Stylometric Analysis of the Sherlock Holmes Canon

I have put together a technical paper regarding my stylometric method for author identification. I provided a summary of results from that method in Shadow Woman. In the technical paper, I get into the gory details of the method and I provide an example calculation, thus providing all the information necessary to anyone who wishes to recreate my work and challenge my results.

Also included in the paper are two separate accuracy analyses, each indicating an author prediction accuracy of 97% or greater.

Also included in the paper a twelve appendices, the last of which provides the calculated author for each of the Sherlock Holmes adventures. I’ll prove a quick summary for those not wanting to read the entire paper, or even scan to the bottom for the final appendix.

Here we go.

Louise and Arthur co-wrote the Sherlock Holmes portion of the first Sherlock Holmes Adventure, A Study in Scarlet. Arthur wrote the Utah narrative of that novel.

Louise wrote each of every other early Holmes adventure, up to and including The Hound of the Baskervilles.

Louise wrote two of the intermediate stories, those collected in The Return of Sherlock Holmes. Arthur wrote two of those intermediate stories.

Jean wrote the remainder of the Sherlock Holmes adventures.

Navigate to the Authorship page of this website for a link to the paper, or simply click here: Stylometric Analysis of the Sherlock Holmes Canon.

Another Arthur Denier

Martin Gardner was not the first person to question the authorship of the Holmes adventures. Almost a decade before Gardner, Sherlockian scholar D. Martin Dakin, in his Sherlock Holmes Commentary (1972), made obvious his suspicions about the authenticity of the last collection of Holmes adventures. In the chapter entitled “The Problem of the Case Book,” Dakin wrote:

This, the last series of Holmesian adventures demands a general consideration […] since serious questions both of authorship and authenticity arise.

Dakin declared two adventures, The Blanched Soldier and The Lion’s Mane (both from 1926), to be “pseudonymous stories” and “pure inventions.” He declared another, The Mazarin Stone (1921), to be “spurious.” He held one story in particular, The Three Gables (1926) in such low regard that he painted it with all three of these insults, declaring, “Never has a feebler story insulted the memory of the great detective.”

Dakin argued that nine of the twelve stories in the final anthology may have been written by someone other than whomever wrote the early stories. He stopped shy of naming names, however, declaring such an undertaking to be “a mystery worthy of Holmes himself.”

To be clear, Dakin wrote his book as a Sherlockian scholar playing a long-accepted game. The game presumes that Holmes and Watson were flesh-and-blood individuals, that Watson wrote the adventures (at least most of them), and that Arthur Conan Doyle was no more than Watson’s literary agent.

Sherlockians are naturally reluctant to accept any challenge to the integrity of the Canon. Walter Pond, for example, responded to Dakin with his essay “A Plea for Respect for the Canon.”

All Sherlockian scholarship must be based on the thesis that […] the stories which the Canon comprises are reports of the great detective’s cases written by Dr. Watson, and that Sir Arthur Conan Doyle was Dr. Watson’s literary agent. Once this thesis is abandoned, the foundation of Canonical research is undermined and Sherlockians will find themselves adrift on uncharted seas.

For this reason, it is deeply disturbing that some commentators have arrogated to themselves the right to determine […] which of the stories are genuine and which are not […] The conclusion would be inevitable that Sir Arthur […] participated in a fraud when he placed the stories for publication. This, of course, is unthinkable.

In just a few posts from now, we shall identify one such adventure, discussed in detail in Shadow Woman, that Arthur clearly did not write. According to Walter Pond, this will set Sherlockians adrift in uncharted seas.

Arthur and Magical Thinking

From 1918 until his death in 1930, Arthur wrote eleven books proselytizing spiritualism, plus numerous pamphlets, articles, and letters to make his case. To spread the word, he lectured throughout Britain, the United States, Canada, Australia, New Zealand, and Africa. He opened his own psychic bookstore.

Strangely, he established a fractious friendship with Houdini, the great magician who spent so much time debunking claims of the supernatural, explaining and demonstrating how so-called psychic phenomena were nothing more than crude magic tricks. Though Houdini refused to reveal how he accomplished his own tricks, he never denied that they were accomplished by natural means.

