[lojban] Re: [lojban-beginners] Re: About the negators

["Jorge Llambías" <jjllambias@gmail.com>]

On Fri, Dec 19, 2008 at 1:33 PM, Michael Turniansky
<mturniansky@gmail.com> wrote:
>
>      This is a very interesting argument.  I've never considered trying to
> glean meaning from how something parses, and I doubt that anyone can.

You can tell many things from the parse tree. For example, the grammar
is explicitly constructed so that {ko'a .e ko'e .a ko'i} will parse as
{(ko'a .e ko'e) .a ko'i} and {ko'a .e ko'e .a bo ko'i} will parse as
{ko'a .e (ko'e .a bo ko'i)}. There was no need to do it that way. In
fact, the grammar would be simpler, fewer rules, if it didn't care
about getting such groupings right.

> Can
> you tell me how "smaji" differs from "ganra" by how they parse?

No, the formal grammar will only tell you that they are both of
"selma'o" BRIVLA. (In fact it will also recognize they are both gismu,
but that's morphology.) The syntax won't tell you anything about the
meaning, which comes from the dictionary. The syntax won't distinguish
between the meanings of {na} and {ja'a} either, they are just members
of NA. But it can inform you about their scope.

> The best
> you can do is tell me that they are both three-argument gismu.

Not even that. The number of arguments is the same for all BRIVLA as
far as the grammar is concerned: every brivla accepts an infinite (or
indefinite) number of arguments.

> "Na" is
> forced by the grammar to be right before a selbri (or connective.  Any other
> legal places, disregarding UI and the like?  I'm not sure).

Strictly, it is part of the selbri, and it can alternate with tags, so
for example you can have {ba na bai ja'a ro roi ja'a na ja'a ja'a na
na vi broda} as one selbri.

> But how you can
> jump from syntax to meaning is beyond me.

The parse tree can be helpful to determine groupings and scopes. Of
course you can always define things so the meaning goes against what
the syntax suggests, but why do that?

> I wish it could be free when
> attached to ku, as PU is, without affecting its meaning, but it's not.  Why
> they made the choice they did to not let it, is beyond me.

Because if {naku} always had scope over every other operator, there
would be no way of saying such simple things as "some x are not
broda".


>> (1) su'o da na broda
>>
>> (3) su'o da ge na broda gi na broda
>>
>> This can only mean "some x is not broda and is not broda", which
>> reduces to "some x is not broda".
>
>   Cute, but that doesn't mean that you can then translate it back as "su'o
> da na broda".

Right. That's the problem.

>  There is no hard and fast transforms for
> arbitrary logical constructs,

Sure there are. Except for the crazy {na} rule, everything else (I
mean everything that has a direct correlate in predicate logic)
behaves according to relatively simple rules.

> any more than, while it is true that "ko'a ge
> du su'ore broda gi du su'oci broda"  can be transformed into the true
> sentence "ko'a du su'oci broda".

{su'ore} and {su'oci} are not basic predicate logic operators,
although they are easy to define in terms of {su'o} only. (And yes,
that transformation happens to be valid.)


>    Dang!!  I said I didn't wish to be sucked back into this discussion that
> we already agreed to disagree on, and there we are again, with a wider
> audience than last time :-(

There are things that are a matter of opinion or aesthetics, and it is
perfectly reasonable to disagree about that. For example, I can think
that {su'o da na broda gi'e na broda} and {su'o da na broda} having
different meanings sucks, and you can think it's perfectly fine. Or,
you can think that yes, it sucks, but we're stuck with it and there's
nothing to do about it because CLL is written as it is written and we
better not change it. We can agree to disagree about all that. But
sometimes you seem to challenge not my sense of aesthetics, or my
attitude towards some official rule, but some basic rules of logic.
You can't really maintain that "There is no hard and fast transforms
for arbitrary logical constructs". The whole point of logic is to have
hard and fast transforms for arbitrary logical constructs, so that we
can determine whether some line of logical reasoning is valid
independently of the semantic content of the predicates involved.

mu'o mi'e xorxes


To unsubscribe from this list, send mail to lojban-list-request@lojban.org
with the subject unsubscribe, or go to http://www.lojban.org/lsg2/, or if
you're really stuck, send mail to secretary@lojban.org for help.