Scriptura
What the Latin Fathers quoted
A first count of the Bible in Migne’s Patrologia Latina, from Tertullian to Innocent III.
1 · The working canon
Each age leaned on different verses
Count the distinct works that cite each verse, using only the citations Migne’s editors identified in the text, and the ages separate cleanly.
To 600
- 119John 1:1In principio erat Verbum
- 102John 1:14Et Verbum caro factum est
- 92John 10:30Ego et Pater unum sumus
- 89John 1:3Omnia per ipsum facta sunt
- 84John 14:6Ego sum via, et veritas, et vita
- 83Phil 2:6qui cum in forma Dei esset
- 74Rom 5:5caritas Dei diffusa est in cordibus nostris
- 73Matt 28:19baptizantes eos in nomine Patris
- 70Gen 1:26Faciamus hominem ad imaginem
- 661 Tim 2:5unus et mediator Dei et hominum
601–1000
- 46John 14:6Ego sum via, et veritas, et vita
- 45Matt 25:34Venite benedicti Patris mei
- 39Matt 11:29discite a me, quia mitis sum
- 34Matt 28:20ecce ego vobiscum sum omnibus diebus
- 34John 1:1In principio erat Verbum
- 34Jas 1:17Omne datum optimum
- 33John 1:14Et Verbum caro factum est
- 31Matt 5:44Diligite inimicos vestros
- 31Matt 28:19baptizantes eos in nomine Patris
- 30Matt 25:41Discedite a me maledicti
1001–1216
- 56John 14:6Ego sum via, et veritas, et vita
- 53Matt 11:29discite a me, quia mitis sum
- 49John 1:1In principio erat Verbum
- 49Phil 3:20Nostra autem conversatio in caelis est
- 47Jas 1:17Omne datum optimum
- 47John 1:14Et Verbum caro factum est
- 42Acts 4:32cor unum et anima una
- 41Gal 6:14Mihi autem absit gloriari, nisi in cruce
- 41Phil 1:23desiderium habens dissolvi
- 40Matt 5:3Beati pauperes spiritu
To 600, the list is Christology and the Trinity: the prologue of John, “I and the Father are one,” the self-emptying of Philippians 2, the one mediator. The controversies of the fourth and fifth centuries are visible in the arithmetic.
601–1000 turns to judgment and the Christian life: the sheep and the goats of Matthew 25, “learn of me, for I am meek,” love of enemies.
1001–1216 reads like a rule of life: “one heart and one soul” (Acts 4:32, the opening of the Rule of Augustine that the canons regular lived by), “our citizenship is in heaven,” “I desire to be dissolved,” “God forbid that I should glory, save in the cross,” “blessed are the poor in spirit.”
2 · The four senses
Where the senses of Scripture are stated
Henri de Lubac’s Exégèse médiévale argued that the Latin Middle Ages read Scripture under four senses: history, allegory, tropology and anagogy. Here the PL is searched for every place that names three or more of them together.
The search finds 269 such places. Two independent codings agree on what kind of passage each is (kappa 0.93), and 132 are statements of theory, not passing lists or applications.
| Earliest statements of all four | d. | PL |
|---|---|---|
| John Cassian, Conferences XIV.8 | 435 | 49, 962 C |
| Eucherius of Lyon, Formulae spiritalis intelligentiae | 449 | 50, 728 B |
| Bede, De schematibus et tropis | 735 | 90, 185 B |
| Bede, De tabernaculo — first in the order history, allegory, tropology, anagogy | 735 | 91, 410 B |
| Amalarius, De ecclesiasticis officiis | 850 | 105, 1035 B |
| Rabanus Maurus, on Exodus | 856 | 108, 147 D |
The figure most often used to teach the senses is Jerusalem: a city of stone, the Church, the soul, the heavenly city, in 28 passages. Next is the four feet of the table in the tabernacle.
3 · Machine and reader
What a count finds, and what only a reader finds
De Lubac read the sources by hand. Setting the machine’s 132 statements against his footnotes shows what each can see.
- He cites 96 of the 132 for exactly this point. For the 11 he does not cite at all, a reader still has to confirm the omission.
- Of his own citations of passages naming three or more senses, the word search finds half. In 98 of the places it misses, relaxing a rule (a wider window, a broader word list) would find them.
- In 64 places, the scheme is carried by a figure, not a word: the three ways of cooking a sacrifice, four casks in a cellar, shadow and image and truth, the three comings of Christ. No vocabulary will ever find these. They are what a reader contributes, and the reason the project keeps a reading layer beside the count.
4 · Which Bible?
The early Fathers quote a different Latin
Before Jerome’s Vulgate, Latin Christians read the Old Latin, and they quoted it. A matcher built on the Vulgate should therefore miss more of the early Fathers, and it does.
| Author died | Verse citations by Migne | Found by Vulgate matching |
|---|---|---|
| to 400 | 22,732 | 48% |
| 401–600 | 67,018 | 58% |
| 601–800 | 15,476 | 70% |
| 801–1000 | 25,592 | 65% |
| 1101–1200 | 48,943 | 66% |
Adding Old Latin texts to the index (Sabatier’s 1743 Versiones antiquae, and the 37 Old Latin manuscripts of John transcribed by the IGNTP) recovers about 4 more points of the early Fathers’ New Testament citations, and 7 on John. Most of the gap remains. Much of it is allusion, not quotation: the early Fathers wove Scripture into their own sentences. They cite differently, not only from a different text.
Method
How this was made
Everything above was computed, not read by a model: the Corpus Corporum transcription of the PL (5,276 works), the Clementine Vulgate, and a citation spine with three streams. The streams are Migne’s parenthetical citations, rare five-word sequences shared with the Vulgate, and ordered word pairs for looser wording. De Lubac’s footnotes were transcribed into a ledger. Only aggregate counts are published here.
These are preliminary findings. Common liturgical formulas (“in the name of the Father, and of the Son…”) are not yet excluded from the matched counts, the Greek Fathers are not yet read, and the 11 uncited passages await a reader. Numbers will move.