1. Introduction
The relevance of linguistic consistency needs to be considered as a working criterion in that most essential of Biblical disciplines, namely textual criticism, and one has to ask how far one ought to expect such consistency from the biblical authors, or from the scribes who copied their writings over the centuries or even from modern editors working from the manuscripts that have chanced to survive. A reader of Luke needs to ask at 22.43–4 whether the content and language of these disputed verses could or could not have been written by the author of the Gospel. A similar query is regularly posed about other textual variants involving longer versus shorter readings such as the endings to Mark, or the pericope of the Adulteress. The precise words allegedly spoken by Jesus at the institution of the Last Supper in Luke, the exact wording of the Paternoster one is to repeat or, of equal importance to certain readers, Paul's teaching on the nature of resurrection in 1 Cor 15.51, the many differences in the narratives of Acts or the variant concerning an issue as fundamental as the Ascension at the end of Luke are all among the most obvious text-critical cruces. Attempted solutions to variants such as these often hinge on the consistency of language and usage expected from the author. These raise their own questions. Are we, twenty centuries after the original compositions were written, able to detect and restore an assumed consistency to our first-century authors, using our arbitrary collection of extant manuscripts? And how far may we plot any changes to their wording that may have taken place? These are of course big questions that span centuries and cultures. What we may expect from a first-century creative author and what we may find to be the practices of a professional scribe or even of a hack copyist need not be identical. The earliest centuries of the written copies of the New Testament books were times of continuous reinterpretation,Footnote 1 One needs to test if an author's consistency survived such use and to remind oneself that the New Testament books were working texts and no mere fossils to be preserved unchanged.
All authors have a consistent and distinctive usage and style.Footnote 2 Their consistency may, indeed should, be determinative in resolving textual variation in our witnesses. An acknowledgement of this consistency and a thoroughgoing application of such a criterion will require many scholars to alter commonly held and traditional text-critical views about the favouring of the ‘best’ manuscripts or the privileging of witnesses of a certain age. It may even mean a reassessment of the role of conjectural emendation. Such issues are dealt with below.
The Introduction to Nestle27 pp. 46*–47* tells readers that the edition provides ‘a well-founded working text together with the means of verifying it or alternatively of correcting it’. (Strangely, such a helpful disclaimer is not in the introduction of Nestle28.) Such advice should serve as a warning to those readers who abjure their responsibilities by ill-advisedly relying on one recently assembled running text. But, when heeding the invitation in Nestle27, textual critics need to assemble an author's distinctive linguistic characteristics. They may do so, utilising as full an apparatus criticus as practicable and noting initially those places where no variation has been recorded from collations. The methodology sometimes dubbed ‘thoroughgoing eclecticism’ concentrates on an author's language and usage and on features of Koine Greek to establish such characteristics and from this analysis to resolve problems in places of textual variation. Doing this often requires three older principles to be rejected and those may be summarised under the headings ‘quantity’, ‘age’ and ‘quality’:
(1) Quantity. The number of manuscripts supporting each variant is of only minor importance. The bulk of witnesses, the so-called Majority Text and the many Byzantine minuscules, need not possess the monopoly of truth, that is, ‘original readings’. In thoroughgoing textual criticism such democracy does not prevail. The earliest readings may have survived in only a handful of witnesses.
(2) Age. Another of the dominant features of modern-day textual criticism is that less emphasis is placed on the age of manuscripts. The relatively recent Coherence Based Genealogical Method (= CBGM), now being applied by the Institute of New Testament Textual Criticism at Münster, Westphalia, has coined the noun ‘tradent’ to describe a manuscript as merely the carrier of a reading, and in many cases such a reading will be considerably older than the palaeographical date assigned to the manuscript transmitting it. The age of a manuscript need be of no special significance, once one remembers that very few extant witnesses predate the acceptance of the canonical status of the books, say in the fourth century. Even in our oldest papyri there are likely to be many stages of copying preceding every current manuscript, and no one knows what or how many changes, deliberate or accidental, happened at each copying.
(3) Quality. Nor is the ‘cult of the best manuscripts’ relevant any longer, i.e. that practice of putting one's specific Alexandrian eggs, or one's Byzantine or other named eggs, into one basket that is then said to be the carrier of the earliest, best, ‘original’ text.
