编纂

清乾隆三十八年（1773年）二月朝廷设立了「四库全书馆」，负责《四库全书》的编纂，由乾隆皇帝的第六子永瑢负责，任命内阁大学士于敏中为总裁，大学士以及六部尚书、侍郎为副总裁。召著名学者纪昀为总纂官，「总理编书之事」。开始编纂这套卷帙浩繁的丛书。陆锡熊、孙士毅、戴震、周永年、邵晋涵等其他学者也参与了编纂。曾参与编撰并正式列名的文人学者达到三千六百多人，而抄写人员也有三千八百人。

《四库全书》收录了当时在全国各地徵收的流通图书、清内廷收藏的图书以及《永乐大典》中辑出来的珍本善本。徵收准则以阐明性学治法、考核典章、九流百家之言为优先，族谱、尺牍、屏障、寿言、唱酬诗文等等则不在考虑之列。据统计，光是徵收得的图书就达13501种，献书最多的鲍士恭、范懋柱、汪启叔、马裕四家赐以内府所印《古今图书集成》一部。当时乾隆还规定，凡从坊肆来的，应该付给一定的费用；若是家藏图书，则装裱印刷；如未曾刊刻，则抄本存留。

这些书籍分为「著录」、「存目」与「禁毁」三类处理：符合部分条件的，被列为「存目」，只存书名，不收其书。「抵触本朝」之书一概「禁毁」。符合收录条件的「著录」，著录书则经过整理、校勘、考证后，按特定格式重新抄写存入，誊写完成后，还要与原本反覆校勘，最后收录的图书有三千四百六十一种。而存目书有著录书的两倍的数量也是因为乾隆想实现观成的愿望才弃多取少。

《四库全书》为了美观与便于识别，采用分色装帧，经部绿色，史部红色，子部月白色（或浅蓝色），集部灰黑色。四部颜色的确定，依春夏秋冬四季而定。《四库全书总目》因为是全书纲领，采用代表中央的黄色。

乾隆四十九年（1784年）四套书陆续完成，全书共抄7部。第一部存于北京紫禁城皇宫文渊阁，第二部存于奉天故宫（今渖阳）文溯阁，第三部存于京郊圆明园文源阁，第四部存于承德避暑山庄文津阁，合称「内廷四阁」（或称「北四阁」），「只准大臣官员、翰林等阅览」。后又以江浙地区为人文渊薮，文风鼎盛，就再钞的正本三部，分别存在镇江金山寺建文宗阁，扬州大观堂建文汇阁，杭州西湖行宫孤山圣因寺建文澜阁，即「江浙三阁」（或称「南三阁」），各藏抄本一部，「为便于士子入院钞阅」。副本存于京师翰林院（其中文渊阁本最早完成，校勘更精、字体也更工整）。

编纂《四库全书》时，清廷为维护统治，大量查禁明清两朝有所谓违碍字句的古籍（见四库禁书）。据统计，在长达10馀年的修书过程中，「荦荦大者文字之狱共有三十四件」。乾隆五十二年（1787年）乾隆皇帝抽查《四库全书》时发现一些书有诋毁清朝的字句，因此下令重检《四库全书》，并最终删除《诸史同异录》等11部。但这11部书虽然从《四库全书》中删除，但是依然存在宫中，没有销毁，这11部书中的9部还流传到今天。

