Voortgang. Jaargang 22

[p. 161]

The great Dutch-Japanese dictionaries in early nineteenth century Japan
Henk de Groot

Abstract - A strict policy of isolation from the outside world restricted Japan's trade with the western world to a tiny Dutch trading post in Nagasaki from the seventeenth century until the middle of the nineteenth century. As a result, Dutch was the only language through which the Japanese were able to learn about western science and technology. Although specialised Japanese-Dutch word lists were compiled, largely for purposes of trade, comprehensive dictionaries did not appear until about 1800, when François Halma's Dutch-French dictionary served as the basis for two major works: the Edo haruma and Nagasaki haruma dictionaries. These two works and the publications they inspired provided the Japanese with tools that greatly improved the speed and accuracy with which translations of western works could be produced.

Introduction

From the middle of the seventeenth century until 1853, the shogunate government of Japan maintained a policy of national seclusion, allowing only one small trading post each for China and the Netherlands in Nagasaki harbour to maintain trade relations with the outside world. The motive for this nearcomplete isolation was to resist religious and colonial pressure from European countries. In order to control the contact between the Dutch and the Japanese, members of the Dutch trade mission were confined to a tiny peninsula in Nagasaki harbour, called Dejima. They were not allowed to learn Japanese, and the Japanese government set up an official Guild of Interpreters. Over the years, a number of Dutch-Japanese word lists and glossaries evolved, which were designed to facilitate the Dutch-Japanese trade. These usually contained only words that were useful in such contexts, such as the names of metals and textiles, or phrases used in polite conversation.

However, two centuries of strictly enforced isolation did not prevent knowledge of western science and technology trickling into Japan via the tiny Dutch trading post, particularly during the first half of the nineteenth century, when official resistance to western influence was reversed as the realisation grew that an understanding of western technology would be essential to maintain Japan's independence. Since this information arrived in Japan mostly through imported Dutch books, for many the study of the Dutch language became the first step on the path to becoming a rangaku¹ scholar.

Towards the end of the eighteenth century, the increasing demand for accurate Japanese translations of western books gave rise to a need for comprehensive works of reference that could provide reliable equivalents of Dutch and

1 Rangaku, lit. ‘Dutch learning’ became the term for the study of European science and technology during this period.

[p. 162]

Japanese vocabulary. This study outlines the motives and processes involved in the creation of several large Dutch and Japanese dictionaries during the final decades of Japan's period of isolation.

A First Attempt

The first compilation of a true Dutch-Japanese dictionary based on an existing European work took place in the second half of the eighteenth century, well over a hundred years after the establishment of a dedicated Interpreters' Guild. This is not entirely surprising, since a ban on the importation of European books was in effect until 1720, and the first recorded importation of books by the Dutch did not take place until a shipment arrived from Java in 1754.² In that year the Interpreters' Guild was presented with two copies of Halma's Dutch-French dictionary, one of Marin's, and a Latin-Dutch lexicon (MacLean 1974:14).

Exactly how the interpreters used these dictionaries is not known,³ but apparently no attempt at using these to compile a Dutch-Japanese dictionary occurred until the late 1760s, when interpreter Nishi Zenzaburõ (?-1768) began compiling such a work, with the help of Pieter Marin's Dutch-French dictionary Groot Nederduitsch en Fransch Woordenboek (1717). He did so alone and in secret, pretending that he was too ill to turn up for work. Unfortunately, he died before he could complete even the letter D, and no trace remains of the work (Katagiri 1985:491).

2 The Portuguese produced a number of printed books on the Japanese language in the early seventeenth century, and a Japanese-Portuguese dictionary appeared in 1603. However, these works were produced by and for Portuguese missionaries, and played no role in language acquisition or translation by Japanese after their expulsion from Japan.

3 MacLean suggests that a number of biology books may still have been kept by the interpreters, which would explain their interest in a Latin dictionary (MacLean, 1974:14).

The Edo haruma

The next attempt at the creation of a dictionary proper occurred in Edo (now Tokyo). There appears to be a connection between Nishi's pioneering attempt and this project. In 1792, after reading Rangaku kaitei, a work which outlined the development of rangaku studies and included an introduction to the Dutch language, published in 1788 by prominent Edo scholar Õtsuki Gentaku (1757-1827), a physician from Inaba province (in the eastern part of Tottori Prefecture) called Inamura Sanpaku (1759-1811) obtained permission to take three years of leave to go and study rangaku at Õtsuki's Shirandõ rangaku academy in Edo. Inamura soon realised that he would need more than three years to learn the Dutch language, and suggested that Õtsuki compile a dictionary that Inamura would be able to use for further study after his return home. Õtsuki felt that he was not qualified enough, and was anyway too busy to undertake such a project. However, Inamura's idea reminded him of a former interpreter called Ishii Shõsuke (1743-?), who had accompanied him in 1786 on his way home from a study trip to Nagasaki. Ishii had told Õtsuki of Nishi Zenzaburõ's un-

