The Three Voices of Poetry
Those who have seen or read The Confidential Clerk and The Cocktail Party will follow with special interest this discussion by T. S. ELIOTof the aims and responsibilities of a poet who is also a playwright. The essay which follows is a slightly condensed version of the lecture which Mr. Eliot delivered at the Annual Meeting of the Book League in London. In book form it will be published by the Cambridge University Press.
by T. S. ELIOT
I SHALL explain at once what I mean by the “three voices.” The first is the voice of the poet talking to himself-or to nobody. The second is the voice of the poet addressing an audience, whether large or small. The third is the voice of the poet when he attempts to create a dramatic character speaking in verse; when he is saving, not what he would say in his own person, but only what he can say within the limits of one imaginary character addressing another imaginary character. The distinction between the first and the second voice, between the poet speaking to himself and the poet speaking to other people, points to the problem of poetic communication; the distinction between the poet addressing other people in either his own voice or an assumed voice, and the poet inventing speech in which imaginary characters address each other, points to the difference between dramatic, quasi-dramatic, and non-dramatic verse.
I wish to anticipate a question that some of you may well raise. Cannot a poem be written for the ear, or for the eye, of one person alone? You may say simply, “Isn’t love poetry at times a form of communication between one person and one other, with no thought of a further audience?”
There are at least two people who might heckle me on this point if they could be present: Mr. and Mrs. Robert Browning. In the poem “One Word More,” written as an epilogue to Men and Women, and addressed to Mrs. Browning, the husband makes a striking value judgment:—
Made and wrote them in a certain volume
Dinted with the silver-pointed pencil
Else he only used to draw Madonnas:
These, the world might view — but one, the volume.
Who that one, you ask? Your heart instructs you. . . .
Would we not? than wonder at Madonnas— . . .
Dante once prepared to paint an angel:
Whom to please? You whisper “Beatrice.”. . .
You and I would rather see that angel,
Painted by the tenderness of Dante,
Would we not?—than read a fresh Inferno.
I agree that one Inferno, even by Dante, is enough; and perhaps we need not too much regret the fact that Rafael did not live to multiply his Madonnas: but I can only say that I feel no curiosity whatever about Rafael’s sonnets or Dante’s angel. If Rafael wrote, or Dante painted, for the eyes of one person alone, let their privacy be respected. We know that Mr. and Mrs. Browning liked to write poems to each other, because they published them, and some of them are good poems. We know that Rossetti thought that he was writing his “House of Life" sonnets for one person, and that he was only persuaded by his friends to disinter them. Now, I do not deny that a poem may be addressed to one person: there is a well-known form, not always amatory in content, called The Epistle. We shall never have conclusive evidence: for the testimony of poets as to what they thought they were doing when they wrote a poem, cannot be taken altogether at its face value. But my opinion is, that a good love poem, though it may be addressed to one person, is always meant to be overheard by other people. Surely, the proper language of love — that is, of communication to the beloved and to no one else—is prose.
Having dismissed as an illusion the voice of the poet talking to one person only, I think that the best way for me to try to make my three voices audible to you, is to trace the genesis of the distinction in my own mind. The writer to whose mind the distinction is most likely to occur is probably the writer like myself, who has spent a good many years in writing poetry, before attempting to write for the stage at all. It may be, as I have read, that there is a dramatic element in much of my early work. It may be that from the beginning I aspired unconsciously to the theatre — or, critics might say with more asperity, to Shaftesbury Avenue. I have, however, gradually come to the conclusion that in writing verse for the stage both the process and the outcome are very different from what they are in writing verse to be read or recited.
Twenty years ago I was commissioned to write a pageant play to be called The Rock. The invitation to write the words for this spectacle — the occasion of which was an appeal for funds for church-building in new housing areas - came at a moment when I seemed to myself to have exhausted my meagre poetic gifts, and to have nothing more to say. To be, at such a moment, commissioned to write something which, good or bad, must be delivered by a certain date, may have the effect that vigorous cranking sometimes has upon a motor-car when the battery is run down. The task was clearly laid out: I had only to write the words of prose dialogue for scenes of the usual historical pageant pattern, for which I had been given a scenario. I had also to provide a number of choral passages in verse, the content of which was left to my own devices: except for the reasonable stipulation that all the choruses were expected to have some relevance to the purpose of the pageant, and that each chorus was to occupy a precise number of minutes of stage time. But in carrying out this task, there was nothing to call my attention to the third, or dramatic voice: it was the second voice, that of myself addressing — indeed haranguing — an audience, that was most distinctly audible.
