Some Second Thoughts on the Majority Text
Related MediaEditor's Note1
In his engaging volume, The Interpretation of the New Testament 1861-1961, Neill remarks, "In historical research there are few axioms; and it is good that periodically every alleged conclusion should be challenged and tested in the light of fresh evidence, or of a change in the premisses [sic] on the basis of which the evidence is weighed." 2 He was speaking of the Synoptic problem, but his words may justifiably be applied to the field of New Testament textual criticism today--at least in the United States.
In the last decade a handful of scholars has risen in protest of textual criticism as normally practiced. In 1977 Pickering advocated that the wording of the New Testament autographs was faithfully represented in the majority of extant Greek manuscripts. 3 This view had been argued in one form or another since John W. Burgon in 1883 sought to dismantle single-handedly the Westcott-Hort theory. 4
What was new, however, with Pickering's approach was perhaps a combination of things: his theological invectives were subdued
Page 271
(especially compared with those of the Textus Receptus-advocating fundamentalist pamphleteers); his theological presuppositions regarding preservation were also played down; his treatment appeared sane, reasonable, and thorough; and he was a graduate of Dallas Theological Seminary. This last point is of no small significance, for in the last several years some if not most of the leading advocates of the majority text view have received their theological training at Dallas Seminary. 5
In 1978 Gordon Fee mounted a frontal attack on the majority text view, especially as articulated by Zane Hodges. 6 The Journal of the Evangelical Theological Society then staged a battle between Fee and Hodges, in which Hodges wrote "Modern Textual Criticism and the Majority Text: A Response," 7 to which Fee responded with "Modern Textual Criticism and the Majority Text: A Rejoinder," 8 to which Hodges responded with "Modern Textual Criticism and the Majority Text: A Surrejoinder." 9 Fee and Hodges have continued to interact with each other's views elsewhere. Most notably, Hodges wrote on the authenticity of John 5:4, 10 to which Fee responded by writing "On the Inauthenticity of John 5:3b-4." 11
Then in 1982 theory was applied to practice. The Greek New Testament according to the Majority Text, edited by Zane Hodges and Arthur Farstad--with the help of Wilbur Pickering and
Page 272
others--was published. Though Hodges and Farstad wanted to publish the text on the centennial anniversary of Westcott-Hort's Greek New Testament (i.e., 1981 12 ), so as to emphasize the difference between the two approaches, 13 the missed conjuncture had no appreciable diminishing effect. A spate of book reviews followed--including a rather lengthy one by Gordon Fee. 14
To be sure, other majority text advocates have published in the last five years, 15 but the focal point of the debate has justifiably been on Hodges and Farstad's book itself. The fact that the second edition of this text was published in 1985 perhaps shows the growing popularity of the textual theory that stands behind it. 16
Page 273
This historical survey raises the question, What issues could be raised that have not already been discussed in detail? 17 Two things, at least: first, the Dallas Seminary-majority text view connection has not been addressed; second, the majority text theory, as it has been displayed concretely in the Majority Text itself, has hardly been noticed. 18 This article addresses these issues as well as a few other points that have been (relatively) neglected.
Dallas Seminary and the Majority Text
As already mentioned, some if not most of the leading advocates of the majority text view are alumni of Dallas Seminary. An inference that has been drawn from this in the evangelical community at large is that Dallas Seminary is monolithic and provincial in its views of textual criticism. However, no faculty member in the New Testament Studies department at the present time embraces the majority text theory of textual criticism. This is because Hodges is no longer on the faculty (he taught from 1959 to 1986). In any case, his view has always been a minority view among the Dallas faculty.
Now this is not pointed out as an argument against the majority text theory. Neither the majority of professors nor the majority of
Page 274
manuscripts is in itself any kind of argument at all. 19 But this is mentioned because those in the evangelical community who are interested in what the faculty of Dallas Seminary are teaching need to be aware that what individual faculty members advocate in print is not necessarily representative of what other faculty members--even the majority of them--embrace.
Some Reflections on the Majority Text: Text and Theory
There are three reasons for considering here the majority text theory as worked out in the Hodges-Farstad Majority Text.
