The genealogy of Jesus given by Luke includes a certain “Cainan” (Luke 3:36) who is indicated to be the son of Arphaxad and father of Shelah. In the Old Testament, Arphaxad was the father of Shelah (Salah), according to Genesis 10:24 and 11:12, and 1 Chronicles 1:18. There is no Cainan between Arphaxad and Shelah in the Hebrew Masoretic text, though he is there in the Septuagint (ancient translation of the OT into Greek, designated LXX).
This has raised the question of where this Cainan came from in the genealogy given by Luke. Should he be there? Was his name accidentally omitted from the Hebrew Masoretic text? Or was it a scribal error in Luke that was eventually inserted into the LXX?
Now, this is not simply a matter of NT textual criticism, as it obviously involves OT textual criticism as well, about which we have not spoken yet. I will present the usual line of argument about this issue, and then present my own view.
The dominant view among evangelicals seems to be that this Cainan does not belong in the genealogy, and is a scribal error introduced into Luke, probably accidentally reinserted from the legitimate Cainan in Luke 3:37. The arguments are:
The earliest manuscripts of Luke omit it.
This error could have easily arisen by homoeoteleuton — a scribe copying “tou Salah” and then looking back at his exemplar looked to the end of a further line which ended with “tou Kainan” and therefore wrote that after “tou Salah,” which he should not have done.
It is not in any of the Hebrew Masoretic text manuscripts.
The Samaritan Pentateuch omits it.
Julius Africanus, writing ca. AD 220, omits it, going straight from Arphaxad to Salah.
It was not in the earliest copies of the LXX, but must have been added some time after AD 220, to assimilate the LXX to Luke.
These arguments may seem prima facie compelling, which is probably why the majority of evangelicals have gone this route, but there are actually serious problems with every one of them:
In fact, only two manuscripts of Luke omit it: P75 and Codex Bezae Cantabrigiensis. P75 is, as is typical of Alexandrian manuscripts, a very sloppy copy, with 257 singular readings, a quarter of which are nonsensical, and a total of over 400 mistakes. It seems that the scribe actually did not even know Greek. Codex Bezae is a V/VI-century manuscript that is notoriously error-ridden. Furthermore, in the genealogy in Matthew, the scribe of this manuscript added in the three names Matthew omitted between Joram and Uzziah in 1:8 (Ahaziah, Joash, and Amaziah), so he seems to be assimilating his copy to the OT.
Homoeoteleuton is not a magic wand that can be waved to proclaim just any claimed textual variant. Homoeoteleuton is an error that occurs when two lines end in the same combination of letters, so that a scribe’s eye jumps from the first group to the second. This error leads to an accidental omission of text (the segment between the two identical combinations), not to a random insertion. Furthermore, there is no identical (or even similar) combination of letters between “TOUEBERTOUSALA” and “TOUIAREDTOUMALELEEL” that could occasion this. So appealing to homoeoteleuton here is truly special pleading.
It used to be considered axiomatic that, wherever the Masoretic text and the LXX differed, the Masoretic text must be correct. This is no longer sustainable. The earliest Masoretic text manuscript we have (and it is not complete) dates to AD 897. The Dead Sea scrolls (DSS), the first of which were discovered in 1947, are Hebrew manuscripts predating the Masoretic manuscripts by more than 1,000 years. It has been assumed that these would support the Masoretic readings against the LXX reading where they differ, but it turned out that the DSS very often sided with the LXX against the Masoretic text. So it cannot be glibly assumed that the absence of “Cainan” in the Masoretic text should be considered decisive. (It is unfortunate that none of the Dead Sea Scrolls include the passages in question.)
There is also no reason to prefer the Samaritan Pentateuch over the LXX. If it was produced from the Masoretic tradition, then it has no independent weight.
It is indeed fascinating that such weight is put on the testimony of Julius Africanus, to the point that the date of his writing (AD 220) is actually used as the terminus post quem for the putative altering of the LXX to introduce Cainan. Why is the date of Julius’ writing compared to the date of the LXX manuscripts? Either the date of Julius’ writing should be compared to the date of the LXX writing, or the date of the earliest LXX manuscript should be compared to the date of the earliest manuscript of Julius. (In both cases, I think the LXX predates Julius by a good 500 years.) This again seems to be special pleading, for how do we know that it was not Julius’ writing that was later altered? Furthermore, while Julius omits Cainan, it is not proof that Cainan was absent from the manuscripts. Also, Julius, inter alia, created a very fanciful attempt to explain the difference between the genealogy in Matthew and Luke, which can at least call into question his reliability. Finally, manuscripts always outweigh the testimony of Fathers as to the form of text.
