On Fri, 7 Nov 2008, Alex Smith wrote:
> You've parsed the rules as
> (rulechange if permittedbyrules), (rulechange if permittedbyrules),
> (monsterrulechange if permittedbyrules); the correct parsing is
> (rulechange if permittedbyrules), (rulechange if permittedbyrules),
> (monsterrulechange if (rulechange if permittedbyrules)).
You've presented a reasonable parsing (though not one I agree with, I stand
by my original). But I'll add an argument that I find strong even if we
accept your premise (consider it a gratuitous argument, in case of appeal
which I don't support, for AFFIRM with a possible error rating).
Given: We accept that the parsing of R2192 is generally (monsterrulechange
if (rulechange if permittedbyrules)).
Question: Is permittedbyrules for R2193 TRUE?
1. By R2141, R2193 *generally may* change rules. However, I've argued
previously that this broad definition says that we should use "may" here
as a broad "possible to" not "permitted to". In other words, it means
that rule changes are the *kind of thing* that a rule *might* do, and
not something that all rules automatically MAY do. (This is my R2152
"weighing of the full implications" of not considering "may" in 2141 to
be MAY). Also, importantly, the inverse of a broad and general "may"
is a broad and general "may not", and this broad "may not" is true if
MAY NOT is true *or* of CANNOT is true. As Murphy pointed out, the term
"generally" is what we assume is the default, but defer to other rules
(even lower-powered ones) if they override the default. Here, either
MAY NOT or CANNOT "generally" prohibits or stops the action, so either
MAY NOT or CANNOT makes the "generally may" into something that is no-
longer assumed, in other words, into a general "may not".
2. By R105, R2193 CANNOT currently, specifically change Rules. Rule
changes can only happen *where permitted*, and this argues that rule
changes must be contained and authorized in a specific location and
way within the rules, e.g. within the rules text somewhere. To make
a rule change where not *specifically* permitted is against R105, therefore
is a CANNOT (and possibly a MAY NOT, again the lower-case can in R105
leaves the exact disposition questionable, but it is not out of context
to treat this lower-case can fairly strongly. But also, as per 1 above,
CANNOT inferred here translates to "may not" in R2141 so we don't need
to argue this point).
3. So R2193 generally may (may as in "might" or "is potentially capable
of") change the rules, but specifically, as the rules are written now,
CANNOT change the Rules.
4. So, the question is, does the "may" in "The Mad Scientist CAN act on
behalf of the Monster to take any action that the Monster may take..."
mean the general, potentially-capable "may" (leading to false as the
R105 CANNOT leads to a R2142 "may not") or a specific, direct, MAY
(leading to possible true iff R105 can is interpreted as CANNOT but not
MAY NOT)?
a. I've presented thorough arguments, based on the broad and general
quality of the rule, for *not* interpreting R2141's may as MAY.
In Rule 2192, it is conspicuous that CAN and SHALL are capitalized
but may is not, showing that this rules was MMI-aware, but
didn't use the "MAY" (exceptio probat regulam). This argues for
interpreting the same general "may" here as in R2141, and as I
discuss above, this generic, non-MMI "may" can be blocked (turned
in to "may not") by a MAY NOT *or* a CANNOT. So by R105, "may" and
therefore "permittedbyrules" is currently false.
b. It is a strong custom, and part of the rules, to interpret rules
specifically as they exist now. When speaking of capability or
permissibility, we do not treat a judgement on "It is possible to
do X" as trivially true based on the argument "it is always possible
to change to rules so as to do X." (I believe there was some
precedent on this a while back but can't recall where). Rather, we
ask about X assuming the Rules are fixed at the time of the CFJ, and
for that may = MAY and CAN = currently false.
-Goethe
