Exegetical Toolkit, Part 2
Textual Criticism
(avg. read time: 12–24 mins.)
Part 1: Introduction and Context
Part 3: Grammatical-Syntactical Analysis
The next tool in the exegetical toolkit deals with establishing what the text you are working with actually says. The translations you work with are inherently interpretive, but they are also based on varied Greek texts. Sometimes this is shown by differences in translating particular texts, but it is also shown by whether certain verses are included or how they are included (i.e., with notes, in brackets, or in the footnotes at the bottom of a page). These differences are driven by textual criticism.
Textual criticism involves the analysis of the manuscripts of a text for one or more of the following purposes. First and foremost is establishing the original or at least earliest recoverable text (depending on who one asks) that is most likely in a given context (until fairly recently, the task was described simply in terms of establishing the original text). Second, concomitantly, the method involves evaluating the plausibility of competing readings of a text. Third, especially in more recent years, there has been an interest in identifying the genealogy to which a given manuscript belongs as a way of explaining the grouping of variants. Fourth, the method can help in discerning what a particular manuscript or family of manuscripts might convey about the history of the text and the communities that used them (this last purpose is peculiar to textual criticism of scriptural texts, as opposed to, e.g., textual criticism of Josephus). In all of these ways, textual criticism concerns variant readings and explanation of the same. It also draws upon a number of disciplines, including philology, the study of particular ancient languages, paratextual features, translation theory (particularly in the case of non-Greek versions), paleography, study of scribal practices, and even historical theology (in the case of the final purpose). The field in which it works includes manuscripts of several varieties across many centuries, including early papyri, codices (including uncials/majuscules [texts written entirely in capital letters] and minuscules), lectionary texts, and patristic quotations.
I have previously addressed introductory matters of textual criticism for the NT and the OT. There are enough corrections and expansions of the NT post that I do not tend to share it, and maybe one day I will return to writing a better version of it. And while I have used textual criticism to some extent in many posts, I have used the tool more extensively for the NT in the following posts:
An Exercise in NT Textual Criticism: The Case of 1 Cor 15:49
Mark 16: How Should It Have Ended?
The “Western Text” of Acts in Codex Bezae and Its Text-Critical Issues
Conflagration and New Creation in 2 Peter 3
The Pericope Adulterae: To Include or Not to Include John 7:53–8:11
Response to Peter Head’s Case Against Mark 16:9–20
What Is Jude Instructing in Verse 22?
Fulfillment of Scripture in the Gospel According to Mark
Gospel Synopsis Commentary, Part 2
Unfortunately, what many people think they know about textual criticism is often based on one of three things: 1) unhelpful or misleading notes from Bible translations; 2) exaggerations from apologists about the manuscripts; 3) exaggerations from skeptics about the manuscripts (less often, there are also exaggerated claims from scholars about certain texts, but I have addressed such things elsewhere). Bible translation notes are fine for making you aware of the variants in the mss, but they are not always good for evaluation or for framing the textual evidence as it is. Apologists often undersell how significant some variations are (although they are correct that most are insignificant) or they oversell various facts about the textual record (such as comparing numbers of mss with other books when not doing like-for-like comparisons and how much text the early church fathers quote).1 Skeptics, usually armed with whatever they read or heard from Bart Ehrman, tend to focus on the differences/variations between mss while overlooking that most of these are insignificant and that, if the problem was as bad as they frame it, we could never identify mss of a common text.
A good resource for correcting a variety of misconceptions is Myths and Mistakes in New Testament Textual Criticism. If you want helpful online analyses of text-critical matters and references for research, you would be well served to read the works of Phoenix Seminary’s Text & Canon Institute, the blog Evangelical Textual Criticism, Timothy Mitchell’s The Textual Mechanic, and James Snapp’s The Text of the Gospels (though he reviews much more than Gospel texts). For an online work on many variants in the Gospels, you can consult Wieland Willker’s extensive Online Textual Commentary on the Greek Gospels.