Rather, one should be prepared to accept readings as ‘original’ in any manuscript or manuscripts that happen to preserve a reading compatible with, and consistent with, the detectable and proven usage of the author. This will be demonstrated with examples shortly, but, first, we need to state what it is that we are privileging. It is the earliest recoverable text, often called the original text, although that word, ‘original’, is fraught with dangers. Text-critics now use the jargon term Ausgangstext, i.e. the initial, published text that may or may not conform to the authorial text; that is what is being sought.Footnote 3 It is the earliest recoverable text from which all known textual variations can reasonably be plotted.
This recent hesitation to use the term ‘original’ arose from Eldon J. Epp's warningsFootnote 4 and is in contrast with Westcott and Hort, who in 1881 could entitle their edited text ‘The New Testament in the Original Greek’ (emphasis added). Such overweening confidence has no place in the textual criticism of today. With 5,000 or so Greek New Testament manuscripts at our disposal, it may be assumed that we are well equipped to locate at virtually every point the original, authorial text although some claim that such a goal is chimeric. Nonetheless, for much of the New Testament text we are able, thanks in part to the evidence provided by undisputed areas of text, to proffer a relatively stable Ausgangstext, which in large measure may represent the authorial Greek of its creative composer. Using as touchstones a proven literary and linguistic style and usage in any given book provides demonstrable and helpful aids for resolving textual variation and (just as importantly) for explaining the origins and, perhaps, motives for subsequent scribal alterations. In most cases that consistency of style will go back to the creative author of the composition.
In an earlier time and culture, when all that the text-critic was apparently trying to achieve was one indubitably authorial and original text, variations from that gold standard were often rejected and ignored as ‘secondary’; now these alternatives are to be seen not as mere blunders or the consequences of maverick rewritings. Rather, much church history and Christian theology may be exposed in an apparatus, in the recognition that the manuscripts supporting even what may commonly be agreed are secondary, later readings were once the canonical, authorised texts of the communities or individuals who used, read and revered the manuscripts containing their scriptures. The current shift away from trying to establish exclusively the author's published text comes alongside an allied recognition that all alternatives merit attention as stepping stones in the dense and varied history of the New Testament.
2. A Small Selection of Variants
2.1 Mark
(a) What name does Mark give John the Baptist? Our editions usually call him βαπτιζων at Mark 6.14 (v.l. βαπτιστης D S W Θ) and 24 (v.l. βαπτιστου A C D W) but βαπτιστης at 6.25 as well as 8.28. Variants exist at 6.25 giving a form of the participle (in L 700 892) and at 8.28 (in 28 565). Consulting the text of NA28, we may well ask why Mark changed the substantive in two contiguous verses (6.24, 25). But thanks to the variants, we can restore a consistent usage throughout Mark.Footnote 5 The newly coined Christian noun βαπτιστής is found in Matthew, Luke and in later writers. The earlier form, based on the verb βαπτίζω, should stand throughout Mark. The change to the name, albeit made erratically, comes from scribes encouraged by later and normal church usage.
Our decision here may help resolve the more important variations at Mark 1.4 when John on first mention may naturally be defined by his distinctively Markan name, ο βαπτιζων. If so, that would allow us to read εγενετο κηρυσσων with this name as its subject (cf. Mark 9.7), giving us a periphrastic ‘John the baptising one was preaching …’ or, probably better in this context, ‘John the baptising one appeared, preaching …’ The reading with the article should be accepted as the Ausgangstext; we also follow the reading of Β 33 in omitting και before κηρυσσων.
(b) There is a problem with the translation of Mark 6.22, which can also be resolved by an appeal to the variants and a recognition of our author's usage. Is the dancing girl Herod's daughter,Footnote 6 or Herodias’ daughter with the same name as the mother,Footnote 7 or is she an unnamed daughter of Herodias? If we follow A C K Θ reading θυγατρος αυτης Ηρωδιαδος, meaning, not necessarily, ‘the daughter of Herodias herself’ but ‘her daughter, that is to say Herodias’ (daughter)’, such a usage agrees with Markan style, which frequently includes similar explanatory parentheses (2.10, 15, 22; 6.14–15; 7.2, 3–4, 19, 26; 8.15; 13.12a; 13.14; 14.36). Scribal attempts to remove the parenthesis or to delete what may have been seen as a redundant pronoun were probably responsible for the other readings. Markan usage should be determinative: the girl is the anonymous daughter of Herodias.
