A statistical approach to the synoptic problem

[Peter Kirby]

See, if not understand, this new web site:

A statistical approach to the synoptic problem (http://www.davegentile.com/synoptics/main)

I've actually contemplated undertaking something similar at some point, but I'm glad that someone has done a lot of the hard work already. I look forward to checking the methods when I delve into stylometry at some point in the (perhaps distant) future. The combination of computers and antiquities is fascinating to me. If the data and the procedure stand up as sound, this would be a convincing argument that deserves publication in a scholarly journal.

best,
Peter Kirby

[GentDave]

Hi,
I put up the website in question, and I'd be glad to try to answer questions. It might help improve the site.

The most important section to understand is the data.
Once you've got that, you're half way there.

Dave

[Yuri Kuchinsky]

Originally posted by GentDave
Hi,
I put up the website in question, and I'd be glad to try to answer questions. It might help improve the site.

The most important section to understand is the data.
Once you've got that, you're half way there.


Hello, Dave,

Long time no talk... :)

Remember, we used to discuss these things a few years back, before they kicked me out from the Synoptic-L? (I guess too much scepticism can be a dangerous thing in NT studies? :))

In any case, I've followed some of your research over the years, but I have one basic problem with it. Namely, how do you deal with all these hundreds of Anti-Markan Agreements (also known as "the Minor Agreements")?

Don't they invalidate the 2ST right from the word go? What more statistical data might there be necessary to see that 2ST is a dud?

Regards,

Yuri.

[GentDave]

Hello again. I was wondering what the connection was to this site. :)

I tend to have a problem with the minor argeements too.
I think the study fairly strongly suggests that Luke used Matthew, and leads me to support the 3ST, but I don't want to overstate the case, either.

[Yuri Kuchinsky]

Originally posted by GentDave
Hello again. I was wondering what the connection was to this site. :)

I tend to have a problem with the minor argeements too.
I think the study fairly strongly suggests that Luke used Matthew, and leads me to support the 3ST, but I don't want to overstate the case, either.

Dave,

I'm well familiar with your study, and how it was set up, because I've been reading the Synoptic-L while you were first presenting your results there (this was all based originally on your school project).

I'm saying that your whole study is fundamentally flawed. You should have eliminated the 2SH right from the beginning, based on the Anti-Markan Agreements alone -- *before* you even started your statistical analysis.

What Anti-Markan Agreements really mean is that there are 1000 good arguments against the 2SH. In the real world (as opposed to the Synoptic Studies World) this is usually called a dud.

If you included 2SH into your statistical analysis, you might as well have included the Tooth Fairy Hypothesis (TFH) in there as well on equal footing...

For example, you say on your webpage,

http://www.davegentile.com/synoptics/main

"I do not believe the 2 Source Hypothesis can be eliminated by this study."

If your study even cannot eliminate the 2SH, what good is it at all, one may wonder? There's a problem there, Dave...

What I'm saying is that, by all rights, the 2SH should have been eliminated even *before* your study...

Regards,

Yuri.

[GentDave]

I didn't eliminate any hypothesis aprior.
The order could have turned up to be Luke => Matthew => Mark
But it turned out to suggest Mark => Matthew => Luke
So, I disagree that I should have not considered the 2SH apriori. It may well be that based on evidence, other than my study, various hypotheses may be rejected, but I did not address that.

I do think that my study can add evidence to the case against the 2SH. But can it prove it?
Let's look at it this way, if the minor agreements, by themselves don't disprove the 2SH, why would we expect my study to do that?

:)

[GentDave]

P.S.

This page has more elaborate conclusions.

http://www.davegentile.com/synoptics/interpretation.html

ending with:

It is possible, maybe even probable, that the true solution is more complicated than the ones considered here. For example, there could have been more than one version of Mark or Matthew. The problem is that as the number of hypothetical documents grows, the number of possible solutions grows exponentially, and the data that can be used to separate them is spread that much more thinly. Personally, I doubt enough data exists to come to any conclusions beyond the basic outline of what happened, with any real confidence.

So, in conclusion, based on this study and other more traditional forms of evidence, not presented here, I believe the 3SH, or some variation of it is most likely the correct solution. The study also provides almost as much support for the FH, and I do not believe the 2SH can be eliminated by this study. However, solutions that do not involve Markian priority have another hurdle to overcome, based on these results, and I think it becomes more difficult to argue that Luke did not use Matthew at all.

[Vorkosigan]

Hi GentDave! A great piece of work. I like the 3SH myself. I think the key piece of evidence lies in your brief discussion of the minor agreements -- that the evidence points to Luke editing them from Matt. What did Mark Goodacre say about this?

Vorkosigan

[GentDave]

Mark Goodacre was very positive about the study. We were working toward getting it included in an upcoming book, but that fell through, because of the work involved in turning it into a publishable piece. That's when I decided to put it on a webpage, so at least it would be available.

[GentDave]

In addition to the minor agreements, the other piece of evidence that points to Luke using Matthew, is that Sondergut Matthew looks similar to the common Matthew/Luke text in "Q".

So material that must be Q on the 2SH, looks like stuff that appears only in Matthew. (Category 202 looks like 200)

[Peter Kirby]

Hey! This deserves publication. I strenuously object to restricting this info to a web page.

best,
Peter Kirby

[Bernard Muller]

quote:

"In addition to the minor agreements, the other piece of evidence that points to Luke using Matthew, is that Sondergut Matthew looks similar to the common Matthew/Luke text in "Q".

So material that must be Q on the 2SH, looks like stuff that appears only in Matthew. (Category 202 looks like 200)"

- What about "Matthew" liking so much Q material that he expanded on it?
- If they are some 700 minor agreements between GLuke & GMatthew against GMark, then "Luke" must have slept many nights with GMatthew under his/her pillow (if "Luke" knew about GMatthew)! But it is rather strange GLuke is not obvious about a prior knowledge of GMatthew and the two gospels have many conflicts between them.
- GentDave, do you have any idea about the nature of these 700 items? I am most interested. Would many be explained by the fact "Luke" & "Matthew" have good knowledge of Greek, contrary to "Mark"?

Best Regards, Bernard

[GentDave]

Peter Kirby wrote:
Hey! This deserves publication. I strenuously object to restricting this info to a web page.

Dave:
The problem is I work in the corporate world, and don't have easy access to a good library for adding footnotes and stuff. I'm also not formally trained in Theology, and am not good with the generally verbose style. I tend to be more scientific and terse.
(Plus I have 2 little ones at home now, taking up my time.)

And given that the math and the text are very different areas of study, there are not a lot of people that can assist.

[GentDave]

Bernard Muller wrote:
- What about "Matthew" liking so much Q material that he expanded on it?

From
http://www.davegentile.com/synoptics/sayings.html

The results for category 200, are also important.

202-200 2E-06
201-200 2E-10

The 202-200 relationship should not be there on the 2SH, at least at first glance. On the 2SH, 200 is Matthew, and 202 is Q. But the 200-202 relationship is strong in spite of the fact that they cover different topics, and at least a large part of 200 is narrative. i.e. Matthew chapters 1 and 2. The only recourse I see for the 2SH is to say that 200 also contains significant amounts of Q material that Luke did not use. For example, there are cases where a “Q” phase is used multiple times in Matthew. This re-use by Matthew tends to draw 200 and 202 together. Mark Goodacre has pointed out that this phenomenon makes the style attributed to Q and the style attributed to Matthew similar. However, I think we would also have to speculate that other parts of 200 were also in Q, and that Luke did not use them even once, in order to account for the strong observed relationship.

Now I’d like to present the way the 3SH would view these relationships. On the 3SH 202 is mostly the saying source S, with a little Matthew mixed in. 102 is more than half Luke, and less than half S. 201 is less than half S and more than half Matthew. 200 is also some S and mostly Matthew.

Based on this we would not expect 102 to look much like the narrative categories of Luke, since there is a difference in genre, a difference in topic, and a difference in authorship for something like half of the material. 102 and 201 would show some symmetry around 202, since both are about half S, and share the same genre and topics. 200 would show a strong relationship to 201, since both are mostly Matthew, with some S. And finally, 200 and 202 would be related, but less strongly that 201-200, since 202 would be mostly S with some Matthew, and 200 would be mostly Matthew, with some S.

Thus, the 3SH most naturally explains the observed results, while the FH and 2SH can not be eliminated.

[GentDave]

Bernard Muller writes:

- If they are some 700 minor agreements between GLuke & GMatthew against GMark, then "Luke" must have slept many nights with GMatthew under his/her pillow (if "Luke" knew about GMatthew)! But it is rather strange GLuke is not obvious about a prior knowledge of GMatthew and the two gospels have many conflicts between them.
- GentDave, do you have any idea about the nature of these 700 items? I am most interested. Would many be explained by the fact "Luke" & "Matthew" have good knowledge of Greek, contrary to "Mark"?

Dave:

I tend to see Luke as liking some of the corrections to Mark's Greek that Matthew made, but not liking a lot of the content changes. Luke did not see Matthew as authoritative, in my view.

I'm familiar with the nature of the MAs, and many could be explained the way the 2SH suggests. However, the study suggests the MAs (cat. 212) are Matthian in style much more than Lukian, suggesting Luke borrowing from Matthew.

I also think the agreements of omission against Mark are too large to explain by completely independant action. I think there must have been a proto-Mark that they both used, or Luke knew Matthew's omissions and tended to agree.

[Yuri Kuchinsky]

[Replying to Dave]

DAVE: ... if the minor agreements, by themselves don't disprove the 2SH

YURI: Why do you think they don't disprove the 2SH?

DAVE: It is possible, maybe even probable, that the true solution is more complicated than the ones considered here.

YURI: That's for sure! So here we agree...

DAVE: So, in conclusion, based on this study and other more traditional forms of evidence, not presented here, I believe the 3SH, or some variation of it is most likely the correct solution.

YURI: Impossible.

DAVE: The study also provides almost as much support for the FH, and I do not believe the 2SH can be eliminated by this study.

YURI: Your three options, i.e. 2SH, 3SH, and FH, are the three Pink Elephants, neither of which has any connection with reality. So your study has determined that one of these Pink Elephants might be marginally preferable to others, while, nonetheless, the other two cannot be eliminated by your study.

More power to you!

Now, speaking about those Anti-Markan Agreements, in real terms, what they demonstrate is that the chances of 2SH being valid are something like 1 in 1,000,000. Now, your study has determined that these chances might actually be more like 1 in 1,000,100.

Again, more power to you!

And here's my own solution to the Synoptic Problem. NONE OF THESE GOSPELS IS THE EARLIEST! And this happens to explain ALL the evidence on the ground.

Best,

Yuri.

[GentDave]

I believe the 3SH outlines the general frame of what happened. I strongly suspect there were more complications, beyond the ability of the study to decern, however.

I also think the original gospel was highly similiar to Mark, although it may not have been Mark. The study provided additional evidence that Matthew and Luke both used something similiar to Mark, if not Mark itself.

I am in general agreement with you about the minor agreements. But without hard numbers, I prefer to understate the case. I think major proponents of the 2SH would agree that the real solution is more complex, but feel they have the basic outline.

So, what would lead you to believe that the 3SH is not at least approximately correct?

[Toto]

Dave - when I try to download your spreadsheet, it asks for a password.

[GentDave]

Toto: when I try to download your spreadsheet, it asks for a password.

Dave: That's odd. I tried it and it didn't ask fo anything.
Did anyone else have a problem?

