Here we face-off T. S. Green (1858) with S. W. Whitney (1892) on the main variants debated between advocates of the Traditional Text (TR) and the Alexandrian Text (RV).
TR: εως ου ετεκεν τον υιον αυτης τον πρωτοτοκον
'...till she had brought forth her firstborn son.'RV: εως ου ετεκεν υιον
'...till she had brought forth a son.'
'The variation which will demand notice in this place, gives occasion, at the outset, to certain general observations, preliminary to the consideration not only of the present instance but of many others of like complexion.
It is clear, from the nature of the case, that the intrusion of glossarial matter into the text must be a gradual process, and, as such, favoured by lapse of time. From this it follows, as a general principle, that documents of a later age would be more extensively infected with such corruption, and that the circumstances of the more ancient are favourable to their purity in this
particular respect. Accordingly, a shorter reading, especially if it be of a kind to call forth glosses, provided it is supported by a few authorities of high antiquity, has at once a strong presumption in its favour : though before such presumption is accepted, it should be ascertained that there is no reason either in the outward
shape of the passage for referring the briefer form to accidental curtailment in transcription, or in its purport for suspecting wilful suppression.
In the present place, instead of the common reading, a shorter one, εως ου ετεκεν υιον, is exhibited by B, Z, and supported by the Syriac (N), by the Old Latin in a, b, c, g1 , as well as the Coptic and Sahidic versions. Another of the same class of Latin documents (g2) adds unigenitum. The remaining mass of authorities have the common form, except that D (secundu manu, 'second hand') and L omit αυτης.
If the text stood originally as it is presented by the few authorities just cited, the bare statement furnished by the words εως ου ετεκεν υιον would leave a blank respecting the subsequent condition of the mother of Jesus, which thought or fancy would not fail to occupy. Another evangelist, indeed, undoubtedly supplies τον πρωτοτοκον (Lk. 2:7); but this term, though it might be regarded as looking towards a certain conclusion, that Mary was the mother of other children, still does not absolutely imply so much and bar the exercise of opinion.
Under these circumstances, the simpler reading, if original, could hardly escape the application of supplementary glosses, perhaps of opposite tendencies; and, since it is supported by clear testimony, the fuller form must fall under the suspicion of having its origin in the accretion of such matter, especially if, as in the present case, this is at once supplied by a parallel passage.
To append in the margin τον πρωτοτοκον from the other Gospel would be a simple proceeding, but having a ready issue in the amplification of the text itself. The Latin addition unigenitum, already noticed, is the bolder expression of an opinion, widely held and stoutly maintained, as may be seen in the comment of Chrysostom.
When these considerations are taken into account, it is unreasonable to acquiesce confidently in the common reading : and, notwithstanding the great preponderance in the amount of the opposing documentary evidence, the few, but ancient, Greek, Syriac, Latin, and other witnesses for the shorter form press
strongly for the conclusion, that the longer reading is the result of assimilation, and that the original shape of the clause was simply εως ου ετεκεν υιον.'
(Developed Criticism, p. 1-2)
'Against the former of these readings it is commonly urged that it is taken from Luke 2:7, where no rival reading exists. This, however, is pure conjecture. There is no proof that it was adopted from Luke; nor can any valid reason be given why it should have been. It certainly could not have been done to afford an argument against the perpetual virginity of Mary, for that was not needed. Besides, the statement that Mary had brought forth her firstborn son was in the text long before the doctrine of her perpetual virginity was originated.