First, before publishing the Majority Text, Hodges was inextricably linked with the advocates of the Textus Receptus. 20 As
Page 275
recently as 1978 Hodges's view was misunderstood by no less a scholar than Fee, who asked, "If they [i.e., Hodges et al.] really mean majority rule, are they ready to give up the TR at such non-superficial variants as Acts 8:37 and 1 John 5:7-8 (where a weak minority of Greek MSS supports the TR)?" 21 In fact even since the Majority Text was printed, Hodges's view has occasionally been confused with a return to the Textus Receptus in toto. 22
Second, previous judgments about the character of the Byzantine text-type can now easily be examined. The Majority Text has facilitated testing of the hypothesis that this text-type is a fuller, smoother, and more conflate text than the Alexandrian text-type or the text of the modern critical editions (i.e., UBS3 [=NA26]).
Third, the second principle of Hodges's theory--that a reconstructed family tree will vindicate the authenticity of the majority text's readings--can also be tested, at least in the pericope adulterae (John 7:53-8:11) and in the Apocalypse (the two places where the Majority Text reflects readings based on stemmatics).
Because of these considerations, this discussion will be restricted for the most part to an interaction with Hodges's text-critical method and results. 23
Page 276
Majority Text Versus The Textus Receptus
In 1977 Pickering predicted that "the Textus Receptus will be found to differ from the Original [=Majority Text] in something over a thousand places, most of them being very minor differences, whereas the critical texts will be found to differ from the Original in some five thousand places, many of them being serious differences." 24
There is much to criticize in the way this prediction is stated; 25 nevertheless the quantitative aspect of Pickering's guess is on the mark. In this writer's examination of Hodges and Farstad's Majority Text 26 he has counted 1,838 differences between it and the Textus Receptus. 27 This is indeed "something over a thousand" differences! Most notably the Majority Text excluded Acts 8:37 and the Comma Johanneum (the Textus Receptus's rendering of 1 John 5:7-8 with its Trinitarian formula). As well, in the last six verses of Revelation, which Erasmus had to translate into Greek from Latin, there are 17 differences between the Majority Text and the Textus Receptus.
The fact of almost 2,000 differences between these two texts, many of them quite significant, is a two-edged sword. On the one hand it should be rather disconcerting to Textus Receptus advocates who have been depending on Hodges's scholarship for some time. On
Page 277
the other hand it cries out for a fresh look, by New Testament students, at the Byzantine text-type, which has been seen only through a glass darkly in the printed editions of the Textus Receptus.
Majority Text Versus Critical Text (I.E., Ubs 3 [=Na26]) 28
To be sure, the Majority Text stands much closer to the Textus Receptus than it does to the critical text. According to this writer's count there are 6,577 differences between the Majority Text and the critical text. But that does not tell the whole story.
Textual variants are customarily placed in one of four categories: omission, addition, substitution, and transposition. The general character of the Byzantine text-type is normally described as smooth, conflated, harmonistic, complete. 29 Therefore one would expect it especially to imbibe in the error of addition. That is, since it is an allegedly later form of text, it must have adapted and adopted earlier traditions. But of the 6,577 differences between the Majority Text and the critical texts, in only 1,589 places is the Majority Text longer than the critical. This is less than one-fourth of the total differences. 30
Further, the Majority Text is sometimes shorter than the critical text. Though this is generally acknowledged, it is severely downplayed--by both friend and foe. Hort, for example, suggests that while "interpolations . . . are abundant," "omissions . . . are rare." 31 Metzger, in a suggestive study on the parallels between the textual criticism of the New Testament on the one hand, and The Iliad and The Mahabharata on the other, quotes with approbation Franklin Edgerton, one of the editors of The Mahabharata:
I have come to believe that any passage, long or short, which is missing in any recension or important group of manuscripts as a whole, must be very seriously suspected of being a secondary insertion . . . probably not one of the some fifty MSS. [that] I have studied for Book 2, nor any of their genealogical ancestors, ever deliberately or intentionally omitted a single line of the text. . . . It appears that no scribe, no redactor, ever knowingly sacrificed a single line which he found in his original. 32
Page 278
Metzger draws the parallel for New Testament textual criticism that the rule that the shorter reading is to be preferred (brevior lectio praeferenda est) is generally sound and that by this canon the Byzantine text-type, in being long, comes up short. 33
On the other side Pickering argues against the canon of the shorter reading. 34 He concludes that "the 'fullness' of the Traditional Text, rather than a proof of inferiority, emerges as a point in its favor." 35 Not once does he suggest that the shorter reading is at times to be preferred or that the Byzantine text-type contains shorter readings. Hodges takes a more cautious approach, saying that one must be agnostic about the principles of internal criticism at the present time. 36 Yet he cites but two studies, both of which are used to demonstrate the invalidity of brevior lectio. 37 The impression one gets, though never explicitly stated, is that the critical text will rarely if ever have a longer reading than the majority text, and the majority text will rarely if ever have a shorter one. 38
Indeed, the battle line almost seems to be drawn at this issue. But what is the evidence? In this writer's count, there are 657 places where the Majority Text is shorter than the critical. Obviously one cannot both invoke (or reject) this canon mechanically and maintain an equally mechanical preference for a given text-type.
