the frequency of using the sword on a battlefield

[Aureliano]

Hello guys,

I recall that quite some time ago I have stumbled upon a text that was telling about some specific archeological-historical research that took place in japan. Apparently, that research was able to determine the approximate amount of injuries and their origins that were sustained by the fighting samurais and their minions during the pre-edo period.

As far as I recall, the researchers even presented numbers such as

70-80% - were injured by arrows.
10-15% - were injured by naginatas, spears and other polearm weapons.
~5% - were injured by swords.

The numbers are only written by my own memory.

I am now looking for the original research, or at least a more detailed review.
Did anyone hear of something like this?
thanks.

[Bruce Mitchell]

I believe that the numbers that you are quoting come from Dr. Karl Friday's essay in Budo Perspectives , edited by Dr. Alex Bennett and sold on this website. I believe that Dr. Friday list his sources, but they are all in Japanese. Hope this helps.

[Josh Reyer]

Look for

Conlan, Thomas. "State of War: the Violent Order of Fourteenth Century Japan" (Ph. D. diss.) Stanford, CA, Stanford University, 1998; and "Innovation or Application? The Role of Technology in War," paper presented at the Association for Asian Studies Annual Meeting, 13 March Boston, MA, 1999.

Suzuki Masaya, 刀と首取り　戦国合戦異説. Heibonsha Shinsho, 2000 and 鉄砲隊と騎馬軍団. Yosensha, 2003

Friday, Karl, "Off the Warpath." Budo Perspectives, ed. Alex Bennett, Auckland, New Zealand: Kendo World Publications, 2005.

Friday's article references the above four. According to his article, Conlan examined 1302 14th century battle reports, and found 721 identifiable wounds, which broke down like this:
73% - arrows
25% - swords
2% - spears

Suzuki examined 175 14th century battle reports and found 554 casulties, broken down thusly:
87% - arrows
8% - swords or naginata
3% - rocks
1% - spears

Examining 1291 reports from the 15th and 16th centuries, Conlan found:
439 arrow wounds
343 gunshot wounds
192 spear wounds
79 injuries by stones
50 sword cuts

Suzuki found for 1501-1560, 620 battle wounds broke down like this:
61% - arrows
21% - spears
16% - stones
3% - swords

while for 1563-1600 (after the introduction of the gun), 584 casualties broke down as:
45% - gunshots
21% - arrows
17% - spears
6% - cutting injuries
5% - rocks
4% - combos of the above

[Aureliano]

Thanks a lot guys!

[rfoxmich]

Dr. Friday (he gave a nice talk about this at MSU a few years back) uses statistics like these to make his point that from a very early stage, the study of swordsmanship was not as a practical fighting method for the battlefield, but as a method for self-development. He did note, however, that the sword was used 'around town' as a personal protection weapon.

[Ellis Amdur]

One question about the statistics that I have is how were the wounds determined? For example, if it is a product of examination of skeletal remains, things would be skewed to a) projectile weapons b) arrow heads remaining in the wounds c) spears that would be more likely to strike bone. Consider this - it is very likely that many warriors were wounded and finished off with tanto. A cut to the neck would result in soft-tissue damage. Would this be found in skeletal remains hundreds of years later.
I think there it is doubtless that the sword was the equivalent to a side-arm in a modern soldier's kit - but I wonder if these statistics do justice to the use of bladed weapons (naginata, sword and short sword/tanto).
Ellis Amdur

[pgsmith]

That's a very good point. It's something that I've wondered about for years, ever since he first wrote down his casualty statistics on Iaido-L some time in the late 90's. Mr. Friday has always mentioned "casualty statistics" in his articles, but where do they come from? I figure that they were either from skeletal remains as you mentioned, or written reports of those treated by army physicians. If skeletal remians, then it would be subject to the possible errors you pointed out. If physicians' reports, then we would be missing all of those that were too badly wounded to be treated by the army physicians.

[Josh Reyer]

Um, it's right there in my post, but all these numbers, at least, come via contemporary battle reports and casualty lists.

