SOURCE: “The Heike Monogatari and the Japanese Warrior Ethic,” in Harvard Journal of Asiatic Studies, Vol. 29, 1969, pp. 93-108.

[In the following essay, Butler argues that the code of the Japanese warrior as presented in the Heike Monogatari is more a creation of the tellers of the tales than historic fact.]

The Heike monogatari has exerted a strong influence on many aspects of the later development of Japanese society. In the political sphere, it is well-known that the accounts of warrior battles contained in the Heike provided a model for the attitudes and standards of conduct of the warrior class until the nineteenth, and even into the twentieth, centuries.¹ The degree of historical validity of these accounts, and their relationship to the actual battles themselves, is less well-known. The present article is an attempt to outline briefly the process by which the narrative passages, which taken as a whole define the warrior code at the time of the Gempei battles of 1180-1185, came to be a part of the Heike monogatari text. The manner in which these passages came to be accepted as historical fact, and how they came to be regarded as superior examples of the code of the Japanese warrior, is also considered.

It is possible to isolate in the standard Heike monogatari text all of the important qualities which later came to be accepted as the ideal and necessary attributes of a warrior. The concepts of personal loyalty, a willingness to sacrifice one's life for one's lord, and the determination to fight on to an honorable death rather than give in to a superior foe, to name but a few, all find frequent and varied expression throughout the sections of the Heike monogatari devoted to the Gempei battles.

It would seem to be a rather simple task to assemble a sufficient number of examples illustrating these concepts, arrange and classify them, and then come up with a model of the feudal warrior ethic as it existed in the late twelfth century. But before one becomes too involved in constructing such a model, it is well to give some attention to the circumstances of the textual development of the Heike monogatari. There exists an extremely large number of variant manuscripts of the Heike, more than a hundred, by recent estimates.² But a clear picture of the exact sequence in which the major variant manuscripts appeared, and the process of interaction between oral and written literature which produced the perfected narrative of the standard Heike text, has only recently begun to emerge.

For about the last six hundred years the text which has been taken as standard has been one dictated in 1371 by the famous biwa hōshi Heike singer, Akashi Kakuichi (d. sixth month, 1371). It has been realized by most people that there had been earlier written versions of the Heike monogatari. In fact, the original Heike monogatari has traditionally been thought to have been written during the time of the Retired Emperor Go-Toba (1198-1221), not long before the Jōkyū War of 1221.³ But little information has been available concerning the textual development of the Heike monogatari during the 150-year period from 1221 to 1371, and consequently the popular audiences after 1371 who heard or read Kakuichi's Heike monogatari text (and its later minor revisions) accepted it as a contemporary account of the events it relates. Even scholars, who knew of the late date of the Kakuichi text, tended to assume that in its essential parts the Kakuichi narrative was identical with the original version. Until quite recently, most scholars have only mentioned in passing that the Heike text has suffered some degree of modification as a result of its textual development, and they then go on to base whatever aspect of the period 1180 to 1185 they may be dealing with on the Heike monogarati narrative as it appears in the Kakuichi text of 1371.⁴

One of the factors which has been conducive to the adoption of such an approach is that it is possible to find in certain of the historical records of the Gempei period many entries which seem to corroborate the Kakuichi Heike narrative. In the past, the historical chronicle of the Kamakura feudal government, the Azuma kagami, has often been cited in asserting the historical accuracy of the Heike monogatari. We know that the Azuma kagami was not compiled before about the year 1270, or almost one hundred years after the Gempei battles. But the general consensus has been that the Azuma kagami entries are based on earlier records kept by the Kamakura warriors and that therefore they should be reasonably accurate. In the early portion of the Azuma kagami there are many entries concerning the major Gempei battles which accord quite closely with similar passages in the Heike. This correspondence tends to reinforce the impression that the major part of the Heike narrative should be accepted as presenting a valid picture of the battles.⁵ But on the basis of a preliminary comparison of the Azuma kagami with the earliest of the preserved Heike monogatari variants, which date from before 1270, the indications are that much of the early part of the Azuma kagami, particularly such sections as the ones describing Mochihito's flight from Kyoto, the Fuji River battle, Ichinotani, and Yashima, is in fact based directly on the accounts as they appear in the early Heike variants rather than on separate historical records. It is therefore rather risky to cite the Azuma kagami to substantiate the validity of the Heike monogatari narrative.

