status: https://faculty.washington.edu/kerim/nomic/cases/#3903
(This document is informational only and contains no game actions).
=============================== CFJ 3903 ===============================
R. Lee's votes on the referendums on proposals 8549 and 8552-8555
were clearly specified.
==========================================================================
Caller: Jason
Judge: G.
Judgement: TRUE
==========================================================================
History:
Called by Jason: 29 Mar 2021 04:08:30
Assigned to G.: 04 Apr 2021 00:32:16
Judged TRUE by G.: 07 Apr 2021 18:47:01
==========================================================================
Caller's arguments:
On the relevant decisions, R. Lee voted "I vote AGAINST all agoran
decisions that add overall text to the ruleset, and PRESENT on the
rest." This is clearly an attempted conditional vote. The question is
whether it is clearly specified under Rule 2127.
Per Rule 2127, they are clearly specified if and only if they were
determinate at the end of the voting period, and they are determinate if
the values can be "reasonably determined [...] from information
reasonably available" (the other possibilities clearly do not apply).
The information that this depends on is certainly "reasonably
available", as it is just the text of the proposals as well as the text
of the rules (assuming we are correct in our understanding of the text
of the rules, which is hopefully the case). This means that this case
depends on whether the vote can be "reasonably determined" from that
information.
In general, evaluating proposals to determine whether they add or remove
text is arguably not "reasonably determined", as it involves evaluating
potentially arbitrary text. However, real distributions do not contain
proposals with arbitrary text, and the proposals in this specific
distribution are, individually, relatively easy to count characters
added/removed on. Perhaps that's sufficient for this specific case to be
considered "reasonable effort", but that's potentially opening a can of
worms for large, perhaps unreasonable, workloads for the Assessor.
Caller's evidence:
On Wed, Mar 24, 2021 at 5:16 PM N. S. via agora-business
<agora-busin...@agoranomic.org> wrote:
>
> I vote AGAINST all agoran decisions that add overall text to the ruleset,
> and PRESENT on the rest.
Rule 2127/11 (Power=3)
Conditional Votes
A conditional vote on an Agoran decision is a vote which indicates
a vote based on some condition(s). A conditional vote is evaluated
at the end of the voting period and, rules to the contrary
notwithstanding, is clearly specified if and only if the value of
the condition(s) is/are determinate at the end of the voting
period. If the conditional is clearly specified, and evaluates to
a valid vote, it is counted as that vote; otherwise, it is counted
as PRESENT.
Any vote which is clearly expressed as a conditional, e.g. "FOR if
<X> is true, AGAINST otherwise", is a valid conditional vote that
evaluates as specified.
A vote endorsing another person is equivalent to a conditional
vote evaluating to the vote specified in that person's valid
ballot on the decision, if any.
For an instant runoff decision, a vote consisting of a list, one
or more entries of which are valid conditional votes, and the
remaining entries of which are valid options, is a valid
conditional vote. Such a vote is evaluated by evaluating each
conditional entry to a list of votes (or an empty list, if it
evaluates to PRESENT either directly or indirectly), and then
concatenating those lists with the specified valid options in the
order they occurred in the original vote.
Rule 2518/0 (Power=3)
Determinacy
If a value CANNOT be reasonably determined (without circularity or
paradox) from information reasonably available, or if it
alternates indefinitely between values, then the value is
considered to be indeterminate, otherwise it is determinate.
--------------------------------------------------------------------------
Judge G.'s Arguments:
In the past, we've accepted many complicated conditionals. In principle,
I don't see why "does this specified proposal add text to ruleset" is more
complicated or error-prone than, say "I vote for the candidate who
transfers me the most coins" in terms of difficulty to the vote collector.
For the latter, for example, the vote collector would need to track any
coin transfer, look at recent reports to know if the transfers succeeded,
and maybe even (if the transfers had previous transfer dependencies) track
more. And errors would be made. Yet I suspect we would accept such a
conditional without thinking about it too much and the poor officer would
be on the hook.
So I don't think we can throw out proposal parsing for conditionals just
because they're proposals or rule changes, or else if we do, we'd throw
out many conditionals that we've accepted in the past.
As an example, looking at one of the proposals in question (8549), the
proposal is of the form "insert after [text] the following [other text]".
I looked up the rule, checked the power, and knew that the proposal in
question would increase the words in any of the rulesets. 3-5 minutes,
and fewer references to check than the coin thing. And probably less
prone to error than the coin thing.
So in principle, proposals aren't special. Some are harder to interpret
than others in the name of interpreting conditionals, That's evaluated
case-by-case - it doesn't force evaluation of "arbitrary" text - some
proposals, just like some conditionals, could be "too hard", but it's
case-by-case).
Now, some of the ones in this batch may be too difficult - but the caller
notes that they're not. Having looked over them for 5 minutes, I conclude
similarly - these are not difficult tasks. TRUE.
I'll note that what's easy in the individual is often difficult in the
aggregate, but that's a matter of social restraint. Doing too many
blanket conditionals is similar to submitting lots of proposals and not
pending them, or calling and retracting lots of cfjs, or whatever. I
don't think a blanket cfj-ban based on the reasonableness of each one (and
these in particular are reasonable) is an appropriate way of handling
that, if it becomes an actual issue.
However, I'll futher note that, given the nature of cumulative effort, the
effort for each individual conditional needs to be truly fairly small.
The standard I'd suggest is that the conditional must be clear and obvious
within a minute or two of checking on facts. So a proposal that
repeals a rule (requiring a word count), or has more than one add or
subtract clause so character count had to be tracked, would be too much.
==========================================================================
