The New Testament and textual criticism – “the shorter reading is preferred” (Part 1)

What is the genuine Greek—what the true Text of the New Testament? Which are the very words which were written by the Evangelists and Apostles of our Lord Jesus Christ under the Inspiration of the Holy Ghost?

The two above questions were posed by Edward Miller in his book, A Guide to the Textual Criticism of the New Testament (1886, p.1 - link to PDF). The answer to these questions is the ultimate goal of New Testament textual criticism.

As the number of extant Greek manuscripts increased, the number of variants within those extant manuscripts also increased. Before the 16th century, all the GNT manuscripts were hand written, but in the year 1514 the first printed edition of the Greek New Testament was produced (volume V of the Complutensian Polygot, which was not published until 1520), and in 1516 Desiderius Erasmus’ Novum Instrumentum Omne was printed and published. Between 1519-1535, Erasmus created four more editions of the Greek New Testament that were also printed and published. Erasmus’ five editions were soon followed by four GNT editions printed by the Parisian Robert Estienne (Latin: Stephanos)—1546, 1549, 1550 and 1551. The 1550 edition was the first GNT to contain a critical apparatus of variant readings; readings that were compiled from fourteen GNT manuscripts and the GNT of the Comlputensian Polygot. The apparatus of Stephanos’ 1550 edition was principally the beginning of textual criticism of the GNT as a distinct discipline.

Following this somewhat humble beginning, the textual criticism of the GNT as a distinct discipline has grown into a massive field of study that includes the textual criticism of non-Greek New Testament translations (especially Latin), and the quotations of the NT by the Church Fathers. Interestingly enough, this field of study became dominated by liberal and nominal ‘Christians’, with one of the top GNT textual critics of our day—Bart Ehrman—becoming an agnostic, and repudiating any notion of the GNT as inspired Scripture from God.

In the 17th century, various methods, rules, and theories began to emerge within the field of GNT textual criticism in the attempt to identify which of the tens of thousands of variant readings found within the thousands of extant GNT manuscripts/texts are the purest representatives of the original texts penned by the apostles and disciples of Jesus Christ—texts that Christians have termed the New Testament.

One of the earliest theories to develop was the identification of manuscripts into groups/text types based primarily—but not exclusively so—on the geographical location where the manuscript was thought to be written. Johann Albrecht Bengel (1687-1752) classified the extant manuscripts he personally was cognizant of into two groups that he termed ‘Asiatic' and ‘African’. The ‘Asiatic' group contained those manuscripts thought to have been written in Constantinople and the surrounding Greek speaking environs. The ‘African’ group being represented by the extant Latin translations and Greek texts like the codex Alexandrinus. Johann Salmo Semler (1725-1791) further developed the theory of groups/text types by classifying the extant manuscripts into three recensions: the ‘Alexandrain', “Eastern/Byzantine' and 'Western'.

A student of Semler’s, Johann Jakob Griesbach (1745-1812), retained the three group distinctions with greater refinements and additions to each of the groups.

This identification of manuscripts into groups/types is classified as one of the 'external evidences' of text-critical methodology. Concerning the ‘evidences' utilized by textual critics, Dr. Epp wrote:

What had emerged in little more than a decade from Mill to Bentley [textual critics of the early 18th century] was a twofold set of criteria, external and internal, that, while partial and rudimentary, formed the foundation of text-critical methodology ever after. These criteria were more clearly defined over time, but basically external evidence assesses factors such as the age, quality, geographical distribution, and groupings of manuscripts and other witnesses, while internal evidence assesses what authors were most likely to write and what scribes were likely to transcribe. During the eighteenth century and through the nineteenth, virtually all notable editors stated a basic, general principle that the text should be formed from the most ancient textual witnesses, and (except for Lachmann) their editions also included a list of internal criteria. Bengel (1725 and 1742) offered twenty-seven canons, Wettstein (1730 and 1751–52) listed eighteen, Griesbach (1796–1806) fifteen, Tischendorf (1869–72, in the prolegomena by Caspar René Gregory)  five, Tregelles (1857–72) nine, and Westcott and Hort (1881–82) also offered some nine, though not in a formal list. (Eldon Jay Epp, “Traditional 'Canons' of New Testament Textual Criticism: Their Value, Validity, and Viability-or Lack Thereof”, in The Textual History of the Greek New Testament: Changing Views in Contemporary Research, 2011, pp. 83, 84)

