William Loader, Sexuality and Gender, and the Ethics of Citation

I have a review of William Loader’s Sexuality and Gender: Collected Essays coming out soon in Reviews in Religion and Theology, a book that raises ethical questions in its content and its composition. I’ll return to the concerns I have with Loader’s writing in a moment.

To be clear, the title suggests that Loader will be talking about issues of sexual orientation and gender paradigms / identity much more than he actually does, although sexual orientation does come up in a few places. Rather, the essays address the ethics of sexuality (as in, being sexual) which are in many ways impacted by perceived gender and gender norms. He covers mostly later Jewish biblical literature, the apocrypha, Philo, and Josephus, as well as the New Testament. The patterns he notes here are discussed more thoroughly in the review. Here, I want to address two issues with the book.

Will the Ancient Gay Jews Please Stand Up!

When Loader does discuss homosexuality, or rather homoerotic practices, he has a tendency to definitively find homoeroticism where texts aren’t exactly clear. These potential type II errors (he fails to reject the hypothesis that ancient texts are talking about homoeroticism even when there’s a strong chance it’s false) have plagued his earlier works on ancient sexual ethics. They are exacerbated in this volume because Loader rarely takes the time to defend his readings, instead citing his earlier works and quickly moving on. A casual reader who does not look up the (sometimes obscure) passage or Loader’s previous discussion may not appreciate how uncertain his reading actually is.

Loader is especially apt to find lesbians when none are mentioned. For example, in his discussion of The Sentences of Pseudo-Phocylides, Loader simply notes “his disapproval of lesbian relations where one woman usurps a male role” (p. 175), elsewhere claiming the text “applies the Leviticus prohibitions by extension to female same-sex relations (190-192)” (p. 354).

I’ve never found this argument convincing. Here are lines 189-192:

Do not outrage your wife by shameful sex acts.

Do not go beyond natural sex acts by unlawful passions

(Nor do male beasts please themselves with sex acts together).

Any female should not imitate the sex acts of men.

Loader is possibly thinking of women using dildos with each other, but they could just as easily use them with their husbands. If so, the whole focus of these lines would be on anal sex: don’t perform it with your wife (189) or with another man (190-91), and don’t let your wife do it to you (192).

The point is, there is almost no allusion to Leviticus 18 (as Loader acknowledges, pp. 174-75). Apparently, Loader thinks of Lev 18:22 only because male homoeroticism is condemned as “unlawful.” That’s it. He hears lesbians in line 192 because they are compared to homoerotic men, and through our own modern paradigm that lumps gay men and lesbians together under the heading of “homosexual,” he infers a strong condemnation of all lesbian behavior in ancient Jewish circles. That seems to be put too much weight on the one ambiguous line. Since real LGBTQ+ people’s lives are impacted by biblical interpretations, Loader has an ethical and scholarly responsibility to unpack the nuance of these ancient texts, which he repeatedly fails to do with texts that can be interpreted as condemning female homoeroticism.

With Romans 1 he admits “one has to be satisfied with degrees of probability” (p. 371) whether lesbians are even addressed, but this comes after a longer argument that claims the passage “reads most naturally to see 1:26 as similarly referring to same-sex relations but here applied to women,” adding, “Female to female eroticism was more widely condemned in the Greco-Roman world than male, so that perhaps Paul chose to begin with the most abhorred” (p. 370). “Most abhorred” is a bit strong. Female homoeroticism is often poo-pooed at best, with very few strong condemnations of it in the literature. It would be more accurate to say that it is the most consistently criticized since, while certain forms of male homoeroticism were condemned much more strongly and at much greater length than anything we see regarding women, Greco-Roman culture had a place for other forms of male homoeroticism. Sometimes these other behaviors are praised right alongside intense condemnations of others. But if we set condemnations of female homoeroticism side by side with condemnations of male homoeroticism, condemnations of women would include some offhand remarks, vague comments, jokes, and the occasional strong statement. Very few wholesale arguments would appear. Condemnations of men would include a much longer, much more thoroughly argued list that reads like an evangelical sermon. Many discussions of male homoeroticism pass over female homoerotic behavior altogether. You wouldn’t get this impression from Loader’s discussions.

Loader opens himself to this sort of critique since he addresses the ethical and ecclesiological application of his findings in several of the later chapters of the book. I can’t say I always agree with his ethical positions, but they are clearly argued and often nuanced. It’s simply that they suggest he’s aware of the ethical implications of his research, which makes the casual way he makes strong but largely unwarranted statements about lesbians a bit harder to swallow.

The Ethics of Citation

There are two scholars Loader relies on that raise questions of whether one can divorce the scholarship from the scholar (although with one far worse than the other). This is important in academia since citations, “impact scores,” and h-indexes raise the authors’ prestige. Their impact influences whether they get important teaching positions, invitations to public talks, book deals, and more. In short, an influential scholar like Loader citing someone – and especially citing them frequently and interacting with them thoroughly – benefits the scholar he cites. It’s not a neutral act.

Furthermore, if there is something problematic about the scholar cited, something publicly known and even relevant to the research cited, and Loader does not acknowledge it or comment on it, the citation acts as a tacit endorsement, or at least a handwaving of these problems.

I bring this up because citation issues came up with two problematic scholars. In the first case, I noticed Loader was heavily relying on Robert Gagnon in discussions of Romans 1. I’ve read some of Gagnon’s work, but the frequent citations piqued my interest. I wanted to learn more. As it turns out, Gagnon is a strong advocate of applying (centuries later) natural law theory to Romans in support of an opinion of heteronormative gender complementarity. He is a biblical scholar, but one who uses his work on ancient sexual ethics in order to argue against the rights of LGBTQ+ people. Much of Gagnon’s applications of his research go directly against Loader’s efforts to read these passage in their historical context. A reader would never get that impression from Loader’s discussions of Gagnon’s work.

By relying on Gagnon so heavily for his examination of NT comments on homoeroticism, Loader has amplified Gagnon’s scholarly prestige and directed readers to works that have a clear political and social agenda. And he has done that without comment on Gagnon’s larger agenda. Novice readers would only get the impression that Gagnon is a leader in this particular subfield (he’s not exactly—many of his exegetical arguments are easy to find elsewhere, after and before his work), and may assume that Loader agrees with Gagnon’s social agenda. Or worse, assume that Gagnon’s (mis)applications of his research follows naturally from reading the biblical text.

Should Loader have cited Gagnon if he relied on him? Yes.

Should Loader have employed other scholars too, so that he engages the (often common) argument rather than a particular scholar? He could have.

Should he have noted even in a footnote the motivation behind Gagnon’s research, and at least claimed it was irrelevant to the specific question he was addressing? I would think so.

To be fair, I was not too turned off by the interaction with Gagnon. I was probably more sensitive to citation issues because I was already turned off by multiple citations of a much more problematic scholar: Jan Joosten.

Joosten did some work on sexual ethics in the Torah, specifically on condemnations of male anal sex in Leviticus 18 and 20. So there are some relevant works there to cite.

However, in 2020 Joosten was convicted of possession of massive troves of child pornography, including depictions of rape. There’s a discussion happening behind the scenes in biblical studies and in academia as a whole whether to cite scholars like Joosten or Richard Pervo, who was also convicted of possession of child pornography in 2001.[1] Despite that, Pervo published many influential books after his conviction, before his death in 2017. I even briefly met him at a Chicago Society of Biblical Research meeting where he was a featured speaker—before I knew anything about his conviction. It’s strange to think that there were no social or professional consequences to a child pornography conviction in such a conservative field. Or maybe it’s because the field is so conservative, I don’t know.

The issue with citations of Joosten is he deals specifically with issues of sexual ethics. Loader not only cites his research approvingly (p. 386, a 2020 article published while Joosten was on trial), but on two occasions cites personal correspondence with Joosten (pp. 37 n. 7, 48 n. 2)! That’s… creepy, dude. You know what else was on the computer he sent those emails from? And that 2020 article has 11 citations in Google Scholar, five of them from Loader.[2]

So what points does Loader make with Joosten’s work? In the article cited, it is that the terms in Lev 18:22 are ambiguous since man and woman are also often the words for husband and wife (always so in Greek), and the “lying” of the man with a man is literally a bed. In other words, don’t f**k a guy in his wife’s bed, an expansion of adultery. The point cited from personal correspondence is only to support a translation Loader gives on other grounds. If the point has to be made (see next paragraph), is the ethical wrong of failing to cite your informal source (in the latter case) greater than the ethical wrong of overlooking such horrific crimes?

