Some thing off topic but useful for story writers

How To Write Detective Stories: Story Writing and Plot

How to Write a Detective Story

From <G. K.'s Weekly>, October 17, 1925

Reprinted in <The Spice of Life>

Let it be understood that I write this article as one wholly conscious that he has failed to write a detective story. But I have failed a good many times. My authority is therefore practical and scientific, like that of some great statesman or social thinker dealing with Unemployment or the Housing Problem. I do not pretend that I have achieved the ideal that I set up here for the young student; I am, if you will, rather the awful example for him to avoid. None the less I believe that there are ideals of detective writings, as of everything else worth doing; and I wonder they are not more often set out in all that popular didactic literature which teaches us how to do so many things so much less worth doing; as, for instance, how to succeed. Indeed, I wonder very much that the title at the top of this article does not stare at us from every bookstall. Pamphlets are published teaching people all sorts of things that cannot possibly be learnt, such as personality, popularity, poetry, and charm. Even those parts of literature and journalism that most obviously cannot be learnt are assiduously taught. But here is a piece of plain straightforward literary craftsmanship, constructive rather than creative, which could

to some limited extent be taught and even, in very lucky instances, learnt. Sooner or later I suppose the want will be supplied, in that commercial system in which supply immediately answers to demand, and in which everybody seems to be thoroughly dissatisfied and unable to get anything he wants. Sooner or later, I suppose, there will

not only be text-books teaching criminal investigators, but text-books teaching criminals. It will be but a slight change from the present tone of financial ethics, and when the shrewd and vigorous business mind has broken away from the last lingering influence of

dogmas invented by priests, journalism and advertisement will show the same indifference to the taboos of today as does today to the taboos of the Middle Ages. Burglary will be explained like usury, and there will be no more disguise about cutting throats than there is about cornering markets. The bookstalls will be brightened with titles

like 'Forgery in Fifteen Lessons,' and 'Why Endure Married Misery?' with a popularization of poisoning fully as scientific as the popularization of Divorce and Birth-Control.

But, as we are so often reminded, we must not be in a hurry for the arrival of a happy humanity; and meanwhile, we seem to be quite as likely to get good advice about committing crimes as good advice about detecting them, or about describing how they could be detected. I imagine the explanation is that the crime, the detection, the

description, and the description of the description, do all demand a certain slight element of thought, while succeeding and writing a book on success in no way necessitate this tiresome experience. Anyhow, I find in my own case that when I begin to think of the

theory of detective stories, I do become what some would call theoretical. That is, I begin at the beginning, without any pep, snap, zip or other essential of the art of arresting the attention, without in any way disturbing or awakening the mind.

The first and fundamental principle is that the aim of a mystery story, as of every other story and every other mystery, is not darkness but light. The story is written for the moment when the reader does understand, not merely for the many preliminary moments when he does not understand. The misunderstanding is only meant as a dark outline of cloud to bring out the brightness of that instant of intelligibility; and most bad detective stories are bad because they fail upon this point. The writers have a strange notion that it is their business to baffle the reader; and that so long as they baffle him it does not matter if they disappoint him. But it is not only necessary to hide a secret, it is also necessary to have a secret; and to have a secret worth hiding. The climax must not be an anti-climax; it must not merely consist of leading the reader a dance and leaving him in a ditch. The climax must not be only the bursting of a bubble but rather the breaking of a dawn; only that the daybreak is accentuated by the dark. Any form of art, however trivial, refers back to some serious truths; and though we are dealing with nothing more momentous than

a mob of Watsons, all watching with round eyes like owls, it is still permissible to insist that it Is the people who sat in darkness who have seen a great light; and that the darkness is only valuable in making vivid a great light in the mind. It always struck me as an

amusing coincidence that the best of the Sherlock Holmes stories bore, with a totally different application and significance, a title that might have been invented to express this primal illumination; the title of "Silver Blaze"

The second great principle is that the soul of detective fiction is not complexity but simplicity. The secret may appear complex, but it must be simple; and in this also it is a symbol of higher mysteries. The writer is there to explain the mystery; but he ought not to be needed to explain the explanation. The explanation should explain itself; it should be something that can be hissed (by the villain, of course) in a few whispered words or shrieked preferably by the heroine before she swoons under the shock of the belated realization that two and two make four. Now some literary detectives make the solution more complicated than the mystery, and the crime more complicated than the solution.

