“jurare in verba magistri.”
“To swear whate'er the master says is true.”
And if on the contrary certain witnesses are found to range themselves continually on the side which is condemned by a large majority of others exhibiting other notes of truth entitling them to credence, those few witnesses must inevitably lose in respectability according to the extent and frequency of such eccentric action.
2. If one Codex (z) is demonstrably the mere transcript of another Codex (f), these may no longer be reckoned as two Codexes, but as one Codex. It is hard therefore to understand how Tischendorf constantly adduces the evidence of “E of Paul” although he was perfectly well aware that E is “a mere transcript of the Cod. Claromontanus[49]” or D of Paul. Or again, how he quotes the cursive Evan. 102; because the readings of that unknown seventeenth-century copy of the Gospels are ascertained to have been derived from Cod. B itself[50].
3. By strict parity of reasoning, when once it has been ascertained that, in any particular instance, Patristic testimony is not original but derived, each successive reproduction of the evidence must obviously be held to add nothing at all to the weight of the original statement. Thus, it used to be the fashion to cite (in proof of the spuriousness [pg 055] of “the last twelve verses” of St. Mark's Gospel) the authority of “Eusebius, Gregory of Nyssa, Victor of Antioch, Severus of Antioch, Jerome[51],”—to which were added “Epiphanius and Caesarius[52],”—“Hesychius of Jerusalem and Euthymius[53].” In this enumeration, the names of Gregory, Victor, Severus, Epiphanius and Caesarius were introduced in error. There remains Eusebius,—whose exaggeration (a) Jerome translates, (b) Hesychius (sixth century) copies, and (c) Euthymius (a.d. 1116) refers to[54] and Eusebius himself neutralizes[55]. The evidence therefore (such as it is) collapses hopelessly: being reducible probably to a random statement in the lost treatise of Origen on St. Mark[56], which Eusebius repudiates, even while in his latitudinarian way he reproduces it. The weight of such testimony is obviously slight indeed.
4. Again, if two, three, or four Codexes are discovered by reason of the peculiarities of text which they exhibit to have been derived,—nay, confessedly are derived—from one and the same archetype,—those two, three, or four Codexes may no longer be spoken of as if they were so many. Codexes B and א, for example, being certainly the twin products of a lost exemplar, cannot in fairness be reckoned as = 2. Whether their combined evidence is to be estimated at = 1.75, 1.50, or 1.25, or as only 1.0,—let diviners decide. May I be allowed to suggest that whenever they agree in an extraordinary reading their combined evidence is to be reckoned at about 1.50: when in an all but unique reading, at 1.25: when the reading they contain is absolutely unique, as when they exhibit συστρεφομένων δὲ αὐτῶν in St. Matt. xvii. 22, they should be reckoned as a single Codex? Never, at all events, can they be jointly reckoned as absolutely two. [pg 056] I would have them cited as B-א. Similar considerations should be attached to F and G of St. Paul, as being “independent transcripts of the same venerable archetype[57],” and to Evan. 13, 69, 124, 346, 556, 561, and perhaps 348, 624, 788[58], as being also the representatives of only one anterior manuscript of uncertain date.
5. It requires further to be pointed out that when once a clear note of affinity has been ascertained to exist between a small set of documents, their exclusive joint consent is henceforward to be regarded with suspicion: in other words, their evidential Weight becomes impaired. For instance, the sympathy between D and some Old Latin copies is so marked, so constant, in fact so extraordinary, that it becomes perfectly evident that D, though only of the sixth century, must represent a Greek or Latin Codex of the inaccurate class which prevailed in the earliest age of all, a class from which some of the Latin translations were made[59].
6. I suppose it may be laid down that an ancient Version outweighs any single Codex, ancient or modern, which can be named: the reason being, that it is scarcely credible that a Version—the Peshitto, for example, an Egyptian, or the Gothic—can have been executed from a single exemplar. But indeed that is not all. The first of the above-named Versions and some of the Latin are older,—perhaps by two centuries—than the oldest known copy. From this it will appear that if the only witnesses producible for a certain reading were the Old Latin Versions and the Syriac Version on the one hand,—Codd. B-א on the other,—the united testimony of the first two would [pg 057] very largely overbalance the combined testimony of the last. If B or if א stood alone, neither of them singly would be any match for either the Syriac or the Old Latin Versions,—still less for the two combined.
7. The cogency of the considerations involved in the last paragraph becomes even more apparent when Patristic testimony has to be considered.
It has been pointed out elsewhere[60] that, in and by itself, the testimony of any first-rate Father, where it can be had, must be held to outweigh the solitary testimony of any single Codex which can be named. The circumstance requires to be again insisted on here. How to represent the amount of this preponderance by a formula, I know not: nor as I believe does any one else know. But the fact that it exists, remains, and is in truth undeniable. For instance, the origin and history of Codexes ABאC is wholly unknown: their dates and the places of their several production are matters of conjecture only. But when we are listening to the articulate utterance of any of the ancient Fathers, we not only know with more or less of precision the actual date of the testimony before us, but we even know the very diocese of Christendom in which we are standing. To such a deponent we can assign a definite amount of credibility, whereas in the estimate of the former class of evidence we have only inferences to guide us.