral philosophy rests on the equation that virtue is freedom is reason. Virtue is not a knowledge of anything; it is rather an ability to impose rational order. We respect others, not as particular eccentric phenomenal individuals, but as co-equal bearers of universal reason; though it is true (and this is Kant's agnosticism) that we do not know reason in ourselves or others, in the same way that we know material objects. Not being purely rational we are not transparent to ourselves. The entire world of our emotions and desires is irrelevant to morality. We turn from the messy and ambiguous region of the emotions to the undoubted clarity of the choice and the act. The only emotion in which Kant interests himself, and which he connects with morality, is the mixed pleasure and pain of Achtung, that is, respect for the moral law, which is an experience of our freedom-our freedom, as it were, biting into our phenomenal being. This great concept is of course the immediate ancestor of the popular existentialist concept of Angst.

Kant's theory of art accords with his theory of morals; but it has one or two special features. Kant makes a distinction, a distinction which has not~in my view fascinated philosophers as much as it should, between the sublime and the beautiful. Strictly, Kant's theory of art as such is concerned only with the beautiful, but the distinction is suggestive, as I shall argue late4 of a total theory of art which uses his concept of the sublime as well. Kant's theory of the beautiful is explanatory of a great deal of Romantic theory and practice. The beautiful is the experience of a conceptless harmony between the imagination and the understanding. Art, as the production of the beautiful, is not a matter of discovering or imparting truths ("what oft was thought but ne'er so well expressed") -it is rather the production of a certain kind of quasi-thing. It is noteworthy that the
249
work of art is conceived by Kant, and mostly by the Romantics, on the analogy of a fairly small perceptual object. Kant of course construed the judgment of taste precisely on the analogy of the perceptual judgment. What is constructed is a self-contained object, strictly purposeless, yet with an air of purpose, existing for its own sake. In art, we enjoy an immediate intuitive inexplicable understanding of a unique quasi-sensible object It is remarkable how friendly and familiar this really very obscure theory seems: we take to it like ducks to water.

In constructing the art-object freedom is not involved, since reason is not in play; but Kant says that the beautiful is an analogon of the good, the enjoyment of art is an analogon of the free rational act, in that it is the construction of something clean, free, empty, self-contained, not contaminated by the messiness of emotion, desire, or personal eccentricity. Art is hygienic; but as it is not an activity of the reason, it is a sort of play and is connected with morals only by analogy.

Now the sublime is a very different matter. As described by Kant the sublime is not connected with art at all. While the beautiful is an experience of the imagination and the understanding in harmony, the sublime is an experience of the imagination and the reason in conflict. Whereas the beautiful reposes us, the sublime rends us. It is an emotional experience resulting from the defeated yet invigorating attempt of reason to compass the boundlessness and formlessness of nature. Confronted with some vast prospect, the starry sky, or the Alps, the imagination and the senses cannot properly take in what lies before them, that is they cannot satisfy the reason, which demands a total complete ordered picture. Yet in being so defeated the reason gains a fresh sense of its own independence and dignity. Since reason is the moral will, the experience of the sublime is a sort of moral experience, that is, an experience of freedom. It is a mark of spiritual power resulting from a serious and amazed survey of the vastness of nature. But since it is not connected with action it is not strictly moral activity. It is the moral will not in action but as it were triumphantly intuiting itself. It is, as Kant puts it, a reminder of our supersensible destiny: an experience analogous to Achtung, a mixture of defeat and victory.

With the theory of the sublime we have the distressing feeling of
250
some vast and wonderful idea being attached to a trivial occasion. Who, one might say, cares what sort of emotions Kant experienced in the Alps? There must be more to it: and I shall suggest later that Kant's view is pregnant with a concept of the tragic, and with a theory of the connection between literature and morality. With the Hegelian descendants of Kant we are in the main stream of Romanticism. For Kant, neither morality nor art was a matter of knowledge, whereas Hegel pictured reality in terms of a developing range of historical and psychological concepts and implied that complete knowledge of it was possible. Reason was not ultimately defeated; it could close the circle of knowledge: no agnosticism here, no sense of limit. Hegel did picture virtue in terms of knowledge in the sense that progress was measured by increasing self-awareness. Virtue was also freedom in the sense of recognition of necessary process, that is freedom as self-knowledge. For Kant virtue is freedom is rational order; for Hegel virtue is freedom is self-knowledge.

In this universe, art is a stage of self-awareness which we ultimately pass beyond. Tragedy too is a product of appearance not of reality. It is the mutual misunderstanding of parts of the whole. From the point of view of the whole itself there is no tragedy. Meanwhile, however, there is the conflict: the self locked in struggle with itself and evolving as a result of the struggle. There is only one being in the Hegelian universe, the whole which cannot allow anything outside itself and which struggles to realize all that is apparently other. This is Hegel's gift to the Romantic Movement, and one from whose effects we have not yet recovered.

The opponents of Hegel may be divided into those who understand him very well and feel a mixture of love and hate as a result, and those who have never read him and would not understand him if they did. We may single out from the former the Kierkegaardian or existentialist Hegelians, who are pure Romantics, and from the latter the Hobbesian empiricists, who are Liberals touched by Romanticism at a certain stage. The Liberal dilemma may be seen as the failure of these two disparate elements to help each other to produce a new post-Hegelian theory of personality.

