The Case for the Existence
of Q
The previous quotation from Koester is an opening preamble for his review of evidence that justifies the assumption
of Q’s existence. He follows that statement with three observations that he claims “argue strongly for the existence
of a Synoptic Sayings Source and its use by Matthew and Luke.” The first of them is as follows:
What Matthew and Luke share in addition
to their common Markan pericopes consists almost exclusively of sayings. The only exceptions are: one miracle story (Matt
8:5-13 = Luke 7:1-10), materials about John the Baptist and Jesus’ baptism (parts of Matt. 3:1-17 = Luke 3:2-9, 16-17,
21-22), and the story of Jesus’ temptation (Matt 4:1-11 = Luke 4:1-13). This requires the assumption of a common source
which consisted mostly of sayings of Jesus and probably of some other shared non-Markan materials.
The first portion of this observation is a simple statement of fact—the double tradition does
indeed consist primarily of Jesus’ sayings. However, in the last sentence, Koester makes a leap without justification—the
fact that the double tradition consists mostly of sayings is alleged to “require the assumption of a common source
which consisted mostly of sayings of Jesus.” Logically speaking, it does
not. It requires the assumption that (a) either Matthew or Luke copied from the other, or (b) both copied from a common source.
Koester’s previous statement argues against Luke’s use of Matthew, but remains silent on Matthew’s use of
Luke. Unless one can dismiss the latter as a possibility, one cannot draw a valid conclusion that Q is the only viable solution.
Koester’s second observation
in favor of the existence of Q is this:
The numerous verbal agreements of these parallel passages
cannot be explained as dependence of either Matthew upon Luke or dependence of Luke upon Matthew because in numerous instances
Luke’s version is evidently the more original one. But there are also passages in which Matthew rather than Luke has
preserved words and phrases which cannot be explained as the product of Matthew’s editorial work.
Here Koester acknowledges that Matthew’s potential use of Luke must be dismissed in
order for the 2DH to stand as a valid theory. To argue this point he notes that there are occasional elements in Matthew’s
double tradition material that cannot have been derived from Matthew’s editing of Luke. Once again, the factual observation
is correct but the inference is not. There are indeed instances in which Matthew records a more original form, and Luke appears
to have recorded a later edited version. However Koester’s inference from this that Matthew did not know Luke is another
unjustified leap.
To illustrate, let us return to the parallelisms we just considered in the previous chapter (The
House Built upon the Rock, and Treasures of Heaven, p.119). Matthew’s versions of these sayings are
presented in the form of a structured parallelism, whereas Luke renders them in prose. It is generally assumed that Matthew’s
parallelisms are the earlier forms of the sayings. May we reasonably infer from this that Matthew could not have been using
Luke as a source? Not at all. Matthew may have been aware of both the parallel form of the saying and Luke’s prose rendering
of it, and opted to include the parallel form for any number of reasons. Perhaps he viewed the structured form as more authentic;
perhaps it was a favored tradition within his community; perhaps the poetic structure had more aesthetic appeal to him. It
does not matter. What does matter is that Matthew’s infrequent use of an early form of a saying which has a later counterpart
in Luke indicates nothing about Matthew’s awareness of Luke. For it is certain that Matthew would have known various
forms of numerous sayings, one version appearing in Luke and another in a different oral or written source. That he would
sometimes opt for variants he may have considered earlier, or more authentic, or more popular in his community than those
recorded by Luke is to be expected. The result would be that, on occasion, Matthew would use an earlier form than Luke. Meanwhile,
his direct editing of Luke would create a body of double tradition sayings in which, as Koester describes, “in numerous
instances Luke’s version is evidently the more original one.”
Koester presents this
observation as one which “argues strongly for the existence of a Synoptic Saying Source and its use by Matthew and Luke.”
In reality, in order to sustain this argument one must presuppose the existence of Q, and presuppose that the data uniquely
fit the theory, all the while ignoring a simpler and more direct explanation of the textual phenomenon. It is circular logic.
