Re: [engelang] termsets (I think)

Well, things were changing post to post and I was changing the terminology along the way so it'd be surprising if anyone comprehended anything.  But this is what we worked out so far.

We start by looking at the logical form and work towards the "encryption".

je ( la grka _le rcte ctka_ )( la mlta _le rcte ctka_ )
"Dogs eat meat and cats eat meat."

which is two formulas parenthesized, coordinated by "je", and having the same predication underlined.  What if we had a particle that acted like a placeholder for a predication shared by more than one "term list"?  Then maybe we could replace the duplicate predicates with particles and postpone the common predication to the end:

je ( la grka he )( la mlta he ) _le rcte ctka_
"Dogs and cats eat meat."

Since the "he" does all the work of connecting the term list to the predication, we drop "je" as superfluous (the "he" rules will write it back in implicitly anyway).

( la grka he )( la mlta he ) _le rcte ctka_
la grka he la mlta he le rcte ctka

If we want a non-default coordinator, then we can use that

ja la grka he la mlta he le rcte ctka
"Dogs [eat meat] or cats eat meat"

jo la grka he la mlta he le rcte ctka
"Dogs [eat meat] iff cats eat meat"

If we want to negate one term list, then we can do that:

la grka ni he la xrma na he le rcte ctka
"Dogs do, and horses don't, eat meat.

We can also add other terms, and we can add coordinators inside term lists so that they terminate with more than one "he".  That's why we settled on calling them *term trees*, of which term list is just a special case.  There's also a way to add term trees after the predication.  All of these things are just rewritings of a right branching tree terminated by several instances of the same predication.

That's the basics.