A blank page is frightening. Something has to be written, but how do you choose the words? Why this word and not that? How to overcome the arbitrariness of writing? One way is to trick yourself, to pretend that what you’re about to write has already been written, that someone has been there already and that you’re only following his traces, trying to reconstruct what must have happened. This is the idea of detective fiction: you begin with an unexplained corpse or an impossible theft; following the clues should tell you what happened, and then how and why. It’s a nice solution: the writer is the detective, presented with a small group of people he doesn’t know; he has to find out who they are, discover their histories and motives. By the time the story is over, he knows everything. At this point another book might begin.
If writers of detective fiction cast themselves as detectives, this isn’t how it seems to their readers, who prefer to think of them as criminals. The writer, thanks to his trickery, creates the illusion that he knew all along who did it. He chooses what information to disclose and when, ideally operating in such a way as to make the puzzle possible in theory for a reader to solve, though not in practice, so that by the end the reader feels hoodwinked but not cheated. This isn’t an easy thing to achieve, but leaving an audience happily hoodwinked is the only conceivable reward for the conman – apart from the money. The trouble is that audiences are rarely happy. Magicians know not to tell you how they did it: writers of detective fiction, on the other hand, have to finish with an ingenious explanation; once the solution has been presented to you, a puzzle ceases to have any point, and you can only begin another. Mystery writers have no choice but to become career criminals. The worst of it is that by leaving the clues that allow the detective to solve the mystery, they’re constantly hanging themselves. This isn’t normal criminal behaviour.
Everyone wants to be the detective: the criminal can’t win, and the detective is smarter. But in the purest detective stories, the most crossword-like, the writer can’t help being a bit of both. This is the puzzle-setter’s game: he has to be able to solve the problem as well as pose it; he has to follow the steps his followers might take, making sure they have just enough information to reach the solution and no more, unless it’s false information designed to lead them the wrong way. Puzzles have rules and conventions: setter and solver have to talk the same language. In these stories, there is usually a corpse; there are clues, alibis and suspects. Since the puzzle has to be soluble, the criminal must be someone known to the detective, not an outsider. Supernatural explanations are prohibited. All the suspects have to be present, so the setting must be circumscribed. Agatha Christie preferred small villages and house parties, but almost anything would do: an island, a plane, a train. G.K. Chesterton’s settings in his Father Brown stories are stranger and simpler: a beheading in a walled garden with no exits, the theft of all the silver of an elite dining club during a meal. He almost removes the need for motive: the perfect criminal commits the perfect crime for its beauty, for the perfection of the puzzle.
Classic English detective fiction suffered for its neatness. It was implausible, frustrating, dull. (Or even nightmarishly addictive, which is my problem even with Chesterton: a much more dangerous thing.) The person usually credited with finding a way out is Dashiell Hammett. Raymond Chandler, Hammett’s more sophisticated successor, made the claim most publicly in the Atlantic Monthly in 1944 in an essay that has come to be seen as his manifesto: ‘Hammett gave murder back to the kind of people that commit it for reasons, not just to provide a corpse; and with the means at hand, not with hand-wrought duelling pistols, curare and tropical fish. He took murder out of the Venetian vase and dropped it into the alley.’ Hammett’s detective-heroes, like Chandler’s, are private eyes, hired to investigate (usually) professional killings in the crooked ganglands of West Coast America. On their way to fingering the culprit, they are threatened, knocked out with blackjacks, tied up and beaten; but they come through in the end and find their man, or (often) woman. There are insurance scams, political murders, altered identities. The new hard-boiled detectives had nothing to do with the nice gentility of their sedentary English competitors.
But Chandler was underselling himself and Hammett: they did more than change the backdrop. They also changed the way the detective story worked – by removing most of the detecting. Take ‘House Dick’, now rescued from the yellowing pages and lurid covers of Black Mask to appear in Nightmare Town, a collection of stories spanning Hammett’s brief career. The narrator, known only as the Continental Op, is filling in for the Montgomery Hotel detective for a day when a maid opens a closet in one of the rooms and three bodies fall out. It’s a promising beginning, but the fingerprints all belong to the maid and there’s no evident motive. Two of the dead men were seen in the lobby at lunchtime discussing golf. Without any clues, the detective has to rely on legwork. So he and his fellow operatives tail all the guests, and find that one, called Orrett, packs a gun and has been making enquiries in some seedy dives after a man with a scar on his cheek called Cudner, a local hood. The Op makes his move, figuring that with a little face-paint Orrett might take him for Cudner, and he might get something out of him. He does: Orrett sees him in a bar and pulls out a gun. ‘I had, I realised, made a mistake – one that might cost me something before we were done. This bird wasn’t hunting for Cudner as a friend, as I had carelessly assumed, but was on the warpath.’ The shooting begins, and the real Cudner appears; Orrett comes off better. ‘Just what there was between them and what bearing it had on the Montgomery murders was a mystery to me, but I didn’t try to solve it now.’ He doesn’t have to. Orrett, in hospital, explains: Cudner was after him and he was after Cudner; Cudner tried to kill him in the Montgomery but got the wrong room. Poirot would never have involved himself in rough work like this, but if he had he’d have been sure to work it all out himself. Hammett’s detectives are only human; and they are probably more likable for it. The interest is in the hell they go through, in what happens next. Along their way they almost leave behind what was most important to the classic English detective story: the past, and the puzzle. In ‘House Dick’, the three accidental corpses are a sideshow; and the puzzle is solved by asking the manufacturer, thus pre-empting potential frustration.
This approach clearly suits readers who don’t like writers with subtle criminal minds. The plain man’s detective thriller, these readers might say, moves forwards. More finicky detective fiction moves backwards, which is a) less exciting and b) bloody-minded. The extreme would be a story like Conan Doyle’s ‘The Adventure of the Blue Carbuncle’, which opens with a diamond found in the craw of a Christmas goose. Holmes begins at the end and works towards the beginning, following in reverse the steps the goose and its prize must have taken, all the way back to the original theft. It’s very plodding. But the peculiarity of the hard-boiled detective thriller is that it doesn’t entirely do away with the puzzle. And the most interesting thing isn’t always what happens next: suspense on its own is as tiring as ingenious detecting. I think Hammett’s real invention is something stranger. It’s all in the shadowing. A story that particularly depends on it is ‘Zigzags of Treachery’. Dr Estep has been shot; his wife has been arrested; the day before the murder a woman had appeared on the scene claiming to be the real Mrs Estep. The death isn’t the primary puzzle. The Op follows the would-be wife (gold-digger or blackmailer?), giving us a brief lesson in the art of tailing on the way, watches the apartment of an associate of hers, sees people passing in and passing out, one of them supported – drunk, or dead? The plot only gives us half the story: we see the exits and the entrances, the in-between places, and wonder what’s happening on stage. The half we didn’t see comes at the end; appearances were misleading. The puzzle, for the shadow, is not what has happened or what will happen, but what’s happening now, on the other side of the closed door. There’s an aspect of things that is usually hidden: it’s their real face.
Hammett was a private detective for three years, with the Pinkerton agency in Philadelphia. When the First World War started, he enlisted, and was stationed with an ambulance company – in Maryland. He had an unlucky war, and fell ill; he never entirely recovered, and when he left hospital, after marrying his nurse, Josephine Dolan, he couldn’t go back to Pinkerton’s. He tried journalism, but found there was more money in mystery magazines, and he became a regular, in particular for Black Mask, the most popular and successful of the pulps. He worked fast, and his audiences loved him. In 1927, he wrote to A.A. Knopf proposing a novel which was to become Red Harvest. They liked it, so he wrote to Blanche Knopf again in March 1928: ‘I’ve another book-length detective story – tentatively entitled The Dain Curse – underway, using the same detective I used in this book, but not using him so violently. I hope to finish that this month. The first serial rights have been sold to the Black Mask.’ Hammett wasn’t quite coming clean: The Dain Curse isn’t a book-length detective story, but three pulp stories loosely tied together, with the same narrator, everyone’s favourite Continental Op, and a series of connected crimes. He was just being persuasive. But he stopped pretending when his publishers got to know him. He wrote to one of the Knopf editors in June 1929, asking: ‘How soon will you want, or can you use, another book? I’ve quite a flock of them outlined or begun, and I’ve a couple of groups of connected stories that can be joined in a whole just as I did with Red Harvest and The Dain Curse.’ He tried for all he could get as long as he could get it quickly: he even suggested a collection of short stories, to be called ‘The Continental Op’. He had 250,000 words of usable material. They didn’t fall for the collection idea at first, as a good publisher shouldn’t. Seven stories came out under that title in 1945, but by then they couldn’t hold out: Hammett was a star. What he produced didn’t matter, so long as it was taken: it kept the ball rolling and allowed him to send cheques to Josephine and the kids; he didn’t see them much. In another letter from 1929 to his editor at Knopf, he wrote: ‘I’ve a title I like – “The Glass Key” – and at least part of a plot to go with it. I’ll probably get going on it next week.’ Next week? This kind of efficiency and carelessness would be decent if his novels had disappeared. But they haven’t.
Good detective stories can’t disappear. This is the real mystery about detective fiction: why is it remembered when it wasn’t written to be? Perhaps it’s the idea of it that is remembered, rather than the sometimes less than satisfying fact. The idea of detecting, of following clues to find out the facts, has given rise to all sorts of clever, self-reflexive non-genre writing. Borges was an addict who wrote pastiches in which the detective works from a prison cell; he also played with the idea of reconstructing what was already there: his Pierre Menard begins rewriting Don Quixote word for word, and his version turns out to be better than the original. Paul Auster, in the New York Trilogy, follows the logic of detection into dizzy madness. But these clevernesses aren’t just sparked by the idea: they are affected by detective fiction’s glamour. And much of this glamour comes from film, and most of that from Bogart. Hammett, it seems to me, invented Bogart; at least, the detective who made Hammett, Sam Spade of The Maltese Falcon (1929), is more like Bogart than even Bogart could manage to be. ‘Think again and think fast. I told that punk of yours that you’d have to talk to me before you got through. I’ll tell you now that you’ll do your talking today or you are through. What are you wasting your time for? You and your lousy secret!’ And he doesn’t let up. I used to think that stardom was in the pauses – in that sense of restraint, of contained explosion – or somewhere in the eyes, but Bogart goes with the explosion, with urgent, threatening breathlessness. Before scriptwriters began taking out the words to let the images do the talking, the stars had to do the work themselves, talking fast and tough.
The original Spade looks nothing like he does in the movie, but that’s because he doesn’t look like anything on earth (‘rather pleasantly like a blond satan’ is how Hammett sums him up).
Samuel Spade’s jaw was long and bony, his chin a jutting V under the more flexible V of his mouth. His nostrils curled back to make another smaller V. His yellow-grey eyes were horizontal. The V motif was picked up again by thickish brows rising outward from twin creases above a hooked nose, and his pale brown hair grew down – from high flat temples – in a point on his forehead.
We can assume, I think, that the point on his forehead is V-shaped. Spade is the first of three people physically described on the first page of The Maltese Falcon, and the descriptions keep coming. Even Spade’s partner, Miles Archer, who appears on page four and is dead by page eight, can’t be skipped. ‘He was of medium height, solidly built, wide in the shoulders, thick in the neck, with a jovial heavy-jawed red face and some grey in his close-trimmed hair.’ You would never have guessed people could come in so many sizes and shapes. There’s a logic to this obsession with exaggerated physical characteristics: a detective has to have learned the trick of describing his man in words that couldn’t apply to anyone else; the elaborate description belongs to the mind of the detective through whose eyes we’re seeing the story. It’s a reflex, his tic. But it goes further than professionalism strictly requires. Here’s Brigid O’Shaughnessy, the book’s love interest and eventual villain, whose body is ‘erect and high-breasted’ and whose eyes are cobalt blue: ‘She wore two shades of blue that had been selected because of her eyes. The hair curling from under her blue hat was darkly red, her full lips more brightly red. White teeth glistened in the crescent her timid smile made.’ Brigid, an incredible spectacle in red, white and blue (she doesn’t translate well to black and white film) isn’t somebody you easily forget – not after the strange rictus of that smile. External appearances naturally have weight in a story that depends on action, but some of those appearances are so arresting that it’s easy to suppose there’s nothing underneath.
Signs are supposed to mean something, particularly in detective fiction. The point of the footprint, the fingerprint, the single hair is that someone – preferably the murderer – must have left them behind. To the ideal detective, everything is a clue: it exists only to tell you about something else. In Christie’s The Murder of Roger Ackroyd, Poirot is interested by the fact of a chair having been moved, which, he says, might not be significant, but of course he thinks it is – and, of course, he’s right. Hammett’s objects, on the other hand, aren’t clues, however suggestive they look. In the final part of The Dain Curse, a man has been pushed off a cliff; his wife, Gabrielle Leggett, on whom suspicion has repeatedly fallen, has run away. In one of the rooms in their house the Op finds a .38 pistol in the middle of the floor; there’s an empty cartridge next to the pistol, and another under a chair. Things look bleak for Gaby, though inexplicable: when the Op goes back to the house, one of the cartridges has disappeared. There’s a rational explanation: Gaby wanted to kill herself, but didn’t dare, and fired at the ceiling; the maid, who found the bullets and worried they might look bad for her mistress, tried to clear away the evidence that might convict her. The cartridges are false leads: they are only mess, meaninglessly themselves. Appearances can be beguiling, but that doesn’t mean they should be believed. Hammett’s detectives are supposed to be plain-speaking men who aren’t taken in by complex appearances: the truth is generally much simpler than the apparent clues suggest. The detective prefers to get his information not by detecting but by asking, threatening, persuading. Buying someone a bottle of rye often helps too.
But this leaves behind the complex, messy appearance of things. Earlier in The Dain Curse, the Op is attacked by what seems to be a ghost. He tries to grapple with the thing but can feel only damp air; it writhes around and can’t be grasped; he feels blood, not realising he was already bleeding; he’s weighed down by something heavy in the air. ‘I couldn’t – though I tried to – tell myself that I did not see this thing. It was there.’ The effect is strange enough to outlast the outlandish explanation. He was in a house belonging to a cult, a scam elaborately designed to manipulate and fleece its clients. The ghost was ‘made by an arrangement of lights thrown up on steam rising from a padded pipe that had been pushed into a dark room through a concealed opening in the wainscoting under a bed’. Maybe, but it felt impossibly real. The executive illusionist at the joint is a woman who is compelling, magnetic, but all appearance: ‘it was as if her face were not a face, but a mask she had worn until it had almost become a face.’ There are sacrificial altars, white robes and bloody daggers – carefully plotted tableaux. This is what the detective fears: not being able to see what is behind the illusion. He always does, of course, but the rattled-through explanations don’t erase the memory of the strange, empty visions that have gone before.
In 1930, with four of his five novels completed, Hammett met Lillian Hellman, who would be his companion on and off for the next thirty years; in the time he knew her she went from star-struck child to toast of Broadway. The legend of The Thin Man, which he had just begun when he met Hellman, is that she talked him out of abandoning it. The parallel seems perfect. The book’s hero, Nick Charles, has taken early retirement from sleuthing to pursue his interest in alcohol. (Fans recommend playing a game as you read: to try to match him drink for drink. This would be dangerous.) Nick’s wife, Nora, who never knew him in his detecting days, persuades him to take on a final case: there’s a family connection and she rather likes the excitement. She’s surprised at his methods, though.
‘But I thought everybody was supposed to be innocent until they were proved guilty and if there was any reasonable doubt, they –’
‘That’s for juries, not detectives. You find the guy you think did the murder and you slam him in the can and let everybody know you think he’s guilty and put his picture all over the newspapers, and the District Attorney builds up the best theory he can on what information you’ve got and meanwhile you pick up additional details here and there and people who recognise his pictures in the papers – as well as people who’d think he was innocent if you hadn’t arrested him – come in and tell you things about him and presently you’ve got him sitting on the electric chair.’
‘But that seems so loose.’
‘When murders are committed by mathematics,’ I said, ‘you can solve them by mathematics. Most of them aren’t and this one wasn’t.’
Here’s a magician explaining his tricks, but it isn’t neat. The dirty world Hammett deals with, which began with the archetypal badland of Izzard in ‘Nightmare Town’ and the Personville (or Poisonville, as its inhabitants like to call it) of Red Harvest, is never neat. No one will condemn himself: everyone has too much to lose – just look at the pre-election intrigues of The Glass Key, or Elihu Wilson in Red Harvest (‘along with those pieces of property he owned a United States senator, a couple of representatives, the governor, the mayor and most of the state’). So the detective has to rely on common sense to see through the fog: a bit of guesswork and a bit of intuition. But he’s also more involved than that: he’s implicated by making something fact, something that wasn’t fact before. If you put the story in the papers everyone will believe it; on its own it won’t stand up, so you have to put a bit of weight behind the shoulder.
Hammett dealt in facts. He also made them for a living. The Continental Op, Sam Spade, Nick Charles and The Glass Key’s Ned Beaumont are four of them. But production ceased. Perhaps he was worried that his facts were empty. In 1931, he wrote to Hellman: ‘My ambition now is to collect enough money to be able to take a couple of months to finish The Thin Man, which, God willing, will be my last detective novel.’ He didn’t mean it to be his last novel: he wanted to try something different, and he was still trying when he died in 1961, leaving part of a book he had finally begun. ‘Tulip’ was published by Hellman in 1966 in a collection called The Big Knockover (a title he had meant to use for another novel he never wrote). It’s oddly discursive, as if when you don’t have clues and mysteries all you can do is talk. Tulip meets Pop, who is trying to write a book. Pop says:
If you’re careful enough in not committing yourself you can persuade different readers to see all sorts of different meanings in what you have written, since in the end almost anything can be symbolic of anything else, and I’ve read a lot of stories of that sort and liked it, but it’s not my way of writing and there’s no use pretending it is.
Poor Hammett. He always committed himself – to his politics, to the people he loved – but what he wrote was condemned to be symbolic, endlessly interpreted. That’s the problem with facts: people don’t always believe them.