Thursday, June 18, 2015

Eliot Elucidates: T. S. Eliot's Detective Fiction Rules

When discussing Golden Age detective fiction people often reference the "rules" for writing it set forth by mystery authors Ronald Knox and S. S. Van Dine; yet the set of detective fiction rules composed by T. S. Eliot, in the Twenties a great mystery fiction fan (not only did he like, as we are often told, Wilkie CollinsThe Moonstone, he enjoyed Golden Age puzzle-masters like R. Austin Freeman, Freeman Wills Crofts, J. J. Connington, Agatha Christie and S. S. Van Dine himself), were not discussed in detail, as far as I am aware, until the publication of my essay, Murder in The Criterion: T. S. Eliot on Detective Fiction. (The original version of this essay appeared in CADS in 2011, while a revised version was published in 2014 in Mysteries Unlocked: Essays in Honor of Douglas G. Greene; on the subject of Eliot's interest in detective fiction see also David E. Chinitz's T. S. Eliot and the Cultural Divide.)

Here are T. S. Eliot's five rules of "detective conduct," set down in 1927 (see The Criterion for the full version).

1. The story must not rely upon elaborate and incredible disguises....Disguises must be only occasional and incidental.

2. The character and motives of the criminal should be normal.  In the ideal detective story we should feel that we have a sporting chance to solve the mystery ourselves; if the criminal is highly abnormal an irrational element is introduced which offends us.

3. The story must not rely upon either occult phenomena, or, what comes to the same thing, upon mysterious and preposterous discoveries made by lonely scientists.

4. Elaborate and bizarre machinery is an irrelevance....Writers who delight in treasures hid in strange places, cyphers and codes, runes and rituals, should not be encouraged.

5. The detective should be highly intelligent, but not superhuman.  We should be able to follow his inferences and almost, but not quite, make them with him.

What do you think of T. S. Eliot's rules?


  1. Excellent rules for the period. And five rules is better than ten or twenty!

  2. I’m a firm believer in rules. If you look carefully at the rules propounded by Eliot and Knox and Van Dine it’s clear that their intention was to encourage creativity rather than to stifle it. They were saying that short-cuts and cheats don’t really work. If you work within a set of rules you’ll have to work harder but the results will be far more satisfying, to both the reader and the writer.

    I prefer Van Dine’s rules because there are more of them!

    The modern psychological crime novel is a dead end because there are no rules. That’s also why most modern “literary” fiction is uninteresting. Writers just start writing and when they get bored they stop.

  3. Well, they are somewhat longer in the original source, but you get the gist here. What's so interesting to me is that Eliot was caught up in the detective fiction craze enough to fashion his own set of "fair play" rules for writing it. They all trend toward that concept: Restraint in disguises and "elaborate machinery," no sci-fi or supernatural aspects, no insane culprits, no superhuman detectives.

  4. What strikes me about this is that it seems to want to disallow a "deus ex machina" solution or ending. And, while the term is no longer bound to its literal meaning, the "deus" reminds one that Eliot (as well as Auden) was quite religious. It's as if there could be nothing miraculous or mysterious (in the mystery play sense) about a mystery. Which, in turn, is interesting, because it seems to amount to an insistence on the secular nature of the mystery story.

    1. That's very well put. It's an emphasis on pure rationalism, isn't it? Of course I've always thought it interesting that the Twenties also saw, on track with the detective novel, the expansion in popularity of the weird story, where the supernatural is most definitely ruled in.

  5. Tom wouldn't have liked the Da Vinci code...