In 1974, Eldon Jay Epp, a respected contemporary textual scholar, wrote: "The 'eclectic' method is, in fact, the 20th century method of NT textual criticism, and anyone who criticizes it immediately becomes a self-critic, for we all use it, some of us with a certain measure of reluctance and restraint, others with complete abandon."[1]

Thus, the RSV (Revised Standard Version), NEB (New English Bible) and NIV (New International Version) are confessedly based upon an eclectic text.

The two great translation efforts of these years—RSV and NEB—each chose the Greek text to translate on the basis of the internal evidence of readings. F C. Grant's chapter in the expository pamphlet on the RSV made this clear. The translators, he says, followed two rules: (1) Choose the reading that best fits the context; (2) Choose the reading which explains the origin of the other readings. Professor C. H. Dodd informed me that the British translators also used these two principles—Hort's Intrinsic Probability and Transcriptional Probability. One of the RSV translators while lecturing to the New Testament Club at the University of Chicago replied to a question concerning the Greek text he used by saying that it depended on where he was working: he used Souter at the office and Nestle at home. One of the British translators in admitting the unevenness of the textual quality of the NEB translation explained that the quality depended on the ability of the man who made the first draft-translation of a book.

Whether in early Christian times or today, translators have so often treated the text cavalierly that textual critics should be hardened to it. But much more serious is the prevalence of this same dependence on the internal evidence of readings in learned articles on textual criticism, and in the popularity of manual editions of the Greek New Testament. These latter with their limited citations of variants and witnesses actually reduce the user to reliance upon the internal evidence of readings. The documents which these rigorously abbreviated apparatuses cite cannot lead the user to dependence upon external evidence of documents. These editions use documents (to quote Housman) "as drunkards use lampposts—, not to light them on their way but to dissimulate their instability."[2]

The statement in the preface to the NIV has already been noted: "The Greek text used in the work of translation was an eclectic one."

The introduction to the Greek text put out by the United Bible Societies, pp. x-xi (1966), says:

By means of the letters A, B, C, and D, enclosed within "braces" { } at the beginning of each set of textual variants, the Committee has sought to indicate the relative degree of certainty, arrived at on the basis of internal considerations as well as of external evidence, for the reading adopted as the text. The letter A signifies that the text is virtually certain, while B indicates that there is some degree of doubt. The letter C means that there is a considerable degree of doubt whether the text or the apparatus contains the superior reading, while D shows that there is a very high degree of doubt concerning the reading selected for the text.

A review of their apparatus and its lack of pattern in the correlation between degree of certainty assigned and external evidence makes clear that it is eclectic. In Acts 16:12 they have even incorporated a conjecture! It will be remembered that this text was prepared specifically for the use of Bible translators. The TEV (Today's English Version) is translated directly from it, as is the Version Popular, etc. The text-critical conclusions of G.D. Kilpatrick, a thoroughgoing eclecticist, were finding expression in A Greek-English Diglot for the Use of Translators, issued by the British and Foreign Bible Society. And so on. Enough evidence has been given to show that eclecticism is a major, if not controlling, factor on the textual scene today.

What Is It?

Wherein does "eclecticism" consist? Metzger explains that an eclectic editor "follows now one and now another set of witnesses in accord with what is deemed to be the author's style or the exigencies of transcriptional hazards."[3]

E. C. Colwell[4] spells it out:

Today textual criticism turns for its final validation to the appraisal of individual readings, in a way that involves subjective judgment. The trend has been to emphasize fewer and fewer canons of criticism. Many moderns emphasize only two. These are: 1) that reading is to be preferred which best suits the context, and 2) that reading is to be preferred which best explains the origin of all others.

These two rules are nothing less than concentrated formulas of all that the textual critic must know and bring to bear upon the solution of his problem. The first rule about choosing what suits the context exhorts the student to know the document he is working on so thoroughly that its idioms are his idioms, its ideas as well known as a familiar room. The second rule about choosing what could have caused the other readings requires that the student know everything in Christian history which could lead to the creation of a variant reading. This involves knowledge of institutions, doctrines, and events. . . . This is knowledge of complicated and often conflicting forces and movements.[5]

(What living person really possesses these qualifications? And how can such rules be applied when neither the identity nor circumstances of the originator of a given variant is known?)