Thirdly, it follows that so far as possible the fact or figure explaining everything should be a familiar fact or figure. The criminal should be in the foreground, not in the capacity of criminal, but in some other capacity which nevertheless gives him a natural right to be

in the foreground. I will take as a convenient case the one I have already quoted; the story of Silver Blaze. Sherlock Holmes is as familiar as Shakespeare; so there is no injustice by this time in letting out the secret of one of the first of these famous tales. News is brought to Sherlock Holmes that a valuable race-horse has been stolen, and the trainer guarding him murdered by the thief. Various people, of course, are plausibly suspected of the theft and murder; and everybody concentrates on the serious police problem of who can have killed the trainer. The simple truth is that the horse killed him. Now I take that as a model because the truth is so very simple. The truth really is so very obvious.

At any rate, the point is that the horse is very obvious. The story is named after the horse; it is all about the horse; the horse is in the foreground all the time, but always in another capacity. As a thing of great value he remains for the reader the Favourite; it is only as a

criminal that he is a dark horse. It is a story of theft in which the horse plays the part of the jewel until we forget that the jewel can also play the part of the weapon. That is one of the first rules I would suggest, if I had to make rules for this form of composition. Generally speaking, the agent should be a familiar figure in an unfamiliar function. The thing that we realize must be a thing that we recognize; that is it must be something previously known, and it ought to be something prominently displayed. Otherwise there is no surprise in mere novelty. It is useless for a thing to be unexpected if it was not worth expecting. But it should be prominent for one reason and responsible for another. A great part of the craft or trick of writing mystery stories consists in finding a convincing but misleading reason for the prominence of the criminal, over and above his legitimate business of committing the crime. Many mysteries fail merely by leaving him at loose ends in the story, with apparently nothing to do except to commit the crime. He is generally well off, or our just and equal law would probably have him arrested as a vagrant long before he was arrested as a murderer. We reach the stage of suspecting such

a character by a very rapid if unconscious process of elimination. Generally we suspect him merely because he has not been suspected. The art of narrative consists in convincing the reader for a time, not only that the character might have come on the premises with no

intention to commit a felony, but that the author has put him there with some intention that is not felonious. For the detective story is only a game; and in that game the reader is not really wrestling with the criminal but with the author.

What the writer has to remember, in this sort of game, is that the reader will not say, as he sometimes might of a serious or realistic study: "Why <did> the surveyor in green spectacles climb the tree to look into the lady doctor's back garden?" He will insensibly and inevitably say, "Why did the author <make> the surveyor climb a tree, or introduce any surveyor at all?" The reader may admit that the town would in any case need a surveyor, without admitting that the tale would in any case need one. It is necessary to explain his presence in the tale (and the tree) not only by suggesting why the town council put him there, but why the author put him there. Over and above any little crimes he may intend to indulge in, in the inner chamber of the story, he must have already some other justification as a character in a story and not only as a mere miserable material person in real life. The instinct of the reader, playing hide-and-seek with the writer, who is his real enemy, is always to say with suspicion, Yes, I know a surveyor might climb a tree; I am quite aware that there are trees and that there are surveyors, but what are you doing with them? Why did you make this particular surveyor climb this particular tree in this particular tale, you cunning and evil-minded man?"

This I should call the fourth principle to be remembered, as in the other cases, people probably will not realize that it is practical, because the principles on which it rests sound theoretical. It rests on the fact that in the classification of the arts, mysterious murders

belong to the grand and joyful company of the things called jokes. The story is a fancy; an avowedly fictitious fiction. We may say if we like that it is a very artificial form of art. I should prefer to say that it is professedly a toy, a thing that children 'pretend' wish. From this it follows that the reader, who is a simple child and therefore very wide awake, is conscious not only of the toy but of the invisible playmate who is the maker of the toy, and the author of the trick. The innocent child is very sharp and not a little suspicious. And one of the first rules I repeat, for the maker of a tale that shall be a trick, is to

remember that the masked murderer must have an artistic right to be on the scene and not merely a realistic right to be in the world. He must not only come to the house on business, but on the business of the story; it is not only a question of the motive of the visitor but of the motive of the author. The ideal mystery story is one in which he is such a character as the author would have created for his own sake, or for the sake of making the story move in other necessary matters, and then be found to be present there, not for the obvious and sufficient reason, but for a second and a secret one. I will add that for

this reason, despite the sneers at 'love-interest' there is a good deal to be said for the tradition of sentiment and slower or more Victorian narration. Some may call it a bore, but it may succeed as a blind.

Lastly the principle that the detective story like every literary form starts with an idea, and does not merely start out to find one, applies also to its more material mechanical detail. Where the story turns upon detection, it is still necessary that the writer should begin from the inside, though the detective approaches from the outside. Every good problem of this type originates in a positive notion, which is in itself a simple notion; some fact of daily life that the writer can remember and the reader can forget. But anyhow, a tale has to be founded on a truth; and though opium may be added to it, it must not merely be an opium dream.