This should be the link

http://www.davegentile.com/synoptics/bib-likely-commented.xls

[Yuri Kuchinsky]

Originally posted by GentDave

I also think the original gospel was highly similiar to Mark, although it may not have been Mark.


Dave,

Sure, it was similar to Mk in so far as it was probably very short, and lacked much in the way of sayings materials.


So, what would lead you to believe that the 3SH is not at least approximately correct?


IMO, there are some basic problems that plague both the FH and the 3SH. Among the biggest ones are,

-- An unquestioning acceptance of the canonical Markan priority.

-- Both FH and 3SH fail to account for an awful lot of material in Lk that is completely original, i.e. not found in either Mt or Mk.

-- Both tend to ignore plenty of material in Lk that appears to be much earlier vis-a-vis both Mt and Mk.

This early Lukan material is the key to understanding the early history of the gospels IMO. In fact, out of the 3 Synoptics, Luke seems to have more early material than any of the others!

I think it's about time that I gathered together in one article some of the evidence that I have for the Lukan priority. There's really quite a lot. So I will post this here pretty soon, and will also put it up on my webpage.

Cheers,

Yuri.

[GentDave]

Yuri: Sure, it was similar to Mk in so far as it was probably very short, and lacked much in the way of sayings materials.

Dave: By traditional means, I would also say it was is substantially the same order as Mark, and that most parts of it were substantailly Mark's wording.

By the study, the document Matthew and Luke used had a vocabulary profile similar to Mark.

So, someting at least pretty Mark-ish, was the source for Matthew and Luke.

Yuri: Both FH and 3SH fail to account for an awful lot of material in Lk that is completely original, i.e. not found in either Mt or Mk.

Dave: How So? Luke worte it himself, or used sources, not used by Matthew and Mark.

Yuri: -- Both tend to ignore plenty of material in Lk that appears to be much earlier vis-a-vis both Mt and Mk.
This early Lukan material is the key to understanding the early history of the gospels IMO. In fact, out of the 3 Synoptics, Luke seems to have more early material than any of the others!

Dave: Are you saying Luke had early material, not used by Matthew and Mark? If so, that does not contrdict my results at all. I'd only say that he re-wrote it into his own style, if he did that.

Or are you saying that Mark and Luke have a common early source that they both used? If so, I'd say that common early document had a vocabulary much more like Mark than Luke.

In fact, the only significant differances, may have been that our Mark is expanded here and there with commentary and details, over the original Mark.

Overall the study drew me more away from a completely independent early gospel, and more toward something very Mark-like as the first.

I'd be interested in whatever details you have about early Luke.

[Peter Kirby]

I didn't read your whole site, GentDave, but perhaps you can answer here: what evidence indicates that Luke used Matthew rather than Matthew used Luke? Evidence from your study, I mean, unless you want to bring in other considerations.

best,
Peter Kirby

[GentDave]

Here are the pages that talk about Luke using Mattew.

http://www.davegentile.com/synoptics/interpretation.html
http://www.davegentile.com/synoptics/sayings.html
http://www.davegentile.com/synoptics/minoragreements.html

In a nut shell, there is a category 212, made up of the minor agreements. It turns up as related to categories like 211, and 210, that are undisputedly Matthew. But 212 is not related to categories that Luke must have written.

Also, category 202 is the agreements in "Q" between Matthew and Luke. These turn up related to 201 (the part of Q in Matthew only) and 200 (Sondergut Matthew), but 202 is not significantly related to the Luke parts of Q (102), or related at all to sondergut Luke. (002)

So, both categories, that a priori might have been Luke or Matthew have a vocabulary style related to Matthew, and not to Luke.

[Yuri Kuchinsky]

Originally posted by GentDave
Here are the pages that talk about Luke using Mattew.

http://www.davegentile.com/synoptics/interpretation.html
http://www.davegentile.com/synoptics/sayings.html
http://www.davegentile.com/synoptics/minoragreements.html

In a nut shell, there is a category 212, made up of the minor agreements. It turns up as related to categories like 211, and 210, that are undisputedly Matthew. But 212 is not related to categories that Luke must have written.

Also, category 202 is the agreements in "Q" between Matthew and Luke. These turn up related to 201 (the part of Q in Matthew only) and 200 (Sondergut Matthew), but 202 is not significantly related to the Luke parts of Q (102), or related at all to sondergut Luke. (002)

So, both categories, that a priori might have been Luke or Matthew have a vocabulary style related to Matthew, and not to Luke.

Dave,

All this material that you're presenting is the product of multiple interpretative decisions that had previously been taken by various parties. Most of this can be described as secondary interpretation, and lots of things could have gone wrong while all those prior interpretative decisions were being made.

For example, the concepts such as "the style of Lk" or "the style of Mt" may already be problematic. In general, I see some parts of Mt as late, while some are early. So since you've mixed all these in your analysis of Matthean style, then your overall findings may not really reflect the true picture of various Matthean styles as used in different parts of Mt.

Likewise, there are probably some early Lukan stylistic elements, and some late ones, as well.

Also, Koester has carefully distinguished between some late Markan stylistic elements that would be different from his early ones. (Vinnie has a summary of this on his webpage.)

So, basically, in order to have a meaningful analysis of all these issues, we really need to come down to the level of primary data. So how about you present here some specific passages in the triple tradition where _you say_ that Lk copied some stuff from Mk and/or Mt (rather than the other way around)?

After we examine some such test cases, perhaps your general methodology can be considerably clarified.

Meanwhile, soon I will present some specific material in Lk that I see as obviously earlier than what we find in Mk and Mt.

Regards,

Yuri.

[GentDave]

The study itself does not go down to the level of specific passages. You know that. There are certainly things to be learned by doing that, but that's not what the study is about. It's about the overall vocabulary profile.

I'm not sure what you mean by previous interpretations. Do you mean the placement of the words into different categories? If so, then there is a small dergee of subjectiveness. I discuss that the authors of the HBB assume the 2SH, so some things in the "Mark/Q overlap" are arguably misclassified. But, for the most part the classification is non-subjective. Up to the point of producing the probabilities there is very little that is subjective or based on assuptions. How to interpret the results is a little more subjective, however.

In any case, I certainly do not want to claim that there are not more details "below the surface". There could be 20 documents, all full of slight modifications of the last, for all I know. What I do know, is that the big picture, at the level of resolution the study is capible of, looks like the 3SH.

As an example, I suppose category (211) which only appears in Matthew could be 2/3 Matthew, 1/3 "someone else", and the minor ageements (212) could be Mathew, and they would still show up as related, and we would totally miss the "someone else" involved.

All it really says is that for the majority of text Mark came before Matthew, and Luke. Also, that Matthew came before Luke. And finally that there is a noticably distinct sayings category.

If you have specific questions about the study, I'd be glad to try to answer them, and I'd be interested in anything you post about Luke.

[Yuri Kuchinsky]

Originally posted by GentDave
The study itself does not go down to the level of specific passages. You know that. There are certainly things to be learned by doing that, but that's not what the study is about. It's about the overall vocabulary profile.


Yes, Dave, but without some specific examples, the idea that the 3SH is valid will not find any ready acceptance, I'm afraid...


I'm not sure what you mean by previous interpretations. Do you mean the placement of the words into different categories? If so, then there is a small dergee of subjectiveness. I discuss that the authors of the HBB assume the 2SH, so some things in the "Mark/Q overlap" are arguably misclassified. But, for the most part the classification is non-subjective. Up to the point of producing the probabilities there is very little that is subjective or based on assuptions. How to interpret the results is a little more subjective, however.


Well, there you go. Even the interpretation of your overall results still remains somewhat subjective...


In any case, I certainly do not want to claim that there are not more details "below the surface". There could be 20 documents, all full of slight modifications of the last, for all I know. What I do know, is that the big picture, at the level of resolution the study is capible of, looks like the 3SH.


But I wish we knew with more certainty what is this Big Picture of yours really the Big Picture of?

Yes, it's a Big Picture of something, but what is it that we are really measuring? That's the central problem here as I see it.


As an example, I suppose category (211) which only appears in Matthew could be 2/3 Matthew, 1/3 "someone else", and the minor ageements (212) could be Mathew, and they would still show up as related, and we would totally miss the "someone else" involved.

All it really says is that for the majority of text Mark came before Matthew, and Luke. Also, that Matthew came before Luke. And finally that there is a noticably distinct sayings category.

If you have specific questions about the study, I'd be glad to try to answer them, and I'd be interested in anything you post about Luke.

I'll be posting the stuff about Lukan priority pretty soon. I can give many specific examples where Lk is clearly the earliest. For example, the Synoptic Anointing scenes. IMHO it's completely impossible that the Lukan Anointing scene was based on the Markan (and/or Matthean) Anointing scene. The reverse was far more likely, and lots of big name scholars happen to agree with me here...

As to your study, basically, I don't think that the application of your type of statistical analysis was really appropriate in this case.

Keep in mind that you started out to compare the 2SH with the other 2 theories. But we already knew from the outset that, based on the Anti-Markan Agreements alone, the chances of 2SH being valid were about 1 in 1,000,000.

OK, suppose you've _totally_ proved that the FH and the 3SH are both somewhat more likely than the currently mainstream 2SH. So how would we put this in the numerical terms? Let's see...

The 2SH chances of being right = 1 in 1,000,000.
The FH chances of being right = 1 in 900,000 (?)
The 3SH chances of being right = 1 in 800,000 (?)

Is this correct?

So are these the sorts of the results that people should see as encouraging for the FH and the 3SH?

Regards,

Yuri.

[Bernard Muller]

Dave wrote:
"I tend to see Luke as liking some of the corrections to Mark's Greek that Matthew made, but not liking a lot of the content changes. Luke did not see Matthew as authoritative, in my view.

I'm familiar with the nature of the MAs, and many could be explained the way the 2SH suggests. However, the study suggests the MAs (cat. 212) are Matthian in style much more than Lukian, suggesting Luke borrowing from Matthew.

I also think the agreements of omission against Mark are too large to explain by completely independant action. I think there must have been a proto-Mark that they both used, or Luke knew Matthew's omissions and tended to agree."

Dave, let me engage in pure unsubstantiated theories right now:
Let's say GMark was introduced early on in Matthew's community.
Then, one of the principals, looking at GMark, noticed the obvious: poor Greek, very undignified for a document which otherwise had values.
Solution: rewriting in part in better Greek (with no major changes or additions)
"Matthew" used that version for his gospel, and somehow, a copy of it (that is 'rewritten GMark' or post-GMark) made it to GLuke community.
One more step: a younger "Matthew" might have been the rewriter, or at least, somebody from the same milieu, sharing the same vocabulary & jargon, that is from a same group of educated Jewish Christians of the same city.
As I contended, a lot of Q was written between GMark & GMatthew, and likely in the same community. If it the case, the Q authors would have access to the post-GMark, explaining the similarities.

Food for thoughts. I do not have a clue how these theories can be proven or disproven, but it seems to me your study would allow them.
How to explain the disappearance of the post-GMark?
Well, it became known this post-GMark was not the authentic original one and therefore did not get recopied, ensuring its end.

I contacted Dave by E-Mail about the resurrection of this thread. So I hope he will share his thoughts on that.

Best regards, Bernard

[GentDave]

Yuri writes: Keep in mind that you started out to compare the 2SH with the other 2 theories.