If Matthew had written only the words given in the RV, we cannot see what possible motive there could be for changing it to the longer reading of the TR. On the contrary, if Matthew wrote the words commonly ascribed to him, it is easy to see that a believer in the perpetual virginity
of Mary might have been tempted to strike out the word πρωτοτοκον. We find Jerome, who contended for the doctrine, though he preserves the reading "her firstborn son" in his Latin [Vulgate] Version, saying in his Commentary on Matthew, in allusion to Helvidius and others who denied the doctrine, that
"from this passage some very perversely infer that Mary had other sons also, saying that none but a person who had brothers would be called a firstborn son."The presence of πρωτοτοκον would very naturally lead a person who believed in the doctrine, but who was less scrupulous than Jerome, to remove the objectionable phrase; for, explain the word as you
will, the evangelist could not, as a historian, have used it if he had regarded Jesus as the only son born to Mary. Matthew afterwards speaks in language in which no one would write who knew that Mary had no other children ; for, if the meaning of words can be depended upon at all, αδελφοι and αδελφαι, in Matt. 12:46, 13:55, 56, mean brothers and sisters in the commonly accepted sense of the words as truly as μητηρ means mother. By thus speaking, the evangelist shows most clearly that, as a historian familiar with the facts in the case, he not only would naturally have written "her firstborn son," but could hardly have written otherwise. In fact, the very presence of υιον, unaccompanied by the article and accepted as a part of the text, is proof conclusive that the longer reading is genuine. After having recorded, in verse 21, the words of the angel to Joseph, "she shall bring forth a son," and again, after quoting, in verse 23, the prophecy concerning Mary, that she should "bring forth a son," Matthew could hardly have gone on in his narrative, and written immediately after, "he knew her not till she had brought forth a son." The article would of necessity have appeared (if αυτης, "her," did not), in connection with υιον, denoting a reference to the son already mentioned as promised and predicted. This difficulty seems to have been long ago seen and felt; hence the Memphitic Version inserts the article, while the Thebaic inserts both the article and "her," and reads, "till she had brought forth her son." Again, the presence of "firstborn" is necessary, in order to bring out the evangelist's idea that Joseph knew not Mary till after the birth of Jesus. The word "till" of itself
does hot show this; it merely indicates that he had no intercourse with her up to that time. But the insertion of "first-born" clearly implies what is indirectly declared in Matt. 12:46, 13:55-56, Mark 6:3, and elsewhere, that Mary had other children, of whom Joseph was the father. It is just what might be expected to have been written by this evangelist. And that it was, the documentary testimony before us leaves no room for doubting. The shorter reading is attested only by the Sinaitic and Vatican manuscripts and one other uncial (Z, 6th century), two cursives, five copies of the Old Latin Version, and the Curetonian Syriac. The common reading, on the other hand, is sustained by C, D, E, K, L, M, S, U, V, Γ, Δ, Π, nearly all the cursives, the Peshito and Philoxenian Syriac, four copies of the Old Latin, the Vulgate, Ethiopic, Slavonian, Armenian, Georgian, and even the two Egyptian Versions ; for, though "firstborn" is not expressed in these, their reading indicates that the word is a part of the genuine text. Then, as to patristic testimony, that is overwhelmingly conclusive, — not less than nineteen or twenty of the Fathers, from the second century downward, testifying in support of the common reading.
Yet Preb. Humphry says,
"There is but little MS authority for the reading which the A.V. here follows." (Commentary on the RV)
If by "authority" he means evidence, we know not what more evidence one could reasonably ask for, whether from manuscripts or from other sources, than we have in proof of the genuineness of this reading. The advocates of the brevior lectio appear to consider Griesbach's canon, the testimony of three uncials, and a surmise as evidence outweighing everything else.
But just here it may be well, in passing, to reply to a query which may have arisen in the minds of some. And that is. Why should any one have omitted these words when they were known to exist in Luke ? In other words. How is it that passages like this and 11:19, for example, could have been changed in Matthew, while corresponding passages in one or more of the other Gospels were left unchanged, and their readings continued unquestioned, or all but unquestioned? The query is a fair one. But it is based upon a false assumption. And it is to this that the whole trouble with the querist is due. We cannot assume that those who are supposed to have made the alteration really knew that similar language existed elsewhere. The presumption is rather that they were not aware of it. These alterations were made at a very early date, — very soon after the apostles' days.
As Dr. Hort says, a transcription including
a 'tolerably free modification of language and even rearrangement of material ... was carried on during the earliest centuries:' (Intro. p.7)
At that time, however, the Gospels were not bound up in one volume, but were written each on a separate parchment or collection of parchments. These were, moreover, expensive, and not easy of attainment. So that, during the first two or two and a half centuries after Christ's death, comparatively few persons, at the most, owned copies of any portion of the NT, and fewer still, copies of the whole. One might be able to become the possessor of one of the Gospels, or at most, though rarely, of two of them, and possibly of one or more of the other books of the NT, without knowing what the rest of the books really contained ; for the making up of the canon of the NT was a slow and progressive work. So that it is not to be wondered at in the least that alterations should have crept into one and not into the other of two passages in different Gospels, which were originally precisely or almost precisely alike. Compare Matt. 7:25 and Luke 6:48.
(The Revisers' Greek Text, p. 58-61)