Stylometry
Stylometry is the mathematical analysis of style in writing. Style can come in the form of word choice and vocabulary, spelling, punctuation, grammatical usage, and other factors. People have been looking at the style of the New Testament books for nearly as long as they’ve been in existence. Early Church Fathers, for example, debated over the style of Hebrews and if Paul was the author or not. However, our very limited ability to navigate Koine Greek meant we would have to find another approach to stylometry. This is where computerized stylometry comes into play.
Computerized stylometric analysis of the New Testament goes back to the 1980s with Anthony Kenny’s A Stylometric Study of the New Testament. . . .
Unsurprisingly, stylometry has come a long way since the 1980s. Recently, researchers have been able to be as precise as knowing whether Donald Trump wrote a given tweet or one of his interns did, and identifying J. K. Rowling as the person behind the pseudonym Robert Galbraith.
(Britt and Wingo, loc 2823, Kindle)
You can read more about the program (Stylo) used by Britt and Wingo on the New Testament and other early “Church Father” writings at
- https://journal.r-project.org/archive/2016/RJ-2016-007/index.html
- https://github.com/computationalstylistics
- https://computationalstylistics.github.io/publications/stylo
So what did their analysis show about Galatians? Here is a composite image of the results of the program’s analysis of various works. On the left you can see that the program, Stylo, groups modern authors correctly. In the middle image we can see the Stylo results for various ancient authors such as Josephus, Origen, and others. Again, the results are what we would expect: noncontroversial works by Josephus are all grouped together, for example. The third image on the right covers results for some of the New Testament letters. The first two chapters of Galatians are closer in style to the letter of 1 Peter and part of 2 Corinthians. (2 Corinthians is widely thought by scholars to be a stitching together of several letters.)
Galatians
I quote here the commentary by B&W on Galatians. Marcion was the second century leader of a form of Christianity that claimed to have been the true followers of the apostle Paul and that the original twelve apostles had failed to carry on the true message of Jesus. Other “proto-orthodox” Christian leaders accused Marcion of falsifying the letters of Paul and argued that Paul and Peter and the other disciples were all united in their theological outlook. Marcion famously went to Rome with his gospel, bringing with him a large sum of money as a gift. In 144 CE the leaders of the church in Rome rejected Marcion’s gospel and his money.
Galatians is a heavily interpolated text, making a clearer stylometric read difficult. The first two chapters of Galatians seem to be a mishmash of writers attempting to explain Paul’s backstory, including striking parallels to Marcion’s personal story. Throughout the second half of the 2nd century and into the 3rd century there are varying accounts of what the letter said at the time, so we know it was still going through edits quite late. On the other hand, in line with what Tertullian says about Marcion “discovering” the letter, the content in chapters 3-6 tends to align more with Marcionite theology aside from verses scattered here and there which seem to contradict the bulk of the content. Such verses are likely from a later early church editor sometime in the process of the creation of our canonical version of the letter.
What seems most likely is that Galatians 3-6 represents the original content of the letter as drafted by someone in Marcion’s church. This was written prior to the falling out with the Roman church, so likely the late 130s or early 140s. Then, after Marcion is excommunicated, more biographical information is added into Paul primarily in chapters 1 and 2. Significant portions of the content in these chapters seem to reflect Marcion’s experiences, and they are likely projections of Marcion’s biographical information back onto Paul. This would explain why the authors look slightly different but still within the larger branch. It could be the same author at a different time or someone else writing a bit later. Eventually, other church leaders affiliated with the Roman church would add verses throughout the text to try to make Paul more acceptable to their theology.
Regardless, both sections of the text come up in the same general group as 1 Corinthians and the majority of the Romans segments. This indicates that no matter how many hands were initially involved in writing the letter, it was written around the same time and by the general same group as the other two letters. These three letters, along with parts of 2 Corinthians makeup whatever the core Paul character might have been.
(Britt and Wingo, loc 3454-71, Kindle)
Britt, Matthew, and Jaaron Wingo. Christ Before Jesus: Evidence for the Second-Century Origins of Jesus. Cooper & Samuels, 2024.
Neil Godfrey
Latest posts by Neil Godfrey (see all)
- Is Everything a Question of Probability? - 2024-12-15 03:04:03 GMT+0000
- The Folly of Bayesian Probability in “Doing History” - 2024-12-13 05:51:46 GMT+0000
- Jesus Mythicism and Historical Knowledge, Part 4: Did Jesus Exist? - 2024-11-27 08:20:47 GMT+0000
If you enjoyed this post, please consider donating to Vridar. Thanks!
The process by which material sympathetic to Marcion written after the split, manages to find its way into the canon of those most antagonistic to Marcion would need to be convoluted.
However I imagine there is a narrow path of possibility here.
If the Valentinians were fans of Paul, perhaps they continued to consume the works after the split and later Valentinians brought this work back into the orthodox fold when some later integrated into the Catholic faith.
Has anyone made an attempt to solve this problem?
I don’t think the process needs to be so very convoluted. All it would appear to have taken is a mediating position such as we find in Acts: Paul is accepted into orthodoxy by means of modifying his position towards circumcision. Similarly, Galatians could be accepted with a few additional qualifying statements followed by the claim that Marcion had taken scissors to the original.
Then I suppose my problem is that I am not aware of any attempt to even understand how the 2nd century religious publishing industry worked, and how they were related to the cults who consumed these works.
One thing that has occurred to me is that these were business enterprises, producing material that people or patrons are paying for.
Your comment seems to belie some previous thought into the processes, that you envisage a sort of horse-trading over the final outcome of the accepted text. It would seem that the aim is conciliatory, rather than an attempt to have the purist form?
There seems like there is a whole avenue of scholarship here. Are we limited to making guesses, or do we have other evidence apart from the texts themselves to shine some light on how the industry worked?
