#592: Reflections on Detection – The Knox Decalogue 1: The Criminal

Decalogue header

So, here we go with an in-depth look at the principles espoused in the Ten Commandments for Detective Fiction as outlined by Ronald Arbuthnott Knox.

This undertaking is going to run for a while, I’ll only get through Rules 1 to 3 in the time remaining this month, so here’s where the links will go to keep everything tidy:

Rule 1: The Criminal

Rule 2: The Supernatural
Rule 3: Secret Passages
Rule 4: Undiscovered Poisons
Rule 5: No Chinamen
Rule 6: No Accidents
Rule 7: The Detective-as-Criminal
Rule 8: Declaration of Clues
Rule 9: The Stupid Friend
Rule 10: Twins

Rule 1, then, runs thus:

The criminal must be someone mentioned in the early part of the story, but must not be anyone whose thoughts the reader has been allowed to follow. The mysterious stranger who turns up from nowhere in particular, from a ship as often as not, whose existence the reader had no means of suspecting from the outset, spoils the play altogether. The second half of the rule is more difficult to state precisely, especially in view of some remarkable performances by Mrs. Christie. It would be more exact to say that the author must not imply an attitude of mystification in the character who turns out to be the criminal.

There are two key streams to this as I see it.  The first is in Knox’s determination of what a detective story is for, and while it’s difficult to agree with his assertion that GAD — which, as has already been stated, was well and truly a strong, recognised, going concern by the time Knox published this list in 1929 — “derives its excitement only from the danger of the criminal getting off scott free, or of some innocent person being condemned in his place” (because, let’s face it, we know in advance that the guilty will be identified — if sometimes allowed to get away with it), I think we can all agree that when we pick up a novel of detection we expect that it will “have as its main interest the unravelling of a mystery; a mystery whose elements are clearly presented to the reader at an early stage in the proceedings”.  This is why Knox’s list is more interesting to me than S.S. van Dine’s, because van Dine insisted on murder as the crime, where Knox appreciates that murder is but one of many mysteries that can be investigated, and the truth behind it detected by rational means.

We want to pick up a novel of detection knowing that we’re going to be told about a group of people — and ‘people’ is an interchangeable term, I’ll get to that — and a crime or mystery in their midst, and that by the end the solution of the crime will have been laid bare, that the guilty party will be one of those people we’ve been told about, and that the milieu in which they exist contains recognisable direction and possibility for the solution of that crime or resolution of the mystery.  When one picks up, say, The Dead Mountaineer’s Inn (1970) by Boris and Arkady Strugatsky and reads of baffling crimes in an isolated, snowbound hotel, this would appear to fit the bill of the novel of detection perfectly.  Come the end of that book — no spoilers, though I’ll advise quite strongly that the classical GAD fan give it a wiiiiide berth — the condition of the above expectation has not been met.  The solutions can in no way be anticipated, not simply through insufficient declaration — that’s Rule 8, we’ll come to that — but because who is responsible and how those crimes are committed is not something you would ever guess from reading the first, say, three-quarters of that book before it takes the turn it does.

In part, then, Knox is simply guarding against similar disappointments.  Declare your criminal early so that what is provided around them is an understanding of how they could — and would — commit the crime in question.  That “mysterious stranger who turns up from nowhere in particular” is to be dismissed simply because while there are thrills to be had in the irrational killer — we were still four years away from Philip MacDonald revolutionising the serial killer genre with X v. Rex (1933) — and the faceless, nameless, unknown who is able to act with impunity is, in many ways, no more fair to the reader-as-detective than any number of criminal conspiracy members who turn up in the book once, do their menacing, and then drift of stage left before Mr. Big, who’s had no hand in anything before the denouement, turns up and expostulates over a whisky or a fine cigar.  Such distance lends itself ease from a writing and plotting perspective, but the cry of delight is “He’s behind you!” for a reason.

