Mary Midgley’s Misleading Essay, “Trying Out One’s New Sword”
©Jordan Sand, 2019
Some years ago, a student in my Japanese history survey brought to my attention a famous
essay by philosopher Mary Midgley titled “Trying Out One’s New Sword.” In this essay,
Midgley presents the problem of how we should evaluate a custom known in Japanese as
tsujigiri, which she describes as practiced by members of Japan’s traditional warrior caste,
the samurai. Tsujigiri, she explains, means to test the blade of one’s sword by killing a
passerby. As morally abhorrent as it is to contemporary readers, Midgley asserts that
tsujigiri was seen as acceptable, even necessary, within samurai culture. She uses this
extreme example of a custom contrasting with our own to pursue the argument that moral
relativism is an untenable position.
Since Midgley was not a scholar of Japan, her essay is not well known among Japanese
historians. However, it has been widely anthologized and taught in courses on ethics at
Anglophone colleges and universities. It presents a moral conundrum in vivid and
accessible form, which doubtless leads to lively discussions in college classrooms. If,
however, instructors and students who read this essay grapple with Midgley’s sample case
on the presumption that it has grounds in the real history of the samurai, then they are
accepting a harmful distortion of Japanese history.
Upon learning of Midgley’s essay, I asked colleagues on the Premodern Japanese Studies
mailing list about it. Their responses are incorporated in my comments below. My initial
question concerned the genealogy of references in European languages to tsujigiri that
might have led to a mid-twentieth century British philosopher’s use of it as a case for her
essay on moral relativism. I soon learned that in a memoir, Midgley had described hearing
the term from Japanese anthropologist Carmen Blacker, apparently when Blacker was an
undergraduate. This would have been in the 1940s. At the time, there was plenty in the
recent behavior of Japanese soldiers with swords to suggest a cultural tradition of
capricious violence, and indeed practices like testing swords on prisoners are documented
among the atrocities of the Asia-Pacific War. Since modern Japanese soldiers were not
samurai, these cases were probably not what prompted Blacker’s remark, but they would
have contributed to the popular perception on which Midgley relies, of a uniquely
bloodthirsty Japanese warrior culture, based on an entirely alien morality.
Since the last quarter of the twentieth century, we historians of Japan have tended to
emphasize the way that the long peace under the Tokugawa shogunate (1600-1868)
transformed swordsmen into bureaucrats. After the early Tokugawa period, few samurai
until the last years of the regime ever used their swords. This story of tamed samurai would
have resonated less at the time that Blacker told Midgley about tsujigiri, when Japanese
wartime brutality was commonly understood as rooted in samurai tradition. The difference
between Midgley’s image and the one many of us teach today thus shows the way that
contemporary context affects what we go looking for in history. And indeed, samurai of
the fourteenth through the early seventeenth centuries were a much rougher lot than during
most of the Tokugawa period. There were also some violent samurai swordsmen around in
the 1860s, who contributed to the collapse of the Tokugawa regime. When Midgley came
1
to write the essay “On Trying Out One’s New Sword” in the 1980s, while most of us
Japanese historians were talking about peaceful Tokugawa samurai-bureaucrats, Western
popular culture still often presented samurai as lawless swordsmen. Midgley could also
readily have confirmed her impression of a samurai cult of violence by reading the work
of several historians on the earlier Warring States period or on the radicalism of the final
years of the shogunate.
Midgley’s essay, however, offers no indication that she in fact read any Japanese history.
Without evidence, she takes the phenomenon to be a normal custom among samurai
generally. In fact, she treats it as normative—morally necessary in their eyes. “Tests were
needed,” she writes, lest the samurai fail in his swordsmanship, which would “injure his
honor, offend his ancestors, and even let down his emperor.” She refers to tsujigiri as a
“custom” of “ancient Japan” without specifying time period. She takes a single word from
an exotic locale, without concern for context or usage, and makes it stand in for a
holistically conceived “culture” without historical evolution. This presents a textbook case
of Orientalism in the Edward Said’s sense.*
Orientalist logic aside, she could have found a better example for her purposes in something
like Aztec human sacrifice or perhaps in some cases of headhunting—cases where indeed
the killing was understood (at least by some of the killers) as integral to sanctioned and
necessary religious and social practices. In fact, even in the way she understands it, tsujigiri
actually makes her case rather poorly, since it is ludicrous to think that Japanese society as
a whole could have accepted this samurai “custom”: the victims could hardly have
acquiesced, since they were purportedly assaulted at random. But this is unrelated to the
historical problem I am discussing.
