Thursday, January 9, 2020

Review - Scribal Skips

            Readers who keep track of recently made English New Testaments may recognize the name of the author of Scribal Skips: 1300 Words That Fell Out of the Bible:  Wayne A. Mitchell is already known as the main developer of the New Heart English Bible.
            In a three-page introduction, Mitchell summarizes a current problem in modern-day textual criticism:  the continued use of the guideline lectio brevior potior, that is, the shorter reading is stronger.  This principle – the first guideline in Griesbach’s list of guidelines, or canons, as stated back in 1796 (and, earlier, by Bengel) – was very heavily qualified, but since then it has been applied in an oversimplified manner, to the point that all that some commentators have needed to say, when offering a case against a particular reading, is that it is longer than its rival.
            As David J. Miller pointed out back in 2006 in the journal The Bible Translator in an article titled The Long and Short of Lectio Brevior Potior, research in the past 40 years has shown that measuring the length of rival textual variants is not a valid way to decide which is original.  James Royse, for example, in a detailed examination of some early papyri (P45, P46, P47, P66, P72, and P75, to be precise), determined that the copyists of those papyri made more omissions (creating shorter readings) than they made additions (creating longer readings), at a rate of 312 omissions and 127 additions – that is, at a ratio of approximately 7:3.   
            Additional research by Juan Hernández on the text of Revelation in three uncial manuscripts, and by Peter Head, in a 1990 article Observations on Early Papyri of the Synoptic Gospels, in the journal Biblica (supplemented and re-confirmed in another article in 2004), supported Royse’s findings; Head concluded his article with a statement that “If early scribes were more likely to omit words and phrases (for whatever reasons) it follows that we should not prefer the shorter reading, but rather prefer the longer reading (other factors being equal).”  Even with full consideration given to the cautionary footnote that Head added to this sentence, to the effect that “Each variant must be assessed on its merits,” at the end of the day, the point seems irresistible that a variant’s brevity is not a merit. 
            And therein lies a problem:  in the creation of the compilation of Westcott & Hort, published in 1881, the editors appear to have routinely thought that the copyists expanded the text; Hort described the tendency of copyists to expand their texts as an “almost universal tendency.”  Similarly, Eberhard Nestle, in his 1901 Introduction to the Textual Criticism of the New Testament, on page 323, stated, “It is a fundamental principle of textual criticism that the lectio brevio is to be preferred.  And if one consults the Preface to the First Edition of the UBS Greek New Testament (on page x of the UBS GNT, 4th ed.), the editors acknowledge that Westcott & Hort’s edition was the initial basis for the UBS Greek New Testament, and Nestle’s work is listed next.  Metzger, in his handbook The Text of the New Testament, stated (4th ed., p. 303), “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.”
            The first exception that Metzger mentioned – cases of possible parablepsis – does not appear to have been given adequate weight.  Many shorter readings that can be explained as results of parableptic error continue to be favored over their longer rivals. This is one reason why the Nestle-Aland Novum Testamentum Graece continues to confidently diverge from the 1881 text at fewer than 700 points.
            What would the text of the New Testament look like if shorter readings that can be accounted for as omissions caused by parablepsis – that is, simple scribal errors that happened when a copyist’s line of sight drifted from one letter (or letters) at one point to another point further along in the text where the same or similar letter (or letters) recurred – were rejected in favor of rival longer readings?  That question may be answered by a careful reading of the portion of Scribal Slips that focuses on the New Testament – that is, everything after page 77. 
            In the part of Scribal Slips that pertains to the New Testament text, each page features a straightforward presentation of two or three passages where there are (a) a shorter reading, and (b) a longer reading which can be explained as the original reading on the grounds that it was accidentally lost via scribal error.  External evidence is thoroughly listed for each reading.  A series of cases in Matthew helpfully exemplifies the kinds of readings that are covered:
            ● Omission of a word:  Matthew 19:11 – witnesses are listed for the longer reading (τον λογον τουτον) and for the shorter reading (τον λογον).  Then the verse is presented in English, with the omitted word within brackets:  “But he said to them, “Not all men can receive [this] saying, but those to whom it is given.”  The idea is that an early copyist’s line of sight drifted from the letters -ον at the end of λογον to the same letters at the end of τουτον.
            ● Omission of consecutive words:  Matthew 19:29 – witnesses are listed for the longer reading (οικιας η αδελφους η αδελφας η πατερα η μητερα η γυναικα η τεκνα η αγρους) and for the shorter reading found mainly in Codex Vaticanus (οικιας η αδελφους η αδελφας η πατερα η μητερα η τεκνα η αγρους).  Then the verse is presented in English, with the omitted words in brackets:  “Everyone who has left houses, or brothers, or sisters, or father, or mother, [or wife,] or children, or lands, for my name’s sake, will receive one hundred times, and will inherit eternal life.”  The idea is that a copyist’s line of sight drifted from η to η, accidentally skipping η γυναικα.
            ● Omission of a phrase:  Matthew 20:16 – witnesses are listed for the longer reading (ουτως εσονται οι εσχατοι πρωτοι και οι πρωτοι εσχατοι πολλοι γαρ εισιν κλητοι ολιγοι δε εκλεκτοι) and for the shorter reading (ουτως εσονται οι εσχατοι πρωτοι και οι πρωτοι εσχατοι).  The idea is that a copyist’s line of sight drifted from the letters at the end of εσχατοι to the same letters at the end of εκλεκτοι.
           
