And is it a modern thing?
My intuition is that this is more frequent in contemporary writers. I have seen it throughout Ringe's two volumes on English, in Owens on Arabic, and (though I can't bring more examples to mind) in many other places.
I can only imagine that people who use it consistently feel "x, which can have developed from y or z" to be more specific than the (to me) more natural "x, which could have developed from y or z". But I think the could/may/might variants are just as clear, assuming we are talking about an x which is known to have actually developed, as is almost always clear from context.