Correctors and Corrections

Contents: Introduction * Noteworthy Corrected Manuscripts * The Significance of Corrections

Introduction

Ancient scribes were at least as aware of scribal errors as moderns. Since all manuscripts were copied individually, each needed to be individually checked for errors. This process eventually came to be standardized.

We don't know how or whether early manuscripts were systematically corrected at the time of copying. In a scriptorium, however, it was the practice to employ a manuscript διορθωτης (literally "one who straightens," which we might loosely render as "guy supposed to make this thing right"). The diorthotes was often a scribe specially trained to find and rectify mistakes, though we often find a scribe acting as his own diorthotes.

Once a manuscript had been corrected, the diorthotes was often expected to mark it finished. So a manuscript might be marked δι or αντιβληθη to show that it had been checked over. This notation seems to be rare in New Testament manuscripts, however.

The diorthotes was often the last scribe to work on a manuscript. (This is particularly true of Byzantine manuscripts.) But manuscripts represented a lot of expense and work; an owner might be reluctant to discard a manuscript simply because its text did not meet the tastes of the times. So we see many manuscripts, including Sinaiticus and Bezae, repeatedly corrected to bring them more in line with the Byzantine text.

Where a manuscript has been corrected, it is customary to refer to the original reading with an asterisk. Thus D* in a critical apparatus indicates that this reading is supported by the original hand of D.

Conventions for the correctors have varied. The simplest is to use additional asterisks to refer to the correctors. Thus, if D* refers to the original hand of D, D** refers to the first corrector, D*** to the second, etc.

The problems with this notation are obvious. If a manuscript has many correctors, simply reading the apparatus is a chore. (Quick! Which corrector is D*******?) In addition, there is an æsthetic difficulty -- D**, despite the presence of two asterisks, refers to the first corrector.

The solution was to use superscripts. So, instead of D**, one would write D^c.

This is, of course, all very well where one corrector is involved. But suppose there are two or three, or even more (as sometimes happened)? In this case, the superscripts were retained, but different symbols used.

In the past, correctors were often referred to by a superscript letter. So ℵ^a referred to a reading from the first corrector of Sinaiticus, while ℵ^b would refer to the second. It is now more normal to refer to correctors by number, making ℵ¹ the first corrector, ℵ² the second, etc. If a manuscript had only a single corrector, of course, the simple ^c notation is retained.

Unfortunately, this system is not always used outside Biblical criticism. The manuscripts of Chaucer, for instance, use superscripts to denote the witnesses -- there are (for example) four manuscripts of the Canterbury Tales in the Rawlinson Collecton at the Bodleian Library in Oxford. These are given the standard symbols Ra¹, Ra², Ra³, Ra⁴. Rather than using a double superscript -- admittedly a severe problem in dealing with New Testament papyri (for those of us with older eyes, it's not easy to tell 𝔓⁴⁶¹ from 𝔓⁴⁶²!), editors of Chaucer use subscripts -- so Ra¹ represents the main text of Ra¹, Ra¹₁ represents a correction made by the original scribe, and Ra¹₂ a correction made by any later corrector.

Also, there is the issue of correctors who can sometimes be told apart and sometimes can't. For example, if a manuscript has two correctors and a corrector writes a new text or fills in a word, you can probably tell which corrector it is based on the handwriting. But if all the corrector does is mark a letter or two for deletion, it's much harder. Maybe you can tell based on the ink color or the pen that is used, but that's much less reliable than looking at actual letters! And there is no universally accepted system for marking a reading as a correction from an undetermined hand.

A distinction is sometimes made between "amateur" and "professional" correctors. This is an unfortunate notation; in the period after the split of the Roman Empire, professional scribes were very nearly the only people who could read and write, and therefore all correctors were professional. If we change the designations to something like "systematic" and "casual," however, the distinction is accurate. A systematic corrector is one who goes over a section of text in detail, comparing it to some sort of exemplar. A casual corrector is one who notices a variant or two, probably in the course of reading, and makes some sort of correction. A casual corrector will make only a few corrections in a manuscript, and may not be dignified with a separate superscript number.

Corrections were made in many ways. If the mistake was noted before the ink had dried and the correction would fit in the space, the first reading might be sponged off and rewritten. Most, though, were spotted later. If words were missing, a mark was made in the text and the addition written in the margin. If the words in the text were wrong, they were so marked, usually by dots above the word, and the correct text in the margin (so, if a scribe wrote "mop" for "map," the reading "mop" would be marked ṁȯṗ and "map" written in the margin or above the line). If the text was simply deleted, then the dots would be written in and no alternative written. The addition of the dots was called "expunctuation."

Ideally a corrector would go over an entire manuscript with consistent diligence, but this was not always so. Often the corrector would experience revisor fatigue. The most famous example of that is found in L^e, where the ancestral manuscript was almost entirely corrected toward the Byzantine text in Matthew but the Alexandrian text was much less affected in the later gospels.

While we usually speak of correctors as working on the text of a manuscript, there are a few instances of correctors working on the artwork as well. An early Syriac bible known as the Rabbula Gospels, for instance, has had many of the paintings retouched, presumably to make them conform more nearly with the opinions and attitudes of the correctors.

The list below describes some of the more noteworthy corrected manuscripts and the scribes who corrected them.

Noteworthy Corrected Manuscripts

The following list describes most of the manuscripts which have experienced noteworthy corrections.

• 𝔓⁶⁶. 𝔓⁶⁶ is, in terms of scribal accuracy, one of the most poorly-written manuscripts known to us. Although it contains only the gospel of John (and portions even of that have been lost), it contains roughly 450 corrections! As Colwell comments ["Method in Evaluating Scribal Habits." now published in E. C. Colwell, Studies in Methodology in Textual Criticism of the New Testament; p. 121], "Wildness in copying is the outstanding characteristic of 𝔓⁶⁶." This means that many of the corrections in the manuscript were early alterations made to correct the scribe's own errors; Colwell [p. 118] reports "𝔓⁶⁶ seems to reflect a scribe working with the intention of making a good copy, falling into careless errors, particularly the dropping of a letter, a syllable, a word, or even a phrase where it is doubled, but also under the control of some other person, or second standard, so that the corrections which are made are usually corrections to a reading read by a number of other witnesses. Nine out of ten of the nonsense readings are corrected, and two out of three of all his singular readings." (It should be noted that Colwell, p. 109, finds no fewer than 482 singular readings in 𝔓⁶⁶; this would imply that two-thirds of the corrections in 𝔓⁶⁶ correct singular readings -- an astonishing proportion. Colwell also reports, p. 111, that "two out of five [of 𝔓⁶⁶'s singular readings] are nonsense readings," leaving 289 "Sensible Singular Readings".;)
  It does appear that 𝔓⁶⁶ was eventually corrected from a different exemplar. The nature of this exemplar is difficult to determine due simply to the mass of nonsense and singular readings requiring correction. Nonetheless, the original text of 𝔓⁶⁶ seems to have been Alexandrian, and the corrections do not seem to have changed this much. (Various scholars have mentioned what they regard as "Western" readings, but most are "Western" only in the false sense "Non-Alexandrian;" many of these readings appear to be simply scribal slips.)
• ℵ. Sinaiticus is one of the most-corrected of all Biblical manuscripts; Tischendorf lists nearly 15,000 alterations (some of them involving multiple changes in the same place), and this is based only on the London portion of the text. At this rate there would have been in excess of 25,000 corrections in the entire manuscript (Old and New Testaments). It is believed that nine correctors (perhaps more) have worked on the manuscript (though not all engaged in the New Testament), dating from the time it was written to perhaps the twelfth century. For reasons of simplicity, however, a rather more limited set of sigla has been used for these correctors:
  • ℵ^a is contemporary with the scribe, or nearly (i.e. fourth century). This corrector made a relatively slight number of changes, not all of them in the direction of the Byzantine text (e.g. this corrector apparently marked Luke 22:43-44 for deletion). Hort, e.g., thought the readings of this scribe to be of value nearly equal to the original readings of the text. Tischendorf believed this copyist was one of the original copyists of the manuscript, specifically, the scribe D who wrote a few random leaves of the New Testament (probably to correct pages he felt incurably flawed).
  • ℵ^b dates probably from the fifth/sixth century. This corrector made many changes in the first few chapters of Matthew (generally bringing it closer to the Byzantine text), but did very little other work.
  • ℵ^c actually refers to a large group of scribes (perhaps five) who worked in the seventh century and made the large majority of the corrections in the manuscript. Often they cannot be reliably distinguished. The most important (and probably the first) of these is known as ℵ^c.a, who did a great deal to conform the manuscript to the Byzantine text (and not infrequently undid the work of ℵ^a). The next phase of corrections, labelled ℵ^c.b, may perhaps have been the work of three scribes, who added a few more Byzantine readings. In addition, the symbols ℵ^c.Pamph is sometimes used to refer to a scribe who worked primarily if not exclusively on the Old Testament (his corrections, in fact, seem to be confined to 1 Kingdoms-Esther), commenting that he was working from a Pamphilian manuscript, while ℵ^c.c and ℵ^c.c* refer to two minor correctors from late in the seventh century; many of their changes are in the Apocalypse. We may ignore ℵ^d; this symbol is not generally used.
  • ℵ^e refers to the last known corrector, who made a few alterations (Tischendorf reportedly lists only three) in the twelfth century.
  The current Nestle-Aland edition has simplified this notation; ℵ^a and ℵ^b are now subsumed under the symbol ℵ¹; all the ℵ^c correctors now appear in the guise of ℵ²; the handful of corrections of ℵ^e are placed under the symbol ℵ^c
• B. The corrections in B are, in a sense, far less significant than those in the preceding manuscripts. There are corrections, but they do not fundamentally change the manuscript's text-type. But in another sense, they affect the entire text of the manuscript.
  Traditionally B has been regarded as having three correctors: B¹, contemporary with the original writing; B², of about the sixth century, and B³, probably of the ninth or tenth century. (A few later corrections are also found.)
  B³ is the most important of these correctors, as this scribe retraced the entire manuscript (except for a handful of words and phrases he regarded as spurious). This scribe added accents, breathings, and punctuation at the same time. Presumably he made some reference to another manuscript during the process (since he did make some few textual changes), but the changes are slight. The primary effect of the retracing was to ruin the beauty of the ancient lettering.
  In the Nestle-Aland text, the readings of the correctors B¹ are labelled B¹, while those of B³ are labelled B².
• C. Codex C is, of course, a palimpsest, which makes it even harder than usual to assess its correctors. The fullest study of the correctors of C was made by Tischendorf, but of course this was done before ultraviolet photography and other modern techniques were available. Robert W. Lyon offered corrections to Tischendorf, but even these are regarded as inadequate. Thus the only fully current information is that offered by the apparatus to the current Nestle-Aland edition -- which is accurate but of course not complete. So all the information here must be considered tentative.
  Traditionally, C is listed as having had three correctors: C¹ (C^a), C² (C^b), and C³ (C^c). C¹ is the symbol used for the diorthotes. However, there are no readings which can be attributed with certainty to this corrector, and many scholars omit this hypothetical scribe from the list.
  The existence of C² and C³ can hardly be denied, however, as each made some hundreds of corrections to the text. (The Nestle-Aland text shows about 251 corrections by C² and about 272 by C³). C² is believed to have worked in the sixth century, possibly in Palestine; C³ worked in the ninth century, perhaps at Constantinople.
  Neither corrector was really thorough. Both seem to have alternated between moderate attention and extreme inattention. This is particularly true of C³, who all but ignored large fractions of the text. For example, C³ offered only three corrections in the Catholic Epistles and only 20 corrections in Mark. The table below summarizes the extent to which the two correctors worked on various parts of the New Testament (the Apocalypse is omitted because NA²⁷ shows only 3 corrections of C in that entire book! All numbers are approximate).
  