To a certain extent, this is the converse observation that I once made about The X-Files — though, really, that’s more Rule 2, see next week — the classic detective fan would get no satisfaction from that show because every single week the answer was vampires or alien abduction, or a murder committed in a sealed room by a man who can stretch himself impossibly thin (as a youngster, Eugene Tooms gave me the willies, man…).  Equally, The Outer Limits posited in its classic opening narration a sort of parallel universe in which we’d see all manner of inexplicable mysteries and demons…and so if Bill Shatner watches a gremlin tearing at the wing of his plane in ‘Nightmare at 20,000 Feet’ (1963) and it turns out to be some guy dressed up to mess with him because his wife wants to get him sectioned so she can have access to the fortune he’s withholding…well, it’s not quite the “journey into a wondrous land whose boundaries are that of imagination” Rod Sterling promised up front.

Which is not to say that demons and gremlins are out of bounds, which is what males “people” an interchangeable term.  “Characters” is better — in the pure sense of “those in the mise en scene” rather than “those fully-rounded individuals” one typically means when the c-word is bandied about — since a novel of detection need not be about people per se, so long as the world into which we’re taken is filled out with sufficient understanding provided about how and why the crime or mystery came to be and was achievable in that setting (Stanley Ellin wrote a great story involving fleas, for pity’s sake).  One needn’t write in a recognisable world at all — scrap gravity as Poul Anderson did in ‘The Martian Crown Jewels’ (1956), or all Newtonian Physics as Isaac Asimov did in ‘The Brazen Locked Room’ (1956) — so long as the reader is brought up to date with what’s unrecognisable about the setting.   Which is not to say that you’ll write a good story by doing so, but that, by telling us about it, at least you’ll have written something that qualifies as detection.  As Knox says:

[O]ne would suppose the first writer of detective fiction to have been a scientist, who, giving up the riddles of his own craft, which either defied explanation or opened out fresh vistas of problems demanding fresh explanations, determined to set himself a problem which he could solve, because he and no other was responsible for the inventing of it…

What we’re hoping to guard against is the scientist — the author — being the only person who can solve it; we’re straying into Rule 8 again a little here, but we’re also establishing a precept that must define the nature of this fiction: you’ve got to give us a chance to know where the solution comes from, both in performance (that is, what is allowable in the universe — walking through walls, magic spells, etc) and provenance (that is, who does the deed).  Hiding behind unclear rules or a sudden retraction of what’s possible in the setting is as bad a piece of writing as is simply hiding the guilty party off-page until it’s time for them to be guilty.  The first is reason enough for me to fail to warm to The Dead Room (1987) by Herbert Resnicow (sorry, TomCat) and the second is…well, it’s #2 in those two streams I mentioned above.


“I’m scared…”

Stream Two is concerned more with Knox’s second half and the idea of “an attitude of mystification” on behalf of the guilty party.  He acknowledges some of the “remarkable performances by Mrs. Christie” unaware, of course, that perhaps the most remarkable performance by Christie in this regard was still to come (as Brad wrote about so superbly here), with that being one of three novels in Christie’s career where we’d be given at times front row seats for the murderer’s thoughts.  It’s examples like this that make people sigh at the notion of hidebound rules and adopt a weary, all-knowing air to impart some aphorism along the lines of “rules were made to be broken”…but, very aware that all the cases I’ve cited thus far came after Knox’s list was written, I’m going to suggest there was a far more teleological intent at work here, and to put this in some appropriate context.

The criminal being someone mentioned in a early part of the story is, in many ways, the key point in GAD — what I will call the Democratisation of Suspicion.

In the suspense and sensation fiction that preceded detection, there was rarely any doubt who the guilty party was.  I mean, spoilers for The Woman in White (1859) by Wilkie Collins — but, like, not really — the second you hear of Sir Percival Glyde in that book, you know full well he’s a bastard who’s been up to all manner of shenanigans:

She was engaged to be married, and her future husband was Sir Percival Glyde. A man of the rank of Baronet, and the owner of property in Hampshire.

In many cases, the suspense of these stories came not from the identity of the villain, but more often than not from the lop-sided social structures that gave an obvious bounder a position of unimpeachable, impregnable security from which to continue to practise his (it was pretty much always his, since the women are usually too busy being oppressed and having a fit of the vapours) nefarious deeds to propagate his wealth, power, and/or influence.  The suspense came, then, from how someone of almost inevitably lower social standing, from a more impecunious background, and without the connections to power represented by the villain, could ever hope to overcome such obstacles and win fair lady’s heart, big house, and associated social standing (though it usually involved death in a fire).  This same bones can be found in pretty much all thriller fiction since the 1800s, from The Count of Monte Cristo (1844) to Die Hard (1988) and beyond.

