Tuesday, May 29, 2007

Nicholas Perrin, Thomas: The Other Gospel: Reflections

Nicholas Perrin, Thomas, the Other Gospel has recently appeared in the UK and is due to be released in the US in October. I would like to blog some of my reflections on the book here. The book comes with some overwhelming endorsements, e.g. Scot McKnight, "a stunning achievement that can set the standard for the next generation of scholarship", and to an extent these will affect the reading experience. They set up very high expectations, and it may be that those who remain unpersuaded by Perrin's case will react more strongly than they might otherwise be inclined to. As will become clear from my remarks below, I am not persuaded by its central thesis, and I will spend most of this blog post explaining why that is so.

At the outset, I should underline that the book has some very good features. In particular, I am going to find the book useful in teaching because there has been a lack of books on the Gospel of Thomas from the non-independence perspective that are at the same time accessible to undergraduates. Perrin's book uses translations of the texts and explains terms when he first uses them, synoptic problem, soteriology, ecclesiology etc. There is footnoting but it is about halfway between monograph level and introductory book level, so it's relatively unobtrusive for the beginner but present for the more advanced student. There is a fairly full bibliography too, but it's select and not comprehensive. Naturally I would throw some others into the bibliography that Perrin keeps out, e.g. Snodgrass and Tuckett would always find a place for me, as classics for the so called "dependence" argument with which Perrin is sympathetic.

On the whole I like the way that the book is pitched. It does not patronise or condescend to the reader, and it is clearly aimed at scholars as well as students. Nevertheless, there are moments where some undergraduate students may get lost, where the argument becomes quite detailed and nitty-gritty without the patient step-by-step that marks it out elsewhere, and there are moments where several scholars will be frustrated because Perrin does not engage in the kind of detailed analysis that they will wish to see. Such comments may be a little unfair given that Perrin is trying to write for a broad audience, and this is a tough balancing act to achieve, but I think the format allows him sometimes to plough a path through the space in between the introductory and the scholarly and to meet neither. I stress there the word "sometimes". Often, Perrin achieves this difficult balancing act quite well.

The book is divided into two halves, "What they are saying about the Gospel of Thomas" and "What they should be saying about the Gospel of Thomas". The first part takes a chapter each to discuss Stephen Patterson, Elaine Pagels and April deConick. The second part is an exposition of Perrin's view of Thomas, in which chapter 4 (The Syriac Gospel of Thomas) expounds his 2002 published dissertation, Thomas and Tatian, chapter 5 (Challenging the Apostolic Line) looks at the place of Thomas in the Christianity of the late second century and chapter 6 (The Jesus of the Gospel of Thomas) looks at the depiction of Jesus in Thomas, with special reference to Hermeticism, and there is a conclusion reflecting on the (non)value of Thomas in Christianity today.

I don't have a lot to say about the first half of the book. On the Forbidden Gospels blog, April DeConick has laid out a series of notes in which she comments on Perrin's representation of her work (Cautionary Note 1: Nick Perrin, Thomas, The Other Gospel, Cautionary Note 2: DeConick on Orality and Literacy, Cautionary Note 3: DeConick on the Historical Jesus, Cautionary Note 4: DeConick on Accretions, Cautionary Note 5: DeConick on Methodology). Nick Perrin is now responding to those posts over on Euangelion, Nick Perrin Responds to April DeConick. I look forward to spending more time with DeConick's posts and Perrin's response in due course. The dialogue between them should help to clarify some of the issues here.

The way that Part 1 of the book functions is to set up what Perrin thinks is wrong about current Thomasine scholarship, so it is not a survey of the full scene, unlike, say, the What are they saying about . . . series of books. This has some value, especially in terms of structure and readability -- we know exactly where we are in the book and how we should be reacting in each half. There is a downside, though, in that this kind of approach does not allow one to appreciate the broader sweep of current Thomas scholarship. In particular, I'd want to include Risto Uro, Ismo Dunderberg and Antti Marjanen in any major discussion of "what they are saying about" Thomas. Perrin does refer to these figures occasionally in his footnotes, but there is no discussion of any of them. This difficulty is particularly focused in two areas. First, Perrin discusses the relationship between John and Thomas a good deal, discussing the issue in the context of dealing with both Pagels and DeConick, but Ismo Dunderberg has written extensively (and perceptively) on this topic, and I'd have liked to have heard his insights brought in here. And second, Risto Uro has written two superb pieces on the relationship between Thomas and the Gospels, applying and developing the term secondary orality which Snodgrass had first used in relation to Thomas, but neither the articles nor the term get a mention here, as far as I could see, and this is a shame given the extensive discussion of the question of Thomas's relationship to the Gospels in the book. I suppose that what concerns me here is that the newcomer to Thomas studies could get something of a "Perrin vs. the world" feeling in reading this book, not realizing that there is a much broader range of positions in Thomas scholarship than one might pick up from the rhetorical strategy of the two-part "What they are saying . . ." and "What should be said . . ."