嘉庆八年（1803年）由纪昀主持《四库全书》最后一部分官修书籍的补遗工作，进一步完善《四库全书》。

特点

古籍辑佚

清朝在编纂《四库全书》的过程中寻回和修复了不少早已失传的中国古籍，梁启超对此评论道：「此二百馀年间总可命为中国之『文艺复兴时代』，特其兴也，渐而非顿耳……吾辈尤有一事当感谢清儒者，曰辑佚。书籍经久必渐散亡，取各史艺文、经籍等志校其存佚易见也。肤芜之作，存亡固无足轻重；名著失坠，则国民之遗产损焉。乾隆中修《四库全书》，其书之采自《永乐大典》者以百计，实开辑佚之先声。此后兹业日昌，自周秦诸子，汉人经注，魏晋六朝逸史逸集，苟有片语留存，无不搜罗最录。其取材则唐宋间数种大类书，如《艺文类聚》、《初学记》、《太平御览》等最多，而诸经注疏及他书，凡可搜者无不遍。当时学者从事此业者甚多，不备举。而马国翰之《玉函山房辑佚书》，分经史子三部，集所辑至数百种，他可推矣。遂使《汉志》诸书、《隋唐志》久称已佚者，今乃累累现于吾辈之藏书目录中，虽复片鳞碎羽，而受赐则既多矣。」

收录西学

《四库全书》对西方科学书籍作出较高的评价：「欧罗巴人天文推算之密，工匠制作之巧，实逾前古。其议论夸诈迁怪，亦为异端之尤。国朝节取其技能而禁传其学术，具存深意。」，同时将西方宗教与西方科学作了区分，强调「节取其技能」，因为《四库》的作者们明白传教士的来华目的是为了传播天主教，当时的学者已经发现西方存在传播宗教与价值观的倾向，介绍西学不过是传教的手段，因此《四库全书》的作者们在赞扬西学的同时指出「盖欲借推测之有验，以证天主堂之不诬，用意极为诡谲」。

《四库全书》收录了27种西方传教士的著述，包括西洋的数学、天文、仪器及机械等方面的著作，被收录的书籍包括有《泰西水法》、《西儒耳目资》、《坤舆图说》、《乾坤体义》等。

成书之后

《四库全书》完成至今的两百年间，中国历经动乱，《四库全书》也同样饱经沧桑，多份抄本在战火中被毁。因此《四库全书》至今仅存4套：

• 文宗阁本、文汇阁本在咸丰三年（1853年）太平天国运动期间被毁；

• 文源阁本在咸丰十年（1860年）英法联军攻占北京、火烧圆明园时被毁，今法国保留部分残本；

• 清代翰林院底本在光绪廿六年（1900年）庚子拳乱时，位于英国使馆北邻的翰林院遭纵火，毁于拳民围攻东交民巷外国使馆期间；

• 文渊阁本原藏北京故宫，后经上海、南京转运至台湾，现藏台北市国立故宫博物院（也是保存较为完好的一部）。

• 杭州文澜阁藏书楼于咸丰十一年（1861年）在太平军第二次攻占杭州时倒塌，所藏《四库全书》散落民间；后由藏书家丁氏兄弟收拾、整理、补抄，才抢救回原书的四分之一，于光绪七年（1881年）再度存放入修复后的文澜阁。文澜阁本在经历1915－1923年的乙卯补抄和1923－1925年的癸亥补抄后大体恢复旧观，现藏杭州浙江省图书馆。

• 文溯阁本于民国十一年（1922年）险些被卖给日本人。1966年10月，因中苏关系紧张，为保《四库全书》安全，林彪要求将文溯阁《四库全书》秘密从渖阳运至兰州，藏于戈壁沙漠中，现藏兰州甘肃省图书馆。近些年，辽宁有关人士要求归还，以「书阁合璧」；甘肃亦建藏书楼加强保护。对于文溯阁本是否要归还渖阳，两省一直未能达成一致。

• 避暑山庄文津阁本于1931年移入国立北平图书馆（现中国国家图书馆）收藏。（这是目前唯一一套原架原函原书保存的版本）

民国初期，商务印书馆影印了《四库全书珍本初集》。1986年，台湾商务印书馆影印出版了文渊阁本《四库全书》，上海古籍出版社曾将之缩印。1999年，香港的迪志文化出版有限公司分别与上海人民出版社以及香港中文大学在中国大陆及香港出版发行文渊阁本《四库全书》电子书版。2006年，迪志文化出版有限公司出版全新《文渊阁四库全书》电子版的内联网版和网上版。2009年，台湾商务印书馆由国立故宫博物院授权，按照典藏之文渊阁本原书原寸，以仿古样式影印出版，即内文、开本、纸质、布面、装帧方式等皆与原版相同。

