Tuesday, January 07, 2020

The Greek Manuscripts of the Comma Johanneum (1 John 5:7–8)

Warning: long post because there’s a lot of detail here.

Although it is one of the easiest text-critical decisions, a lot of attention often goes to the Comma Johanneum (henceforth, CJ), the addition at 1 John 5:7–8 (addition in italics): “For there are three that bear record in heaven, the Father, the Word, and the Holy Ghost: and these three are one. And there are three that bear witness in earth, the Spirit, and the water, and the blood: and these three agree in one. (1 John 5:7-8 KJV).” A lot of people who take a textus receptus position vigorously defend the CJ because of its theological value. This was actually one of my first tastes of the implications of textual criticism—I was doing evangelism with my John MacArthur NKJV Study Bible back when I was much younger and found this amazing verse that ‘proves’ the Trinity, until my friend told me I couldn’t use it because it wasn’t original to 1 John but not to worry because there was ample proof of the Trinity elsewhere in Scripture. As it turns out, a mere appeal to 1 John 5:7–8 without also having a Trinitarian interpretation of the passage does not automatically ‘prove’ the Trinity, because Oneness Pentecostals (=a branch of Pentecostalism that denies the Trinity) who use the KJV also appeal to this passage as a proof of their anti-Trinitarian doctrine—they claim that the phrase “and these three are one“ teaches their “oneness” doctrine (a great example here).

The Tyndale House Greek New Testament gives special treatment to the CJ by breaking its normal pattern of citing very few witnesses, and in the last few days, I decided to examine each of the Greek manuscripts that contain some form of the CJ to learn a bit more about each of them. The following are some of my findings. There are 10 Greek manuscripts that have the CJ, but only three of them have it in the same form as in Stephanus’ 1550 edition and Scrivener’s edition reprinted by the TBS—these three are 221marg, 2318 and 2473. All ten of these manuscripts are indexed for 1 John 5:7–8 at the INTF’s VMR, so you are free to verify them yourself. Regarding scribes and manuscript acquisition histories/provenance, I got all of that information from a combination of catalogues of manuscripts in those libraries (see the short bibliography at the end of this post) and shelfmarks/information given at the online Liste.

Before I get there, I want to mention that sometimes an eleventh Greek manuscript is cited. GA 635 is sometimes cited as having the CJ in the margin, but it does not.

GA 635 
I can’t make out all the words, but I’m seeing το πν(ευμ)α το αγιο(ν) και ο π(ατ)ηρ (και)...the rest is more difficult. αιματος? It’s probably an obvious solution, and I’m happy to update the post if someone has a better image or can make more sense of it. There are a number of notes like this in the margins of 635, and there isn’t room for the whole CJ here anyway, so I didn’t want to spend too much time on it.

I discuss below the 10 manuscripts with the CJ, with the first eight in approximate chronological order of the inclusion of the CJ. I want to mention from the start that I have drawn attention to the Roman Catholic provenance of a few of these. I’ll explain more at the end why I do so, but the short version is that some of the most vehement defenders of the CJ are Protestant textus receptus advocates who subscribe to the Westminster or London Baptist confessions and claim doctrinal purity via affirmation of these confessions, yet in order to defend the CJ by appealing to the Greek manuscripts, they have to appeal to manuscripts from the tradition from which their own tradition broke away, and in some cases, manuscripts that were made (or marginal notes that were added) after that break.

1. GA 629 (1362–1363)

GA 629, f. 105v, images at https://digi.vatlib.it/view/MSS_Ott.gr.298
629 is dated 1362–1363 and is the earliest known Greek manuscript of the CJ. There are earlier manuscripts, than 629 but in those, the CJ is still later, so, say in the year 1400, those manuscripts themselves existed but they did not have the CJ yet, where as in the year 1400, 629 did.