More recently Colwell seems to be less enchanted with the method.

The scholars who profess to follow "the Eclectic Method" frequently so define the term as to restrict evidence to the Internal Evidence of Readings. By "eclectic" they mean in fact free choice among readings. This choice in many cases is made solely on the basis of intrinsic probability. The editor chooses that reading which commends itself to him as fitting the context, whether in style, or idea, or contextual reference. Such an editor relegates the manuscripts to the role of supplier of readings. The weight of the manuscript is ignored. Its place in the manuscript tradition is not considered. Thus Kilpatrick argues that certain readings found only in one late Vulgate manuscript should be given the most serious consideration because they are good readings.[6]

J.K. Elliott, a thorough-going eclecticist like Kilpatrick, says of transcriptional probabilities:

By using criteria such as the above the critic may reach a conclusion in discussing textual variants and be able to say which variant is the original reading. However, it is legitimate to ask: can a reading be accepted as genuine if it is supported by only one ms.? There is no reason why an original reading should not have been preserved in only one ms. but obviously a reading can be accepted with greater confidence, when it has stronger support. . . .

Even Aland with his reservation about eclecticism says: "Theoretically the original readings can be hidden in a single ms. thus standing alone against the rest of tradition," and Tasker has a similar comment: "The possibility must be left open that in some cases the true reading may have been preserved in only a few witnesses or even in a single relatively late witness."[7]

Among what Elliott calls "positive advantages of the eclectic method" is the following:

An attempt is made to reach the true or original text. This is, of course, the ultimate aim of any textual critic, but the eclectic method, by using different criteria and by working from a different standpoint, tries to arrive at the true reading, untrammeled by discussion about the weight of ms. support. . . .[8]

No wonder Epp complains:

This kind of "eclecticism" becomes the great leveller—all variants are equals and equally candidates for the original text, regardless of date, residence, lineage, or textual context. In this case, would it not be appropriate to suggest, further, that a few more conjectural readings be added to the available supply of variants on the assumption that they must have existed but have been lost at some point in the history of the textual transmission?[9]

What shall we say of such a method; is it a good thing?

What About It?

An eclecticism based solely on internal considerations is unacceptable for several reasons. It is unreasonable. It ignores the over 5,000 Greek MSS now extant, to say nothing of patristic and versional evidence, except to cull variant readings from them. In Elliott's words, it "tries to arrive at the true reading untrammeled by discussion about the weight of ms. support." It follows that it has no principled basis for rejecting conjectural emendations. It has no history of the transmission of the text. Therefore the choice between variants ultimately depends upon guesswork. This has been recognized by Colwell.

In the last generation we have depreciated external evidence of documents and have appreciated the internal evidence of readings; but we have blithely assumed that we were rejecting "conjectural emendation" if our conjectures were supported by some manuscripts. We need to recognize that the editing of an eclectic text rests upon conjectures.[10]

F.G. Kenyon[11] called conjectural emendation "a process precarious in the extreme and seldom allowing anyone but the guesser to feel confidence in the truth of its results."[12] Although enthusiasts like Elliott think they can restore the original wording of the text in this way, it is clear that the result can have no more authority than that of the scholar(s) involved. Textual criticism ceases to be a science and one is left wondering what is meant by "sound principles" in the NIV preface.

Clark and Epp are right in calling eclecticism a secondary, tentative, and temporary method.[13] As A.F.J. Klijn[14] says, "This method arrives at such varying results that we wonder whether editors of Greek texts and translations can safely follow this road."[15] This procedure seems so unsatisfactory, in fact, that we may reasonably wonder what gave rise to it.

What Is Its Source?