Kierkegaard, as we know, fought against the swallowing up of the individual human person in the Hegelian system. He fought for
251
the conception of a private individual destiny: the root idea of existentialism, that the individual human existence is not enclosed by a world of essences. However, Kierkegaard is profoundly Hegelian. He retained and used with wonderful versatility the clear, dramatic, solipsistic picture of the self at war with itself and passing in this way through phases in the direction of self-knowledge. Kierkegaard, like the modern existentialists, is anti-system, and even, like them, anti-bourgeois; yet psychologically speaking he may be described as "totalitarian," in the sense that he is concerned with the whole man and with his isolated struggle for salvation. The individual described by Kierkegaard is alone, except for the mystery of religion. Kierkegaard's agnosticism (like that of Kant, as we see when we look at that a second time) is dramatic rather th~n resigned. We do not know all-but "all" is magnetically there to be known. Our solitude is completed, the circle is closed, by our relation to a veiled deity. The deity and the solitary self between them enclose the whole of reality.

When we turn from the existentialist Hegelians to the Hobbesian empiricists we breathe a very different air. Whereas Kantian aesthetics is familiar to us as cultured persons and consumers of art, the universe of discourse of Hobbes and Locke is familiar to us as political beings and more generally as ordinary users of common sense. Here are the familiar ideas of our Liberalism, ideas we take for granted, and whose undisputed simplicity makes the Liberalism of the English-speaking world different from other varieties. In the world as envisaged by Hobbes, Locke, and Hume there is a plurality of persons, who are quite separate and different individuals and who have to get along together. Moreover, implicitly for Hobbes and explicitly for Locke, that which has a right to exist, that which is deserving of tolerance and respect, is not the rational or good person, but the actual empirically existing person whatever he happens to be like. The agnosticism out of which this tolerance sprang was the undramatic commonsensical agnosticism of Locke rather than the dramatic agnosticism of Kant and Kierkegaard. It just is in fact rather difficult to understand other people and to be certain what is the right thing to do: one is fallible, so one must be patient. This agnostic tolerance was developed most explicitly by Mill, who was indeed touched but by
252
no means overwhelmed by the Romantic Movement Mill retains an eighteenth-century sense of society. His individual is eccentric, unique, holy, pregnant with genius, but not alone.

This is the tradition out of which a criticism of Romanticism should have come, but did not. It was partly that Romanticism was from the start absolutely entwined with the Liberal tradition and must have seemed to be its lifeblood. But there were also features in that tradition itself, thinking of it as something with older roots and a separate being, which made it ultimately unable to find as it were a footing, a vantage point from which to criticize, even to see the Romantic Movement. The empiricists had, it is true, a certain commonsensical picture of society. They recognized a range of virtues corresponding to the range of human impulses and social situations; whereas the Hegelians ultimately recognized only one virtue, progress in self-mastery. But the empiricist picture of society was undynamic and naive; and this was true even of Mill. The real impetus of the philosophical movement with which Liberalism was connected was not primarily moral or political, it was scientific; and we may say that the simple and powerful Liberal ideas to which we are so attached occurred in a way as a by-product and as a result of what we may call the phenomenal 1uck of our English-speaking societies. What most of all concerned the empiricists, and what drove their theories onward, was the construction of a picture of the material world, the development of that atomic empiricism which has had so strong a hold on our philosophical imagination, and which is only now vanishing from the scene, partly as a result of the development of science as a separate discipline, and partly as a result of the criticism by Wittgenstein of assumptions held in common by Hume and by Bertrand Russell; and with this parceling out of our intellectual tradition which has been happening of late, we are become uneasily aware of its extreme simplicity and poverty in certain respects. It was indeed the Romantic Movement itself which gave it an appearance of color and body.

Now before turning about, and attempting to use a consideration of literature in a more telling diagnosis of these ills, it is still necessary to say something about the most recent trends in philosophy. Here again I must be, on a very complex subject, very brief. For my pur-
253
poses, there are two important philosophies, existentialism (here I take the work of Sartre as typical) and linguistic empiricism (the tradition of Moore and Wittgenstein). Both of these, I shall argue, show certain symptoms which may be connected with symptoms of decline in our literature.

Existentialism and empiricism (as I shall call it for short) share a number of motives and doctrines. Both philosophies are against traditional metaphysics, attack substantial theories of the mind, have a touch of puritanism, construe virtue in terms of will rather than in terms of knowledge, emphasize choice, are markedly Liberal in their political bias, are neo-Kantian. But in other ways they are very different.

The inspiration of empiricism is a scientific inspiration which expresses itself in an extreme desire for precision of meaning. Such a desire cannot but be to some extent inimical to words; and one cannot be surprised at the increasing importance and popularity of mathematical logic, which is now regarded in many quarters as the basis of philosophy. On the other hand, empirical concepts have their being in empirical languages, and in their clarification appeal is made to the conception of "ordinary language," thought of as a really existing network, within which meanings are established for general inspection, so that perfect clarity and conclusiveness can be attained by the making of logical points, i.e. points depending on conventionally accepted definitions. This technique, which produces satisfactory results when used to clarify our concepts of the material world, is less satisfactory when used to clarify moral concepts.