Let me now turn to the history of usage of the term tsujigiri, whose presumed normativity
forms the crux of Midgley’s argument. More work needs to be done to document the term,
but I will show here at least that it has evolved in usage and that in no actual case for which
I can find reliable contemporaneous reference was it used to describe an act deemed normal
or socially acceptable.
Before learning the Carmen Blacker story, I had thought there might be a longer history of
references in Western texts or some nineteenth-century locus classicus on tsujigiri. But I
have turned up little in the common early Western writing on Japan. Neither Engelbert
Kaempfer nor Franz Siebold appears to mention it. Kaempfer and Siebold provide us the
two most detailed accounts of Tokugawa Japan by European visitors. A googlebooks
search for instances of the Romanized tsujigiri or tsuji-giri before 1900 yielded five hits:
*
Her ignorance of samurai history is further attested by several things in the essay,
including the fact that she imagines that the samurai would be trying a “new sword,” when
old swords were much more prized; by her reference to the samurai’s fear he would “let
down…his emperor” (only a few samurai at certain revolutionary moments in history
understood their loyalty as directed to the emperor); and by her use of the phrase “ancient
Japan” itself, since standard practice in the field is to reserve this term for the period before
the rise of the samurai.
2
definitions in two editions of the Hepburn Japanese-English dictionary (1874 and 1894)
and three references in German works. Two of these German works were translations of a
Tokugawa-period legal compendium. The third, an article on the Japanese sword published
in a journal of ethnology by George Müller-Beeck in 1882, relates of the swords of famous
sixteenth-century swordsmith Muramasa that samurai would try them out through tsujigiri,
slashing beggars or animals in the street, or have the executioner test them (F. George
Müller-Beeck, “die japanischen Schwerter,” Zeitschrift für Ethnologie, 1882;
googlebooks). This is the earliest instance I have found of tsujigiri referring specifically to
trying out a blade. It is intriguing that the word appears here in particular connection with
Muramasa swords, since these swords were widely believed in the Tokugawa period to be
cursed, and were particularly sought by anti-Tokugawa swordsmen at the end of the period
for precisely this reason. The 1874 Hepburn dictionary gives “killed or murdered in the
streets.” The 1894 Hepburn dictionary gives “killing or murdering in the streets to try one’s
sword or skill.” I can’t be sure of the significance of this amendment in Hepburn’s later
edition but it seems worth noting that the German ethnological work, which coupled the
killing and the sword-testing, appeared in the period between the two editions. More
investigation would be needed to determine whether there is a larger lineage of Western
texts discussing tsujigiri.*
It is easy to imagine the word gaining some currency among foreigners in bakumatsu, the
last years of the shogunate, when sword-happy shishi (anti-Tokugawa imperial loyalists)
were cutting foreigners and officials down in the streets. The historical encyclopedia
Kokushi daijiten relates that Kumamoto samurai and imperial loyalist Kawakami Gensai
was known as “Tsujigiri Gensai” because he did lots of slashing while in Kyoto. The entry
also refers to him as a terorisuto. This was a time of terrorism, so the label is doubtless
warranted. There is nothing to indicate that Tsujigiri Gensai cut anyone down in order to
test his sword or that his behavior was sanctioned by custom. On the contrary, his violence
served quite modern political goals.
Carmen Blacker may have had some older precedent in mind. All she is reported to have
told Midgley is that classical Japanese had a word meaning “to try out one’s new sword on
a chance wayfarer.” Blacker/Midgley’s definition matches the one in the modern dictionary
Nihon kokugo daijiten and other standard reference works that I viewed in the
JapanKnowledge database. Just to be clear, we must note that this definition has two
elements: (1) to cut down a passerby in the street and (2) to do it in order to test a sword or
one’s swordsmanship. Literally translated, tsuji-giri is “street-crossing-cut,” which should
simply mean only element (1), so at some point the definition acquired additional
specificity, which included the motive in the attack.
