How to Kill Someone Using a Piece of String

Looking at Shinobi no mono's famous poison-thread assassination technique

Children making figures from string, as drawn by Kawanabe Kyōsai 河鍋暁斎 (1831-1889) in A Children’s Handbook of String Pictures (Kyokumusubi osana tehon 曲結雅画手本), 1863. Source

The specific rabbit-hole that led to the article I just published was my attempt to track down the evidence for the claim that the robber Ishikawa Goemon had tried to kill the warlord Oda Nobunaga by dripping poison down a thread into his mouth while Nobunaga was sleeping. This, for the record, never actually happened. It is in reality the climactic scene of a fictional 1962 ninja movie, Shinobi no mono, and it was, as I showed last time around, an addition taken from an earlier 1925 detective novel, made during the process of adapting the original novel version of Shinobi no mono for the screen.

The thing is, the poison-thread technique is a really cool method of murder, regardless of whether it actually happened or not. It’s seen quite a few homages over the years, the most famous being in the 1967 James Bond film You Only Live Twice. This is where I first saw the technique back in my teens, and I’d guess the same probably holds for most of my readers too:

I suspect that You Only Live Twice provided the inspiration for the use of the same technique in the 1997 film Grosse Pointe Blank, deployed by John Cusack’s jaded professional assassin in the opening part of the movie (7:30 or so, though it’s dark and quite hard to see). The technique also pops up again in 2021 in the hilariously-titled Japanese anime The World’s Finest Assassin Gets Reincarnated in Another World as an Aristocrat (watchable here if you can stand the ads, at 13:40 or so).¹

This is all good fun from a film motif point of view, but I’m also bound to point out that quite a few ‘ninja’ writers have jumped to the conclusion that the poison-thread technique as depicted in Shinobi no mono was in fact a real method of murder used by the historical shinobi.² Tom Lockley, for instance, suggests in his horribly credulous African Samurai that Yasuke and his Jesuit companions in Japan would have needed to watch out for ‘ninja’ assassins, specifically for “poison administered to the open mouth of a sleeping victim by thread in the dead of night.”³ The British author Roald Dahl, of Charlie and the Chocolate Factory fame, also seems to have assumed that the James Bond poison-thread thing had been used regularly in Japan, judging by his use of the present tense (“involves”) in an article in Playboy from 1967:

[t]he manner of the killing was interesting and complex – a sly, silent Japanese method that involves a long length of cotton thread and a tiny little bottle.⁴

Since the Dahl quote above appears in the published academic article, this means I have managed to include a contextually appropriate, fully relevant citation to Playboy magazine in a peer-reviewed publication.

I have quite literally been reading Playboy for the articles.

This may well be the finest moment of my career.

The Extreme Realism of Complete Fiction

Anyway, for the record, it wasn’t just Western audiences who were impressed by Shinobi no mono’s poison-thread scene. The Japanese film critic Ogawa Tōru, for instance, wrote in his 1963 review of the film that the technique was one example of how the word ninjutsu was undergoing a shift from meaning basically “magical powers” to being a system of practical, real-world, ‘scientific’ techniques. Of the poison-thread thing specifically, Ogawa recorded his impression of it as being “so realistic and demonstrably practical that one thinks, ‘ah yes, of course.’”⁵

Let me pause to note here that one of the weirdest things about the reception of the film Shinobi no mono is the constant insistence on its supposed “realism.” Quite a lot of viewers, it seems, go out of their way to assert that the film is in some sense a “realistic” depiction of what they believe historical ‘ninja’ to have been. Animeigo’s 2007 English-language box set, for instance, claims that the movie contains “realistic Ninja techniques” (emphasis in the original), while the Amazon blurb for the 2009 collector’s edition of the movie says that the film is: “[c]onsidered by many critics to be the best and most realistic films [sic] about the ninja.”

“Realistic” is a very slippery term, though, especially when applied to film. One man’s realism is another man’s distortion, and the term can be used in many different ways, not all of which overlap completely in their meaning. A more precise term in the context of Shinobi no mono would probably be “verisimilitude,” in the sense that the film is not fantasy, with characters having magic powers, but rather is set in a reasonable approximation of real-world medieval Japan. The confusion appears to arise when copy writers use terms like “realistic ninja techniques,” because although I think that’s intended to mean “ninja techniques that could theoretically be done in real life” (as opposed to magic), it’s easy to see how a viewer could read that as a claim that the film shows ‘ninja’ doing things they are known to have actually done in real life.⁶

Needless to say, this latter is not a supportable claim, as Shinobi no mono is historical fiction from beginning to end. No matter how “realistic” a work of historical fiction may be, it’s still fiction, and nothing in there can be taken as a reliable depiction of real events without corroboration from actual historical sources. This is a basic and obvious point, but one that seems to escape ‘ninja’ writers over and over, as shown by Andrew Adams’ inclusion of Goemon’s assassination attempt as historical fact in his 1970 Ninja: The Invisible Assassins.

