One or Two Generations

After the Tokugawa made the final steps of unifying Japan, swordsmiths adopted more clear traditions of signing swords and dating them became much more common. The information they left behind and the fact that we’re dealing with “near history” makes it easier to understand swordsmith lineages. 

When it gets into the Muromachi period and earlier, things get a bit more murky. Many signatures were lost, dates are few and far between, and period specific references can contradict each other. 

In the modern period, with swords accessible to everyone and importantly with the work of the NBTHK passing Juyo blades and publishing them, the picture has become more clear. We owe a lot to Fujishiro Yoshio who’s work in the early 1900s on reference materials is more often right than wrong. So I’ll start this discussion with some of his general thoughts on this matter of one or two generations.


Fujishiro was a skeptic and he tended to not believe in theories that involved splitting major swordsmiths like Kanemitsu and Rai Kunitoshi into first and second generations. These are the most clear examples as the work of both smiths is plentiful, there are many signed and dated examples available, and both were heads of very successful schools. They were famous in their time and famous until now. 

Both of these smiths had shifts in their work styles, and this change in style is generally the reason why people began to believe in two generations. It’s not a logical thought at all though: you learn whatever you learn from your teacher. 

Thus, your initial style should be your teacher’s style. Any dramatic changes in style are the inspired work of a master evolving his own work into something new, rather than a raw student fresh to his name and starting from scratch.


Kanemitsu’s work style begins looking just like Kagemitsu in kataochi gunome and with some suguba works. During his time swords evolved to very large sizes and with the shift of sugata and the influence of Soshu, his work moved into notare based forms and chikei became emphasized and the size of the blades grew. This change is very conspicuous around Enbun (1356) so sometimes work after this period is called Enbun Kanemitsu and those works have also been attributed to a second generation Kanemitsu who was a student of Masamune.

In looking at the dates though and all of the Osafune works, one smith, Osafune Nagashige shows the first influences of Soshu den around 1332-1334 when his style changes and Honma Junji believed him to have worked directly with Masamune. Nagashige historically was thought to be a younger brother of Chogi, but now with work dates available for everyone to see, the relationship is clearly inverted. Nagashige seems then to be the jumping off point for Soden Bizen and brought the Soshu style to Osafune where it was picked up by Kanemitsu and later Chogi. 

So there is no second generation Kanemitsu but rather it was a master evolving his style based on some new techniques imported from Kamakura.

Previously, the Kanemitsu of the Enbun era was taken as the nidai, but I think this is a continuation from the shodai Kanemitsu and is the same person. It is difficult to believe that he “Returned to being a pupil of Sōshū Masamune at the age of 42 in Gen’ō”. Rather, I think the several facts are that Kanemitsu became the head of the so-called Sōshū Den.

— Fujishiro Yoshio

At the time Fujishiro wrote this, as mentioned above, Kanemitsu was thought to be two generations so it is a pioneering thought to believe that it was a change in style. The statement that he became to be the head of the Soshu den, I think means that he became the top master of the style. During the end of Kanemitsu’s tenure, just before the Muromachi period begins, there are no great smiths working in Kamakura anymore. Akihiro was the last great Soshu smith and was in charge of a waning school, while Kanemitsu’s at the same time was flourishing with many talented students. After Kanemitsu and for a long time, there were almost no good smiths anywhere in Japan other than in Osafune, until the end of the Muromachi period.