            Similarly in Mark, Mitchell pin-points readings which are shorter in the NIV, ESV, and NLT than in the KJV, NKJV, and MEV, and shows how simple scribal inattentiveness – not some devious Gnostic agenda – caused the loss of some text.  Examples:
            Mark 2:16 – Parablepsis accounts for the loss of the words “and drinks”
            Mark 10:24 – Parablepsis account for the loss of the words “for those who trust in riches”
            Mark 11:26 – Parablepsis accounts for the loss of the entire verse

            Even though this may give readers an illuminating view at the possible impact of scribal errors upon the text, the spotlight is not as bright as it could be.  In just two chapters of Matthew (chapters 19 and 20), readers are not shown how parablepsis may be the cause of . . .
            ► the non-inclusion of οτι in 19:8 in Vaticanus and Bezae
            ► the non-inclusion of αγαθέ in 19:16 in ℵ B L D et al
            ► the non-inclusion of ει μη εις ο Θεος in 19:17 in ℵ B L D et al
            ► the non-inclusion of ου μοιχεύσεις ου κλέψεις in 19:18 in ℵ*
            ► the cause of the non-inclusion of εισελθειν in 19:24 in ℵ L f1 157 et al
            ► the non-inclusion of παρα ανθρώποις in 19:26 in ℵ*
            ► the non-inclusion of οτι in 20:12 in ℵ B D et al
            ► the non-inclusion of οι in C* and H in 20:12, and
            ► the non-inclusion of 20:31 (the entire verse) in GA 2* and 157. 
But if all such instances of possible parablepsis and haplography had been included, this would be a much longer book.  The samples that Mitchell has recorded should be sufficient to cause some readers to exclaim, “Leapin’ lizards; why on earth are we relying on the Alexandrian Text; it’s full of holes?!” or similar expressions of dismay.
            There is not a lot of argumentation in this book, just straightforward presentations of evidence which may or may not constitute a genuine case of a haplography-induced scribal skip at any given point.  For instance, regarding Luke 4:5, Mitchell accurately reports that À* B L 1241 et al do not have the words ο διαβολος εις ορος υψηλον, which appear in most manuscripts immediately after the words και αναγαγων αυτον.  The idea is that a copyist’s line of sight drifted from the letters -ον at the end of αυτον to the same letters at the end of υψηλον, but one could also argue that the words have been inserted here as a harmonization to the parallel-scene in Matthew 4:8.
            Some readers might initially suspect, as they observe over and over that the shorter reading is found in the flagship manuscripts of the Alexandrian text, and the longer reading is found in the Textus Receptus, that perhaps Mitchell’s presentation is a surreptitious defense of the base-text of the King James Version.  However, the tendency for the Alexandrian Text to have the shorter reading is only a tendency; the Alexandrian text is not always shorter.  Mitchell points out that the Textus Receptus appears to be missing 39 words, in passages such as John 12:19 (where P66 À B also do not include ολος after κοσμος), James 4:12 (where the Textus Receptus, like most manuscripts, does not include και κριτης (“and Judge”) after νομοθέτης (“Lawgiver”), and Jude verse 25 (where the Textus Receptus, like most manuscripts, does not include the phrase, “through Jesus Christ our Lord”).    