So on the one hand, there are only three citations all on minor translation issues – or at least, Loader doesn’t make much of them. He doesn’t base a whole chapter on Joosten’s work. On the other hand, they are very minor points, which raises the question of whether making them was worth citing a voracious hoarder of violent child pornography. I myself had a citation of Joosten’s pre-publication article on Leviticus in a piece I was working on that I simply removed because… ew. It would have supported my reading, sure, but there were other ways to make the argument. I won’t be adding a twelfth citation to Google Scholar’s list.

In these discussions, the idea has also come up of simply acknowledging the conviction to mitigate the additional professional status given by scholarly citations and to avoid any appearance of validating or ignoring his crimes. “I got this point from so-and-so. It should be noted that this scholar was convicted of a heinous act, which he admitted and pled guilty to.” Something like that.

I don’t have the answers here. I don’t know that Loader failing to make these points is the solution, or acknowledging the conviction in a footnote. I can’t even definitively say that citing this sort of criminal a few times in 400 pages without comment is unequivocally bad. But given that he cites personal correspondence and has cited the article in question on five separate occasions in a year suggests that Loader simply overlooks the issue. I draw attention to it only as a blatant test-case for these sorts of ethical questions, and to note that I as a reader found it to be a distraction from Loader’s broader points.


[1] See for example Mark G. Brett, “Social Inclusion and the Ethics of Citation, JBL 140/4 (2021): 819-25, here p. 825.

[2] At least one citation is to a “forthcoming” article, so may have been cited before Joosten’s arrest and conviction, but then the essay is basically a blog post so easily emended.

Review of Matthew D.C. Larsen, Gospels Before the Book

Matthew D.C. Larsen, The Gospels Before the Book

Matthew D.C. Larsen’s Gospels Before the Book challenges New Testament scholars to reframe how they think about authorship and publication in the first century. Larsen examines how “publication” actually worked then, and how it differs from how it works now. The spread of the printing press in the West would not happen for thousands of years, but it fundamentally reset how we think about the moment of publication since there are handwritten (or hand-typed) notes and drafts before publication, and a new, distinct, and uniform product meant for widespread circulation after it. The point of publication becomes definitive: the text is now out in the world.

But that wasn’t the case in the first century.

In that early context, there were handwritten drafts prior to “publication” (the making-public of the text), and there were handwritten drafts afterward, maybe only a few produced at a time. People made their own copies of a text from others’ if they couldn’t find a copy to purchase from a vendor, and the author-produced manuscript would be indistinguishable from the “bootlegged” one in material form, even if changes were made to the content during copying – whether deliberate or due to copyist error.

Furthermore, the line from rough notes to finished text was more continuous and a bit fuzzier. Larsen highlights how writers shared notes with others, incomplete, rough texts for insiders who could fill in the gaps. These notes could then occasionally be published or used to create a new text without the writer’s permission (sometimes after he died). Authors could even request post-publication changes to people they knew owned their texts. As Larsen traces, texts that looked like notes—again, incomplete, rough in language and style, often lacking meta-textual elements like author or title—basically invited other writers to improve them, finish them off, and publish them. Sometimes the later author credited the original writer(s), sometimes he didn’t.

Larsen applies this understanding to the many changes made to the Gospel of Mark in the first century. Effectively, he presents the text of Mark as unpublished notes that were adapted and modified by multiple later writers, including:

  • The writers of at least four different endings to Mark, which seemed to lack a proper one: the shorter ending, the longer ending (Mark 16:9-20), the longer ending with the Freer logion, and the Codex Bobiensis ending.
  • The incorporation and improvement of the notes into the Gospels of Matthew and Luke (Larsen examines the former in much more detail).

Larsen also traces how there is little evidence for paratextual data like title or author until over a century later, how the language and style is unrefined, and how the stories are stitched together episodically with less development than in Matthew. His main contribution here is to examine Mark as a collection of notes grouped loosely by themes or key-words. Larsen finds five main blocks of material, bounded by inclusios:

  1. Mighty deeds and opposition (1:21-3:6) (keywords: sabbath, synagogue)
  2. Hearing and doing / sea and discipleship (3:7-6:30) (keywords: twelve, sending)
  3. Eating with insiders and outsiders (6:42-8:21) (keywords: bread, eating)
  4. Seeing clearly, following Jesus (8:22-10:52) (keywords: “on the way,” seeing)
  5. Jesus death and the fate of the temple (11:1-16:8) (keywords: temple, Jerusalem)

Larsen’s approach allows the reader to let tensions in characterization and christology stand. The collector had simply arranged his notes thematically but maybe hadn’t smoothed over the tensions and contradictions as the authors of Matthew and Luke later would.

For my part, I have been involved in a number of discussions regarding the “Synoptic Problem” lately, which seeks to understand how the gospel authors made use of each other’s texts. One issue in the Farrer Hypothesis (Matthew used Mark, Luke used Matthew and Mark) and Matthean Posteriority (Farrer, but switch Matthew and Luke) is the different treatments of Mark compared to the other text. E.g., in the Farrer, why does Luke use Matthew so differently than Mark? Part of the appeal of the Two-Document Hypothesis is that Matthew and Luke can treat “Q” more or less the same way they treat Mark. They’re consistent in this model. In either of the others, they aren’t.

Larsen’s study opens the possibility that the perception of a text as “authored” versus written might have played a role. Under Farrer, if Mark was perceived as unfinished notes, but the more refined Matthew was perceived as an authored, finished text, then Mark would be more open to revision and use—but to use Matthew in the same way would be to plagiarize, to appropriate a text assigned to an author. It would be to steal rather than to make use of raw materials.

It may even help to explain, again from a Farrer perspective, why Luke might break up Matthew’s well-formed blocks of teaching material and spread them throughout Mark. Effectively, Luke may be under pressure to do something original with the teaching material or he would only be copying Matthew. His patron, Theophilus, may not be happy with a Matthew rip-off, so Luke needs to reappropriate this “authored” material more extensively than the “unauthored” notes found in the text of Mark. But this is my take, not Larsen’s.

Overall, I greatly recommend Larsen’s book to people working on the gospels, especially in the area of the Synoptic Problem.

American Jesus-Was-A-Pussy Christian Theology

A little while ago, several of my Christian friends hate-shared a picture of a car with a seemingly confused sticker on the back:

The sticker reads, “If Jesus had a gun, he’d be alive today.” And yes, obviously the person’s from Florida.

The caption on the photo read, “In this age of proud ignorance, I challenge anyone to find a photo more ignorant.” My friend’s caption read, “It is because Jesus laid down his rights that he is still alive today!” Someone left a gif in the comments from the SNL sketch, “Djesus Uncrossed”

“Djesus Uncrossed,” SNL (NBC).

My comment on the original photo challenged the idea that the problem is ignorance. It is rather an alternative christology. To express how the actual Christians who actually think this way about Jesus despite how the cross is presented in the New Testament, I call this Jesus-Was-A-Pussy (JWAP) theology. It’s not terribly uncommon in the US, and it seems to be getting more popular.

The challenge, the stumbling block if you will, of Jesus in the NT is that he died like a thug on the cross — “thug” approximating how Romans would think of it. Like a slave or conquered foreigner, a low-life. And he does so willingly. Paul acknowledges how this fact causes trouble for his followers and takes time to counter it (1 Cor 1:23-25). Matthew even has Jesus reject the idea that he should fight what is happening despite having an army of angels to fight for him, while being arrested (Matt 26:53). More passages could be added (e.g. Jesus as a powerful but bloodied lamb in Revelation), so clearly a case could be made that Jesus — despite his divine power (according to Christian belief) — did not fail by dying on the cross.