[p. 163]

completed project, and had expressed a wish to continue with Nishi's work. Õtsuki introduced Ishii to Inamura Sanpaku, and the two borrowed Gentaku's copy of François Halma's Woordenboek der Nederduitsche en Fransche Talen⁴ and began the long task of translating it (Numata J. et. al. 1976:384). Before long, Ishii was called back for duties by his master in Shirakawa, in Fukushima Prefecture, but he obtained permission to take Õtsuki's Halma dictionary with him, and returned to Edo with the draft of a dictionary in the following year. Exactly to what extent Nishi Zenzaburõ's initial efforts are connected with this work is unclear. In his preface to the Edo haruma, as the dictionary was eventually called,⁵ Õtsuki Gentaku claims to have seen Nishi's work, and tells us how Ishii had a wish to ‘continue with Nishi's work’. This appears to imply that Ishii simply continued where Nishi had left off. However, Nishi Zenzaburõ had based his translation on Pieter Marin's dictionary, whereas the Edo haruma is based on that of Francois Halma.⁶ It would seem, therefore, that Ishii started anew with his own translation, merely taking up the idea, and perhaps employing some of Nishi's approaches and methods. Since it is not known what became of Nishi's unfinished manuscript, we do not know to what extent, if at all, Ishii relied on Nishi's translations for his entries of the first few letters of the alphabet.⁷

Once Ishii had completed his draft, Inamura Sanpaku, Udagawa Genshin (1769-1834) and others proceeded with corrections and additions. Ishii's draft had contained some twenty to thirty thousand entries, but by 1796 Inamura and his team, often working through the night until dawn, had extended this to about eighty thousand. They then used wood-carved movable type letters to print thirty copies of the dictionary containing the Dutch entries only, and wrote in the Japanese translations by hand (Saitõ 1985:58). It must have been an enormously time-consuming task, but as a result of this students and translators of Dutch texts in Japan on the threshold of the nineteenth century finally had a proper dictionary to work with.⁸

4 Kerlen identifies this work as Halma's first edition of 1717 (Kerlen 1996:86).

5 Haruma is a Japanese rendition of the name Halma. The Edo haruma is also known as Haruma wage (‘Halma translated into Japanese’).

6 The question has been asked why French-Dutch dictionaries were imported into Japan, and not a descriptive Dutch dictionary, which surely would have been more useful. The answer is that such a dictionary did not exist. The original purpose of dictionaries was to provide equivalents of words in different languages. The first proper Dutch language dictionary was the Nederduitsch taalkundig woordenboek by Pieter Weiland, which was published progressively in eleven volumes between 1799 and 1811.

7 A translation of Õtsuki's preface to the Edo haruma can be found in Goodman 2000:140.

8 In the original Edo haruma, a number of linguistic terms were listed without translation (notably woordbuiging, ‘declension’) indicating that at the end of the eighteenth century grammatical expertise was still in its infancy among the rangaku scholars in Edo.

Hendrik Doeff and the Nagasaki haruma

Yet the Edo haruma, although it was the first of the great Dutch-Japanese dictionaries, did not quite gain the widespread acceptance that was accorded the Duufu haruma (which is sometimes also referred to as the Nagasaki haruma, to contrast it with the Edo haruma) dictionary.⁹ Hendrik Doeff (1777-1835) first arrived in Nagasaki in 1799 in the capacity of chief bookkeeper to the Dutch trade mission, and became opperhoofd (chief of the trade mission) in 1803. As a result of events in Europe the Dutch were unable to dispatch ships to Japan for an extended period, and Doeff found himself stranded on Dejima without a successor until the arrival of Jan Cock Blomhoff in 1817.¹⁰ The lack of trading

9 Duufu is a Japanese rendition of the family name of Hendrik Doeff, the main compiler of the dictionary.

10 The VOC, and later the Dutch government, were responsible for the living expenses of the Dutch traders in Nagasaki. Doeff survived his lengthy period of isolation through a combination of frugality and loans from the Japanese (Doeff 1833:70 and 177).

[p. 164]

activities left him with a great deal of spare time, and he eventually embarked on the project of compiling a complete Dutch-Japanese dictionary.

In the postscript to his book Herinneringen uit Japan (‘Memories from Japan’) he describes his motivation and approach:

The experience that the Japanese interpreters speak the Dutch language very poorly, and that many words in translation are assigned an incorrect meaning, made me adopt the notion of compiling a Dictionary of both languages. I was of the opinion at the time that I had progressed sufficiently in the Japanese language to take on this labour, for which no precedent existed as yet, with good results. There were a number of small booklets to assist the Dutch functionaries and interpreters, but these were small and deficient. As my guide I therefore took the Nederduitsch en Fransch Woordenboek by Francois Halma.¹¹

Doeff states that he started on his dictionary twelve years after his arrival in Japan, that is, in 1811.¹² When he left, six years later, he had completed a first draft of the work. Initially he worked on it largely by himself, but in the year before his departure he was informed that the Shogun wished to obtain a copy of the work. In his introduction to the dictionary, he relates:

For the preparation of a copy for His Imperial Majesty,¹³ His Excellency the Governor of Nagasaki has appointed eleven interpreters. However, if the work were to be sent off in its present state, errors would certainly be found, which in a work of this size is unavoidable. Therefore, being of the opinion that the work needs correction, I employed as first correctors junior interpreter Nakayama Tokujuurõ and junior interpreter Yoshio Gonnosuke, the latter of whom, as the most proficient in the Dutch language among the members of the Interpreters' Guild, is indispensable to this work.¹⁴

In addition to the eleven interpreters, the Dejima ward headman was also appointed to the project, because, Doeff writes:

...he had a very good understanding of the Japanese language (something which even with the interpreters was not always the case).¹⁵

Although in the introduction to his dictionary Doeff makes reference to ‘several [other] Dutch-Japanese dictionaries’, the comment in his 1833 memoirs regarding ‘small and deficient booklets to assist the Dutch functionaries and interpreters’, as well as his observation that ‘no precedent existed as yet’ for such a project, quoted above, indicate that this did not refer to the Edo haruma, and suggests that he never knew about the existence of this rival work. This discounts Boxer's theory that Doeff's dictionary was based on the Edo haruma (Boxer 1950:66).

11 ‘De ondervinding, dat de Hollandsche taal door de Japansche tolken zeer slecht gesproken, en dat aan vele woorden in de vertaling eene geheel verkeerde beteekenis gegeven wordt, deed mij het eerste denkbeeld ter vervaardiging van een Woordenboek der beide talen opvatten. Ik meende toen ver genoeg in de Japansche taal gevorderd te zijn, om dezen arbeid met een goed gevolg op mij te nemen, waarvan nog geen voorbeeld bestond. Men had er wel eenige boekjes tot behulp voor de Hollandsche ambtenaren en tolken, met de noodzakelijkste beteekenissen, doch dezelve waren klein en gebrekkig. Tot handleiding nam ik dus het Nederduitsch en Fransch Woordenboek van François Halma.’ (Doeff 1833:263-4). In his introduction to the Duufu haruma, Doeff identifies Halma's second edition (1729) as his source.

12 Doeff first visited Japan briefly in 1799, and did not begin his posting on Dejima until 1800. This has given rise to speculation that ‘twelve years after my arrival indicate’ is the year 1812. However, in the foreword to his dictionary, which is dated 1816, Doeff states that he had begun his project ‘5 or 6 years earlier.’ The two statements combined unambiguously identify 1811 as the year when Doeff began compiling his dictionary.

13 Although the Shogun was not of imperial blood, he was seen by the Dutch traders as being the (secular) Emperor of Japan, hence Doeff's use of the term ‘His Imperial Majesty’ here.

14 ‘Om dit werk voor zijne Keijzerlijke majesteit af te schrijven heeft den weleedele Achtb. heer nagazakijsche gouverneur elf tolken benoemd, dog indien men hetzelve zodanig als het was zond zouden daar wel fouten in gevonden worden, hetwelk in zoo een omlastig werk onvermijdelijk is, en ik dus noodig geoordeeld hebbende om dit werk te verbeteren heb ik den temporeel ondertolk Nj: tokzuro, en den onder het geheele tolkencollegie in de hollandsche taal allerbekwaamste en tot dit werk onontbeerlijke vice ondertolk Yos: gonnoskij, tot eerste verbeeteraars [...] geemployeerd.’ The Dutch used by Doeff in this passage may strike the reader as somewhat unusual. Doeff wrote this introduction for the benefit of Japanese readers, and explains in his memoirs that, in order to satisfy the interpreters and avoid critical comment, he would sometimes ‘adapt’ his Dutch to make it easier for the Japanese to understand (Doeff 1883:101).

15 ‘... die de Japansche taal zeer goed verstond (iets, dat zelfs bij de tolken niet altijd het geval is)’ (Doeff 1833:264).

[p. 165]

Unlike the Edo haruma, the Dutch part of which was printed, Doeff's dictionary was entirely handwritten. At the time of Doeff's departure, there were officially two copies in existence: the work itself, deposited with the Interpreters for further development, and its first draft,¹⁶ which Doeff publicly donated to the son of a senior interpreter as a farewell present, in order to allay suspicions that he might attempt to smuggle a copy of his work out of the country. That, however, is exactly what he did: as the work progressed, Doeff secretly produced a second draft, which he hid among his luggage. Tragically, on his way back to The Netherlands his ship sank in flames, and he lost most of his possessions, including his dictionary.¹⁷

After Doeff's departure for The Netherlands in 1817 the interpreters continued with the development of the work, and completed the work in 1818 or 1819, possibly with the assistance of Blomhoff. At the behest of the Japanese authorities, the interpreters subsequently embarked on further corrections and additions, and finally completed the work in 1833 (Koga 1966:85). Copies were then presented to the Shogunate, the governor of Nagasaki and the governmental bureau for translation of western books, the Bansho shirabesho, in Edo. None of these original manuscripts of the work appears to have survived, but the work is said to have contained over a hundred thousand entries.