Linguistic moral philosophy, of which one of the most influential exponents is the American philosopher Stevenson, operates by seeking the meaning of moral concepts in the moment of moral choice, through studying the role which words such as "good" play in choice situations. What is given by this study is, with variations of detail, as follows: Moral situations are those in which we give generally comprehensible reasons for choices of a certain degree of importance. Our choices, together with our reasons, display our values, and the moral concept (e.g. the word "good") is the instrument of commendation whereby we point out what is to be chosen. Note that this picture is conventional, behavioristic, and Liberal in tendency. It is
254
conventional: the agent is seen as a being subject to rules, surrounded by a civilized society, surrounded in short by the network of ordinary 1anguage, that is, for these purposes, by the network of moral conceptual activity at its most common and universally accepted level. It is behavioristic; and here an anti-substantialist theory of the mind joins hands with a certain Liberal puritan austerity. Since inner acts of the mind only have identity through their conventional connections with outer acts, we may say that morally speaking a man is what he observably does. As in the philosophy of Kant, we turn away from the chaos of empirical inwardness to the clarity of overt action. What is a man "feels" is of no interest to us, and even what he believes is of no interest except in so far as his beliefs are defined by his actions. The picture is Liberal (neo-Kantian) in its concern with reasoned choice. Ordinary Language Man, as we may call the person here envisaged, is not overwhelmed by any structure larger than himself, such as might be represented by a metaphysical belief or by an institution; As a moral agent he is completely free, choosing between acts and reasons on his own responsibility; and it has been a major preoccupation of empiricist moral philosophy to depict the agent as totally free and self-sufficient. Even the presence of others is felt, if at all, simply as the presence of rational critics. This man is alone with a loneliness similar to the loneliness of Kant's man.

In the philosophy of Sartre we find the same solitary moral agent, and the same emphasis on the moment of choice, but displayed in terms of a dramatic Hegelian psychology. One might say that whereas Ordinary Language Man represents the surrender to convention, the Totalitarian Man of Sartre represents the surrender to neurosis: convention and neurosis, the two enemies of understanding, one might say the enemies of love; and how difficult it is in the modern world to escape from one without invoking the help of the other. Sartre's man is like a neurotic who seeks to cure himself by unfolding a myth about himself. Ordinary Language Man is at least surrounded by something which is not of his own creation, viz. ordinary language. But Totalitarian Man is entirely alone. How well we know this man from the pages of modern literature. He suffers from Angst, which is Achtung minus confidence in universal reason, that is, with its dignified and exultant aspect removed. He makes his choices against
255
the apocalyptic background of the modern world-an apocalyptic world picture favors a total creed-and if he is sincere he knows that he is always in an extreme situation. He is stripped to essentials. Sartre says, speaking of existentalist literary works: "It is always the whole man that is in question."

This man on the one hand mistrusts his inner life and finds it insubstantial; to attribute substance to it is to fall into insincerity. On the other hand, he dramatizes his situation in a myth. He is Hegel's man, who is a clear-cut piece of drama rather than an individual. He is also Hegel's man who abhors the contingent or accidental. (La Nausee, horror of the contingent.) According to Sartre, a desire for our lives to have the form and clarity of something necessary, and not accidental, is a fundamental human urge. In the world inhabited by Totalitarian Man there are other people, but they are not real contingent separate other people. They appear as organized menacing extensions of the consciousness of the subject. A potentially or apparently separate center of significance is necessarily a menace to a Hegelian, something to be internalized in a battle of consciousnesses just as discrepant centers in oneself are overcome by reflection.

Virtue, for total man, is sincerity, courage, will: the unillusioned exercise of complete freedom. Virtue is not knowledge, since to pretend to any stable knowledge relevant to morality would be a case of bad faith; just as the ordinary bourgeois social virtues, which take society for granted, are cases of bad faith. Existentialism shares with empiricism a terror of anything which encloses the agent or threatens his supremacy as a center of significance. In this sense both philosophies tend toward solipsism. Neither pictures virtue as concerned with anything real outside ourselves. Neither provides us with a standpoint for considering real human beings in their variety, and neither presents us wjth any technique for exploring and controlling our own spiritual energy.' Ordinary Language Man is too abstract, too conventional: he incarnates the commonest and vaguest network of conventional moral thought; and Totalitarian Man is too concrete, too neurotic: he is simply the center of an extreme decision, man stripped and made anonymous by extremity.

I take these two philosophies, linguistic empiricism and Sartrian existentialism, as representative of the wisdom which philosophy has
 256
now to offer to the Liberal tradition. I consider them here as symptoms rather than positively as influences, though I think existentialism has in fact been influential well outside academic circles. ~ The philosopher often clarifies and crystallizes something which exists in a less coherent form in the general consciousness: that is, I take the general consciousness today to be ridden either by convention or by neurosis; and there are many features in both these current philosophies in which we can recognize ourselves: There are of course critics, particularly critics within the existentialist tradition, of many of the points I have mentioned. One might think of Gabriel Marcel, and of Simone Weil; but such critics have remained isolated since they do not at all represent what we take ourselves to be like. Nor, and I speak briefly here of something that would need to be discussed at greater length, does it seem to me that the Christian religion has been able to present us in recent times with any satisfying or powerful picture of ourselves and each other. What we take ourselves to be like is, I think, successfully portrayed by Ordinary Language Man on the one hand, and Totalitarian Man on the other. And I shall argue presently that this regrettable situation is to be intimately connected, both as cause and as effect, with the decline of our prose literature.

Here I conclude my philosophical introduction and turn to consider literature, ho~ing ito use certain philosophical conceptions in the diagn9sis of certain literary ills. Let me say here again that I am not a critic. I am doing what philosophers do, that is putting up an abstract structure to edify, explain, and provoke reflection. In this case mostly the last. And I shall not be unduly dismayed if it is pointed out that this or that particular work does not fit conveniently into my structure. I shall be content if something is clarified, even if something is discussed.
If we take it that a dominant philosophy pictures the consciousness of the age, and if we take it that the dominant philosophy of the nineteenth century, outside England and America, was the philosophy of Hegel, and if with this in mind we turn to look at the nineteenth century novel, we get an agreeable surprise. There are of course plenty of reasons (obvious ones in our civilization) why the nineteenth century novel, although it shared with Hegelianism a historical sense,
257
certainly a social sense, is, in the respect that interests me, so un-Hegelian. All the same it is remarkable, and in ways entirely relevant to its characteristic and preeminent merits, how very un-Romantic the great nineteenth-century novel is. I am here using the word "Romantic" in my own somewhat narrow sense, which will I hope very soon be made clearer. The feature that most interests me in the un-Hegelian nature of those great novels is simply this: that they contain a number of different people.

