By W. M. Hawley

Japanese Sword Club of Southern California – Talk #37 – Nov. 13, 1970

It is one thing to read the characters on a sword tang and something

else again to say for sure who made it!  As we all know, 90% of all big

name signatures are fakes, as well as a good percentage of lesser


Often they are very superior blades compared to those make by average

smiths, but lacking verification of a signature, their origin may be

doubtful.  Then, of course, there are the out and out fakes and copies

imitating the style but seldom the quality of those of the great masters

So, the problem is to be able to recognize a fake signature.

Signatures cut with a chisel exhibit as may or more characteristics than

those written with a pen or brush.  The variables are as follows:

  1. The chisel.  Each smith had his own favorite chisel for cutting a

signature.  Now-what shape point?  Was it V shaped or U shaped and what

angle V or how wide a U?  A wide angle V point held at a low angle to

the work could cut a shallow groove the same width as a narrow V shape

that was hit harder at a steeper angle, producing a much deeper groove.

The depth of the groove would not show up on an oshigata rubbing, only

the width.  Neither would the roundness of a U shape of the same width.

The difference between a V and a U might be apparent in a photograph but

not the depth of the cutl.  In badly rusted old tangs you would not be

able to see either.  Direct comparison with a genuine blade would be

necessary to prove this point.  So here we have a number of

characteristics that would not be in hand writing ro show up in a

rubbing depth and shape of the cut.

  1. At what angle did the smith bold the chisel and how hard did he tap

it with what weight hammer?  We don’t have to know these three variables

but they would definitely affect the cuts.  How many taps with the

hammer did he use to cut a line ½ long?  If the strokes were heavy this

might show up on an oshigata, but if light it probably would not

magnifying glass on the sword itself would be necessary to reveal this


  1. Most important of all, and easy to see is, how did he form the

strokes?  Length of strokes, shape and angle of dots, curved to straight

lines, shape and angle of the hook on the end of a line, are all just as

individual as in brush or pen writing, as are the width and taper the

strokes and these are just as hard to imitate even if you could know all

the points covered in 1 and 2 above.  Also, these were a matter of

unconscious habit and were not affected by his age or changing chisels.

These all show up in oshigata and are ample to show up all but the

cleverest forgeries.  For this part we do not have to have actual

genuine blades to compare with as pictures of oshigata in the books will

serve.  Suck works as the Juyos, the various Taikans, and the two

volumes of Fujishiro’s Nihon Toko Jiten are available even if you don’t

own them.  The more pictures you can find the better knowledge you will

have of the peculiarities of a man’s signature so that irregularities

should be easy to spot.  Natural variations will show up but the shape

of the strokes will remain pretty much the same.  For the kind of

changes that occur in a man’s signature over a period of years, study

part 2 of the Osaka Shinto Zufu which shows year by year progressions of

a lot of smiths.

Even clever forgers had their own habits and chisels and a wrong book,

curve, or weight of stroke will give them away.  If you have assess to

the current sword magazines from Japan, note the true and false

signatures shown side by side with sometimes very slight discrepancies

pointed out.

I would seem that certain men specialized in forging signatures or

certain smiths as the forger’s own characteristics may show up in

several fakes.  A study of true and false signatures of Kiyomaro and

Naotane bear this out.  In the big work “Minamoto Kiyomaro,” huge

blow-ups of his signatures show an even swelling or tapering of his

horizontal strokes, while all the fakes show bumps at the end of each

stroke.  Fake signatures of Naotane are extremely close to the genuine

except in the kakihan where the top zigzag lines of the genuine are

crowded together but appear much more open on the fakes.

In order to research signatures, you need as many examples as possible

that can reasonable be genuine, Disagreement among experts is mostly

confined to unsigned blades and a few Juyo certifications have been

repudiated.  However, some of the very old books, while considered

reliable in general, are completely useless when it comes to checking

the fine points of strokes.  This is because the oshigata first had to

be copied by brush, then carved in wood blocks, then printed, and

sometimes re-copied, re-carved, and re-printed for later editions.  It

would be impossible to go through all these operations and retain

anything like the photographic quality necessary for comparison of

chisel strokes.  Such works as the Honcho Kajiko, Honcho Gunkiko, Shinto

Meijin, and all the other 16th to 18th century wood block books are

useless for this purpose, even assuming that all examples shown were

genuine, which would be doubtful.

