[EM] Reply to Demorep. Ambiguity issues.

[MIKE OSSIPOFF]

>One or more examples of real votes will suffice to show if the claim(s) of 
>a
>method is(are) ambiguous.

Of course an example is needed. But the example isn't enough,
by itself, to show that a criterion is ambiguious. What's
needed is claims, by someone defending the criterion's current
wording, and by someone claiming ambiguity for it, of what
it means to meet a criterion.

Markus has claimed that such
a meaning must include a requirement that the criterion's
requirement be met no matter which N-candidate permutation of
the example's candidates are taken as the N candidates referred
to by letters in the criterion. That seems quite incompatible with
the wording of the criteria, and It's never occurred to me that
anyone believed that that should be required.

Anyway, I've written, in my reply to Markus tonight
(September 17th 2000, 4:33:02 GMT) a claim for what it means
to meet a criterion that names candidates by letters. This
meaning is more general than yesterday's, and simpler, and
makes more sense.

The reason why I wrote it differently yesterday is that I was
working from Markus's claim. He said that,for every permutation of
N candidates from the example, the criterion's requirement must
be met for those N candidates--and so I said instead of requiring
it to be met for _all_ of the permutations of the example's
candidates simultaneously, it should only have to be met by
each one of those permutations, taken one at a time. But that's
an awkward 2-step meaning, where we pick an example, and then
, additionally, start picking permutations of its candidates.

Tonight's meaning is more general, where we instead try examples
such that each example has an A, B, C...  That's briefer, simpler,
neater. And it also happens to work without contradictions or
inconsistency for criteria that say that a certain group of voters
must have a certain kind of a way to vote. You could say I
contrived the meaning to work with that kind of a criterion, but
I point out that this meaning is simpler, briefer, neater.

I asked Markus to post what he thinks it means to meet a criterion
that names some candidates by letter designations, but he hasn't
posted one. Presumably, if he did, it would be one that incorporates
his requirement that the criterion's requirement be met
simultaneously with respect to every one of the N-candidate
permutations that could be taken from the example's candidates
to be the N candidates that the criterion refers to by letters.

I suspect that maybe such a meaning might run into contradictions
or inconsistencies when trying to say what kind of an example
is needed. In any case, Markus hasn't suggested a meaning, and
so the one that I wrote tonight is the only one so far.

But I hope Markus doesn't claim that, because his type of meaning
can't be defined without inconsistencies or contradictions, that
means that those are problems of my criteria. No, they're only
problems of that type of a compliance meaning for criteria.
And if Markus's type of meaning has those problems, then he can't
say that that's what it should mean to meet a criterion that
refers to candidates by letter designations. That, in addition
to the more obviousness & reasonableness of my claimed meaning,
is another reason why Markus' requirement is unreasonable.

Mike Ossipoff

_________________________________________________________________________
Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com.

Share information about yourself, create your own public profile at 
http://profiles.msn.com.