But of course, many American Evangelical men think Jesus did fail on the cross based on their model of masculinity, whether they stick a sign on the back of their truck or not. In their “tough guy” masculinity, “men” are muscular and fundamentally autonomous. They don’t let anyone tell them what to do, even if it will save their lives or someone else’s. We found this out the hard way during the pandemic when asking them to wear a simple mask to keep themselves from infection, filling up rare hospital beds, but also to keep the vulnerable around them from getting infected and dying. This caused conservative men around the country to lash out violently. The idea that Jesus would suffer and die for others, for sinners and people not in the in-group, is just so out of step with their idea of masculinity. What is he, some cuck lib?

The Jesus of the New Testament was a pussy. He should have blasted every last Roman with laser cats or something. Pew pew! So they set out to create a new, non-biblical Christ who acts like a man, or would have if he had access to an AR-15 and 10,000 rounds of ammunition!

This isn’t anything new in American Evangelicalism. Warner Sallman’s famous and widely reproduced Head of Christ was painted because his patron was sick of skinny, weak Christs hanging like pansies on a cross. He wanted a more muscular, manly (and clearly white) Jesus.

Head of Christ, Warner Sallman

JWAP theology is just the natural evolution of that attitude. American Evangelical men openly reject a central image of Christ in the New Testament while claiming to be Bible-based Christians. It isn’t ignorance. It’s rejection. They don’t want your compassionate, caring Christ. They want someone they can go hunting with, someone who takes what he wants, someone who is Lord and makes sure you know it through force.

They want Emperor worship, just like some first century Christians did.

Or in NT symbolic terms, they want to worship the Beast, “For who can fight against him?” (Rev 13:4).

Review of Randall Balmer, Bad Faith: Race and the Rise of the Religious Right

Cover of Balmer, Bad Faith

I’ve just submitted a review for Randall Balmer’s Bad Faith: Race and the Rise of the Religious Right to Reviews in Religion and Theology, and as I often do, I’d like to follow up on some specific points here.

First, let me say that Bad Faith is a phenomenal little book—inexpensive, quick-paced and engaging, and it completely overturns a paradigm of rightwing political history with solid reporting. Balmer explodes the myth that the Religious Right emerged in response to the Roe v Wade decision in 1973. Balmer traces mounds of evidence for at “best” indifference to abortion, at “worst” active support for legal access to abortion among Evangelicals before and after Roe. Paul Weyrich, one of the architects of the Religious Right, tried and failed to make abortion an issue that motivated Evangelicals to go to the polls throughout the 1970s. Instead, groups like the Southern Baptist Convention continued to voice support for keeping the government out of women’s pants.

What motivated the rise of the Religious Right according to Weyrich, and backed by Balmer’s reporting, was the IRS denying tax-exempt status to private schools that discriminated based on race. So-called “segregation academies” sprang up in the 1960s after segregation in public schools finally ended. White Christians wanted desperately to continue isolating themselves and their children from Black students, but, you know, private schools are expensive.

So they wanted the federal government to subsidize racist education. When the Supreme Court affirmed the IRS’s right to deny tax-exempt status to schools that refused to serve all students, Weyrich latched onto the issue. He disingenuously reframed the issue as a matter of religious freedom (to segregate based on race), just as modern Southerners reframe the Civil War as a conflict over states’ rights (to deny people personal freedom based on race and exploit them for labor). This issue—the belief that White Christians should be able to deny Black people opportunities, and the government should help them pay for the privilege—motivated Evangelicals to vote out one of their own, Jimmy Carter, and to elect a pro-abortion governor of California to the presidency: Ronald Reagan.

During the 1980 campaign, Reagan dog-whistled about the IRS overreaching its authority to Evangelical audiences. He avoided the issue of abortion. As chief executive, Reagan failed to defend the IRS policy in the Supreme Court. Instead, after the tax-exempt status of private Christian schools was struck down 8-1 anyway, Reagan nominated the lone dissenter to become the Chief Justice of the Supreme Court. Reagan, who once referred on tape to Black people as “those monkeys from those African countries,” continued his racist dog-whistling and racist policies throughout his two terms, and Balmer draws a through-line of racism and racist dog-whistling from Reagan to Trump—the current darling of the Religious Right. The myth of abortion or religious (read: Evangelical Christian) freedom giving rise to the Religious Right emerges later as a way to mask the racist past of the political movement.

One point that stuck out to me appears in Balmer’s postscript. He notes that during the 1980 presidential campaign, Jimmy Carter was attacked for denying tax exemption to Bob Jones University, a still segregated school in South Carolina.

Did they enslave Black people in South Carolina? I can’t seem to remember.

Anyway, putting aside that the president generally does not exert direct control over IRS decisions like this, the claim was that it happened ‘on Carter’s watch,’ that four more years of Carter would mean further infringements on Evangelicals’ religious freedoms… to, ahem, discriminateagainstBlackpeoplewhyisitimportantwhattheparticularissuewas. As Balmer points out, Bob Jones University did lose its tax exempt status in 1976, but on January 19, 1976. Over a year before Carter was sworn in. This happened on Republican Gerald Ford’s watch, not Carter’s.

It reminded me of Republican criticisms of Barack Obama for bailing out the banks… a policy signed by George W. Bush. Yes, it happened in 2008, but the election is not until November, and the new president did not take office until 2009. Republicans knew people didn’t like the bank bailout, and they wanted that anger directed at Obama, facts be damned.

With Carter, the leaders of the Religious Right who made Bob Jones an issue undoubtedly knew that Carter had nothing to do with it, but it revved up the base and they needed Carter to take the fall. So the leaders pinned it on the wrong guy, and their followers did not bother to put in the 0.3 milliseconds of thought needed to figure out that their politico-religious leaders were outright lying to them, manipulating them into voting against an Evangelical who wouldn’t support their discriminatory positions because of his Christian faith, and for someone who would be more likely to force Black Americans in the South to subsidize the same private schools that wouldn’t even let them on the grounds. It’s not just disappointing that political leaders will try this sort of sleight of hand, but that it apparently continues to work even when google is a thing.

This is just one tidbit of the trail of racist manipulations and motivations Balmer tracks from the 1960s to the 2010s. Bad Faith is an excellent book, and I enthusiastically recommend it to any reader interested in the topics. The review in RRT has more details, so look out for that soon. In light of rightwing censorship of discussions of race across Southern states right now, I also recommend the book to any teachers / professors covering the role of religion in politics – and politics in religion.

Sounds Like BS 7: Abortion “in” the Bible

I’ve just posted a new episode of Sounds Like BS that addresses the fact that the Bible actually mentions abortion only once — well, depending on how we interpret the passage of course.

In Jeremiah 20, the prophet wishes he’d never been born, but specifically that the man who announced his birth had instead “killed me in the womb” (Jer. 20:17). If this refers to abortion (it’s not clear whether Jeremiah wishes the man had killed his mother while pregnant, which is a whole other thing), then it would be the only reference to the procedure in the Bible.

This silence on the issue of abortion leaves Jewish and Christian readers of the Bible on the hook for what they think about the procedure. There are NO clear passages condemning (or endorsing) abortion. So in the video, I look at six common ways that people read abortion into the Bible:

  • In Jeremiah 1, Yahweh says he knew Jeremiah since before he was born, which some anti-abortion readers interpret to mean that Jeremiah (and by inference through him, all fetuses) had full personhood in the womb. I’ve addressed this passage before here.
  • In Numbers 5, Moses describes a ritual designed to determine whether a woman has cheated if she’s not caught, which some readers interpret as a levitical priest forcing her to drink an abortifacient tea.
  • In Exodus 21, a case law seems to imply that fetuses are not covered by “a life for a life.”
  • More broadly, some argue that abortion is covered by the commandment not to murder, although murder language is rarely broad enough to include fetuses — and sometimes it’s so broad that it implicates Jesus and God in the deaths of fetuses if that’s how one chooses to read it.
  • Michal Gorman argues that condemnations of sorcery (pharmakeia) carry an oblique reference to abortifacient teas or vaginal suppositories. This is highly unlikely, although not impossible.
  • Finally, Clement of Alexandria draws out moral implications from Luke’s vocabulary, which gives us an excuse to look at ancient ideas of when life began.

