#506: The Men Who Explain Miracles – Episode 9.2: Laying Down the Laws – The Knox Decalogue & Detection Club Oath

Van Dine Knox header

Last week we looked at S.S. van Dine’s take on what a detective novel should be, this week Ronald Knox has his say.

The Knox Decalogue is possibly more famous than Van Dine’s Twenty Rules — I imagine largely on account of The Disavowed Controversial One, and yes, we address that — and as it’s also a shorter list it’s much easier to discuss.

However,  there’s actually a lot more to Knox’s list than many people may realise.  This page conveniently contains the Van Dine and Knox lists, as well as Raymond Chandler’s commandments that get a mention in this episode, but the full glory of Knox’s undertaking can be found here at the GADetection wiki,  It’s worth a look if you only know the “short-form” version, and casts a few new ideas into the hat regarding a great deal of Knox’s intent, and especially The Disavowed Controversial One (but, as I say, we address that).

So, below we discuss Knox’s work on Sherlock Holmes, and how it might not necessarily have helped where his Decalogue was concerned, look at some features of Knox’s strictures, and the proliferation of detective fiction just as Knox seemed to think the jig was up.  And then I draw a parallel with the oath  that members of the Detection Club take, with a little help from a very special guest.  And then Dan unveils the masterplan he’s been working on for some time now.  And all this in a mere 34 minutes.

Is it possible?  Listen ahead to find out…


Previous The Men Who Explain Miracles episodes:

1. Rim of the Pit (1944) by Hake Talbot — A spoiler-heavy discussion

2. An interview with YA author Robin Stevens

3. On republishing Murder on the Way! (1935) by Theodore Roscoe

4. The Ed Hoch ’15 Best Impossible Crime Novels’ list of 1981

4.1 Books 15 to 11

4.2 Books 10 to 6

4.3 Books 5 to 1

5. Choosing our own 15 favourite impossible crime novels

5.1 JJ’s list

5.2 Dan’s list

6. An interview with Martin Edwards

7. The Ages of John Dickson Carr [w’ Ben @ The Green Capsule]

7.1 Part 1

7.2 Part 2

8. The Impossible Crimes of Paul Halter

8.1 Part 1

8.2 Part 2

9. The Laws of Detective Fiction

9.1 S.S. van Dine’s Twenty Rules for Writing Detective Stories


33 thoughts on “#506: The Men Who Explain Miracles – Episode 9.2: Laying Down the Laws – The Knox Decalogue & Detection Club Oath

  1. I wonder if the various rules, not only by Van Dine and Knox, but the lectures on locked rooms (Carr) and dying messages (Queen/Ross) and the various meta-moments found in other authors like Christie (Partners in Crime, The Clocks) didn’t serve as publicity for an oft-maligned but extremely popular genre. Yes, rules suggest legitimacy, but we see the tongue in cheek attitude held toward them by the creme de la creme of mystery authors. I do think JJ’s interpretation of “The Chinaman Rule” lays the controversy to rest once and for all.

    I’ll have to give the Locked Room Rules idea some considered thought before I fly off the handle with silly suggestions. The first thing that comes to mind, however, is simplicity and/or clarity of presentation. I confess that often I simply cannot follow the situation being laid down before me. Rarely do I attempt to figure out the solution to these things because they’re too complex for me to picture. There are a couple of strategies, however, that I do recognize and solve right away. That’s how I got Mr. Splitfoot figured out on the spot, as it were. I won’t say any more, but if you ever read that one, I figure you’ll catch it, too.

    Thanks for letting a bit of my essence travel overseas for this one. Sadly, this may be my only opportunity to recite the oath. Let this be the first of many times that my voice mixes with yours in homicidal harmony.


    • That was a fantastic recital! A performance not soon to be forgotten. Thank you Brad. And thanks to JJ for witholding the surprise from Dan to make it even better (and of course for the excellent podcast!)


        • Listyening to oneself is just not in the human DNA, is it? The editing of these episodes brings out my Full Cringe Mode at times when I have to re-hear what I’ve already said, looking for a point where I eventually stop talking and let Dan get a word in edgeways.

