Catholic Culture Podcasts
Catholic Culture Podcasts

On Criticism

by C.S. Lewis


This essay, On Criticism, is included in the larger work, Of Other Worlds: Essays and Stories by C.S. Lewis.

Larger Work

Of Other Worlds: Essays and Stories

Publisher & Date

First Harvest, 1975

I want to talk about the ways in which an author who is also a critic may improve himself as a critic by reading the criticism of his own work. But I must narrow my subject a little further. It used to be supposed that one of the functions of a critic was to help authors to write better. His praise and censure were supposed to show them where and how they had succeeded or failed, so that next time, having profited by the diagnosis, they might cure their faults and increase their virtues. That was what Pope had in mind when he said, 'Make use of every friend—and every foe.' But that is not at all what I want to discuss. In that way the author-critic might no doubt profit, as a critic, by reviews of his critical work. I am considering how he could profit, as a critic, by reviews of his non-critical works: his poems, plays, stories, or what not; what he can learn about the art of criticism by seeing it practiced on himself; how he can become a better, or less bad, critic of other men's imaginative works from the treatment of his own imaginative works. For I am going to contend that when your own work is being criticized you are, in one sense, in a specially advantageous position for detecting the goodness or badness of the critique.

This may sound paradoxical, but of course all turns on my reservation, in one sense. There is of course another sense in which the author of a book is of all men least qualified to judge the reviews of it. Obviously he cannot judge their evaluation of it, because he is not impartial. And whether this leads him, naively, to hail all laudatory criticism as good and damn all unfavorable criticism as bad, or whether (which is just as likely) it leads him, in the effort against that obvious bias, to lean over backwards till he under-rates all who praise and admires all who censure him, it is equally a disturbing factor. Hence, if by criticism, you mean solely valuation, no man can judge critiques of his own work. In fact, however, most of what we call critical writing contains quite a lot of things besides evaluation. This is specially so both of reviews and of the criticism contained in literary history: for both these always should, and usually try to, inform their readers as well as direct their judgment. Now in so far as his reviewers do that, I contend that the author can see the defects and merits of their work better than anyone else. And if he is also a critic I think he can learn from them to avoid the one and emulate the other; how not to make about dead authors' books the same mistakes that have been made about his own.

I hope it will now be clear that in talking about what I think I have learned from my own critics I am not in any sense attempting what might be called an 'answer to critics'. That would, indeed, be quite incompatible with what I am actually doing. Some of the reviews I find most guilty of the critical vices I am going to mention were wholly favorable; one of the severest I ever had appeared to me wholly free from them. I expect every author has had the same experience. Authors no doubt suffer from self-love, but it need not always be voracious to the degree that abolishes all discrimination. I think fatuous praise from a manifest fool may hurt more than any depreciation.

One critical fault I must get out of the way at once because it forms no part of my real theme: I mean dishonesty. Strict honesty is not, so far as I can see, even envisaged as an ideal in the modern literary world. When I was a young, unknown writer on the eve of my first publication, a kind friend said to me, 'Will you have any difficulty about reviews? I could mention you to a few people. . . .' It is almost as if one said to an undergraduate on the eve of a Tripos, 'Do you know any of the examiners? I could put in a word for you.' Years later another man who had reviewed me with modest favor wrote to me (though a stranger) a letter in which he said that he had really thought much more highly of my book than the review showed: `but of course,' he said, 'if I'd praised it any more the So-and-So would not have printed me at all.' Another time someone had attacked me in a paper called X. Then he wrote a book himself. The editor of X immediately offered it to me, of all people, to review. Probably he only wanted to set us both by the ears for the amusement of the public and the increase of his sales. But even if we take the more favorable possibility—if we assume that this editor had a sort of rough idea of what they call sportsmanship, mid said, 'A has gone for B, it's only fair to let B have a go at A'—it is only too plain that he has no idea of honesty towards the public out of whom he makes his living. They are entitled, at the very least, to honest, that is, to impartial, unbiased criticism: and he cannot have thought that I was the most likely person to judge this book impartially. What is even more distressing is that whenever I tell this story someone replies—mildly, unemphatically—with the question, 'And did you?' This seems to me insulting, because I cannot see how an honest man could do anything but what I did: refuse the editor's highly improper proposal. Of course they didn't mean it as an insult. That is just the trouble. When a man assumes my knavery with the intention of insulting me, it may not matter much. He may only be angry. It is when he assumes it without the slightest notion that anyone could be offended, when he reveals thus lightly his ignorance that there ever were any standards by which it could be insulting, that a chasm seems to open at one's feet.

