2010-03-05

E.P. Sanders’ Test for Authenticity of the Sayings of Jesus

Creative Commons License

This work is licensed under a Creative Commons Attribution 4.0 International License.

by Neil Godfrey

Following Professor James McGrath’s advice to pay particular attention to E. P. Sanders’ discussion of methodology (pp.3-22 in Jesus and Judaism) I am here have a look at one of the main tests for the sayings material.

Sanders does not discuss any methodology for testing authenticity of biographical events in Jesus’ life. The closest he comes to this is an a priori analysis of the plot narrative: “If Jesus did X then he probably also did or meant or tried to do Y”, or “Since B happened to Jesus then it was probably because he did something like A.”

He does discuss a methodology for the sayings material. It is the criterion of double dissimilarity.

The principle test which scholars have recently used for assessing authenticity is the test of double ‘dissimilarity’ . . .

The test is this: material which can be accounted for neither as traditional Jewish material nor as later church material can be safely attributed to Jesus. (p. 16)

Sanders is well aware of difficulties with this.

There are well-known difficulties in applying the test. We know first-century Judaism very imperfectly, and knowledge about the interests of the church between 70 and 100 CE . . . is slender indeed.

Despite these two deficiencies in our knowledge, Sanders refers to the “Let the dead bury their dead” saying as one example where this test can be successfully applied. (But even this saying has been contextualized within pre-Christian Judaism’s and the early church’s beliefs in the sharp divide between the life of the true community of faith and the spiritual death of those outside. Fletcher-Louis, JSNT 26.1 (2003).)   Nonetheless, Sanders continues:

Yet a problem remains. The test rules out too much.

It is this very “problem” of the test that is addressed by Robert Price, and as outlined earlier in 5 (more) commandments for historians:

And if the criterion of dissimilarity is valid, then we must follow unafraid wherever it leads. (Price in The Historical Jesus: Five Views, p.59)

Price argues that since every saying attributed to Jesus in the Gospels was written by “church” scribes and for church needs, it follows, by the criterion of dissimilarity, that every saying we have of Jesus is a creation for church needs.

Price appears to be directly responding to Sanders when he remarks that this criterion has been watered down by many scholars on the grounds that, applied consistently, it leaves virtually no sayings left to attribute to Jesus. But of course, we cannot justify a complaint about a method solely on the grounds that it does not yield the results we want.

Back to Sanders.

[T]he remaining material does not interpret itself or necessarily answer historical questions. It must still be placed in a meaningful context . . . .

Since historical reconstruction requires that data be fitted into a context, the establishment of a secure context, or framework of interpretation, becomes crucial. There are basically three kinds of information which provide help in this endeavour: such facts about Jesus as those outlined (. . . nos. 1-6); knowledge about the outcome of his life and teaching (cf. nos. 7 and 8); knowledge of first-century Judaism. (pp.16-17)

The numbers here refer to Sanders list of “facts about Jesus’ career and its aftermath which can be known without doubt. . . . The almost indisputable facts . . . are these” (p.11):

  1. Jesus was baptized by John the Baptist
  2. Jesus was a Galilean who preached and healed
  3. Jesus called disciples and spoke of their being twelve
  4. Jesus confined his activity to Israel
  5. Jesus engaged in a controversy about the temple
  6. Jesus was crucified outside Jerusalem by the Roman authorities
  7. After his death Jesus’ followers continued as an identifiable movement
  8. At least some Jews persecuted at least parts of the new movement (Gal. 1.13, 22; Phil. 3.6), and it appears that this persecution endured at least to a time near the end of Paul’s career (II Cor. 11:24; Gal. 5.11; 6.12; cf. Matt. 23.3; 10.17)

There is no evidence for any of the above apart from the Gospel narratives themselves. Some are disputed by some scholars (e.g. points 2, 3 and 5). And if such “facts” can only qualify as being “almost indisputable”, that hardly leaves them in the same class of what normally is taken as historical fact, such as Caesar crossing the Rubicon or the existence of itinerant sophists teaching around Athens from the fifth century b.c.e.

