Who First Identified that Beza's Conjecture at Revelation 16:5 was a conjecture?

 A reader recently emailed with a question, which I am paraphrasing in the following way:

Who is the first person to claim that Beza was engaging in conjectural emendation by replacing hosios with esomenos? 

This is a difficult question. 

As early as 1764, in Anthony Purver's "A New and Literal Translation of the All the Books of the Old and New Testament; with Notes Critical and Explanatory," vol. II, p. 332, we find the following:

"Ver. 5 holy] shalt be only on Beza's slender Authority, that being not in all the various Readings of Mills, Kuster, Wetsten, Etc. though Doddr. says other Copies have it so, which however he does not follow. By Wesl. the Gracious one, but wrong.

In 1861, William Webster's "The Greek Testament with Notes Grammatical and Exegetical," vol. II, p. 816 states:

A.V. has ' and shalt be ; ' reading ὁ ἐσόμενος from Beza on insufficient authority 

Saying, "Beza's slender authority" and "on insufficient authority" are not quite fully saying that the conjecture was a conjecture.  The reason for hesitancy was, of course, that Beza's own annotations most naturally indicate that Beza had amended in some way based on a manuscript.

In 1881, "The Holy Bible ... with Explanatory and Critical Commentary and a Revision of the Translation ..." (Cook, ed.) (p. 463)(similarly, in the 1890 printing, p. 463), states:

The third edition of Stephens (A.D. 1550) was the basis both of the editions of Beza (Geneva, 1559, 1565, 1582, 1589, 1598,-- see Scrivener l. c. p. 390) and of the Textus Receptus. Beza's edition of 1589² (or 1598) was taken as the basis of our Authorized Version, by the Translators of 1611:  and thus the English Version of the Apocalypse represents a Greek text which does not rest upon the same authority as that of the other Books of the New Testament³ E.g. in ch. xvi. 5, the conjectural reading of Beza's last three editions (ἐσόμενος for ὅσιος, which rests on no authority whatever) is still represented in the words of the Authorized Version "and shalt be." 

This is one of the earliest English-language references to Beza's conjecture, as such, but the study of the New Testament is not limited to English. 

In 1833, Commentationes Latinae, Tertiae Classis, Instituti Regii Belgici, Volumen Quartum, (Disputation with Valckenaerii, p. 17) says:

c. XVI. 5. Καὶ Ὁ ἐσόμενος Ende Die zijn fal. Complutenses, ERASMUS et STEPHANUS, ediderant, quod omnes agnoscunt versiones, et codices Gr. fere omnes: καὶ ὁ ὅσιος, estque ea lectio, ab ELZEVIRIIS recepta, huic loco valde apta. Recte igitur eam, in I. et II. editione, secutus fuerat BEZA: sed in III. IV. et V. eius in locum substituit lectionem a PISCATORE et BELGIS praelatam, hac addita annotatione: Legitur vulgo καὶ ὁ ὅσιος, ostendente articulo, praeter omnem loquendi morem, depravatam esse scripturam. Vulgata vero, sive articulum legit, seu non legit, nihilo rectius vertit ὅσιος, sanctus, male extrita particula καὶ, prorsus necessaria, ut δίκαιος et ὅσιος connectantur. Sed quum Johannes reliquis omnibus locis, ubi Jehovae nomen explicat, sicuti diximus supra 1. 4. addere consueverit tertium, nempe καὶ ̔Ὁ ἐρχόμενος, cur istud h. 1. praeteriisset? Itaque ambigere non possum, quin germana sit scriptura, quam ex vetusto bonae fidei manus. codice restitui, nempe Ὁ ἐσόμενος. 

Haec BEZA, quibus vero hodie nemini persuadebitur. Nec persuasum fuit ROLANDO, qui Graece scripsit: καὶ ὁ ὅσιος, Belg. en die heylige. Eamque germanam lectionem, in veteri quoque belg. versione expressam, in notis commemorant BELGAE. Intelligitur autem e BEZAE annotatione, alteram lectionem, unius codicis auctoritate, non autem, ut VALCKENAERIO visum fuit, (p. 334.) nuda coniectura assumtam esse a nostratibus. Cf. ELUIT, p. 703.

