What place of nesting bridi {ce'u} refers to?

[la gleki]

http://dag.github.com/cll/7/11/

11.1)  le ka ce'u gleki
       the property-of (X being-happy)
       the property of being happy
       happiness

and

11.2)  le ka gleki ce'u
       the property-of (being-happy about-X)
       the property of being that which someone is happy about

These examples are pretty clear. No ambiguity. Now let's open Chapter 11.4

http://dag.github.com/cll/11/4/

4.9)   la djan. cu zmadu la djordj. le ka mi prami ce'u
       John exceeds George in-the property-of (I love X).

This is something very strange.

Let's imagine that I'm a boy and I meet a girl. I tell her

{do melbi mi lo ka ce'u clani}

Does {ce'u} refer to {do} or {mi}?

There are at least two ways this sentence can be understood.

1. I like tall girls. She is tall. I mean "I like that you are tall". i.e. {ce'u} refers to {lo melbi}.

2. I'm a tall boy. She is shortish. I mean "I like that you are short (given that I'm tall)". i.e. {ce'u} refers to {lo se melbi}.

[selpa'i]

Am 30.09.2012 09:56, schrieb la gleki:

These examples are pretty clear. No ambiguity. Now let's open Chapter 11.4

http://dag.github.com/cll/11/4/

4.9)   la djan. cu zmadu la djordj. le ka mi prami ce'u
       John exceeds George in-the property-of (I love X).

This is something very strange.

The ce'u in zmadu3 refers to both zmadu1 and zmadu2, individually.

Let's imagine that I'm a boy and I meet a girl. I tell her

{do melbi mi lo ka ce'u clani}

Does {ce'u} refer to {do} or {mi}?

It refers to melbi1. melbi2 is just a subjective observer. The definition makes that somewhat clear "x1 is beautiful in aspect x3".

mu'o mi'e la selpa'i

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

doị mèlbi mlenì'u
   .i do càtlu ki'u
ma fe la xàmpre ŭu
   .i do tìnsa càrmi
gi'e sìrji se tàrmi
   .i taị bo pu cìtka lo gràna ku

[Jacob Errington]

It'd be nice if the gimste mentioned to which higher-level sumti places the ce'u referred to, much like guaspi's gimste, but that isn't the case.

On 30 September 2012 00:56, la gleki <gleki.is...@gmail.com> wrote:

http://dag.github.com/cll/7/11/

11.1)  le ka ce'u gleki
       the property-of (X being-happy)
       the property of being happy
       happiness

and

11.2)  le ka gleki ce'u
       the property-of (being-happy about-X)
       the property of being that which someone is happy about

These examples are pretty clear. No ambiguity. Now let's open Chapter 11.4

http://dag.github.com/cll/11/4/

4.9)   la djan. cu zmadu la djordj. le ka mi prami ce'u
       John exceeds George in-the property-of (I love X).

This is something very strange.

Let's imagine that I'm a boy and I meet a girl. I tell her

{do melbi mi lo ka ce'u clani}

 

Does {ce'u} refer to {do} or {mi}?

It definitely refers to melbi1. If x3 is the aspect in which x1 is beautiful, it's pretty clear that it's the x1 that is referred to by the argument place in clani.

.i mi'e la tsani mu'o

[la gleki]

How do you know that? Just a common sense? If so it's a real problem. Someone is developing lojban in terms of NLP but your answers clearly show that currently it's impossible.

Don't you think that we should mark all gismu that have such "problems"?

[John E Clifford]

Neuro-Linguistic Programming?  Natural Language Processing? Neither seems to fit with the present situation being impossible with respect to it.

---

From: la gleki <gleki.is...@gmail.com>
To: loj...@googlegroups.com
Sent: Monday, October 1, 2012 10:24 AM
Subject: Re: [lojban] What place of nesting bridi {ce'u} refers to?

--
You received this message because you are subscribed to the Google Groups "lojban" group.
To view this discussion on the web visit https://groups.google.com/d/msg/lojban/-/Vl8RFZbT5ZsJ.
To post to this group, send email to loj...@googlegroups.com.
To unsubscribe from this group, send email to lojban+un...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/lojban?hl=en.

[la gleki]

On Monday, October 1, 2012 7:33:20 PM UTC+4, clifford wrote:

Neuro-Linguistic Programming?  Natural Language Processing? Neither seems to fit with the present situation being impossible with respect to it.

u'i

i u'u

Natural Language Processing.