Apart from the obvious fact that writing to order is not the same thing as writing to please oneself, I learned only that verse to be spoken by a choir should be different from verse to be spoken by one person; and that, the more voices you have in your choir, the simpler and more direct the vocabulary, the syntax, and the content of your lines must be. This chorus of The Rock was not a dramatic voice; though many lines were distributed, the personages were unindividuated. Its members were speaking for me, not uttering words that really represented any supposed character of their own.
The chorus in Murder in the Cathedral does, I think, represent some advance in dramatic development: that is to say, I set myself the task of writing lines, not for an anonymous chorus, but for a chorus of women of Canterbury — one might almost say, the charwomen of the Cathedral. I had to make some effort to identify myself with these women, instead of merely identifying them with myself. But as for the dialogue of the play, the plot had the drawback (from the point of view of my own dramatic education) of presenting only one dominant character; and what dramatic conflict there is takes place within the mind of that character. The third, or dramatic voice, did not make itself audible to me until I first attacked the problem of presenting two (or more) characters, in some sort of conflict, misunderstanding, or attempt to understand each other, characters with each of whom I had to try to identify myself while writing the words for him or her to speak. You may remember that Mrs. Cluppins, in the trial of the case of Bardell v. Pickwick, testified that “the voices was very loud, sir, and forced themselves upon my ear.” “Well, Mrs. Cluppins,” said Serjeant Buzfuz, “you were not listening, but you heard the voices.” It was in 1938, then, that the third voice began to force itself upon my ear.
IN a lecture on “Poetry and Drama,” delivered three years ago and subsequently published in the Atlantic, I said: —
In writing other verse [i.e. non-dramatic verse], I think that one is writing, so to speak, in terms of one’s own voice: the way it sounds when you read it to yourself is the test. For it is yourself speaking. The question of communication, of what the reader will get from it, is not paramount . . .
There is some confusion of pronouns in this passage, but I think that the meaning is clear; so clear, as to be a. glimpse of the obvious. At that stage, I noted only the difference between speaking for oneself, and speaking for an imaginary character; and I passed on to other considerations about the nature of poetic drama. I was beginning to be aware of the difference between the first and the third voice, but gave no attention to the second voice. I am now trying to penetrate a little further into the problem. So, before going on to consider the other voices, I want to pursue for a few moments the complexities of the third voice.
In a verse play, you will probably have to find words for several characters differing widely from each other in background, temperament, education, and intelligence. You cannot afford to identify one of these characters with yourself, and give him (or her) all the “poetry" to speak. The poetry (I mean, the language at those dramatic moments when it reaches intensity) must be as widely distributed as characterization permits; and each of your characters, when he has words to speak which are poetry and not merely verse, must be given lines appropriate to himself. When the poetry comes, the personage on the stage must not give the impression of being merely a mouthpiece for the author. Hence the author is limited by the kind of poetry, and the degree of intensity in its kind, which can be plausibly attributed to each character in his play. And these lines of poetry must also justify themselves by their development of the situation in which they are spoken. Even if a burst of magnificent. poetry is suitable enough for the character to which it is assigned, it must also convince us that it is necessary to the action; that it is helping to extract the utmost emotional intensity out of the situation.
The poet writing for the theatre may, as I have found, make two mistakes: that of assigning to a personage lines of poetry not suitable to be spoken by that personage, and that of assigning lines which, however suitable to the personage, yet fail to forward the action of the play. There are, in some of the minor Elizabethan dramatists, passages of magnificent poetry which are in both respects out of place — fine enough to preserve the play for ever as literature, but yet so inappropriate as to prevent the play from being a dramatic masterpiece. The best known instances occur in Marlowe’s Tamburlaine.
