History of the Canons
If, as seems increasingly clear from research into scribal habits, the long-held transcriptional canons are wrong, where did the canons of transcriptional probability come from? James Royse has written about presentations of the transcriptional canons in modern textual handbooks as follows: ‘Regrettably, though, most presentations of these canons are not – as far as one can tell from the exposition – based on the actual knowledge of documents of which Hort speaks, but rather appear to rest upon a priori reflections on how scribes behaved (or must have behaved)’.
However, the truth about the history of the transcriptional canons is more complicated than Royse’s characterisation of the current situation might suggest. The original framers of the rules were, of course, not unacquainted with actual documents, nor were their rules based simply on idle reflection. Rather, it would appear that the canons were formulated by early textual scholars as a direct result of some of their most important work. Here is how it appears to have happened.
Although there were earlier critical apparatuses, it was John Mill’s 1707 Greek New Testament which, more than any other, led to the development of canons of transcriptional probability. Mill’s Greek New Testament reproduced Stephanus’ ‘Textus Receptus’ (TR), but its critical apparatus contained about 30,000 variant readings from manuscripts, versions and fathers. This large apparatus provoked the full range of reactions to textual criticism still seen today, from defences of the traditional text to attacks on scripture by sceptics. From a text-critical perspective, the most important response was that of Gerhard von Maestricht who produced a list of forty-three textual canons in his Greek New Testament of 1711. Gerhard’s purpose was to defend the TR; his canons were designed to discredit the variant readings that Mill’s apparatus had brought to light. But Gerhard’s canons prompted the counter-reactions of J.A. Bengel and J.J. Wettstein, who produced their own canons, including the first statements of Prefer the Harder Reading (lectio difficilior potior) and Prefer the Shorter Reading (lectio brevior potior) respectively.
The transcriptional canons thus arose as new manuscripts were discovered and collated against the base text of the TR. Many of these readings tended to show up as being shorter, harder, harsher and less harmonized than the TR. Since it was assumed that earlier manuscripts were nearer to the original text, most early textual critics went a step further and reasoned that scribes must have added to, polished up, harmonized and attempted to improve the text to produce the TR; hence the canons.
Tregelles and the Logic of the Canons
To take an example of the logical steps by which textual critics formulated the transcriptional canons, we shall briefly consider the argument (if that is the right word) of S.P. Tregelles. His 1854 work, An Account of the Printed Text of the Greek New Testament with Remarks on its Revision upon Critical Principles, proceeds by means of an historical sketch of the various editors who published editions of the Greek New Testament. Tregelles was far from a disinterested chronicler, however, of textual history. His writing has a moral tone to it; the heroes of the discipline are those who championed ancient evidence while the villains of the piece are those who produced or perpetuated the TR. Tregelles’ history is one long argument: his desire is to establish the text upon the authority of ancient witnesses.
When Tregelles leaves the history of the discipline behind and sets out the principles by which textual decisions must be made, he again spends another 45 pages(from 174 to 219)repeating his primary rule: we must follow only ancient evidence. Having now argued his main point for over 200 pages, Tregelles devotes about a dozen paragraphs to a variety of transcriptional canons. Here is how he starts:
‘When once the position has been definitively taken, that the ancient evidence is that which we must especially regard, other considerations affecting various readings must have their place, in order to judge between the ancient authorities, when they differ among themselves. … As copyists were always more addicted to amplification than the contrary, as a general rule it must be said, that less evidence is sufficient (other things being equal) in favour of an omission than of an insertion; especially if the insertion is one which might naturally be suggested’.
Tregelles offers no evidence for his dogmatic belief that copyists are always more addicted to addition than omission, and so we might well question on what basis he makes his assertions about the behaviour of copyists. (He does offer an analogy from nature: ‘It can hardly be too habitually remembered, in criticism, that copyists were always more accustomed to add than to omit. Those who know nothing of criticism or of ancient books, biblical or classical, often imagine the contrary; but such is not the fact. Of course careless transcribers may omit; but, in general, texts, like snowballs, grow in course of transmission’ ). However, in reality, Tregelles has given us two notable clues as to where he derived his canons from. The two main features of his book give us all the explanation we need.
Firstly, Tregelles’ transcriptional canons are based on his belief in the superior authority of ancient witnesses, the point he has spent most of his book arguing, and the additional fact (unstated but easily demonstrated) that more recent witnesses have a longer text than ancient ones. Tregelles’ transcriptional canons are therefore but an extension, an elaboration, of his primary rule. Tregelles takes this logical step almost unconsciously. Tregelles’ second reason is also held on implicit grounds: his trust in the authority of the heroes of the discipline whose history he has spent the most part of the book retelling. Tregelles has set up his narrative in such a way as to show us who we should respect: people like Bengel and Griesbach. Tregelles unhesitatingly repeats their transcriptional dictums with added dogmatism.
<h2 > The Textus Receptus and the Canons
It is important to realise the connection between the transcriptional canons and the sometimes bitter polemic between defenders of the TR and those promoting readings of newly collated, more ancient, manuscripts. Early textual critics were vilified for departing from the TR. Being only human, their own arguments for a New Testament text based on evidence from earlier manuscripts were not altogether free from frustration with those who refused to accept the many faults of the TR. We notice this ongoing conflict whether we look at early textual critics like Gerhard von Maestricht responding to Mill’s apparatus, Bengel and Wetstein responding to Gerhard or 19th century critics like Tregelles. It is the debate over the TR and the preferential treatment which most early textual critics gave to more ancient witnesses which overshadows (or rather, as we shall see, clouds) the issue of scribal habits.