The Duufu haruma was copied by hand several times, and was eventually published (see below). In his autobiography, Fukuzawa Yukichi (1834-1901), who was later to become one of Japan's greatest educationalists, describes how he and his fellow students at the famous Tekijuku school in Osaka waited in the ‘Doeff room’ for their turn to be allowed access to the work (Kiyooka 1966:82),¹⁸ and relates how some feudal lords would pay students of Dutch sixteen mon per page to write copies of Doeff's dictionary (Kiyooka 1966:83). In his introduction to the dictionary, Doeff explains that he employed what he termed ‘vulgar Nagasaki style’¹⁹ for his Japanese entries so that all the interpreters, both old and young, would be able to understand it. Unwittingly he thus provided a major impetus to a growing trend in Japan towards unification of the written and the spoken language (Sugimoto in Blussé et. al. 125). Numata describes the work as ‘both the largest and the best Dutch-Japanese dictionary available [at the time]’ (Numata 1989:115). It remained the standard such dictionary until well into the twentieth century.

Thus, although the Edo haruma had appeared over thirty years earlier and was initially produced in larger numbers, it was Doeff's dictionary which became the staple reference work in most language schools. The reasons for this are likely to be the larger number of entries, the fact that both Dutch and Japanese native speakers were involved in the project, the inclusion of example phrases and its use of colloquial Japanese.

16 ‘Mijn oorspronkelijk opstel of klad...’ (Doeff 1833:265).

17 Initially, following his return to The Netherlands, Doeff made only passing mention of his dictionary in his book Herinneringen uit Japan. It was not until controversy arose over the authorship of his dictionary (see below) that he added a somewhat heated postscript to his book, in which he describes the reasons and circumstances of the compilation of the dictionary, offers proof of his authorship, and demands credit for his labours. The implication of this is that, although the compilation of such a work must have been a mammoth task indeed, Doeff did not at the time think that he and his assistants were doing anything that would be of much interest to anyone in The Netherlands, certainly not as momentous as the many political and commercial events and transactions which form the main topic of his book.

18 The Tekijuku School building in Osaka is now a museum, and the Duufu haruma is on permanent display there.

19 ‘... in de platte Nagazakij stijl...’

[p. 166]

The haruma dictionaries in The Netherlands

Some controversy surrounded Doeff's dictionary in the Netherlands. J.F. van Overmeer Fisscher (1800-1848), a clerk and subsequently warehouse master on Dejima from 1820 to 1829, chanced across the dictionary in 1823, and spent the next six years quietly copying the entire work (Van Overmeer Fisscher 1833:93). In 1825 Von Siebold praised Van Overmeer Fisscher for his singlehanded work on the dictionary in a report to the governor-general of the Dutch East Indies. When Van Overmeer Fisscher, upon his return to the Netherlands, presented his dictionary to King William I without mentioning that he had copied it from Doeff's, Doeff was sufficiently incensed to contact the incumbent opperhoofd on Dejima and request a copy of the introduction to his dictionary, signed by fourteen interpreters and sealed, as proof that the work was his. This document, he stated in the epilogue to his book Herinneringen uit Japan, was ‘available for inspection’ at his home (Doeff 1833:267).

Hendrik Doeff died in 1835, two years after the interpreters in Nagasaki completed his dictionary, and therefore never knew the extent of the acclaim that his work subsequently received among the students of Dutch in Japan. The postscript of his book conveys the impression of a frustrated man, angry at the attempts by others to claim credit for his dictionary and bitter at the fame and recognition Von Siebold, who was twenty years his junior, received on his return after a mere six years in Japan, while the tangible results of his own achievements lay in a shipwreck at the bottom of the Indian Ocean. Had he known that as a result of his dictionary his name was later to become a household word in rangaku circles, he might have died a happier man.

There are two versions of the Edo haruma in the library of Leiden University, a partially printed copy and a hand-written one.²⁰ Although they are both identified as such on their respective title pages, confusion persists as to the true identity of these works. This seems to stem from a description provided by Von Siebold in his Latin treatise on Japan, Isagoge, which shows that he was under the impression that the Edo haruma was Doeff's work (Von Siebold 1841:21). Serrurier, in his catalogue of Japanese books in the Leiden University library, copied the mistake (Serrurier 1896:20), and it has from time to time been repeated until the present day.²¹

20 Kerlen mentions only one, the hand-written copy (Kerlen 1996:86).

21 A major Dutch television documentary series marking four hundred years of Dutch-Japanese relations in the year 2000 is a recent example. The same error probably led Sugimoto (1991:88) to believe that a romanized dictionary in the Leiden collection is a manuscript from Doeff's hand (it is in fact Blomhoff's-see below), and applied this as proof that Doeff used romanized Japanese in his original manuscript, later to be transcribed into Chinese characters and kana by the Nagasaki interpreters. It is not known whether or not Doeff used rõmaji in his first draft.