From long ago there was one generation of Kanemitsu, but the origin of the story that there were two generations around Kenmu and Enbun was for the reason that the styles were different. This was not because the style changed when it continued from the teacher to the pupil, but after making the style of the teacher one’s own, due to the demands of the times, it changed during one generation of the swordsmith who brought it forth.
— Fujishiro Yoshio
This is a bit more awkward phrasing of what I said above. Your initial style is your teacher’s style. It doesn’t change when handed to you, but it changes during your own work period naturally, if it changes at all. So, he clearly understood this and uses this logic to assist with the shooting down of a two smith theory for Kanemitsu.
The following examples show Kanemitsu signing and working in the suguba and kataochi gunome of his father Kagemitsu. The first example even shows a sansaku style boshi, and the second example the horimono style we see on Kagemitsu tachi.
Signed Kanemitsu in Kagemitsu style (note Sansaku boshi)
Kanemitsu in Kagemitsu style (note typical Horimono and suguba) dated 1334

Contrast these with the Enbun style. The boshi has become a new original style, the hamon becomes based on notare, and the kissaki enlongates. Previously this would be the archetype of second generation work, but this is the ultimate masterpiece of Kanemitsu hybridizing Bizen with Soshu traditions.

Kanemitsu after Soshu influence: o-kissaki, notare hamon and 1360 date

Rai Kunitoshi

Rai Kunitoshi helped us a lot by leaving behind a fair number of dated blades, and as well in a couple of times he wrote down his age at 75 years on a dated blade from 1315. This allows us to know his year of birth: 1240. His last dated blade is 1321 so he worked to a remarkable 81 years of age. European life expectancy must have been about 35-40 years around this time, so it is very remarkable to understand that he was not only alive but working and productive at this advanced age. This also sets the table for a fact that we cannot dismiss long lifespans and work periods for Japanese swordsmiths and immediately cut them into generations because of this. 

The teacher of Rai Kunitoshi is Rai Kuniyuki. His work period begins around 1240 and if he had his son around 20-30 years of his own age, that would put his birth year around 1200-1210. 

Rai Kunitoshi would have been working in his shop from a young boy, and possibly been making his own swords around 20 years old. The first signature we see for this smith is signed in two characters as 國俊 (Kunitoshi) which resembles his father’s signature. Midway in his life he added the character Rai to make a three character signature 来國俊 (Rai Kunitoshi). 

Now, people usually say the style between both signature types is different and so they were two smiths. The two character signature is referred to as Niji Kunitoshi (literally meaning two-character Kunitoshi) and the work looks like Kuniyuki with choji midare, while Rai Kunitoshi is in suguba.

These are gross simplifications however. There are works of Niji Kunitoshi which are signed and evolving into suguba and there are works of Rai Kunitoshi which are in choji midare and signed.

Signed Niji Kunitoshi showing transitional mix of choji based in suguba.
Signed Rai Kunitoshi in choji midare
1281 dated Rai Kunitoshi but only nengo preserved, so attributed to Niji Kunitoshi, based in suguba

This final piece is the last dated work of Rai Kunitoshi when 81 years old. It is generally considered his masterpiece work and recalls the Niji Kunitoshi style and probably is a joint work with Rai Kunimitsu who may have done the signature. If made O-suriage I think this would be attributed to Niji Kunitoshi based on the style, and it illustrates that this style belongs to Rai Kunitoshi.

Signed Rai Kunitoshi, choji midare mixed with suguba


Furthermore, the signature does not change radically overnight. The KUNI character in the signature is a giveaway as the three horizontal strokes begin in the upper corner of Niji Kunitoshi and migrate down over time to be spread uniformly over the full vertical space. When RAI is first added to the signature it is simply a prefix to the Kunitoshi as is found in Niji Kunitoshi and then over time the KUNI evolves. So we can see some natural progression in this.

The signature marked Intermediate is taken from a blade with the three character signature and you can see that the KUNI is identical to the work of Niji Kunitoshi to the left of it. 

The 1289 dated blade is a tanto that I owned and passed Juyo after the death of Dr. Honma. The work was not known at the time of the 20th century books being written. It is the earliest known date of a three-character signature for this smith, and at the time he made that blade we know he was 46 years old due to his preserved age on the 1315 blade.  

There are dates on the two-character works as well, the last of which is 1278 and he would have been 38 years old at this time. 

