Synoptic Hypotheses and Authors

Introduction

Much of the text of Mark, Matthew, and Luke is very similar to, and in some places identical to, text in either one or both of the other two gospels, and for this reason they are collectively referred to as the Synoptic (‘seeing together’) Gospels, or just the ‘synoptics.’ The existence of these synoptic parallels raises the issue of how they came to exist, and a very good general statement of the problem (usually referred to as the synoptic problem), with suggested solutions, is given by Stephen C. Carlson on the Synoptic Problem Website. He comments:

It is rare for two independent reporters of the same event to share more than a few words in common, but the synoptic gospels often feature a substantial number of agreements in their exact words. For example, in one passage about John the Baptist, Matthew and Luke agree for 61 out of 63 Greek words of a presumably Aramaic speech. Generally, the verbatim agreement between Matthew, Mark, and Luke runs about 50% of the words, but, by contrast, their agreement with John in parallel episodes falls to about 10%.

Because of these agreements between Mark, Matthew, and Luke a number of questions come to mind: How did the agreements arise? Were the words in one gospel copied from another, or did the common words come from other (unknown) sources? Where words are the same in two or all three gospels, can we tell which was written first? There are various hypotheses that attempt to answer these questions, generally involve copying and/or editing between Mark, Matthew and Luke, possibly supplemented by additional unknown source material, typically either earlier material from which one or more of the synoptics was derived or a document containing material common to two of them.

Although the synoptic problem is not ‘solved,’ the majority opinion is that Mark was the first of the synoptics to be written (this is known as Markan Priority), with Matthew and Luke following, and this is the case in the currently most accepted hypotheses. However, other orders (there are six of this form) still have their proponents, because the constant problem has been the great difficulty of falsifying, or 'ruling out' any particular hypothesis. In general this is because each synoptic gospel contains text for which we have no known source, and because so much of their contents can be considered to be simply due to editorial choice. This discussion proposes a mechanism through which it should be possible to determine that the author of one synoptic had to have known the contents of at least one (and possibly both) of the other two.

Note: In the discussions below ‘aMark’ refers to the author of Mark, and similar notation is used for the authors of Matthew, Luke, and any other author of a ‘named’ piece of text.

Synoptic Hypotheses

Unless otherwise specified, synoptic hypotheses assume that the author of a later gospel may have been influenced by the content of an earlier one, and while some hypotheses require only that A => B and A => C, i.e. that gospel A was first and that B and C both depend only on A, many specify a dependency of the form A => B => C and A => C, so that C depends on both A and B.

Without a non-synoptic source (sometimes known as a ‘second source’), there are six possible ways in which a dependency of the form A => B => C and A => C could have existed, all of which have been studied as possible solutions of the synoptic problem. They are usually referred to by the names given below, and the first shown here is currently one of the two most favored hypotheses:

1. Mark => Matthew => Luke and Mark => Luke The Farrer / Goulder (FH) or Mark without Q (MwQH) hypothesis
2. Mark => Luke => Matthew and Mark => Matthew The Wilke or Mark-Luke model
3. Matthew => Mark => Luke and Matthew => Luke The Augustinian hypothesis (AH)
4. Matthew => Luke => Mark and Matthew => Mark The Greisbsach or Two-Gospel (2GH) hypothesis
5. Luke => Mark => Matthew and Luke => Matthew The Lockton or Luke-Mark model
6. Luke => Matthew => Mark and Luke => Mark The Busching or Luke-Matthew model

There are also multiple variations on the above hypotheses that include other ‘stages’ in the development of one or more of the synoptic gospels, e.g. by including proto (earlier) or deutero (later) versions of a gospel, or by allowing other sources, and in some cases these hypotheses do not then need so many dependences between the gospels themselves. For example:

1. pG => Mark, pG => Matthew, pG => Luke The Proto-Gospel Hypothesis
2. pMt => Matthew, pMt => Luke, pMt => Mark, Luke => Mark The Proto-Matthew Hypothesis

One other hypothesis that includes an additional source has gained a very strong following and is the other of the two currently most favored hypotheses:

Mark => Matthew, Mark => Luke, Q => Matthew, Q => Luke The Mark-Q or Two Source (2SH) Hypothesis

Because one gospel author may have written a parallel (i.e. identical or very similar) to text he saw in an earlier gospel, text originating in Mark, Matthew and Luke can be categorized according to how many of the synoptic gospels contain parallels of that text. There are seven combinations depending on which of the synoptic gospels contain a particular parallel: Three refer to text that exists in only one gospel (i.e. was not re-used by the other two authors), three to text that exists in two gospels, and one to text in all three gospels. All but the two containing the smallest number of words have commonly used names:

1. Text unique to Mark (Sondergut Mark or SMk)
2. Text unique to Matthew (Sondergut Matthew or SMt)
3. Text unique to Luke (Sondergut Luke or SLk)
4. Parallel text in Mark and Matthew but not Luke
5. Parallel text in Mark and Luke but not Matthew
6. Parallel text in Matthew and Luke but not Mark (Double Tradition or DT)
7. Parallel text in Mark, Matthew, and Luke (Triple Tradition or TT)

If these combinations are considered as divisions of the text of the synoptic gospels, then taken together the seven divisions contain all of the synoptic gospel text. The implication here may be that the parallels were created by one gospel author re-using something that he saw in another gospel, but it is of course possible that either two or all three of the authors of what we know now as the synoptic gospels saw and re-used something that they jointly saw in an unknown document, via which any or all of the parallels could have been created without the authors having any knowledge of the other synoptics.

What Did the Authors Know?

As shown above there are six possible orders in which the authors of the three synoptic gospels could have written their gospels, and assuming that each author knew the earlier gospel or gospels (and for now assuming there was no second source) the following diagram indicates which of the ‘divisions’ each of the authors could by his actions create.

(Gospel1/2 etc. means that portion of those two gospels that is common, and Gospel1- and Gospel2- represent the portions of those Gospels that are not in Gospel1/2)

The author of the last synoptic gospel to be written (Author3) is the only one of the three authors to be able to identify the above seven divisions into which we now categorize the text of the synoptic gospels, because by his actions he created those divisions. Author1 just writes his gospel and has no knowledge of what Author2 and Author3 will or will not do with his text. Author2 can both write his own text and add to it text (possibly edited) from Gospel1, after which only some of Gospel1 is unique. At this point these three categories do not match the divisions that we see, because Author3 has yet to subdivide those categories into text he adds to his own gospel and text he does not.

For example, only Author3 could know which text in category Gospel2- he would ignore, in the process creating Sondergut2, and which new text of his own he would add, so creating Sondergut3. Author2 would of course know which portions of Gospel1 he did or did not use, but at the time of writing may not have known about Author3, or perhaps did but did not know how Author3 would treat those same portions of Gospel1. Using this format the following diagrams show how the seven divisions of the text that we see would have been created on the assumption of three of the major synoptic hypotheses.

The Greisbsach or Two-Gospel (2GH) Hypothesis

On this hypothesis the order of the creation of the synoptic gospels is Matthew => Luke => Mark, with aLuke knowing Matthew but not Mark, and aMark knowing both Matthew and Luke.

(Here the ‘-‘ signs, such as in Matthew-, indicate that this excludes text in common with one or both other gospels, such as Matthew/Luke)

There are some particular consequences of this order:

1. Although aLuke knew which text was common to his gospel and Matthew (because he created it), this text was not just the Double Tradition because it still included text that aMark would later turn into Triple Tradition text when he included it in his own gospel.
2. Although aLuke knew which text of Matthew he ignored, it was not at that point all Sondergut Matthew because aMark would later include some of it in Mark.
3. The text that aLuke added was not all what we know as Sondergut Luke because aMark would later include some of it in Mark.

The Augustinian Hypothesis (AH)

On this hypothesis the order of the synoptic gospels is Matthew => Mark => Luke, with aMark knowing Matthew but not Luke, and aLuke knowing both Matthew and Mark.