Dave: This kind of statement leads me to believe you don't understand what I did. I didn't set out to compare any set of hypotheses. After the results were in, I evaluated what popular theories were consistant with the results and which ones were not consistant with the results.

Reduced to its basics the study says 4 things:
1) For the majority of text that Mark and Matthew have in common Mark is more original.
2) For the majority of the text Mark and Luke have in common Mark is more original
3) For the majority of the text that Matthew and Luke have in common outside of Mark, Matthew is more original.
4) There are "sayings" that are somewhat distinct in their vocabulary.

Note that I said the *majority* of the text, not *all* the text.

There are an infinite number of hypotheses consistant with that, if we allow unlimited complexity. The simplest theory that is consistant with the study is the 3SH, but others that are more complex are also consistant.

The study does eliminate some ideas, like the idea that Mark was a composit based on a finished Matthew and Luke, but it does not eliminate many other ideas.

I'm basicly sympathetic to the idea that Luke has some early source, but the study adds no support for that. However, one could hypothesis that there was an early source that Mark tended to copy word for word, and Luke tended to re-write in his own vocabualry. This would be fully consistant with the study. Another possibility would be that in most cases Mark was more original, but in a few cases Luke was. There are many possibilities. Again, any theory that fits the 4 poins above, would fit wih the results.


Yuri also writes: As to your study, basically, I don't think that the application of your type of statistical analysis was really appropriate in this case.

Dave: These are very general complaints you are making. "There are assumptions". "the statistical analysis was not appropriate here". But, unless you have something specific, like an assumption that was incorrect, or something that invalidates the limited conclutions that I claim, then I don't find this very productive.

[GentDave]

Bernard,

That idea is generaly known as deutro-Mark.
see:
http://www.mindspring.com/~scarlson/synopt/
under dMk

My study would tend to argue against that. If that were true we'd expect the minor agreements to be unrelated to Matthew's style or Luke's style, but they do show up as related to Matthew.

[Bernard Muller]

Reduced to its basics the study says 4 things:
1) For the majority of text that Mark and Matthew have in common Mark is more original.
2) For the majority of the text Mark and Luke have in common Mark is more original
3) For the majority of the text that Matthew and Luke have in common outside of Mark, Matthew is more original.
4) There are "sayings" that are somewhat distinct in their vocabulary.
Dave

So where the 700 bits where GLuke & GMatthew agree against GMark would fit in? In 3), that is what is considered Q? Where else?

Does Q, if postulated a document, and if compiled/written between GMark & GMatthew, and if put together in the same milieu, in the same city, from people sharing the same vocabulary/syntax as GMatthew, would show similarities with GMatthew, of the same nature you identified in your study?

OK, I see the deutero-Mark is shot down. I did not like the idea but I was trying to figure out from where the 700 items might come.

Best regards, Bernard

[GentDave]

Bernard,

The study suggests that what is commonly called "Q" (202) has a vocabulary profile related to sondergut Matthew (200), (or the parts of Matthew found only in Matthew.) It also suggests the "minor agreements" (212) are related to other words in the triple tradition found only in Matthew. (211)

In other words, both the "minor agreements", and "Q" look at least somewhat Matthian in style.

For clearification, "Q" is large sections of text found in Matthew and Luke, but not in Mark. The minor agreements are in sections of text that are in all three gospels, and where the exact wording of Luke and Matthew agree against Mark.

===

Could one author have a style that could not be seperated from another? I suppose they could. The vocabulary of Mark, Matthew, and Luke as they stand is distinct enough that the study can easily tell them apart, however, any one of them could contain work by more than one author. The "style of Matthew" could potentially just be the style that results from two differant authors in colaboration, for example.

I doubt that "author-sayings" and "author-Matthew" would share the same vocabulary, however. And even if the same person did write them, the differance in genera might set them apart. That's why the study can't firmly say there was a saying source. It might show up as distinct, even if written by "Matthew".

[Bernard Muller]

My study would tend to argue against that. If that were true we'd expect the minor agreements to be unrelated to Matthew's style or Luke's style, but they do show up as related to Matthew.
Dave

Another rumination of mine. I did look at your pages, believe me, even if nothing is clear in my mind.

Let's say, in Matthew's community, someone wrote the deutero-GMark. Then a large amount of Q material was generated in the same community, knowing about deutero-GMark, and therefore using some of the same syntax (possibly to have the sayings seemingly coming from the same author), and that got compiled within the hypothetic Q document. Matthew, then, for his own material, stayed close to the syntax of Dtr-GMark and Q, for the aforementioned reasons. In other words, Matthew tried to pretend his gospel was a second expanded/corrected/better edition of GMark (rather dtr-GMark), suddenly rediscovered.
Luke, working from dtr-GMark and Q, reproduced the MA's, but for his/her own material, broke loose, without conforming to the syntax in dtr-GMark & Q.

Would that be acceptable according to your findings, as I think it would (if I got something right!)?

Best regards, Bernard

[Mathetes]

Maybe this belongs to another thread, but...

In the 3SH, I understand that Luke uses Mark, Q and Matthew as sources. In the parts that are common to GMk and GMt, Luke is closer to GMt. It is hard to believe, then, that he was using the cannonical Matthew as we know it.

If this were true, the Luke redactor had a strange way of dealing with the Matthew source. He had to do the following:

- Throw away Matthew's genealogy, and invent a competing one.
- Throw away the Magi/Herod/Innocents stories, and put in his own almost completely independent infancy stories.
- When it comes to Jesus ministry, however, he copies Matthew more than he copies Mark.
- After the resurrection, he throws away Matthew's accounts and writes his own, in Jerusalem instead of Galilee.

This "selective copying" does not make much sense to me. Maybe the version of Matthew that he had did not contain the beginning and the end?

[Bernard Muller]

If this were true, the Luke redactor had a strange way of dealing with the Matthew source.
Mathetes

Yes, the 3SH does not make sense, because:
If "Luke" knew GMatthew, why GLuke does not incorporate more Matthean material. There are no reason for "Luke" to reject a lot of Matthean stuff, some of it actually would be very suited for "Luke" (such as Mt25:35-45) and many corrections "Matthew" made on GMark would have helped "Luke".
On the Q side, "Luke" incorporated sayings which hurt the Gentile cause of the author. One example is
Lk16:17 "But it is easier that the heaven and the earth should pass away than that one tittle of the law should fail." (parallel Mt5:18)
So why did "Luke" not use favorable Matthean material when he/she used unfavorable Q material?
The argument is valid also if Q is believed to have originated from within GMatthew (FH).
Among the conflicts between "Luke" and GMatthew, the circumstances leading to the death, and the description of the death itself of Judas could have been easily harmonized by "Luke" if he/she had known about GMatthew.

Best regards, Bernard

[GentDave]

Bernard,

On the assumption that being produced in the same community can lead to the same style, your solution would probably work.

I'm disinclined to make that assumption, but I can't rule it out. Keep in mind that most of the words we are dealing with are very common ones like "has", "was", etc. I'm not sure why two authors in the same community would have a tendency to use those words with similar frequency, but its not impossible.

Regarding your latest post:
Why did Luke not use more of Matthew?

What if Luke did not believe that Matthew was, in general, historically accurate, but Luke believed Mark and Q were?
Luke then only used Matthew's Greek, here and there, and used a very few select pericope from Matthew. (What the 2SH calls the Mark/Q overlap)

[GentDave]

Mathetes,

Actually, on the 3SH, in the triple tradition, Luke is almost always using Mark almost exclusively, just as on the 2SH. He may occationally borrow some wording from Matthew, but mostly he used Mark's words, or his own words. Luke selects a very few select bits from Matthew to use. (The Mark/Q overlap and some of Matthew's "Q").

Basicly, the 3SH claims that Luke knew Matthew, but made extreamly limited use of Matthew.

[Bernard Muller]

A) Let's say, in Matthew's community, someone wrote the deutero-GMark. Then a large amount of Q material was generated in the same community, knowing about deutero-GMark, and therefore using some of the same syntax (possibly to have the sayings seemingly coming from the same author), and that got compiled within the hypothetic Q document. Matthew, then, for his own material, stayed close to the syntax of Dtr-GMark and Q, for the aforementioned reasons. In other words, Matthew tried to pretend his gospel was a second expanded/corrected/better edition of GMark (rather dtr-GMark), suddenly rediscovered.
Bernard
:banghead:

Dave, I would like a straight answer from you.
I am not asking you to approve the above theory, far from that.
My question is: can this hypothesis fit confortably within your findings? If not, what would be the problem(s)?

B) I'm disinclined to make that assumption, but I can't rule it out. Keep in mind that most of the words we are dealing with are very common ones like "has", "was", etc. I'm not sure why two authors in the same community would have a tendency to use those words with similar frequency, but its not impossible.
DAVE
:banghead:

Are you talking about my own quote above? I wonder.
Here, I made a point that those authors (some of the Q ones and "Matthew") would intentionally used some of the verbose as the one into a hypothetical deutero-GMark. Which "Luke" would follow, at least for Markan & Q material, but adopt his/her own style for the Lukan one.

C) What if Luke did not believe that Matthew was, in general, historically accurate, but Luke believed Mark and Q were?
DAVE

I have a big problem with that, as you guessed already.
I cannot understand: if "Luke" did not like GMatthew, if "Luke" thought GMatthew was not accurate, why would "Luke" pick up some 700 MA's from GMatthew? That does not make sense.
It makes more sense that "Luke" picked up that from a DTR-GMark, and parts of Q which followed the syntax of DTR-GMark. And "Matthew", for whatever reasons, decided to write in the way of DTR-GMark and Q.
That would explain most of the MA's, and why GLuke is close to GMatthew (Matthean stuff included) on Q and Markan material, with the MA's.
Would that scenario explain what you found in your studies?
Again, I do not ask you to endorse the aforementioned hypothesis.

D) Luke then only used Matthew's Greek, here and there, and used a very few select pericope from Matthew. (What the 2SH calls the Mark/Q overlap)
DAVE
Does your study allow the overlaps to be part of Q, and generated after GMark, or better, after DTR-GMark?

Best regards, Bernard

[GentDave]

Bernard,

A & B) You asked if the hypothesis fits with my results. The answer is a qualified yes, given what I consisder a somewhat unlikely but cerainly possible assumption. The assumption is that a number of authors from the same community would produce documents with the same vocabulary style. That's probably the best answer I can give.

====

C) On the next point, I may be wrong, but I think you have an incorrect idea regarding the relationship between the three gosples in the triple tradition. In the triple tradition material, Mark is the middle term. That is there are large numbers of agreement between Matthew and Mark, and large numbers of agreements between Mark and Luke. There are also places where all three texts are differant, and places where they are argee. But there are relatively few agreements between Matthew and Luke against Mark. In short, Luke is unquestionably closer to Mark's text than to Matthew's text.

Almost none of the MAs that come to mind involve any significant factual information. The greatest bulk of them are KAI => DE that is Mark used the word KAI, and Luke and Matthew tend to use DE, both of which mean "and".

Mark wrote in very very basic Greek. Luke I believe has something like 3 times the vocabualry of Mark. Suppose Luke is setting out to do two things, get the facts as straight as possible, and write an elegant document. It is quite possible that Luke would ignore Matthew as far as most facts go, but pick up a turn of phrase here and there. Mark and Q may have been around for 30 years, at the time Luke is writting. Matthew may be contemporay with Luke, and may give Luke the inspiration to attempt what he does.

