For some time now there has been an unusual informal trans-Atlantic alliance. It is between those who hold to the neo-Griesbach position and those who maintain a version of the legacy of Austin Farrer on the source question. While it is not absolute, the center of gravity of the neo-Griesbach view is in North America, while the Austin Farrer position is strongest in the United Kingdom. In recent years Michael Goulder, albeit with his over inimitable perspectives and accents has emerged as the most vocal partisan on behalf of the Austin Farrer legacy.

There are several key areas where the neo-Griesbachians and Goulder are at odds. Markan priority is the most obvious point of contention. Goulder’s view of the compositional development of the Gospels which essentially negates Matthew and Luke (critically examined) as valid depositories of material for a reconstruction of the Jesus of history is another. But there is one basic area where the interests of the neo-Griesbachians and Goulder coalesce, and which has precipitated this informal alliance. It is the question of Q. Both groups are adamant that the hypothesis of Q is not necessary and, indeed, is a figment of scholarly imagination. Meanwhile, what appears to be the Juggernaut of Q rolls on from victory to victory.¹ Severely undermanned, on both sides of the Atlantic, these two small groups have raised their flags in stout opposition to the advancing forces. The two groups desperately need to stand together. But is this fragile alliance destined to fall apart because of the different ways they come at the Q question and the separate reasons they find for rejecting it? Can this alliance of convenience fine enough common ground not only to put up a good fight against the Q onslaught but, finally, to win the day.

Now, at a very opportune time, comes an outstanding analysis of the contribution of Goulder to New Testament studies. This analysis of Goulder’s work by Mark Goodacre spans Goulder’s academic career enabling us to view the major contours and emphases of his work.

The sub-title of Goodacre’s book² is "An Examination of a New Paradigm." This is well named. Because, at the center of Goulder’s many-sided thrusts into different areas of research (including his recent attempt to revive a version of the F.C. Baur thesis) has been an endeavor to overthrow the standard Two Source paradigm and replace it with one that still accepts Markan priority but rejects Q. Although very different in the direction their respective careers have taken, in this instance, Goulder and W.R. Farmer share something in common. Both have made cutting-edge contributions in many area of theological scholarship over the course of their careers. But at the end of the day, both will ultimately be remembered for their contributions to the source question; Farmer for his rejection of the priority of Mark, and Goulder for his attack on Q.

Essentially Goodacre’s book is divided into three parts.³ Part one deals directly with the Q question by: (1) assessing Goulder’s contention that the characteristic vocabulary of Matthew and the so-called Q of the TSH (Two Source Hypotheses) is essentially the same; and (2) that the Minor Agreements between Matthew and Luke against Mark provide sufficient hard literary evidence in favor of characteristic Matthean terminology appearing in the Lukan parallels to falsify the TSH. Part Two deals with Goulder’s analysis of the non-Markan and non-Matthean material (in terms of his source hypothesis) in Luke. In this analysis of what has traditionally been known among Two Source Theorists as the L material, Goulder comes down hard on the conclusion that this is mainly Lukan composition. Here, Goodacre deftly notes that Goulder not only spends a lot of time refuting the notion of hypothetical written sources but also remains tone deaf to the possibility of oral tradition.⁴ Finally, in Part 3, Goodacre assesses Goulder’s use of the Lectionary Thesis.

In a brief response it will be necessary for us to address Goodacre’s contribution selectively. Consequently, we will not enter into a dialogue with him with respect to his analysis of the Lectionary Thesis. Also, we will deal only in passing with the treatment of the so-called L material. This is a critical area in Lukan studies, and much remains to be done here; but our focus will be on the Q question. Thus the issue of whether Luke used non-Matthean (and non-Markan on Goulder’s terms) material or composed most of it is not directly relevant for the immediate purposes of this paper.

