Observatory
What this harvest holds, and what it cannot tell you. Every figure below is a count of one catalogue on one day. It measures how a library indexes Origen; it does not measure the field.
One population, on every figure here and in the Explorer alike. A count on this site is a count of the 1 402 records where Origen is the subject or holds a section of the argument. Of the 1 632 harvested, 225 mention him only and 5 are held outside the count: they stay in the index, they answer a search, and they enter no figure. The three sets are drawn below before anything is counted.
Loading the harvest…
Three sets, kept apart
A count is readable only when it says what it counts. Each record is classed against the vocabulary, and no class is deleted: the noise stays in the index, where it can be measured instead of hidden.
What the third set holds. The rule of the first wave filed a record as not about Origen whenever its metadata did not name him. On a catalogue where librarians have already attached every record to Origen's authority record, that rule was too severe: it swept in studies where he is one witness among several, and the figure it produced had to be read as a ceiling on the noise rather than as a measurement of it. The instructions were rewritten and the harvest classed again. On that curated perimeter the class now floors at mentioned only, and not about Origen is kept for what it was meant for. The third set now holds 5 records where the first wave held 223, and the second holds 225 where it held 9. A record of that second set is retrieved by a search, listed under the count, and added to no figure. Of the 5 held aside, 5 are classed: 4 are editions of texts by Origen catalogued as studies of him and 1 is a title where the word is not the name. None is held aside for want of a tag: a record the deduplication left behind goes back to the tagger, and the build refuses to publish while one is missing. The Method page says what that costs.
Scope of the harvest
One source, one day. The editions of Origen's own works are set aside: they are primary sources, not scholarship about him.
Publication over time
By decade of publication, stacked by language, on the counted records. Those without a year, and the handful printed before 1900, are not in this series; the note under the chart says how many. The figures are given as a table under the columns.
Records per decade
The figures behind this chart
Languages and formats
Language and format as coded in the catalogue record, not as inferred from the title. Counted on the same 1 402 records as every other figure, which is why these totals run below the size of the harvest.
Language of publication
Format
Themes
Sixteen domains of the controlled vocabulary. A publication may carry up to five theme leaves; it is counted here under the first one, the one that files it on the map, and it remains findable under the others.
Counted on the same 1 402 records as the figures above: a record that merely mentions Origen thickens no theme, no work and no angle, and is listed rather than counted wherever the Explorer returns it.
Domain of the leading theme
Works of Origen
Works named in the record or clearly implied by it. Most publications name none: they treat a doctrine, a reader, a milieu.
Works most often studied
Angle of approach
What these figures cannot say
Stated with the numbers, not after them.
- One catalogue. IxTheo is a theological index built in Tübingen. It reads German and European theology closely and other regions less closely. A count of its records is a count of its indexing policy.
- Uneven indexing. Subject headings are richer on recent records, which pushes every subject count toward the well-indexed decades.
- The 2020s are open. The last column stops in mid-2026 and will keep filling.
- Automatic classing. Themes, works, angles and relevance are assigned by a program against a closed vocabulary, checked against a set of records tagged by hand. The rate of records flagged for review is published above; where the agreement figure stands, and why it carries no number this month, is on the Method page.
- No reading or download figures anywhere. No source holds them reliably, and an invented one would be worse than none.