4

AN EVALUATION OF THE W-H THEORY

The Basic Approach

Should the New Testament be treated just like any other book? Will the procedures used on the works of Homer or Aristotle suffice? If both God and Satan had an intense interest in the fate of the New Testament text, presumably not. But how can we test the fact or extent of supernatural intervention?  Happily we have eyewitness accounts to provide at least a partial answer. Hort said that "there are no signs of deliberate falsification of the text for dogmatic purposes," but the early Church Fathers disagree. Metzger states:

Irenaeus, Clement of Alexandria, Tertullian, Eusebius, and many other Church Fathers accused the heretics of corrupting the Scriptures in order to have support for their special views. In the mid-second century, Marcion expunged his copies of the Gospel according to Luke of all references to the Jewish background of Jesus. Tatian's Harmony of the Gospels contains several textual alterations which lent support to ascetic or encratite views.[1]

Gaius, an orthodox Father who wrote between A.D. 175 and 200, names Asclepiades, Theodotus, Hermophilus, and Apollonides as heretics who prepared corrupted copies of the Scriptures and who had disciples who multiplied copies of their fabrications.[2]

Surely Hort knew the words of Origen.

Nowadays, as is evident, there is a great diversity between the various manuscripts, either through the negligence of certain copyists, or the perverse audacity shown by some in correcting the text, or through the fault of those, who, playing the part of correctors, lengthen or shorten it as they please (In Matth. tom. XV, 14; P. G. XIII, 1293).[3]

Even the orthodox were capable of changing a reading for dogmatic reasons. Epiphanius states (ii.3b) that the orthodox deleted "he wept" from Luke 19:41 out of jealousy for the Lord's divinity.[4]

Subsequent scholarship has tended to recognize Hort's mistake. Colwell has done an instructive about-face.

The majority of the variant readings in the New Testament were created for theological or dogmatic reasons.

Most of the manuals and handbooks now in print (including mine!) will tell you that these variations were the fruit of careless treatment which was possible because the books of the New Testament had not yet attained a strong position as "Bible." The reverse is the case. It was because they were the religious treasure of the church that they were changed.[5]

The New Testament copies differ widely in nature of errors from copies of the classics. The percentage of variations due to error in copies of the classics is large. In the manuscripts of the New Testament most variations, I believe, were made deliberately.[6]

Matthew Black says flatly:

The difference between sacred writings in constant popular and ecclesiastical use and the work of a classical author has never been sufficiently emphasized in the textual criticism of the New Testament. Principles valid for the textual restoration of Plato or Aristotle cannot be applied to sacred texts such as the Gospels (or the Pauline Epistles). We cannot assume that it is possible by a sifting of 'scribal errors' to arrive at the prototype or autograph text of the Biblical writer.[7]

H.H. Oliver gives a good summary of the shift of recent scholarship away from Hort's position in this matter.[8]

The fact of deliberate, and apparently numerous, alterations in the early years of textual history is a considerable inconvenience to Hort's theory for two reasons:  it introduces an unpredictable variable which the canons of internal evidence cannot handle, and it puts the recovery of the Original beyond reach of the genealogical method.[9]

To illustrate the second point, Hort's view of early textual history may be represented by figure A whereas the view suggested by the Church Fathers may be represented by figure B. The dotted lines in figure B represent the fabrications introduced by different heretics (as the early Fathers called them).         

              Original                                                   Original

                    O                                                                 O

 


              o                  o                   o                            o                  o                   o

 


       o     o     o    o     o     o     o     o    o                  o    o    o      o    o   o        o     o       o

 


    o    o  o  o    o  o  o    o  o   o   o   o  o   o           o   o  o  o   o   o  o  o  o   o   o     o     o    o 

                                            Figure A                                                                       Figure B

 

            Genealogy cannot arbitrate the conflicting claims posed by the first line of descendants in Figure B.[10] Further, in Colwell's words, this method (genealogy)

rested on identity in error as the clue to common ancestry.  These errors were unintentional changes which can be identified objectively as error. Agreement in readings of this kind seldom occurs by chance or coincidence. The New Testament copies differ widely from copies of the classics at this point. The percentage of variations due to error in copies of the classics is large. In the manuscripts of the New Testament, on the other hand, scholars now believe that most variations were made deliberately.[11]

The reconstruction of family trees is seriously complicated by the presence of deliberate alterations. And those are not the only difficulties under which genealogy labors.