So it becomes quite hard to imagine a scenario where Niji Kunitoshi can be the father of Rai Kunitoshi, given that his own father is Rai Kuniyuki who is a middle Kamakura period smith. And we know Rai Kunitoshi has to be born in 1240 himself. As well, the signature does slowly evolve from Niji style to Sanji style, and there are Sanji works like this one that are in Niji Kunitoshi work style. There is no other rational explanation for this other than that it is simply one man making a progression in style over the course of his lifetime, walking a path, and evolving.

The theory that two kanji Kunitoshi and Rai Kunitoshi were separate persons is a theory that ignores the style changes of the times. In the Shōwa Gonen Book section of the Kanchiin Hon, beneath a Kunitoshi two kanji signature nakago, it says, “Rai Magotarō Nyūdō, kogiri yasuri, becomes wide sugu yakiba”, and this is a record which coincides with the time when Kunitoshi and Kunimitsu nado were living.
What became clear in later years was that it was discovered that his works covered a long period of time, due to works inscribed with Shōwa Yonen Nanajūgosai (Shōwa 4, 75 years of age) and Bunpō Gannen Seinen nanajūhachi (Bunpō 1 age 78). Also, the two character Kunitoshi period coincides with the prime of his life, and backs up the one generation theory.  
— Fujishiro Yoshio
At the end of his life, Kunitoshi had three master students working for him and signing blades with his name. So, while we can consider him productive in his 70s, probably the majority of the work was being handled by Kunimitsu, Kuninaga and Kunitsugu. These smiths used code in how they signed Rai Kunitoshi to distinguish which smith did the work. 

Fujishiro figured this out with Rai Kunimitsu and Rai Kunitsugu, and I extended it to find Rai Kuninaga in the same body of work.

Though Fujishiro set the groundwork for this, Tanobe Michihiro of the NBTHK pretty much shut the door on the two smith theory and has stated there is only one Rai Kunitoshi smith with two styles over his long life. 

In spite of this, the NBTHK continues the tradition of attributing to Niji Kunitoshi and Rai Kunitoshi as if they were separate smiths. This is done to keep consistent with tradition and old attributions, as well as to recognize that this is just a theory still (though I think it is conclusive). It does help as well to separate out the period of a particular work. Niji Kunitoshi attributed blades will generally be from the first part of his work and closer to middle Kamakura, while Rai Kunitoshi attributed blades will be from the last half of his life and more like suguba in style.


Nagamitsu is another smith who has been broken into two traditionally. We can track the same evolution in style as with Rai Kunitoshi, where Nagamitsu inherits choji midare from Mitsutada, it eventually becomes suguba. The wide and powerful middle Kamakura shape also becomes more elegant. I think that the Soshu den may even have been concieved by Yukimitsu, Norishige and Masamune as a counterpoint to these elegant and beautiful styles coming from Osafune and Kyoto as it is deliberately vibrant and masculine in this time of elegant blades.

Nagamitsu further confused the situation by adding Sakonshogen to his signature in his later years. So traditionally it was thought that this Sakonshogen Nagamitsu was the second generation.

This held up until fairly recently, and it is common to find blades attributed at Juyo Token to Nidai Nagamitsu or Sakonshogen Nagamitsu when mumei. This habit stopped about 30 years ago as it became more clear that it is one smith with an evolving style, at the same time the door shut on the two smith theory of Rai Kunitoshi.

So when you look at those older Juyo papers and they say second generation Nagamitsu, understand them now to mean it is first generation (only generation) Nagamitsu but from the second half of his life.