[Arnt Richard Johansen]

On Sun, Sep 30, 2012 at 12:56:59AM -0700, la gleki wrote:

> These examples are pretty clear. No ambiguity. Now let's open Chapter 11.4
> http://dag.github.com/cll/11/4/
>
> 4.9) la djan. cu zmadu la djordj. le ka mi prami ce'u
> John exceeds George in-the property-of (I love X).
>
> This is something very strange.
> Let's imagine that I'm a boy and I meet a girl. I tell her
>
> {do melbi mi lo ka ce'u clani}
>
> Does {ce'u} refer to {do} or {mi}?

Neither. {ce'u} does not refer to anything, and that's sort of the whole point. The clause refers to a “property” of being tall *in the abstract*, not that of someone in particular being tall.

* lo ka ce'u clani
being long

* lo ka clani ce'u
being a dimension of length

* lo ka clani fi ce'u
being a standard of length

* lo ka mi clani
my height

* lo ka do clani
your height

* lo ka lo penbi cu clani
a pen being long

* lo ka clani
being long (OR being a dimension of length OR being a standard of length)

> There are at least two ways this sentence can be understood.
> 1. I like tall girls. She is tall. I mean "I like that you are tall". i.e.
> {ce'u} refers to {lo melbi}.
> 2. I'm a tall boy. She is shortish. I mean "I like that you are short
> (given that I'm tall)". i.e. {ce'u} refers to {lo se melbi}.

The definition of {melbi} is:
x₁ is beautiful/pleasant to x₂ in aspect x₃ (ka) by aesthetic standard x₄.

Although English is not my first language, I'm pretty sure that when something is beautiful in some aspect, that aspect is an aspect of that which is beautiful, not of that which perceives the beauty. So interpretation 1 is the correct one.

--
Arnt Richard Johansen http://arj.nvg.org/
Let's have some real examples from a real, non-English language.

[selpa'i]

Am 01.10.2012 22:04, schrieb Arnt Richard Johansen:
> On Sun, Sep 30, 2012 at 12:56:59AM -0700, la gleki wrote:
>
>> These examples are pretty clear. No ambiguity. Now let's open Chapter 11.4
>> http://dag.github.com/cll/11/4/
>>
>> 4.9) la djan. cu zmadu la djordj. le ka mi prami ce'u
>> John exceeds George in-the property-of (I love X).
>>
>> This is something very strange.
>> Let's imagine that I'm a boy and I meet a girl. I tell her
>>
>> {do melbi mi lo ka ce'u clani}
>>
>> Does {ce'u} refer to {do} or {mi}?
> Neither. {ce'u} does not refer to anything, and that's sort of the whole point. The clause refers to a “property” of being tall *in the abstract*, not that of someone in particular being tall.

Yes, but the ce'u place gets filled later by one of the sumti in the
parent bridi, and that is what the question was about.

> * lo ka mi clani
> my height
>
> * lo ka do clani
> your height
>
> * lo ka lo penbi cu clani
> a pen being long

These are very non-standard, and probably many would call them
incorrect. The only way I can see to make these work is to say that "lo
ka mi clani" is to be interpreted as "lo ka mi no'u ce'u clani", or else
the "ce'u" would end up in a different sumti place and the meaning would
change drastically.

> * lo ka clani
> being long (OR being a dimension of length OR being a standard of length)

Right, but much more than with "ke'a", where this is handled much more
loosely, the convention is that "ce'u" fills the first empty slot.

[la gleki]

So if there are no {ce'u}  specified we must understand it as the first and only the first slot is filled with omitted {ce'u}. If two or more of the slots of the  nested brivla are filled we must specify all of them (like in {mi e do simxu lo ka ce'u ce'u prami}), right?

That would make sense however this is something that must be included into CLL 2.0.

IMO {ce'u}-izing gimste is also a must. 

btw, dont you think that we can use {ce'u} in {mi djica lo ka/nu *ce'u* citka lo plise} instead of {vo'a/mi}?

[Jacob Errington]

It's funny that you're thinking about this now. It's reminding me of myself when I "discovered" this phenomenon some months ago.

I believe that {ce'u no'u ko'a} is wrong. The bridi in which the property-abstraction appears provides the value for the ce'u-place. Personally filling a ce'u-place changes the abstraction type from ka to du'u.

 

> * lo ka clani
>    being long (OR being a dimension of length OR being a standard of length)

Right, but much more than with "ke'a", where this is handled much more
loosely, the convention is that "ce'u" fills the first empty slot.

