The NA28 is Here. But Don’t Scrap Your 27th Just Yet.

Peter Williams on the Nestle-Aland Novum Testamentum Graece 28

Nestle-Aland, Novum Testamentum Graece, 28th rev. ed., Deutsche Bibelgesellschaft, 2012. 14 + 94* + 4 + 890 pp. $59.99

If I were just allowed one book to assist my study of the New Testament, this edition, with its 114 year history, would be it. The Institute for New Testament Textual Research in Münster (INTF) is to be congratulated for its immense labours in producing what is arguably the single most useful volume for NT studies of all time.

The 28th edition of the Novum Testamentum Graece differs from all previous editions in being simultaneously an online publication. We are even promised downloadable editions for PC, Mac, and smartphones in 2013. This review primarily pertains to the printed edition (hereafter NA28) and also highlights differences from the 27th edition (hereafter NA27).

1. The main changes

The text of NA28 differs from NA27 in relatively minor ways in 34 passages in the Catholic Epistles because NA28 follows the text of the second edition of the Editio Critica Maior, and the ECM is currently only available for these books. In accord with the ECM the word ἐπιστολή is also added to the title of all seven Catholic Epistles. For the rest of the NT, ‘the text of the 27th edition has been retained’ (p. 54*), though some small exceptions to this claim will be noted below.

The ECM of the Catholic Epistles is based not only on a thorough analysis of the witnesses, but also on the INTF’s own Coherence-Based Genealogical Method, which provides a sophisticated system for checking the consistency of one’s textual decisions. The revision of the apparatus is more thorough in the Catholic Epistles because there the ECM’s analysis of the manuscripts is available.

In the rest of the New Testament the apparatus has been revised in various ways, which generally increase accuracy, readability, and transparency:

  • References have been checked and different witnesses appear in the apparatus. Thus in 2 Timothy 1:10 manuscript 81 was cited by NA27 in support of txt’s Χριστοῦ ’Ιησοῦ and is now cited in support of the opposite variant ’Ιησοῦ Χριστοῦ; the disgraced manuscript 2427 has been removed from the apparatus of Mark; the latest papyri (numbers 117-127) are cited; some manuscripts previously subsumed under the Majority text symbol unless they disagreed with the Majority text, are now consistently cited separately.
  • The distinction – found confusing by some – between consistently cited witnesses of the first and second order has been dropped.
  • Conjectures have been dropped from the apparatus.
  • The references to versional readings have been considerably reduced, especially in light of criticisms that NA27 was overly certain of the readings behind versions.
  • References to correctors, especially of Codex Sinaiticus, are significantly more precise than previously.
  • The apparatus now contains much more information on variants of accentuation which affect the sense. Sometimes this newly gathered information stands rather strongly against the printed text, e.g., 1 Corinthians 3:14, where only mss 81 and 1505 and allegedly the Latin support NA28’s future μενεῖ and an impressive range of witnesses support the present μένει. In Hebrews 5:12 only mss Ψ and 81 (again!) support NA28’s τινά against the well attested reading τίνα.

2. Comparative statistics

  NA27 NA28 Percentage change (nearest integer)
Approximate weight of blue edn 489g 517g +6%
Pages 920 1002 +9%
Pages of biblical text 680 789 +16%
Pages of Catholic Epistles 44 50 +14%
Pages of Mark 62 75 +21%
Lines of Mark’s text 1487 1484 0%
Lines of Mark’s apparatus 1060 1378 +30%
Variant units in James 204 213 +4%


In Mark’s gospel, we see that the 21% increase in pages has not been accompanied by any more lines of text. The number of words per line is essentially identical, despite NA28’s slight change of font. Moreover, the interline spacing and the margins of the printed text have not changed. The increased number of pages is therefore caused by the 30% growth in the apparatus. Add to that the fact that the apparatus font is larger in NA28 and the interline spacing of the apparatus greater and you will see why NA28 gives a different visual impression from NA27. NA28’s apparatus takes up considerably more of the page.

