is one of the insuperable difficulties that confront the translator when he tries to convey the beauty of Shōnagon’s prose in a language as remote from Heian Japanese as modern English. Should he reproduce each okashi and each ito by a given English equivalent, however monotonous and banal the result may be for Western readers? Or should he conceal the repetitiveness of Shōnagon’s style by searching for synonyms or even by leaving out some of her favourite words when they seem to add little to the meaning? In broader terms, should he reproduce her sentences with the greatest possible mechanical accuracy, or try to suggest the poetic quality of her language at the cost of obscuring certain characteristic elements of her style? One possibility would be to produce both a literal and a literary version; but even the most long-suffering publisher could hardly welcome that solution.
As usual in translation, one must compromise between the two extremes. When in doubt, I have tended to be ‘free’. This is partly because the language of The Pillow Book, in which the most laconic phrasing is often combined with seeming redundancy, is peculiarly resistant to literalism. Any ‘accurate’ translation would impose terrible ordeals on all but the most determined. Since Shōnagon’s book is noted for the limpid beauty of its language, a translation that adhered to the exact wording of the original, faithfully reproducing each particle, each repetition, each apparent ambiguity, would from a literary point of view be totally inaccurate. A language that afforded as little pleasure to the Japanese as the following passage does to English readers would hardly have preserved The Pillow Book from oblivion for a thousand years:
the manner in which [they] did such things as deliver [honourable] letters and move about and behave was not awkward-seeming and [they] conversed and laughed[.] even wondering indeed when in the world [I] would mix thus was awkward[.]
While I have not aspired to convey the beauty of Shōnagon’s prose, I have at least tried not to obscure it entirely by the stark, graceless literalism and the ‘rebarbative barricades of square brackets’ that Mr Vladimir Nabokov, for one, appears to recommend. When the need to put Shōnagon’s sentences into readable English has obliged me to take unusual liberties with her text, I have appended a more or less literal translation in the notes. Students and other readers who require a close translation of the entire book should refer to Les Notes de chevet de Sei Shōnagon (Paris, 1934), in which Docteur André Beaujard has conscientiously retained everything that was possible from the original and indicated necessary additions by a liberal, though not always consistent, use of brackets.
In translating the quoted and original poems from The Pillow Book, I have abandoned all attempts to be literal and have tried instead to give their general meaning and to suggest a certain poetic rhythm. I have not preserved the line or word patterns of the poems unless they seemed to lend themselves naturally to those forms in English. There can be no literature in the world less suited to translation than classic Japanese poetry; and it is only because verse is such an integral part of The Pillow Book that I have ventured on an undertaking that is unlikely ever to succeed. Docteur Beaujard has provided ad verbum, prosaic versions, arranging them all, line by line, in the forms of the original poems.
My complete translation (Oxford University Press and Columbia University Press, 1967) is based primarily on the Shunsho Shōhon version as edited by Kaneko Motoomi in 1927 and on the Nihon Koten Bungaku Taikei edition of the Sankanbon version edited by Ikeda Kikan and Kishigami Shinji in 1953. Publication in a single volume necessitated certain cuts. As a devotee of Sei Shōnagon I found it hard to excise passages of her book; but in the hope that this new edition would make her work available to many more readers I removed the necessary number of pages from my original translation and from the accompanying notes. Most of the cuts are lists, especially lists of place names, words, titles, and the like that are interesting mainly to the specialist. Though Sei Shōnagon would certainly have disapproved of such tampering with her text, which she might well have included in her list of Presumptuous Things, I am confident that I have not jettisoned a single passage of outstanding interest or beauty.
With a few exceptions I have avoided making any additions to the text. Japanese authors, especially those writing in the classical language, omit personal names and pronouns as much as possible; in direct quotations the identity of the speakers is usually left to the reader’s imagination. All this has to be supplied if the text is to be comprehensible in English. When Shōnagon does identify her characters, she usually refers to them by their titles or offices. This helps to date the sections, but can result in great confusion since people frequently changed their posts; a gentleman who appears in one section as a Chamberlain, for example, may be described a few pages later as an Imperial Adviser and ‘Chamberlain’ may now refer to an entirely different person. In my translation I normally identify men by their given names (e.g. Korechika) rather than by their titles (e.g. Major Counsellor).
I have headed each of the sections with a title. In the lists these are the first words given by Shōnagon herself (e.g. ‘Hateful Things’); in the other sections they are the first words of my translation (e.g. ‘Once during a Long Spell of Rainy Weather’). I have also added my own numbers for each section. I have not indicated these various additions by square brackets; if brackets were used consistently, that is, if they enclosed every single word and punctuation mark not in the original, almost each sentence would have a dozen or more pairs and to read the text would be a suffering for all but the most resolute students.
In translating titles, government offices, and the like I have normally followed the nomenclature in R. K. Reischauer’s Early Japanese History, but I have occasionally altered his terms when they seemed cumbersome or misleading. Except when it was essential for clarifying puns, I have usually not translated proper names. This is not a result of ‘Translator’s Despair’ but because I wished to avoid the type of fake exoticism that can result from identifying the Emperor’s residence, for example, as ‘the Pure and Fresh Palace’. Names should not be made to sound more colourful in translation than they do to the reader of the original Japanese. For the same reason months are identified by their numbers (e.g. Fifth Month), which are clearer, though admittedly less poetic, than literalisms like ‘Rice-sprouting Month’. I have, however, given direct translations of the hours (‘Hour of the Monkey’, ‘Hour of the Sheep’, etc.) since there is no simple Western equivalent for the zodiacal system of timekeeping. My translations of trees, flowers, birds, and the like are often approximations; I have preferred to use words that correspond more or less to the Japanese original and that have a similar degree of familiarity (e.g. ‘cypress’ for hinoki and ‘maple’ for kae no ki) rather than technically exact equivalents (Chamaecyparis obtusa and Acer pictum), which would be meaningless to most non-specialists.
Some fifty sections of The Pillow Book, representing about two fifths of its total length, can be dated by methods that are on the whole reliable. It would have been a simple matter to rearrange these sections in chronological order, possibly putting them all together in the first part of the book, which would then become a sort of diary. I have, however, preferred to retain the confused time-sequence of the traditional texts, not because this was necessarily the order in which Shōnagon arranged her book, but because any systematic reorganization would be arbitrary and possibly misleading. If one thing is clear about the writing of The Pillow Book, it is that Shōnagon was not keeping a daily, or even a monthly, record of events. To suggest that this was her intention would falsify the spirit of her work. Readers who wish to peruse the datable sections in chronological sequence can do so by consulting Appendix 5 (Chronology).
The notes are numbered consecutively from 1 to 584 and have been placed separately, in order to avoid encumbering the text. Shōnagon and her courtly contemporaries, who expected their world and its customs to continue as long as civilized society lasted, would no doubt have been shocked to find that such a large quantity of annotation and scholarly accessories was necessary to explain an informal, seemingly simple collection of lists, descriptions, and anecdotes; but without supplementary material of this kind much of The Pillow Book