I do not want to pause for long at this point, since the thing which I am indicating is well known to you and obvious. There is in these novels a plurality of real persons more or less naturalistically presented in a large social scene, and representing mutually independent centers of significance which are those of real individuals. What we have here may be called a display of tolerance. A great novelist is essentially tolerant, that is, displays a real apprehension of persons other than the author as having a right to exist and to have a separate mode of being which is important and interesting to themselves. We may decide later that "tolerance" is too mild a word for this capacity at its highest. But "tolerance" is a word which links nineteenth-century literature with Liberalism. Here one may see the Liberal spirit at its best and richest, disporting itself in literature, and not yet menaced by those elements of romanticism which later proved, if I am right, so dangerous. The great novels are victims neither of convention nor of neurosis. The social scene is a life-giving framework and not a set of dead conventions or stereotyped settings inhabited by stock characters. And the individuals portrayed in the novels are free, independent of their author, and not merely puppets in the exteriorization of some closely locked psychological conflict of his own.

The literary work itself is not in the grip of necessity-how soon we sense this in the cases where it is. The great novelist is not afraid of the contingent; yet his acceptance of the contingent does not land him in banality. In respect of this quality, and of others, the writer with whom we are most tempted to compare this novelist is Shakespeare.

The persons whom I have here in mind, and whom I have called the great novelists, are of course Scott, Jane Austen, George Eliot, Tolstoy, especially Tolstoy-one could add other names, but these
258
suffice to make the point. I realize it is paradoxical to call, for instance, Scott an un-Romantic writer; but I do not mind the paradox so long as the meaning is clear. In calling these novelists the great ones I do not exclude other types of greatness-though it is part of <my thesis that this is probably the greatest sort of greatness. It is true that we find in the nineteenth century other remarkable novelists (Dostoevsky, Melville, Emily Bronte, Hawthorne) to whom we would not want to deny a first place, and to whom the title "Romantic," in my sense, could more readily be applied: writers who give the impression of externalizing a personal conflict in a tightly conceived self-contained myth; and it would be perverse to argue that they are great in spite of their Romanticism. This is not the place to analyze their merits. But in attempting to elucidate a standard based on the work of what I have called the un-Romantic writers, one may at least provide oneself with a tool to explore that contrast further. Many reasons might be given for the particular qualities of the nineteenth-century novel: reasons which might connect it with particular, now-vanished historical and s6cial conditions. I am not concerned with these either. What I want to discover and assert is a value which I think belongs, or has belonged since at least the eighteenth century, to prose literature as such, and which does not cease to be a value ~hen it becomes more difficult of attainment. We should always beware of doctrines of necessity which show us (with professions of regret) the eminently desirable, the good, as being, alas, the impossible.

I want now, after this brief glance at the nineteenth-century novel, to look at what may be called Romanticism in decline, and to look in particular at a certain literary movement which shows us the Romantic view of art no longer controlled by the forces which had contributed to produce the great novels. I mean the movement connected originally with Symbolism, and represented in our tradition in varying ways by such persons as T. S. Eliot, T. E. Hulme, L A. Richards, and others. I shall take this movement as a clear and selfconscious symptom of a wider general trend; and I think the influence of this movement on modern literature has in fact been considerable. This is a matter more usually discussed in relation to poetry;
259
but I think we may see the power of Symbolist ideas in the development of prose literature as welL

The Symbolists, as I shall call this heterogeneous group for short, professed of course to be opponents of Romanticism. A position common to Eliot and Hulme, and which they took to be an anti-Romantic position, could be summed up in Hulme's statement that perfection since the Renaissance had been erroneously conceived in human terms. This had resulted in an art and a philosophy which was vague, emotional, formless, and messy. The Symbolists were certainly anti-humanist; and in common with the linguistic empiricists and with the Sartrian existentialists what they most abhorred was messiness. They shared especially with the empiricists an extreme desire for precision and clear definition. "Everything is what it is and not another thing," Butler's remark which G. E. Moore prefixed to Principia Ethica could equally have been a Symbolist motto; and indeed Hulme hailed Moore as an ally.

But the Symbolist desire to be precise and to escape from the messy took a special form. As Hulme said, "We must find beauty in small dry things." The Symbolists were fascinated by the senses, particularly the sense of sight. What they wanted were small, clean, resonant, and self-contained things of which the image or symbol was the type. What is beautiful must be separate, conceived on the analogy of a sensible object. Indeed to create a sensible object-"to hand over sensations bodily"-but one purged of its contingency, was the Symbolist ideal. What they wanted, one might say, was to have the sensible world, but with the help of art to know it intuitively and not discursively. Prose literature was the form of art which lay furthest from their concerns; and, like the empiricists, and for roughly the same reasons, they were uneasy about the discursive nature of language. Language was a pis aller. The ideal of a work of art was "not to mean, but to be." Art, including literature, should be the creation of unique self-contained things.

