Relationship between Thomas and the Synoptics

In the comments Mike K says:

I take it you view Thomas as largely independent of the Synoptics so that there similarities and differences may be explained as a result of different processes of oral transmission from the original eyewitnesses? I was just reading Andrew Gregory’s “The Reception of Luke and Acts in the Period before Ireneaus” and he is very cautious about accepting any literary dependence as demonstrable unless it meets Koester’s criteria of whether the redaction is present, but he too finds evidence of Lukan redaction in the Greek fragments of Thomas (so difficult to blame on later scribal harmonization in the Coptic version). So I was wondering if you think it may be possible that whoever put together this compilation of sayings in the 2nd century was familiar with the Synoptics in some way, yet perhaps many of the sayings still reached the author independently from oral tradition?

This is too big to answer in the comments, so I am bringing it up to a post of its own.

I don’t know that I want to be that definite, but that’s certainly the way I am leaning, yes. As a person of faith, I believe that Jesus really existed, that there were many eyewitnesses to the various parts of his ministry and that they shared their stories of their encounters with Jesus with friends, family, colleagues. I would also like to believe that the early Christians were, in general, people of good will and integrity who told others what they genuinely believed to be true and accurate accounts of what they saw, heard and experienced, rather than deliberately reshaping material to convince others to their way of thinking. A lot of the redaction criticism theory sounds too cold and calculating to me: author X took author Y’s version and edited it so that it fitted in with his theology sounds like a very deliberate thing to me.

I think that the level of verbatim correspondence between some parts of the Synoptics are long enough for us to be pretty sure either that one version was copied from the other or that there was some common source with which both authors were familiar. If what scholars beginning with Gerhardsson suggest is correct, it may well be that this source was oral material learned verbatim, rather than necessarily a written copy but I suspect that alterations were made because the version that was available did not line up with the author’s memory of the event (or the account that s/he had heard from an(other) eyewitness), rather than something more deliberate. How we understand an event andwhat we remember of an event are strongly shaped by what we believe about the world, so our memories do tend to line up with our beliefs and therefore theologies but mostly this shaping is unconscious. McIver summarises the evidence here quite well (see my review of his Memory, Jesus and the Synoptic Gospels) although I don’t agree with the conclusions he draws about what it means for reliability.

Another possible explanation, however, of how there are different versions of what appears to be the same story in circulation is that Jesus actually told basically the same story several times with slightly different details because he was telling them to a different audience in different settings. In other words, it is possible that the reason that Matthew, Mark and Luke have parallel stories in different settings is not that they wrote their gospels to achieve particular theological purposes that worked best if they put them in different orders, combinations and settings and but because Jesus actually used them in different orders, combinations and settings and the ones that fitted best with a particular eyewitness’s theological emphasis were the ones that were remembered by that person. Quite a few of the changes that we see that are named as ‘redaction’ could equally be the result of people retelling a story in their own words and it doesn’t matter whether the story is one they have read somewhere or one they have heard somewhere. My article  “How Accurate Are Eyewitnesses? Bauckham and the Eyewitnesses in the Light of Psychological Research.” Journal of Biblical Literature 129, no. 1 (2010): 177-97 summarises the literature on both eyewitness testimony and human memory.

So, in short, yes, I think what you suggest is possible, Mike, but because we are working with Greek text of the Synoptics and largely Coptic text of Thomas and some of the Greek text we do have puts the sayings in a different order to the Coptic text, I don’t think we canbe at all certain about which of the various theories is correct. I think that on the basis of the evidence we have, it is perfectly possible that Thomas was written in total isolation from the Synoptics, on the basis of teaching that he learned verbatim from Jesus. That he used the sayings he used because they were the ones that he remembered best over the years because they were the ones that helped him to make sense of his world and his life.  I think it is exceedingly unlikely that the author picked up texts of the Synoptics, selected his/her favourite bits, changed them so that they would produce the spin he wanted and made up a whole pile of other stuff so that he could convince a group of gullible people of the veracity of his/her particular theological system (whatever that is). What we need in order to make a more definitive statement about which of the various theories of composition is most likely is an early copy of a full Greek text. Even then, however, if you accept that Jesus tried to get his disciples to learn his sayings verbatim, you don’t need to have a theory of textual dependence even with really significant verbatim sections.