Instead, we will focus directly on the Q question. Procedurally, we will examine closely Goodacre’s analysis of the basic sub-structure of Goulder’s argument that Q is essentially Matthean composition. Then, since Goodacre is so fond of tests, we will put his analysis of Goulder’s linguistic treatment of two key "Q passages" to a test. Our way of conducting the test will be to use the resources of Tevis⁵ and Collison⁶ as determinative factors with respect to what constitutes linguistic characteristics of Matthew and Luke respectively, and we will apply their results rigorously as a check on Goodacre’s results. Our goal will be two-fold. First, although we may be operating from a different source paradigm and using different tools then Goulder and Goodacre, we wish to determine whether our results finally come out at about the same place. Second, we wish to decide whether we can agree that Goulder has truly undermined the case for Q.

In this context, the Minor Agreements are brought on stage. Basically, we are on Goulder’s side with respect to the Minor Agreements. But because of his acceptance of Markan priority, we believe that he has unnecessarily complicated the issue. On a neo-Griesbachian hypothesis the Minor Agreements are a non issue. Luke utilized Matthew as a major source; and Mark edited source material from Matthew and Luke thus leaving a trail of both positive agreements of Luke and Matthew against his text and at the same time, agreements of omission in Matthew and Luke against his text. Since this avenue of focusing on Markan composition as a way of dealing with the Minor Agreements is not open to Goulder he rests his attack on attempting to show that in the text of Luke there are undeniable instances of Matthean linguistic terminology in the positive agreements of the text of Matthew and Luke against Mark. According to Goodacre, Goulder links the first argument with a second line of attack by attempting to show that these same words and phrases in Luke not only are characteristic of Matthew but, also, are uncharacteristic of Luke.⁷ Goodacre deftly shows that it is difficult for Goulder to provide pure instances of this test. Even the famous Minor Agreement at Mk 14:65 fails the strictest version of this test since the language there appears to be closer to Luke than Matthew.⁸ Yet it remains true in this case, and in such other instances as the appearance of Nazara in Mt 4:13//Lk 4:16, that it is difficult to see how a literary connection between Matthew and Luke can be denied. Nevertheless, (even though we perceive the Minor Agreements may be used more effectively than Q, for example, in the whole mess of the Mark-Q overlaps) at this juncture we will not pursue the issue of the Minor Agreements further.

 We are now ready to address directly the thorny methodological issues with respect to determining how Goulder attempts to go about disproving the existence of Q.

 The Definition of Q

A fundamental issue is the definition of Q. Historically, the strict bedrock view of Q is that it is a hypothetical source in Greek which can be reconstructed from the verbatim agreements between Matthew and Luke where there is no parallel pericope in Mark. Of course, there has been notorious slippage in this definition. Regularly in the literature we are treated with everything from the flat equation of Jesus Tradition with Q to Mark-Q overlaps etc. In order for us to be scientifically rigorous, however, we must insist on the bedrock definition. Where does Goulder stand? Does Goodacre hold him accountable in this key area?

Goulder, of course, does not believe in Q. But, apparently for irenic purposes, he regularly uses the terminology. Sometimes this can be confusing to the reader. In volume 1 of Luke: A New Paradigm (LNP) in the Table of Abbreviations Goulder defines Q as "Matter source common to Luke and Matthew."⁹ That is ambiguous enough. But then Goulder makes a key distinction between QC (words in the so-called Q passages common to Luke and Matthew) and QD (words in Q passages differing from Luke and Matthew). It appears that this distinction is primarily descriptive and functional for purposes of discussion with other parties. It is also accompanied by the use of such terminology as "Q sections" as though this were an area upon which all universally agreed. But given the widespread disagreement as to what is the actual text of Q one wonders why Goulder continued to use and even refine this terminology.

Surprisingly, Goodacre registers few difficulties with Goulder’s terminology. Mainly he takes the position that one needs to analyze the work of an author on his own terms and he generally goes along with Goulder’s usages. For Goodacre, the QD passages can be somewhat expansive. Thus, twice in his work Goodacre lists Mt 5:47 as QD.¹⁰ The entire verse of Mt 5:47 has no parallel in Luke. Aside from appearing in a Q passage one may ask why is this Q at all? Evidently both Goulder and Goodacre appear to be defining Q by the scope of an entire pericope not by verbatim agreement between Matthew and Luke. This needs to be tidied up.

