Andreas Kolbe wrote:
But if we use a *different* style, it should still be traceable to an educational or scholarly standard, rather than one we have made up, or inherited from 4chan. Would you agree?
Yes, and I dispute the premise that the English Wikipedia has failed in this respect.
As I've noted, we always must gauge available images' illustrative value on an individual basis. We do so by applying criteria intended to be as objective as possible, thereby reflecting (as closely as we can, given the relatively small pool of libre images) the quality standards upheld by reputable publications. We also reject images inconsistent with reliable sources' information on the subjects depicted therein.
We don't, however, exclude images on the basis that others declined to publish the same or similar illustrations.
Images widely regarded as "objectionable" commonly are omitted for this reason (which is no more relevant to Wikipedia than the censorship of "objectionable" words is). But again, we needn't seek to determine when this has occurred. We can simply apply our normal assessment criteria across the board (irrespective of whether an image depicts a sexual act or a pine tree).
David Levy