Genealogy

We have already noted Hort's definition and supposed use of genealogy. However, scholars have so far isolated only a few  parent-child sets among all 5,000 plus manuscripts.[12] How then did Hort go about plotting the genealogical descent of the extant MSS? M.M. Parvis answers: "Westcott and Hort never applied the genealogical method to the NT MSS, . . ."[13]  Colwell agrees.

That Westcott and Hort did not apply this method to the manuscripts of the New Testament is obvious. Where are the charts which start with the majority of late manuscripts and climb back through diminishing generations of ancestors to the Neutral and Western texts? The answer is that they are nowhere. Look again at the first diagram, and you will see that a, b, c, etc. are not actual manuscripts of the New Testament, but hypothetical manuscripts. The demonstrations or illustrations of the genealogical method as applied to New Testament manuscripts by the followers of Hort, the "Horticuli" as Lake called them, likewise use hypothe-tical manuscripts, not actual codices. Note, for example, the diagrams and discussions in Kenyon's most popular work on textual criticism, including the most recent edition. All the manuscripts referred to are imaginary manuscripts, and the later of these charts was printed sixty years after Hort.[14]

How then could Hort speak of only "occasional ambiguities in the evidence for the genealogical relations,"[15] or say: “So far as genealogical relations are discovered with perfect certainty, the textual results which follow from them are perfectly certain, too, being directly involved in historical facts; and any apparent presumptions against them suggested by other methods are mere guesses against knowledge”[16] when he had not demonstrated the existence of any such relations, much less with "perfect certainty"?

Another challenge to genealogy is "mixture."

The second limitation upon the application of the genealogical method to the manuscripts of the New Testament springs from the almost universal presence of mixture in these manuscripts. . . .

The genealogical diagram printed above (p. 110) from Westcott and Hort shows what happens when there is no mixture. When there is mixture, and Westcott and Hort state that it is common, in fact almost universal in some degree, then the genealogical method as applied to manuscripts is useless.

Without mixture a family tree is an ordinary tree-trunk with its branches—standing on the branches with the single trunk—the original text—at the top. The higher up—or the further back—you go from the mass of late manuscripts, the fewer ancestors you have!

With mixture you reverse this in any series of generations. The number of possible combinations defies computation, let alone the drawing of diagrams.[17]

Other scholars have agreed that the genealogical method has never been applied to the New Testament, and they state further that it cannot be applied. Thus, Zuntz says it is "inapplicable,[18] Vaganay that it is "useless,"[19] and Aland that it "cannot be applied to the NT."[20] Colwell also declares emphatically "that it cannot be so applied."[21] In the light of all this, what are we to think of Hort when he asserts:

For skepticism as to the possibility of obtaining a trustworthy genealogical interpretation of documentary phenomena in the New Testament there is, we are persuaded, no justification either in antecedent probability or in experience. . . . Whatever may be the ambiguity of the whole evidence in particular passages, the general course of future criticism must be shaped by the happy circumstance that the fourth century has bequeathed to us two MSS of which even the less incorrupt must have been of exceptional purity among its own contemporaries.[22]?

After demolishing the genealogical method, Colwell concludes his article by saying, "yet Westcott and Hort's genealogical method slew the Textus Receptus. The a priori demonstration is logically irrefutable."[23] However, the a priori demonstration cannot stand in the face of an a posteriori demonstration to the contrary. Colwell himself, some twelve years prior to this statement, recognized that the "a priori demonstration" to which he here refers has been refuted.

The universal and ruthless dominance of the middle ages by one texttype is now recognized as a myth. . . .

The complexities and perplexities of the medieval text have been brought forcibly to our attention by the work of two great scholars: Hermann von Soden and Kirsopp Lake. . . .

This invaluable pioneer work of von Soden greatly weakened the dogma of the dominance of a homogenous Syrian text. But the fallacy received its death blow at the hands of Professor Lake. In an excursus published with his study of the Caesarean text of Mark, he annihilated the theory that the middle ages were ruled by a single recension which attained a high degree of uniformity.[24]

Actually, Hort produced no "demonstration" at all­—just assumptions. Since the genealogical method has not been applied to the MSS of the New Testament it may not be used as an integral part of a theory of NT textual criticism. If it was Hort's genealogical method that "slew the Textus Receptus" then the TR must still be alive and well—the weapon was never used. But Hort claimed to have used it, and the weapon was so fearsome, and he spoke of the "results" with such confidence, that he won the day.

Since Westcott and Hort, the genealogical method has been the canonical method of restoring the original text of the books of the New Testament. It dominates the handbooks. Sir Frederic Kenyon, C.R. Gregory, Alexander Souter, and A.T Robertson are a few of the many who declare its excellence.[25]

The situation is essentially the same today, and the warning Colwell gave in 1965 is still valid.

Many years ago I joined others in pointing out the limitations in Hort's use of genealogy, and the inapplicability of genealogical method—strictly defined—to the textual criticism of the NT. Since then many others have assented to this criticism, and the building of family trees is only rarely attempted. Therefore we might assume that the influence of Hort's emphasis upon genealogical method is no longer a threat. But this assumption is false.

Hort`s brilliant work still captivates our minds. So when confronted by a reading whose support is minimal and widely divorced in time and place, we think first and only of genealogical relationships. Hort has put genealogical blinders on our eyes. . . .[26]

Present-day scholars, exegetes, and translators continue to act as though the genealogical method not only can be, but has been, applied to the NT MSS, and to base their work on the supposed results.  But what about those "results"?

Text-types and Recensions

Although Hort claimed absolute certainty for the results of genealogical evidence as described by him, it is clear that the "results" were a fabrication. How could there be results if the method was never applied to the MSS? A contemporary of W-H protested that such claims would only be allowable if the textual critic had first indexed every principal Church Father and reduced MSS to families by a laborious process of induction.[27]

Still, Hort's "results" became accepted as fact by many—George Salmon spoke of "the servility with which his [Hort] history of the text has been accepted, and even his nomenclature adopted, as if now the last word had been said on the subject of New Testament criticism. . . ."[28]

Subsequent scholarship

Subsequent scholars have been obliged to reconsider the matter by the discovery of the Papyri and closer looks at MSS previously extant.  Parvis complains:

We have reconstructed text-types and families and sub families and in so doing have created things that never before existed on earth or in heaven. We have assumed that manuscripts reproduced themselves according to the Mendelian law. But when we have found that a particular manuscript would not fit into any of our nicely constructed schemes, we have thrown up our hands and said that it contained a mixed text.[29]

Allen Wikgren shows that sweeping generalizations about text-types in general and the "Byzantine" text and Lectionaries in particular, should not be made.[30] Colwell affirms:

The major mistake is made in thinking of the "old text-types" as frozen blocks, even after admitting that no one manuscript is a perfect witness to any text-type. If no one MS is a perfect witness to any type, then all witnesses are mixed in ancestry (or individually corrupted, and thus parents of mixture).[31]

After careful study of P46, Zuntz makes certain observations and concludes:

One would like to think that observations like these must put an end to time-honoured doctrines such as that the text of B is the 'Neutral' text or that the 'Western' text is 'the' text of the second century. If the factors of each of these equations are meant to be anything but synonyms, they are wrong; if they are synonyms, they mean nothing.[32]

Klijn doubts "whether any grouping of manuscripts gives satisfactory results,"[33] and goes on to say:

It is still customary to divide manuscripts into the four well-known families: the Alexandrian, the Caesarean, the Western and the Byzantine.

This classical division can no longer be maintained. . . .

If any progress is to be expected in textual criticism we have to get rid of the division into local texts. New manuscripts must not be allotted to a geographically limited area but to their place in the history of the text.[34]

After a long discussion of the "Caesarean" text, Metzger says by way of summary that "it must be acknowledged that at present the Caesarean text is disintegrating."[35] Two pages later, referring to the impact of P45, he asks, "Was there a fundamental flaw in the previous investigation which tolerated so erroneous a grouping?" Evidently there was. Could it be the mentality that insists upon thinking in terms of text-types and recensions as recognized and recognizable entities?[36] Those few men who have done extensive collations of manuscripts, or paid attention to those done by others, as a rule have not accepted such erroneous groupings.[37]

H. C. Hoskier, whose collations of NT MSS are unsurpassed in quality and perhaps in quantity, commented as follows after collating Codex 604 (today's 700) and comparing it with other MSS:

I defy anyone, after having carefully perused the foregoing lists, and after having noted the almost incomprehensible combinations and permutations of both the uncial and cursive manuscripts, to go back to the teaching of Dr. Hort with any degree of confidence. How useless and superfluous to talk of Evan. 604 having a large "Western element," or of its siding in many places with the "neutral text." The whole question of families and recensions is thus brought prominently before the eye, and with space one could largely comment upon the deeply interesting combinations which thus present themselves to the critic. But do let us realize that we are in the infancy of this part of the science, and not imagine that we have successfully laid certain immutable foundation stones, and can safely continue to build thereon. It is not so, and much, if not all, of these foundations must be demolished.[38]