The Farrer / Goulder (FH) or Mark without Q (MwQH) Hypothesis

On this hypothesis the order of the synoptic gospels is Mark => Matthew => Luke, with aMatthew knowing Mark but not Luke, and aLuke knowing both Mark and Matthew.

Common to the Above Hypotheses

These three hypotheses are all supposed to be able to give rise to exactly the same results, e.g. so that in all three cases the content of the hypothesized Sondergut Mk agrees with what we know (because we can see it) actually is Sondergut Mk. However, there are problems with this: For example, it is only the author of the last synoptic gospel to be written who can make decisions regarding what he is going to put in his gospel based on where the two previous gospels have parallel text, and what he is going to ignore.

Consequently, those decisions are likely to lead to different results depending on the order in which the gospels were written, even though all three hypotheses can create the same seven divisions of the text. Using Mark as an example, on the 2GH, AH, and MwQH Sondergut Mk is created in three different ways:

1. On the 2GH Sondergut Mk is simply text that aMark decided to add to whatever text he took from Matthew and Luke;
2. On the AH Sondergut Mk is that part of the text that aMark added to whatever he took from Matthew, that aLuke did not add to his gospel;
3. On the MwQH Sondergut Mk is that part of the text of Mark that aMatthew added to his own gospel, that aLuke did not.

As it is highly unlikely that these three different text-selection processes would result in exactly the same text in SMk, the text that we actually see in each of the seven divisions may hold clues as to the order in which that text was created. For example, in each of the divisions that contains text from Mark (SMk, Mark/Luke, Mark/Matthew, TT) there may be textual artifacts that tell us whether the text appears to have been written first and then later split apart when being used in Matthew and/or Luke, or it appears to have written in small pieces intended to be inserted into something written earlier. For example, is Sondergut Mk just a collection of left-overs that neither aMatthew nor aLuke re-used, or was it written as an add-on after aMark had already gathered together what he wanted to re-use from Matthew and Luke?

The above term ‘textual artifact’ is not meant to refer to the use of particular words by each author, e.g. suggesting that the text written by each author can be recognized by their commonly used (or not) words, a study of which is termed Stylometric Analysis. Instead, it is meant to refer to characteristic ways in which an author structures his text. As an example, all three synoptic gospels contain doublets, which are pieces of text a second version of which (a parallel) exists in the same gospel, and which may or may not have parallels in the other two gospels. Doublets can be created either by an author simply re-using something he had written previously, or by writing something in his gospel and then elsewhere in his gospel adding a parallel version of the same thing taken from another source, and a study of these textual artifacts in the synoptic gospels may provide clues to the order in which they were written.

In general, if one of the synoptic gospels contains multiple instances of a textual artifact unique to only one of the seven divisions of the text identified above, then the author of that gospel must have been able to distinguish that division from the other six, i.e. that gospel must have been the last of the three to have been written. For example, if there was a textual artifact that occurred multiple times in Luke, and in every instance it was present in or connected to the double tradition text in Luke, then we would have to conclude that aLuke could distinguish the double tradition from the rest of his gospel, meaning that Luke must have been the final gospel to be written.

The Mark-Q Hypothesis

There is a scenario that has not yet been examined. Suppose the writers of the second and third synoptic gospels each made use of the first, but neither knew each other. Then, and using Mark as the first gospel, we would have:

These are the same as the first copying ‘events’ in the MwQH and Wilke or Mark-Luke hypotheses respectively, and because neither aMatthew nor aLuke knew of the existence of each other’s gospel there would be no non-Markan text common to both gospels, i.e. no double tradition text. Instead the only text the two could have in common would be a portion of Mark, which would not necessarily be SMk because the text included in Matthew would still include some of the text of Mark that aLuke would later (on the MwQH) choose to include in his gospel, with a similar situation regarding the early stages of the Wilke or Mark-Luke model. After these two events none of the divisions in the gospels correspond to any of the seven that we know.

