Re: [lojban] Re: tersmu 0.2

On Sat, Nov 8, 2014 at 2:10 PM, Martin Bays <mbays@sdf.org> wrote:

* Saturday, 2014-11-08 at 11:28 -0300 - Jorge Llambías <jjllambias@gmail.com>:

> So following CLL, without logical connectives "broda .i [tag] bo brode"
> makes three independent claims:
>
> (1) broda
> (2) brode
> (3) lo nu broda cu xo'i [tag] (for tense tags)
> lo nu brode cu xo'i [tag] (for non-tense tags)

I read CLL as explicitly claiming a relation between events
corresponding to the two sides, not leaving one of them up to context
like this. So I read it as having
(3') {[tag] le nu broda kei brode} (tense)
{[tag] le nu brode kei broda} (non-tense)

Agreed, but that means the role of the brode event for tenses and the broda event for non-tenses is left up to context, since the tag only guarantees the x1 role of its underlying relation. Of course for many tags the role for the other event will often also be quite obvious, but it's not a part of the general logic of tags. With (3) I wanted to indicate that, but (3') also works.

> CLL analyzes the relatively easy cases (causals) where the tag-claim
> presupposes the subordinate claims in any case,

Implies rather than presupposes, I'd say (the distinction matters in
embedded contexts).

OK. In any case, something that may apply to a particular tag but not to another.

Because CLL only considers such cases, I think in the non-tense case we
can feasibly read it as only claiming (2) and (3'), with (1) being an
implication of (3') in the cases it considers.

I suppose we could, but it doesn't feel right.

Reading its account of the distinction between 9:7.1 and 9:7.7, where
the former is of form (3') and the latter is a non-tense connection, it
seems quite clearly to be saying that all it adds to (3') is a claim
of form (2).

Yes, I agree CLL seems to be saying that, although the distinction it makes in that particular example seems rather tenuous.

> but this may be problematic for other tags such as "se ba'i" or "se
> cau", where the negation of one of the subordinate claims would have
> to be presupposed. So for example:
>
> (a) mi na klama lo zarci .i ba'i bo mi stali lo zdani
> I don't go to the market. Instead, I stay home.
>
> (a1) mi na klama lo zarci
> (a2) mi stali lo zdani
> (a3) lo nu mi stali lo zdani cu basti (lo nu mi (ja'a!) klama lo zarci)
>
> That works fine, because the change from "na" to "ja'a" is not part of
> anything explicit in (3).

This doesn't work with (3'), and hence not with CLL as I'm reading it.

(3') gives:

ba'i lo nu mi stali lo zdani kei mi na klama lo zarci

With replacement my staying at home: it is not the case that: I go to the market.

I don't think it's too unreasonable for ba'i to provide a relation such that basti1 relates to the negation of basti2 rather than to basti2 directly, since this relation has to be contextual anyway. So I wouldn't say it doesn't work with (3'), at worst it takes more work than just assuming that "lo nu mi na klama lo zarci" has to take the basti2 role.

> But what if we wanted to say "mi stali lo zdani" first. Do we say:
>
> (b) mi stali lo zdani .i se ba'i bo mi klama lo zarci
>
> (b1) mi stali lo zdani
> (b2) mi klama lo zarci (?!)
> (b3) lo nu mi klama lo zarci cu se basti (lo nu mi stali lo zdani)
>
> or do we say:
>
> (c) mi stali lo zdani .i se ba'i bo mi na klama lo zarci
>
> (c1) mi stali lo zdani
> (c2) mi na klama lo zarci
> (c3) lo nu mi na klama lo zarci cu se basti (lo nu mi stali lo zdani) (?!)
>
> Neither (b) nor (c) seems to work well with (1), (2), (3), so we either
> have to discard (2), amend (3) somehow.

Meanwhile with (1)+(2)+(3'), this definition of {ba'i} gives an
immediate contradiction between (1) and (3').

And here we have no escape because the role of lo nu mi na klama lo zarci has to be basti2.

> or dictate that tags like "se ba'i" don't make sense as bridi
> connectives. CLL says that tags like "bau" don't make much sense,
> which is reasonable because a bridi doesn't really describe
> a language, but if "se ba'i" were not to make sense it would be for
> a very different reason.

Yes. Another option, if we do want to keep (1), which given the surface
form would I agree be nice, would be to take it as part of the semantics
of non-tense tags that {[tag] [[sumti]] broda} implies {broda}. That
would rule out this use of {ba'i} entirely.

Or we could still have {[tag] [[sumti]] broda} implying {broda}, but then (3') not always equivalent to (3) and it would be plain (3), or possibly

(3'') lo nu broda cu xo'i [tag] do'e lo nu brode (for tense tags)
lo nu brode cu xo'i [tag] do'e lo nu broda (for non-tense tags)

that expands the tag connective. There doesn't seem to be a strong reason to disallow (a).

> In any case, all of that applies to ".i [tag] bo". ".i [jek tag] bo" is
> related, but has to be analyzed separately.
>
> "broda .i brode" makes two independent claims.
> "broda .i [jek] brode" makes one claim.
> "brode .i [tag] bo brode" makes three independent claims according to CLL
> (I would prefer it rather made two.)
>
> How many independent claims does "broda .i [jek tag] bo brode" make, and
> what is it or what are they?
>
> I'd like it to make just one claim, the jek-logical connection between (1)
> and (3) above. CLL seems to say that it makes two independent claims:
> "broda .i [jek] brode" and (3). But this doesn't work very well when "jek"
> is anything other than "je".

Evidence in CLL is very slim for non-je tensed logical connections. The
only example I know of is the rather strained {mi bevri le dakli .ebake
le gerku .acabo le mlatu} (10:17). I think what's there reads most
naturally as being a case of "TT-skimming" as described earlier, but
I don't think it's enough to force this interpretation.

I don't remember the details of TT-skimming. Was it something like that the tag only kicks in when both connectands are true, and is otherwise ignored?

> > > > > ca ro nu mi xagji kei mi klama lo zarci .e ba bo lo zdani
> > > > > > -> ca ro nu mi xagji kei da poi nu mi klama lo zarci zo'u ge da
> > > > > > fasnu gi ba da mi klama lo zdani
> >
> ca ro nu mi xagji kei ko'a fasnu
> .i ko'a nu ge ko'e fasnu gi ko'i fasnu ba ko'e
> .i ko'e nu mi klama lo zarci
> .i ko'i nu mi klama lo zdani
>
> Every time I'm hungry, X happens.
> X is Y happening and Z happening after Y
> Y is my going to the market
> Z is my going home

I don't see that this helps at all. Now I have to ask: if ko'e is
a constant (kind of an) event, what does {ko'i fasnu ba ko'e} mean?

Oh, I see I didn't expand it enough. OK:

ca ro nu mi xagji kei ko'a fasnu

,i ko'a nu ge ko'e fasnu gi ko'i fasnu

;i ko'i nu ko'o balvi ko'e

.i ko'e nu mi klama lo zarci

.i ko'o nu mi klama lo zdani

mu'o mi'e xorxes