Re: proposals

[jorge@PHYAST.PITT.EDU]

> > Not all BAIs can be sensibly used as connectives, only those whose
> > associated selbri have the same type of arguments in x1 and x2: du'i,
> > mau, me'a, ki'u, ri'a, ba'i, etc.
>
> pc: And here is the havoc we wreak.  Either we have to introduce a new
> lexeme or we legitimate a vast array of meaningless BAI-as-conjunction forms.

They are already legitimate, in forethought form: {bai gi ... gi ... }

(And in any case, the array of meaningless forms is not limited to BAIs.
There are lots of other constructions that are meaningless.  Numbers and
anaphora are a good source for those.)

> >         la migelitos prami la mafaldas maubo la susanitas
> >         Miguelito loves Mafalda more than Susanita.
> >
> > It is better than the form with {.e}, because no absolute claim
> > is made, only the comparative one.
>
> pc:  But I have to admit (not even grudgingly) that this makes a very
> nice case for something like what is proposed. Deuce, I think; time for
> more evidence.

More evidence:

        ko'a klama la paris ba'ibo la romas
        He goes to Paris instead of Rome.

Here {.e} would even be incorrect.

We can already say it in forethought:

        ko'a klama ba'igi la paris gi la romas


> > >         la djan ? la maris zmadu la xeris ? la selis le ka prami
> >
> > I'd love to know what the "?" would be replaced with there. I can't
> > think of anything.
>
> pc: I take this as evidence that the connector for ordered pairs -- or
> perhaps its was for sequences generally -- has disappeared from the cmavo
> list.  Well, I said I liked the old list better and find myself deploring
> the last few years more and more.

Well, there are {ce} and {ce'o}, I think you mean those. The way they are
defined, it seems that they make a set out of the connected parts, but
since I think sets are mostly useless for language purposes, I wouldn't
mind using them like that:

     la djan ce'o la maris zmadu la xeris ce'o la selis le ka prami


> > I could get a similar sense with an unortodox use of "fa'u":
>
> >     la djan zmadu la xeris le ka [ke'a] prami la maris fa'u la salis
> >     John exceds Harry in their loving Mary respectively Sally.
>
> pc: How unorthodox (extent and manner)?

It's hard to say what's the orthodoxy here.  My understanding of {fa'u}
is that it should appear at least twice in the sentence, so that we know
what goes with what.  For instance:

        la djan fa'u la xeris prami la maris fa'u la selis
        John resp. Harry loves Mary resp. Sally.

        la djan fa'u la xeris pu fa'u ba klama le zarci
        John resp. Harry went resp. will go to the store.

If it shows up only once, then I have to understand it as being
distributed with some {ro}:

        le re prenu cu klama la paris fa'u la romas
        Each of the two persons goes to Paris resp. Rome.
        (i.e. one goes to Paris and the other to Rome.)

Or for more than two:

        le reno prenu cu klama la paris fa'u la romas
        Each of the twenty persons goes to Paris resp. Rome.
        (i.e. some go to Paris and some to Rome.)

(The main difference in this case with using logical "or" is that at
least one has to go to Paris and at least one to Rome, at least that's
how it seems to me.)

That is my speculation.  If there is only one {fa'u} in the sentence, I
don't see a clear rule, although there seem to be some sensible possible
interpretations.

> > But if we are not focusing only on the lovers, but also on the loved
> > ones, we could use termsets, plus my proposed {maubo} connector to get:
> >
> >   nu'i la djan la maris nu'u maubo nu'i la xeris la selis nu'u prami
> >   [ John  Mary ] more than [ Harry  Sally ] loves.
>
> pc: Does _nu'i_ fix order?

Yes.

> If not, this will miss the boat.  If it does
> preserve order, then it is what I needed above at "?", though with a
> different grammar.
>
> > (Using {emaubo} instead of {maubo}, that should be grammatical,

I was wrong. {nu'i...nu'u}s can only be connected with logical
connectives, so it doesn't work.

> >  Also, if we want to attach a relative clause, we
> > need the lambda variable to be explicited.
>
> pc:  I don't see this one: where would we attach a relative clause in an
> abstraction?

Well, for example:

        ko'a ko'e zmadu le ka ke'a prami
        He exceeds her in loving.

        ko'a ko'e zmadu le ka ke'a poi prenu cu prami
        He exceeds her in being a person who loves.


> >  I would argue that since it is already possible
> > to define the origin with the "vectors", then the magnitude should be
> > defined by the "tensors".
>
> pc:Two factors.  One is that the patterns is already established in at
> least two of three cases (and I meant it to be establ9shed for all three
> when this system was being set up) and so -- the parity part -- the same
> pattern applies in the third case.

This factor is "that's how things are now". That I recognize, otherwise
I wouldn't have proposed the new interpretation.

> Second, the origin is essenial in all
> of these calculations, the metric and the direction are not: "before" or
> "to the left" or "thirty feet" or "three minute"  make no sense for
> locating something unless there is a starting point (note that, as
> locators, the last three all need a "from/of <origin>" phrase).

That's not a big problem.  When used as tenses, the origin can't be
given directly either, and we still use them.  The origin is either the
here and now, or the one established by {ki}.  The fact that it is not
explicited doesn't mean that there is no starting point.

> Of
> course, if the origin is established in one place, the other factors does
> not need to do it again, so that other factor can be a free floating
> form, not a sumti tcita at all, but we need all the current sumti tcita
> to establish vector/tensor from otherwise unstated origins.

Why all?

> pc: Transmission error, I wrote "removing" and, as noted, what is lost is
> a way of fixing an axis for tensors standing alone.

That is not lost. The axis can always be fixed either using one of the
directionals (PU, FAhA), or with the origin fixer {ki}.

Jorge