Hiroki Kikuchi, "Letting the Copy out of the Window: a history of Copying Texts in Japan", The East Asian Library Journal 4, no. (2010): 120-157

copying texts in japan    127

(1643–1724),  who  was  the  lord  (daimyō )  of  the  Kaga  domain,  showed  strong 

interest in all kinds of old documents and writings, and his collection project 

extended to the Sanjōnishi Bunko. Tsunanori supported Kinfuku financially and 

his daughter married Kinfuku; at the same time Tsunanori started researching 

the Sanjōnishi Bunko, made title lists, and copied several texts by hand. Later, in 

exchange for the access he was granted, Tsunanori offered to repair deteriorating 

rare books. In the process of Tsunanori’s research, one of his largest discoveries 

in the Sanjōnishi Bunko was Sanemikyō-ki manuscripts, numbering about sev-

University of Tokyo. Photograph from the collection of the Historiographical Institute. 

Compare with a traced copy of the same text shown in figure 6 below.

---

128    hiroki kikuchi

enty scrolls. At first members of the Sanjōnishi family could not even determine 

whose diary record these manuscripts were. Tsunanori borrowed them, identified 

them as Sanemikyō-ki, compiled a list of these manuscript scrolls, and repaired 

damaged scrolls. The list he made attracted many aristocrats to the texts because 

only a very small part of the copy of Sanemikyō-ki had been known previously. 

As an adult Kinfuku became interested in his ancestor’s diary record and copied 

a part of Sanemikyō-ki, which was gradually recopied and spread among other 

aristocractic houses. (See figure 5.)

 

About one hundred years later, in the early nineteenth century, a remarkable 

project to copy the Sanemikyō-ki was begun in the Tebori-Sanjō family. This family 

was the main branch of the Kan’in clan, under which the Sanjōnishi family also 

fell. Although the Tebori-Sanjō already possessed a recopied version of Kinfuku’s 

Sanemikyō-ki copy, the set was still incomplete. Tebori-Sanjō Saneoki (1756–1823) 

borrowed the remaining sixteen scrolls of Sanemikyō-ki manuscripts directly from 

the Sanjōnishi. Under Saneoki’s management, his son Kimiosa (1774–1840) and 

grandson Sanetsumu (1802–1859) were engaged in the copy work. Hino Suke-

naru (1780–1846) also cooperated with the Tebori-Sanjō’s copying project. Later 

Sukenaru introduced Kuze Michiaya (1782–1850) into the work as well. Thus, 

the copy in the Tebori-Sanjō was carried out as a group project. (See figure 6.)

 

Furthermore, at this time, the Tebori-Sanjō copied not only the text, but 

also the whole style of manuscripts, which included the exact shape of each letter 

with its calligraphic character and even drew the shape of the worm-eaten holes 

found in the paper of the original. Such a copy style is called eisha (traced copy). 

Since Sanjō Sanemi was not known as an excellent calligrapher, the exact copy 

would be of no use as a calligraphic sample. If the main goal were to research 

court ritual, only the text of Sanemikyō-ki without the calligraphic imitation would 

have been sufficient. Why, then, did Saneoki make an eisha copy of Sanemikyō-

ki. (Compare figures 4 and 6.) Here we can confirm that, in the early modern 

period, the main point of research for diary records had drifted into philological 

(shoshi-gaku) issues in the aristocracy. The word “philology” or “philological” might 

not seem to be appropriate for use in this essay since in Western scholarship this 

concept has come to imply the study of classical texts and translation. However, 

in this essay these words will be used as the translation of shoshi-gaku, which has 

been developed as komonjo-gaku (diploma study or paleography) or shiryō-gaku 

(historiography), all of which imply careful consideration of the material aspects 

of manuscripts and their transmission or function. In Japanese scholarship this 

---

copying texts in japan    129

field is not only appreciated as the basis for the writing of history, but it is also 

anticipated that it will develop into an independent field of study.

 

Before  the  Tebori-Sanjō  copying  project  started,  another  copy  of 

Sanemikyō-ki  had  been  completed  by  Ōgimachi-Sanjō  Kin’nori  (1774–1880). 

Though this copy has not been found, it is supposed that it was produced as an 

eisha copy as well, because some parts of Tebori-Sanjōs’ second version, which 

was copied from Kin’nori’s copy, have kept the style of the traced copy. Aristo-

crats who were interested in diary records very much appreciated Tebori-Sanjō’s 

and  Ōgimachi-Sanjō’s  eisha-style  versions.  For  example, Takatsukasa  Masahiro 

(1761–1849) borrowed these eisha versions soon after their completion.

17

 

The  common  admonition  among  the  aristocracy  was, “don’t  let  [the 

copy]  out  of  the  window”  (Sōgai  ni  idasu  bekarazu),  which  means  that  a  text 

should be kept within a collection, out of sight, and thus maintained for use in 

a certain house exclusively. Nevertheless, despite the prohibition against giving 

a copy to other families and lineages, once a set of copies was produced from an 

original, it was in turn recopied by many other houses in order to build their 

own libraries for research on ritual and for other purposes. For example, Kajūji 

Tsuneitsu (1748–1805) recopied the Ōgimachi-Sanjō version because Tsuneitsu 

was Kin’nori’s father-in-law. In spite of the fact that Kin’nori asked Tsuneitsu 

never to show the copy to others, Tsuneitsu secretly showed this copy to the 

Takatsukasa family, as mentioned above. Later, the Tebori-Sanjō also borrowed 

the Ōgimachi-Sanjō copy and made another version. It is not difficult to assume 

that Hino Sukenaru, who cooperated with the Tebori-Sanjō’s copying project, 

mediated between the Ōgimachi-Sanjō and the Tebori-Sanjō because Sukenaru’s 

wife and Kin’nori’s wife were sisters and both were Tsuneitsu’s daughters. Thus, 

Sukenaru not only gave advantage to the Tebori-Sanjō, but also benefitted from 

association with that clan. Sukenaru is thought to have introduced the Tebori-

Sanjō version to the Yanagiwara family because Sukenaru’s mother came from 

the Yanagiwara.

18

 Thanks to Sukenaru’s cooperation, the Yanagiwara, by gathering 

material from the various versions, was able to complete one of the best copies 

of Sanemikyō-ki.

 

Thus while interest in texts such as Sanemikyō-ki increased more and more 

over time, diary records were not published in printed form in the premodern era.

19

 

Though many were produced in the Edo period, they were always copied by hand 

through connections to relatives and other relationships in accord with the con-

straints of the traditional precedent of “not letting the copy out of the window.”

20

---