|
Determining Literary Dependence
Basic overview of some methods and pitfalls...
There are a few simple rules that enable us to determine if there exists a literary dependency between a set of texts. They are simple in theory, but difficult when applied to the actual texts themselves. A first rule entails comparing the texts. If the wording, order and content of works is so similar and extensive that coincidence must be ruled out, the exegete must posit a literary relationship between the texts. If the texts do appear to have a literary dependence we do not know the nature of this relationship. One author may have known the other author's work but even this is vague. Did he have a copy of it, or did he see a copy of it in the distant past? Has the author merely heard about it or has the content of the source became conventional knowledge? Or maybe both sources shared a common source---one that we might not now possess. Determining the relationship between the texts is the difficult part. We know there is a relationship between the gospels of Matthew, Mark and Luke. The majority of Mark is reprinted in Matthew and Luke, sometimes verbatim. The order, wording and extent of this material cannot be attributed to oral memory or coincidence. Somehow the texts knew one another. The material that Matthew, Mark and Luke share is called "triple tradition material." Both Matthew and Luke are significantly more extensive than the text of Mark. They share an extensive amount of non-Marcan material known to scholars as the "double-tradition". In this case as well the material is so similar that either a shared written source or dependence upon one another must be posited. The late Raymond Brown summed up the situation succinctly in his Introduction to the New Testament, "Mark has 661 verses (vv.); Matt has 1,068, and Luke has 1,149. Eighty percent of Mark's vv. are reproduced in Matt and 65 percent in Luke. The Marcan material found in both the other two is called the "Triple Tradition". The approximate 220-235 vv. (in whole or in part) of nonMarcan material that Matt and Luke have in common is called the "Double Tradition." In both instances so much of the order in which that common material is presented, and so much of the wording in which it is phrased are the same that dependence at the written rather than simply at he oral level has to be posited." 1 The nature of the relationship is difficult to assess. For the triple tradition material, did Matthew use Mark or Luke? Did Mark use Luke and Matthew used Mark? Any combination can be imagined. What is important is how well the theorized line of dependence stacks up with the textual data. The double tradition can be explained in three ways: 1) Matthew and Luke shared a common sayings source. 2) Matthew knew and used the text of Luke. 3) Luke knew and used the text of Matthew. Making the issue more difficult is the notion that Matthew may have used a sayings source and Luke may have known Matthew and that sayings source. This would have to be detected through painstaking labor which entails analyzing tiny textual remnants. Something similar to this is found in the modern two document hypothesis. The two document hypothesis claims that Mark wrote his gospel first and that Matthew and Luke both had a copy of Mark in front of them when they wrote their gospels. In addition to Mark, both used a now lost sayings Gospel which scholars have affixed the title Q (Quelle in German which means "source"). The triple-tradition material is attributed to Mark and the double-tradition material is attributed to Q. But proponents of this theory have noticed Mark and Q overlaps. There are instances where Mark would appear to have the same content of Q. See the link for more information and examples of Mark and Q overlaps. Some scholars such as E.P. Sanders (SSG) have critiqued the two document hypothesis due to Mark//Q overlaps. If the Mark-Q overlaps become too extensive then one is forced to posit the textual dependence of Mark and Q and this would effectively undercut the rationale for accepting the two document hypothesis over other theories such as the Greisbach hypothesis. Thus the distinctions between sources is not always neat an trimmed. There can be considerable overlap through a range of different channels.
Determining the nature of the textual relationship? John Dominic Crossan devotes a small section to this issue in the Birth of Christianity. He asks, "How can you tell if one gospel is or is not dependent upon another? What happens, for example, if you have twin texts that are too similar in order and / or content for coincidence to serve as an explanation? There are four possible explanations why two texts might be too similar for sheer coincidence. The first two involve a common source (oral or scribal) used independently of one another by later authors. The second two involve literary dependence (direct or indirect) of a later author on an earlier one."2 1. Common Oral Matrix Luke 12:35 (RSV) Let your loins be girded Didache 16:1 Let your lamps be not quenched Crossan commented on this classic example as follows: "The core of the oral matrix is those twin symbols of readiness in no set order. Actualizations must then choose either lamps/loins or loins/lamps as sequence, and either positive ("be") or negative ("be not") as formulation. I prefer that explanation to direct dependence of either text on the other."3 A common literary source is not needed for these passages.
