Re: [lojban] BPFK Section: Non-logical Connectives

On Tue, Feb 17, 2015 at 3:24 AM, Gleki Arxokuna <gleki.is.my.name@gmail.com> wrote:

2015-02-17 2:31 GMT+03:00 Jorge Llambías <jjllambias@gmail.com>:

I think when discussing "joi" it's useful to bear in mind it has (at least) two different definitions (as does "loi"):

(joi1) ko'a joi ko'e = lo gunma be ko'a jo'u ko'e
(joi2) ko'a joi ko'e = ko'a jo'u ko'e + indication that when this sumti is used as the argument of a predicate, the predicate should not distribute over the referents of the sumti.

So {joi2} is for "plural" non-distributive sumti variable type, {jo'u} is for distributive and {je/.e} is for vague?

No, "jo'u" is not (necessarily) distributive. And ".e" is shortcut for a sentence connective, so of course it has to be distributive.

But distributiviy is not really a property of a sumti, it's a property of a tersu'i. If "ko'a broda" entails "ro ko'a broda", then broda1 is distributive. This doesn't mean that "ko'a" is distributive, it means broda1 is distributive.

Asking whether "ko'a jo'u ko'e" is distributive or not is like asking whether "ko'a" is distributive. The question doesn't really make sense. "ko'a jo'u ko'e" is just like "ko'a" or like "ko'e", it's something with referents. What any predicate says about those referents is not up to "jo'u".

"ko'a .e ko'e broda" is completely different, because it's a shortcut notation for ""ko'a broda .ije ko'e broda". "ko'a .e ko'e" doesn't mean much by itself since ".e", like all logical connectives, needs a selbri to work with.

That's why "joi2" is weird, "ko'a joi2 ko'e" creates a new sumti with all the referents of "ko'a" and "ko'e", but then it also wants to tell us how these referents will be treated by anything predicated about them. If you say "ko'i goi ko'a jo'u ko'e" and "ko'o goi ko'a joi ko'e", I'm not sure if "ko'i" and "ko'o" are any different. They have the exact same referents, but does "ko'o" carry also the information that it can only be used with a non-distributive predicate?

May be instead use {joi} for {joi1} only and use {ce} for {joi2}?

I think both "joi1" and "joi2" are pretty useless, so I don't have an opinion on that. If it was up to me I would forget about both those uses and reserve "joi" for "jo'u", which is the more basic and needed one.

I don't think usage can help here. Most of it would probably be wrong usage.

Yes, usage is probably all over the place. "Masses" have also been used to make other distinctions besides emergent entity (joi1) and non-distributivity (joi2), so this is only part of the mess, but this affects "loi" more than "joi", since the "loi djacu" sense doesn't really have a proper counterpart in "joi".

However, I think that BPFK should strictly specify the meaning of {joi} or even better to specify to how to express plural type ("set") like casnu1, simxu1, how to express masses (if needed), and distributivity.

All sumti are plural by default. If singular reference is significant it has to be specified as a special case, for example attaching "noi pa mei" to the sumti, or with an explicit inner "pa" in a description. In general Lojban doesn't mark number. casnu1 and simxu1 are tersu'i, not sumti, so it's up to the definition of the predicate to specify what it says about its arguments. I think for example that casnu1 can be filled with a single entity consisting of people ("lo (pa) kamni cu casnu") or about many entities each of whom is a person ("lo (so'i) prenu cu casnu") and both are acceptable. "Mass" means so many different things in Lojban-speak that before deciding how to express them we need to specify what we are talking about. One of them is "lo gunma be", "a group consisting of". Distributivity is done with "ro": "ro ko'a broda" says that each of the referents of "ko'a" satisfies "broda" by itself. "ko'a .e ko'e broda" is short for "ko'a broda .i je ko'e broda".

With (joi1) "ko'a joi ko'e" refers to a single entity that has two constituents. In this case there's no point in talking about distributivity since there's only one thing involved, so nothing to distribute.

I think masses ({lo gunma}) and non-distributive sumti variable type are different things and should not be reconciled in one connective.

Indeed. In fact neither of them should be done with connectives.

mu'o mi'e xorxes