            In a six-page conclusion, Mitchell maximizes the implications of the data he has presented.  The texts of the manuscripts that are routinely referred to as “oldest and best” in some Bible-footnotes are by no means safe from scribal skips; Mitchell indicts P44vid, P45, P66, P75, and À B D L et al as containing echoes of an early parableptic error in John 10:13 in which ο δε μισθωτος φευγει was accidentally skipped after σκορπίζει at the end of verse 12.  Besides proposing that the Hebrew text of the Old Testament (in the Masoretic Text) has lost 1,279 words due to haplography, he also proposes that the Greek text of the New Testament (in the Nestle-Aland compilation, 28th edition) has lost 327 words (in addition to which 100 words are bracketed), including four entire verses (Mt. 27:35, Mark 11:26, Mark 15:28, and Luke 23:17), due to scribal skips.
            Finally, in an interesting appendix, Mitchell comments about some differences between the form of some Old Testament passages and New Testament quotations of them.  These include considerations of Genesis 11:13 (compared to Luke 3:36), Genesis 47:31 (compared to Hebrews 11:21), Exodus 2:14 (compared to Acts 7:28), Psalm 8:2 (compared to Matthew 21:16), Isaiah 28:16 (compared to First Peter 2:6), Hosea 14:2 (compared to Hebrews 13:15), and more.  Often the quotation in the New Testament is identical or similar to the reading in the Septuagint.    

            Almost certainly, Mitchell’s evidence-presentations will not be persuasive in every case.  (For example, it seems to me that the longer form of Luke 11:2-4 is explained much more easily as a harmonization to the Lord’s model prayer in Matthew 6:9-13, than the shorter form is explained as due to a series of remarkable scribal skips.)  But in many cases, Mitchell has supplied a strong basis to reject “prefer the shorter reading,” and to put in its place, “prefer the reading with the simplest explanation.”   Very often a shorter reading is explained by a scribal slip with more simplicity than a longer ending is explained as the result of conscious scribal effort. 

            It is not as Mitchell provides only vague evidence for the textual losses that he proposes have occurred.  He (or his sources) apparently sifted through the contents of 28 papyri, over 80 uncials, over 110 minuscules, at least 18 Old Latin manuscripts, four editions of the Vulgate, seven Syriac (or Aramaic) texts, eight Coptic sources, five additional versions (Armenian, Georgian, Ethiopic, Gothic, Old Church Slavonic), and the works of 62 patristic writers – all of which/whom are listed near the beginning of the book, giving readers the tools to trace the losses.  That’s far more materials than were cited in the apparatus of the Tyndale House Greek New Testament.
            Mitchell acknowledged near the end of his book that it is not exhaustive.   (A future edition could be re-titled with  “1,500” or  “1,600” in place of  “1,300.”)  First Peter 2:2, for example, could be profitably added to the list of passages where a difference among English translations may have its origin in an ancient scribe’s inattentiveness.   But Scribal Skips is already a resource which investigators of the differences among modern English translations, and their base-texts, will find highly informative and useful.

            Scribal Skips: 1,300 Words That Fell Out of the Bible is available to purchase at Amazon as a Kindle e-book for $7.99, or as a paperback for $14.99.  
 
            [Pages 1-77 of Scribal Skips cover Old Testament passages, and are not covered in this review.  But Mitchell’s h.t.-implying comparisons, which he lets speak for themselves, are often very interesting, and merit consideration from investigators of the Old Testament text, whether regarding large readings (at passages such as First Samuel 10:27-11:1 and First Samuel 14:41) or small ones (at passages such as Malachi 1:7 and 3:2).]


[P.S.  Another man named Mitchell, Jonathan Mitchell, has also made an English New Testament, using a super-amplified technique.  Wayne A, Mitchell is not the same person.]



5 comments:

Matthew M. Rose said...