The modern books give accurate reproductions leaving only the sometimes

pertinent question of how expert was the “expert” who said the blade was

genuine?  Here again, signatures would have a better chance than mume


Now, we are back to the problem of source material.  We of the West

cannot compete with the big Japanese appraisal groups who have enormous

libraries and thousands of genuine blades for direct comparison.  But

still, it will help and save the cost of sending swords to Japan for

appraisal if we can spot the more obvious fakes.  If you still want to

send it you can state “signature probably false but who did make it?”

Which will save your face if it is a phony!

Certainly we all should want to study, regardless of whether we own an

almost “National Treasure” or a Sukesada we can’t pin down because he

didn’t add his personal name.  So, the only answer is to acquire as many

books as possible that show pictures of tangs.  Some are certainly

expensive but there are a lot of inexpensive ones also being turned but

in Japan, that are in the $5 to $10 bracket.  Assuming that you have

acquired some of these, there is still the problem of finding the

picture you want without a knowledge of Japanese, or a whole day of

searching.  Indexes are the answer.  An index is being prepared for

“Tanto”, but many more are needed.  For those of you owning some volumes

of the Juyo Token nado Zufu or the two volumes of the Nihon Toko Jiten,

these are indexed in my book “Japanese Sword smiths”.  For myself, I am

indexing one, all of my library that contain pictures, and using a code

letter for each work, note each picture of a tang, opposite the man’s

name in my book.  Anyone who can figure out names and dates well enough

to use my book should be able to do this to almost any modern book.  If

you want to benefit mankind- make your index available to club members

by mimeographing it.

Two good sources are the Koson Oshigata and the Umetada Meikan which

reproduce the scrap books of oshigata gathered by those two men who were

experts in earlier days.  Both need indexing.

Now back to reading and evaluating inscriptions, We have to remember

that there are genuine swords to which the name was added later by

someone who recognized the work and added the name with or without

trying to imitate the signature, in order to make the blade easier to

sell.  Properly, such attributions should have been done in gold by a

recognized appraiser and signed with his name and/or kakihan, but many

tried to fake the signature.  Certification of these has to ignore the

fake signature.  Certification of these has to ignore the fake


Signatures added at a much later date often exhibit a different color of

rust in the chisel marks.  Another thing to watch is a hole through a

character of a signature.  Generally, new holes were added when a blade

was shortened, but a blade with signature and only one hole right

through a character is obvious nonsense!  Or even several holes if the

lowest one pierces a character.  When blades were shortened, the new

holes (always drilled not punched) had to be higher not lower.  When a

hole pierces a character, use a glass to see if burrs from the chisel

were pushed into the sides of the hole – this is a dead give-away.

For the most part, fake signatures were intended to up-grade the value

of a blade, so were intended to indicate the most famous men if several

generations existed.  However, later generations sometimes thought they

were good enough to pass off a blade as that of a famous predecessor and

cut an inscription that was only recorded to him.  Generally the blade

itself will give this away and a check of signature characteristics will

confirm it.  So the wording of an inscription does not always pin it

down to the only one listed in the books.  Always check the other


A signature badly incrusted with rust may require some cleaning, but

don’t disturb it if it is obviously very, very old.  First lay the tang

on a hard wood block.  Lay a piece of thin leather or thick cloth over

the inscription and tap lightly with a small hammer.  Not hard enough to

distort the metal but enough to break up the layers of rust.  A chisel

made of bone or bamboo may help to lift of the layers of rust.  New red

rust often indicates a deliberate attempt to make a tang look old.  It

won’t stick very tight and usually comes off with a wire brush or coarse

steel wool.  After cleaning, oil or wax the tank to prevent further


One final word – signature is a lot easier to fake than the blade

itself, so a thorough study of the blade should precede the research on

the signature.  Then, if the school and probable date are comparable

with the inscription, it is time to go to work on the signature.

Of course study is the key to understanding swords in general and the

same applies to the peculiarities of signatures.  Start with the swords

you own and check our every stroke of every character in the whole

inscription.  Then read up on the man, making notes on everything

pertaining to him, then go on either generations of the line, then to

pupils, etc. with special attention to the outstanding points of

difference in the signatures.  Sometimes one line or dot will be enough

to identify a generation.

Return to Articles