Yet raw statistics can be tantalizingly deceptive. If the Byzantine text's "additions" are frequently due to harmonization or
Page 279
conflation, 39 while the Alexandrian text (which usually, though not always, stands behind the critical text) "adds" an article here or a pronoun there (which could easily drop out via homoioteleuton or for stylistic reasons in the Byzantine tradition), then the significance of these statistics is greatly altered. 40
Hodges Versus Hodges: Inherent Contradictions
One inconsistency has already been mentioned which applies to majority text advocates in general--as well as, to some degree, to reasoned eclecticists. That inconsistency is that too dogmatic an appeal to the superiority of shorter or longer readings in toto actually softens the dogmatic appeal to a preferred text-type.
In addition there seem to be four areas of inherent contradiction to Hodges's general theory. Before looking at them it may be helpful to examine the first principle of his method, namely, that mathematical statistics are in some way relevant to the supposition of "majority rule."
Hodges and Farstad say, "(1) Any reading overwhelmingly attested by the manuscript tradition is more likely to be original than its rival(s). . . . (2) Final decisions about readings ought to be made on the basis of a reconstruction of their history in the manuscript tradition." 41 Elsewhere Hodges adds, "Under normal circumstances the older a text is than its rivals the greater are its chances to survive in a plurality or a majority of the texts extant at any subsequent
Page 280
period." 42 He then shows, through the mathematical calculations of his brother, David Hodges, 43 that in any generation of normal transmission, the reading of the autograph will survive in the majority of manuscripts. In a hypothetical genealogy, three copies are made directly from the original. Two of them are good, one bad. Unfortunately the last generation listed on his diagram seems to contradict his thesis, for there are 13 good copies and 14 bad ones! 44
Perhaps this is why David Hodges adds, "A one-third probability of error is rather high, if careful workmanship is involved." 45 After making such adjustments, he argues, "Consequently, the conclusion is that, given the conditions described, it is highly unlikely that the erroneous reading would predominate to the extent that the majority text predominates." 46 An integral part of David Hodges's calculations is the supposition that the correct reading can arise from a faulty reading just as easily as a faulty reading can arise from the correct one. 47
This statistical demonstration has four basic problems 48 (either irrelevancies or inconsistencies): (1) "Reading" and "text" are confused, 49 giving a distorted picture of how the statistics are applied
Page 281
to the New Testament. 50 (2) The statistics are relevant only for potential variation of a singular nature (e.g., transposition of one number for another), and hence they cannot accurately be applied to the problem of New Testament textual criticism. 51 (3) The supposition that a good reading can arise from a bad reading just as easily as the reverse does not take into account the theological-literary nature of the New Testament. 52 (4) If it were true that a good reading could
Page 282
easily arise from a bad one, the result would be a hopeless labyrinth from which a stemmatic reconstruction (Hodges's second principle of textual criticism) could not possibly be accomplished. 53
Majority versus genealogy. Though the statistical demonstration is certainly Hodges's best-known argument for the supremacy of the majority text, it is not his last word. His coup de grce, as it were, is not a theoretical genealogy, but applied stemmatics, for he says, "Final decisions about readings ought to be made on the basis of a reconstruction of their history in the manuscript tradition." 54
Though not explicitly stated, since Hodges has been vigorous to defend both the majority text and the genealogical method, 55 the distinct impression arises that he is convinced that a reconstructed family tree will vindicate majority rule. In other words his second principle should validate his first.
