============================  Appeal 2276a  ============================


Panelist:                               Elysion
Decision:                               OVERRULE/FALSE


Panelist:                               root
Decision:                               OVERRULE/FALSE


Panelist:                               ais523
Decision:                               OVERRULE/FALSE

========================================================================

History:

Appeal initiated:                       20 Nov 2008 16:46:59 GMT
Assigned to Elysion (panelist):         25 Nov 2008 20:35:05 GMT
Assigned to root (panelist):            25 Nov 2008 20:35:05 GMT
Assigned to ais523 (panelist):          25 Nov 2008 20:35:05 GMT
ais523 moves to OVERRULE/FALSE:         30 Nov 2008 17:50:00 GMT
Elysion moves to OVERRULE/FALSE:        04 Dec 2008 22:23:19 GMT
root moves to OVERRULE/FALSE:           12 Dec 2008 06:46:30 GMT
Final decision (OVERRULE/FALSE):        12 Dec 2008 10:35:48 GMT

========================================================================

Gratuitous Arguments by ais523:

I don't agree with either line of argument, actually. In my
interpretation of what happens, "is" specifies a value for the voting
limit; the Notes rule specifies a way to increase it, which takes
precedence, and therefore sets its own value, which is 1 more than it
would have been otherwise. Caste says the voting limit is 1: Notes says
the voting limit is 1+50, and Notes wins. For something which is
platonically set with an "is", an increase clearly also platonically
sets it to a higher value than its previous platonic value; as the value
isn't a floating property like a switch, there's no other interpretation
of "increase" that makes sense.

========================================================================

Gratuitous Arguments by Murphy:

I recommend REMAND with instructions to explicitly evaluate the two
competing interpretations:

  +S) 2126 takes precedence, so 2156 implicitly defines the initial
      limit and 2126's increases stick.

  -S) Even though 2126 takes precedence, 2126 only attempts to operate
      once and 2156 attempts to operate conditionally, so 2126's
      increases happen but 2156 comes along afterward and resets things.

========================================================================

Gratuitous Arguments by omd:

An increase that remains in effect for an infinitesimal amount of time
is no increase at all.

========================================================================

Gratuitous Arguments by G.:

R2156, in defining that a voting limit *is* a caste level (as opposed
to saying "is set to" a caste level) constitutes a whole and complete
definition of a particular voting limit.  As a definition, R2156 governs
the properties of the voting limit that are possible to exist (R1586)--
it is no more possible to set a voting level to deviate from a player's
caste than it would be possible to set a voting limit to be a chunk of
green cheese.  Such a definition constitutes an implicit claim of
precedence for the purposes of R1030.  Since R2156 and R2126 are of the
same power, and R2126 does not itself claim to have greater precedence,
this claim should be sufficient to establish R2156's authority in the
matter.

========================================================================

Gratuitous Arguments by ais523:

This actually argues in comex's favour. It is indeed possible to set a
voting limit to a different value. This does not create an implicit
claim of precedence for R2156 though; far from it! It means that R2126
has the claim of precedence, as it's claiming to do something R2156
claims is impossible. Imagine R1 saying "Goethe CAN deregister by paying
1 Stem" and R2 saying "Goethe CANNOT deregister"; R1 clearly takes
precedence here, due to the direct conflict. If it had been R1 saying
"Goethe CAN deregister by paying 1 Stem" and R2 saying "Every second,
Goethe becomes unable to deregister", then it's no longer so clear;
there might not be a conflict due to deregistration-ability becoming
some sort of flip-state there. When there's an obvious direct conflict
between two rules, as there is here, there isn't an implicit claim of
precedence on either; there's an explicit clash, and the more powerful
rule, or the rule with the lower number, takes precedence.

