It would be a serious omission indeed to close this enumeration of Tests of Truth without adverting to those Internal Considerations which will make themselves heard, and are sometimes unanswerable.

Thus the reading of p??t?? (masculine or neuter) which is found in Cod. B (St. Luke xix. 37) we reject at once because of its grammatical impossibility as agreeing with d???e?? (feminine); and that of ?a?d?a?? (2 Cor. iii. 3) according to the witness of A?BCDEGLP on the score of its utter impossibility69. Geographical reasons are sufficiently [pg 066] strong against reading with Codd. ?IKN? ??at?? ?a? ??????ta in St. Luke xxiv. 13 (i.e. a hundred and threescore furlongs), to make it of no manner of importance that a few additional authorities, as Origen, Eusebius, and Jerome, can be produced in support of the same manifestly corrupt reading. On grounds of ordinary reasonableness we cannot hear of the sun being eclipsed when the moon was full, or of our Lord being pierced before death. The truth of history, otherwise sufficiently attested both by St. Matthew and Josephus, absolutely forbids a?t?? (?BDL?) to be read for a?t?? (St. Mark vi. 22), and in consequence the wretched daughter of Herodias to be taken to have been the daughter of Herod.

In these and such-like instances, the Internal reasons are plain and strong. But there is a manifest danger, when critics forsake those considerations which depend upon clear and definite points, and build their own inventions and theories into a system of strict canons which they apply in the teeth of manifold evidence that has really everything to recommend it. The extent to which some critics are ready to go may be seen in the monstrous Canon proposed by Griesbach, that where there are more readings than one of any place, that reading which favours orthodoxy is an object of suspicion70. There is doubtless some reason in the Canon which asserts that “The harder the reading, the less likely it is to have been invented, and the more likely it is to be genuine,” under which de?te??p?t? [pg 067] (St. Luke vi. 1) must receive additional justification. But people are ordinarily so constituted, that when they have once constructed a system of Canons they place no limits to their operation, and become slaves to them.

Accordingly, the true reading of passages must be ascertained, with very slight exception indeed, from the preponderating weight of external evidence, judged according to its antiquity, to number, variety, relative value, continuousness, and with the help of the context. Internal considerations, unless in exceptional cases they are found in strong opposition to evident error, have only a subsidiary force. Often they are the product of personal bias, or limited observation: and where one scholar approves, another dogmatically condemns. Circumstantial evidence is deservedly rated low in the courts of justice: and lawyers always produce witnesses when they can. The Text of Holy Scripture does not vary with the weathercock according to changing winds of individual or general opinion or caprice: it is decided by the Tradition of the Church as testified by eye-witnesses and written in black and white and gold in all countries of Christendom, and all down the ages since the New Testament was composed.

I desire to point out concerning the foregoing seven Notes of Truth in Textual Evidence that the student can never afford entirely to lose sight of any of them. The reason is because although no doubt it is conceivable that any one of the seven might possibly in itself suffice to establish almost any reading which can be named, practically this is never the case. And why? Because we never meet with any one of these Tests in the fullest possible measure. No Test ever attains to perfection, or indeed can attain. An approximation to the Test is all that can be expected, or even desired. And sometimes we are obliged to put up with a very slight approximation indeed. Their strength resides in their co-operation.