The reference to "Valckenaerio" seems to refer to Lodewijk Caspar Valckenaer, a Dutch classicist (1715-1785).  I could not immediately locate the specific printed edition (or maybe not even the correct work), but I did find the same point expressed here:

Apocal. c. XVI, 5, in istis, ὁ ὢν, καὶ ὁ ἦν, καὶ ὁ ὅσιος⋅ legendum καὶ ὁ ἐσόμενος, pro ὅσιος, suspicati sunt Piscator et Beza: vera suspicio visa fuit Belgis Interpretibus; dederunt enim: die is, ende die was, ende die zyn sal. Suam, ut puto, coniecturam expresserunt eiusdem 

Ludovici Caspari Valckenaerii Opuscula philologica, critica, oratoria nunc primum coniunctim edita ; accedunt indices · Volume 2, (p. 241)

Update: I found what I believe to be the actually cited work, which as the same statements: (p. 334)

Tiberus Hemesterhusii et L.C. Valckenari Orationes (1784)

I'm not sure who literally is the first who came out and said that he believed it to be a conjecture, in so many words. However, I don't claim to have done an exhaustive research of the topic.

*** 

Same Day Update:

The "NT Conjectures" page notes a variety of historic comments about Beza's conjecture (link to page). That list does not (at the time of this post) mention Valckenaer, but it does mention  Johann Christoph Wolf, Curae 4 (11735), 566–567, which says (in 1735):

καὶ ὁ ἐσόμενος Beza: ex vetusto, ut ait, bonae fidei ms. codice. nec tamen nihil dedisse coniecturae videtur. cur enim in nullo alio Apocalypseos loco illum codicem annotationes Bezae memorant? aut cur Rob. Stephani margo nil hoc loco variat? Bezam tamen sequitur H. Stephanus, Pareus, Piscator, Amama: ac plenius. adhibita parenthesi, (καὶ ὁ ἐσόμενος,) καὶ ὅσιος E. Schmidii textus habet. sed Bezam reprehendit Marckius ad h. l. et gravius Zeltnerus in Diss. de Evangelio Tetragrammato p. 29.

καὶ ὁ ἐσόμενος  Beza: from an ancient, as he says, manuscript of good faith. Nor, however, does he seem to have contributed nothing from conjecture. For why, in no other place of the Apocalypse, do Beza’s annotations mention that codex? Or why does Robert Stephanus’ margin vary nothing in this place? Yet H. Stephanus, Pareus, Piscator, Amama follow Beza: and more fully, with the parenthesis applied, (καὶ ὁ ἐσόμενος,) καὶ ὅσιος the text of E. Schmid has. But Marckius criticizes Beza at this place, and more seriously Zeltner in the Dissertation on the Tetragrammaton Gospel, p. 29.

So, this is actually an earlier example than what I provided in the original post. 

Also we find Adriaan Kluit, Vindiciae 1.3 (1771), 703–704 , which says (in 1771):

De Recepta Lectio heeft: δίκαιος, Κύριε, εἶ, ὁ ὤν, καὶ ὁ ἦν καὶ Ὁ ὍΣΙΟΣ—. Gij zijt rechtveerdig, heere, die is, en die was, en de heilige, dat gij dit geoordeeld hebt. Maar de Ned. Overz. — die is en die was en die zijn zal, dat gij — Kantteekening: “Andere lesen, ende de Heylige. —” Doch deze aanteekening voldoet geenzins. Want niet alleen, dat andere lezen de heilige, maar ’t is opmerkelijk: daar is noch nooit een eenig MS. vertoont, waarin dit onzer Nederd. Overz, die zijn zal, ὁ ἐσόμενος, te vinden is. En niet alleen dit, maar zelfs alle Uitgaven, van eerst tot lest hebben ὅσιος, heilige, ’t is alleen beza, die in Edit. 3–5. zich de vrijheid aangematigd heeft om hier ὁ ἐσόμενος te zetten, gelijk Wolff h. l. toont; maar de Recepta Lectio behoudt noch heden bij allen ὁ ὅσιος. De meeste MSS. leiden ons intusschen daarhenen, om hier te lezen eenvoudig: Δίκαιος εἶ ὁ ὢν καὶ ὁ ἦν ὅσιος, ὅτι ταῦτα ἔκρινας· of ὁ ἦν, ὁ ὅσιος, gelijk H. I, 8.