Now it is plain that this view is none other than Kant's theory of the beautiful, served up in a fresh form, and as such it is something which had been inside the Romantic Movement from the start. The motive here, as in the case of Kant, is a fear of contingency,
260
a yearning to pierce through the messy phenomenal world to some perfect and necessary form and order. An adoration of necessity, more or less concealed, has always been a characteristic of Romanticism, coexisting in the earlier days with the wilder untidy life-loving more purely Rousseausque elements, but in later times proving itself more powerful. What is feared is history, real beings, and real change, whatever is contingent, messy, boundless, infinitely particular, and endlessly still to be explained; what is desired is the timeless non-discursive whole which has its significance completely contained in itself.  One might say of the Symbol that it is an analogon of an individual, but not a real individual. It has the uniqueness and separateness of an individual, but whereas the real individual is bound-less and not totally definable, the symbol is known intuitively to be self-contained: it is a making sensible of the idea of individuality linder the form of necessity, its contingency purged away. Plato mistrusted art because it imitated what was various and unreal; the symbolists desired an art which would have satisfied Plato.

It is not at all surprising, when we consider the matter further, to discover that the Symbolist trend is intolerant and anti-Liberal. Its fear of contingency and history is a fear of the real existing messy modern world, full of real existing messy modern persons, with individual messy modern opinions of their own. To this Hulme and Eliot would oppose institutions and dogma, the presumed clarity and cleanliness of the 'medieval world when, to use Eliot's phrase apropos of Dante, "thought was orderly and strong and beautiful." It is odd and sobering that we may discover this extreme horror of the details of modern life expressed not only by Eliot but also by Tolstoy. In What is Art? Tolstoy asks: "Strip the best novels of our time of their details and what will remain?" And to the hopelessly contingent detail-ridden novel Tolstoy opposes the pure simplicity of the parable or the folk-tale: the symbol, one might say, of a religious perception. It is as well that Tolstoy did not practice what he preached until later in life. And Eliot, in a mood not un like that of Tolstoy, says that he would prefer an illiterate audience to one that is but half-educated.

Eliot of course quite explicitly connects his moral criticism of what he calls the Liberal view of personality with his aesthetic criticism of post-seventeenth-century literature which suffers from the "dis-
261
sociation of sensibility." Romantic poetry lacks hardness, is not able directly to present us with things, because Romantic Liberal persons are concerned with the emotional expression of their own personalities, and not with the being and authority of the thing that confronts them. This Romantic cult of personality and denial of external\ authority Eliot traces back through humanism, the Romantic movement, the puritans, to the work of Shakespeare, and it is at this point one may attempt to define one's dislike of his attitude.

With certain aspects of Eliot's criticism of the Liberal view of personality I am in sympathy, as I shall explain shortly. What I like about his criticism is that he accuses the Liberal view of a failure to emphasize the discipline involved in realizing that something real exists other than oneself. What I dislike about his criticism is his view of what it is that we are thus to discover outside ourselves, what we are to oppose to the individual. That which should be respected by the individual is according to Eliot, in literature, that is in poetry, "the thing," and life, in morality, "the institution," or "the dogma." In neither case is it another person whom we should thus treat as separate and real. This is objectionable both because Mr. Eliot's faith in institutions seems to me excessive: I cannot agree with his view that "it is better to worship a Golden Calf than to worship nothing"; and because Mr. Eliot seems so ready to throw overboard what, with the theoretical assistance of Hobbes, Locke, Mill, and others we have established in our tradition, a respect for the individual person as such, however eccentric, private, messy, and generally tiresome he may be. And one aspect of respecting something is being interested enough in it to try to understand it.

This is where literature, especially prose literature, comes in or should come in. I have spoken of Shakespeare as being the greatest exponent of what I called, giving it too humble a name, that tolerance which we find also in the great novelists. The pages of Shakespeare abound in free and eccentric personalities whose reality Shakespeare has apprehended and displayed as something quite separate from himself. He is the most invisible of writers, and in my sense of the word the most un-Romantic of writers. Now it is impossible to give close attention to the novel, either as writer or as reader, without facing the problem of freedom. Let us see how Mr. Eliot faces it in this
262
connection. I am unable to be confident from Mr. Eliot's writings that he has ever enjoyed and admired any novel. He does however in After Strange Gods make some interesting comments on the nineteenth-century novel. He gives qualified approval to Jane Austen, Dickens, and Thackeray, because with them "personality . . . was more nearly in its proper place. The standards by which they criticized their world, if not very lofty ones, were at least not of their own making." But with George Eliot the modern decline begins: she exhibits "the dreary rationalism of the epoch," and "is of the same tri~ as all the serious and eccentric moralists we have had since." It is not surprising that given a choice between neurosis and convention Mr. Eliot should prefer convention. Some of the reasons for which he does so are ones which may be respected, as I shall soon argue.

But it is significant, and from my point of view unpardonable, that he should cast his vote against George Eliot. For she, at a level at times almost equal to that of Tolstoy, displays that godlike capacity for so respecting and loving her characters as to make them exist as free and separate beings. It is this freedom which Mr. Eliot resents, and which he sees simply in terms of a messy Romantic self-assertion. But this is to confound the categories. The discipline involved in creating characters such as those of George Eliot is the reverse of an uninhibited self-assertion: What Mr. Eliot really dislikes here in the modern world, as shown in the independent and unconventional stand-point of some of the characters and of their author. Jane Austen, who exhibits a similar discipline in the creation of separate persons, does not incur his censure because she places her free people within an old conventional world. Mr. Eliot fails to distinguish between two senses of "free": "free" as meaning "independent of the author" and "free" as meaning "independent-minded." Yet in fact if we consider his whole position he is really committed to disliking freedom in both these senses, and it is only because he is not interested enough in the novel and does not take it seriously enough, that he fails to see that Jane Austen is just as great a menace, from his point of view, as George Eliot. That is, to put my point in another way, one cannot altogether separate these two senses of "free." A society which can produce great novelists and which can appreciate great novelists
263
is a society in which tolerance and respect for the existence of other persons is likely to flourish, with all that that implies of independence of mind. So Mr. Eliot is right too from his point of view, to be afraid of Shakespeare and to see in this loving toleration of, indeed delight in, manifold different modes of being, a beginning of the modern world.