If you have enough competence in Greek and knowledge of how to read mss (including the markings therein), there are sites and tools available online for interacting with texts. These include the Institut für Neutestamentliche Textforschung (INTF) and their virtual manuscript room, as well as the Center for the Study of New Testament Manuscripts. Remarkably, Wikipedia can actually provide some helpful orientation and numerous links to images of individual manuscripts.
In terms of judging between variants and explaining how those variants arose, textual critics have devised a number of tests for the readings, mostly working from some or all of the following assumptions:
In terms of the original or earliest recoverable reading, only one can be original, and textual difficulties cannot be solved by conjecturing variants with no attestation. (Of course, this is sometimes not the case with scholars, as there have been conjectural emendations proposed.)
The original/earliest recoverable reading is the one that best satisfies external criteria, internal criteria, or a balance of both.
The genealogical principle of how mss are related is crucial in this regard in contributing to the explanation of how certain readings arose. Variants thus must be considered in the context of tradition.
Pride of place is given to the Greek manuscripts with the patristic quotations and other versions (mostly in Latin, but also in Syriac, Coptic, Armenian, Georgian, Gothic, Ethiopic, and Slavonic, as well as others) being secondary or supplemental. A reading only attested in non-Greek manuscripts would generally not be preferred over one that is present in Greek manuscripts.
It is better to “weigh” manuscripts than simply count them. This principle is often misunderstood so that it is invoked to attribute immense significance to one or two manuscripts and little significance to many others (which results in some favoring a reading of only one or two mss against multitudes of others that attest to another reading). Rather, it emphasizes independent testimony from a variety of witnesses over a simply higher quantity of related or dependent witnesses.
External criteria for analyzing readings include the following:
Age of manuscript (determined primarily by paleography)
Quality of manuscript
Geographical distribution (Is the reading widespread or localized to witnesses in a certain area)
Number of manuscripts
Internal criteria for analyzing readings include the following:
The reading that best explains other variants is to be preferred.
The more difficult reading (lectio difficilior) is to be preferred, since it is, prima facie, more likely that scribes would make readings easier rather than more difficult.
The shorter reading (lectio brevior) is to be preferred, since it is, prima facie, more likely that scribes would expand texts and combine readings than shorten a text.
The reading more in line with the author’s style and theology is to be preferred. Another version of this criteria is that the reading more in line with the immediate context is to be preferred.
In the particular case of the Synoptic Gospels, the reading that is not harmonized or is less parallel is to be preferred, since it is, prima facie, more likely that scribes would harmonize similar texts from each Gospel than that they would create difference.
It is worth noting, however, that scholars now more widely recognize that none or almost none of these criteria are absolute. Concerning the external criteria, the oldest manuscripts, though closer in time to the autographs, are often not the best; in fact, papyri like P46 and P66 contain a number of spelling errors that gave rise to many others that copied from them, not to mention other questionable variants. Quality judgments must always be tempered based on the specific manuscript and specific parts thereof, and there is a risk in circular reasoning in judging the quality. Geographic distribution is helpful if the production centers of these manuscripts operated independently, but even then, some scribal groups operated according to higher standards than others, and some may have been more consistent about, say, using proofreaders. Number is, of course, not a strong criterion, but it should not be dismissed either, as it is unfortunately not unheard of for critical texts like the Nestle-Aland to favor readings attested in only one ms, two mss, or three mss. In any case, while it helps to have a census of witnesses for a reading, it is by no means decisive.
Concerning the internal criteria, the principle of the difficult reading was once especially influential, but its weaknesses have become more apparent over time. While scribes may have a tendency to smooth out difficult readings, sometimes a reading is difficult precisely because it does not make sense in the context and thus should not be preferred. Furthermore, this supposition about scribal tendencies, while prima facie probable, does not always apply, as there are several cases where what is supposed to be the easier reading is in a distinct minority of mss. Furthermore, as I have tried to show elsewhere, sometimes what is described as the more difficult reading may be based on a misunderstanding of how the reading came about and thus whether it can properly be considered more difficult.