Related publications

Since only thirty copies of the Edo haruma were produced, while no more than a few manuscript copies of the Duufu haruma existed, they were inaccessible to most people. In response to this problem, both works were used as the basis

[p. 167]

for a number of later published dictionaries. Inamura Sanpaku, who had got himself into financial difficulties through his efforts to have his dictionary printed (Hesselink 1995:216), eventually left Edo, and moved to Kyoto in 1806, where he became a major force in the development of rangaku in the Osaka and Kyoto area. One of his students, Fujibayashi Fuzan (1781-1836), published a concise version of the Edo haruma, called Yakken (‘A Key to Translation’) in 1810. Having witnessed many attempts by new students to hand-copy the Edo haruma, only to see them give up because of the enormous amount of time it took to produce a complete copy, Fuzan and a friend, a medical student called Komori Genryõ (1782-1843), assembled some thirty thousand entries from the Edo haruma into a new compilation. When they found that this new work was quite useful, they obtained permission from Inamura Sanpaku to publish their work.²² Initially a hundred copies were printed, and the work was reprinted in 1824. In addition to selecting only a portion of the complete Edo haruma entries, Fuzan also added a considerable number of corrections to the translations and some new entries. The work contains an appendix for medical words, and a companion volume in the form of a brief treatise on the Dutch language and script and some symbols, entitled Rangaku kei (‘A Path to Dutch Learning’).

A revised version of Yakken appeared under the title Kaisei zõho yakken (‘Corrected and Enlarged Yakken’) in 1860.²³ Its editor was Hirota Kenkan (1818-1888), who had earlier collaborated with Sakuma Shõzan in the 1840s on the revision for publication of the Duufu haruma, which was eventually published under the title Oranda jii (see below). Hirota in fact based many of his improvements and some nine thousand new entries on this latter work, as a result of which Kaisei zõho yakken is a hybrid work that can justly and uniquely claim to be based on both the great ‘rival’ dictionaries Duufu haruma and Edo haruma. Unfortunately, by the time of its arrival the great switch to German, French and English had begun in earnest, and Kaisei zõho yakken faded into obscurity, its qualities unrecognised.

The Duufu haruma itself became the basis for two further dictionaries, Oranda jii (‘Dutch vocabulary’) and Rango tsuu (‘Understanding the Dutch Language’). Oranda jii had originally been revised for publication in the 1840s by Sakuma Shõzan (1811-1864), a samurai intellectual whose concerns about foreign invasion had brought him to the study of Dutch gunnery and military strategy, particularly coastal defence. Ironically, he was unable to get official permission for publication, because the judiciary in Edo at the time was dominated by anti-foreign elements, and publication and study of Dutch materials by the public at large was discouraged (Sansom 1950:258). It was Katsuragawa Hoshuu (1827-1880), a member of an influential family of rangaku scholars, who finally succeeded in obtaining official permission to publish (Numata et. al. 1984:127).

22 Fujibayashi Fuzan in his foreword to Yakken, quoted in Numata et. al. 1984:713.

23 Although dated 1857, the work is not thought to have been completed until 1860 (Saitõ 1985:175).

[p. 168]

The work appeared in four volumes between the years 1855-1857, during a wave of publications on the Dutch language that followed the arrival of Commander Perry's American warships off the coast of Japan in 1853. Although the foreword states that the work is based on Doeff's dictionary, Sakuma explains that Marin's dictionary was also referred to for revision purposes.²⁴ The presence of different writing styles and copying errors indicates that a number of people, not all of them experts, were involved in the preparation of the manuscript. For example, in some places French elements have inadvertently been included among the Dutch entries.

At least four copies of Oranda jii found their way to The Netherlands. Of the three in the Leiden University collection, two look unused, but the third shows a certain amount of wear and tear, and has handwritten notes added. These notes generally concern themselves with the readings of Chinese characters, suggesting that they have been added by, or for the benefit of, students of Japanese in the Netherlands. The works were no doubt bought with this purpose in mind. A small red stamp on one of the wrappers reveals that the purchase price was twelve tael.

In 1857 a second publication based on the Duufu haruma appeared. This work, which appeared under the title Rango tsuu, was the result of the efforts of rangaku scholar Maki Bokuchuu (1809-1863), who had been a student of prominent Edo rangaku scholar Mitsukuri Genpo (1799-1863). In the introduction, the compiler states that he has attempted to produce a corrected edition of the Duufu haruma. Only three parts are extant, containing the entries for A-D, E-K and L-O respectively. It is not known whether a volume of entries for the remaining letters of the alphabet was ever produced. It has been suggested that Rango tsuu may have been combined with Sakuma's work to form Katsuragawa's publication of Oranda jii (Numata et. al. 1984:753). However, the two works are quite different in format. Furthermore, the first volume of Oranda jii appeared in 1855, two years before the publication of Rango tsuu.

In his introduction, Maki relates that he obtained the Duufu haruma ‘from the hands of someone from Nagasaki’. According to that (unidentified) person, this work was ‘a manuscript [of the dictionary] for Doeff to take back to The Netherlands’. Apparently the fact that Doeff had smuggled a copy of his dictionary out of the country was by then no longer a secret, and indeed may have been known by the Japanese all along. That Maki had access to this manuscript is not altogether surprising, since his mentor, Mitsukuri Genpo, was attached to the governmental bureau for translation of western books, the Bansho shirabesho, where one of the Duufu haruma manuscripts had been deposited after it was handed over to the authorities in 1833.

