Probability, Chance and Coincidence
( Originally Published 1912 )
ARISTOTLE indulges in an often-quoted paradox to the effect that, in drama, the probable impossible is to be preferred to the improbable possible. With all respect, this seems to be a some-what cumbrous way of stating the fact that plausibility is of more importance on the stage than what may be called demonstrable probability. There is no time, in the rush of a dramatic action, for a mathematical calculation of the chances for and against a given event, or for experimental proof that such and such a thing can or cannot be done. If a thing seem plausible, an audience will accept it without cavil; if it seem incredible on the face of it, no evidence of its credibility will be of much avail. This is merely a corollary from the fundamental principle that the stage is the realm of appearances, not of realities, where paste jewels are at least as effective as real ones, and a painted forest is far more sylvan than a few wilted and drooping saplings, insecurely planted upon the boards.
That is why an improbable or otherwise inacceptable incident cannot be validly defended on the plea that it actually happened : that it is on record in history or in the newspapers. In the first place, the dramatist can never put it on the stage as it happened. The bare fact may be historical, but it is not the bare fact that matters. The dramatist cannot restore it to its place in that intricate plexus of cause and effect, which is the essence and meaning of reality. He can only give his interpretation of the fact; and one knows not how to calculate the chances that his interpretation may be a false one. But even if this difficulty could be overcome; if the dramatist could prove that he had reproduced the event with photographic and cinematographic accuracy, his position would not thereby be improved. He would still have failed in his peculiar task, which is precisely that of interpretation. Not truth, but verisimilitude, is his aim; for the stage is the realm of appearances, in which intrusive realities become unreal. There are, as I have said, incalculable chances to one that the playwright's version of a given event will not coincide with that of the Recording Angel: but it may be true and convincing in relation to human nature in general, in which case it will belong to the sphere of great art; or, on a lower level, it may be agreeable and entertaining without being conspicuously false to human nature, in which case it will do no harm, since it makes no pretence to historic truth. It may be objected that the sixteenth-century public, and even, in the next century, the great Duke of Marlborough, got their knowledge of English history from Shakespeare, and the other writers of chronicle-plays. Well, I leave it to historians to determine whether this very defective and, in great measure, false vision of the past was better or worse than none. The danger at any rate, if danger there was, is now past and done with. Even our generals no longer go to the theatre or to the First Folio for their history. The dramatist may, with an easy conscience, interpret historic fact in the light of his general insight into human nature, so long as he does not so falsify the recorded event that common knowledge cries out against him.'
Plausibility, then, not abstract or concrete probability, and still less literal faithfulness to recorded fact, is what the dramatist is bound to aim at. To understand this as a belittling of his art is to misunderstand the nature of art in general. The plausibility of bad art is doubtless contemptible and may be harmful. But to say that good art must be plausible is only to say that not every sort of truth, or every aspect of truth, is equally suitable for artistic representation or, in more general terms, that the artist, without prejudice to his allegiance to nature, must respect the conditions of the medium in which he works.
Our standards of plausibility, however, are far from being invariable. To each separate form of art, a different standard is applicable. In what may roughly be called realistic art, the terms plausible and probable are very nearly interchangeable. Where the dramatist appeals to the sanction of our own experience and knowledge, he must not intro-duce matter against which our experience and knowledge cry out. A very small inaccuracy in a picture which is otherwise photographic will often have a very disturbing effect. In plays of society in particular, the criticism " No one does such things," is held by a large class of playgoers to be conclusive and destructive. One has known people despise a play because Lady So-and-so's manner of speaking to her servants was not what they (the cavillers) were accustomed to. On the other hand, one has heard a whole production highly applauded because the buttons on a particular uniform were absolutely right. This merely means that when an effort after literal accuracy is apparent, the attention of the audience seizes on the most trifling details and is apt to magnify their importance. Niceties of language in especial are keenly, and often unjustly, criticized. If a particular expression does not happen to be current in the critic's own circle, he concludes that nobody uses it, and that the author is a pedant or a vulgarian. In view of this inevitable tendency, the prudent dramatist will try to keep out of his dialogue expressions that are peculiar to his own circle, and to use only what may be called everybody's English, or the language undoubtedly current throughout the whole class to which his personage belongs.