========================================================================

Gratuitous Arguments by ais523:

On Thu, 2008-11-20 at 13:04 -0800, Kerim Aydin wrote:
> An important missing piece in your argument is that I used R1586 to
> argue for definitional preference. R2 is actually in conflict with R1586
> as well as R1.  In particular from R1586:
>                                                          "then that
>       entity and its properties continue to exist to whatever extent
>       is possible under the new definitions."
>          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> It is not possible for the Voting Index to be Green Cheese under its
> definition as a number.  As R2 would allow me to set it against the R1
> "new definitions", R2 is in conflict with R1586 (and in the current case,
> by numerical precedence, R1586 wins).
This argument fails utterly: quote the first half of that sentence
{{{
      If the documents defining an entity are amended such that they
      still define that entity but with different properties,
}}}
and the then clause fails to apply. No way is spending a Note to
increase voting limit /amending/ the rules that define voting limit.

Also, you seem to have numerical precedence backwards, although I don't
think it affects your argument.

> Finally, we're *both* making implicit claims.  You're claiming the fact
> that someone CAN set a numerical index to Green Cheese is an implicit
> definition that the numerical index has a defined green cheese state.
> I'm saying that the definition implicitly claims precedence (which only
> works at the same Power to override numerical precedence) and forbids
> a green cheese state under R1586 as it is not possible under the new
> definitions.  I still find my implicit claim far less of a stretch than
> yours.
R1030:
{{{
      If two or more Rules with the same Power conflict with one
      another, then the Rule with the lower ID number takes
      precedence.

      If at least one of the Rules in conflict explicitly says of
      itself that it defers to another Rule (or type of Rule) or
      takes precedence over another Rule (or type of Rule), then such
      provisions shall supercede the numerical method for determining
      precedence.

      If all of the Rules in conflict explicitly say that their
      precedence relations are determined by some other Rule for
      determining precedence relations, then the determinations of
      the precedence-determining Rule shall supercede the numerical
      method for determining precedence.

      If two or more Rules claim to take precedence over one another
      or defer to one another, then the numerical method again
      governs.
}}}

Implicit claims don't affect precedence; R1030 explicitly requires
precedence claims to be explicit. There is definitely no comparison of
strength of implicit claims! If there is a claim on both sides, or a
claim on neither side, go numerical; and R2126 beats R2156, with R1586
being irrelevant.

========================================================================

Gratuitous Arguments by omd:

Several interpretations have been thrown out, including (extended from
Murphy's post):

 +S) 2126 takes precedence, so 2156 implicitly defines the initial
     limit and 2126's increases stick.

 -S) Even though 2126 takes precedence, 2126 only attempts to
     operate once and 2156 attempts to operate conditionally, so
     2126's increases happen but 2156 comes along afterward and
     resets things.

 +G) Even though 2126 takes precedence, it defers to Rule 2156
     because the latter is defining a term used by the former (and
     Rule 754 takes precedence over both).

 -G) Rule 2156 defines voting limit as caste, so an attempt to
     increase voting limit is in fact a failed attempt to increase
     caste.

 +A) Rule 2126's "increase" should be interpreted such that upon an
     increase, Rule 2126 defines the voting limit as one higher than
     it would otherwise be, in the fashion of an RPG.

Wooble's judgement agreed with -S, but e gave the issue only a
perfunctory treatment.

========================================================================

Gratuitous Arguments by Murphy:

-G') Rule 2156 defines voting limit as caste, so an attempt to increase
     voting limit on one proposal alone is an attempt to simultaneously
     change and not-change caste, thus ineffective.

========================================================================

Gratuitous Arguments by ais523:

A very quick summary of the arguments (I'll flesh them out
later):

+G No rule 754 doesn't say that, and anyway you didn't
consider the effect of rule 683 in this; rule 754 merely
says everything's talking about the same thing.

-G/-G': Voting limit isn't caste now, but caste at the
start of the voting period. Trying to change that
retroactively isn't something that can sanely be implied
into a rule, especially given the "by default" in rule
754.

+S/-S: The rules don't try to define a switch here, we
shouldn't imply one into them unless absolutely necessary.
(Actually, rule 2156 is ambiguous, and the alternative
interpretation makes it clear that +S is correct, but
nobody's suggested that interpretation before and it would
be against game custom.)

+A: The only correct one, see my earlier arguments for now,
I'll explain in more detail later.