If I exclude this matter of honesty from my main subject it is not because I think it unimportant. I think it very important indeed. If there should ever come a time when honesty in reviewers is taken for granted, I think men will look back on the present state of affairs as we now look on countries or periods in which judges or examiners commonly take bribes. My reason for dismissing the matter briefly is that I want to talk about the things I hope I have learned from my own reviewers, and this is not one of them. I had been told long before I became an author that one mustn't tell lies (not even by suppressio veri and suggestio falsi) and that we mustn't take money for doing a thing and then secretly do something quite different. I may add before leaving the point that one mustn't judge these corrupt reviewers too harshly. Much is to be forgiven to a man in a corrupt profession at a corrupt period. The judge who takes bribes in a time or place where all take bribes may, no doubt, be blamed: but not so much as a judge who had done so in a healthier civilization.

I now turn to my main subject.

The first thing I have learned from my reviewers is, not the necessity (we would all grant that in principle) but the extreme rarity of conscientiousness in that preliminary work which all criticism should presuppose. I mean, of course, a careful reading of what one criticizes. This may seem too obvious to dwell on. I put it first precisely because it is so obvious and also because I hope it will illustrate my thesis that in certain ways (not of course in others) the author is not the worst, but the best, judge of his critics. Ignorant as he may be of his book's value, he is at least an expert on its content. When you have planned and written and re-written the thing and read it twice or more in proof, you do know what is in it better than anyone else. I don't mean 'what is in it' in any subtle or metaphorical sense (there may, in that sense, be 'nothing in it') but simply what words are, and are not, printed on those pages. Unless you have been often reviewed you will hardly believe how few reviewers have really done their Prep. And not only hostile reviewers. For them one has some sympathy. To have to read an author who affects one like a bad smell or, a toothache is hard work. Who can wonder if a busy man skimps this disagreeable task in order to get on as soon as possible to the far more agreeable exercise of insult and denigration. Yet we examiners do wade through the dullest, most loathsome, most illegible answers before we give a mark; not because we like it, not even because we think the answer is worth it, but because this is the thing we have accepted pay for doing. In fact, however, laudatory critics often show an equal ignorance of the text. They too had rather write than read. Sometimes, in both sorts of review, the ignorance is not due to idleness. A great many people start by thinking they know what you will say, and honestly believe they have read what they expected to read. But for whatever reason, it is certainly the case that if you are often reviewed you will find yourself repeatedly blamed and praised for saying what you never said and for not saying what you have said.

Now of course it is true that a good critic may form a correct estimate of a book without reading every word of it. That perhaps is what Sidney Smith meant when he said 'You should never read a book before you review it. It will only prejudice you. I am not, however, speaking of evaluations based on an imperfect reading, but of direct factual falsehoods about what it contains or does not contain. Negative statements are of course particularly dangerous for the lazy or hurried reviewer. And here, at once, is a lesson for us all as critics. One passage out of the whole Faerie Queene will justify you in saying that Spenser sometimes does so-and-so: only an exhaustive reading and an unerring memory will justify the statement that he never does so. This everyone sees. What more easily escapes one is the concealed negative in statements apparently positive: for example in any statement that contains the predicate 'new'. One says lightly that something which Donne or Sterne or Hopkins did was new: thus committing oneself to the negative that no one had done it before. But this is beyond one's knowledge; taken rigorously, it is beyond anyone's knowledge. Again, things we are all apt to say about the growth or development of a poet may often imply the negative that he wrote nothing except what has come down to us—which no one knows. We have not seen the contents of his waste paper basket. If we had, what now looks like an abrupt change in his manner from poem A to poem B might turn out not to have been abrupt at all.

