Chinese Bibliographies (Mulu): How Catalogues Classified Books and Knowledge

A mulu is more than a list of titles

An entry connects several kinds of evidence

The word mulu 目录 can refer broadly to a catalogue or bibliography, but a useful entry does more than name a book. It may join title, attributed author, number of juan, category, editorial note and the circumstances in which a work was seen or reported. The Oxford Handbook chapter on libraries, catalogues and lost writings places cataloguing beside library formation and textual transmission. Read every field as a claim with a history, not as a transparent label attached forever to one stable object.

Bibliography, shelf inventory and sale list answer different questions

A court bibliography may map inherited learning, a private catalogue may describe one collection, and a bookseller's list may advertise what could be supplied at a particular moment. None is automatically a census of all books then in existence. The guide to Chinese private libraries explains collection catalogues, while the article on bookshops and book markets treats commercial lists as time-bound offers. Purpose must be identified before absence or abundance is interpreted.

Dynastic bibliographies preserve books—and editorial decisions

The Hanshu Yiwenzhi is a transmitted historical bibliography

The bibliographic treatise in the Hanshu is among the foundational witnesses for early Chinese book history. It reaches us as a chapter in a transmitted dynastic history, not as the physical shelf list of a surviving Han library. Robert Eno's guide to traditional bibliography introduces the Yiwen zhi sections of the standard histories as research tools. Their titles, attributions and divisions preserve precious evidence, but each record belongs to a compilation, redaction and transmission history.

A lost title can remain visible through its record

Historical catalogues name both surviving and lost works. Daniel Patrick Morgan's comparison of the Hanshu, Suishu and Jiu Tangshu bibliographies uses authorship, date and indigenous categories while openly documenting disagreements and omissions. A catalogue entry may therefore be the last trace of a work, yet it cannot supply the work's full contents. Conversely, omission does not prove that no copy, alternate title or uncatalogued text existed.

Classification makes an argument about knowledge

Categories moved as compilers reorganized the textual world

A book's category is evidence of how one catalogue positioned it. Morgan shows subjects moving, splitting or merging between Han, Sui and Tang bibliographic treatises, and warns that a bibliographic category is not automatically a modern genre or discipline. Readers should compare neighbouring headings and the catalogue's whole architecture. The excerpting system of Chinese classified books and the ordering of dictionaries and rhyme books offer related examples of knowledge being made searchable through historically specific structures.

The Four Branches became influential, not timeless

The Four Branches—Classics, History, Masters and Collections (jing, shi, zi, ji)—became a highly influential framework. The University of Southern California's rare-book guide to the Four Branches summarizes their broad contents and emergence. Yet subcategories, placements and boundaries changed across catalogues. A Daoist text, technical manual or literary collection should not be forced into an imagined universal shelf plan; classification records an editorial choice made for a particular corpus.

Descriptive notes turn cataloguing into criticism

A tiyao can describe and judge

An annotated catalogue may summarize a book's author and contents while also evaluating its reliability, style or place in a lineage of texts. The annotation is therefore not neutral metadata. It can transmit learning, expose an editor's standards and shape later reputations. Prefaces and afterwords perform related framing work, as the guide to Chinese xu and ba explains, but a catalogue note speaks from the structure and institutional authority of the catalogue that contains it.

The Siku Zongmu emerged through revision

National Taiwan University research on three versions of Siku front annotations distinguishes draft annotations, revised front notes and finalized catalogue entries, and finds differences in title, authorship, volume count and evaluation. This layered process matters: the familiar note is a finished editorial product, not one scholar's untouched first reaction. Comparison among stages can reveal how description and judgement were standardized, negotiated or changed.

The Four Treasuries joined collecting, selection and order

The project was both a collection and a classification enterprise

The Library of Congress record for the Siku quanshu identifies a vast Qing imperial compilation organized in four sections. Stefano Gandolfo's study, ‘To Collect and to Order’, emphasizes the scale and complexity of the project as both text collection and ordered knowledge. The accompanying catalogue should therefore be read within imperial collecting, copying and editorial selection—not simply as a modern neutral bibliography placed beside a finished library.

Inclusion, exclusion and description are separate acts

A work could be described even when it was not copied into the main collection, and an entry's judgement could influence later readers independently of the book itself. These distinctions prevent a common error: equating a catalogue title with a surviving, complete, uniform edition. The histories of individual collected works and poetry anthologies show how selection and arrangement can construct an author or canon even when the surviving materials are unstable.

One title may conceal several material witnesses

Title, juan, edition and copy are not interchangeable

Catalogues may record the same or similar title with different volume counts, attributions or locations. A title can gather recensions, supplements, rearrangements or later reprints; a physical copy can be incomplete while the entry describes an intended whole. Before identifying two records as the same book, compare author statement, chapter count, prefaces, table of contents, opening and closing lines, edition information and the actual copy's condition. The guide to illustrated Chinese books likewise shows how recutting and detached leaves change evidence.

The blank space around a record also matters

A missing category, unrecorded edition or absent author may reflect loss, selection, access, cataloguing policy or the compiler's purpose. It should not be converted immediately into a claim that a work never circulated. Catalogues are especially valuable when their silences can be compared with biographies, quotations, surviving copies and other bibliographies. Their limits are part of the evidence, not defects that can be repaired by guesswork.

Modern catalogues extend the tradition with new rules

A union catalogue aggregates records rather than merging every object

Modern library records add identifiers, edition statements, copy notes, subjects and holdings. Harvard Library's guide to Chinese bibliographies and union catalogues points researchers toward tools that bring records from many institutions together. A shared title in several catalogues may represent multiple copies, different editions or duplicated metadata. Search results are a discovery layer: the scanned title page, catalogue note and physical or digital witness still need comparison.

Traditional and modern fields should be translated cautiously

A traditional category can guide browsing, but it does not map perfectly onto a present discipline or subject heading. Likewise, a modern keyword can retrieve records that historical compilers would have separated. Preserve the original wording and hierarchy before translating it into current terms. This protects historical difference while allowing catalogue data to be compared, counted or mapped with explicit assumptions.

Read a bibliography through eight accountable questions

Begin with compiler, institution and purpose

Ask who compiled the catalogue, for which institution, at what date, from what materials and for which readers. Then identify whether it is a dynastic bibliography, an imperial project, a private-collection catalogue, a sales list, a modern library record or another form. Check what an entry actually records—title, author, juan, category, annotation, edition or copy—and keep those levels separate.

Treat the record as a guide, not the book itself

Use eight questions: 1) what kind of catalogue is this; 2) who made it; 3) what corpus could they inspect; 4) how are categories organized; 5) what fields does each entry contain; 6) does it describe a title, edition or copy; 7) what changed across catalogues; and 8) what does absence fail to prove? This method links bibliographic practice to the wider history of Chinese Poetry and Literature without treating catalogues as transparent shelves.