I have suggested that Mr. Eliot does not really like novels. In fact I think that he does not really like prose except when it is used for didactic purposes. This is very important. Eliot remarks, apropos of his own work, "In one's prose reflection one may be legitimately occupied with ideals, whereas in the writing of verse one can only deal with actuality." And I suspect that he believes in general that prose is not well able to deal, in that sense, with actuality. Such a belief is of course entirely in the Symbolist manner. The Symbolist ideal of significance, the ideal of the resonant self-contained work of art which made itself as like as possible to a thing, made difficulties for the ordinary discursive features of language. Mallarm6, for instance, went to extremes to overcome these difficulties. Language was, as it were, too "spread out" for the Symbolists. It was only of value to them if they could tie it into a knot. It was "not to mean, but to be." Clearly from this point of view, poetry is of far more use than pros~ in fact, prose might just as well be written off, banished altogether from the realm of art, and left to do the jobs of explanation and instruction which it is so well fitted to perform.

It is significant that this view is quite explicitly stated by Sartre in his book What is Literature, where he distinguishes between "the word which is lived and the word which is met." Poetry is the word which is met, which lies outside us, separate, thing-like, and complete. Prose is the word which is lived, the language which we inhabit and which we must treat as a tool, and use for making disclosures. That is, poetry is conceived in terms of the Kantian Romantic conception of the work of art, whereas prose should be thought of as useful, informative, and essentially didactic. Sartre imagines that by making this distinction he is defending prose; he is defending it by asserting that it must give information and not attempt to be poetry. There is a place for this assertion, as a protest for instance against certain degenerations, attempted suicides as it were, of prose
264
language. 'But I shall argue that this is not the best way to defend prose, and is indeed merely another way of betraying it.

We may put Mr. Eliot's attitude to language in terms of a distinction which I made earlier, and which may serve as a formula to explain the Symblist-Romantic position, and to give a clue to much that is, in my view, not well with modern literature. Eliot criticizes the Liberal view of personality because it seems to him to encourage a complete self-absorption of the individual. With this criticism I have some sympathy. But one must ask what it is outside the individual the reality of which ought to be attended to. Eliot answers: things and institutions. Our attention to things properly takes the form of art, and where language is concerned takes 'the form of poetry. Our attention to institutions properly takes the form of didactic reflective prose writing. Eliot does not say that what we should attend to outside ourselves is other persons. So it is not surprising that he makes no place for imaginative prose literature which is par excellence the form of art most concerned with the existence of other persons. And with developments such as these, which conscious form in critics as wide apart as Eliot and Sartre, the stage is set for the decline of the novel. The Romantic theory of art, now passing triumphantly into its last phase, seems to condemn the novel to being either a poem in disguise or else a piece of informative prose, a pamphlet, a human document, or a piece of journalism.

If we turn now to look at modern literature in England and America it is not, I think, difficult to see the pattern which I have indicated. The modern novel, the serious novel, does tend toward either two extremes: either it is a tight metaphysical object, which wishes it were a poem, and which attempts to convey, often in mythical form, some central truth about the human condition~or else it is a loose journalistic epic, documentary or possibly even didactic in inspiration, offering a commentary on current institutions or on' some matter out of history. We are offered things or truths. What we have lost is persons. Modern French literature offers notable examples in both kinds. It is interesting that Sartre and Simone de Beauvoir, both of whom have admitted that they look to art to transform the contingent into the necessary, have been quite willing to surrender to contingency where their own novels are concerned. They have
265
here readily sacrificed form and "universal significance" to the formless and the deliberately ephemeral. This is done partly as a result of the nature of their talents, but partly, and this they say themselves, with a definite didactic purpose. So we have on the one hand a novel like The Mandarins, which is enormous, formless, topical, and often close to being brilliant journalism, and on the other hand, a novel like The Stranger of Camus, which is a small, compact, crystalline, self-contained myth about the human condition, as economical, resonant, and thing-like as it is possible to make any piece of imaginative prose writing to be. There is little point in multiplying examples, though one could amuse oneself by doing so at length; and one thing which would, I think, emerge is that on the whole the small metaphysical novels are better than the social epics. Neurosis pays better dividends than convention. The tightly conceived thing-like books are on the whole better written, more imaginatively conceived, and altogether more inspired and ambitious than the others. Plenty of reasons could be suggested for this. Whereas society in the nineteenth century was either a reassuring place where one lived, or else an exciting, rewarding, interesting place where one struggled, society today tends to appear, by contrast, as menacing, puzzling, uncontrollable, or else confining, and boring. And on the other hand, behind and through society we see the whole apocalyptic scene, the traveling rocket, the hydrogen bomb, and all the things which precisely make us want to think in terms of the human condition and the total man.

