It's probably high time we moved on from our "photo-ontological" musings, as people who are not interested in the topic must be getting bored by now. But before we go I'd like to add one more idea and one recommendation to the mix.
Actually, the recommendation is only a "what he said." Those who are talking in the comments about Tom Wolfe's The Painted Word and From Bauhaus to Our House (both great fun and easily recommendable, although not entirely applicable to today's art world and its main concerns) should check out Michel Hardy-Vallée's recommendation of David Davies' Art As Performance. Davies begins his book with several long quotes from Wolfe—and a gentle critique of him, which is also, um, desirable. Art As Performance is a somewhat more difficult book (although still very readable) but, in the end, it's a more rewarding one.
The following is just something I want to throw up for your consideration. It seems to me that a number of readers are confused between photographic rendering that needs to be interpreted and an alleged departure from the "straight," or real, or reportorial principle.
Photographs show (or "tell us") certain things and don't tell us other things; but all photographs still need to be interpreted, often based on the viewer's knowledge of photographic techniques and means, and all such interpretations are provisional, with greater or lesser foundations of evidence. In that respect they're like any other attempt to arrive at truth, whether you're an historian trying to interpret what happened in the past or a reporter trying to put across what happened in a news event or a judge and jury trying to weigh the truth-value of evidence to come to a determination of guilt or innocence. No truth is absolute, perhaps, but that doesn't mean there is no truth.
The provisional aspect of the facts contained in a "straight" photograph do not mean that there are no facts there. This can be proven in a number of ways, but perhaps the most obvious way is that you can extract facts from pictures that the photographer didn't specifically intend to point out, or perhaps wasn't even aware of. Astronomers were able to pinpoint the exact moment that Ansel Adams's "Moonrise, Hernandez, New Mexico" was taken, but that wasn't because Adams was specifically trying to show what time it was by astronomical means when he took the picture.
In John Lehet's barn photograph two posts ago, for example, the primary fact I'm pretty sure of is that there was a barn there. That has a pretty high evidence-quotient even in his "expressive" photograph and I'm fairly certain of it. I'm also pretty sure I've gotten a reasonably accurate idea of the relative positions of the barn, shed, and tree, and that the barn has a metal roof. Now, I personally know what "infrared" or partial-infrared rendition looks like, so I'm not fooled into believing that the field and tree are covered with snow (although the evidence of this is somewhat less sure); someone who has never seen an infrared photograph might be immediately convinced that the meadow is snow-covered, and we could argue about what the fact in the photograph actually is. One of us would be right and one would be wrong—there was either snow present in the meadow when the picture was taken or there wasn't—but neither of us could be 100% sure of our conclusion based solely on the picture. We'd have to admit that our conclusions were provisional. (Cf. Roger's Fenton's cannonballs.)
Now consider what the photograph tells us about...oh, I don't know, let's say, what State it was taken in. There, the evidence is minimal or nonexistent. We can make a stab at a guess (e.g., "it probably wasn't Hawaii"), but it would just be speculation. It was taken somewhere, but exactly where is a fact that the picture itself doesn't obviously contain.
The point is that none of the aspects of technical photographic rendering put forth by people as being "unreal" are actually departures from "straight" photography, necessarily. Whether a photograph is luminance-only (i.e., black-and-white or monochrome), whether a sky has a creamy tone or some objects are out of the depth of field or rushing water is motion-blurred—all such things merely affect the quality of the evidence the photograph offers, and perhaps increase or decrease the necessity of interpretation. You can still pull some facts out of the picture based on interpretation (having seen unnumbered pictures of motion-blurred running water, for instance, I am unlikely to conclude that someone has gone and hurled dry ice into a stream), and other facts remain fugitive or absent. But that's the situation with any photograph. And it's the same way with any other attempt to determine truth, whether it's a statement in a history textbook or an assertion by a witness in court.
The change that has overtaken "photography" is simply that a larger number of facts contained in new pictures are now less secure. If the moon in "Hernandez" had been Photoshopped in, or moved, then the time of day based on its position would no longer be a fact contained in Adams's picture. If you can't reasonably determine from the picture whether the moon was Photoshopped in or not, then "a" fact about the time of day is still retrievable, but its quality as evidence plummets—it might have been such-and-such a time of day, or it might not have been. You might not care; Adams might not have cared. But the difference is real.
The only reason this matters is that the only thing that is interesting about many photographs, or what is most interesting about them, is the evidence they give of something that actually was. In Paul Strand's famous picture of the blind beggar woman, we can have a high degree of confidence that her sign actually said "BLIND" and that her left eye actually appeared to be open and peering at something in that wonderful furtive manner. If our confidence in either or both of those facts goes from 98% to, say, 30%, then the enigmatic, emblematic, numinous masterpiece becomes more like a mild joke with about the same level of interest as a Hallmark card or a New Yorker cartoon. The fact that we can't tell what color her cloak was isn't something that makes the photograph unreal, and it's no reason to throw up our hands in a fit of relativism; it's just one fact that the photograph doesn't happen to contain, is all. One of many (she might actually have been totally blind, for instance; the picture doesn't tell us that).
That's just the way it is. This is nothing that digital photographers need to have a chip on their shoulders about, but neither does having such a chip on one's shoulder change anything. Our newfound digital freedom gives us a lot—an awful lot—but it also taketh away; and I think it's wise to be aware of that.