Does the lost Gospel Q Exist? | Dr. James McGrath Vs Dr. Mark Goodacre

[Secret Alias]

If the Marcionite Evangelion is earlier than canonical Luke (which it clearly is), and the two-source hypothesis for the Evangelion is valid

I had dinner with Trobisch recently and didn't bring up this point (mostly because I am much more polite in public than I am on board's like this). I really don't think Tertullian is working from a Marcionite gospel or canon. I think he is adapting a previous work against Marcion presuming as its starting point that Marcion corrupted Luke and that canonical Luke as the "true gospel of Marcion" can used to disprove Marcion. I see no or very little evidence for Tertullian claiming to actually be working from Marcion's gospel. Not sure how useful "data science" would be in that case.

And I am saying this not as someone disputing that Marcion's gospel is the earliest (I acknowledge this) but as someone who feels Against Marcion misrepresents the Marcionite gospel.

[Paul the Uncertain]

My impression is that Q adherence reflects a robust tolerance for potential false positives (erroneously "detecting" a document that never existed). Some choice needs to be made, and why a consensus formed favoring false-negative avoidance isn't clear.

If we have both (1) indicators that Luke used Matthew, and (2) indicators that Luke could not have used Matthew, you naturally have to weigh the evidence for both, lean towards one while proposing a explanation for the other. Both sides do this.

Goodacre cites (1) as evidence that Q is not needed, and then proposes alternative explanations for (2). He seems to think all the changes in the Q material (i.e. #2) seem consistent with Luke's material, so Luke must have made those changes as he copied Matthew.

Q-Source proponents do the opposite, citing (2) as evidence for Q while proposing alternative explanations for (1). They seem to think the alternating primitivity of the Q material cannot be accounted for without an original 3rd source, not to mention the numerous surprising omissions in Luke (if he knew Matthew). Regarding (1), those could be accounted for by resorting to later scribal interference over the course of decades of copying, among other things.

Both approaches, in my opinion, are legitimate, and it's just a matter of figuring out which one is more convincing. What doesn't seem convincing is Goodacre's idea that we should favor his approach simply because it doesn't include additional entities.

I was trying to stay within a narrow frame regarding Q: in what sense does Professor Goodacre appeal to Occam's Razor, when there are so many ways that the term has been used? Without him here to explain, we are somewhat stuck.

I think there is an interesting methodological concern when hypotheses which are constrained by what exists compete with hypotheses which need satisfy no hard constraints beyond being well-formed. Can the following situation arise in something whose features have never been observed, and which is designed to fit the evidence well?

the numerous surprising omissions in Luke (if he knew Matthew)

It seems intuitive to me that real things with fixed features are more likely to have surprising features than hypothetical things crafted to avoid surprise (and in the case of Q crafted as a defeater to a family of seriously possible hypotheses, that information leaked between Matthew and Luke). To the extent that "the best explanation" corresponds to the relative absence of surprising features, then the possibly real may have an advantage over the categorically real.

I don't know if that is what Professor Goodacre was getting at. Whether or not it was, I haven't worked through whether the concern would hold up to scrutiny.

In a separate statement in the video, Professor Goodacre expressed doubt that NT scholars would detect a (counterfactual) GMark that didn't survive antiquity except in the uses GMatthew and GLuke had made of GMark. He didn't explain that remark either, nor could I find an explanation in his paper that he mentioned in the video. Possible "false positive tolerance" in the guild (e.g. confident acceptance of hypothetical Q as a real work) was a thought of my own about what sort of issues might come up in evaluating the guild as a "detector" of lost (if ever existing) documents.

[Ken Olson]

I may be able to shed some light on how Goodacre thinks Occam's razor ought to be used. He (and I) are heavily influenced by Michael Goulder, who used the term several times in his book Luke: A New Paradigm (1989). Goulder took the term paradigm from Thomas Kuhn, and he argues for a new paradigm (Luke's use of Mark and Matthew) to replace the old paradigm (Matthew and Luke independently used Mark and Q). In his first reference to Occam in the book, he wrote:

On the [old] paradigm we might have hoped that Q would preserve for us some authentic, individual tones of the Baptist; but the fact is, as I have illustrated, that the Baptist not only speaks with the same tones and phrases as Jesus, but with the same tones and phrases as the Matthaean Jesus. Some simple-hearted followers of Occam might be beguiled by this into reducing the number of hypotheses, since we now have too many. Since Q’s vocabulary and Matthew’s seem to be the same, and since sophisticated defenders of the paradigm will allow that Q is post 70, and so in the same decade as Matthew, and since Q also shares most of Matthew’s theology, it looks as if either Q or Matthew could go. Either Matthew wrote Q, or Q wrote Matthew. (pp. 14-15).

Goulder's point is that to be credible a hypothetical source has to be demonstrably different from an extant source, otherwise why hypothesize it in the first place? A hypothetical source can always offer at least as good an explanation as an extant source because we can always hypothesize it to have any characteristic the extant source has.

Frequently when i have argued that a certain characteristic of Luke's source is a characteristic of Matthew people who are not specialists in the synoptic problem (and some that are) ask: "How do you know it wasn't in Q?"

I don't. I can never know something was not present in a hypothetical source. My response, though, is usually to ask: so if Q looks like Matthew, how do you know the document you've been calling Q was not Matthew? Can you demonstrate that it was necessarily different from Matthew? If not, why insist that it's different from Matthew?

Best,

Ken

[vocesanticae]

We are compelled to believe in the existence of Mark, Matthew and Luke because we have manuscripts. Their existence is therefore demonstrable. We do not have manuscripts of Q.

When mythicists use similar logic against the existence of Jesus they are met with ridicule.

Mark, Matthew, Luke and John were (together) the officially sanctioned "gospel" (singular) of the Church. Four as one is an absurd proposition. It is absurd to suggest that anyone NATURALLY understood four texts to be one gospel. But more than that we have testimony from antiquity that the Marcionites thought these texts were forgeries (De Recta in Deum Fide). They are forgeries. The Marcionites were right. So what are we left with? A choice between an undemonstrable source text and demonstrable forgeries. Bad choices. But it is an even worse choice to mistake canonical Mark for a source text for all gospels. It is only a source text for the forgeries of Matthew and Luke.

I don't get why we just can't admit we will never unravel the origins of the gospel. What the ur-gospel looked like. At least with any degree of exactness. It's part of life to accept limitations. The hubris of humanities scholarship. We would choose a certain nothing than an uncertain something. (roughly the equivalent of lieber will noch der Mensch das Nichts wollen, als nicht wollen.)

Disambiguating voices, even underlying voices, is more and more possible these days with Natural Language Processing.

BeDuhn, Gramaglia, Daniel Smith, and I all concur that Marcion's Evangelion has two clear underlying sources. If the Marcionite Evangelion is earlier than canonical Luke (which it clearly is), and the two-source hypothesis for the Evangelion is valid (which it is), then that is likely the way we can reconstruct the earlier underlying gospel strata/texts in this movement. The Qn source is highly coherent in terms of its overarching Aesopian framing, its combination of Aesopian logia/fabulae and pharmakos narrative, its persistent focus on the poor, its repeated emphasis on women leaders, and its linguistic patterns. These patterns, across many features, prove distinct from the two main sections that cluster/overlap with early Mark, which is focused on controversies over miracles and ritual piety, male characters, etc. As just one linguistic example, lemmas with "en" (episolon nu) in them have a statistically significant variation between these two portions of the Evangelion, with extremely high frequency rates in Qn, and extremely low frequency rates in early Mark.

Rigorous data science can clarify far beyond the muddled mess that currently obtains.

If scientists today can map black holes that formed billions of years ago billions of light years away (!), then we sure as heck should be able to reconstruct significant voices echoing to us from and through the texts of 2000 years ago.

The admission that a person or group of people haven't been able to achieve scientific-historical clarity in the pass should never be confused with the proposition that a new scientific achievement cannot occur, now or in the future.

I may be misunderstanding you, but you seem to conflate your own theory of “Qn” and what it “means,” with the methods of computational linguistics and open source knowledge that (very admirably) you promote.

There is a logical chasm between this claim, which anyone must admit —

Disambiguating voices, even underlying voices, is more and more possible these days with Natural Language Processing.

And this cluster of claims, which contains some very loaded exegesis —

If the Marcionite Evangelion is earlier than canonical Luke (which it clearly is), and the two-source hypothesis for the Evangelion is valid (which it is), then that is likely the way we can reconstruct the earlier underlying gospel strata/texts in this movement. The Qn source is highly coherent in terms of its overarching Aesopian framing, its combination of Aesopian logia/fabulae and pharmakos narrative, its persistent focus on the poor, its repeated emphasis on women leaders, and its linguistic patterns. These patterns, across many features, prove distinct from the two main sections that cluster/overlap with early Mark, which is focused on controversies over miracles and ritual piety, male characters, etc.

I really don’t object to the substance of this portait, as a theory. It may be as correct as any portrait can be of earliest Christianity (or whatever we should call it). My problem is that you seem to think that this account just falls out of the “science” as the only possible account. In the recent youtube, your disparage standard accounts as “pet” theories, but seem to think that yours is not a pet theory. But the “Aesopian” and “pharmakos” characterizations, as well as the strong claims about class and gender, are very much interpretations or meanings that seek to define. They are in a different category from those “linguistic patterns” that are susceptible of quantitative analysis. I am trying to make a basic appeal to the distinction between the hermeneutics of a cultural artifact like the Bible, and any scientific comprehension of its outward, empirical, quantifiable attributes.

As for the Marcionite text in particular, you declare that “the two-source hypothesis for the Evangelion is valid.” I can see the outline of such a theory in BeDuhn 2013, except that Jason is quite careful to distinguish between what he knows, and what he thinks a reasonable theory (of a Q within *Ev) would look like. But you seem confident that the two-source idea has already been demonstrated and widely accepted.

What am I not understanding?

My open science book of a half million words details the five hypotheses and offers hundreds of pages of proofs with meticulous word counts, cluster analysis, binomial distributions probabilities, etc. I'm also the first person to normalize and enrich (with lemmatization and POS and morphological tagging) all of the Evangelion reconstruction data from past editors, allowing for myself and others to analyze the data and correlate it with the canonical version scientifically for the first time. Refute any of the hypotheses or any of the proofs if you can.

[Paul the Uncertain]

Yes, Ken, that's very helpful, I think. Thank you also for the Goulder "backstory."

A hypothetical source can always offer at least as good an explanation as an extant source because we can always hypothesize it to have any characteristic the extant source has.

That is pretty much what I suspect, and if that is what Professor Goodacre was getting at, then I think it would be at least something that would deserve to be addressed.

Effective rebuttal might be that the hypothetical satisfies constraints that are in some sense comparable with the constraints inherent in being something that actually exists with fixed features.

It seems too sweeping (just at the impression level) that no comparison between a hypothetical and something actually observed could ever be justifiable. It does, however, seem that caution would be warranted.

[mlinssen]

My open science book of a half million words details the five hypotheses and offers hundreds of pages of proofs with meticulous word counts, cluster analysis, binomial distributions probabilities, etc. Refute any of the hypotheses or any of the proofs if you can.

Hypothesis 2. When Luke has a parallel in Matthew and/or Gos. Thomas and those parallels are explicitly corroborated by GMcn, then this confirms their existence in Qn

Hi Mark, your version 3.02 names 8 Thomas logia throughout the reconstruction (footnotes 167, 246-247, 393, 452, 500, 509-511, 522, 631) and I count 9 on page 79. Perhaps I have missed some, then I apologise in advance

I have gone by Klinghardt's reconstruction as his reconstruction contains the highest percentage of LK2 3-24, namely 73.7% - whereas yours contains 33.3% there. What I have found is 57 Thomasine parallels:



These 57 Thomasine parallels in *Ev are only 4 less than in canonical Luke; now, even though your material amounts to roughly half of Klinghardt's, I would have expected more than 8 references to Thomas in your reconstruction.
As BeDuhn mentions 15 parallels and Klinghardt 5, you hold the middle position and likely have evaluated both their findings; I'd appreciate it very much if you could comment on the numbers presented

[vocesanticae]

My open science book of a half million words details the five hypotheses and offers hundreds of pages of proofs with meticulous word counts, cluster analysis, binomial distributions probabilities, etc. Refute any of the hypotheses or any of the proofs if you can.

Hypothesis 2. When Luke has a parallel in Matthew and/or Gos. Thomas and those parallels are explicitly corroborated by GMcn, then this confirms their existence in Qn

Hi Mark, your version 3.02 names 8 Thomas logia throughout the reconstruction (footnotes 167, 246-247, 393, 452, 500, 509-511, 522, 631) and I count 9 on page 79. Perhaps I have missed some, then I apologise in advance

I have gone by Klinghardt's reconstruction as his reconstruction contains the highest percentage of LK2 3-24, namely 73.7% - whereas yours contains 33.3% there. What I have found is 57 Thomasine parallels:



These 57 Thomasine parallels in *Ev are only 4 less than in canonical Luke; now, even though your material amounts to roughly half of Klinghardt's, I would have expected more than 8 references to Thomas in your reconstruction.
As BeDuhn mentions 15 parallels and Klinghardt 5, you hold the middle position and likely have evaluated both their findings; I'd appreciate it very much if you could comment on the numbers presented

I haven't done a careful comparison of the Thomas traditions with Marcion's Gospel. It's on my list to do for future improvement cycles. Thank you for bringing this to my attention. In my view, the Thomas traditions definitely need to be included fully in the signals mapping, not least because they often contain earlier/simpler forms of traditions than what are found in the canonical texts. While Goodacre and Gathercole on the whole are correct that Thomas as we have it shows clear reliance on the canonical forms of these texts, that doesn't mean that 100% of its signals stem from the canonical signals. As with most of these early texts, we are dealing with multi-directional contaminations.

[mlinssen]

My open science book of a half million words details the five hypotheses and offers hundreds of pages of proofs with meticulous word counts, cluster analysis, binomial distributions probabilities, etc. Refute any of the hypotheses or any of the proofs if you can.

Hypothesis 2. When Luke has a parallel in Matthew and/or Gos. Thomas and those parallels are explicitly corroborated by GMcn, then this confirms their existence in Qn

Hi Mark, your version 3.02 names 8 Thomas logia throughout the reconstruction (footnotes 167, 246-247, 393, 452, 500, 509-511, 522, 631) and I count 9 on page 79. Perhaps I have missed some, then I apologise in advance

I have gone by Klinghardt's reconstruction as his reconstruction contains the highest percentage of LK2 3-24, namely 73.7% - whereas yours contains 33.3% there. What I have found is 57 Thomasine parallels:



These 57 Thomasine parallels in *Ev are only 4 less than in canonical Luke; now, even though your material amounts to roughly half of Klinghardt's, I would have expected more than 8 references to Thomas in your reconstruction.
As BeDuhn mentions 15 parallels and Klinghardt 5, you hold the middle position and likely have evaluated both their findings; I'd appreciate it very much if you could comment on the numbers presented

I haven't done a careful comparison of the Thomas traditions with Marcion's Gospel. It's on my list to do for future improvement cycles. Thank you for bringing this to my attention. In my view, the Thomas traditions definitely need to be included fully in the signals mapping, not least because they often contain earlier/simpler forms of traditions than what are found in the canonical texts. While Goodacre and Gathercole on the whole are correct that Thomas as we have it shows clear reliance on the canonical forms of these texts, that doesn't mean that 100% of its signals stem from the canonical signals. As with most of these early texts, we are dealing with multi-directional contaminations.

Thank you very much!
I'm really looking forward to that inclusion

[vocesanticae]

If the Marcionite Evangelion is earlier than canonical Luke (which it clearly is), and the two-source hypothesis for the Evangelion is valid

I had dinner with Trobisch recently and didn't bring up this point (mostly because I am much more polite in public than I am on board's like this). I really don't think Tertullian is working from a Marcionite gospel or canon. I think he is adapting a previous work against Marcion presuming as its starting point that Marcion corrupted Luke and that canonical Luke as the "true gospel of Marcion" can used to disprove Marcion. I see no or very little evidence for Tertullian claiming to actually be working from Marcion's gospel. Not sure how useful "data science" would be in that case.

And I am saying this not as someone disputing that Marcion's gospel is the earliest (I acknowledge this) but as someone who feels Against Marcion misrepresents the Marcionite gospel.

Adv. Marc. 4.1, 4.2, 5.1, and passim in books 4 and 5 are quite clear he is working directly from Marcion's scriptural canon, not just the Evangelion, but also the Antitheses and Apostolos. Indeed, the entire treatise reads as a verse by verse refutation of that text, criticizing it for lacking a named author, for changing the wording and occasionally its order, etc. If you take Roth's text as the basis, then Tertullian directly attests to 90% of the Evangelion content that can be restored. If Klinghardt's, then 50% of the verses. For me and BeDuhn, it's about 80%. Whatever position one takes, at any point in the scholarly spectrum, there is complete agreement that T is a direct, albeit hostile witness of Marcion's scriptures, at least in the form in which he received them.

To the best of my knowledge, every scholar who has looked carefully and published on Tertullian, Tertullian's AdvMarc and Marcion's Gospel concurs on this point. The only point of debate in the literature related to this is the language of the Marcionite text, whether T had it in Greek, Latin, or (as Vinzent told me he now thinks) bilingual Greek and Latin, akin to several other early bilingual Greek-Latin manuscripts of the Pauline corpus.

Do you know of any scholar in any publication who has challenged the position that T had direct access to the Marcionite scriptures? If so, I'd be eager to know about it. If not, then it seems that your position is baseless contrarianism.

[Secret Alias]

If you take Roth's text as the basis, then Tertullian directly attests to 90% of the Evangelion content that can be restored

Gag me with a spoon (as we used to say).