(c) Mark 10.1. We read in NA28 that ‘crowds’ not ‘a crowd’ approach Jesus. That is a unique occurrence of the plural of this noun in Mark. But there is a variant (not in NA28) giving the singular. We are looking not only at οχλος/ οχλοι, either of which could have been written accidentally for the other by an inattentive scribe, but also at the associated verb. Rather, this is a deliberate change: συμπορευονται … οχλοι read by א B and συμπορευεται … οχλος read by D Θ W fam.13 pauci Lvt aur k q. From what I observe, most commentaries on Mark fail to discuss this variant.Footnote 8 In looking for consistency we see that οχλος (singular) occurs elsewhere ca. 38 times in Mark and nearly all of them are firmly established. In Luke, Acts, John and Revelation the noun is also singular; in Matthew it is frequently plural (with its thirty plurals and twenty singulars, including v.ll. at Matthew 8.18; 15.31). As an indication that there is no inherent difference in meaning, we may see from our synopses that the singular at Mark 4.1pr. parallels the plural in Matthew 13.2: cf. similarly at Mark 3.9 // Matthew 12.15 plural and v.l. om.; Mark 5.24 // Lk 8.42 plural; Mark 5.31 // Luke 8.45 plural etc.
I suggest we read the singular noun and verb at Mark 10.1 as being consistent with Mark elsewhere and comparable to Mark 3.20. The plural form is likely to have arisen as a scribal harmonisation to the parallel in Matthew 19.2 or ad sensum (picking up αυτους in Mark 10.1). Once this change was made, few readers thereafter would have noticed the inconsistency with the rest of Mark thereby created.
2.2 Two Proper Names
(a) Jerusalem. The New Testament in general, and most prominently Acts, fluctuates between the two different names for Jerusalem, the Hellenised form Ιεροσολυμα and the Hebrew name, transliterated as Ιερουσαλημ. Printed editions of the Greek New Testament present a chaotic picture with an apparently feckless and reckless use of the two within the same writing. Consequently, exegetes struggle to explain this apparent capriciousness as the author seems to flit from one form of the name to the other without rhyme or reason. Variants commonly substitute one form for the other. There are too many examples to merit space now but we need to say that these are precisely names that an author like Luke may well have carefully chosen for appropriate use at each point in his writings. Authors may be fastidious using a particular name in certain contexts or when addressing a particular language group. I once attempted to establish the firm, undisputed occurrences of both names for Jerusalem to resolve which noun should stand appropriately in differing contexts in Acts.Footnote 9 More recently Jenny Read-Heimerdinger and Josep Rius-Camps have used a different methodology, concentrating on the so-called Alexandrian text-type in Acts and the ‘Western’ text-type to resolve these troubling discrepancies.Footnote 10 We arrive at somewhat differing conclusions but we both apply a consistency of usage to determine the reading to be chosen at each point of variation. Without some effort to resolve the confusion by selecting the appropriate reading at each variant, one would have to admit that there is anarchy in the author's usage or in the manuscript tradition.
(b) Another proper name that cries out for a consistent usage is Ναζαρηνος/Ναζωραιος. Whatever the etymology of the two forms of the name, the evangelists assume that they refer to Jesus’ home town – he is ‘the man from Nazareth’. Personal preference by the evangelists seems to have determined which was used. Mark seems to prefer Ναζαρηνος, although both forms appear as variants at Mark 10.47; Matthew, John and Acts prefer Ναζωραιος.Footnote 11 According to the printed Greek New Testament, Luke apparently used both. We find the vocative Ναζαρηνε at 4.34, which is firm according to our printed editions, as well as Ναζωραιος at Luke 18.37 and 24.19, but both of these have variant readings (printed in NA28), the first of which (18.37) parallels Mark 10.47! Consistency of usage urges us in one direction in Luke's Gospel albeit with hesitation, because his second volume, Acts, has Ναζωραιος.