The principle of the shorter reading makes intuitive sense, as scribes may add marginal notes, as well as their own impressions of what should be there syntactically or theologically, which are then sometimes added into the text proper. However, readings may be shorter because of accidental omissions (even entire lines could be omitted if two proximate lines ended the same way). Or they may be shorter as a way of correcting grammar from what was perceived as an “overly full” grammatical construction. Or, in the case of the Western and Eastern traditions of Acts, the shorter reading could possibly be a different edition. In general, it is by no means given that developments simply happen from simpler to more complex.2 Sometimes, there may be reasons for simplifying what was once more complex.
The principle of context and style makes sense, because scribes may substitute similar words based on their own stylistic preferences or dominant preferences of the time in which they write. But in some cases, we simply do not have enough text to be able to make strong pronouncements of an author’s style when a different term or phrase appears, and sometimes we overestimate how different prepositional constructions might be for the purpose of an author’s theology. Furthermore, as I also discuss in my own work, this sense of contextual fit can at times be based on misreading what the variant is actually saying.
The Gospel principle also makes sense in a way, as scribes and teachers of the Gospels (notably, Tatian) are all too keen to harmonize texts. But even this principle cannot be used by itself, lest we operate on the assumption that only texts that produce Synoptic Gospels as different as possible are to be preferred. Furthermore, if scribes did engage in harmonizing, we cannot assume it was all in one particular way. If we go by citations alone, the Gospel according to Matthew seems to have been the most popular of the Gospels to use in the early church, but the harmonized readings are not always attempting to align texts from Mark and Luke to Matthew. Sometimes, the move goes in another direction, and sometimes the divergent reading was a deviation from a case where the Synoptics originally agreed on a textual level. (Of course, I have shown in Part 2 of my Gospel Synopsis Commentary an example of how there were harmonizations of other Gospels with Matthew.)
The principle of best explanation is the most important of the criteria. But it is also important here not to apply this criterion in a simplistic fashion, as, for example, I have noted in my work on Mark 16. It helps to actually apply the criteria for determining the best explanation, as I have outlined in my own work elsewhere, of background plausibility, explanatory scope, explanatory power, simplicity/parsimony, and illumination. It should also be acknowledged that what constitutes a best explanation is a provisional conclusion made with varying degrees of confidence that may be reaffirmed with further study or undermined with either further evidence or actual due consideration for an alternative.
Application to Rev 1:1–8
Revelation has a complicated history of transmission (on which see here and here for some more info), one result of which is that it is not attested nearly as early and as well in other early versions of the East as many other books in the NT (though it is better attested in Western versions). It was also not as widely used in Greek churches for a stretch of time. But we still have a substantial number of witnesses for the text of Revelation. For Greek mss specifically, 310 mss attest to some portion of Revelation or the entirety thereof. Most of these come from the latter half of the medieval era (owing both to the lifespans of mss in most areas that are not Egypt/the Levant and to the relatively greater stability of conditions for scribes in later times to produce a large number of mss). That tally also includes fragments of seven papyri, one of which (P98) may date from the second century (or the third at the latest) and two others of which (P18 and P47) are from the third or fourth century. Our earliest codices also include Revelation, and they are important not only for their testimony to the text, but also for what they indicate about the contexts in which the mss were composed. In total, we have seven papyri, twelve majuscules, and 291 minuscules, although a few of these are not currently accessible.
Revelation is also illustrative of how complex textual traditions can be. Although it is not as if manuscripts are so different as to be unrecognizable as descended from the same book, even the Byzantine texts, which have often been treated as a practical monolith otherwise (especially among the later representatives), show evidence of different branches of transmission in Revelation. One branch is referred to as the “Koine” text. Another is the “Andreas” text, representing the many Greek mss also accompanied by the commentary of Andrew of Caesarea. Furthermore, there is another branch more recently identified as the Complutensian text. These are often treated as distinct textual families, but further research is needed to confirm this.
Although there are a number of critical texts available today (probably the most popular of which are the “Nestle-Aland” Novum Testamentum Graece currently in its 28th edition and the companion volume of the United Bible Societies’ Greek New Tesament currently in its 5th edition), only recently have those involved with the more thorough Editio Critica Maior (ECM) completed their multivolume work on Revelation. I have not been able to purchase it because my money tree has not yet sprouted, much less blossomed. If you have more disposable income than I do, or if you have access to a good academic library, you will want to consult this for text-critical work on Revelation. The same applies to the Text und Textwert der griechsichen Handschriften des Neuen Testaments VI.3 The aforementioned INTF is incorporating notes from the ECM on Revelation as it has done for other texts available in the series thus far, but it will not provide all the information available in the volumes.
Interestingly, one of the earliest copies of Revelation—P18 (also see here)—features fragments of 1:4–7. It was written on the other side of a scroll that also contains the ending of Exodus. Thus, fortunately, we will be able to consult at least one papyrus witness for our study here. While its text from the third or fourth century, insofar as it is preserved, largely agrees with the critical texts and the majority of texts, there are some notable respects in which it is differentiated. We see here how the nomina sacra abbreviations were in use, but they were either not standardized yet or the scribe did not follow the standard every time. And so in v. 4 he uses the first two letters of “Jesus” and “Christ,” rather than the first and last letters of each. At the same time, the original scribe used the first and last letters of θεου (θυ) in v. 6, a form in disagreement with the consensus of witnesses, which a corrector made to be the dative θω in agreement with the consensus. Another error is the feature of a pronoun ημ–ν (presumably, ημιν), against the vast, vast majority of witnesses across languages to the accusative that makes better grammatical sense (ημας). For whatever reason, the order of “glory and strength/might” has also been reversed here to “strength/might and glory,” in contrast to nearly every other witness.
Similarly, the early codex Sinaiticus (א) has some idiosyncratic or near-idiosyncratic readings and spellings. One that cannot be attributed to a mere spelling issue is in v. 1. Here, the text uses ἁγίοις where other mss feature δούλοις in 1:1. This is clearly a later variant introduced either by the scribe or a predecessor, as it is matched by another variant introduced at the end in 22:21, wherein the benediction in Sinaiticus calls for the grace of the Lord Jesus Christ to be with “the saints” rather than the less specific “all.”
The ms also features some spelling errors compared to other mss, which are not exactly uncharacteristic for Sinaiticus. Verse 1 features Ιωαννει instead of Ἰωάννῃ. Verse 4 features Ἰωανης instead of Ἰωάννης. And v. 5 features βασιλειων instead of βασιλέων. Another spelling issue that is less peculiar to Sinaiticus, but still a distinctly minority one is the use of ιδεν instead of εἶδεν in v. 2. While both of these texts mostly agree with the consensus of other witnesses where they can be compared, they still illustrate that the earliest text is not always the best/most likely one. Sometimes, early mss feature unique or remarkably rare readings that are less plausible than more common readings.
Still, these mss show how most of the variants in these first eight verses are of varieties where only slight attestation diverges from the consensus, and/or only slight variations of spelling or word division make up the variants. As such, we will not be reviewing every variant in these first eight verses. But some variants either are more interesting, have more witnesses, or have more impact on wording, and so they will be reviewed here. In any case, none of the variants in question introduce major alterations of the message.
For example, only two mss cited in ECM refer to this apocalypse as “the apocalypse/revelation of John” rather than “the apocalypse/revelation of Jesus Christ.” These two mss are from the thirteenth and fifteenth centuries (MSS 1852 and 2845, respectively). It is not difficult to figure out what the most likely original opening words for Revelation were. The other two cases simply conformed the opening words to the title of the book.