Koester continues:
Indeed, in some instances what is certainly original in a particular saying
may occur partially in Luke and partially in Matthew. A striking example of this is [Matthew 7:23 = Luke
13:27] The second half of this saying is a quotation of Psalm 6:8. But while the first words of the sentence from this psalm
are accurately preserved only in Luke, the last words of the quotation have an exact parallel only in Matthew. One must assume
that there was a common source used by both authors and that this common source quotes the sentence exactly as it occurred
in Ps. 6:8.
This argument illustrates the extent to which advocates of the Q theory may be inattentive to alternative
explanations for observed data. If one presupposes the existence of Q, one will believe with Koester that Q must have been
the source used by both authors that contained the sentence exactly as it occurred in Psalms 6:8. On the other hand, those
unconvinced of Q’s existence might suppose the common source to have been the Septuagint, which is certain to have been used by both Matthew and Luke. Here we might imagine that Matthew sees the quotation in Luke,
and in cross-referencing a Greek translation of the Psalm finds a preferred way to present the saying. Yet Koester interprets
the citation of Ps 6:8 as evidence that Matthew and Luke must necessarily have drawn from Q.
The third of Koester’s observations for the existence
of Q is this:
The sequence in which certain groups
of sayings occur in the Gospel of Luke often reveals an association and composition of sayings that is more directly related
to the process of the collection of oral materials, while Matthew interrupts or disturbs such sequences whenever his motivations
as an author of literature are evident. In his version of Jesus’ speech for the “sending of the disciples”
(Matt 9:37-11:1), Matthew parallels Luke in the reproduction of a series of sayings which instruct missionaries with respect
to their conduct. But he repeatedly interpolates materials which belong to other contexts and often do not fit the genre of
an older collection of originally oral sayings.
Q/Luke 10:2-12 exhibits all the features
of an early collection of rules for the conduct of the missionary. Its composition most likely took place in the oral transmission
of such regulations, and Q still reflects the loose connection of such a unit of tradition. That Matthew’s text is the
result of a secondary redaction, revealing the use of written sources, is evident in the manner of his composition. He
employed one primary source, i.e., Q, still intact in Luke’s version,
in which he included additional materials which were mostly drawn from his other major source, i.e., Mark, then adding materials
drawn from other contexts of both sources. In the case of Mark, the materials which Matthew used from his collection of rules
about missionaries appear in the same sequence. But in the case of Q, Matthew changed their original order.
Once again, the missing logical element is that which would require
us to presume that Matthew used Q instead of Luke. For there would have been no difference in Matthew’s results had
he drawn from Q, or from the Gospel of Luke in which Q allegedly remains “intact.” The fact is that Koester describes,
though perhaps inadvertently, the process by which Matthew conflated Mark and Luke, reorganizing and editing Luke’s
materials to contextualize them into a Markan framework. Absent a proof that Matthew did not know Luke, this description of
Matthew’s editorial procedure lends no weight to the theory that the authors drew from Q. In circular fashion, the existence
of Q must be presupposed in order to make the case for its existence.
In summary, Koester’s observations that are alleged to argue
strongly for the existence of a Synoptic Sayings Source do not in fact do so. Each of his observations can be interpreted
and resolved within the context of the simpler alternative that Matthew conflated Mark and Luke. Koester is by no means unusual
in this. Many Q advocates exhibit the same tendency to seize upon questionable data and ascribe to them decisive evidentiary
value beyond that which they can sustain under critical examination. Consider this observation by Arland Jacobson which appears
at the beginning of his discussion on the literary unity of Q:
…we may note one small but
very striking example of the distinctive usage of Q over against that of Mark. In Q, the quotation formula, “I say to
you,” never occurs with the word “truly” ….. In fact, the word “truly” does not occur
anywhere in Q; at least, there is no double attestation of it. But Mark has fourteen instances of the “I say to you”
formula, and in all but two he has “truly.” In at least
one case the absence of “truly” is easily explained: it is used in the previous verse. The consistency of Markan
usage is as dramatic as the fact that “truly” is never found in Q, even though it occurs often in Matthew and
six times in Luke. Not only does this illustrate the difference in usage between Mark and Q, but
it is also a potent argument for the Two-document hypothesis.
If one already accepts the 2DH on other grounds, this argument is compelling. It appears on
its surface to be a substantive affirmation of the theory. However, those not yet convinced of Q’s existence will dig
a bit deeper. Is this really a potent argument for the 2DH? Can the observation be explained apart from the existence of Q?
An answer is to be found
in Luke’s redactional use of Mark. Of the thirteen instances of the phrase “Truly,
I say to you” that are found in Mark, only three are duplicated by Luke. In point of fact, Luke does not appear
to favor the use of this phrase. He typically either deletes it or modifies it when he is drawing a saying from a Markan text
in which it appears. There is one striking occurrence to the contrary, in which Luke adds this phrase to a Markan saying where
it does not exist in Mark (Mark 6:4=Luke 4:24). However, Luke most frequently omits truly as part of the quotation
formula, if not the formula altogether.
If the phrase “truly I say to you” was a uniquely popular quotation formula
within the community behind the Markan tradition, we may reasonably surmise that the non-Markan sources Luke drew from may
not have featured it. Furthermore, on those occasions in which it might have appeared in his non-Markan sources, Luke would
have tended to omit it as he often did with his Markan source. Thus, the fact that truly is relatively scarce in
Luke is unremarkable. Moreover, if Matthew drew his double tradition material directly from Luke rather than Q, there is no
mystery in the fact that the non-Markan material he copied over from Luke did not contain the word truly.
This would have produced a double tradition in which truly does not appear. Thus, once
again, we find an observation that is alleged to be a “potent argument” for the 2DH is really no argument at all.
The data can be comfortably explained without any reference to Q.
One of the remarkable aspects of the debate on the Synoptic Problem
is that scholars are very close to recognizing that the theoretical Q Gospel as it is envisioned by the IQP, and the double
tradition as it appears in Luke are practically identical. As Koester notes above, there is general accord among Q theorists
that Luke best preserves the sequence of the Q sayings as they originally appeared in Q, while Matthew liberally reorganizes
the material. Kloppenborg observes:
Matthew has clusters of double tradition materials
that in Luke are scattered, but nonetheless, Matthew presents the sayings in Lukan order, as if he had scanned Q, lifting
out and collecting sayings as he found them in Q.”
Little imagination is required to conclude that Matthew could as easily have scanned Luke directly
to achieve the same result. There is also a broad consensus that Luke far more frequently records an earlier version of a
Q saying than does Matthew. These observations are consistent with the theory that Matthew was drawing upon Luke.
It is interesting to note that if the author
of Matthew used Luke, then by definition he was the final redactor of “Q” as it is defined in modern scholarship.
Matthew’s editorial decisions to reproduce some portions of Luke’s non-Markan text and exclude others would have
defined the boundaries of the double tradition. Under this scenario the “Q document” as reconstructed by the IQP
should have a perceptible Matthean bias. Thus Burton Mack’s observation is intriguing:
If one were to ask which of the narrative gospels most nearly represents an ethos toward which the
community of Q may have tended, it would be the Gospel of Matthew.”
Under the 2DH an ideological affinity between the primitive Q Gospel
and the Gospel of Matthew is nothing more than a coincidence. Under the theory of Matthean posteriority, Matthew defined the
content of the double tradition, and therefore an “ethos” toward the Gospel of Matthew would have been unavoidable.
Traditional Arguments in Support of the 2DH
The observations cited
above do not represent the full complement of arguments that have been put forth in support of the 2DH. Many observations
have been made which have collectively persuaded the majority of scholars of the probability of Q’s existence. These
observations, for the most part, either illustrate the superior resolving power of the 2DH as compared to theories promoting
Luke’s use of Matthew, or argue for the status of Q as a written document rather than an oral source. Since Matthean
posteriority has not been viewed as an alternative, it has gone unnoticed that the textual phenomena advanced in support
of the 2DH can always be resolved by Matthean posteriority. The case for the existence of Q consists of a random assortment
of anecdotal observations that individually and collectively are insufficient to support the theory, for they all beg the
fundamental question of Matthew’s knowledge of Luke. Other than the arguments reviewed thus far, the common arguments
for Q’s existence are these:
Peculiar phrases in common. It has been observed that Matthew and Luke on occasion
use certain phrases in common in the double tradition that constitute unusual grammatical constructs. It is suggested that
both Matthew and Luke reproduced theses phrases from Q. However, it is evident that if Matthew had a propensity to copy unusual
constructs from Q, he would have had an equal propensity to copy them from Luke. Indeed, Matthew tends to reproduce his sources
verbatim unless there is a specific reason to alter them for ideological reasons, or to eliminate extraneous language. If
Luke had incorporated peculiar grammatical constructs that were present in written sources he had before him, or if he had
introduced them himself in the process of transcribing oral traditions, it is not unlikely that Matthew would have taken them
over directly. The result would be a double tradition that featured unusual phrases introduced into the Gospel record by Luke.
Common
sequence of double tradition pericopae. Many of the double tradition sayings appear
in the same sequence in both Gospels. It is often alleged that this indicates the use of a common written source by both evangelists.
However, if both authors achieved a similar sequence of sayings by progressively scanning and copying from Q, then Matthew
would have achieved the same result by progressively scanning and copying from Luke. This argument, like those of frequent
verbal agreements and the presence of peculiar phrases, could be relevant in establishing the “written” nature
of the Q source if and only if the independence of Matthew and Luke could be established on other grounds. Absent such a proof,
these phenomena are as easily explained by both theories.
The presence of the doublets. Do the doublets constitute evidence of Q’s existence? No. Let us first acknowledge that it is not often evident why
the authors chose to duplicate certain sayings. On occasion they may have been duplicated for special emphasis, or to create
a literary frame or set of “parentheses” in the text for interpretive purposes. Some appear to represent two slightly
different forms of the sayings that the authors may have wished to document. Perhaps they wanted to present them in two different
contexts. It is also possible that on occasion, given the vast inventory of raw material, they were simply duplicated in error.
Given the array of sources available to the authors, the simple fact that doublets are present in both Gospels is not particularly
surprising. What is more noteworthy is that both authors duplicate a number of the same sayings. What might be the most logical
accounting of this phenomenon?
Despite
some conceivable rationales, it does not seem probable that both authors would
independently duplicate the same sayings no matter what reason they may have had to do so. The circumstances under
which this may have occurred are worth pondering. The 2DH argues that the two authors, either by design or error, independently
drew common sets of doublets, one each from Mark and Q respectively. Practically speaking, this does not have much intuitive
appeal.
Consider,
then, that Matthew drew upon both Mark and Luke. In this situation, Luke would have created a set of doublets in his own Gospel
by drawing one saying from Mark, and another from one of the many oral and written sources he had at his disposal. Whether
this was done by design or error is immaterial. In either case, this produces an interesting scenario: Matthew was using two
sources in which a number of sayings appeared three times—once in Mark and twice in Luke. If Matthew compiled his Gospel
using these two sources, the odds of duplicating some of the sayings that appeared in triplicate, either intentionally or
erroneously, would have been increased to some degree. Under this scenario it would not be surprising to find that Matthew
would produce a greater array of doublets than does Luke, with the Lukan doublets comprising a subset of those found in Matthew.
Such is indeed the case. Since the phenomenon of the doublets can be explained by assuming Matthew’s direct use of Luke,
it cannot be cited as evidence for the existence of Q.
Luke is never aware of the modifications and expansions
that Matthew makes to his Markan source. This statement of fact is a compelling argument against the proposal
that Luke knew Matthew. Hence, it is among the important observations that favor the 2DH over all theories which argue that
Luke was dependent upon Matthew. However, it has no evidentiary value in weighing the 2DH against Matthean posteriority. Clearly,
if Matthew was the last of the three Synoptics to be composed, Luke would have had no awareness of Matthean expansions to
Mark.
Luke
would have removed all Q sayings from their Markan context in Matthew. Q theorists
point to several phenomena that suggest peculiar editorial behavior on the part of the author of Luke, were he to have relied
upon Matthew as a source. This is one of them. However, this argument only casts doubt upon one direction of potential
dependence between Matthew and Luke, that of Luke upon Matthew. There is no difficulty in imagining that Matthew used
Luke as a source and in so doing, intentionally recast Luke’s non-Markan sayings into Markan contexts. Indeed, based
upon several observations already discussed, this is precisely what he did.
The Gospel of Thomas.
Some scholars have argued that the discovery of Thomas increases the likelihood of Q’s existence as a discrete collection
of sayings. Thomas certainly does illustrate that at least one collection of Jesus’ sayings had been compiled prior
to the end of the second century. However, the common inferences drawn from this fact—that Thomas must represent just
one example of a genre of Jesus literature, that its lack of narrative form and replication of certain double tradition sayings
suggest composition in the mid-first century, that Q must therefore have been another example of the same literary genre—are
not adequately founded.
Two
objections may be made in response to the Thomas argument. First, it is not obvious that Thomas and Q are examples of a common
genre of sayings literature. Thomas consists entirely of sayings that are explicitly attributed to Jesus. Jesus is the supreme
oracle in Thomas; his name appears 105 times in this work. On the other hand, the IQP’s proposed Q text is a miscellaneous
assortment of wisdom, philosophical, and apocalyptic sayings that are, for the most part, attributed to no one. Diverse fragments
of narrative structure appear in Q, whereas there are none in Thomas. Remarkably, the name of Jesus appears only five times
in Q via double attestation. Four of those are in pericopae manifesting evident narrative form (three in the temptation pericope,
and one in the healing of the centurion’s son). Only in one instance is Jesus’ name used in the context of an
explicit saying attribution in a form similar to those found throughout Thomas. Therefore, Q does not portray Jesus as an
inspiring oracular figure in the manner of Thomas. Oddly enough, the name of John the Baptist occurs more frequently in the
IQP’s Q text than does the name of Jesus. To suggest that Q and Thomas are examples of a common genre of Jesus sayings
literature is to mischaracterize the evidence. On this observation alone it is difficult to see any substantive nexus between
Thomas and Q on the grounds of literary genre that would bear upon the probability of Q’s existence.
Second, there is circularity
in the reasoning related to Thomas as well. The document cannot be dated with certainty to the first century. Though its discovery
inspired speculation regarding the existence of a first century “Sayings Gospel genre”, the only potential indicators
of the existence of such a genre are Thomas and Q. Since Thomas may be a second century work and the evidence that Q existed
is tenuous, there is no firm evidence that a Sayings Gospel genre existed during or prior to the composition of the narrative
Gospels. At some point, theory needs to be grounded in factual data. Since the data are not available, the idea that Thomas
sheds some relevant light upon the existence and/or nature of Q is without merit.
Literary unity
in the double tradition. Another of the circular argu-ments used to sustain the
2DH is this: If Q was indeed a discrete Gospel, then the text of Q should bear signs of literary unity; therefore the discovery
of “unities” in the text may in turn be used to confirm Q’s existence. Thus, we find claims such as Jacobson’s,
wherein the absence of the word truly in the double tradition
could be cited as a substantive argument for the 2DH. The logical fallacy in the argument from literary unity is the failure
to ask if there are other means by which evidence of literary cohesion may have come to exist in the double tradition other
than by its alleged origin in Q.