2016年8月9日，历时12年方告完成的影印版文津阁《四库全书》重新入藏文渊阁，该影印版制作时间长达12年，全部采用手工宣纸，包装上也恢复了传统的装订工艺，书封采用耿绢。整套书用6144个金丝柚木盒，重新放到了空置已久的文渊阁128个书架上。

评论

• 鲁迅在批评明清和民国文人时表示：「现在不说别的，单看雍正乾隆两朝的对于中国人著作的手段，就足够令人惊心动魄。全毁，抽毁，剜去之类也且不说，最阴险的是删改了古书的内容。乾隆朝的纂修《四库全书》，是许多人颂为一代之盛业的，但他们却不但捣乱了古书的格式，还修改了古人的文章；不但藏之内廷，还颁之文风较盛之处，使天下士子阅读，永不会觉得我们中国的作者里面，也曾经有过很有些骨气的人……清朝的考据家有人说过，「明人好刻古书而古书亡」，因为他们妄行校改。我以为这之后，则清人纂修《四库全书》而古书亡，因为他们变乱旧式，删改原文；今人（民国）标点古书而古书亡，因为他们乱点一通，佛头著粪：这是古书的水火兵虫以外的三大厄。 」

• 历史学家吴晗慨叹：「清人纂修《四库全书》而古书亡矣！」。

• 唐弢：《四库全书》...使作者的思想和躯壳一齐淹没，是近于消极的取消。至于改窜，是要借作者的名声，替改者说好话，使读者暗受其影响，看不出底细来。所以这不但是毁尸灭迹，而且还借刀杀人，要古人也和自己一样，让他们进反省院，写悔过书，从此变成正统。然而《四库全书》不但挂著钦定的招牌，便是删改和摈斥，在乾隆，也是不加深讳的。他的给《四库》总裁们的谕旨里，就常常明白地指示著：哪一种应该删改，哪一种应该厘订，大都说得很清楚。他似乎并没有料到：他的那些谕旨，在将来，正是自己篡改古书罪案的证人。

• 史学家陈寅恪在《陈垣元西域华人化考序》中指出清朝所禁毁的书籍几乎都是晚明清初的书，其他书籍甚少：「然清室所最忌讳者，不过东北一隅之地，晚明初清数十年间之载记耳。」

• 费正清：通过这项庞大工程，清廷实际上进行了一次文字清查（文学上的「宗教裁判」）工作，其目的之一是取缔一切非议外来统治者的著作。编纂人在搜求珍本和全整文本以编入这一大文库时，也就能够查出那些应予取缔或销毁的一切异端著作。他们出善价收集珍本，甚至挨家挨户搜寻。该禁的图书是研究军事或边务的著作以及有反夷狄之说的评议，还有就是那些颂扬明朝的作品。……正如L. C. 古德里奇所论证的，这是最大规模的思想统治。

• 史景迁：编纂《四库全书》还具有检视文献的目的，藉此搜查私人藏书，并严惩那些收藏有轻视满人内容的书籍的人。这类书，以及包含有害于中国国防的地理和游记类书籍也被销毁。这些毁书行动之彻底，使得我们知道的被乾隆的文化顾问纳入禁毁之列的两千多种图书从此销声匿迹。

• 独立学者周英杰：你当然可以说它是一项前无古人的伟大事业，但也完全可以说它是「阉割」中国传统文化的一项「缺德工程」。

• 历史学者张宏杰：「经过康熙雍正两代文字狱，清王朝思想专制已经大大加强。乾隆皇帝又通过修四库全书，开展「消灭记忆」运动。」

• 中国国家新闻出版广电总局规划发展司副司长李建臣认为，编纂《四库全书》过程也是对中国传统文化最大的一次戕害。只要是被认为对清朝统治不利的著作和图书，就会毫不犹豫地统统毁掉，被销毁的各类图书典籍在3000馀种，6万馀卷，15万册，70多万部。这导致中华民族的一些优秀思想文化成果和善本、孤本、秘本、禁毁本等一大批珍贵史料典籍永远消失，甚至对明代皇家档案和有关明朝类的史书、笔记等也进行了系统销毁。由于《四库全书》是国家强行徵集，民间所珍藏的典籍无一漏网，尽皆被毁。