This brings us face to face with a key issue; namely, if we abandon the bedrock definition of Q, would it even be possible to settle upon an agreed text for the so-called Q passages let alone determine whether this text was once a separate source? One of the most vulnerable areas of Q research has always been the inability of its advocates to ever agree on what it is.¹¹ It seems that both Goulder and Goodacre may have taken the opportunity to underscore this point more than they did.

 Is Matthew Equivalent to Q?

Goulder asks us to consider the following instance.¹² In the temptation narrative in Mt 4:5, Matthew has the devil take Jesus into the holy city (eiV thn agian polin). Goulder claims, rightly, that this is a linguistic usage of Matthew.¹³ Goulder then goes on to note that the parallel expression in Lk 4:9 is eiV Ierousalhm the special spelling of which is Lukan.¹⁴ The inevitable conclusion is that in situations like this it is impossible to reconstruct the text of Q because one does not know the Vorlage of the text.¹⁵ At best, one can offer as a conjecture a third linguistic possibility or, attempt a reconstruction of Q, based on a determination as to which particular author (Matthew or Luke) has the less characteristic word or phrase; and go with that as the reconstructed text.¹⁶ Such a procedure is patently unpersuasive. Goulder sees this clearly and goes on the attack. The Tendenz of the one wishing to reconstruct Q is to choose vocabulary that is unlike Matthew and Luke. But, Goulder notes, there are a number of instances in pure bedrock ‘Q passages’ where a linguistic characteristic of Matthew is parallel in Luke in the same contextual passage.¹⁷ This would seem to call into question severely the idea of reconstructing Q from words and phrases most unlike those in Matthew and Luke. In Goulder’s view a better assessment of this linguistic evidence would be that either Q and Matthew are the same or that Q’s style is indistinguishable from Matthew.¹⁸ If Goulder can make this case, in our judgment, he has rescued gospel studies from a methodological howler, and in the process has sunk Q.

How would one go about evaluating fairly Goulder’s claims that the essence of the ‘Q Text’ is really the text of Matthew? It is here that Mark Goodacre enters into the picture.

 Goodacre on QC Words and Characteristic Vocabulary

As a basis for evaluating Goulder’s arguments Goodacre sets forth a basic proposal that involves the use of linguistic characteristics.¹⁹ He proposes that one could take the QC words (bedrock Q) and submit them to a basic analysis. The analysis would be to determine the number of Matthean and Lukan characteristic linguistic expressions which appear in bedrock Q (QC). Since, conceptually, the Matthean vocabulary and Q are the same one would necessarily expect to find considerably more Matthean linguistic expressions in this vocabulary than Lukan expressions. For Luke, on this hypothesis, would be copying a basic Vorlage (i.e., the text of Matthew). Goodacre then canvasses the entire "Q sections" (44-83) to determine whether this is so.

Before evaluating Goodacre’s conclusions a crucial issue must be raised as to what determines Goulder’s criteria for the acceptance of a linguistic characteristic of Matthew. In respect to this issue we will make two points and then briefly comment on Goulder’s results.

First, there is the question as to whether Goulder has distorted his formulation and listing of the Matthean linguistic characteristics by his acceptance of Markan priority. Has Goulder engaged in a fair construal of the linguistic data? According to Goodacre, in Midrash and Lection in Matthew (MLM)²⁰ Goulder determined Matthean linguistic characteristics on the basis of a combination of both the so-called Q and M sections of Matthew.²¹ Only if a vocabulary item occurs twice the times in Matthew as it does in Mark is it held to be Matthean.²² This appears to allow the Markan phraseology to hold a veto over what counts for Matthean terminology. Presumably, Goulder’s position still holds for LNP, although the focus is now only on the Matthean characteristics that appear in the bedrock Q (QC) passages.²³ It would seem to be more preferable to isolate the linguistic characteristics of Matthew on grounds that are independent of both Mark and Luke. Pleasingly, Goodacre voices assent to such an idea.²⁴