As he has applied stemmatics only to the pericope adulterae and the Apocalypse, we can test this "validation" in only these places. For John 7:53-8:11 Hodges has constructed a family tree of the extant Greek manuscripts, using von Soden's data. He sees the manuscripts as belonging to seven subgroups, according to their distinctive readings. 56 A group of approximately 250 manuscripts, given the label M6, "is viewed as the original form of the pericope from which all other groups are descended." 57
Page 283
A look at the textual apparatus here reveals a startling fact: Of the 30 textual problems listed, the editors, on the basis of their stemmatic reconstruction, have adopted at least 15 readings supported by a minority of manuscripts. 58 In other words for the pericope adulterae, the Majority Text, in half its readings, is a minority text.
One might object, however, that every reading adopted by the editors appears within the majority text, even if it is not the predominant reading of that text-type. But this would be something of a bait-and-switch response: Does the majority text mean to Hodges a text-type per se, or does it mean the majority of manuscripts? This writer's distinct impression is that Hodges would not regard what others call the Byzantine text a text-type at all. 59 If so, then in no sense do these 15 minority text readings represent the majority text as Hodges uses the term. Of course they are representative of the Byzantine text--and Hodges has done an invaluable service by providing a provocative stemma that apparently traces the pericope back to its roots in the Byzantine tradition. 60
Admittedly 15 minority text readings in a volume called The Greek New Testament according to the Majority Text may seem hardly significant (even if these 15 variants do comprise half the readings in the test passage). Yet it must be remembered that the pericope adulterae is one of only two places where Hodges has applied his stemmatic principle--the principle he believed would vindicate majority rule. Nevertheless the stemmatic method should be given a full hearing. In the Apocalypse, where stemmatics have been applied for 22 chapters, one can see more clearly how well stemmatics have vindicated the majority text.
Hodges's stemma for Revelation is based on two magisterial pillars: the complete manuscript evidence collated by Hoskier 61 and
Page 284
the stemmatics of Schmid. 62 Hoskier supplied the raw data (the readings of the manuscripts), and Schmid interpreted the data by grouping the manuscripts into genealogically related families.
In the Hodges-Farstad Majority Text, Schmid's groups (slightly modified) are cited, but not the individual manuscripts that compose them. One can get a rather artificial impression then as to the number of manuscripts supporting each variant. Consequently Hoskier must be consulted to see where individual manuscripts line up.
This writer's comparison of Hoskier with the Majority Text has revealed 152 minority text readings that have been adopted by the editors of the Majority Text. 63 This is 15 percent of all textual problems. 64 Thus over 150 times in the Apocalypse, Hodges and Farstad have invalidated their own first principle of textual theory, that original readings will survive in the majority of manuscripts. 65
Page 285
Such minority text readings demonstrate Hodges's deep integrity; he has not altered the stemmatic evidence to save the majority's neck. In fact in many respects Hodges is moving toward a critical text and away from a purely majority text as he practices his genealogical method. On this score it seems ironic that the leading majority text advocate has produced a text that is undermining the majority text school.
Though it would be too much to label his genealogical principle the "Hodgian fallacy," 66 it must be recognized that the more the stemmatics principle is applied to the Hodges-Farstad text, the less it will deserve the name Majority Text. Ultimately it would, in the interest of truth, need to be called something like "The Intra-Byzantine Stemmatics Greek New Testament."
Normal rate of copying versus stemmatics. What stands behind the "majority rule" principle (as well as the statistical demonstration) is the idea of a normal rate of copying: "The manuscript tradition of an ancient book will, under any but the most exceptional conditions, multiply in a reasonably regular fashion with the result
Page 286
that the copies nearest the autograph will normally have the largest number of descendants." 67 Hodges's stemmatic reconstruction not only contradicts "majority rule," as well as coming into conflict with the statistical demonstration, 68 but also seems to mitigate the "normal rate of copying" hypothesis. This last point is so because in order to reconstruct a family tree by placing hundreds of extant manuscripts into less than a dozen groups, 69 one would have to posit, it seems, that concentrated copying was done in particular places (such as scriptoria) and particular times. Hodges's argument that a single hypothetical examplar stands behind one or more groups of manuscripts 70 is no different from Hort's argument that a single archetype stood behind the Byzantine text-type. As such, it destroys any notion of a normal rate of copying.
