ColossiansOne

Counted, not asserted

What the corpus shows once you count it.

A compendium that attributes every claim to a dated text can be asked questions no unsourced demon list can answer. These are three of them. Each figure is computed from the corpus at build time rather than written down, so none of it can drift out of step with the data underneath.

Corpus last revised 29 August 2026.

Finding one

Two of these texts are not two texts

Counting how many figures each pair of source texts has in common should produce a gentle slope. It does not. The Lesser Key of Solomon, Book I: Ars Goetia and Pseudomonarchia Daemonum share 72 entities. The next closest pair in the entire corpus shares 17. A cliff that steep is not two independent witnesses agreeing — it is one list copied and extended.

  1. The Lesser Key of Solomon, Book I: Ars Goetia and Pseudomonarchia Daemonum

    72 shared

  2. The Egyptian Book of the Dead and The Pyramid Texts

    17 shared

  3. Dictionnaire Infernal and Hebrew Bible & Deuterocanon

    15 shared

  4. Hebrew Bible & Deuterocanon and The Baal Cycle (Ugaritic Tablets)

    11 shared

  5. Theogony & Works and Days and The Iliad

    11 shared

  6. Hebrew Bible & Deuterocanon and New Testament

    10 shared

  7. Theogony & Works and Days and The Homeric Hymns

    10 shared

  8. De Coelesti Hierarchia (The Celestial Hierarchy) and Summa Theologiae, Treatise on the Angels

    9 shared

  9. De Coelesti Hierarchia (The Celestial Hierarchy) and Homilies on the Gospels, Homily 34

    9 shared

  10. Summa Theologiae, Treatise on the Angels and Homilies on the Gospels, Homily 34

    9 shared

  11. Hebrew Bible & Deuterocanon and The Catholic Encyclopedia (1913)

    8 shared

  12. The Homeric Hymns and The Iliad

    8 shared

This is the site marking its own homework. The attestation score treats each attesting text as evidence, and already admits in prose that the Solomonic pair is its weakest joint. The chart is what that admission looks like measured: any figure whose score rests on those two texts alone has one source of evidence, not two, and the score currently says otherwise.

Finding two

The distance between a name and the book you know it from

Every profile on this site is ordered oldest-first because the gap is the argument. Measured across the corpus, 72 of the 270 figures attested in more than one dated text are named in texts over a thousand years apart. The 28 longest are below. Asherah runs 3,291 years — from The Baal Cycle (Ugaritic Tablets) (c. 1400-1200 BCE) to Territorial Spirits & Strategic-Level Spiritual Warfare (1989-1996) — and is the same name throughout.

  1. Asherah
    c. 1400-1200 BCE to 1989-1996, 3291 years, across 3 texts
  2. Astaroth
    c. 1400-1200 BCE to 17th c.; ed. 1904, 3204 years, across 6 texts
  3. Nergal
    c. 1300–800 BCE to 1818; illustrated ed. 1863, 2863 years, across 3 texts
  4. Asmoday
    c. 1000-400 BCE to 1913, 2613 years, across 8 texts
  5. Hecate
    c. 700 BCE to 1818; illustrated ed. 1863, 2563 years, across 4 texts
  6. Nephthys
    c. 2350–2100 BCE to c. 100 CE, 2450 years, across 3 texts
  7. Horus
    c. 2350–2100 BCE to c. 100 CE, 2450 years, across 3 texts
  8. Isis
    c. 2350–2100 BCE to c. 100 CE, 2450 years, across 3 texts
  9. Osiris
    c. 2350–2100 BCE to c. 100 CE, 2450 years, across 3 texts
  10. Anubis
    c. 2350–2100 BCE to c. 100 CE, 2450 years, across 3 texts
  11. Set
    c. 2350–2100 BCE to c. 100 CE, 2450 years, across 3 texts
  12. The Lying Spirit
    c. 900–100 BCE to 1998, 2398 years, across 2 texts
  13. The Queen of Heaven
    c. 900–100 BCE to 1989-1996, 2391 years, across 2 texts
  14. The Absalom Spirit
    c. 900–100 BCE to 1989, 2389 years, across 2 texts
  15. The Ahab Spirit
    c. 900–100 BCE to 1989, 2389 years, across 2 texts
  16. The Spirits of Affliction
    c. 900–100 BCE to 1973, 2373 years, across 3 texts
  17. The Familiar Spirit
    c. 900–100 BCE to 1973, 2373 years, across 2 texts
  18. The Leviathan Spirit
    c. 900–100 BCE to 1973, 2373 years, across 2 texts
  19. Alastor
    458 BCE to 1818; illustrated ed. 1863, 2321 years, across 2 texts
  20. Dagon
    c. 900–100 BCE to 1913, 2313 years, across 3 texts
  21. Belial
    c. 900–100 BCE to 1913, 2313 years, across 7 texts
  22. Moloch
    c. 900–100 BCE to 1913, 2313 years, across 4 texts
  23. Beelzebub
    c. 900–100 BCE to 1913, 2313 years, across 6 texts
  24. Satan
    c. 900–100 BCE to 1913, 2313 years, across 5 texts
  25. Lucifer
    c. 900–100 BCE to 1913, 2313 years, across 6 texts
  26. Gabriel
    c. 900–100 BCE to 1913, 2313 years, across 5 texts
  27. Michael
    c. 900–100 BCE to 1913, 2313 years, across 6 texts
  28. Bael
    c. 900–100 BCE to 17th c.; ed. 1904, 2304 years, across 4 texts

