I think it an abuse to use the novel as a pulpit or a platform, and I believe readers are misguided when they suppose they can thus easily acquire knowledge. It is a great nuisance that knowledge can only be acquired by hard work. It would be fine if we could swallow the powder of profitable information made palatable by the jam of fiction. But the truth is that, so made palatable, we can’t be sure that the powder will be profitable, for the knowledge the novelist imparts is biassed and thus unreliable; and it is better not to know a thing at all than to know it in a distorted fashion. There is no reason why a novelist should be anything but a novelist. It is enough if he is a good novelist. He should know a little about a great many things, but it is unnecessary, and sometimes even harmful, for him to be a specialist in any particular subject. He need not eat a whole sheep to know what mutton tastes like; it is enough if he eats a chop. Then, by applying his imagination and his creative faculty to the chop he has eaten, he can give you a pretty good idea of an Irish stew; but when he goes on from this to broach his views on sheep-raising, the wool industry and the political situation in Australia, it is wise to accept them with reserve.

  The novelist is at the mercy of his bias. The subjects he chooses, the characters he invents and his attitude towards them are conditioned by it. Whatever he writes is the expression of his personality and it is the manifestation of his innate instincts, his feeling and his experience. However hard he tries to be objective, he remains the slave of his idiosyncrasies. However hard he tries to be impartial, he cannot help taking sides. He loads his dice. By the mere fact of introducing a character to your notice early in his novel, he enlists your interest and your sympathy in that character. Henry James insisted again and again that the novelist must dramatize. That is a telling, though perhaps not very lucid, way of saying that he must arrange his facts in such a manner as to capture and hold your attention. So, if need be, he will sacrifice verisimilitude and credibility to the effect he wants to get. That, as we know, is not the way a work of scientific or informative value is written. The aim of the writer of fiction is not to instruct, but to please.

  (2)

  There are two main ways in which a novel may be written. Each has its advantages, and each its disadvantages. One way is to write it in the first person, and the other is to write it from the standpoint of omniscience. In the latter, the author can tell you all that he thinks is needful to enable you to follow his story and understand his characters. He can describe their emotions and motives from the inside. If one of them crosses the street, he can tell you why he does so and what will come of it. He can concern himself with one set of persons and series of events, and then, putting them aside for a period, can concern himself with another side of events and another set of persons, so reviving a flagging interest and, by complicating his story, give an impression of the multifariousness, complexity and diversity of life. The danger of this is that one set of characters may be so much more interesting than the other, as, to take a famous example, happens in Middlemarch, that the reader may find it irksome when he is asked to occupy himself with the fortunes of persons he doesn’t in the least care about. The novel written from the standpoint of omniscience runs the risk of being unwieldy, verbose and diffuse. No one has written it better than Tolstoy, but even he is not free from these imperfections. The method makes demands on the author which he cannot always meet. He has to get into the skin of every one of his characters, feel his feelings, think his thoughts; but he has his limitations and he can only do this when there is in himself something of the character he has created. When there isn’t, he can only see him from the outside, and then the character lacks the persuasiveness which causes the reader to believe in him.

  I suppose it was because Henry James, with his solicitude for form in the novel, became conscious of these disadvantages that he devised what may be described as a sub-variety of the method of omniscience. In this the author is still omniscient, but his omniscience is concentrated on a single character, and since the character is fallible the omniscience is not complete. The author wraps himself in omniscience when he writes: ‘He saw her smile’; but not when he writes. ‘He saw the irony of her smile’; for irony is something he ascribes to her smile, and, it may be, without justification. The usefulness of the device, as Henry James without doubt very well saw, is that since this particular character, in The Ambassadors, Strether, is all-important, and it is through what he sees, hears, feels, thinks, surmises that the story is told and the characters of the other persons concerned in it are unfolded, the author finds it easy to resist the irrelevant. The construction of his novel is necessarily compact. The device, besides, gives an air of verisimilitude to what he writes. Because you are asked to concern yourself primarily with one person, you are insensibly led to believe what he tells you. The facts that the reader should know are imparted to him as the person through whom the story is told gradually learns them; and so the reader enjoys the pleasure of the elucidation, step by step, of what was puzzling, obscure and uncertain. The method thus gives the novel something of the mystery of a detective story, and so that dramatic quality which Henry James was always eager to obtain. The danger, however, of divulging little by little a string of facts is that the reader may be more quick-witted than the character through whom the revelations are made and so guess the answers long before the author wishes him to. I don’t suppose anyone can read The Ambassadors without growing impatient with Strether’s obtuseness. He does not see what is staring him in the face, and what everyone he comes in contact with is fully aware of. It was a secret de Polichinelle and that Strether should not have guessed it points to some defect in the method. It is unsafe to take your reader for more of a fool than he is.