Modern literature presents us with the triumph of neurosis, the triumph of myth as a solipsistic form. Our social epics lack creative vitality, and are more concerned with exploration of institutions than with creation of character. While in our metaphysical novels, which represent what is best and most influential in our literature, the hero is alone, with no company, or with only other parts of himself for company. Here Hegel is still king and we have Romanticism in its final, purest, and most undiluted form, where the struggle between persons is really a struggle within the mind of a single character. In such works we feel the ruthless subjection of the characters to the will of their author. The characters are no longer free. The author does not even want them to be free. If they were free they would get in his way. His book is an attempt to work out his own salvation
266
by an exercise in self-discovery. I described the Symbol or image of the Symbolists as the making sensible of an idea of individuality, as being an analogon of the real individual. We may notice that with the dominance of what I have called neurotic Romantic literature the real individual has tended to disappear from the novel, and his place has been taken by the symbolic individual who is the literary work itself The naturalistic conception of character has largely disappeared from the intention of the novelist and from the apparatus of the critic. And indeed how few characters from recent novels can we remember as personalities. We most remember personalities in those cases where a single person has swallowed up the entire book: we remember the hero of The Stranger, we remember the hero of The Catcher in the Rye. But this is because of an overwhelming presence, not of course necessarily in the autobiographical sense, of the author. What we recall is the author himself, or else something very significant about him. Whereas when we think of the works of Tolstoy or George Eliot, we are not remembering Tolstoy and George Eliot, we are remembering Dolly, Kitty, Stiva, Dorothea and Casaubon.

I cannot feel that we should be resigned to this situation. Indeed the danger is rather, as one looks about, not simply that we should become resigned to it, but that we should positively glory in it. But if we do this we abandon something important which is the special property of prose literature. Poetry may tell us truths or it may attempt to be a thing. It may be The Vanity of Human Wishes or Bateau Ivre. But the novel has got to face the special problem of the individual within the work. It has got to solve that problem; and, if I am right, to solve it by denying freedom to the fictional individual either by making him merely part of his creator's mind, or by treating him as a conventional social unit, is likely to be a sort of failure. To speak of failure her has nothing disgraceful about it. Almost every work of art is a failure. The point is that this particular type of failure is one that we ought never to cease worrying about. What sort of works are most effortlessly and naturally written under present social conditions is quite another matter; and it is another matter too that there are major works which disregard the rules I am suggesting. Though I would venture tosay that ultimately we judge the greatest novelists by the quality of their awareness of others, and that for the novelist this is at the highest level the most crucial test. But
267
whatever one may think about the temper of the age, and a~out the value of what I have called neurotic works, it remains important to resist certain critical assumptions, and to keep alive certain standards of value and comparison. Unless we do this, and do it, I think, in something like the way I am suggesting, we shall be unable to understand or explain why it is that Walter Scott and Tolstoy are to a staggering degree better than the most praised of contemporary novelists. And what we cannot explain we may cease to believe. We shall lose our sense of distance; and this will be a pity.

Nowadays we no longer demand of people in books that they should be like real people, except in some minimal sense of verisimilitude in a book of a documentary type. And we may be tempted to forget how almost impossibly difficult it is to create a free and life-like character, or to feel that this particular effort is worth working. This ready surrender of something which it seems to me is the essential power of prose literature is also dangerous at a moment when science leads us to think of personality in a technical and departmentalized fashion, and when philosophy has left us without a way of envisaging the whole person, and has even popularized a very partial view. Mr. Eliot was right to denounce the shallowness of the Liberal concept of personality. And there is a sense here in which literature, for our sake as well as its own, is called on to bear witness.

To ask the novelist to bear these things in mind is not to ask him to become a didactic writer. On the contrary, as I have suggested, didactic writing is, paradoxically, a nemesis of the Romantic view of art, and involves the surrender of the values of which I speak. That literature must be either play (production of self-contained things) or else didactic (discursive statement of truths) is a fallacy which dates from Kant, and which is of the essence of Romanticism. This view, which condemns prose to being either poetry manque or journalism, is I think a misleading view even when applied to poetry. Applied to prose it can be very dangerous indeed. Prose literature can reveal an aspect of the world which no other art can reveal, and the discipline required for this revelation is par excellence the discipline of this art. And in the case of the novel, the most important thing to be thus revealed, not necessarily the only thing, but incomparably the most important thing, is that other people exist.

We may indeed look back to Kant, not only for the source of the
268
error, but for the clue to its solution. Kant contrasted the experience of beauty, the reposeful contemplation of the purposeless self-contained quasi-object, with the experience of the sublime, the upsetting glimpse of the boundlessness of nature; the latter only was a spiritual or moral experience, and for Kant it had nothing to do with art. But one's theory of art must account for the fact that experience of art is spiritual experience. This is more obviously true of some arts than of others, and whether one wants to make it universally true, i.e. to say that any art experience is spiritual experience, is too large a question to pursue here. For purposes of my argument, it seems clear enough that experience of the art of the novel is spiritual experience; and where spirit fails, I would want to argue, art fails. Kant creates the error and suggests the cure; for the theory of the sublime can be transformed into a theory of art. The sublime is an enjoyment and renewal of spiritual power arising from an apprehension of the vast formless strength of the natural world. How close this is indeed to being a theory of tragedy, if we think of the spectator as gazing not at the Alps, but at the spectacle of human life.

It is indeed .the realization of a vast and varied reality outside our-selves which brings about a sense initially of terror, and when properly understood of exhilaration and spiritual power. But what brings this experience to us, in its most important form, is the sight, not of physical nature, but of our surroundings as consisting of other individual men. It is the spectacle of this manifold, if we can actually apprehend it, which is not easy, which brings the exhilaration and the power and reminds us, to use Kant's words, of our supersensible destiny. Obviously its apprehension is not solely the concern of art. But art, of certain types at least, is intimately connected with this apprehension, fails where the apprehension fails, and where it succeeds has spiritual power derived from this source.