[Ellis Amdur]

Well, that begs a lot of questions:
1. Conlan - 13th-14th century. 721 identifiable wounds for the whole period. So how many "unidentifiable" wounds were screened out. Next, were the wounds of higher ranked warriors reported more frequently than lower ranked warriors? Were arrows targeted against those of higher rank, or were they shot in volleys?
2. Suzuki's numbers are quite similar - are the 175 wounds part of the 721 of Conlan, or another group altogether? Same questions otherwise apply.
3. The most puzzling is the 2nd Conlan: Examining 1291 reports from the 15th and 16th centuries, Conlan found: 439 arrow wounds, 343 gunshot wounds, 192 spear wounds, 79 injuries by stones, 50 sword cuts. That's 1103 wounds in 1291 reports. That's less than one death per report! Who was included, who were left out? What is this a cross-section sampling, or the only reports he could find. It is certain that over two centuries, a mere 1103 deaths are not all that were recorded. So what was the sampling procedure. Were the stones from slings or were they large objects. (If the latter, were they finishing blows - crushing the skull of a downed, injured man, or stones cast down a battlement from a siege.).
4 -5. Same questions as above.
I have absolutely no doubt that these statistics show trends. I noted some 20 years ago, when I first wrote the first draft of Old School that requests of generals from Korea in the Chosen Eki were for more spears and guns, never more swords.
Still, these statistics, with their paucity of detail, remind me of psychological statistics, where a certain therapy claims 30% improvement, but has carved out of the study, the 80% who dropped out early because the treatment was so noxious.
So I would be intrigued to find out more about who died, from what class they were in, and who and what kind of person did not have their death explicitly recorded in such reports.
best
Ellis Amdur

[b8amack]

You're confusing history with religion. You go on the evidence you DO find, not find reasons to discount it and continue believing as you wish without proof.

Also... you wrote Old School? I love that movie...

[Josh Reyer]

Ellis,

I have some training in psych stats, as well, and I agree with much of what you say. Still, the stats provided here are from Friday's cites in his article. I imagine the original studies written up by Conlan and Suzuki go into much more detail.

One thing I took from these stats, though, is not the uselessness of the sword on the battlefield, but rather that it was used, and not to an insignificant degree. Heck, 6% of 584 is 35 -- so even from clearly non-comprehensive survey, 35 men had to use their sword in anger. The actual number is surely much more. On a thread regarding Friday's article on E-Budo, I made the comparison to modern warfare. 500 years from now, a historian looking over some of the war records of the late 20th/early 21st century would certainly find that a great majority of casualties were caused by bombings, missile strikes, and heavy artillery. He would be mistaken, though, if he then assumed that the use of personal firearms was not given heavy emphasis in training (and actually used), and hand-to-hand/close quarters combat hardly regarded at all.

[stealth_monkey]

You're confusing history with religion. You go on the evidence you DO find, not find reasons to discount it and continue believing as you wish without proof.

Also... you wrote Old School? I love that movie...

I think you're confusing statistics with pop-science. The first thing you learn in any scientific discipline is to determine and eliminate possible causes of bias in your data set. Ellis brings up some very valid points and I 100% agree with him that those statistics show some strong trends, but the exact figures should not be taken at face value. Without getting our hands on the same data sets that were reported and the original author's exclusion criteria we can't really draw any major conclusions from the numbers provided.

[b8amack]

Sure he does. But then he offers a bunch of conjecture. I can argue that spontaneous combustion which self-extinguishes before reaching the bones would not be reflected in the skeletal remains, or that magic spells that stir the brain up like a big tub of jello can also not be discounted, for the same reason. Or fear induced heart attacks from the sight of giant dinosaurs.... and so on and so forth. What's the point? Should we not instead give Conlan some credit for having actually done his research, rather than just seeking to poke holes?

[hyuna]

Should we not instead give Conlan some credit for having actually done his research, rather than just seeking to poke holes?
The problem is that without "seeking to poke holes," it isn't clear that the author of a paper did their research. If you don't approach academic claims with skepticism, then you will never notice if people are basically just making stuff up. Moreover, each academic paper is just one more step up -- looking for weak points is how you figure out where to take the next level of research. To that point, poking holes and giving credit for doing the research aren't actually in conflict with each other. Someone can be wrong and still materially advance knowledge. Indeed, this is the typical state of things (see, for example: http://www.newscientist.com/article/dn7915 )

The default position is "we don't know." Anyone who asserts that they do know, whether it be that swords are more significant than spears, or vice-versa, or if giant dinosaurs roamed medieval Japan, bears the burden of proof. We shouldn't take Conlan at face value, and neither should we take anyone else's conjecture at face value. Ellis says that it's likely that people were finished off with tanto, and you are absolutely correct in pointing out that there is no obvious reason to believe that to be the case. That doesn't, however, invalidate the questions.