Eclecticism grew out of the Westcott and Hort (hereafter W-H) theory of textual criticism. Epp gives a useful summary statement of that theory, for our immediate purpose:

. . . the grouping of manuscripts led to the separation of the relatively few early manuscripts from the mass of later ones, and eventually the process reached its climactic point of development and its classical statement in the work of Westcott and Hort (1881-1882), and particularly in their (actually, Hort's) clear and firm view of the early history of the NT text. This clear picture was formed from Hort's isolation of essentially three (though he said four) basic textual groups or text-types. On the basis largely of Greek manuscript evidence from the middle of the 4th century and later and from the early versional and patristic evidence, two of these, the so-called Neutral and Western text-types, were regarded as competing texts from about the middle of the 2nd century, while the third, now designated Byzantine, was a later, conflate and polished ecclesiastical text. . . . This left essentially two basic text-types competing in the earliest traceable period of textual transmission, the Western and the Neutral, but this historical reconstruction could not be carried farther so as to reveal—on historical grounds—which of the two was closer to and therefore more likely to represent the original NT text.[16]

. . . the question which faced Westcott-Hort remains for us: Is the original text something nearer to the Neutral or to the Western kind of text? . . . Hort resolved the issue, not on the basis of the history of the text, but in terms of the presumed inner quality of the texts and on grounds of largely subjective judgments of that quality.[17]

Hort, following the "ring of genuineness," preferred the readings of the "Neutral" text-type (today's Alexandrian) and especially those of Codex B, while some subsequent scholars have preferred the readings of the "Western" text-type and of Codex D, on the same basis. Although Hort professed to follow external evidence—and he did in fact follow his "Neutral" text-type, by and large—his prior choice of that text-type was based on internal (subjective) considerations.[18] Still, the general impression was given that the W-H theory was based on external (manuscript and historical) evidence.

But various facets of the theory came under attack soon after it appeared in 1881, and with the conflicting voices came confusion. It is this confusion that has given rise to eclecticism. Thus, Elliott frankly states: "In view of the present dilemma and discussion about the relative merits of individual mss., and of ms. tradition, it is reasonable to depart from a documentary study and to examine the N.T. text from a purely eclectic standpoint."[19] In R.V.G. Tasker's words, "The fluid state of textual criticism today makes the adoption of the eclectic method not only desirable but all but inevitable."[20] Metzger cites dissatisfaction "with the results achieved by weighing the external evidence for variant readings" as the cause.[21] Epp blames "the lack of a definitive theory and history of the early text" and the resultant "chaotic situation in the evaluation of variant readings in the NT text."[22] Colwell also blames "manuscript study without a history."[23] The practice of pure eclecticism seems to imply either despair that the original wording can be recovered on the basis of external evidence, or an unwillingness to undertake the hard work of reconstructing the history of the text, or both.

But most scholars do not practice pure eclecticism—they still work essentially within the W-H framework. Thus, the two most popular manual editions of the Greek text today, Nestle-Aland and UBS (United Bible Societies), really vary little from the W-H text.[24] The recent versions—RSV, NEB, etc.—also vary little from the W-H text.

Why is this? Epp answers:

One response to the fact that our popular critical texts are still so close to that of Westcott-Hort might be that the kind of text arrived at by them and supported so widely by subsequent criticism is in fact and without question the best attainable NT text; yet every textual critic knows that this similarity of text indicates, rather, that we have made little progress in textual theory since Westcott-Hort; that we simply do not know how to make a definitive determination as to what the best text is; that we do not have a clear picture of the transmission and alteration of the text in the first few centuries; and, accordingly, that the Westcott-Hort kind of text has maintained its dominant position largely by default. Gunther Zuntz enforces the point in a slightly different way when he says that "the agreement between our modern editions does not mean that we have recovered the original text. It is due to the simple fact that their editors . . . follow one narrow section of the evidence, namely, the non-Western Old Uncials."[25]

Clark agrees with Zuntz: "All are founded on the same Egyptian recension, and generally reflect the same assumptions of transmission."[26] Clark also gives a sharper focus to one aspect of Epp's answer.