earliest text naming the figurelatest

What travels is the name. The rank, the offices and the form are usually the late text’s own invention, which is why a long bar is not evidence of a continuous tradition — it is evidence of a durable word. The full timeline places every text rather than every figure.

Finding three

Which powers every culture invents, and which one genre made up

Each office in the corpus is mapped to one entry, so the same job can be counted across texts and cultures. Ordered by how many of the 22 traditions claim it, 18 offices reach at least three-fifths of them — conducts souls appears in 21, while 3 never leave one or two. The first group is a job description humans keep needing filled. The second is a convention of the genre that wrote it down.

  1. Conducts Souls
    21 trad
  2. Governs Nature
    20 trad
  3. Executes Judgment
    20 trad
  4. Causes Tempests
    20 trad
  5. Causes Sickness
    20 trad
  6. Guards & Wards
    19 trad
  7. Grants Victory in War
    19 trad
  8. Burns & Slays
    18 trad
  9. Deceives & Tempts
    17 trad
  10. Binds Spirits
    17 trad
  11. Commands Legions
    16 trad
  12. Reveals Past, Present & Future
    15 trad
  13. Reveals Divine Secrets
    15 trad
  14. Sows Discord
    15 trad
  15. Changes Shape
    15 trad
  16. Torments
    14 trad
  17. Heals
    14 trad
  18. Records Deeds
    14 trad
  19. Teaches the Liberal Sciences
    12 trad
  20. Teaches Handicrafts
    12 trad
  21. Raises or Speaks with the Dead
    11 trad
  22. Grants Riches
    10 trad
  23. Intercedes
    10 trad
  24. Procures Love
    9 trad
  25. Grants Favour & Dignity
    9 trad
  26. Answers Truly
    9 trad
  27. Bears Messages
    8 trad
  28. Discovers Hidden Things
    7 trad
  29. Transports Swiftly
    6 trad
  30. Declares Thefts
    6 trad
  31. Teaches Astronomy
    5 trad
  32. Teaches Herbs & Stones
    5 trad
  33. Builds & Destroys
    5 trad
  34. Praises Unceasingly
    5 trad
  35. Reveals Treasure
    4 trad
  36. Restores Lost Dignities
    4 trad
  37. Teaches Rhetoric & Logic
    3 trad
  38. Grants Languages
    3 trad
  39. Confers Invisibility
    3 trad
  40. Gives Familiars
    2 trad
  41. Reconciles Enemies
    2 trad
  42. Transmutes Metals
    2 trad

three-fifths of traditions or morea quarter or morenarrower

What these do not show

The corpus is not the record

Everything here is bounded by the 63 texts on this site. A wider corpus would lengthen some spans and add overlaps; it would not shorten or remove any, so every figure is a lower bound.

Shared names are not shared figures

Two texts naming Astaroth are counted as overlapping even though one means a Canaanite goddess and the other a duke of hell. That is the point of the overlap chart, and a limitation of it.

Offices are mapped by hand

Grimoires phrase the same office a dozen ways, and collapsing them into one entry is an editorial judgement. A different mapping would move the middle of the third chart, though not its ends.

How to cite this

ColossiansOne, “Findings: what counting the sources shows”, https://colossiansone.com/findings (corpus revised 29 August 2026).

The underlying texts are public domain. The compilation, the scoring and this analysis are the site’s own. Corrections that identify a misattributed claim or a mis-dated source are the most useful thing anyone can send.