*
I have found one earlier example in a Western dictionary: the Portuguese-Japanese
dictionary Vocabvlario da Lingoa de Iapam, published in Nagasaki in 1603, of which I
have a later French translation. This dictionary gives the definition: “Those who run by
night and kill at crossroads.” Here, as in the 1874 Hepburn dictionary, there is no
indication of the motive for killing.
3
The Kokushi daijiten encyclopedia’s entry for tsujigiri was written by distinguished
Tokugawa scholar Yoshiwara Ken’ichirō. Yoshiwara first gives the definition containing
both elements, then says that tsujigiri sometimes also occurred as part of a robbery (which
would suggest something different from merely “testing” a sword). The earliest citation he
presents is from the Taiheiki, a military epic chronicling events in the first half of the
fourteenth century and probably authored sometime late in that century. This text describes
the followers of an evil shogun cutting down young and old in the streets of the capital.
The author of the epic refers to tsujigiri in this passage to demonstrate the horrific
immorality of the time. No mention is made of testing swords. Yoshiwara next lists a
prohibition of tsujigiri in 1597, in which tsujigiri is grouped with robbery. He then
describes a period of frequent incidents of tsujigiri in the shogunal capital of Edo while
Shogun Iemitsu was away visiting Kyoto in 1648, another incident that was punished, and
finally the record of the official sentence for committing tsujigiri in the classic
compendium of shogunal criminal law, Osadamegaki hyakkajō (1748), which called for
the offender to be carried around the capital in chains before being executed. Yoshiwara’s
encyclopedia article points us to the large historical document collection Koji ruien (also
available in JapanKnowledge), which has a one-page entry on tsujigiri. Excerpting a work
called Setsuya kandan, that entry tells the story of the rash of tsujigiri in Edo while Shogun
Iemitsu was in Kyoto in 1648. I could not identify the publication date of Setsuya kandan.
It relates that during Iemitsu’s absence, night after night there were incidents in the Banchō
Ogawa-chō area in which hatamoto (shogunal bannermen) cut down townspeople in the
street. The town magistrate struggled to stop the violence. Troubled, the shogunal
counselors sent messengers repeatedly to the shogun in Kyoto. Sakai Tadakatsu, who was
accompanying the shogun, told him it could easily be stopped. Sakai had Itakura Shigenori
tell the hatamoto that the shogun had heard about the incidents around the hatamoto district
and said that this was because the hatamoto were ignorant of the proper way of samurai
(budō no kokorogake utoki), that they should examine one another, and if these incidents
were repeated, it would be considered the fault of all the hatamoto. The incidents reportedly
stopped immediately.
This account mentions no motive for the slashings. It does not indicate that the hatamoto
were trying out their swords, although they may have been doing so. The entry in the same
Kokushi daijiten for a different term, tameshigiri (lit. “test-cutting”), authored by
Shigematsu Kazuyoshi, meanwhile, begins with this definition: “to cut the body of an
executed prisoner in order to test the edge of a sword. In the early Edo period it was done
by tsujigiri.” It then indicates that tameshigiri was officially performed by machi dōshin
(shogunal police), then became the special work of the executioner Yamada Asaemon,
whose clan inherited the practice of testing the swords of elite samurai on the corpses of
executed criminals.
In a work entitled Ō-Edo shitai kō (A Study of the Place of Corpses in the Edo Capital,
1999), Tokugawa historian Ujiie Mikito has examined references to the practice of
tameshigiri in seventeenth and eighteenth-period texts. He finds that many samurai
witnessed the cutting of corpses for sword practice, particularly at the execution grounds,
and describes accounts from early in the period of elite samurai who took relish in showing
their sword skills on prisoners’ bodies (in some instances possibly living prisoners,
4
although he does not clarify this). However, he mentions only one case in which a samurai
intentionally killed someone in the street to try out his sword. The swordsman in this case,
Tokugawa Mitsukuni, describes having killed an outcaste under a bridge to test his sword
(tameshigiri), prompted by a dare from a companion. The episode is recorded in
Mitsukuni’s memoir as a shameful incident of his youth. Mitsukuni was born in 1628, so
the incident would probably have occurred in the 1640s or early 1650s. Ujiie’s main finding
is that during the extended peace following this time, most samurai became repulsed by
the testing of swords not only on living humans but on prisoners’ corpses as well.