“Realism” Cuts Both Ways

While doing the research for what would ultimately become the JLL article, I realized that I needed to do some due diligence regarding the possibility that the poison-thread actually was a shinobi thing. Maybe, just maybe, Ranpo’s Stalker in the Attic had taken it from some obscure account I didn’t know about. Reading through Ranpo’s own work, though, that didn’t appear to be the case. Ranpo had no obvious connection to ‘ninja’ in his own work, although in a weird coincidence he did serve as a mentor to the novelist Yamada Fūtarō, one of the most prolific ‘ninja’ novelists of the 1960s. As a tribute to Ranpo, Yamada would later publish a short story called The Stalker of Iga, the title an obvious reference to Ranpo’s The Stalker in the Attic.⁷

Helpfully enough Ranpo did in fact leave a fairly detailed account of how he’d come up with the plot for The Stalker in the Attic. The below quote is a long one, the TL:DR being that Ranpo, trying to devise a ‘locked-room’ murder mystery, originally came up with the idea of shooting someone through a knothole in the ceiling, then let the idea ferment for a few years until eventually it became The Stalker in the Attic:

So, regarding how The Stalker came about…this idea of a knothole-in-the-ceiling had been wriggling around in the corners of my mind for a couple of years, and so at first I thought up something along the lines of the plot below, then since I couldn’t find a way in which I could write it, I set it to one side. The plot I had come up with was as follows.

It was of a kind with my earlier A Room With No Entrance, in that an individual has been shot dead in a closed and locked room in their lodgings or whatever, but there is of course no evidence that any criminal has secretly secured entry, nor is there any opening by which a pistol could have been fired into the room. But it’s not a suicide, and no firearm is found at the scene. The sort of mysterious case where what happened seems at first glance to be impossible.

The key to the whole thing is that the murderer, in the course of helping out with the New Year’s cleaning at the house, has lifted up the tatami mats on the second floor, and has chanced to discover that there is a knot-hole in the second-floor floorboards; that right below that same knot-hole, in other words in the ceiling-boards of the room below, there is also another hole, and that through these two holes one can see all the way through down into to the room below. So the murderer makes use of this chance discovery; when the victim is sleeping in a spot directly below where the two holes line up, he lifts up the tatami mats, not making a sound, and fires his pistol from above.

That was the plot, pretty much. But there was something that just felt off about it. It was kind of interesting as a formula but it was also impractical no matter how you looked at it, so I just dropped the idea and left it; but then, finding myself in a tight spot when I couldn’t come up with a good basic idea for a story, I refused to completely give up and kept kneading away at it, and gradually it morphed into what eventually became The Stalker.⁸

Since Ranpo mentions no other prior source, the implication would be that the poison-thread thing was 100% Ranpo’s idea.

What’s interesting, especially given claims about Shinobi no mono’s supposed “realism,” is that Ranpo also expressed some doubts as to whether the poison-thread technique could have worked in real life. (For context, the poison the murderer uses in The Stalker in the Attic is liquid morphine):

I got a lot of criticism for the part where the killer drips poison down from the hole in the ceiling, but even I myself, as the author, had some problems with it. I had to devise a chance scenario whereby the ceiling hole and the victim’s mouth would come together in the same vertical line, or I had to have the killer judge the distance by using the draw-string of his trousers, and include so much in the way of additional explanation that my arms got stiff from all of it. And some people pointed out a few things to me regarding the poison itself. Morphine is normally sold in the form of a white powder, for instance, and I had written that the killer used morphine, but not in the same form as it is normally used, that sort of thing, or they pointed out that a few drops of morphine in liquid form would not be enough to kill a person, and so on. Of course, these were obviously errors on my part.⁹

What did Ranpo mean by “I got a lot of criticism?” Well, it seems that some of his fellow detective fiction authors had gone on record as stating they thought that the poison-thread technique could never work in the real world, plausibility being quite important in detective fiction. In 1935, for instance, Ranpo’s fellow novelist Kōga Saburō criticized the murder depicted in The Stalker in the Attic as “an absolute impossibility,”¹⁰ for the following reasons:

1. It would be extremely difficult to position yourself directly above the mouth of a sleeping individual.

2. A sleeping individual would reflexively turn away when touched by the first drop of poison and move out of reach.

3. You could not guarantee that the victim would actually swallow the poison.

4. There exist hardly any real-world poisons lethal enough to guarantee death with only a few drops.¹¹

Much as with the kusari-gama, then, even if we grant that the poison-thread technique was actually associated with shinobi, it’s not clear that it would have been very effective.

How Long is a Piece of String?

How could the poison-thread technique be assailed as unrealistic in a work of detective fiction, Ranpo’s 1925 Stalker in the Attic, but hailed as a “realistic ninja technique” thirty-seven years later in the ninja movie Shinobi no mono? I think the answer lies in the differences between reader expectations with detective fiction versus historical fiction, ‘ninja’ fiction being a sub-category of the latter.

