I doubt Toki Pona, were it to acquire a community in one place, would remain isolating. It's already developed arbitrary meanings for many compounds, so my guess is it would become agglutinative and acquire voiced (pre-nasalized) consonants. If the CBB is anything to go by, that is the most common fate of isolating languages. The newbie starts with an isolating language, and stress patterns change it into agglutinative languages. But for the sake of argument, TP e and en are already easy to mistype, and some persons do reduce all post-tonic syllables to a schwa. Without a speaking (rather than typing) community, it's hard to tell how various individuals treat secondary stress. Going with Mandarin tones (which I've been using over in Romlang to derive a Mandarinesque Romlang), jan pi tokipona > jap1 to4po1 or ja1 pi2to41. e2 and e1 would come from e and en, assuming e doesn't drop out altogether. Would jan unpa > ja1u1pa2 or ja2nu1pa1? Maybe I'll cook something up in the jan nasa section of the toki pona forum, but of course, I can't claim it under a Creative Commons license.