A second difficulty arises with respect to Goulder’s creation of a new term for the literature: the "semi-Matthean" characteristics. Goulder accepts and uses Hawkins’ list of Matthean linguistic characteristics.²⁵ But Goulder adds the additional category of semi-Matthean words which he claims is an extension of Hawkins’ categories.²⁶ He means by this two things: (1) those vocabulary items which occur twice as often in Matthew than in Mark and those vocabulary items which occur more often in Matthew than in Luke: and (2) those words and phrases which are inserted redactionally by Matthew into an agreed Markan context or an O.T. citation.²⁷

In response, we have already raised questions about the appropriateness of his formulation of the first category. The second category is also problematic for similar reasons: it presumes the correctness of Goulder’s source hypothesis. Goodacre rather reservedly states

Goulder’s criteria are not quite as straightforward as Hawkins, however, and his lists need to be treated with caution. While his vocabularies are supposed to be tools which provide help in the demonstration of his thesis in one respect they take the thesis for granted.²⁸

As noted above, we are puzzled that after pointing out these shortcomings in Goulder’s methodologies, Goodacre still ploughs ahead presuming its validity for the sake of his test.²⁹ Nevertheless, even in terms of using Goulder’s questionable compilation of linguistic characteristics Goodacre’s test produces startling results. The data he compiles (44-83) does not vindicate the contention that the QC vocabular items are in essence Matthean. Far too many Lukan linguistic characteristics emerge. In some cases (Mt 11:2-19//Lk 7:18-35), they even outnumber the Matthean characteristics. Goodacre’s summary speaks for itself

It was stated at the outset that we should not be surprised to find some Lukan vocabulary. If Luke did use Matthew, one would have expected him to have taken over some ‘Luke-pleasing’ words. But clearly to discover a large number of characteristically Lukan words and phrases among QC words would not be satisfactory for Goulder’s thesis. This study has shown that there is a large number of Lukan words and phrases in Q, frequently in individual pericopes as much characteristic of Luke as there is characteristic of Matthew.

So there we have it. According to Goodacre, Goulder has not made his case that the essence of bedrock Q (QC) is the text of Matthew. This is not a negative argument for Q. But it does not undermine Q either. Goulder has made a stout-hearted effort. But the case against Q must be placed on more objective grounds.

In our judgment, Goodacre has already directed us toward the answer: the need to compile an objective list of the characteristic linguistic words and phrases of each synoptic writer not presuming any source hypothesis.³⁰ Then, we could take the bedrock Q (QC) words, check to see the number of Matthean linguistics characteristics vis-a-vis those of Luke. Finally, as a way of determining the direction of literary dependency, we would only accept as impregnable evidence those recurrent words and phrases which are characteristic of author A but are paralleled in the same context by author B where there seems to be evidence of copying; and at the same time, we determine that this word or phrase is not a characteristic of author B and he never uses it in passages that are not paralleled with author A. In these cases it can clearly be determined that author B would be dependent on author A.³¹

Since we are examining the work of an author who likes to conduct tests, in the last section of this paper we will conduct our own test. We will examine two of the ‘Q passages’ that feature prominently in Goodacre’s analysis of Goulder. One (Mt 3:1-12//Lk 3:1-9, 15-17) seems to favor Goulder. The other (Mt 11:2-19//Lk 7:18-35: 16:16) according to Goodacre seems to present the strongest evidence against his position. We will subject these passages to the Zeller text and seek to determine whether the nature of the linguistic indicators is determinative for the direction of literary dependence. Again, it must be remembered that this is a test of some of the evidence--not all of it. Nevertheless, the test may have usefulness as a point of comparison with Goodacre’s results.

 Analysis of Linguistic Characteristics of Mt 3:1-12//Lk 3:1-9; 15-17

Goodacre notes that Hawkins has one vocabulary item that is characteristic of the QC words in Mt 3:1-12. It is sunagw. Tevis has determined that the collocation of sunagw + eiV (thn) apoqhkh in Mt 3:12, 6,26, and 13:30 is characteristic of the text of Matthew.³² Although Tevis is not certain that all these expressions come from the same writer, this collocation provides evidence that Luke is dependent upon the text of Matthew, since only at Lk 3:17//Mt 3:12 is there a parallel in Lk.

A second recurrent linguistic expression is found in Mt 3:7. It is gennhmata ecidnwn + rhetorical questions. This occurs in Mt 3:7, 12:34, and 23:33. Tevis considers that this is the phraseology of one writer.³³ The only parallel in Lk is at Lk 3:7//Mt 3:7. This provides further evidence that Luke is dependent upon Matthew.

A third collocation is o rhqeiV dia Hsaiou tou profhtou legontoV found in Mt 3:3. There are at least twelve occurrences of this phraseology in Matthew.³⁴ A broken form of this appears only in Lk 3:4. The evidence is piling up in favor of Luke’s dependency on Matthew. A fourth collocation is found in Mt 3:10: pan oun dendron mh poioun karpon kalon ekkoptetai kai eiV pur Balletai. Almost verbatim this phrase comes in Mt 7:19.³⁵ It is paralleled in Lk 3:9b. Although not as strong as some of the other indicators, this is additional potential evidence of Luke’s use of Matthew. It may also be linked with the motif of "cutting off" and "casting away" found in Mt 5:30 and 18:8.

On the other hand, the collocation of upodeiknumi with tiV and umin occurs three times in Lk (Lk 3:7; 6:47; 12:5). It occurs only once in Mt at 3:7//Lk 3:7. It is an indicator of Matthew’s dependence upon Luke. However, the use of keimai (Mt 3:10//Lk 3:9) does not, since it is used by Matthew in material that is not paralleled in Luke.

We conclude that, on balance in this unit, where there seems to be evidence of copying, that the direction of literary dependency is strongly in favor of Luke using Matthew. Thus we are in agreement with the conclusions of Goulder (Goodacre), even setting aside most of what are deemed the semi-Matthean or semi-Lukan words.

 Analysis of Linguistic Characteristics of Mt 11:2-19//Lk 17:18-35; 16:16

Goodacre has noted that Mt 11:2-19//Lk 7:18-35; 16:16 provides a good example of a pericope which has more Lukan linguistic characteristics than Matthew in an extended unit of QC words.³⁶ (We would prefer to add to his definition of an extended unit: where there is strong evidence of copying.) This would seem to be an excellent passage to balance the analysis of the previous unit which seems to be weighted strongly in favor of vocabulary items that include Lukan dependence upon Matthew. As with the previous passage we commence with noting the Matthean linguistic characteristics that show up in Luke. Goodacre claims that there are four linguistic items appearing in Mt 11:2-19 that meet Hawkins criteria.³⁷ They are skandalizomai en, outoV estin, omoiow anqrwpoV (anhr) + noun.

First, the use of skandalizomai is noted by Hawkins in his list of Matthean characteristic words and phrases.³⁸ This is verified by Tevis.³⁹ This phraseology is characteristic of Matthew and occurs in Luke in one passage (Lk 7:23) in contextual parallelism with Matthew. It constitutes strong evidence in favor of Lukan dependence upon Matthew.

Second, the use of outoV estin in Lk 9:35 and 20:14, without a parallel in Matthew, means that this phraseology cannot be used as strong evidence of literary dependence of Luke on Matthew even though it meets Hawkins criteria for a characteristic linguistic expression of Matthew. Neither strictly does omoiow and anqrwpoV (anhr) + noun.