If we put aside the Azuma kagami, the one other major source most often cited to support the Heike narrative is the Gyokuyō, the kambun diary which was kept as a day-to-day record of the events of the period as observed in Kyoto by the Fujiwara nobleman Kujō Kanezane (1149-1207). There was a definite and legitimate influence from the Gyokuyō at the time the original Heike monogatari was written, but comparison of the Gyokuyō and the early extant Heike manuscripts reveals that the type of narrative element which finds substantiation in the Gyokuyō is not such passages as those that treat of the actions and mentality of the Heike and Genji warriors in battle. Instead, the sections which can be shown to be based on the historical Gyokuyō account are for the most part merely those which relate specific dates, or deal with such matters as who was where at what time, what kind of court ceremonies were held when and where, who was promoted to what office in the civil government, and the like.⁶ Kujō Kanezane was a civil official who did not leave Kyoto. With the exception of a few scathing attacks on warriors such as Kiso Yoshinaka, or accounts describing his own negotiations with the Genji leader Yoritomo concerning his later appointment as sesshō or “Regent,” Kanezane did not deign to mention much concerning matters pertaining to warriors.

So, if we get no help from the Azuma kagami or from the Gyokuyō in determining the validity of the warrior ethic as it appears in the Heike monogatari, we are left with little choice other than to do it on the basis of the variant Heike manuscripts themselves. Fortunately our understanding of the textual development of the Heike monogatari has improved greatly during the last decade, and it is now possible not only to appreciate much better the Heike as a work of epic literature but also, by comparative study of the early manuscripts, to gain a much more accurate insight into the events of the Gempei wars as related in this work.

In a recent study I have cited several major variants of the Heike monogatari which are crucial to any consideration of the Heike.⁷ Of these, four are of particular importance to an understanding of the development of the warrior ethic within the Heike text. The earliest of these manuscripts is known by the title Shibu kassenjō daisamban tōjō Heike monogatari, or “Heike monogatari, the third battle of a battle-record in four parts.” This is a version of the Heike monogatari written in a kambun style which verges on pure Chinese. On the basis of internal evidence I have shown that this text was written between the years 1218 and 1221. With the aid of these dates and also as a result of detailed textual comparison with the other early Heike variants, I have suggested that the Shibu kassenjō text is in fact the original written form of the Heike monogatari. The style of the Shibu kassenjō text is much closer to that of a straight historical chronicle than is the style of any of the other Heike variants, and by comparison with the Gyokuyō, it can be shown that the greater part of the factual information in the Shibu kassenjō is based directly on the kambun entries of the Gyokuyō.⁸ The second major variant is the Yashiro Heike monogatari text. The Yashiro text was written sometime after 1242, but probably before 1300. It represents the earliest attempt to revise the kambun narrative of the Shibu kassenjō original Heike monogatari for use as a fixed text for memorization and recitation by rhapsode-type biwa hōshi reciters. The Yashiro text is written in Japanese, rather than kambun, and uses the katakana phonetic script with Chinese characters. The third major variant is the Kamakura Heike text. The Kamakura text was made at some time during the period from about 1300 to 1340, and it represents an expanded and more sophisticated revision for recitation of the original Heike narrative. The Kamakura text is also in Japanese and uses the katakana script. The final major variant necessary for an understanding of the Heike monogatari is the Kakuichi text of 1371. In essence, the Kakuichi text is a combination of the best elements of both the Yashiro and the Kamakura divisions of textual development. Because of the perfection of its narrative, which uses the hiragana script rather than katakana, from about 1400 the Kakuichi text became accepted as the standard version of the Heike monogatari. Copies of the earlier variants such as the Shibu kassenjō, the Yashiro, and the Kamakura texts, dropped from sight. These texts were relegated to the archives of Buddhist temples and private collections, where they remained until discovered by manuscript collectors of modern times.

