Conventional wisdom usually claims that Athanasius of Alexandria was the first Christian commentator to use the descriptor “canonical” in reference to biblical writings.^[1] As one Bart Ehrman pointed out in an old article, there is a kind of theological appeal to this: Athanasius, the great defender of Christological orthodoxy, also gave shape to the Bible as we know it, the first author to list the twenty-seven-book NT of today. Fundamental Christology gets to go hand-in-hand with our Bible in the grand sweep of historical theology.

Putting aside his precise list of books for a moment—and if one is more interested in this angle, I recommend the superb book by Edmon Gallagher and John Meade on the topic—I want to focus on the mere idea and terminology of “canonical” in reference to the Bible, because I don’t believe Athanasius actually invented this notion. In fact, I think it’s more than a century older.

Athanasius’ preferred term is κανονιζόμενα, “canonized.” And yes, if you go digging through the Greek databases, you won’t find a secure dating before him that uses iterations of that term with respect to the biblical texts. Open and shut, yes? Well, no, as it turns out Origen (d. 253) uses similar language in several of his writings, but these only survive in Latin translation.

Some experts immediately groan at any technical argument depending on Origen’s Latin translation, because his later translation into Latin was hotly disputed in antiquity itself. For example, one of his translators, Rufinus of Aquileia, was accused of massaging Origen at various points to make the polymath’s theology more palpable for a late fourth-century audience. Moreover, if one compares some of Rufinus’ translations to their surviving Greek originals, it does appear that he occasionally made certain “moves” in bringing the original text up-to-date theologically.

Here’s how Rufinus translates a portion of Origen’s commentary on Song of Songs, where Origen consider how Christians should handle apocryphal literature:

It’s above us to pronounce on such things. But it is plain that many [apocryphal] passages were produced and included in the New Testament by both the apostles or the evangelists. We never find these passages in the writings that we consider canonical, yet they are found in apocryphal texts, and the adopted passages are obviously displayed from the same texts. But indeed, no place should be given to apocryphal writings. For “the eternal boundaries should not be moved, which our father have established [Prov. 22:28].^[2]

Most scholars would assume that Rufinus translated anachronistically, inserting later biblical terminology back into Origen’s text. If this were the lone example, that would be reasonable skepticism, given the accusations against Rufinus and given that the language of “canonical” (canonic-) does not appear in Origen’s Greek.

We have, however, a second Latin passage of Origen from a much later, anonymous translator. “We are not unaware,” notes Origen,

that many of the secreta (multa secretorum = πολλά ἀποκρύφων?) have been forged by the impious and those speaking “iniquity in excess.” And indeed, the Ypythians use certain forgeries, but those of Basilides use others. It is fitting, therefore, to consider carefully that we do not receive all of the secreta which are brought forth in the name of the saints because of the Jews, who perhaps forged certain things to the destruction of the truth of our scriptures, confirming false doctrines. . . . Yet because of them who are not able like the money-changers to discern between the words whether they are true or false, and are not able to guard themselves carefully, so that those among you hold to the truth, “yet let them abstain from every evil appearance” [1 Thess. 5:22]. No one ought to use books that are beyond the canonized scriptures for the confirmation of doctrine.^[3]

Here, we have Origen again warning about apocryphal books (secreta) and contrasting them with the “canonized” texts (canonizatae scripturae). That I can find, the Latin form canonizat- is extremely rare. For the entire period 200–735, the Library of Latin Texts only has a single match (two, if we count an instance of canonizatur). By contrast, the same search parameters reveal 550 occurrences of canonic-and cognates. Much like secreta must be a hyper-literalism for ἀπόκρυφα, something similar I argue is play with canonizat-. If the anonymous sixth-century translator were “cleaning up” Origen’s text here, he would most likely have chosen canonic-, (which he in fact does once in a different passage). The “money-changers” analogy from 1 Thess. 5:22 was also a favorite of Origen’s in matters of literary criticism, as was Prov. 22:28 in the previous passage.^[4]

There is also in important ideational through-line: in both passages, Origen draws bright lines around non-canonical books, insisting they cannot have theological authority. In Rufinus’ own original Exposition of the Creed, after listing his biblical canon, he writes,

These are the books that the fathers have included within the canon and from which they intended that we base our declarations of faith. It should be know, however, that there are also other books that are not called “canonical” but “ecclesiastical” by the elders.

Here, Rufinus includes the texts disputed by later Catholics and Protestants that are variously called, “intertestamental,” “deuterocanonical,” or (most unhelpfully) “the Apocrypha.” Rufinus continues, saying again that the “elders” certainly wanted these texts read in church, but they were not admissible evidence for “confirming the authority of faith.”^[5] Let the record show that the later Protestants did not invent their of “the Apocrypha” de novo.

Compare again each of these three passages about non-canonical writings.

• the anonymous translator: nemo uti debet ad confirmationem dogmatum libris, qui sunt extra canonizatas scripturas.
  
• Rufinus’ translation of Origen: dari maioribus non placuit locum nec admitti ad auctoritatem.
  
• Rufinus’ original words: non tamen proferri ad auctoritatem ex his fidei confirmandam.
  

The most logical connection between all three is Origen himself, who was an important influence on Rufinus; only an exceptionally convoluted explanation could work Rufinus’ influence back into the later, anonymous translator. In other words, we seem to be handling Origen’s own ideas here, which I think lends further weight to the translators’ faithfulness when they use the language of “canonical.”

In sum, this would mean that terminology of “canonical” scripture goes back much further than Athanasius. Nor would it be the first time a “later” doctrinal category showed in his work, as has been argued with homoousios.^[6] It also makes me wonder whether Origen himself was the “inventor” or whether he himself was drawing on earlier sources. Certainly, the “canon” was used by earlier sources to refer to the rule of faith. Perhaps someone had already made the conceptual expansion.

1. See most famously, his Festal Letter 39 from February 367. ↑
   

2. Supra nos est pronuntiare de talibus. Illud tamen palam est multa vel ab Apostolis vel ab evangelistis exempla esse prolata et Novo Testamento inserta, quae in his scripturis, quae canonicas habemus, numquam legimus, in apocryphis tamen inveniuntur et evidenter ex ipsis ostenduntur assumpta. Sed nec sic quidem locus apocryphis dandus est; ‘non’ enim ‘transferendi sunt termini aeterni, quos statuerunt patres’ nostri. ↑
   

3. Haec omnia diximus discutientes sermonem, non ignorantes quoniam multa secretorum ficta sunt ab impiis et“iniquitatem in excelsum” loquentibus. Et utuntur quidem quibusdam fictis Ypythiani, aliis autem qui sunt Basilidis. Oportet ergo caute considerare, ut nec omnia secreta quae eruntur in nominee sanctorum suscipiamus propter Iudaeos, qui forte ad destructionem veritatis scripturarum nostrarum quaedam finxerunt confirmantes dogmata falsa, nec omnia abiciamus quae pertinent ad demonstrationem scripturarum nostrarum. Magni ergo viri est audire et adinplere quod dictum est: “omnia probate, quod bonum est tenete.” Tamen propter eos, qui non possunt quasi trapezitae inter verba discernere utrum vera habeantur an falsa, et non possunt semetipsos caute servare, ut verum quidem teneant apud se, “ab omni autem specie mala abstineant.” nemo uti debet ad confirmationem dogmatum libris, qui sunt extra canonizatas scripturas.
   

4. Megan Hale Williams, The Monk and the Book: Jerome and the Making of Christian Scholarship (Chicago: University of Chicago Press, 2006), 237.
   

5. Haec sunt quae patres intra canonem concluserunt et ex quibus fidei nostrae adsertiones constare uoluerunt. Sciendum tamen est quod et alii libri sunt, qui non canonici sed ecclesiastici a maioribus appellati sunt . . . . Quae omnia legi quidem in ecclesiis uoluerunt, non tamen proferri ad auctoritatem ex his fidei confirmandam.
   

6. Ilaria L.E. Ramelli, “Origen’s Anti-Subordinationism and Its Heritage in the Nicene and Cappadocian Line,” Vigiliae Christianae 65, no. 1 (2011): 21–49. ↑