BTW, I'm looking forward to the treatise on the use of magic spells and giant dinosaurs in mass combat.

[b8amack]

So am I. Preferably one of those manga history texts, like they have for kids, here.

[Raffa]

Well, we can try a comparison with middle age european warfare.
Pole arms and arrows - Excellent in battle formation, almost useless in close combat.
Swords - _Bad choice in battle formation, excellent in close combat.

So I suppose that a samurai would had used his bow or long weapon (nagitata, yari) while in a formation, and after it, if the battle would go into a general melee, or, for whatever reason in close combat, put aside his "long reach" weapons and drew his sword....

[Ellis Amdur]

Consider the number of ryu that focused on armoured grappling - yoroi kumiuchi, kogusoku, koshi no mawari. In fact, I do not disagree at all with the general thesis that Friday cites. I've actually written in Skoss' Koryu Bujutsu series that many of the classic ryuha can be viewed as "officer training" rather than "basic training," that it was training for excellence at arms to enable the aristocratic warrior to fulfill a leadership role.
However, when one considers the grappling schools, in which one gets one's hands dirty, so to speak, and which were often the provenance of lower-ranking rural bushi or goshi, then some new questions arise. Why did these schools develop in the heart of the Sengoku period if they, too, were unrealistic - when the waza of such schools as Takenouchi-ryu and Yagyu Shingan-ryu are eminently practical. The stabs with a a tanto in Araki-ryu torite-kogusoku, for example, are almost always directed to soft tissue areas.
In the early war-tales, which were not realistic, to be sure, we do have a cultural meme - that the only figures of importance were the aristocratic warriors. Only their deaths are described. The foot-soldiers, who were often given "tsukai-ste" (use and throw away) weapons, were of little importance. Unmounted, they were the ones who most often ended up rolling in the muck in melee situations when battlefield formations broke down. If the cultural meme of the importance of the bushi class was transferred to record keeping and battlefield reports, how were their deaths described?

[pgsmith]

Sorry Josh, I didn't take the time to actually more than glance at your post. However, it still really doesn't tell me where those numbers come from. While I realize that the Japanese tended to be more than a bit anal about record keeping, I have an extremly hard time believing that they would have someone go through and write down what wounds were on all the bodies lying on the ground after a battle. Therefore, the next logical thing (in my mind at least) would be medical reports. I am sure that the army's physicians would keep track of who was treated for what wounds. This then would be a record of those that were only wounded during the battle, not those that actually died outright. Alternatively it could be as Ellis suggests; that it was a record of the fate of the higher ranking aristocracy only.

Ellis, I can't remember now who it was, but a number of years ago I had Araki ryu kogusoku described to me as "how to hold the fellow close so you could wiggle your yoroidoshi around under his armor." :)

[Theodore]

Look for

Conlan, Thomas. "State of War: the Violent Order of Fourteenth Century Japan" (Ph. D. diss.) Stanford, CA, Stanford University, 1998; and "Innovation or Application? The Role of Technology in War," paper presented at the Association for Asian Studies Annual Meeting, 13 March Boston, MA, 1999.

Suzuki Masaya, 刀と首取り　戦国合戦異説. Heibonsha Shinsho, 2000 and 鉄砲隊と騎馬軍団. Yosensha, 2003

Friday, Karl, "Off the Warpath." Budo Perspectives, ed. Alex Bennett, Auckland, New Zealand: Kendo World Publications, 2005.

Friday's article references the above four. According to his article, Conlan examined 1302 14th century battle reports, and found 721 identifiable wounds, which broke down like this:
73% - arrows
25% - swords
2% - spears

Suzuki examined 175 14th century battle reports and found 554 casulties, broken down thusly:
87% - arrows
8% - swords or naginata
3% - rocks
1% - spears

Examining 1291 reports from the 15th and 16th centuries, Conlan found:
439 arrow wounds
343 gunshot wounds
192 spear wounds
79 injuries by stones
50 sword cuts

Suzuki found for 1501-1560, 620 battle wounds broke down like this:
61% - arrows
21% - spears
16% - stones
3% - swords

while for 1563-1600 (after the introduction of the gun), 584 casualties broke down as:
45% - gunshots
21% - arrows
17% - spears
6% - cutting injuries
5% - rocks
4% - combos of the above

Rocks seem to outnumber swords. Is there an iwa-do school?

