Peterson: A Recounting of My Accounting of the Difficulty of Counting

Having blogged about the discrepancy between the number of manuscripts in Myths and Mistakes and the number recently given by INTF, I asked Jacob Peterson, the author of the chapter in question, to weigh in. I’m happy to share his thoughts here. —ed.

---

I had several people send me the INTF blog post by Katie Leggett and Greg Paulson in the days after it was posted and now I’ve been asked to write a short addendum here about Peter’s posts (here and here). Having been out of the field essentially since my chapter in Myths and Mistakes was published, this feels a bit like I’ve been granted some lasting and undue authority as an adjudicator on the issue at hand. At best, I have some thoughts for your continued consideration.

Preliminarily, I want to reiterate an overriding theme from my chapter—as Peter did in his most recent piece—that at the level of disagreement we have between my count and INTF’s the stakes are incredibly low and inconsequential, no matter how interesting the issue might be as an intellectual exercise. As I’ll show below, we are arguing over around 100 manuscripts out of over 5000. Beyond that, I have only a couple things to contribute that I think bring my numbers and theirs much closer. And, of course, I’ll offer updated guidance at the end.

If I have an actual qualm with the blog from INTF, it is in the conclusion to use the maximal count of 5700 as the total number of manuscripts when just two sentences earlier it was stated that “you can add together all destroyed manuscripts with ones presumed missing, and subtract this number from the total number of manuscripts, which results in 5,541 manuscripts.” I spent a fair amount of space in my chapter dealing with these categories and problematising issues around lost/destroyed/missing manuscripts, especially those for which we do not have images or the images are illegible (see especially p. 58). If the question is how many manuscripts are there and not how many were there, then surely a “destroyed” manuscript ain’t there (pardon my mostly suppressed Texas grammatical heritage).

Moving forward with their 5,541 number, we also need to account for the passage of time since I wrote my chapter. I submitted the draft in December 2017 with a limited subsequent edit in April 2019 to include the famed “First Century Mark” into my various charts, especially since I had a chart on the preservation of Mark in the first four centuries. In the nearly 5 years since that edit, the last issued numbers in the Liste have changed as follows. 

Category	M&M Count	INTF Blog count	Additions
Papyri	139	141	2
Majuscules	323	326	3
Minuscules	2940	3019	79
Lectionaries	2483	2555	72

That’s a total of 156 newly utilised numbers in the Liste, which roughly aligns with the 167 stated in the INTF blog to have been added since January 2019 (they indicate that some lectionaries were assigned to previously vacated numbers). If we subtract those 167 newly catalogued manuscripts from INTF’s 2023 number of 5541, we arrive at a count of 5,374 manuscripts. That is more than close enough to my approximation of 5300 to make me happy and justify the extrapolations I was doing.

In trying to explain the difference between my count and INTF’s, Peter points out that I got it correct for majuscules but says that “since minuscule and lectionaries turn out to have a far smaller rate of difference [from the Liste numbers due to loss], Peterson’s extrapolation undershot the total by hundreds.” I don’t contest what the Liste records as the state of things today, and certainly do not in any way malign them for an inability to perfectly reflect an ever-changing landscape of manuscripts spread around the world. Rather, as I contend in the chapter and its footnotes (especially p.68 fns. 44 and 45), minuscules and lectionaries probably do experience a rate of loss not too dissimilar to the majuscules but keeping track of them is much more difficult and the payoff is not worth the effort. 

Thus, my count relies on the Liste being, justifiably, out of sync with the real-world conditions of the much larger and harder to monitor class of minuscules and lectionaries. This is certainly a theoretical point—and a bit anecdotal from working in, visiting, and contacting dozens of monasteries—and it was made in the broader service of my point to use prudent and round numbers rather than sensational ones. 

Thus, we might say INTF and I were offering two different types of counts. My extrapolation was an approximation based on a readily ascertainable (and might I say accurate!) rate of loss within the majuscules to give a likely total of how many manuscripts there actually are, whereas INTF is offering a “state of the Liste” count that indicates how many manuscripts are currently catalogued and maybe still survive.

Whatever the case may be for this last point, I’m glad to see that my now-outdated numbers were not far off the mark. I’m similarly pleased to see that some of my comments have now been incorporated into the Liste, although doubtfully under the influence of my chapter. I have in mind here the de-listing of P99 and recognition that 0229 was not destroyed, among others. We should all be thankful for INTF’s ongoing work to maintain the Liste, not least so that I can pop in every once in a while to say a few things. 

In conclusion, with the updated catalogue and in view of my own advice about numbers, I’m happy to update my own guidance regarding how we should answer the question of how many manuscripts there are of the New Testament: 5500.

Explaining a Discrepancy in the Number of Greek NT Manuscripts

In my post last week, I pointed to a blog by the folks at INTF giving the number of known Greek NT manuscripts. Their number was 5,700. But some eagle-eyed readers may remember reading in Jacob Peterson’s chapter in Myths and Mistakes that there are more like 5,100–5,300. 

Why the discrepancy?

The short answer, I believe, is that Peterson used an extrapolation that turned out not to hold true across all categories. To get his total, Peterson went through the majuscules carefully to note places where manuscripts had been counted twice or added without warrant. From the original number of 323 (the highest majuscule number assigned when he wrote his chapter), he found 41 had been stricken from the Liste (p. 58). 

He further noted cases where the actual manuscript’s whereabouts are lost or unknown and that brought the number down from 282 to 261. From there it was a simple extrapolation to the other categories (see p. 68, esp. n. 44). From a total of 5885 numbers in the Liste, the result would be 5138. Given that minuscules and lectionaries might not suffer as much from the problem of double counting, he was happy to say it might be closer to 5,300.

As it turns out from INTF’s recent work, that number is not steady and the minuscules and lectionaries have suffered far less than the majuscules from the problems of mis-cataloguing. By comparing Peterson’s numbers to INTF’s, we can see the difference clearly.

	Highest Number in Liste	MSS Removed from Liste	Rate of difference
Papyri	141	6	4%
Majuscule	326	43	13%
Majuscule (Peterson)	323	41	13%
Minuscule	3019	159	5%
Lectionary	2555	133	5%

As you can see from this, Peterson and INTF got the same 13% rate of difference for majuscules. But since minuscule and lectionaries turn out to have a far smaller rate of difference, Peterson’s extrapolation undershot the total by hundreds. 

So, that’s the why. What about the so what?

Two things. First, Peterson notes 21 majuscules whose whereabouts are unknown. Incorporating them would have taken him below 5,100. I don’t see that category factored into INTF’s numbers. We could have a good debate about whether those should be counted or not. But part of Peterson’s goal was to question the usefulness of trying to attain an exact number.

That leads to the second point. The reason Peterson didn’t go through the minuscule and lectionary data is because there’s no point apologetically. As he says, “Arriving at a range this precise for the thousands of minuscules and lectionaries would be a monumental task and one that, as will hopefully become clear in what follows is not worth undertaking—at least not for apologetic purposes” (p. 58). Why? Because, as he says in his conclusion, “finding one more minuscule is not going to convince someone Christianity is true” (p. 68). 

Personally, I’m happy to use INFT’s numbers going forward. Perhaps a future edition of M&M could be updated to note all this. Which brings me to the conclusion of Peter Rodgers’s very nice review of M&M in the latest Filologia Neotestamentaria, “It is clear from my last few paragraphs that a wider audience should be in view than simply apologists for this fine collection of essays by upcoming scholars in New Testament textual research. It is hoped that periodic updates in further editions will address a broader range of issues for a more diverse readership.”

How Many Greek NT Manuscripts Are There Really?

That’s a question I get asked a lot. It’s a question many Christian apologists and skeptics of Christianity ask. And, it’s the question that Katie Leggett and Greg Paulson answered very carefully and very helpfully last month at the INTF blog. I’m tempted not to give you their answer here so that you have to read their entire post. 

But first, why is answering this so hard? Isn’t it just a matter of tallying up the highest manuscript number assigned in each of the four categories of papyri, majuscule, minuscule, and lectionary in the official list of manuscripts (known as the K-Liste)? It’s not. For at least three reasons. 

The first is the difficult question of what counts as a manuscript. Do amulets count? Should they count? What the Liste has included over the years has changed. 

Second, manuscripts can go missing, change hands without the list-keepers knowledge, or they can be lost or destroyed. If they were once on the list but are now unaccounted for, do they count? 

Third, what do we do when the same physical artifact contains portions of the New Testament in different hands from different centuries? Are we counting the resulting artifact or the varying copying events (for lack of a better term) that led to the one artifact? 

The point is not that these questions can’t be answered but that they must be answered before we can answer the original question. (By the way, these issues are all helpfully addressed in Jacob Peterson’s chapter in Myths and Mistakes.)

With all that said, what number do Leggett and Paulson arrive at? Drum roll, please... 5,700.

This is the number I will now be giving to people when they ask and it’s the number I would encourage you to use. Whenever you do, let me encourage you give the necessary caveats about the wide range in date, quality, and size of these 5,700 manuscripts. They are not all created equal! Furthermore, the number is not stable given new discoveries, the movement of manuscripts, and the ongoing identification of duplicates and the like. Still, it is very helpful to have a count from the same source that makes the official list we all use. Kudos to Leggett and Paulson!

Update (1/11/24): see the follow up posts on this here and here.