<h2 > Three Problems with the Canons: 1. Methodological
Three problems arise from the way in which these early textual critics allowed the debate over the TR to shape their transcriptional canons: one methodological, one logical and one practical. Firstly, the process of determining scribal habits was ill-defined and thus methodologically far from satisfactory. Early textual critics did not draw up clearly formulated principles by which to determine scribal errors, nor did they publish comprehensive lists of variant readings which established patterns of scribal corruption. Instead, they simply observed deviations from the TR, using Mill’s apparatus or their own collations. The TR with all its faults is hardly an acceptable basis from which to determine textual principles. Even so, for all its faults, the TR coincides with the text of virtually all witnesses in many places in the New Testament. Variations from the TR cannot be assumed ‘original readings’, for they are often simply scribal errors themselves. So, even when these early textual critics cited particular readings in defence of the canons ‘presumably as evidence – the evaluation of one reading as the original and another as arising by a scribal error is frequently suspect from a methodological point of view, and so one is left wondering why the direction of scribal error may not have been other than is stated’.
<h2 > 2. Logical
Secondly, the rules of scribal habit contained a logical problem in that they formed part of a circular argument. Textual critics simply noticed that earlier manuscripts had a leaner, meaner text than the TR. From this observation (and the assumption of the superiority of earlier manuscripts) the rules of scribal habits were framed. To restate the argument in a logical formulation: (1) the TR is a late and degenerate text, (2) the TR is observed (relatively speaking) to be full of additions, improvements, harmonizations and smoothings, (3) scribes must have added, improved, harmonised and smoothed the New Testament text to produce the TR, (4) hence, the TR is proved to be a degenerate text, seeing it is guilty of typical scribal corruptions. The scribal canons helped to dethrone the TR by proving its degeneracy while, at the same time, the scribal canons depended upon the assumption of the degeneracy of the TR. This was less logical syllogism, more logical circle. Whatever truth it contained, it proved nothing.
As an example of such circular reasoning, consider a statement of Hort’s under the heading of ‘Absence of Interpolations in B’. He writes,
‘In the New Testament, as in almost all prose writings which have been much copied, corruptions by interpolation are many times more numerous than corruptions by omission. When therefore a text of late and degenerate type, such as is the Received Text of the New Testament, is consciously or unconsciously taken as a standard, any document belonging to a purer stage of the text must by the nature of the case have the appearance of being guilty of omissions; and the nearer the document stands to the autograph, the more numerous must be the omissions laid to its charge. If B is pre-eminently free from interpolations, Western, Alexandrian, or Syrian, it cannot but be pre-eminently full of what may relatively to the Received Text be called omissions’.
Hort attempts to convert one of the more embarrassing features of his favourite manuscript Vaticanus, its frequent omissions, into an additional proof of its general excellence. His proof of Vaticanus’ virtue here rests upon the preference for the shorter reading, but the canon itself rests upon the assumed superiority of manuscripts like Vaticanus that tend to omit.
<h2 > 3. Practical
A third problem of a more practical nature flows from this circular logic: the problem of how to get textual critics to take studies into transcriptional probability seriously. The attitude of some textual critics to the accumulating (and as we shall see, disconfirming) evidence about scribal habits is one of Olympian indifference. Others, however, faced with mounting evidence that the transcriptional canons are endangered by evidence, have resorted to arguing that either the canons or the disconfirming evidence have somehow been misinterpreted.
Thus, in response to Royse’s 1981 study which called Prefer the Shorter Reading into serious question by showing that in six early major papyri, scribes made 337 omissions and 130 additions (a ratio of 72:28), Moisés Silva made his own study of the scribal habits of P46, Sinaiticus, Alexandrinus and Vaticanus in Galatians. Silva’s study found that these manuscripts omitted 68 times while only adding 25 times (a ratio of 73:27). Despite Silva’s own fresh evidence of the failure of lectio brevior potior, he urged the continuing validity of the canon, arguing that Griesbach qualified his canon by excepting small omissions and cases of homoeoteleuton. Silva also made an earlier study of the scribal habits of P46, Sinaiticus, Alexandrinus, Vaticanus and Claromontanus in Philippians in which omissions again outnumbered additions by 76 to 51 (60:40). Despite the results again confirming Royse’s conclusion, Silva suggested otherwise, arguing that if all words that were mere ‘grammatical markers’ were ignored, then ‘scribes were indeed more likely to add than to omit’ (p159).
Silva thus argues that despite the fact that his own studies found that scribes overwhelmingly tend to omit, we should continue to hold that scribes tended to add. It is not often that we are presented with such a clear example of Orwellian thinking, and readers wishing for a more technical refutation of Silva's methodology and argument may consult Royse's 2008 monograph and Andrew Wilson's 2011 Filologia Neotestamentaria article on scribal habits, but for the moment, it is worth asking a question: What amount of evidence should convince us to abandon lectio brevior potior if the fact that scribes appear to omit three-quarters of the time and only add one quarter does not? Whatever the reason behind the reluctance to follow where the evidence leads, turning a blind eye to the problem will become ever more difficult as more studies make the indefensibility of the current canons increasingly obvious.
 Notably those of Robert Stephanus in 1550, which showed the readings of 16 manuscripts in the margins of his Greek New Testament, Brian Walton whose Polyglot of 1655-57 added the witness of Codex Alexandrinus and 15 other manuscripts to those of Stephanus, and John Fell in 1675 who claimed to show in his apparatus the witness of over 100 manuscripts.
 ibid, p88, second footnote
 The Text of Galatians: Evidence from the Earliest Greek Manuscripts, in Scribes and Scripture: New Testament Essays in Honor of J. Harold Greenlee, ed. D. A. Black, Winona Lake, Ind., Eisenbrauns, 1992, pp 17-25