It may be here pointed out that there are three different planes on which plausibility may, or may not, be achieved. There is first the purely external plane, which concerns the producer almost as much as the playwright. On this plane we look for plausibility of costume, of manners, of dialect, of general environment. Then we have plausibility of what may be called uncharacteristic event of such events as are independent of the will of the characters, and are not conditioned by their psychology. On this plane we have to deal with chance and accident, coincidence, and all " circumstances over which we have no control." For instance, the playwright who makes the " Marseillaise " become popular throughout Paris within half-an-hour of its having left the composer's desk, is guilty of a breach of plausibility on this plane. So, too, if I were to make my hero enter Parliament for the first time, and rise in a single session to be Prime Minister of England there would be no absolute impossibility in the feat, but it would be a rather gross improbability of the second order. On the third plane we come to psychological plausibility, the plausibility of events dependent mainly or entirely on character. For example to cite a much disputed instance is it plausible that Nora, in A Doll's House, should suddenly develop the mastery of dialectics with which she crushes Hel-mer in the final scene, and should desert her husband and children, slamming the door behind her?
It need scarcely be said that plausibility on the third plane is vastly the most important. A very austere criticism might even call it the one thing worth consideration. But, as a matter of fact, when we speak of plausibility, it is almost always the second plane the plane of uncharacteristic circumstance that we have in mind. To plausibility of the third order we give a more imposing name we call it truth. We say that Nora's action is true or untrue to nature. We speak of the truth with which the madness of Lear, the malignity of Iago, the race-hatred of Shylock, is portrayed. Truth, in fact, is the term which we use in cases where the tests to be applied are those of introspection, intuition, or knowledge sub-consciously garnered from spiritual experience. Where the tests are external, and matters of common knowledge or tangible evidence, we speak of plausibility.
It would be a mistake, however, to imagine that because plausibility of the third degree, or truth, is the noblest attribute of drama, it is therefore the one thing needful. In some forms of drama it is greatly impaired, or absolutely nullified, if plausibility of the second degree, its necessary preliminary, be not carefully secured. In the case above imagined, for instance, of the young politician who should become Prime Minister immediately on entering Parliament : it would matter nothing with what profundity of knowledge or subtlety of skill the character was drawn : we should none the less decline to believe in him. Some dramatists, as a matter of fact, find it much easier to attain truth of character than plausibility of incident. Every one who is in the habit of reading manuscript plays, must have come across the would-be playwright who has a good deal of general ability and a considerable power of characterization, but seems to be congenitally deficient in the sense of external reality, so that the one thing he (or she) can by no means do is to invent or con-duct an action that shall be in the least like any sequence of events in real life. It is naturally difficult to give examples, for the plays composed under this curious limitation are apt to remain in manuscript, or to be produced for one performance, and forgotten. There is, however, one recent play of this order which holds a certain place in dramatic literature. I do not know that Mr. Granville Barker was well-advised in printing The Marrying of Anne Leete along with such immeasurably maturer and saner productions as The Voysey Inheritance and Waste; but by doing so he has served my present purpose in providing me with a perfect example of a play as to which we cannot tell whether it possesses plausibility of the third degree, so absolutely does it lack that plausibility of the second degree which is its indispensable condition precedent.
Francisque Sarcey was fond of insisting that an audience would generally accept without cavil any postulates in reason which an author chose to impose upon it, with regard to events supposed to have occurred before the rise of the curtain; always provided that the consequences deduced from them within the limits of the play were logical, plausible, and entertaining. The public will swallow a camel, he would maintain, in the past, though they will strain at a gnat in the present. A. classical example of this principle is (once more) the Oedipus Rex, in which several of the initial postulates are wildly improbable : for instance, that Oedipus should never have inquired into the circumstances of the death of Laius, and that, having been warned by an oracle that he was doomed to marry his mother, he should not have been careful, before marrying any woman, to ascertain that she was younger than himself. There is at least so much justification for Sarcey's favourite principle, that we are less apt to scrutinize things merely narrated to us than events which take place before our eyes. It is simply a special instance of the well-worn improbability in the antecedents of a play by Ibsen or Sir Arthur Pinero, by Mr. Galsworthy or Mr. Granville Barker, on the plea that it occurred out-side the frame of the picture. Such a plea might, indeed, secure a mitigation of sentence, but never a verdict of acquittal. Sarcey, on the other hand, brought up in the school of the " well-made " play, would rather have held it a feather in the playwright's cap that he should have known just where, and just how, he might safely outrage probability.' The inference is that we now take the dramatist's art more seriously than did the generation of the Second Empire in France.