In light of all this, I propose the following hypothesis: that at some point after the
lawlessness of the sixteenth and early seventeenth centuries, the words tsujigiri (“slashing
in the street”) and tameshigiri (“slashing to test”) somehow became conflated. There seem
to me two possible explanations for the conflation. The first is that the shogunate caused
the semantic slippage by introducing formalized tameshigiri in part as a response to
hatamoto violence. Early Edo-period cases like the hatamoto slashings of 1648 could have
been caused by any number of things apart from the desire to test swords. This was a period
of gang warfare between hatamoto and non-samurai gangs called machi yakko. The
slashings might have been acts of revenge, part of drunken brawling or robberies, or pure
sadism. Since the practice of tameshigiri on corpses at the execution grounds was
formalized after this time, and Shigematsu’s entry for tameshigiri states flatly that it
replaced tsujigiri, it is easy to imagine formalized tameshigiri being presented by the
authorities as a substitute for tsujigiri at the same time that they cracked down on slashing
in the streets. This would have retrospectively conferred a small degree of martial honor
on the outlaw behavior of their own retainers by saying, in effect, “we recognize that a
samurai must test his sword; here is the proper means to do it.” My point is that the idea
that one must “slash a person in the street” for the purpose of maintaining the blade and
one’s skills, could well be a retrospective justification, regardless of the original reasons
for the street killings. This would explain why the term tsujigiri subsequently, perhaps
during the comparatively peaceful mid-Tokugawa period, came to refer specifically to
killing to “test one’s sword.” The second possible explanation is that the conflation is even
later, dating to the late nineteenth century, and derives from a new reading of samurai
tradition based in part on the swashbuckling bravado of late Tokugawa imperial loyalists.
For a brief period in the tumultuous 1860s, skill with the sword was suddenly important
again for samurai, and some hotheaded men among them became genuinely eager to test
theirs out, having never used them before. Over the two decades following the Meiji
Restoration of 1868, these men’s impulses were eventually suppressed, and legitimate use
of swords came to be monopolized by the modern police and military. By the 1880s, the
“way of the sword” was beginning to be mythologized as part of the national past, and
Japanese antiquarians, together with European ethnologists, might readily—if, perhaps,
mistakenly—have conjoined “street-slashing” and “test-cutting” as part of a portrait of the
samurai as a man whose loyalty to his weapon and his lord exceeded conventional morality.
Even if in fact there were historical cases among samurai in which desire to test one’s
sword was the only reason for a murder, I have found no evidence that it was ever accepted
practice. At least in terms of word usage, the addition of the explanatory element, “in order
to test a sword,” to the merely descriptive “slash in the street” may be entirely apocryphal.
5
The true lineage of ways of imagining and defining the act of tsujigiri has yet to be fully
traced. Mary Midgley heard a one-line definition that conflated it with tameshigiri. The
conflation had indeed become standard at some point before she heard it, as it appears in
modern Japanese dictionaries. Perhaps the young Carmen Blacker, who would go on to
become an erudite scholar of Japan, knew a reason to describe it as an ancient practice in
Japan. But more likely she was just relating to Midgley the modern dictionary definition,
which Midgley then conjured into the evidence of an alien and disturbing Japanese martial
ethic for purposes of her thought experiment. Although it is certain that some samurai
slashed down people in the street during particularly lawless times in history, Midgley is
surely mistaken to imagine that it was ever normal and accepted, let alone that it was a
moral imperative.
I should add that despite its being ahistorical, I actually like Midgley’s essay. I am in
fundamental agreement with her rejection of absolute moral relativism. We cannot allow
the claim of cultural difference to trump all moral judgment. But to make this point she
invented a “samurai culture.” This fictional system of customs and morals serves her point
well, so perhaps she should be allowed to have it for the sake of argument, even if it is
purely imaginary. But as with everything else claimed casually for the sake of an argument,
it comes at a cost. The cost in this instance is that students in college classrooms, as they
debate the question of moral relativism that Midgley poses so artfully, also struggle with
the question of how Japan’s traditional military elite could have valorized such inhuman
cruelty. They should not be struggling with this second question, since it is not based in
historical reality.
**
Someone’s scan of a reprint of Midgley’s full essay is available here:
http://www.ghandchi.com/IONA/newsword.pdf
6