          And, man, some of the stuff that gets cut out…gee whizz, those points are just embarassingly bad at times. If anyone listens to this and hates it, take solace in the fact that it could have been so much worse 😆


    • I wonder if the various rules…didn’t serve as publicity for an oft-maligned but extremely popular genre

      That’s a fascinating idea. I read early Crofts and see the genre being beaten into shape, but once you get past a certain date I just sort of assume that everyone knew what they were doing and what they were about and so these entreaties in the text of a book are just someone stretching their wings and showing they know what they’re about. But the notion of self-publicity for the genre, to sort of self-adjudicate as well as promote, is extremely appealing.

      Which I supose then raises the question from your point: do we know the detective novel was maligned at the time? We’re all fans — as much as we disagree about the finer points of novels and authors, we’ll happily submit that the novel of detection done right was a wonderful thing — but was there still an aspect of misunderstanding or dislike for it at the time? Or was it again just another way for the genre to take form, and for that form to be cemeneted in the wider consciousness?

      Huh. What a wonderful idea., Brad.


  2. Superb declamation from Brad there.

    I think you’ve both got me thinking more about the necessity of at least some rules in detective fiction if it’s not to descend into nonsense or parody. Maybe the perceived lack of seriousness is in the mode or idiom of delivery or expression as opposed to the intent?

    Locked rooms? I’d need to think a fair bit on that. One thing I do like is strong sense of atmosphere in a locked room/impossible setting – I find a hint of the grotesque goes a long way in such situations.


    • Yeah, the intent is very important, isn’t it? I’m especially taken by the ending of that oath — “May other writers anticipate your plots, may strangers sue you for libel”, etc. It’s — humorously, but at the same time deliberately — wishing the worst possible outcomes upon someone where their writing is concerned, but I imagine everyone has a big smile on their face as it’s said and would commiserate with any fellow-author this happened to…so long as they’re stiking to the rules (when the commiseration may well happen more in public than private…though, of course, I speculate on behalf of a group of people I do not know).

      I don’t think there’s a definiitvie answer, and part of me loves that I don’t think I’ll ever get to the bottom of this humorour/serious threshold. It’s a supremely difficult balance to walk, and somehow the genre creators managed it.


  3. I loved Brad’s cameo – was anyone ever expelled from the club for breaking their oath? Looking forward to the Locked Room Rules episode.


  4. Oh my goodness! Brilliant! That certainly was a twist!
    Thank you Brad 😀

    Well, er, what was I going to say? Right. At least personally, every time I’ve come across an “apology” for rule 5, it has been that same one – avoiding Yellow Peril – yet the writer apologising seemingly hasn’t had the full text, since they use “probably” and “likely”.

    I’d like to mention a sentence before Knox reaches the rules:
    “The true essence of a detective story … is that in it the action takes place before the story begins. … the interest of the detective romance centres in the question ‘What has happened?'”. What an interesting statement. And often true, even. I have thoughts about this in relation to the detective videogames that this comment box is too small to contain 😉
    Anyway, I wanted to single this out. Even in a book where we see the event happen in “real time”, even if the detective happens to be present already, even if the plot begins to spiral and lots more excitement happens (thinking of some Carrs here), the plot is not resolved until the initial event, now long in the past, is resolved.
    I’ve singled out this quote too early I think. A lot of this has other excellent quotable sentences…

    Oh, he spoils The Secret of Chimneys in this. Can’t say I’m too heartbroken, though.

    After reaching the end, he makes his existential prediction, “all the possible combinations will have been used up.” I find myself thinking this sometimes. Particularly when it comes to impossible crimes, actually. I recently went through the solutions of Adey’s Locked Room Mysteries (avoiding books want to read but haven’t yet by skipping the whole page just to make sure), and was disheartened to see 90% of them involve some trick with throwing the key back into the room. 8% involve the criminal opening the door themself. 2% are not that. I suppose that’s what you get when you have to avoid the Carr and Hoch sections cause you haven’t read them all yet. It was a bit disheartening though.