So if there are no {ce'u}  specified we must understand it as the first and only the first slot is filled with omitted {ce'u}. If two or more of the slots of the  nested brivla are filled we must specify all of them (like in {mi e do simxu lo ka ce'u ce'u prami}), right?

That would make sense however this is something that must be included into CLL 2.0.

IMO {ce'u}-izing gimste is also a must. 

btw, dont you think that we can use {ce'u} in {mi djica lo ka/nu *ce'u* citka lo plise} instead of {vo'a/mi}?

Sure, but what of wanting things that don't involve djica1 at all?

I've thought long and hard about this, as evidenced by the abstractions paper that I wrote on the wiki. In general, when an abstraction place doesn't make sense if you exclude the broda1 from it, then it requires ka. 

*{.i mi kakne lo nu do citka lo plise} (example from an old mriste thread of mine)

is nonsense, which prompts me to believe that kakne2 is a ka or equivalent.

However {.i mi djica lo nu do klama lo zarci} makes perfect sense, even though the djica1 is not inside the djica2.

The gimste being more prescriptive about abstraction types would be great, but good luck finding enough people who support the cause of more rigid types. (AFAIK there are at most two or three, myself and probably you included.)

[selpa'i]

Am 02.10.2012 10:26, schrieb la gleki:

> * lo ka clani
>    being long (OR being a dimension of length OR being a standard of length)

Right, but much more than with "ke'a", where this is handled much more
loosely, the convention is that "ce'u" fills the first empty slot.

So if there are no {ce'u}  specified we must understand it as the first and only the first slot is filled with omitted {ce'u}. If two or more of the slots of the  nested brivla are filled we must specify all of them (like in {mi e do simxu lo ka ce'u ce'u prami}), right?

That is the current standard, yes. An omitted ce'u was originally intended to fill any place that made sense, just like ke'a. I think I prefer the current standard of ce'u filling the first empty slot.

That would make sense however this is something that must be included into CLL 2.0.

Depending on how much support this convention has, it will be.

IMO {ce'u}-izing gimste is also a must.

By ce'u-izing, do you mean making any abtraction that requires you to refer to an outside sumti a ka-abstraction? If so, tsani would certainly be very happy. I see two slight disadvantages to this approach and one advantage. I don't know which one overweighs.

The advantage is that a lot of "nonsense" sentences become easier to identify as being "nonsense", e.g. "mi kakne lo nu ta pelxu" wouldn't happen anymore. However, even such sentences can easily be made to make sense by saying that the kakne1 appears in some other place, probably do'e. Of course, this is an indicator that ce'u was the right thing to begin with, so this is not a strong counter-position.
Personally, the disadvantages I see are mainly these two:

1. ka-abstractions are not events, so they are much more awkward to use / incompatible with other places that don't accept ka. "lo se zukte be do mi pluka" becomes wrong when zukte2 is a ka.
2. Too many ce'u. What do you do with nested ka-abstractions and ce'u? It causes some additional work in some cases to indicate which of the ce'u one is currently using. Subscripting is not very elegant in my opinion. (This becomes worse when ce'u can also appear in a nu-abstraction, which is one reason I'm not sure about it). This wouldn't be an extremely common problem, but it's potentially there.

(Relatedly, I'm also not a huge fan of boolean ka, but that might not be of importance here.)

btw, dont you think that we can use {ce'u} in {mi djica lo ka/nu *ce'u* citka lo plise} instead of {vo'a/mi}?

tsani already answered this, but note that it's usually not accepted to use ce'u in a nu. Personally, I'm still undecided about ce'u in a nu, however. In any case, you'd usually either have "lo ka ce'u" or "lo nu [vo'a]".

[la gleki]

On Tuesday, October 2, 2012 3:25:47 PM UTC+4, tsani wrote:

It's funny that you're thinking about this now.

xajmi la'a ije ku'i nalgleki ju'o

And I only got convinced that

{ka=su'u ce'u}

Such understanding would give us a nice abbreviation, though.

By {ce'u}-izing gimste I mean that currently each gismu is described in jvs in

1. glossword

2. keywords

3. definition

4. notes

I propose that we add the fifth item

5. ce'u formula.

For zmadu it will be

3 (1,2)

which means that the abstraction of the first place has the first slot (i.e. ce'u) referring equally to the first and the second slots of the nesting brivla.

For kakne we have

2(1)

i.e. lo se kakne has ce'u inside that refers to lo kakne.