It would be wrong to leave this point without saying that, however it may be with reviewers, academic critics seem to me now better than they ever were before. The days when Macaulay could get away with the idea that the Faerie Queene contained the death of the Blatant Beast, or Dryden with the remark that Chapman translated the Iliad in Alexandrines, are over. On the whole we now do our homework pretty well. But not yet perfectly. About the more obscure works ideas still circulate from one critic to another which have obviously not been verified by actual reading. I have an amusing piece of private evidence in my possession. My copy of a certain voluminous poet formerly belonged to a great scholar. At first I thought I had found a treasure. The first and second page were richly, and most learnedly annotated in a neat, legible hand. There were fewer on the third; after that, for the rest of the first poem, there was nothing. Each work was in the same state: the first few pages annotated, the rest in mint condition. 'Thus far into the bowels of the land' each time, and no further. Yet he had written on these works.

That, then, is the first lesson the reviewers taught me. There is, of course, another lesson in it. Let no one try to make a living by becoming a reviewer except as a last resource. This fatal ignorance of the text is not always the fruit of laziness or malice. It may be mere defeat by an intolerable burden. To live night and day with that hopeless mountain of new books (mostly uncongenial) piling up on your desk, to be compelled to say something where you have nothing to say, to be always behindhand—indeed much is to be excused to one so enslaved. But of course to say that a thing is excusable is to confess that it needs excuse. I now turn to something which interests me much more because the bottom sin I detect in the reviewers is one which I believe we shall all find it very difficult to banish from our own critical work. Nearly all critics are prone to imagine that they know a great many facts relevant to a book which in reality they don't know. The author inevitably perceives their ignorance because he (often he alone) knows the real facts. This critical vice may take many different forms.

1. Nearly all reviewers assume that your books were written in the same order in which they were published and all shortly before publication. There was a very good instance of this lately in the reviews of Tolkien's Lord of the Rings. Most critics assumed (this illustrates a different vice) that it must be a political allegory and a good many thought that the master Ring must `be' the atomic bomb. Anyone who knew the real history of the composition knew that this was not only erroneous, but impossible; chronologically impossible. Others assumed that the mythology 'of his romance had grown out of his children's story The Hobbit. This, again, he and his friends knew to be mainly false. Now of course nobody blames the critics for not knowing these things: how should they? The trouble is that they don't know they don't know. A guess leaps into their minds and they write it down without even noticing that it is a guess. Here certainly the warning to us all as critics is very clear and alarming. Critics of Piers Plowman and the Faerie Queene make gigantic constructions about the history of these compositions. Of course we should all admit such constructions to be conjectural. And as conjectures, you may ask, are they not, some of them, probable? Perhaps they are. But the experience of being reviewed has lowered my estimate of their probability. Because, when you start by knowing the facts, you find that the constructions are very often wholly wrong. Apparently the chances of their being right are low, even when they are made along quite sensible lines. Of course I am not forgetting that the reviewer has (quite rightly) devoted less study to my book than the scholar has devoted to Langland or Spenser. But I should have expected that to be compensated for by other advantages which he has and the scholar lacks. After all, he lives in the same period as I, subjected to the same currents of taste and opinion, and has undergone the same kind of education. He can hardly help knowing—reviewers are good at this sort of thing and take an interest in it—quite a lot about my generation, my period, and the circles in which I probably move. He and I may even have common acquaintances. Surely he is at least as well placed for guessing about me as any scholar is for guessing about the dead. Yet he seldom guesses right. Hence I cannot resist the conviction that similar guesses about the dead seem plausible only because the dead are not there to refute them; that five minutes conversation with the real Spenser or the real Langland might blow the whole laborious fabric into smithereens. And notice that in all these conjectures the reviewer's error has been quite gratuitous. He has been neglecting the thing he is paid to do, and perhaps could do, in order to do something different. His business was to give information about the book and to pass judgment on it. These guesses about its history are quite beside the mark. And on this point, I feel pretty sure that I write without bias. The imaginary histories written about my books are by no means always offensive. Sometimes they are even complimentary. There is nothing against them except that they're not true, and would be rather irrelevant if they were. Mutato nomine de me. I must learn not to do the like about the dead: and if I hazard a conjecture, it must be with full knowledge, and with a clear warning to my readers, that it is a long shot, far more likely to be wrong than right.

