In my previous post I suggested that assessing the historical evidence for the existence of ‘ninja’ is not a particularly interesting or productive way to approach the topic. That does not mean that I intend to completely ignore the question of historical evidence. One of the aims of this newsletter is to demonstrate that most ninja history is unutterably awful, after all, and I can’t do that without touching at least occasionally on questions of historical evidence.
Examples of Shinobi in Historical Sources
A large part of the case for the historical ninja is based on appearances in premodern Japanese sources of the word shinobi. Pretty much all serious advocates of the historical ‘ninja’ will concede that the precise word ninja (忍者) doesn’t seem to have been used in medieval and early modern Japan; in fact, it’s possible, as Stephen Turnbull suggests in his 2017 book, that the word ‘ninja’ was not in widespread use even in Japan until the late 1950s.1 You do occasionally come across the graphs 忍者 in premodern Japanese sources, but as those of you who have studied the language will know, Japanese is tricky. There are often multiple ways to read the same graphs, and it’s generally accepted that the graphs 忍者 were probably read (that is, pronounced) as shinobi no mono, meaning literally ‘one who sneaks.’ You do also see the noun shinobi (literally, ‘sneaker’) as a stand-alone noun occasionally, but more often shinobi appears as part of a verb and so is indicating an action, not an identity.
Here’re a couple of examples from the Taiheiki (‘Chronicle of Great Peace’), a war chronicle compiled some time in the late 14th century that covers the civil wars of the early to mid-14th century in Japan.2 The first example is an imperial prince, who is warning the Emperor that the warriors of the Hōjō clan in eastern Japan have come to Kyoto intending to exile the Emperor and execute the prince:
“I have secretly learned of eastern envoys who have just come to the capital; they are here to exile Your Majesty to some distant province and to put me, Prince Son’un, to death. Sire, this very night you must steal away (shinobitamau beshi) and make for the city of Nara.”3
今度東使上洛の事内々承候へば、皇居を遠国へ遷奉り、尊雲を死罪に行ん為にて候なる。今夜急ぎ南都の方へ御忍び候べし。
Here shinobi is being used as a verb, shinobitamau (with an honorific suffix tamau, because it’s referring to the actions of the Emperor), meaning basically to sneak away and hide, as it’s translated above. I include the above example to illustrate that people in Japan might indulge in a spot of shinobi-ing for any number of possible reasons, and also, given that the person doing the prospective shinobi-ing is the Emperor of Japan himself, to show that not everyone who shinobi-s is a shinobi in the sense of ‘ninja.’
(Although now that I think of it, ‘the emperor of Japan was also a ninja’ would be an awesome pitch for a movie.)
What we would ideally be looking for, then, would be examples of shinobi in a narrower military or paramilitary sense, of people sneaking into castles or enemy encampments to engage in sabotage, arson, assassination, or something along those lines. And these kinds of cases do indeed exist; we can find one in the same source, the Taiheiki. This particular case, dated to 1338, is often mentioned by ninja historians as being an early example of a ‘ninja’-ish shinobi. For context, the situation is that the Ashikaga clan, who will go on to found a dynasty of Shoguns, are facing off against a group of rival Nitta clan troops in the fortified Hachiman shrine near Kyoto.4 And then:
One night, under cover of wind and rain, they [the Ashikaga] had an unusually skilled shinobi (ichimotsu no shinobi) get into the Hachiman shrine and set fire to the main hall.5
或夜の雨風の紛に、逸物の忍を八幡山へ入れて、神殿に火をぞ懸たりける。
That’s the whole of the relevant passage, by the way – just one line. Turnbull mentions this repeatedly in his earlier ninja histories, and the Japanese scholar Yamada Yūji argues that this line is evidence that shinobi was understood as early as the 14th century as referring to a distinct role in warfare.6 I think Yamada may be reading more into this one line than is warranted, but I’ll discuss that point in detail in a later post. (I should also note that the Taiheiki itself is widely known to be heavily romanticized and unreliable as history, but we’ll let that go for now).
Another example comes from the Diary of Tamon-in (Tamon-in nikki), a diary kept by a Buddhist monk in the 16th century. This has an entry dated to the 26th day of the 11th month of 1541:7
This morning a force of Iga warriors sneaked into (shinobi-irite) Kasagi Castle and set a few fires...8
今朝伊賀衆笠置城忍ヒ入テ少々放火
The entry adds a few more details, that the Iga warriors set fire to a few other buildings as well, captured one of the structures within the castle, and that there were between seven and ten of them. That’s about it, though. You can read the entry for yourself here (bottom-right of the page) if you have the language abilities, thanks to the National Diet Library’s amazing digital collections. Bear in mind it’s not in modern Japanese but in classical, which is significantly harder.
Were Shinobi a Thing?