    Alright, one final quote – the outro to the whole thing. I think it should be read.
    “But I must say no more. It is not for the anthologist to crush, heavy – handed, all the juice and sweetness out of the flowers he has garnered before the bouquet is handed over. More especially, when he deals with such frail blooms as mystery stories, that yield their scent but once, and then, desiccated, lose all save a botanical interest. When I was about to read one of the greatest detective novels of our time, my present collaborator said to me, ‘I wonder if you will think it fair?’ – no more than that; and I, greatly to be pitied, solved the problem at the first sentence. I will be silent, and let the reader cut the pages for himself with the sharper edge of his own intelligence. ”
    The number of times that’s happened…

    Whew, long comment… if I have any thoughts about impossible crime rules, I might try a separate comment…


    • I’ve always wondered about that “I solved it in the first sentence” thing — like, really, Ronald? Sure, hyperbole, and, sure, we’ve all read and/or watched something that has been transparent to us at and early stage, but I don’t see what benefit there is — and this is partly why I don’t think the essay and rules are intended to be humorous — in dismissing the genre that quickly and lazily.

      Now, sure, the intended audience would likely have been the sorts of people who were interested in mystery stories, but following and essay in which you lay out the precepts of the mystery story, call into question the validity of some of the mysteries in the forgoing book, and say that the genre is reaching a point where no new ideas and combinations will exist before too long, how much good work are you actually doing to then, in possibly the cruellest way, wipe out (sure, he doesn’t name the book, but still) a supposed great work in that genre? Like, don’t shit where you eat, Ronald., y’know?

      The more I dwell on this, the more I wonder how much Knox actually liked mystery novels. His own might even bear examination in light of this discussion.


      • I thought the ending was more of a recognition of how easy it is to spoil things, that perhaps even in telling you when you can solve the case (or that you can’t solve the case) he is too close to giving things away.
        And, I singled out that final passage because I did recognise it. “I wonder if you will think it fair” tells you to be on your guard, and look for something in particular. I read it as him saying that the novel actually was great, just that with the extra information he was given he had a key surprise taken away.
        Truth be told I think I admire his honesty when writing this, even if he is a bit grumpy. I think he is more charitable and self aware than you might be saying.


        • I dunno — “I wonder if you will think it’s fair?” seems to me to be more a “Well, it could be seen as fair, but I suppose it’s how you seek to interpret that”. Brings to mind, say, [no spoilers herein] John Dickson Carr’s The Case of the Constant Suicides, which brandishes the key clue in your face in a very clear way, but once you get to the answer you could arguably be justified in thinking you hadn’t been prepared in spite of the mention of X, Y and Z — things are said, but it requires you to know some specific background on the way to realising the connection between them…is that fair?

          To complain about this and then go on to say “I solved the book in the first line” — the first line, not simply “I solved it” or “this had me looking out for stuff” — that seems like he’s outright dismissing the efforts of the author in being over-clear and insufficiently intelligent in their construction. I’ve read plenty of lazily-constructed mysteries, and solved a bunch very. very early, but I’d never say I solved any novel’s mystery in the first line. Now, yes, hyperbole, but why go to that degree, unless to communicate utter disdain? The only way you could make it worse would be to say “I solve it by looking at the cover” excet that people have done this on account of some inexplivable cover art choices that are in no way the fault of the author. The first line is very much the author’s fault, and to draw attention to that seems deliberately curmudgeonly to me.


          • Which is why I will never quite forgive Christianna Brand for her introduction to the reprint of a certain title where she basically gave the whole game away by bragging about how she did this very thing. (You probably know which title I mean!) It was Brand more than the book itself that clued me into the killer right at the start! 😦


            • See also when reviews say things like “Reminds me of the trick used in…” or “If you’ve read X then you might see this coming”.


  5. Because there are only twenty or so solutions, I’d say the rules would be quite limited. For instance, if you say the parameters of the room must be clearly defined, all the solutions involving secret passages and floorboards would be violations. And just what is a locked-room murder anyway? Death of Jezebel is not technically a locked-room murder case. The doors to the theater were not locked and the tower was not locked. (obviously) When we say locked-room, do we mean an impenetrable or inaccessible area? That has less of a ring to it – An inaccessible-area murder mystery!

    I imagine the best way to make a list might be to look at the worst of the worst and see what went wrong. Rube Goldberg contraptions annoy me. Also, the laws of physics should not magically disappear. Shutting a window tightly can’t cause a lamp to fall over and stab someone in the chest. Can the key be missing or does it have to be produced? The deception must be done for a logical reason? It’ll be interesting to see what you two come up with.