Gathercole on the composition of Thomas (6)

I have finally finished the book and propose to deal with the remaining 8 chapters in one post. The other option would be to look at each chapter in detail and that would take too long.

Chapter 6  is entitled Thomas and the Synoptics: A Method for Assessing Influence and proposes 6 stages in a method for assessing the influence of the Synoptics on Thomas. Gathercole outlines it as follows:

  1. Influence from the Synotpics on Thomas will be evident where Thomas reproduces redactional material.
  2. Where there is influence, taking the direction of that influence to be Synoptics –> Thomas (rather than Thomas –> Synoptics) can be justified on various grounds.
  3. The influence from the Synoptics can only reliably be seen in Thomas’s reception of Matthean and Lukan redaction of Mark.
  4. The sample of Thomas sayings to be analysed is thus restricted to places where there are parallels with Mark and at least one of the other Synoptics.
  5. Various options are discussed for how influence might take place, including combinations of oral and literary factors.
  6. Finally, the quesiton is raised of when in Thomas‘s compositional and transmission history any influence of the Synoptics might have been exerted (p 145).

On the basis of point 4, Gathercole only looks at 20 sayings – those that have a parallel in both Mark and at least one of Matthew or Luke.

Having amplified the six points, he then applies the method in chapters 7 & 8, Matthew in the Gospel of Thomas and Luke and the Gospel of Thomas. In doing this, he picks up Elaine Pagels’ suggestion that GTh 13.3 (where Jesus asks the disciples what they think he is and Matthew answers that he is a wise philosopher) is a reference to the content of Matthew’s gospel, although he rejects the notion that Peter’s comment about Jesus being like a righteous angel does not need to be a reference to either Mark or Gospel of Peter, because Peter is an unsurprising interlocuter, whereas Matthew is not. The notion that GTh 13.3 is a reference to Matthew’s gospel appears to be a key piece of evidence because he mentions it quite frequently. On the basis of his methodology, 11 of the 20 sayings are examples of Thomas‘s reception of Matthean or Lukan redaction of Mark.

Section II ends with chapter 9, The Composition of the Gospel of Thomas, in which Gathercole surveys the previous chapters and concludes that:

attempts to exclude the influence of the Synoptics from the Gospel of Thomas are unsuccessful. There is in Thomas what one might term ‘significant’ influence indentifiable from Matthew and Luke. The influence is significant not because the redactional elements … which appear in Thomas are remarkably extensive in any particular places, but rather because these redactional traces appear in eleven out of twenty sayings in which they might be identified ( p 223).

He notes that it is not possible to know by what method this influence was exerted – whether it was oral, literary or “secondary orality”, but it is clear from the foregoing and succeeding parts of the book that his definite preference is that it is literary influence.

I find his arguments significantly less convincing than he does, especially for literary influence. He continually uses words such as ‘significant’ and ‘striking’ for extremely short strings of correspondence, similar wording that is not the same and the occasional shared use of an unusual word. Research on human memory has demonstrated that even members of highly writing-dependent cultures can reproduce strings of 15 words or more verbatim from memory and people living in oral and verbomotor cultures can do significantly better than this, so strings of 6-7 words, even with one of 13, are not convincing evidence of literary influence. While it is, of course, possible that the author of Thomas was familiar with written versions of both Matthew and Luke but chose to change the wording to suit his own ends, the evidence provided is not enough to rule out: oral transmission; a common source, either oral or written; or transmission of traditions that sprang from two different eyewitnesses to Jesus’ teaching. Gathercole dismisses the first two of these and does not address the third. Indeed, it appears that he has done very little reading in the area of oral transmission and none at all in human memory and eyewitness testimony. The latter two are, at least in my opinion, very important given that we hold that the gospels are, in general, eyewitness accounts of the life and teaching of Jesus which were transmitted orally in Christian communities for several decades before they were recorded in written form.

Part III looks at Thomas and other early Christian literature. Chapter 10 examines two passages in Romans and one in 1 Corinthians and concludes that Paul influences Thomas.  Chapter 11 looks at the phrase ‘the world is not worthy’ in the Epistle to the Hebrews and GTh 56, 80 and 111 and concludes that Thomas was influenced by the Epistle to the Hebrews. It seems to me that the argument he provides and dismisses for the expression simply being a pre-existing multilingual Jewish expression is stronger. Chapter 12 is entitled A note on the “two wys” tradition and GTh 25 and in it he concludes that Thomas should be included in the group of texts influenced by a hypothetical re-existing “two ways” source.