Many years ago, back when I still had patience for that sort of thing, I read The Castle of Otranto (1764) by Horace Walpole, under the impression that it might shed some light on the origins of the mystery story.  The instigation of its plot concerns a young man dying mysteriously in a way that implies foul play on the morning of his wedding, and his father deciding that he’ll marry the intended bride instead.  It develops from there into a sort of po-faced version of The Princess Bride (1973) by William Goldman, full of improbable nonsense and sudden revelations about identity that become increasingly dizzyingly unlikely, though it’s insane enough to warrant a look if you’re curious.  The second Manfred appears on the page — all discontent with his sickly son and horny for his sickly son’s betrothed, Isabella — you know full well that he’s the bastard in this tale, and again the suspense is to be wrought from whether he, a titled member of the landed gentry, has sufficient gall and influence to get what he wants, or whether Isabella — lacking these things, and being a woman to boot — will be saved from his clutches.

Part of the genius of Charles Dickens was in taking this principle and removing the social inequality.  No-one reads, say, David Copperfield (1850) and trusts Uriah Heep at first meeting, with his blowing into the nostrils of a nag as if putting under some kind of charm.  When it turns out Heep has, well, er — do I really need to avoid spoilers for David Copperfield!? — been a bit underhanded, no-one is shocked excepting those who viewed such unprincipled behaviour as unlikely coming from a lackey and menial who should really Know Their Place.  And Arthur Conan Doyle built, I’d argue, on Dickens’ careful flattening out of the land, ensuring that the titled could now easily be the victims as much as the malefactors, even if they were often victims at the hands of other titled betters and not some grubby hoi polloi who was too busy tugging their forelock to even contemplate crime against their social superiors.  Edgar Allan Poe — whose work will be the subject of examination on this blog next month — equally took to ensuring that the crimes he has C. August Dupin and others investigate occurred at all social levels, and that the victims were not merely subject to the whims of the elite so that crime would forever take on a morganatic aspect.

What Dickens, and Doyle, and Poe, and others too numerous to mention failed to do, though, was to make any of this wrong-doing especially surprising.  If you can read Doyle’s ‘A Case of Identity’ (1891) and be shocked by the revelation concerning Mr. Hosmer Angel — still the best name in the Holmes canon — or the conduct of Charley Goodfellow in Poe’s ‘Thou Art the Man’ (1844), I have a bridge covered in flying pigs I’d like to sell you.  You start those stories and, as outlined above, you know exactly who the antagonist is, and probably exactly what they’ve either done or are going on to do.  Now, yes, calm down — this is doubtless on account of modern awareness of genre tropes, and at the time such ideas would have been more surprising to contemporary readers.  I know.  Let’s be friends, okay?  The point is, that if it’s obvious to us now, there’s absolutely no reason that sort of thing wasn’t already getting old by the time detection came around, and that Knox wasn’t rightly conscious of the linger hold Victorian sensation fiction had on these novels birthed, as it were, from the same motivations.  As Knox says:

[T]he stories become cleverer and cleverer, but the readers are becoming cleverer and cleverer too; it is almost impossible nowadays to think out any system of bluff which the seasoned reader will not see through. Thus, in the old days, when a woman was found very uncomfortably bound to a chair, with her mouth gagged, and possibly only just recovering from the effects of an anaesthetic, we used to suppose, not unnaturally, that she had been tied up like that by villains. Now we assume as a certainty that she is in league with the villains, and all the tying – up business was merely a plant; we have had the old bluff worked off on us so many times that it fails to take us in. Again, when the room is found covered with fingermarks or the lawn with foot – prints, we know at once that these are false clues, arranged by the criminal so as to throw suspicion on an innocent person.


