I wrote this to clarify my own thoughts on writing murder mysteries, but it grew somewhat.

Crime, setting, clues, plot. Four factors that influence each other when creating the outline of a mystery story.

"Crime" is, of course, the criminal event, which is usually a murder. There's no great significance in the crime being a murder - it's just part of the genre. It could just as easily be theft, rape or fraud. Indeed, it could be an event which is not technically illegal but is noteworthy, such as the bride absconding from a wedding, or someone taking a surprising job. However, I'm going to assume the crime is a murder.

The murder has a Setting. I don't just mean the place where it happens; I mean all the aspects of a location, building, era, culture etc. which limit what the killer, victim, witnesses, suspects and detective can do.

If the murder takes place in the present day, in the attic of a guest house, then this places limits on how the killing could be achieved, what the witnesses could have seen, what clues could be left, and how the detective could find them.

The killer can't use antigravity boots to get into the attic, can't persuade the victim to smother themselves with a pillow, and can't disguise themselves as the victim well enough to fool all the other residents for a week.

The Clues are a collection of individually incomplete indicators of how the crime really occurred. Together, they have to logically show both that the perpetrator committed the crime in a certain way, and that it could not have been committed by anyone else in the same way, and that it could not have been committed by anyone in some other way.

Sometimes clues indicate motive, but thankfully this is currently out of fashion. Any mystery that is solved simply by showing which one of the suspects had a motive is just plain rubbish. Dorothy L Sayers wrote that once you know the only possible way a crime could have been committed, that will tell you who is the only person who could have committed it. And once you know that, motive is either obvious or trivial.

Maybe she was overstating the case, but in my opinion, motives are very easy to construct, and can be just tagged onto the rest of the story as an afterthought, if you are so inclined.

Some clues are red herrings - they point to some perpetrator or method other than the "real" one. Red herrings are tricky to write well, because there need to be extra clues which constitute incontrovertible proof that they're the false clues and the others are the "real" ones. If you don't find a way in the story to convincingly discount red herrings, there is more than one possible solution at the end of the book and - shock horror - the detective might be wrong.

The good mystery writer hides real clues in a mass of irrelevant detail and a few red herrings, but doesn't withhold them from the reader.

The fourth factor, Plot, is the way the clues are laid out in front of the detective, and is also constrained by Setting. If the detective is a Miss Marple-like character, they have to go around each of the characters, interviewing them informally, hoping that some clues will slip out in their testimony. Miss Marple can blag her way into the hotel room where the murder occurred and nose around for a few minutes, but she can't get a search warrant or batter down the door. Also, she can't aggressively interrogate anyone, unlike DCIs Taggart or Tennyson.

There is a fifth factor - character. The personalities and quirks of the characters, including the detective and victim. Some character traits, such as a habit or a way of walking, can be clues. However, most characterisation is there to turn the puzzle into a story. It makes the story more readable - and bulkier - but doesn't add much to the mystery.

Having said that, some murder mysteries are highly novelistic, with extensive characterisation - even including subplots that have nothing to do with the mystery. The fashion for novelistic mysteries - or even novels where the mystery is just an excuse for the other plots - is quite strong as the moment, perhaps because all the straightforward puzzles have already been done.

So, we have Crime, Setting, Clues and Plot. Plus the parenthetical issues of Character and Motive. When constructing a mystery story, you can start with any one of the four main factors, seeing what it suggests about the others, and letting the implications bounce back and forth, sometimes changing the initial premise.

For example, let's say I want begin with the Crime and work outwards. I begin with the premise that an elderly woman, a patient in a care home, is found dead one morning. She's been poisoned, but has no obvious enemies, was not wealthy, and had not visitors for the last six months.

Already this gives me some elements of the setting - a care home for the elderly, containing a varied community of old people (mostly women), plus carers, administrators and visitors. Let's say the victim was maintained there by her family, who would have the cash to do so, but, as she's not rich, neither are they.

This gives me a plausible motive - the family, running out of money because the old woman is living longer than expected, kill her to stop themselves going bankrupt. That's the "why" sorted out, but what about the "how"?

Let's say the old woman, who I've decided on a whim to call Mary, was sent a box of chocolates, one of which was poisoned. it might take her a week of occasionally eating random chocolates from the box to find the poisoned one, but eventually she did. So the killer wanted her dead, but wasn't on a strict timetable.

If Mary's family just visited out of the blue after six months and gave her a box of chocolates, soon after which she died, that would throw suspicion instantly onto the family. So they need to find another way to get the chocolates to her.

How about: Mary had been receiving romantic letters for months from what she thought was a male admirer? He could ask her to burn all his letters - to eliminate evidence - and she wouldn't be suspicious if he sent her a little gift in the post.

But if Mary is 80, is this likely? It's possible, but not probable. So let's make Mary 60, attractive and mentally active, but forced into the home by her parents because she couldn't walk easily and had a heart condition?

And what if she'd had a fling with a married man when she was 40? A man who appeared to be writing her romantic letters 20 years later, now that his wife had died. The killer, knowing about the fling, could easily make Mary think he was writing to her again.

This suggests a clue for the detective to find. Mary keeps the letters from her old flame, and the police track him down - to find he's living with his domineering wife, and still having affairs with women his own age at the bowls club. But handwriting analysis shows that whoever sent Mary the letters, it wasn't him.

This tells the detective that someone who knew about Mary's past had killed her. Which excludes the carers and most of her friends at the home, because she never discussed the affair with them.

The handwritten letters show nothing interesting forensically, and the handwriting is too idiosyncratic and inconsistent for analysis, and doesn't match any of the suspects. The variability of the handwriting is a clue, which I've just retrofitted into the story.

So now I have a crime, a setting, some clues, and the barest outlines of plot, the development of which have modified the original crime.

But what about non-family members who knew about Mary's past, and might want her dead for their own bizarre reasons? Mary had a confidant at the home, a fellow inmate (call her Sarah) she trusted with knowledge of her past exploits. Could Sarah have written the letters? No, because she has Parkinson's disease, which makes her hands too unsteady.

Could she have dictated the letters to a nurse, on the pretext that they were just conspiring to cheer up poor old Mary who's family never visited her? I'm sure there could be a red herring where the nurse does write Sarah's letters for her, but they're to her solicitor, and concern her will.

Now how can I tie all this up into a plot? Well, it gives my detective (a police inspector, not an amateur) a set of people to interview, all of who can drop clues. Specifically:

* The resident who found the body
* Some administrators of the home
* Some carers, including the one Sarah used
* Sarah
* Sarah's solicitor
* Mary's family - let's say her daughter and the daughter's husband.
* Mary's old flame

Interviews with these gives a plot outline something like this:

* The administrators know nothing. Neither do most of the carers, or the resident who found the body.

* One of the carers tells the detective that Mary had been receiving letters, and she was always cheerful afterwards, writing back immediately.

* Sarah confirms that the old flame was back in Mary's life. On a second interview she says that the letters she dictated were to her solicitor. The solicitor confirms this, as does the carer who helped write the legal letters.

* The old flame maintains he hasn't heard from Mary in 20 years. A handwriting expert shows that neither he nor his wife wrote the letters.

* Mary's daughter maintains (truthfully) that neither she nor her husband had visited Mary for six months, but admit that they had pushed her into the care home to be rid of her. The husband lets slip that he's in debt - which suggests they can no longer afford to keep Mary in the home.

* The postmark on the letters to Mary are from a town nowhere near where any of the characters live. However, the husband does drive through there every week as part of his job.

* The handwriting on the letters is highly variable and doesn't match any of the suspects, but if some words or phrases were written by the daughter and some by her husband, a match is possible.

* Between them, they did it.

Okay, it's not great, but I think it's passable as a whodunnit puzzle. There's undoubtedly some holes that can be picked, but I only came up with this one as an exercise. The real one will need more time to work out.

1 comment: