Re: xor questions

[John Cowan <cowan@locke.ccil.org>]

la lojbab. pupu cusku di'e

> >>ma du lu'a lo tcati onai lo ckafi lu'u poi se vasru le patxu
> >>what is the identity of [tea XOR coffee] which is contained in the pot.

la xorxes. pu cusku di'e

> >As with all other tries, {lo glare} is a true and unhelpful answer.
> >(Or probably {lo lenku} by now.) Any question with {ma} will suffer
> >from this, I suspect.

la lojbab. cusku di'e

> Why is either lo glare or lo lenku valid there.  I have asked for a member
> of the set or at leats something identical to a member of the set.  maybe I
> need to play around with "me" or "cmima" or even "pamei" if we want to
> formally restrict ourselves to members of a specific set.  But I don't see
> how "lo glare" is a member of a set that does not specifically include "lo
> glare" in its membership.

Because "du" is object identity: it claims that its sumti refer to the same
object.  When you say "ma du X" then any sumti that co-references X is
a correct answer, even if unhelpful because of its over-generality.

For example, if you ask me "Is it Cicero or Caesar who is the author of the
First Oration Against Catiline?" then I may truthfully reply
"Tully", because Cicero is Tully (that is, the names "Cicero" and "Tully"
co-refer), even if you don't know that.  With equal truth, and even
greater unhelpfulness, I may reply "A Roman" or "A man" or "An object",
since Cicero is all these things: it is correct to assert that a Roman
is a member of the set {Cicero, Caesar}.

The only way to avoid this problem is to move to a meta-level, asking
about words rather then objects.  In that case, the question becomes
"What word from the set {"Cicero", "Caesar"} correctly names the author
of the First Oration etc."  Note that the members of the set here are
words, not objects: it would not be correct to assert that "a Roman"
was a member of the set {"Cicero", "Caesar"}.


-- 
John Cowan	http://www.ccil.org/~cowan		cowan@ccil.org
			e'osai ko sarji la lojban