. . . the Westcott-Hort text has become today our textus receptus. We have been freed from the one only to become captivated by the other. . . . The psychological chains so recently broken from our fathers have again been forged upon us, even more strongly. . .

Even the textual specialist finds it difficult to break the habit of evaluating every witness by the norm of this current textus receptus. His mind may have rejected the Westcott-Hort term "neutral," but his technical procedure still reflects the general acceptance of the text. A basic problem today is the technical and psychological factor that the Westcott-Hort text has become our textus receptus. . . .

Psychologically it is now difficult to approach the textual problem with free and independent mind. However great the attainment in the Westcott-Hort text, the further progress we desiderate can be accomplished only when our psychological bonds are broken. Herein lies today's foremost problem with the critical text of the New Testament.[27]

In spite of the prevailing uncertainty and dissatisfaction, when it comes right down to it most textual critics fall back on W-H—when in doubt the safe thing to do is stay with the party line.[28]

Elliott, mentioned earlier, deliberately tried to set the party line aside, and the result is interesting—his reconstruction of the text of the Pastoral Epistles differs from the Textus Receptus 160 times, differs from W-H 80 times, and contains 65 readings that have not appeared in any other printed edition. A review of his reasoning suggests that he did not altogether escape the psychological grip of W-H, but the result is still significantly different from anything else that has been done.[29]

Elliott's effort underscores, by contrast, the extent to which UBS, NEB, etc. still hew to the W-H line. To really understand what is going on today we must have a clear perception of the W-H critical theory and its implications. Its importance is universally recognized.[30] J.H. Greenlee's statement is representative. "The textual theory of W-H underlies virtually all subsequent work in NT textual criticism."[31]

So, to a discussion of that theory I now turn.

[1]E.J. Epp, "The Twentieth Century Interlude in New Testament Textual Criticism," Journal of Biblical Literature, XCIII (1974), p. 403.

[2]E.C. Colwell, "Hort Redivivus: A Plea and a Program," Studies in Methodology in Textual Criticism of the New Testament, E.C. Colwell (Leiden: E.J. Brill, 1969), pp. 152-53. Tasker records the principles followed by the NEB translators: "The Text to be translated will of necessity be eclectic, . . ." (p. vii).

[3]Metzger, The Text, pp. 175-76.

[4]The late Ernest Cadman Colwell might well have been described as the dean of New Testament textual criticism in North America during the 1950s and 1960s. He was associated with the University of Chicago for many years as Professor and President. Some of his important articles have been collected and reprinted in Studies in Methodology in Textual Criticism of the New Testament.

[5]Colwell, "Biblical Criticism," pp. 4-5. For words to the same effect see also K. Lake, The Text of the New Testament, sixth edition revised by Silva New (London: Rivingtons, 1959), p. 10 and Metzger, The Text, pp. 216-17.

[6]Colwell, "Hort Redivivus," p. 154. Cf. pp. 149-54.

[7]J.K. Elliott, The Greek Text of the Epistles to Timothy and Titus, ed., Jacob Geerlings, Studies and Documents, XXXVI (Salt Lake City: University of Utah Press, 1968), pp. 10-11. Cf. K. Aland, "The Significance of the Papyri for Progress in New Testament Research," The Bible in Modern Scholarship, ed. J.P. Hyatt (New York: Abingdon Press, 1965), p. 340, and Tasker, p. viii.

[8]Elliott, p. 11.

[9]Epp, p. 404.

[10]Colwell, "Scribal Habits in Early Papyri: A Study in the Corruption of the Text," The Bible in Modern Scholarship, ed. J.P. Hyatt (New York: Abingdon Press, 1965), pp. 371-72.

[11]Frederick G. Kenyon was an outstanding British scholar during the first half of this century. He was Director and Principal Librarian of the British Museum and his Handbook to the Textual Criticism of the New Testament is still a standard textbook.

[12]F.G. Kenyon, Handbook to the Textual Criticism of the New Testament, 2nd ed., 1926, p. 3.

