the expectations of the type of reader to whom it is addressed.
Thus, assuming that good detective fiction must be good fiction in general terms, we may dismiss those qualities which it should possess in common with all other works of imagination and give our attention to those qualities in which it differs from them and which give to it its special character. I have said that the satisfaction which it is designed to yield to the reader is primarily intellectual, and we may now consider in somewhat more detail the exact nature of the satisfaction demanded and the way in which it can best be supplied. And first we may ask: What are the characteristics of the representative reader? To what kind of person is a carefully constructed detective story especially addressed?
We have seen that detective fiction has a wide popularity. The general reader, however, is apt to be uncritical. He reads impartially the bad and the good, with no very clear perception of the difference, at least in the technical construction. The real connoisseurs, who avowedly prefer this type of fiction to all others, and who read it with close and critical attention, are to be found among men of the definitely intellectual class: theologians, scholars, lawyers, and to a less extent, perhaps, doctors and men of science. Judging by the letters which I have received from time to time, the enthusiast par excellence is the clergyman of a studious and scholarly habit.
Now the theologian, the scholar and the lawyer have a common characteristic: they are all men of a subtle type of mind. They find a pleasure in intricate arguments, in dialectical contests, in which the matter to be proved is usually of less consideration than the method of proving it. The pleasure is yielded by the argument itself and tends to be proportionate to the intricacy of the proof. The disputant enjoys the mental exercise, just as a muscular man enjoys particular kinds of physical exertion. But the satisfaction yielded by an argument is dependent upon a strict conformity with logical methods, upon freedom from fallacies of reasoning, and especially upon freedom from any ambiguities as to the data.
By schoolboys, street-corner debaters, and other persons who are ignorant of the principles of discussion, debates are commonly conducted by means of what we may call "argument by assertion." Each disputant seeks to overwhelm his opponent by pelting him with statements of alleged fact, each of which the other disputes, and replies by discharging a volley of counterstatements, the truth of which is promptly denied. Thus the argument collapses in a chaos of conflicting assertions. The method of the skilled dialectician is exactly the opposite of this. He begins by making sure of the matter in dispute and by establishing agreement with his adversary on the fundamental data. Theological arguments are usually based upon propositions admitted as true by both parties; and the arguments of counsel are commonly concerned, not with questions of fact, but with the consequences deducible from evidence admitted equally by both sides.
Thus the intellectual satisfaction of an argument is conditional on the complete establishment of the data. Disputes on questions of fact are of little, if any, intellectual interest; but in any case an argument—an orderly train of reasoning—cannot begin until the data have been clearly set forth and agreed upon by both parties. This very obvious truth is continually lost sight of by authors. Plots, i.e., arguments, are frequently based upon alleged "facts"—physical, chemical, and other—which the educated reader knows to be untrue, and of which the untruth totally invalidates conclusions drawn from them and thus destroys the intellectual interest of the argument.
The other indispensable factor is freedom from fallacies of reasoning. The conclusion must emerge truly and inevitably from the premises; it must be the only possible conclusion, and must leave the competent reader in no doubt as to its unimpeachable truth.
It is here that detective stories most commonly fail. They tend to be pervaded by logical fallacies, and especially by the fallacy of the undistributed middle term. The conclusion reached by the gifted investigator, and offered by him as inevitable, is seen by the reader to be merely one of a number of possible alternatives. The effect when the author's "must have been" has to be corrected by the reader into "might have been" is one of anti-climax. The promised and anticipated demonstration peters out into a mere suggestion; the argument is left in the air and the reader is balked of the intellectual satisfaction which he was seeking.
Having glanced at the nature of the satisfaction sought by the reader, we may now examine the structure of a detective story and observe the means employed to furnish that satisfaction. On the general fictional qualities of such a story we need not enlarge excepting to contest the prevalent belief that detective fiction possesses no such qualities. Apart from a sustained love interest—for which there is usually no room—a detective novel need not, and should not, be inferior in narrative interest or literary workmanship to any other work of fiction. Interests which conflict with the main theme and hinder its clear exposition are evidently inadmissible; but humour, picturesque setting, vivid characterization and even emotional episodes are not only desirable on aesthetic grounds, but, if skilfully used, may be employed to distract the reader's attention at critical moments in place of the nonsensical "false clues" and other exasperating devices by which writers too often seek to confuse the issues. The Mystery of Edwin Drood shows us the superb fictional quality that is possible in a detective story from the hand of a master.
Turning now to the technical side, we note that the plot of a detective novel is, in effect, an argument conducted under the guise of fiction. But it is a peculiar form of argument. The problem having been stated, the data for its solution are presented inconspicuously and in a sequence purposely dislocated so as to conceal their connexion; and the reader's task is to collect the data, to rearrange them in their correct logical sequence and ascertain their relations, when the solution of the problem should at once become obvious. The construction thus tends to fall into four stages: (1) statement of the problem; (2) production of the data for its solution ("clues"); (3) the discovery, i.e., completion of the inquiry by the investigator and declaration by him of the solution; (4) proof of the solution by an exposition of the evidence.
1. The problem is usually concerned with a crime, not because a crime is an attractive subject, but because it forms the most natural occasion for an investigation of the kind required. For the same reason—suitability—crime against the person is more commonly adopted than crime against property; and murder—actual, attempted or suspected—is usually the most suitable of all. For the villain is the player on the other side; and since we want him to be a desperate player, the stakes must be appropriately high. A capital crime gives us an adversary who is playing for his life, and who consequently furnishes the best subject for dramatic treatment.
2. The body of the work should be occupied with the telling of the story, in the course of which the data, or "clues," should be produced as inconspicuously as possible, but clearly and without ambiguity in regard to their essentials. The author should be scrupulously fair in his conduct of the game. Each card as it is played should be set down squarely, face upwards, in full view of the reader. Under no circumstances should there be any deception as to the facts. The reader should be quite clear as to what he may expect as true. In stories of the older type, the middle action is filled out with a succession of false clues and with the fixing of suspicion first on one character, then on another, and again on a third, and so on. The clues are patiently followed, one after another, and found to lead nowhere. There is feverish activity, but no result. All this is wearisome to the reader and is, in my opinion, bad technique. My practice is to avoid false clues entirely and to depend on keeping the reader occupied with the narrative. If the ice should become uncomfortably thin, a dramatic episode will distract the reader's attention and carry him safely over the perilous spot. Devices to confuse and mislead the reader are bad practice. They deaden the interest, and they are quite unnecessary; the reader can always be trusted to mislead himself, no matter how plainly the data are given. Some years ago I devised, as an experiment, an inverted detective story in two parts ["The Case of Oscar Brodski"]. The first part was a minute and detailed description of a crime, setting forth the antecedents, motives, and all attendant circumstances. The reader had seen the crime committed, knew all about the criminal, and was in possession of all the facts. It would have seemed that there was nothing left to tell. But I calculated that the reader