If {ka=su'u ce'u} then we'll have very nice and compact way of expressing such an important and useful concept of natlangs as infinitives.

I don't like how gua\spi works here cuz as i can see it demands using another gismu every time you change the value of {ce'u}.

However, many European Indoeuropean languages allow very intuitive ways of expressing {ce'u} and {vo'a}.

I want to eat an apple    = I want that I eat an apple [a really awkward lojban-style sentence]

{mi djica lo ka citka??)  = mi djica lo nu mi/vo'a citka lo plise

So {ka=su'u ce'u} can be of great use after we ce'u-ize gimste.

[selpa'i]

Am 03.10.2012 17:54, schrieb la gleki:

If�{ka=su'u ce'u} then we'll have very nice and compact way of expressing such an important and�useful concept of natlangs as infinitives.

I don't like how gua\spi works here cuz as i can see it demands using another gismu every time you change the value of {ce'u}.

Okay, what you want is *exactly* what gua\spi is doing, so I think you misunderstand something. I wouldn't add those "ce'u formulae" as a seperate field. Instead, I would add them right to the definition. Gua\spi does this:

zmadu : x1 exceeds x2 in property/quantity x3:1+2

I'm not sure what you mean by

I don't like how gua\spi works here cuz as i can see it demands using another gismu every time you change the value of {ce'u}.

Can you show me what you mean?

mu'o mi'e la selpa'i

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

doị m�lbi mlen�'u
   .i do c�tlu ki'u
ma fe la x�mpre ŭu
   .i do t�nsa c�rmi
gi'e s�rji se t�rmi
   .i taị bo pu c�tka lo gr�na ku

[.djo,is.]

On Wednesday, October 3, 2012 8:54:13 AM UTC-7, la gleki wrote:

I want to eat an apple    = I want that I eat an apple [a really awkward lojban-style sentence]

{mi djica lo ka citka??)  = mi djica lo nu mi/vo'a citka lo plise

People seem to use {nu} or {ka} a lot of times when {zu'o} would be more useful. At least my understanding is that {zu'o} has a {ce'u} place, as "giving" {lo zu'o dunda} and "receiving" {lo zu'o te dunda} are certainly two different activities.

[selpa'i]

zu'o is a subtype of nu, which is a general abstractor for events. nu unites/stands in for pu'u, zu'o, mu'e and za'i. It saves you a syllable and the trouble of picking the right one of those four or of having to narrow it down. If zu'o had a ce'u, then so would nu.

[.djo,is.]

On Wednesday, October 3, 2012 4:42:40 PM UTC-7, selpa'i wrote:

Am 04.10.2012 01:08, schrieb .djo,is.:

On Wednesday, October 3, 2012 8:54:13 AM UTC-7, la gleki wrote:

I want to eat an apple    = I want that I eat an apple [a really awkward lojban-style sentence]

{mi djica lo ka citka??)  = mi djica lo nu mi/vo'a citka lo plise

People seem to use {nu} or {ka} a lot of times when {zu'o} would be more useful. At least my understanding is that {zu'o} has a {ce'u} place, as "giving" {lo zu'o dunda} and "receiving" {lo zu'o te dunda} are certainly two different activities.

zu'o is a subtype of nu, which is a general abstractor for events. nu unites/stands in for pu'u, zu'o, mu'e and za'i. It saves you a syllable and the trouble of picking the right one of those four or of having to narrow it down. If zu'o had a ce'u, then so would nu.

Shoot. Looks like I've been totally misusing it.

[la gleki]

On Wednesday, October 3, 2012 8:21:32 PM UTC+4, selpa'i wrote:

Am 03.10.2012 17:54, schrieb la gleki:

If�{ka=su'u ce'u} then we'll have very nice and compact way of expressing such an important and�useful concept of natlangs as infinitives.

I don't like how gua\spi works here cuz as i can see it demands using another gismu every time you change the value of {ce'u}.

Okay, what you want is *exactly* what gua\spi is doing, so I think you misunderstand something. I wouldn't add those "ce'u formulae" as a seperate field. Instead, I would add them right to the definition. Gua\spi does this:

zmadu : x1 exceeds x2 in property/quantity x3:1+2

I'm not sure what you mean by

I don't like how gua\spi works here cuz as i can see it demands using another gismu every time you change the value of {ce'u}.

Can you show me what you mean?

Although it's not the right place to discuss it here please translate to gua\spi

1."I want to eat an apple".

2."I want you to eat an apple".

btw, what is your opinion about {ka=su'u ce'u}?