More significant is the additional clause at the end of v. 2 (with some variations in spellings, word divisions, and word order): και ατινα εισιν και ατινα χρη γενεσθαι μετα ταυτα (“and what(ever) are and what(ever) must happen after these things”). Nestle-Aland assigns this reading to the “Andreas” branch of Byzantine texts, which is mostly correct but still oversimplified.4 It is simple enough to determine that this is a late addition by the fact that only later medieval mss attest to it (the earliest witness, 2074, is from the tenth century, and most are from multiple centuries after that). It is practically rephrasing a clause from v. 1, so nothing substantial is added in any case.
One other case, which is bracketed in Nestle-Aland, comes from v. 6 and concerns whether τῶν αἰώνων should be added after the prepositional phrase εἰς τοὺς αἰῶνας. 212 Greek witnesses favor this form of the text with three others varying only the spelling of τῶν αἰώνων. Additionally, most Latin witnesses, Georgian mss, Slavonic mss, Armenian mss, Ethiopic mss, Sahidic mss, and Syriac mss provide indirect testimony for this fuller phrase. Twenty-nine others do not have this part of the text extant or are not readable. Three others omit more of the surrounding text. Of those that can be accessed, that currently leaves thirty-nine that omit the words bracketed in Nestle-Aland. However, that minority includes some early witnesses like P18 and A, as well as 025 and various minuscules. Additionally, some Latin witnesses and Bohairic mss indicate this shorter phrase.
The only translational difference made here is whether the phrase would effectively be “forever” or “forever and ever.” Similar issues about whether the fuller phrase was preferable or if it was redundant arose in the transmissions of the phrase in Rom 16:27; Heb 13:21; 1 Pet 4:11; and 5:11. Adding to the perceived lack of clarity of what should be preferred is that the phrase could just as well be used in the condensed form and in the fuller form without significant controversy in textual transmission (see, e.g., Gal 1:5; 1 Tim 1:17; Rev 4:9–10). The fuller form is used without significant textual controversy elsewhere in Revelation (1:18; 4:9–10; 5:13; 7:12; 10:6; 11:15; 14:11; 15:7; 19:13; 20:10; 22:5). Some take this to imply that the phrase in 1:6 was expanded to conform to the other instances, but this is an unintuitive in the absence of extensive evidence for correctors going back and adding the phrase after the pattern was recognized (which would mean waiting until reading either 1:18 or all the way for 4:9–10 so that a proper “pattern” can begin to be perceived). It is also rather unintuitive that only the first use of the prepositional phrase would not fit the pattern. Nor was there a consistent pattern in using the shorter or fuller phrases in doxological contexts or otherwise among the earliest Christians so that the text could be explained in that way. More likely, for one reason or another (aesthetic sense, perception that there was accidental repetition, and so on), these mss omitted the phrase in the earlier use only to realize later that this was an intentional feature of John’s. Thus, most of the scribes in question from whom we have more text left these other uses alone.
The only other major variant that affects multiple words is in v. 8, as nearly a hundred Greek mss include after “I am the Alpha and the Omega” some form of the apposition, “the beginning and the end/goal.” This phrase appears alongside other references to the Alpha and the Omega in 21:6 and 22:13. Most Greek mss omit the phrase, as do many Old Latin witnesses, Armenian mss, Georgian mss, Slavonic mss, Syriac mss, and Ethiopic mss. On the other hand, Vulgate mss, some Slavonic mss, some Syriac mss, and at least one Arabic ms feature this phrase. Interestingly, there are significant overlaps between mss that had the fuller phrase in v. 6 and omitted this phrase in v. 8, as well as between mss that used the shorter phrase in v. 6 and included this phrase in v. 8. In both cases, the majority reading is more likely.