He is the son of Mitsutada, and is of the legitimate line of the Osafune Kaji, which is the one to which the majority of the Bizen Kaji belong. He is called Saemonnojō, and as for Sakonshōgen, the prevalent story
is that this was the nidai, but I think they are the same person.
His works are largely tachi, and he is said to have received the Gō of Junkei in his
latter years, there are points that I am not convinced about in regard to these mei kanji and the works which have the style of Ko-Bizen, or I wonder if these are the early mei of Nagamitsu. In the beginning, his blades were like those of Mitsutada and were tempered with ō-choji, and later he changed to a samishii hamon of sugu choji or ko-choji suguba nado.
— Fujishiro Yoshio
Again Fujishiro rejects the commonly held idea at his time of two smiths and unites the work under one smith with a style change over time. 
He brings up another issue which is that there is a smith who signs Junkei and the question is still a bit open on this smith and who he is. Some have said it is Nagamitsu in retirement, and Fujishiro brings up that it might be Nagamitsu at the beginning of his life. When I asked Tanobe sensei about this, he said that Junkei is not one of the names of Nagamitsu and it’s a separate smith. It may be that associating the two a few centuries ago was a way of selling some Junkei swords, and at this point in time it is certainly not correct to assume that a Junkei attribution or signature means Nagamitsu.
He changed from ō-choji to sugu choji, then in his later years when he took the name of Sakonshōgen, he made suguba and sugu ashi iri. Rai Kunitoshi nado of this same period changed from an exuberant choji ha to the suguba type. This was not a period of deterioration in technology due to changes in ownership (generation), but was something that was demanded. This is clear when you see the prosperity of Nagamitsu and the Kunitoshi Mon.
— Fujishiro Yoshio

Part of the reason for this change may be due to the fact that both Rai Kunitoshi and Nagamitsu had large numbers of master-level students and followers who were capable of forging blades in these styles with high quality results, and these would be signed by the master or by the student in the master’s name. It may have been easier to make a regular and standard style that everyone could follow, but this is just me wondering about the issue.

Compare the various styles of Nagamitsu.

Signed Nagamitsu in Mitsutada style (oldest)

This next blade is signed with the Sakonshogen signature and a 1303 date. The suguba would make it a very typical work and with the signature change and style differences you can see the basis for an argument about a second generation.

Sakonshogen Nagamitsu, 1303 date.

However work has been found that shows the earlier choji based style and the Sakonshogen signature, showing that there are intermediate steps between the two and they are all by the same smith. This has a 1294 date.

Sakonshogen Nagamitsu, 1294 date.


This is another very talented smith who has been broken into two generations. I wrote a separate article on him, but there is another smith called Ko-Motoshige who is an Aoe school smith and signed only with two characters. Certainly this smith’s work combined with normal Motoshige looking like it has some Aoe influence, confused the issue for determining generations.

Motoshige’s initial work style is classical similar to Kagemitsu and then evolves around Enbun into a larger sugata and more like Aoe. Possibly this is a counterpoint to the work of Kanemitsu and the other Osafune smiths changing at this time. 

The two people, the older Ko-Motoshige, and the Motoshige who was one of the three disciples of Sadamune, are made out to be separate persons, but the denial of the three disciples of Sadamune is for the same reason as that of the ten disciples of Masamune. The older Motoshige is seen as being of the early period, and is thought to have been rather a senpai.

The reason the styles of the earlier Motoshige and the Motoshige after Kenmu are different are due to the changes of the times. His works can be seen covering a time span of over fifty years, from Kagen to Enbun and Jōji, and the long life of Motoshige is the subject of legends. Accurate proof cannot be offered from the point of records for during the kotō period, but during the shintō period, there are many records of kaji with long working lives of fifty or sixty years.

— Fujishiro Yoshio

To clarify some things about this statement, first is that the translator of this book made a mistake when he translated Ko-Motoshige, and misread the Ko character for Rai, so in the english translation it says Rai Motoshige. Fujishiro is referring to Ko-Motoshige who today is understood to be an Aoe smith.

So the first step in this statement is to clarify that there is indeed an older Motoshige but that this work is distinct from the main body of the Osafune work. It is the Osafune work that people want to divide into two smiths and this part he goes on to say is not correct.

Motoshige in Kagemitsu style, dated 1326.
Motoshige dated 1351, more influence from Aoe.