On the other hand, the construction hlqen gar IwannhV, which occurs twice in Matthew (11:18; 21:32) and once in a parallel passage in Lk 7:33 does meet our criteria for the direction of dependency being from Matthew to Luke. The only other phraseology that meets the criteria is o nomoV kai oi profhtai. This phraseology occurs in Mt 5:17; 7:12; 11:13; 22:40. It appears only in Lk in the contextual parallel of Lk 16:16.⁴⁰ Upon examination of the list of "semi-Matthean" items the only one that met our criteria was kalamoV (Mt 11:7//Lk 7:24).

We now turn to the listing of characteristic expressions of Luke found in Matthew. Goodacre finds seven words or phrases, in this passage, which meet the Hawkins criteria.⁴¹ However, following the stronger criteria, which we have articulated earlier, we do not accept esqiw + pinw, or dikaiow, or pempw. However, prosdokaw (Mt 11:3//Lk 7:19,20) does meet the Zeller criteria. Likewise, euaggelizomai (Lk 7:22//Mt 11:5) is very strong evidence of Matthew’s dependence on Luke.⁴² Also, prosfwnew (Lk 7:32//Mt 11:16) falls into this category; and filoV, (Lk 7:34//Mt 11:19).

Among Goodacre’s compilation of the "semi-Lukan" words only saleuw (Lk 7:24//Mt 11:7) meets the strict criteria.

In our own analysis, peri ou (Lk 7:27//Mt 11:10) also could be listed as a bedrock Q word, that appears in both Matthew and Luke, and is a linguistic characteristic of Luke.⁴³

 Summary

It is evident in this instance the balance (6 to 4) is mildly in favor of Luke’s dependence upon Matthew. Goodacre is proven to be correct in his assessment that this passage is "problematic for Goulder’s thesis."⁴⁴ But, on the other hand, in our judgment, no firm conclusion can be drawn about the direction of literary dependence from this one passage. As we have noted above, all of the bedrock Q (QC) passages need to be looked at, and a global evaluation of them needs to be made before final conclusions about the direction of literary dependence can be drawn. When Zeller made this global assessment in the nineteenth century he found it came out in balance, in favor of Luke being dependent on Matthew. This data is consistent with the view that Luke used Matthew as his major source and not Q.

Throughout the past decade scholarship on Q has moved in two widely different directions. One group of scholars, a major expression of which are those engaged in the International Q Project, are attempting to consolidate Q research into a virtual industry. They are actively engaged in the production of texts, concordances, and commentaries on Q.⁴⁵ On the other hand the chorus of opposition to the existence of Q, on both sides of the Atlantic, while not exactly a crescendo, is also steadily increasing its volume. This state of affairs will have to be addressed and sorted out by responsible gospel scholarship.

The contribution of Mark S. Goodacre, in assessing the legacy of Michael Goulder, provides an opportune time for us to determine the present status of the strength of the anti-Q argument. Goulder’s central thrust against Q has been to claim that Q’s vocabulary and that of Matthew (compiled according to his source critical presuppositions), is essentially the same. Therefore, one does not need Q. Goodacre claims that Goulder has overstated his case. Analysis of bedrock Q vocabulary does produce more Lukan linguistic characteristics than are warranted by Goulder’s thesis; and the evidence deduced from a study of the Minor Agreements, although leaning in Goulder’s direction, is not absolutely decisive.

In our judgment Goodacre has made a thorough and fair analysis of Goulder’s work. However, there are two areas in which the analysis may have been improved. The first is in the area of the blanket acceptance of Markan priority. This not only has negative implications for Goulder’s work in the Minor Agreements but also skews his assessment of what constitutes an objective list of the linguistic characteristics of Matthew. And second, the acceptance of the compilation of the list of "semi-Matthean" and "semi-Lukan" words was done too easily. This is not only a questionable category, but seems to presume a particular source hypothesis. In other words, Goulder’s work has not been as objective as he claims.

Nevertheless, the test which we ran in conjunction with Goodacre’s assessment of the bedrock Q material reveals that the neo-Griesbachians and those who hold to the Farrer legacy are not that far apart in their united stand against Q. More work needs to be done in the area of what counts objectively as linguistic characteristics and what constitutes decisive evidence in the use of these linguistic characteristics to show definitively the direction of literary dependence. Also, the whole area of Mark-Q overlaps needs to be addressed more aggressively; and finally, much more can be said on the Minor Agreements.

