This summer’s idiosophical conceit is that the novelist Charles Williams 1 communicated pointers to occult rituals in his novels in concentric circles, mimicking the degrees of initiation into a secret society. That he put hints of his occult practices into his novels is well established. (See for instance, the introduction section of Gwen Grant’s 2021 paper.2) What I’m looking at is the structure of how he did so. If this idea checks out, the books will show stylistic or structural features arranged in three levels.

As an initial hypothesis, we could look for a structure like

  1. The Outer Circle: A reader who knows nothing about the occult3 can read the novel as a murder mystery or an Indiana-Jones-style adventure story and not worry about it. But a reader who’s interested can find clues that there’s something deeper in the book, if they’re willing to dig. This is obviously true — is it done by machine-detectable stylistic cues?
  2. The Initiate: Williams’s novels are “an attempt at active communication of an established body of knowledge in symbolic form.” 4 Some occult symbols aren’t as obvious as an ancient relic or an alchemical formula. Esoteric practices were popular in Victorian England; in the 1930s and 40s memories remained as cultural touchstones. Collecting a list of these and searching for their presence in the text could reveal patterns. I’m not talking about the object of power that activates the plot. Rather, there will be tools or ideas that the characters use to get out of the plot alive.
  3.  The Adept: Each novel has a character who possesses an inner tranquillity that gives them almost supernatural power over events. Sørina Higgins calls them “jedi masters“. Perhaps these characters personify the innermost circle.

The singular feature of this hypothesis is that the inner circles have already been done, to a great extent. Previous scholars have written entire books about the cool parts. Grant: “Gavin Ashenden’s Alchemy and Integration (2006), which focuses on alchemical symbolism in Williams’ novels and, more recently, Grevel Lindop’s biography Charles Williams: The Third Inkling (2016), which provides a fully integrated discussion of Williams’ occult activities alongside the other events of his life. Aren Roukema’s Esotericism and Narrative (2018) has continued this exploration to include Kabbalistic symbolism, the role of the imagination, and specifically the influence of The Fellowship of the Rosy Cross rituals in Williams’ novels.”

So all I have to do is the first stage. Williams is famous for “the obscurity of his prose style.” It’s reasonable to suppose that the more obscure parts of the books are a signal to the reader that a deeper meaning lies beneath. Can we measure obscurity? That would have to be defined specially, but maybe complex occult ideas require complex writing. We can measure complexity of a text with canned algorithms. There are lots of them. Probably the most famous is the Flesch-Kincaid formula that measures the effective “grade level” of a text.

Figure 1 shows the results if we break up all of Williams’s novels into chapters and measure each one separately. There’s a magenta dot for Williams’s chapters. Dots in other colors are measurements of four comparison novels from the same historical period: two adventure stories and two murder mysteries. The adventure stories are Captain Blood by Rafael Sabatini and Meet the Tiger by Leslie Charteris (the first story about “The Saint”). The mysteries are The Murder at the Vicarage by Agatha Christie (the first “Miss Marple” story) and Clouds of Witness by Dorothy L. Sayers (a “Lord Peter Wimsey” story). I guessed that an obscure novelist like Williams should make a clear contrast with pop-culture fiction. I was wrong.

horizontal dot-plots of the grade level for each chapter of the selected books. Each book's distribution has a peak and a long right tail. Left tails can be long or short.

Fig. 1. Chapter-by-chapter reading grade level

All the novels vary from chapter to chapter in the complexity of their prose, but some trends are clear. The Miss Marple story is the easiest to read, with its chapters tightly bunched between 4th and 5th grade. Lord Peter Wimsey is second easiest.5 The two adventure novels are indistinguishable from Williams’s novels using this measure. That’s going to take some explaining. Another unexpected thing: note that Figure 1 is sorted by writing date — Williams’s later novels are written at a lower grade level than he used earlier in his career.

At one level, this is easy to understand. Before the occult came into the calculation, Charles Williams wanted to make money by selling books. Charteris and Sabatini sold very well a few years earlier, so it makes sense to follow their lead. Shadows of Ecstasy was written first, but War in Heaven was published first. It’s easy to see the shift to simpler writing between the two; the difference may have been important to the publisher.

Looking a little deeper, it would be useful to have a more direct metric for comparison. We find that when we restrict ourselves to novels from a short period in one place, lots of the ways to measure complexity of a text agree with each other. For this sample, grade level is 99% correlated with sentence length, which is much easier to calculate, so let’s switch to that. It turns out that averaging over chapters makes us miss an important feature of Williams’s prose style. Most of the time, Williams writes very directly. Here’s a randomly-selected paragraph, from The Place of the Lion, ch. XIV:

She went into the kitchen and made herself some sandwiches, considerably to the maid’s astonishment. She even attempted a little conversation, but she was feeling so shy that it was not altogether a success. The maid, she realized, was very much on her guard. That was the kind of world that Damaris Tighe had hitherto insisted on making all round her, a world where people were watchful and hostile. She looked at it humbly while she finished the sandwiches; then she went upstairs to her father.

While not as simple as Christie’s prose6, this is easy to read, and it’s what Williams does most of the time. But occasionally he cuts loose. Figure 2 shows what I mean. It shows Meet the Tiger and The Greater Trumps together. The horizontal axis is the number of words into the novel. The vertical axis is the average sentence length, where the average is computed over a sliding window 1000 words wide.

The rolling average looks like a seismograph, with a giant earthquake in the CW novel.

Fig. 2. Rolling average sentence length in two novels.

The Saint’s adventures are documented with a steady sentence length of 20-25 words, with occasional sallies that stretch sentences out 30 or more, usually for humorous effect.7 Williams, indicated by the red curve, mostly uses paragraphs of short sentences, but these are punctuated with occasional sallies into epic verbosity. That giant spike was caused by a 242-word monster of a sentence, describing a box of marionettes corresponding to the cards of the tarot.

So this is the first place to look. The next step is to correlate the long sentences with some more intuitive marker of occult content.

 

Notes

  1. Notorious victim of the entry barrier to literary research
  2. Grant, Gwen. “Sound and Ritualistic Language in Charles Williams’ War in Heaven.” Aries, vol. 23, no. 2, Dec. 2021, pp. 169–205. DOI.org (Crossref), https://doi.org/10.1163/15700593-20211006.
  3. On account of it being occult.
  4. Grant, p. 182
  5. After I deleted a couple pages of French that the computer marked as “rare words”.
  6. I gave a little shriek; it sounded so awful, and then–the person laughed. Laughed! Distinctly. And that was all. I heard him hang up the receiver.
  7. One paragraph I recall, from much later, began “The present writer, whose Spartan economy of verbiage can not have escaped the notice of even the least attentive reader, …