This absence of any way to determine the historical origin of the many of the Bible’s narratives has not gone unnoticed among some scholars:

[A]ll the reports about [Jesus] go back to the one source of tradition, early Christianity itself, and there are no data available in Jewish or Gentile secular history which could be used as controls. Thus the degree of certainty cannot even be raised so high as positive probability. (Schweitzer, Quest, p.402)

Twentieth-century scholarship, with its faith in history, assumed a historical Jesus as its starting point. It shared Schweitzer’s personal dilemma: a choice between a Jesus who fits modern visions of Christianity and Mark’s failed prophet. But they always assumed there was a historical Jesus to describe. (p. 7, The Messiah Myth (2005) by Thomas L. Thompson)

So far, historical research by biblical scholars has taken a … circular route …. The assumption that the literary construct is an historical one is made to confirm itself. Historical criticism (so-called) of the inferred sources and traditions seeks to locate these in that literary-cum-historical construct. (Philip R. Davies, In Search of ‘Ancient Israel’, pp.35-37 — in other words, scholars have just assumed that the narrative originated in historical events)

Laziness is common among historians. When they find a continuous account of events for a certain period in an ‘ancient’ source, one that is not necessarily contemporaneous with the events , they readily adopt it. They limit their work to paraphrasing the source, or, if needed, to rationalisation.Liverani, Myth and politics in ancient Near Eastern historiography, p.28.

So what is left of he principle test for establishing the historicity of any of the sayings of Jesus?

If applied consistently, we must exclude all sayings that

  1. could have derived from traditional Jewish material and later assigned to Jesus
  2. could have been created to meet church needs, and assigned to Jesus to add to their authority

That doesn’t leave much, if anything, left over.

And that little that might be salvaged needs to be found in a secure historical context. Yet neither Sanders nor McGrath can offer any methodology by which to determine whether or not the narrative has a genuine historical origin. They can only assume this. And once assumed, they use a priori arguments about the plot of that narrative to “establish” which story details are more or less likely to belong to that (presumed) history.

It is relatively easy, on the other hand, to demonstrate that the gospel narratives have borrowed heavily from other narratives, in particular those of the Jewish scriptures.

In the former case (the presumption of historicity) we must assume a source; in the latter (adaptations from other narratives) we can see the sources.

The following two tabs change content below.

Neil Godfrey

Neil is the author of this post. To read more about Neil, see our About page.

Latest posts by Neil Godfrey (see all)



If you enjoyed this post, please consider donating to Vridar. Thanks!


21 thoughts on “E.P. Sanders’ Test for Authenticity of the Sayings of Jesus”

  1. ‘The test is this: material which can be accounted for neither as traditional Jewish material nor as later church material can be safely attributed to Jesus.’

    How can this test be evaluated to see if it actually works?

    1. I don’t think this is a testable claim, barring new evidence. If a cache of 1st century Christian documents was suddenly discovered, you might be able to test this claim against the new documents. Depending on what’s in the documents.

      And I don’t think that Sanders is even stating the criterion of dissimilarity correctly here. I may be quite wrong about this, but isn’t the actual criterion more along the lines of “material which can be accounted for neither as traditional Jewish material nor as later church material has a higher probability of being attributable to Jesus than material that can be accounted as traditional Jewish material or as later church material.”? You can’t really “safely” do anything here without some kind of external corroboration, can you? You can only suggest that one of these things is more probable than the other.

      And the giant, massive elephant in the room is this:

      There are well-known difficulties in applying the test. We know first-century Judaism very imperfectly, and knowledge about the interests of the church between 70 and 100 CE . . . is slender indeed.

      How can you possibly “safely” categorize something as “neither traditional Jewish material nor as later church material” if you’re going to admit that we have poor knowledge of what Judaism was like at the time AND we have poor knowledge of what the church needs of early Christianity were like in the time frame the sayings were put to paper? You can’t possibly know with that level of certainty what bin the sayings belong to if you don’t even know what shape the bins are. How can you speak in absolutes instead of talking about probabilities in a situation like this?

      1. Historians can only work with the evidence we have. Obviously new discoveries could radically change our understanding – as has happened plenty of times in the past. But we can either make our best judgment based on available evidence, or simply not say anything except in the case of those figures and events that we are fortunate to have extensive evidence about.

      2. Exegesis of a narrative cannot magically conjure up evidence for the historical reality of the narrative.

        [A]ll the reports about [Jesus] go back to the one source of tradition, early Christianity itself, and there are no data available in Jewish or Gentile secular history which could be used as controls. Thus the degree of certainty cannot even be raised so high as positive probability. (Schweitzer, Quest, p.402)

        Sanders is merely attempting to calculate what narrative details are more or less likely to make sense of the larger plot, given the assumption that the narrative originated largely from “traditions” going back to an historical Jesus. Historicity itself is an assumption.