The link to video is below, so please check it out and let me know what you think.

Here, I want to touch on something I didn’t have time to address in the already long video: arguments from silence. Maybe the fact that the Bible says nothing about abortion is itself meaningful?

The first type of argument from silence one sees out and about is that abortion wasn’t a thing yet by the time the biblical texts were written. Of course God wouldn’t condemn abortion if it hadn’t been invented yet — don’t want to give anyone bad ideas.

The claim is simply false. As I note in the video, we have evidence of medical abortions from Egypt dating to the time when–traditionally–the Israelites were still living there (ca. 1550 BCE). By the first century, there were many known ways to end a pregnancy, some of them more effective than others. In a medical manual written by Soranus of Ephesus (ca. 1st-2nd century CE), Gynecology 1.60-65, the physician discusses dozens of techniques to perform an abortion. These range from presumably ineffective means like wearing magical amulets to more likely sound ones such as physical trauma, exhausting a woman (e.g. having her work excessively), malnutrition and dehydration, medical abortions in the form of teas, ointments, or suppositories (previously forbidden in the Hippocratic Oath), and primitive surgical abortions. Most of these, especially surgical techniques, posed a physical threat to the woman, so they were options of last resort.

The best option was the use of abortifacient herbs. Herbal remedies constituted an early form of medical abortion, similar to the modern “abortion pill” (RU486, not to be confused with the “morning-after pill,” which is not an abortifacient but merely prevents implantation). Several herbal abortifacients have been found scientifically to be somewhat effective. Teas and vaginal suppositories were often safer and less traumatic than mechanical options. However, one had to depend on 1) knowing the herb would work, 2) knowing the proper dosage (herbs such as pennyroyal could be lethal if the dose were too high; Riddle, Eve’s Herbs 47), and 3) having access to it. The most sought out herb was silphium, which could not be domesticated and was difficult to grow outside of Cyrene (in modern Libya), so was also extremely expensive—worth its weight in silver at the time of the NT (Pliny, Nat. Hist. 19.15). Because it was such a good abortifacient, silphium was picked to extinction within a few centuries (see Riddle, Eve’s Herbs 35-63).

Nevertheless, miscarriages, even intentional ones were especially dangerous before hospitals, blood transfusions, or antiseptics were available. The safest option still carried significant risk, and condemnations of abortion in medical tracts focus on the threat to the health of the woman rather than the morality of the procedure. There is plenty of evidence that abortion was a widespread and well-known practice. For ignorance to carry the argument, one would have to show that abortion was hidden from the dozens of NT writers (although probably all male…) as they traveled around the Mediterranean and Near East. Perhaps, but it seems unlikely.

The second argument from silence speculates that early Jews and Christians were just so anti-abortion that the biblical writers had no need to condemn the act — obviously one wouldn’t want to get an abortion.

There are variants on this argument regarding the nature of Christian communities and why abortion would not be common among them specifically: e.g. they were largely Jewish, or Middle Eastern, or just ancient, and these cultures valued having children so much that they would never terminate a pregnancy. Or early Christians in particular desired children due to ethical or sub-cultural values. Or they simply practiced celibacy when they wanted to avoid having children…

The theories multiply like rabbits.

There are two counter-arguments. First, some early but non-biblical Christian texts do explicitly forbid abortion, including the second century CE texts, The Didache (which may contain sections of early material) and The Epistle of Barnabas. Of course, they forbid the procedure, so we might say they provide evidence that early Christians would not receive or perform abortions. The rules may have functioned as boundary markers: they terminate pregnancies; you will not. But it may also have served as an exhortation, in the same way that rules against lying or sex before marriage do: we expect some of you to violate these rules from time to time, but try not to. They define what we should not do, not what we don’t do. Which purpose the rules serve is not immediately obvious (and they could serve both). What is evident is that at least some early Christians felt the need to address the issue, which only highlights that none of the familiar texts did.

Second and related, we find many condemnations of behaviors in which we would not expect virtuous Christians to engage. Consider Paul’s vice lists, places in his letters where he rattles off things that followers of Christ should not do. Among them:

Galatians 5:19-21: The works of the flesh are evident, which are sexual immorality, impurity, licentiousness, idolatry, sorcery, enmities, strife, jealousy (or zealotry), rages, contentions, dissensions, factions, envies, drinking, partying, and things like these.

Romans 1:28-31: God gave them up to a depraved mind, to do things that are not fitting, being filled with unrighteousness, wickedness, greed, malice; full of envy, murder, strife, deceit, maliciousness; gossips, slanderers, God-haters, insolent, arrogant, boastful; inventors of evil things, disobedient to parents, foolish, untrustworthy, heartless, unmerciful.

Is the argument that early Christians were sexually immoral, idolatrous, factious drunks who worked as prostitutes, murdered people, and hated God—but would never get an abortion? It seems unlikely that Paul only addresses behaviors the community actually engaged in. Notice in these cases, Paul does set up boundary markers: we don’t do these things—they do. The language is not descriptive, but normative. Paul takes the time to condemn anger, drinking, gossip, being untrustworthy, wanting what someone else has, being disobedient to one’s parents… but not abortion. Entire occupations in America are devoted to gossip (think tabloids or TMZ). Indeed, according to a 2015 LifeWay poll, many women do not seek counselling from their churches when deciding whether to end a pregnancy because they fear they will be the subject of gossip in their churches. A good number of our political leaders openly practice many of these vices without any punishment at the polls for creating factions, being boastful or arrogant, or showing rage.

We might argue that abortion was simply so horrible that it would have been obvious to Paul’s audience that it was vicious. There would be no need to include it in vice lists, much less discuss it at length. So it is worse than murder? Than hating God? This misses the point of the vice lists. Such things are included because they are horrible, because they are things we shouldn’t do. Yet Paul never bothers to condemn abortion as something horrible we don’t do. The silence could just as easily support an argument for indifference to abortion as for implicit condemnation. But I would not go that far.

That brings us to the third type of argument from silence floating around on the internet: maybe they didn’t condemn it because it was ok to get an abortion.

It is possible that NT authors say nothing about abortion because they simply did not care. In other words, they were indifferent to the practice, or they accepted it. If this were the case, there would be no reason to condemn abortion, nor even to bring it up, for the same reason they don’t condemn eating olives or enjoying a sunny day.

However, as I mentioned, the second-century Didache condemns abortion explicitly. The Christian text is very old, but it is not in the Bible. If the Didache is from a group in Syria close to the one that produced the canonical Gospel of Matthew, as some argue, then they cared enough about abortion to include it in a list of rules guiding the community’s behavior. They did not condemn abortion in the gospel they would send out to other communities, and which eventually did become part of the NT. Maybe as far as they knew, Jesus said nothing about abortion. Maybe they didn’t want to alienate outsiders. Maybe they thought their stance on abortion was implicit in the other rules they do have Jesus utter in Matthew.

We can’t say, so we shouldn’t.

The Didache gives evidence of at least one very early group that cared about abortion, but did not feel the need to press the issue into a gospel like Matthew. There may have been other cases of the same, so a pro-abortion argument from silence remains just as shaky as the others. What we can say in the end is that the absence of data, the silence on early Christian attitudes toward abortion is far too open to interpretation to dictate a moral position.

Hey, I Did a Thing: New article in JSNT, “The Eyewitnesses in Their Own Words: Testing Richard Bauckham’s Model Using Verifiable Quotations”

Pre-publication proofs of an upcoming article.

I just submitted the corrected proofs to the Journal for the Study of the New Testament testing the hypothesis Richard Bauckham lays out in Jesus and the Eyewitnesses. Bauckham contests the impression given by form criticism that stories about Jesus including his sayings changed quickly from telling to retelling during the oral phase, leaving very little that is historically reliable in the written gospels. His fundamental observation is that eyewitnesses to these events and people who learned the sayings directly from Jesus, like the disciples, lived for many if not all of the roughly 40 years between Jesus’s death and the first gospel (Mark). These eyewitnesses could act as a control on developments of their stories as either people came to them in Jerusalem, or they went out to other cities. So far, so good.

