Re: [lojban] Mathy person interested in concept, unsure where to begin.

We are at a bit of cross purposes here. Team 1 is using the full potential of a modern theoretical grammar, one that would derive every sentence of a language from some formula, not just sentences of some set trivially matching the structure of the logic. For most languages and, indeed, for most sentences in those languages, several non-equivalent formulae may give rise to the same sentence (most languages are syntactically ambiguous). Lojban is planned to avoid this: a given sentence can come from only one formula (up to equivalence - speaking of which, of course, equivalent sentences in Lojban derive from the same formula or equivalent ones). This means that every logically significant feature of the formula must be represented somehow in the sentence and. If that representation is shortcut somehow, that shortcut must be marked to allow a unique reconstruction. The logic > Lojban process and the Lojban > logic are of course distinct but presumably developed together very closely, as is the surface grammar available for ordinary use (PEG at present). So the tests proposed are not trivial.

Sent from my iPad

On Sep 24, 2014, at 14:32, Jorge Llambías <jjllambias@gmail.com> wrote:

On Wed, Sep 24, 2014 at 4:16 PM, 'John E. Clifford' via lojban <lojban@googlegroups.com> wrote:
Well, in theory in general, every sentence is derived from a formula. The question is whether we can automatically derive that formula from the sentence.

With today's Lojban, in general we cannot. (For a restricted set of Lojban sentences, we can.)

The best (only?) way to test this is to see whether the formula team 2 derives from a sentence supplied by team 1 is the same formula that team 1 used to derive the sentence in the first place.

I don't think that would prove much. If team 1 and team 2 are minimally competent, that test would be passed with flying colors. Indeed that test could even be passed by two automatic converters, since conversion from Logic to Lojban can be done automatically, and the reverse can also be done automatically when starting from the restricted set of Lojban that the first part would generate.

The true test would be to start with a general Lojban sentence, then convert it to a formula (that's the hard part), and then see how well the automatically generated sentence from that formula matches the original sentence.

A weaker test is to see whether the automatically derived formula corresponds to the intuitive reading of the sentence as provided by proficient speakers who re logically sophisticated, etc.

There is currently no automatically derived FOPL formula for Lojban sentences in general. (Only for a restricted subset of Lojban there is.)

mu'o mi'e xorxes

--
You received this message because you are subscribed to the Google Groups "lojban" group.
To unsubscribe from this group and stop receiving emails from it, send an email to lojban+unsubscribe@googlegroups.com.
To post to this group, send email to lojban@googlegroups.com.
Visit this group at http://groups.google.com/group/lojban.
For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "lojban" group.
To unsubscribe from this group and stop receiving emails from it, send an email to lojban+unsubscribe@googlegroups.com.
To post to this group, send email to lojban@googlegroups.com.
Visit this group at http://groups.google.com/group/lojban.
For more options, visit https://groups.google.com/d/optout.