        I do not need to make this assumption when reading any other ancient historians we normally take as valid historians. External controls give me some measure of confidence with them.

      3. But we can either make our best judgment based on available evidence, or simply not say anything except in the case of those figures and events that we are fortunate to have extensive evidence about.

        Granted. But when a scholar is using his “best judgment” when making an assessment based on the available evidence saying that something is “probable” or “more probable than these other alternatives” is very different from saying that something is “certain” or “safe to assume”. Especially when a scholar is then going to take those things that he’s labeled as “certain” and build hypotheses on top of them.

        It just really seems that more certainty is asserted than is actually warranted from the evidence. It’s not just that new evidence could show up to destroy the certainty, it’s that an assumption seems to be being made that “if we don’t have evidence for something being church material or early Judaism material then it MUST BE SAFE TO ASSUME it to be something that Jesus actually said (until new evidence shows up to prove otherwise)”. That’s actually a hard core assumption to make given how little actually seems to be known about the early Church and 1st century Judaism.

      4. JER
        If a cache of 1st century Christian documents was suddenly discovered, you might be able to test this claim against the new documents. Depending on what’s in the documents.

        CARR

        Just imagine if we could find such a thing.

        At present, for example, the baptism of Jesus by John the Baptist is speculative.

        Indulge me for one second.

        Suppose we could find a new Gospel, where an anonymous person had written that this was the work of a disciples.

        Suppose this Gospel had a different list of the Twelve disciples

        Far fetched, I grant, but just imagine.

        Just think how this would confirm the historicity of the Twelve.

        Just suppose this newly discovered Gospel had no baptism of Jesus by John the Baptist.

        At last the smoking gun would have been found! Final proof that the baptism happened and early Christians were embarrased by it.

        At present, all our Gospels claim Jesus was baptised by John the Baptist.

        But this can not be confirmed to the satisfaction of historians unless we found a Gospel where no such event happened.

        Of course, this is just wishful thinking, Without such a find, we have to work with the evidence we have, so the baptism remains speculative until and unless we find a Gospel where there is no mention of it.

  2. How is it an assumption to apply the criterion of double-dissimilarity? It “rules out too much” in the sense that it inevitably excludes most authentic material, since no figure in history is totally discontinuous from what went before and what came after. But that which is discontinuous is more likely to be authentic than inauthentic, since there is no obvious or known reason why it would have been invented, nor is it likely to have simply been taken over from the broader context.

    But that was clear from Sanders, and from everyone else who has discussed historical methodology. Perhaps you’d care to cite some scholar of historiography not connected with Biblical studies who challenges this way of sifting through the evidence? Or perhaps you’d like to propose better criteria yourself?

    1. Double-dissimilarity (dd) can isolate certain texts, but it cannot of itself tell us about the origin of any text. (Besides, dd cannot even isolate texts if used consistently, following Price’s argument. And the one example cited by Sanders — ‘let the dead bury their dead’ — is not securely isolated according to later studies anyway.)

      Merely isolating a text from two contexts is not establishing its origin in a particular person whose words were transmitted through decades before being recorded. It merely prompts us to ask where else the text may have originated. It may even lead us to ask if we know enough about either of the two contexts we have prised it from. Why not ask if we can learn more about the needs or interests of the community that preserved them, rather than assume that they had no particular relevance for the preserving community?

      If we can find other candidate sources for that “isolated” text (as we apparently can with the one example Sanders cites, “let the dead bury their dead”) then why not tentatively accept them?

      The various criteria of authenticity used are designed to work with the assumption of basic historicity of the contents of the narrative. To claim that the criteria establish historicity is circular reasoning.

      The criteria have their place, but it’s not to establish the historicity of whole persons and events in the absence of external controls. See my reply to Smijer below for the sort of evidence that does establish historicity.

  3. “there is no obvious or known reason why it would have been invented”

    But, per Sanders himself, we have “We know first-century Judaism very imperfectly.” So, no, of course the reason isn’t known, or, well, it would be known, and you, and most other historians, would probably be mythicists. And it’s not “obvious” in the same way that a lot of the twists and turns of history (and here, we’re almost dealing with meta-history, the way ancients contextualized and developed their own historiographical or pseudo-historiographical narratives) are rarely if ever “obvious,” save with the advantage of hindsight. That’s why we’ve been trusting a professional guild to give us these insights for all these years.

    And to approach from another angle, neither is it known or obvious why the crucifixion of a Galilean peasant rabble rouser should give rise to what invention that did take place or why it took the form that it did. I just don’t see how you think this is such a telling point against the possibility of invention. It strikes me as a bare argument from incredulity, and one, moreover, that could be equally well leveled at much of the analysis of mythmaking and invention in the works of mainstream NT scholars.

    1. If you have a better approach than allowing those with appropriate training to make the best judgment they can based on the available evidence, I’d love to hear it. No one is suggesting that this is a perfect system, and historians would be the first to emphasize that it doesn’t provide anything like certainty. But that doesn’t mean that all views fit the available evidence equally well, and if we stop trying for the best fit to the evidence, then pretty much anything goes, since few scenarios are strictly speaking impossible, provided one is willing to assume that evidence that is no longer extant would have provided a very different impression than the evidence that has survived.

      We make do with what we have in a court of law, and in history. We sometimes get it wrong, since the evidence we have isn’t often complete and unambiguous. I don’t think that’s an argument against either the criminal justice system or historical inquiry, both of which are attempts to deal with precisely those less-than-ideal scenarios that are typical.

      1. No, don’t misunderstand me. I’m just pointing out that to say “it’s not obvious why” this or that only highlights the need that we all agree exists for critical historical reconstructions by those scholars with the time, the resources and the temperament to do the work. If any of this (including the mainstream analysis) were just obvious, “those with the appropriate training” would be superfluous.

        What I am suggesting is that 1.) an interested and reasonably well-educated amateur, while unlikely to overturn an established paradigm single-handedly, may still bring questions and perspectives to the debate that have not been adequately treated in the literature, and related, 2.) that guilds such as “those with the appropriate training” to do biblical studies sometimes become hidebound and overly committed to the assumptions that guided past members of the guild. Hector Avalos has a lot to say in this regard.

        When you stand on the shoulders of giants, sometimes it’s advisable to take a glance down and see if the giant is still paying attention.

  4. At least some Jews persecuted at least parts of the new movement (Gal. 1.13, 22; Phil. 3.6), and it appears that this persecution endured at least to a time near the end of Paul’s career (II Cor. 11:24; Gal. 5.11; 6.12; cf. Matt. 23.3; 10.17)

    There is no evidence for any of the above apart from the Gospel narratives themselves. Some are disputed by some scholars

    Item 8 there… well, evidence for that comes from the epistles of Paul much more than the Gospels and Acts. And, we have it from the pen of a participant in the action being witnessed to. This confession doesn’t appear to have been coerced. Not all confessions, even those voluntarily given, are true. But you *do* have to give some evidential weight to first-person accounts of embarrassing deeds, don’t you?

    Item 6 does not rely entirely on the gospels and epistles, either. It is testified to by Tacitus, and later by Lucian.

    I’m suspicious myself of historical claims made in theological material, but I wonder if mythicists don’t sometimes get a little too dismissive of the New Testament as a source of historical evidence. It makes sense to view NT material critically. It makes less sense to dismiss NT evidence out of hand.

    1. NT evidence is not dismissed. It is evaluated alongside other comparable literature. What is dismissed is the assumption that its self-testimony is indeed a historical reality. We can’t begin by assuming this. We need to have a reason for accepting this. That’s not being hyper sceptical. It’s how we approach any documentary evidence in history (except, it seems sometimes, in biblical history). Ancient documents in particular are renowned for their novellas, their false use of authorial names and settings for any number of reasons, their interpolations and forgeries. Those that we know we can basically trust as genuine attempts at history (not that we need to believe every detail they contain) give us confidence because we can see in them evidence for their credibility. Their authors identify themselves, and they inform readers very often of their sources, and how they decided between competing claims, etc. Not always, but very often. (Sometimes in the case of Jewish writing we have the voice of anonymous authority associated with sacred works. — a neat device for introducing something new in a culture that valued the old, as discussed by Bernard Levinson in his work on Deuteronomy.)

      But more importantly than any of that is that we have external controls that inform us that such histories are indeed about real people and events. As Scwheitzer himself admitted, we have nothing comparable to verify the Gospel narratives.

      If historians discover diaries claiming to be by Hitler, or a telegram claiming to be from the Kaiser of Germany, they don’t just assume their authenticity. They need to analyze the evidence through a range of external controls — facts known independently of the documents and by which they can assess certain claims of the document.

      I can read Josephus with some measure of confidence that I am reading a view of more or less real historical happenings. This is because we have non-Jewish/partisan accounts of the events he writes about. We also have primary evidence of some of the events. But we don’t trust Josephus on everything. Like other ancient histories such as Herodotus and Israel’s Primary History (Gen to 2 Kgs), his history also begins with mythical tales before merging into something he thinks more real. But we have good defensible reasons for reading much of his account of the War as “history”.

      There are no external controls by which we can make similar assessments of the narrative in the Gospels. That does not mean we dismiss the Gospels. It means we study the gospels in the light of the evidence we do have, and see where that leads.

      Ditto for the epistles. Rosenmeyer (linked in my archive of book reviews/comments) demonstrates the popularity of epistolary fictions of the time. But I am not saying that this proves Paul’s letters are fiction. It is something that needs to be considered, but it cannot be assumed. More significant in the study of Paul’s letters is reading them in the light of the ideas we know were available to him (not through Gospel assumptions). And that’s where we see Paul says a lot that is too often silenced by Gospel historicist presuppositions.

      1. NT evidence is not dismissed. It is evaluated alongside other comparable literature.

        Good then. I just hadn’t seen it. It seemed to me that the biggest refutation of the 8 points was that they were evidenced mainly in the NT. It looked to me like the only argument you presented was “only evidenced in the [NT]”, when that wasn’t even true for point 6, and wasn’t persuasive for point 8. In fairness, I haven’t read Sanders to see how he supports the 8 points, if at all. And I am suspicious of certainty claims for point 3 myself.

        It is unfortunate that there is so little documentary evidence outside the NT for which to try to study the historical Jesus. I don’t get involved much with the Myth debate because I don’t know much about it. If it claims that there is insufficient evidence to hold fair certainty that Jesus even existed, then I guess I have to side with the consensus. If it claims that there is insufficient evidence to know much about what Jesus was like apart from having been executed by Pilate, then I’m more sympathetic. But having read and reasoned somewhat on the subject, I have a hard time doubting some of the basic questions about Jesus – such as his Jewishness, his “involvement in a temple controversy”, his preaching of a “Kingdom of God” of some sort…. It is extremely hard for me to explain the existence of the various NT & nancanonical (Christian / Jesus community) documents to myself absent a Jesus that meets a few of those requirements.

  5. ‘I don’t think this is a testable claim, barring new evidence. If a cache of 1st century Christian documents was suddenly discovered, you might be able to test this claim against the new documents.’

    This is an admission that there can never be any non-Christian mentions of Jesus from the first century.

    The whole ‘double dissimilarity’ test is an admission that there is no record of Jesus outside the church, and not even a possibility of external controls on the life of Jesus as depicted in the Novels.

    1. Not really. The “testability” claim I was suggesting was that if a new cache of 1st century Christian documents showed up and indicated that some/many/all of the sayings that had been attributed to Jesus by the method above were actually covering the needs of the sect that the new documents belonged to, it would mean that the methodology used was a poor one for sussing out sayings that could be “safely” attributed to Jesus. (And that’s not really “testability” so much as it is “falsifiability” I realize today).

      I would not want to eliminate the possibility that a cache of non-Christian documents or other archaeological evidence might someday show up that mention Jesus. There are non-Christian sources that if they existed could go far towards establishing the Gospel narrative as “based on a True Story” – records from Pilate’s involvement in the crucifixion for example. It’s improbable that documents like that would have survived if they ever existed in the first place, but if they were to turn up and prove authentic that would be an example of documentation from outside of the church that could provide an external check on the narrative.

  6. Notice that mainstream Biblical scholars like James Dunn regard this double dissimilarity test as obvious rubbish, when they cite the following passage as proof that Jesus existed.

    Romans 15:3 ‘For even Christ did not please himself but, as it is written: “The insults of those who insult you have fallen on me.” For everything that was written in the past was written to teach us, so that through endurance and the encouragement of the Scriptures we might have hope.’

    Or perhaps people who refute Jesus-myths are simply clueless about what their own methodology is supposed to be, as Dunn blatantly junks this ‘criterion of double dissimilarity’ when he is given the task of refuting mythicism.

    If there really were criteria that mainstream Biblical scholars used, then Dunn would never have cited that passage as proof that Jesus existed, as it obviously trashes Sanders and McGrath’s claims that mainstream Biblical scholars use carefully worked out criteria that they agree on.

    The plain truth is that there is *no* methodology for finding a historical Jesus, despite McGrath’s claims that he is just doing what all other historians do.

Leave a Reply to James F. McGrath Cancel reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Discover more from Vridar

Subscribe now to keep reading and get access to the full archive.

Continue reading