Twice More— Thomas and the Synoptics: A Reply to Simon Gathercole,The Composition of the Gospel of Thomas,and Mark Goodacre,Thomas and the Gospels

In the late second century Irenaeus of Lyons promulgated one of the most enduring decisions in the history of Christianity: there shall be four gospels, no more, no less. Matthew, Mark, Luke and John were sufficient and complete—fewer gospels would not be sufficient; more would be too many. Why? Irenaeus’s argument went thus: ‘There are four zones of the world … four principal winds … four faces on each of the cherubim … and four covenants between God and the human race’. Thus, there should be precisely four gospels, no more, no less (Against Heresies 3.11.8). Irenaeus’s opponents championed other gospels: of Peter, of Truth and, possibly even, of Thomas. However, Irenaeus believed them to be theologically dangerous. And so he argued: four winds, four gospels, no more, no less.

The four-gospel canon has perdured through 18 centuries, through the theological shifts of Augustine, Aquinas, the Reformation and even Liberal Theology. When the Deists of the eighteenth century first began to search for a historical teacher behind the miracle-working savior, they did not argue about the sources. Their research was confined to the fourfold gospel canon, no more, no less. Irenaeus’s arguments about the four winds and the fourfold faces of the cherubim were now long forgotten. Another assumption—unproven—had replaced the arguments of Irenaeus: these gospels were the earliest and therefore the most reliable. Theological authority and historical priority were met in the fourfold gospel canon.

Then Q appeared in the sources. What was Q? Was it part of the four, or was it something in addition to the four? In any event, whether part of, or in addition to, the four Gospels, it was, by definition, earlier than them. This was, and still is, the dilemma posed by Q. It problematized for scholars the principle of the fourfold gospel canon. What if historical priority and theological authority could not be wed after all? Fortunately, Q was still lost, and thus not yet really real.

Then the Gospel of Thomas was discovered. Here was a gospel with so much of the same material as Matthew, Mark and Luke, and sounding so much like John. As to the date of this new rival, it was elusive. Many of its sayings appeared to be earlier; its parables, for example, were unembellished with the secondary allegorical features that so often turn up in the Synoptic versions. This also posed a dilemma, only more forcefully now than Q, for Thomas was a real gospel. What if historical priority and theological authority could not be wed after all?

What does Irenaeus and the fourfold gospel canon have to do with these two new works of scholarship? The evidence for deciding the question of the relationship between Thomas and the Synoptics has not changed much in 50 years. Most of the details examined in the books by Simon Gathercole and Mark Goodacre have been discussed in the literature since the 1960s. What each brings to the discussion is a new perspective, a new framing of the question, which is a proper exercise of the historian. Each chooses to frame the question as narrowly as possible, focusing on those sayings with close Synoptic parallels and the individual, sometimes single, words that connect them. To present a different perspective, I have chosen instead to frame the question as broadly as possible: what is the context in which this question matters? Why is Eerdmans ready to publish an entire monograph—a very good one, I must add—on this arcane little topic? Irenaeus and the fourfold gospel canon—this is why the question matters. There is a very large and interested audience eager to be shown why Thomas is, after all, not the fifth gospel. What is at stake is the wedding of theological authority with historical priority. This is not to say that these are the stakes for Goodacre or Gathercole, whose own theological commitments are left unstated and deemed irrelevant to the historical arguments being proffered in their books. Indeed, they are irrelevant. Historical arguments must stand or fall on their own merits, regardless of where they might lead one theologically or ideologically. Still, to pretend that the context within which these books make their argument is not fraught with theological weight is naïve. If the history of this discussion has taught us anything, it is that good arguments can be made on either side of the issue, and persuasiveness often depends on how one is willing to be persuaded.

Simon Gathercole frames the question similarly to how early participants in this debate framed it, indeed how I framed it in 1993 (Patterson 1993), and more recently in my contribution to the Tuckett Festschrift (Patterson 2011: 783-808). He asks, how often does Lukan or Matthean redaction of a saying show up in the Thomas version? He is looking for Synoptic fingerprints in the text of Thomas. When I last worked through all the parallels, adding to my own list of suspected cases of influence the instances counted by Christopher Tuckett in his separate accounting (Tuckett 1988: 132-57), I found this to be true in 16 out of 95 cases, or 17% of the time. However, that figure includes a good number of instances where a possible Q Vorlage is in play, so that Matthean and/or Lukan redaction might be too much a matter of speculation. Therefore, in a separate count using only those cases where there is a Markan original for control, I came up with 3 instances out of 25 cases. ¹ By counting fewer parallels and crediting more evidence of Synoptic influence on Thomas, Gathercole arrives at something like 50%; in about half of the parallels there is evidence for Thomas having known the Synoptics (see especially p. 210). I stand by my smaller figures, as I think he credits too much weak evidence. A few examples will illustrate.

The Greek version of Gos. Thom. 5.2 appears to agree verbatim with Luke’s revised version of Mk 4.22 (see Lk. 8.17). This would imply Thomas’s knowledge of Luke, as Gathercole argues (pp. 186-87). This may well be an instance that I should have added to my list. However, only about half of the Thomas saying is extant in the manuscript, so its precise wording is unknown. The first clause is formulated exactly like Luke—and Mark—so it tells us nothing. The second clause is, perhaps, more telling. Luke reformulates it as a relative clause, and Thomas appears to have an identically worded relative clause. However, the relative form is found also in Q’s version, so that is not uniquely Lukan. And Thomas’s verb in this clause must be reconstructed. All that remains of it are the letters φανε. It might originally have been the same as Luke—φανε[ρὸν γενήσεται—but this might not be the case. The possibilities for filling the lacuna also include φανε[ρὸν ἔσται, or even φανε[ρωθήσεται. We simply do not know.

Or, some examples from Matthew: Gathercole argues that Thomas knew Matthew because Gos. Thom. 13 mentions the apostle Matthew (pp. 169-78). This is not very convincing, in my view. The fact that a gospel bears his name probably indicates that Matthew was a fairly well-known apostolic figure, even apart from the Gospel of Matthew. Knowing of the apostle Matthew does not mean that Thomas’s author knew, let alone used, the Gospel of Matthew. In the case of Gos. Thom. 14.5, Gathercole thinks that, since Thomas specifies that ‘whatever goes into your mouth’ will not defile you, thus agreeing with Matthew’s version of this saying (see Mt. 15.11; ‘mouth’ is missing in Mk 7.15), Thomas must have known Matthew (pp. 178-79). That both Matthew and Luke specify the mouth for eating is not very much of a coincidence. (Where else shall food go?) So, we disagree about ambiguous evidence. Nevertheless, in the end Simon Gathercole proposes a solution that is more or less right. We will return to that finally, but first, Goodacre’s approach.

Mark Goodacre’s approach is more novel. In addition to noting instances of Synoptic redaction that turn up in Thomas, Goodacre is interested in cases where there is verbatim agreement between Thomas and one or another of the Synoptics. He argues, perhaps rightly, that focusing on the question of where Thomas reproduces Synoptic redaction may have obscured a more obvious fact, namely, that occasionally the level of verbatim agreement between Thomas and the Synoptics rises to a level that could not be accounted for by merely supposing a shared oral tradition. Here are the instances that he musters for consideration (see pp. 26-65):

This is an impressive list. But what does it mean? Out of 94 possible parallels there are 9 instances where verbatim agreement rises to a level that, in Goodacre’s view, could not be accounted for by shared oral tradition alone. I might, then, add these to the earlier list that I compiled of sayings where Synoptic influence might be detected. That would bring the total to 25 sayings out of a possible 94, or 26% of the time. Would that indicate that the author of Thomas was using the Synoptic Gospels?

Before deciding this, we should look more closely at Goodacre’s list. First, items 6-9 are available only in Coptic; Goodacre’s comparison is on the basis of a modern Greek retroversion by scholars who actually did know the New Testament parallels in Greek. Therefore, if the question is whether the authors of this Greek text knew the Synoptic Gospels, then the answer is yes. But this does not really answer our question. Secondly, items 1, 3, 4, 7, 8 and 9 are, in my view, from Q; item 2 might also be from Q. Though the point is moot with Goodacre (who rejects the Q hypothesis), verbatim agreement in these cases might indicate familiarity with Q rather than the Synoptic Gospels themselves. Finally, and most importantly, all of these examples are very brief, memorable sayings.

Goodacre believes that these brief sayings constitute ‘diagnostic shards’ that tip the hand of a plagiarist. His argument is based on a modern analogy. When a professor reads an elegant phrase in a student paper, Googles it, and finds it in a book or essay to which the student will have had access, she suspects plagiarism. It matters not whether the student has copied whole paragraphs word for word. A tell-tale phrase—a ‘diagnostic shard’—is enough to convict. This may all be granted. But are these nine sayings in Thomas enough to convict the author of Thomas of plagarism?

I am doubtful. Goodacre’s analogy does not seem an apt one to me. There are two things that we all presume to be true today that are not at all true of the ancient world. First, everyone in our university culture is literate and has access to books. Secondly, there is no oral tradition in modern university culture, at least not for the things that one might use in an academic paper. In antiquity, by contrast, most people were not literate and books were relatively rare. In fact, most tradition was oral. This makes a difference. When I read a student paper I know that she has access to books and that she does not have access to oral tradition. A tell-tale phrase therefore raises only one question: Did she get it from a book? Google search. Done. When reading an ancient text, the conditions are much different. I know that the author was literate, but I do not know whether he would have had access to many texts. A few, to be sure. But I do not know which texts, and I certainly cannot suppose he had access to many texts, let alone one particular text. On the other hand, I do know that he had access to oral tradition. If he is associated with the followers of Jesus, I can assume that he has heard many of the sayings and stories familiar to us from other texts from this tradition. Now a tell-tale phrase is not so easy to interpret. Is it from another text, or is it from an oral tradition that has also influenced other texts? Since oral tradition is common and texts relatively rare, I think one should tilt towards oral tradition, unless there is compelling evidence to suggest a literary source.

An analogous situation rarely arises in the modern world because oral tradition is rare. There is, however, one remaining pocket of orality in modernity: jokes. Of course, there are joke books in circulation, but I do not own any of them. Most of us do not. We simply remember jokes. I do not remember them very well. This one I do remember, though, probably because I liked it, and thus repeated it many times until it was grooved into my memory. I can even recall the oral source. I heard it first on one of Garrison Keillor’s ‘Joke Show’ episodes of the radio show A Prairie Home Companion. Now, the Prairie Home Companion website has transformed this once-oral medium into a mixed oral/digital form, but I have never looked up this joke on the website. I just remember it. From when? I don’t know. It must be 10 to 15 years ago. The joke goes like this:

A skeleton walks into a bar and says to the bartender, ‘Give me a beer … and a mop’.

Now, everyone who reads this article is going to remember this joke (trust me). And someday you are going to be at a large family gathering and your cousin Jane is going to tell you this joke. Are you going to accuse her of plagiarizing it from this article of mine? What if you were to note certain details to her, that her joke has a ‘skeleton’, ‘into a bar’, a ‘beer’ and a ‘mop’, all verbatim, and (now the pièce de résistance in your argument), she, too, says ‘walks into a bar’ and not ‘goes into a bar’. She could have said ‘goes’ but instead she said ‘walks’, ‘just like in Patterson’s article. She must have stolen it from Patterson’, you say. She and your aunt Mary are going to look at you like you are nuts, and rightly so. It is oral tradition. It could have come from anywhere. She might even be a fan of Garrison Keillor.

Now, that is not the only ‘walks into a bar’ joke I have heard on Garrison Keillor’s radio show over the years. Indulge me, just for a moment, in treating you to a little collection of them:

A horse walks into a bar. The bartender says, ‘Why the long face?’

A skeleton walks into a bar and says to the bartender, ‘Give me a beer … and a mop’.

A grasshopper walks into a bar. The bartender says to him, ‘Hey, we have a drink named after you’. The grasshopper says, ‘You have a drink called “Ned”?’

A rope walks into a bar and orders a drink. The bartender says, ‘We don’t serve rope here’. So the rope goes outside, takes a comb to one of his ends, and teases it all out so that it is good and frayed. Then he ties himself in a knot and goes back into the bar and orders a drink. The bartender says, ‘Hey, aren’t you that rope that was just in here?’ The rope says, ‘I’m a frayed knot’.

Later, at your big family gathering, you now overhear your cousin Jane in the next room telling jokes. First there is the horse joke, then the skeleton joke (again!), then the grasshopper joke (the same name, ‘Ned’), and finally the long involved rope joke. Later that evening you confront her in the hallway: ‘So you did!’ She says, ‘Yea, pretty good article, wasn’t it?’

In a situation where orality is truly in play, what evidence do you need to show dependence? Cousin Jane tells one joke. You have nothing on her. Maybe later you will overhear her telling another bar joke, another one from my little collection. Maybe even a third. Now you are really suspicious. But still, you have nothing. She just likes bar jokes. Now you hear her tell two or three more jokes, ones not on the list. So that’s it—she just likes bar jokes. You misjudged her. She is not reading historical Jesus scholarship on the side. Then you hear her relate the whole collection—same wording, same order. Now you know that she really is a crypto-Jesus scholar. She must like bar jokes, which is why she remembered Patterson’s list. But she definitely got the list from Patterson: same wording, same order.

Goodacre thinks that Thomas must have taken the poverty beatitude from Luke and Matthew because they have the same wording: μακάριοι, πτωχοί, and the pièce de résistance, βασιλεία τῶν οὐρανῶν (just like Matthew). I am not convinced. This is a brief, very formulaic logion with very few moving parts. This applies to all of Goodacre’s examples. The longer they get, the less verbatim agreement you find (in Foxes Have Holes [Gos. Thom. 86 = Mt. 8.20//Lk. 9.58], verbatim agreement in the original languages is only about 60%). To convince me that Thomas plagiarized these very brief, familiar sayings from the Synoptic Gospels, there would have to be significant verbatim agreement spread throughout the collection and a common order. Thomas has the ‘Poor’ beatitude, the ‘Hungry’ and the ‘Persecuted’, but not clustered, like in Matthew and Luke, not in the same order, and with surprisingly varied wording. In other words, the kind of evidence that tells us that Matthew and Luke took (plagiarized?) the beatitudes from Q (viz., common wording and order) is lacking in Thomas. Goodacre thinks that this raises the bar too high (see p. 19), that in the case of Thomas, a small amount of evidence is enough to show thorough literary dependence upon the Synoptic Gospels. Here is why I disagree:

First, it must be said that the Gospel of Thomas relies without question on traditions that are independent of the canonical Gospels. About half of Thomas’s sayings find no parallel in the Synoptic Gospels at all. Among these are many that are Synoptic-like in both content and style—wisdom sayings and parables and the like. I have called these ‘Synoptic cousins’ and by my count there are 26 of them. They are the equivalent of the Matthean and Lukan Sondergut. There are also many sayings in Thomas that have Synoptic parallels, but with wording so different that one could hardly assume, let alone prove, dependence. I called these ‘Synoptic siblings’ and by my count there are 15 of them. I will not live or die by any of these identifications, but the principle, I think, is clear: for the Synoptic cousins, the Synoptic siblings, and all the other, more distinctive, sayings in Thomas, there can be no question of dependence on the Synoptic Gospels. They come from an autonomous Thomas tradition. The question, then, really has to do with what I called the ‘Synoptic twins’, those sayings with close Synoptic parallels, and the focus of the two books under review. The question is, where do these sayings come from?

They could all come from the Synoptic Gospels. But what is the likelihood of this? If we know that Thomas had other sources for Synoptic-like material, we may also reasonably assume that the Synoptic Gospels would not have been the only possible source for these Synoptic twins. These sayings could also have come from other sources. Since the Synoptic fingerprint shows up so seldom in them, my assumption is that most of them do come from other sources. As Gathercole sees the Synoptic fingerprint much more frequently than I—about 50% of the time—he could rightly assume that more of them come from the Synoptics than I do. But if I read him correctly, he thinks that they all come from the Synoptic Gospels, ‘on the principle that a blood sample reflects the entire circulation’ (p. 213). I do not think that this principle is at play at all. Thomas is a complex book, a list of many layers, most of which can no longer be detected. The sayings in this collection could come from many different sources—a mix of blood types, if you will. The human body may not tolerate different blood types, but ancient texts do. If some come from, or are influenced by, the Synoptic texts, this does not mean that every Synoptic parallel shares that history. Given the relatively sparse evidence of borrowing in a relatively limited number of cases, I would assume that few sayings have this history.

On the matter of exactly how Thomas came by these few sayings, we may have more in common than one might guess. Gathercole accepts the theory that I first encountered in the work of Ernst Haenchen, and later of Klyne Snodgrass and Risto Uro, namely, that these sayings came to Thomas by way of secondary orality. ² That is, the Synoptic versions of these sayings, once performed aloud, would then have been repeated by others, and in this way become part of the oral tradition whereby the author of Thomas would have encountered them. That Thomas came across some of its Synoptic twins by this route seems right to me. It could even account for the absence of clearer markers of Synoptic history in most of these sayings. But assuming that this was the source for all the Synoptic twins would only be likely if we could assume no other sources for this material, as though once a Synoptic author wrote it down, all other forms of it then disappeared. This is not very likely.

I have recently tried to move this discussion beyond the impasse under which it has suffered for decades by pointing out that Thomas is a complex book that poses challenges to scholars that are not easily met (Patterson 2011: 784-87). We have just one copy of Thomas for most of its sayings, in translation, from a collection of manuscripts that is relatively late. We do not have the original Gospel of Thomas—not by a mile. Even if tomorrow we were to discover a Greek version of all 114 extant sayings in our Nag Hammadi version, that would only help a little. For in reading through the Gospel of Thomas it is abundantly clear that this is a layered text. Pointing to the many flaws in DeConick’s recent iteration of this theory ³ does not undermine it in the least. Many have commented over the years on the importance of this observation. ⁴ These complications in the text itself make it impossible to offer blanket statements about the date of Thomas, its theology and (especially) its various sources. ⁵ We must be humble in approaching this text, and as precise as we can be. It is possible that one Markan parallel in Thomas comes from the same pool of oral tradition that Mark drew from, and another was added to the collection by an Egyptian scribe who loved the Gospel of Mark and wanted to see a few more of his favorite sayings of Jesus, as he read them in Mark, included in this anthology. All of this is possible. It is only a matter of where the evidence leads us.

For all their merits (only a portion of Gathercole’s excellent monograph is devoted to this topic), I do not think that these two new books have managed to move the needle very much. Gathercole takes an older approach and simply tries to add to the evidence for Synoptic influence so that skeptics might see it as ‘significant’ at last. But most of the new evidence is not new. I doubt whether his efforts will finally tip the scales for those who have until now harbored reservations. Goodacre tries something new, a new model and framework. He is right. The verbatim agreements have not received much discussion in the past, but perhaps that is because it is difficult to know what to make of them. We find them only in the shortest of sayings, where there are very few moving parts and not much room for variation. In the longer sayings and parables, where variation in numerous oral retellings would make verbatim agreement very unlikely, we find almost none. Has the plagiarist here covered his tracks, or is this just the expected result of shared oral traditions?

So, we are back to the larger framework with which I began. Wading into the details still produces ambiguous results. Most scholars are still going to be left to make this call based on instinct. For those whose instincts come from Irenaeus and centuries of assuming that the four canonical Gospels are the oldest, and therefore the most reliable gospels, these new books will offer considerable comfort. Goodacre has even added the taint of ‘plagiarism’ to the ‘fifth gospel’. Thomas, he argues, used the Synoptic sayings as an ‘authenticating device’ to give credibility to his ‘newer, stranger’ material. In my view, all of this only makes sense in a post-Irenaean world, when everyone knows which gospels were first and thus the most authentic, and which came later and needed added authentication. In truth, we really do not know which gospels came first, and which later (Mark before Matthew and Luke, yes, but otherwise, no). None is truly apostolic, and to ancient ears all of this talk about Jesus would have been strange and implausible. Centuries later, the familiar canonical Gospels do not seem strange. It is as though they have been there since the beginning. But this is the illusion that started with Irenaeus. I have not yet seen convincing evidence that Thomas presents the historian with a situation that is substantially different from the canonical four. Irenaeus preferred them as earlier and more original, but historians must be wary of this orthodox conceit.