On Saturday, October 6, 2012 8:21:14 PM UTC+4, ts= ani wrote:
Why is having a {ce'u= } in djica *so* important? We've gotten along *just fine* using {zo'= ;e} there, even though it refers to djica1 in a lot of cases.=C2=A0

zo'e doesn't refer to first places of nes= ting bridi.=C2=A0
It's equally unnatural to say {mi djica lo = nu mi sipna} by repeating {mi} two times. If onlu we had some analogue to {= ce'u} or {ri} referring to the previous sumti even if it's {mi}.

Indeed, {zo'e} can refe= r to *anything* (with some exceptions). If it's clear that it's {mi= }, then it's {mi}. Remember, *this has never been a problem before*. To= really solve this problem, we could redefine djica to use only ka+ce'u= , but that would violently break usage, and leaving us with {pacna} for old= -djica.

.i mi'e la tsani mu'o

=C2=A0

There are some selbri, say kakne, whe= re {zo'e} in the abstraction kakne2 breaks the meaning in some way, suc= h as {mi kakne lo nu do citka lo plise}. IMO selbri that get *weird* when t= here's no ce'u should have {ka}.=C2=A0

Furthermore, nu+ce'u is strange to me, because {ce&= #39;u} marks an argument slot in a function, but {nu}, et al., abstractions= are never function-abstractions. As I detail in my analysis of abstractors= , there're two classes of abstractors, namely function-abstractors and = non-function-abstractors. It gets a bit fuzzy eventually, because some abst= ractors are function when they contain {ce'u}, but can equally not cont= ain {ce'u}. {ni} is an example of such an abstractor:

{.i mi zmadu do lo ni xendo} vs {.i lo ni mi prami do c= u zmadu lo ni do nelci lo mlatu kei du bu} (du bu is the identity function)= .

The advantage of not allow {ce'u} inside {nu= } is that {nu} are as a result completely self-contained entities. {lo nu m= i do cinba} forms one single object that doesn't depend on the containi= ng bridi.=C2=A0

As for {su'u}, it turns out that it's simply a = vague abstractor, and that's it. It's a stand-in for any regular ab= stractor, and its type is completely context-dependent. Although this has n= ever really happened as far as I know, it would, however, be possible to cr= eate new types of abstractors by means of su'u2.

.i mi'e la tsani mu'o
On 5 October = 2012 21:47, la gleki <gleki.is...@gmail.com>= wrote:

On Friday, October 5, 2012 7:36:18 PM UTC+4, selpa'i wrote: Am 05.10.2012 16:41, schrieb la gleki:
>
>
> On Thursday, October 4, 2012 11:47:16 PM UTC+4, selpa'i wrote:
>
> =C2=A0 =C2=A0 Am 04.10.2012 17:02, schrieb la gleki:>
> =C2=A0 =C2=A0 =C2=A0 >
> =C2=A0 =C2=A0 =C2=A0 >> =C2=A0 =C2=A0 1."I want to eat = an apple".
> =C2=A0 =C2=A0 =C2=A0 >>
> =C2=A0 =C2=A0 =C2=A0 >> =C2=A0 =C2=A0 The normal way uses an= infinitive compound:
> =C2=A0 =C2=A0 =C2=A0 >>
> =C2=A0 =C2=A0 =C2=A0 >> =C2=A0 =C2=A0 ^:i \ji /daw crw \xo p= lyw
> =C2=A0 =C2=A0 =C2=A0 >>
> =C2=A0 =C2=A0 =C2=A0 >> =C2=A0 =C2=A0 But you can also use a= n explicit infinitive:
> =C2=A0 =C2=A0 =C2=A0 >>
> =C2=A0 =C2=A0 =C2=A0 >> =C2=A0 =C2=A0 ^:i \ji /daw \vo crw \= xo plyw
> =C2=A0 =C2=A0 =C2=A0 >>
> =C2=A0 =C2=A0 =C2=A0 >> =C2=A0 =C2=A0 2."I want you to = eat an apple".
> =C2=A0 =C2=A0 =C2=A0 >>
> =C2=A0 =C2=A0 =C2=A0 >> =C2=A0 =C2=A0 ^:i \ji /gu pli \ju ^v= o crw \xo plyw
> =C2=A0 =C2=A0 =C2=A0 >>
> =C2=A0 =C2=A0 =C2=A0 >> =C2=A0 =C2=A0 or
> =C2=A0 =C2=A0 =C2=A0 >>
> =C2=A0 =C2=A0 =C2=A0 >> =C2=A0 =C2=A0 ^:i \ji ^ju /gu pli \c= rw \xo plyw
> =C2=A0 =C2=A0 =C2=A0 >>
> =C2=A0 =C2=A0 =C2=A0 >
> =C2=A0 =C2=A0 =C2=A0 > But I have a clear feeling that in both = sentences the same semantic
> =C2=A0 =C2=A0 =C2=A0 > prime can be used. And this prime descri= bes "desire".
> =C2=A0 =C2=A0 =C2=A0 > Lojban can replace {ce'u} with anyth= ing. Natlangs can do the same.
> =C2=A0 =C2=A0 =C2=A0 > gua\spi can't. {to zoi gy. I don'= ;t want to criticize gua\spi
> =C2=A0 =C2=A0 anymore.
> =C2=A0 =C2=A0 =C2=A0 > gy. toi}
>
> =C2=A0 =C2=A0 You cannot replace ce'u at all or else it's = gone and it's not a
> =C2=A0 =C2=A0 ka-abstraction anymore (or not a well-formed one).
>
>
> True. Still the same brivla can be used. Unlike gua\spi.