“Go on…”

And so with detective fiction, the purpose became to remove that inevitability, to take away from simply The Most Suspicious Person the eventual revelation that they’ve been up to shady stuff.  Now, thanks to the social ironing done by Dickens and Poe, the possibility of blame could lie anywhere, with anyone we’ve met at any previous point in the book.  Hence, suspicion had been democratised, because now we had the task of knowing that wrongdoing could come from any angle, from any quarter, and the acknowledged game on the author’s part was to hide this while knowing the reader is on their side looking out for any indications to which a palpable sniff of mistrust may be attributed.  And the longer the reader has had to make this link, why, the more delightful the surprise come the end.  Yes, the author could exclude this person from the narrative…but, as outlined above, where’s the enjoyment in that?  The real test of skill, and the true appreciation of an author’s talent, is in providing plenty of chance to allow suspicion to linger and, in the most artful cases, to be fomented and then dismissed erroneously.

Where the author “must not imply an attitude of mystification in the character who turns out to be the criminal” comes into this is, I hope, obvious, since it’s far and away the shortest cut to giving someone the appearance of innocence: allowing us to believe that someone would be thinking on how these crimes could possibly be done when they themselves are the killer and so know precisely how they were done…to what end would that be the case?  Only if, meta-fictionally, they were aware that they were a character in a detective story and that the armchair detective at home was only now having access to their thoughts.  Arguments could be made that this sort of book was indeed written in The Demolished Man (1953) by Alfred Bester, but I’d suggest firstly that anyone so inclined consider how wide they’re willing to throw the ‘mystery’ net and secondly that the entire first section of this essay dealt with exactly the possibility of a universe in which such things were done.

In conclusion, then, I suggest that Knox’s observation of this rule was about establishing the baseline from which the detective novel should be built.  Next week we’ll look at the way authors would often straddle a genre, and how while the uncertainty over what would result could be seen as a form of game-playing, Knox’s intention — and it’s one that I share — was that the precise nature of what a reader is getting up front should be clear, that the game should be initiated with an equal footing on both sides, and then it may be enjoyed and appreciated fully by both players.  In unburdening the genre of the trappings of its sensation-riddled past, Knox frees us up for a world in which it is not merely glowering uncles and gloomy denizens of moody castles who have the facility to commit crimes for their own gain, and instead hands the poison, the dagger, the poker, the incriminating letter to even the least of us, to empower and surprise along the way…and to then face the consequences of those actions, that jealousy, that rage.  Even your occasional luncheons with the Chief Constable can’t save you now…

13 thoughts on “#592: Reflections on Detection – The Knox Decalogue 1: The Criminal

  1. I dunno, I dunno, I dunno . . . I see Knox here as a critical arbiter of popular detective fiction, possibly a literary guide, but I wouldn’t give him credit for actually shaping the GAD experience. That had happened much earlier! In other words, Knox gave voice to “rules” that others had already invented, perfected, and toyed with. You gave the wrong example for Dickens, since Uriah Heep is of that standard characterization where the name reveals all. (There’s a tutor in Our Mutual Friend named Bradley Headstone who, it would surprise you not at all to learn, is a depressing fellow indeed!) But turn to Bleak House and nestled within all that metaphorical fog is a whodunnit with multiple suspects, the indefatigable Inspector Bucket investigating, and a twist ending. And if The Woman in White is pure Gothic melodrama, surely The Moonstone is more of a closed circle mystery.

    So I don’t disagree with the rule, as evidenced by my flinging the occasional volume against the wall that has the audacity to withhold introduction of the murderer until the end. I just think that Knox is stating the obvious here. I do appreciate the distinction between Van Dine and Knox as to the former ruling out all but murder as a viable centerpiece (see Collins above, as well as that bloody Gaudy Night), but can I help it if i feel cheated if there’s not a murder?

    P.S. Please forgive me if this sounds like one of those early Superman fanboys who complained to the editor about Panel 4 or page 16 mis-coloring the hero’s cape blue, but you seem to have confused The Outer Limits with Rod Serling’s infinitely superior The Twilight Zone. Rarely did the former transcend its sci-fi plot mechanics with any sense of irony or hidden meaning. The fact that Serling could make us think in not one, but two episodes, through the emotive laxative that was William Shatner is nothing short of miraculous!


    • Oh, hey, I’m not suggesting that Knox is responsible for shaping the genre — did you not read last week’s post? All I’m saying is that these are the principles he’s observing and reflecting on. I even said in the above that the fundamentals of GAD were observed fully by this point. Knox isn’t shaping anything, that’s never going to be an argument of mine here.