The criminal must be someone mentioned in the early part of the story, but must not be anyone whose thoughts the reader has been allowed to follow. All supernatural or preternatural agencies are ruled out as a matter of course. Not more than one secret room or passage is allowable. No hitherto undiscovered poisons may be used, nor any appliance which will need a long scientific explanation at the end. No Chinaman must figure in the story. No accident must ever help the detective, nor must he ever have an unaccountable intuition which proves to be right. The detective must not himself commit the crime. The detective must not light on any clues which are not instantly produced for the inspection of the reader. The stupid friend of the detective, the Watson, must not conceal any thoughts which pass through his mind; his intelligence must be slightly, but very slightly, below that of the average reader.
• Twin brothers, and doubles generally, must not appear unless we have been duly prepared for them.

PS: Here is the oath, composed by G. K. Chesterton, of membership in the famous British Detection Club: "Do you promise that your detectives shall well and truly detect the crimes presented to them using those wits which it may please you to bestow upon them and not placing reliance on nor making use of Divine Revelation, Feminine Intuition, Mumbo Jumbo, Jiggery-Pokery, Coincidence, or Act of God?"

The criminal must be somebody mentioned in the story. (This is absolutely essential, otherwise the book cannot be called a detective story. The other bit about 'sharing thoughts' is too strict, but a writer should still be cautious because an outright authorial deception must be avoided.) Supernatural elements are allowable for atmospheric or plot reasons, but they must play no part in the actual solution of the mystery. Secret passages or hidden rooms are all right (if the setting allows it), but do not deserve to be used as an explanation of the murder method. Avoid unknown Amazonian arrow poisons or newly invented Death-Ray machines, unless as an author you are qualified (scientifically) to justify it (i.e., if Newton had written a mystery based on his laws of Optics, that would be OK, but don't presume to invent a poison if you don't even know that aspirin can be fatal.) Do not use 'foreigners' or other aliens as major characters unless you have some real understanding of their culture and mind-set, and they have some relevance to the plot beyond exotic obfuscation. Avoid accidental solutions, as they are hardly fair in a story of deduction and the presentation of real clues. And please do not inflict on the poor reader one of those mid-book "Mon dieu, how could I not have seen that before" exclamations which sit like undigested food until the end of the mystery. The criminal should not be someone you have intentionally presented as totally trustworthy. (If he/she is a liar, at least provide some clue to give the reader a chance to spot that.) All clues must be revealed, although it is perfectly legitimate to disguise them. (But I would draw the line at basing a clue on some misspelling of a word, American vs. British usage, for example, because most books are hardly proofread any more.) There should but doesn't have to be a 'Watson' or some observing point of view that sees but misinterprets the events under investigation. (Only common sense, otherwise where is the drama?)
• Do not try to fool the reader with improbable impersonations, such as a woman posing as a man or vice versa and getting away with it by consummate acting ability, especially when they are deceiving people who know them well. (This doesn't even work in Shakespeare.) Especially avoid wigs and false whiskers!

Do not try to confuse the reader with elaborate timetables based on train schedules, etc., as there is no guarantee that things like that would ever work out for even the carefullest murderer. (Sod's or Murphy's Law.) Avoid having your Prime Suspect turn out to be the culprit after all, because this is ultimately disappointing (unless you are clever enough to totally reshuffle motives and alibis). Do not present an 'impossible crime' situation without at least attempting to verify its plausibility by experiment. Also try to avoid using an accomplice to abet the criminal's illusion. (That's OK for stage magicians with their assistants, but spoils a mystery plot where the villain has to deceive the detective, almost, but without cheating. It makes a lot of sense, too, if you are a villain, not to risk collaboration.) The murderer should never turn out to be somebody incapable of committing the crime, at least as presented in the lead-up (i.e., invalids in wheelchairs, morons, a person in an intensive-care ward, an astronaut who happened to be in orbit at the time). A conspiracy involving a hired hit-man, or a mysterious Illuminati cartel, does not belong in a true detective novel. This also includes situations where several suspects are independently up to no good and just happen to be on the scene at the relevant time. (Sod's Law, again, and a very mechanical manipulation of coincidence for supposedly dramatic purposes -- this won't fool anybody and should be dismissed as mere padding.) No faking of fingerprints or other forensic details. In spite of their portrayal, even the police a hundred years ago were not as incompetent as they were made out to be. Nowadays, if you want to commit a murder, forget trying such a thing, unless you can afford a good lawyer to screw up the expert witnesses at your trial! If you are going to talk down to the reader (who is an ignoramus, whereas you are a genius), via your detective, make sure your facts are correct. Twaddle about Egyptology (curse of the pharaoh, etc.) is unacceptable. Informative facts about some obscure subject, however, are beneficial. Do not present your detective as an ineffectual fool or allow him or her to show any signs of not being superior to the reader or the 'Watson' (except to the extent that the detective can have misjudgements and miscalculations for the sake of 'bonding' with the reader). An incompetent detective is an actor in a comedy, not a detective story. Get your details of real police policies and forensic science up to date as far as you can. Unless the book takes place in the classic stranded house-party tradition, there is no way an author can get away with ignoring public procedures, no matter how gifted the detective.
• Finally, a personal peeve: Don't have a large cast of characters and refer to them all by their Christian names, such as Evelyn, Jane, Meg, Charles, and Chris. Who in the hell are you talking about?

