Background

That Hideous Strength (1945), the third novel in CS Lewis’s Ransom Trilogy (1938–45), is frequently described as “a Charles Williams novel by CS Lewis” (Walter Hooper, CS Lewis: A Companion and Guide (1996), 231; Brian Horne, “A Peculiar Debt: The Influence of Charles Williams on C.S. Lewis” (1990), 93–95). Between 1930 and 1945, Charles Williams published seven novels: War in Heaven (1930); Many Dimensions (1930); The Place of the Lion (1931); The Greater Trumps (1932); Shadows of Ecstasy (1933); Descent into Hell (1937); All Hallows’ Eve (1945). In each of these novels, the supernatural impinges — sometimes violently — on the ordinary phenomenological world, and characters undergo spiritual growth or change because of the supernatural incursions.

Lewis was undeniably influenced by Williams, whom he admired, but an easy equation of the two as authors may be questioned or modified. Lewis first read a novel by Williams in 1936: the first novel of his trilogy (Out of the Silent Planet) was published in 1938 and the second (Perelandra) in 1943, but these novels are not described as being “Williams novels by Lewis”. The difference in reception may be because the first two novels are set on Mars and Venus, while That Hideous Strength has an earthly setting, like all seven of Williams’ novels (Hooper 231; Horne 94). In addition (as Lewis argued) the combination of the marvellous, fantastic, or supernatural with the ordinary was not a new thing, and Lewis and Williams, in much of their fiction, were continuing an existing tradition (CS Lewis, “The Novels of Charles Williams” (1982), 46–7, 49).

There are certainly elements in That Hideous Strength that are similar to elements in Williams’ novels. Some are particular acts of homage, others arise from shared interests, and others arise from the combination of supernatural and ordinary in the lives of the characters — and hence will share characteristics with all fiction of that kind. But this is still not enough either to praise or dismiss That Hideous Strength as “a Williams novel by Lewis”.

This is as far as ordinary literary critical argument can take us. But maybe some of the techniques of digital humanities can add weight to one or the other side of the debate. Stylo, an R package for stylometic analysis, may be able to show whether That Hideous Strength is closer to the other two novels of Lewis’s trilogy or to Williams’ novels. (Of course, Williams’ novels — published over sixteen years, but with the first five over four years — may prove not be stylistically homogenous: this would complicate and enrich the questions and answers that would follow.) And, whatever Stylo shows, exploration using desktop software such as Wordsmith Tools and/or web-based Voyant Tools would enable those showings to be explicated.

Preliminary findings:

The consensus of various sampling methods in Stylo is that the first two novels in Lewis’s trilogy cluster together, while That Hideous Strength clusters more closely with novels by Williams. These clusters vary with re-runs of random samples, suggesting that different parts of the novel exhibit different levels of closeness to Williams’ novels.

Preliminary word counts using Voyant Tools also show That Hideous Strength having more in common with Williams’ novels, while the first two of Lewis’ novels differ from all the other novels under consideration, and in similar ways. Word counts suggest that Out of the Silent Planet and Perelandra contain a somewhat smaller proportion of dialogue than the other novels. They also suggest that they contain a greater proportion of similes and metaphors. Questions:

The paper will conclude with the beginnings of answers to at least some of the following questions.

Granted that That Hideous Strength is stylistically closer to Williams’ novels than Out of the Silent Planet and Perelandra, in what ways is this so? How does That Hideous Strength differ from the other novels in Lewis’s trilogy? How is it similar to Williams’ novels – and is it equally close to all of them? Do different parts of the novel (which moves between three distinct settings and groups of characters) have clearly different relations to the other novels under consideration? Does Stylo generate the same results for the novels if they are compared in terms of grammatical structure (using n-grams and part-of-speech tagging) rather than lexical content?

How do the novels compare in terms of the ratio of narrative/exposition to dialogue? How do they compare in terms of their use of comparative language such as similes, metaphors, and analogies (this question may be applied to dialogue, narrative/exposition, or both)? Wordsmith Tools, especially the “Tagging and Markup” tools, will facilitate the answering of these questions.

Once these questions are answered, it will be possible to ask again whether, or in what way, or to what degree, That Hideous Strength might be considered “a Charles Williams novel by CS Lewis”. The discussion might then return to the domain of literary criticism — but greatly enriched; and the analyses may generate hitherto unsuspected areas for exploration within digital humanities.