  Book/Section	C2	C3
  Matthew	33	42
  Mark	48	20
  Luke	31	42
  John	49	89
  Acts	21	24
  Catholics Epistles	26	3
  Pauline Epistles	41	51

  
  The text of C³ is almost purely Byzantine. That of C² is more complex. The Byzantine element is still dominant, but there are occasional readings which go against the Majority Text. Few of these agree with the earliest Alexandrian witnesses, but they are often shared with late Alexandrian manuscripts.
• D^ea/05. Codex Bezae is unique. (Oh, you knew that?) No other manuscript departs so far from the New Testament norm. It is a testimony to the value of manuscripts, and the effort required to make them, that it was preserved and repeatedly corrected, rather than thrown away.
  Scrivener counts a total of fifteen correctors who worked on the manuscript; nine worked on the Greek side (the others confined their attention to the Latin or the margins). The earliest of these is contemporary with the writing (the original scribe occasionally sponged and/or scraped away errors); the last dates from the eleventh or twelfth century. Gregory summarizes the earliest of these as follows: "The first one made about 181 changes in a careful beautiful hand in the sixth century. The second was probably of the seventh century, and made about 327 changes, besides adding some spiritus and accents and other signs. The third, it may be towards the end of the seventh century, made 130 changes, and the fourth, of the same age, 160 changes, mostly in Acts" (The Canon and Text of the New Testament, p. 352).
  Scrivener, naturally enough, designated the various correctors by the letters A through M (the use of twelve letters -- I/J are treated as one -- is explained by the fact that correctors E and G worked only on the Latin side). In Tischendorf's edition this was simplified; D^A becomes D¹, D^B and D^C retain their symbols; the rest are subsumed as D². In the Nestle text this is further simplified; the early correctors D^A, D^B, D^C, and D^D are summarized as D¹; the middle correctors (D^F, D^H, D^J, D^K, and D^L, all of around the ninth century) are given the symbol D², and the eleventh/twelfth century corrector D^M becomes D^c.
• D^p/06. Codex Claromontanus resembles Codex Bezae in many ways. It is a diglot, it dates from about the sixth century -- and it has been heavily corrected. Tischendorf distinguished nine correctors, though only four were really significant. These four he assigned the symbols D^b (D**, seventh century?), D^c (D***, ninth century; whom Tischendorf regards as actually the fourth corrector. It should be noted that Tischendorf often marked corrections D^{b et c}, indicating that this corrector agreed with D^b), plus the nearly-contemporary correctors D^d (D****) and D^nov, which must be after the ninth century. (In the Nestle-Aland text, D^b becomes D¹, D^c becomes D², and D^d and D^nov together constitute D^c.)
  Of these, the most significant was the ninth century corrector (Nestle-Aland's D²), who, according to Scrivener, made "more than two thousand critical changes in the text, and added stops and all the breathings and accents." The text used by this corrector, as might be expected, was almost entirely Byzantine.
• H^p (015). H is not as noteworthy for its corrections as for their claimed source. Originally written in the sixth century, some centuries later a second hand went over the manuscript adding accents and breathings as well as badly retracing letters. Of greater interest is a note affixed to the end of Titus. This claims that the manuscript was corrected from a manuscript written by Pamphilius and kept at Cæsarea. (The wording of the note is εγραψα και εξεθεμην κατα δυναμιν στειχηρον^. τοδε το τευχος παυλου του αποστολου προς εγγραμμον και ευκαταλημπτον αναγνωσιν. των καθ ημας αδελφων^. παρων απαντων τολμης συγγνωμην αιτω. ευχη τη υπερ εμων^. την συνπεριφοραν κομιζομενος^. αναβληθη δε η βιβλος^. προς το εν καισαρια αντιγραφον της βιβλιοθηκης του αγιου παμφιλου χειρι γεγραμμενον αυτου). This note is dated by Tischendorf to the seventh century -- i.e. to a date after the manuscript was written. However, it seems almost certain that the note is either wrong or misunderstood. It is highly unlikely that a Pamphilian manuscript would have a purely Byzantine text -- but the handful of surviving corrections in H that involve a change of text (as opposed to spelling, accents, etc.) -- will be seen to be almost invariably Byzantine, with the others being perhaps from the Lectionary (1799 also has lectionary readings). Readings marked * are not in the Nestle apparatus, and so have been given in full; for the other variants listed here, the reader is referred to NA²⁷:
  • 1 Cor. 10:28 -- H* with ℵ A B C* D F G P 33 81 365 630 1175 1739 1881; H^c with K L Byz
  • 2 Cor. 11:28 -- H* with P⁴⁶ ℵ B D F G 0243 33 81 1175 1739 1881; H^c with I^vid K L 0121 Byz
  • *Col. 1:29 -- H* δυναμει with P⁴⁶ ℵ A B C D F G K L P 330 436 1739 Byz vg; H^c adds θεου (I know of no other support for this reading)
  • Col. 2:7 -- H* with ℵ* 33 81 1175 1739 1881; H^c with B D² K L Byz
  • *Col. 3:4 -- H* οταν with P⁴⁶ ℵ A B C D F G K L P (330 οταν ουν) 436 Byz vg; H^c 1799 αδελφοι οταν read (from the lectionary?)
  • 1 Tim. 1:13 -- H* with ℵ A D* F G I P 6 33 81 365 1175 1739 1881; H^c with D² K L Byz
  • 1 Tim. 1:17 -- H* with ℵ* A D* F G 33 1739; H^c with ℵ² D¹ K L 1881 Byz
  • 2 Tim. 2:3 -- H* with ℵ A C* D* F G I P 33 81 365 1739 1881*^vid; H^c with C³ D¹ K L Byz
  • Heb. 1:3 -- H* with ℵ A B D¹ P 33 81 1175; H^c with (P⁴⁶) D(*)^,2 K L 0243 1739 1881 Byz
  • Heb. 10:34 -- H* with P¹³ P⁴⁶ ℵ* A D H* 33 1739^c?; H^c with ℵ^c D² K L 1739* 1881 Byz
  • Heb. 10:38 -- H* with P⁴⁶ ℵ A 33 1739; H^c with P¹³ D² I K L 1881 Byz
• 424. 424 is the only minuscule known to have been heavily corrected. There were actually three stages of correction (denoted simply 67** in Tischendorf, and 424** by Souter, etc., but in K. Aland et al, Text und Textwert der griechischen Handschriften des Neuen Testaments the hands are distinguished as 424¹, 424², and 424³). Of these, the second set of correctors are by far the most important, introducing thousands of changes (especially in Paul, but also in the Catholics; the Acts are relatively unaffected).
  Even more interesting than the fact of these extensive corrections is their nature: instead of its corrections moving the manuscript toward the Byzantine text (as has taken place in every other heavily corrected manuscript), the changes in 424 move it away from the Byzantine text and toward the text of Family 1739 (especially toward 6).

