You had me wondering, especially since I found it hard to see right off the bat due to the significant difference in overall contrast - and it was easy enough to try. As to that difference in overall contrast, I think it's indeed quite pronounced and creates a different impression of the same image. I think that's what maybe you were responding to, and I find it a meaningful difference. Although of course in the end, it's still the same image. If we're trying to understand the composition (if there's such a thing as rationally understanding a composition), I don't think it matters all that much in the end.
My two cents is that all elements contribute to the image and if the characters in the corners would have been removed, the image would have worked differently and I expect it would have been less liable at keeping the viewer's attention. The people not directly involved in the action (although strictly speaking this is only
maybe true for the guy top left and the two people along the top edge looking out to the right) trigger questions about context and as such tie the central action into that broader context, most of which we can't see, and as a result, these questions remain unanswered. I suppose people like open questions, so it helps the image. That's the social bit at least.
From a graphic/image-wise perspective I think if you remove any of the people in the corners, the 'visual logic' (as
@gary mulder calls it) breaks down.
Of course, regardless of how we apply our hermeneutics to an image like this, it'll always remain tautological. The image has survived several selection processes and as a result remains with us today. Apparently it 'works'. We are then left to guess at the question why it works, but the answer always boils down to "because it does."
For me, an exchange like this is similar to comments (amateur) photographers sometimes make about each other's work - "you should have cropped half an inch here" or "it would have been better if that element wasn't included in the frame" etc. Such remarks always make me think "then it would have been a different photograph and we're now looking at
this one - so go out and make your own version of it if you feel it should be done differently." I've said it before in threads like this one - you like the image or you don't. Maybe you can express a little why you think you like it, but the response is primarily an emotional one. At the very least, if you try to rationalize an emotional response, a lot is lost in the process. At worst, you totally miss the point.
Sorry if that's not very constructive. Evidently I'm Ok with discussions like these; I find it interesting to see what others come up with, although in all honesty, the golden nuggets in such discussions are IMO kind of few and far between. But that's quiet personal. Simply engaging in the discussion can bring some feelsgoodman-vibes as well, I suppose.
PS: this also tangentially touches upon the hypothetical 500-word essay
@cliveh referred to. By all means and write it; please don't ever let me keep anyone from doing so. At the same time, my take on it is that Cartier-Bresson was a visual artist and not a writer, so he chose to do the 'essay' in the form of an image. He told the story in its original form, apparently pretty much as he intended to (given that it survived several selections), in just 1/60 of a second and an 8x10" surface area. We could try to understand it and re-tell the same story in 500 words, or 5000, but it would be a different story, and "we're now looking at
this one."
PPS: at some point in his career (and not a particularly high point at that), Devin Townsend realized that people are just sacks of cells making sounds towards each other. In discussions like these, I often have to think of that notion. It's kind of silly, rather nihilist too, and perhaps not very exciting or stimulating - but also there's an inescapable truth to it in my view.