Index ← 3792 CFJ 3793 3794 → text
===============================  CFJ 3793  ===============================

      Rance’s master switch is set to Gaelan.

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

Caller:                        Gaelan

Judge:                         Alexis
Judgement:                     FALSE

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

History:

Called by Gaelan:                                 16 Jan 2020 06:16:40
Assigned to twg:                                  22 Jan 2020 15:51:57
Judged FALSE by twg:                              26 Jan 2020 12:12:24
Motion to Reconsider filed:                       26 Jan 2020 20:51:37
twg recuses emself:                               07 Feb 2020 00:19:21
Assigned to Alexis:                               08 Feb 2020 21:20:45
Judged FALSE by Alexis:                           15 Feb 2020 23:08:23

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

Caller's Arguments:

1885/9: For the purpose
     of such a auction, to transfer a zombie to a player is to set that
     zombie's master switch to that player.

2551/3: When [the bid is paid to the auctioneer],
     if the auctioneer CAN transfer the items in that lot to that
     winner at will, e immediately does so; otherwise, e SHALL do so in
     a timely fashion.

So, the question is: CAN Agora, as an autonomous entity, set a switch? I 
don’t see any language that implies it can, but perhaps there’s an 
argument to be made that Agora obviously is able to change its own state?

--------------------------------------------------------------------------

Gratuitous Arguments by G.:

The standing precedents are CFJs 3693-3694.

--------------------------------------------------------------------------

Gratuitous Arguments by Falsifian:

Jason Cobb pointed out Gaelan's 22-coin bid was NttPF [0], so e did not 
win any lots in the auction, so this should be FALSE.

- Falsifian

[0] https://mailman.agoranomic.org/cgi-bin/mailman/private/agora-
business/2020-January/041890.html

--------------------------------------------------------------------------

Judge twg's Arguments (prior to Motion to Reconsider):

Falsifian and Jason Cobb are perfectly correct; Gaelan's final bid on the
zombie auction in question was not among the three highest, so e did not
win any lot. (Because of eir phrasing, though, e did still transfer the
amount of eir "bid" to Agora. Sucks to be em.)

I judge this CFJ trivially FALSE.

However, it's worth considering the matter that the caller _intended_ to
inquire into. Is it generally POSSIBLE for players to flip master switches
as the result of a zombie auction?

As G. correctly points out, this exact question was the subject of CFJ
3694. Despite the complexity of the discussion surrounding that case, e
has kindly supplied us with the reasoning adopted by CFJ 3694's judge:

> The CFJ is FALSE, because a necessary implication of the Rules is that
> Agora CAN transfer a zombie pursuant to a properly initiated Rule 1885
> zombie auction.
>
> Under Rule 2545 (power=2), “An Auction is a way for entities to give
> away items in exchange for a currency.”  By necessarily implication,
> if a Rule with high enough power authorizes some player to initiate an
> auction with a particular item as a lot, then that Rule also
> necessarily authorizes that player to initiate a process that would
> “give away [that] item[] in exchange for a currency.”  That’s what an
> auction _is_ under the Rules.
>
> Under Rule 1885 (power=2), at the start of the month, “the Registrar
> CAN put [a] zombie [that meets certain conditions] (along with any
> other zombies that fulfill the same conditions) up for auction.”  By
> necessary implication, read in conjunction with Rule 2545, that means
> that a zombie can be transferred pursuant to that auction. Otherwise
> it wouldn’t be an auction at all—a way of transferring an item for
> currency. It would be something else entirely.
>
> This interpretation is consistent with the best interests of the game
> and ordinary language. There’s no reason to adopt a contrary
> interpretation, which would break auctions and zombies.

Although unconventional, I do not find flaw with the judge's logic. By
defining an "Auction" as "a way for entities to give away items" (text
that is still present in the rule today), Rule 2545 ("Auctions")
effectively makes a blanket statement that entities CAN give away items
through auctions. Thus, the condition in Rule 2551 ("Auction End") that
"the auctioneer CAN transfer the items in that lot to that winner at will"
is always true (provided that the items are possessed by the auctioneer,
which is always the case with official zombie auctions).

However, it should be noted that the result of this wording is clearly
unintentional, as it might for example permit a player to auction even
_fixed_ assets, including eir own zombie(s), by means of Rule 2584 ("Free
Auctions"). This court therefore recommends that language be introduced to
Rule 2551 explicitly enabling Agora to transfer zombies, and that Rule
2545 be rephrased to remove its unfortunate permissiveness.