    • If defining the parameters of the room excludes secret passages, I’m all for it. It’;s possibe the prepared for a secret passage by discussion dimensions, say, or making observations about the siz of a space, and I think that’s an important and valid piece of definig the parameters. I remember reading one mystery with a secret passage solution and thinking “But, there’s not enough room in the space you’ve described for a passgae to fit“…that’s the sort of thing that should be avoided.

      It’s mainly about ensuring that any aspects of the room are clear in their specifics where the reader is concerned. There’s much to no dislike in The Sinister Student by Kel Richards — here’s my review if you’re interested — but at the top of the list is the fact that the window locks itself because it’s opened with the victim’s left hand, whereas if if were opened by someone using their right hand it wouldn’t self-lock…and it’s just not explained what that means. That, my friend, that is not laying out your parameters.

      Rube Goldberg stuff vexes me, yeah, as does “The victim was stabbed outside the room and locked themselves in to protect themselves from the killer”, as does “a piece of string was placed around the head of the bolt and pulled once the door was closed to lock it”. Each have their place in the blocks of a solution — an episode of Jonathan Creek uses the second in a minor way very well — but if either of them are the totality of the solution, it’s unlikely I’ll be reading that author again.


  6. I can think of several novels by Carr that feature a killer that comes out of the blue at the end (I’m sure JJ’s read the two I’m thinking of). In both cases, it’s actually incredibly memorable, although for reasons that I obviously can’t get into here.

    I love the idea for the locked room rules – I’ll have to give that some thought. One that comes to mind immediately: the author really has to sell the impossibility. Grind it into the reader’s brain that there is absolutely no way that the crime could have been committed. Spend chapters dismissing every possible solution that the reader can think of.

    Too often I’ve seen an impossibility treated almost as a piece of trivia and forgotten about as the plot pivots to focus on other elements. Also, don’t provide an easy out, such as a partly opened window. That disrupts the entire puzzle, even if your clever solution doesn’t involve the open window at all.


    • There’s one impossible crime novel in particular where an open widow has no bearing on the crime…and, goddamn, it frustrates me so much. I know that’s often used as a piece of misdirection, but that needless suspense of whether or not it’ll turn out to be an impossible crime is infuriating. If I’m reading something that posits an impossible situation, I want it to be clear about that impossibility.

      Grind it into the reader’s brain that there is absolutely no way that the crime could have been committed. Spend chapters dismissing every possible solution that the reader can think of.

      Someone needs to read Death of Jezebel…


  7. Okay, I’ve got one! And it popped up a few minutes ago in the book I’m currently reading (The White Priory Murders), so I’ll let HM tell it to you himself:

    “The first thing is to determine the murderer’s motive. I don’t mean his motive for murder, but for creating an impossible situation. That’s very important, son, because it’s the best kind of clue to the motive for murder. Why’d he do it? Nobody but a loony is goin’ to indulge in a lot of unreasonable hocus-pocus just to have some fun with the police.”

    There must be a reason for the impossible nature of the crime!


  8. Since Brad mentioned the motive for the criminal creating an impossible situation, that reminded me of something.

    I’ve read a few reviews recently where the reviewer has mentioned that a crime in the book they were reading was not an impossible crime, but could easily have been made into one.
    What are your thoughts on this? If an author can set up an impossible crime, should they? Is it worth adding an impossible crime if it’s not meant to be the focus of a book? If it isn’t the focus, it won’t really get the time it deserves – as in thegreencapsule’s comment above. Or it may well end up using the old old tropes of key-on-a-string etc. So for a rule could you say that the story should focus on the impossibility as its main aspect?

    Mostly, since I haven’t read the books being discussed, I want to know what kind of situations could be modified so easily. Have you ever come across things like this in your reading?


    • “I’ve read a few reviews recently where the reviewer has mentioned that a crime in the book they were reading was not an impossible crime, but could easily have been made into one.”

      Hrm, being one of those reviewers who say this regularly I feel moved to reply. 🙂

      I love impossible crimes, and I think that a mystery becomes better simply by including one. In most of the cases where I use the above comment, I generally feel that the author has already created a situation where their solution could actually be used for the impossible situation with perhaps just a tiny bit of tweaking.