[selpa'i]

Am 04.10.2012 09:34, schrieb la gleki:

I'm not sure what you mean by

I don't like how gua\spi works here cuz as i can see it demands using another gismu every time you change the value of {ce'u}.

Can you show me what you mean?

Although it's not the right place to discuss it here please translate to gua\spi

So I have to translate this for you, but you don't like how it works? I'm not sure you can properly judge it then.

1."I want to eat an apple".

The normal way uses an infinitive compound:

^:i \ji /daw crw \xo plyw

But you can also use an explicit infinitive:

^:i \ji /daw \vo crw \xo plyw

2."I want you to eat an apple".

^:i \ji /gu pli \ju ^vo crw \xo plyw

or

^:i \ji ^ju /gu pli \crw \xo plyw

btw, what is your opinion about {ka=su'u ce'u}?

I'm not sure how to answer that question. You can say that ka is du'u ce'u, I don't know what the advantage is in defining ka in terms of su'u, because to me su'u could then mean nu or ni, which means that ka could become nu ce'u or ni ce'u. So why not use a more precise abstractor? Also, what is the point anyway? ka has at least one ce'u in it, that's pretty clear to almost everyone. Why do you need to use su'u here?

mu'o mi'e la selpa'i

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

doị mèlbi mlenì'u
   .i do càtlu ki'u
ma fe la xàmpre ŭu
   .i do tìnsa càrmi
gi'e sìrji se tàrmi
   .i taị bo pu cìtka lo gràna ku

[la gleki]

On Thursday, October 4, 2012 4:36:54 PM UTC+4, selpa'i wrote:

Am 04.10.2012 09:34, schrieb la gleki:

I'm not sure what you mean by

I don't like how gua\spi works here cuz as i can see it demands using another gismu every time you change the value of {ce'u}.

Can you show me what you mean?

Although it's not the right place to discuss it here please translate to gua\spi

So I have to translate this for you, but you don't like how it works? I'm not sure you can properly judge it then.

I have a general idea how it should be translated into gua\spi.

The problem of translation might be of a philosophical kind.

1."I want to eat an apple".

The normal way uses an infinitive compound:

^:i \ji /daw crw \xo plyw

But you can also use an explicit infinitive:

^:i \ji /daw \vo crw \xo plyw

2."I want you to eat an apple".

^:i \ji /gu pli \ju ^vo crw \xo plyw

or

^:i \ji ^ju /gu pli \crw \xo plyw

But I have a clear feeling that in both sentences the same semantic prime can be used. And this prime describes "desire".

Lojban can replace {ce'u} with anything. Natlangs can do the same. gua\spi can't. {to zoi gy. I don't want to criticize gua\spi anymore. gy. toi}

btw, what is your opinion about {ka=su'u ce'u}?

I'm not sure how to answer that question. You can say that ka is du'u ce'u, I don't know what the advantage is in defining ka in terms of su'u, because to me su'u could then mean nu or ni, which means that ka could become nu ce'u or ni ce'u. So why not use a more precise abstractor? Also, what is the point anyway? ka has at least one ce'u in it, that's pretty clear to almost everyone. Why do you need to use su'u here?

Well, I used su'u because wasn't sure about {du'u}. And {* lo ka lo penbi cu clani } had bewildered me.

I just wanna know whether it would be correct to add the definition of {ka=du'u ce'u} into jvs or not.

[selpa'i]

Am 04.10.2012 17:02, schrieb la gleki:>
>
>> 1."I want to eat an apple".
>>
>> The normal way uses an infinitive compound:
>>
>> ^:i \ji /daw crw \xo plyw
>>
>> But you can also use an explicit infinitive:
>>
>> ^:i \ji /daw \vo crw \xo plyw
>>
>> 2."I want you to eat an apple".
>>
>> ^:i \ji /gu pli \ju ^vo crw \xo plyw
>>
>> or
>>
>> ^:i \ji ^ju /gu pli \crw \xo plyw
>>
>
> But I have a clear feeling that in both sentences the same semantic
> prime can be used. And this prime describes "desire".
> Lojban can replace {ce'u} with anything. Natlangs can do the same.
> gua\spi can't. {to zoi gy. I don't want to criticize gua\spi anymore.
> gy. toi}

You cannot replace ce'u at all or else it's gone and it's not a
ka-abstraction anymore (or not a well-formed one). What natlangs can and
can't do has little relevance when discussing Lojbanic topics such as
ka-abstractions.

