[YG Conlang Archives] > [engelang group] > messages [Date Index] [Thread Index] >

[Date Prev] [Date Next] [Thread Prev] [Thread Next]

Re: [engelang] Xorban: Termsets

On Sat, Sep 29, 2012 at 10:32 AM, Jorge Llambías <jjllambias@hidden.email> wrote:
 
Ignoring aftertrees for now, I think this *might* remove the foretree ambiguity:

formula := core-formula | closed-foretree-formula | open-foretree formula
closed-foretree-formula := closed-foretree formula | term
closed-foretree-formula
foretree := open-foretree | closed-foretree
closed-foretree := JA foretree foretree
open-foretree := HE | term open-foretree
core-formula := CCA | term core-formula

term := NA | LA formula | JA formula

Assuming that's unambiguous (and that it generates everything we want
it to generate), we now need to add HI in such a way that it doesn't
create ambiguities with parentheticals.

co ma'a xrxe


It's interesting to consider this problem first while leaving out closed foretrees.  Leaving out CFTs, your putative grammar for formulas shrinks to:

=================================

formula := core-formula | open-foretree formula
open-foretree := HE | term open-foretree
core-formula := CCA | term core-formula
term := NA | LA formula | JA formula

=================================

With that, starting with a formula, if we produce a core-formula, we can subsequently get any number of terms:

    formula
    core-formula
    term core-formula
    term (term core-formula)
    ... [etc.]

... but we can never go back and get a tree.  Alternately we can get any number of open foretrees:

    formula
    open-foretree formula
    open-foretree (open-foretree formula)
    ... [etc.]

... but we can't get a term directly from the formula until we commit to producing a core-formula, and the terms will be on the right side of the foretrees.

I think that works so far.  I'm going to make a quick intuitive leap here: why the distinction between closed and open trees?  Why not now just generalize the open foretrees to all foretrees?  As in:

=================================

formula := core-formula | foretree formula
foretree := HE | term foretree | JA foretree foretree
core-formula := CCA | term core-formula
term := NA | LA formula | JA formula

=================================

Re whether the rules generate generate everything we want: I think that the rules generate exactly what we want, namely zero or more implicitly conjoined foretrees per "clause".  "Subordinate clauses" can have their own foretrees, but they would have to be embedded inside terms.  Revising my older example:

la ma djna he la ma lce he lu ra grka he ra mlta he fu ju glka xkra ju stda jnvaku.
"John and Alice, that all dogs and all cats are happly black, firmly believe."

So far, so good (?)


--
co ma'a mke

Xorban blog: Xorban.wordpress.com
My LL blog: Loglang.wordpress.com