While I haven’t yet figured out where 629 came from, there is a note on f. 2 that identifies it as being the property of “Joannis Angeli Ducis ab Altaemps”. This would be Giovanni Angelo D’Altemps, who, if that Italian Wikipedia article is correct, seems to have inherited at least part of his library from his grandfather, Cardinal Mark Sittich von Hohenems Altemps, who himself was the nephew of Pope Pius IV. Today, 629 is in the Vatican Library as Ottob. gr. 298. That is all to say that as much as I can tell about the original provenance of this manuscript (and the manuscript does pre-date its known owner by more than 200 years), by the time of later editions of the TR, it can be tied to a Roman Catholic family.
GA 629, f. 2r
629 is a Latin-Greek diglot, which should raise suspicions about its witness to the CJ. After all, the contention is that the CJ was translated into Greek from Latin and that’s how it ended up in a few Greek manuscripts, and that looks exactly like what happened here. Notice above that there are no definite articles in the three heavenly witnesses, but Latin does not have definite articles like Greek, so the lack of articles here in Greek suggests that it is a translation from Latin. Also note that this manuscript is not a witness to the form of the CJ as it appears in the TR. In fact, in these 23–25 words, I count 8 differences between the CJ here and the CJ as it appears in the Trinitarian Bible Society’s textus receptus (at least according to the TBT TR on Bibleworks). They are απο του ουρανου, απο της γης, the article(3x), the order of Πνεῦμα Ἅγιον, οὗτοι in the TBT and εἰς τὸ in 629:

ἐν τῷ οὐρανῷ πατήρ λόγος καὶ τὸ Ἅγιον Πνεῦμα καὶ οὗτοι οἱ τρεῖς ἕν εἰσι καὶ τρεῖς εἰσὶν οἱ μαρτυροῦντες ἐν τῇ γῇ

629: απο του ουρανου πατήρ λόγος καὶ Πνεῦμα Ἅγιον καὶ οἱ τρεῖς εἰς τὸ ἕν εἰσιν καὶ τρεῖς εἰσὶν οἱ μαρτυροῦντες απο της γης

In short, 629 appears to be a thoroughly Roman Catholic manuscript that has an unusual form of the CJ as a translation from Latin into Greek, probably (and this is purely my guess at this point) to harmonise the differences between the Latin and Greek columns. An interesting study would be for someone to compare these two columns in more detail to see if they appear to be harmonised.

2. Codex Montfortianus (GA 61)

GA 61, f. 439r, catalogue record here
GA 61 is the infamous manuscript on whose authority Erasmus added the CJ to his third edition of the Greek New Testament (1522). No, Erasmus didn’t promise to include the CJ if someone could give him a Greek manuscript with it, and no, 61 wasn’t made to force Erasmus’ hand.

Still, we can make a few observations about the manuscript—and for all this and more, see Brown, A. J. “Codex 61 (Montfortianus) and 1 John 5,7–8.” and Grantley McDonald, Biblical Criticism in Early Modern Europe: Erasmus, the Johannine Comma and Trinitarian Debate (Bibliographic information at the end of this post).

First, it has a Franciscan provenance. Even if the Franciscan Observant Francis Frowick was merely the earliest known owner of the manuscript and not its copyist, the manuscript was definitely copied by a Franciscan, probably in the window of 1495–1521 (date based on the watermarks of the paper and that Erasmus seems to have known about it by 1521). We know it is a Franciscan production because the copyist wrote “Jesus, Mary, Francis” in it, which is apparently a thing that Franciscans often did (it is on this basis that Ussher identified the manuscript as Franciscan according to Brown [p.44]).
GA 61, f. 198v, “Jesus, Mary, Francis”
Also interestingly, we can tell that 61 is a copy of 326 in the Catholic Epistles. GA 326 doesn’t have the CJ, and there are a number of places where 61 diverges from the text of 326 to ‘Latinise’ the manuscript an introduce Vulgate-derived readings. Of course, that’s more what you would expect from a Western, Latin tradition making Greek manuscripts. Another piece of evidence that the 61 likely reflects a translation from Latin into Greek is that it also lacks definite articles here. I only count four differences between the CJ in 61 and the CJ in the TBT edition: three definite articles and the order of Πνεῦμα Ἅγιον, but four is still a lot for a length of text this size.