24 It is ironic that Halma and Marin, whose names are so closely interwoven with Dutch language studies in Japan, had an intense and public dislike for each other. Halma in particular went to extremes in his campaign to discredit his rival; in the second edition of his dictionary (1729) he referred to Marin as ‘die verwaande taalbrabbelaar’ (‘that conceited producer of gibberish’) and, employing a pun based on a French interpretation of Marin's name, proclaimed him to be ‘un nouveau Monstre Marin’ (‘a new Seamonster’). He was clearly very pleased with this trouvaille, because he went so far as to commission an artist to visualise his insult in the frontispiece, a magnificent etching depicting a victorious knight in armour (presumably Halma himself) standing triumphantly atop a defeated sea monster, while the surrounding onlookers hold their noses to ward off the stench that is still emanating from the monster's mouth.

[p. 169]

In 1822, a part-translation of L. Meijer's Woordenschat, bevattende in drie delen, de verklaring der basterdwoorden, kunstwoorden en verouderde woorden²⁵ (12th ed. Dordrecht 1805) was published in Edo under the twin titles Nieuwe-Gedruct Bastaardt Woorden-Boek²⁶ and Basutaado jisho. The translation was based on the first volume of Meijer's dictionary, which contained loan words. Published by the Nakatsu feudal lord Okudaira Masataka (1781-1855) (who had sponsored the publication of a Japanese-Dutch glossary called Rango yakusen some twelve years earlier), this work was compiled and translated by Masataka's personal physician Õe Shuntõ (1787-1844), who had studied under pioneer rangaku scholar Maeno Ryõtaku and had spent some six or seven years studying in Nagasaki. The preface in Dutch from the hand of Masataka himself states that the work was subsequently checked and corrected by Baba Sajuurõ (1787-1822), a talented Nagasaki interpreter who in 1810 had been summoned by the government to Edo to work for the translation bureau. The Dutch in the preface itself, however, is poor even by Japanese standards, and was clearly not checked by anyone with any Dutch language expertise. The work is in two volumes and contains well over seven thousand entries. Although it has no connection with the haruma dictionaries, its alphabetical arrangement places it firmly within the family of proper dictionaries. Furthermore, it was based on a much more recent Dutch work than the haruma dictionaries had been, and there can be no doubt that it played a useful role alongside the great dictionaries in the final decades of the Edo period.

25 ‘A Vocabulary in Three Volumes, Containing the Explanations of Loan Words, Technical Terms and Archaic Words’.

26 ‘New Printed Dictionary of Loan Words’.

Further Dejima initiatives

Three other Europeans compiled dictionaries during their stay in Japan during the latter years of the Edo period: Blomhoff, Von Siebold and Van den Broek. However, there is an important difference between their work and that of their predecessors. Their motive was not to provide the interpreters with an effective tool to help them improve their performance; they compiled their dictionaries either for themselves or for students of Japanese in the Netherlands. Jan Cock Blomhoff (1779-1853), who succeeded Doeff as opperhoofd to the trade mission in Nagasaki in 1817, used the same Dutch-French dictionary by Halma that Doeff had used, to compile a dictionary with the Japanese entries in romanized script.²⁷ Blomhoff enlisted the assistance of two young interpreters, Araki Toyokichi²⁸ and Kikutani Yonezõ for the Japanese translation.

In connection with this work too, a misunderstanding has persisted since the days of Von Siebold, who reported that a dictionary by Weiland had provided the basis for Blomhoff's work (Von Siebold 1841:22). Although this notion has been adopted by bibliographers and historians up until the present day, on inspection it is clear that Blomhoff's source was certainly not Weiland, but either the second or the third edition (they are very similar) of Halma's Dutch-

27 This is the work Sugimoto thought was Doeff's. See note 21 above.

28 Toyokichi is perhaps better known as the artist of a portrait of Hendrik Doeff, which he made in 1810 or 1811 (Boxer 1951:108-9).

[p. 170]

French dictionary. It seems likely that Blomhoff used the same copy that Doeff had used, since it would have been in the permanent collection of the trade mission chief. Occasional similarities in the Japanese entries indicate that the Duufu haruma may also have been consulted, but Toyokichi and Yonezõ clearly had their own ideas regarding many of the Japanese translations.

Blomhoff was on Dejima from 1809 to 1813 as warehouse master and from 1817 to 1823 as opperhoofd. His dictionary was probably compiled during his second posting. He would have had considerably less spare time on his hands than Doeff had had during his many years of isolation from the Netherlands, and the dictionary was never completed. Nevertheless, he still managed to compile well over thirty thousand entries. One clue as to how he accomplished this lies in the work's handwriting styles. Other dictionaries in manuscript generally show that the Dutch entries were listed first, with the Japanese translations added later. Often this would be in a different hand, indicating that the tasks were divided among various collaborators. Thus, in the case of the Duufu haruma, Doeff himself copied most of the Dutch entries from Halma, after which he left it to the interpreters to fill in the Japanese translations.