In summary, Gathercole finds that Thomas was originally written in Greek, is dependent on both Luke and Matthew as well as some of the Pauline corpus and the Epistle to the Hebrews (and probably other sources as well). Gathercole makes it clear that he is not arguing that the author of Thomas sat down with the texts of the Synoptics, Paul and Hebrews in front of him and copied and pasted as he saw fit, but it is also clear that he prefers the notion that there is literary rather than oral dependence of Thomas on the Synoptics. What this actually means, though, is rather unclear. The book would have benefitted from a careful definition of dependence (as would most works addressing this issue). Sometimes he suggests that this might be as vague as once having heard them read and remembering them, at other times, the reader gets the impression that he thinks that the author may have read them. In the latter chapters, he tends to use ‘influence’ rather than dependence, and this seems to be a better way of talking about what he is claiming. I think that  he succeeds in making it obvious that we simply do not have sufficient evidence to be able to be dogmatic about any theory of  composition of Thomas and that theories about source, dating, original language etc are all intertwined, but I do not find his arguments for his position nearly as compelling as he does.

Thomas as script consultant? (a temporary diversion from Gathercole)

What would a stageplay of the Gospel of Matthew look like if they used the author of the Gospel of Thomas as script consultant? An awful lot like Godspell, I suspect. We watched it yesterday because it was Easter Day, so the local TV stations were running some vaguely Christian programming. As I watched, I kept thinking two things:

  • If I didn’t already know about the life and death of Jesus, this would make very little sense
  • This reminds me very much of the Gospel of Thomas

Like Thomas, Godspell provides the audience with very little context. A man appears. He blows a shofar and people who are unhappy with their lives follow him. He baptises Jesus (dressed as a sad clown) in the local fountain and they go off to a secluded place where Jesus says things, fairly much a propos of nothing. Strange and probably fairly incomprehensible things if you don’t have the back story. We have no idea of the timeframe in which the teaching happens, no context from within which to understand any of the sayings, no miracles, no narrative at all. All very Thomasine.

I could make reasonable sense of it all because I knew the stories behind the sayings, but otherwise it would have been quite mystifying, I think. This reminded me of Thomas. Like the people in the English-speaking world in the second half of the twentieth century (Godspell’s implied audience), Thomas’ implied readers knew the backstory and therefore had no need of it to make sense of the sayings and anyone who didn’t needed to seek until they found it – not the author’s problem, really.  I don’t think this stripped down characteristic helps with the dating of Thomas, despite what a number of scholars want to suggest.

Of course, Godspell then moves out of  ‘sayings collection’ mode so that Jesus can be betrayed – in Godspell the movie, by the John the Baptist character – and die and be carried away, but no resurrection here, either. Jesus is simply presented as a teacher – as someone who could be described as the righteous messenger and wise philospher that Peter and Matthew proclaim him to be in GTh 13.  The screenplay is much closer to the gospel account than is Jesus Christ Superstar (no tapdancing Herod singing “prove to me that you’re divine – change my water into wine”, no disciples singing “then when we retire, we can write the gospels and they’ll still talk about us when we’ve died”), but it still moves the temptations to the night of Jesus’ betrayal and death and has a number of the people around Jesus saying words that the gospels put on Jesus’ lips. It also adds some material that’s only found in Luke. At the time when Superstar and Godspell were box office stage plays, the churches preferred Godspell because it was closer to the Bible, but they condemned both as not presenting the Gospel fully or accurately. Again, a lot like Thomas and the church fathers, really. And maybe, like the writers of Superstar and Godspell, the author of Thomas wasn’t trying to provide a full, accurate account of Jesus.

Gathercole on the composition of Thomas (5)

Chapter 5 is entitled Responses to Arguments for Independence and contains a discussion of the weaknesses that Gathercole perceives in the arguments for Thomas being independent of the Synoptics. I find a number of the things he says in it puzzling or surprising and often wish that he had provided examples to back up his statements.