ἐν τῷ οὐρανῷ πατήρ λόγος καὶ τὸ Ἅγιον Πνεῦμα καὶ οὗτοι οἱ τρεῖς ἕν εἰσι καὶ τρεῖς εἰσὶν οἱ μαρτυροῦντες ἐν τῇ γῇ

ἐν τῷ οὐρανῷ πατήρ λόγος καὶ Πνεῦμα Ἅγιον καὶ οὗτοι οἱ τρεῖς ἕν εἰσιν καὶ τρεῖς εἰσὶν οἱ μαρτυροῦντες ἐν τῇ γῇ

In short, 61 is a manuscript of Catholic (Franciscan) provenance that has a series of what looks like private owners, suggesting that it was either not made for church use or never made it to church use, copied by a scribe who diverged from his exemplar in order to introduce Latin readings into his text rather than copying what was there in the Greek. (Also, there’s a nice little section on GA 61 and some of its philological marginalia by co-blogger Peter Malik in Myths and Mistakes in New Testament Textual Criticism).

3. 429marg (date: after 1522)

GA 429marg, f. 177r

GA 429 is itself 14th century, but the marginal addition of the CJ happened after 1522. We know that because it was copied from Erasmus’ third edition. It contains the exact form of the CJ as in Erasmus’ third edition (without articles; UPDATE: also note that there are no nomina sacra here), but Erasmus added definite articles in his 4th edition and kept them in his 5th. GA 429 is also a member of the Harklean group, but it’s the only member of its group to contain the CJ. If a single manuscript diverges from all the other members of its subgroup, it’s a more than reasonable assumption that it is the one that changed the text. More than that, 429 has a series of notes in it, some of which explicitly cite Erasmus as the source of the note, including one on the facing page.
GA 429, f. 176v. It’s blurry, but you can see “Eras” there.
For more on this manuscript, see Klaus Wachtel’s monograph Der byzantinische Text (full title and bibliographic info at the end of this post). In short, the marginal note in 429 was copied from Erasmus’ third edition and is an addition that causes 429 to diverge here from the group of manuscripts to which it belongs. 429marg is not a witness to a pre-Erasmian CJ.

4. 918 (date: probably 1573–1578)

GA 918, f. 390r
GA 918 is another manuscript with a Catholic provenance. It is one of three manuscripts (the other two are 61 and 429marg) that have a CJ in exactly the same form as Erasmus’ third edition, so Erasmus is likely the source of the CJ here. If that were not enough, Wachtel (Der byzantinische Text) places 918 in “Group 453”. It is the youngest member of that group by a couple of centuries and the only member to have the CJ, so per my remarks above, it’s beyond a reasonable doubt that the CJ is an addition derived from Erasmus. We might have expected the Complutensian Polyglot here given its Spanish Catholic provenance (see below), but there are a number of differences between Erasmus’ third edition and the Complutensian Polyglot in the CJ, and 918 matches Erasmus perfectly. Consequently, 918 is not a witness to a pre-Erasmian CJ.

We can say a bit more about 918. We know that the scribe is one Nicolás de la Torre (sometimes written as Νικόλαος Τουρριανός; see the reference in Ernst Gamillscheg and Dieter Harlfinger, Repertorium der griechischen Kopisten 800–1600). Nikolaos was born in Crete, but he worked for Philip II of Spain at the Library of El Escorial beginning in 1573. Nikolaos’ dated manuscripts have a range of 1562–1586, but given the ties to El Escorial (where the manuscript remains to this day), and given what we know about Nikolaos employment there and travels elsewhere, I would expect that he made the manuscript between 1573 and 1578.

Conclusion: GA 918 is a manuscript of Spanish Catholic provenance from the 1570s that broke from its textual tradition by adding the CJ from Erasmus’ third edition.

5. 2473 (1634)

GA 2473, f. 301r
I haven’t yet found much about 2473 except that it’s dated 1634, and Wachtel considers it to be a copy of one of the later editions of the textus receptus based on its textual affinity (Der byzantinische Text, 320). Still, the King James Version already existed by the time this manuscript rolled around.

6. 2318 (1700s)

GA 2318, f. 394r
GA 2318, f. 394v
GA 2398 is the only manuscript for which the CJ spans a page turn. Palaeographically, it dates to the 1700s. It came to the Romanian Academy Library from the “Central Seminary”, but I haven’t tracked down more than that yet. It is a commentary manuscript, and the first page (which is Romans) identifies the commentary as that of Oecumenius. Of course, the identification of the actual author of these sorts of things is not always easy. If it is Oecumenius, he seems to have advocated the impeccability of Mary, but I haven’t checked to see if this manuscript advocates that. Wachtel thinks the biblical text is copied from one of the later editions of the textus receptus (Der byzantinische Text). At any rate, it seems to have a clearly non-Protestant provenance in the 1700s.