Let us turn, then, to the second part of the book, where Perrin forwards his own ideas on Thomas. The centre of gravity here is Chapter 4, "The Syriac Gospel of Thomas", which functions largely as a lucid summary of his Thomas and Tatian of 2002, which argued that Thomas was composed in Syriac, that this is demonstrated by the preponderance of catchwords found in a reconstructed Syriac text over against Coptic or Greek, and that Thomas's primary source was Tatian's Diatessaron. I have read that book and was unpersuaded by its thesis, so I was interested to see whether I might be persuaded second time round, all the more so given the stunning endorsements for the current book, and the passage of time since the publication of Thomas and Tatian. That passage of time has brought a good number of reviews, some highly critical of the book, but Perrin does not engage with any of them. He mentions Luomanen (p. 93) but does not mention any of the reviews of his book. I would regard those by Parker, Poirier, Shedinger, Williams and Jan Joosten (see also Quispel, Morrice and McL Wilson) as making some key criticisms of the thesis that need to be seriously addressed if it is to stand up. Perrin's inclination simply to summarise the thesis without engaging with his critics inevitably detracts from the persuasiveness of the restatement. Let me try to isolate some of the features that caused concern about Thomas and Tatian and which are not addressed in any detail here.


Perrin argues that Thomas was originally composed in Syriac because the number of catchwords in his retroversion (502) is far greater than the number in the extant Coptic text (269) or in retroverted (+ P.Oxy) Greek (263). The obvious danger here, and one of which Perrin is aware, is what he calls rather amusingly calls "fudging" (p. 87), but which I would want to call the problem of experimental bias, i.e. the one conducting the experiment is the one reconstructing the text. There is no control. So we are not really comparing like with like -- the experimenter's own retroversion is compared with an extant text. One of the best treatments of this issue is in Peter Williams review of Thomas and Tatian (EJT 13:2 (2004) 139-40), from which I quote:
Though this conclusion may seem impressively supported, in fact recurring problems in his reconstructions considerably reduce its support. Firstly, the reconstructions are not straightforward. Thus from the Coptic word 'earth' (saying 9) and the Coptic word 'world' (saying 10) he reconstructs the Syriac word 'earth', despite the fact that Syriac has a perfectly good word for 'world' (pp. 65-66). When it suits Perrin to render Coptic 'world' by Syriac 'earth' it is so rendered (p. 78), but on other occasions the Coptic word 'world' is rendered by Syriac 'world' (p. 83). The author is thus selecting the words used in his retroversion in order to create catchwords. Similarly tendentious renderings from Coptic back to Syriac are 'corpse' rendered by 'flesh' (p. 106), 'evening' rendered by 'night' (p. 115), and 'belongings' rendered by 'house' (p. 124). A significant proportion of the catchwords discovered can be accounted for in a similar way. (139-40).
In the same review, Williams speaks of "scores of technical errors" (140) that cause the thesis to fail, but on which I am unable to comment given my deficiency in Syriac, but Jan Joosten's review of Thomas and Tatian (Aramaic Studies 2.1 (Jan 2004) 126-130) makes the same points with several examples and concludes:
In the end, the compilation appears to be almost entirely useless. Nothing proves that the network of Syriac catchwords ever existed outside of Perrin’s imagination. (128)
Perrin's answer to the perceived problem of "fudging" is to assert the statistical improbability of certain patterns of words occurring in the text by accident (pp. 87-8), but this avoids engaging with the most important question, which is not about "a blend of speculation and luck" (p. 87), but is rather about experimental bias, the selection of specific retroversions that make catchword links where other retroversions would not have done. Given the extent to which Perrin's case relies on the retroversion + catchword argument, criticisms of the earlier book need to be taken seriously. It may be that a good counter-argument can be made, but if so, it needs to be made rather than ignored.

The Diatessaron and Johannine Material

In his review of Perrin's Thomas and Tatian (TC 8, 2003, http://rosetta.reltech.org/TC/vol08/Perrin2003rev.html), David Parker wrote:
It strikes me as very strange that, if the author was working from a harmony, he produced a text with no discernible Johannine source material. Since, if this was the form of Gospel narrative known to him, he would have had no criteria for discerning Johannine material, this result would have to be an incredible coincidence.
I had hoped that Perrin might deal with this problem in the book, but it is not mentioned, so I flag it up again here. While the Gospel of Thomas features extensive parallels with the Synoptic Gospels, it has no major parallels with the Gospel of John (only arguably a phrase here, a thought there -- nothing clear cut or extended). If Thomas's major source is the Diatessaron, this is surprising. It won't do to say that Thomas simply removed the Johannine material, because then we would be back to Thomas's familiarity with the four canonical Gospels, and the apparent economy of the Thomas and Tatian theory would be negated. Perrin's thesis is that Thomas's knowledge of the Gospels is mediated via the Diatessaron, and if his knowledge of the Gospels is mediated via the Diatessaron, why does he not feature Johannine material?

The Order of Diatessaronic Material

One of Perrin's arguments in favour of Thomas's dependence on the Diatessaron is based on the order of the Synoptic material that appears in Thomas. The argument is stressed in both Thomas and Tatian and in Thomas, the Other Gospel. There is a problem with the argument here. In all but one of the examples of Thomas following the Diatessaron's order, he is in fact also following the Synoptic order. In other words, the Diatessaron is almost always unnecessary to explain Thomas's order (I'll return to the exception below). In both books, Perrin speaks as if parallels with the Synoptic order are good enough to establish the point, but they are not. Where Thomas's order parallels the Synoptic order, it is unnecessary for it to be mediated through the Diatessaron. This difficulty was addressed in Parker's review of Thomas and Tatian (6):
His own suggestions are limited to comments on places where Thomas has the same ordering of material as Tatian. This is rather confusingly introduced with the statement that "at some points Thomas does indeed follow the order of the canonical and Diatessaronic tradition" (p. 185). But there is a problem: where examples he adduces--such as Matt 5.14b and Matt 5.15--consist of contiguous material within a given Gospel, it is hard to see why anybody should want to claim that this is evidence for having followed a harmony.
The point is not dealt with in the new book; rather, the new book simply works with the same underlying assumption, that parallels in order with the Synoptics are evidence for Thomas's use of the Diatessaron. Take, for example, the following statement:
At points the Gospel of Thomas does follow the order of both the synoptics and the Diatessaron: Gos. Thom. 8-9, 32-33, 42/43-44, 47, 65-66, 68-69, 92-93 and 93-94. (95)
My point is that the "both . . . and" is true but irrelevant. The relevant piece of data is that Thomas here follows the order of the Synoptics.

However, Perrin provides one additional example where Thomas agrees with the Diatessaron against the Synoptics. The example is:

Thomas 44 // Matt. 12.32 // Luke 12.10
Thomas 45.1 // Matt. 7.16 // Luke 6.44
Thomas 45.2-4 // Matt. 12.35, 34b // Luke 6.45

Perrin helpfully sets out the example in a clear table (96). It is difficult to quantify such things, but I would tend to feel that we would need a lot more than one example to make a strong case for Thomas's dependence on the Diatessaron. Nevertheless, if it were a very strong example, it would at least give us some important evidence, so it is worth looking in a little more detail. This is how Perrin states the case:
If Thomas were imitating the sequence of Matthew 12 this would explain Gos. Thom. 44 and 45.2-4, but would not explain the insertion of 45.1 (= Matt. 7.16). Luke as a source would explain the wording of Gos. Thom. 45, but would not explain the collocation of Gos. Thom. 45, as Matthew 12 does. The case for Thomas's dependence on Matthew or on Luke has its merits as well as its problems. The best explanation is that the hand behind Gos. Thom. 44-45 drew on a harmonization of Matthew and Luke as reflected in the Diatessaron, where, judging by the eastern witness of Ephraem and the western witness of the Middle Dutch harmony, the words of Matthew 12.32-35 seem to have attached themselves precisely at this point of the Sermon on the Mount. (95).
There are two problems with the case here. First, Thomas's familiarity with Matthew and Luke is adequate on its own to explain the order; there is no need to appeal to the Diatessaron. As Perrin's chart makes clear, there is a simple map of parallels here. Matt. 12.32-35 is in mind throughout; Luke 6.45 is parallel to the last two verses in that passage, and Luke 6.44 is parallel to Matt. 7.16. The harmonist proceeds naturally from Matt. 12.32-35 to Luke 6.44-45 to Matt. 7.16. There is nothing out of place or surprising here. The fact that these passages appear together in Aland's Synopsis and Huck-Greeven's Synopsis is not because they are dependent on the Diatessaron, or Thomas, but because there is a clear and natural pattern of parallels.

Second, Perrin's comments here assume that these parallels are undoubtedly found together in the Diatessaron, but this is by no means clear. The reference to "the eastern witness of Ephraem" is an error; as David Parker points out in his review of Perrin's earlier book (9), the passage does not even appear in Ephraem's commentary, and as he goes on to note:
The Persian Harmony contains only the material found in Matthew 12.33ff. The Arabic follows the order Luke 6.44 - Matthew 7.17f - Luke 6.45. We thus already find a dearth of Eastern witnesses to fulfil the Petersen criterion, accepted by Perrin.
In other words, it is not clear, in this one case where Perrin attempts to demonstrate an agreement between Thomas and the Diatessaron against the Synoptics, that the ordering is Diatesseronic.

The Importance of Oxyrhynchus

If I have been critical of Perrin for not always having engaged with his critics, there is one area where he does respond to some criticism of his earlier book. In a section headed "Objections considered" (97-99), he notes that some have objected that his thesis provides only a small window for the writing of Thomas. It has to be between the writing of the Diatessaron, dated to 173, and the dating of the first extant text of Thomas, P.Oxy.1, which Grenfell and Hunt dated to roughly 200. This gives a tight period for all the following to have taken place: (a) Thomas becomes familiar with the Diatessaron; (b) Thomas writes his Gospel in Syriac in Edessa; (c) Thomas is translated into Greek; (d) A Greek manuscript copy finds its way to Egypt. These things all take place within about 25 years. Perrin mounts a good defence of his thesis here, and notes that this is adequate time for Thomas to be disseminated, and he encourages readers not to be too dogmatic about the year 200. I think that that is right; palaeography is not a precise science and we can certainly allow a generation either side of that 200 date, say 175-225. There is still a lot to squeeze into one generation or so, but it is not impossible. P.Oxy. 1 could be as late as 225, but it could also be as early as 175.

I remain concerned, however, about an issue here which Perrin does not address, and which adds a difficulty for the Syriac Thomas hypothesis, the issue of verbatim agreement between the P.Oxy. fragments of Thomas and the Synoptics. There are good several examples of verbatim agreement in Greek, and one is very strong, and it is a fact widely ignored in Thomas studies. Thomas 26 in P. Oxy. 1.1-4 features a thirteen word verbatim agreement with Luke 6.42 (position of ἐκβαλεῖν agreeing with Matt. 7.5b), a fact all the more striking in that it is a very literary Greek construction, ὁ + phrase + noun, and that the piece in question is only fragmentary. This evidence needs accounting for on the Syriac Thomas hypothesis.

Overall, there remain too many question marks over the thesis of Thomas and Tatian, in spite of a helpful restatement in Thomas, the Other Gospel. What would help would be some critical engagement with the reactions to the first book.

Most of my reflections have focused on Chapter 4, which is the heart of the book, since it is the restatement of the case for the Syriac Thomas dependent on the Diatessaron. But the remaining two chapters of the book deserve mention, especially Chapter 5, "Challenging the apostolic line", in which Perrin has some interesting and helpful reflections in Logion 13. He follows others (e.g. Francis Watson) who see Simon Peter and Matthew here as cyphers for the Gospels of Mark and Matthew, and Thomas commenting negatively on them. Perrin situates Thomas in the late second century, contemporary with Irenaeus, and in tension with the combination of Mark's and Matthew's Gospels. "It is the first two Gospels together that stand opposed to Thomas" (p. 115, emphasis original). There is a lot of material of interest and insight in this chapter, but again, Perrin's late dating actually sits a bit less comfortably with his thoughts here than an earlier dating, in the 140s, for example, would do. In the 170s, we have Irenaeus's stress on the fourfold Gospel (which Perrin cites) and we have Tatian writing a harmony of the four Gospels. And then Perrin's Thomas stresses only two of the Gospels, Mark and Matthew, and has extensive parallels to only three of them, the Synoptics. A Thomas that first emerges in the 140s makes better sense; all three Synoptics are known to him; Mark and Matthew have some status, as in Papias, but Luke is a relative newcomer. John is newer still, and is not mentioned by name in Logion 13, nor is it the subject of extensive parallels. Nevertheless, this chapter is a very helpful contribution to the debate on Thomas and it is one I look forward to engaging with further in my own research.

The book is relatively free of typographical errors, but I spotted a few:

p. 95: the only sense I can make of the flow of argument on this page is to assume that paragraphs two and three have become switched around by mistake.

p. 61, n. 30: the publication date for Wrede is 1901, not 1910.

p. 133: "as late the sixth century" should be "as late as the sixth century"


Andrew Criddle said...

The problem with Perrin's work is that he moves directly from presenting strong evidence that Thomas is a/ at least partly based on some form of harmony and b/ at least partly based on Syriac gospel texts to claiming that hence Thomas is based on the Diatessaron.

IMO he needs to engage more closely with the complex problems both of harmonies predating the Diatessaron and Syriac Gospel texts predating the Diatessaron

J. B. Hood said...

Mike Bird has posted a response by Perrin to deConick's response.

Eric Rowe said...

Thanks for your review of this.

I've had nagging problems with his catchwords thesis too. Unfortunately, I haven't started my Syriac studies yet, and even when I will have, I think critiquing his reconstruction in a detailed way would require a command of both Syriac and Coptic that goes well beyond even that possessed by many eminent scholars of early Christian literature.

I do know enough about learning ancient languages to know that the breadth of vocabulary we know well enough to use in composition or in translating out of one ancient language into another is always going to be smaller than the vocabulary we know well enough to recognize when we see it. In his Syriac reconstruction of Thomas, two thirds of the logia have no parallel in the Old Syriac or Diatessaron, meaning that in reconstructing those logia in Syriac Perrin had to rely on that lower level of vocabulary knowledge. Perrin very well may be outstanding in Syriac. But how many words does he really know well enough to reproduce in composition? Maybe 3,000 if he's very good at it? At any rate, his available vocabulary was necessarily much smaller than an ancient author whose native tongue was Syriac. It was also a smaller vocabulary than the Coptic vocabulary commanded by the author/translator of Coptic GThom. It seems to me that this weakness would guarantee that his reconstruction would have words used with greater frequency than an original Syriac actually would have had, thus increasing the apparent significance of catchwords.

Add to this the fact that the triliteral roots of the Semitic languages result in more words that are cognates of other words, and the fact that it would be hard for Dr. Perrin to reign in his own subliminal temptation to make the catchword theory work, and from the perspective of someone who doesn't know Syriac, the argument seems suspicious.

In fairness to him, though, I certainly have no problem with assuming that GThomas is a unified work. Those who divide it up into various strata or gradually accruing logia seem simply to assume the basic disunity of the work. I tend to think the burden of proof is on them to demonstrate disunity. And they haven't seemed to want that burden.

Anonymous said...

Jan Joosten writes a very good (slamming) review of "Thomas and Tatian" in Aramaic Studies 2.1 (Jan 2004) 126-130

Mark Goodacre said...

Thanks for all these comments. Steph: I've now read the Joosten review and have added reference to it to the post.

Anonymous said...

Great review - many thanks.


Anonymous said...

1. Fudging the translation into Syriac to suit the theory.
“So we are not really comparing like with like -- the experimenter's own retroversion is compared with an extant text.”

The answer to Perrin’s ‘maximum correspondence’ Syriac retroversion would be to do a ‘minimum correspondence’ version and see whether his theory still holds under these conditions. Otherwise he is just being criticised in general terms, without clarity as to whether the criticisms are overwhelming or just marginal.

2. Why so little Johannine material in Thomas?
“Perrin's thesis is that Thomas's knowledge of the Gospels is mediated via the Diatessaron, and if his knowledge of the Gospels is mediated via the Diatessaron, why does he not feature Johannine material?”

This is a fair point which Perrin needs to address.

3. Order
“the new book simply works with the same underlying assumption, that parallels in order with the Synoptics are evidence for Thomas's use of the Diatessaron. Take, for example, the following statement:
At points the Gospel of Thomas does follow the order of both the synoptics and the Diatessaron: Gos. Thom. 8-9, 32-33, 42/43-44, 47, 65-66, 68-69, 92-93 and 93-94. (95)
My point is that the "both . . . and" is true but irrelevant. The relevant piece of data is that Thomas here follows the order of the Synoptics.”

This misunderstands Perrin. The above quotation in his book is immediately preceded by the words “Whereas it is regularly claimed that the sayings in Thomas do not reflect the sequence of the canonical gospels and this in turn often becomes, as it does for Patterson, the basis for arguing for the collection’s independence, it must be said (as Patterson concedes) that this claim is not completely true. At points the Gospel of Thomas etc ....”

All that Perrin is saying here is that where Thomas follows the synoptics (which is a problem for Patterson’s theory) his own theory readily supplies the solution: Diatessaron follows synoptics here. So he is claiming that Thomas follows Diatessaron both when it follows synoptics and when it doesn’t. If he can demonstrate that, his theory is very strong indeed.