How have the very great dramatic poets — Sophocles, or Shakespeare, or Racine — dealt with this difficulty? This is, of course, a problem which concerns all imaginative fiction - novels and prose plays — in which the characters may be said to live. I can’t see, myself, any way to make a character live except to have a profound sympathy with that character. Ideally, a dramatist, who has usually far fewer characters to manipulate than a novelist, and who has only two hours or so of life to allow them, should sympathize profoundly with all of his characters: but that is a counsel of perfection, because the plot of a play with even a very small cast may require the presence of one or more characters in whose reality, apart from their contribution to the action, we are uninterested.
I wonder, however, whether it is possible to make completely real a wholly villainous character — one toward whom neither the author nor anyone else can feel anything but antipathy. We need an admixture of weakness with either heroic virtue or satanic villainy, to make character plausible. Iago frightens me more than Richard III; I am not sure that Parolles, in All’s Well That Ends Well, does not disturb me more than Iago. And I am quite sure that Rosamond Vincy, in Middlemarch, frightens me far more than Goneril or Regan. It seems to me that what happens, when an author creates a vital character, is a sort of give-and-take. The author may put into that character, besides its other attributes, some trait of his own, some strength or weakness, some tendency to violence or to indecision, some eccentricity even, that he has found in himself. Something perhaps never realized in his own life, something of which those who know him best may be unaware, something not restricted in transmission to characters of the same temperament, the same age, and least of all, to the same sex. Some bit of himself that the author gives to a character may be the germ from which the life of that character starts. On the other hand, a character which succeeds in interesting its author may elicit from the author latent potentialities of his own being. I believe that the author imparts something of himself to his characters, but I also believe that he is influenced by the characters he creates. It would be only too easy to lose oneself in a maze of speculation about the process by which an imaginary character can become as real for us as people we have known. I have penetrated into this maze so far only to indicate the difficulties, the limitations, the fascination, for a poet who is used to writing poetry in his own person, of the problem of making imaginary personages talk poetry. And the difference, the abyss, between writing for the first and for the third voice.
THE peculiarity of my third voice, the voice of poetic drama, is brought out in another way by comparing it with the voice of the poet in nondramatic poetry which has a dramatic element in it. -and conspicuously, in the dramatic monologue. Browning, in an uncritical moment, addressed himself as “ Robert Browning, you writer of plays.” How many of us have read a play by Browning more than once; and, if we have read it more than once, was our motive the expectation of enjoyment? What personage, in a play by Browning, remains living in our mind? On the other hand, who can forget Fra Lippo Lippi, or Andrea del Sarto, or Bishop Blougram, or the other bishop who ordered his tomb? It would seem without further examination, from Browning’s mastery of the dramatic monologue, and his very moderate achievement in the drama, that the two forms must be essentially different. Is there, perhaps, another voice which I have failed to hear, the voice of the dramatic poet whose dramatic gifts are best exercised outside of the theatre? And certainly, if any poetry, not of the stage, deserves to be characterized as “dramatic,” it is Browning’s.
In a play, as I have said, an author must have divided loyalties; he must sympathize with characters who may be in no way sympathetic to each other. And he must allocate the “poetry” as widely as the limitations of each imaginary character permit. This necessity to divide the poetry implies some variation of the style of the poetry according to the character to whom it is given. The fact that a number of characters in a play have claims upon the author, for their allotment of poetic speech, compels him to try to extract the poetry from the character, rather than impose his poetry upon it. Now, in the dramatic monologue we have no such check. The author is just as likely to identify the character with himself, as himself with the character: for the check is missing that will prevent him from doing so — and that check is the necessity for identifying himself with some other character replying to the first. What we normally hear, in fact, in the dramatic monologue, is the voice of the poet, who has put on the costume and make-up either of some historical character, or of one out of fiction. His personage must be identified to us — as an individual, or at least as a type—before he begins to speak. If, as frequently with Browning, the poet is speaking in the role of an historical personage, like Lippo. Lippi, or in the role of a known character of fiction, like Caliban, he has taken possession of that character. And the difference is most evident in his “Caliban upon Setebos.”In The Tempest, it is Caliban who speaks; in “Caliban upon Setebos,”it is Browning’s voice that we hear, Browning talking aloud through Caliban. It was Browning’s greatest disciple, Ezra Pound, who adopted the term “persona" to indicate the several historical characters through whom he spoke: and the term is just.