No, it can not. If you "ce'u-ize" the gimste, for instanc= e by saying=20
that djica2 ba a ka (which is a bad example, but it illustrates the=20
point), then you will not be able to use it for "I want you to bro= da",=20
because that's a different predicate that doesn't involve yours= elf in=20
the abstraction.

Well, I= 9;m not sure if we should ce'u-ize gismu with {nu} =C2=A0abstractions.<= /div>
But if do this for {djica} then it would be
1.{mi djica= lo nu ce'u citka}

2.{mi djica lo nu do citka}

That's all I = want. But gua\spi's /daw/ can't do that.

W= e could also say {mi djica lo nu ce'u citka i do na go'i} =3D "= ;I want to eat but you don't" (if {go'i} is able to update the= value of {ce'u} in the previous sentence, of course)

This is a *str= enght* of gua\spi; its predicates are=20
semantically much clearer.

Well, well, I don't want som= eone to stop learning gua\spi =C2=A0because "gua\spi is a crap. I know= , la gleki told me". :)

>
> =C2=A0 =C2=A0 What natlangs can and
> =C2=A0 =C2=A0 can't do has little relevance when discussing Lo= jbanic topics such as
> =C2=A0 =C2=A0 ka-abstractions.
>
>
> Then gua\spi has little relevance too.

What? Gua\spi is not a natlang, and you brought up Gua\spi in the first= =20
place. Gua\spi's entire gimste is ce'u-ized, that's what it= looks like.

>
>
>
> =C2=A0 =C2=A0 In Lojban, djica2 is a nu, not a ka. You could say t= hat djica should be
> =C2=A0 =C2=A0 polymorphic and allow both nu and ka, but I don'= t think that's what
> =C2=A0 =C2=A0 you're saying, is it? (I don't know *what* y= ou are saying).
>
>
> My only complaint =C2=A0that we have a nice shortcut of saying {du= 'u ce'u}
> but we don't have one for {nu ce'u}.

But ka is not a shortcut for du'u ce'u... ka is what you get if= you have=20
a du'u abstraction and add a ce'u to it.

>
> =C2=A0 =C2=A0 Why is it a nu? Because you can djica things that do= n't involve
> =C2=A0 =C2=A0 yourself. (Gua\spi's _daw_ is restricted to desi= ring to do or be
> =C2=A0 =C2=A0 something, hence it's always like a Lojban ka. A= nd that's why the
> =C2=A0 =C2=A0 second
> =C2=A0 =C2=A0 example uses a different predicate.)
>
> =C2=A0 =C2=A0 Again, what is the difference between the Lojban and= the gua\spi
> =C2=A0 =C2=A0 sentence?
>
>
> Hopefully no semantic difference. Looks like Lojban just gives mor= e
> freedom in recombining the same words without drawing in extra pre= dicates.

Okay, but that wasn't even your original point. And as I tried to= =20
explain above, you get seperate predicates if you ce'u-ize the gims= te.=20
One will be

x1 wants to be/do x2 (ka)

the other will be

x1 wants/wishes/desires that x2 (nu) happen

Is that what you want or not?

>
>
> Let's stop arguing and let's ce'u-ize gimste :).
>
> How the new ce'u-ized definitions of gismu should look like in= ur opinion?

Just look at gua\spi's gimste. It did everything right in that rega= rd,=20
but you have to remember that gua\spi is not Lojban, and not everything= =20
can be copied 1:1.

mu'o mi'e la selpa'i

--=20
pilno zo le xu .i lo dei bangu cu se cmene zo lojbo .e nai zo lejbo

do=E1=BB=8B m=C3=A8lbi mlen=C3=AC'u
=C2=A0 =C2=A0 .i do c=C3=A0tlu ki'u
ma fe la x=C3=A0mpre =C5=ADu
=C2=A0 =C2=A0 .i do t=C3=ACnsa c=C3=A0rmi
gi'e s=C3=ACrji se t=C3=A0rmi
=C2=A0 =C2=A0 .i ta=E1=BB=8B bo pu c=C3=ACtka lo gr=C3=A0na ku

--
You received this message because you are subscribed to the Google Groups &= quot;lojban" group.
To view this discussion on the web visit https://groups.google.com= /d/msg/lojban/-/WyFuvnW2QSEJ.