However, Bauckham bolsters his base hypothesis with three expansions that are a tad easier to contest:

  • First, he speculates that the disciples and other groups in the early church controlled Jesus traditions, actively working against developments. Someone who deviated too far from the story as eyewitnesses or authorized tradents remembered it would be corrected.
  • Second, they exerted particular control over the sayings material, as evident in the similarities in the sayings of Jesus between gospels even when other narrative details vary. He speculates that the disciples may have trained to memorize Jesus’s sayings or even write them down while he was alive so they could maintain them as accurately as possible.
  • Third, the gospels reflect the way the eyewitnesses told these stories since the authors were all either eyewitnesses themselves (John), close to eyewitnesses (Mark and Matthew), or consulted eyewitnesses and tradents as they composed (Luke and Acts).

In other words, people like the disciples consciously memorized Jesus’s sayings, worked hard not to change them as they told and retold the stories, and these habits are reflected in the gospels and Acts.

In the article, I do make some criticisms of Bauckham’s (sometimes unspoken) assumptions in drawing these three hypotheses, but my main goal is to test them. Instead of using intertextual data (comparing how the same story is told in different gospels), I use intratextual data: verifiable direct quotations, or places where a character directly quotes a statement that appeared earlier in the narrative. (Obviously, here I’m drawing on material that I examined in my book, Direct Internal Quotation in the Gospel of John). In the article, I ask whether the disciples are depicted as quoting Jesus (including influential sayings) more closely than other quotative scenarios.

The short answer is no. The disciples paraphrase Jesus when they quote him as much as they do anyone else (so Jesus is not an influential variable), and the disciples paraphrase Jesus when they quote him as much as other characters who would not be eyewitnesses do when they quote him (so the disciples aren’t influential variables either). There’s some nuance in there, which I note in the article, but basically all the characters feel free to alter statements that they quote directly, even when they are influential sayings spoken by Jesus.

Let me give one example of a prediction that Jesus makes about Peter, which appears in all four gospels but is only quoted directly later on in the Synoptics.

The writer who comes closest to quoting Jesus literally is Matthew, so let’s look at his case first. In Matt 26:34, Jesus tells Peter:

‘Amen, I tell you that on this night, before a rooster crows, three times you will deny me.’

After this happens, Peter remembers what Jesus said (Matt 26:75), and although the narrator re-quotes him, I agree with Bauckham that the phrasing is probably meant to reflect Peter’s recollection:

‘Before a rooster crows, three times you will deny me.’

The quotation is not exact: Peter subtracts “Amen, I tell you that on this night” from what Jesus said, but what remains reflects the phrasing Jesus used exactly.

Now let’s compare Mark’s pair of sayings:

Mark 14:30: ‘Amen, I tell you that you today, this night, before twice a rooster crows, three times you will deny me.’

Mark 14:72: ‘Before a rooster crows twice, three times you will deny me.’

Like Matthew, Peter’s recollection in Mark drops the introduction (here, “Amen, I tell you that today, this night”) but also changes the word order. The new word order emphasizes the rhyme of “twice” (dis) and “thrice” (tris), which is indeed appropriate to orally transmitted material — rhyming makes memorization easier — but shifts the phrasing slightly away from what Jesus is depicted as actually saying.

Luke’s Peter paraphrases more extensively:

Luke 22:34: ‘I tell you, Peter, a rooster will not crow today until three times you will deny knowing me.’

Luke 22:61: ‘Before a rooster crows today, you will deny me three times.’

Here the intro is dropped (“I tell you, Peter”), but there are other subtractions (“knowing”), transpositions, and extensive substitutions (e.g. “before a rooster crows” for “a rooster will not crow… until”). Luke’s phrasing is more similar to Mark’s and Matthew’s phrasing, but it is very different from what he depicted Jesus as saying earlier. There are few signs that verbatim recall of what Jesus said is important to Peter even when he is regretting that a prediction about him came true. And this is typical of verifiable quotations in all four gospels and Acts.

Now, this is perhaps just the way Peter told this story, and maybe we could credit the shifts in wording to Peter’s emotional state or to individual flaws. But in other cases, Peter “misquotes” Jesus in front of the authoritative bodies that Bauckham believes controlled the wording of these sayings, and yet they do not. As just one quick example, in the Gospel of John, Jesus tells the disciples in quick succession:

John 16:7-10: ‘When [the Paraclete] comes, he will convict the world … concerning righteousness, because to the Father I go (ὅτι πρὸς τὸν πατέρα ὑπάγω) and you no longer see me.’

John 16:16: ‘A little while and you no longer (οὐκέτι) see me, and again a little while and you will see me.’

Since Jesus says this a few hours before he’s killed, I agree with most scholars that Jesus is referring to his death and resurrection. Twice he emphasizes that they will no longer see him, evidently since they are currently looking right at him.

The second prediction of his departure throws the disciples into a tizzy, resulting in a fumbling string of quotations:

John 16:17-18: So his disciples said to one another, ‘What is this he says to us, “A little while and you do not (οὐ) see me, and again a little while and you will see me”, and, “Because I go to the Father” (ὅτι ὑπάγω πρὸς τὸν πατέρα)?’ Then they were saying, ‘What is this he says, the “little while”? We don’t know what he’s saying!’

The disciples are fairly close in either case, but each time they change Jesus’s phrasing in ways that reflect their own concerns rather than his. In the first case, they change his word order from “to the Father I go” to “I go to the Father.” Jesus emphasizes his destination, “to the Father.” They focus on the fact of his leaving, which is what has disturbed them for three chapters. Jesus does not correct them. They do not correct each other.

In the second case, they make a tiiiiiiny substitution, replacing Jesus’s “no longer” (οὐκέτι) see him with οὐ — they don’t see him. At a narrative level, John may do this to address the story level in which the disciples currently see Jesus but soon will no longer see him, and the audience who simply do not see Jesus but hope to “in a little while.” That way, a single saying can address Jesus’s death and resurrection, but also his current apparent absence on earth and his future return (“in a little while” is a common apocalyptic description of when the End will come). But this sort of application to the current situation of the church is what the form critics suggested was the cause of changes in Jesus sayings, and here it happens right in the text.

More to the point, Jesus doesn’t stop them! He does not correct their change in his word order in his first saying, and he even adopts their substitution in his second saying.

John 16:19: Jesus knew that they wanted to ask him, and he said to them, “Are you discussing among yourselves that I said, ‘A little while, and you do not (οὐ) see me, and again a little while, and you will see me’?

A number of copyists ‘correct’ Jesus or the disciples in different ways so the adverbs match, and the NRSV just goes ahead and ignores the critical texts, giving “no longer” in 16:19 so Jesus doesn’t ‘misquote’ himself. Yet not only does Jesus not correct the disciples when they quoted him to his face, he adopts their alternative phrasing. John fails to portray either the teacher with his students or the students with each other behaving the way his model requires. The early audience of John would have listened as the disciples altered Jesus’s phrasing over and over without judgment, without correction.

It is not to say that we should expect a one-to-one mapping of the historical disciples’ behavior onto their character representations in the gospels and Acts. Bauckham could be right and the writers of the gospels just didn’t care. But whatever the case, the disciples are not portrayed as the careful preservers of Jesus’s words that his model requires.

For more on this phenomenon and obviously more detail in my argument, check out the JSNT article when it comes out.

Postscript: I wanted to point out one case where a radical paraphrase of a Jesus saying is in fact corrected, something I only had time to fit into a footnote in the article. At the end of the gospel of John, we read regarding the Beloved Disciple:

John 21:22-23: Jesus told him, ‘If I want him to remain until I come, what is that to you? You follow me!’ So this saying went out among the brothers, ‘That disciple will not die [or: is not dying].’ But Jesus did not say to him, ‘He will not die,’ but, ‘If I want him to remain until I come, what is that to you?’