2.3 Three Examples from the Pastoral Epistles
(a) 1 Timothy 3.16. This is probably the best known of the textual variants in these letters.Footnote 12 It concerns whether the hymn quoted here opens with the pronoun ‘who’, which lacks an antecedent, or with the name ‘God’ which seems to be the noun expected. The variants are:
o D* Latin Fathers (The UBS Greek New Testament 5th edition (= UBS5) includes Severian, Hilary, Pelagius, Augustine)
ος א* A* C* F G 33 365 1175 Didymus Epiphanius
[Note that the first three manuscripts are early and that the readings of the original hands have been changed by subsequent scribes.]
ΘεοςFootnote 13 א3 (12th century!) Ac C2 D2 Ψ Maj.
[Prior to the late 4th Century no Patristic writer knows this reading.]
ος is likely to have been original; the v.l. o assumes the originality of ος. The change to o was made on grammatical grounds as the antecedent is the neuter noun immediately preceding. ος may reflect a constructio ad sensum: the mystery is the incarnation of God in Christ. More probably, though, the masculine pronoun without an antecedent is due to the verse's having been taken out of its original context in an early Christian hymn where the verse preceding could have shown an appropriate antecedent. A supplementary but often overlooked argument favouring ος is that Θεος in the nominative is arthrous throughout the Pastorals, e.g. at 1 Tim 4.3; 2 Tim 1.7; 2.25; Tit 1.2. Were we to have argued for the originality of Θεος then that would have created an uncharacteristic anarthrous use of this noun.
(b) 2 Timothy 4.8 om. πασι D* 6 1739* 1881 Lat syp and many Latin fathers; πασι cett. The omission could have been accidental, having been caused by homoeoteleuton (καΙΠασΙΤοις). Note that NA28 and UBS5 print πᾶσιν as the text contra NA27 and UBS4. If the movable nu were indeed written then the omission by homoeoteleuton would have been less likely. But, more probably, om. πασι was a deliberate omission. The word was regularly avoided by scribes (pace Metzger, Commentary, p. 581, who claims that copyists often added the word to enhance an account). In the Pastorals omissions of πας may be found in manuscripts at 2 Tim 1.15; 2.10; 4.21. Πας + article + substantive is a Semitism found in the Pastorals at 1 Tim 6.10; 2 Tim 3.12; 4.17. Here πασι(ν) is original and it balances ου μονον δε εμοι.
(c) At Tit 1.4 we find the following variants:
(i) χαρις και ειρηνη א C* D F G Ψ Lvt Lvg syp copt Chrys
χαρις ελεος ειρηνη A C2 Κ L Maj.
(ii) XY IY א A C D* 088 0240 33 81
IY XY 1739 1881
KY IY XY D2 F G K Maj.
[cf. 1 Tim 1.2: χαρις ελεος ειρηνη απο θεου πατρος (ημων in א2 D2 Maj.) και XY IY του KY ημων and 2 Tim 1.2 χαρις ελεος ειρηνη απο θεου πατρος και XY IY (in א2 A D F G K L Maj.) or KY IY XY (in א 33) or IY XY (in 629 1739 1881) του σωτηρος υμων]
On (i) Metzger, Commentary, at Tit 1.4 says that the shortest form is original because it conforms to Pauline usage, where χαρις και ειρηνη is common. But our author is not Paul. So the longer form may be original in the opening greetings in all three letters and the usage would then be consistent within the Pastorals. When we turn to (ii) the problem of the choice between a longer and a shorter version of the divine names may be resolved by stating that the original is the shortest reading and that an expansion is a pious addition. Brevior lectio potior is seldom a reliable maxim, but here it is likely to be correct, i.e. omit Kυριου.Footnote 14
2.4 Two Examples from Revelation
Revelation is a good place to practise thoroughgoing textual criticism. The Apocalypse flies in the face of the textual criticism normally practised by mainstream eclectic critics in the rest of the New Testament. The old, now discredited and largely abandoned, text-types barely worked for Revelation in the same way as it was claimed they helped us for the rest of the New Testament. Even adherents of the Byzantine text are forced to adopt differing principles as their beloved Majority Text is often divided in Revelation.
