On Oct 28, 2008, at 21:32, Robin Lee Powell wrote:
On Tue, Oct 28, 2008 at 10:30:16PM -0300, Jorge Llambías wrote:On Tue, Oct 28, 2008 at 10:24 PM, Robin Lee Powell <rlpowell@digitalkingdom.org> wrote:I don't believe, but am not certain, that that is possible. That is, I believe that if there is more than one way to generate a string, you've describe an ambiguous language, and that such a description would not, in fact, describe Lojban.See example 3 in <http://en.wikipedia.org/wiki/Context-free_grammar>You win.
It seems to me that if you have such an ambiguous grammar, then it is, if not actually wrong, then misleading: the parse tree has degrees of freedom which are not actually part of the structure of the language.
For example, the trivial grammar SELBRI -> BRIVLA | SELBRI SELBRIwill generate only valid selbri, but non-left-leaning 'parse' trees may suggest tanru structure which isn't actually present (due to lack of ke...ke'e).
(On the other hand, you can have an unambiguous grammar whose trees go the wrong direction; e.g. right-associative tanru.)
(An unambiguous grammar for the same set of strings would be SELBRI -> BRIVLA | SELBRI BRIVLA An unambiguous grammar for the same set of parse trees would be SELBRI -> SBATOM | SELBRI SBATOM SBATOM -> BRIVLA | "ke" SELBRI "ke'e" ) -- Kevin Reid <http://homepage.mac.com/kpreid/> To unsubscribe from this list, send mail to lojban-list-request@lojban.org with the subject unsubscribe, or go to http://www.lojban.org/lsg2/, or if you're really stuck, send mail to secretary@lojban.org for help.