Jesus predicts that Peter will die a violent death, then scolds him for asking about the Beloved Disciple. “The brothers” interpret his hypothetical about the disciple remaining “until I come,” which only makes the needed rhetorical point if Jesus means his appearance at the End, as a declaration that the disciple will remain until the End. In light of that interpretation, they paraphrase Jesus as saying, “That disciple will not die,” and the gospel corrects this paraphrase as too far off the mark, not necessarily because it is too lexically distant from Jesus’s saying–that would be fine if the interpretation were valid. Instead, it is because they are wrong — either John has died or he is dying, but either way, he’s not immortal. Believing Jesus had predicted he would live forever is therefore dangerous. So John corrects the paraphrase by quoting more literally (except for, “You follow me!”).

To be fair, beliefs in John’s physical immortality continued for millennia despite this correction. Acts of John 115 (2nd century) has John assumed bodily to heaven so he “remains” in the sense that he does not die, even if he’s, you know, gone. Giotto’s Assumption of Saint John (1320) shows that the idea of John’s ascent survived for over a millennium.

Giotto, The Assumption of Saint John

Acts of John at least seems to acknowledge the correction in John 21, but other folk legends of John wandering the earth and popping up randomly at people’s houses survived into 19th-century America, where they were quite popular. In fact, these legends are canonized in Mormon scripture when three Nephites join old John in wandering the earth as immortals:

“ye have desired the thing which John, my beloved, who was with me in my ministry, before that I was lifted up by the Jews, desired of me. Therefore, more blessed are ye, for ye shall never taste of death; but ye shall live to behold all the doings of the Father unto the children of men, even until… I shall come in my glory with the powers of heaven” (Book of Mormon, 3 Nephi 28:6-7; cf. Mormon 8:10–11; DC 7; 78)

The Doctrines and Covenants expand even more:

And the Lord said unto me: John, my beloved, what desirest thou? For if you shall ask what you will, it shall be granted unto you. And I said unto him: Lord, give unto me power over death, that I may live and bring souls unto thee. And the Lord said unto me: Verily, verily, I say unto thee, because thou desirest this thou shalt tarry until I come in my glory, and shalt prophesy before nations, kindreds, tongues and people. And for this cause the Lord said unto Peter: If I will that he tarry till I come, what is that to thee? For he desired of me that he might bring souls unto me, but thou desiredst that thou mightest speedily come unto me in my kingdom. (DC 7)

A.E. Fife collected 19th-century stories about the 3 Nephites and John the Beloved roaming the American countryside, collecting souls as it were (“The Legend of the Three Nephites among the Mormons,” JAF 53/207 [1940]: 1–49).

Which highlights an important point: the wording of a saying is not as important as the idea it’s trying to convey. John corrected the wording. It didn’t matter. John’s still out there, like a benevolent Jason Vorhees.

Review of Robert McIver, Verbatim and Gist Parallels Between the Gospels: Coded Greek Synopsis and Selected Statistics (2nd edition)

I’ve recently submitted a review of Robert McIver’s Verbatim and Gist Parallels Between the Gospels: Coded Greek Synopsis and Selected Statistics, a multi-faceted synopsis of the four biblical gospels, to Review of Biblical Literature. There are several distinct features of the synopsis that contribute greatly to considering the text in new ways.

First of all, there are several synopses focusing on different types of parallels. McIver begins with a synopsis of the Synoptic Gospels (Matthew, Mark, and Luke). He pairs off passages, so if there are three parallel versions of a story, one will see a block pairing Matt / Mark, Matt / Luke, and Mark / Luke. McIver also underlines exact matches in wording (under a rubric he describes in the introduction), highlighting verbatim parallels and making it easier to see passages that are closer or more distant in phrasing.

Importantly for me, he includes tables of statistics based on his counts of words and verbatim agreements. For example, in the first table, he lists:

  • His passage numbers
  • The passages compared
  • The number of words in each text
  • The number of verbatim agreements in the text
  • The percentage of each passage made up of verbatim agreements
  • And the lengths of the longest sequence of verbatim agreements if one allows for zero, one, two, or three differences.

McIver presents the material this way based on his understanding of how to differentiate literary dependence (marked by long sequences of verbatim agreement) and mnemonic dependence (marked by high percentages of verbatim agreements, but no long sequences). He then adds tables that focus on specific texts and specific features, like the long sequences of verbatim agreements — which should help researchers interested in a specific text or in a specific type of agreement between them.

He follows with a synopsis of John paired with Synoptic passages, and tables relating the same features.

McIver also includes synopses focused on other features, like internal parallels (say, the feeding of the 5,000 / 4,000), or only the texts with the highest gist parallels or longest verbatim sequences. Again, not only are these portions useful for people with particular research interests, they got me thinking quite a bit. For example, the passages with long verbatim sequences often only have two parallels, while all but one of the passages with high percentages of words in common but no long sequences of verbatim matches had three or more parallels. Does this align with McIver’s distinction between literary dependence and memory? Or does it suggest that literary competition is at play — was there more pressure to add variation when there are more versions of a story out there? I haven’t resolved how I think about this yet, but then again, I had never considered the question before reading McIver’s synopsis.

I will warn you: there are many typos and printing errors in the English portions of McIver’s book — mine even had one on the cover (see above). But so far I have seen none in the Greek, and I do not suspect that his statistics are off. The second edition came out just a year after the first, and I suspect the quick turnaround is to blame. You may not always agree with the way McIver counted, but there’s little reason to suspect his counts are off.

More details will obviously appear in the full review, but long story short: I would recommend McIver’s synopsis of the biblical gospels as a highly useful tool for studying gospel parallels. At about $16 for the paperback volume and much more compact than many synopses, it is also worth considering if you are using it in a college course on the gospels.

Jermo van Nes Applies Regression Analysis to the Question of the Disputed Letters of Paul

I’ve just come across an article written by Jermo van Nes from 2018 in ZNW, “Hapax Legomena in Disputed Pauline Letters: A Reassessment.” Here, van Nes uses regression analysis to address the use of unique words in determining which of the letters credited to Paul were written by the apostle, and which were forged. The article is creative, measured, and should be taken seriously by Pauline scholars.

So what’s the issue?

Thirteen letters are ascribed to Paul in the New Testament, but scholars do not always trust that all of them were actually written by Paul himself. Some may have been written by a disciple of Paul’s in his name or on his behalf, and others just blatantly forged. So Paul’s letters are divided into two camps: the “undisputed” letters (Romans, 1 and 2 Corinthians, Galatians, Philippians, 1 Thessalonians, and Philemon) and the “disputed” letters (the other six). Hebrews doesn’t count: it’s anonymous and almost no one thinks Paul wrote it.

There are many reasons to question the authenticity of the disputed letters (some more than others), but one reason has been the frequency of hapax legomena, or unique words in those six letters. Van Nes defines hapaxes as “words that are univocal in meaning and that are used exclusively, whether once or more, in one Pauline letter and not in any of the other twelve” (p. 122). The claim often made by Pauline scholars is that the relative frequency of such words is much higher in the disputed letters than in the undisputed.

In other words, the vocabulary of the disputed letters is frequently unlike anything else in the letters we are certain were written by Paul. Van Nes seeks to test this proportionally, and to create a baseline, he uses regression analysis.

Van Nes counts the total number of words in the undisputed letters as his x-variable, and correlates it to the total number of unique words in the same letter as his y-variable. There is a very strong correlation between the two. The correlation isn’t perfect—the dots are not all lined up perfectly in a row. But none of them strays far from the line, either.

Once we have a best-fit regression line, we can also compare other results by creating confidence intervals around the regression line. Any point within these confidence intervals shows typical variation of the type that is evident in the original seven letters. Any point that outside the confidence intervals is significantly different from the undisputed letters, whether using too many unique words (if above the interval) or too few (if below).

I will only put van Nes’s first graph showing the undisputed here:

As you can see, none of the disputed letters is significantly different from the undisputed letters in terms of the use of the unique words except for 1 and 2 Timothy, which use significantly more unique words. On the one hand, that seems to put to rest the use of unique words to argue againt Pauline authorship of, say, 2 Thessalonians (which uses unique words slightly less often than average). On the other, van Nes does not let the issue rest for 1 and 2 Timothy either, instead questioning some ways that his method may be over-counting unique words and reanalyzing once these are subtracted away. I’ll let you read his article for the rest.

