There is a possibility that Mk 1:1-3 were not original, or perhaps that the original beginning of Mark was lost at the same time that the original ending was lost (e.g. by the loss of some or all of the outermost sheet of a codex). All three verses have significant variants, about which Wieland Willker comments in A Textual Commentary on the Greek Gospels Vol 2 Mark:
A third alternative, not backed up by manuscript evidence though, is that the beginning of Mk is (as the ending) somewhat corrupt. This has been argued by several authors (see verse 1 for references). Some argue that verses 2-3 are an early gloss. Compare E. Güting (TC Mark, 2005, p. 53-55): "The beginning of the Gospel of Jesus Christ was John the Baptizer, who, in the desert, was proclaiming a baptism of repentance ...". If verses 2-3 are a harmonization to Mt/Lk, it is a rather sophisticated one, because verses 2 and 3 come from different places. Lachmann suggested that the two verses are the result of a conflation.
The references given by Willker include the following:
J.K. Elliott "Mark 1:1-3 – A later addition to the Gospel?" NTS 46 (2000) 584-8 [argues for a lost beginning, verses 1-3 non-Markan]
N. Clayton Croy "Where the Gospel text begins: A non-theological interpretation of Mk 1:1" NovT 43 (2001) 106-127 [also argues that the beginning has been lost, verse 1 is non-Markan]
Tommy Wasserman "The 'Son of God' was in the Beginning (Mark 1:1)" JTS 62 (2011) 20-50
Both Elliot and Croy argue for non-Markan beginnings for Mark. For example, Croy writes:
To sum up, then, Mark 1:1 is unlikely to be a title, grammatically independent of what follows, but it is equally unlikely to be a main clause, having its continuation in 1:2-3 (or 4). This dilemma might be solved, however, if vs. 1 is a post-Markan gloss.
… The proposal is simple: the beginning of Mark’s Gospel is defective. Something obviously preceded Mark 1:2 but has been lost. A superscription was added in various forms by redactors probably in the second century, not as a title but as a manuscript marker with the sense: “the Gospel begins here.” Some such marker was necessary since a manuscript beginning with Mark 1:2 would have been nearly intolerable on the grounds of grammar, style, and coherence.
While Wasserman disagrees, he does site an amulet as an early witness to the omission of “the Son of God” from Mk 1:1:
At the Annual Meeting of the Society of Biblical Literature in New Orleans, 2009, Geoffrey Smith announced a new early papyrus from Oxyrhynchus as witness to the short reading. It contains Mark 1:1–2, and the first verse reads: arch tou euaggeliou ihsou tou cristou. The second definite article in front of cristou is unique in the Greek manuscript tradition. On the whole, however, the text in the two verses is akin to that in Codex Koridethi (Y 038). Smith cautiously assigned the yet unedited papyrus to the third/fourth centuries... the text is introduced by this curious appeal on the first line: anagnwti thn archn tou euaggelliou kai ide (‘Read the beginning of the gospel and see’), which is clearly set off from the subsequent gospel text by indention, line spacing, and very different line length.
… Apart from P45, which is not extant in the opening chapters of Mark, there are no early papyrus witnesses to the Gospel of Mark. Therefore one might be tempted to assign great weight to this witness. As an amulet, however, this papyrus does not belong to the New Testament textual tradition proper and will therefore not likely be registered with a Gregory–Aland number. On the other hand, it may still be significant for the reconstruction of the New Testament, not least by virtue of its age. Otherwise, there are at least two other Greek papyrus amulets with the incipit of Mark, and they in fact attest to the long reading: P. Berlin inv. 6096 (fourth century?) and PSI VI 719 (fourth–fifth centuries, from Oxyrhynchus).
This amulet, which may be the earliest extant ms of Mk 1:1-2, is now known as P.Oxy LXXVI 5073, as indicated by Joseph E. Sanzo:
The editors date this artifact to the late third or early fourth century CE primarily on the basis of paleography (Smith and Bernhard, “Mark 1 1-2: Amulet,” 21). If the editors have dated this amulet correctly, P.Oxy LXXVI 5073 would be considerably earlier than any extant amulet with the Gospel incipits. Given that the Marken incipit is missing from two other ancient amulets (i.e., P. Cairo CG 10696 [no. 16] and BGU III 954 [no. 15; the Lucan incipt is also missing here]) and its wording seems to be confused with Johannine language in another amulet (PSI VI 719 [no. 10]; for discussion, see n. 55 in this chapter above), the early date of P. Oxy LXXVI 5073 is surprising and, I believe, warrants further study.
Larry Hurtado agrees with Sanzo, stating: “Given the scarcity of extant manuscript evidence for Mark in the first three centuries, even this curious fragment is worth attention for text-critical purposes,“ and notes these interesting variants in the text of the amulet:
The text witnesses to the opening line of Mark as “the beginning of the gospel of Jesus Christ” (i.e., without “Son of God”). Both “Jesus” and “Christ” are written as nomina sacra (ιηυ and χρυ respectively), and, interestingly, the Greek definite article precedes “Christ”… The text also witnesses to the reading “Isaiah the prophet” (in v.2).
In addition to the existence of the difficult to explain variants noted by Willker et al., another factor that points to some of Mk 1:1-3 not being present in the copy or copies seen by aMatthew and aLuke is their apparent lack of knowledge of some of this text, at least in the beginning of Mark, with neither of them using Mk 1:1 and 2b. However, both Matthew and Luke do contain re-located parallels to Mk 1:2 later, at Mt 11:10 and Lk 7:27 respectively, although neither give the name of the prophet here.
From the mss evidence it is clear that Mk 16:8 was (and still is) taken to be the original ending by many people, resulting in the addition of the long and short endings in various mss. It is even more likely that if there was damage at the beginning of Mark it would be recognized as such, with consequent attempts to supply suitable additional text. Undoubtably if most of Mk 1:1-3 were missing this would have been ‘fixed’ in different ways in copies of the faulty ms, leading to the situation we see today in which Mk 1:1-3 are present in some form in all mss, albeit with variants that are hard to understand if these verses were original. Croy concludes:
If Mark’s Gospel is, in fact, defective, the question naturally arises, what is missing? What has been lost from the beginning of Mark’s Gospel? From what has already been said above, the material most obviously lacking in Mark’s beginning is introductory information about John the Baptist, i.e. a statement of the “fulfillment” that would precede the statement of the “prophecy.” There probably was some reference to the temporal setting of John’s ministry (cf. Matt 3:1a; Luke 3:1-3; and Mark 1:9 in regard to Jesus). Presumably, the lost beginning did not duplicate information about John’s origin or call that would introduce and constitute a “fulfillment” of 1:2-3. The lost portion may have been no more than a few verses or a few paragraphs. Greater precision in stating its content or extent is not possible.
References
Croy, N. Clayton: Where the Gospel text begins: A non-theological interpretation of Mk 1:1 NovT 43 (2001) 106-127
Elliott, J.K: Mark 1:1-3 – A later addition to the Gospel? NTS 46 (2000) 584-8
Hurtado, Larry: New Oxyrhynchus Manuscripts, July 2, 2013
Sanzo, Joseph E: Scriptural Incipits on Amulets from Late Antique Egypt, 2014
Snapp, James, Jr: A Defense of "In the Prophets" in Mark 1:2, 2010, also Mark 1-2, parts one, two, three and four (October 2017)
Wasserman, Tommy: The 'Son of God' was in the Beginning (Mark 1:1) JTS 62 (2011) 20-50