[Kim Taylor]

I suppose one could just ask the fellow?

http://www.bowdoin.edu/faculty/t/tconlan/index.shtml

Kim.

[Kim Taylor]

Oh why not double post eh? After a few moment's perusal of Dr. Conlan's site and opening up a paper I've found one very good eason why detailed records of wounds would have been kept by the warriors of the time.


Warriors did not fight for a lord; instead their deeds required compensation. Each warrior

who chose to fight submitted petitions for reward to their commander after any military

engagement. These petitions record the date and location of battle. Each warrior listed all the

damages incurred, including wounds and death, which required compensation. They were

submitted to a general, or his underlings, who vigorously inspected them. Some even added

notations of "deep," "shallow," and "medium" to those wounds recorded. After each document

was verified, it was signed by a general and returned. Warriors used these pieces of paper as proof

of their military service when submitting claims for remuneration. If they did not receive adequate

compensation, they would take their services elsewhere and fight for a rival commander capable of

offering better.

This is from http://www.fas.harvard.edu/~rijs/pdfs/conlan.pdf which also notes the following:


...one can discern that the battles of the fourteenth century were fought

with waning intensity. This period witnessed two peaks of heightened military activity, which

closely corresponded with political events. The first occurred during the initial six years of warfare

while the second lasted from 1350 until 1355. During the years of 1332-38, casualty rates reached

33 percent (19 percent killed and 14 percent wounded), with swords accounting for one third of all

wounds, which suggests frequent hand-to-hand combat. Nevertheless, few were willing to come

to blows and die for a cause, particularly from the Kannø disturbance of 1350 onward, where one

sees casualty rates of only 23 percent (6 percent killed and 17 percent wounded), with swords

accounting for approximately ten percent of all wounds.

He cites his dissertation for the numbers. One would suppose that a PhD committee might raise similar questions on the data that we have seen here, but one could obtain a copy of the dissertation easily enough and check it.

Kim

[pgsmith]

Thanks for that Kim, it's appreciated. However, it does point out the fact that these are records of not immediately fatal injuries received, and not battlefield deaths. You college edumacated folks sure know how to track down the facts eh? My way tends to be to discuss it endlessly on the internet until such a time as I can catch someone with real knowledge over a beer and pump them for information. :)

[Kim Taylor]

Thanks for that Kim, it's appreciated. However, it does point out the fact that these are records of not immediately fatal injuries received, and not battlefield deaths. You college edumacated folks sure know how to track down the facts eh? My way tends to be to discuss it endlessly on the internet until such a time as I can catch someone with real knowledge over a beer and pump them for information. :)

The quote quite specifically mentions that these are records of death as well Paul. "Each warrior listed all the damages incurred, including wounds and death, which required compensation." One's fellow family or friends would presumably describe the death which would merit some compensation.

The man's name in google was enough to get to his University page. It used to be a skill to find information in the libraries but now it's just a matter of google and sometimes a credit card to get papers that were weeks in the getting ten years ago.

But again, why even bother to get his thesis when one could simply send him an email and ask what one wants to know?

Kim.

[pgsmith]

The quote quite specifically mentions that these are records of death as well Paul. "Each warrior listed all the damages incurred, including wounds and death, which required compensation." One's fellow family or friends would presumably describe the death which would merit some compensation.
While I feel that it's possible that a warrior's friends or family would describe it for compensation, it still seems to me that it is only a very few of the deaths that occured that were recorded in the records. I'm sure it would be much more clear to me if I were to read all of the papers that were cited in the research, but my desire for history is just not that strong.


The man's name in google was enough to get to his University page. It used to be a skill to find information in the libraries but now it's just a matter of google and sometimes a credit card to get papers that were weeks in the getting ten years ago.
That's both a good and bad thing in my opinion. It's good in that so much more information is easily available than it was just a few short years ago. It's bad though in that people have come to expect instant gratification as far as information is concerned, and many don't even take the time to research a little bit to find out if what they read on Wikipedia is even accurate or not.