Arthur, being unable to conceive of how Houdini accomplished his many escapes, simply concluded that Houdini was one of the great mediums of all times who accomplished his feats by supernatural means. In fact, Arthur deemed his insight to be a great discovery, one of which the world should be made aware. To Houdini’s biographer, Harold Kellock, Arthur wrote, “I think, however, that you may take the words ‘An Unsolved Mystery’ off your cover. It is I who have solved the mystery of Houdini and I have no more doubt that he used psychic powers than I have that I am dictating this letter.”

Arthur was fooled not only by young women with unusual appendages, by girls with scissors and a camera, and by a dog that wouldn’t stop yipping, he was fooled most thoroughly by himself. The creator of Sherlock Holmes must have been someone less silly and gullible than Arthur Conan Doyle.

Arthur’s Fairy Foolishness

In 1922, Arthur put his name on the cover of The Coming of the Fairies. Edward Gardner contributed substantially to the book, but Arthur did not acknowledge him as co-author. The two intended that their book would prove the existence of fairies worldwide, focusing in particular on five photographs taken by young cousins Elsie Wright and Frances Griffiths. Three of those photographs are shown below.

The first of the five photographs, taken by Elsie Wright in 1917, shows Frances Griffiths with the alleged fairies. (Public Domain)

“The series of incidents set forth in this little volume,” wrote Arthur, or his ghost writer, “represent either the most elaborate and ingenious hoax ever played upon the public, or else they constitute an event in human history which may in the future appear to have been epoch-making in its character.”

The fourth photograph, "Fairy Offering Posy of Harebells to Elsie." (Public Domain)

Of the two options offered by Arthur, most ingenious hoax ever or epoch-making event, Arthur of course went for the latter.

The second of the five photographs, showing Elsie with a winged gnome. (Public Domain)

The photos, as it turned out, were neither an epoch making event nor the most elaborate and ingenious hoax of all time. For the first photo, for example, the girls simply cut out (or otherwise copied) some dancing girls from a Claude Arthur Shepperson illustration (found in Princess Mary’s Gift Book,1914) and posed with them.

Three dancing figures

 

Montage of Cottingley Fairies and illustrations from "Princess Mary's Gift Book." (Wikimedia Commons)

Near the end of her life, Elsie explained: “What we did was a long hat pin that we put down the back like that, and stuck the tape at the back like that, […] and then we wormed that down into the earth,” she said. “The thing was that they said they could see that the fairies were moving when the photographs were taken, but that’s because they were in a breeze.”

“I never even thought of it being a fraud,” said Frances. “It was just Elsie and I having a bit of fun. And I can’t understand why they were still taken in. They wanted to be taken in. People often say to me, ‘Don’t you feel ashamed that you made all these poor people look fools; they believed in you.’ But I don’t, because they wanted to believe.”

Elsie added, “Two village kids and a brilliant man like Conan Doyle – well, we could only keep quiet.”

So far, after only four posts, we have Arthur being completely taken by two young girls, an old and excitable fox terrier, and a beautiful young woman with an ectoplasmic phallus. That’s hardly Sherlockian behavior.

Martin Gardner’s argument seems to have merit. The skeptical and observant mind of Sherlock Holmes doesn’t seem to be the offspring Arthur’s uncritical and easily deceived neural network.

Arthur and the Nigerian Prince

In the previous post, I discussed how a medium, Eva C., fooled Arthur into fondling her ectoplasmic phallus. I mentioned that as the first example of Arthur’s unbounded gullibility because it’s my favorite. My second favorite I’ll describe here. It has to do with Arthur concluding, based on zero evidence, that some unknown number of dogs have brain power equivalent to humans.

Long after Louise dropped Holmes down Switzerland’s Reichenbach Falls, long after she was dead even, Arthur toured the world with his second wife, Jean, to spread the word of spiritualism, the belief that we can communicate with the dead via gifted people known as mediums. Regarding his trip down under, he recorded the events in his book Wanderings of a Spiritualist (1921). In that book, he recorded his encounter with particular fox terrier that had “a power of thought comparable, not merely to a human being, but even […] to a clairvoyant.” The dog had, allegedly, often demonstrated his remarkable ability by barking out the number of coins in a person’s pocket. When Arthur put it to the test, the dog began barking but would not stop. Arthur nonetheless expressed confidence in the dog’s supernatural abilities, attributing its poor performance to age and excitement. Arthur concluded his anecdote with a distinctly non-Sherlockian musing: “One wonders how many other dogs have human brains without the humans being clever enough to detect it.”