7. 177marg (c. 1785)

GA 177, f. 174r
GA 177 is fun. The manuscript dates to the 11th century, but the CJ dates to 1785. Notice, it’s introduced by a chapter and verse reference [EDIT: I was interpreting the "v" as a Roman numeral V indicating the chapter, but now that I think about it, it could be an abbreviation for versus. Either way, it clearly designates the marginal addition as verse 7, and the verse number is the later feature anyway], which firmly places it later than 1551, when Stephanus first introduced the modern chapter/verse system.

The fun part is that we know who wrote the note and when:
GA 177, f. 224v
The ink is different, and some of the letters are a little stylized, but this looks to me to be clearly the same hand. A lot of the letters are distinctively the same (ρ, ς especially), and neither sample of handwriting has very many accents/breathing marks. The author of the note is Ignatius Hardt (1749–1811), priest of the city of Munich, in the year 1785, month of June, the 20th.

Hardt also published catalogues of manuscripts in Bavaria (see one of them here). I found one reference to him as a Weltpriester, and he worked in the Electoral Court Library in Munich. As far as I can tell, those two elements point to a Catholic provenance. The Electoral Court Library appears to have acquired a large collection from the Jesuits in 1773, so maybe this was one of those? I haven’t looked that far into it, since the name, date and verse number of the CJ here more or less disqualify it as a genuine Greek New Testament witness to be taken seriously when establishing the original text. GA 177marg is not mentioned in the Text und Textwert volumes, though 177 is cited as one of the 500 or so manuscripts that do not have the CJ. This is pure conjecture on my part, but I imagine it’s because the chapter and verse numbers strongly suggest that it was not copied from any ancient Greek manuscript tradition but loosely and from a printed source.

I should also note that the form of the CJ here is unique among manuscripts. Of the editions I checked (admittedly not many), it’s closest to the Complutensian Polyglot. It doesn’t have v.8 at all. I give a comparison below of 177marg and the Trinitarian Bible Society’s textus receptus, with differences underlined:

ἐν τῷ οὐρανῷ πατήρ  λόγος καὶ τὸ Ἅγιον Πνεῦμα καὶ οὗτοι οἱ τρεῖς ἕν εἰσι καὶ τρεῖς εἰσὶν οἱ μαρτυροῦντες ἐν τῇ γῇ

ἐν οὐρανῷ πατήρ λόγος καὶ Πνεῦμα Ἅγιον καὶ οἱ τρεῖς εἰς τὸ ἕν εἰσιν

Although I can’t say much about the provenance of the manuscript, the provenance of the CJ here is a Catholic priest in Munich making the note so recently that by the time he did it, the United States already existed as a country.

8. 221marg (after c. 1850?)

GA 221, f. 150r
GA 221 was also a complete surprise when I looked into it. It is the oldest manuscript (10th century) with the youngest CJ. It’s in Oxford, and Klaus Wachtel (Der byzantinische Text, 319–320) observed that Henry Coxe’s catalogue of manuscripts printed in 1854 explicitly states that this manuscript lacks 1 John 5:7.

Coxe, Henry. Catalogi codicum manuscriptorum bibliothecae Bodleianae: Pars tertia. Oxford: Typographeo Academico, 1854, p. 100
The implication of the note that 221 explicitly lacks the CJ is clear: the CJ looks to be added after that catalogue was prepared. The one published in 1854. I mean, for all I know Charles Spurgeon was alive in time to see this manuscript before somebody added in the CJ.
There are two more manuscripts with the CJ in the margins, but I couldn’t place the dates accurately other than that the CJ is added in later hands.