There is a tendency in biblical studies to make quantitative claims (e.g. Ephesians wasn’t written by Paul because there are so many uncharacteristic words) without carrying out any sort of quantitative analysis. Van Nes tests whether the use of unique words can differentiate between the disputed and undisputed letters, and his use of regression analysis shows that it cannot.

Now, van Nes does overstate the case when he says, “This type of analysis has not been applied to New Testament studies” (p. 123).

(After all, I had applied regression analysis to arguments from order for Q five years prior to van Nes’s article, but I get what he’s saying.[Point of Clarification: I’m not accusing van Nes of anything untoward — he is the first as far as I know to use regression analysis in this particular way to answer questions about the authorship of a NT work; his phrasing simply gives the impression that he is the first to apply regression analysis to any question in NT studies, which is not the case–I wasn’t even the first])

But otherwise, I think his critical testing of the claim that unique words is enough to differentiate using regression analysis is spot on.

Now, I would add that we’ve looked at the use of unique words here before with regard to the Story of the Adulteress in John, and we’ve seen that unique words is not a sensitive enough metric to differentiate some non-Johannine passages from John. This is consistent with van Nes’s data, but it warns us that just because the disputed letters are not significantly different in this regard, it does not follow that they were written by Paul. Instead, scholars who advocate for the pseudonymity of the six disputed letters need to refine their arguments and their methods. Hopefully van Nes’s article will spur them in the right direction.

Review of Joshua W. Jipp, The Messianic Theology of the New Testament

I’ve just submitted a review of Joshua Jipp’s The Messianic Theology of the New Testament to Reviews in Religion and Theology. The book is a theology of the New Testament, which Jipp expresses some reserve about even writing since the many texts of in the NT evince such diversity in their theologies and the way they express them. Still, Jipp argues that Jesus’s role as a royal, Davidic, and anointed figure provides a consistent framework throughout (most of) the books of the NT*. He divides his study into two parts: first, a series of exegetical chapters moving his way through the four gospels (with Acts attached to Luke), the works of Paul, Hebrews / 1 Peter / James, and Revelation. Then a second part unpacks the theological (really christological, but hey, Nicea) implications of the NT’s messianic christology.

The review in RRT will carry much more detail of course, and overall my take is very positive. Jipp writes well—he’s clear and accessible, and he makes a strong case for royal, Davidic motifs running through the NT. One really couldn’t ask for more out of him.

But there are some weak points in his reasoning, or rather, points where he pushes the evidence further than it warrants. And often this is due to a failure of Bayesian reasoning. So here, I’d like to highlight this particular weakness in Jipp’s reasoning, again not because he wrote a bad book or because overall he’s wrong, but only because attending to this flaw would strengthen his argument and leave him less open to attack.

When I say Jipp occasionally fails at Bayesian reasoning, I mean he uses a high conditional probability of one claim to imply a high conditional probability for the converse, which does not logically follow.

Wait, what’s that now?

Part 1: Um, what’s Bayesian reasoning?

When interpreting the meaning of a biblical passage, there are often a range of possibilities. Generally, we’d like to opt for the most likely meaning from whichever perspective we’re approaching the text, i.e. the meaning with the highest probability.

Okay, so an empirical or experimental probability of an event is simply how many events turn out the way you want them to, out of how many total ways the event could turn out. The probability that a dog is a mammal is 100% because you would divide the number of dogs that are mammals by the number of dogs—and these numbers are the same. Meanwhile, the probability that a mammal is a dog is very small since you divide the number of mammals that are dogs by the total number of mammals—and that second number is much, much bigger.

A conditional probability shifts from the empirical probability because you’ve got additional information. So, if we wanted the probability that a mammal is a dog given that it is a pet, the numbers shift up dramatically, maybe even to a plurality when compared to all mammal pets. If you already know someone owns the mammal as a pet, it is simply much more likely to be a dog than if all you knew was that it was a mammal, when it could be a gorilla, whale, elephant, etc.

What Bayes recognized (and formalized) was that the direction of conditionality matters. The probability that a mammal is a dog given that it is a pet is relatively high, but not extremely high. There are cats, hamsters, gerbils, teacup pigs, all sorts of pets that are mammals and not dogs. Say the probability that a mammal is a dog given that it is a pet is around 40%.

Now let’s reverse it. The probability that a mammal is a pet given that it is a dog is very high, certainly in the US, where the overwhelming majority of dogs are pets. Even if we discount dogs in shelters or stray dogs, we might estimate that the proportion of dogs that are pets is closer to 90%.

If you knew the mammal was a pet, you have a fair chance it’s a dog (40%).

If you knew the mammal was a dog, you have a really good chance it’s a pet (90%).

The two conditional probabilities do not give the same result, and in this case we might be as much as 50% off confusing the one with the other.

What Bayesian analysis cautions us against is that confusion. We have to attend to the direction of conditionality because it is easy to get the two mixed up. After 9/11, many Americans were afraid of Muslims. In their heads, the probability of being Muslim given that someone was a terrorist was very high. But if all you know is someone is a Muslim, then the informationt that is given is that he is Muslim, not that he is a terrorist. You’re assessing the converse, the probability that someone is a terrorist given that he is Muslim, which was abysmally small. Even if you imagined there were 1,000,000 Muslim terrorists in the world (there weren’t, but say you believe this), then the probability that a person is a terrorist given that he is Muslim would be 0.1%, or 99.9% of Muslims you met were not terrorists. There was no reason to fear a Muslim was a terrorist even if 100% of terrorists were Muslim!**

Part 2: Bayesian reasoning applied to biblical studies

I’ve discussed Bayesian reasoning in this context before, and Christoph Heilig uses it much more thoroughly than I do in his biblical work.

It came up in a recent video I did on Paul’s use of malakoi in 1 Cor 6:9, and how translators render this as related to homosexuality somehow. These translators rightly note that the probability that a man would be called soft given that he bottoms during anal sex with other men is high, in fact almost certain. So they translate malakoi as something like “homosexuals.” But we are not given the information that these men engage in certain homoerotic practices. We are given the information that they are called soft. So the real question is: what is the probability that a man bottoms in sex with other men given that he was called soft? That probability is rather low. As I detail in the video, Paul could be commenting on a number of nonsexual vices (lack of restraint, being too rich, dressing too nicely, being too emotional, etc.) or heteroerotic sexual vices (masturbating, bottoming with a woman, being too attracted to women, etc.). If gay sex were the issue, he could be a sex slave or a male prostitute. He could even be a rape victim. So when we add up all the uses of “soft” as an insult, probably less than 10% apply to men who have sex with men. That’s not much of a basis to translate malakoi as homosexuals, and it reflects a failure of Bayesian reasoning.

We need to attend to the textual data as we have it, and ask: what is the probability of my reading given what the text says? We are on much less stable ground asking: what is the probability that the text would say this given my reading? My reading is not a given—I’ve imposed it on the text.

The text is the given.

Part 3: Jipp’s messianic reading of the NT

In the exegetical chapters, Jipp seeks out passages in various texts that associate Jesus with a royal, Davidic messiah. Many of these are solid—especially when Jesus is just called a king or son of David. But others are somewhat questionable.

One “sign” that Jesus is being marked as the Davidic messiah is the presence of the Spirit. Here, for example, is a line from Jipp’s exegesis of John:

“We have seen that Israel’s messianic traditions often depict the Lord’s anointed as endowed with God’s authority to enact his rule and that this relationship between God and messiah enabled the king to operate as an agent of God’s Holy Spirit.”

Messianic Theology, p. 128 (emphasis added)

In other words, Jipp notes that the probability that the Spirit will be invoked given that a text comments on a figure’s royal, Davidic status as Yahweh’s messiah is high. He then uses the presence of the Spirit as a sign that a text is commenting on Jesus’s royal, Davidic nature or role.

But what is given? And what is Jipp concluding? It is not given information in all passages that Jesus is the royal messiah, but rather that the Spirit hangs out with Jesus. Jipp is concluding that Jesus is marked as royal and Davidic. So what he wants is the probability that a text is commenting on a figure’s royal, Davidic role given that the Spirit is present.