The "text-types" themselves

To take the "text-types" one by one, Kenyon says of the "Western" text:

What we have called the d-text, indeed, is not so much a text as a congeries of various readings, not descending from any one archetype, but possessing an infinitely complicated and intricate parentage. No one manuscript can be taken as even approximately representing the d-text, if by "text" we mean a form of the Gospel which once existed in a single manuscript.[39]

Colwell observes that the Nestle text (25th edition) denies the existence of the "Western" text as an identifiable group, saying it is "a denial with which I agree."[40] Speaking of von Soden's classification of the "Western" text, Metzger says: "so diverse are the textual phenomena that von Soden was compelled to posit seventeen sub-groups of witnesses which are more or less closely related to this text."[41] And Klijn, speaking of "a 'pure' or 'original' Western Text" affirms that "such a text did not exist."[42] K. and B. Aland speak of “the phantom ‘Western text’” and replace it with “D text”, referring to Codex Bezae.[43] In fact, it has been many decades since any critical apparatus used a cover symbol for the so-called “Western” text.

As for today's "Alexandrian" text, which seems essentially to include Hort's "Neutral" and "Alexandrian," Colwell offers the results of an interesting experiment.

After a careful study of all alleged Beta Text-type witnesses in the first chapter of Mark, six Greek manuscripts emerged as primary witnesses: À B L 33 892 2427. Therefore, the weaker Beta manuscripts C D 157 517 579 1241 and 1342 were set aside. Then on the basis of the six primary witnesses an 'average' or mean text was reconstructed including all the readings supported by the majority of the primary witnesses. Even on this restricted basis the amount of variation recorded in the apparatus was dismaying. In this first chapter, each of the six witnesses differed from the 'average' Beta Text-type as follows: L, nineteen times (Westcott and Hort, twenty-one times); Aleph, twenty-six times; 2427, thirty-two times; 33, thirty-three times; B, thirty-four times; and 892, forty-one times. These results show convincingly that any attempt to reconstruct an archetype of the Beta Text-type on a quantitative basis is doomed to failure. The text thus reconstructed is not reconstructed but constructed; it is an artificial entity that never existed.[44]

Hoskier, after filling 450 pages with a detailed and careful discussion of the errors in Codex B and another 400 on the idiosyncrasies of Codex À, affirms that in the Gospels alone these two MSS differ well over 3,000 times, which number does not include minor errors such as spelling, nor variants between certain synonyms which might be due to "provincial exchange."[45]  In fact, on the basis of Colwell's suggestion that a 70% agreement be required so as to assign two MSS to the same text-type, Aleph and B do not qualify. The UBS and Nestle texts no longer use a cover symbol for the "Alexandrian" text-type.

Of the "Byzantine" text, Zuntz says that "the great bulk of Byzantine manuscripts defies all attempts to group them."[46] Clark says much the same.

The main conclusion regarding the Byzantine text is that it was extremely fluid. Any single manuscript may be expected to show a score of shifting affinities. Yet within the variety and confusion, a few textual types have been distinguished. . . . These types are not closely grouped like the families, but are like the broad Milky Way including many members within a general affinity.[47]

Colwell's emphatic statement to the same effect has been given above. The work of Lake referred to by Colwell was a collation of Mark, chapter eleven, in all the MSS of Mt. Sinai, Patmos, and the Patriarchal Library and collection of St. Saba at Jerusalem.  Lake, with R. P. Blake and Silva New, found that the "Byzantine" text was not homogeneous, that there was an absence of close relationship between MSS, but that there was less variation "within the family" than would be found in a similar treatment of "Neutral" or "Caesarean" texts. In their own words:

This collation covers three of the great ancient collections of MSS; and these are not modern conglomerations, brought together from all directions. Many of the MSS, now at Sinai, Patmos, and Jerusalem, must be copies written in the scriptoria of these monasteries. We expected to find that a collation covering all the MSS in each library would show many cases of direct copying.  But there are practically no such cases. . . . Moreover, the amount of direct genealogy which has been detected in extant codices is almost negligible. Nor are many known MSS sister codices. The Ferrar group and family 1 are the only reported cases of the repeated copying of a single archetype, and even for the Ferrar group there were probably two archetypes rather than one. . . .

There are cognate groups—families of distant cousins—but the manuscripts which we have are almost all orphan children without brothers or sisters.

Taking this fact into consideration along with the negative result of our collation of MSS at Sinai, Patmos, and Jerusalem, it is hard to resist the conclusion that the scribes usually destroyed their exemplars when they had copied the sacred books.[48]

J.W. Burgon,[49] because he had himself collated numerous minuscule MSS, had remarked the same thing years before Lake.