(a) Rev 10. 2, 8, 9 and 10. There is a nest of variants, based on the word ‘book’, βιβλίον, with this and other diminutive forms of this noun throughout this chapter. Thus we have v.ll. βιβλιδαριον, βιβλιδιον and βιβλαριδιον. Our printed editions are inconsistent in this choice of noun. Do we follow blindly a majority of manuscripts, or accept the readings in the oldest manuscripts, or stick by our favourite manuscripts, despite their fickleness in such matters, and allow the printed text thereby created to have all the inconsistencies revealed in this chapter in NA28, for example? The answer to all three questions is ‘No.’ We argue that our author was consistent in his use of the word ‘book’, assuming (as we may) that all forms are identical in meaning.Footnote 15 I suggest the diminutive form with suffix -αριδιον commends itself throughout. Atticising scribes in particular would have avoided this post-classical suffix and altered it to the classical -ιδαριον or to -ιδιον or even -ιον. Thanks to a restoration of the Koine form, consistency can therefore be restored to the chapter.
(b) Rev 14.8. The dividing of manuscripts over the readings αλλος αγγελος δευτερος and αλλος δευτερος αγγελοςFootnote 16 ensures that this variant is relatively conspicuous even in apparatus critici as truncated as that in the Nestle editions. Again, may author's established usage be appealed to in order to resolve the dilemma? It seems as if the author of Revelation places a second adjective to follow a noun that is preceded by αλλος (as at 14.9 and see also 6.4; 10.1; 15.1). Thus an appeal to consistency seems justified here but it should also be applied in other variation units where the manuscripts are less evenly divided in their attestation.
Obviously, as an increasing number of manuscripts is read and collated, it may well be that some erstwhile firm and fixed examples will become insecure, once hitherto unknown variants are exposed.Footnote 17 In practice, though, I observe that very few genuine new variants are emerging from recent collations. Old variants may receive buttressing and that is often very welcome, especially in places where a well-known variant is but sparsely represented in our manuscript tradition. En passant it is worth reminding ourselves of Metzger's two studies, one of textual variants known to and commented upon by Jerome and the other by Origen.Footnote 18 In both articles Metzger is able to point to places in these patristic writers’ comments where they allude to a variant which is said to be found in the bulk of manuscripts known to them but which, due to the vagaries of transmission and survival, is today found in only a minority of witnesses. The lesson to draw from this is that an allegedly weak reading, i.e. one supported by very few witness, deserves as much attention as a reading supported by ‘many’ manuscripts.
3. Conjectures
Many scholars may accept the principle that a reading which is consistent with the writer's style is likely to be ‘original’. But if that reading transpires to have allegedly weak support from the manuscripts (either in terms of their number or assumed importance), then doubts may surface in those editors who are unwilling to print such a reading as the Ausgangstext in their critical edition. Such doubts are evident in the UBS and NA editions, as may be seen in Metzger's Commentary, which tells us, disturbingly often, that if a reading which fits in with the author's style has support deemed to be ‘weak’, then the UBS committee avoided it and they printed in its stead another reading considered ‘strong’ merely because of its manuscript support.Footnote 19 Elsewhere, their hesitation has resulted in the ubiquitous brackets that confusingly and irritatingly disfigure these editions. They occur in those places where the editors could not decide between a shorter and a longer text and thereby have exposed their dilemma. In many such cases an investigation into the author's usage could and should have been decisive.
Allegedly weak evidence is one thing, but where we have no Greek support that is quite another matter.Footnote 20 A recurring question is the extent to which one may use conjectural emendation as a legitimate tool in a New Testament text-critic's armoury, just as it is in a classicist's. Nowadays I am less inclined to be dogmatic in opposing conjectures and have even left myself exposed as one who has proposed in an articleFootnote 21 that Mark 1.1–3 is secondary to that Gospel even though all manuscripts contain the verses.