In Lojban, djica2 is a nu, not a ka. You could say that djica should be
polymorphic and allow both nu and ka, but I don't think that's what
you're saying, is it? (I don't know *what* you are saying).
Why is it a nu? Because you can djica things that don't involve
yourself. (Gua\spi's _daw_ is restricted to desiring to do or be
something, hence it's always like a Lojban ka. And that's why the second
example uses a different predicate.)

Again, what is the difference between the Lojban and the gua\spi sentence?

>
>>
>> btw, what is your opinion about {ka=su'u ce'u}?
>>
>> I'm not sure how to answer that question. You can say that ka is
>> du'u ce'u, I don't know what the advantage is in defining ka in
>> terms of su'u, because to me su'u could then mean nu or ni, which
>> means that ka could become nu ce'u or ni ce'u. So why not use a more
>> precise abstractor? Also, what is the point anyway? ka has at least
>> one ce'u in it, that's pretty clear to almost everyone. Why do you
>> need to use su'u here?
>
>
> Well, I used su'u because wasn't sure about {du'u}. And {* lo ka lo
> penbi cu clani } had bewildered me.
> I just wanna know whether it would be correct to add the definition of
> {ka=du'u ce'u} into jvs or not.

That seems pointless and it's not a real definition either. You can't
blindly replace "ka" with "du'u ce'u". It has been said many times that
"a ka-abstraction is a du'u-abstraction that contains at least one
ce'u". This is one view you can have, I don't know if everybody would
agree with it. We don't just add definitions for the cmavo, that's the
job of the - currently idle - BPFK.
Sorry, I don't understand your motivation here. Try to be a bit more
clear, please.

[la gleki]

On Thursday, October 4, 2012 11:47:16 PM UTC+4, selpa'i wrote:

Am 04.10.2012 17:02, schrieb la gleki:>
 >
 >>     1."I want to eat an apple".
 >>
 >>     The normal way uses an infinitive compound:
 >>
 >>     ^:i \ji /daw crw \xo plyw
 >>
 >>     But you can also use an explicit infinitive:
 >>
 >>     ^:i \ji /daw \vo crw \xo plyw
 >>
 >>     2."I want you to eat an apple".
 >>
 >>     ^:i \ji /gu pli \ju ^vo crw \xo plyw
 >>
 >>     or
 >>
 >>     ^:i \ji ^ju /gu pli \crw \xo plyw
 >>
 >
 > But I have a clear feeling that in both sentences the same semantic
 > prime can be used. And this prime describes "desire".
 > Lojban can replace {ce'u} with anything. Natlangs can do the same.
 > gua\spi can't. {to zoi gy. I don't want to criticize gua\spi anymore.
 > gy. toi}

You cannot replace ce'u at all or else it's gone and it's not a
ka-abstraction anymore (or not a well-formed one).

True. Still the same brivla can be used. Unlike gua\spi.

What natlangs can and
can't do has little relevance when discussing Lojbanic topics such as
ka-abstractions.

Then gua\spi has little relevance too.

In Lojban, djica2 is a nu, not a ka. You could say that djica should be
polymorphic and allow both nu and ka, but I don't think that's what
you're saying, is it? (I don't know *what* you are saying).

My only complaint  that we have a nice shortcut of saying {du'u ce'u} but we don't have one for {nu ce'u}.

Why is it a nu? Because you can djica things that don't involve
yourself. (Gua\spi's _daw_ is restricted to desiring to do or be
something, hence it's always like a Lojban ka. And that's why the second
example uses a different predicate.)

Again, what is the difference between the Lojban and the gua\spi sentence?

Hopefully no semantic difference. Looks like Lojban just gives more freedom in recombining the same words without drawing in extra predicates.

 

 >
 >>
 >>     btw, what is your opinion about {ka=su'u ce'u}?
 >>
 >>     I'm not sure how to answer that question. You can say that ka is
 >>     du'u ce'u, I don't know what the advantage is in defining ka in
 >>     terms of su'u, because to me su'u could then mean nu or ni, which
 >>     means that ka could become nu ce'u or ni ce'u. So why not use a more
 >>     precise abstractor? Also, what is the point anyway? ka has at least
 >>     one ce'u in it, that's pretty clear to almost everyone. Why do you
 >>     need to use su'u here?
 >
 >
 > Well, I used su'u because wasn't sure about {du'u}. And {* lo ka lo
 > penbi cu clani } had bewildered me.
 > I just wanna know whether it would be correct to add the definition of
 > {ka=du'u ce'u} into jvs or not.