      One really good example of this is Sapper’s “Mystery of the Slip-Coach” (recently published in the anthology “Blood on the Tracks”). It’s just a hair’s width from being a true impossible crime, and Sapper wouldn’t even have had to change his solution to make it into one.

      However, it goes without saying that if the author would have no clue how to resolve the impossibility in at least a vaguely creative way, then yes, certainly the story is better off as it is.


      • Thanks for responding – I think it was one of your reviews I saw that comment most recently. I’ll have to check out that collection so that I can see just what you mean.


      • The lack of creativity in a lot of solutions is very annoying, I agree. I don’t know who decides they’re being clever in putting in a solution that they devise with the least possible effort or coverage of a genre, but these people should be stopped. We need a sort of Wolfrom Alpha for detective novels where you can type in your idea and see how hoary it is:

        “How many mystery stories resolve themselves by the narrator not having told the audence something?”
        Too Many

        “How many mystery stories use gender neutral names to surprise readers with a man instead of a woman (or vice versa)?”
        It is 2019, please do not write this book



    • I don’t think there’s any “should” or “should not” here — everyone is free to write preciselyt what kind of book they like — but it’s interesting to see some authors walk right up to the line and simply not shuffle even a toe over it to venture into impossibilities. I assume they’re just unaware of what they’re doing, and it’s just because of a reader’s heightened awareness of possibilities that some people pick up on it.

      Consider the number of books where you could go “Ah, man, if they just changed tiney detail X then issue A would be resolved” — sometimes authors have to keep an eye on so many things that elements creep into the setup by accident. Carr needlessly tells a lie at the start of one book, and if you remove that line it improves the story immensely; Paul Halter mentions two or three tiny details that, if removed, would tidy the overall strucutre of about three or four books; Max Afford does it, Rupert Penny does it, Philip MacDonald and Norman Berrow over-share information and could have witheld clues that would render the mystery far more agreeable; plenty of authors have witheld clues that could easily have been included and improved their plots hugely…sometimes people just don’t realise what they’re doing, and for whatever reason it was either never pointed out to them or it was and they decided to leave it anyway. So it’s not just impossibilities, it’s almost every aspect of a plot.

      Regarding impossibilities, what I find frustrating is the way some are so unclear, how it could be that you can ignore — say — the open window, but it’s only when you get to the solution that you know for a fact ot had no impact and was, therefore, a genuine impossible crime semi-disguised as a “normal” one. There cold be narrative reasons for this — say, the killer wishes to leave the possibility of another method of death so as for the real method, which points right at them, to be undiscovered — but even that goes undiscussed and so the whole setup has been this frustrating will they-won’t they thing, and, while there’s a way to do it well, most times this is done they pay-off is not worth the uncertainty. No, I don’t need impossible crimes everywhere, but if you’re doing one there are ways to address it in the narrative. But then, as I say, it’s also true of every other element of a plot, too.


  9. I think there’s a tendency to regard Van Dine, Knox, the Detection Club and others as competing Moseses (Mosesi?), offering different (though largely overlapping) tablets of detective fiction rules, and thereby setting the parameters of the Golden Age. However, as you said in your podcast, they were actually already looking back, and I believe that such rules lists were merely a trendy formalization of till-then tacit understandings, responding to commonly shared (by both readers and writers) complaints of mystery stories of the times. And while the union of these lists produces a collection of rules quite numerous, an intersection of them reveals an attempt to address a very small number of types of complaints (outside of those directed against all types of fiction):

    “The clues weren’t all there.”
    “The author lied.”
    “The solution was too far-fetched.”
    “The solution was too obvious.”

    Further, these complaints result from the failure of works to fulfill a (n again quite small) number of common expectations that govern nearly all of the “rules”— expectations that greatly predate the Golden Age (indeed, date back at least as far as Aristotle’s Poetics [“these incidents are of greatest impact then they are unexpected yet in consequence of one another”]). Indeed, nearly all of the “rules” reflect these 5 tacit expectations:

    EXPECTATION 1 – The solution will be sufficiently indicated by the narrative which precedes it.

    There’s admittedly a vagueness to that precept (primarily due to the word “sufficiently”), but that is actually quite fitting, as there as an equally (though largely unrecognized) vagueness to the usual complaints against those works which don’t fulfill it:

    “The clues weren’t all there.”
    “I didn’t have a fair chance to solve it”
    “There weren’t enough clues.”