I risk the generalization also, which may indeed be far too sweeping, that dramatic monologue cannot create a character. For character is created and made real only in an action, a communication between imaginary people. It is not irrelevant that when the dramatic monologue is not put into the mouth of some character already known to the reader—from history or from fiction — we are likely to ask the question “Who was the original?” About Bishop Blougram people have always been impelled to ask, how far was this intended to be a portrait of Cardinal Manning, or of some other ecclesiastic? The poet, speaking, as Browning does, in his own voice, cannot bring a character to life: he can only mimic a character otherwise known to us. And does not the point of mimicry lie in the recognition of the person mimicked, and in the incompleteness of the illusion? We have to be aware that the mimic and the person mimicked are different people: if we tire actually deceived, mimicry becomes impersonation. When we listen to a play by Shakespeare, we listen not to Shakespeare but to his characters; when we read a dramatic monologue by Browning, we cannot suppose that we are listening to any other voice than that of Browning himself.
In the dramatic monologue, then, it is surely the second voice, the voice of the poet talking to other people, that is dominant. The mere fact that he is assuming a role, that he is speaking through a mask, implies the presence of an audience: why should a man put on fancy dress and a mask only to talk to himself? The second voice is, in fact, the voice most often and most clearly heard in poetry that is not of the theatre: in all poetry, certainly, that has a conscious social purpose — poetry intended to amuse or to instruct, poetry that tells a story, poetry that preaches or points a moral, or satire which is a form of preaching. For what is the point of a story without an audience, or of a sermon without a congregation? The voice of the poet addressing other people is the dominant voice of epic, though not the only voice. In Homer, for instance, there is heard also, from time to time, the dramatic voice: there are moments when we hear, not Homer telling us what a hero said, but the voice of the hero himself. The Divine Comedy is not in the exact sense an epic, but here also we hear men and women speaking to us. And we have no reason to suppose that Milton’s sympathy with Satan was so exclusive as to seal him of the Devil’s Party. But the epic is essentially a tale told to an audience, while drama is essentially an action exhibited to an audience.
Now, what about the poetry of the first voice — that which is not primarily an attempt to communicate with anyone at all?
I must make the point that this poetry is not necessarily what wc call loosely “lyric poetry.” The term “lyric” itself is unsatisfactory. We think first of verse intended to be sung — from the songs of Campion and Shakespeare and Burns, to the arias of W. S. Gilbert, or the words of the latest “musical number.” But we apply it also to poetry which was never intended for a musical setting, or which we dissociate from its music: we speak of the “lyric verse “of the metaphysical poets, of Vaughan and Marvell as well as Donne and Herbert. The very definition of “lyric,” in the Oxford Dictionary, indicates that the word cannot be satisfactorily defined:—
Lyric: Now the name for short poems, usually divided into stanzas or strophes, and directly expressing the poet’s own thoughts and sentiments.
How short does a poem have to be, to be called a “lyric”? The emphasis on brevity, and the suggestion of division into stanzas, seem residual from the association of the voice with music. But there is no necessary relation between brevity and the expression of the poet’s own thoughts and feelings. “Come unto these yellow sands” or “Hark! hark! the lark" are lyrics—are they not? — but what sense is there in saying that they express directly the poet’s own thoughts and sentiments? London, The Vanity of Human Wishes, and The Deserted Village are all poems which appear to express the poet’s own thoughts and sentiments, but do we ever think of such poems as “lyrical”? They are certainly not short. Between them, all the poems I have mentioned seem to fail to qualify as lyrics, just as Mr. Daddy Longlegs and Mr. Floppy Fly failed to qualify as courtiers —
Because his legs have grown too short;
The other cannot sing a song,
Because his legs have grown too long!
It is obviously the lyric in the sense of a poem “directly expressing the poet’s own thoughts and sentiments,”not in the quite unrelated sense of a short poem intended to be sot to music, that is relevant to my first voice — the voice of the poet talking to himself — or to nobody. It is in this sense that the German poet Gottfried Bonn, in a very interesting lecture entitled Problems der Lyrik, thinks of lyric as the poetry of the first voice: he includes, I fool sure, such poems as Rilke’s Duinese Elegies and Valery’s La Jeune Parque. Where he speaks of “lyric poetry,”then, I should prefer to say “meditative verse.”
What, asks Herr Benn in this lecture, does the writer of such a poem, “addressed to no one,” start with? There is first, he says, an inert embryo or “creative germ” (ein dumpfer schöpferiacher Keim) and, on the other hand, the Language, the resources of the words at the poet’s command. He has something germinating in him for which he must find words; but he cannot know what words he wants until he has found the words; he cannot identify this embryo until it has been transformed into an arrangement of the right words in the right order. When you have the words for it, the “thing” for which the words had to be found has disappeared, replaced by a poem. What you start from is nothing so definite as an emotion, in any ordinary sense; it is still more certainly not an idea; it is —to adapt two lines of Beddoes to a different meaning - a
Crying with frog voice, “what shall I be?”
I agree with Gottfried Benn, and I would go a little further. In a poem which is neither didactic nor narrative, and not animated by any other social purpose, the poet may be concerned solely with expressing in verse — using all his resources of words, with their history, their connotations, their music — this obscure impulse. He does not know what he has to say until he has said it; and in the effort to say it he is not concerned with making other people understand anything. He is not concerned, at this stage, with other people at all: only with finding the right words or, anyhow, the least wrong words. He is not concerned whether anybody else will ever listen to them or not, or whether anybody else will ever understand them if ho does. He is oppressed by a burden which he must bring to birth in order to obtain relief. Or, to change the figure of speech, he is haunted by a demon, a demon against which he feels powerless, because in its first manifestation it has no face, no name, nothing; and the words, the poem he makes, are a kind of form of exorcism of this demon. In other words again, he is going to all that trouble, not in order to communicate with anyone, but to gain relief from acute discomfort; and when the words are finally arranged in the right way — or in what he comes to accept as the best arrangement he can find—he may experience a moment of exhaustion, of appeasement, of absolution, and of something very near annihilation, which is in itself indescribable. And then he can say to the poem: “Go away! Find a place for yourself in a book — and don’t expect me to take any further interest in you.”
I don’t believe that the relation of a poem to its origins is capable of being more clearly traced. You can read the essays of Paul Valery, who studied the workings of his own mind in the composition of a poem more pcrsevoringly than any other poet has done. But if, either on the basis of what poets try to tell you, or by biographical research, with or without the tools of the psychologist, you attempt to explain a poem, you will probably be getting further and further away from the poem without arriving at any other destination. The attempt to explain the poem by tracing it back to its origins will distract attention from the poem, to direct it on to something else which, in the form in which it can be apprehended by the critic and his readers, has no relation to the poem and throws no light upon it.
I should not like you to think that I am trying to make the writing of a poem more of a mystery than it is. What I am maintaining is, that the first effort of the poet should be to achieve clarity for himself, to assure himself that the poem is the right outcome of the process that has taken place. The most bungling form of obscurity is that of the poet who has not been able to express himself to himself; the shoddiest form is found when the poet is trying to persuade himself that he has something to say when he hasn’t.
SO FAR I have been speaking, for the sake of simplicity, of the three voices as if they were mutually exclusive: as if the poet, in any particular poem, was speaking either to himself or to others, and as if neither of the first two voices was audible in good dramatic verse. And this indeed is the conclusion to which Herr Benn’s argument appears to lead him: he speaks as if the poetry of the first voice — which he considers, moreover, to be on the whole a development of our own age — was a totally different kind of poetry from that of the poet addressing an audience. But for me the voices are most often found together - the first and second, I mean, in non-dramatic poetry; and together with the third in dramatic poetry too. Even though, as I have maintained, the author of a poem may have written it primarily without thought of an audience, he will also want to know what the poem which has satisfied him will have to say to other people. There are, first of all, those few friends to whose criticism he may wish to submit it before considering it completed. They can be very helpful, in suggesting a word or a phrase which the author has not been able to find for himself; though their greatest service perhaps is to say simply “this passage won’t do” — thus confirming a suspicion which the author had been suppressing from his own consciousness. But I am not thinking primarily of the few judicious friends whose opinion the author prizes, but of the larger and unknown audience — people to whom the author’s name means only his poem which they have read. The final handing over, so to speak, of the poem to an unknown audience, for what that audience will make of it, seems to me the consummation of the process begun in solitude and without thought of the audience, the long process of gestation of the poem, because it marks the final separation of the poem from the author. Let the author, at this point, rest in peace.