But again, why even bother to get his thesis when one could simply send him an email and ask what one wants to know?
Yep, easy enough to do. However, it leaves out the whole "over beers" part! :)

[Jonathan]

Being introduced to new reading material and research is never a bad thing, echoing sentiments of appreciation for the paper. Yay for forums.

[UnimportantHero]

When looking at any statistic, it is also important to question how many people were using those weapons. There are always more archers on a medieval or feudal battlefield than melee combatants, because in medieval/feudal warfare the archers were not target shooters. They fired upward, and allowed arcing rain to come down on the enemies. They also had ammunition. Sooo ... yes, arrows wounds will be the most common cause for taking someone out. Same goes for polearms which have more people trained in their use on the battle field, to help with counteracting mounted soldiers. Swordsmen are going to hang back while the archers level the numbers, then they will continue to hang back while soldiers with polearms handle the mounted soldier and leading footsoldiers, and then the swordsmen will go in. It is just good sense, right?

(*all said with the disclaimer that i have no training in medieval or feudal field combat or warfare, and that my opinion is based wholly on personal conjecture)

:)

[Kim Taylor]

Just finished listening to a podcast of a speech that Conlon gave at his University. He mentioned a few very interesting factoids and conjectures. For instance he mentions that in the 1300s the word we translate today as "loyalty" meant something more like "services rendered" in the sense of doing a job and requiring pay for said job. So the loyalty of the warrior in that age was actually more accurately described as "work for hire" of the warrior. One fought for pay. Conlon notes that the idea of the loyal samurai didn't show up until centuries later.

The records that we have been talking about above seem to be very common, and very accurate as to wounds. Conlon himself is the reader of the original sources rather than citing other authors on this.

Worth a listen, it's at the same website I mentioned above.

Kim.

[MithLuin]

If the records were very commonplace, then it's not unreasonable to think we have a fair sampling. But it is important to remember when dealing with events over 500 years ago that it is a sample, only. It's not like you can look up the exact number killed/wounded in battle like you could for, say, the war in Afghanistan. Even if detailed records were kept at the time, they have to be preserved, and it's the nature of history that much is lost in the intervening centuries. So much of what we know of medieval European history is at the mercy of some manuscript that someone found laying around in an attic or library. History isn't like contemporary news reports; there is definitely a healthy margin of error in any statistics like the ones presented.

But at the same time, the whole point of studying what remains is to gain a more accurate understanding of that time period. The types of wounds reported in the records we have available would certainly suggest something about the style of combat. I would not translate '6% of injuries were caused by swords' as 'swords were unimportant'. Swords are close range weapons. Arrows, guns, missiles - projectile weapons, which can be wielded from far away - are always going to cause more damage and have higher casualty rates. The long range weapons are used until you close to shorter ranges. Someone who carries their sword about town for personal defense is unlikely to leave it home when entering a battlefield...but that same person isn't going to fight with a sword at bow-and-arrow distance! The importance of the English archers at Agincourt didn't mean the English lacked knights or other weapons.

Consider modern warfare. What causes the most damage to US troops in Iraq and Afghanistan are IEDs and helicopter crashes. If you think that the insurgents aren't using AK-47s on a daily basis....well, you'd be wrong. Marines don't want to have to use their KA-BAR to kill someone...but they are still armed with that (and other weapons, too, of course). It's just much more of a last resort. If you're to the point of hand-to-hand combat, something has gone terribly wrong to lead to that situation.

[chidokan]

There are similar statistics in the rebuilt Osaka castle, again swords score low, missiles high. This stands to reason.... difficult to hit someone with a sword who is standing 50 meters away due to height of the walls and the moat.... far easier with bow or gun! Spears were the common poor infantry weapon (easy to make and cheap) which explains the higher ranking...once inside, swords/close combat weapons come to the fore.

[MartialArtsGirl]

BTW, I'm looking forward to the treatise on the use of magic spells and giant dinosaurs in mass combat.

LOL that really made me laugh! :)

[rainmaker]

Well... Thiniking about 2025, it would be

30% gun
60% Macdonald's trans fat
7% Car, wives run over hunsbands..
3% ???? I am sure you can fill with something...