Though Arthur witnessed not a scintilla of evidence that the dog was in any way exceptional, he concluded nonetheless that the dog was not only as smart as a human, but clairvoyant as well. He based his conclusion on nothing more than the amazing claims of a dog owner he had not previously met.

Had Nigerian princes been trolling for easy marks in Victorian times, Arthur would have been at the front of the queue.

How it All Started

More than three decades ago, I read a book by Martin Gardner called Science: Good, Bad, and Bogus (1981). It was a collection of his essays, and one of them in particular caught my attention. It was “The Irrelevance of Conan Doyle.” In that essay, Gardner argued that Arthur could not have created any character as rational and scientifically minded as Sherlock Holmes because Arthur was simply too gullible. Gardner did not provide any specific examples of Arthur’s gullibility, but he did allege that Arthur “spent the last twelve years of his life in a tireless crusade against science and rationality. … There is scarcely a page in any of Doyle’s books on the occult,” Gardner continued, “that does not reveal him to be the antithesis of Holmes. His gullibility was boundless. His comprehension of what constitutes scientific evidence was on a level with that of members of London’s flat-earth society.”

This interested me. Though Gardner provided no examples of Arthur’s credulity, I had no trouble discovering many on my own, and I’ll relate one of them here. In his book The History of Spiritualism (1926), Arthur discusses the astounding power of ectoplasm. Spiritualism is the belief that those of us remaining on earth can communicate with our forebears who have passed into the spirit world, but we can do so only through an especially gifted person called a medium. Ectoplasm is the material through which the spirits make themselves visible to us. In Spiritualism’s Victorian heyday, the best mediums exuded ectoplasm, hence spirits, from their “natural orifices,” to use Arthur’s delicate words. Of ectoplasm, Arthur wrote:

The substance itself emanates from the whole body of the medium, but especially from the natural orifices and the extremities, from the top of the head, from the breasts, and the tips of the fingers. […] The substance occurs in various forms, sometimes as ductile dough, sometimes as a true protoplasmic mass, sometimes in the form of numerous thin threads, sometimes as cords of various thickness, or as a broad band, as a membrane, as a fabric, or as a woven material, with indefinite and irregular outlines.

Arthur believed that female mediums could exude “psychic rods” from whatever natural orifice they might have beneath a tabletop, and that these rods could become firm enough to lift the table. He described this titillating phenomena in his Edge of the Unknown (1930).

In the Belfast experiments this same ectoplasm was used for the making of rods or columns of power, which protruded from the body of the unconscious girl, and produced results such as raps, or the movement of objects, at a distance from her. Such a rod of power might be applied, with a sucker attachment, under a table and lift it up, causing the weight of the table to be added to that of the medium, exactly as if she had produced the effect by a steel bar working as a cantilever and attached to her body.

Most mediums were female. They were investigated by some of the most famous scientists of the time, virtually all of whom were male. Given Victorian sensibilities, and given that ectoplasm was exuded from all the “natural orifices,” this gender dichotomy might have posed a problem of propriety. Fortunately, the female mediums and the male investigators were willing, for the sake of research, to overcome all obstacles. If science demanded that the medium perform in the nude or take her investigator as a lover, so be it. If science demanded that Arthur fondle the ectoplasmic phallus of a female medium, how then could Arthur decline? He describes one such fondling in his History of Spiritualism.

The author has frequently seen ectoplasm in its vaporous, but only once in its solid, form. That was a sitting with Eva C. under the charge of Madame Bisson. Upon that occasion this strange and variable substance appeared as a streak of material six inches long, not unlike a section of umbilical cord, embedded in the cloth of the dress in the region of the lower stomach. It was visible in good light, and the author was permitted to squeeze it between his fingers, when it gave the impression of a living substance, thrilling and shrinking under his touch. There was no possibility of deception on this occasion.

Finding so many examples of Arthur’s self-admitted silliness helped me better understand Gardner’s too-gullible-to-be-Sherlock’s-creator suspicion. Gardner, however, provided no solid proof of Arthur’s literary fraud, nor did Gardner offer any alternative to Arthur as Sherlock’s creator. He left that work to whomever among his readers might take up the gauntlet.