One of the ‘canons’—i.e. rules, principles—of the ‘internal evidences’ of textual criticism found in the works of the majority of textual critics (one can add Bruce Meztger, Kurt and Barbara Aland to the above list by Epp), is lectio brevior potior—the shorter reading is preferred. A detailed description of lectio brevior potior was provided by Johann Jakob Griesbach. The following is Bruce Metzger’s English translation, with his brief introduction:

Among the 15 canons of textual criticism that Griesbach elaborated, the following (his first canon) may be given as a specimen:

The shorter reading (unless it lacks entirely the authority of the ancient and weighty witnesses) is to be preferred to the more verbose, for scribes were much more prone to add than to omit. They scarcely ever deliberately omitted anything, but they added many things; certainly they omitted some things by accident, but likewise not a few things have been added to the text by scribes through errors of the eye, ear, memory, imagination, and judgement. Particularly the shorter reading is to be preferred, even though according to the authority of the witnesses it may appear to be inferior to the other,—

a. if at the same time it is more difficult, more obscure, ambiguous, elliptical, hebraizing, or solecistic;

b. if the same thing is expressed with different phrases in various manuscripts;

c. if the order of words varies;

d. if at the beginning of pericopes;

e. if the longer reading savours of a gloss or interpretation, or agrees with the wording of parallel passages, or seems to have come from lectionaries.

But on the other hand the longer reading is to be preferred to the shorter (unless the latter appears in many good witnesses),—

a. if the occasion of the omission can be attributed to homoeoteleuton;

b. if that which was omitted could have seemed to the scribe to be obscure, harsh, superfluous, unusual, paradoxical, offensive to pious ears, erroneous, or in opposition to parallel passages;

c. if that which is lacking could be lacking without harming the sense or the structure of the sentence, as for example incidental, brief propositions, and other matter the absence of which would be scarcely noticed by the scribe when re-reading what he had written;

d. if the shorter reading is less in accord with the character, style, or scope of the author;

e. if the shorter reading utterly lacks sense;

f. if it is probable that the shorter reading has crept in from parallel passages or from lectionaries. (Bruce Metzger, The Text of the New Testament: Its Transmission, Corruption, and Restoration, 2nd  ed., 1968, p. 120.)

Metzger himself was an advocate of lectio brevior potior; note the following:

In general, the shorter reading is to be preferred, except where parablepsis arising from homoeoteleuton may have occurred or where the scribe may have omitted material that he deemed to be superfluous, harsh, or contrary  to pious belief, liturgical usage, or ascetical practice. (Compare Griesbach's fuller statement of this criterion, p. 120 above.) [The Text of the New Testament: Its Transmission, Corruption, and Restoration, 2nd ed., 1968, pp. 209, 210.]

And from the ‘Introduction’ of Metzger’s, A Textual Commentary on the Greek New Testament, we read:

2. In general the shorter reading is to be preferred, except where

(a) Parablepsis arising from homoeoarcton or homoeoteleuton may have occurred (i.e., where the eye of the copyist may have inadvertently passed from one word to another having a similar sequence of letters); or where

(b) The scribe may have omitted material that was deemed to be (i) superfluous, (ii) harsh, or (iii) contrary to pious belief, liturgical usage, or ascetical practice. (1975, corrected edition, p. xxvii)

Metzger’s colleagues, Kurt and Barbara Aland, also accepted lectio brevior potior as one of their, “TWELVE BASIC RULES FOR TEXTUAL CRITICISM”:

11. The venerable maxim lectio brevior lectio potior ("the shorter reading is the more probable reading") is certainly right in many instances. But here again the principle cannot be applied mechanically. It is not valid for witnesses whose texts otherwise vary significantly from the characteristic patterns of the textual tradition, with frequent omissions or expansions reflecting editorial tendencies (e.g., D). Neither should the commonly accepted rule of thumb that variants agreeing with parallel passages or with the Septuagint in Old Testament quotations are secondary be applied in a purely mechanical way. A blind consistency can be just as dangerous here as in Rule 10 (lectio difficilior). [Kurt and Barbara Aland, The Text of the New Testament, trans. by Erroll F. Rhodes, second edition, 1989, p. 281]

Though Griesbach, Metzger and the Alands delineate criteria for exceptions to the  lectio brevior potior canon/rule, in the vast majority of cases when a textual reading has shorter and longer variants, the shorter reading is adopted the preferred one.