In the introduction, Gathercole notes that while the previous four chapters can stand alone, their findings also have three significant implications for how we understand the relationship between Thomas and the Synoptics. The first two are straightforward and simply summaries of what has gone before:

  • a putative early Aramaic Thomas would make a relationship between it and the Synoptics unlikely, but a Greek original makes a relationship between the four more likely.
  • if it were possible to demonstrate that Thomas and the Synoptics contained divergent translations from Aramaic originals, independence would be more likely, but the conclusions drawn in chapters 2 and 3  suggest that the case is not particularly compelling.

The third surprised me.

  • He says that the discussion in chapter 4 of the similarities bettween the Greek and Coptic texts “showed that the content of Thomas was reasonably stable across the century or two separating the Greek fragments and the Coptic version. The implications of this point for our discussion are traced further in ensuing chapters (p 129).” Perhaps I have not read chapter 4 thoroughly enough, but I have been unable to find any statement about the stability of the content until this point in the book. It is therefore not clear to me exactly how he feels that he has demonstrated this.

Gathercole also indicates that:

these chapters will not argue for anything like total dependence upon the Synoptics, as if all the author or editor of Thomas knew was Mark, Matthew and Luke and nothing else (I am not aware of any scholar who has argued for that position.) Clearly Thomas is – on any reckoning – at least partially independent of the Synoptics, as it is virtually incredible that the editor of Thomas invented all the material not paralleled with the Synoptics. The presence both of non-Synoptic but Synoptic-like material in Thomas and of other quite different sayings clearly points towards partial independence, but neither of these can be regarded as indicating the independence of what is paralleled in the Synoptics (pp 129-130). [underlining added]

I am not sure what it is that Gathercole is trying to convey in the underlined sentence.  Clearly, those sayings in Thomas which do not appear in any of the Synoptics cannot possibly be held to be dependent on them, but the source of any non-parallel material is irrelevant.  FWIW, if you hold that the extra material is not authentic Jesus tradition, then the most likely explanation is that the editor and/or his/her community did, in fact, invent it. If you hold that it has the possibility of being authentic Jesus tradition then it must ultimately stem back to some alternative eyewitness source – and John 21:25 certainly suggests that there is plenty of authentic Jesus material that has not appeared in the canonical gospels.

Having said this, Gathercole then identifies four main areas of argument in favour of Thomas being independent and addresses each in turn. I am using his own headings:

  1. Do the differences in order imply the independence of Thomas and the Synoptics? He makes four ponts. He first raises the question of why Thomas would break up and re-order Matthew, which he says is essentially a non-problem, arising from the highly scribal mentality of the early Thomas scholarship. “When this scribal mentality is abandoned, however, the objection ceases to have any force (p 131).” With this comment, he moves on to the next issue, but the situation is by no means as clear cut as he suggests. While it is true that in cultures where material is transmitted orally rather than in written form, the preservation of text in exact verbatim form is not as high a priority as it is in a scribal culture, one of the features of oral transmission is grouping items ways that make them easy to memory. Deliberately dragging a nice, neat, easy to remember section such as Matthew 13 apart is counter-intuitive, unless you wish to accept an explanation such as Perrin’s Syriac catchwords as providing a new way of remembering such a large body of material. His second counter is to quote Tuckett: ‘someone somewhere must have changed or created either the synoptic order or GTh’s order to produce the other (probably with a number of stages in between’ (“The Gospel of Thomas: Evidence for Jesus?”, NTT 52 (1998) 23-24). This is, of course, only true if one is dependent on the other. Given the lack of verbatim agreement and, in some cases, significant differences even in the gist of the parallels, another possibility is that the two come from different eyewitness accounts of different events in Jesus’ teaching where he used slightly different versions of stories in somewhat different order. His third is to remind us that Wilson identified several cases where adjacent sayings in Thomas ‘are also juxtaposed in the Synoptics” (p 131) and asks if this is purely accidentally. It almost certainly isn’t, but if what we have is accounts from different eyewitnesses recounting the same events, it would be expected that there would be overlap as well as difference. Fourthly he draws attention to the difference in genre and suggests that in a collection of sayings, one might expect the order not to be as important as it is in something that is clearly narrative. While this is true, it goes against his earlier argument that Thomas and the Synoptics are of the same genre. Most of these points would benefit from the provision of an example to illustrate how he reaches his conclusions, rather than just a bare statement of what he holds to be fact but which appears to me to be open to question.
  2. Do form-critical factors suggest the priority of Thomas‘s versions? This section looks at the various form-critical ‘rules’ used to indicate that one piece of text is older/earlier than another and demonstrates  that some contradict one another and they are by no means as watertight as their supporters suggest. While I agree with what he says in most of this section, his cases would be stronger with the use of examples. As it stands, much of his argument consists simply of quoting the opinions of other scholars and, as usual, he has ignored the evidence from oral transmission and human memory, which this time would support his case.
  3. If Thomas is dependent on the Synoptics, why is there no extensive verbatim correspondence? Here, we are promised a discussion of individual sayings in later chapters, shown again the longest example of correspondence between Greek texts and told that by adopting the correspondence between the Synoptics as the ‘norm’ scholars who favour the independence argument are placing the burden of proof artificially high. Again, the psychological literature on eyewitness testimony and human memory would suggest that this is not the case.
  4. Does the absence or insignificance of Thomas’s appropriation of redactional feature in the Synoptics show that there is no literary relationship? Gathercole obviously argues that it doesn’t, because he has made it very clear that he is convinced that there is a literary relationship between them. He indicates that he will deal with these issues more fully in the next three chapters but in the meantimes makes a number of remarks. I find the points he makes less than obvious without concrete examples. At this point, all I feel that he does successfully is to indicate that there is sufficient evidence of potential redaction so that we cannot rule out the possibility of literary relationship. This hardly needs stating, given the debate in scholarly circles over the past 50 or so years.

In view of my comments above and in previous posts, I am afraid I cannot agree with the first sentence of the conclusion to this chapter “In sum, there is not really a single argument for the thoroughgoing independence of Thomas which has any force (p 143).” He promises that in the following chapters he will demonstrate “that there is actually good reason to suspect the opposite, namely that Matthew and Luke do influence Thomas (p 144).” I hope that this involves more than simple citations of the opinions of others.

To this point, the book has very much the feel of a text for which the author has been given a strict word/page limit which only allows a sketch of the material being presented. It is a relatively slim volume, but this is clearly the result of the nice paper on which I commented in the first post. There are 270 pages of text, plus introduction and indices, and I notice that many of the thicker paperbacks on my shelves have fewer pages than this.

Gathercole on the composition of Thomas (4)

Chapter 4 is entitled “Positive evidence for a Greek-language origin” and in it, Gathercole addresses six areas:

  1. The material evidence of the manuscripts: Here, Gathercole says that we have no manuscript evidence of a Semitic version of Thomas but there are three fragments of  Greek copies. Although he notes that an argument from silence needs to be viewed with caution, he suggests that the material evidence is sufficient to suggest that a Greek composition should be the default position.
  2. Level of correspondence between items of Greek vocabulary in Greek and Coptic Thomas: Gathercole looks at those sections of Coptic Thomas where there is also an existing Greek version in the P. Oxy fragments and  lists 27 Greek loanwords in the Coptic text. In only three cases is there a different word used in the extant Greek text and two of these are particles, which (as was noted previously) are least predictably rendered in other Greek-to-Coptic translations.  He ends this section with “This is a fairly remarkable statistic, making a Greek Vorlage – and one which is fairly similar to our extant Greek fragments – almost certain (p 108).” Again, I feel that he overstates the case, for three reasons. First, as he says himself in the next section, it is quite common for texts that we are sure were originally composed in Coptic to have many Greek loanwords.  Second, the presence of Greek loanwords could simply indicate that the translator spoke  Greek more fluently than Coptic. Third, loanwords usually enter a language because the target language doesn’t already have a word that expresses the concept effectively and this is likely to be the case with at least some of the words cited eg sabbaton, sarx.
  3. Additional features of Greek loanwords in Coptic Thomas: Gathercole draws attention to Stephen Emmel’s index to the Coptic Gnostic Library, which contains 372 instances of “words borrowed from Greek” that are not proper nouns. This, as he says, does not in itself support a Greek original or Vorlage since it was quite common for “native Coptic works” to contain a high proportion of Greek vocabulary. He provides  examples, however, where the Greek is unusual, and also points to the survival of inflected Greek forms. Again, however, this kind of oddity could result from the translator being more fluent in Greek than in Coptic.
  4. Greek Gospels: Here, Gathercole argues that the genre of Thomas is “overwhelmingly” Greek: because of its designation as a gospel; because it is referred to as a gospel in the patristic references; and because it “was intended as a collection of saving words. Even if it is not a Gospel in the canonical sense, it is probably a Gospel when considered on its own terms (p 110).” He then goes on to demonstrate that, although the area is “messy and difficult to penetrate” the majority of scholarly opinion is that the original language of the gospels that we have is Greek – that “the Gospel genre is overwhelmingly a Greek-language genre (p 115)” and thus that the original language for a gospel is most likely to be Greek. This hinges on the assertion that Thomas  is a typical Greek gospel. In fact, however, if we are to accept Richard Burridge’s argument (and most scholars do) that the canonical gospels are a sub-genre of the Graeco-Roman genre bios (Burridge, R.A. What Are the Gospels?: A Comparison with Graeco-Roman Biography: Wm. B. Eerdmans Publishing, 2004), Thomas simply doesn’t make the cut. Most obviously lacking of Burridge’s criteria is the narrative about Jesus’ life and, in particular, the account of his death. Thomas may fit into the genre of Greek sayings-source, but these are not exactly thick on the ground to provide comparisons.