Historical fiction is a funny genre in some ways, because the reader expects that the broad sweep of events will be “historically accurate” but knows at the same time that other parts of the plot will be fictional. The tricky bit is that the general reader usually doesn’t know precisely where the line is drawn in any given work, which bits are “accurate” and which are fictional. This is what might be termed the “based on” effect, which leads to the unanswerable question “How accurate is Shogun?”

Detective fiction works rather differently. Here the point is less “accuracy” - nobody thinks that Murder on the Orient Express actually happened - and more to provide a sense of mystery around the murder, which the reader then tries to solve before the novel reveals the answer as the end. To be believable as a mystery, the method of murder has to be at least superficially plausible in the real world, because the modern detective novel demands verisimilitude. The glowing phantom hound in Conan Doyle’s The Hound of the Baskervilles, for instance, turns out in the end not to be supernatural at all but painted with phosphorus, because as a genre modern detective novels generally avoid supernatural explanations for their murders.

I think this is probably why the same poison-thread technique could be judged so differently in two different genre contexts. Detective fiction demands the method of murder be plausible to the last degree, so the poison-thread thing gets criticized for its real-world implausibility. ‘Ninja’ fiction, by contrast, insists that there really was some real-world correlate to the action on the page or screen, but leaves the reader more or less to guess which parts are fact and which are fiction, and so the poison-thread thing gets taken as historical fact.

What’s fascinating, from a socio-cultural standpoint, is that the very notion of fiction - that people can make things up - seems to have been entirely alien to most of the first generation of ‘ninja’ writers. It doesn’t seem to have occurred to Adams and Draeger that Shinobi no mono might simply be an entertaining fictional story rather than a faithful depiction of historical truth. Seemingly convinced that ‘ninja’ as depicted in Shinobi no mono were a matter of historical fact, they seem to have assumed that the poison-thread thing must have been a real shinobi method. And as they say, you know what happens when you assume.

For next time, we’ll continue with some of the other material that didn’t make it into the published article and digress a bit into film history. We’ll take a closer look at what happened with Roald Dahl’s involvement in You Only Live Twice, and how the ninjas got into the movie in the first place.

1. Japanese: Sekai saikō no ansatsusha, isekai kizoku ni tensei suru 世界最高の暗殺者、異世界貴族に転生する.

2. Plus a few reprises in overtly fictional contexts as well.

3. Lockley and Girard, African Samurai (Hanover Square Press, 2019), p. 92.

4. Roald Dahl, “007’s Oriental Eyefuls” Playboy June 1967, p. 88. In case anyone from Arizona is reading this, I paid for the copy out of my own pocket. Didn’t really fancy explaining to an audit committee why I was using state funds to buy back issues of Playboy.

5. Ogawa Tōru 小川徹 (1923-1991), “Ninja wa teikōsha tarienu ka” 忍者は抵抗者たり得ぬか [Can the Ninja Be A Figure of Resistance?] Eiga hyōron 20.2 (January 1963), p. 19. The Japanese for the quote is なるほどと思われるほど実証的でリアルである。 Ogawa’s article is really interesting, by the way, and I want to write about it in more detail at some point. As the title suggests, he understands the ‘ninja’ as an inherently political figure, and uses the film to muse on how and when violence can be justified as a response to state oppression. As I’ve written previously, plenty of Japanese critics immediately grasped the potential cultural and political meaning of the ‘ninja,’ something that went over the heads of basically every Western ‘ninja’ writer until Stephen Turnbull’s mea culpa book in 2017.

6. This reminds me a bit of the way ‘ninja’ writers tend to exploit the ambiguity of the word “legendary,” which can equally mean “real and very famous” or “completely fictional but very well-known.”

7. Iga no sanposha 伊賀の散歩者. One of the characters in the story is called Hirai Hosaemon 平井歩左衛門, which, in case you don’t get the joke, is a reference to Ranpo; Hirai was Ranpo’s birth surname, while the ‘ho’ (歩) of Hosaemon is the same graph as the ‘po’ of Ranpo (乱歩).

8. “Ranpo danshō” 乱歩断章 [Ranpo’s Odds and Ends] in Edogawa Ranpo zenshū [Complete Works of Edogawa Ranpo] (2003), 24:605-606. Not going to include the Japanese here because it’s too long.

9. “Ranpo danshō,” p. 606.

10. Kōga Saburō 甲賀三郎 (1893-1945); “an absolute impossibility” is zettai fukanōji 絶対不可能事.

11. Kōga Saburō, “Tantei shōsetsu kōwa” 探偵小説講話 [Lectures on the Detective Novel] in Kōga Saburō tantei shōsetsu sen #3 Ronsō misuterī sōsho 104 (Tokyo: Ronsōsha, 2017), p. 316.