There are probably a few dozen other similar cases where the word shinobi or a variant thereof appears in the context of a clandestine operation, but I’m not sure it would add all that much to our discussion to cover each and every one of them. One of the things you may have noticed is that in two of the three above cases, shinobi is a verb, meaning it’s describing what someone is doing, not necessarily what they are. Another thing that most such accounts have in common is that they tend to be very short and lacking in detail. This means it’s often challenging to know exactly how to interpret the passages in question, and in fact I’ll come back to the case of the Taiheiki’s “unusually skilled shinobi” in the next post to show that interpretation is often not as straightforward as it first seems.
One of the risk factors that we inevitably run into is of reading these sources anachronistically. We live in a post-1960s media environment, more than sixty years after the pop-culture ‘ninja’ was first introduced to Japan and fifty years after it blew up in the West during the late 1970s. So we, in 2024, know exactly what a ‘ninja’ is, or think we do, anyway; and this means that when we read of accounts of clandestine military actions in medieval Japan, there’s going to be the risk that our brains will fill in details that aren’t there. We might think that everyone now knows that Iga Province was the heartland of the historical shinobi, and so Eishun, the author of the Tamon-in Diary, must have understood the Iga raiders as ‘ninja’ in the same way that we do. But was Iga known as a hotbed of shinobi in 1541? And if it was, did Eishun know that? Since the Tamon-in Diary doesn’t say anything on those points, we would be branching out into risky territory in assuming either without concrete evidence. To get students to do good history, I have to drill into them one exceedingly obvious yet seemingly hard-to-grasp point: people in the past did not think in the same terms that we do.
So the key question is: when we see a mention of someone described as a shinobi, can we interpret that as being an identity or specific role? And if we can, does the concept of the shinobi map closely enough onto the modern concept of the ‘ninja’ that we can read them as essentially the same thing? The answers, I would suggest, are 1) “maybe” and 2) “almost certainly not.” Let me deal with these in order.
There are cases where reading shinobi as denoting a role or identity may be justifiable. Two relevant examples come from the 16th and very early 17th centuries, actually both from non-Japanese sources. In 1512, an account in the official history of the Korean Joseon dynasty records that there was a Japanese man in a Korean prison named Yoshirō. This Yoshirō stated that in the city of Hakata there were certain people who were skilled at disguising themselves to gain access to military camps and other structures. The graphs used to refer to the people in question, 時老末, would be shilaomo in modern Mandarin, but were probably something like shinomi or shinobi in the pronunciation of the time.9 Ninety-odd years later in 1603 we have an entry in the Nippo jisho (‘Dictionary of Japanese and Portuguese’), a very early Portuguese-Japanese dictionary of incredible value to linguists, because it records a lot of Japanese as it was spoken around 1600. In the Nippo jisho the word xinobi appears, which is defined as:
espias que no tempo da guerra de noite, o rescondidamente trepão nas fortalezas, ou entrão nos arroyais a uer o que paβa.10
My 17th-century Portuguese is a bit rusty, but I think a reasonable translation would be:
spies who, in times of war, use night or stealth to steal into fortresses or enter camps to see what is happening.
I think it would be reasonable to grant based on the above that a shinobi does seem to have been a thing, at least by the mid-16th century, and that “spy” or “infiltrator” might be a pretty good translation.
AHA! SEE? NINJA WERE REAL AFTER ALL!
Well…not necessarily.
For Shinobi, Read Ninja?
This is where my second answer, “almost certainly not,” comes in. The bit about the shilaomo in Korea cited above comes from Aragaki Tsuneaki’s book chapter Did ‘Ninja’ Exist? (Ninja to wa jitsuzai suru ka), in which Aragaki also observes that you cannot necessarily draw a line from these mentions of shinobi all the way straight through to the modern ‘ninja.’11 I would ask you to think back to the metaphor I used last time around, of a matchstick of historical evidence holding up a two-ton elephant of claims about the ninja. There are simply too many aspects of the modern ‘ninja’ missing from the medieval historical sources for us to be able to say that the modern ‘ninja’ is fully or even mostly supported by historical evidence - note, for instance, that none of the above sources mention black clothing, assassinations, shuriken, ‘ninja masters,’ low social status and/or antagonism with samurai, kunoichi, special tools, smoke bombs, availability as mercenaries, etc etc etc. You can’t mentally substitute the modern concept of ‘ninja’ every time you see the word shinobi in a premodern Japanese text; that virtually guarantees that you will misread the text in question.
What’s more, to put all of this in some sort of perspective, we’re not exactly swimming in historical evidence here. On the contrary, we’re having to make tentative interpretations of a relative handful of very bare-bones accounts over several centuries; it’s not like we have a full and complete picture of what shinobi were and did at all. Nor, to be honest, does the weight of historical evidence suggest that shinobi were a particularly important part of medieval warfare. A scholar specializing in medieval warfare in Japan could probably go her whole career coming across two or three mentions of shinobi tops.