How does a knowledge of the textual development of the variant Heike monogatari texts aid in understanding the basis of the development of the warrior ethic in Japan? The best way in which this question may be answered is by reference to a particular passage as it appears in the four major Heike monogatari variants themselves.

In Book 7 of the standard Kakuichi text of 1371 a detailed account is given of the battles in 1183 north of Kyoto between the attacking Genji force of Kiso Yoshinaka, and the defending forces of the Heike.⁹ This section opens with the statement that Yoshinaka was poised in the north with 50,000 warriors, ready to attack the Heike in Kyoto. We are told that the Heike, led by Kiyomori's grandson Koremori, raised a force of 100,000 and advanced to meet Yoshinaka. The validity of these figures need not concern us, since they are probably inflated, but at any rate, in the present account, after a passage describing how the two opposing warrior groups are divided, we find Yoshinaka, with a force of 40,000, confronting the main Heike force of 70,000 in the Tonami mountains. Yoshinaka decides that since his force is so much smaller, he must rely on superior strategy to defeat the Heike. Accordingly he devises a plan of boxing up the Heike warriors in Kurikara Canyon, at the foot of the mountains. After the initial encounter, the Heike force is deceived into thinking that Yoshinaka's force is much stronger than their own, and at the time of the major charge by Yoshinaka, the entire Heike force turns and flees in confused retreat without joining battle. The Heike force flees into Kurikara Canyon, and Yoshinaka, as planned, boxes them in. Of the 70,000 Heike only 2,000 manage to escape with their lives. After this encounter, Yoshinaka takes his troops to help defeat the smaller Heike force of 30,000 at Mt. Shiho, and this Heike force also flees. With this defeat, the section describing the major encounter between the Genji and the Heike ends.

In the narrative up to this point the battle has been presented only in general terms, and there is no mention of fighting between individual warriors, or of any details which would provide a direct insight into the actual code of the warrior as it was expressed in action. On the contrary, by placing the emphasis in the narrative on the fact that the entire Heike force of 100,000 retreats without fighting at the first sign of the Genji battle flags, the Heike monogatari at this point seems to be advocating cowardice as the better part of valor, and this of course is the exact opposite of the warrior code as it is known in later ages.

But the narrative of the Kakuichi Heike text of 1371 does not stop here. After the description of the major battles at Kurikara Canyon and Mt. Shiho, there is a series of accounts of battles between individual warriors. It is in these later sections that the narrative elements commonly accepted as reflecting the warrior code of the Gempei wars emerge.¹⁰ First there is a section which bridges the gap between the general battle and the individual acts of heroic sacrifice which are to come. In this section we are told that before the Heike warriors set out from the capital, a certain Saitō Bettō Sanemori had tested twenty of the leading Heike warriors, who had their origins in the East, by suggesting that since the Genji were obviously going to win, they should desert the Heike and align themselves with the Genji forces of Kiso Yoshinaka. The men are about to decide to do this, when one among them remonstrates with Sanemori saying that it is most disgraceful for a warrior to change from one side to the other according to the will of fortune, and he urges them all to die in fighting for the Heike. Saitō Bettō Sanemori replies that he had only been testing the loyalty of the warriors, and that they should all make a vow to die for the Heike. In an aside, the Heike text tells us that tragic as it may be, all twenty of the men kept their vow and died at the time of the battles in the north with Yoshinaka.

Here we have a classic expression of a concept which later became central to the code of the warrior in Japan. A warrior should not serve two lords, and should choose death in battle rather than go over to the enemy. Following this section in the Kakuichi text it is related that after fleeing from Kurikara Canyon and Mt. Shiho, the remnants of the Heike force stopped to rest at Shinohara in Kaga. Yoshinaka pursues them, and it is at this point, coming after the previous brief section stressing the loyalty of a warrior, that we get three actual examples of the warrior code at its best.