What is pictured here is very like Kant's idea of the sublime-and yet importantly different too. Kant's man stands alone confronting the mountains or the sea and feels defiant pride in the free power of his reason. His reason, it is true, is at that moment frustrated and conscious of its inability to achieve complete understanding; but there is nothing humbling or regrettable about this frustration. On the contrary, it brings with it a larger consciousness of the dignity of
269
rationality. Whereas the man that I have in mind, faced by the manifold of humanity, may feel, as well as terror, delight, but not, if he really sees what is before him, superiority. He will suffer that undramatic, because un-self-centered, agnosticism which goes with tolerance. To understand other people is a task which does not come to an end. This man will possess "spirit" in the sense intended by Pascal when he said: "The more spirit one has the more original men one discovers. Ordinary people do not notice differences between men." And a better name for spirit here is not reason, not tolerance even, but love.

When Mr. Eliot praised Jane Austen because her standards, though not lofty, were not of her own making, he was, I have suggested, confusing the question of her achievement as an artist (her ability to see something other than herself) with the question of her attitude to the society she described (which he likes because it is a conservative one). An artist as great, but not as conservative, George Eliot, meets with his disapproval. Yet when he was thinking of his own work, and when he was thinking simply as an artist and not as a politician, Mr. Eliot understood the matter very well, as when he said: "The progress of an artist is a continual self-sacrifice, a continual extinction of personality." This is perfectly true. Art is not an expression of personality, it is a question rather of the continual expelling of oneself from the matter in hand. Anyone who has attempted to write a novel will have discovered this difficulty in the special form which it takes when one is dealing with fictitious characters. Is one going to be able to present any character other than oneself who is more than a conventional puppet? How soon one discovers that, however much one is in the ordinary sense "interested in other people," this interest has left one far short of possessing the knowledge required to create a real character who is not oneself. It is impossible, it seems to me, not to see one's failure here as a sort of spiritual failure. And to look at virtue, morality, as it appears in this context, may perhaps enlighten us concerning its nature in general. Here a diagnosis of literary ills brings us back to moral philosophy and to the philosophy of Liberalism.

Virtue is not essentially or immediately concerned with choosing between actions or rules or reasons, nor with stripping the personality
270
for a leap. It is concerned with really apprehending that other people exist. This too is what freedom really is; and it is impossible not to feel the creation of a work of art as a struggle for freedom. Freedom is not choosing; that is merely the move that we make when all is already lost. Freedom is knowing and understanding and respecting things quite other than ourselves. Virtue is in this sense to be construed as knowledge, and connects us so with reality. The Kantians were wrong to exclude knowledge from virtue, and the Hegelians were wrong to make virtue into a self-knowledge which excluded others. The knowledge and imagination which is virtue is precisely the kind which the novelist needs to let his characters be, to respect their freedom, and to study them themselves in that most significant area of their activity, where they are trying to apprehend the reality of others. The artist is indeed the analogon of the good man, and in a special sense he is the good man: the lover who, nothing himself, lets other things be through him. And that also, I am sure, is what is meant by "negative capability."

I have suggested that it is necessary to detach Liberalism from Romanticism. To do this we must be willing consciously to defend against science, against philosophy, against political theories, against even in some forms literature, a conception of the whole human being, the contingent eccentric fellow, the fellow whom John Stuart Mill lovingly envisaged but whom he was unable philosophically to protect, as having a right to exist. Here prose literature can help our health by not abandoning the naturalistic idea of character; and that this will also be in its own best interests as an art will not be accidental. Literature, morals, and politics must all concern themselves with reality. David Hume uttered perhaps the most important half-truth in the history of political philosophy when he said that something could be true in politics which was false in fact. It seems that what is false in fact cannot go on indefinitely being true in politics; and we have lived for long enough on the notion of the Kantian rational being.

I want to bring this to a close by quoting two remarks of Henry James. The first is his well-known remark concerning Balzac and his characters; that Balzac did not love these people because he knew them, he knew them because he loved them. And this expresses the
271
essence of what I have here wanted to say. The second remark is this, made by James in a letter. He quotes a passage which describes two people falling in love, from a novel by Pierre Loti, and then he says of the passage: "Perhaps you will find in it something of the same strange eloquence of suggestion and rhythm as I do: which is what literature gives when it is most exquisite and which constitutes its sovereign value and its resistance to devouring time.

This, which reads almost strangely now, comes to remind us that novels are after all written in words. I have suggested that we are still suffering from the results of the Romantic attack on words. The novelist who is either poet or journalist is not using prose as he should. A literature that is written in words, like a literature which really envisages persons, is disappearing. And with Henry James's remark we may turn at last to what finally differentiates art from life, the question of form. Form is the temptation of love and its peril, whether in art or life: to round off a situation, to sum up a character. But the difference is that art has got to have form, whereas life need not. And any artist both dreads and longs for the approach of necessity, the moment at which form irrevocably crystallizes. There is a temptation for any novelist, and one to which if I am right modern novelists yield too readily, to imagine that the problem of a novel is solved and the difficulties overcome as soon as a form in the sense of a satisfactory myth has been evolved. But that is only the beginning. There is then the much more difficult battle to prevent that form from becoming rigid, by the free expansion against it of the individual characters. Here above all the contingency of the characters must be respected. Contingency must be defended for it is the essence of personality. And here is where it becomes so important to remember that the novel is written in words, to remember that "eloquence of suggestion and rhythm" of which James spoke. A novel must be a house fit for free characters to live in; and to combine form with a respect for reality with all its odd contingent ways is the highest art of prose.