WikiDiscuss

WikiDiscuss


posts: 149

webmaster@lojban.org scripsit:

> I (Robin Powell) am of the opinion that the "set" places
> in various gismu are un-necessary, and should be abolished (with
> the obvious exception of gismu that are specifically about sets).
> Any specification of a group should be acceptable in these places.

I don't think that's going to fly unless we have a list of which
places you want to change. Sets were used there *because* they were
the singularist view of a group.

--
Andrew Watt on Microsoft: John Cowan
"Never in the field of human computing cowan@ccil.org
has so much been paid by so many http://www.ccil.org/~cowan
to so few!" (pace Winston Churchill) http://www.reutershealth.com


posts: 1912


John Cowan:
> > I (Robin Powell) am of the opinion that the "set" places
> > in various gismu are un-necessary, and should be abolished (with
> > the obvious exception of gismu that are specifically about sets).
> > Any specification of a group should be acceptable in these places.
>
> I don't think that's going to fly unless we have a list of which
> places you want to change.

I find 22 gismu with places reserved exclusively for sets:
slilu, bridi, kampu, simxu, steci, mupli, fadni, rirci,
cnano, ralju, cuxna, sisku, kancu, girzu, ciste, liste,
porsi, pluta, kruvi, linji, plita, kurfa.

Other gismu have places that mention sets, but they also allow
other things there. For example, the place structure of {fenso}
says that you can saw individuals together, or if you prefer
you can saw a set together.

> Sets were used there *because* they were
> the singularist view of a group.

But since there are examples of non-distributive places where
the gi'uste allows normal individuals, there is no reason for
these places to be restricted exclusively to sets.

mu'o mi'e xorxes






__
Do you Yahoo!?
New and Improved Yahoo! Mail - 100MB free storage!
http://promotions.yahoo.com/new_mail


posts: 14214

On Fri, Aug 13, 2004 at 01:33:52PM -0700, Jorge Llamb?as wrote:
>
> John Cowan:
> > > I (Robin Powell) am of the opinion that the "set" places
> > > in various gismu are un-necessary, and should be abolished (with
> > > the obvious exception of gismu that are specifically about sets).
> > > Any specification of a group should be acceptable in these places.
> >
> > I don't think that's going to fly unless we have a list of which
> > places you want to change.
>
> I find 22 gismu with places reserved exclusively for sets: slilu,
> bridi, kampu, simxu, steci, mupli, fadni, rirci, cnano, ralju, cuxna,
> sisku, kancu, girzu, ciste, liste, porsi, pluta, kruvi, linji, plita,
> kurfa.
>
> Other gismu have places that mention sets, but they also allow other
> things there. For example, the place structure of {fenso} says that
> you can saw individuals together, or if you prefer you can saw a set
> together.
>
> > Sets were used there *because* they were the singularist view of a
> > group.
>
> But since there are examples of non-distributive places where the
> gi'uste allows normal individuals, there is no reason for these places
> to be restricted exclusively to sets.

What he said.

More: jbini, bende (not kidding; read the notes), traji (the broken x4
place).

Note, interestingly, that girzu has a set-required place, but gunma does
not. So it's not even consistent, that I can see. Amazingly, cmima is
actually *unclear* as to whether the second argument must be a set.

Some words say "(x2, if a set, is completely specified)", which I'm fine
with.

Note that, unlike xorxes, I *love* sets, and I love that Lojban has them
as a primitive type (although we're missing an operator). But requiring
sets as opposed to other types of groups seems silly in most cases.

-Robin


posts: 1912


Robin:
> More: jbini, bende (not kidding; read the notes), traji (the broken x4
> place).

{jbini} already allows for non-sets though. I missed {bende} because the
restriction is not in the definition field. I don't know why I missed
{traji}.

> Note, interestingly, that girzu has a set-required place, but gunma does
> not.

I am happy with {gunma} now. x1 is the reified group, a single thing,
and x2 are the individuals as individuals, many things, a non-distributive
place. I couldn't make any sense of the place structure before
understanding plurals a la McKay.

> So it's not even consistent, that I can see. Amazingly, cmima is
> actually *unclear* as to whether the second argument must be a set.

{cmima} clearly says it's not only for sets. The difference between
{cmima} and {se gunma} is that x1 is distributive in {cmima} (each
member) and non-distributive in {se gunma} (all members together).

> Note that, unlike xorxes, I *love* sets,

I have nothing against lo cmaci selcmi.

> and I love that Lojban has them
> as a primitive type

I don't like that.

> (although we're missing an operator).

You can use {ku'a gi ... ginai ...} for it, if it's the one
I think you mean.

mu'o mi'e xorxes




__
Do you Yahoo!?
Yahoo! Mail Address AutoComplete - You start. We finish.
http://promotions.yahoo.com/new_mail


posts: 14214

On Fri, Aug 13, 2004 at 02:29:51PM -0700, Jorge Llamb?as wrote:
>
> Robin:
> > More: jbini, bende (not kidding; read the notes), traji (the broken
> > x4 place).
>
> {jbini} already allows for non-sets though.

It's unclear.

jbini bin bi'i between
x1 is between/among set of points/bounds/limits x2
(set)/amidst mass x2 in property x3 (ka)

x2 (a complete set, generally ordered) defines the
bounds/limits/range for x1
;

The word "mass" is in there, but then it says "complete set".

> > So it's not even consistent, that I can see. Amazingly, cmima is
> > actually *unclear* as to whether the second argument must be a set.
>
> {cmima} clearly says it's not only for sets.

I don't see the clarity you're seeing.

> > and I love that Lojban has them as a primitive type
>
> I don't like that.
>
> > (although we're missing an operator).
>
> You can use {ku'a gi ... ginai ...} for it, if it's the one I think
> you mean.

I mean set subtraction. Assuming X - Y == X ku'a !Y, then yeah, that
works.

-Robin