In Blomhoff's dictionary, however, although two writing styles are apparent, these never occur on the same page. The Dutch and Japanese entries have clearly been written side by side by the same person. Several errors in the Dutch that are unlikely to have been made by a Dutch native suggest that the two writing styles belong to the young interpreters Toyokichi and Yonezõ, and that Blomhoff's role was limited to a supervisory position.

Thanks to a short note signed by Toyokichi attached to one of the volumes we know his handwriting, and can therefore identify the sections written by him, and subsequently infer which were entered by Yonezõ. Almost the entire first half of the work was written by Toyokichi, and it may be that Yonezõ was not recruited until a later stage to help out. The two interpreters appear to have initially disagreed on some aspects of romanization. Yonezõ preferred to employ the letter v where normally the letter f or h is used before vowels in Hepburn spelling. Thus, where Toyokichi wrote fito for ‘person’ and fune for ‘ship’, Yonezõ would write vito and vune respectively. Later on in the work, however, Yonezõ conformed to Toyokichi's spelling conventions.

The use of Dutch instructions among a number of the Japanese entries suggests that it was Blomhoff's intention to provide Dutch students of the Japanese language with a romanized dictionary. Unfortunately, the market for Blomhoff's dictionary was limited. The work is stored in the library of Leiden University in an almost pristine condition, and was obviously never used to any great extent.

[p. 171]

In addition to the attempt by Van Overmeer Fisscher to claim credit for his dictionary, Doeff experienced further anguish through rumours that were circulating to the effect that Von Siebold had also compiled a Dutch and Japanese dictionary. In the foreword to his book he expresses the fear that Von Siebold too might take credit for his (Doeff's) years of work (Doeff 1833:vii). He need not have worried. Von Siebold did indeed compile a dictionary, but it has nothing whatsoever in common with Doeff's work. In a single binding of no more than eighty-four pages Von Siebold arranged his entries in the Japanese iroha sequence.²⁹ In his Isagoge he describes how the work was compiled. A samurai scholar from Hizen province whom he names as Totoroki Buhitsiro compiled an initial list of Japanese words for him, taking care to select only those words which would be of the greatest use. The ‘Chinese elements’ (presumably the Chinese characters) were then verified by Oka Kenkai (1799-1839), a Dutch-style physician from Suõ Province (present-day Yamaguchi Prefecture), who had travelled to Nagasaki to study at Von Siebold's Naritaki Academy. Finally, Von Siebold enlisted the help of interpreters Yoshio Gonnosuke and Tsuujirõ, who helped him with the Dutch translation and suggested many of their own entries, which were written on small strips of paper and glued into the book at the appropriate places (Von Siebold 1841:23). Although he had a reputation as a conscientious and systematic scholar, Von Siebold freely mixed Dutch and German in the entries to his dictionary, which suggests that he compiled the work purely to assist him in his own private study of the Japanese language.³⁰ It is dated Bunsei 11 (1828), the year the so-called Von Siebold Affair brought Von Siebold's many activities in Japan to an abrupt halt.³¹

Finally, in 1854, J.K. van den Broek (1814-1866), who was posted on Dejima as physician from 1852 to 1857, began a series of attempts to compile a dictionary. He was assisted by interpreter Shizuki Ryõta (1802-1868), who gave him a copy of the Japanese-Dutch glossary Kaisei zõho bangosen, which had been published in 1847 by Mitsukuri Genpo. Van den Broek's motivation was to give himself more opportunity to communicate with Japanese authorities directly, without the intervention of the interpreters, whose services he had come to distrust (Moeshart 2003:68). He began with a Dutch-Japanese glossary, with which he got as far as the letter R. He appears to have used the Duufu haruma as well as works on Japanese grammar by Rodriguez and Léon de Rosny as resources for this project. Later, after his departure from Japan, he seems to have changed his mind as to the format of his dictionary, and eventually produced a complete Japanese-Dutch dictionary in the iroha sequence. Plans for publication were never realised, and Van den Broek's manuscripts were discovered in the basement of a public library in Arnhem, the Netherlands, in 2001 by historian Herman Moeshart (Moeshart 2003:193-197).

29 Iroha is a Heian period (794-1183) poem in which the characters of the phonetic kana syllabary each appear once. This kana sequence was often used in Japanese works of reference.

30 Serrurier (1896:20) records this work under the title Oranda jisho (‘Dutch dictionary’), but no such title is visible on or in the work itself.

31 A contraband map of Japan was found among personal goods that Von Siebold attempted to take back from Japan to The Netherlands. He was interrogated, placed under house arrest, and finally expelled from Japan in January 1830. He did, however, return in 1859 and spent a further two years in Japan (Rubinger, 1982: 117-8).