Now as a counterpoint to this unifying of two-generation theories, Muramasa was long thought to be one smith and someone from the Kamakura period. This is how it appears in folk tales of the Edo times though sword scholars probably understood it better. For folk tales it was easier to amalgamate the various Muramasa generations into one figure and place him in the Kamakura time to make a dramatic counterpoint to Masamune since the blades had so much bad mojo associated with them. So in this case, we think there are between 3 and 4 generations of Muramasa looking at the work and certainly there are at least two. This illustrates that it is not a uniform rule that where there is a two-generation theory we should condense it down to one.

Yukimitsu, Norishige, Masamune

From time to time it behooves people to create multiple generations of these smiths. There are not a lot of signed works available from any of them, and dated work is restricted to three blades by Norishige one of which has the jidai eroded away. So we know only that Norishige has a couple of dates from around 1312-1319. 

Those are valuable because in combination with old books and dates on Shintogo Kunimitsu works it helps us piece together when the Soshu revolution was taking place.

Yukimitsu in particular has been something of a safe harbor for difficult to make Soshu attributions. The works of these smiths overlap a bit more than people would want. Some Norishige are very close to Masamune, as are some Yukimitsu, and Yukimitsu also overlaps with Sadamune, and Norishige with Go Yoshihiro. 

When a blade is confusing then the fallback attribution has habitually been made to Yukimitsu as a default. As a result any work that comes from Edo with a Yukimitsu attribution needs to be closely inspected to determine if it is better attributed to one of the other Soshu masters. 

I think because of this habit, we have ended up with the statements from old books that Yukimitsu has a wide-ranging style. Where I think it is more true that the wide-ranging attributions given to Yukimitsu are the source of this thought.

This means on any given Yukimitsu there are two types we need to look for: a slam-dunk-certainly-Yukimitsu type of attribution, and then, Yukimitsu-as-conservative-Soshu-attribution. They do mean different things, and sometimes for the second type it can be open to argument as Masamune, Norishige, Go Yoshihiro, Sadamune, Shintogo Kunihiro, or Hiromitsu as potential alternatives. Which way to go depends on the style of the blade. As such this kind of blade can be very interesting to study as the question as to who made it is still open and it is fair to come up with your own theories. The first type is also extremely precious because they show the true level of skill of Yukimitsu and the fact that he came first among the smiths who developed the Soshu den after Shintogo Kunimitsu.

Now what is particularly interesting was the discovery in 2017 of an old document in a library in Japan which lays out the Soshu lineage. This document is dated 1351 and the knowledge would be contemporary. It talks about Masamune being a recently living master smith along with Rai Kunitoshi and Awataguchi Yoshimitsu and others that there have been no debates about. 

But in this document it seems to imply that Yukimitsu became the second generation Shintogo Kunimitsu. So it may be that when Shintogo died, Yukimitsu inherited the school and this agrees with his traditional place as being a bit older and coming before Norishige and Masamune.

Contrast the two following Yukimitsu examples, both of which I owned and passed Tokubetsu Juyo. They are very different, both in the hamon and sugata. The first is earlier work and is attached more closely to Yamashiro works like Awataguchi but the nie is more vivid and thick. The style is conservative but strong, and tapers to almost an ikubi kissaki. 

Yukimitsu katana in most typical style based on Shintogo

The second, the kissaki elongates and the body becomes wide with a narrowing shinogi and this evokes the Nanbokucho Soshu sugata. Dr. Honma was uncertain about the old attribution to Yukimitsu and I believe he held the door open to reattributing it to Sadamune.

Yukimitsu katana in atypical style, possibly Sadamune as alternative

Why second generations were invented

There have been second generations invented for Yukimitsu, Masamune and Norishige however. The desire to do this is twofold, the first being that it is an attempt to explain a blade that looks-like-but-isn’t-quite-there. 