Goulder may not have sunk Q, but his torpedoes have scored several direct hits. He has made it list very badly. It is up to the rest of us to put it out of its misery.

1. The reference to Juggernaut is, of course, an echo of Michael Goulder's, "Is Q a Juggernaut?" JBL 115/4 (1996) 667-681.
2. Mark S. Goodacre, Goulder and the Gospels: An Examination of a New Paradigm. JSNTS Sup 133 (Sheffield: Sheffield Academic Press, 1996).
3. The opening chapter on "Goulder and his Critics" is a scintillating piece that gives an admirable overview of how the Goulder project unfolded. It is replete with analyses of his successes and failures. This respondent has read Goulder for a number of years, but, nevertheless, never saw so clearly that Goulder's various and sundry projects were mainly informed by a grand vision of making viable a version of Christian origins first set forth by the F.L. Baur School. One cannot resist pointing out how ironies abound here. Neo-Griesbachians claim that a factor in the nineteenth century eclipse of their hypotheses was the excesses of the Baur School which utilized Griesbach. Now in the late twentieth century, a strong Markan priortist seeks to revive many of the historical assumptions of a school that used the Griesbach paradigm.
4. Goodacre, "Goulder and the Gospels," 285.
5. Dennis G. Tevis, An Analysis of Words and Phrases Characteristic of the Gospel of Matthew.: Unpublished dissertation SMU 1983 (Ann Arbor, Michigan: UMI Dissertation Series, 1983).
6. J.G. Franklyn Collison, Linguistic Usages in the Gospel of Luke: Unpublished Dissertation SMU 1977 (Ann Arbor, Michigan: UMI Dissertation Series, 1977).
7. Goodacre, "Goulder and the Gospels," 89.
8. Goodacre, "Goulder and the Gospels," 107.
9. Michael Goulder, Luke: A New Paradigm. 2 Vols. JSNTS Sup 10 (Sheffield: JSOT Press, 1989).
10. Goodacre "Goulder and the Gospels," 58, 76.
11. Note the devastating rehearsal of this history which hardly inspires confidence in the venture at all given by E. Earle Ellis, "Gospel Criticism: A Perspective on the State of the Art," in the Gospel and the Gospels. ed. Peter Stuhlmacher (Grand Rapids: W.B. Eerdmans Publishing company, 1991) 35-36. In this article Ellis relies some upon the classic article of S. Petrie "Q is only what you make it," Nov T 3 (1959) 28-33. The title of the article tells all.
12. Goulder, "Is Q a Juggernaut?" 671-672.
13. See Tevis, "Words and Phrases of Matthew" # 333.
14. 2/0/27+36 [=occurrences in Matthew/Mark/Luke & Acts].
15. Of course, minimally by giving us an instance in a parallel text in the same literary context that features the linguistic characteristics of both Matthew and Luke respectively, apropos above, Goulder has shown why it will always be impossible to reconstruct the text of Q, even if on other grounds one knew that it existed. Without the Vorlage how could one ever know the wording of Q at this point?
16. Goulder, "Is Q a Juggernaut?" 671.
17. Ibid.
18. Ibid. 671-672. Goulder goes on to formulate his argument as saying, "that the occurrence of so many rare and striking phrases in Q that are also expressions of Matthew destroys the principal criterion for reconstructing Q."
19. Goodacre, "Goulder and the Gospels," 42-48.
20. M.D. Goulder, Midrash and Lection in Matthew. (London: SPCK, 1974).
21. Goodacre, "Goulder and the Gospels," 42.
22. Ibid., 45.
23. Curiously, as Goodacre notes ("Goulder and the Gospels," 45), Goulder demands a word to appear twice as many times in Matthew over Mark to be Matthean, but he only demands a mere plurality of Matthean words over Luke to be considered Matthean.
24. Goodacre, "Goulder and the Gospels," 47.