2. Another type of critic who speculates about the genesis of your book is the amateur psychologist. He has a Freudian theory of literature and claims to know all about your inhibitions. He knows what unacknowledged wishes you were gratifying. And here of course one cannot, in the same sense as before, claim to start by knowing all the facts. By definition you are unconscious of the things he professes to discover. Therefore the more loudly you disclaim them, the more right he must be: though, oddly enough, if you admitted them, that would prove him right too. And there is a further difficulty: one is not here so free from bias, for this procedure is almost entirely confined to hostile reviewers. And now that I come to think of it, I have seldom seen it practiced on a dead author except by a scholar who intended, in some measure, to debunk him. That in itself is perhaps significant. And it would not be unreasonable to point out that the evidence on which such amateur psychologists base their diagnosis would not be thought sufficient by a professional. They have not had their author on the sofa, nor heard his dreams, and had the whole case-history. But I am here concerned only with what the author can say about such reviews solely because he is the author. And surely, however ignorant he is of his unconscious, he knows something more than they about the content of his conscious mind. And he will find them wholly overlooking the (to him) perfectly obvious conscious motive for some things. If they mentioned this and then discounted it as the author's (or patient's) 'rationalization', they might be right. But it is clear that they have never thought of it. They have never seen why, from the very structure of your story, from the very nature of story telling in general, that episode or image (or something like it) had to come in at that point. It is in fact quite clear that there is one impulse in your mind of which, with all their psychology, they have never reckoned: the plastic impulse, the impulse to make a thing, to shape, to give unity, relief, contrast, pattern. But this, unhappily, is the impulse which chiefly caused the book to be written at all. They have, clearly, no such impulse themselves, and they do not suspect it in others. They seem to fancy that a book trickles out of one like a sigh or a tear or automatic writing. It may well be that there is much in every book which comes from the unconscious. But when it is your own book you know the conscious motives as well. You may be wrong in thinking that these often give the full explanation of this or that. But you can hardly believe accounts of the sea-bottom given by those who are blind to the most obvious objects on the surface. They could be right only by accident. And I, if I attempt any similar diagnosis about the dead, shall equally be right, if at all, only by accident.

The truth is that a very large part of what comes up from the unconscious and which, for that very reason, seems so attractive and important in the early stages of planning a book, is weeded out and jettisoned long before the job is done: just as people (if they are not bores) tell us of their dreams only those which are amusing or in some other way interesting by the standards of the waking mind.

3. I now come to the imaginary history of the book's composition in a much subtler form. Here I think critics, and of course we when we criticize, are often deceived or confused as to what they are really doing. The deception may lurk in the words themselves. You and I might condemn a passage in a book for being 'labored'. Do we mean by this that it sounds 'labored'.? Or are we advancing the theory that it was in fact 'labored'.? Or are we sometimes not quite sure which we mean? If we mean the second, notice that we are ceasing to write criticism. Instead of pointing out the faults in the passage we are inventing a story to explain, causally, how it came to have those faults. And if we are not careful we may complete our story and pass on as if we had done all that was necessary, without noticing that we have never even specified the faults at all. We explain something by causes without saying what the something is. We can do the same when we think we are praising. We may say that a passage is unforced or spontaneous. Do we mean that it sounds as if it were, or that it actually was written effortlessly and currente calamo? And whichever we mean, would it not be more interesting and more within the critics' province to point out, instead, those merits in the passage which made us want to praise it at all?

The trouble is that certain critical terms—inspired, perfunctory, painstaking, conventional—imply a supposed history of composition. The critical vice I am talking about consists in yielding to the temptation they hold out and then, instead of telling us what is good and bad in a book, inventing stories about the process which led to the goodness and badness. Or are they misled by the double sense of the word Why? For of course the question 'Why is this bad?' may mean two things: (a) What do you mean by calling it bad? Wherein does its badness consist? Give me the Formal Cause. (b) How did it become bad? Why did he write so ill? Give me the Efficient Cause. The first seems to me the essentially critical question. The critics I am thinking of answer the second, and usually answer it wrong, and unfortunately regard this as a substitute for the answer to the first.

Thus a critic will say of a passage, 'This is an afterthought.' He is just as likely to be wrong as right. He may be quite right in thinking it bad. And he must presumably think he has discerned in it the sort of badness which one might expect to occur in an afterthought. Surely an exposure of that badness itself would be far' better than an hypothesis about its origin? Certainly this is the only thing that would make the critique at all useful to the author. I as author may know that the passage diagnosed as an afterthought was in reality the seed from which the whole book grew. I should very much like to be shown what inconsistency or irrelevance or flatness makes it look like an afterthought. It might help, me to avoid these errors next time. Simply to know what the critic imagines, and imagines wrongly, about the history of the passage is of no use. Nor is it of much use to the public. They have every right to be told of the faults in my book. But this fault, as distinct from an hypothesis (boldly asserted as fact) about its origin, is just what they do not learn.