After describing the preliminary encounter at Shinohara, an account of the fortunes of one Takahashi no Hōgan Nagatsuna is given. He joins battle with 300 of the Genji with a mixed force of 500 warriors from various provinces who have straggled into the Heike camp. Although Takahashi fights valiantly, his men are a motley lot and most of them flee. Takahashi himself is finally forced to retreat, but his intention is to try to find some of his own loyal personal retainers and once again to join battle with the Genji. As he is searching for his men, he encounters a young Genji warrior named Yukishige. Takahashi starts to kill him, but then remembers that his own son who had been killed in battle the previous year was about the same age as Yukishige, and he decides to spare the young Genji warrior. But Yukishige, for his part, is filled with warrior spirit. Even though Takahashi has just spared his life, Yukishige waits for a chance to take Takahashi by surprise and kill him. This he does, with the aid of three of his retainers who happen along, and thus ends the great warrior Takahashi.

The second example of a warrior in action concerns a certain Arikuni, who penetrated far into the Genji ranks with a force of 300. The Heike monogatari narrative is particularly vivid at this point, and is worth quoting in part.

Arikuni, having penetrated deep within the enemy ranks, exhausted all his arrows and had his horse shot from under him. Then afoot, he drew out his sword and fought on, killing many warriors. Finally, pierced by seven or eight shafts, he met his death, still on his feet. Their great general thus fallen, his forces all fled in retreat.¹¹

The final example appearing in this section of the Kakuichi text is one involving Saitō Bettō Sanemori, the warrior who appeared earlier as the one who admonished his twenty men to die fighting. This passage is rather long, and there is no need to trace it out in detail. The importance of it as it bears on the warrior ethic is that after the Heike forces have been completely routed and are retreating as fast as their horses will carry them, Sanemori alone among the Heike warriors refuses to retreat, and despite his advanced age, constantly turns back to engage the enemy and protect the rear of the retreating Heike. He is finally killed, but not before he has displayed by personal example the manner in which a true warrior should act in battle. The Sanemori section ends the portion of the Kakuichi Heike monogatari text dealing with the battles in the north against Yoshinaka. Not long after this time, Yoshinaka and his forces enter the capital, and the scene soon shifts to an account of the battles in the west—Ichinotani, Yashima, and finally Dannoura, where the Heike meet their end.

On the basis of the four sections coming after the general account of the Kurikara Canyon battle in the Kakuichi text of 1371, we can discern several of the qualities previously mentioned which later ages held up as shining examples of the Japanese warrior code. In the earlier Sanemori section we saw that loyalty to one's lord, even if the cause is hopeless, was defined as one of the cardinal virtues. In the Takahashi passage, we get the young warrior Yukishige appearing as the cold-hearted warrior who thinks only of killing the enemy, even if the enemy, as in this particular case, has shown him mercy. The actions and feelings of Takahashi in this passage are almost identical to those of Kumagai Naozane in the famous Atsumori section of the battle of Ichinotani, related in Book 9. Naozane, however, while regretting his action, does in the end kill Atsumori.¹² In the Takahashi section, Takahashi's actions should probably be interpreted as an example of the fact that if one's enemy is weak, the true warrior should take advantage of this and accomplish his objectives with no thought of human feelings. The Arikuni section, while short, presents an example of the warrior fighting on to the bitter end, and in the final Sanemori section once again we get the picture of the true warrior refusing to retreat before superior odds and sacrificing his own life so that his lord might live.

All of these actions are examples of the qualities that have been emulated by Japanese warriors of later ages, and are also the ones which together with a liberal dosage of Confucian teachings were developed into the code of Bushidō by such theorists as Yamaga Sokō (1622-1685) during the Tokugawa period. If these events as related in the Kakuichi text of the Heike monogatari were only true or at least bore some demonstrable resemblance to the actual situation, there would be no problem, and a survey of the warrior ethic in the Heike monogatari would end at this point. But when we look at the Shibu kassenjō text of 1221, and also the other early Heike variants, the picture is quite different.

In the Shibu kassenjō text there is a general description of the battles at Kurikara Canyon and at Mt. Shiho which is more or less identical to that of the Kakuichi text. The numbers of warriors involved are stated as being slightly less, but for the most part the account is the same. But following this general account of the major battles, in which the Heike are described as fleeing without attempting to fight, we get a very different sort of narrative. First, the sections of the Kakuichi text devoted to Sanemori and his twenty warriors, the adventures of Takahashi and Yukishige, and the death in battle of Arikuni are completely missing. We do get the final section describing the death of Sanemori in battle, but even here the emphasis is entirely different. In the Shibu kassenjō version Sanemori is not presented as having remained behind to protect the rear. Instead, he is fleeing like all the rest, and it is only because he is old that he is overtaken by a younger Genji, who dispatches him with ease and takes his head to Yoshinaka. In short, at no point in the Shibu kassenjō account, which is the version of the Heike mongatari written at a time closest to the events it relates, is there any developed presentation of individual warriors acting in a way we have come to believe they did on the basis of the narrative of the standard Kakuichi text of 1371.¹³

When we consider the Yashiro text, we can see some expansion of the narrative, but it also is still far from presenting any sort of concrete picture of a developed warrior ethic. The story about Arikuni has been added, in a rather terse form, but both he and Sanemori are still presented as having become separated from the main group of fleeing Heike, and fighting for their lives out of necessity, rather than choice.¹⁴

It is only when we get to the Kamakura text, which was not written until about 1300, more than one hundred years after the Gempei battles, that we find the narrative developed to a point approaching that of the Kakuichi version of 1371. In the Kamakura text, in addition to the story of Sanemori, there are also the stories of both Arikuni and Takahashi. The only major difference with the Kakuichi text is that the final brief general battle at Shinohara is not described at length.¹⁵ But even more importantly, in the Kamakura text we find a major shift in emphasis. No longer is Sanemori presented as an old warrior who has fallen behind. Instead we find him possessing all the attributes of a great hero in the best tradition of the warrior ethic. Sanemori stays behind by choice, to defend his lord who must live to fight another day. The other warriors, Takahashi and Arikuni, have also taken on heroic attributes. In the earlier Yashiro text Arikuni is presented not as having charged into the enemy ranks, but rather as having fallen behind along with Sanemori in attempting to flee and as having been finally overtaken by the enemy who shoot his horse from under him and then kill him. In the Kamakura text Arikuni appears as a stout-hearted warrior, one of the leading generals in the battle, who charges into the enemy ranks in much the same manner as he does in the account quoted from the Kakuichi text.

There is much more that can be said about the evolution of the Heike monogatari narrative throughout the variant texts, but it should be clear from the brief outline given that the elements of the Heike narrative which in the past have commonly been accepted as representative of the warrior spirit at the time of the Gempei wars of 1180 to 1185 did not enter the Heike text until about 1300 at the earliest—or more than a century after the actual events. If we examine all the other battle sections of the Heike monogatari as they appear in the four major variants, the situation is the same. All of the battles begin with a general description of the large battle itself. Then the general account is followed by a section in which the actions of the individual warriors are described. The Shibu kassenjō text, however, contains the general account of the battle, but lacks the appended sections of the actions of the individual warriors. Similarly, it is only at the stage of the Kamakura text of about 1300 that we find the accounts of the individual warriors appearing in a fashion similar to that of the Kakuichi text.

There should be little doubt that the details of the heroic actions of the warriors as they are presented in the Kamakura and Kakuichi texts are entirely fictional. There are no acceptable historical records which corroborate any of the Kakuichi text narration of the specific incidents of the Gempei battles. Since little or none of this detail appears in the original Shibu kassenjō Heike text and very little of it in the Yashiro, it is difficult to imagine that suddenly about the year 1300 someone would discover complete historical accounts of these battles and insert them into their appropriate places in the Heike narrative. But the fact remains that these accounts in the standard Heike text have, with only minor reservations, been accepted over the centuries as being more or less historically accurate. There is an inherent, built-in feature of the standard Kakuichi Heike monogatari narrative which to a large extent explains this rather unusual phenomenon. The remainder of the present article is devoted to outlining how this came about.

