Knowledge Organization in Obsidian

Knowledge organization in Obsidian is best treated as a practical, adaptable system—not a one-time “perfect” setup. The core message is that personal knowledge management should borrow proven ideas from library and archival practice (classification, cataloging, and controlled access points), then reshape them to fit how a specific person works. The payoff is less time spent reinventing structure and more reliable retrieval when notes multiply.

The talk begins by rejecting the idea that any single file structure or tagging scheme will solve everyone’s problems. Instead, it offers a rapid history of how Western institutions organized knowledge—starting with Mesopotamian tablets labeled with descriptors, moving through ancient Greek and Roman library practices that used prefatory scrolls and author-related metadata, and then into periods where organization often reflected the “whims” of whoever controlled collections. After the printing press, the information flood triggered new classification efforts, including Gabriel Naudé’s seven-category library scheme (theology, medicine, juris law, history, philosophy, mathematics, and amenities), designed partly to make shelves look orderly.

Modern library systems emerged as collections grew and became too large for ad hoc methods. In the United States, federal investment after the burning of Washington’s library led to the Jefferson collection being acquired without a ready sorting system, prompting improvised approaches. Later, major classification frameworks took hold: Dewey Decimal Classification (still widely used in public libraries), the Library of Congress Classification (built from Cutter Expansive Classification and now the most extensive hierarchy), and Universal Decimal Classification by Paul Otlet and Henri La Fontaine—described as faceted and analytical-synthetic, enabling database-style queries using number strings (for example, filtering by subject and time range).

From that history, the talk draws a key distinction: classification answers “where things go,” while cataloging answers “what things are.” In personal note systems, “classification” maps to folder structures and shelf-like placement, while “cataloging” maps to the internal structure of notes—titles, metadata, and linked content that make retrieval possible. The speaker emphasizes that humans naturally remember locations (a “memory palace” effect), which is why folder trees and consistent placement work.

The practical guidance then shifts to how to implement these ideas in Obsidian. Folder structures can vary by project: one part of a vault might use a decimal-style scheme, while another uses a different hierarchy better suited to the content. The talk also recommends reusing established classification systems rather than building from scratch, with specific examples like Cutter and Library of Congress subject headings (including “see also” and narrower terms). Finally, it challenges simplistic tag-first thinking. Tags are compared to uncontrolled reader indexing—useful but prone to inconsistency and misspellings—so the talk argues for more structured retrieval via linking and “access points” (library-style entry terms). As a creative twist, it discusses using emojis as semi-structured status or note-type markers, noting that emojis carry real semantics and could be treated as a more constrained alternative to free-form tags.