    The problem with the question of clue sufficiency (and the primary reason I rail against the notion of “fair play” in regard to it) is that people speak as if there were an objective standard, a threshold above which clueing is sufficient. Yet no one can identify what that standard is. It seems to me— based on critical responses to puzzle plot stories— to lie somewhere between some level of indication of the truth (i.e. at least one clue), and the exhaustive logical elimination of all other possibilities
    (total deductive provability). But where it lies on that spectrum I cannot say (and wouldn’t want to try, as I personally don’t believe such an objective standard exists).

    The reason I insist the standard is not a matter of total deductive provability is not merely because I don’t believe any detective story has ever reached that standard (though I do suspect that to be true), but rather because most of the detective stories that are most highly regarded for their clueing don’t even aspire to it. For instance, I myself would call the clueing of Five Little Pigs rich and extremely (subjectively) satisfying, and many genre critics would agree. But while a behavioral discrepancy clue such as “Everything tastes foul today” packs a satisfying wallop, there is no effort whatsoever to logically discount all other possibilities for it (which I don’t think could be done anyway). Indeed, Five Little Pigs is composed almost entirely of such unprovable clueing, yet is rarely the target of “the clues weren’t all there” complaints.

    No, what I believe most people mean when they say that “all the clues were there” is merely that “the truth seemed retrospectively inevitable”— “seemed” being the operative word. And that expectation of retrospective inevitability hails back to Aristotle.

    (Incidentally, such rules as “The reader must be granted all the clues given the detective” are essentially less comprehensive variations of “the reader is given all clues necessary”; for, we can presume that if the reader is not afforded all the clues given to the detective, he has not been given “all the clues necessary.” Moreover, merely being given all the clues afforded the detective itself offers no minimum guarantee, for many a detective story has a detective “solve” the case on little or no evidence which the reader was also given, fulfilling the lesser condition, but still not satisfying the reader that he was given “all the clues necessary”).

    Ultimately, many of Van Dines rules:

    “The reader must have equal opportunity with the detective for solving the mystery“

    “The culprit must be found through logic, not unmotivated confession“

    “The culprit must be prominent in the tale“

    “The solution should, in retrospect, have stared the reader in the face”

    All really demanding the same thing— that the truth of the solution be sufficiently indicated to the reader prior to that solution— though what constitutes sufficiency isn’t (and cannot be) delineated.

    EXPECTATION 2 – The solution will not only allow for, but also account for all that precedes it.

    This expectation is a variation on the general expectation of all fiction that no elements included in a fictional work should be inessential, that they all should serve some necessary function toward the primary purpose of the story. The “checkov’s guns” principle. And in the case of detective fiction, that primary purpose is presumably to arrive at a solution to the presented puzzle that satisfies by being surprising and yet also the apparently inevitable outcome of all that came before.

    Thus, I believe that Expectation 2 is the primary basis of the various rules prohibiting “love interest,” “character analysis,” “scholarly digressions,” etc. Presumably, the problem with such elements is that their presence in the story is not accounted for or deemed justified by the solution. And so, beyond just being unwarranted (as they might be considered in other non-detection works), their presence might be regarded as unethical obfuscations—a cheap form of misdirection.

    However, as you pointed out on your podcast, characterization can become integral to clueing, and indeed I believe there was point in the late Golden Age (I’m thinking the mid-1940’s) where clues and characterization became so interdependent (in certain works of Carr, Christie, Brand and others) that clueing and characterization essentially became indistinguishable from each other. And I’d further suggest that several such works were among their authors’ finest. Perhaps if the rulemakers had qualified their rules to disallow “unnecessary emphasis on love interest” and “excessive levels of character analysis” these rules wouldn’t sound so silly. Even so, I find Van Dine’s prohibition of “extended scholarly digressions” hilariously hypocritical, as no other author has ever wasted much of my reading time with erudition that had no crucial solution payoff.

    EXPECTATION 3 – The solution will surprise.