  Since novels have for the most part been written from the standpoint of omniscience, it must be supposed that novelists have found it on the whole the most satisfactory way of dealing with their difficulties; but to tell a story in the first person has also certain advantages. Like the method adopted by Henry James, it lends verisimilitude to the narrative and obliges the author to stick to his point; for he can tell you only what he has himself seen, heard or done. To use this method more often would have served the great English novelist of the nineteenth century well, since, partly owing to methods of publication, partly owing to a national idiosyncrasy, their novels have tended to be shapeless and discursive. Another advantage of using the first person is that it enlists your sympathy with the narrator. You may disapprove of him, but he concentrates your attention on himself and so compels your sympathy. A disadvantage of the method, however, is that the narrator, when, as in David Copperfield, he is also the hero, cannot without inpropriety tell you that he is handsome and attractive; he is apt to seem vainglorious when he relates his doughty deeds and stupid when he fails to see, what is obvious to the reader, that the heroine loves him. But a greater disadvantage still, and one that no authors of this kind of novel have managed entirely to surmount, is that the hero-narrator, the central character, is likely to appear pallid in comparison with the persons he is concerned with. I have asked myself why this should be, and the only explanation I can suggest is that the author, since he sees himself in the hero, sees him from the inside, subjectively, and, telling what he sees, gives him the confusions, the weaknesses, the indecisions he feels in himself; whereas he sees the other characters from the outside, objectively, through his imagination and his intuition; and if he is an author with, say, Dickens’s brilliant gifts, he sees them with a dramatic intensity, with a boisterous sense of fun, with a keen delight in their oddity, and so makes them stand out with a vividness that overshadows his portrait of himself.

  There is a variety of the novel written on these lines which for a time had an immense vogue. This is the novel written in letters; each letter, of course, is written in the first person, but the letters are by different hands. The method had the advantage of extreme verisimilitude. The reader might easily believe that they were real letters, written by the persons
they purported to have been written by, and come into his hands by a betrayal of confidence. Now, verisimilitude is what the novelist strives to achieve above all else; he wants you to believe that what he tells you actually happened, even if it is as improbable as the tales of Baron Münchausen or as horrifying as Kafka’s The Castle. But the genre had grave defects. It was a roundabout, complicated way of telling a story, and it told it with intolerable deliberation. The letters were too often verbose and contained irrelevant matter. Readers grew bored with the method and it died out. It produced three books which may be accounted among the masterpieces of fiction: Clarissa, La Nouvelle Héloise and Les Liaisons Dangereuses.

  There is, however, a variety of the novel written in the first person which, to my mind, avoids the defects of the method and yet makes handsome use of its merits. It is, perhaps, the most convenient and effective way in which a novel can be written. To what good use it can be put may be seen in Herman Melville’s Moby Dick. In this variety, the author tells the story himself, but he is not the hero and it is not his story that he tells. He is a character in it, and is more or less closely connected with the persons who take part in it. His role is not to determine the action, but to be the confidant, the mediator, the observer of those who do take part in it. Like the chorus in a Greek tragedy, he reflects on the circumstances which he witnesses; he may lament, he may advise, he has no power to influence the course of events. He takes the reader into his confidence, tells him what he knows, hopes or fears, and when he is non-plussed frankly tells him so. There is no need to make him stupid, so that he should not divulge to the reader what the author wishes to hold back, as happens when the story is told through such a character as Henry James’s Strether. On the contrary, he can be as keen-witted and clear-sighted as the author can make him. The narrator and the reader are united in their common interest in the persons of the story, their characters, motives and conduct; and the narrator begets in the reader the same sort of familiarity with the creatures of his invention as he has himself. He gets an effect of verisimilitude as persuasive as that which the author obtains who is himself the hero of his novel. He can so build up his protagonist as to arouse your sympathy and show him in an herotid light, which the hero-narrator cannot do without somewhat exciting your antagonism. A method of writing a novel which conduces to the reader’s intimacy with the characters, and adds to its verisimilitude, has obviously much to recommend it.

  I will venture now to state what in my opinion are the qualities that a good novel should have. It should have a widely interesting theme, by which I mean a theme interesting not only to a clique, whether of critics, professors, highbrows, bus-conductors or bar-tenders, but so broadly human that its appeal is to men and women in general; and the theme should be of enduring interest: the novelist is rash who elects to write on subjects whose interest is merely topical. When they cease to be so, his novel will be as unreadable as last week’s newspaper. The story the author has to tell should be coherent and persuasive; it should have a beginning, a middle and an end, and the end should be the natural consequence of the beginning. The episodes should have probability and should not only develop the theme, but grow out of the story. The creatures of the novelist’s invention should be observed with individuality, and their actions should proceed from their characters; the reader must never be allowed to say: ‘so and so would never behave like that’; on the contrary, he should be obliged to say: ‘That’s exactly how I should have expected so and so to behave.’ I think it is all the better if the characters are in themselves interesting. In Flaubert’s L’Éducation Sentimentale he wrote a novel which has a great reputation among many excellent critics, but he chose for his hero a man so null, so featureless, so vapid that it is impossible to care what he does or what happens to him; and in consequence, for all its merits, the book is hard to read. I think I should explain why I say that characters should be observed with individuality: it is too much to expect the novelist to create characters that are quite new; his material is human nature, and although there are all sorts and conditions of men, the sorts are not infinite, and novels, stories, plays, epics have been written for so many hundreds of years that the chance is small that an author will create an entirely new character. Casting my mind’s eye over the whole of fiction, the only absolutely original creation I can think of is Don Quixote, and I should not be surprised to learn that some learned critic had found a remote ancestry for him also. The author is fortunate if he can see his characters through his own individuality, and if his individuality is sufficiently out of the common to give them an illusive air of originality.