--------------------------------------------------------------------------

Judge Alexis's Arguments:

## History

For a brief recap of this case, the fundamental question here is
whether, upon paying the fee to win a zombie auction, Rance's master
switch was actually set to Gaelan.

This case has gone through a rich history and varied arguments. The
issue at the heart of this case was raised before in CFJs 3693 and
3694, with CFJ 3693 having been withdrawn due to a technical flaw and
CFJ 3694 representing the substantive judgement on the question "It is
generally IMPOSSIBLE for a zombie to be transferred to the winner of a
lot in a zombie auction."

CFJ 3793 was assigned to twg, who found that this case was also
trivially false as Gaelan was not one of the three highest bids in the
auction in question, so the case must clearly have the value FALSE. I
find no reason to disagree. Therefore, I judge CFJ 3793 FALSE and
award myself blue glitter.

However, Judge twg went into obiter on what the correct judgment would
be notwithstanding the trivial error, and found, based on the
precedent of CFJ 3794, that indeed the zombie generally could be
transferred.

In argument after the judgement, I raised Rule 2125, and in particular
its requirement that "A Regulated Action CAN only be performed as
described by the Rules, and only using the methods explicitly
specified in the Rules for performing the given action." The judgment
in CFJ 3794 held that Rule 2545's text "An Auction is a way for
entities to give away items in exchange for a currency." meant that,
by necessary implication, it must be capable of effecting that
transfer. Consequently, Rule 1885's provisions that a zombie auction
can be initiated by the Registrar create a necessary implication that
there exists a mechanism to transfer zombies in the auction. Judge D.
Margaux also pointed out that this was consistent with the best
interests of the game and ordinary language. However, I argued that
this did not meet the standard of an explicitly specified method from
Rule 2125.

As a result of my argument, a Motion to Reconsider was filed, and a
few additional arguments were given. twg argued that the method of an
auction was specified explicitly, though perhaps not entirely clearly.
E argued that R2545's text, quoted above, explicitly states that the
auction provides a method, noting that "way" is synonymous to
"method". G. did note that "clear" and "unambiguous" are not the same,
citing CFJ 3659, but pointed out also that the word "explicit" "seems
to embody clarity". E later gave an argument that the rules are
explicit that a zombie auction is the method of executing a transfer
of assets, even if the internal mechanisms of the auction rules are
broken. E compared to the dependent action rules, whereby a definition
that something can be done "without objection" is explicit regardless
of whether or not the dependent action rules are broken or not.

I had thought that I had submitted additional argument that the text
in R2545 could be interpreted as merely being generally explanatory,
but I cannot seem to find such argument.

## The Function of Rule 2125

So in order to address this case, we must remind ourselves of the
provisions of Rule 217. The interpretation of the rules is guided
first by their text, followed by game custom, common sense, past
judgements, and the good of the game to be applied where the text of
the rules is silent, inconsistent, or unclear. I find that, in the
context of zombie auctions specifically, every one of these factors
points to them functioning and allowing zombies to be transferred.
However, because the text of the rules takes precedence, it must be
considered first.

There is an interesting conundrum here, as there is a question of
where in this sequence to consider whether the auction rules are
"explicit". Do we evaluate the text of all the rules in a single step,
then proceed to consider the other R217 factors on the rules as a
whole, or do we do a two-phase process whereby we first attempt to
learn what "explicit" means, then apply its text to the auction rules?
This arguably creates a circular instance of the same problem, applied
to R217 itself, however, I believe that it can be short-circuited by
observing that it would very much not be in keeping with common sense
or the best interests of the game for R217 to have anything other than
uniform application across all cases, for its very existence would be
undermined. Its interpretation could change based on the nature of the
situation to which it is applied, including on rules of lower power,
something it expressly considers to be problematic.

It follows by similar logic, and by common sense, that R217 applies
similarly to other interpretive rules. Thus, when dealing with rules
which affect or are affected by the interpretation of other rules, we
must apply the R217 factors to those rules independently of the R217
factors of the rule(s) to which the interpretation is being applied.
Once the interpretive rule is correctly interpreted, it can be applied
to the text of the subject rule(s), and at this stage the R217 factors
will again kick in.