  5. Greek originals of Nag Hammadi tractates: Gathercole shows that the weight of scholarly opinion is that the majority, if not all, of the texts in the Nag Hammadi library were originally composed in Greek. He concludes “if the scholarly consensus on the rest of Codex II is right, this is at least circumstantial evidence in favour of a Greek original for Thomas. It is strong evidence for a Greek Vorlage to the present Coptic translation, and the more evidence for Greek one finds in all this, the higher the burden of proof on Semitic theories (p121).” This seems to me to be an overstatement of the strength of circumstantial evidence.
  6. Close similarity to early Greek parallels: Here, Gathercole points to the similarity between the Greek text of Thomas and the Greek texts of a number of other Gospels, both canonincal and non-canonical. The major problem with the case he presents is that there is no example where the level of correspondence that he reports is adequate to satisfy the criteria of copying rather than familiarity developed by McIver and Carroll. Even in their earlier, less stringent work (McIver, Robert K., and Marie Carroll. “Experiments to Develop Criteria for Determining the Existence of Written Sources, and Their Potential Implications for the Synoptic Problem.” Journal of Biblical Literature 121, no. 4 (2002): 667-87), they conclude that there needs to be verbatim correspondence in a string of at least 16 words. Gathercole’s longest example is 13 words and most are much shorter, so he can really only suggest that the author of one text was familiar with the other at the level of oral transmission.  Thus, his conclusion: “Unless one regards a Greek Thomas as the ultimate source of the canonical versions, we are left with the strong likelihood that Thomas incorporated known Greek tradition” again overstates the case, since the evidence could again be accounted for by the development of parallel eyewitness traditions (see previous post). A second problem is that it is quite clear that the P. Oxy fragments which which he is working are not the original source of Coptic Thomas. P. Oxy 655 contains the prologue and sayings 1-7, then saying 24, which could be explained if it were a collection of someone’s favourite sayings from Thomas.P. Oxy 1, however, contains sayings 26-33 with lines 2-3 of saying 77 interpolated between saying 30 and saying 31, which suggests that there was a Greek version of Thomas in which the sayings appeared in a different order to that of Coptic Thomas. Thus it seems to me that one cannot assume that the conclusions about similarity with the canon reached from the Greek text necessarily apply to the Coptic text.

Gathercole concludes that the evidence provided means “that a Greek Vorlage to the Coptic version of Thomas is a virtual certainty, with proposals for a translation into Coptic from another language being highly speculative (p 125).”  As is obvious from my various comments, I do not think the case he makes is a strong as he suggests.

He continues “Moreover, the close parallels in phraseology between the Greek texts of Thomas and other Gospels are perhaps the strongest evidence for the incorporation of Greek tradition at the stage of Thomas’s composition (p 125).” The psychological research literature on human memory suggests, however, that the parallels he mentions are not particularly close. This will, I think, have significant implications for what follows, since he indicates that this will be an important factor in the case he will build “that Thomas is likely to be dependent upon Mathew and Luke, as well as upon some other early Christian literature.”