D) The overlaps are found in Mark, so they had to be in existance at the time of Mark. Maybe I misunderstood that question.

The interesting fact about the overlap section is that they are found in all 3 gospels, but whereas normally Mark is the middle term, here Matthew is the middle term. Mark-Luke agreements against Matthew are rare here. On the 3SH these represent the few instances where Luke actually choose to use Matthew's idea's and text, instead of Mark's.

So according to the 3SH in most of the triple tradition the pattern is:
Mark => Matthew
Mark => Luke (with an occational turn of phrase picked up from Matthew)
but in the overlap the pattern is
Mark => Matthew => Luke

I would also speculate that in the Q sections the pattern is mostly:
Q => Matthew
Q=> Luke
but occationally:
Q => Matthew => Luke

[GentDave]

Bernard,

Let me suggest another idea, a proto-Matthew.

Suppose we have Mark and Q.
Proto-Matthew combines them but the saying are still mostly in an ill arrainged block.

Luke then uses Mark, proto-Matthew, and perhaps the original Q.

Matthew later comes along and produces a new version of his own document, giveing us Matthew.

So:

Mark + Q => p-Mat
Mark + Q? + p-Mat => Luke
p-Mat => Matthew

That fits completely with the study.

[Yuri Kuchinsky]

Originally posted by GentDave
Yuri writes: Keep in mind that you started out to compare the 2SH with the other 2 theories.

Dave: This kind of statement leads me to believe you don't understand what I did. I didn't set out to compare any set of hypotheses. After the results were in, I evaluated what popular theories were consistant with the results and which ones were not consistant with the results.


OK, Dave, so let me rephrase it then. You _ended up_ your study by comparing the 2SH with the other 2 theories. But in reality -- i.e. in the real world -- 2SH had already been excluded even before you began your study. Because the 1000 Anti-Markan Agreements exclude 2SH from consideration under normal conditions, where empirical evidence and logic are respected.

What I'm saying is that, assuming 2SH, your category 212 (Anti-Markan Agreements between Mt and Lk) shouldn't be there at all. Just by being there, 212 constitutes a death blow to the 2SH.

So what do you have left then? You have FH and 3SH, neither of which has any realistic hope of success, since they are burdened by too many shortcomings of their own.

So what's the value of your study then?


Reduced to its basics the study says 4 things:
1) For the majority of text that Mark and Matthew have in common Mark is more original.


Might be...


2) For the majority of the text Mark and Luke have in common Mark is more original


So what about those 1000 Anti-Markan Agreements (the Lukan side of them)? How can Mk be more original than Lk in these 1000 triple tradition passages? This sounds highly unlikely to me.


3) For the majority of the text that Matthew and Luke have in common outside of Mark, Matthew is more original.


But that's in contradiction with the whole Q scholarship, dozens of detailed studies. Maybe it's _your_ analysis that's in the wrong here, in such a case?


Note that I said the *majority* of the text, not *all* the text.


And the Q scholars have demonstrated in many detailed studies that, for the *majority* of the text that Matthew and Luke have in common outside of Mark, Luke is more original.

Regards,

Yuri.

[Yuri Kuchinsky]

Originally posted by Bernard Muller

I cannot understand: if "Luke" did not like GMatthew, if "Luke" thought GMatthew was not accurate, why would "Luke" pick up some 700 MA's from GMatthew? That does not make sense.


That's right, Bernard. This is just one of those things that make no sense at all, and cast plenty of doubt on the FH/3SH.

OTOH, there emerges a very strong probability that these Anti-Markan Agreements of Mt and Lk go back to some sort of a proto-gospel, that preceded all 3 canonical Synoptics.

Regards,

Yuri.

[GentDave]

Yuri writes:

"And the Q scholars have demonstrated in many detailed studies that, for the *majority* of the text that Matthew and Luke have in common outside of Mark, Luke is more original."

Dave:

I don't think that is the case. Luke's order is generally supposed to be more original, but not necessarily his wording.

Now, this is where we get to various possible interpretations of the results of the study, but what I believe is happening there is that sondergut Matthew contains parts of Q that Luke did not use, so that sondergut Matthew is really a composite of the styles of Q and Matthew.

[GentDave]

Yuri writes:

So what do you have left then? You have FH and 3SH, neither of which has any realistic hope of success, since they are burdened by too many shortcomings of their own.

So what's the value of your study then?

Dave:

I'd like to try to sepearte this in to two differant issues.
1) The study
2) the 3SH
As I've tryed to indicate, they are not one and the same. What the study does is provide very objective mathematical evidence eliminating certain things that are possible a priori. For example
it provides strong objective evidence against the Griesbach Hypothesis (2GH). The 3SH is only the simplest hypothesis that it can say nothing against.

Now, seperately, I suppose we could discuss the 3SH. While I'm perfectly willing to admit it is probably more complex than that, I think you would have a very difficult time showing to any high degee of certainty that it does not represent what happened.

[GentDave]

--------------------------------------------------------------------------------

2) For the majority of the text Mark and Luke have in common Mark is more original

--------------------------------------------------------------------------------

Yuri writes:

So what about those 1000 Anti-Markan Agreements (the Lukan side of them)? How can Mk be more original than Lk in these 1000 triple tradition passages? This sounds highly unlikely to me.

Dave:

I don't think I follow you here. The MAs represent a very small portion of the total text. But there is a large amount of text where Luke and Mark are in rough agreement. The question is, in these sections, "Which is the original, and who made the changes?" The study says that at least most of the time Mark is the original, and Luke made the changes. It does not say that *all* the time Mark is the original.

In other words the focus here is on agreements between Luke and Mark. And the question is, "Are those areas of agreement more like text found only in Mark, or are they like text found only in Luke?" The study answers that they are more like Mark.

The MAs are agreements between Matthew and Luke, and there the study asks, "Do these agreements look more like text found only in Luke, or text found only in Matthew?", and the answer it gives is "Matthew".

[Bernard Muller]

DAVE wrote:
Let me suggest another idea, a proto-Matthew.
Suppose we have Mark and Q.
Proto-Matthew combines them but the saying are still mostly in an ill arranged block.
Luke then uses Mark, proto-Matthew, and perhaps the original Q.
Matthew later comes along and produces a new version of his own document, giving us Matthew.
So:
Mark + Q => p-Mat
Mark + Q? + p-Mat => Luke
p-Mat => Matthew
That fits completely with the study.

Yes Dave, I can relate to that:
First, according to your own definition, this Proto-Matt can as well be, word by word, mt-Mark plus mt-Q. By 'mt', I mean as rewritten by Matthew. Same text, and I do not care if mt-Mark & mt-Q would show on the same scroll (as contiguous) or two.
I am not changing your Proto-Matt, just emphasizing the two parts of it.
So now we would have:
Mark + Q => (mt-Mark + mt-Q)
Mark + Q? + (mt-Mark + mt-Q) => Luke
(mt-Mark + mt-Q) => Matthew

I noticed you are not certain that Luke had (original) Q.
Trying to see what you are thinking, I gather that in some cases you would prefer Luke working on original Q, but in most cases, Luke would redact from mt-Q. Of course, that would be according to your study.
May I suggest that Luke may have only one Q document; this Q document, in some parts, not showing Matthean tendency, but in most parts, showing this Matthean coloring. That would explain:
Mark + Q? + (mt-Mark + mt-Q) => Luke
OR
Mark + (Q?/mt-Q) + mt-Mark => Luke

Then why Q would look more Matthean in GMatthew than in GLuke?
The obvious solution would be that Matthew rewrote parts of (Q?/mt-Q), making it look more like fully mt-Q.
So I do think Matthew & Luke worked from the same version of Q, and only one version for each.

How to explain the Matthean coloring in (Q?/mt-Q)?
My hypothesis is many parts of Q, with the Markan overlaps were written in the same community, by the same "school" of people to which Matthew was part, and after GMark was known here. As a matter of fact, younger Matthew himself could have issued some of the Q material then.

But that would not explain everything and it is too hypothetical.
What about a reverse process: Matthew was fond of expressions and words from Q, used those a lot in his Matthean material (and sometimes Mat-Q). Consequently, Q would show strong affinity with sondergut Matthew, because Luke did not do that and therefore Lukan material looks more distant of Q.
Here are examples I found from one of my page:
"gnashing of teeth": Mk = 0, Mt = 6, Lk (Q) = 1, Jn = 0
"outer darkness": Mk = 0, Mt = 3, Lk (Q) = 1, Jn = 0
"wailing": Mk = 0, Mt = 3, Lk (Q)= 1, Jn = 0
"weeping": Mk = 0, Mt = 5, Lk (Q)= 2, Jn = 0
In that page, I explain why Matthew used that often, because the treatment of "undesirables" was dear to him and a major theme.

As you can see (probably know already), Matthew used with gusto some expressions appearing in Q, but Luke abstained from it.
Furthermore, I recall that Matthew faithfully recopied, for a doublet (that is a Markan saying which has a Q couterpart, redacted differently) both occurences of it (GMark's version & Q's). But the GMark version is replaced by the Q version (which Luke does not do). That happens several times. Essentially Matthew "standarized" on Q. (I can document that if you request it). I think Mark Goodacre might have been referring something similar to you, to which you partially agreed.

All that combined would tip your result into having Q looking Matthean.

Relevant pages from my site:
Q (http://www.concentric.net/~Mullerb/q.shtml)
Matthew's coloring (http://www.concentric.net/~Mullerb/appdx.shtml)

Best regards, Bernard

[GentDave]

Bernard writes:

What about a reverse process: Matthew was fond of expressions and words from Q, used those a lot in his Matthean material (and sometimes Mat-Q). Consequently, Q would show strong affinity with sondergut Matthew, because Luke did not do that and therefore Lukan material looks more distant of Q.
Here are examples I found from one of my page:
"gnashing of teeth": Mk = 0, Mt = 6, Lk (Q) = 1, Jn = 0
"outer darkness": Mk = 0, Mt = 3, Lk (Q) = 1, Jn = 0
"wailing": Mk = 0, Mt = 3, Lk (Q)= 1, Jn = 0
"weeping": Mk = 0, Mt = 5, Lk (Q)= 2, Jn = 0
In that page, I explain why Matthew used that often, because the treatment of "undesirables" was dear to him and a major theme.

Dave:

Yes I agree. That is how the 2SH explains those. And Mark Goodacre has pointed out that that implies the style of Q and Matthew are somewhat similar.

Another way of putting it is that sondergut Matthew contains elements of Q that Luke did not use. (In this case just extra copies of things Luke did use)

That could be all that is going on. That's why I say that the study does not completely eliminate the 2SH. It's very suspicious that both the MAs and Q look Matthian, but its not absolutely conclusive that Luke used Matthew.

[Bernard Muller]

1)
DAVE wrote:
On the next point, I may be wrong, but I think you have an incorrect idea regarding the relationship between the three gospels in the triple tradition. In the triple tradition material, Mark is the middle term. That is there are large numbers of agreement between Matthew and Mark, and large numbers of agreements between Mark and Luke. There are also places where all three texts are differant, and places where they are argee. But there are relatively few agreements between Matthew and Luke against Mark. In short, Luke is unquestionably closer to Mark's text than to Matthew's text.
:banghead:

But Dave, that would shot down the idea of a Proto-Matthew (or a Deutero-Mark & Deutero-Q, both with Matthean overtone). I thought that was what you meant in:
"Mark + Q => p-Mat
Mark + Q? + p-Mat => Luke
p-Mat => Matthew
That fits completely with the study."
Can you clarify?