It is difficult to assess whether the overall number of variants has changed between the two editions because, in the interests of readability, NA28 has recorded as separate variants many cases where NA27 had recorded subvariants. There may be a slight increase in variants in the Catholic Epistles, but outside these epistles the choice of variants seems to have remained basically constant aside from deletion of conjectural and some versional variants and the separate presentation of erstwhile subvariants.

3. Versional citations

The versional citations are much reduced and their overall accuracy thereby greatly improved. Nevertheless, there are still instances where versions are cited when their witness is actually ambiguous. For instance, how can one cite the Peshitta in support of δέδοται αὐτῇ as opposed to αὐτῇ δέδοται in 1 Corinthians 11:15? Is it really reasonable to infer that the Latin versions in Colossians 2:12 read βαπτισμῷ not βαπτίσματι, when they appear to use baptismus as an equivalent of βάπτισμα elsewhere (e.g., Matthew 3:7)? The Syriac should not be cited to support variants relating to the order ‘Jesus Christ’ versus ‘Christ Jesus’ as in 2 Timothy 1:10. These are just the tip of the iceberg of notes which could be questioned. This illustrates that there is still a substantial work to carry out on the translation technique of the early versions and its relationship to textual criticism though the INTF is already publishing such investigations.

4. Don’t throw away your copy of NA27

With the general removal of readings supported only by versional witnesses NA28 may give users less of an impression of the textual character of witnesses than its predecessor. At Matthew 1:7-8 readers of NA27 learned that syc (the Curetonian Syriac) inserted the names of Uzziah, Joash, and Amaziah into Jesus’s genealogy or that in Matthew 1:21 syc read ‘world’ for ‘his people’. The problem with NA27’s notes was their implication by using Greek that syc attested a Greek variant, which might or might not have been the case. With NA28’s removal of these notes, the new edition is arguably more accurate, but the reader learns less about syc. Similarly NA27’s notes giving a variety of versional interpretations of ἐπιούσιον in Matthew 6:11 are omitted from NA28. Thus NA27 seems generally to give more information relating to the history of reception of the NT text than NA28. So alongside NA28’s improved accuracy there is some associated loss of information.

The references to pc (pauci) and al (alii) which in NA27 indicated varying numbers of unidentified manuscripts have been dropped. This is in line with the general strategy of improving the quality of data and ridding the edition of fuzzy information. A downside to this is that one now no longer knows whether a particular reading is supported by a small or rather substantial body of manuscripts. For quick orientation to the witnesses in a variation unit I will reach for my NA28 first, but might still regularly consult NA27.

What I will most sorely miss from NA27 is the subscriptio at the end of each of Paul’s letters. NA27 records interesting manuscript traditions about the geographical origin and occasion of Paul’s letters. This is gone in NA28. It is a shame for New Testament scholars to be able to forget so easily that, whereas a range of scholars now reach the mutually convenient conclusion that Galatians was one of Paul’s earliest letters (this generally suits radicals as well as conservatives), the earliest traditions set the epistle in Rome, and therefore later in Paul’s career. One may consider this datum wrong, but one should not be ignorant of it. For such information keep your NA27 in reach.

5. Orthographical and typographical changes

The apparatus in NA28 is easier to read than in NA27. The latter’s confusing abbreviations of readings have been dropped, and overall there is less Latin in the edition.

The moveable nu has been globally supplied where it was lacking in NA27. Thus πᾶσι becomes πᾶσιν (Matthew 2:16); βλέπουσι becomes βλέπουσιν (Matthew 18:10); ἔδοξε becomes ἔδοξεν (Luke 1:3). Manuscripts are inconsistent, but if in Hebrews 9:1 our earliest witnesses (P46 01 and 03) all have εἶχε was there really any good reason for NA28 to print εἶχεν? As standardization is also introduced in the apparatus, in Matthew 15:36 ms 700 went from being NA27’s only witness for the reading ελαβε to being NA28’s only witness for the reading ελαβεν.