Almost all other manuscripts contain corrections, of course. But few if any contain corrections such as those found in the manuscripts listed above, which actually change the nature of the manuscript. Descriptions of these manuscripts are therefore omitted.

The Significance of Corrections

Most corrections in most manuscripts merely correct slips of the pen. These are usually obvious, and have no textual significance. But the manuscripts listed above are another matter. ℵ, D, and 424 in particular were clearly corrected against manuscripts of completely different types.

This forces us to look at exactly what we know about those other manuscripts. Can we exactly reconstruct their texts? The answer is no. If the corrector leaves a reading alone, we cannot be certain that the manuscript he worked from actually agreed with the manuscript in our hands. The corrector may simply have ignored the alternate reading, either accidentally or on purpose. The useful readings are the corrections, not the uncorrected portions. (There is an analogy to this in Shakespeare criticism, in the many cases where the handful of witnesses are partially but not fully independent. Where the semi-independent witnesses agree, the reading is actually less weighty than where the disagree, because the agreement may be coincidence in error, but where they disagree, it is nearly sure that at least one witness is correct.)

So, in assessing a corrected text, we should examine first the corrections in isolation and only then the text as corrected. 424 proves this point well: Examining the corrections shows us that the direction of the corrections was toward the 1739 type of text; looking at 424-as-corrected shows a clear kinship to 6. This is obvious when examining both; it is less clear when examining either in isolation.