These are examples of the general point that for any of the synoptic authors to be able to identify the seven textual divisions being discussed here they must have known both of the other two synoptic gospels. However, if they knew one of the other synoptic gospels and another source then there is the possibility that, depending on the contents of that other source, they could know at least some of the divisions that we recognize. For example, if any of them additionally used all the text of a non-synoptic source that contained exactly what we now know as of one of the divisions then they would know that division, although they would not know that it was that division. For example, if aLuke knew Mark and also incorporated everything from a source containing just the text that we see as Sondergut Luke then aLuke would know his own Sondergut, but strange as it may seem in this scenario he would not know that it was his own Sondergut because he could not know (for example) whether aMatthew would use any of it.

The most likely candidate for “a non-synoptic source that contained exactly the text that we now know as the text of one of the divisions” is the source known as Q, which on the traditional definition in the Mark-Q hypothesis contains almost exclusively parallels of the double tradition text. However, it is often allowed that Q actually contains some text that overlaps with Mark, and some people believe that it also could have contained some of what we see as Sondergut Matthew and/or Luke. On the Mark-Q hypothesis the order of the synoptic gospels is Mark => Matthew and Mark => Luke, with aMatthew and aLuke knowing Mark and Q but not knowing each other, so for this hypothesis the diagrams shown above have to be augmented to include Mark and Q as sources for both Matthew and Luke. However, this is not similar to including another source in sequence, because in each case there is only one copying ‘event’ that we know of, ending with both aMatthew and aLuke creating their respective gospels using just Mark, Q and their own materials as sources. How aMatthew and aLuke prioritized or used Mark, Q and anything else they may made use of is not known, so their uses of Mark and Q have to be shown below as happening contemporaneously.

(As indicated above the text of Q is not fixed, so ‘Other’ here refers to any part of Q that is not double tradition (DT) text, most likely text in common with Mark (usually referred to as Mark-Q Overlaps), but also possibly non-DT text in common with Matthew or Luke but not both)

As previously shown the textual divisions known to aMatthew and aLuke do not match those that we see, for example although both authors excluded some of Mark, they did not exclude the same text and hence ‘SMk+’ is not what we know as SMk. The one possible exception is the DT in the event that Q contains just the DT, i.e. that ‘Other’ actually contains no text. This means that on this hypothesis none of the gospel authors could have seen six of the seven divisions that we (knowing all the synoptic gospels) are able to distinguish, with the possible exception being the DT. The corollary of this is that if the text of either Matthew or Luke can be shown to contain artifacts unique to any of the divisions then (unless the artifacts are unique to the double tradition), it cannot be accounted for on the Mark-Q hypothesis.

Summary

In the first three scenarios described above the author of the last of the synoptic gospels to be written is the only one who knew both of the other two gospels, and as a result he is the only one of the three authors who is capable of knowing the seven textual divisions into which we subdivide the text. In the fourth scenario none of the authors is capable of knowing those divisions, with the possible exception of the double tradition. These observations give rise to the following:

- If Mark contains a statistically significant number of examples of a textual artifact in one or more of the seven textual divisions but none in the others, then Mark was written last;
- If Matthew contains a statistically significant number of examples of a textual artifact in one or more of the seven textual divisions but none in the others, then Matthew was written last;
- If Luke contains a statistically significant number of examples of a textual artifact in one or more of the seven textual divisions but none in the others, then Luke was written last.

Given the possibility that a ‘statistically significant number of examples’ could still by occur by chance, and hence on the above all three gospels could appear to have been written last, a more useful expression of these findings would be:

The greater the number of statistically significant examples of different textual artifacts that exist in one or more of the seven textual divisions (with other divisions having no examples) in a synoptic gospel, the more likely it is that that gospel was written last.

Conclusions

The seven different categories into which the text of the synoptic gospels are commonly divided (based on which of the gospels contain parallel versions of that text) came into existence as a direct result of the last of the synoptic gospels having been written. As a result only the author of that gospel could have known the divisions of the text that we see, and only he could have made repeated use of some textual artifact in the text of one division but not another. The existence of such use of a textual artifact in a synoptic gospel would therefore point to it having been the last one to have been written.