Many people have emailed to alert me to Errol Morris's latest article in the Times. In this one he's concerned himself with how our—well, how Susan Sontag's— interpretations of pictures might be built around incorrect assumptions. Since the larger argument hinges on which picture was taken first and which second, Morris spends a great deal of time and energy analyzing that question, and ends by asking readers to help.
Perhaps appropriately, many of the comments (there are now well over 700 of them, making the entirety difficult to read through) are mistaken. Many people simply don't know what they're looking at—while thinking that they do.
First, I think all the many comments analyzing the alleged "shadows" are mistaken. Below is an enlargement (to pixel visibility) of the large TIFF of the print made available by the Library of Congress. As you can see, there are no clear sunlight shadows cast by the cannon balls at all. As a darkroom expert with 25 years of experience, I'm reasonably certain that what people are interpreting as "shadows" are in fact the result of one of the two pictures having received less exposure than the other. Albumen glass-plate photography is a double-negative process, so when the plate (negative) is exposed a little more, the print then has to be exposed a little more also, since the negative that received more exposure has more density. This naturally affects the tonal relationship of the so-called "shadow" areas of the picture (the dark tones), which expand or contract somewhat along with less or more negative exposure, respectively. But this effect is not evidence of actual specular shadows created by the sun. Just looking at the detail of the cannon balls in the "ON" photo confirms that there are no definitive shadows to analyze (analyzing the shadows, n.b., is different from analyzing the lighting).
Secondly, several comments (for example, #296) try to analyze the prevailing lighting by looking at the image of the sky. Although a perfectly natural impulse, here again is an example of the perils of non-expert interpretation.
Albumen plates were highly blue sensitive, which meant that many times when the plate was exposed for the ground features, the sky was exposed past the threshold of reciprocity and became unnaturally mottled. For this reason, 19th-century landscape photographers often painted out the sky on the plate itself (think White-Out!) and the sky then prints as a featureless white (or rather, it prints as "paper base," which was often a cream or sepia color). See American frontiers: The Photographs of Timothy H. O'Sullivan, 1867-1874, by Joel Snyder of the University of Chicago (Aperture, 1981) for more depth on this topic.
I’ve enlarged the Library of Congress TIFF of the "ON" Fenton print in Photoshop, and I can’t see if Fenton actually did this or not in these particular pictures, but that might just mean that he did it skillfully—the featureless white sky, so typical of 19th century landscapes, is a strong clue that it was done.
Ironically, an excellent example of this is Alexander Gardner’s picture of the dead Confederate sniper at Devil’s Den, another picture that is frequently cited as having been set up by the photographer (there is considerable evidence that the body was moved and that the weapon is not a sharpshooter's rifle). If you look at the TIFF of that picture from the Library of Congress, you can clearly see the brush marks where he painted out the sky on the negative. So any attempt to analyze the weather conditions on the day the Fenton pictures were taken by looking at the "skies" in the picture are most likely to be fruitless, if not actively misleading.
I do find one major anomaly in the article itself: after he quotes a lengthy passage from one of Fenton's letters, the author summarizes "the facts as expressed" in the letter, five in number, but he leaves out the fact that is most conspicuous—which is that Fenton is witnessing cannon balls bounding up the road and coming to rest near him, in one case literally at his feet (which he then picks up to bring back to his wife as a souvenir). Why Errol Morris would ignore this evidence of Fenton's own testimony is mystifying, unless it's just the typical critic's attitude that photographers are trained monkeys too stupid to know what's in their own pictures.
As for Errol Morris's main question—which of the two photographs was made first—I find myself 65% convinced by the excellent analysis by Wendy Ju of Stanford, although I think it's likely that the truth cannot be known to any reasonable degree of certainty. Incidentally, perhaps the world's leading expert on comparative photographic views has posted a comment—it's #703. And he agrees with Wendy Ju.
ADDENDUM: I think the larger point here is that we need to be very conservative about extrapolating from conclusions based on assumptions. Incidentally, I don't think (as some of the commenters in the Times thread claimed) that investigations like this one are trivial. We can't investigate thoroughly every single statement of claim ever made by a scholar or critic in the course of them making an argument, but that makes it more important, not less, that we investigate one or two every once in a while; it has the effect of keeping writers honest and making them think twice before they jump to a conclusion or state a claim—and of keeping our own skepticism sharp.
Very few people come to the same conclusion I do, which is that OFF came first and ON was the later exposure because the valley was under bombardment at the very time Fenton was there, and Fenton and his assistant had the camera set up just beyond the range of the cannon. Maybe that wasn't so; but the comparison of the photographs makes it seem unlikely that the cannonballs were placed by Fenton trying to make his picture more dramatic, because in the ON picture there are more cannonballs in road and in the ditch, and if the two of them were out there rearranging things for pictorial effect, a) why wouldn't they take balls from the ditch and put them in the road, and b) why would they add more balls to the ditch instead of placing them somewhere more visible?
In any case, I think it's really curious that both Errol Morris and so many commenters ignored what Fenton himself says in his letter to his wife: "...I brought the van down & fixed the camera & while leveling it another ball came in a more slanting direction touching the rear of the battery as the others but instead of coming up the road bounded on to the hill on our left about 50 yards from us & came down right to us stopping at our feet. I picked it up put it into the van & hope to make you a present of it." Notice that he says "instead of coming up the road"—in other words, other cannonballs were coming up the road (and it's perfectly logical that the road would be what the gunners were aiming at) and this anomalous ball he's talking about is remarkable to him because it rolled over the hill on the left and made it all the way down to where he is standing. This makes it seem evident to me that the road was being bombarded during the time that he's there with his camera set up. It also explains why he bothered to stay at the site for an hour or an hour and a half after taking the first exposure.
Of course, if that's so obvious, then why isn't everyone else saying the same thing? So I could well be wrong. (I always have tended to grant more credit to photographers themselves than many viewers—and critics—are willing to do.) All we really know is that the greater weight of the photographic evidence itself supports OFF before ON insofar as we can tell. —MJ