“The Received Reading has: δίκαιος, Κύριε, εἶ, ὁ ὤν, καὶ ὁ ἦν καὶ ὁ ὅσιος—. ‘Thou art righteous, Lord, who is, and who was, and the Holy One, because thou hast judged these things.’ But the Dutch Version: — ‘who is and who was and who shall be, because thou—’ Marginal note: ‘Others read, and the Holy One.—’ Yet this note by no means suffices. For not only that others read ‘the Holy One,’ but it is remarkable: there has never yet been produced a single manuscript in which this, of our Dutch Version, ‘who shall be,’ ὁ ἐσόμενος, is to be found. And not only this, but even all editions, from first to last, have ὅσιος, ‘holy’; it is only Beza, who in Editions 3–5 has assumed the liberty to put here ὁ ἐσόμενος, as Wolff in this place shows; but the Received Reading still retains with all, ὁ ὅσιος. The majority of manuscripts, meanwhile, lead us thither, to read here simply: Δίκαιος εἶ ὁ ὢν καὶ ὁ ἦν ὅσιος, ὅτι ταῦτα ἔκρινας· or, ὁ ἦν, ὁ ὅσιος, just as [in] I, 8.”

This comes very close to saying that it is a conjecture, and clearly endorses Wolf, who does say that.

(Thanks to Nick Sayers, who reminded me about the NT Conjectures page.)

Jan Krans on Beza's Emendation of Revelation 16:5

Dr. Jan Krans, who has provided the definitive treatment of conjectural emendation by Erasmus and  Beza, provided an article on conjectural emendation as it relates to Revelation in the Editio Critica Maior of Revelation (VI/3.1, pp. 419-20).  In this article, Dr. Krans addresses Beza and Revelation 16:5.  The following is my own English translation of Krans' German original (the translation has not been approved in any way by Dr. Krans, and I have omitted the two footnotes for this section of the text):

A fairly famous case that I did not mention in my 2006 doctoral thesis is Revelation 16:5. Here, in his third edition (1582), Beza replaced ὁ ὅσιος with ὁ ἐσόμενος, even in his Greek text and in his translation, not just in his Annotations. At the time, I thought that the reading ὁ ἐσόμενος was not a conjecture, because Beza himself said that he knew of an old manuscript with this reading. But now I know that in all probability such a manuscript does not exist and never has existed. As we will see, it is quite possible that Beza deceived himself as well as everyone else. Let us first take a closer look at Beza's remark:

"And he will be, καὶ ὁ ἐσόμενος. Usually καὶ ὁ ὅσιος is read, but the article is contrary to all rules of language and shows that the reading is corrupt. The Vulgate, whether it reads the article or not, has by no means a more correct "Sanctus" ("holy"), ὅσιος, whereby the particle καί is wrongly omitted, although it is necessary to connect δίκαιος and ὅσιος. But as we said above on Rev. 1:4, John was accustomed to add a third element in all other places where he extended the name of Jehovah, namely ὁ ἐρχόμενος. Why then would he have omitted this in this place? Therefore I cannot doubt that the original reading is the one which I have restored on the basis of an ancient and reliable manuscript, namely ὁ ἐσόμενος. Here is not ὁ ἐρχόμενος, as in the four places above (1:4, 1:8, 4:8 and 11:17), probably for the reason that there it is Christ as the coming judge, while in this vision is presented as already sitting on the judge's seat and delivering legal orders. "

Beza here presents several arguments to change the reading. It is clear, however, that this is different from earlier passages, such as in Revelation 1:4. The reason for his conjecture is thus given: the author of the Apocalypse seems to be inconsistent, because otherwise the expression ὁ ὢν καὶ ὁ ἦν is always ended by ὁ ἐρχόμενος. This pattern is restored by Beza, and at the same time changed, for the sake of context, for his conjecture here has ἐσόμενος instead of this ἐρχόμενος, because Christ in the previous passages (Rev 1:4,8; 4:8; 11:17) was the coming judge, but now already sits on the judgment seat.

