First of all, I may have been too quick to criticize things like the
formulese in Bruce's letter. That particular speciment of formulese
wasn't the long or complicated kind that I was concerned about the
possible appearance of, so it was premature for me to comment about
The comment was motivated by ability that journal articles sometimes
have, to say in a page of formulese what could be said in a few lines
of English. Voting systems is a subject that can be easily discussed
in English & in which, it seems to me, everything can be said in
English. But formulese would be justifiable in instances where it's
clearer or appreciably more precise, or a lot briefer.
Anyway I was concerned that the list could be diverted from
its goals to translating lots of formulese, but, as I said,
no such problem had really showed up, so it was premature to
bring it up.
Standards & Criteria:
I don't want to seem stubborn--I've written criteria, yes/no tests,
to measure compliance with the standards that I've proposed.
As I've said, a standard can be a criterion, but it needn't be.
A criterion can be a standard if it's important enough to someone
to propose it as one.
Here's an example to show that a standard needn't be a criterion:
A widely agreed-upon standard for judging politicians is honesty.
It's enough merely to say that honesty is a standard for judging
politicians. With that, I can say "I claim that politician X
meets the honesty standard better than politician Y, and I believe
that you'll agree when I tell you what I know about X & Y."
True, I can't prove, in most cases, that X is more honest than Y,
since they could be dishonest in different ways, impossible to
compare quantitatively. But I could still show you that, as you
& I perceive dishonesty, X is less dishonest than Y.
That's useful, and that kind of a "vague" standard is also useful
in comparing voting systems. "Vague" is a perjorative word which,
it seems to me, reflects the mathematician's prejudice against
what can't be quantitatively proved. Some of my statements of
my standards aren't criteria, yes/no tests, but are still useful.
I've, as I said, also stated some y/n tests for measuring compliance
with my standards, and those yes/no tests, criteria, can also be
called "standards" too.
Of course we could write a criterion for honesty: A politician shouldn't
be so dishonest that he'd steal from his own mother. With that, we
can say that a particular politician does meet that honesty criterion,
that he doesn't, or that sufficient information isn't available to
say one way or the other. Obviously the politician who fails that
criterion is easier to identify than the one who meets it.
Criteria are useful, but so are standards exressed other than as