Re: [lojban] Re: what {me lo broda ku} means and why not allow {selbri NOI}

On 4 April 2013 12:24, la gleki <gleki.is.my.name@gmail.com> wrote:

However, "You sing, which I didn't know" just implies that {lo nu ...} in the beginning was dropped. But NOI states that there is a terminator of that {lo nu}.

The can only allow dropping articles when it doesn't lead to ambiguity like
{broda cu brode} definitely means {lo broda ku brode} and
{do sanga noi mi na pu djuno} means {co'e lo du'u do sanga ku noi mi na pu djuno ke'a}.

Or we can resort to the panoply of other methods that exist to add on bridi-related comments in afterthought, such as sei or, simply, another sentence.

e.g. .i do sanga sei mi na djuno

e.g. .i do sanga .i la'e di'u mi na se djuno

Adding NOI to selbri breaks much existing usage (not that I'm opposed to breaking usage, but some are) and it has a handful of distinct possible interpretations.

1) Reusing the selbri and making another sentence out of that selbri.

To realize this interpretation, we'd need to either ratify no'oi, po'oi, and me'au or introduce a GOhA that is a sort of selbri-ke'a.

Let go'ai be the selbri-ke'a:

e.g. .i do mi cinba noi lo mi bruna lo by. pampe'o cu go'ai

"You kiss me, which my brother does to his girlfriend."

With latro'a's and my experimental cmavo, that becomes:

.i do mi cinba no'oi lo mi bruna lo by. pampe'o cu me'au ke'a

"You kiss me, which my brother does to his girlfriend."

Using ke'a to represent the infinitive of the bridi (i.e. the pure selbri) make it possible to use it as an argument of another selbri expecting an infinitive, such as zdile or pluka:

e.g. .i do mi cinba no'oi ke'a pluka mi

"You kiss me, which pleases me."

This makes it much the same as a sei construct.

Let ke'a be bound the vague abstraction of the containing bridi:

e.g. .i do mi cinba sei ke'a mi pluka

"You kiss me -- that pleases me."

The annoying thing about sei is that it only supports a degenerate bridi-tail, and all the formal sumti must thus appear in the bridi head. Using no'oi/po'oi doesn't have this restriction. On the other hand, sei can appear anywhere, whereas no'oi/po'oi must appear after the selbri. Both have their pros and cons.

Finally, po'oi can be used to restrict bu'a-series selbri-variables and {mo}:

e.g. .i do mo po'oi na'e fadni

"What're you doing that isn't typical?"

e.g. .i lo mi bruna lo mi mensi cu bu'a po'oi mi ji'a me'au ke'a by .e my

"My brother and sister are related in some way whereby I too am related to both of them."

2) Simply create an ordinary sumti relative clause with ke'a bound to "lo su'u <the bridi>".

In my honest opinion, having selbri-NOI have this interpretation is counterintuitive. The problem is that we're considering too-simple cases so we don't realize how silly selbri-noi having this feature would be. It would be more useful to allow {vau NOI} to have this feature.

e.g. .i do tcidu lo se cukta vau noi ke'a se mukti lo nu tolzdi

"You read a book, which is motivated by your being bored."

(This is mostly redundant to fi'o and sei constructs, so it's pretty useless, really, and it breaks some usage.)

If we were to let selbri-NOI have this feature, all the sumti would be required to be in the bridi-head. All this really does is make a handful of naive natlang->lojban glosses correct.

3) A "freer" version of gi'e.

"This is the house that I used to live in." "This is the dog that I take care of." "This is my friend whom I've known for a long time." What do these all have in common? Nouns. This interpretation has ke'a bind to the x1 of the main bridi, which is basically gi'e's domain. Again, this basically just makes a bunch of naive glosses correct. The only real advantage is that the ke'a can be nested deeply. Really though, this can be achieved with me'au lo ka + a deeply nested ce'u.

e.g. .i ta mi pendo poi ze'u ba'o slabu

"That's my friend whom I've known for a long time."

e.g. .i ta zdani poi mi pu zu citka lo cirla ne'i ke'a

"That's the house in which I ate cheese long ago."

Those two are easily redundant to the following that use gi'e instead:

.i ta pendo mi gi'e ze'u ba'e slabu

"That's my friend and I've known them for a long time."

.i ta zdani gi'e pu zu jai ne'i te citka lo cirla mi

"That's a house and the place in which I ate cheese long ago."

(The second was reasonably more complex.)

For more complex cases still, where the ke'a is deeply nested, using gi'e becomes impossible, lest jai be thoroughly (ab)used.

.i li vo li re su'i re du noi mi vedli lo li'i mi cilre lo du'u li bi vu'u vo lo'o ji'a ke'a du kei ca lo nu mi jai nanca ke'a

[I can't produce a reasonable English translation for a reason I give a bit later.]

Using jai and gi'e, this becomes:

.i li vo du li re su'i re gi'e jai se vedli mi fai lo ka ko'a ce'ai mi cilre lo du'u li bi vu'u vo lo'o ji'a ko'a du kei ca lo nu mi jai nanca ko'a

Using me'au + lo ka, it becomes:

.i li vo du li re su'i re gi'e me'au lo ka ko'a ce'ai mi vedli lo li'i mi cilre lo du'u li bi vu'u vo lo'o ji'a ko'a du kei ca lo nu mi jai nanca ko'a

Regardless of these transformations, I have to admit that using selbri-NOI like this is *weird* in cases where the selbri isn't a noun-type selbri (having a definition of {x1 is a <something> ...}).

In sum, interpretation one has already been realized with experimental cmavo by latro'a and me, interpretation two is redundant to sei- and fi'o-clauses as well as TAG sentence connectives, and interpretation three is just plain silly, and typically redundant to clever gi'e constructs. I thus strongly disapprove of extending the grammar of NOI to handle any of these interpretations as they simply aren't useful enough. All they really do is make a certain number of naive assumptions about relative clauses true. The most common type of these naive glosses are of the third interpretation and are simply malgli, in my opinion. The speaker simply has to get used to using gi'e or making more than one sentence when nesting becomes a problem.

.i mi'e la tsani mu'o