Majority versus majority. 71 Kilpatrick remarks, "Hodges' and Farstad's view must explain two features, first that there is no evidence for Hort's Syrian text before the fourth century, and second that the dominant text of the second and third centuries is so different." 72 The fact of no early Byzantine manuscripts is a well-worn issue. 73 Nevertheless three important questions are rarely brought into the discussion. First, why is it that not only are there no early Byzantine manuscripts (i.e., before the late fourth century), but also the Byzantine text-type, as far as the extant manuscripts demonstrate, did not become the majority until the ninth century? 74 Does this not indicate that the principle of "majority rule" changes from century to century? 75 Second, why do majority text advocates count only Greek
Page 287
manuscripts? Is it because inclusion of the Latin Vulgate, for example, with more than 8,000 extant copies (compared to less than 5,400 Greek manuscripts)--and a text-form closer to the critical text than to the majority text--would demolish their theory? 76 Third, what would happen to the majority text theory if a cache of thousands of New Testament manuscripts--whose textual affinities were different from the Byzantine text-type--were to be discovered? Could the majority text view survive the blow of a "Greek Ebla"? 77 Far from achieving certainty about the wording of the autographs, the so-called "majority text" seems to be built on shifting sands.
Genealogical method ultimately dependent on internal criteria. Hodges inveighs against the canons of internal criticism, speaking of them as "very broad generalizations about scribal habits," and arguing that "all such generalizations tend to cancel each other out." 78 Hodges attacks Hort on the grounds that "Hort's study of manuscript history and his investigation of documents is predicated above all on the internal evidence of readings!" 79 Hodges concludes by stating:
Modern textual criticism is psychologically "addicted" to Westcott and Hort. Westcott and Hort, in turn, were rationalists in their approach to the textual problem[s] in the New Testament and employed techniques within which rationalism and every other kind of bias are free to operate. The result of it all is a methodological quagmire where objective controls on the conclusions of critics are nearly nonexistent. It goes without saying that no Bible-believing Christian who is willing to extend the implications of his faith to textual matters can have the slightest grounds for confidence in contemporary critical texts. 80
It is well known that Hort's "approach to the textual problems in the New Testament" was the genealogical approach. That is,
Page 288
he reconstructed the relationship of the text-types according to readings contained in them. His view of the superiority of a and B squarely rested on his own judgments as to the superiority of their readings. 81 This approach, Hodges maintains, is the result of rationalism.
Thus the majority text advocates--Hodges included--prefer a more objective approach, one based on external rather than internal evidence. This is the motive behind Hodges's first principle of "majority rule." Yet Hodges's second principle of stemmatics--on which "final decisions about readings ought to be made" 82 --is none other than the genealogical method. Kilpatrick points out that in this regard "the two editors are more rigorous than Hort." 83 Even Hodges and Farstad admit that the genealogical method "remains the only logical one. If Westcott and Hort employed it poorly, it is not for that reason to be abandoned." 84
But perhaps Hodges's genealogical method is more objective than Hort's. Let us again hear what he says about it:
A valid stemma must have the power to explain the descent of the readings in a natural way. Each hypothesized intermediate archetype must show itself to be the starting point of more than one reading which appears below it on the stemma, but not above. . . . Moreover, the readings found high on the stemma should quite easily be seen as the natural progenitors of readings lower down which developed from them. In particular there ought to be some readings treated as original which are noticeably superior to their rivals. 85
One might ask, On what basis are the readings judged to be superior? And how does this differ from Hort's dictum, "Where then one of the documents is found habitually to contain these morally certain or at least strongly preferred readings, and the other habitually to contain their rejected rivals, we can have no doubt . . . that the text of the first has been transmitted in comparative purity, and that the text of the second has suffered comparatively large corruption"? 86
If Hodges responds that Hort never really applied the genealogical method to individual manuscripts (a point which is quite
Page 289
true), 87 this does not thereby obviate the problem that Hodges's genealogical method is still founded on the subjectivity of internal criteria. In his insightful study on the genealogical method, Colwell lists several problems with this approach: "It is doubtful if it can be applied to New Testament manuscripts in such a way as to advance our knowledge of the original text of the New Testament"; 88 "genealogical method can trace the tree down to the last two branches, but it can never unite these last two in the main trunk--it can never take the last step"; 89 "when there is mixture, and Westcott and Hort state that it is common, in fact almost universal to some degree, then the genealogical method as applied to manuscripts is useless"; 90 "there is no unmixed text in existence, nor any manuscript with an unmixed text"; 91 "in a field where no manuscripts have parents, where centuries and continents separate witnesses, the genealogical method is not of primary importance." 92 These statements show that the genealogical method is hardly objective, especially when it is applied to specific manuscripts (as in Hodges's approach). 93 At bottom, then, Hodges's stemmatic reconstruction is inextricably tied to the subjectivity of internal criteria which he so adamantly condemns.