This brings us, however, to an important fact, which must by no means be overlooked. There is a large class of plays or rather, there are several classes of plays, some of them not at all to be despised the charm of which resides, not in probability, but in ingenious and delightful improbability. I am, of course, not thinking of sheer fantasies, like A Midsummer Night's Dream, or Peter Pan, or The Blue Bird. They may, indeed, possess plausibility of the third order, but plausibility of the second order has no application to them. Its writs do not run on their extramundane plane. The plays which appeal to us in virtue of their pleasant departures from probability are romances, farces, a certain order of light comedies and semi-comic melodramas in short, the thousand and one plays in which the author, without altogether despising and abjuring truth, makes it on principle subsidiary to delightfulness. Plays of the Prisoner of Zenda type would come under this head : so would Sir Arthur Pinero's farces, The Magistrate, The Schoolmistress, Dandy Dick; so would Mr. Carton's light comedies, Lord and Lady Algy, Wheels within Wheels, Lady Huntworth's Experiment; so would most of Mr. Barrie's come-dies; so would Mr. Arnold Bennett's play, The Honeymoon. In a previous chapter I have sketched the opening act of Mr. Carton's Wheels within Wheels, which is a typical example of this style of work. Its charm lies in a subtle, all-pervading improbability, an infusion of fantasy so delicate that, while at no point can one say, " This is impossible," the total effect is far more entertaining than that of any probable sequence of events in real life. The whole atmosphere of such a play should be impregnated with humour, without reaching that gross supersaturation which we find in the lower order of farce plays of the type of Charlie's Aunt or Niobe.
Plausibility of development, as distinct from plausibility of theme or of character, depends very largely on the judicious handling of chance, and the exclusion, or very sparing employment, of coincidence. This is a matter of importance, into which we shall find it worth while to look somewhat closely.
It is not always clearly recognized that chance and coincidence are by no means the same thing. Coincidence is a special and complex form of chance, which ought by no means to be confounded with the every-day variety. We need not here analyse chance, or discuss the philosophic value of the term. It is enough that we all know what we mean by it in common parlance. It may be well, however, to look into the etymology of the two words we are considering. They both come ultimately, from the Latin " cadere," to fall. Chance is a falling-out, like that of a die from the dice-box; and coincidence signifies one falling-out on the top of another, the concurrent happening of two or more chances which resemble or somehow fit into each other. If you rattle six dice in a box and throw them, and they turn up at haphazard say, two aces, a deuce, two fours, and a six there is nothing remarkable in this falling out. But if they all turn up sixes, you at once suspect that the dice are cogged; and if that be not so if there be no sufficient cause behind the phenomenon you say that this identical falling-out of six separate possibilities was a remarkable coincidence. Now, applying the illustration to drama, I should say that the playwright is perfectly justified in letting chance play its probable and even inevitable part in the affairs of his characters; but that, the moment we suspect him of cogging the dice, we feel that he is taking an unfair advantage of us, and our imagination either cries, " I won't play ! " or continues the game under protest.
Some critics have considered it a flaw in Shakespeare's art that the catastrophe of Romeo and Juliet should depend upon a series of chances, and especially on the miscarriage of the Friar's letter to Romeo. This is not, I think, a valid criticism. We may, if we are so minded, pick to pieces the course of action which brought these chances into play. The device of the potion even if such a drug were known to the pharmacopoeia is certainly a very clumsy method of escape from the position in which Juliet is placed by her father's obstinacy. But when once we have accepted that integral part of the legend, the intervention of chance in the catastrophe is entirely natural and probable. Observe that there is no coincidence in the matter, no interlinking or dovetailing of chances. The catastrophe results from the hot-headed impetuosity of all the characters, which so hurries events that there is no time for the elimination of the results of chance. Letters do constantly go astray, even under our highly-organized system of conveyance; but their delay or disappearance seldom leads to tragic results, because most of us have learnt to take things calmly and wait for the next post. Yet if we could survey the world at large, it is highly probable that every day or every hour we should somewhere or other find some Romeo on the verge of committing suicide because of a chance misunderstanding with regard to his Juliet; and in a certain percentage of cases the explanatory letter or telegram would doubtless arrive too late.