• 季羡林评论《四库全书》为「我国现存最大的一部官修丛书，是清乾隆皇帝诏谕编修的我国乃至世界最大的文化工程。全书分经、史、子、集四部，收书3503种，79309卷，存目书籍6793种，93551卷，分装36000馀册，约10亿字。相当于同时期法国狄德罗主编《百科全书》的44倍。清乾隆以前的中国重要典籍，许多都收载其中。由于编纂人员都是当时的著名学者，因而代表了当时学术的最高水平。乾隆编修此书的初衷虽是「寓禁于徵」，但客观上整理、保存了一大批重要典籍，开创了中国书目学，确立了汉学在社会文化中的主导地位，具有无与伦比的文献价值、史料价值、文物价值与版本价值」。

• 史学家龚书铎评价：「乾隆年间《四库全书》的编纂，为华夏文明的延续做出了不可磨灭的贡献，为后人保留了许许多多珍贵的书籍和资料，但是，在编纂过程中，也免不了发生许多令人遗憾的事情」。

• 梁启超：「乾隆中修《四库全书》，其书之采自《永乐大典》者以百计，实开辑佚之先声。此后兹业日昌，自周秦诸子，汉人经注，魏晋六朝逸史逸集，苟有片语留存，无不搜罗最录。」

• 哈佛大学历史学教授、新清史学派学者欧立德指出，乾隆从一开始就想通过编纂《四库全书》来对学术和文学的发展加以审查的说法有失公允。乾隆的初衷是「蒐集和保存古今之图书」，以塑造出自己是「文化巨人」的正面形象，尽管乾隆期间的文字狱带来种种乱象，但并非是乾隆一个人的责任，他指出，许多文字狱的判罚虽然以朝廷的旗号进行，但这并不能说明大规模的审查得到了乾隆本人的授意，虽然这和乾隆的「报复心」有关，但「底层官员的热情和野心以及地方上发生的那些与地主士绅相关的琐碎纷争至少也应该承担同样的责任」。乾隆不出面阻止的原因，是因为清朝皇帝的皇权实际上并不是无限的，乾隆无力阻止官僚扩大化文字狱。哈佛大学历史学家、汉学家孔飞力以「叫魂案」为例，指出「清代皇权并非真正意义上的独裁，皇权与官僚制度存在不稳定的依存关系，皇权必须随时提防来自官僚制度的惰性威胁」。中国历史学教授高王凌在也提出过相似的概念「反行为」，高王凌认为清代官僚对于皇权只是在表面服从的前提之下，在暗地里却干著只可意会不可言传的猫腻行为，从而将皇帝的意志「反过来」。

相关典籍

• 《四库全书总目提要》/《四库全书总目》/《钦定四库全书总目》二百卷

• 《四库全书简明目录》二十卷

• 《四库全书荟要》一万二千册

• 《四库全书考证》一百卷

• 《武英殿聚珍版丛书》

History

Creation

Late in the 18th century, the Qing dynasty set about a momentous task, the creation of the Siku Quanshu. The Qianlong Emperor ordered the creation of the Siku Quanshu in 1772. Local and Provincial officers were in charge of locating and collecting important books. The Qianlong Emperor encouraged owners of rare or valuable books to send them to the capital, however few actually did due to concerns about the Literary Inquisition. Towards the end of 1772, seeing that only a limited number of people actually handed in books, the Qianlong Emperor issued imperial decrees stressing that books would be returned to their owners once the compilation was finished and that owners of the books would not be persecuted if their books contained anti-Manchu sentiment. Less than three months after the issue of this decree, four to five thousand books were handed in.

By March 1773, an editorial board (composed of hundreds of editors, collators, and copyists) was created in Beijing to gather and review books brought to them. This board included over 361 scholars, with Ji Yun and Lu Xixiong (陆锡熊) as chief editors. There was around 3,826 scribes who copied every word by hand. These copyists were not paid in coinage but in government positions after they had transcribed a set amount of the encyclopedia. It took over a decade until the encyclopedia was completed and all seven copies were distributed.

