Re: [lojban] Re: tersmu 0.2

On Mon, Nov 17, 2014 at 11:50 PM, Martin Bays <mbays@sdf.org> wrote:

Aside:
The kind "going shopping" isn't a going shopping, surely?

In a metalinguistic universe of discourse that contains both a kind and the instances of the kind (i.e. not the universe of discourse of our speaker), then I would say that the kind is not one of its instances, no. But if the question is whether the referent of "lo nu klama lo zarci" satisfies the predicate "nu klama lo zarci", then surely the answer must be yes.

That there's no easy way in lojban to differentiate between going
shopping and goings shopping is a real problem, I feel.

I think the distinction is necessarily metalinguistic, so no linguistic differentiation would really help. There's no problem having predicates that mean "kind", "instance", their relation, relation between them and a property, and so on, but these predicates would only be mostly useful for linguistic analysis, i.e. for metalinguistic discourse, not particularly useful to say something directly, because as soon as you introduce them into the discourse, there will be other metalinguistic kinds/instances that apply to these linguistic ones.

Is claiming that an event-kind occurs at a particular time not
equivalent to claiming that an event instance of the kind occurs at that
time?

Your second claim may be ambiguous. Do you mean claiming that there's an instance that occurs, or do you mean there being an instance of which one claims that it occurs? The universe of discourse that contains instances is richer, which means it allows you to make more fine distinctions, but it also forces you to make these distinctions. The coarser grained universe of discorse with kinds doesn't allow/force these distinctions.

Could you clarify something else for me about the kinds translation:
if ko'a and ko'i are event-kinds, does {ca ko'u ko'a balvi ko'i} imply
{ca ko'u ko'a .e ko'i fasnu}?

Probably. That would depend on the details of the semantics of "balvi", which being a time relation is of course tricky to tense. But to make sense of it, ko'u would probably have to be long enough for ko'i to occur and then ko'a to occur at ko'u. Say something like "this spring the swallows came after school started".

ko'u = this spring

ko'a = the swallows come

ko'i = school starts

If so, why also explicitly declare ko'a to fasnu? But if not, I think
the kinds translation's meaning might be quite different from that
I attribute to the original sentence.

I'm thinking the first connectand is independent of the tag-carrying connectand. The tag-carrying connectand might also imply the first connectand, but that's not necessarily the case for all tags.

Also if so, actually - I wouldn't understand {ca ko'u broda gi'e ba bo
brode} to imply that {ca ko'u brode}, but rather that {ca ko'u ba
brode}.

Yes, but ba what? "ba ko'u" or "ba lo nu broda"?

> Another point against the quantifier reading is that if you change ".e ba
> bo" to "na .a ba bo", then we would seem to need to change "su'o" to "ro"
> in the quantified expansion: "ca ro nu mi xagji kei mi klama lo zarci na .a
> ba bo lo zdani" -> "ca ro nu mi xagji kei ro da poi nu mi klama lo zarci
> zo'u ga nai da fasnu gi ba da mi klama lo zdani" or something like that,
> whereas with the kind-reading you use the same expansion "ca ro nu mi xagji
> kei ko'a goi lo nu mi klama lo zarci zo'u ga nai ko'a fasnu gi ba ko'a mi
> klama lo zdani".

Interesting. So this is like a literal translation of "if I go to the
market, after going to the market I go home", analysing "going to the
market" as a reference to a kind rather than as a reference to a witness
going in the antecedent?

It's what one would have to say in a more fine grained universe containing instances, but it's not a direct translation in the sense that a direct translation I think would have to be based on the same universe of discourse.

So you now prefer this approach to your previous suggestion of using
{nu na broda} when analysing {broda .i [jek] [tag] bo brode}?

I'm not very committed to anything at this point. I'm only saying that given the choice between the "lo nu"-version and the "su'o nu"-version I would go with the "lo nu"-version.

broda .i [jek] [tag] bo brode -> broda .i [jek] [tag] lo nu (na) broda cu brode

sounds reasonable to me for tenses. The "(na)" doesn't really kick in for "je" or for "naja", so these two examples don't cover all cases. For non-tenses I want to read it as:

broda .i [jek] [tag] bo brode -> broda .i [jek] broda [tag] lo nu brode

but that may be even more problematic for "jonai" type connectives.

So I guess this kind of reasoning would have
{broda .i je nai ca bo brodo} mean something like "broda occurs, but
broda never occurs simultaneously with brodo"? Whereas I would have
expected it to mean something more like "broda occurs some time when
brodo doesn't".

I hope you are not thinking that by rejecting an equivalence of "lo" with "su'o" I'm somehow embracing an equivalence with "ro". "lo" is neither "su'o" nor "ro" (or it is both), because "lo" sees all instances as one, so there's no distinction to be made between "su'o" and "ro" in this coarse grained view. I think it just means "broda occurs, but not simultaneously with brodo".

mu'o mi'e xorxes