========================================================================

Panelist ais523's Arguments:

There is only one interpretation of rule 2126, 2156, 683
and 754 that does not lead to a contradiction, which obeys
the letter of all the rules in question, which doesn't go
strongly against game custom or rely on an unintended
ambiguity; that is, that rule 683 sees ehird's voting
limit as 1 (per rule 2156) increased by 50 (per rule 2126),
and thus as 51. Therefore, the AFO's assessment of proposal
5956 was correct, and it was adopted. This panel hereby
rules OVERRULE/FALSE on CFJ 2276a.

========================================================================

Gratuitous Arguments by ais523:

The really controversial one is CFJ 2276a. So far, 6 possible
interpretations of what happened have been suggested, which I repeat
here for reference:

 +S) 2126 takes precedence, so 2156 implicitly defines the initial
     limit and 2126's increases stick.

 -S) Even though 2126 takes precedence, 2126 only attempts to
     operate once and 2156 attempts to operate conditionally, so
     2126's increases happen but 2156 comes along afterward and
     resets things.

 +G) Even though 2126 takes precedence, it defers to Rule 2156
     because the latter is defining a term used by the former (and
     Rule 754 takes precedence over both).

 -G) Rule 2156 defines voting limit as caste, so an attempt to
     increase voting limit is in fact a failed attempt to increase
     caste.

 +A) Rule 2126's "increase" should be interpreted such that upon an
     increase, Rule 2126 defines the voting limit as one higher than
     it would otherwise be, in the fashion of an RPG.

 -G') Rule 2156 defines voting limit as caste, so an attempt to increase
     voting limit on one proposal alone is an attempt to simultaneously
     change and not-change caste, thus ineffective.

Here are the relevant excerpts of the relevant rules:

Rule 754(2):
{{{
       (2) A term explicitly defined by the Rules by default has that
          meaning when used in any Rule of equal or lesser power, as
          do its ordinary-language synonyms not explicitly defined by
          the rules.
}}}
Excerpt from rule 683:
{{{
      The voting limit of an eligible voter on an Agoran
      decision is one, except where rules say otherwise.
}}}
Rule 2126(7):
{{{
      (7) A player CAN spend one Note to increase another player's
          voting limit on an ordinary proposal whose voting period is
          in progress by 1.
}}}
Excerpt from rule 2156:
{{{
      The
      voting limit of an eligible voter on an ordinary decision is eir
      caste at the start of its voting period, or half that (rounded
      up) if the voter was in the chokey at that time.
}}}

I want to cover the +G interpretation first, as it was the subject of an
extensive argument between me and Goethe. The argument here is that rule
754 makes definitions take precedence over actions. First, I want to
take rule 683 into account. Unarguably, it is rule 683 that defines
voting limit; it's defined as 1 except where rules say otherwise. In
this case, only rules 2126 and 2156 say otherwise, so they're the
relevant rules. Goethe claims that rule 754(2) means that definitions
take precedence over actions; but it says no such thing. (The "by
default" is a big clue against, but even without that, all the rule is
saying is that text defined by the rules can be referenced by the rules,
obvious but it's relevant because it's establishing a precedence order
of definitions.) So rule 754(2) establishes that "voting limit" means
the same things in rules 683, 2126, and 2157; but this has not been
disputed. (Goethe also suggested the possibility that rule 2156 defines
a switch with only one possible value; I'll consider this later along
with -S, which is similar but different.)

I'd like to rebut -G and -G' next. The relevant phrase here is "at the
start of its voting period"; this pretty clearly establishes that voting
limit and caste are different things. If it had said "The voting
limit ... is eir caste", then it would be unclear whether it was
defining a synonym for caste (what these interpretations rely upon) or
something different; even then, though, the existence of rule 683
implies that voting limit is something separate. The interpretation that
-G and -G' put on rule 2126 in particular is absurd, though; it would
mean that people could spend notes to /retroactively/ change eir Caste
at the start of the voting period. Implying retroactive behaviour into a
rule that doesn't ostensibly act retroactively is something highly
against game custom, and in this case no rule suggests that that is what
is happening; rule 683 makes it pretty clear that the value of voting
limit is not the same thing as caste is, even though rule 2156 defines
voting limit to have the same value as caste.