Matthew 3:7-10 and Luke 3:7-9 are extremely similar. Despite divergent introductions, Jesus' saying in both, which encompasses over sixty words in Greek, is verbatim the same in Matthew and Luke (with the exception of four minor changes). As Crossan wrote, "Those twin versions are not the independent actualizations of an oral matrix but the very, very faithful reproductions of a written source."4 Crossan would assign this passage to Q as he is a proponent of the two document hypothesis. That both sayings are verbatim with diverse introductions appears to suggest the two authors obtained the same saying and placed it in a different setting. Other scholars explain the double tradition material by having Matthew dependent upon Luke or as is far more common for Q skeptics, Luke being dependent upon the gospel of Matthew. Luke took the saying from Matthew but gave it his own setting. This latter scenario is entirely plausible. There is no reason against the evangelists being free to move material they borrowed from other sources and place it in their own setting. For example, in Matthew the parable of the lost sheep is directed to the disciples, while in Luke it is told against the Pharisees (Matthew 18:10-14 and Luke 15:4-7). There are also instances where the same incident is placed in different locations in the gospel texts (e.g. Jesus' cleansing cleansing of the temple occurs in the beginning of the gospel according to John).
3. Direct literary Dependence Crossan writes, "But what if some common source, be it oral matrix or written text, is not an adequate explanation? What if one text is dependent on the other? How do you tell when that is so and which one has used the other? . . .There are certainly no easy answers to that question (as if the shorter or longer were always first, or the better-written or worse-written were always first). The process is much more complicated in theory and much more controversial in practice. [Direct literary dependence] must be supported by two mutually supportive arguments: one is genetic relation and the other is redactional confirmation. Genetic relationship means that certain elements of order and content that are characteristically Markan are found in Matthew and Luke We are not talking of general tradition common to all three gospels but of specific editorial aspects of Markan sequence or style whose presence in those other two texts indicates copying."5 Verbatim agreements tell us that some literary relationship must exist. Crossan is speaking of determining the direction of the relationship between two texts once we know there exists one. Crossan goes on to point out the classic example of Markan sandwiches or Markan intercalation. The intercalation is a distinctive Markan compositional device. As Crossan noted, "The device has two elements. First, literary presentation: Event A begins (A1), then Event B begins and finishes (B), and finally, Event A finishes (A2). Second, theological meaning: the purpose of the intercalation is not mere literary show; it presumes that those two events-call them the "framing event" and the "insert event"-are mutually interactive, that they interpret one another to emphasize Mark's theological intention. It is this combination of literary structure and theological import that makes those intercalations peculiarly if not uniquely Markan."6 He presented six (some scholars add more) widely accepted cases of this phenomenon in the gospel of Mark and went on to say that, "What reassures me that the device does not move from either Matthew or Luke into Mark but vice versa is the fact that . . . of Mark's nine sandwiches, Matthew retains Mark's A-B-A pattern five times and Luke retains it four times . . . it is the presence of such specific, personal, or compositional elements, be they order or content, topic or style, that is the surest evidence of the dependence of one text upon another."7 Crossan went on to note that numerous examples of such material need to be checked in case after case so that the argument is finally cumulative. He also stated that there are no absolute conclusions on dependence and nothing is beyond debate. But all scholars must adopt some conclusions regarding dependence or they could not academically study early Christian literature. Crossan also goes on to point out a support of genetic relationship: redactional confirmation. He said, "It is not really a second proof but a way of testing some postulated genetic relationship. If, for reasons such as those given in the preceding example, you postulate Matthean and Lukan dependence on Mark, you should be able to explain every omission, addition, or alteration in Matthew and Luke over their Markan source. Because, of course, we still have Mark"8 To simply mention it as an example, one potential problem with Crossan's specific instance is the possibility that Mark obtained the intercalation from Luke or Matthew and really liked the device so the author utilized it more frequently. The intercalation has to be first shown to be a Marcan device before it can be used to demonstrate the direction of synoptic relations. It is possible to demonstrate this. For example, we would call this a Marcan fingerprint if upon evaluating all the synoptic intercalations we found Mark to be the middle term. Or if we found some textual or grammatical fingerprints left in Matthew and Luke from their stripping of the intercalation if they retained the Markan material.
4. Indirect Literary Dependence A more complicated view is indirect dependence. For example, it is possible that the Gospel of Thomas was not written with any canonical texts in front of it but was very familiar with this canonical material through other sources dependent upon it, oral preaching and so forth. As John Meier put it, the canonical gospels, "inevitably . . . "contaminated" and modified the oral tradition that existed before and alongside themselves."9 Of utmost importance to scholars here would be dating and popularity of the texts. If Thomas dates to 140 A.D. and direct literary dependence is ruled out, indirect dependence would be more likely as the popularity of the canonical Gospels was growing (esp. Matthew). But if Thomas dates to 100 or earlier then it would be much more difficult to imagine that all four canonical Gospels "inevitably" lead to the corruption and modification of Thomas Christianity at a large enough scale for Thomas to be called dependent upon them. The period in between these dates would (presumably) present scholars with a more complex scenario. It is of course very possible and likely for floating oral sayings and independent traditions to be found in the second century texts but the later one goes the more likely dependence (direct or indirect) on the canonical Gospels becomes. There are general pitfalls here. Similar order may simply indicate common knowledge of an event or for a sayings Gospel, a common theme or that both authors shared a common source (oral or written). Order is usually bolstered when in tandem with wording. The two-document hypothesis is a prime example of this. The order and wording is so similar that written dependence (rather than an oral matrix) must be posited.