Now, what you will rarely hear from textual critics who have embraced the, “venerable maxim lectio brevior lectio potior (‘the shorter reading is the more probable reading’)”, is that many of the assumptions/criteria used support the lectio brevior potior canon/rule are problematic. The rest of the post will focus on one of those assumptions/criteria: scribes were much more prone to add than to omit.

If my memory serves me correctly, the first time I saw a challenge to the axiom that scribes were much more prone to add than to omit was in Dr. Harry Sturz’s book, The Byzantine Text-Type & New Testament Textual Criticism. From Dr. Sturz’s book we read:

Kilpatrick, in his evaluation of the text behind the TR, includes a discussion on conflation, in which he examines variant readings eclectically, and finds that in many instances the longer reading should be preferred as the original reading. He concludes the discussion on homoeoteluton with the following observations:

This list is ... sufficient to show both how prevalent this kind of mistake is and how frequently the Textus Receptus and its allies preserve the original reading. Westcott and Hort of course rejected their evidence and chose the shorter text even when it clearly impaired the meaning as at Mark x. 7.

lt is worth considering how this came about. One of the canons of  textual critics in modern times has been lectio brevior potior . ... On the other hand if we substitute the maxim, 'the longer text, other things being equal, is preferable', have we any reason for thinking that this is more mistaken than the conventional lectio brevior potior? We are used to this last but the fact that it is traditional is no argument for its being true. Nonetheless, Westcott and Hort do not seem to have thought of challenging it.

There are passages where reasons can be given for preferring the longer text and there are others where we can find reasons for preferring the shorter. There is a third category where there does not seem to be any reason for deciding one way or the other. How do we decide between longer and shorter texts in this third category? On reflection we do not seem able to find any reason for thinking that the maxim lectio brevior potior really holds good. We can only hope that a fuller acquaintance with the problems concerned will enable us increasingly to discern reasons in each instance why the longer or the shorter reading seems more probable.

Cited from Kilpatrick's essay: "The Greek New Testament Text of Today and the Textus Receptus," Chap. VIII in The New Testament in Historical and Contemporary Perspective, Essays in Memory of G. H. Macgregor, ed. by Anderson and Barclay (Oxford: Basil Blackwell, 1965), p. 196. [Sturz, Harry A. The Byzantine Text-Type & New Testament Textual Criticism, p. 89]

After reading the above quote provided by Sturz, I wanted to get a broader context, as well as find out what was removed and replaced by the ellipses, so I ordered the referenced book.

Before getting to the greater context of Kilpatrick's essay (referenced above by Sturz), I will provide the full texts of those sections that contain the ellipses.

Sturz’s quote: This list is ... sufficient to show both how prevalent this kind of mistake is and how frequently the Textus Receptus and its allies preserve the original reading.

Full text: This list is far from exhaustive but is sufficient to show both how prevalent this kind of mistake is and how frequently the Textus Receptus and its allies preserve the original reading.

Sturz’s quote: lt is worth considering how this came about. One of the canons of  textual critics in modern times has been lectio brevior potior . ... On the other hand if we substitute the maxim, 'the longer text, other things being equal, is preferable', have we any reason for thinking that this is more mistaken than the conventional lectio brevior potior?

Full text: It is worth considering how this came about. One of the canons of textual critics in modern times has been lectio brevior potior. We may limit this to the form, 'the shorter text, other things being equal, is preferable', in deference to the investigations of A. C. Clark and others which have revealed how widespread has been the prevalence of ὁυ and other causes of omission. On the other hand if we substitute the maxim, 'the longer text, other things being equal, is preferable', have we any reason for thinking that this is more mistaken than the conventional lectio brevior potior?

Now, the broader context. The first paragraph of the essay presents two “basic contentions of Westcott and Hort” that Kilpatrick believes are seriously flawed. The following is the full paragraph:

One of the basic contentions of Westcott and Hort was that the Syrian text¹, the text that appears with variations in A², the Textus Receptus, and the vast majority of later witnesses, is a secondary text based on the older Neutral  and Western texts. A second basic contention was that the Western text was in general inferior to the Neutral text. The Neutral text alone preserved the Greek New Testament in something like its original purity and so served as the foundation of Westcott and Hort’s edition.

In the next paragraph, Kilpatrick provides valuable information on the “second basis contention" of Westcott and Hort:

Hort’s views on the Western text were soon challenged. Scholars like F. C. Burkitt and C. H. Turner showed that, if readings were examined on their merits, the Western text was often right and the Neutral or Egyptian text often wrong. Their contentions have been widely accepted and an increasing number of readings in D³ and its allies are recognized as probably original.

