[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [lojban] semantic parser - tersmu-0.1rc1
On Sat, Nov 26, 2011 at 2:14 PM, Martin Bays <mbays@sdf.org> wrote:
>
> (would you agree that "guhek selbri-3 gik selbri-6" is a more sensible
> grammar of guheks than the official one, by the way? I spent some time
> trying to make sense of the official "guhek selbri gik selbri-6", but
> failed - mostly because of the weird interactions with {co})
I think connectives should allow the same things for both connectands,
so I'm unhappy with both "guhek selbri gik selbri-6" and
"gek sumti gik sumti-4".
I haven't thought about it enough to have an opinion on what level
they should connect, other than that it should be the same on both
sides.
> I actually was, until I changed it a few days ago, handling selbri3
> after tailterms. What made me change was {broda be da brode ro da} - it
> seems counterintuitive to have the first {da} in the scope of the
> second, which is what I seemed to be forced to do when handling
> tailterms first.
>
> Do you see another way to avoid that? Or do you consider it correct?
I still don't know how you handle tanru, but it seems to me that
"broda be da brode" should be handled like "brodi be da", where the
exact derivation of "brodi" from "broda" and "brode" is probably
undefinable.
As for "brodi be da ro da", I suppose the first "da" has to be bound
by "ro", but I'm willing to reconsider.
> But I'm afraid I do intend to let conservatism take priority in
> this case, at least for now.
Is there any indication that CLL gives giheks tighter scope than
tail-terms? (I don't remember either way.)
>> You can use "na'e (ke)" instead of "na" to get your meaning easily,
>
> You have {na'e} giving logical negation?
What's the alternative?
>> With your gi'e-rule, my na-rule makes little sense. We both agree that
>> na is subordinate to gi'e, so if gi'e has tight scope, na must too.
>
> Hold on. I understand your gi'e rule as giving e.g.
> na broda gi'e na brode da
> -> ge na broda da gi na brode da ,
Did you leave out a "vau"?
I have:
na broda gi'e na brode su'o da
-> ge na ku zo'u broda gi na ku zo'u su'o da zo'u brode da
na broda gi'e na brode vau su'o da
-> ge na ku zo'u su'o da zo'u na broda da gi na ku zo'u su'o da zo'u brode da
> so there's still the issue of whether those {na}s scope over the {da}s.
If you have "su'o da" with scope over "gi'e", there is no doubt that
it will also have scope over "na".
>> It's clear that one rule will allow some things being shorter, and
>> another rule will allow other things being shorter. I don't think a
>> single example with no semantic content will be convincing either way.
>
> Sure. But note that I can get your rule by replacing the {na} with
> a {na ku} after the selbri, so just one syllable extra. Going the other
> way seems harder in some cases - e.g. the example I just gave.
I don't follow this particular example, but in general yes, although
your argument would be better if you use say "ka'e" rather than "na",
because we do have "na'e" for tight scope negation.
>> > Another argument would involve your agreeing with me that
>> >
>> > broda je brode da
>> > Prop: EX x1. (broda( ,x1) /\ brode( ,x1))
>> >
>> > is correct... do you?
>>
>> I agree it's EX x1 brodi(x1), and what you offer for brodi is probably
>> the obvious choice. But tanru are weird, so I won't commit to it. In
>> any case, I suppose you don't need the exact form of "brodi" for your
>> argument, but I don't consider na and tag to be part of the tanru, so
>> selbri attached "na" is different from tanru-making "je".
>
> Hmm. Why do you call this a tanru? The relevant grammar for
> selbri connection is selbri-4 through selbri-6. It's true that the usual
> use (and the only documented use, to my knowledge) of selbri connection
> is in a seltau, but I don't see why we should consider the connectives
> themselves as anything more complicated than boolean operators.
You're probably right. We just don't have a clear rule on what the
place structure of "broda je brode" is wrt the place structures of
broda and brode.
mu'o mi'e xorxes
--
You received this message because you are subscribed to the Google Groups "lojban" group.
To post to this group, send email to lojban@googlegroups.com.
To unsubscribe from this group, send email to lojban+unsubscribe@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/lojban?hl=en.