25. John C. Hawkins, Horae Synopticae: Contributions to the Study of the Synoptic Problem. (Oxford: Clarendon Press, 2nd Revised edition, 1909). Hawkins gives lists of the characteristic vocabulary of each synoptic writer. his criteria for determining that list were limited to two principles: (1) that a word or phrase occurs four times in a particular gospel writer; and (2) that it does not occur in one of the other synoptics at all or at least as twice as often in a particular gospel as the other two combined.
26. Goulder, Midrash and Lection. 122.
27. Ibid., 122, 476. Also Goulder has done something similar with Luke. Goodacre, "Goulder and the Gospels," 45-46, notes that Goulder does use the terminology 'semi-Lukan' toward the end of MLM but not in LNP. In any case, Goulder treats as Lukan: (1) a vocabulary item that is introduced as redaction into his text; or (2) a vocabulary item that is markedly more frequent in Luke than in Matthew or Mark, and (3) is used in at least three different contexts (See LNP 2, 800). Still, there is some question as to what Goulder means by "markedly."
28. Goodacre, "Goulder and the Gospels," 44.
29. It must be said that, on his own terms, Goulder has compiled an amazing amount of interesting data in his lists of Lucan Vocabulary in LNP 2, 800-809. The neo-Griesbach Research team has used it cautiously but regularly in the work Beyond the Q Impasse: Luke's Use of Matthew. (eds.) Allan J. McNicol with David C. Dungan and David B. Peabody (Valley Forge, Pennsylvania: Trinity Press International, 1996).
30. Goodacre, "Goulder and the Gospels," 47.
31. Essentially, this is a test that is commonly used to determine the direction of literary dependence from one author to another among the neo-Griesbachians. It is discussed at some length by D.B. Peabody, Mark as Composer. NGS 1 (Macon, Georgia: Mercer University Press/ Peeters, 1986) 168-171. Peabody traces this test back to some published work on linguistic characteristics in 1843 by Edward Zeller. Subsequently, we will refer to this in this paper as the Zeller test. Also, one other important methodological point needs to be made. It is understandable that an author, who is copying the work of another, may well select a favorite linguistic expression which he sees in the work of the other author. This factor needs to be taken into consideration. The best way to resolve the matter would be to look at the entire body of material where there seems to be evidence of copying between two authors. Then, most likely, the statistics of the entire body of the text will clearly reveal the direction of literary dependency.
32. Tevis, "Words and Phrases of Matthew " 506.
33. Tevis, "Words and Phrases of Matthew," 219.
34. Tevis, "Words and Phrases of Matthew," 17.
35. Tevis, "Words and Phrases of Matthew," 224, considers it comes from a single writer.
36. Goodacre, "Goulder and the Gospels," 83.
37. Goodacre, "Goulder and the Gospels," 67.
38. Hawkins, "Horae Synopticae," 33.
39. Tevis, "Words and Phrases of Matthew," 307. Tevis formulates it as as skandalizw + en + dative pronoun. He has four usages in Mt 11:6; 13:57; 26:31, 33.
40. It is true it occurs in Lk 24:44 in a passage that is not paralleled with Mt. However, there Luke adds the Psalms. It is generally accepted that this is Lukan terminology for references to scripture.
41. Goodacre, "Goulder and the Gospels," 67.
42. Collision, "Linguistec Usages of Luke," 51.
43. Collision, "Linguistic usages of Luke," 140. The actual structure is peri + a form of autouV. It occurs ten times in Luke and is only paralleled here in Matthew.
44. Goodacre, "Goulder and the Gospels," 68.
45. A good example is the emergence of the first volume of the The Database of the International Q Project entitled Q 11:2b-4 in the series Documenta Q: Reconstructions of Q through Two Centuries of Gospel Research Excerpted, Sorted, and Evaluated. eds. Shawn Carruth and Albrecht Garsky (Leuven: Peeters, 1996).

Allan J. McNicol Institute for Christian Studies

July 10, 1997