Structure and Sources: Mainly keep in mind that the plot comes first and that the actions of the characters are 'retrofitted' into it, which is how a detective story differs from a crime novel where of course the characters themselves drive the plot. Any central pivot, such as an expertise about some unusual subject, is up to the author -- as long as it is accurate. The Need for Unity: In other words, make the story fit the devised crime. A person -- detective, suspect, witness -- should not act out of 'character' just because the plot demands it. In that case, it is better just to redesign the character. The Detective: This is almost axiomatic -- one must have a detective who is distinctively defined, preferably a series detective (which saves having to create a new one for each book -- easier that way both for the writer and the reader and engenders a familiarity that ensures comfort and a market for new books). Initially defining a detective whom readers can 'identify' with as a familiar friend is one of the hardest things, apart from plotting, for a detective novelist to do, but once done removes the burden of re-explication. Watson or Not?: Discouraged now because it is trite, but if you have to have one, make him a total opposite of the detective -- e.g., Archie Goodwin vs Nero Wolfe. Viewpoint: Standard literary practice, whatever the genre. There has to be a consistency of delivery for the story, no matter what technique is used (first person, omniscient, point-of-view, whatever). The Crime: There really must be a murder, or at least a major felony -- otherwise, what's the point? Who's ripping off the hand towels at the Dorchester Hotel is hardly the business of a mystery novel. The Title: "The best advice to the author faced with the selection of a title is not to worry about it." Having a good title and basing the book on it is like the tail wagging the dog. 'Nuff said. The Plot: Keep it flowing from one thing to another and don't get sidetracked into dead-ends. Well, that's common sense for all fiction. "Had I But Known": That has always been a bugaboo among mystery fans since Mary Roberts Rinehart and earlier. It is the ditzy heroine sneaking into Bluebeard's chamber even if she has been repeatedly warned not to. It is the kid's action when told "Whatever you do, don't climb on the railway tressle." Does not belong in a detective story. (Can be fun enough in a gothic romance or horror novel or something of the sort.) Emotion and Drama: Of course for dramatic reasons there has to be some of this for the sake of interesting the reader, but for the most part remember that this is a novel of detection, not a love story. The Puzzle Element: Don't make that the whole story; this is not a crossword puzzle. Background and Setting: Basically, the author should be familiar personally with the location. If you were in Aruba for three hours on a cruise ship trip, don't set your novel in Aruba based on that. Use real settings when possible, for verisimilitude, and be accurate. And, PS, don't borrow somebody else's setting, such as Wuthering Heights. Characters and Characterization: Not all of the players need to be fully defined -- puppet roles are fine (cops, servants, etc.) -- but at least the detective, the murderer, and preferably the victim should be convincingly realized. Is this obvious or what? But a lot of formula mysteries totally ignore this precept. Style: Avoid corniness, pretentiousness, and overwriting. (Duh...) The Devices of Detection: Don't be so elaborate as to make the dnouement incomprehensible. Beware of ignorance of the simple rules of evidence and forensics. (Then follows a whole list of things to avoid, like tobacco ashes, locked rooms, footprints, etc., but that is just HH's judgment based on what were clichs then. If it works, then it's OK, right?) Physical Boundaries: This is basically advice on how long a mystery novel or story should be. Times change -- sometimes very lengthy, sometimes very short, now lengthy again (because of the high cover cost of a book these days -- padding out an extra couple hundred pages, which isn't that more expensive production-wise, makes the reader think it's worth the money).
• Some General Considerations: Basically extols the existence of bodies like The Detection Club in England, which encouraged new ventures in this genre, and was a professional forum for both established and hopeful writers. MWA encourages this now in the US (but not so much back then).

By lisaj66

How to write mystery stories