The Nestle texts until recently always tantalised us with their display of (relatively modern) conjectural readings, especially those published by scholars from the Low Countries. NA25 included in its apparatus 220 such conjectures from 87 different scholars; NA26–7 pruned this to 129 conjectures by 73 scholars. Now, with our latest edition, the 28th, all conjectures have been expunged from the apparatus. Ironically, though, a reading with no Greek support is maintained as the NA txt at Acts 16.12, and now a conjecture has been newly introduced in the Editio critica maior (= ECM) and allied texts at 2 Peter 3.10.Footnote 22 (Some versional evidence supports a negative at 2 Peter 3.10; some versional evidence – namely a few Latin manuscripts and, according to earlier editions of the UBS text, the Provençal and Old German – supports the reading printed as txt at Acts 16.12).
Other conjectures have been proposed. Several commentators on 1 Cor 14.34–5, unhappy at its teaching and alert to the differing positioning of the verses in the manuscript tradition (often a sign of scribal tampering), have felt obliged to appeal to conjectural emendation by omitting these verses and to deny their Pauline authorship, even though no manuscripts actually omit the verses. Editors who wish to omit vv. 34–5 may claim that the verses do not agree with Pauline teaching on women elsewhere, thus applying the criterion of authorial coherence to justify the emendation.
Ryan Wettlaufer recently wrote a convincing thesis to encourage text-critics like myself to reconsider our stance on conjectural emendation.Footnote 23 His thesis concentrated on the Epistle of James. Among examples from that epistle he argued in favour of the conjecture πολυλαλοι rather than πολλοι at Jas 3.1, and in favour of φθονετε, a conjectural reading, first proposed by Erasmus, instead of φονευετε at Jas 4.2. Dale Allison's recent ICC commentary on JamesFootnote 24 finds both readings tempting. (Wettlaufer also proposes conjectures at Jas 1.1 and 2.1 as well as at the nonsensical Jas 4.5.)
Many of the deliberate changes made by scribes during the early centuries of hand-copying were in fact the conjectural emendations of their day, because scribes (or the ecclesiastical authorities behind them) tried to enhance, correct, improve and change the text before them in ways that are no different from the scholarly emendations proposed by recent academics. We still need to identify all such deliberate changes and treat them as secondary to the sought-after Ausgangstext, while recognising their importance for church history and the developments in Christian theology. NA28 is, however, probably right to discard modern emendations from its apparatus. (The place to discuss those is in learned commentaries.)
In a review of Wettlaufer's published thesisFootnote 25 I concluded that we must allow all authors occasionally to be imprecise and unclear. Not every writer, even those whose work was later branded as canonical, necessarily always wrote sense. Jas 4.5 is indeed nonsense but may we not equally allow our author further occasional obscurity? Can we permit our biblical authors to nod? Does not every writer sometimes write something illogical or wrong? Paul at 1 Cor 6.5 seems to have written ανα μεσον του αδελφου, which is nonsense. We know what Paul means: he asks how one may differentiate between brothers. But του αδελφου in the singular stands firmly in all manuscripts. Are we at liberty to correct our author by putting the noun in the plural? And what should one do about another verse in 1 Cor, namely 4.6, with its inscrutable το μη υπερ α γεγραπται? There have been several inspired and over-ingenious guesses to solve that problem. None has gained universal approval, and this is characteristic of many such proposals throughout the New Testament. Ought we not admit that from time to time we do not know what a text means? That admission comes as an alternative to jumping to accept any plausible conjecture as a panacea for a questionable reading.