So much for the poem which is primarily a poem of the first voice. I think that in every poem, from the private meditation to the epic or the drama, there is more than one voice to be heard. If the author never spoke to himself, the result would not be poetry, though it might be magnificent rhetoric; and part of our enjoyment of great poetry is the enjoyment of overhearing words which are not addressed to us. But if the poem were exclusively for the author, it would be a poem in a private and unknown language; and a poem which was a poem only for the author would not be a poem at all. And in poetic drama, I am inclined to believe that all three voices are audible. First, the voice of each character —an individual voice different from that of any other character: so that of each utterance we can say, that it could only have come from that character. There may be from time to time, and perhaps when we least notice it, the voices of the author and the character in unison, saying something appropriate to the character, but something which the author could say for himself also, though the words may not have quite the same meaning for both. That may be a very different thing from the ventriloquism which makes the character only a mouthpiece for the author’s ideas or sentiments.
Is not the perpetual shook and surprise of these hackneyed lines evidence that Shakespeare and Macbeth are uttering the words in unison, though perhaps with somewhat different meaning? And finally there are the lines, in a play by one of the supreme poetic dramatists, in which we hear a more impersonal voice still than that of either the character or the author.
Shall make me live.
And now I should like to return for a moment to Gottfried Benn and his unknown, dark psychic material — we might say, the octopus or angel with which the poet struggles. I suggest that between the three kinds of poetry to which my three voices correspond there is a certain difference of process. In the poem in which the first voice, that of the poet talking to himself, dominates, the “psychic material” tends to create its own form — the eventual form will be to a greater or less degree the form for that one poem and for no other. It is misleading, of course, to speak of the material as creating or imposing its own form: what happens is a simultaneous development of form and material; for the form affects the material at every stage; and perhaps all the material does is to repeat “not that! not that!" in the face of each unsuccessful attempt at formal organization; and finally the material is identified with its form. But in poetry of the second and in that of the third voice, the form is already to some extent given. However much it may be transformed before the poem is finished, it can be represented from the start by an outline or scenario. If I choose to tell a story, I must have some notion of the plot of the story I propose to tell; if I undertake satire, moralizing, or invective, there is already something given which I can recognize and which exists for others as well as myself. And if I set out to write a play, I start by an act of choice: I settle upon a particular emotional situation, out of which characters and a plot will emerge, and I can make a plain prose outline of the play in advance — however much that outline may be altered before the play is finished, by the way in which the characters develop. It is likely, of course, that it is in the beginning the pressure of some rude unknown psychic material that directs the poet to tell that particular story, to develop that particular situation. And on the other hand, the frame, once chosen, within which the author has elected to work, may itself evoke other psychic material; and then, lines of poetry may come into being, not from the original impulse, but from a secondary stimulation of the unconscious mind. All that matters is, that in the end the voices should be heard in harmony; and, as I have said, I doubt whether in any real poem only one voice is audible.
You may well, by now, have been asking yourselves what I have been up to in all these speculations. Have I been toiling to weave a labored web of useless ingenuity? Well, I have been trying to talk, not to myself— as you may have been tempted to think — but to the reader of poetry. I should like to think that it might interest the reader of poetry to test my assertions in his own reading. Can you distinguish these voices in the poetry you read, or hear recited, or hear in the theatre? If you complain that a poet is obscure, and apparently ignoring you, the reader, or that he is speaking only to a limited circle of initiates from which you are excluded— remember that what he may have been trying to do, was to put something into words which could not be said in any other way, and therefore in a language which may be worth the trouble of learning. If you complain that a poet is too rhetorical, and that he addresses you as if you were a public meeting, try to listen for the moments when he is not speaking to you, but merely allowing himself to be overheard: he may be a Dryden, a Pope, or a Byron. And if you have to listen to a verse play, take it first at its face value, as entertainment, for each character speaking for himself with whatever degree of reality his author has been able to endow him. Perhaps, if it is a great play, and you do not try too hard to hear them, you may discern the other voices too. For the work of a great poetic dramatist, like Shakespeare, constitutes a world. Each character speaks for himself, but no other poet could have found those words for him to speak. If you seek for Shakespeare, you will find him only in the characters he created; for the one thing in common between the characters is that no one but Shakespeare could have created any of them. The world of a great poetic dramatist is a world in which the creator is everywhere present, and everywhere hidden.