Is Mr. Mitchell really implying that the Hebrew Text is corrupted by haplography due to homoeoarcton in Genesis 1:6, 1:8, etc.? Are we to believe that no Hebrew scribe (or hearer) ever noticed such an immediate departure from the original Text in the first book of Moses? These variations can also be explained by dittography via HA. Not to mention the fact that the editor/s of the LXX could have simply expanded the text here with the intention of making it more uniform with the surrounding context and wording. This comes across as very poor judgement.

Wayne Mitchell said...

Concerning Lk 4:5, as noted by one recent textual commentary, the support for the omission is not good and the shorter text is awkward without the words. “Taking him up” to what? Copyist haplography often results in a text that doesn’t make sense. The manuscript evidence for a copyist two-letter homoioteleuton (on-on) is evident and explains the origin of the shorter reading with it’s problematic text.

“In comparison of the Mss of Lk 11:2-4 and the parallel at Mt 6:9-13, all Mss of the longer version of Lk 11:4 lack a doxology, which is strong evidence against a supposition of harmonization to Mt. The longer version also has strong support, which includes the early uncials A C D W. In comparison to other Mss, the locations where text is omitted in the shorter version suggests that it is a consequence of three separate instances of copyist haplography.” [1]

Matthew Rose has brought up questions on Genesis 1:6 and 1:8. Along with the other examples listed for chapter 1 it is apparent to those who have looked at the structure of this chapter and all the manuscripts that there are significant problems in MT due to copyist haplography. For example, the evidence suggests that wayehi-ken “And it was so” fell out of vv.6, 20 and 26 in MT from two-letter copyist homoioarcton: wy-wy, and that two (vv.6 and 26) were restored by MT at the wrong locations (at vv.7 and 30), while the one at v.20 was not restored. [2]

In Genesis 1:8, wayyar elohim ki-tov “And God saw that it was good” has fallen out of v.8 in MT from a two-letter copyist homoioarcton: wy-wy. [3]

As Brug noted, while the NT has undergone significant text criticism in light of all the manuscripts, “[g]enerally, until recently, Old Testament textual scholars did not try to produce an eclectic Hebrew text which incorporates readings from many different manuscripts…” [4]

The discoveries in the field of biblical textual criticism have demonstrated that there is no manuscript or textual stream that is free of copyist mistakes. The manuscripts must be compared and the tools of an unbiased textual criticism appropriately applied in order to restore the authentic text.



1. W. Mitchell, Scribal Skips (4th ed.), 136 fn 14.

2. E.A. Speiser, Genesis (Anchor Bible), 6; J. Cook, The text-critical and exegetical value of the Dead Sea Scrolls. HTS Teologiese Studies/Theological Studies (2016) 72(4), 3; D. N. Freedman and D. Miano, Slip of the Eye: Accidental Omission in the Masoretic Tradition,” in The Challenge of Bible Translation. G. G. Scorgie, M. L, Strauss, S. M. Voth, (eds.), 278-79.

3. J. Cook, The text-critical and exegetical value of the Dead Sea Scrolls, 3; Freedman and Miano, Slip of the Eye, 279.

4. J. F. Brug, Textual Criticism of the Old Testament (2014), 2; Cf. B. Albrecktson, “Masoretic or Mixed?,” Textus XXIII (2007), 33-50.

Matthew M. Rose said...

Wayne Mitchell,

You state: "Genesis 1:6 and 1:8. Along with the other examples listed for chapter 1 it is apparent to those who have looked at the structure of this chapter and all the manuscripts that there are significant problems in MT due to copyist haplography. For example, the evidence suggests that wayehi-ken “And it was so” fell out of vv.6, 20 and 26 in MT from two-letter copyist homoioarcton: wy-wy, and that two (vv.6 and 26) were restored by MT at the wrong locations (at vv.7 and 30), while the one at v.20 was not restored."

This is a purely arbitrary evaluation of the situation. Considering that there are other possible explanations for the errors within Genesis ch. 1, and that the burden of truth is upon those who would alter the Hebrew MT; I don't suppose that you could supply a full Critical Apparatus on the above mentioned variant readings?

Wayne Mitchell said...

