Until someone shows me a good theoretical argument to the contrary, I view merism to be a post-syntactic interpretation process.
In terms of valency (verb semantics plus syntax), I see no difference between "he created fish and plants" (obviously not a merism) and "he created the heavens and the earth" or "he created good and bad".
Thanks. That's what I was really hoping for when I suggested an application of the double complement search, i.e., how a trained linguist would describe merism. And your description just confirms what I already thought, even though you expressed it much more elegantly. Even if there are a few merisms in the HB, post-syntactic interpretations of them differ so wildly that I wonder if the term is of much use at all. Even your two examples are open to interpretation.
I guess what I was getting at is I see no difference either, and if lexical items identified as merisms are regularly separated by syntax (in adjoining clauses, phrases, etc.), then merism is definitely diminished in value.
I know this is a bit off topic, but against the background of Egyptian cosmology your interpretation of בראשית (אשר) ברא . . . would benefit from השמים and הארץ being viewed as distinct entities. But of course, another case of post-syntactic interpretation process (also something I'm actually writing about).
Anyways, again, thanks for the linguistic comments, irrespective of whether you follow or agree with me. My heart leaps when I read these comments.