I don't think the idea is really that first proposition. The idea is more along the second line of yours I quoted. Roy won't be "the real point guard." Roy has above average court vision and passing for a shooting guard. Bayless may have below average court vision and passing for a point guard. I'm not convinced that each player needs to be the platonic version of their position, so long as the "right amounts" of passing, scoring, rebounding and defense are there. In terms of court vision and passing, if Roy's above-average-ness at his position makes up for Bayless' below-average-ness at his, that seems like a fine model to work with.
I think distribution of "point guard" duties over both guards can work fine. Neither Bayless or Roy may be ideal point guards, but if they can both play-make to some extent, I'm willing to believe that that will be equivalent to the "traditional" model of a point guard who largely handles all the play-making duties and a shooting guard who doesn't do much distribution.
I don't think who specifically does what is the key. I think the key is that the team, overall, has all the things it needs. In an extreme case, if the team gets 10 RPG from its shooting guard and its power forward gets 4 RPG, I don't think that's a problem--all that matters is that the team gets all the rebounding it needs. In the case of the Blazers, I think the team will get all the passing it needs, as Bayless and Roy are probably at least average for NBA backcourts in terms of passing and Aldridge and Oden are probably above average for NBA front courts in terms of passing. Not having a classic point guard doesn't seem like a major issue.