[p. 172]

Conclusion

While without a doubt the great Dutch-Japanese dictionaries of the nineteenth century formed a key component of the accelerated progress that was made in the area of Dutch language studies during the first half of the nineteenth century, the issue of cause and effect is not so simple. We have seen that the first Dutch dictionaries were donated to the Interpreters' Guild in 1754. Some ten years later interpreter Nishi Zenzaburõ began working on his translation, but he did so without the blessing or cooperation of the Guild. Rumour of Nishi's abortive attempt reached Õtsuki Gentaku in 1786, but astonishingly, he did not pursue the matter himself or set such a process in motion until Inamura Sanpaku approached him with the same idea more than six years later. Even then, the Edo haruma only came about because Inamura was faced with the prospect of having to continue his studies in his hometown, that is, far removed from his mentors and others who might be able to assist him, and felt that such a dictionary would be useful to him. Doeff, on the other hand, had little awareness of, or interest in, the academic pursuits of the rangaku scholars in Edo, and merely wished to fill in his abundance of spare time with something productive. He stated, as noted above, that the appalling quality of the interpreters' Dutch prompted him to take on the project. Both initiatives, therefore, were born of ambitions that were fairly narrow in proportion to their outcomes.

The two great dictionaries, the Edo haruma and the Duufu haruma, can therefore hardly be said to be the products of great vision. Each began as an initiative on a personal level, but took on a life of its own as the scale and consequences of the respective projects became apparent. In the case of the Edo haruma this resulted in the decision to manufacture thirty copies of the finished work. By the time news of Doeff's little enterprise reached the authorities, almost twenty years had elapsed, and the Edo haruma's obvious usefulness had created a demand that far exceeded availability. As a result, Doeff's dictionary became the object of a requisition order from the Shogunal authorities in Edo, and a dozen people were officially assigned to its completion. Subsequent growing demand prompted the production of several handwritten copies and a number of published condensed versions. Despite their somewhat small-minded beginnings, there can be no doubt that these works contributed immeasurably to the speed and accuracy of the translations that were being produced at the time, and must be seen as a significant factor in the spread of Western knowledge in Japan in the first half of the nineteenth century.

[p. 175]

Bibliography

Blussé, L., Remmelink, W., Smits, I. (ed.) 2000, Bridging The Divide. Leiden: Hotei Publishing.

Boxer, C.R. 1950, Jan Compagnie in Japan. The Hague: Martinus Nijhoff.

Doeff, H. 1833, Herinneringen uit Japan. Haarlem: De Erven Francois Bohn.

Goodman, G.K. 2000, Japan and the Dutch 1600-1853. Richmond, Surrey: Curzon Press.

Hesselink, R.H. 1995, ‘A Dutch New Year at the Shirandõ Academy.’ In: Monumenta Nipponica Vol 50 No. 2: 189-234. Tokyo: Sophia University.

Katagiri, K. 1985, Oranda tsuuji no kenkyuu. Tokyo: Yoshikawa Kõbunkan.

Kerlen, H. 1996, Catalogue of Pre-Meiji Japanese Books and Maps in Public Collections in the Netherlands. Amsterdam: J.C. Gieben.

Kiyooka E. (transl.) 1966, The Autobiography of Fukuzawa Yukichi. New York: Columbia University Press.

Koga J. 1966, Nagasaki yõgakushi. Nagasaki: Nagasaki bunkensha.

MacLean, J. 1974, ‘The Introduction of Books and Scientific Instruments into Japan, 1712-1854.’ In: Japanese Studies in the History of Science, Vol 13: 9-68.

Moeshart H. 2003, Een miskend geneesheer; Dr. J.K. van den Broek en de overdracht van kennis van westerse technologie in Japan 1853-1857. Unpubl. Dissertation, Leiden University.

Numata J., Matsumura, A., Satõ, M (ed.) 1976, Yõgaku jõ. Tokyo: Iwanami shoten.

[p. 176]

Numata J., Ishiyama, H., Umetani, N., Õmori, M., Katagiri, K., Sakai, S., Yabe, I. (ed.) 1984, Yõgakushi jiten (‘Dictionary of the History of Western Learning’). Tokyo: Yushodo Press Ltd.

Numata J. 1989, Western Learning. Tokyo: The Japan-Netherlands Institute.

Overmeer Fisscher J.F. van 1833, Bijdrage tot de kennis van het Japansche Rijk. Amsterdam: J. Müller & Co.

Rubinger R. 1982, Private Academies of Tokugawa Japan. New Jersey: Princeton University Press, 1982.

Saitõ, S. 1985, Nihon ni okeru orandago kenkyuu no rekishi. Tokyo: Daigaku Shorin.

Sansom G.B. 1950, The Western World And Japan. Alfred A. Knopf, New York.

Serrurier, L. 1896, Bibliothèque Japonaise. Leiden: E.J. Brill.

Siebold, P.F. von 1841, Isagoge. Leiden: Apud Auctorem.

Sugimoto T. 1991, Kokugogaku to rangogaku, ‘Japanese Linguistics & Dutch Linguistics’, Musashino Shoin, Tokyo.