Now those many MSS were executed demonstrably at different times in different countries. They bear signs in their many hundreds of representing the entire area of the Church, except where versions were used instead of copies in the original Greek. . . .  And yet, of multitudes of them that survive, hardly any have been copied from any of the rest. On the contrary, they are discovered to differ among themselves in countless unimportant particulars; and every here and there single copies exhibit idiosyncrasies which are altogether startling and extraordinary. There has therefore demonstrably been no collusion—no assimilation to an arbitrary standard—no wholesale fraud. It is certain that every one of them represents a MS, or a pedigree of MSS, older than itself; and it is but fair to suppose that it exercises such representation with tolerable accuracy.[50]

Kurt Aland[51] sums it up:

P66 confirmed the observations already made in connection with the Chester Beatty papyri. With P75 new ground has been opened to us. Earlier, we all shared the opinion, in agreement with our professors and in accord with NT scholarship, before and since Westcott and Hort, that, in various places, during the fourth century, recensions of the NT text had been made, from which the main text-types then developed. . . . We spoke of recensions and text-types, and if this was not enough, we referred to pre-Caesarean and other text-types, to mixed texts, and so on.

I, too, have spoken of mixed texts, in connection with the form of the NT text in the second and third centuries, but I have always done so with a guilty conscience. For, according to the rules of linguistic philology it is impossible to speak of mixed texts before recensions have been made (they only can follow them), whereas, the NT manuscripts of the second and third centuries which have a "mixed text" clearly existed before recensions were made. . . . The simple fact that all these papyri, with their various distinctive characteristics, did exist side by side, in the same ecclesiastical province, that is, in Egypt, where they were found, is the best argument against the existence of any text-types, including the Alexandrian and the Antiochian. We still live in the world of Westcott and Hort with our conception of different recensions and text-types, although this conception has lost its raison d'être, or, it needs at least to be newly and convincingly demonstrated. For, the increase of the documentary evidence and the entirely new areas of research which were opened to us on the discovery of the papyri, mean the end of Westcott and Hort's conception.[52]

(I have quoted men like Zuntz, Clark and Colwell on the "Byzantine" text to show that modern scholars are prepared to reject the notion of a "Byzantine" recension, but the main lesson to be drawn from the variation among "Byzantine" MSS is the one noted by Lake and Burgon—they are orphans, independent witnesses; at least in their generation. The variation between two "Byzantine" MSS will be found to differ both in number and severity from that between two "Western" MSS or two "Alexandrian" MSS—the number and nature of the disagreements between two "Byzantine" MSS throughout the Gospels will seem trivial compared to the number (over 3,000) and nature (many serious) of the disagreements between Aleph and B, the chief "Alexandrian" MSS, in the same space.)

A recent return

Both Colwell[53] and Epp[54] take issue with Aland, claiming that the papyri fit right in with Hort's reconstruction of textual history. But the existence of an affinity between B and P75 does not demonstrate the existence of a text-type or recension. We have just seen Colwell's demonstration and declaration that an "Alexandrian" archetype never existed. Epp himself, after going on to plot the early MSS on three trajectories ("Neutral," "Western," and "midway"), says:

Naturally, this rough sketch should not be understood to mean that the manuscripts mentioned under each of the three categories above necessarily had any direct connections one with another; rather, they stand as randomly surviving members of these three broad streams of textual tradition.[55]

The point is, although different manuscripts exhibit varying affinities, share certain peculiarities, they each differ substantially from all the others (especially the earlier ones) and therefore should not be lumped together. There is no such thing as the testimony of a "Western" or "Alexandrian" text-type (as an entity)—there is only the testimony of individual MSS, Fathers, Versions (or MSS of versions).

In disagreeing with Aland (see notes 52 and 54), Epp declared that our extant materials reveal "only two clear textual streams or trajectories" in the first four centuries of textual transmission, namely the "Neutral" and "Western" text-types.[56] He also suggested that P75 may be considered as an early ancestor for Hort's "Neutral" text, P66 for Hort's "Alexandrian" text, and P45 for Hort's "Western" text.

But he himself had just finished furnishing counter evidence. Thus, with reference to 103 variation units in Mark 6-9 (where P45 is extant), Epp records that P45 shows a 38 percent agreement with Codex D, 40 percent with the Textus Receptus, 42 percent with B, 59 percent with fl3, and 68 percent with W.[57] How can Epp say that P