Friday, October 7, 2016

Over The Truth

It's time for my annual finally-it's-not-a-review post! I really should try doing more of these feature posts...

Longtime readers of this blog will have noticed that I am a big fan of puzzle plot mysteries. There's a reason why I keep mentioning writers like Queen and Arisugawa, and why I wrote a rambling piece on clues in mystery fiction. To me, mystery fiction is at its best when it's a game between the author and the reader, where the writer has laid out a logical problem for the reader to solve. Obviously, this challenge has to be fair. While I don't believe in either Knox' nor Van Dine's overly specific rules, I do think a puzzle plot mystery has to be fair: it must be possible for the reader to logically point out the solution of the problem in the story, be it a murder or something more innocent like a code. Note that it doesn't need to be realistic: only fair. Mysteries set in science fiction or fantasy settings can be as fair as mysteries set in hyperrealistic settings. Mystery fiction is at its most exciting when it goes beyond the in-universe story of a detective VS a criminal, and transcends to a meta-level duel between the author and the reader.

The win conditions for the author to this game are obvious. The author wins if the reader fails in solving the case correctly, provided that the story contained enough hints so it can be reasonably expected a reader could solve the problem. Revealing without any warning that Sally was in fact a ghost who could float through walls in a natural realistic story is probably not fair. If the story is set in a world where ghost do live, and we find a gravestone with Sally name's on it, then you could make the argument it might've been fair. Of course, 'fairness' is a very subjective thing. When can we say something was adequately hinted at? The author's job is of course to run awfully close along the line of [unsolvable] and [solvable]: it needs to be difficult enough that people believe they can solve it, but not be disappointingly easy. One of the things I heard at the Mystery Club made a big impression on me and it's still one of the things I keep in my head whenever I read a mystery story: It's easy to write an unsolvable mystery story that baffles the reader. It's difficult to write a solvable mystery that still entertains the reader.

If the reader failing in solving the case is the win condition for the author, it stands to reason that solving the mystery is the win condition for the reader. But that brings me to my main question today. What do we considering 'solving' the mystery? The Stereotypical Example: Suppose you're watching a whodunit mystery drama with a somebody, who says this at the beginning: "I bet you the butler did it. It's always the butler, and look at how suspicious he looks", and ninety minutes later, it is revealed it was indeed the butler, do you consider this solving the case? (Never mind the fact that a mystery drama with a guilty butler is actually quite rare) The main problem of a whodunit is right in its nomer. So can you say someone solved the mystery if they correctly identified the culprit? Too often have I seen people saying they solved a mystery not based on logical reasoning, but for 'meta' reasons. From 'you know he's the killer because he acts so nice' to 'you know she's the killer because it couldn't be that couple, and there's nobody left'. Is this solving a case?

No, of course not. If we consider the puzzle plot mystery a game of logic, guessing isn't going to be a correct answer (and don't forget the fact that most people like to forget about hindsight bias). Solving a puzzle plot mystery is like a (very limited) math test: there is a correct way to arrive at the solution, and the key is to have both the method and the solution. I'm pretty sure that you don't get full points at an exam if you just write down an answer, even if correct, on the answer sheet, if you get any points at all. I remember at the Mystery Club, we had whodunit games: participants were given the first part of a short whodunit story, up to a Challenge to the Reader, and you had one hour to read and solve it. If you thought you got it, you'd need to go to the writer of the story and explain the method through which you identified the culprit and eliminated the other suspects. So picking a suspect at random was never an option. You needed to identify the logical path the writer had laid out for you and reveal everything they had hidden in the story in order to 'win'. To me, this was the most game-like form of the puzzle plot mystery and I loved it.

But how much of the path must you have explored before you can really say you solved the mystery?  What if you only identified part of the hints the author laid out across the story? Would you say you solved the mystery? If a story features multiple fake solutions (each with their own proper chain of deductions leading to them), and you manage to deduce all of them, but stumble upon the final, true solution, how much of the mystery have you actually solved? A lot, or almost nothing? The win conditions for the reader will differ per reader, I guess, but I am curious as to what those conditions are. Are you easily satisfied with your own performance, or do you consider everything but perfection as absolute failure?

On a side note, what puzzles me also are people who comment about how they knew right away who the murderer was in series like Ace Attorney, which is in fact mostly influenced by Columbo and like Columbo, seldom a whodunit, but a howdunit. If the mystery never was about hiding the identity of the culprit, why bother solve it, I'd say...

To give two examples of 'solving a mystery' from The Decagon House Murders and The Moai Island Puzzle (Disclosure: I translated both novels): I didn't solve The Decagon House Murders. Guessing who it was, is actually not very difficult from a certain point on, because the pool of suspects has been thinned out near the end, but arriving ata a logical answer as to why X is the murderer is actually quite hard. Some might even say that it is impossible to logically arrive at the solution in this book, though I'd have to argue otherwise: it is actually possible to logically deduce who is very likely the murderer based on hints and facts (who knew what and when) scattered across the narrative: the thing is that this chain of logic is not explained in the novel itself. As for The Moai Island Puzzle, I guessed the identity of the murderer, and I also got a nice chunk of the (amazing) chain of reasoning that leads to that person, but I never felt like I solved the mystery, as it was just a guess + partial suspicions/loose bits of deductions.

Anyway, enough rambling for today. What do you 'solving the mystery' means, and do you actually try to do that when you consume mystery fiction? Obviously, there are more types of mysteries than whodunits, and most stories are actually a mix of several elements, so what is the percentage you need to 'solve' in order to win the game? Thoughts to bring along as I start reading a new (old) book.


  1. Usually, I attempt to disentangle a plot and try to answer all of the questions before reaching the final chapter, which is what makes detective stories fun to read.

    However, not every mystery is solvable. Even if they technically play fair with the reader. Some of them provide a clue or two, but those clues either hinge on the color of someone's tie or a slip of the tongue. You practically have to be clairvoyant to solve those.

    But if you've a mystery novel that genuinely plays fair with the reader, you only have to answer two questions to win the game: identify the murderer or murderers (the who) and explain how the crime was accomplished (the how). The why is optional (bonus points) unless it's tightly woven into the who-and why of the plot (e.g. the locked room from The Moai Island Puzzle).

    That being said, I wonder if I can lay claim to having solved John Russell Fearn's Thy Arm Alone. Very early on, I was able to visualize exactly how the victim died, which is the crux of the plot, but was craftily dragged away from my solution and only came back to it when Fearn (more or less) allowed it.

    So, yeah, that's another complication when deciding whether you solved a mystery or not.

    1. I remember talking about whodunits at the Mystery Club, and a writer describing it as a leading the reader down a path. He was a person who wanted the readers to solve his mysteries, so he'd lay out a logical path, imagining that the reader would make this or that connection, and thus experience this, and so on. It's this 'path' I usually imagine when I think of actual fair play mystery, more so than the single-clue ones, which indeed have a history of being rather hit-or-miss. If you look at the genre like a game, you *want* a game, not something that ends after one move.

      The experience with the whodunit games at the Mystery Club really did change my feelings on looking at the genre as a game though. Usually, every reader has their own definition of 'solving a mystery', like explained in the post, but with the whodunit games, it was really an open game between reader and writer, and you needed to get all the important points in order to win (eliminate all other suspects, and identify the culprit). When I read a story now, I'll often solve parts, but seldom everything, but I still have a tendency to feel like I did solve a great deal.

      With your Thy Arm Alone example for example, I'd be inclined to feel like I did solve it, if you still managed to get back in time for the conclusion. The problem with mystery fiction is of course it's often about information flow, so even if you have a good idea of what happened early on, often there is some information roadblock that isn't removed until the end.

    2. Leading the reader down a path of logic is one way of looking at it, I suppose. You can compare that approach to, what we call over here, a speurtocht, but I always preferred the Carrian interpretation of the detective story being a Grand Game.

      A duel between the cunning author and a quick-witted reader, in which the former has to fool the latter without playing his cards too close to his chest.

      The author has to create "a ladder of clues" and "a pattern of evidence," forming a coherent design, but also find ways to distract the reader from discerning the pattern too early. All the while, the reader has to slouch across the pages, like Columbo, attempting to poke holes in all of these clever schemes. Of course, you rarely solve these intricate, fair-play puzzles for a full 100%, but getting halfway there often makes you feel as if you did your part in solving the mystery. For example, I'm usually very satisfied with myself when I figure out the locked room angle of a story (or even come up with an alternative explanation).

      Yes, the single-clue stories can be aptly described as one-move chess games and it would take a gem of a clue to make such a story on par with the ones that have that ladder of clues or the path of logic.

      I was not lured away in the Fearn novel by a wad of confusing or unfair information, but how the whole scene was presented and tempered with by the culprit. It completely disagreed with my picture of the crime, but, once the layers of misdirection were peeled away, I saw my initial solution emerge underneath. I say it was a successful piece of misdirection.

    3. It's funny when you think of mystery videogames; they are by definition designed to be solvable by the player. They are designed with a clear path (with red herrings), something you can easily turn into a flowchart. Nobody is going to buy a game they can't complete. In most mystery games, the narrative won't even move forward unless the player solves something. Yet they are often a lot simpler than conventional mystery fiction, precisely because players need to be able to solve them. So there's usually a very big gap between the difficulty levels, or what is considered 'solvable', in 'normal' mystery books, and games. Of course, most scenario writer for mystery games aren't professional, specialized mystery writers, but still, it's funny to see how the same design idea can lead to different results, just because one medium is active and the other (more-or-less) passive.

      If that was your experience with the Fearn novel, I'd agree it was a great piece of misdirection. Leading you away from the general direction of a solution is one thing, but to absolutely get you away from the right solution is something else!

  2. After reading so many novels by the same author, do you sometime feel like you can sort of predict who the culprit is earlier in the novel base on the author style of writing? I certainly feel that way sometime while I was bridge reading Ellery Queen. There are some exception of course, especially in Queen's earlier novels; The Dutch Shoe Mystery being the prime example where the culprit just came out of nowhere (I still think this is not a fair play by the way).

    Speaking of The Dutch Shoe Mystery (and this is something I wanted to discuss ever since I read that book, which has been a while to be honest), how do you feel about a mystery where technically where all the "hint" are technically fair and point to the suspect, however, the suspect was rarely mentioned in the book among the chapters as if he/she was a minor character? Fair or not fair? It has been a while since I read the book so I don't remember all the details, but I remember feeling cheated when I finished the book.

    1. Oh yeah, authors do often have a preference for certain culprits or tricks. Ayatsuji's House series often revolves around the same, fundemental idea, even if in completely forms.

      I don't remember the details of how it went in Dutch Shoe, but in general, I think I wouldn't feel cheated if the novel was written in the logical school like practiced by Queen or Arisugawa, which often takes a lot of time and effort to prove why the other suspects are eliminated and why X is the one. The 'basically a minor character who turns out to be the culprit' trope can feel a bit disappointing, I agree, even if it's become a common trope with the Father Brown stories, but I think that if a story at least puts the effort in convincingly striking off the 'major' characters and pointing adequately at the real culprit, I don't have any real problems with it.

  3. Some really cool discussions going here :)
    There's not much else I could add, but for me personally, if I find myself managing to solve/figure out the core/most baffling part of the mystery being presented, I would usually consider myself having solved the main riddle that the author was going for. I don't think I had ever solved a story 100% and even if I did figure out the main trick, I would usually be still surprised by the amount of fine details/good foreshadowing that a masterful author can put into his work.

    Going a bit on a tangent, do you plan on reviewing Aosaki Yuugo's other books? I recently read his Black Umbrella mystery and Library Murder Case in Chinese, and man, had a real blast with the Ellery Queen style chained logic. Definitely one of those budding authors that I will be following from now on.

    1. I recently read the second book in the series, the aquarium one, but the review isn't scheduled to appear until somewhere next year ^^' (But in short: this one was a bit *too* complex/detailed for its own good) I'm reading the pocket releases, and the aquarium one was released just a few months ago, so it will take a while before I'll get to the library one.

  4. In attempting to determine whether I have achieved a win, I take into account the fact the both the author and the reader have certain unfair advantages.

    The author has a certain advantage in that he can lay as many false trails as he likes, and then differentiate the correct trail with a well-hidden clue. In other words, the author has an inherent advantage because he is in full control of the plot.

    On the other hand, the reader has the advantage of meta-analysis; after a while he knows the author's favorite ploys. Further, after you have read enough mysteries you develop a feel for how facts can be hidden and the limited number of solutions that are possible.

    My opinion therefore is that since both the author and the reader have an unfair advantage, I consider that I have solved the mystery if I figure out the guilty party by some logical solution even if it is not the author's solution, simply because the selection of the murderer can be an arbitrary choice on the author's part. In other word's Detective Conan's statement that one truth prevails is nonsense; life is to complicated for that, and that is why murder trials can be very messy affairs. Human perception and cognitive equipment are far from perfect so the determination of the truth is not so easy. One truth prevails in fiction because the author wants it that way.

    I have always considered the Ellery Queen novels to be the best fair play novels because I have a pretty good win/loss ratio with them; If you pay attention they really are solvable. I think the reaction the author is going for is to have the reader admit at the end that he could have solved, i.e., that it really was "fair play" and you should have figured it out but did not. The closer the book actually is to fair play, the better the author and the better the reader. This is what makes for the most satisfying experience.

  5. With the reader's meta-analysis method, it can turn out into a 'I know that the readers knows that...' loop, so I at least try being careful for falling into a trap ^^.

    I remember reading a great whodunit once (the whodunit game I explained in the post), where the writer said he purposely aimed at a 50% solution rate. So he expected half of us to correctly identify the culprit and reasons for that. And it was great! With the writer's inherent advantage as you call it, it's pretty easy to make something complex by adding things here and there, but I was surprised to hear someone was carefully adjusting his story to get a certain success/fail rate (which mostly works because we did this with a club on a regular basis, so he had a good idea of what the average difficulty level should be, but still.

    Yeah, Queen is really great at fair play. I think I explained in a post on clues last year, but the way the 'logical path' is set out in those novels really results in stories you can solve with mostly *diligence* and effort (combining fact and clues in a logical way), while with Christie and Carr, I still often feel many of them depend on *inspiration*/sudden flashes of insight which might come or not.

    1. I agree with you on Christie and Carr. I read Death-Watch a long time ago. And I would have to say that while the clues were there, they were so obscure and difficult to interpret that there was little reasonable chance (for me)to solve the case by the application of reason to the problem.

  6. What I really enjoyed about the Decagon House Murders is that the trick was 'fair-play', but it involved a logical leap that would be very unlikely for the average reader to solve.


    The way Ayatsuji presents to you two characters and then in the end tells you they were one and the same is brilliant; technically solvable if you remove yourself from the way you would normally read books, but very unlikely to be solved. Some might consider this unfair, or be disappointed that they could not solve it, but I don't actually read mystery books to solve them. For me solving them is a cool plus, but I would prefer to be surprised by something that makes sense but that I would have never seen coming. In short, I want my mystery books to be solvable, but not in a way that I could do it easily i.e. I always want to lose. This is why I consider Decagon to be one of my favourite books. (That's not to say solving the mystery doesn't have its own charms.)

    PS Good to see another 'feature post' - always enjoyable.

    1. Your comment reminds me of the inherent contradiction to the genre that Ace Attorney's Takumi also pointed out in an essay: people want mystery fiction to be solvable, but also to baffle them. It's really difficult to find that exact spot where you're fooled, but still think it's fair enough (more than one vague hint, for example).

    2. This comment has been removed by the author.

    3. How interesting - could you post a link to that interview, or is it on the Gyakuten Saiban Library and I've just missed it? I actually think that the Ace Attorney games deal with this contradiction (ba-doom tish) pretty well. Being of a completely different medium to novels, they can more easily surprise you while still being fair as you work through the case with the writer guiding you as opposed to having all the facts laid out and then everything revealed to you. In that way they stop being the battle between the writer and reader (because as you mentioned in another comment they have to be solvable). I think the whole thing about the fine line you mention just goes to show how difficult writing good mystery fiction is; you can't just come up with a good trick (hard in and of itself), you've also got to mask it well.

      Speaking of, I actually just bought the newly published (in the UK at least) English translation of A Midsummer's Equation, which I haven't actually read before. I seem to remember you being quite lukewarm on the trick in that one, but I'm still quite excited about it. Hopefully it's alright - if anything his novels are always quite readable.

    4. It's this essay.

      It's been a really long time ago since I read that book (I think I read it when it was first released, so some 5 years ago), but yeah, I remember I wasn't as fond of it as the other two full-length novels in the Galileo series. That said, Higashino's novels are basically always entertaining reading materials, so you'll never go really wrong with him.

  7. I think I understand why games like ace attorney or danganronpa are so compelling compared to most novels. In AA it's your job to defend the client, or else they will be guilty even though they are innocent. And in danganronpa, finding the culprit is actually a matter of life and death. So there is a sense of urgency and a lot of tension.

    Whereas, if we take a Conan Soyle or SS Van dine novel or whatever, the detectives actually don't have any obligation to investigate. So there is no real danger for them, nothing is in jeopardy. They don't even have a time limit to complete their investigation.

    1. 'Stakes' is definitely a problem with a lot of series, as you know it's unlikely the detective-protagonist is going to die.

      Closed circle situations, where the whole cast is locked up in an area with an unknown assailant, usually work best when there isn't a character around who wears plot armor. That's why it's difficult to be afraid for Hajime in the Kindaichi Case Files after decades of being in closed circle situations.

      Though I wouldn't say it's all about stakes. Both Ace Attorney and Danganronpa work differently than most novels, because usually mystery novels are rather... end-loaded: with most of the 'exciting' stuff happening at the very end. Compare to Columbo, Ace Attorney and Danganronpa, where you clearly follow the protagonist as they solve one smaller mystery after another, which eventually add up. Few mystery novels do the bit-by-bit approach, which often result in slower mid-sections.

    2. I'm reading this one novel by Giles Blunt called By the Time You Read This and I have a feeling it'll be like that. Literally one or two relevant things have happened in the first 150 pages and it's just a snoozefest. The way the author talks about things is realistic though. I guess he was in the mood of setting up the story's atmosphere for 300 pages.

    3. It can work if the writing is entertaining, but as I as a reader are definitely look for 'the game', so the author needs to be good if they want to keep my attention.