Message-Id: <199501230116.AA14796@nfs1.digex.net>
Date:         Mon, 23 Jan 1995 01:13:58 +0000
Reply-To: ucleaar <ucleaar@ucl.ac.uk>
Sender: Lojban list <LOJBAN@CUVMB.BITNET>
From: ucleaar <ucleaar@ucl.ac.uk>
Subject:      replies re. ka & mamta be ma
To: Bob LeChevalier <lojbab@access.digex.net>
In-Reply-To:  (Your message of Sun, 22 Jan 95 14:56:05 EST.)
Status: RO

Goran:
> > > There is no x2 in {ka}, at least in my vlaste.
> > There is in my list, dated 6.1.93 and labelled as baselined at the
> > start of the file.
> Mine is dated 06/13/94. No x2.

HAS SOMEONE CLANDESTINELY ABOLISHED X2 OF {ka}?

> > I wouldn't know what a property is, then.
> I thought before that it's property of x1 of being broda. Now I'm not so
> sure, and I've come to ju'ocu'i think that it's the property of each of
> the arguments of satisfying the relationship. {lo ka mi klama le zarci}
> is my going-to-the-market-ness and the market's being-gone-to-by-me-ness

I'm still not clear. I've offered what to me is a fairly clearish
characterization of the meaning: the properties responsible for
something being categorizable as broda. So my going-to-market-ness is
the same as your-going-to-the-marketness, I'd have thought. Furthermore,
if ka refers to properties of each of the arguments then it refers to
n (n=number of arguments) clusters of properties, which would be very
confusing.

> > I haven't changed my mind. So far, I think, your versions have just
> > pertained to lohe glipre.
> Then put {rolo} instead of {lo'e}. If that still doesn't satisfy you,
> then say for god's sake what you want to express, because I just keep
> guessing your intentions.

In general, if an Englishman likes whisky his taste is acquired. As I've
said {lohe glipre poi vusnei la .uiskis} or {lohe ka vusnei la .uiskis.
kei be lo glipre} would do the job. So there isn't anything I've been
saying is still unsayable.

Jorge:
> > > > > > How would you say "the mothers of Jorge and And"?
> > > > > Perhaps {lei mamta be la xorxes a la and}.
> > > > That could refer to just your mum.
> > > Yes, you'd have to rely on the specificity of {lei}.
> > I.e. an actually-nonveridical usage. I see.
> No! A veridical usage: "The mass of those who are mothers of you or me."

You're right of course. I hadn't noticed the change of .e to .a.
I could change the example to "siblings of Jorge and And" (nonspecific
"siblings") to make the point that the problem remains (not that you
suggested otherwise).

> If you want to be really drastic: {ro lo re mamta be la xorxes a la and}

Wouldn't work for the siblings example, of course.

> > > > Ax, x is a member of {Xorxe, And}: I met the mother of x.
> > > >    I met lo mamta be xohu ro luha luhi la xorxes ce la and
> > This is exactly the purpose of xohu [so named by me, but proposed by
> > pc]. Normally as you work left to right through the sentence you add
> > elements to the right end of the (implicit) prenex, so if X is after
> > Y then X is within the scope of Y. But when you hit something marked
> > with xohu you add it to the start of the implicit prenex instead.
> > Neat, eh? That's how xohu works in general.
> What I had understood before was that {xo'u} lets you jump from one
> prenex (the one in the embedded clause) to another (the one in the
> outside clause). Now you are using it for something different.

Well if that were the case, the issue would still arise of where in the
higher prenex you jump to - the start of the prenex so far, or the end
of the prenex so far. My version - "jump to the start of the highest
prenex" generalizes one of those two solutions.

> You are requiring that the {ro} be actually transported to the prenex,
> and ignored in the body. This never happens. In general, if you explicitly
> write the word {ro} in the body, you override the prenex, so
>  ro da su'o de zo'u de prami ro da
>  For every x, there is a y, such that y loves every x.
> Is equivalent to:
>  de prami ro da
>  There is a y that loves every x.

By "implicit" prenex I meant the sequence of quantifiers in the logico-
semantic form derived from the sentence.
So, given the sentence {de prami ro da}, you build the log-sem form
   Ex Ay: x prami da
While, given the sentence {de prami xohu ro da}, first you encounter
{de}, and build:
   Ex
then you encounter {xohu ro da}, and instead of adding "Ay" after the
"Ex", you instead add it to the front:
   Ay Ex
So {de prami xohu ro da} means "Everyone is loved by someone; everyone
has a lover".

You, as a lover of flexibility, should surely appreciate such a cmavo.


> > > > Can you give construction-independent rules for interpreting
> > > > {keha} and {makau}?
> I'll try again for {makau}. It is a specific sumti that is not necessarily
> identified by the speaker. Let me give another example. Say there's a blue
> house, and I want to say "I like the colour of that house". A literal
> translation would be {mi se pluka le se skari be le zdani}. But this is
> not what I usually would mean. All that says is that I like the colour
> blue. What I mean is that I like a property of the house, not that I like
> a certain colour. A better translation would then be {mi se pluka
> le ka ke'a skari makau kei be le zdani}.

O wow. Acute. You're wasted on physics.

Are you sure do don't just want {mi se pluka le nu le zdani cu skari da}?
I still don't understand {makau} here!

> It would not be the same to say {mi se pluka le ka ke'a skari da kei
> be le zdani}, because that says that I like that the house has a colour
> (ie that it is not transparent, I suppose).

Okay, point taken. "I like the house having a colour". The solution
could simply be to make x2 of skari specific. Use {keha} or {le cohe}
or something:
  mi se pluka le nu le zdani cu skari le cohe
  mi se pluka le ka keha skari le cohe kei be le zdani

"I like it that the house has that colour".

> > Is there a word that means "bridi with one empty tersumti" - i.e. what
> > the complement of LE is? This is what the complement of ka is.
> In syntactic terms, the complement of LE is a sumti-tail, which indeed
> is a selbri with one of its tersumti singled out, namely the x1.
> I think that {ke'a} (or some other lambda variable) is the way to single
> out a tersumti for {ka}. The alternative (forcing it to be the x1, like
> for LE) is in my view unnecessarily restrictive.
If this ever gets resolved I expect it will be in the way you propose,
since adding interpretation rules for {keha} involves less change to
the grammar than changing the syntax of {ka} and {lihi} to make them
require a sumti-tail.

> If something can be expressed in two ways, which way is "logically
> necessary", and which is an optional add-on? You could say that each is
> logically necessary if you eliminate the other.

Well first before you consider logical necessity you set aside
considerations of flexibility. Then you see if either construction
is used elsewhere in the grammar: if one has greater generality,
use it.

I suppose a good example is SE versus FA. SE could not be got rid
of without a hefty redesign of the syntax, whereas FA could be
dispensed with. But FA gives lots of flexibility, so is a useful
add-on.

-----
And