My previous remarks having disappeared, let me try again.
Sorry, Tepan, but using prepositional phrases to modify nouns is as old as can be and it has been adapted to the ‘pi’ rule for at least half a decade, if not longer. That pu never uses the construction and so never needs to apply the rule (which it doesn’t notice in other contexts) is o evidence that it is wrong, even for puers. The language restricted to pu is a pretty sparse language indeed. But we know that Sonja thinks of it as much larger and along pretty much these lines, though she often has trouble formulating what she wants (the ‘jan pi li lasa e waso telo’ structure, for example, which she gave up but did not take up the obvious -- and common -- ‘jan pi alasa pi waso telo’).
Speaking of unpu, you in. fact cannot say "ona li tu e telo ni: ona li lon anpa, tan telo ni: ona li lon sewi.” because it is neither a sentence nor a string of sentences. In short, ou cannot embed a display inside a sentence then continue the sentence after it (this violate the antirecursion clause). I don’t know a handy way to do this legally, maybe ‘ ona li tu e wan telo tan wan ane telo. wan li anpa. ante li sewi’, which loses some zip.
I agree there are better ways to say “plants having seeds” than ‘kasi pi jo kiwen’ but it is perfectly legitimate. To be sure, it comes from ‘kasi li jo e kiwen’, but that is just the case with all modifiers: ‘kasi laso’ come from ‘kasi li laso’ and so on. Tepan doesn’t do derivations, whih may be why some things he says are a little forced.