Summary
Three major points were made in this article: (1) The Majority Text differs from the Textus Receptus in almost 2,000 places, suggesting that the Byzantine text-type has been seen only through a glass darkly in the printed editions of the Textus Receptus. (2) The
Page 290
Majority Text, differing from the critical text in over 6,500 places, has over 650 readings shorter than the critical text; such readings call out for an exhaustive evaluation. (3) In "Hodges versus Hodges" five points were noted: (a) The statistical demonstration of majority rule for the New Testament transmissional history, though ingenious, seemed to be irrelevant for it did not deal with the phenomenon of a literary document. (b) Hodges's second principle of stemmatics, as applied in the pericope adulterae and in Revelation, overturned, in large measure, his principle of "majority rule" (thus rendering The Greek New Testament according to the Majority Text something of a misnomer). (c) Hodges's reconstructed family tree also contradicts the "normal rate of copying" canon for it seems to imply abnormal (i.e., heavy) copying in particular places and at particular times. (d) The "majority rule" principle does not take into account the majority of Greek manuscripts in the first eight centuries, nor the versions, nor any future cache of manuscripts. (e) The genealogical method (Hodges's final vindication of "majority rule") ultimately depends on internal criteria and as such vitiates any statements about an objective method. 94
Footnotes
1 This essay is a revised and abridged version of a paper entitled "Some Reflections on the Majority Text," available for a small fee in the Dallas Theological Seminary Book Room. The longer version contains more extensive technical data.3 Wilbur N. Pickering, The Identity of the New Testament Text (Nashville: Thomas Nelson, 1977).
4 John W. Burgon, The Revision Revised (London: J. Murray, 1883).
10 "The Angel at Bethesda--John 5:4," Bibliotheca Sacra 136 (January-March 1979): 25-39.
11 Evangelical Quarterly 54 (1982): 207-18.
13 Personal conversation with Zane Hodges in 1980.
21 Fee, "Modern Textual Criticism and the Revival of the Textus Receptus," p. 23.
24 Pickering, The Identity of the New Testament Text, p. 177.
30 And hence the category of "additions" is actually smaller than the average category of variation.
31 Westcott and Hort, Introduction [and] Appendix, p. 135.
34 Pickering, The Identity of the New Testament Text, pp. 79-83.
37 Hodges, "A Defense of the Majority-Text," pp. 16-17.
41 Hodges and Farstad, eds., Majority Text, pp. xi-xii.
42 Hodges, "A Defense of the Majority-Text," p. 4.
54 Hodges and Farstad, eds., Majority Text, p. xii.
64 There are 986 textual problems listed in the Majority Text 's Revelation.
67 Hodges, "The Greek Text of the King James Version," p. 344.
71 Harold W. Hoehner of Dallas Seminary is to be credited with the seminal form of this section.
72 G. D. Kilpatrick, review of Majority Text, in Novum Testamentum 26 (1984): 85-86.
78 "A Defense of the Majority-Text," p. 16.
81 Westcott and Hort, Introduction [and] Appendix, pp. 19-72, 90-145.
82 Hodges and Farstad, eds., Majority Text, p. xii.
83 Kilpatrick, review of Majority Text, p. 86.
84 Hodges and Farstad, eds., Majority Text, p. xii.
87 Cf. Colwell, "Genealogical Method: Its Achievements and Its Limitations," pp. 109-10, 112.
Copyright 1997 by Dallas Theological Seminary and Galaxie Software. If you are interested in owning the Theological Journal Library on CD-ROM, visit the Galaxie Software Home Page. |
Related Topics: Textual Criticism