Discourse-sensitive {zo'e} and {zi'o}

2/14/2014

This is another idea I had a few nights ago. According to the CLL, all ommited sumti places are filled with {zo'e} (leaving aside {ka}-abstractions).

A number of people like the idea of treating elided places as being {zo'e} or {zi'o} (the earliest proposal I can remember being by And Rosta). I like the idea, but I would go a step further, or rather, I'd like to make things a bit more dependent on the current discourse. What's the problem with having all omitted places be {zo'e}? Well, it fills in values in places we might not even want to fill. This is especially true for bloated gismu, but in general it can be "unclean".

I propose the following procedure for determining what fills an omitted place:

1. At the start of discourse, all places are filled with {zi'o}. {mi citka} as the first sentence of discourse (or a conversation) means {mi citka zi'o}, "I am eating." The difference is that {mi citka zo'e} could mean "I am eating cereal", but {mi citka zi'o} is true no matter what I eat. In this regard, {zi'o} is like a {su'o da} without scope. In another way it's like {ro da}: {ro da zo'u: ga nai mi citka da gi mi citka zi'o}. The important point is that {zi'o} is more general than {zo'e}, and is therefore the better default value for neutral contexts.
2. Once a place of a predicate has been filled, that place will afterwards be filled with {zo'e}. As the first {citka}-sentence of a given discourse, {mi na citka} means "I'm not eating anything." ({mi na citka zi'o}). Once someone says {mi citka lo nanba}, {mi na citka} will be {mi na citka zo'e}, and can (but needn't) be interpreted as "I'm not eating bread".
3. Under certain circumstances, a place can be reset to {zi'o}. Presumably, anytime a place gets filled with a non-{zo'e} non-{zi'o} value, it reverts back to {zi'o} after the sentence ends. I.e. {da}, {ce'u} and possibly {ke'a} reset their places to {zi'o} afterwards.

This flexible procedure makes {zi'o} and {zo'e} behave more intelligently, and allows us to be general more easily when it is more likely to be desired, while making us be explicit a little more often when we actually mean specific things. This will be facilitated by the new anaphora system which we will surely have to adopt sooner or later. (I'm sort of cooking one up at the moment)

6 Comments

gleki

2/14/2014 07:06:50 pm

y y y
i cinri
i ei mi piksku no da
i si'au lo banli
i li'a cylyly zo'u zo zo'e cu jai se ciksi tai lo nalxamgu i mu'a xu lu lo co'e li'u smuni frica lu zo'e li'u
i ku'i lo do se stidi zo'u mi na birti zo'e gi'e ku'i mi ganse lo nabmi
i a'o banli
i sa'u ei ma'a co'a pilno

arpis

5/17/2014 03:05:29 am

I am very much against this (though, by the fact that i am not fluent enough in lojban to explain it, you can guess how much this should matter).

One of my favorite things about lojban is that every selbri has places whether it wants them or not, and this contributes to its meaning. {citka} always has a thing being eaten; it's just not the same concept if there is "eating" going on without food being involved.

One of the first places I ran into this distinction meaningfully was {lo se cevni}. {lo cevni}, in lojban, does not exist without worshipers/believers, or at least is not a {cevni}; that is, in lojban, a god without worshipers is no god at all, by definition (using the phrase literally). {lo cevni be zi'o} is a different kind of entity: a god that can exist whether worshiped or not. This difference can have interesting social, philosophical, and theological implications, and I like it.

gleki link

5/17/2014 03:43:34 am

This is against CLL. {zi'o} creates new predicates whereas acc. to this proposal it serves the purpose of a discourse places manager.

May be a separate service cmavo should be used instead.

arpis

5/17/2014 10:57:54 pm

I don't follow; are you claiming that my description is against the CLL (as I would infer from your first sentence) or that the proposal is against the CLL (as I would infer from the second sentence)?

I personally don't see what problem this suggestion is trying to solve. The three that come to mind are: a) too-many placed selbri; b) lying by intentionally misinterpreting zo'e; and c) the degree of dependence on context not evident in the text.

I don't think c) is a problem at all; b) isn't a problem in practice, but can be solved formally by saying that neither the speaker nor the listener determines the referent for {zo'e}, but The Context, and it is possible for a speaker to misspeak (use {zo'e} to mean something other than what The Context determines) or the listener to misunderstand (interpret zo'e as something other than what The Context determines), or even (in fact, often, except in cases of correct understanding or willful deception) both; and I don't think a) is a big problem in practice, but I have a pet solution in case I ever construct a {logji bangu} that involves shifting responsibilities from many places onto BAI and fi'o.

ovaloid

7/13/2016 06:56:16 am

What do you like about it?
It would be bias against religious believers in what is supposed to be a culturally neutral language.

And Rosta

9/5/2014 09:50:15 am

I think the formal rule should be "null = zi'o or zo'e", but 1--3 would be an excellent discourse convention to follow. To reify 1--3 into a formal rule would raise a load of edge cases, mostly involving the issue of what criteria determine whether two predicates count as the same.

Discourse-sensitive {zo'e} and {zi'o}

Leave a Reply.