In particular, passing this lojbanic form back through tersmu would
yield something involving {srana} whose equivalence to the original form
involving f(x) is far from immediate. Worse, there wouldn't even be
a fixed point for the lojban form - since {pe} without a quantifier is
taken as effectively introducing a {noi} clause, {ro da zo'u fy pe da
broda da} would become {ro da zo'u ge fyno pe da ge'u srana da gi fyno
pe da ge'u me fy me'u .i ro da zo'u fyno pe da ge'u broda da}, which on
another pass through would blow up those {pe} clauses even further...
Hmm. I've been adopting {lo broda} == {zo'e noi broda} as absolute
dogma, so it's really making a side-claim that the referent(s) broda(s).
You think a more accurate dogma would be
{lo broda} == {zo'e noi ca'e broda}?