The third paragraph, sets the tone for the rest of the essay:

No such change has taken place in opinion about Syrian text. Few attempts have been made to show that any of its distinctive readings are original and Hort's account of its origins and characteristics have not been challenged by the majority of textual critics.

Kilpatrick then goes on to present substantive cases for three assessments: first, "the great majority of [N. T. textual] variants came into being before A. D. 200”; second, many of the Syrian/Byzantine text distinct readings are much older than what most textual critics have accepted as a truism (as per Westcott and Hort’s theory); and third, the lectio brevior potior canon/rule has serious flaws (see above quotes).

Kilpatrick’s first and second assessments are inextricably linked. Note the following:

Professor H. Vogels has suggested that, apart from errors [unaware copying vs. deliberate], the the great majority of [N. T. textual] variants came into being before A. D. 200. This seems reasonable. Many readings can be shown to be in existence before that date: few demonstrably came into being after it. On this hypothesis most readings distinctive of the Syrian text will be older than A. D. 200 even if the selection of these readings in that text appear later. Consequently we cannot condemn these variants a a product of the depravity of the fourth century. (p.190)

He then adds:

One of Hort’s complaints against the Syrian text was that it as characterized by conflate readings. In principle the presence of conflate readings in the New Testament need not surprise us. The evidence of the critical apparatus suggests that they are to be found up and down the Greek text. There are, however, two questions we must try to answer: (i) are conflate readings distinctive of the Syrian text? And (ii) are all readings that look like conflate readings really conflate? (p. 190)

Kilpatrick then provides examples of 'conflation' within Westcott and Hort’s so-called ‘Neutral' text by comparing certain readings from ℵB⁴ that are shorter than the same readings found within the Western and/or Syrian texts.

He then concludes the following:

From these examples we can see that not all apparent conflate readings are really conflate. Sometimes they present the original text and, when they do so and are peculiar to the Syrian text, then the Syrian text must have credit for preserving the correct reading. Likewise real conflations occur in other witnesses apart from the Syrian text and it would be mistaken to argue that conflate readings were characteristic of this text. Thus the argument from conflation does not serve to condemn the Syrian text in the way that Westcott and Hort had supposed. (pp. 192, 193)

On pages 194-196, Kilpatrick examines “some Syrian readings on their merits, seeing that we cannot dismiss the Syrian text as obviously secondary on grounds of conflation or harmonization” [when compared to the so-called ‘Neutral’ text]. In this section he argues that the Textus Receptus sometimes preserves Semitic expressions that are longer readings than those found in ℵB, suggesting that the ℵB readings were purposeful changes made to the texts to conform to Classical Greek, rather than Koine Greek used to retain the original Semitic expressions found in the New Testament.

He follows his examples with the following bold assessment:

These three instances of the superiority of A and the Textus Receptus justify us in looking afresh at readings that are characteristic of these witnesses and considering each on its merits. From the time of Westcott and Hort to Syrian or Byzantine or as proper to the Textus Receptus to condemn it outright. There have been exceptions to this practice such as those of van Soden, Vogels, and Bover, but there have been few formal attempts at a justification of them.

To contribute to any such justification it is necessary to show in the main categories of variants the Syrian text is sometimes right. (p. 195)

Kilpatrick then provides eight such examples (pp. 195, 196), which is followed by the quote from Sturz—cited earlier in this post—that started with, “This list is ... sufficient to show both how prevalent this kind of mistake is and how frequently the Textus Receptus and its allies preserve the original reading.”

Unfortunately, few textual scholars have adjusted their sacred canons/rules—maintaining the status quo as reflected in the quotes from Metzger and the Alands provided above—retaining the lectio brevior potior axiom as a weapon in their criticisms of the Syrian/Byzantine text type.

In my next post (the Lord willing) I will delve into the substantive critique of the lectio brevior potior canon/rule by one recent textual critic who has taken Kilpatrick’s assessments seriously.

Grace and peace,

David

Notes:

1. The ‘Syrian text’ (also know as the Antiochian, Byzantine, Constantinopolitan, Ecclesiastical, Majority, Traditional) is one of the four textual families/types identified by Westcott and Hort; the other three were termed the Alexandrian, Western, and Neutral.

2. A = Alexandrinus codex (5th century)

3. D = Bezae Cantabrigiensis codex (5th century)

4. ℵB = Sinaiticus codex and Vaticanus codex (both 4th century; most textual critics now include ℵB in the Alexandrian text-type)