A fundamental observation must be made at this point: The double
tradition is clearly not a random collection of miscellaneous pericopae, regardless of its literary origin. If Q
did not exist, Luke would have compiled his non-Markan materials from an assortment of collected oral traditions and written
notes. The process of selecting, organizing and editing this material would have imparted some degree of ideological and grammatical
form to the material. If Matthew were then to have used Luke as a source, his subsequent scanning, selection, and editing
of the non-Markan materials in Luke would have added further contours and definition to what would eventually be interpreted
as the double tradition. It is not realistic to assume that the double tradition, having gone through two editorial filters
under this scenario, would not manifest any characteristics of literary and ideological unity except those which might be
attributable to an origin in Q.
Thus, Q theorists face a daunting task. In order to cite literary unity in the double tradition as
evidence of the existence of Q, it must be demonstrated that ideological and/or grammatical content is present that cannot
have been the result of the material passing through two successive editorial filters. The “Q document” must be
shown to manifest literary integrity as a work by a single redactor. Unfortunately for Q theorists, the data do not support
this proposition. To the contrary, scholars have found it necessary to posit three successive redactions of the Q document
in order to adequately explain the diversity of its content. Nevertheless, Q theorists persist in the attempt to identify
alleged unities in the text. Jacobson believes he has identified a deuteronomistic motif in the double tradition:
The suggestion is at hand that it is the deuteronomistic tradition which provides the theological
framework for the redaction of Q, and thus is the theological basis for its literary unity.
Kloppenborg attempts to mitigate this claim to some degree while at the same time affirming
Jacobson’s conclusion:
It will be noticed that deuteronomistic influence is in fact restricted to a relatively few passages;
large portions of Q (including 6:20b-49 and 10:2-12, 16, notwithstanding Jacobson’s attempt to label either a “call
to repentance”) lack this motif. Nevertheless, Jacobson’s observations are of utmost significance since they are
coupled with a redaction-critical judgment that this theology dominated one stage of Q redaction. Hence, although he has not
attempted to prove that deuteronomistic theology pervades the whole of Q, Jacobson successfully demonstrates that at one point
in its literary evolution, Q was organized and redacted from a coherent theological perspective. This redaction lends to the
collection an important theological unity.
Here is an example of the Q theorist seeing data through Q-colored
lenses. When Q’s existence is assumed, the data must be resolvable under
the Q hypothesis. If Q existed as a single literary document, one may assume that it was created within a coherent ideological
context. Q theorists thus reach to identify that context, being tethered by the reality of the data at hand. In Kloppenborg’s
statement the tension between observable fact and desired outcome is palpable. The facts as acknowledged by Kloppenborg are
that “large portions of Q lack [the deuteronomistic] motif” and “deuteronomistic
influence is restricted to a relatively few passages.” Q theorists assume that these relatively few passages were
compiled at a discrete redactional stage in Q’s chronological development. Yet, it is the obvious lack of literary unity
in the double tradition that requires the assumption of multiple redactional stages to begin with. From the premise that one
subset of the Q corpus was redacted under the influence of deuteronomistic thought, Kloppenborg makes a breathtaking leap
to the conclusion that this “lends to the collection an important theological
unity.” No, it does not. There is no logical nexus between the data and the conclusion. Even Kloppenborg’s
reference to the double tradition as a collection is rhetorically unjustified
as there is no evidence that it ever existed as a collection until Luke collected it and Matthew defined its boundaries. Yet
to those already convinced of Q’s existence, a literary unity must exist—it cannot be otherwise. Thus, it is simply
declared to exist, despite the paucity of evidence and frailty of the logic.
Within the context of Matthean posteriority, the data that suggest
literary unities can be resolved by assuming that Luke assembled a variety of oral and written traditions that had evolved
independently over time, bearing no literary association with one another until being incorporated into Luke’s Gospel.
Matthew’s subsequent application of, among other things, an “anti-universalist” editorial filter to Luke
resulted in a double tradition that, as a discrete collection of sayings, is somewhat provincial in outlook. It is as remarkable
for what it does not contain as what it does. Though one might claim that the lack of Pauline influence lends the material
a tenuous theological “unity,” the phenomenon is not related to the existence of Q.