The problem remains of how Cainan got into the LXX, if it was not originally there. The assertion that it was added into the LXX to assimilate it to Luke overlooks a huge problem: Genesis 11:12-13 in the LXX reads “And Arphaxad lived a hundred and thirty-five years, and begot Cainan. And Arphaxad lived after he had begotten Cainan four hundred years, and begot sons and daughters, and died. And Cainan lived a hundred and thirty years and begot Sala; and Cainan lived after he had begotten Sala three hundred and thirty years, and begot sons and daughters, and died.” So it is not simply a matter of inserting the lone name Cainan to assimilate with Luke. For some reason, the putative inserter invented all these time spans out of whole cloth. This suggestion seems risible.
So, without any other consideration, it seems clear that if the choice is between Cainan being original and dropped from the Masoretic text on the one hand, and Cainain being a an error in Luke 3:36 on the other, that case for the former is very much stronger.
Now add to this the case from textual criticism. As we demonstrated before, any reading that is in the overwhelming majority of the manuscripts is virtually certain to be the original and should be taken as such unless there is overwhelming evidence to the contrary. In this case there isn’t, so I would say that Cainan in Luke 3:37 is not an error but is the original reading.
However, here is my take on the matter:
Everyone seems to assume that Cainan must be in BOTH the Masoretic text AND Luke, or in NEITHER. It’s either BOTH or NEITHER. (So either the Masoretic text is wrong or Luke contains a scribal error and, under the influence of Westcott-Hort thinking, evangelicals stand ready to cast Cainan out of Luke, though it is in virtually every manuscript; it is the NT text that is considered less reliable.) What no one seems to consider is that it doesn’t have to be BOTH or NEITHER; it could be that the Masoretic text correctly does not have it, and Luke correctly does have it i.e. they are both right. Follow:
First, genealogies need not be complete. In Hebrew thinking, all lineal descendants of a man are considered his sons, and begotten by him, not only the immediate next generation. This should be obvious from the genealogy in Matthew 1. The very first verse reads
The book of the genealogy of Jesus Christ, the son of David, the son of Abraham.
Jesus was obviously not the immediate next generation after David, nor was David the immediate next generation from Abraham, yet Jesus is called the “son” of David, and David the “son” of Abraham. Then the chapter continues and fills in the gaps among these three men, but not completely: 1:8b reads
… Jehoshaphat begot Joram, and Joram begot Uzziah.
In fact, the immediate descendant of Joram was Ahaziah, whose immediate descendant was Joash, whose immediate descendant was Amaziah, and it was he whose immediate descendant was Uzziah. (For use of the term “beget” in reference to distant, not immediate, descendants, see 2 Kings 20:18/Isaiah 39:7.)
Thus, it should be clear that genealogies may not include all names in them, but may skip generations. Furthermore, this is not something that can always be detected by cross-comparison with other parts of Scripture. It seems to me that there must be another place where there are gaps that seem to have been universally overlooked (as far as I can tell), and one that is found in more than one book of the Bible. Ruth 4:21-22 reads,
Salmon begot Boaz, and Boaz begot Obed; Obed begot Jesse, and Jesse begot David.
This is same order is given in 1 Chronicles 2:11b-15. This same order is given in Matthew 1:5-6a, but with additional information, viz.
Salmon begot Boaz by Rahab, Boaz begot Obed by Ruth, Obed begot Jesse, and Jesse begot David the king.
Now, if Boaz was the son of Rahab, he must have been born around 1400 BC, while David was born ca. 1041 BC. There are only two names between Boaz and David (Obed and Jesse), a span of ca. 350 years. There cannot have been only three generations across 350 years at this time of human history, so there must be unnamed members of the genealogy here. Q.E.D.
What this means is that there could well have been a link between Arphaxad and Salah (i.e. Cainan) who is simply not named or mentioned in the OT, just as there are somewhere among Boaz, Obed, and Jesse. The NT, then, gives us this information in Luke 3:36 that the OT does not give us.
Are there places in which the NT provides additional information, and even names, about OT events that are not give in the OT? Yes. Consider the additional information about the call and movements of Abraham (Acts 7:2-4) and about Enoch (Jude 14-15). We are even given the specific names of men who opposed Moses (the magicians of Egypt?) in 2 Timothy 3:8 (“Jannes and Jambres”), though these names are not given in the OT. So the idea of Luke 3:36 giving information in the NT that is not specified in the OT is consistent with what we see elsewhere.
Finally, it should be noted that the insertion of Cainan in the genealogy does not alter the historical timeline; we still have Arphaxad’s age at which he begot Salah, so the timing does not change.
In sum then, the Westcott-Hort mindset seems to have invaded the evangelical church to such an extent that we are all too cavalier about throwing out a part of the NT text that is in all but two corrupt manuscripts, and in doing so undermine the trustworthiness of the text (for if this reading, present in all but two corrupt manuscripts, can be discarded, how can we know that other readings, though they are present in virtually all manuscripts, are part of God-breathed Scripture?) This is done to deal with a supposed problem which, if my understanding is correct, is not a problem at all.