Anyone who has read the battle sections of the Kakuichi text cannot help but have been impressed by the detailed and vivid nature of the description of the warrior and his actions. These passages seem to have a certain living force of their own, which tends to pull the audience into the narrative and causes one to feel that he himself has been a part of what has taken place. Space limitations prohibit a detailed discussion of how the narrative of the entire twelve books of the Heike monogatari reached its perfected written form, but it is possible to state that much of the striking effect of the battle tale narrative in the Kakuichi text is accomplished by the use of narrative techniques which had their origins in orally composed battle tales sung by a class of singers who originally accompanied the warriors to battle.¹⁶

The details of orally composed tales, no matter what the country or what the subject matter, are by definition largely fictional. Oral tale singers utilize a very definite method of composition based on the use of formulas and a formulaic method of elaborating themes. The fundamental features of such a method of oral composition have been defined by Professor Albert Lord of Harvard in his book The Singer of Tales.¹⁷ The oral singer traditionally takes his subject matter from historical events and personages, but the specific content of his story is entirely fictional. He does not recite his tale from memory, as was the case of the recitation of the written Heike monogatari after 1371. Rather, each time a true oral tale singer tells his tale he recreates it anew by means of oral formulaic techniques of composition. The oral singer is able to establish a special rapport with his audience by means of this method of composition, and even though the audience might know in one part of their minds that the details of the tale sung by the singer cannot possibly be true, by the very nature of the art the audience is drawn into the story, and by the end is left with the distinct feeling that what they have heard is an accurate version of the historical event as it actually happened.

We can see the hand, or rather the mind and voice, of the oral tale singer at work in all of the battle tale sections of the Kamakura and Kakuichi Heike monogatari texts. The largest themes contained in these two texts are the overall battles themselves. But it is in the latter sections of the battles, the ones describing individual warriors, that we find descriptions which betray a stereotyped pattern which establishes them as having been developed originally by means of the process of formulaic composition and thematic elaboration which is the hallmark of the oral tale singer. This kind of oral composition can be classified on the basis of the themes related. By way of illustration a short section from the Kakuichi text is quoted below, which in addition to displaying formulaic language, contains classic examples of two of the shorter oral tale themes, those of “dressing the hero,” and “naming one's name,” which appear frequently throughout the Heike monogatari.

The setting is the Battle at the Bridge at the time Prince Mochihito and his Genji supporters are fleeing from Kyoto in 1180. A monk-warrior from among the temple troops supporting Mochihito steps out to fight, and is described as follows:

From among the monk-warrior group Tsutsui no Jōmyō Meishū wearing black leather-threaded armor over a dark blue “victory” robe, and a five-plated helmet, and with a black-lacquered sword at his side, and a quiver of twenty-four black-feathered arrows hung at his back, together with a lacquered bow and his favorite white-handled halberd in his hands, advanced onto the bridge. In a mighty voice he named his name, saying: “You have long heard of me, now take a good look. I am Tsutsui no Jōmyō Meishū, known to all of Mii Temple as a warrior worth a thousand men. Let those who will advance, and we shall see the outcome,” and he then mercilessly let fly his twenty-four arrows. Immediately twelve warriors fell dead and eleven were wounded, and still one arrow remained in his quiver. He threw down his bow and stripped off his quiver. Then, kicking off his foot-gear, he sprang barefoot onto the beams of the bridge, and charged across. None would advance to meet him, and Jōmyō proceeded on, as though strolling the wide streets of the capital.¹⁸

The theme of “dressing the hero” as it is applied to the monk-warrior Meishū in this passage is almost identical with a similar theme used repeatedly in the Iliad and the Odyssey, and it also resembles one used in the Yugoslav oral tale tradition as described by Albert Lord.¹⁹ Both the themes of “dressing the hero” and “naming one's name” are repeated over and over in the Heike monogatari, and it is on the basis of such minor themes as these and others that the larger themes of entire battles are put together. Furthermore, it is the kind of specific detail that a singer of oral tales is able to include as a result of the process of formulaic oral composition which builds a theme, that contributes the realistic element to the narrative. In addition to exact descriptions of his armor, in the passage quoted we are told that Meishū was carrying not just a halberd, but rather a “white-handled” halberd. And not only this, but it was his “favorite” white-handled halberd. Although such detail-producing formulas are repeated over and over throughout the Heike monogatari, this fact should not lead one to assume that their use is in any way stereotyped or hackneyed to an extent which would limit the range of personalized description in any given case. Concerning the monk-warrior Meishū, the “dark” and “black” images in the early formulas describing his robe, sword, and arrows, when viewed in comparison with dozens of other such passages in the Heike, particularize Meishū as an especially “ferocious” warrior, one indeed who is “worth a thousand men” (another recurring formula) in his own distinctive way. The point to make is that while oral formulaic repetition distinguishes the Heike monogatari as being at once different from a normal written work of literature, the technique itself admits of a particularized use which transcends a mere mechanical repetition of set, stereotyped phrases, and this in turn works favorably towards creating the special aura of authenticity which surrounds all such epics emanating from an oral tradition.

It is by the method of stringing the smaller themes (or motifs, as they are sometimes called), which themselves are based on formulaic composition, together into larger themes that the type of battle narrative we find in the perfected Heike monogatari is produced. There is an overall governing principle to this method in the orally composed literature of any country. In the case of the Heike monogatari, this governing principle demanded that after a general description of a large battle, there should be a series of accounts describing the actions of individual warriors. But the actions as they are described are not factual. Rather they are defined and limited by the range and extent of the oral formulas and themes available to the singer in his tradition. As an example of how this works out, it will be recalled that the actions of Takahashi in sparing the life of Yukishige are essentially the same as those of Kumagai Naozane in his encounter with Atsumori. In terms of oral composition both Takahashi and Kumagai Naozane are in fact the same person. Their actions might be described in oral terms as the theme of “the older warrior sparing the life of a younger warrior because of resemblance to his own young son.” There is no way of knowing whether Takahashi or Kumagai Naozane actually performed actions similar to those attributed to them in the Kakuichi text of 1371. It is certain, however, that no matter what their actions, the details were not like what is described.

The descriptions of the actions of warriors in the other battles recounted in the Heike monogatari can also be reduced to general stereotyped themes of a similar type, which can be shown to be the kind of themes that the oral singer has at his disposal. As for why these oral themes appear in the late Kamakura text but not in the earlier Shibu kassenjō and Yashiro texts, this has to do with the manner in which the different texts were composed. First, it can be shown that it takes considerable time for a tradition of oral tales such as these to develop to perfection. The available evidence points to the interpretation that these oral battle tales were not current in Kyoto at the time both the original Shibu kassenjō text and the Yashiro revision were made. But it also appears that the Kamakura text was made at the Buddhist center of Shoshazan in Harima province. The indications are that after the last of the Gempei battles, a group of oral battle tale singers settled at Shoshazan and that it was on the basis of the tales they sang that the existing Heike monogatari narrative was revised to include these elements.²⁰ This process produced the text we now know as the Kamakura text of the Heike monogatari. The Kamakura text was subsequently revised further by Kakuichi, using techniques of oral composition, and this resulted in the final perfected Kakuichi version of 1371.