+S and -S both take the argument that rule 2156 defines some sort of
implicit switch, differing about how exactly this switch works. First, I
would like to point that rule 2156 is ambiguous: if it's read as "At the
start of its voting period, the voting limit of an ordinary voter on an
eligible decision is eir caste...", then +S is pretty obviously the
correct definition with a relatively obvious implicit switch. However,
this reading isn't the one that most players have come up with reading
the rule, and thus game custom and rule 217 would imply that it isn't
the reading that should be taken. (Note that this CFJ leads to the same
result but different reasoning if it is, by the way.) I would favour
this reading of the rule if it were the only reading consistent with the
older rule 2126, but as it happens, it isn't.

With the intended reading of rule 2156, we have to consider whether it
does in fact define a switch. It's been argued that it's a switch with
only one possible value, a switch that is changeable by the rules at
will, or some sort of elastic switch which is platonically set to a
particular value at every opportunity; however, none of these
definitions seem reasonable; the ambiguity that has blown up here is
good evidence of this. Still better evidence, though, is rule 683
itself, which defines voting limit but clearly doesn't define a switch.
Also, we should be wary of implying switches into the rules where they
clearly don't imply there is one; many things specified in much the same
way, like pledgeness, are not switches but instantaneously evaluated
things. I think the main reason that +S and -S were suggested was that
there were no clear alternatives suggested at the time; when, as in this
case, the rules do not define a switch, it is clearly preferable to
interpret them as not defining a switch.

This leaves us with +A; I believe this is the correct interpretation, as
it's the only interpretation which fits the letter of the rules and game
custom. We have -G and -G' which imply synonyms that aren't there and
retroactive changes; +G which relies on a perverse reading of rule 754,
and also ignores rule 683; and +S and -S which define a switch with
unknown parameters which isn't mentioned by the rules; none of these
fits the literal definition of what the rules say. +A does, and also
fits both game custom, and the spirit of the rules in question (this
last is a bonus, but obviously, and especially where scams are
concerned, the letter of the rules in question takes precedence over
their spirit as long as it's unambiguous); it also does not rely on any
precedence resolution mechanism, as none of the rules contradict each
other in this interpretation (Power Controls Mutability is relevant, but
only insofar as it requires me to point out that rule 683 is the rule
defining voting limit here but it's explicitly looking at other rules
and using their subordinate definitions; again, uncontroversial I think,
although some people missed it during the initial arguments). The
reasoning goes like this:

Rule 683 says (in this case) that voting limit is according to what
rules say.

Rule 2156 says that voting limit (in this case) is equal to caste at the
start of the voting period (possibly halved).

Rule 2126 says that voting limit on a particular proposal CAN be
increased by spending notes (i.e. attempts to increase it by spending
notes succeed).

So suppose there's an attempt to increase voting limit on a particular
proposal. Rules 2126 and 2156 don't do anything directly to the voting
limit; they only matter because rule 683 looks at them to see what
happens. When rule 683 looks up the voting limit, what does it see? It
sees that ehird's voting limit is 1, increased by one 50 times, i.e. 51.
Although this definition of "increase" is different from the definition
used in "increase caste", it is a perfectly common English definition;
in fact, similar definitions are used elsewhere in the rules ("lowered"
in rule 1668, "decreases" in rule 2019). Importantly, there are no
contradictions anywhere here, each rule is working as expected and
exactly as it literally states, and there are no contradictions
anywhere.

I don't know of any alternative interpretations for what happened; but I
have yet to see a plausible argument as to why +A is wrong. (No, Goethe,
rule 754 does not say what you think it says, and even if it did it
wouldn't matter.)

========================================================================

Panelist Elysion's Arguments:

I support this. Referencing rule 683 is what won me over, BTW.

========================================================================

Panelist root's Arguments:

[support actually given by CotC Murphy]

========================================================================