  And just as behaviour should proceed from character, so should speech. A woman of fashion should talk like a woman of fashion, a street-walker like a street-walker, a racing tout like a racing tout and an attorney like an attorney. (It is surely a fault in Meredith and Henry James that their characters invariably talk like Henry James and Meredith respectively.) The dialogue should be neither desultory nor should it be an occasion for the author to air his views; it should serve to characterise the speakers and advance the story. The narrative passages should be vivid, to the point, and no longer than is necessary to make the motives of the persons concerned, and the situations in which they are placed, clear and convincing. The writing should be simple enough for anyone of fair education to read with ease, and the manner should fit the matter as a well-cut shoe fits a shapely foot. Finally, a novel should be entertaining. I have put this last, but it is the essential quality, without which no other quality avails. And the more intelligent the entertainment a novel offers, the better it is. Entertainment is a word that has a good many meanings. One item is that which affords interest for amusement. It is a common error to suppose that in this sense amusement is the only one of importance. There is as much entertainment to be obtained from Wuthering Heights or The Brothers Karamazov as from Tristram Shandy or Candide. The appeal is different, but equally legitimate. Of course, the novelist has the right to deal with those great topics which are of concern to every human being, the existence of God, the immortality of the soul, the meaning and value of life; though he is prudent to remember that wise saying of Dr. Johnson’s that of these topics one can no longer say anything new about them that is true, or anything true about them that is new. The novelist can only hope to interest his reader in what he has to say about them if they are an integral element of the story he has to tell, are essential to the characterisation of the persons of his novel and affect their conduct – that is, if they result in action which otherwise would not have taken place.

  But even if the novel has all the qualities that I have mentioned, and that is asking a lot, there is, like a flaw in a precious stone, a faultiness in the form that renders perfection impossible to attain. That is why no novel is perfect. A short story is a piece of fiction that can be read, according to its length, in anything between ten minutes and an hour, and it deals with a single, well-defined subject, an incident or a closely related series of incidents, spiritual or material, which is complete. It should be impossible to add to it or to take away from it. Here, I believe, perfection can be reached, and I do not think it would be difficult to collect a number of short stories in which this has in fact been done. But a novel is a narrative of indefinite length; it may be as long as War and Peace, in which a succession of events is related and a vast number of characters are displayed through a period of time, or as short as Carmen. Now, in order to give probability to his story, the author has to narrate a series of facts that are relevant to it, but that are not in themselves interesting. Events often require to be separated by a lapse of time, and the author for the balance of his work has to insert, as best he can, matter that will fill up this lapse. These passages are known as bridges. Most writers resign themselves to crossing them, and they cross them with more or less skill, but it is only too likely that in the process they will be tedious. The novelist is human and it is inevitable that he should be susceptible to the fashions of his da
y, since after all he has an unusual affectivity, and so is often led to write what, as the fashion passes, loses its attractiveness. Let me give an instance: until the nineteenth century novelists paid little attention to scenery, a word or two sufficed to enable them to say all they wanted to about it; but when the romantic school, and the example of Chateaubriand, captivated the public fancy, it grew modish to write descriptions for their own sake. A man could not go down a street to buy a tooth-brush at the chemist’s without the author telling you what the houses he passed looked like and what articles were for sale in the shops. Dawn and the setting sun, the starry night, the cloudless sky, the snow-capped mountains, the dark forests – all gave occasion to interminable descriptions. Many were in themselves beautiful; but they were irrelevant: it took writers a long time to discover that a description of scenery, however poetically observed and admirably expressed, was futile unless it was necessary – that is, unless it helped the author to get on with his story or told the reader something it behoved him to know about the persons who take part in it. This is an adventitious imperfection in the novel, but there is yet another that seems inherent. Since it is a work of considerable length, it must take some time to write, weeks at least, generally months and occasionally even years. It is only too likely that the author’s inventiveness will sometimes fail him. Then he can only fall back on dogged industry and his general competence. It will be a marvel if by these means he can hold his readers’ attention.