I discussed earlier that our printed editions have a single example of Markan inconsistency where they print the plural of οχλος at Mark 10.1 against some thirty-four or so firm examples of the singular. I discovered that there is a variant reading giving the singular and I argued that this is original because it preserves Mark's consistency of usage. But what if there had been no variant giving the singular at 10.1? Would I have been at liberty to rewrite Mark and make him consistent with himself, conjecturing and printing the singular form of noun and verb there? The alternative would be that one would have to admit, albeit very reluctantly, a reading that goes against the author's normal practice. From that question, theoretical as far as Mark 10.1 is concerned, we turn now to an aberrant stylistic feature where no variant exists. Reading through the Pastoral Epistles one may be struck by the recurrence of ταυτα standing first in a sentence or clause and preceding the verb (e.g. 1 Tim 3.14; 4.6, 15; 5.7, 21; 6.2, 11; 2 Tim 1.12; 2.2, 14; Tit 2.15; 3.8). One would then be pulled up short when encountering at 1 Tim 4.11 παραγγελλε ταυτα. We would have to ask ourselves why the pronoun should follow its verb only there. This is especially strange when one sees comparable commands at 1 Tim 5.7 with the same wording but in the expected sequence, and at 1 Tim 6.2; Tit 2.15 where the recipients are similarly commanded to provide authoritative instruction. May an author's consistency allow us to reverse the order of the words in 1 Tim 4.11 and thus to print a conjectural emendation? To my knowledge there is no variant putting these words into the reverse, i.e. the ‘correct’, order.
A comparable case occurs with the introductory words to each of the seven letters to churches in Revelation; these are formulaic (Rev 2.1, 8, 12, 18; 3.1, 7, 14) and the wording is identical in NA28, namely τω αγγελω της εν (Εφεσω) εκκλησιας γραφον, but at 2.1 (A C), 6 (A), 12 (2050), 18 (A); 3.1 (046) the manuscripts within brackets here, some highly respected by many text-critics because of their age, read τω not της. Versional evidence seems to support τω2 at 3.7 and at 3.14. It is reasonable to expect an author to be consistent in such stereotypical formulae. Scribes though are clearly erratic here. Manuscript A throughout the seven formulae has της four times, τω thrice! The usage της … εκκλησιας to mean ‘to the angel of the church at …’ (rather than ‘to the angel who is of the church at …’) looks right, but is that what later scribes thought as they, admittedly irregularly, changed the texts in that direction, i.e. from τω to της? Grammar should help us, and it favours the originality of τω throughout. The rule is that prepositional phrases such as εν Εφεσω may precede anarthrous nouns, but ought not to stand between an article and its noun. That ruling would militate against της in all seven places. If such a phrase were in an attributive position, as here, it must follow the noun and the appropriate article (which would be τω) has to be repeated.
May grammar and a recognition that this author has been consistent be determinative for the originality of τω here, despite the alleged thinness of the Greek support in five places and its absence in the other two? If a version be deemed an inappropriate witness when discussing Greek articles, then we would be obliged to conjecture τω at 3.7, 14. If one accepts the arguments here one lays oneself open to accepting a reading with no Greek witnesses. Are we allowing conjectural emendation in pursuit of this Holy Grail of consistency? At the very least, the barrenness of the manuscript attestation for τω from the earliest Christian centuries permits the opinion that an original form of the wording has not survived 100 per cent of the time. Admittance of a conjecture as the Ausgangstext, though, requires careful thought.
4. Conclusion
Obviously, textual criticism, like many other disciplines within Biblical studies, has changed and developed over the past decades, but this is not the place to rehearse that history. Suffice it to say that modern technology is being profitably harnessed to textual criticism, possibly more than to most other branches of the discipline, the digitising of an increasing number of manuscripts readily coming to mind. The CBGM works with electronically produced diagrams, and is a way of coping with our vast quantities of manuscripts, especially minuscules. It has superseded other earlier attempts to harness this overwhelming wealth of material.Footnote 26 Alongside this methodology, innovative speedy dissemination of research materials means that the old fragmentation of the discipline has ended. Research is now universally shared and democratically enjoyed to everyone's benefit. The Centre for New Testament Textual Research in Westphalia may still be the Mecca for text critics but the Institute for Textual Scholarship and Electronic Editing at the University of Birmingham plays an amicable role alongside Münster's centre. Other institutions in Dallas (The Center for the Study of New Testament Manuscripts under Dan Wallace at Dallas Theological Seminary, located online at www.csntm.org) and in New Orleans (The H. Milton Haggard Center for New Testament Textual Studies under Bill Warren at the Baptist Theological Seminary and online at www.nobts.edu.cntts) are also highly significant. These developments may aid our use of the resources available to editors and readers of a Greek New Testament but by themselves cannot solve the dilemmas exposed in this paper.