Concerning the problems previously noted in the MT edition for Genesis 1:

Gen 1:6: “This clause is correctly reproduced here by LXX but misplaced in Heb. at the end of vs. 7.” [1]

“[T]he equivalent of the Hebrew formula [And it was so] does not occur in verse 6 following the Wortbericht, but in verse 7 after the Tatbericht. This applies also to the sixth work (vs 20), where it is not used at all in the MT.” [2]

“The typical formulaic phrase (“and it was so”) appears at the end of verse 7 in the Hebrew version but at the end of verse 6 in the Greek. The reading of the Greek tradition is consistent with the order of events on the other creative days, while that of MT is not...[T]he placement of [and it was so] at the end of verse 7 makes no sense, and no convincing explanation has been given as to why the author would have put the phrase there, particularly when it is at variance with the established modus operandi and disturbs the flow of discourse. LXX’s testimony, therefore, is to be preferred here.”

“In LXX, we find evidence of another [“and it was so”] in verse 20 that is not in MT or SP. We would naturally expect the formula to appear in this position, and a double-letter homoioarkton (waw-yod) urges us to accept the longer reading.”… “At the end of verse 26, we are surprised not to find another [“and it was so”]...A two-letter homoioarkton (waw-yod) could have caused it.” [3]

Gen 1:8: “This Rabbinic passage also includes attempts to explain why the ending formula for the Tatbericht [And God saw that it was good] in verse 8 is omitted in MT.” [4]

”The LXX preserves the familiar phrase [“and God saw that it was good”] immediately before the evening-morning formula. MT does not contain the clause. The two-letter homoioarkton caused the copyist’s eye to skip ahead, and the phrase was lost...[T]here are eight acts of creation. We would expect a [“and God saw that it was good”] formula...to coincide with each one of them.” [5]

Lets examine the situation in MT for Genesis 1 a little further. In 1:9 the “And X happened” is missing in MT but present in LXX, OL and in a few extant words in DSS 4Qgen(k). A three-letter copyist homoioarcton (wyq-wyq) is responsible for the removal of eight words in MT. [6][7]

“[T]he rigid structure of the creation account makes it more likely that the phrase [Gen 1:9b] is original and was lost by homoeoarchton in M[T] etc”. [8]

In Genesis 1:14 most of the “Let X happen” is missing in MT from three separate instances of homoioarcton, removing eight words, but the words can be found in LXX, LXX Mss, DSS, SP and OL. [9]

There are even more problems in the MT copy of Genesis 1 that could be discussed. When the structure of Genesis 1 is examined and all manuscripts are compared it is evident that text is missing in MT. Translators should no longer be using an uncorrected MT edition for this chapter. The Bible reader deserves better.


1. E.A. Speiser, Genesis (Anchor Bible), 6.
2. J. Cook, The text-critical and exegetical value of the Dead Sea Scrolls. HTS Teologiese Studies/Theological Studies (2016) 72(4), 3.
3. D. N. Freedman and D. Miano, Slip of the Eye: Accidental Omission in the Masoretic Tradition,” in The Challenge of Bible Translation, 278-79, 281-82.
4. J. Cook, The text-critical and exegetical value of the Dead Sea Scrolls, 3.
5. Freedman and Miano, Slip of the Eye, 279.
6. D. N. Freedman and S. D. Overton, “Omitting the Omissions: The Case for Haplography in the Transmission of the Biblical Texts,” in ‘Imagining’ Biblical Worlds (2002), 102.
7. R. S. Hendel, “Qumran and a New Edition of the Hebrew Bible,” in The Bible and the Dead Sea Scrolls (2006), 151-52.
8. Discoveries in the Judaean Desert XII. Qumran Cave 4: Genesis to Numbers (1995), 76.
9. D. N. Freedman and D. Miano, Slip of the Eye, 279-81.

Matthew M. Rose said...

Wayne Mitchell,

You're welcome to come to your own conclusions (obviously), or follow the conclusions of others for that matter. And yet you seem to put forth your opinion (or the opinions of others; which you share) as fact.

If I may ask one important question: is there anywhere in the Old or New Testament where an omission could be accounted for by an instance of haplography (due to homoeoteleuton or homoeoarcton) that you believe the shorter reading is original?