We also have to note that the usage of the word shinobi almost certainly would have changed over time, especially between the pre-1600 era and the post-1600 era. It’s very, very important to understand that post-1600 Japan was no longer at war, and as warfare drifted further and further into the past, so the warrior past came to be romanticized, re-imagined, and taken up as a topic for literary creation. It’s not a coincidence that we see a big increase in the number of mentions of the word shinobi during the two hundred and fifty years of peace of the Edo period (1600-1867), as Japanese historians and literary writers alike began to retell and reimagine their recent and not-so-recent past. By the early 19th century, the scholar (actually, he was a blind scholar – it’s a cool story) Hanawa Hokiichi could write the following:
It is my understanding that a ‘shinobi no mono’ is what is called an ‘intelligence agent’ (kanchō),’ and so some call them ‘clandestine agents’ (kanja) or ‘ spies’ (chōsha). Now, their role is to travel in secret to other provinces to observe the enemy disposition, or to blend in with enemy for a time to spy out weaknesses, or to sneak into an enemy castle and set fire to things, or act as assassins… 12
按忍者はいはゆる間諜なり故に或は間者といひ又諜者とよふさて其役する所は他邦に潜行して敵の形勢を察し或は仮に敵中に随従して間隙を窺ひ其余敵城に入て火を放ち又刺客となり […]
Whether Hanawa’s account was fully accurate or not is an open question. Note that Hanawa’s description above is also one of very, very few premodern accounts that suggest the shinobi no mono was an assassin, a function that is now almost automatically linked with the ‘ninja’ but for which there is little to no historical support (and yes, I’ll get to this in a later post).
Lastly, it’s also worth noting that during the Edo period, the term shinobi seems to have acquired a new and somewhat different meaning, essentially that of ‘secret policemen,’ as it was used by agents of the Tokugawa government who were charged with monitoring the population and rooting out potential disorder.13 This seems to be reasonably well-attested to, and while you’ll see some ninja histories argue that this meant that ‘the ninja came to serve the Tokugawa government,’ I’m not sure that’s an accurate reading of the situation.
Where Has This Gotten Us, Then?
So where does all of this put us? Has it shown that there is indeed historical support for the ‘ninja’ as we now understand them? Well, yes and no, though I’d be inclined to say mostly ‘no.’ I wrote in the previous post that I don’t think the modern ninja is 100% supported by historical evidence; based on what I have seen so far, I’d say the proportion would realistically be closer to 10%. But more than that, I’d say that the above shows us that approaching the ninja primarily as an historical phenomenon is simply the wrong way to do it. Doing ‘normal’ history does not allow us to answer most of the questions about the historical ‘ninja,’; on the contrary, it just raises tons of additional questions we can’t answer with that approach.
This, for me, is why I say that questions of historical evidence are arguably the least interesting thing about the ninja phenomenon, because a purely historical approach is quite far down on the list of things that really shed light on the topic.
Turnbull, Unmasking the Myth (2017), p. 8.
Taiheiki 太平記. The title is really a huge misnomer because it’s a long and quite famous account of the ongoing Japanese civil wars between 1330 to 1367. We don’t know exactly when it was written or who wrote it, though it was probably compiled at some point in the late 14th century. As I’ll be repeatedly noting, Taiheiki is heavily romanticized and not really a reliable historical source, but this doesn’t stop people using it anyway.
This is my translation; a different one is available in Helen McCullough, trans., Taiheiki p. 55.
Ashikaga 足利; Nitta 新田; Hachiman 八幡.
Translation mine. This will generally be the case for translations from Japanese unless I indicate otherwise.
Yamada Yūji 山田雄司.
“11th month,” not ‘November.’ Premodern Japanese sources use the lunar calendar, so 11th month wouldn’t actually correspond to the month of November.
Tamon-in nikki 多聞院日記. Written by the monk Eishun 英俊 (1518-1596). Text is from Eishun, Tamon-in nikki Part 1 (dai 1-kan), (Sankyō Shoin, 1935), p. 256.
This is quoted in Aragaki Tsuneaki 荒垣恒明, “Ninja to wa jitsuzai suru no ka” 忍者とは実在するか [Did ‘Ninja’ Exist?] in Watanabe Daimon 渡邊大門, ed., Sengoku no zokusetsu o kutsugaesu 戦国の俗説を覆す [Upending Popular Theories on the Warring States Period] (Tokyo: Kashiwa Shobō, 2016), p. 252.
Kamei Takashi, ed., Nippo jisho (日葡辞書): Vocabvlario da lingoa de Iapam (Tokyo: Benseisha, 1978 - a reprint, obvs), p. 304a. This is quite widely known and quoted in Aragaki pp. 251-2, Turnbull 2017 p. 10, and probably lots of other sources as well. The text is available in digital form here, but I think you have to create an account.
“One has to exercise due caution in drawing a facile link to ‘ninja’ in general, just because the word shinobi has made an appearance” (「忍び」と出てきたからといって、それを安易に忍者一般と結びつけることには十分注意しなければならない。) Aragaki, p. 261.
Hanawa Hokiichi 塙保己一 (1746-1821). This is also quite a widely-known passage, cited in Aragaki p. 257, Turnbull 2017 p. 10, and several other sources.
See e.g. “The Shogun’s Shinobi,” which is Ch. 7 of Turnbull’s 2017 Unmasking the Myth, pp 87-95.