I do not think it is quite accurate to say that this represents a conformity of 1:8 to 21:6 and 22:13, since I have already noted how unintuitive that kind of explanation is, especially in the absence of extensive evidence of correction in this fashion. It is more likely that this became a regular feature of doxological formulae invoking the Alpha and the Omega, and so the text was conformed to the habit, which had in turn been shaped by the latter two texts. Otherwise, it is unclear why the phrase should be omitted in the majority of mss if it were original. Conversely, it makes sense that John’s text should have the descriptor grow from the beginning with “the Alpha and the Omega” (1:8) to “the Alpha and the Omega, the beginning and the end/goal” (21:6) to “the Alpha and the Omega, the first and the last, the beginning and the end/goal” (22:13). This makes sense as the subtle design of an author rather than of a later scribe or later scribes who got many of their fellows to go along with the idea.
Every other case with significant variance concerns individual words. One of these is whether or not “God” (θεου) appears after the ἀπό in v. 4. 135 Greek witnesses attest to the genitive term, but the earliest witnesses are well into the medieval era (the only uncial with this reading, 046, is from the tenth century). All the earliest witnesses that have this part of the text do not feature it, and the absence is also supported by most Old Latin witnesses, Vulgate mss, Coptic mss, Slavonic mss, and at least one Arabic ms. While there is no apparent reason for it to have been omitted if it were original, all the addition really accomplishes is a grammatical bridge between the participial phrase and the preposition that normally takes a genitive as its object (therefore, seemingly, correcting the grammar of the text), as well as further, redundant specification of the identity of “the one who is, the one who was, and the one who is coming” (1:4, 8; 4:8; cf. 1:18; 4:9–10; 10:6; 11:17; 15:7; 16:5).
The other variants to discuss are all in v. 5. One comes from a minority of mss attesting to the presence of the preposition ἐκ in the middle of “the firstborn of the dead.” Because the witnesses are a distinct minority and late, and since Latin witnesses (especially among Vulgate mss) and others (including Syriac, Bohairic, Armenian, and others) tend to favor the absence of the preposition, it makes more sense for the preposition to have been added than omitted. However, it should be noted that a minority of witnesses omit the preposition in the similar phrase of Col 1:18. But since that is more frequent among witnesses for this text, it appears that the explanation is the text being due to expansion conforming this phrase to one known from Col 1:18, a popular NT source for christological statements.
A group of variants in v. 5 concerns both the participle as λύσαντι vs. λούσαντι and the preposition attached to it (ἐκ or ἀπό). Early versions cannot help adjudicate on the preposition, since they are so similar in sense. And while λύσαντι more naturally pairs with ἐκ and λούσαντι more naturally pairs with ἀπό, there are variants attesting to pairings of each participle with each preposition. The participles are distinguished by only one letter, but their difference in meaning is between releasing/setting free (λύσαντι) and washing (λούσαντι). Both terms fit presentations of Jesus’s work in the NT more broadly and Revelation in particular (5:9; 7:14; 22:14). However, this would be the only use of the term λούω in Revelation if it were original. Although the majority favor the “wash” reading, as do Vulgate mss, Coptic mss, Georgian mss, Armenian mss, and some Slavonic mss, the “release” reading is favored by all the earlier mss, a good portion of other Greek mss, and a similar or wider distribution of versions with Old Latin, Slavonic, Armenian, Syriac, and Ethiopic witnesses. The latter reading is also favored by Victorinus of Pettau in his commentary on Revelation. On balance, the external and internal factors favor λύσαντι with λούσαντι being a slight lengthening of the same in later mss.
There are technically other matters we might comment on with 1:7. However, those concern the use of the OT. For that reason, I hold off on addressing those until the part of this series on the NT use of the OT. But to conclude, here is the Greek text of Rev 1:1–8:
1 Ἀποκάλυψις Ἰησοῦ Χριστοῦ ἣν ἔδωκεν αὐτῷ ὁ θεὸς δεῖξαι τοῖς δούλοις αὐτοῦ ἃ δεῖ γενέσθαι ἐν τάχει, καὶ ἐσήμανεν ἀποστείλας διὰ τοῦ ἀγγελου αὐτοῦ τῷ δούλῳ αὐτοῦ Ἰωάννῃ, 2 ὃς ἐμαρτύρησεν τὸν λόγον τοῦ θεοῦ καὶ τὴν μαρτυρίαν Ἰησοῦ Χριστοῦ ὅσα εἶδεν. 3 Μακάριος ὁ ἀναγινώσκων καὶ οἱ ἀκούοντες τοὺς λόγους τῆς προφητείας καὶ τηροῦντες τὰ ἐν αὐτῇ γεγραμμένα, ὁ γὰρ καιρὸς ἐγγύς. 4 Ἰωάννης ταῖς ἑπτὰ ἐκκλησίαις ταῖς ἐν τῇ Ἀσίᾳ· χάρις ὑμῖν καὶ εἰρήνη ἀπὸ ὁ ὤν καὶ ὁ ἦν καὶ ὁ ἐρχόμενος καὶ ἀπὸ τῶν ἑπτὰ πνευμάτων ἃ ἐνώπιον τοῦ θρόνου αὐτοῦ 5 καὶ ἀπὸ Ἰησοῦ Χριστοῦ, ὁ μάρτυς ὁ πιστὀς, ὁ πρωτότοκος τῶν νεκρῶν καὶ ὁ ἄρχων τῶν βασιλέων τῆς γῆς. Τῷ ἀγαπῶντι ἡμᾶς καὶ λύσαντι ἡμᾶς ἐκ τῶν ἁμαρτιῶν ἡμῶν ἐν τῷ αἵματι αὐτοῦ, 6 καὶ ἐποίησεν ἡμᾶς βασιλείαν, ἱερεῖς τῷ θεῷ καὶ πατρὶ αὐτοῦ, αὐτῷ ἡ δόξα καὶ τὸ κράτος εἰς τοὺς αἰῶνας τῶν αἰώνων· ἀμήν.
7 Ἰδοὺ ἔρχεται μετὰ τῶν νεφελῶν,
καὶ ὄψεται αὐτὸν πὰς ὀφθαλμὸς
καὶ οἵτινες αὐτὸν ἐξεκέντησαν,
καὶ κόψονται επ’ αὐτὸν πᾶσαι αἱ φυλαὶ ττῆς γῆς. Ναί, ἀμήν.
8 Ἐγώ εἰμι τὸ ἄλφα καὶ τὸ ὦ, λέγει κύριος ὁ θεός, ὁ ὤν καὶ ὁ ἦν καὶ ὁ ἐρχόμενος, ὁ παντοκράτωρ.
In terms of the former, I mean they sometimes refer to thousands of Greek mss of the New Testament, when this is a cumulative total of all manuscripts of each of the twenty-seven books (whereas there are, in fact, more mss of the Gospels than Revelation, and so on), and they sometimes compare more inflated totals with toals for texts that are not counted the same way. Lists of texts from critical editions of classic works tend to list only the mss used, rather than the entirety of all mss of the works in question. In terms of the latter, you sometimes hear claims that all but eleven verses are quoted somewhere by the patristic sources or that we could thus reconstruct almost all of the text even if without the actual mss. The problem is that these quotes are not all quotes, they are not all complete, and they are generally not of such extent and order that we could actually reconstruct the text in order.
Indeed, this assumption is often false. See James R. Royse, Scribal Habits in Early New Testament Greek Papyri, NTTSD 36 (Leiden: Brill, 2008), esp. 705–36.
Markus Lembke et al., Text und Textwert der griechsichen Handschriften des Neuen Testaments VI: Die Apokalypse; Teststellenkollation und Auswertungen, ANTF 49 (Berlin: de Gruyter, 2017).
Nineteen of the witnesses cited in ECM are of the “Andreas” branch, but it is not in all of the “Andreas” texts. 35 is a mixed text, but the corrector who added this as a marginal gloss was probably influenced by that branch of texts. 104 mostly aligns with the “Koine” branch, but it is at a lower level of agreement than many other texts. Other texts generally identified as Koine or as being more mixed with this reading include 367, 620, 632, 2071, 2436. 1637 and 2723, which have this reading, are considered Complutensian.