I don’t know what basis there is for postulating your “business model”. Certainly it cost money to copy and distribute texts, but I don’t know of any reason to think further than that.
As for “conciliatory” literature, yes, it is widely understood, I believe, that Luke in particular wrote “catholicizing” works that functioned to embrace as wide a spectrum as possible. I have seen it suggested that Marcion was attempting something similar — and that would explain why some of his reconstructed texts do not always seem to align perfectly with what his doctrine was supposed to be.
I am reluctant to use words like “industry” because I think they inject anachronistic models into how processes worked in the Roman world.
I think the texts do serve as reasonable evidence for a catholicizing program under Luke’s name: Acts depicts a Paul and Peter who are on the same page doctrinally, unlike the impression we get from the epistles — and the epistles seem to be unmentionable by Luke and one suspects it is because he wants to present different Paul, one who is “at one” with Peter and the Jerusalem apostles.
My point about ‘business model’ is that it seems to me that these enterprises were not cheap to run. Without saying anything about the motivation behind the endeavor, they still had to be financially viable. Somehow it had to be financed.
I am trying to think outside the text here. What role did these texts play in their community? How did it get financed?
The authors of our NT literature were all part of an educated elite, an upper class. They had money behind them (and slaves) to have acquired their education in literature and philosophy that is so evident in the NT writings. They attended readings of Josephus, rubbing shoulders with fellow elites. They were also familiar with dramatic literature and poetry. Such literate persons who saw themselves as philosophers or authors tended to have hangers-on, followers of lesser education who were keen to do the bidding of their “idol” (patron). If their patron became a Christian they would (as a rule) follow — such was the custom.
Local ecclesias or churches appear to have been modelled after mutual benefit and artisan guilds. Members with similar interests (say, bakery workers or potters) would have associations that required all members to contribute a modest sum that would go towards paying for the needs (such as the funerals) of members as they arose. It appears from the NT writings that if an entire church area was in need other churches would pitch in to help out with emergency relief funds.
But we know there were also wealthy individuals (as mentioned above — we have the products of their education in our bibles today) and a prize value of the era was for wealthy persons to generously donate as specific needs arose. In the non-Christian community private benefactors prided themselves on donating money for local civic projects such as a theatre construction or a public festival. It follows that one would expect well-to-do persons in the Christian community to use their slaves and wealth and their regular hangers-on (i.e. “clients” — who sought out wealthy patrons to serve for the benefit of “free meals” or other rewards) to meet the needs of copying tracts, stories, messages, …. and opening up and providing a large house where members could meet and hear the literature read, etc.
Thanks – that is a helpful picture.
This analysis by Britt and Wingo seems to confirm that independently made by Robert M Price in his book, ” The Amazing Colossal Apostle” of 2012. I suggest that all other ideas on the nature and history of “Paul” have been superseded.
To some extent, yes that is so. But I want to do a more careful study of what data was actually used before drawing firm conclusions. In the case of Galatians in particular we have wide ranging arguments for interpolations. If they are valid then that surely must corrupt the results of a stylometric analysis. Nor should we replace wholesale arguments made on the basis of human studies with machine studies. (Sometimes even my often helpful ChatGPT gives me wildly inaccurate feedback.) There are also different ways of interpreting the same data, in some cases. Do different styles within a document mean subsuquent corruption or should they be interpreted as collaborative composition of some kind from the get-go, for example.
I have always been wary of stylometric claims. They are useful in a sense of “there’s something interesting going on here” but imitation can also explain the same data, and the sample sizes for these letters are small and can’t account for the great bulk of classical texts and authors that are gone forever.
A tangential point that I don’t see discussed much is that age can change both writing style and content. Writing skill and style may get better or worse or evolve in a weird parallel direction, and positions on theology and philosophy can likewise shift. The idea that an author has one style only -and that it never changes – is a strange one to me.
For example, I’ve always had a strong impression that the Mark we have was written by a young man – a clever one, but not yet a masterful stylist. Perhaps later. If that gospel was written/edited/whatnot at the age of 25, say, and he went back and attempted to edit it at 55, would his style have changed or improved in those 30 years? Would the edits be obvious, or would he have the skill to conceal them and match his older parsing? If he went on to write other texts and his skill improved or shifted but they are lost to time, would we be able to match them back to Mark even if we had them?
In a modern textbook with many editions, the author may in a preface be at pains to point out what has changed (and perhaps at excessive length), but some chapters always tend to end up more stale than others. Alexander Bain’s 19th century composition textbook comes to mind – the last edition, many years after the first, is completely different and open-minded in mindset and approach compared to its draconian first edition. He mellowed! The style still matches, but the content is almost 180 degrees from where he started.
And of course, it’s the first editions that everyone cites as his positions…
I generally agree with your points — and one positive detail with Britt and Jingo is that they do acknowledge how style changes as we age.
Yes, stylometric analysis does indicated that “something is going on”, as you say — and for that reason I am not comfortable with dismissing it outright, either.
I find myself comparing …..
Pennebaker, James W. The Secret Life of Pronouns: What Our Words Say About Us. New York: Bloomsbury Press, 2013.
Hi Neil
This snippet from one of the quotes piqued my interest
“Throughout the second half of the 2nd century and into the 3rd century there are varying accounts of what the letter said at the time, so we know it was still going through edits quite late.”
I am not familiar with this. Do you have other posts that summarize the different states of Galations?
Thanks.
Don
The sort of evidence we have comes from “church fathers” such as Tertullian who are critiquing what Marcion has said about Paul’s writings. Sometimes their discussions leave a gap that enables us to conclude a particular verse was unknown to both Tertullian and Marcion, for example. That’s the sort of evidence being referred to in that quote.