Skip to main content

About This Project

The Propaedia

The Propaedia, or "Outline of Knowledge," is one of three parts of the Encyclopaedia Britannica (15th edition). Created under the direction of Mortimer J. Adler, it provides a systematic classification of all human knowledge organized into 10 Parts, 41 Divisions, and 177 Sections.

Unlike a traditional encyclopedia arranged alphabetically, the Propaedia presents knowledge hierarchically - showing how subjects relate to one another within the broader landscape of human understanding.

This Web Edition

This site turns the Propaedia into an interactive reading guide. The original knowledge hierarchy is preserved, but every Section can now open four different reading collections:

  • Oxford Very Short Introductions - curated books matched to Sections and outline paths
  • BBC In Our Time - mapped radio episodes and programme pages attached to the same outline
  • Wikipedia Vital Articles - mapped articles from Wikipedia's curated list of essential topics
  • Macropaedia articles - the original Britannica reading references from the print edition

Each Section page also includes the original hierarchical subject outline with cross-references linking related Sections across the ten Parts.

Homepage Modes

The homepage now has three different ways into the outline, each built for a different style of study:

  • Whole Outline - a coverage-first path that suggests the next unread book, article, or episode based on how much new territory it opens across the outline
  • Selected Fields - Adler's circle navigator, where you can centre one Part, compare it with another, and surface bridge readings that connect both
  • Specific Topic - direct browsing by Part, Division, and Section, with essays and recommendation panels tied to the exact outline area you are exploring

In the Selected Fields mode, when one Part is placed at the centre and another is rotated to the top, the panel below the circle shows how they connect through cross-referenced Sections and through bridge readings that were independently recommended in both Parts.

Reading Tools

A reading preference setting in the sidebar stores your default reading type (Oxford VSI, Wikipedia, BBC In Our Time, or Macropaedia) across the outline pages.

Section, Division, and Part pages each have their own reading-type selector inside the recommendation panel, and the homepage has separate selectors in Whole Outline and Selected Fields. Changing any of these selectors updates the shared default, so the same reading type is shown first elsewhere whenever it is available.

A reading checklist lets you mark books, articles, and episodes as done. Progress is saved locally in your browser and reflected across the site - checked items are excluded from homepage paths and bridge recommendations, and checkboxes appear on individual reading pages.

Reading Recommendations

The Oxford VSI, Wikipedia, and BBC In Our Time recommendations are matched to Sections using build-time scoring plus AI-assisted mappings. Macropaedia recommendations come from the original Britannica references already attached to the Propaedia. Each mapped recommendation includes a rationale explaining why it was matched. Clicking an outline item filters the list to that branch of the outline and re-ranks the matched items inside the current selection.

How relevance is calculated

Each book, article, or episode receives a score based on how closely it relates to the Section you are viewing. The scoring gives the strongest weight to direct title matches, then adds weight for subject terms, keywords, mapped outline paths, and overlap with the language used in the Section title and outline.

When you click an outline item to filter the list, the scoring also considers the sub-topics beneath that item. A resource that touches several child topics ranks above one that only matches a narrow fragment of the current branch.

The relevance percentage shown on each card is relative - the best-scoring resource in the current view is always shown as 100%, and the others are scaled proportionally.

How Whole Outline recommendations work

The Whole Outline tool ignores items you have already marked done and asks a coverage-first question: which unread item adds the largest number of currently uncovered Parts, Divisions, Sections, or Subsections for the selected reading type?

After each suggested step, the covered area is updated and the next step is recalculated from what remains. If two items add the same amount of new coverage, the tie is broken in favour of the one whose new coverage is spread more widely across the outline, and then the one attached to more Sections overall.

How Part and Division recommendations are aggregated

The recommendation panels on Part and Division pages are not separate AI runs. They are rollups built from the Section-level matches already attached to the outline.

On a Part page, a reading ranks higher when it appears across more Divisions in that Part, covers more Sections, and spreads that coverage evenly rather than being concentrated in one corner. On a Division page, the same idea is applied one level down: readings are ranked by how many Sections in that Division they cover, with extra weight for broader mapped-path coverage.

The relevance percentage shown on those cards is relative to the strongest item in the current Part or Division list, not a global score across the whole site.

Source Data

  • Oxford Very Short Introductions - 793 titles in the current catalogue fetched from Oxford Academic on March 15, 2026. Where the source metadata includes it, individual book pages show the book's edition and release year.
  • BBC In Our Time - 1091 episodes fetched from BBC programme pages on March 23, 2026.
  • Wikipedia Vital Articles - 998 articles last fetched on March 16, 2026.
  • Macropaedia - original article references preserved from the printed Propaedia/Britannica pairing.

Your Reading Progress

Your Progress

No readings checked off yet.

Your progress is stored in this browser. Export a backup to keep it safe or transfer it to another device.

Technical Details

This site is built with Astro and Preact, uses static site generation for fast loading, and is hosted on GitHub Pages. Search is powered by Pagefind. All data is processed at build time - there are no external API calls or server dependencies at runtime.