The fact that {zo'e} gets its meaning from context, specifically from the place it fills, is enough to undercut the definition of {lo}, since the referent (andf the class from which the referent is drawn) can then change with each use. But, further, the referent of {zo'e} has to be one that makes the head bridi true else it fails in its fundamental purpose, to fill the missing pieces in a claim (which is not going to be falsified -- we hope -- by what is not said). This does not mean that nothing but platitudes can be said in Lojban, only that the nonplatitudinous parts are explicit. What {zo'e} contributes (if anything), depends upon its scope (not resolved, but presumably shortest) and is only a problem in its supposed role in {lo}. This is an argumentum ad
absurdum against that role.
Insofar as parser expansions use MEX they are suspect, as is all of MEX ex officio. Once the core is properly dealt with (and this, again ex officio, does not involve MEX) MEX can be examined to see what way of dealing ith it fits (or can be made to fit) with the basics.
On Wednesday, October 8, 2014 6:18 PM, Jorge Llambías <jjllambias@gmail.com> wrote: