>The Four-Partition Schema

A filesystem's directory layout is never neutral. This one encodes a theory of knowledge in the order its partitions mount.

The name is the schema. Siku Quanshu — 四庫全書 — means "Complete Books of the Four Repositories." The four treasuries (庫) are the four partitions into which the editors classified the whole of Chinese written knowledge: Classics (經部), History (史部), Masters (子部), Collections (集部). The scheme wasn't invented for the project — it had been standard since the early Tang, when the Sui History's bibliography (隋書·經籍志) formalized a classification that had been developing for centuries. The Siku project applied it at unprecedented scale.

The partitions are not sorted alphabetically, by size, or by date. They are sorted by epistemic status — by what kind of knowledge each holds and what authority that knowledge carries. The ordering is a philosophical claim with a mount sequence.

The Classics (經部) mount first because they hold the foundational texts — the works through which, the preface says, the sages "embedded instruction within practical affairs." Everything else is downstream the way case law is downstream of a constitution. Inside, the categories map the classical curriculum: the Yijing and its commentaries (易類), the Shangshu (書類), the Shijing (詩類), the Three Rites, the Spring and Autumn Annals, the Four Books, plus music, philology, and general studies.

And the Yijing sorts first within the Classics — first within the root partition, therefore first in the entire filesystem. Not arbitrary: the Yi Lei preface holds that the Changes is the root of all pattern-knowledge, the text where the sages first articulated the principles every other classic elaborates. Pattern recognition mounted at the root of the tree.

History (史部) mounts second — the record of human affairs across time. If the Classics state principles, History documents their application and misapplication: the concrete log of what happened when humans governed, fought, and failed. The subcategories reveal how the tradition modeled historical knowledge: Standard Histories (正史類) first, then chronicles, then narratives-by-episode, then the miscellaneous and unofficial.

But the partition runs wider than a Western "history" shelf. It absorbs Geography (地理類) — geographic knowledge as administrative knowledge, essential to governance. It absorbs Government Documents (政書類) — institutional manuals, legal codes, tax records. And it absorbs Bibliographies (目錄類), which means the Zongmu Tiyao classifies itself as a historical work. The catalog knows it is a document of its own moment, comprehensive for its time, not timeless.

The Masters (子部) is where it gets interesting. The character 子 is the honorific for the great pre-imperial thinkers — Kongzi, Mengzi, Laozi, Zhuangzi. But by the Siku project the partition had expanded to hold nearly every specialized knowledge that didn't fit elsewhere. The subcategories read like a map of the intellectual world: Confucian Masters first, then Military Science (兵家類), Legalism (法家類), Agriculture, Medicine, Astronomy and Mathematics, Divination and Numerology (術數類), Arts, Encyclopedias, Fiction, and finally Buddhism and Daoism.

It is the catch-all — everything neither canonical, nor historical record, nor literary art. And it most clearly exposes the value judgments baked into the schema. Philosophy sits beside medicine, strategy beside agriculture, divination beside fiction — grouped not by content but by structural position: all bodies of specialized knowledge built by particular masters and their lineages, legitimate but subordinate.

The placement of divination (術數類) inside the Masters partition tells you something Western categories would hide. In a Western library, divination shelves under "occult" or "folklore" — flags for marginality. In the Siku schema it sits next to medicine, astronomy, and military strategy as a legitimate branch of specialized knowledge.

The preface is explicit: it traces the tradition from the Yijing's hexagrams through the Han masters, notes which approaches are well-founded and which are superstitious accretions, and sets evaluation criteria. The editors dismiss plenty of works as nonsense — but they treat the field as real, a domain with a legitimate history and standards of rigor. These were the most rigorous bibliographers in Chinese history, not credulous enthusiasts. The classification was a considered judgment, not a lapse.

Collections (集部) mount last, and the position matters. It holds literary works — individual authors' collected writings, anthologies, literary criticism, drama. Last not because the editors dismissed literature, but because in this schema literary works are expressions rather than foundations. Classics provide principles; History the record; Masters specialized knowledge; Collections the literary articulation of all three. Some of the warmest, most detailed entries in the whole catalog are reviews here — but the structural verdict stands: literature serves, it does not lead.

The sequence — Classics, History, Masters, Collections — is a theory of knowledge expressed as a directory tree: principles first (經), tested and documented through experience (史), specialized fields developing from that foundation (子), individuals articulating it all through literary expression (集).

This is fundamentally unlike Western classification. Dewey Decimal (1876) and Library of Congress (~1900) organize by subject as a flat landscape of topics, equally reachable from any starting point. The Chinese four-part schema is hierarchical and directional: it has a root and a direction, and it makes an explicit claim about which knowledge is primary and which derivative. The Western systems answer "where is this book?" The Siku schema answers "what kind of knowledge is this?" — retrieval versus orientation.

To move a text from one partition to another was to change its epistemic status. The schema didn't just store knowledge. It told you, by mount order, what each thing was for.