Earthquake Research and the Day of Jesus' Crucifixion

Usually speaking, these stories crop up around Easter so it's a surprise to see this one appearing now.  As usual on these occasions, it's a shame that journalists did not check in with Biblical scholars before going to press with a problematic article.  Here's the piece on Discovery News:

Quake Reveals Day of Jesus' Crucifixion

It's been debated for years, but researchers say they now have a definitive date of the crucifixion.

Jennifer Vegas

The article focuses on a detail in Matthew's story of the crucifixion:

Matt. 27.51: And behold, the veil of the temple was torn into two, from top to bottom.  And the earth shook and the rocks split open.

Geologist Jefferson Williams has investigated earthquake activity to see if he can use it to pinpoint the date of Jesus' crucifixion:

To analyze earthquake activity in the region, geologist Jefferson Williams of Supersonic Geophysical and colleagues Markus Schwab and Achim Brauer of the German Research Center for Geosciences studied three cores from the beach of the Ein Gedi Spa adjacent to the Dead Sea. 

Varves, which are annual layers of deposition in the sediments, reveal that at least two major earthquakes affected the core: a widespread earthquake in 31 B.C. and an early first century seismic event that happened sometime between 26 A.D. and 36 A.D.

Now I should say at this point I have not been able to consult Williams's work directly, but from this report, it becomes apparent that the research pinpoints the crucifixion to the decade between 26 and 36.  Given that the traditions locate Jesus' death during the time when Pontius Pilate was prefect, and given that Pilate was prefect from 26-36, this is not big news.  In fact, the precise correlation between Pilate's governorship and the window for the earthquake seems so striking that I wonder whether there is some confusion over the reporting.

Nevertheless, the article goes on to pinpoint the date of Jesus' crucifixion using other means, primarily the work of Colin Humphreys and Graeme Waddington, which dates back to 1985 (see The Date of the Crucfixion).  At least on the basis of the reports in the article, then, the earthquake research is only able to locate the crucifixion within the ten year period 26-36, but the older Humphreys and Waddington article is required for more precision.  

Several bullet-points are offered in order for the pinpointing:

• All four gospels and Tacitus in Annals (XV,44) agree that the crucifixion occurred when Pontius Pilate was procurator of Judea from 26-36 AD.
• All four gospels say the crucifixion occurred on a Friday.
• All four gospels agree that Jesus died a few hours before the beginning of the Jewish Sabbath (nightfall on a Friday).
• The synoptic gospels (Matthew, Mark, and Luke) indicate that Jesus died before nightfall on the 14th day of Nisan; right before the start of the Passover meal.
• John’s gospel differs from the synoptics; apparently indicating that Jesus died before nightfall on the 15th day of Nisan.

The "all four gospels" is the kind of thing that might sound impressive to someone not acquainted with scholarship on the Gospels because it gives the impression of multiple independent attestation.  However, it is consensus in New Testament scholarship that Matthew and Luke knew Mark and were dependent upon Mark for their crucifixion narratives, so this is not independent attestation.  Views differ a little on John, but many (like me) think that John knew the Synoptics too. 

The first three bulletpoints are taken over from the article by Humphreys and Waddington, which also uses the rhetoric of "all four gospels".  The latter two bullet points contain errors.  The Synoptics appear to place the death of Jesus on the day of Passover, 15 Nisan, and not on 14th.  They depict Jesus engaging in the Passover meal at sunset, when the day begins, and being crucified that same day.  John does indeed differ from the Synoptics, but not in the way claimed here.  John depicts Jesus' death as occurring not on the day of Passover (15th), but on the day before (14th).  So either Williams is confused or the journalist is confused or both.

Typically, the same errors are taken over without any checking in other versions of the report, e.g. the Daily Mail.

However, the real problem with this kind of work is that it fails to take seriously the nature of the texts that are being studied.  What Matthew appears to be doing here is to rewrite his Marcan source in typical Matthean fashion.  Of all the evangelists, Matthew is the one who likes to add earthquakes to his accounts.  There is one again at the resurrection (Matt. 28.2).  Indeed, one should probably be wary of using the term technical term "earthquake" to describe all of these -- it is the evangelist's way of saying that the earth was shaking and something dramatic was happening.  He describes the big storm (Matt. 8.23-27) as a great earthquake (seismos) on the sea.

To take Matthew's "earthquake" as a geological report is to misread his account.  The story he is presenting here is one of those that very few New Testament scholars would take seriously as history.  It's even read with caution by the most conservative scholars, and for good reason.  The Discovery report ends its quotation of Matt. 27.51-2 with the tombs opening, but if it had continued its quotation, the reader would have seen how Matthew goes on to recount what some people call the Zombie Pericope, when bodies come out of the tombs, walk around and meet people.  This is not history but legend.

Was the Last Supper a Passover Seder?

The most recent Bible Archaeology Review features an excellent article by Jonathan Klawans, "Was Jesus' Last Supper a Seder?".  I won't here summarize the article because it is available for all to read on the Bible Archaeology Review website but I would like to make a few comments on the details.  I will simply quote the passages I would like to comment on, and will then add my comments.

Of course a number of New Testament scholars—the Jesus Seminar comes to mind—tend to doubt that the Gospels accurately record very much at all about Jesus, with the exception of some of his sayings. Obviously if the Gospels cannot be trusted, then we have no reason to assume that there ever was a Last Supper at all. And if there was no Last Supper, then it could not have taken place on Passover.

Well, there is also the testimony of Paul in 1 Cor. 11.23-26, which most take as the earliest witness to some kind of Last Supper tradition.  Of course it does not help us a lot with the question of dating, though it parallels the Gospel traditions by locating the meal on "the night that Jesus was handed over".  And there are interesting Passover motifs elsewhere in the letter, especially 1 Cor. 5.7, "Christ our paschal lamb has been sacrificed" (which Klawans later mentions in another context).  The letter may have been written around Passover time since Paul wishes to stay at Ephesus "until Pentecost" (16.8).

While three of the four canonical Gospels strongly suggest that the Last Supper did occur on Passover, we should not get too comfortable based on that. The three Gospels that support this view are the three synoptic Gospels—Matthew, Mark and Luke. As anyone who has studied these three Gospels knows, they are closely related. In fact, the name synoptic refers to the fact that these three texts can be studied most effectively when “seen together” (as implied in the Greek etymology of synoptic). Thus, in fact we don’t really have three independent sources here at all. What we have, rather, is one testimony (probably Mark), which was then copied twice (by Matthew and Luke).

This is a common claim, and it is important to explain to newcomers that it indeed simplifies things to make it three vs. one.  Matthew and Luke know and use Mark's Gospel therefore they probably derive their dating from Mark.  All that is quite right.  However, some caution is necessary.  Eliminating Matthew and Luke as witnesses on the basis of their familiarity with Mark assumes that their only source of information was Mark.  That may be the case -- and I am very much a minimum sources man, especially when it comes to the Passion -- but we should bear in mind that Matthew and Luke may sometimes have been familiar with parallel traditions.  In other words, the likelihood that Matthew and Luke have Mark as a literary source does not automatically rule out the possibility of familiarity with other traditions.  They are sometimes willing to modify Mark in line with alternative traditions, for example Luke in his eucharistic discourse, which has parallels with Paul, or in his resurrection account, which has appearances in Jerusalem that cohere with the presence of the leaders in Jerusalem at a later date.

It's worth adding too that John may not be an independent witness either.  If John knows the Synoptics (the best evidence for which is, in my opinion, his account of the Anointing), we have to reckon with the possibility that he has simply adjusted the timing for theological reasons, with no parallel tradition about the dating.