      Bleak House…yeah, I acknowledge that I could have chosen better, but that’s very much the exception within Dickens’ output, and indeed within the genre of this writing at the time. I’m gonna call that the overwhelming majority of Bad Guys were durned obvious, and the democratising of suspicion was the key principle in the development of detection. Obvious it may be to us now, but in context — as the weeks ahead will show — there’s a lot going on at the time which made this formalising interesting even if not original.


    • Critic Jan Broberg at one point said that a mystery without a murder was all right in a short story, but a novel without one seemed a bit lacking. I think I might agree.

      Liked by 1 person

  2. The second aspect (presenting the inner thoughts of the killer) interests me greatly. It can’t happen with first-person (single narrator) unless the dude is really not talkative with himself, lying diarists aside. There’s memory loss/mental illness, but that is an unreliable narrator–not a very fair way to play.
    However, first person with multiple narrators (say switching every chapter) might not break the spirit of the rule–especially if the technique is applied throughout and enough characters get their own chapter. (democratization of PsOV??) As I’m not nearly as well-read as most of the readers of this blog, I’m sure there are tons of examples of this.

    It goes without saying that third-person narration (limited) wouldn’t enter the head of the killer and therefore does not apply to Knox’s principle. Even with my limited reading, I’ve run across this technique more times than I can remember.


    • There’s a great (terrible) example of a person being “unaware” who the killers is in the Halle Berry/Bruce Willis movie Perfect Stranger: Berry spends scenes creeping on her own around terrified that a killer is after her…and at the end it turns out that she is the killer (yeah, that’s a spoiler, I just gave you two hours of your life back). But, like, if she’s the killer…for whose benefit was she pretending to be scared?

      I feel a bit like this when some shenanigans are tried with admission in the same way: obviously the killer will pretend ignorance when others are present, but when the actions of a person not being observed are justified purely on the basis that it builds suspense for the reader…that’s a paddlin’.

      Liked by 1 person

    • I think what Knox means is you cannot have Brad think, I wonder who killed JJ when it was Brad who killed JJ. Or similar thoughts, such as guilty Brad wondering if Ken were guilty. There is one novel that flouts this rule in an extremely clever fashion, but to name it would be spoilerization. We all know which Christie flouts the cruder form of the rule …

      Liked by 1 person

    • Great poll! I have already voted. Coincidentally, I recently started a EQ-focused blog. Click on my name if you’re interested. I also suggest that someone should post the poll on the Facebook GAD group to maximize the number of votes.


  3. I think there’s some more to be said about the very first bit of the commandment — “The criminal must be someone mentioned in the early part of the story” — because there are some good detective stories that manage to break this rule without leaving you feeling disappointed by the appearance of a stranger in the last act.

    There are a couple of Freeman Wills Crofts novels (“The Cask” and “The Sea Mystery”) in which the early part of the story concerns the discovery of the body and the detective’s reconstruction of its trail, and only at the mid-point of the book are any suspects encountered. That’s because an early Crofts mystery is generally not a case of picking the guilty party from a group of plausible suspects, but of unpicking the mechanism of a complex cover-up in order to find any plausible suspects at all.

    And there’s a novel by John Rhode, which I won’t name for obvious spoiler reasons, in which the murderer is not mentioned at all, until the very last few pages, and yet it is quite fair — as you read the book you gradually come to realise that, given everything else that has been discovered in the course of the investigation, a certain person surely exists, and if that person exists, they have a very compelling motive for the crime. Rhode is commonly considered a rather dull and un-innovative writer, but has anyone else attempted this kind of device?


    • Gareth, you might be interested in this post on your opening point that I wrote back in August.

      There’s a Christie novel which does a similar kind of thing to what you describe in that Rhode book — effectively leading you to believe in the existence of someone who would benefit from the crime and so must have committed it…but the difference is that, well, we’re not introduced to them especially late-on, shall we say (although, depending on your definition, we’re also introduced to them very later indeed). But of course Christie did it. I take your point that it’d be rare enough in the genre, and I now really want to read that Rhode!


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 )

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.