Re: resolution of open Lojban issues - limited distributiuon message

[Veijo Vilva <veion@xiron.pc.helsinki.fi>]

Putting it all together:

terms_80                :  term_81a
                        |  terms_80  term_81
			;
			
term_81a                :  term_81
			|  GEK_807 term_81 links_161 GIK_816 term_81 links_161
                        ;
			
term_81                 :  sumti_mod_chain_82
			|  GEK_807 term_81 links_161 
			    GIK_816 term_81 links_161 BO_508
                        |  NA_KU_810
			;
			
sumti_mod_chain_82      :  sumti_or_mod_84
                        |  sumti_mod_chain_82 CEhE_680 sumti_or_mod_84
			;

sumti_or_mod_84         :  sumti_90
                        |  mod_head_490 gap_450
                        |  mod_head_490 sumti_90
                        |  mod_head_490 CEhE_680 selbri_130 tail_terms_71
			;


Notes:

   1) there is no possibility of grouping within a chain

   2) The last rule for sumti_or_mod_84 makes terms like "zace'e nanca
      lire" possible. VAU is required at the end of tail_terms if the
      chain continues. "vau ce'e" can only be avoided by adding a new
      cmavo (ce'u ?). distance+refpoint could then be expressed as
      follows:

              vace'e mitre limu ce'u le vorme 

      The added rule would be:

         |  mod_head_490 CEhE_680 selbri_130 tail_terms_71
               CEhU_681 sumti_90

      This would give the shortest possible expression for the most
      common case. An added bonus is that a following ce'e doesn't
      tie to the sumti but to the whole phrase. 

I'm ready to buy this model in case you still don't buy a limited relative 
clause model with NOI_585 replacing CEhE_680 in the last rule.
I think, however, that a rule of either type is the only Zipfean solution
-- anything building on a sumti structure adds unnecessary complexity.

  Veijo