By 1782, the special guide to the Siku Quanshu was completed, the Siku Quanshu Zongmu Tiyao, which contains accurate information of bibliography of the 3,593 titles in the Siku Quanshu itself. Additionally, it also contains bibliography information of 6,793 other books that are not available in the Siku Quanshu. The Siku Quanshu Zongmu Tiyao would not be published until 1793 and when released, the Siku Quanshu Zongmu Tiyao became the largest Chinese book catalog of the time.

Compilation

The compilation of the Siku Quanshu started with the Siku Quanshu Zongmu Tiyao. Fully compiled in 1773, editing would begin shortly, with the first workable drafts being completed in 1781.This would include bibliographic explanation of all the works fully included in the final Siku Quanshu, as well as a large number of works that are included in title only. The Siku Quanshu contains 4 series: Confucian Classic which contains important works of Confucius, Belles-Lettres which contain literary works ranging from personal letters to poetry or writing meant for the masses, and finally Historiography and Masters which houses works from scholar and the content can range from scientific works and military works.

A large number of these edits were made to limit the effects of embellishments and other inaccuracies found in local records. Personal documents, often entailing the actions noteworthy local people, that could be verified through preexisting government documents were often included into the Siku Quanshu Zongmu Tiyao for consideration for inclusion into the finished Siku Quanshu. However, documents that could not be verified were often included in title only, and were criticized by the compilers as unfit for full inclusion into the finished collection. Not even officially sponsored writers, such as local gazetteers, were safe from the scrutiny of the official compilers, leading to criticisms of adding or using ambiguous sources to elevate local figures to be more significant than they actually were.

Medical knowledge was often documented through case-style narratives first seen in twenty-five instances in Sima Qian's Records of the Grand Historian. These instances would form prototypical templates for future medical accounts. Medical accounts from then on took a narrative voice with a secondary analytical tone focusing on a blend of storytelling, vocational knowledge, and historical recording. However, as time progressed, the vocabulary used to record the medical cases started to differ from author to author. By the time of the Qing dynasty, however, the language used to create and define the medical cases had begun to resettle, allowing for ease of inclusion of these texts into the Siku Quanshu.

Case-based recording and discourse of philosophy also was a target for compilation. Similar to how medical knowledge had a prototypical template for future works, Huang Zongxi's writings in the field had largely served a similar purpose. However, despite popular discourse among scholars of this era, philosophical writing had suffered immensely from two problems. Firstly, was a lack of clear definition regarding philosophical writings as a whole, giving rise to two separate, but equal definitions. "Archival" would mean that the philosophical work would be defined as a scholarly article. Whereas, "cultural" would mean the literature would be rearranged as a Buddhist Kōan, though any writing presented this way would be interpreted in a more literal fashion as compared to the traditional rhetorical question. Chinese philosophical writing's other problem at this time would be a lack of bibliographical classification, largely in part due to authors and previous compilers not considering any philosophical work as part of a historical record. As such, compilers for the Siku Quanshu redefined the classifications of several compilations that made it into the published copies, and set boundaries based on the author's biographical history and intent of their writing in an attempt to remedy these dilemmas.

The Qianlong Emperor made reviews on works that were currently being compiled, and that their opinions on the work reviewed were often conveyed through direct comments or imperial edicts. This, in turn, colored the official compilers criteria for works suitable for inclusion in the Siku Quanshu, to more closely align with those of the emperor. However, the emperor often commented poorly on works from or about their political rivals, especially opponents holding Anti-Manchu sentiments, running in contrast to stories from locally published sources. This can be exemplified in the compilers' handling of the story of Zhang Shicheng, and his rival Zhu Yuanzhang. In this particular case, the Qianlong Emperor sought to discredit the previous dynasty by highlighting the cruelty of early Ming dynasty rule. Ming-era rule would be contrasted by comparison to Qing-era policies, which appeared more palatable in comparison to harsh judgement. The Qing dynasty acknowledged the legitimacy of the former Hongwu Emperor, but by portraying him in this way, sought to solidify the legitimacy of their own dynasty by raising doubts of the latter dynasty's rule. Moreover, the compilers did not see Zhang Shicheng's rule as legitimate, but as a natural response to the narrative tyranny of the people under the Ming dynasty.

Distribution

The Qianlong Emperor commissioned seven copies of the Siku Quanshu to be made. The first four copies were for the Emperor and were kept in the north. The Qianlong Emperor constructed special libraries for them. They were located in the Forbidden City, Old Summer Palace, Shenyang, and Chengde. The remaining three copies were sent to the south. They were deposited into libraries within the cities of Hangzhou, Zhenjiang, and Yangzhou. All seven libraries also received copies of the 1725 imperial encyclopedia Gujin tushu jicheng.

The copy kept in the Old Summer Palace was destroyed during the Second Opium War in 1860. The two copies kept in Zhenjang and Yangzhou were completely destroyed while the copy kept in Hangzhou was only about 70 to 80 percent destroyed, during the Taiping Rebellion. The four remaining copies suffered some damage during the Second Sino-Japanese War. Today, those copies can be located in the National Library of China in Beijing, the National Palace Museum in Taipei, the Gansu Provincial Library in Lanzhou, and the Zhejiang Library in Hangzhou.

Censorship

The Qianlong Emperor did not keep his promises to return the books. Any books that did not make it into the Siku Quanshu risked becoming part of the Siku Jinshu. The Siku Jinshu is a catalogue of over 2,855 books that were rejected and banned during the completion of the Siku Quanshu. An additional four to five hundred other books were edited and censored. A majority of the books that were banned were written towards the end of the Ming dynasty and contained anti-Manchu sentiment. The Siku Jinshu was the Qianlong Emperor's attempt to rid the Qing dynasty of any Ming Loyalists by executing scholars and burned any books that gave direct or implied political attacks against the Manchu.

Contents

A page from the Siku Quanshu.Each copy of the Siku Quanshu was bound into 36,381 volumes (册), with more than 79,000 chapters (卷). In total, each copy is around 2.3 million pages, and has approximately 800 million Chinese characters.

Complete Catalogue

The scholars working on the Siku Quanshu wrote a descriptive note for each book which detailed the author's name along with their place and year of birth. Next, after they determined what parts of the author's work would go into the compilation, they would analyze the main points of the author's argument. This short annotation, which reflected their own opinion, would be put in the beginning of the Siku Quanshu and formed the Complete Catalogue. The Complete Catalogue was divided into four sections or kù (库; translated to "warehouse; storehouse; treasury; repository"), in reference to the imperial library divisions. The name, Siku Quanshu, is a reference to these four sections. These four sections are:

• Jīng (经 "Classics") Chinese classic texts

• Shǐ (史 "Histories") histories and geographies from Chinese history

• Zĭ (子 "Masters") philosophy, arts, sciences from Chinese philosophy

• Jí (集 "Collections") anthologies from Chinese literature

44 Sub-Categories

The books are then divided into 44 sub-categories, or lèi (类). The Siku Quanshu collection includes most major Chinese texts, from the ancient Zhou Dynasty to the Qing Dynasty, covering all domains of academia. It also lacks any Western or Japanese texts. Included within these 44 sub-categories are: the Analects of Confucius, Mencius, Great Learning, Doctrine of the Mean, I Ching, Rites of Zhou, Classic of Rites, Classic of Poetry, Spring and Autumn Annals, Shuowen Jiezi, Records of the Grand Historian, Zizhi Tongjian, The Art of War, Guoyu, Stratagems of the Warring States, Compendium of Materia Medica, and other classics.

Authors in the Siku Quanshu

Two of Zhao Yiguang's works are housed in the Wang Qishu, they were the Jiuhuan Shitu (九圜史图) and the Liuhe Mantu (六匌曼图). They were part of the Siku Quanshu Cunmu Congshu (四库全书存目丛书).