    I don’t believe there are any rules on these lists that specifically mandate that a whodunit solution must surprise or even attempt to, but it is certainly an expectation upon which the perceived success of such a work rests, and the source of major disappointment when it is not fulfilled. Thus, I think this expectation is the source of those various rules against “cliched,” “shopworn,” and “hackneyed” deception devices (Van Dine’s #20)- the problem with such devices being merely that, due to their overuse, they have lost their power to deceive, and thus destroy the solution’s power to surprise.

    EXPECTATION 4 – The author will not make false statements.

    Over the brief history of detective fiction, the reader understandably came to demand that— especially in a genre so loaded with lies and deception— there be at least one voice upon which he could place his trust. And being the one voice consistently bridging the fictional universe with that of the reader— the only one apparently cognizant of the reader’s existence— it is natural that the author be held as that one trusted voice. Though there have been many variations of narrative trickery since Roger Ackroyd and even before, it is ultimately that single expectation that governs reader acceptance.

    Van Dine’s embargo on deception on the part of the author is rather ludicrous— the reader expects (and always has) that the author to try to deceive him, and accepts that. What he does not accept is the author lying to him, making actual false statements. And such lies can only be considered inacceptable if they are from the voice of the author himself, i.e. a third person narrator (omniscient or limited).

    The surprise of Roger Ackroyd was clearly borne of a tendency of readers to confuse a first person narrator’s voice with that of the author, but once reminded of that distinction, there is no reason why there ever should have been any controversy. In the case of Van Dine’s disapproval, I suspect it was a matter of sour-grapes outrage because he hadn’t thought of a way to pull off the device himself. I myself consider the triumph of Roger Ackroyd not a matter of Christie’s “invention” of the idea (which we know to have predated her use of it, anyway), but rather her three-fold “insurance” against outrage: not only was the deception upon the reader committed by a character within the fictional world (not the voice of the author), and additionally the deception committed via omission rather than outright lies but, perhaps as importantly, Christie supplies her culprit with an entirely believable psychological motivation for wanting to deceive the reader. All bases covered.

    There have been many tour-de-forces of unreliable narration since that time— from Carr’s She Died a Lady to Ian McEwan’s Atonement, but I don’t think anyone has ever gone as far to hedge their bets as Christie did on that occasion. (Ironically, though Christie was remarkably “clean” with Ackroyd with regard to Expectation 4, such other less-controversial works as And Then There Were None and Death on the Nile are actually not so guilt-free in that respect— though admittedly only in minor, almost unnoticeable ways).

    I pointed out earlier that the numerous “rules” are governed by a very small number of expectations, and here I find that that number may indeed even be reduced, as Expectation 4 can be regarded as simply an extension of Expectation 2. After all, if we are told by a source we trust (the author) that X is true, and it we learn in the solution that x was not true, the solution is not accounting for the falsehoods in the prior narrative. In such a case, the puzzle’s syllogism may be valid, but not sound, as one of its premises (X) is not true.

    EXPECTATION 5 – The fictional universe will function under the the same natural laws as ours does, unless the work prepares us for a discrepancy.

    I believe the reason readers object to far-fetched solutions (whether they involve the introduction of the supernatural or merely extreme coincidence or unlikely motivation) is that these solutions are perceived as a cop-out deus ex machinas, a substitute for ingenuity. Where ingenuity is clearly evident, such complaints are much rarer. For example, though there have been complaints about the ending of Carr’s The Burning Court, very few people would call it a matter of “cheating,” as the novel has already exhibited an extraordinary amount of puzzle plot ingenuity. The final supernatural twist is seen as a stunning bonus reversal, not an easy out.

    The “unless the work prepares us for a discrepancy” qualification is also key. A work that combines the supernatural with puzzle plotting is generally accepted, if that work clearly posits the reality of the supernatural as a foundational premise (such as Dorothy Macardle’s “Uneasy Freehold,” the film “The Sixth Sense,” even the Harry Potter novels). It is only when an author suddenly introduces dematerialization as a way out of a locked room at the denouement— because he clearly had no other idea of how to get out— that the reader considers it a cheat.

    The same applies with extreme and unusual psychology. Perhaps if Christie had convinced me that a certain character had an extraordinarily poor memory I might have believed that she would have married the same man twice without realizing it. Perhaps. As it was, it was one of the few Christine novels that had me crying “bullshit!”— because people don’t usually think or behave like that in our world.