We all remember how, in Mr. Hardy's Tess, the main trouble arises from the fact that the letter pushed under Angel Clare's door slips also under the carpet of his room, and so is never discovered. This is an entirely probable chance; and the sternest criticism would hardly call it a flaw in the structure of the fable. But take another case : Madame X. has had a child, of whom she has lost sight for more than twenty years, during which she has lived abroad. She returns to France, and immediately on landing at Bordeaux she kills a man who accompanies her. The court assigns her defence to a young advocate, and this young advocate happens to be her son. We have here a piling of chance upon chance, in which the long arm of coincidence 1 is very apparent. The coincidence would have been less startling had she returned to the place where she left her son and where she believed him to be. But no! she left him in Paris, and it is only by a series of pure chances that he happens to be in Bordeaux, where she happens to land, and happens to shoot a man. For the sake of a certain order of emotional effect, a certain order of audience is willing to accept this piling up of chances; but it relegates the play to a low and childish plane of art. The Oedipus Rex, in-deed which meets us at every turn is founded on an absolutely astounding series of coincidences; but here the conception of fate comes in, and we vaguely figure to ourselves some malignant power deliberately pulling the strings which guide its puppets into such abhorrent tangles. On the mod-ern view that " character is destiny," the conception of supernatural wire-pulling is excluded. It is true that amazing coincidences do occur in life; but when they are invented to serve an artist's purposes, we feel that he is simplifying his task altogether beyond reason, and substituting for normal and probable development an irrelevant plunge into the merely marvellous.
Of the abuse of coincidence, I have already given a specimen in speaking of The Rise of Dick Halward (Chapter XII). One or two more examples may not be out of place. I need not dwell on the significance of the fact that most of them occur in forgotten plays.
In The Man of Forty, by Mr. Walter Frith, we find the following conjuncture of circumstances : Mr. Lewis Dunster has a long-lost wife and a long-lost brother. He has been for years in South Africa; they have meanwhile lived in London, but they do not know each other, and have held no communication. Lewis, returning from Africa, arrives in London. He does not know where to find either wife or brother, and has not the slightest wish to look for them; yet in the first house he goes to, the home of a lady whose acquaintance he chanced to make on the voyage, he encounters both his wife and his brother! Not quite so startling is the coincidence on which Mrs. Willoughby's Kiss, by Mr. Frank Stayton, is founded. An upper and lower flat in West Kensington are inhabited, respectively, by Mrs. Brandram and Mrs. Willoughby, whose husbands have both been many years absent in India. By pure chance the two husbands come home in the same ship; the two wives go to Plymouth to meet them, and by pure chance, for they are totally unacquainted with each other, they go to the same hotel ; whence it hap-pens that Mrs. Willoughby, meeting Mr. Brandram in a half-lighted room, takes him for her husband, flies to his arms and kisses him. More elaborate than either of these is the tangle of coincidences in Mr. Stuart Ogilvie's play, The White Knight
Giulietta, the ward of David Pennycuick, goes to study singing at Milan. Mr. Harry Rook, Pennycuick's most intimate friend, meets her by chance in Milan, and she becomes his mistress, neither having the least idea that the other knows Pennycuick. Then Viscount Hintlesham, like Pennycuick, a dupe of Rook's, meets her by chance at Monte Carlo and falls in love with her. He does not know that she knows Rook or Pennycuick, and she does not know that he knows them. Arriving in England, she finds in the manager, the promoter, and the chairman of the Electric White Lead Company her guardian, her seducer, and her lover. When she comes to see her guardian, the first person she meets is her seducer, and she learns that her lover has just left the house. Up to that moment, I repeat, she did not know that any one of these men knew any other ; yet she does not even say, " How small the world is! " Surely some such observation was obligatory under the circumstances.
Let us turn now to a more memorable piece of work; that interesting play of Sir Arthur Pinero's transition period, The Profligate. Here the great situation of the third act is brought about by a chain of coincidences which would be utterly unthinkable in the author's maturer work. Leslie Brudenell, the heroine, is the ward of Mr. Cheal, a solicitor. She is to be married to Dunstan Renshaw; and, as she has no home, the bridal party meets at Mr. Cheal's office before proceeding to the registrar's. No sooner have they departed than Janet Preece, who has been betrayed and deserted by Dunstan Renshaw (under an assumed name) comes to the office to state her piteous case. This is not in itself a pure coincidence; for Janet happened to come to London in the same train with Leslie Brudenell and her brother Wilfrid; and Wilfrid, seeing in her a damsel in distress, recommended her to lay her troubles before a respectable solicitor, giving her Mr. Cheal's address. So far, then, the coincidence is not startling. It is natural enough that Renshaw's mistress and his betrothed should live in the same country town; and it is not improbable that they should come to London by the same train, and that Wilfrid Brudenell should give the bewildered and weeping young woman a commonplace piece of advice. The concatenation of circumstances is remarkable rather than improbable. But when, in the next act, not a month later, Janet Preece, by pure chance, drops in at the Florentine villa where Renshaw and Leslie are spending their honeymoon, we feel that the long arm of coincidence is stretched to its utter-most, and that even the thrilling situation which follows is very dearly bought. It would not have been difficult to attenuate the coincidence. What has actually happened is this : Janet has (we know not how) become a sort of maid-companion to a Mrs. Stonehay, whose daughter was a school-friend of Leslie's; the Stonehays have come to Florence, knowing nothing of Leslie's presence there; and they happen to visit the villa in order to see a fresco which it contains. If, now, we had been told that Janet's engagement by the Stone-hays had resulted from her visit to Mr. Cheal,.. and that the Stonehays had come to Florence knowing Leslie to be there, and eager to find her, several links would have been struck off the chain of coincidence; or, to put it more exactly, a fairly coherent sequence of events would have been substituted for a series of incoherent chances. The same result might no doubt have been achieved in many other and neater ways. I merely indicate, by way of illustration, a quite obvious method of reducing the element of coincidence in the case.
The coincidence in The Second Mrs. Tanqueray, by which Ellean meets and falls in love with one of Paula's ex-lovers, has been very severely criticized. It is certainly not one of the strong points of the play; but, unlike the series of chances we have just been examining, it places no excessive strain on our credulity. Such coincidences do occur in real life; we have all of us seen or heard of them; the worst we can say of this one is that it is neither positively good nor positively bad a piece of indifferent craftsmanship. On the other hand, if we turn to Letty, the chance which, in the third act, leads Letchmere's party and Mandeville's party to choose the same restaurant, seems to me entirely justified. It is not really a coincidence at all, but one of those everyday happenings which are not only admissible in drama, but positively desirable, as part of the ordinary surface-texture of life. Entirely to eliminate chance from our representation of life would be a very unreason-able austerity. Strictly speaking, indeed, it is impossible; for even when we have worked out an unbroken chain of rational and commensurate causes and effects, it remains a chance, and an unlikely chance, that chance should not have interfered with it.
All the plays touched upon in the last four para-graphs are in intention realistic. They aim, that is to say, at a literal and sober representation of life. In the other class of plays, which seek their effect, not in plodding probability, but in delightful improbability, the long arm of coincidence has its legitimate functions. Yet even here it is not quite unfettered. One of the most agreeable coincidences in fiction, I take it, is the simultaneous arrival in Bagdad, from different quarters of the globe, of three one-eyed calenders, all blind of the right eye, and all, in reality, the sons of kings. But it is to be noted that this coincidence is not a crucial occurrence in a story, but only a part of the story-teller's framework or mechanism a device for introducing fresh series of adventures. This illustrates the Sarceyan principle above referred to, which Professor Brander Matthews has restated in what seems to me an entirely acceptable form --namely, that improbabilities which may be admitted on the outskirts of an action, must be rigidly excluded when the issue is joined and we are in the thick of things. Coincidences, in fact, become the more improbable in the direct ratio of their importance. We have all, in our own experience, met with amazing coincidences; but how few of us have ever gained or lost, been made happy or unhappy, by a coincidence, as distinct from a chance ! It is not precisely probable that three brothers, who have separated in early life, and have not heard of one another for twenty years, should find them-selves seated side by side at an Italian table-d'hτte; yet such coincidences have occurred, and are credible enough so long as nothing particular comes of them. But if a dramatist were to make these three brothers meet in Messina on the eve of the earthquake, in order that they might all be killed, and thus enable his hero (their cousin) to succeed to a peerage and marry the heroine, we should say that his use of coincidence was not strictly artistic. A coincidence, in short, which coincides with a crisis is thereby raised to the nth power, and is wholly inacceptable in serious art. Mr. Bernard Shaw has based the action of You Never Can Tell on the amazing coincidence that Mrs. Clandon and her children, coming to England after eighteen years' absence, should by pure chance run straight into the arms, or rather into the teeth, of the husband and father whom the mother, at any rate, only wishes to avoid. This is no bad starting-point for an extravaganza; but even Mr. Shaw, though a despiser of niceties of craftsmanship, introduces no coincidences into serious plays such as Candida or The Doctor's Dilemma.