2)
DAVE wrote:
Almost none of the MAs that come to mind involve any significant factual information. The greatest bulk of them are KAI => DE that is Mark used the word KAI, and Luke and Matthew tend to use DE, both of which mean "and".

QuestionS:
a) Here (triple tradition material), how many MA's are we are talking about, grand total.
b) Here again, how many MA's of the type KAI => DE, are we talking about?
c) Is there any reasons why Greek writers, in late 1st cent., would prefer DE to KAI?
d) Do we know what other writers in these days were using preferably?
e) Here again, do you think there is enough MA's here (kind & quantity) to make a case that Luke had either a Proto-Matthew, a "Matthean" Deutero-Mark (as Matthew) or even GMatthew?

3)
DAVE wrote:
Mark wrote in very very basic Greek. Luke I believe has something like 3 times the vocabualry of Mark. Suppose Luke is setting out to do two things, get the facts as straight as possible, and write an elegant document. It is quite possible that Luke would ignore Matthew as far as most facts go, but pick up a turn of phrase here and there. Mark and Q may have been around for 30 years, at the time Luke is writting. Matthew may be contemporay with Luke, and may give Luke the inspiration to attempt what he does.
:mad:
Why Luke, very fluent in Greek, would look at GMatthew for his/her redaction, and hardly nothing else? I cannot understand that at all.
And I would not assume that most of Q had been around as long as you think. That might affect your thinking and your interpretation.
Inspiration? To change KAI to DE? Let's be reasonable.

4)
DAVE wrote:
The overlaps are found in Mark, so they had to be in existance at the time of Mark. Maybe I misunderstood that question.

:confused:
OK, my apology. The Markan overlaps are pieces of rewritten Markan material either appearing on their own in both GMatthew and GLuke, or as insertion in "Q" blocks such as JB's baptism or Jesus & Beelzebub. That's my definition & hopefully the right one, but I can be corrected.
But wait a minute: Are we talking about the pieces as they appear in GMark, or as they appear in Q? Because the redaction is quite different.

5)
DAVE wrote:
The interesting fact about the overlap section is that they are found in all 3 gospels, but whereas normally Mark is the middle term, here Matthew is the middle term. Mark-Luke agreements against Matthew are rare here. On the 3SH these represent the few instances where Luke actually choose to use Matthew's idea's and text, instead of Mark's.

:)
Of course, if the overlaps were fabricated from the parallel in GMark, and put in Q, and from there in GMatthew & GLuke, then we would have agreement between GMatthew & GLuke against GMark!!!
And with the Q overlaps being from the same milieu as Matthew, there is a better chance these overlaps to be closer of GMatthew than GLuke.
Why bring 3SH for that? With all the problems it involves!
Because my front door is partially open, that's no reason to drive a truck through it.
Because I need sugar, that's no reason to bring the whole grocery store in my kitchen.
We have to be pragmatic, before going into absurdities.
How many overlaps do we have? Maybe twenty?
Dave, do you have a number for those?
Can you give a list?

6)
DAVE wrote:
So according to the 3SH in most of the triple tradition the pattern is:
Mark => Matthew
Mark => Luke (with an occasional turn of phrase picked up from Matthew)
but in the overlap the pattern is
Mark => Matthew => Luke


I would love to go over those "occasional turn of phrase picked up from Matthew"
Can you specify when & where & how they occur?
How many we are talking about?
I am anxious to know how you define 'overlap'.

Best regards, Bernard

[Bernard Muller]

Dave wrote:
Yes I agree. That is how the 2SH explains those. And Mark Goodacre has pointed out that implies the style of Q and Matthew are somewhat similar.)

OK. When I see Matthew multiplying in his own material expressions he found in Q, I have reason to believe the style of Matthew comes from Q, rather than the other way around.

Dave wrote:
Another way of putting it is that sondergut Matthew contains elements of Q that Luke did not use. (In this case just extra copies of things Luke did use). That could be all that is going on.
This is not another way of putting it!
This is an endorsement of 3SH, which you do by assuming sondurgut Matthew contains part of Q not used by Luke. Where is the evidence for that?

Dave wrote:
That's why I say that the study does not completely eliminate the 2SH. It's very suspicious that both the MAs and Q look Matthian, but its not absolutely conclusive that Luke used Matthew.

"does not completely eliminate the 2SH".
Nothing I saw so far would do that:
We know now why Q would look Matthean, don't we?
And many of the MA's apparently comes from the Mark-Q overlaps. But if those overlaps were fabricated/modified from GMark, then of course they would generate all kinds of MA's!

And what about the KAI and DE?
YOUR quote:
"Almost none of the MAs that come to mind involve any significant factual information. The greatest bulk of them are KAI => DE that is Mark used the word KAI, and Luke and Matthew tend to use DE, both of which mean "and"."

I just made an approximate survey:
Mark: KAI/DE = 3.3
Matthew: KAI/DE = 1.65
Luke: KAI/DE = 1.75
So yes, Matthew & Luke used more DE than Mark as compared with KAI. Of course there are huge amounts of those, used in many verses.
For example:
For Matthew, KAI about 600 verses, DE about 400 verses
For Luke, KAI about 820 verses, DE about 500 verses.
The question is: Did Luke copy on GMatthew for the use of DE rather than KAI?
Looking at 'Acts', a Lukan work, we have KAI about 670 verses, DE about 500 verses, ratio KAI/DE = 1.35
So Luke, when free from dealing with reference texts, had a tendency to use more DE relative to KAI than GMatthew ever did. So maybe the great number of DE in GMatthew comes from Luke !!! ;)

Best regards, Bernard

[GentDave]

Wow, a lot of stuff here. I'll try to cover them in seperate posts.

Bernard:

1)

quote:
--------------------------------------------------------------------------------
DAVE wrote:
On the next point, I may be wrong, but I think you have an incorrect idea regarding the relationship between the three gospels in the triple tradition. In the triple tradition material, Mark is the middle term. That is there are large numbers of agreement between Matthew and Mark, and large numbers of agreements between Mark and Luke. There are also places where all three texts are differant, and places where they are argee. But there are relatively few agreements between Matthew and Luke against Mark. In short, Luke is unquestionably closer to Mark's text than to Matthew's text.
--------------------------------------------------------------------------------


But Dave, that would shot down the idea of a Proto-Matthew (or a Deutero-Mark & Deutero-Q, both with Matthean overtone). I thought that was what you meant in:
"Mark + Q => p-Mat
Mark + Q? + p-Mat => Luke
p-Mat => Matthew
That fits completely with the study."
Can you clarify?


Dave:

What I say in that quote is just an obeservation. That's not something there is any disagreement about, its something that any proposed solution needs to try to explain. In the triple tradition Mark is the middle term.

As far as how I see the 3SH, or the proto-Matthew idea working here:

In either case, I see Luke relying mostly on Mark for the triple tradition.

[GentDave]

Bernard:

QuestionS:
a) Here (triple tradition material), how many MA's are we are talking about, grand total.
b) Here again, how many MA's of the type KAI => DE, are we talking about?
c) Is there any reasons why Greek writers, in late 1st cent., would prefer DE to KAI?
d) Do we know what other writers in these days were using preferably?
e) Here again, do you think there is enough MA's here (kind & quantity) to make a case that Luke had either a Proto-Matthew, a "Matthean" Deutero-Mark (as Matthew) or even GMatthew?

Dave:

You seem to have found some counts of these. Let me just add some totals from the study, that might be of intrest. The total of all the words in all the categories in the study is: 25,843
Sondergut Luke: 5755
The number of words of triple agreement: 1493
The number of words in the MAs: 207
Note: This involves judgement, what is an MA and what is an overlap?
Matthew-Mark agreement against Luke: 992
Luke-Mark agreement against Matthew: 531

Just as an aside, my own impression of the solution, without relying on the study is that Luke had early versions of both Matthew, and Mark, and a saying source. Not something I'm prepared to rigorously defend, just where my head is.

[GentDave]

Bernard:

Why Luke, very fluent in Greek, would look at GMatthew for his/her redaction, and hardly nothing else? I cannot understand that at all.
And I would not assume that most of Q had been around as long as you think. That might affect your thinking and your interpretation.
Inspiration? To change KAI to DE? Let's be reasonable.

Dave:

I think Luke as he wrote each bit, would have probably looked at all of his sources, and then procede to put it in his own words.
That is, the procedure I see for each pericope is:
1) Review Mark, Matthew, Q, and any other source.
2) Write the section in Luke's words, using Mark as the outline.
A result that follows Mark with a tad of Matthew here and there does not seem unreasonable.

However, that's a judgement call. If it seems unreasonable to you, there is not much I can do about that.

What I meant by "inspiration", is that Luke may have been satisfied with Mark and Q, but then Matthew shows up. Luke thinks that the idea of producing one document using both Mark and Q, in more elegent language, is a good idea, but he just does not like the way Matthew has gone about it.

[GentDave]

Bernard:
OK, my apology. The Markan overlaps are pieces of rewritten Markan material either appearing on their own in both GMatthew and GLuke, or as insertion in "Q" blocks such as JB's baptism or Jesus & Beelzebub. That's my definition & hopefully the right one, but I can be corrected.
But wait a minute: Are we talking about the pieces as they appear in GMark, or as they appear in Q? Because the redaction is quite different.

Dave:

It sounds like we are talking about the same thing. The Beelzebub bit is the archtype for me. On the 2SH one would have to say Mark used Q, or Q used Mark, or they both used a common source.

On the 3SH or FH Luke is following Matthew here, who followed Mark.

Personally, this looks to me like the whole thing is more complicated. Using my proto-Mark and proto-Matthew idea. I'd say proto-Mark did not have it. A saying or two of it are in Q. It appears first in near its current form in proto-Matthew. Luke and current Mark are both dependent on that prto-Matthew version.

But that's mostly speculation.

[GentDave]

Bernard:
Of course, if the overlaps were fabricated from the parallel in GMark, and put in Q, and from there in GMatthew & GLuke, then we would have agreement between GMatthew & GLuke against GMark!!!
And with the Q overlaps being from the same milieu as Matthew, there is a better chance these overlaps to be closer of GMatthew than GLuke.
Why bring 3SH for that? With all the problems it involves!
Because my front door is partially open, that's no reason to drive a truck through it.
Because I need sugar, that's no reason to bring the whole grocery store in my kitchen.
We have to be pragmatic, before going into absurdities.
How many overlaps do we have? Maybe twenty?
Dave, do you have a number for those?
Can you give a list?

Dave:
Unfortunately the data in the HHB is not broken up into a "overlap" category. That is the biggest problem I have with the data. Questions I have:
Does Mark in the overlap look like Mark other places?
Does Q in the overlap look like Q other places?
How do we tell an overlap from an MA?

There are a limited number of these, so I don't think they have a large impact on everything else, but I really would have liked a seperate treatment.

So the study really can't comment on the "overlap" specificly, it can comment on the MA's.

[GentDave]

Bernard:

I would love to go over those "occasional turn of phrase picked up from Matthew"
Can you specify when & where & how they occur?
How many we are talking about?
I am anxious to know how you define 'overlap'.


Dave:

I don't have counts. But we could try an example, I suppose.
I just fliped to a random page.

Mark 2:3
And they came bringing to him a paralytic.
Matthew 9:2
And behold, they brought to him a paralytic.
Luke 5:18
And behold, men were bringing on a bed, a man who was paralyzed.

