Q: You find parallels between Mark's empty tomb narrative and Orphic mystery narratives, and between Luke's Emmaus narrative and the epiphany of Romulus, but despite the parallels, aren't there still a lot of differences?
A: Of course. But adapted myths acquire meaning precisely because of what is left out and what is kept in, as well as by what is changed. In other words, Mark deliberately left out of his account everything in the Orphic narrative that he rejected, and kept in everything that still had direct parallels with the gospel message. And then he changed details specifically to convey how his message was different from the Orphics (pp. 161-63). The same goes for Luke's transvaluation of the Romulus narrative (pp. 180-82) and so on (like possible parallels to the Osiris myth: p. 159). That is the whole point of including such parallels: certain readers would immediately get the parallel (or be taught it in secret initiations) and then they would understand what it is that Mark is really saying. Mythic elements are in that respect just like words: the words are the same and carry the same meaning, but when you select and rearrange them, you say something different. Readers or initiates would see the elements, the symbols, as words with distinct cultural meanings, and would see their careful selection and rearrangement as what was being said with those symbols. For more on this, see the chapter by Evan Fales, "Taming the Tehom" (pp. 307-48).
Another example of Carrier’s circular scholarship. Of course, if you already knew that x was literarily dependent on y, then, and only then, could you claim that x “deliberately left out of his account everything in the y-narrative that he rejected, and kept in everything that still had direct parallels with the x-message.”
Unfortunately for Carrier’s argument, what is missing is evidence for direct parallels in the first place.
He appeals to deliberate omissions to harmonize his theory with the actual state of the record. But absent independent evidence that x is, in fact, literarily dependent on y, his harmonization is a classic case of backward reasoning.
If he already knew that x and y were truly parallel, in some genealogical relationship, then he might be able to account for the disanalogies by appeal to selective editing—but if the only evidence he as to work with are x and y as they stand, and if he must appeal to selective editing to harmonize his theory with the actual state of the evidence, then the evidence itself does not support his theory. Rather, his theory is trimming the evidence to agree with his preconceived hypothesis.
More desperate still is his rearguard appeal to esoteric teaching, without which even the original audience would be unable to discern the alleged parallels.
This is a backdoor admission that alleged parallels are absent from the canonical text itself.