Skip to content

The structure of a classification

Why nine Types, not "fiction vs non-fiction"

It seems intuitive to divide books into "fiction" and "non-fiction". In practice the dichotomy is insufficient and generates edge cases that a classification system must handle explicitly:

Work Fiction? Non-fiction? The problem
Memoir uses narrative arc, characters based on real events Hybrid — belongs to neither side cleanly.
Poetry can be narrative, lyrical, or essayistic in verse — Does not coincide with either macro-category.
Theatre a creative work (script) structured as dialogue, not prose Structurally distinct from both Fiction and Poetry.
Sacred Text — not essayistic in the secular sense A religious foundation; lookup, not linear reading.
Comics narrative, but told through sequential art — The primary medium is visual, not textual.
Catalogue — neither narrative nor essayistic An inventory, a structured list.
Textbook — an essay with a didactic purpose Folded into Essay as a modifier, not its own type.

This is why the system defines a closed list of 9 Types that capture the real modes in which editorial content presents itself. The Type is the first and most stable classification level — it describes what the book is as an editorial object, not what it is about:

  1. Fiction — prose narrative
  2. Poetry — verse
  3. Theatre — stage scripts
  4. Comics — sequential art
  5. Essay — analysis, criticism, theory
  6. Memoir — personal memory with narrative style
  7. Manual — technical-operational instructions
  8. Travel — itineraries and place description
  9. Reference — point lookup (encyclopedia, dictionary, sacred text)

The full definition of each Type is in the ontology reference.

The list is closed: new Types are not added casually. If a work does not fit, the answer is a modifier or a new genre, not a new Type. This stability is what makes the Type a reliable first-level partition and a translatable constant.

The hierarchy of levels

A classification is a hierarchy of levels:

  • Type — mandatory. One of the 9 closed Types.
  • Genre — mandatory. The discipline or subject (Mystery, Physics, Music).
  • Additional levels — optional, added when the content requires further specification (Mystery/Investigation, Music/Jazz, Essay/Literature/History/Italy).

A genre can also serve as a subgenre and vice versa: Fiction/Mystery/History (a story set in the past) versus Essay/Literature/Mystery/History (an essay on the history of the mystery genre). The position in the hierarchy, not the term itself, determines the meaning.

Dynamic modifiers

Some levels are not fixed genres but placeholders the classifier fills at classification time. They are slots that adapt to the content:

Modifier Example use Filled with
<Cuisine> Manual/Cooking/Recipe/<Cuisine> Country, regional style, dietary type, or cooking type
<Faith> Reference/Religion/<Faith> A religion name (Christianity, Islam, …)
<Continent> Travel/<Continent> One of 6 continents
<Country> Essay/Literature/History/<Country> A sovereign state name
<Argument> Essay/<Argument> An Art, a Discipline, or an AllowedType
<Sport> Manual/Sport/<Sport> A sport name
<Animal> Manual/<Animal>/Training A species common name
<Instrument> Manual/Music/<Instrument> A musical instrument name
<Language> Manual/<Language> A natural language name
<ProgrammingLanguage> Manual/Computer Science/<ProgrammingLanguage> A programming language name
<Culture> Essay/Mythology/<Culture> An ancient/distinct cultural tradition

Some modifiers draw from a fixed vocabulary (continents: 6 values; faiths: 12 values). Others are open (countries, sports, animals, instruments) — the classifier supplies the value, and the system accepts it. This distinction matters for the data model: fixed vocabularies can be validated and translated as a closed set; open ones are free-form tags that get translated opportunistically. See the ontology reference for the full vocabularies and ADR 0018 for how this maps to storage.

The structure described here is governed by three universal rules — Autonomy, Substantivization, and Instrumental disambiguation — which constrain how levels are named and disambiguated.