Mark:
KAI ERCONTAI FERONTES PROS AUTON PARALUTIKON
Matthew:
KAI IDOU PROSEFERON AUTW PARALUTIKON
Luke:
KAI IDOU ANDRES FERONTES EPI KLINHS ANQRWPON OS HN PARALELUMENOS

Luke picks up IDOU, behold, from Matthew.

=================

222 211 112 212 221 122 121

IDOU 2 18 6 5 2 0 0

=================

Category 211 has a count of 18. Matthew likes adding this word to Mark's text. Matthew never ommits it if Mark has it. (122=0, 121=0)

The MA's (212) have a count of 5. Luke picks up some of Matthew's additions of "behold". Of the 13 times Luke uses the word in the triple tradition, 5 of them are exactly where Matthew added the word.

[GentDave]

quote:
--------------------------------------------------------------------------------
Dave wrote:
Another way of putting it is that sondergut Matthew contains elements of Q that Luke did not use. (In this case just extra copies of things Luke did use). That could be all that is going on.
--------------------------------------------------------------------------------

Bernard:

This is not another way of putting it!
This is an endorsement of 3SH, which you do by assuming sondurgut Matthew contains part of Q not used by Luke. Where is the evidence for that?

Dave:

category 200 (sondergut Matthew) and 202 (Q Mt-Lk agreements) are strongly related. On the simple 2SH, they should not be. However, if 200 (sondergut Matthew) contains parts of Q that Luke skiped, then the 2SH can explain the relation between 200 and 202.

[GentDave]

Bernard:

"does not completely eliminate the 2SH".
Nothing I saw so far would do that:
We know now why Q would look Matthean, don't we?
And many of the MA's apparently comes from the Mark-Q overlaps. But if those overlaps were fabricated/modified from GMark, then of course they would generate all kinds of MA's!

Dave:

I don't claim the study can prove Luke used Matthew. I do think it provides some additional evidence in that direction, however. The 200-202 relation is strong. The MAs studyed are not mostly from the overlap. In fact, to be clasified as "MA", the HHB authors had to decide that the item was an MA and not an overlap. The size of the agreement was the primary criteria, as far as I can tell.

[Bernard Muller]

From BERNARD:
First, I want to thank you Dave for answering all my posts, more so when I feel I am given the silent treatment from others on this list.
DAVE wrote:
Sondergut Luke: 5755
The number of words of triple agreement: 1493
The number of words in the MAs: 207
Note: This involves judgement, what is an MA and what is an overlap?
Matthew-Mark agreement against Luke: 992
Luke-Mark agreement against Matthew: 531

Correct me if I am wrong Dave, but I am under the impression Matthew rewrote about half of his Markan material, and produce the other half word by word. However Luke rewrote all (or almost all) his/her Markan material. If it is the case, that would explain why we have more agreements Matt-Mark against Luke than Luke-Mark against Matthew.

I want to add up here, my impression is Luke did some "smooth" rewriting of Q material (but without changing the meaning), when Matthew was a lot more "brutal" on the same material. If it is true, it is bound to make Q look more Matthean than Lukan!
That would be quite unfair: one writer (Matt) modifies Q more than the other one (Luke), so Matt is credited a hand on Q (according to your study), and the other one, Luke, is accused to change Q a significant lot.
So the consequence is the complete reverse of the initial hypothesis! Isn't it ironic!
Let's look at that in a different way: I have a rare text and I modify it, add up on it, with my own style & syntax. Another also uses the same rare text, but changes nothing.
No other version of this text is known. The initial text got lost.
What would be the conclusion?
According to your computer program (and with the (unproven) assumption the other guy HAD to work from the rewrite of the rare text, and NOT the initial text), the other guy is the corrupter!

I also notice the relative small number of MA's. Would that number include the DE replacing the KAI of GMark?

I'll have to study your next replies, and make some research on my own, because of the specifics you brought (I love specifics!).

Best regards, Bernard

[Yuri Kuchinsky]

Originally posted by GentDave
:

Let me just add some totals from the study, that might be of intrest. The total of all the words in all the categories in the study is: 25,843
Sondergut Luke: 5755
The number of words of triple agreement: 1493
The number of words in the MAs: 207
Note: This involves judgement, what is an MA and what is an overlap?


Dave,

So, according to the HHB Synoptic Concordance, the number of the Anti-Markan Agreements (AMAs) is only 207?

But that's ridiculous! Even F. Neirynck, a hard core 2SH stalwart that he is, listed over 750 of them in his 1974 study, [Leuven University Press, 1974]! And the real number is around 1000...

Of course we also have to consider tons of _negative_ AMAs, as well, i.e. all those places where Mt and Lk both happen to miraculously "delete" the very same Markan words and phrases in unison.

So I conclude that the HHB Concordance is flawed right from the start, and thus it cannot provide an accurate picture of the Synoptic relationships.

Also, contrary to what you said, it's not true that the majority of the AMAs are just KAI => DE cases.

Regards,

Yuri.

[Yuri Kuchinsky]

Originally posted by GentDave
Yuri writes:

"And the Q scholars have demonstrated in many detailed studies that, for the *majority* of the text that Matthew and Luke have in common outside of Mark, Luke is more original."

Dave:

I don't think that is the case.


Sure it is the case!


Luke's order is generally supposed to be more original, but not necessarily his wording.


Luke's order is one thing (and also very significant), but his wording is also generally taken to be more original by the Q scholars. See that section in my ORIGINALITY OF LUKE article,

http://www.iidb.org/vbb/showthread.php?threadid=60324

These are the very basics of Q scholarship...


Now, this is where we get to various possible interpretations of the results of the study, but what I believe is happening there is that sondergut Matthew contains parts of Q that Luke did not use, so that sondergut Matthew is really a composite of the styles of Q and Matthew.

Well, I don't believe in Q, in any case, so none of this really persuades me. The only part of Q scholarship that I do accept is that the Lukan Sayings Material is more original than that of Mt.

Regards,

Yuri.

[GentDave]

Yuri:

So, according to the HHB Synoptic Concordance, the number of the Anti-Markan Agreements (AMAs) is only 207?

Dave:

No. The study only covers about 800 of the most common Greek words, and actually KAI is left out, because they thought it was *too* common to tabulate. Plus, how many MAs you get depends on what you count as Mark-Q-overlap and what you count as MAs.

Yuri:

Of course we also have to consider tons of _negative_ AMAs, as well, i.e. all those places where Mt and Lk both happen to miraculously "delete" the very same Markan words and phrases in unison.

Dave:

Yes. I consider those negative agreements to be some of the best evidence available for a proto-Mark. You can see a related thing going on in the study, that I comment on in my proto-Mark section. I think it is quite reasonable to suppose those negative agreements were added to Mark after Matthew and Luke used its predicessor.

[GentDave]

Yuri,

I can't really comment much on other's work.

The study shows 102 just less that significantly related to 202.
201 is just more than significantly related to 202.
200 is significantly related to 202.
201 is very significantly related to 200.
102 is not significantly related to Luke's other categories.

Now how we interpret all that would depend on what hypothesis we are evaluating. What I would propose is that Luke may be closer on the true *sayings*. However, the section commonly called "Q", also contains things that originate in Matthew and were copied by Luke, and were not in "Q". On these things Luke copies Matthew, (Or proto-Matthew), so Matthew is more original.

[GentDave]

Bernard:

I want to add up here, my impression is Luke did some "smooth" rewriting of Q material (but without changing the meaning), when Matthew was a lot more "brutal" on the same material. If it is true, it is bound to make Q look more Matthean than Lukan!
That would be quite unfair: one writer (Matt) modifies Q more than the other one (Luke), so Matt is credited a hand on Q (according to your study), and the other one, Luke, is accused to change Q a significant lot.
So the consequence is the complete reverse of the initial hypothesis! Isn't it ironic!
Let's look at that in a different way: I have a rare text and I modify it, add up on it, with my own style & syntax. Another also uses the same rare text, but changes nothing.
No other version of this text is known. The initial text got lost.
What would be the conclusion?
According to your computer program (and with the (unproven) assumption the other guy HAD to work from the rewrite of the rare text, and NOT the initial text), the other guy is the corrupter!

Dave:

O.K. lets work with that. We also need to take into account that we have other samples of Luke's and Matthew's writting from the non-Q (triple-tradition and sondergut) sections.

Let's say Q is heavily modified by Matthew, and lightly modified by Luke. What would we find?

Category 202 will be the style of Q, this is where Matthew and Luke agree.

Category 102 will also be the style of Q. It will mostly reflect Luke's retaining the original Q, where Matthew made changes.

Category 201 will be the style of Matthew. Since this reflects Matthew's changes to Q.

So:
1) 202 and 102 should be related.
2) 201 and 202 should not be related.
3) 201 should be related to Matthew.
4) 102 should not be related to Luke.

#4 is true.
#3 is maybe as true as can be expected.
#2 is not true.
#1 is not true.

You can make this fit with a version of the 2SH, but it more naturally fits with the 3SH. Some of "Q" is really the saying source, and some of "Q" is rellay stuff Luke grabed from Matthew.

[Bernard Muller]

DAVE wrote:
O.K. lets work with that. We also need to take into account that we have other samples of Luke's and Matthew's writting from the non-Q (triple-tradition and sondergut) sections.
Let's say Q is heavily modified by Matthew, and lightly modified by Luke. What would we find?
Category 202 will be the style of Q, this is where Matthew and Luke agree.
Category 102 will also be the style of Q. It will mostly reflect Luke's retaining the original Q, where Matthew made changes.
Category 201 will be the style of Matthew. Since this reflects Matthew's changes to Q.
So:
1) 202 and 102 should be related.
2) 201 and 202 should not be related.
3) 201 should be related to Matthew.
4) 102 should not be related to Luke.

#4 is true.
#3 is maybe as true as can be expected.
#2 is not true.
#1 is not true.

OK, let me massage that, but I want to keep your thoughts and study results 100%. The reason I am spelling everything is I want no misunderstandings.

#4: About "Luke's
(allegedly, because we do not know of original Q. Please note my correction is at my disadvantage)
retaining the original Q, where Matthew made changes.". Your study shows most occurrences of that (102) are related to Luke's style.
If it is so, did you consider Luke made quite a few extrapolations, extensions on Q passages to his/her liking (as the ones about giving or cancelling debts)? I noted that long ago. And Luke had reason to adopt the same Q style to make the Lukan additions look original. More so if Luke (and likely his/her community), as generally agreed, had more respect for Q than for GMark (and the reverse for Matthew). Furthermore, they are a lot more insertions in Q by Matthew than by Luke, which would confirm the point. You do not insert in a text you (or your community) highly consider and know.

#3: You seem to be leaning towards the change in Q here would be according to Matthew's style. Of course, that's OK by me.

#2: 202 & 201 are not related. That means Q passages would have GMatt & GLuke agreeing, but these passages are different of the ones agreeing with Luke's style.

#1: 202 & 102 are not related. That means Q passages would have GMatt & GLuke agreeing, but these passages are different of the ones agreeing with Matthew's style.

It seems to me we cannot draw any conclusion on #1 & #2, which neutralize each other.

For #4, I also agree that Luke might have rewritten Q passages in his/her style.

So in conclusion, I see Q passages (202) being faithfully transcribed by both Matthew & Luke.
I see other passages (102) either rewritten by Luke or in the style of Luke because of the Lukan additions on the same themes: "reverse effect".
I see still other passages (201) rewritten by Matthew (or in the style of Matthew because of the Matthean additions using the same expressions: "reverse effect").
The only thing tipping the balance so slightly is your comment for #3 "is maybe as true as can be expected", which I take as not as true as for #4.

Maybe I missed something, and I hope I did not misrepresent you, but I do not see any clear evidence for whatsoever.
Actually, and taking the "reverse effect" out of consideration, this is what could be expected: Each gospeler either being faithful or making changes on Q material in their own style.
Who made the most changes? We would have to assess the "reverse effect" first for both and look at the details & at the "coloring" of both gospelers.

Best regards, Bernard

[Bernard Muller]

DAVE wrote:
Mark 2:3
And they came bringing to him a paralytic.
Matthew 9:2
And behold, they brought to him a paralytic.
Luke 5:18
And behold, men were bringing on a bed, a man who was paralyzed.

Mark:
KAI ERCONTAI FERONTES PROS AUTON PARALUTIKON
Matthew:
KAI IDOU PROSEFERON AUTW PARALUTIKON
Luke:
KAI IDOU ANDRES FERONTES EPI KLINHS ANQRWPON OS HN PARALELUMENOS

Luke picks up IDOU, behold, from Matthew.

=================
222 211 112 212 221 122 121

IDOU 2 18 6 5 2 0 0
=================

Category 211 has a count of 18. Matthew likes adding this word to Mark's text. Matthew never ommits it if Mark has it. (122=0, 121=0)

The MA's (212) have a count of 5. Luke picks up some of Matthew's additions of "behold". Of the 13 times Luke uses the word in the triple tradition, 5 of them are exactly where Matthew added the word.


I notice Mark use IDOU in ten verses, starting at 1:2.
Luke used IDOU 55 times, Matthew 59 times (or 55/59 verses)
Luke used IDOU 9 times in Chapt. 1 & 2, which have no equivalent in GMark or GMatthew.
So right from there, I say Luke could have picked up IDOU from GMark; more, Luke seemed to be fond of IDOU!

5 MA's on IDOU? Wouldn't that be a coincidence?
With 45 five IDOU left in his/her bag, Luke was bound to put some IDOU where Matthew put them.

More, these IDOU's were put in place where the author (Luke or Matthew) wanted to raise attention, that's why they would use "BEHOLD!". So Luke & Matthew, by judgment call, picked up five common spots separately for IDOU's (but Luke missed the other Matthew's spots (13 times!)).

Also to be considered: Luke covers only 50% of GMark, so we have 15 IDOU's when Mark has ten for the whole gospel. Luke is using some 3 times more IDOU's (density wise) on Markan material than Mark's. And with Matthew's using even more IDOU's on the triple tradition, that would increase the chance of coincidence.
But why a perfectly Greek literate person would look for pointers from another text, on such words like IDOU and DE?

Best regards, Bernard

[Bernard Muller]

DAVE wrote:
category 200 (sondergut Matthew) and 202 (Q Mt-Lk agreements) are strongly related. On the simple 2SH, they should not be. However, if 200 (sondergut Matthew) contains parts of Q that Luke skiped, then the 2SH can explain the relation between 200 and 202.

Now you are saying 200 and 202 are strongly resembling each other in style.
The relation can be explained by Q & GMatthew made in the same city, by the same group of people.
AND
Matthew using liberally expressions found in Q (as noted by Mark Goodacre & myself) creating the reverse effect.

"However, if 200 (sondergut Matthew) contains parts of Q that Luke skiped, then the 2SH can explain the relation between 200 and 202"

Why? If Q Mt-Lk agreements and sondergut GMatthew are similar, why a transfer of some of sondergut GMatthew to Q would be required?
Why adding more of the same thing to the same would change anything?
:confused:

Best regards, Bernard

[GentDave]

========
1) 202 and 102 should be related.
2) 201 and 202 should not be related.
3) 201 should be related to Matthew.
4) 102 should not be related to Luke.

#4 is true.
#3 is maybe as true as can be expected.
#2 is not true.
#1 is not true.
==========

Let me rephrase to eliminate double negatives.
202 and 102 are not significantly related.
201 and 202 are significantly related.
201 is significantly related to sondergut Matthew (200)
102 is not related to Luke.

[GentDave]

Bernard:

Why? If Q Mt-Lk agreements and sondergut GMatthew are similar, why a transfer of some of sondergut GMatthew to Q would be required?
Why adding more of the same thing to the same would change anything?


Dave:

To make the 2SH work well with the results, I would suppose that what we call "sondergut Matthew (200)" contains some things written by Matthew, and some things written by author-Q that Matthew used and Luke did not. 202 would be all author-Q.

Since 202 was all Q, and 200 was partly Q, they might show some relation.

On the other hand, if 200 was only written by Matthew, I think it is less likely that it would be related to 202.

[GentDave]

Bernard:

I notice Mark use IDOU in ten verses, starting at 1:2.
Luke used IDOU 55 times, Matthew 59 times (or 55/59 verses)
Luke used IDOU 9 times in Chapt. 1 & 2, which have no equivalent in GMark or GMatthew.
So right from there, I say Luke could have picked up IDOU from GMark; more, Luke seemed to be fond of IDOU!

Dave:
I focused only on the triple tradition. The statistics reffer to that material. Focusing on only the triple tradition keeps the size of the text the same in all 3 gospels. Since Luke is larger than Mark, it is harder to compare apples to apples if we look at all occurances. We could do that but its harder to just eyeball. In the triple tradition Mark uses it 4 times. Matthew 27 Luke 13. So, no, Luke is not oppsed to the word.

Bernard:

5 MA's on IDOU? Wouldn't that be a coincidence?
With 45 five IDOU left in his/her bag, Luke was bound to put some IDOU where Matthew put them.

Dave:

No. It is not at all likely that 1/3 of Luke's additions would end up in the same places, when there are probably 100s of places they could have ended up.

But what the study does is calculate some real probabilites, across many words. The study concludes that the probability of the relation between 211 and 212 being only coincidence is about 5*10^-7 ( .0000005).

[GentDave]

More on IDOU

IDOU is 5 of 207 words in category "212" = 2.4%
IDOU is 18 of 1248 words in category "211" = 1.4%
IDOU is 104 of 25,843 words in all categories. = .4%
IDOU is 30 of 5755 words in sondergut Luke(002) = .5%

So Luke has a frequency typical of the synoptics as a whole. But 211 and 212 both show eleveated frequencies. This by itself is unlikely to be coincidence, and taken together with the other words in the study, coincidence is virtually eliminated as a possibility.

[Bernard Muller]

DAVE wrote:
========
1) 202 and 102 should be related.
2) 201 and 202 should not be related.
3) 201 should be related to Matthew.
4) 102 should not be related to Luke.

#4 is true.
#3 is maybe as true as can be expected.
#2 is not true.
#1 is not true.
==========

Let me rephrase to eliminate double negatives.
202 and 102 are not significantly related.
201 and 202 are significantly related.
201 is significantly related to sondergut Matthew (200)
102 is not related to Luke.

I am abandoning on that one and others. I do not have any certitude of what you are trying to say. Did it occur to you that not everybody is a computer analyst?

DAVE wrote:
No. It is not at all likely that 1/3 of Luke's additions would end up in the same places, when there are probably 100s of places they could have ended up.
But what the study does is calculate some real probabilites, across many words. The study concludes that the probability of the relation between 211 and 212 being only coincidence is about 5*10^-7 ( .0000005).

According to these results, or only the five IDOU's, I do not know why you entertain the possibility of 2SH. There is no chance for that. Luke had to know about GMatthew hundreds of times, at least! And not for important stuff, just things like the location of IDOU's, not all of them, only five.

DAVE wrote:
IDOU is 5 of 207 words in category "212" = 2.4%
IDOU is 18 of 1248 words in category "211" = 1.4%
IDOU is 104 of 25,843 words in all categories. = .4%
IDOU is 30 of 5755 words in sondergut Luke(002) = .5%
So Luke has a frequency typical of the synoptics as a whole. But 211 and 212 both show elevated frequencies. This by itself is unlikely to be coincidence, and taken together with the other words in the study, coincidence is virtually eliminated as a possibility.

So for these 5 IDOU's, which fit in only 207 words (another surprise!), Luke had to know about GMatthew!
So category 212 is barely existing but you make all kind of conclusion from it, as being related to sondergut Matthew. How could it be used as a reference? I suppose if the 5 IDOU's were not there, a big chunk of tiny 202 would move somewhere else.

Computer study like that can go beserk in a nanosecond, when you do not associate it with the human element: such as Luke and Matthew placing their own IDOU's into appropriate special host locations, relatively few in number, in order to enliven their narration OR the reverse effect Mark & I talked to you about.
Another problem is, if a result is generated wrong, and use to compute other results, due to the cascading effect, all the rest will be erroneous. When I see your 207 words of 202, I am very concerned, more so because a lot depends on 202.

Best regards, Bernard

[GentDave]

Bernard:

According to these results, or only the five IDOU's, I do not know why you entertain the possibility of 2SH. There is no chance for that. Luke had to know about GMatthew hundreds of times, at least! And not for important stuff, just things like the location of IDOU's, not all of them, only five.

Dave:

I personally do not believe the 2SH, certainly not in its simple form. However, saying that the relation between 212 and 211 is not pure chance, does not say for sure what did cause it. A proto-Mark? A Deutro-Mark? A proto-Matthew? Matthew?

Actually, the 2SH says that Luke and Matthew both authored the the MAs indepently. So you could say the problem for the 2SH is more the problem of explaining why Luke is not related to the MAs, than explaining why the MAs are related to Matthew.

Dave:

Regarding your other comments: If we were going to project how many IDOUs would be found in the 212 category, based on the frequency in the synoptics or in Luke, we would expect only 1. But we find 5, which is much more typical of Matthew's style. It provides a piece of concrete evidence the the MAs are in Matthew's style. That is just one vocabulary item. Studying the other 800 vocabulary items provides additional evidence, to the point of being conclusive.

Both people and computers are capible of making mistakes. I'm open to the possibilities that I have in fact made one. However, at this point, I don't believe I have. If you discover some specific mistake I have made, please bring it to my attention.

[Yuri Kuchinsky]

Originally posted by GentDave
More on IDOU

IDOU is 5 of 207 words in category "212" = 2.4%
IDOU is 18 of 1248 words in category "211" = 1.4%
IDOU is 104 of 25,843 words in all categories. = .4%
IDOU is 30 of 5755 words in sondergut Luke(002) = .5%

So Luke has a frequency typical of the synoptics as a whole. But 211 and 212 both show eleveated frequencies. This by itself is unlikely to be coincidence, and taken together with the other words in the study, coincidence is virtually eliminated as a possibility.

Hi, Dave,

Since I'm quite interested in the Anti-Markan Agreements (your category 212), I think it would be interesting to see some more such cases as IDOU in the 212 category.

So what are the most common words in 212? Why don't you post a couple more here with their detailed distribution in all other important categories? I suppose you should have such data handy?

Regards,

Yuri.

[Bernard Muller]

Dave wrote:
Regarding your other comments: If we were going to project how many IDOUs would be found in the 212 category, based on the frequency in the synoptics or in Luke, we would expect only 1. But we find 5, which is much more typical of Matthew's style. It provides a piece of concrete evidence the the MAs are in Matthew's style. That is just one vocabulary item. Studying the other 800 vocabulary items provides additional evidence, to the point of being conclusive.

So the MA's are more in Matthew's style. What does that prove? Luke's style, at a few times, was close to Matthew' style, because we have the MA's in the triple tradition, such as the IDOU's.
Conclusion: Luke was (much) less in Matthew's style than Matthew. Make sense.
And why the triple tradition material would look Matthean? Because Matthew stayed close of Mark's style (to the point of copying word by word) when Luke did rewrite almost everything. Then Matthew adopted Mark's style (and more so words as IDOU!) for his own stuff and the TrTrad, more so than Luke, who used more of his/her own style (as usual); so the reverse effect.
And why would Q look Matthean? Because Matthew used a lot more expressions from Q than Luke: reverse effect.
That's the way I see it.

Dave wrote:
A Deutro-Mark? A proto-Matthew?
That's possible, certainly not absurd as the whole of GMatthew.
But before deciding, I would want to see the whole ball of wax, all words, all verses related to the relatively few MA's. And then take in account the authors' patterns, such as about why they would use the IDOU's on Markan material and then, what would be the most likely locations to do so for each. You see the five IDOU's with cold blood (as a computer analyst, I may add), but it may be argued, that if more can be gathered about the two authors' thinking (still by looking at the primary evidence), there is a much greater chance (than one) to have five so-called coincidences for IDOU's.

Dave wrote:
Both people and computers are capable of making mistakes. I'm open to the possibilities that I have in fact made one. However, at this point, I don't believe I have. If you discover some specific mistake I have made, please bring it to my attention.

Humm! I would first look at reverse effects. That would be the same as a fractional circular argument (or unwanted feedback). Of course, I am concerned that the soft rewriting of Luke on TrTrad would greatly influence your findings, more so since your study is based on words, one by one.
I do not know about word based research, which I think can be very misleading.
If word based research is used for MA's, as I said, if there is an abnormal amount of coincidences on one word, I would look at the human element to see if it can be explained. Having Luke takes his/her cues on DTR-Mark or PRT-Matthew (or GMatthew) to know where to place five IDOU's is rather stupid in my way of thinking.
Isolate Q material in one category, including the Q version of the Mark-Q overlaps.
I would look also at the obvious coloring & bias & likings of the authors, bound to affect the wording for each, and not related to any sourcing (and affect reverse effect).
The big question would be, and still is: did Matthew and Luke have a deutero-Mark (or proto-Matthew with Q included) instead of plain GMark. Better try to answer that, rather than getting involved in absurd situation, like Luke had the whole of GMatthew, which brings a lot more problems than it solves (even if they are hard driving scholars who want that: that sells books).

Best regards, Bernard

[GentDave]

Yuri,

Working on the request. I should be able to post a good example tonight. I don't have my books here.

[GentDave]

Bernard:

So the MA's are more in Matthew's style. What does that prove?

Dave:

I suppose that depends on what the question is. Other evidence from the study, and traditional arguements, say that Matthew used Mark. If this is true, then category 211 is additions to Mark by Matthew. Since 212 reflects the style of 211, we would say the minor agreements appear to be authored by Matthew, and that would argue for Luke's use of Matthew, or proto-Matthew, and against say a Deutro-Mark.

Now, if we were trying to argue against the 2SH, in its basic form, we'd really want the answer to a differant question. "Are the MA's also in the style of Luke?" The study does not address that question directly. It finds no evidence Luke *is* related, but it does not prove it *is not*. I'm going to get some specific examples for Yuri, that help with that.

Bernard:

Luke's style, at a few times, was close to Matthew' style, because we have the MA's in the triple tradition, such as the IDOU's.

Dave: The way the study works, the style of the autor is defined by how frequently he tends to use certain words.

Bernard:

And why the triple tradition material would look Matthean? Because Matthew stayed close of Mark's style (to the point of copying word by word) when Luke did rewrite almost everything.

Dave:

But, when I am looking at Matthew's style, in say category 211, we are looking specificly at words that Matthew did not take from Mark.

Bernard:

You see the five IDOU's with cold blood (as a computer analyst, I may add), but it may be argued, that if more can be gathered about the two authors' thinking (still by looking at the primary evidence), there is a much greater chance (than one) to have five so-called coincidences for IDOU's.

Dave:

Some people dislike the idea that people and their choices can be described statisticly. Marketers and advertisers do it. So do financial markets. I'm sure if we went case by case, we could come up with "just so" stories for any number of them. But that would not tell us how likely each senerio was, nor gather evidence across many examples. What you reffer to as "cold-blood", I would call "objective". It removes as far as possible subjective human impressions, and looks to the extent that is possible and hard facts.

Bernard:

Of course, I am concerned that the soft rewriting of Luke on TrTrad would greatly influence your findings, more so since your study is based on words, one by one.


Dave: Not sure what you mean here. Do you mean that late harminisers may have made Luke look a little more like Matthew? That is indeed a possibility.

Bernard:

Isolate Q material in one category, including the Q version of the Mark-Q overlaps.

Dave:

That material is in three categories.

102 = Q found only in Luke
202 = Q found in both Matthew and Luke
201 = Q found only in Matthew

Bernard:

The big question would be, and still is: did Matthew and Luke have a deutero-Mark (or proto-Matthew with Q included) instead of plain GMark. Better try to answer that, rather than getting involved in absurd situation, like Luke had the whole of GMatthew, which brings a lot more problems than it solves (even if they are hard driving scholars who want that: that sells books).

Dave: Well the study does not answer that with certainty. It provides evidence in the direction of Matthew. However given

1) The overlaps in Luke seem to be based on Matthew. (traditional means)
2) The MA's seem Matthian.
3) Q looks alot like sondergut Matthew
4) Matthew exists, we have no direct evidence of other documents.

I would think the burden of proof would lie on the person claiming that Luke did not use Matthew. I would think that person would have to show that Luke using Matthew could not explain certain features of the written record, and therefore a more complex hypothesis is needed.

[Bernard Muller]

Dave wrote:
If this is true, then category 211 is additions to Mark by Matthew. Since 212 reflects the style of 211, we would say the minor agreements appear to be authored by Matthew, and that would argue for Luke's use of Matthew, or proto-Matthew, and against say a Deutro-Mark.

Later, in the same post, Dave wrote:

But, when I am looking at Matthew's style, in say category 211, we are looking specificly at words that Matthew did not take from Mark.

Dave, I always wondered how you determined 211 (or 112). **It is by words not in GMark**.
I suppose 200 (or 002) would be the same, also for 210 (or 012).
Am I right?

BTW, I doubt 012 has anything in it, in view Luke used 50% of GMark when Matthew used 90%. The chance are slim Luke used any Mk verses from the 10% missing in GMatthew.

So now we have: 212 reflects 211, and 211 is based on Matthean words not in GMark. That means wherever Matthew dropped a non-Markan word in the TrTrad, the affected sentence (I assume, or more, or less!), or rather its words, would go into the 211 category (total 1248 words). Then from that, you would look at the key words in it (as IDOU) and compare that with 212 key words (as IDOU again). The same of course from 112.
And if I am (finally) right, one (non-Markan) key word would enable a bunch of other keywords.

I found the info that Matthew used 137 words not found anywhere else in the NT. I do not know about GLuke but it is probably a lot less (because not worth mentioning).
If it is the case, and representative of non-Markan words in GMatthew and GLuke, then 211 (& 200 and 210) is going to unable a lot more key words in favor of the Matthean cause than 112 (and 002 but with the corresponding of 210, that is 012, being quasi-inexistent).

My questions are:
How many non Markan keywords in GMatthew are used to determine the Matthean flavor?
How many non Markan keywords in GLuke are used to determine the Lukan flavor?
How many words are in 112?
More, can you give the number of words in each category?
More, can you answer precisely my questions, for a change?
And I am still waiting for:
The list of all keywords (such as IDOU) which triggered MA's, their respective numbers and associated verses, or part of them (adding up to 207 words).

Dave wrote:
Now, if we were trying to argue against the 2SH, in its basic form, we'd really want the answer to a differant question. "Are the MA's also in the style of Luke?" The study does not address that question directly. It finds no evidence Luke *is* related, but it does not prove it *is not*.

I do not see why the issue had not been addressed.

Dave wrote:
The way the study works, the style of the author is defined by how frequently he tends to use certain words.

That may be a big weakness in your study, possibly misleading in some area. Frequency of words is one thing, but maybe not enough to support some of your conclusions. As a matter of fact, the author using the same key words more often than the other may become predominent just because of that. And if Luke used a wide vocabulary, and not repeating his key words too often, that would probably take the author below the radar screen in many cases.

Dave wrote:
Some people dislike the idea that people and their choices can be described statistically. Marketers and advertisers do it. So do financial markets. I'm sure if we went case by case, we could come up with "just so" stories for any number of them. But that would not tell us how likely each scenario was, nor gather evidence across many examples. What you refer to as "cold-blood", I would call "objective". It removes as far as possible subjective human impressions, and looks to the extent that is possible and hard facts.

Ya, I got your point. But your program can be the equivalent of these methodologies scholars used. They look attractive, seems to make sense but lead to divergent results. Your program can make sense to you, but it may be not perfect. Just a tiny flaw at the start (the butterfly effect), or the peculiar nature of the evidence, can drive some of your conclusions wild.

Dave wrote:
1) The overlaps in Luke seem to be based on Matthew. (traditional means)
2) The MA's seem Matthian.
3) Q looks alot like sondergut Matthew
4) Matthew exists, we have no direct evidence of other documents.

I would think the burden of proof would lie on the person claiming that Luke did not use Matthew. I would think that person would have to show that Luke using Matthew could not explain certain features of the written record, and therefore a more complex hypothesis is needed.

I read your pages many times, but you do not provide enough data to follow you and check you out, step by step.
I am questioning your conclusion from the MA's and Q. I still maintain they are mainly due to feedback (reverse effect), that is Matthew adopting aggressively key words from GMark & Q, more so than Luke, making the MA's & Q look Matthean. Of course, not knowing about the key words in question, the verses involved, etc., I cannot be more specific than I am now or in my previous posts. Maybe my question is: are you sure it is not the case?

I think your assumption might be the two gospelers were scribal compilers, working from the same material, and putting their self aside, did their desk work as part of their regular job (cold blooded!). Then maybe your study might be OK. But the gospelers were not scribes. They had different agenda & bias & mood & audience & (and likely sex!), reflecting in the key words they were using and the frequency of them.

Best regards, Bernard

[GentDave]

Bernard:

Dave, I always wondered how you determined 211 (or 112). **It is by words not in GMark**.

Dave:

Well, I didn't that was done by the authors of the HBB synoptic concordance. Also, let's be clear the English word "word" has multiple meanings. To counted in 211, the *occurence* of the word must be in Matthew, without a parallel in the other 2 gospels. The *vocabulary item* may ocur in all 3 gospels.

Here are links to the data description, and the information from my site.

http://www.davegentile.com/synoptics/synoptics.txt
http://www.uni-bamberg.de/ktheo/nt/forschung/synconbe.htm

Each instance of a word in the synoptics is placed into one of a number of
"synoptic categories". Each category is designated by a three-digit number.
"210", for example.

The first digit gives information about Matthew, the second refers to Mark,
and the third to Luke. The digit "0" indicates no parallel.
The digit "1" indicates a parallel exist, but does not contain the key word.
The digit "2" indicates the parallel contains t