9. 88marg

GA 88, f. 54r
GA 88 itself is 12th century, but the hand of the note is later. We know that the manuscript came to the Biblioteca Nazionale in Naples (where it is Ms. II. A. 7) from the “Bibliotheca Farnesianae,” which would be the collection of Cardinal Alessandro Farnese (not to be confused with another A. Farnes relevant to manuscript studies), would would become Pope Paul III (note: see Counter-Reformation). There is a Latin note on f. 2r of the manuscript that notes that the CJ is omitted and uses chapter/verse numbers, though it does not prove when the CJ was added. Still, it would not surprise me if the addition post-dates printed editions. The only ligature is κ(αί), and the letters are otherwise cleanly separated. I might expect more ligatures from a native Greek hand of a pre-printing press era. Also, the α in the marginal addition looks to me far more like a Latin a than a Greek α, but I could be wrong on that. More than that, there are interestingly no nomina sacra used, which is highly unusual except in a printed book. Though I can’t say much about its 12th-century provenance, 88 does have strong ties to Counter-Reformation Roman Catholicism by the time the CJ seems to have been written in it.

GA 88, f. 2r

10. 636marg

GA 636, f. 74r
GA 636 is a 15th-century manuscript, but again the CJ looks like a later hand. It lacks the definite articles, so it could be an independent translation of the Latin. It’s textually similar to Erasmus’ third edition except that it lacks οὗτοι, which Erasmus has. We know that the manuscript came to the Biblioteca Nazionale in Naples (where it is Ms. II. A. 9) from an Augustinian Monastery associated with what is today the Church of San Giovanni a Carbonara in Naples. That seems to give the manuscript itself a 15th-century Roman Catholic provenance, though it is unclear when the marginal addition was written. It does have numerous Greek and Latin notes, both in the margins and interlinearly, and an interesting study might be for someone to go through 636 and try to identify the source of these notes.

GA 636, sample of notes, f. 4r


I wanted to give a survey of the manuscripts here that goes slightly beyond merely mentioning them. If you’re just dealing with numbers and vague generalities, it’s really easy to lose sight of the significance of what each manuscript is. To someone who doesn’t know how to evaluate evidence, 10 manuscripts of the CJ might look like it’s even more or better/stronger evidence than places where modern editions go with a reading that has fewer manuscripts in support, but when you are so focused on a big idea that you neglect to look at the evidence you are claiming to support your idea at specific points, you don’t see things like what I have pointed out here. More than that, a lot of people don’t want to just “trust the scholars”. That’s another discussion, but I wanted to lift the curtain a bit and show why the scholars can arrive at some of the conclusions we do. I’ve tried to walk through in a few cases why I think these are the reasonable observations about the manuscripts and from those, the reasonable conclusions about where the data leads (note: yes, data is a Latin neuter plural, but let’s not forget that in both Greek and Latin, a neuter plural subject can take a singular verb).

Maybe I have been reading too much from textus receptus advocates, but it struck me that some of the arguments I hear from them actually work against the textus receptus position once you take the time to step away from the grand claims and look at how the specifics about manuscripts fit in with those grand claims. I often hear from people who want to do away with modern textual criticism that the textus receptus is based on manuscripts with known provenance (and I think they mean “approved texts used by the Church”) whereas the Oxyrhynchus papyri were discarded as rubbish, etc. (Yet I don’t think I’ve ever seen any textus receptus advocate making this claim acknowledge the lectionary markings in Codex Bezae, which are as sure indicators of church use as anything.) In my experience, some of the manuscripts whose provenance we do know are the ones that should be most quickly rejected by people whose position on the text is derived from a particular reading of the Westminster Confession or London Baptist Confession. This is why I made all the references to Roman Catholic provenance in this post. If Protestants (specifically, those who actively align themselves with the Protestant Reformation and the Puritans and claim to have the correct view of the text based on confessional statements made in the 1600s) are citing the Greek manuscript evidence for the CJ, they are appealing to manuscripts produced by, owned by, and used by those whom their own theological predecessors rigorously opposed.

To bring the ‘Evangelical’ back into Evangelical Textual Criticism, one of the most liberating things to me is knowing that Jesus is powerful enough to save me—to grant forgiveness of my sins and reconcile me to God, and God’s promises to preserve his word are so strong that I am not powerful enough to screw it up. As a result, I don’t have to scramble to explain away things that are difficult for my position in order to try to preserve the integrity of my position, because that would be exhausting. When it comes to manuscripts, I can let the chips fall where they may because Jesus is king (and they’re his chips, after all!). As my dad always says, “The truth don’t hurt unless it ought to”. I can take rest and comfort in the knowledge that I must do my best with what I’ve been given, and if I make a mistake, well I’ve been wrong before (just ask anyone who was at my 2016 SBL paper! [EDIT: or ask the friend who found a typo in this very sentence before I updated the post just now to fix it]), and I’m sure I’ll be wrong again, but no matter what, Christ is enough, and I am not powerful enough to thwart his purposes.

A few of the catalogues and sources I used (in case anyone wants to double-check my facts):

Aland, Kurt, and Barbara Aland, eds. Text und Textwert der griechischen Handschriften des Neuen Testaments: IV. Die synoptischen Evangelien: 2. Das Markusevangelium: 1:2. Resultate der Kollation und Hauptliste. ANTF 27. Berlin: De Gruyter, 1998.

Brown, A. J. “Codex 61 (Montfortianus) and 1 John 5,7–8.” In Novum Testamentum Ab Erasmo Recognitum, IV, Epistolae Apostolicae (Secunda Pars) et Apocalypsis Iohannis, edited by A. J. Brown, 30–111. ASD, VI–4. Leiden: Brill, 2012.

Coxe, Henry. Catalogi codicum manuscriptorum bibliothecae Bodleianae: Pars tertia. Oxford: Typographeo Academico, 1854.

Gamillscheg, Ernst, and Dieter Harlfinger. Repertorium der griechischen Kopisten 800–1600: 1. Handschriften aus Bibliotheken Grossbritanniens: A. Verzeichnis der Kopisten. Veröffentlichungen der Kommission für Byzantinistik, III/1 A. Vienna: Österreichischen Akademie der Wissenschaften, 1981.

Litzica, Constantin. Biblioteca Academiei Románe: Catalogul Manuscriptelor Greceşti. Bucharest: Instit. de Arte Grafice “Carol Göbl” S-r I. St. Rasidescu, 1909.

McDonald, Grantley. Biblical Criticism in Early Modern Europe: Erasmus, the Johannine Comma and Trinitarian Debate. Cambridge: Cambridge University Press, 2016.

Mioni, Elpidius. Catalogus codicum graecorum Bibliothecae Nationalis Neapolitanae: Volumen 1, 1. Rome: Istituto Poligrafico dello Stato, 1992.

Pierleoni, Ginus. Catalogus codicum graecorum Bibliothecae Nationalis Neapolitanae: Volumen 1. Rome: Istituto Poligrafico dello Stato, 1962.

Wachtel, Klaus. Der byzantinische Text der katholischen Briefe: eine Untersuchung zur Entstehung der Koine des Neuen Testaments. ANTF 24. Berlin: de Gruyter, 1995.


  1. Great write-up Elijah! I especially enjoyed your personal note at the end.

  2. “in both Greek and Latin, a neuter plural subject can take a singular verb.” But your verb is in ENGLISH!

    1. Don't forget that my native dialect is one that gets subtitles on American reality TV. I'm just saying I have good Classical precedent for it.

  3. Thanks Elijah,

    The only evidence for the CJ of any consequence here is ms.629. The rest could just as easily be looked upon as testimony against the reading generally speaking (i.e. post Erasmus mss. and/or later marginal insertions are of little weight here, in fact they are very telling). The primary rebuttal/plea is usually an overestimated evaluation of the Patristic and versional evidences, coupled with a personal preference for the KJV and the more prominent TR editions; the combination of which screams of a blatant confirmation bias.

    What I think most TR advocates miss here is the fact that the omission of the CJ is *also* contained within the TR tradition (e.g. Erasmus ed. 1,2). Why not follow those TR editions here and stay on the solid ground--and overwhelming weight of the all but universal Greek manuscript tradition?

    1. And of course there are other examples out there where the reading of the Textus Receptus has no Greek support whatsoever. That is, Greek manuscripts containing the TR reading that pre-date Erasmus' first edition. Revelation 17:16 is but one example.

  4. When you say a CJ "matches exactly" are you disregarding the moveable nu in εἰσὶν?

  5. You're a gifted blogger. I have joined your feed and anticipate looking for a greater amount of your awesome post. Additionally, I have shared your site in my informal communities!

  6. Looking at 636, there's some sort of erased note in the middle of the inside front cover (image 40 on the NTVMR) – this might help if only we had better images.