That seems pointless and it's not a real definition either. You can't
blindly replace "ka" with "du'u ce'u". It has been said many times that
"a ka-abstraction is a du'u-abstraction that contains at least one
ce'u". This is one view you can have,

I do. {ka} not always  = {du'u ce'u}, sure. May be it can be expressed as {du'u [...] ce'u}.

I don't know if everybody would
agree with it. We don't just add definitions for the cmavo, that's the
job of the - currently idle - BPFK.
Sorry, I don't understand your motivation here. Try to be a bit more
clear, please.

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?

[selpa'i]

Am 05.10.2012 16:41, schrieb la gleki:
>
>
> On Thursday, October 4, 2012 11:47:16 PM UTC+4, selpa'i wrote:
>
> Am 04.10.2012 17:02, schrieb la gleki:>
> >
> >> 1."I want to eat an apple".
> >>
> >> The normal way uses an infinitive compound:
> >>
> >> ^:i \ji /daw crw \xo plyw
> >>
> >> But you can also use an explicit infinitive:
> >>
> >> ^:i \ji /daw \vo crw \xo plyw
> >>
> >> 2."I want you to eat an apple".
> >>
> >> ^:i \ji /gu pli \ju ^vo crw \xo plyw
> >>
> >> or
> >>
> >> ^:i \ji ^ju /gu pli \crw \xo plyw
> >>
> >
> > But I have a clear feeling that in both sentences the same semantic
> > prime can be used. And this prime describes "desire".
> > Lojban can replace {ce'u} with anything. Natlangs can do the same.
> > gua\spi can't. {to zoi gy. I don't want to criticize gua\spi
> anymore.
> > gy. toi}
>
> You cannot replace ce'u at all or else it's gone and it's not a
> 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 instance by saying
that djica2 ba a ka (which is a bad example, but it illustrates the
point), then you will not be able to use it for "I want you to broda",
because that's a different predicate that doesn't involve yourself in
the abstraction. This is a *strenght* of gua\spi; its predicates are
semantically much clearer.

>
> What natlangs can and
> can't do has little relevance when discussing Lojbanic topics such as
> 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
place. Gua\spi's entire gimste is ce'u-ized, that's what it looks like.

>
>
>
> In Lojban, djica2 is a nu, not a ka. You could say that djica should be
> polymorphic and allow both nu and ka, but I don't think that's what
> you're saying, is it? (I don't know *what* you are saying).
>
>
> My only complaint that 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
a du'u abstraction and add a ce'u to it.

>
> Why is it a nu? Because you can djica things that don't involve
> yourself. (Gua\spi's _daw_ is restricted to desiring to do or be
> something, hence it's always like a Lojban ka. And that's why the
> second
> example uses a different predicate.)
>
> Again, what is the difference between the Lojban and the gua\spi
> sentence?
>
>
> Hopefully no semantic difference. Looks like Lojban just gives more
> freedom in recombining the same words without drawing in extra predicates.

Okay, but that wasn't even your original point. And as I tried to
explain above, you get seperate predicates if you ce'u-ize the gimste.
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 regard,
but you have to remember that gua\spi is not Lojban, and not everything
can be copied 1:1.

[la gleki]

Well, I'm not sure if we should ce'u-ize gismu with {nu}  abstractions.

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.

We could also say {mi djica lo nu ce'u citka i do na go'i} = "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 *strenght* of gua\spi; its predicates are
semantically much clearer.

Well, well, I don't want someone to stop learning gua\spi  because "gua\spi is a crap. I know, la gleki told me". :)

[Jacob Errington]

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. 

There are some selbri, say kakne, where {zo'e} in the abstraction kakne2 breaks the meaning in some way, such as {mi kakne lo nu do citka lo plise}. IMO selbri that get *weird* when there's no ce'u should have {ka}. 

Furthermore, nu+ce'u is strange to me, because {ce'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 abstractors are function when they contain {ce'u}, but can equally not contain {ce'u}. {ni} is an example of such an abstractor:

{.i mi zmadu do lo ni xendo} vs {.i lo ni mi prami do cu 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 mi do cinba} forms one single object that doesn't depend on the containing bridi. 

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 abstractor, and its type is completely context-dependent. Although this has never really happened as far as I know, it would, however, be possible to create new types of abstractors by means of su'u2.

.i mi'e la tsani mu'o

--

You received this message because you are subscribed to the Google Groups "lojban" group.