[13]Epp, pp. 403-4. Cf. K.W. Clark, "The Effect of Recent Textual Criticism upon New Testament Studies," The Background of the New Testament and its Eschatology, ed. W.D. Davies and D. Daube (Cambridge: The Cambridge University Press, 1956), p. 37. In a paper presented at the forty-sixth annual meeting of the Evangelical Theological Society (Nov., 1994), Maurice A. Robinson reinforces the serious deficiency that "neither 'reasoned' nor 'rigorous' eclecticism offers a consistent history of textual transmission. . . ." (p. 30). The seriousness of this deficiency may be seen from the results. UBS3, a confessedly eclectic text, repeatedly serves up a patchwork quilt. For example, in Matthew there are at least 34 places where its precise rendering is not to be found, as such, in any single extant Greek MS (cf. R.J. Swanson, The Horizontal Line Synopsis of the Gospels, Greek Edition, Volume I. The Gospel of Matthew [Dillsboro, NC: Western North Carolina Press, 1982]).

[14]A.F.J. Klijn is a well-known textual scholar who has specialized in the study of the "Western text-type."

[15]A.F.J. Klijn, A Survey of the Researches into the Western Text of the Gospels and Acts; part two 1949-1969 (Leiden: E.J. Brill, 1969), p. 65.

[16]Epp, pp. 391-92.

[17]Ibid., pp. 398-99.

[18]Metzger states that "Westcott and Hort's criticism is subjective." The Text, p. 138. See also Colwell, Studies in Methodology in Textual Criticism of the New Testament (Leiden: E.J. Brill, 1969), pp. 1-2.

[19]Elliott, pp. 5-6.

[20]Tasker, p. vii.

[21]Metzger, The Text, p. 175.

[22]Epp, p. 403.

[23]Colwell, "Hort Redivivus," p. 149.

[24]See K.W. Clark, "Today's Problems with the Critical Text of the New Testament," Transitions in Biblical Scholarship, ed. J.C.R. Rylaarsdam (Chicago: The University of Chicago Press, 1968), pp. 159-60, for facts and figures. Also see Epp, pp. 388-90. G.D. Fee has charged that my treatment of eclecticism is "hopelessly confused" ("A Critique of W. N. Pickering's The Identity of the New Testament Text: A Review Article," The Westminster Theological Journal, XLI [Spring, 1979], p. 400). He feels that I have not adequately distinguished between "rigorous" (my "pure") and "reasoned" eclecticism and have thereby given a distorted view of the latter. Well, he himself says of the reasoned eclecticism which he espouses, "Such eclecticism recognizes that W-H's view of things was essentially correct, . . (Ibid., p. 402)." My statement is, "But most scholars do not practice pure eclecticism—they still work essentially within the W-H framework (p. 28)." Are the two statements really that different?

The fairness of this assessment may be illustrated from the works of both Fee and Metzger (whom Fee considers to be a practitioner of reasoned eclecticism). In his "Rigorous or Reasoned Eclecticism—Which?" (Studies in New Testament Language and Text, ed. J.K. Elliott [Leiden: Brill, 1976]), Fee says: "Rational eclecticism agrees in principle that no MS or group of MSS has a prima facie priority to the original text (p. 179)." But on the next page he says of Hort: "if his evaluation of B as 'neutral' was too high a regard for that MS, it does not alter his judgment that compared to all other MSS B is a superior witness." Metzger says on the one hand, "the only proper methodology is to examine the evidence for each variant impartially, with no predilections for or against any one type of text (Chapters, p. 39)," but on the other hand, "readings which are supported by only Koine, or Byzantine witnesses (Hort's Syrian group) may be set aside as almost certainly secondary (The Text, p. 212)."

But Fee has more to say. "An even greater error [than my 'distortion' discussed above] is for him to argue that Elliott's method is under 'the psychological grip of W-H' (p. 29) ("A Critique," p. 401)." He goes on to explain that Elliott and W-H are on opposite ends of the internal/external evidence spectrum because "it is well known that W-H gave an extraordinary amount of weight to external evidence, just as do Pickering and Hodges (Ibid.)." And yet, on another occasion Fee himself wrote: "it must be remembered that Hort did not use genealogy in order to discover the original NT text. Whether justified or not, Hort used genealogy solely to dispense with the Syrian (Byzantine) text. Once he has [sic] eliminated the Byzantines from serious consideration, his preference for the Neutral (Egyptian) MSS was based strictly on intrinsic and transcriptional probability [emphasis Fee's] ("Rigorous," p. 177)." And again: "In fact the very internal considerations for which Kilpatrick and Elliott argue as a basis for the recovery of the original text, Hort used first [emphasis Fee's] for the evaluation of the existing witnesses (Ibid., p. 179)."

It seems to me that these latter statements by Fee are clearly correct. Since Hort's preference for B and the "Neutral" text-type was based "strictly" on internal considerations, his subsequent use of that text-type cannot reasonably be called an appeal to external evidence. In sum, I see no essential difference between "rigorous" and "reasoned" eclecticism since the preference given to certain MSS and types by the "reasoned" eclecticists is itself derived from internal evidence, the same considerations employed by the "rigorous" eclecticists. I deny the validity of "eclectic method" in whatever guise as a means for determining the identity of the NT Text. (I do agree with Z.C. Hodges, however, that any and all Traditional Text readings can be defended in terms of internal considerations, should one wish to.)

[25]Epp, 390-91. Cf. G. Zuntz, p. 8. Epp reinforces an earlier statement by Aland: "It is clear that the situation with which our present day method of establishing the New Testament text confronts us is most unsatisfactory. It is not at all the case that, as some seem to think, everything has been done in this field and we can for practical purposes rest satisfied with the text in use. On the contrary, the decisive task still lies ahead." "The Present Position of New Testament Textual Criticism," Studia Evangelica, ed. F.L. Cross and others (Berlin: Akademie—Verlag, 1959), p. 731.

[26]Clark, "Today's Problems," p. 159.

[27]Ibid., pp. 158-60. Cf. M.M. Parvis, "Text, NT.," The Interpreter's Dictionary of the Bible (4 Vols.; New York: Abingdon Press, 1962), IV, 602, and D.W. Riddle, "Fifty Years of New Testament Scholarship," The Journal of Bible and Religion, X (1942), 139.

[28]Cf. Clark, "Today's Problems," p. 166, and especially Colwell, "Scribal Habits," pp. 170-71.

[29]Elliott's results are interesting in a further way. He does his reconstruction "untrammeled" by considerations of manuscript support and then traces the performance of the principal manuscripts. Summarizing his statement of the results, considering only those places where there was variation, Codex Aleph was right 38% of the time, A was right 38% of the time, C right 41%, D right 35%, F,G right 31%, and the bulk of the minuscules (Byzantine) was right 35% of the time (pp. 241-43). He claims that doing a reconstruction his way then enables one to trace the behavior of individual MSS and to show their "illogical fluctuations." Such a tracing is based upon his own subjective evaluation of readings but the illogical fluctuations can be seen empirically by comparing the collations of a variety of MSS.

[30]See, for example, K. Aland, "The Significance of the Papyri," p. 325; Colwell, "Scribal Habits," p. 370; Metzger, The Text, p. 137; V. Taylor, The Text of the New Testament (New York: St. Martin's Press Inc., 1961), p. 49; K. Lake, p. 67; F.G. Kenyon, Handbook to the Textual Criticism of the New Testament (2nd ed.; Grand Rapids: Wm. B. Eerdmans Publishing Co., 1951), p. 294; Epp, "Interlude," p. 386, and Riddle, Parvis and Clark, noted above (fn. 27).

[31]J.H. Greenlee, Introduction to New Testament Textual Criticism (Grand Rapids: Wm. B. Eerdmans Publishing Co., 1964), p. 78.

Previous Chapter
Table of Contents
Next Chapter