Near the end of John 6:15 – following John’s account of the feeding of the 5,000 – there is an interesting variant-unit: did John write that Jesus withdrew again (ἀνεχώρησεν πάλιν) to the mountain by himself, or that Jesus withdrew (ἀνεχώρησεν) (without “again”) to the mountain by himself, or that Jesus “fled again” φεύγει πάλιν) to the mountain by himself?
John does not use ἀνεχώρησεν anywhere else, so perhaps there is something to be said in favor of φεύγει: the idea is that copyists changed the wording of Jesus’ movement to avoid giving the impression that Jesus was running away like Brave Sir Robin. On the other hand, John does not use φεύγει anywhere else, either.
Constantine von Tischendorf was willing to let Codex Sinaiticus overrule all other Greek manuscripts in existence in John 6:15, in his eighth edition of the Greek New Testament. Scholars less entranced with À have not concurred, arguing (as Metzger reports in his Textual Commentary) that φεύγει originated as a Western reading. It had to be introduced as a sort of paraphrase fairly early, inasmuch as it is supported by Tertullian, De Rebaptismate, Ambrose, Jerome (who placed fugit in the Vulgate, agreeing with several Old Latin copies) and Augustine. (Tertullian’s reference is not airtight, however: in On Idolatry XVIII, he simply says that Christ “shrank back from being made a king,” without explicitly quoting John 6:15.)Ἀνεχώρησεν is read in all text-types: Alexandrian (P75, B, L), Western (D), Byzantine (A, K, Λ, Π, N) and Caesarean (f1). But did John write only ἀνεχώρησεν or ἀνεχώρησεν πάλιν? Manuscripts without πάλιν here include W, Δ, Ψ, and most Byzantine copies. The Peshitta, likewise, does not support πάλιν. The word could perhaps have been skipped accidentally, if a copyist’s line of sight slid from the ν at the end of ἀνεχώρησεν to the ν at the end of πάλιν. Or, having just read John 6:3’s account of Jesus going up into the mountain, early copyists couldn’t resist harmlessly adding πάλιν – but one would have to believe this happened at least twice for πάλιν to drift into A, D, Π, N, and Y.
The Curetonian Syriac, interestingly, supports a conflate reading: fled and withdrew again.
Readers of the Nestle-Aland Novum Testamentum Graece compilation would never detect that the reading found in the Byzantine Text, W, and the Peshitta exists: the N-A apparatus mentions the reading in À but not the reading of most manuscripts – an omission that is difficult to explain except as an effect of the editors’ pro-Alexandrian bias. The UBS apparatus doesn’t mention the non-inclusion of πάλιν either; nor does they Tyndale House GNT’s apparatus.
The non-original reading φεύγει is more interesting than it looks – not only because it was temporarily favored by a leading textual critic in the 1800s, but because it reveals that retro-translation sporadically affected the Greek text in the exemplar of the opening chapters of John in Sinaiticus – the same sort of Western phenomenon that has unfortunately affected the NIV 2011’s base-text of Mark 1:41.