To view this discussion on the web visit https://groups.google.com/d/msg/lojban/-/WyFuvnW2QSEJ.

[la gleki]

On Saturday, October 6, 2012 8:21:14 PM UTC+4, tsani 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. 

zo'e doesn't refer to first places of nesting bridi. 

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}.

[Jacob Errington]

On 6 October 2012 09:33, la gleki <gleki.is...@gmail.com> wrote:

On Saturday, October 6, 2012 8:21:14 PM UTC+4, tsani 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. 

zo'e doesn't refer to first places of nesting bridi. 

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 refer 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.

To view this discussion on the web visit https://groups.google.com/d/msg/lojban/-/JzUcrDw0IJAJ.

[la gleki]

On Sunday, October 7, 2012 6:45:40 AM UTC+4, tsani wrote:

On 6 October 2012 09:33, la gleki <gleki.is...@gmail.com> wrote:

On Saturday, October 6, 2012 8:21:14 PM UTC+4, tsani 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. 

zo'e doesn't refer to first places of nesting bridi. 

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 refer 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 don't want to redefine anything and to limit djica to ka+ce'u structures.

mi djica lo nu ro remna ku citka

mi djica lo nu/ka ce'u citka

^ ^ that is my proposal. However, {ce'u} or {ka+ce'u} in the second sentence might be replaced with something else (I don't care with what but it must be a variable, not a constant like {mi}).

[la gleki]

I also have to say thatit's simpkly strange not to have a simple way to express logophor in Lojban.

Actually neither {vo'a} nor {ri} can do the trick.

{vo'a} won't refer to the nesting bridi we need if there are more than two nesting levels.

e.g. in {do djuno lo du'u mi djuno lo du'u vo'a stati} vo'a refers to do but we need it to refer to mi.

In {.i ta mlatu .i djica lo nu ri sipna kei fa lo remna} ri refers to the cat whereas we want it to refer to lo remna.

The only way is to use {goi} which produces awkward sentences.

Let's take two examples.

1. I'm afraid to revenge [on people]

2. I'm afraid of revenge.

1. lo remna goi ko'a terpa lo nu ko'a venfu

2. lo remna terpa lo nu venfu 

The differrence is pretty clear.

Let's add two more examples.

3. lo remna ku kakne lo ka venfu

4. lo remna ku snada lo ka venfu 

doi la tsani, your phrase *this has never been a problem before* can be applied to any language.

Yes, you can always cheat both in Lojban and natural language.

for me the difference between first sentence and 3. and 4. is pretty clear. 1. just looks akward.

[selpa'i]

Am 08.10.2012 13:11, schrieb la gleki:
> I also have to say thatit's simpkly strange not to have a simple way to
> express logophor in Lojban.
> Actually neither {vo'a} nor {ri} can do the trick.
>
> {vo'a} won't refer to the nesting bridi we need if there are more than
> two nesting levels.

> e.g. in {do djuno lo du'u mi djuno lo du'u vo'a stati} /vo'a /refers to
> /do/ but we need it to refer to /mi./

do djuno lo du'u mi djuno lo du'u lo no'a cu stali

> /
> /
> In {.i ta mlatu .i djica lo nu ri sipna kei fa lo remna} /ri/ refers to
> the cat whereas we want it to refer to /lo remna/.

ta mlatu .i djica lo nu vo'a sipna kei fa lo remna

>
> The only way is to use {goi} which produces awkward sentences.
>
> Let's take two examples.
> 1. I'm afraid to revenge [on people]
> 2. I'm afraid of revenge.
>
> 1. lo remna goi ko'a terpa lo nu ko'a venfu
> 2. lo remna terpa lo nu venfu

1. lo remna cu terpa lo nu vo'a venfu
2. lo remna cu terpa lo nu [vo'a] se venfu

You rarely need the "vo'a" or "lo no'a" in such simple sentences,
specifying explicitly only happens when confusion might arise.

Lojban does not have a static solution for logophoric pronouns, but it
has the tools for a dynamic one, using vo'a, no'a and nei.

mu'o mi'e la selpa'i

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

doị mèlbi mlenì'u
.i do càtlu ki'u
ma fe la xàmpre ŭu
.i do tìnsa càrmi
gi'e sìrji se tàrmi
.i taị bo pu cìtka lo gràna ku

.

[la gleki]

Well, I hope this will be the last message in this thread.

Sorry for not running this thread properly and for poorly formulated questions and ideas.

I have been overwhelmed by other non-lojbanic stuff lately.

anyway ki'esai