This has been done for various reasons in various smiths. It even happens today when a dealer tries to sell something sketchy, and says, “maybe it’s a second generation” and uses this to explain deficiencies in a signature or in technique or quality. Traditionally this is the primary reason it seems to have been done, or sometimes it’s been used to try to move the “first” generation back in time to increase its value.

Blind men and the elephant

Once this gets done and becomes a habit, someone records it in a book, and later people reference the book as if it’s fact. The point to remember is that old books are people trying to make sense out of foggy information, and a small part of a bigger picture. The result is that you are dealing with the blind men and the elephant metaphor.

In this metaphor, three blind men are brought up to an elephant and asked to describe it. There are various tellings of this metaphor, in one of them, the first blind man says that an elephant is quite like a snake. The second blind man says no, an elephant is like a wall of canvas. The third says you are both wrong, an elephant is like the trunk of a tree.

So of course if we step back and look at what is going on, they are all right, just that they are feeling different parts of the elephant and so are perceiving small parts of a greater truth. The inability to step back and see the big picture, leads them to making wrong conclusions: they are partially right. But their inability to see the whole picture leads to invalid conclusions and disagreement between “experts.”

As such it’s always important to try to take this step back and see the bigger picture. Old books are the blind men and they are making conclusions based on incomplete information. They do however supply important data, and where the truth lies is in trying to unify what they represent, especially where the information they lay out overlaps with other reports. 

Historians use this approach of seeing where contemporaneous accounts confirm each other. It is also important though to winnow out re-reporting of the same information.

For instance, once an author makes the mistake of reporting two generations of Kanemitsu, any authors that follow him in time will simply re-publish the bad information. Then in our time, we can look back and say something really wrong: look at all of these authors reporting two generations of Kanemitsu, it must be true. 

We can reach that conclusion if we don’t do our homework and realize that each one of those authors is simply faithfully re-recording a mistake that came from higher up the chain.

So to help sort it out, we need to look for:

  1. Hard evidence (the actual work coming to light in the modern period)
  2. Older references (the closer to the time of the event, Usually makes reporting more accurate and less distorted)
  3. First-hand references if at all possible, and contemporaneous reporting (two authors of the same time period, not referring to each other’s work)

Those are the tools of historians for sifting out historical events.

Modern scholarship

As well in the modern period, we need to understand that scholarship is evolving. Swords are still being discovered and each discovery advances our knowledge. 

This means even in the case of Juyo and Tokuju blades, as well as in modern books, we need to look at the date of publication and realize that the picture can change between then and now due to new discoveries sharpening the focus of our knowledge.

Also we need to understand that people who wrote books in the Muromachi period, and Edo period, even at the end of the 1800s, they did not have access to swords like we do today. The great daimyo collections were broken down in the 1900s and sold off. Many swords that got taken out of Japan by American soldiers came back as Juyo Token, and it is this act of pulling them out of the dark and bringing them back again in the light, that has opened up a lot of new information to study.

We need to understand these authors of those periods as the blind men trying to describe an elephant. They may have part of the truth and then come to an invalid conclusion.

Fujishiro is right more often than he is wrong, and I think in a lot of ways he was visionary in what he wrote. But he didn’t see everything, and what we can go and see at an NBTHK Juyo exhibition or in a typical museum of today or even at the sword show in San Francisco is not a normal experience for pre-WWII authors.

This is another reason then that we get some conservative judgments that may frustrate some collectors. At this point in time, judges are aware that the situation is dynamic and not static and may require revision in the future.

Dr. Honma wrote into some of his Juyo commentary that the blade was being accepted on the strength of an old attribution and qualified that it needed to be studied more in the future. So he kept the door open for some changes to these attributions and understood that scholarship is a dynamic field. This is not always the case as we do encounter slam-dunk examples, but I bring it up so that people should be aware that the edge cases may need to be updated from time to time. It is more than valid to read one of these signals he gives of uncertainty and take that as a gateway to study and come to your own conclusion as an alternate explanation.