"The narrative requirement seems unusually constrictive. Doesn't it exclude a wide range of poetry that is purely subject-based? If so, why was that an exclusion you were willing to accept as editor?"
As I said, my definition was wildly unpopular with the sf poetry community. As tablesaw says, it excludes "a wide range of poetry that is purely subject-based." Absolutely; it was intended to.
For a poet who wants to write a poem about gravity, there's the ancient tradition of the science poem; the poet who wants to write a poem about unicorns has the fantasy poem genre available; the poet who wants to write a poem about blobs of slime has the horror poem genre available; and for everything else there is the cover term "speculative poetry." I wanted the science fiction poem to be defined in such a way that it would stand out from the pack of "speculative" poems and could be recognized unambiguously as an sf poem. I wanted it to be possible for people to say, "It's a splendid poem and I'm glad I had the opportunity to read it -- but it's not a science fiction poem." It was and is my opinion that that's the only way to build and establish the sf poetry genre and the reputation of its poets.
My definition -- the chunk of language has to meet the definition of a poem (mine and Grinder's, or some other recognized definition), it has to have a science component, and it has to have a narrative component -- was just a proposal. A hypothesis. I'd be delighted to see it replaced by a better definition that would accomplish the same purpose; my goal wasn't to establish my definition but to further the cause of science fiction poetry. So far as I know, no one has yet proposed an alternative definition. It's possible that there've been scores of alternative proposals; I can't keep up with even a fraction of the speculative poetry publications. But no one has approached me and said, "Suzette, this is my proposal for an alternative to your definition of the sf poem" or contacted me and said, "Suzette, there's a proposal for an alternative to your definition of the sf poem in the current issue of [some poetry journal or other publication]."
This is analogous to what happened with my LAadan language. I constructed it as a hypothesis. The hypothesis was "This is what a human language designed to adequately express the perceptions of women would look like." Many people objected to LAadan, and said the hypothesis was all wrong; many people said it wasn't at all the way they thought a language designed to express women's perceptions would look. But I'm still waiting for someone to propose an alternative.
The thing about a hypothesis is that it stands until it has been disproved, refuted, or replaced. If another hypothesis accomplishes at least the same amount of work at least as well, you have two competing hypotheses. If the second hypothesis can demonstrably do related work that the first one can't do, or can do some of that work more effectively and efficiently, then it has replaced the first hypothesis. That's how science is done; it has nothing at all to do with me as a person, or my opinions about poetry, or how well I write, or anything of that kind. I believe that science fiction poetry, for its own sake, should be written from within that framework.