According to John, Jesus died just when the Passover sacrifice was being offered and before the festival began at sundown . . . Any last meal—which John does not record—would have taken place the night before, or even earlier than that. But it certainly could not have been a Passover meal, for Jesus died before the holiday had formally begun.

This is a minor error here -- John 13.2,4 speak of a last supper, though indeed one that is before the Passover.

The article also features a sidebar,  but it does not render well in the web version -- it is practically incomprehensible.  What I would recommend here is Felix Just SJ's table, The Death of Jesus in Mark vs. John, which is clear and accurate.

But what of Klawans's major conclusion, that the Last Supper was most likely not a Passover Seder?  I think this may be right.  I have argued in favour if a liturgical origin of the Passion narratives in the Gospels, that they were developed in the context of worshipping communities.  This had the effect of fixing the dating of the narratives to the times when the communities were celebrating the events of Jesus' Passion, whether on 14th Nisan (the Quartodecimans, perhaps in continuity with some first century communities) or in a Passion vigil on the Thursday night / Friday after Passover, a custom that eventually came common but which may have had roots in the first century.  The difficulty with the celebration of events on particular days is that those celebrating quickly come to think of the day of celebration as the day on which the events happened.  My guess is that Jesus died at around Passover time, that this was remembered, that his death and resurrection were then celebrated around this time, but that the precise details were forgotten.

Klawans may well be right to lean towards the Johannine dating, though.  Although the Synoptics clearly locate Jesus' death on the Passover (Mark 14.12, 16, 17), there are remnants of the alternative view in Mark 14.2.  It is two days before the Passover (Mark 14.1) and the chief priests and scribes want to execute Jesus but they say, "Not during the festival, lest there will be a riot among the people".  I suspect that this note assumes a Johannine style chronology according to which Jesus was arrested and executed on 13/14 Nisan, the day before Passover.  The same underlying chronology is I think present in the idea that they were rushing to get Jesus' body buried in time for the Sabbath (Mark 15.42), an artificial idea if this is in fact Passover day, but a coherent notion if John is right that Sabbath and Passover coincided.

NT Pod 22: The Anonymity of the Gospels

I released the latest episode of the NT Pod at the weekend. The topic is The Anonymity of the Gospels.

This semester, I am teaching New Testament Introduction here at Duke and I am running the NT Pod alongside classes, providing episodes related to the topics that arise in class. This one came out of the lecture on the authorships and dates of the Gospels. It is a topic I have blogged on here before, and you can read some of my thoughts on it in a little more detail in The Dating Game VIII: John, Thomas and Authorial Self-Representation (cf. the disciples writing the Gospels), as well as in forthcoming publications, of course.

For the next couple of episodes of the NT Pod, I am planning to talk about the Synoptic Problem. Well, it had to happen sooner or later, didn't it?

SBL Dating Paper

I realized this evening that in the shake up of my websites and blogs earlier this year, I forgot to relocate my SBL paper on "Dating the Crucial Sources in Early Christianity". I have now done so:

Dating the Crucial Sources for Early Christianity (MS Word)

Dating the Crucial Sources for Early Christianity (PDF)

It's a draft so it remains "work in progress". I am not sure at this stage how long I will keep it on the web. I will almost certainly not publish it in its current format but will borrow some parts of it for expansion elsewhere.

I have noticed a few other things that have gone missing from the web in the transition, and I will be uploading those in due course too.

Death and the Dating of Thomas and the Gospels

Although we should be careful of tracking over-simple lines of inevitable evolution in our early documents, there are occasions where one can see a reflection of a document's general dating by oberving shifting perspectives.  I recently suggested that one such example is the ever increasing presence of authorial self-representation.  Another example occurs over the question of death in the early Gospels.  In Mark and Matthew, where death is envisaged it is violent death in the present.  And where they speak about the future, natural death is scarcely ever in view.  Instead, people are snatched away at the eschaton, or go to their judgement.  With the later Luke, though, natural death begins to appear, notably on two occasions in the L parable material, the Rich Fool (Luke 12.13-21) and Dives and Lazarus (Luke 16.19-31), both of which feature rich men dying, and not at the end of the age.  In Dives and Lazarus, the rest of the world continues on its ordinary way while the protagonists are in Hades and Abraham's bosom respectively.

Thomas, typically, is further along the same trajectory.  Although there are references to violent death (e.g. Logion 98), the references to natural death are now more common than they were in the Synoptics, as in these sayings (Lambdin's translation):

59: Jesus said, "Take heed of the living one while you are alive, lest you die and seek to see him and be unable to do so."

109: Jesus said, "The kingdom is like a man who had a hidden treasure in his field without knowing it. And after he died, he left it to his son. The son did not know (about the treasure). He inherited the field and sold it. And the one who bought it went ploughing and found the treasure. He began to lend money at interest to whomever he wished."

Logion 59 occurs in a cluster of material in which life and death is a key thread, from logia 58-61 and again in 63.   Logion 109 shows us that as in Luke, natural death is now a feature of the parable material.  Indeed Thomas's parallel to the Rich Fool (Luke 12.15-21 // Thomas 63) ends with the narration of the man's death ("that same night he died") rather than the death being implied in God's address, as in Luke.  

Perhaps the clearest example of the same phenomenon occurs in Thomas's version of a the double tradition saying Matt. 24.40-1 // Luke 17.34-5.  Luke's version of the Matthean saying is closest to Thomas's but both Matthew and Luke speak of people being "taken" rather than dying:

Luke 17.34-5: "I tell you, in that night there will be two in one bed.  One will be taken and the other left.  There will be two women grinding together.  One will be taken and the other left." 

Thomas 61: Jesus said, "Two will rest on a bed: the one will die, and the other will live."

Thomas, even more than Luke, comes from a time where natural deaths have found their way into the representation of Jesus' teaching.  It's one small sign among several others that Thomas belongs to a slightly later historical context.

Dating Mark After 70: Revisited

I am back in England for Christmas, but as usual I have brought my blogging machine with me and as time allows, I will continue to blog over the Christmas period, though probably with a little less regularity than usual what with eating, drinking and watching TV to do.
--

One of the central arguments in my recent paper on Dating the Crucial Sources in Early Christianity (Handout; Blog Series on Dating) was that Mark's focus on the temple, in which prophecies of its destruction, alongside a narrative climax that stresses its connection with Jesus' death, makes best sense in a post-70, post-destruction context. I suggested that arguments about whether or not Jesus actually prophesied the destruction of the Temple were largely beside the point. What I wanted to stress was the narrative function of prophecies like this in a text like Mark. It is all about the way in which the reader is led to recognize successful prophecy, and how that successful prophecy functions to legitimate the words of the speaker, and the text where they are found.

The inevitable difficulty, however, with an argument like this is that people do not actually hear the argument about narrative function and instead only hear phrases like "ex eventu", phrases that trigger a particular kind of response along the lines that "Jesus could have prophesied the destruction of the Temple", as if the argument had been "Jesus could not have prophesied the destruction of the Temple". Now if the discussion of the dating of Mark's Gospel is only allowed to constrain itself to the issue of whether or not Jesus could or could not have prophesied the destruction of the Temple, the argument is unlikely to move forward. It gets stuck on questions that unhelpfully draw in the writer's own prejudices about what Jesus could or could have done, could or could not have said. It is a bit like those discussions of Gospel miracle stories that get stuck on whether or not a given miracle could have occurred when the writer is attempting to reflect on its function in its narrative context.

What I am suggesting is that the way forward in this context is to by-pass the historical Jesus questions and to focus instead on literary context and narrative function, to notice that the clearest parallels to what is happening here in Mark are found in texts that post-date the events that are being prophesied. Josephus reports Jesus ben Ananias's prophecies of doom because they turn out to be accurate predictions of what in fact happened. The very point of narrating them is that the reader says, "Ah-ha -- they did not listen to Jesus just as they failed to listen to the prophets of old." Indeed, the story of the persecution of the prescient prophet (try saying that before breakfast) is one that provides a model for both Josephus and the Synoptic evangelists -- it is the old Deuteronomistic history's means of showing that the punishment of exile was an unavoidable consequence of the people's failure to hear the prophets' warnings.

I discovered Adam Winn's book about Mark late in the process of writing my paper on "Dating" and he has a helpful passage here that bears on the topic:

Much of the debate surrounding the authenticity of this prophecy has centered on whether is is an authentic Jesus tradition. The logic works in the following way: 'If it can be shown that this prophecy is an authentic Jesus tradition, it cannot be considered a vaticinium ex eventu and, therefore, Mark can be dated prior to the destruction of Jerusalem." But here, we suggest that this prophecy's identity as an authentic Jesus tradition is only indirectly related to Mark's date. Mark could have just as easily recorded an authentic Jesus tradition at a point after the temple's destruction as before it and doing so would make the tradition no less authentic. The days in which we concluded that Mark simply recorded all the tradition that was available to him are long past. We have come to recognize Mark as a creative and selective author who intentionally shaped his material. The prophecy then must be considered Mark's own prophecy that comes from either his (possibly authentic) sources or his own imagination. The focus of the debate over Mark's date of composition, therefore, should not be on whether this saying is an authentic Jesus tradition, but on whether Mark recorded (or created) this prophecy (essentially adopting it as his own) at a time before or after the temple's destruction (Adam Winn, The Purpose of Mark's Gospel: An Early Response to Roman Imperial Propaganda (WUNT, 245; Tübingen : Mohr Siebeck, 2008), 57-8).

It is, in other words, a question about the success of prophecy. The prophecies embedded in this narrative that stresses the destruction of the temple tell the reader about the prophet's authority. They are, by their nature, retrospective, celebratory, confirmatory of the speaker's authority and prescience. The point here is as it is in the Hebrew Bible: the prophets told them so, and just look at what happened.

Problems with studying memory in antiquity

On The Forbidden Gospels blog, April DeConick follows up her earlier post (discussed here, More SBL Dating Discussions) with a fresh post entitled SBL Memories 3: Become More Scientific. The first half of the post appears to be aimed at a kind of fundamentalist view according to which Jesus' words were recorded with verbatim accuracy in the Gospels, a view in which (of course) I have no stake or interest, so I will pass over it. About half-way in to the post, though, April turns to the section of my paper on the "missing middle" in Thomasine parallels with the Synoptics, drawing special attention to Thomas 57, the Parable of the Wheat and the Tares. April's disagreement with me here is based in part on a critique of an imagined model that I do not work with and which I regard as untenable:

What is the evidence that writers who have a literary document in front of them from which they are copying ever leave out the middle because they are rushed? Just based on logic, I would think that literary copying would be otherwise. That the copyist would be more careful to preserve the material he is using, that he is working slowly, that he can stop and go back and double check, and that he can erase and correct. Such is not the case, however, when an author is relying on human memory, when he cannot double check a written source.

Few writers who think that Thomas is familiar with the Synoptics are using a model of scribal copying, whereby the author of Thomas has a literary document in front of him. (Perrin may be an exception here, but see my comments on his work). The way I imagine the process is of familiarity with the Synoptics by means of memory through regular reading aloud (by himself or others).

April goes on to make several interesting observations about studying memory, and it is here that the real interest in the post lies. I will withhold any lengthy discussion until I have had the chance to read April's new article on the topic (I have the book in which it appears on order), but I will make a comment about the classic she mentions, F. C. Bartlett, Remembering: A Study in Social Psychology (Cambridge: Cambridge University Press, 1995; original ed., 1932). Bartlett's fascinating experiments with memory bear on the topic, not least given that one of his examples (72-3) features a "missing middle" similar to the one I have been illustrating in Thomas. It is worth mentioning that Bartlett was testing his subjects' memory of a written text. He gave his subjects a text called "The War of the Ghosts" and asked them to read the text a couple of times, and then he tested them for recall of the piece after selected periods of time, with interesting results, including a good example of an individual retelling the story without its middle section (in his first retelling, 20 hours after his reading of the text). I might claim that this coheres with my own view on Thomas's knowledge of the Synoptics, but alas, I have to confess that I can't help thinking that it does not recreate anything like the conditions that may be in view in antiquity, with communal texts read aloud by the literate to the community over a period of time, not a single unfamiliar text read by a modern individual and then recalled. Given my curmudgeonly scepticism on the transference of such studies to investigations into memory in antiquity, I am loathe to make anything of Bartlett's studies for our studies of issues like Thomas's familiarity with the Synoptics, even though they would help me, but I will concede one important point. Analysis of the way that moderns attempt to recall texts can at least stimulate our reflections on antiquity, even if that reflection ends up being about contrast more than comparison.

More SBL Dating Discussion

April DeConick responded to my paper, Dating the Crucial Sources in Early Christianity (Handout; Blog Series on Dating), at this year's SBL Annual Meeting, in the second meeting of a new consultation on Cross, Resurrection and Diversity in Early Christianity, and now she offers some useful reflections on that response on her Forbidden Gospels blog, SBL Memories 2: Dating Our Sources. One of the things I like about April DeConick's writing is that she often gets me thinking -- she has a great way of approaching subjects from a distinctive angle.

In the current post, April speaks of division in the academy and I think the implication is that we are on different sides of that divide. One side uses "older models" that are now "being seriously questioned" while April emphasizes "three major shifts in the field" that "must be taken very seriously" (though she goes on to enumerate four). In spite of the talk about division, there are actually several areas here where we are in agreement. In each of the four categories, I will begin with our agreement and then make clear where we differ. Let me add a quick word too on "older models", to use April's term. I too am in favour of questioning older models, though they are not always the same ones that April wishes to question; sometimes April works with some older models that I wish to question. And sometimes, of course, the old wine is good.

(1). As April mentioned at the session, she agrees with my post-70 dating for Mark and so too for Matthew, Luke and John. This is an important agreement because it establishes a working model for dating the crucial works, Paul well before 70, the Gospels after. April's sketch, however, expands to inclusion of hypothetical sources and earlier versions of documents, which I avoided in favour of discussing the materials to which our texts bear witness. Unlike me, April is conservative on the existence of Q, and even speaks of different versions of the text, and their provenance. I am also sceptical about the existence of kernel Thomas. I have not done enough work on James and the Didache to express a firm opinion on whether or not they post-date 70. So April's pre-70 block is much more richly populated than mine. I would love to be able to share her confidence in that area, but I remain sceptical about the survival of key materials from the earliest period.

A further difference is that the model I discussed in the paper was a genealogical one. Where April organizes documents into groups, I attempted to sketch sequence, Galatians post-dating 1 Corinthians, Matthew post-dating Mark, Luke post-dating Matthew, John post-dating the Synoptics and so on. The reason that this kind of work might be helpful is that it can map the evolution and development of ideas from one literary work to another. Working on relative dating in this way is tough because we simply have to get our hands dirty engaging in study that a lot of us would rather avoid, Synoptic Problem, Pauline chronology, John's relationship to the Synoptics and so on. But the potential pay off is major and the work is worthwhile.

(2). April's second point, about textual criticism, echoes my own warnings on this subject, which I am happy to repeat:

It is easy to engage in this kind of discussion without thinking through the broader issues of what it means to talk about “texts” and “literary works” in antiquity. It is a somewhat hackneyed to point out the obvious facts that none of the autographs have survived and that there were no printing presses, but textual critics rightly remind us to behave like we actually know that that is the case. Too often, we lapse into treating our scholarly constructs as if they are the actual artefacts that they can only aspire to be. At the very least, we need to keep reminding ourselves in discussions like this that we are not dealing with fixed points and known entities but with reconstructions and approximations. (3)

Nevertheless, with the appropriate cautions in place, it is also worth reminding ourselves that the texts are all we have. The situation is no different here than it is for any other set of ancient texts. We are dealing with manuscript witnesses. Indeed, in many respects the situation is a great deal better for scholars of early Christianity because of the relative earliness of the textual evidence as well as the richness of the manuscript deposit. We have to work with what we have, and what we have is pretty good.

April offers a valuable caution against "basing our conclusions on 'same' words here and there", but as I mentioned in the session in response, this is why we should also look at patterns of agreement and disagreement, tracing parallels in structure, order, theme, motif and imagery as well as the more minor parallels in wording. My own caution in this area, previously expressed in discussions of the Synoptic Problem, and especially of the Minor Agreements, is that one should be wary of appealing to conjectural emendation as a means of resisting texts that are difficult for one's theory, not least given the fact that absent textual evidence is as likely to have caused further problems for one's preferred theory as it is to have provided solutions.

(3)-(4): I will take these together since they are closely related. I sympathize with the desire for memory experiments but I am highly sceptical of our ability to recreate the necessary conditions for providing useful information on the way that memory worked in the first century. As I mentioned in the session, one of my favourite television programmes is Doctor Who, and in a recent episode, the doctor and Donna went to Pompei in 79. I would have loved to have joined them and to conduct some experiments there. But as I also mentioned in the session, there are indeed useful experiments that we can do, using the texts that we have. As some of my readers will know, I have been an advocate for developing tests on Synoptic (and other related) theories with a view to seeing whether they work or not. Given our current state of knowledge, and tools available, serious work on the ancient texts we have is preferable to experiments on our contemporaries.

Dating Sacred Texts on the Basis of Fulfilled Prophecy

In a recent post, the ever intriguing N. T. Wrong discusses the Scholarly Dating of Daniel to After the "Prophecies" were "Fulfilled". Here in the biblioblogging community, we are all on first name terms, so I hope the bishop will not mind my calling him Tom. Tom quotes and then argues against a character who sees "The practice of late-dating the books of the Bible . . . as a position of faith on the part of those scholars who do so"; Tom pays special attention to John J. Collins on Daniel, and the quotation is worth repeating here:

The issue is not whether a divinely inspired prophet could have foretold the events which took place under Antiochus Epiphanes 400 years before. The question is whether this possibility carries any probability: is it the most satisfactory way to explain what we find in Daniel? Modern critical scholarship has held that it is not.
- John J. Collins, Daniel, First Maccabees, Second Maccabees, with an Excursus on the Apocalyptic Genre (Wilmington: Michael Glazier, 1981): 11-12.

Tom's timely post coheres with what I have been arguing here (especially in Dating Game VI and Dating Game VII) with reference to the predictions of the temple's destruction in the Gospels. Allow me to quote from a section of my forthcoming SBL paper (31-32 in the current draft):

One of the standard arguments against the idea that Mark shows knowledge of the destruction of Jerusalem is the reassertion of the text’s own character here as prediction. In his Introduction to the New Testament, David A. DeSilva suggests that “The primary reason many scholars tend to date Mark’s Gospel after 70 CE is the presupposition that Jesus could not foresee the destruction of Jerusalem – an ideological conviction clearly not shared by all.” But this kind of appeal, while popular, tends not to take seriously the literary function of predictions in narrative texts like Mark. Successful predictions play a major role in the narrative, reinforcing the authority of the one making the prediction and confirming the accuracy of the text’s theological view. It is like reading Jeremiah. It works because the reader knows that the prophecies of doom turned out to be correct. It is about “when prophecy succeeds”.

My concern about the popular appeal to what Jesus could or could not have done is that it does not take seriously the real issue, which has nothing to do with making a judgement about the historical Jesus. Rather, it is about observing the literary function of successful prophecy in the narrative in which it appears. The prediction only gains traction because the reader is saying, "Hey, yes! I know what that's about!" The issue is parallel to the one discussed here by Tom Wrong, and I am grateful to have his Daniel discussion to inform my own. James McGrath weighs in on Exploring our Matrix, with similarly helpful observations.

SBL Paper on Dating the Crucial Sources in Early Christianity

My paper for a session at the SBL is now available online here. I was invited to write this paper for a new consultation on the Cross, Resurrection, and Diversity in Earliest Christianity (I have also been asked onto the steering committee). Session details are at the previous link, along with other papers for this consultation, and the paper is available here:

Dating the Crucial Sources for Early Christianity (MS Word)

Dating the Crucial Sources for Early Christianity (PDF)

Regular readers may spot some overlap with my blog sketches on the same topic, though the paper is longer and more detailed and on the whole post-dates the blog posts.

The Dating Game VIII: John, Thomas and Authorial Self-Representation

With the momentous events of last night still fresh in our minds, I hope readers will not mind my returning to the question of dating, as we approach the conclusion of this series.

In the most recent post, we looked for corroborating evidence that Matthew and Luke indeed post-dated 70, something that their dependence on a post-70 Mark would of course lead us to expect. In this post, I would like to turn to the Gospels of John and Thomas. Is there knowledge of the destruction of the temple here too? I think that there is, though their greater distance from 70 may be reflected in the fact that there are fewer references now to the destruction of the Temple. In Thomas’s case, this is also no doubt a function of its genre (Sayings Gospel in which narratives about the Temple are of course absent) and theological proclivity (the relative lack of so-called apocalyptic eschatology). Nevertheless, both texts allude to the destruction of the Temple, John in 2.19-20, “Destroy this temple and I will raise it again in three days . . . “ and Thomas in logion 71, “I shall destroy this house and no one will be able to rebuild it.”

The more blatant signs, though, of the relative lateness of John and Thomas lie in their attempts at authorial self-representation. Where earlier Gospels like Mark and Matthew are anonymous and avoid attempting to project an authorial presence to lend authority to their work, the author of the Fourth Gospel makes claims to have been present, most notably in 19.35 and of course 21.24, “This is the disciple who testifies to these things and wrote them down (καὶ ὁ γράψας ταῦτα). We know that his testimony is true,” similar in style and literary function to the Incipit of Thomas, “These are the secret sayings which the living Jesus spoke and which Didymos Judas Thomas wrote down.” In both, the authorial self-representation legitimizes the message of the book in a way absent from the earlier Gospels but found explicitly in later texts like the Apocryphon of James. John’s claim enables the author to establish his Gospel’s authority – he knows that the things he reports are true because he was there. In Thomas, there is a further step: the author was present and, moreover, he was privy not just to the public teaching but also the secret teachings (Incipit, Thomas 13).

There is a trajectory among these early Christian texts, from the absence of authorial self-representation in Mark and Matthew, to hints in Luke and Acts (with the first person found in Luke 1.1-4 as well as in the “we” passages in Acts), to the marked but nevertheless still unnamed authorial presence in John, to the explicit self-representation of Didymos Judas Thomas in its Gospel’s Incipit, a naming that also leads the reader to pay special attention to Thomas 13. The same texts likewise witness to a growing consciousness of predecessor texts, from the πολλοί of Luke’s preface, to the many other books that could fill the world in the last verse of John, to the twelve disciples sitting around writing their books at the Last Supper in the Apocryphon of James.

These observations depend in part on the work of Ismo Dunderberg, “Thomas and the Beloved Disciple” in Risto Uro (ed.), Thomas at the Crossroads: Essays on the Gospel of Thomas (Studies of the New Testament and Its World; Edinburgh: T & T Clark, 1998), 65-88, especially 80-88, though his use of the term "authorial fiction" (derived from John Kloppenborg) is not ideal. The term “authorial self-representation” is preferable because it characterizes the process more precisely and less prejudicially, and uses terminology familiar in literary criticism.

The Dating Game VII: Dating of Matthew and Luke

In the previous post in this series, we looked at the case for Mark's dating in the post 70 period, suggesting that the predictions of the destruction of the temple function to underline the authority of Jesus as the one who successfully predicted what the reader knew had now happened. The repeated and pervasive emphasis on the temple and its destruction is most plausible in this post-70 period. It is worth investing time on this question because if Mark was written after 70, and if Matthew, Luke and John are all familiar with Mark, then they too post-date 70. But does a closer look at the later Gospels correlate with this picture? For J. A. T. Robinson (Redating the New Testament), it was the lack of reference to 70 anywhere in the New Testament that proved decisive in his attempts at redating:

The fall of Jerusalem in AD 70, and with it the collapse of institutional Judaism based on the temple - is never once mentioned as a past fact. It is, of course, predicted; and these predictions are, in some cases at least, assumed to be written (or written up) after the event. But the silence is nevertheless as significant as the silence for Sherlock Holmes of the dog that did not bark.

The claim is unimpressive, though, given that most of the documents in question are either written in the pre-70 period (Paul’s letters) or set in the pre-70 period (Gospels-Acts). What is remarkable is that documents set a generation before 70 appear to speak so clearly about the destruction of the Temple. For Robinson,

That Jesus could have predicted the doom of Jerusalem and its sanctuary is no more inherently improbable than that another Jesus, the son of Ananias, should have done so in the autumn of 62.

The problem for this perspective is that Jesus ben Ananias’s prophecy occurs in a document that post-dates 70, Josephus’s Jewish War. As with Mark, it is important to ask the question about the literary function of the prediction in the narrative, here in a document that climaxes with the story of Jerusalem’s destruction. Indeed, a comparison between Jesus ben Ananias in Josephus and Jesus of Nazareth in Matthew and Luke provides further striking parallels. The oracle Matthew 23.37-39 // Luke 13.34-35 has marked similarities with the oracle in Jewish War 300-1, the same threefold focus on the people, the city, the temple. Jesus ben Ananias cries “a voice against Jerusalem . . .” and Jesus laments “Jerusalem, Jerusalem”. Jesus ben Ananias singles out “the holy house” and Jesus says “Behold your house is forsaken.” Jesus ben Ananias raises “a voice against this whole people” just as Jesus exclaims, “how often would I have gathered your children.” Moreover, the same context in Josephus features a portent of voices being heard in the temple saying “we are departing from hence” (μεταβαίνομεν ἐντεῦθεν, War 6.299), similar to the implication here in Matthew and Luke that God has left the temple – “Behold your house is forsaken and desolate” (Matt. 23.38). Such prophecies and portents function similarly in each of the texts and they point to a post-70 dating.

The Dating Game VI: Was Mark written after 70?

In the previous post in this series, we concluded by looking briefly at James Crossley’s commendable effort to rethink the dating of Mark. If that attempt is unsuccessful, it is nevertheless worth asking how secure the standard scholarly dating is. One of the values of challenges to the consensus is that they can send us scurrying back to the texts to think again about the issues and to reexamine our reasons for coming to particular views. My own thinking on the subject has been strongly influenced by three recent studies which successfully reinforce the grounds for locating Mark in the aftermath of 70, Brian Incigneri’s The Gospel to the Romans, H. M. Roskam’s The Purpose of the Gospel of Mark in its Historical and Social Context and John Kloppenborg’s article “Evocatio Deorum and the Date of Mark”. Although these three disagree with one another on the details (e.g. the precise referent of Mark 13.14), all agree on the significance of the key text:

Mark 13.1-2, Do you see these great buildings? Not one stone will be left upon another which will not be torn down.

For many, so blatant a prediction of the destruction of the Temple in Jerusalem settles the question of Mark’s date – it is written in full knowledge of the disastrous events of 70. For Kloppenborg,

“The fact that this seems to correspond so precisely to what occurred invites the conclusion that it was formulated (or reformulated) ex eventu” (431).

For Roskam,

“The evangelist could not have presented the prediction of the destruction of the temple as an utterance of Jesus with such firmness unless he was very certain about its fulfilment” (86).

Objections to this view are ably discussed by Incigneri (Chapter 3, "No stone Upon another"), who stresses Mark’s “over-arching concentration on the Temple” (154), the destruction of which is so important in his narrative that it is implausible that it was still standing when Mark wrote.

One of the standard arguments against the idea that Mark shows knowledge of the destruction of Jerusalem is the reassertion of the text’s own character here as prediction. To take one example among many, David A. DeSilva, in his Introduction to the New Testament, suggests that

The primary reason many scholars tend to date Mark’s Gospel after 70 CE is the presupposition that Jesus could not foresee the destruction of Jerusalem – an ideological conviction clearly not shared by all (196).

But this kind of appeal, while popular, tends not to take seriously the literary function of predictions in narrative texts like Mark. Successful predictions play a major role in the narrative, reinforcing the authority of the one making the prediction and confirming the accuracy of the text’s theological view. It is like reading Jeremiah. It works because the reader knows that the prophecies of doom turned out to be correct. It is about “when prophecy succeeds”.

The text makes sense as Mark’s attempt to signal, in a post-70 context, that the event familiar to his readers was anticipated by Jesus, in word (13.2, 13.14) and deed (11.12-21) and in the symbolism of his death, when the veil of the temple was torn in two (15.38). The framing of the narrative requires knowledge of the destruction of the temple for its literary impact to be felt. Ken Olson has alerted me (especially in a paper read at the BNTC three years ago) to the importance of Mark 15.29-30 in this context. It is the first of the taunts levelled when Jesus is crucifie:

So! You who are going to destroy the temple and build it in three days, come down from the cross and save yourself!

For the irony to work, the reader has to understand that the Temple has been destroyed; the mockers look foolish from the privileged perspective of the post-70 reader, who now sees that Jesus’ death is the moment when the temple was proleptically destroyed, the deity departing as the curtain is torn, the event of destruction interpreted through Gospel narrative and prophecy.

The point that is generally missed in the literature, especially that which comes from a fairly conservative perspective, relates to the attempt to understand the literary function of the predictions of destruction in Mark's narrative. John Kloppenborg is one of the few scholars who sees the importance of the literary function of the predictions, noting the role played by the literary motif of "evocation deorum" echoed here in Mark, e.g.

This raises a crucial distinction between omens and rituals that (allegedly) occurred before the events, and their literary and historiographic use in narrative (446).

Discussions about whether the historical Jesus was or was not prescient may be interesting, but in this context they miss the point. The theme of the destruction of the temple is repeated and pervasive in Mark's narrative, and it becomes steadily more intense as the narrative unfolds. Jesus' prophecies in Mark attain their potency because "the reader understands" their reference.

The Dating Game V: Document and tradition

So far in this series of posts, we have been looking at the relative ordering of the crucial documents, focusing on the sequence of the documents without attempting to pin them to particular decades. As far as the canonical Gospels are concerned, we are looking at an order like this: Mark > Matthew > Luke > John. The time comes, though, when we need to attempt to pin these texts to points in time. As anyone familiar with New Testament studies will know, the dating issue is determined by a pivotal question: do the documents post-date the destruction of Jerusalem in 70CE? Since Mark is the first in this sequence of documents, dating Mark would be a very helpful way of moving forward. If Mark post-dates 70, so do Matthew, Luke and John.

Before tackling that question, however, there are some necessary reminders. The discussion is inevitably clouded by the complications of textual tradition (observable) and textual tradition (hypothesized); I have spoken already (Preliminaries) about some of the difficulties involved with a document's evolution and range of dates and the inevitable difficulties that that causes the historian. Nevertheless, it is possible to speak reasonably about the dating of the documents as long as one bears these kinds of difficulties in mind. History, and especially ancient history, often needs to deal in approximations. It is a heuristic and not a descriptive discipline, and reasoned discussion of the date of given documents is achievable provided one proceeds with care.

It is important, for example, to distinguish clearly between the date of a given document and the date of the traditions within it and to avoid allowing document dating to get bound up with tradition history. How, then, should we conceive the question of dating a document? It should refer, I think, to the date of the given document as a observable, substantive entity with recognizable parameters such that it distinguishes itself from other documents. Matthew, for example, is recognizably Matthew and not Mark, even though it contains a lot of Mark. Luke is recognizably Luke; it is not Matthew and it is not Mark even though it contains a lot of the shape and the substance of those documents. In this kind of discussion, then, we need to be clear about what it is we are trying to date. We are dating the documents to which our texts bear witness, and not prior oral traditions, written traditions, or hypothetical earlier versions of the document in question. In this context, we are not investigating the dating of elements within the larger, later document; we are attempting to date the document itself.

A document can be no earlier than its most recent datable tradition. This is why, when we come to Mark, the question of its knowledge of the destruction of the temple is so important. If Mark is familiar with the events of 70, the presence of traditions earlier than 70 will be irrelevant. A good example of an approach that recognizes the distinction between the date of the document and the history of its traditions is Gerd Theissen's work on Mark's Apocalypse and Passion Narrative. He argues for versions of Mark 13, and of Mark 14-15, that date from the late 30s or early 40s, but thinks that Mark itself was written after 70.

It is in this context that I find myself reflecting on James Crossley's recent book, The Date of Mark's Gospel, published in the series I edit called Library of New Testament Studies in 2004. Crossley argues against the consensus that Mark should be dated somewhere in the region 65-75CE, suggesting instead that Mark's knowledge of Jewish Law, and the assumptions he makes about it, make best sense at a very early point, as early as mid to late 30s or early 40s. There are many things I like about James -- I am all in favour of young, attractive British blogging scholars who are willing to stick their necks out against the consensus on important issues, and who publish in LNTS. So I wish I were able to agree with the thesis of The Date of Mark's Gospel, but I don't. One of the book's virtues, I think, is that it effectively strengthens the case for a law observant Historical Jesus and Crossley's arguments to that end are effective. I am not persuaded, though, that James succeeds in narrowing the gap between Jesus and the author of Mark. As David Gowler points out in his review of James's subsequent Why Christianity Happened, "Jesus' Torah observance could still have been adequately represented by Mark in the 60s" (CBQ 69 (2007), 815-6 [816]). The notion that the originating circumstances of the tradition correlate directly with the perspective of the evangelist is problematic given the possibility that Mark is sometimes a faithful retailer of traditional material. Or, to put it another way, it is always going to be a tall order to demonstrate that assumptions apparently made in given traditions are identical with assumptions made by the author of the document in which they appear.

Moreover, where there are clear signs of Marcan redaction, they point away from Crossley’s thesis. In the key passage about hand-washing in Mark 7, the narrator’s framing of the material explains that hand-washing before eating food is something practised by “the Pharisees and all the Jews” (καὶ πάντες οἱ Ἰουδαῖοι). This does not set up the debate as an intra-Jewish one of the kind that Crossley’s thesis requires. The practice of hand-washing is established as something that all Jews do, and which Jesus’ disciples do not do (7.2, 5), setting up a contrast that Jesus’ words then speak into, a contrast that makes good sense on classic form-critical grounds. For Crossley, the reference here to “all the Jews” is a Marcan exaggeration, but this concedes the ground about the accuracy and precision of Mark’s knowledge of Judaism that is a major and necessary element in his case.

The Dating Game IV: What about John?

It is easy to find oneself spending so much time with the vexed question of the chronological sequence of the Synoptic Gospels, and the big Synoptic Problem questions that arise, that one can forget about the Gospel of John and the Gospel of Thomas. To an extent, the relative neglect of these Gospels in this context is understandable. Whereas the Synoptic Gospels are clearly related on the literary level, there is no consensus about whether the same is true of John and Thomas. Does John know the Synoptics? Some say yes, some no. Does Thomas know the Synoptics? Again, opinions are divided. Clearly, this series of blog posts is not the place to solve this fascinating problem, but I would like to suggest a couple of ways in which we might be able to sketch out the possible lines of relationship between the Synoptics and Thomas and the Synoptics and John. I am currently in the middle of a book on the relationship between Thomas and the Synoptics, so the full argument will have to wait for its publication in the (I hope) not too distant future. As for John and the Synoptics, let me just say that I am persuaded by the evidence set out by several including C. K. Barrett and Frans Neirynck concerning John's knowledge of the Synoptics and that I would like to add an observation that may be of relevance.

So let us go to John and focus on an issue related to the phenomenon of fatigue in the Synoptics, discussed in the previous post. One of the indicators of familiarity with prior texts is a rewriting of elements in those texts in such a way that the author inadvertently creates anomalies or inconcinnities. One of the clear examples of this phenomenon in John occurs in his story of the anointing of Jesus by Mary in John 12.1-8. The story is parallel to Matthew 26.6-13 // Mark 14.3-9. The Johannine incident is clearly the same as the Synoptic incident: (1) It takes place in Bethany (2) just before Passover, (3) at a dinner where a woman has a jar of very expensive perfume of pure nard (Mark 14.3, ἀλάβαστρον μύρου νάρδου πιστικῆς πολυτελοῦς; John 12.3, λίτραν μύρου νάρδου πιστικῆς πολυτίμου; (4) she anoints Jesus; (5) there are complaints about the costliness of the perfume (τριακοσίων δηναρίων) which could have been given to the poor (καὶ ἐδόθη πτωχοῖς, John 12.5; καὶ δοθῆναι τοῖς πτωχοῖς, Mark 14.5); (6) Jesus says "Leave her. . . The poor you will always have with you . . . But you will not always have me" (ἄφες αὐτήν . . . τοὺς πτωχοὺς γὰρ πάντοτε ἔχετε μεθ' ἑαυτῶν ἐμὲ δὲ οὐ πάντοτε ἔχετε, John 12.7-8; ἄφετε αὐτήν . . . πάντοτε γὰρ τοὺς πτωχοὺς ἔχετε μεθ' ἑαυτῶν καὶ ὅταν θέλητε δύνασθε αὐτοῖς εὖ ποιῆσαι ἐμὲ δὲ οὐ πάντοτε ἔχετε, Mark 14.6-7); (7) Jesus interprets the anointing in connection with his burial (John 12.7, Mark 14.8).

John appears to have crafted this account on the basis of the Marcan narrative; the structure, the story, the wording have substantial links. The only major fresh elements in John are the naming of the woman as Mary, contextually determined by his resetting of the account as a postlude to the Lazarus story, and the naming of the one who complains as Judas, which itself may be derived from Mark 14.10-11, which comes straight after the anointing, and links Judas with an unhealthy interest in money. But there is one element in John that appears not to be found in Mark, Mary's wiping Jesus' feet with her hair (καὶ ἐξέμαξεν ταῖς θριξὶν αὐτῆς τοὺς πόδας αὐτοῦ, John 12.3). This detail appears to come from Luke 7.38 (καὶ ταῖς θριξὶν τῆς κεφαλῆς αὐτῆς ἐξέμασσεν. . . ), where it forms part of Luke's story of the anointing, 7.36-50, his version of the Marcan // Matthean anointing. As there, it is an anointing by an anonymous woman in the house of a man called Simon, though Luke relocates it at an earlier point in the narrative, as often (cf. the Rejection at Nazareth, brought forward to Luke 4.16-30; Paul's first visit to Jerusalem, brought forward to Acts 9.25-6 and the Jerusalem Council, brought forward to Acts 15 from its "true" location in Acts 18.22), a move that necessitates some reworking of the details, especially the stress on the forthcoming death and burial. It is now a story about a "sinner", whose hair hangs down.

The anointing in each of the Synoptic accounts makes sense. In Mark and Matthew, Jesus' head is anointed with perfume. No hair is mentioned, no feet are are mentioned. In Luke, the woman wets Jesus' feet with her tears, an act of repentance, and she wipes them with her loose "sinner's" hair before she anoints them with perfume. But John's reminiscence of the Lucan detail about the wiping of Jesus' feet with her hair creates an anomaly. First, there is no reason for Mary, in John, to be wearing her hair like a "sinner", which is the point of the Lucan story. Second, because there are no tears in John, Mary's wiping of Jesus' feet with her hair means that the perfume ends up on her hair and not on Jesus. Jesus is the one who is supposed to be being anointed. This appears to be an example of John's secondary use of prior texts that has generated narrative inconcinnity and which helps us, therefore, to sketch John into a relationship of post-dating the Synoptics.

The Dating Game III: Getting the Synoptic Gospels in Order

Getting the Gospels in order is one of the fundamental issues in the dating the crucial sources in early Christianity. Before attempting to work out whether we can locate the Gospels in any particular decade, there is preliminary work to be done, to see whether we can get them into sequence in relation to one another. The issue is separable into several separate questions, all of them controversial, all of them interesting, (1) the Synoptic Problem, (2) the question of John's knowledge of the Synoptics, (3) the question of Thomas's knowledge of the Synoptics. There are still other additional questions that we could add, like the relationship of the Gospel of Peter or the Didache to the others, but to make the task manageable, at least in an introductory discussion, it is worth focusing on the texts generally regarded in the scholarship as crucial to the task at hand.

Let us begin with the Synoptic Problem. I have written a couple of books on this topic and it is pointless for me to pretend that I am beginning fresh here so let me instead summarize my conclusions and then offer a special illustration of how I think we can stack up the Synoptic Gospels in sequence.

(1) Mark is the first Gospel and it was used as primary source by both Matthew and Luke. The Priority of Mark is rightly the consensus view in Gospel scholarship. Its major contemporary competitor, the Griesbach (Two-Gospel) Hypothesis does not adequately account for much of the Synoptic data, especially the combination of Mark's alleged omissions from and additions to the combined witness of Matthew and Luke, which generate a curious profile for Mark the redactor. (See further The Case Against Q, Chapter 2 and The Synoptic Problem: A Way Through the Maze, Chapters 3-4).

(2) Luke is dependent on Matthew as well as Mark. This theory (the Farrer theory) dispenses with the need to posit a hypothetical document, Q, to explain the extensive verbatim agreement between Matthew and Luke that is not mediated by Mark. This is the thesis of my Case Against Q, summarized also for introductory students in the last chapter of The Synoptic Problem: A Way Through the Maze (see also the Case Against Q website). It is an argument against a major element in what is currently the majority view in Gospels scholarship, the Two-Source Theory, which argues that Matthew and Luke used Mark independently of one another, which necessitates Q. Arguments for Luke's independence of Matthew are unconvincing, and evidence of Luke's familiarity with Matthew needs to be taken seriously. To take just one area, it is commonly said that Luke's re-ordering of Matthew's discourses is inexplicable, but it makes good sense when one pays attention to Luke's redactional habits with respect to Mark, and his narrative habits overall.

The direction Mark > Matthew > Luke can be observed by paying attention to an important but underestimated indicator of the genealogy of documents, the phenomenon of editorial fatigue. I have argued (Fatigue in the Synoptics) that Matthew's and Luke's dependence on Mark, and Luke's dependence on Matthew, can be seen in the way in which each evangelist will, on occasion, begin by making changes to a pericope, only to lapse into the wording of the source as time goes on, creating minor contradictions. Thus we can see Matthew using Mark in the story of the death of John the Baptist (Mark 6.14-29 // Matt 14.1-12), beginning the pericope by changing Mark's Herod "the king" to his own more accurate "Herod the tetrarch", only to lapse into calling him "king", with Mark, half-way through the passage. Moreover, he adjusts the plot of the story. Where in Mark, Herodias wants John killed, Matthew has Herod himself desiring to kill John, but then Matthew retains Mark's notice that Herod grieved John's death.

Luke appears secondary to Mark in the Feeding of the Five Thousand story (Matt 14.13-21 // Mark 6.30-44 // Luke 9.10-17), which he begins by resetting it in "a city called Bethsaida", which causes an inconcinnity when he later repeats, with Mark, "we are in a desert place here" (Luke 9.12).

The same phenomenon of editorial fatigue also shows Luke to be secondary to Matthew. In the Parable of the Talents / Pounds, Luke, who loves the 10:1 ratio, begins with a major change: ten servants, not three; and with one pound each. Yet as the story progresses, Luke gets drawn back to the plot of the Matthean parable, with three servants, "the first", "the second" and "the other". The wording moves steadily closer to Matthew's as the parable progresses.

I offer these brief examples of the phenomenon of fatigue to draw attention to the possibilities for using literary criticism to theorize about the direction of dependence among related documents. What will be of interest next will be to explore the still more vexed questions of the relationships between the Synoptics, John and Thomas.

The Dating Game II: Getting Paul's letters in order

One of the easy mistakes in the dating game is to shoot too quickly for absolute dates, to look in a given document for hints that might help us to pin it to a a specific date. Some of our documents, though, are not of the nature that will allow us to pin them to a particular decade, let alone a particular year, and in such circumstances, it is important to try to get them into the right relative order, to make sure that we are stacking them up in the right order with respect to one another. Our general reluctance to do this may have something to do with our general reluctance to get our hands dirty doing serious work on the Synoptic Problem, or to do the related, equally difficult work on other big issues that make some of us recoil, Pauline chronology, John's familiarity (or not) with the Synoptics, Thomas's use (or not) of the Synoptics. But if we are to make progress on dating our crucial sources, these are the kinds of specialist areas that we need to invest in.

Let us take what is perhaps the most straightforward area first, the issue of getting Paul's letters in order. We are lucky here to have a degree of consensus on the parameters and general shape of the question. We agree, broadly, that Paul's letters were written in the 50s, with the late 40s the very earliest we can go. And we have general agreement on the basics of how to frame Paul's life. No one seriously thinks that 1 Thesslaonians is a late letter, or that Romans is an early one. If there are serious disagreements about the integrity of 2 Corinthians, and the authenticity of 2 Thessalonians, there is nevertheless broad consensus that the order of the undisputed letters goes something like this:

1 Thessalonians
Galatians
1 Corinthians
2 Corinthians
?Philippians
?Philemon
Romans

It is easy to be sure about Romans. Paul is explicit that he has preached the gospel fully in a circle from Jerusalem to Illyricum (15.19), and that he is on the way now to Jerusalem with the collection for the saints there, with a view to heading off next to Rome and then to Spain (Rom. 15.23-9). Now the collection provides us with the most helpful basic piece of sequential dating material because it is mentioned, at different states of development, in three other epistles, all of which predate Romans.

Gal. 2.10: Only they would have us remember the poor, which very thing I was eager to do.

1 Cor. 16.1-4: Now concerning the collection for the saints: you should follow the directions I gave to the churches of Galatia. 2 On the first day of every week, each of you is to put aside and save whatever extra you earn, so that collections need not be taken when I come. 3 And when I arrive, I will send any whom you approve with letters to take your gift to Jerusalem. 4 If it seems advisable that I should go also, they will accompany me.

2 Cor. 9.1-4: Now it is superfluous for me to write to you about the offering for the saints, for I know your readiness, of which I boast about you to the people of Macedonia, saying that Achaia has been ready since last year; and your zeal has stirred up most of them. But I am sending the brethren so that our boasting about you may not prove vain in this case, so that you may be ready, as I said you would be; lest if some Macedonians come with me and find that you are not ready, we be humiliated - to say nothing of you - for being so confident.

This is a fine example of the way in which sequential biography mentioned in documents can help us to date those documents. Clearly the collection is at an early point in 1 Cor. 16 -- Paul has recently instructed the Galatians about it, and he is only beginning now to talk to Achaia about it; presumably he has not yet begun to talk to Macedonia about it. By 2 Cor. 9 it has advanced much further. At least a year has passed; Paul is expecting Achaia to be ready, and Macedonia is ready too. So 1 and 2 Corinthians are placed in their expected sequence with respect to one another, but both also earlier than Romans.

There is actually one more opportunity the material here provides, but it is an invitation often and surprisingly refused. The major, marked difference between 1 Cor. 16 on the one hand and 2 Cor. 8-9 and Rom. 15 on the other is that Galatia has dropped out. Where Paul, when he was writing 1 Corinthians, had expected the Galatians to participate, they are out of the picture by the time that he was writing 2 Corinthians, something further confirmed by their absence from Romans. The crisis in Galatia, therefore, appears to have taken place between the writing of 1 and 2 Corinthians. This is when Paul lost the allegiance of the Galatians who had turned to what Paul saw as "another gospel" and getting circumcised (Galatians). The order of Paul's letters, then, goes something like this:

1 Thessalonians
1 Corinthians
Galatians
2 Corinthians
?Philippians
?Philemon
Romans

Getting the relative dating of 1 Corinthians and Galatians right illustrates the value of dating questions in the study of Christian origins. The hints provided by Paul's biography for establishing that 1 Corinthians precedes Galatians correlate with other factors of interest in the study of Paul. What is the source of his gospel? Is it through human agency (1 Corinthians 15.1-11) or directly from God (Galatians 1.6-12)? What about his use of Jesus material? Is it a coincidence that his earlier epistles, 1 Corinthians and 1 Thessalonians, are rich in Jesus material but his later epistles are not? What about Paul, the Law and justification? Is it significant that the earlier epistles, 1 Thessalonians and 1 Corinthians, are light (to put it mildly) on the forensic language while the later epistles (Galatians, Romans, Philippians 3) feature it heavily?

The Dating Game I: Preliminaries

I have been asked to present a paper at this year's SBL Annual Meeting in Boston on "Dating the Crucial Sources" for a new consultation entitled "Cross, Resurrection and Diversity in Early Christianity". Simon Gathercole is speaking in the same session; April DeConick is responding to me and Stephen Patterson is responding to Simon. John Kloppenborg will be in the chair.

It is a pleasure to be invited to give a paper, and a special pleasure to do so at one of the inaugural meetings of a new consultation. When one is invited to present, it will usually be in an area that overlaps with one's area of expertise, but which at the same time provides a fresh challenge, as here. I have written a little about dating issues but not a lot. It is something that interests me, but this is providing me with an opportunity to spend some time thinking through things in a more systematic and serious way. But where does one begin with such a broad topic? One of the pitfalls of attempting to date early Christian documents is to shoot too quickly for absolute dates, to try to pinpoint each text to a specific moment without doing the prior work on working out the relationships of documents to one another. Thus however much we might find matters like the Synoptic Problem not to our taste, it is essential to get on top of such things if we are to get some feeling for the most plausible relationship of documents to one another. It is a necessary prior step before attempting to fix documents to a specific date or range of dates.

Before that, though, several more preliminaries and important reminders:

(1) What is a document? I once wrote an article on Q called "When is a text not a text?". Although it dealt specifically with the hypothetical document Q, it got me thinking about the broader issues of what we mean when we talk about "texts" and "documents" in antiquity. Of course we all know that we do not have autographs and we know that there were no printing presses, but textual critics rightly remind the rest of us to behave like we actually know that that is the case. Too often, we lapse into treating our scholarly constructs as if they are the actual artefacts that they are only aspiring to be. At the very least, we need to keep reminding ourselves in discussions like this that we are not dealing with fixed points and known entities but with reconstructions and approximations.

(2) A Document's Evolution: there is a related issue here, that the more we become text-critically sensitive, the more we are inclined to reflect on the evolution of the documents we think we know. When we try to date Mark's Gospel, what are we dating? Something that approximates to our scholarly reconstructions of Mark 1.1-16.8 or something akin to what the vast majority of witnesses have, a Mark that goes on beyond 16.8? When we try to date John, are we imagining a version with or without the pericopae adulterae, with or without Chapter 21? When we date Thomas, are we dating textual antecedents to the Oxyrhynchus fragments, where where Coptic Thomas's Saying 77 is found with Saying 30, or constructs more akin to the Coptic, or both or neither? Even in our print culture, a document's history is often about a date range rather than a fixed point in time. When I refer to John Knox's Chapters in a Life of Paul, do I date it to its original influential edition in 1950 or the revised version of 1987, in which he reacts to his own critics of his earlier work? (And to make it still more complicated, we could insist too that even the 1950 edition featured revised versions of articles written in the 1930s). The point here is that sometimes our attempts to date documents precisely ignore what we know to be the case, that documents are not static entities even today, let alone in antiquity.

(3) Text and Tradition: There is a further related issue that often causes confusion. We sometimes speak as if a document is as early as the traditions it contains. Or, to put it in another way, we confuse tradition history with a document's dating. Thus a document first penned in the year 80CE might contain good traditions from the early 30s. One first penned in the 60s might be full of historically dubious legends. We should be careful to make sure that in attempting to date a document we are not simply dating the traditions contained in that document.

It is not my intention, though, just to talk about the difficulty of the task at hand, but rather to make sure that certain warnings are in place before embarking on the journey ahead. I want to make clear that where I do talk about dating documents, I am doing so in full knowledge that there are difficulties here, and that we speak in a shorthand that sometimes has to bypass complex issues to which we will have to return.