    It is a matter of sufficiently establishing these unusual premises in the puzzle proper, so that we don’t see them as an excuse for lacking ingenuity. And thus, much as I regard Expectation 2 as covering Expectation 4, I see Expectation 5 as a narrower-scoped reflection of Expectation 1: If an extraordinary-reality premise is sufficiently indicated prior to the solution, its use in the solution will not be objected to.

    I do feel many of the specific Golden Age rules date badly, or now seem unnecessary restrictive. And the “competitive game” analogy they’re aligned doesn’t hold up to scrutiny either—upon careful reflection, the differences between detective stories and competitive games are more numerous and fundamental than their similarities.

    However, the expectations that served as the basis and catalyst for the rules still hold: that the relationship between puzzle and solution should be non-arbitrary (the solution is both indicated by the puzzle and accounts for it) and yet still surprise.

    And no Chinamen.


    • With regard to expectation 5, the way that I interpret the ending of that Carr novel is that, since it is written from the character’s point of view, we are not obliged to treat it as real, since the character could be deluded. The ending still casts a new light on some of the events of the novel, but you are not forced to bring in the supernatural to explain it (unless you prefer to do so).


      • I absolutely agree— it’s something someone (I can’t remeber who) describes as the “Janus resolution”— offering two possibilities from which the reader can choose. The key point, I believe, is that the supernatural explanation does not appear as an “easy out” to escape a corner the author has placed himself in otherwise, an impossibility he cannot explain rationally. And certainly in The Burning Court, Carr has fully done his job to provide an intricate, well-clued rational puzzle plot solution, purposely leaving one niggling discrepancy to push our loyalties just enough toward the supernatural explanation. So that I, personally, don’t know which to to believe, but find both fully satisfactory.


  10. In the podcast you say, “if I wrote a novel where the solution was to do with someone who wasn’t present through the entire book to the last minute, it wouldn’t be a good story.” But like all the other rules, this too can be broken if the writer does so with enough skill. There’s a novel by John Rhode where you gradually realise that a certain person must exist (though they are not mentioned until very close to the end), and that if they exist, then they have a jolly good motive for the crime. This seemed to me very well done, and neither unfair nor unsatisfactory.

    I have a couple of rules for impossible mysteries, or indeed for mysteries in which it is merely difficult to explain how the crime was committed.

    The first is that the mystery should have a degree of robustness or depth that allows the characters in the book to consider and investigate obvious possibilities like keys and bolts being operated from the outside, witnesses lying or being mistaken, clocks being changed, and so on. It’s only through the characters’ investigation of the space of possibilities that we can really understand the constraints that the solution has to meet. For example, a story that turns on footprints ought to discuss the possibility that someone has worn someone else’s shoes, and the possibility that there might have been other footprints that were swept or raked away. If these possibilities cannot be discussed (because they would give too much away or advance the plot too quickly) then the mystery is not deep enough. This is a fault of a good few of John Rhode’s novels, where it takes far too long for obvious possibilities to be raised, because there is nothing very interesting behind them.

    The need for robustness was discussed on this blog back in 2016: “But one thing which I think authors often overlook and which doesn’t get much discussion is that there needs to be enough play in the puzzle for the characters to actively discuss it. Far too many mysteries introduce impossibilities early on and then barely discuss them until the solution, because the author knows that the setup is too flimsy to withstand much investigation. False solutions aren’t just useful because they allow for twists and double-twists, they also help prevent impossible crimes from being ostracised in the middle of their own plots.”

    The second is that the mystery should not rely on the investigators omitting steps that a reasonably conscientious person in their position would have taken, unless this is clued and explained. For example, if a police sergeant says to his constables at an early stage in the investigation of the crime scene, “Go over this room for any secret entrance whatever, understand? Tear the place to bits if you like. See if anybody could get up the chimney…” then in the absence of other hints, the reader ought to be able to conclude that the constables would not have missed an object “seven feet broad by four feet high” if it was hidden in the chimney. Or if the victim’s throat has been cut, but the detective fails to preserve and check the clothing of the suspects for blood transfer, the reader ought to be able to conclude that this part of the investigation has been omitted by the author because it is irrelevant to the case, not because the detective missed this line of inquiry.

    (Like many fair-play rules, this falls under Scott K. Ratner’s “expectation 1”.)


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.