In conclusion, two final comments may be made concerning how a knowledge of the oral basis of the battle sections applies to understanding the Japanese warrior ethic as it developed in conjunction with the textual evolution of the Heike monogatari. First, it is because of the techniques of oral composition that we get our warriors, who in the Shibu kassenjō text are presented in very human terms, changed into stereotyped heroes exhibiting all of the qualities later accepted as proper for a warrior. In this respect, the warrior ethic as contained in the Heike monogatari is not the result of an attempt at a factual description of the actions of actual warriors in the Gempei battles. Rather it should properly be attributed to the skill of the oral tale singers as they developed their tales in later years. We therefore have the paradox of the Japanese of later ages modeling their actions not on those of the Gempei warriors as they actually were, but rather upon the ideal warrior as conceived by oral singers who formed their heroes by means of formulaic techniques of oral composition. Secondly, it is as a result of the combination of oral tale narrative of this type, which itself has elements conducive to its acceptance as historical fact, with the completely factual elements derived from the Gyokuyō and other court records, that there was produced the perfected Kakuichi version of 1371 of the Heike monogatari, which has an imprint of reality sufficient to make it acceptable to later Japanese as an historically accurate and valid account.

Notes

1. See Edwin O. Reischauer and John K. Fairbank, East Asia: The Great Tradition (Boston, 1960), 534-544.

2. An early attempt at classification of the variant manuscripts appears in Takahashi Teiichi, Heike monogatari shohon no kenkyū (Fuzanbō, 1943), 569 pp.

3. See my article, “The Textual Evolution of the Heike Monogatari,” HJAS 26(1966).16-17.

4. Two of the best such studies are Ishimoda Shō, Heike monogatari (Iwanami, 1955); and Sasaki Hachirō, Heike monogatari no kenkyū (Waseda, 1948), 3 vols.

5. For a short general discussion of the relationship between the Azuma kagami and the Heike, see Minoru Shinoda, The Founding of the Kamakura Shogunate, 1180-1185 (New York, 1960), 9-12.

6. In his monumental study, Heike monogatari ryakuge (Hōbunkan, 1929), Mihashi Tokugen has pointed out many of the passages in the standard Heike text which have a basis in the Gyokuyō and other court records of the period.

7. “The Textual Evolution of the Heike Monogatari,” 8 ff.

8. The most recent study of the Shibu kassenjō and Gyokuyō connection is Shida Itaru, “Rekishi sono mama tor ekishibanare—Shibu kassenjōbon Heike monogatari o megutte”, Bungaku (Nov. 1966).11-21.

9. Heike monogatari (Iwanami, 1959, 1960), 2 vols., II, 61-82.

10. Heike monogatari, ii, 75-82.

11. Heike monogatari, ii, 78.

12. Heike monogatari, ii, 219-222. Translated in Anthology of Japanese Literature, Earliest Era to Mid-Nineteenth Century, ed. Donald Keene (New York, 1955), 179-181.

13. Shibu kassenjōbon (Shōwa Women's University MS.), 7.7a-27b. Atsumi Kaoru, in Heike monogatari no kisoteki kenkyū (Sanseidō, 1962), 353-356, was the first to point out the evolution of these passages through the four variant texts. Her discussion and the charts appearing on pp. 407-408 were of great help in the early stages of preparing the present article.

14. Yashiro Heike monogatari (Kadokawa, 1966), 511-516.

15. Kamakurabon Heike monogatari (MS.), 7.15a-18b.

16. For a general account of the oral basis of the Heike monogatari, see my article, “The Heike Monogatari and Theories of Oral Epic Literature,” Bulletin of the Faculty of Letters, Seikei University 2(Tokyo, 1966).37-54.

17. The Singer of Tales (Cambridge, 1960).

18. Heike monogatari, i, 310. The equivalent passage in the Shibu kassenjō text appears in Nomura Seiichi, “Shibu kassenjobon Heike monogatari kan-shi”, Bungaku, (Nov. 1966).96. This Shibu kassenjō passage, which does not display the oral thematic technique in developed form, is discussed in Yamashita Hiroaki, “Heike monogatari no katari”, Bungaku, (Nov. 1966).7-10.

19. James I. Armstrong, “The Arming Motif in the Iliad,” American Journal of Philology, 79.344; The Singer of Tales, 92.

20. “The Textual Evolution of the Heike Monogatari,” 37. See also note 144.