5. Common Sense
How controversial is this in practice? Scholars disagree on the nature of a number of texts. To highlight some examples of the diverse views found in scholarship today: John is dependent upon the synoptics; John is independent of the synoptics; John is indirectly dependent upon the synoptics; Matthew and Luke used Mark and Q; Matthew and Luke used Mark but Luke used Matthew (not Q), Mark and Luke used Matthew; The Gospel of Thomas is dependent upon the canonical gospels; The gospel of Thomas is independent of the canonical gospels; The gospel of Peter is independent of the canonical gospels; The gospel of Peter is dependent upon the canonical gospels; The gospel of Peter is indirectly dependent upon the canonical gospels; and so on. In the case of the New Testament the majority of exegetes do hold to Markan priority. Most scholars believe Mark was written first and that Matthew and Luke used the text of Mark when writing their own gospels. Within this position there are two distinct and significant schools of thought. The first is the two-document hypothesis view. In addition to using Mark, Matthew and Luke also made use of another written source, now lost, and designated as Q by scholars. The second school of thought accepts Mark without Q and believes that the best explanation of the double-tradition material is found by Luke being dependent upon Matthew. Despite years of argumentation, technological breakthroughs like computers which aid in research and statistical comparisons, this issue is still not resolved. Greisbach proponents and the Mark without Q proponents are significant minorities with top notch historical Jesus and New Testament scholars embracing them. We still have not fully resolved these issues yet though some consensus has emerged. What is problematic is that scholars now believe gospels were fluid at all stages of their development. It is possible to have layers (e.g. Thomas, Q) stages (John, Mark) to have documents combined (e.g. Pauline letters) and to have layers and stages interacting with other layers and stages. There is also the issue to textual harmonization by pious copyists through the years. If something indicates Luke knew Matthew could it not indicate the copyist knew Matthew--the most prominent of the gospels in the early church? If you sit down and comb through all the material at times some of the arguments can be poorly constructed and look rather tenuous. For example, it has been suggested that Mark's poor Greek is an indication that this text was written first. It is supposed to be easier to see how Matthew and Luke would improve on the poorer Greek of Mark. But this is not a good method for determining the direction of synoptic relations. People rewrite material in their own style. It could work both ways. In the second and subsequent centuries, many of the apocryphal gospels have 'worse' Greek than Mark (worse according to the Attic standard). As EP Sanders notes, " Many authors, and no doubt many readers and hearers, preferred more colloquial and less elegant prose." Different prose for different audiences. Other examples are sometimes found when evaluating the gospel of Thomas. Many scholars such as John Meier (11 Marginal v.1) have constructed very poor arguments for its dependence upon the canonical gospels. Meier's position consisted of assuming a second century gnosticism in Thomas which yields a later dating of the gospel and this allowed him to go through and find synoptic parallels in Thomas. By this time the gospels had, as Meier writes, "inevitably . . . "contaminated" and modified the oral tradition that existed before and alongside themselves." The process granted what it needed to actually demonstrate. The methods scholars themselves use can often be very blunt and poorly developed. After listing potential verse showing Matthean redaction in Thomas Meier writes, "only one of the passages I have listed would have to be Matthew's own or reflect Matthew's redaction to prove beyond a doubt that Thomas knows and uses Matthew's gospel to compose his own." Meier has articulated nonsense. If only one redactional element were found out of all the 114 sayings in Thomas, Meier would hardly be justified in concluding that Thomas was dependent upon Matthew. Another possibility given only one or even a few instance would be that Thomas was corrupted at a later date by this canonical intrusion which assimilated the passage to the text of Matthew. This is a well known process. I am not picking on Meier here. I only use him as an example of a famous, intelligent, and well studied historical Jesus scholar who makes blunders when it comes to canonical and non-canonical dependence. Scholars disagree on a lot of views about the literary relationships of various texts and they often do so for ill-advised reasons. I do not say this to dishearten the reader. Lots of good scholarship is out there. But we must be willing to accept the possibility that in some instances we cannot always give an answer. Does the gospel of Thomas date early or late? Did Q exist? Did it have layers? Can we reconstruct them? Was there an earlier version of Mark? We could ask countless questions and all sides of the issue have good arguments and competent scholars promoting them. We may even be able to find satisfactory reasons for holding to certain positions. Yet it is hard for Jesus scholars to offer frequent judgments of non liquet because if they cannot nail down their sources then they cannot discuss a historical Jesus and they would be out of a job. Despite any potential threats of unemployment that might be looming, the evidence should be what leads people to their destination and they should not be afraid to say I don't know when necessary. The synoptic problem remains an unresolved, interesting and lively realm of inquiry.
1.Raymond Brown’s Introduction to the New Testament,
p. 111 |