That probability is substantially lower.

The Spirit is associated with judges (Judges 3:10; 6:34; 11:29, etc.), priests (e.g. Hag 1:14; Rom 15:16), and especially prophets (just effing everywhere). Sometimes the presence of the Spirit says nothing about the human acting in the scene, but about God’s actions there (e.g. Gen 6:3 of everyone). If we add all these up, the probability that the presence of the Spirit is commenting on someone’s royal nature does not make up a majority, maybe not even a plurality. So it becomes questionable to take the presence of the Spirit as pointing to Jesus’s royal role. It is coherent with that characterization, but not necessarily indicative of it.

Another lesser example of this is the idea of anointing as indicating a royal office. As Jipp recognizes, priests and other functionaries were anointed. Furthermore, he almost completely ignores Jesus’s actual anointings by women in all four gospels, preferring to understand Jesus’s baptism (by a dude?) as a metaphorical anointing, even though this language is never used of Jesus’s baptism in the NT. Presumably, Jipp understands the scenes with women as anointings for burial (Mark, Matthew, John) or for cleanliness / a good foot rub (Luke). He also gets into trouble when NT texts talk about the anointing of a believer. Granted, a major theme of Jipp’s book is that believer’s participate in Jesus’s kingly rule—but they are neither the King, nor the Messiah.

So, while I would agree with Jipp that language about anointing most often points to a royal office, it is not certain. I would leave open the possibility that a given text may point to Jesus’s role as a priest, for example, if that was an option on other grounds.

Finally, Jipp uses Jesus’s role as fostering (a better) righteousness, i.e. proper Torah-observance, as a messianic vocation. Two pieces of evidence support this ideal image of the king. First is “the Law of the King” in Deut 17:14-20, which says that “if” the Israelites ever choose a king, he may not 1) be a foreigner, 2) have a lot of horses, 3) have a lot of wives (take that, Solomon!), and he must have a copy of the Torah made by the priests to learn from so he can be an observant king. Granted, pretty much all the Davidic kings failed in this from David on down—barring Josiah—but it reflects the ideal king from a deuteronomistic perspective. Second, Jipp argues that the psalter pairs royal psalms with Torah psalms (e.g. Pss. 1-2; 18-19; 118-119), again associating the king with Torah observance. So, there is a strain of Jewish tradition linking the ideal messiah to proper observance of the Torah, providing an example of righteousness to his people.

But again, Jipp takes claims that Jesus, for example, fulfills all righteousness (Matt 3:15) as indicative of Davidic themes, not simply coherent with them. If Christians held to the deuteronomistic model, or more likely, were influenced by the psalms, they probably connected righteousness to Jesus’s Davidic character. But what we know is that some NT texts associate Jesus with righteousness. So if a figure is associated with righteousness, what is the probability that figure is Davidic? My guess is not terribly high. Priests are associated with righteousness, and the prophets continuously take on a role of criticizing the unrighteousness of kings—of guiding them to proper righteousness. In the first century as well, we have the emergence of the rabbis, who interpreted the Torah to guide the people to a better righteousness. I’m not sure of any royal, Davidic claims for the Qumranic Teacher of Righteousness, but cases for a rabbinic, prophetic, or even priestly role could be made. So having Jesus guide followers toward righteousness could portray Jesus as a rabbi (a title he also gets in the gospels; Matt 26:25, 49; Mark 9:5; 11:21; 14:45; John 1:38, 49; 3:2; 4:31; 9:2; 11:8; 20:16), as a priest, or as a prophet. The theme is coherent with an idealized Davidic role, but hardly indicative of it at a probabilistic level.

To be fair to Jipp, his case does not rest on these three issues. There is much stronger evidence for a Davidic core to NT christology. Furthermore, it is not as if the probability is zero that a text focuses on Jesus’s Davidic character given that it introduces the Spirit, or the theme of anointing, or of righteousness. I only mean to point out that on an individual level, the probability that each case is pointing to a royal theme is low, so not an indicator of this character, only coherent with it. Jipp would also be justified in highlighting the cases where one of these themes shows up in the middle of a bunch of explicitly Davidic language, making it more likely that the text is not suddenly switching to a prophetic or priestly image in the middle of all that. As I said above, I would recommend Messianic Theology of the New Testament to interested scholars and devotional readers.

*Jipp does not even address four texts—2 and 3 John, 2 Peter, and Jude—and others are only addressed to the extent that they are in the Pauline corpus.

**Now they’re more likely to be white, Christian American men with a gun fetish and a history of violence toward women, but since Americans know many of these types of men, they don’t confuse every white Christian American man with a terrorist. Still, even though the probability is high right now that a terrorist in America is one of these men, the probability is not high that one of these men is a terrorist.

Sounds Like BS, ep. 6: The Synoptic Problem

The Synoptic Problem comes down to this: if three students all turned in assignments as close in wording and content as the Synoptic Gospels, they’d all get hit for plagiarism. Many parts are just too similar. But with students, you can question them to find out what happened — who copied off of whom? The gospel writers have been dead for centuries, so we can’t ask them. Instead, gospel scholars try to piece together who wrote first and who lifted whose material based on a close reading of the texts.

In this episode of Sounds Like BS, I present the Synoptic Problem in basically three stages.

Stage 1: Why do we think there is plagiarism in the Synoptic Gospels?

The Synoptics have similarly constructed scenes set within a very similar plot, told with very similar language. I try to demonstrate this by comparison with John. I know most Synoptic scholars will chafe at this — most would rather pretend John doesn’t exist. But John is useful in that it is 1) a first-century gospel 2) from within the same social circles that 3) nevertheless tells Jesus’s story in its own way. In other words, there evidently wasn’t overwhelming pressure to tell Jesus’s story in the particular way that Mark or Matthew does, but all three Synoptics choose to anyway.

Stage 2: How many relationships could there be?

I set out two initial assumptions here.

Assumption 1: The gospels share a literary relationship. That means I basically dismiss similarities due to oral tradition or memory. Close agreements in wording are best credited to one author reading another text. (That doesn’t mean that oral tradition or memory played no role, and it doesn’t mean that they only played a role where texts disagree, just that according to most Synoptic scholars, there is a literary relationship in there somewhere).

Assumption 2: I (mostly) ignore hypothetical (but now lost) source texts or earlier drafts of gospels. One benefit is that we avoid complex and unfalsifiable scenarios like Boismard’s.

The main benefit is that the number of possible solutions balloons beyond control, so it’s better to keep the possibilities to graphs involving just the three Synoptic Gospels.

Unfortunately, on just these assumptions, there are still 37 possible graphs, including sequential cases (where, say, Mark copies off of Matthew, and Luke copies off of Mark without ever seeing Matthew), independent cases, cumulative tradition cases, and collaborative cases (where two or more writers mutually influence each other as they work).

Field of 37 possible solutions to the Synoptic Problem

So we add a third assumption.

Assumption 3: Mark wrote first. This position is now reaching a consensus in Synoptic scholarship, and I detail a couple of the less technical reasons for that. This lowers our number to eight possibilities, with three of them the main contenders today.

From left to right: The Farrer model, the Matthean Posteriority model, and the Two-Document model (minus one of the documents).

Stage 3: The Three Main Contenders

Here I briefly look at the strengths and weaknesses of each of the models.

For the Farrer Hypothesis, it fits all three of our assumptions and helps to explain both differences and similarities between Matthew and Luke. Yet Luke’s editorial work on Matthew does not leave many Synoptic scholars satisfied.

For Matthean Posteriority, it also fits all three of our assumptions without the esthetic problems. Yet it hasn’t received much attention, positive or negative, which means we most likely haven’t found the major issues with the model yet.

For the Two-Document Hypothesis, it explains the major differences between Matthew and Luke, especially their different beginnings and endings, and their different placements of teaching material into Mark. However, it does require a fourth document (or series of small documents), which they call Q. In other words, it violates one of our assumptions.

In the next episode, I hope to talk with some Synoptic scholars from each of these camps to get a deeper understanding of why they find their solution the best, what weaknesses they find in the other solutions, and why it’s important to actually solve the Synoptic problem.

Here’s the video: