My recent musings on the coincidences that went into Mystery Science Theater 3000 being a success got to me to thinking about coincidences in murder mysteries. The general rule is, of course, that coincidences may not help the hero of a story, and this was codified in Fr. Knox’s decalogue in rule number six. It would be a fool’s errand to try to count up which rule was most often broken, but I suspect it might be this one.

I should clarify that I mean broken but not to the benefit of the story. Agatha Christie managed to break several of the rules in ways that produced a good story, but not this one. (There are two examples I can think of in Agatha Christie’s work that involve coincidences, one in Poirot and one in Miss Marple. In the case of Poirot, she even went to the trouble of saying that Poirot considered the case a failure because he would not have solved it except for the coincidence.)

Having said that, I don’t think it’s impossible to use coincidences in mystery stories. One tolerable example of this is a coincidence which brings the detective in to the case. A good example of this is the Lord Peter Wimsey novel Unnatural Death. Lord Peter learns of the case by the accident of being seated in a restaurant next to someone who was telling a friend about it. He then weedles his way into an acquaintance with the man who told the story and sneakily gets enough information about it out of the man that he can begin investigating. Thus even in this coincidence Lord Peter has to do work to really get started.

This kind of coincidence is tolerable, I suspect, because it’s just a somewhat exaggerated form of the sorts of coincidences which are necessary for the detective to be involved at all. If Sherlock Holmes is to be called into a case, the murder must take place in London, or at least in England. If a man murdered another in the central African jungle in the cleverest possible way, Sherlock Holmes would never hear of it. This is even clearer in terms of time; if a man in the 1980s murders another, Sherlock Holmes could not possible have heard of it, at least Holmes as written by Conan Doyle. Nor would a fiendish plot ever come to the attention of Holmes which happened upon a whaling ship at sea which was lost in a storm before it ever reached port, with all hands dying. For a detective to embark upon a case, many things need to be coincident with his location in time and space. To add on top of this someone happening to talk about the mystery at lunch with a friend at a table next to the detective is just more of the same.

So what are we to make of the sort of coincidences which are more than this but less than just giving the detective the solution?

One of the more difficult ones are coincidences which look like they help the detective but are actually misleading. Probably the best example I can think of, here, is in the story Have His Carcass. Harriet finding the fresh blood seems to be helpful in pinning down the time of the murder with unusual precision but actually confounds the investigation almost until the end of the story. It definitely was quite interesting in that story, though I think it would be difficult to pull off well.

Then there are the coincidences which only seem to be clues, but actually aren’t.

These are often quite interesting when they happen prior to the detective getting on the scene. Red herrings are probably the most obvious example of this. Finding out that the maid’s earring was in the parlor where the body was found because the butler had been stealing jewelry and secretly hiding it in the chandelier above the door (which was never used) is, properly speaking, untangling a coincidence from the main problem.

Red Herrings are not the only such coincidence, of course. Sometimes things look weird for the murderer to have done because the murderer did not do them, but at the same time the person who did is not available. There might be a book missing from the library because someone—perhaps a neighbor—borrowed it a week ago and no one (still alive) knew that or noticed it then. It’s possible that someone was mistaken about which book is missing, and the person who borrowed it didn’t say anything because they were asked about the wrong book and weren’t told why they were asked, so couldn’t tell that there might be a mistake. Perhaps the police are withholding the evidence that the book is missing because they don’t want to tip off the murderer that they know, and so the person who could have easily told them didn’t know to come forward. All of these would work well in a story.

Then we come to the cases of coincidences that do actually help the detective, though they are not merely handing him the solution. Can these work?

I want to say that they can—the safe answer is to never say never—but it’s hard to think of how it can be done. One obvious answer is for the help to be trivial. The problem with that solution is: then why bother at all?

I suspect that the answer has to be something that preserves the detective working hard and being the only person who could solve the crime even with the luck. I suspect that the best way for this to work would be for the detective to manufacture his luck. That is, it is only through his knowledge and effort that he was in the place to receive the luck at all.

A good example of this would be reasoning that if there was evidence to prove who did it, it would be of a particular kind that would then have fallen in a particular place. Since it is not there to be found, if it ever was there it must have been picked up by a particular kind of person and so if he circulates word among these people—or interviews them, or some such—the evidence will fall into his lap. I have a memory that Sherlock Holmes did this, perhaps more than once. I can’t place the story, but I have a memory of more than one person coming, hat in hand, saying that he heard that Mr. Holmes was looking for someone who saw something-or-other, and he did, and getting rewarded for it.

The other, I suspect inferior, kind of luck would be something coming completely out of the blue, but only the detective understands its true significance. An example which comes to mind, though it is a very imperfect example, since it wasn’t discovered by luck, would be the evidence given by the nanny in the Poirot story Five Little Pigs. The nanny thinks that the evidence she has proves the guilt of Caroline Crale (which is why she withheld it), when Poirot knows that it proves Caroline’s innocence. If that kind of evidence were to come to the detective, even by accident, I think it would still work.

To bring this back to where I started: I think that coincidences are acceptable only when something unusual and special went into taking advantage of them. This is very much true of Mystery Science Theater 3000. Yes, a lot of unusual circumstances came together to make it possible, but it was a special group of people who took advantage of those circumstances and made it happen. Most people would not have made something great in the same circumstances.