Here is an example which is specially important because I am quite sure the judgment which the critic was really making was correct. In a book of essays of mine the critic said that one essay was written without conviction, was task-work, or that my heart was not in it, or something like that. Now this in itself was plumb-wrong. Of all the pieces in the book it was the one I most cared about and wrote with most ardour.1 Where the critic was right was in thinking it the worst. Everyone agrees with him about that. I agree with him. But you see that neither the public nor I learns anything about that badness from his criticism. He is like a doctor who makes no diagnosis and prescribes no cure but tells you how the patient got the disease (still unspecified) and tells you wrong because he is describing scenes events on which he has no evidence. The fond parents ask, 'What is it? Is it scarlatina or measles or chicken-pox?' The doctor replies, 'Depend upon it, he picked it up in one of those crowded trains.' (The patient actually has not traveled by train lately.) They then ask, ‘But what are we to do? How led are we to treat him?' The doctor replies, 'You may be quite sure it was an infection.' Then he climbs into his car and drives away. Notice here again the total psychological disregard of writing as a skill, the assumption that the writer’s psychological state always flows unimpeded and undisguised into the product. How can they not know that in writing as in carpentry or tennis-playing or prayer or lovemaking or cookery or administration or anything else there is both skill and also those temporary heightenings and lowerings of skill which a man describes by saying that he is in good or bad form, that his hand is 'in' or 'out', that this is one of his good days or his bad days?

Such is the lesson, but it is very difficult to apply. It needs great perseverance to force oneself, in one's own criticism, to attend always to the product before one instead of writing fiction about the author's state of mind or methods of work: to which of course one has no direct access. 'Sincere', for example, is a word we should avoid. The real question is what makes a thing sound sincere or not. Anyone who has censored letters in the army must know that semi-literate people, though not in reality less sincere than others, very seldom sound sincere when they use the written word. Indeed we all know from our own experience in writing letters of condolence that the occasions on which we really feel most are not necessarily those when on which our letters would suggest this. Another day, when we felt far less, our letter may have been more convincing. And of course the danger of error is greater in proportion as our own experience in the form we are criticizing is less. When we are criticizing a kind of work we have never attempted ourselves, we must realize that we do not know how such things are written and what is difficult or easy to do in them and how particular faults are likely to occur. Many critics quite clearly have an idea of how they think they would proceed if they tried to write the sort of book you have written, and assume that you were doing that. They often reveal unconsciously why they never have written a book of that kind.

I don't mean at all that we must never criticize work of a kind we have never done. On the contrary we must do nothing but criticize it. We may analyze and weigh its virtues and defects. What we must not do is to write imaginary histories. I know that all beer in railway refreshment rooms is bad and I could to some extent say 'why' (in one sense of the word: that is, I could give the Formal Cause)—it is tepid, sour, cloudy, and weak. But to tell you 'why' in the other sense (the Efficient Cause) I should need to have been a brewer or a publican or both and to know how beer should be brewed and kept and handled.

I would gladly be no more austere than is necessary. I must admit that words which seem, in their literal sense, to imply a history of the composition may sometimes be used as merely elliptical pointers to the character of the work done. When a man says that something is 'forced' or 'effortless' he may not really be claiming to know how it was written but only indicating in a kind of short-hand a quality he supposes everyone will recognize. And perhaps to banish all expression of this kind from our criticism would be a counsel of perfection. But I am increasingly convinced of their danger. If we use them at all, we must do so with extreme caution. We must make it quite clear to ourselves and to our readers that we do not know and are not pretending to know how things were written. Nor would it be relevant if we did. What sounds forced would be no better if it had been dashed off without pains; what sounds inspired, no worse if it had been arduously put together invita Minerva.

I now turn to interpretation. Here of course all critics, and we among them, will make mistakes. Such mistakes are far more venial than the sort I have been describing, for they are not gratuitous. The one sort arise when the critic writes fiction instead of criticism; the other, in the discharge of a proper function. At least I assume that critics ought to interpret, ought to try to find out the meaning or intention of a book. When they fail the fault may lie with them or with the author or with both.

I have said vaguely 'meaning' or 'intention'. We shall have to give each word a fairly definite sense. It is the author who intends; the book means. The author's intention is that which, if it is realized, will in his eyes constitute success. If all or most readers, or such readers as he chiefly desires, laugh at a passage, and he is pleased with this result, then his intention was comic, or he intended to be comic. If he is disappointed and humiliated at it, then he intended to be grave, or his intention was serious. Meaning is a much more difficult term. It is simplest when used of an allegorical work. In the Romance of the Rose plucking the rosebud means enjoying the heroine. It is still fairly easy when used of a work with a conscious and definite 'lesson' in it. Hard Times means, among other things, that elementary state education is bosh; Macbeth, that your sin will find you out; Waverley, that solitude and abandonment to the imagination in youth render a man an easy prey to those who wish to exploit him; the Aeneid, that the res Romana rightly demands the sacrifice of private happiness. But we are already in deep waters, for of course each of these books means a good deal more. And what are we talking about when we talk, as we do, of the 'meaning' of Twelfth Night, Wuthering Heights, or The Brothers Karamazov? And especially when we differ and dispute as we do, about their real or true meaning? The nearest I have yet got to a definition is something like this: the meaning of a book is the series or system of emotions, reflections and attitudes produced by reading it. But of course this product differs with different readers. The ideally false or wrong 'meaning' would be the product in the mind of the stupidest and least sensitive and most prejudiced reader after a single careless reading. The ideally true or right 'meaning' would be that shared (in some measure) by the largest number of the best readers after repeated and careful readings over several generations, different periods, nationalities, moods, degrees of alertness, private pre-occupations, states of health, spirits, and the like canceling one another out when (this is an important reservation) they cannot be fused so as to enrich one another. (This happens when one's readings of a work at widely different periods of one's own life, influenced by the readings that reach us indirectly through the works of critics, all modify our present reading so as to improve it.) As for the many generations, we must add a limit. These serve to enrich the perception of the meaning only so long as the cultural tradition is not lost. There may come a break or change after which readers arise whose point of view is so alien that they might as well be interpreting a new work. Medieval readings of the Aeneid as an allegory and Ovid as a moralist, or modern readings of the Parlement of Foules which make the duck and goose its heroes, would be examples. To delay, even if we cannot permanently banish such interpretations, is a large part of the function of scholarly, as distinct from pure, criticism; so doctors lab our to prolong life though they know they cannot make men immortal.

Of a book’s meaning, in this sense, its author is not necessarily the best, and is never a perfect, judge. One of his intentions usually was that it should have a certain meaning: he cannot be sure that it has. He cannot even be sure that the meaning he intended it to have was in every way, or even at all, better than the meaning which readers find in it. Here, therefore, the critic has great freedom to range without fear of contradiction from the author's superior knowledge.

Where he seems to me most often to go wrong is in the hasty assumption of an allegorical sense; and as reviewers make this mistake about contemporary works, so, in my opinion, scholars now often make it about old ones. I would recommend to both, and I would try to observe in my own critical practice, these principles. First, that no story can be devised by the wit of man which cannot be interpreted allegorically by the wit of some other man. The Stoic interpretations of primitive mythology, the Christian interpretations of the Old Testament, the medieval interpretations of the classics, all prove this. Therefore (2) the mere fact that you can allegorize the work before you is of itself no proof that it is an allegory. Of course you can allegorize it. You can allegorize anything, whether in art or real life. I think we should here take a hint from the lawyers. A man is not tried at the assizes until there has been shown to be a prima-facie case against him. We ought not to proceed to allegorize any work until we have plainly set out the reasons for regarding it as an allegory at all.

[Lewis, apparently, did not finish this essay for at the foot of the existing manuscript are the words:]

As regards other attributions of intention
One's own preoccupations
Quellenforschung. Achtung—dates


1. Lewis, I am quite certain, is talking about the essay on William Morris in Rehabilitations and Other Essays (1939).

Of Other Worlds: Essays and Stories may be purchased from Amazon.

This item 9115 digitally provided courtesy of