But what about the "old and reliable manuscript" to which Beza refers? I do not claim to have solved the riddle, but I will make a suggestion for a possible sequence of events. It is fortunate that the hand copy of his second edition from 1565, which Beza used during the preparation of the third edition (1582), has survived. There are only a few entries in the Apocalypse, but here in Revelation 16:5 Beza has underlined ὁ ὅσιος and written in the margin ὁ ἐσόμενος, i.e. his conjecture, but nothing more. A similar entry can be found in Revelation 1:4, where he has underlined τοῦ and noted in the margin "In vetusto c[odice] legatur θεοῦ", "In an old manuscript it says θεοῦ." The difference is that here it is already explicitly a manuscript was mentioned, although Beza, in characteristic fashion, does not care which manuscript is meant. What I now suspect - also a kind of conjecture - is the following: In Revelation 16:5, Beza added ὁ ἐσόμενος early on, not from a manuscript, perhaps not even as a conjecture, but as the reading he would have expected, according to the pattern described above. Later, let us say when preparing the third edition for printing itself, i.e. around 1582, he simply forgot where this ὁ ἐσόμενος came from, and, following the example of Revelation 1:4, assumed that it was the same "old and [now because he liked the reading] reliable manuscript". This alone can explain why Beza referred to a manuscript that in all probability never existed. 

Had Beza known that ὁ ἐσόμενος was actually nothing more than his own guess (or exegetical remark), he would probably never have included the reading in his text. There is a certain irony to the whole story. Be that as it may, one can see here again the expectation of coherence and its devastating effect. The case is well known because it plays a role in the ongoing discussion on the Internet about the King James Version and the Textus Receptus (comparable to that about Erasmus and the conclusion of the Apocalypse). "O Beza, why did you do that?" asks one person. A question to which, in my opinion, there is only one answer: so that future generations can be left with a clear, all too clear example of the historical contingency of the Textus Receptus.

I tend to agree with Dr. Krans' analysis.  The reading "θεοῦ" is a variant found in Stephanus' margin.  We have good reason to believe that Beza had access to Stephanus' textual apparatus as it appears in the 1550 edition.  It is also believed that Beza may have had access to a more detailed unpublished apparatus prepared by Henri Stephanus.  

Stephanus identifies the manuscript in question as 15(ιε), namely GA 82, a tenth century miniscule.

(source, 476 - 477)

GA 82 is available online.  The relevant portion of the manuscript is this:

(source)

GA 82 differs slightly from the text of Beza (pre-1582) in terms of having "ο ων. και ος ην οσιος" (INTF transcription).  Similarly, GA 82 differs from the main text of the Stephanus 1550.

(source, 494 - 495)

Stephanus' margin does note a variant at Revelation 16:5 and points to the same manuscript: 15(ιε), namely GA 82.  In this case, though, I believe Stephanus intended to direct the reader to the fact that GA 82 does not insert "Lord" into the text.  

Of course, it is possible that Henri Stephanus' more complete apparatus (assuming this in fact existed and that Beza had it) may have mentioned the absence of the "και ο" before the "οσιος".  These are details that Beza's explanation aims to address, at least in part.

That leaves the question of whether possibly Henri Stephanus himself in his more complete apparatus suggested the conjectural emendation adopted by Beza.  We must admit, however, that this is simply speculation, as the more complete apparatus (if it existed) seems to have been lost to time.

Based on my reading of Dr. Krans' thesis, I believe it would be accurate to say that Beza generally only identified the two codices in his possession: Codex Claromontanus and Codex Bezae Cantabrigiensis. Neither of these manuscripts includes Revelation.  

I don't think Dr. Krans' take is too harsh, namely that Beza did not particularly care which manuscript had a given reading.  So, it makes sense that his handwritten notes mention a manuscript, but not which manuscript (the note referenced by Dr. Krans at Revelation 1:4 is reproduced below):

Whether Beza was then tripped up by the excessive brevity of his note at Revelation 16:5 (reproduced below) or whether Beza misread the notes of Henri Stephanus, we may never know.

It was interesting to note that Dr. Krans had not included Revelation 16:5 in his thesis because Beza asserted he had based his change on a manuscript.  What seems inescapable, however, is that Beza seems to have been wrong in that claim.  As we have no reason to think Beza a liar, we must conclude that he came to this error honestly. 

Update note: Updated 12/3/2024 to correct "co-expectation of heresy" (which was my mistranslation) with "expectation of coherence".