The most frequent change to spelling is that ἀλλά ‘but’ is now universally abbreviated to ἀλλ’ when it precedes a vowel (e.g., Matthew 8:4; 9:13, 18; 16:12; 17:12; 18:22, 30), even before a punctuation mark (Romans 9:7). Thus in 2 Corinthians 7:11 NA27’s ἀλλὰ ἀπολογίαν, ἀλλὰ ἀγανάκτησιν, ἀλλὰ φόβον, ἀλλὰ ἐπιπόθησιν, ἀλλὰ ζῆλον, ἀλλὰ ἐκδίκησιν has become ἀλλ’ ἀπολογίαν, ἀλλ’ ἀγανάκτησιν, ἀλλὰ φόβον, ἀλλ’ ἐπιπόθησιν, ἀλλὰ ζῆλον, ἀλλ’ ἐκδίκησιν despite the advice of P46 01 and 03 and the fact that this ruins the oral effect of the repetition of ἀλλά (cf. changes in 1 Corinthians 6:11).

This global editorial change is not fully integrated with the apparatus so that in 2 Corinthians 4:5 we have ἀλλ’ with no variant immediately before a variant between ’Ιησοῦν Χριστόν in the text and Χριστὸν ’Ιησοῦν in the apparatus. In this case it is not possible to select Χριστὸν ’Ιησοῦν from the apparatus and put it into the text and produce proper Greek.

These spelling changes may interest only a few, such as those investigating oral features of the text. Nevertheless, it strikes me as odd to make a global decision to print the shortened form when in some cases all the earliest witnesses have the long form. For instance, the INTF’s own data tell us that in John 4:23 P66 P75 01 02 and 03 have ἀλλὰ ἔρχεται and yet NA28 prints ἀλλ’ ἔρχεται. Such changes from NA27 only add to the effect of increasing the homogeneity of spelling in the GNT, which is already too homogeneous in our editions. Students learning Greek will also get a misleading impression that only one form is standard morphology.

NA28 and its predecessor use italics for quotations from the Old Testament. The marginal references to the Old Testament have been thoroughly revised, and consequently what is italicized has changed. Matthew 13:32 and 17:10-11 thus read rather differently with the italics of NA27 and without them as in NA28.

Capitalization has changed slightly between the two editions (e.g., NA28’s Χριστός vs. NA27’s χριστός in Matthew 1:16), but there are few such changes. Sentences after a small space, which in NA27 began with a lower case letter, now begin with a capital (e.g., Matthew 25: 13, 19, 22, 24, 26, 29, 34, 41).

6. Increasing Conservatism

Since the text of NA28 outside the Catholic Epistles is the same as that of NA27 (1993) which in turn had the same text as NA26 (1979), which had the same text as the third edition of the United Bible Societies (1975), it is worth noting that the textual decisions for most of the NT were made in the early 1970s. NA29 will be some time hence, and if future Nestle-Aland editions await the publication of the ECM before changes are made to the main text, we may be looking at half a century or more between the original textual decisions and subsequent revisions.

It is understandable that the INTF do not want to make hasty decisions when they have such long-term plans of doing a thorough job. Nevertheless, with increasing numbers of papyri and with better understanding of scribal habits, it is hard not to feel that improvements to the basic text could be widely agreed. The edition is being made with such caution that there is a slight danger that it could become a victim of that natural caution which arises when such a good reputation has been established. I do not think that Eberhard Nestle or Kurt Aland were they with us today would have held back their hand from making interim textual decisions while waiting for the ECM to appear.

As the text has not changed for so long and the apparatus is now larger, more reliable, and more legible than ever before it is especially imperative that NT scholars do not abdicate their own responsibility to consider variants and when necessary to depart from the decisions of these esteemed, hardworking, and praiseworthy editors.