There is one other aspect of this to consider, which is that R2125
certainly has an interpretive aspect in terms of identifying whether
the rules are "explicit". But it derives its nature from a principle
of statutory construction that that which is not included is
explicitly excluded, and thereby from the game custom prohibiting
ISIDTID. Is R2125 itself an interpretive principle, affecting the
conclusions we draw from the rules as a whole, or is it a substantive
rule which exists as merely one element of the system? This is not an
academic question, because of the way that precedence interacts. If
R2125 is an interpretive principle, its precedence conflicts would
only be with other interpretive rules. By contrast, if it is a
substantive rules, it does not affect the interpretation of other
rules directly, but only depends on them, and its precedence conflicts
would be with rules which permit actions to be performed.

I would like to say that R2125's prohibition against performing
regulated actions outside the methods specified in the rules is an
interpretive principle. There are clear hallmarks that it intends to
be one, including its application to the entire ruleset, its lack of a
precedence claim over other power-3 rules, and the following sentence
that explicitly addresses interpretation of the rules. What about the
text, though? Is it silent, inconsistent, or unclear enough to admit
this interpretation? I find that it is, albeit just barely. The
wording "CAN only" is not quite that of a usual prohibition, and seems
more akin to a statement of a higher-level principle: it is intended
to prevent the absurdity of a situation where, effectively, the rules
amount to a nonconstructive proof that an action could be performed by
providing no mechanism. This could only properly work as an
interpretive principle, as in the second paragraph of R217.

On the other hand, the rule could be desugared, following Rule 2152's
definitions, to "Attempts to perform Regulated Actions are
unsuccessful unless...". I believe that this would, fairly clearly, be
a substantive prohibition. But R217 specifically refers to the text of
the rules, not to the text after substituting other definitions in, so
I find that such a substitution cannot eliminate the inherent
unclarity in the text of the rules by changing its grammatical
structure. Even if R2152 were interpreted as trying to modify this
principle---perhaps a bit of a stretch---R217 takes precedence to the
extent of a conflict. This is consistent with past judgments on R2152
involving unusual capitalized terms, such as CFJ 2282, though I could
not find any that explicitly framed it in terms of a precedence
conflict with R217. I should note that this is not a general statement
that R217 takes precedence over the semantic content of definitions.
In fact, this would conflict with its own claim that the text of the
rules, which would necessarily include definitions, takes precedence.
Rather, it only prevents looking at artificially created texts, based
on definitions, in order modify grammatical structure or other
linguistic properties to eliminate the unclarity that exists on the
face of the text. The semantic content of definitions is deferred to,
effectively, by the very fact that those definitions are part of the
text of the rules and therefore must be applied before resorting to
the secondary interpretation criteria.

At this point, the factors clearly tilt to R2125 being an interpretive
principle. Surprisingly, despite much case law on R2125, I could not
find a single instance where it was argued that a rule that failed to
specify a method for an action to be performed simply took precedence
over R2125 by way of having a lower ID number, and thereby made the
action possible but not actually performable. But ISIDTID is a
long-standing game custom, or rather non-custom and it clearly
benefits the good of the game for it to remain as such, codified in
R2125. I'm not sure where common sense lies, honestly, but I know that
it doesn't outweigh the decades of custom.

## General Application of Rule 2125

Now that we have concluded that Rule 2125 is an interpretive
principle, and not a substantive rule, we must figure out how to apply
it generally. It is very clearly a principle of interpretation of the
rules of the whole, and not of any single rule in particular. This is
evident directly from its text, but it is also worth noting that if it
applied to individual rules, then any rule which authorized an action
would thereby prohibit lower-powered rules from also authorizing it.
That would have significant ramifications.

Because its interpretation is a matter of the rules as a whole, not
any individual rule, we must also evaluate the explicitness of the
rules as a whole, rather than only looking at rules in isolation. It
provides that, where no method is explicitly available within the
rules, then the correct interpretation is that there is no method
available, rather than inferring a specific method or entertaining a
weird non-constructive situation where something is possible but
cannot actually be done. The latter part is relevant when rules care
about whether or not something is possible. Under R2125, something
simply is not possible without an associated method.

I think that G.'s example of a dependent action when the dependent
action rules are broken provide a good example of this. If I
understand eir argument correctly, if something caused dependent
actions not to be performable, then a dependent action specification
such as "without objection" would meet R2125's explicitness
requirement, because "without objection" is a method explicitly
specified, and therefore it would be inapplicable in the context.

While I think that R2125's focus on the rules in general, rather than
individual rules, is enough of an answer to this suggestion, also
consider the practical consequences. With dependent actions broken, a
specification such as "without objection" desugars into a broken
definition, and then the result is that a rule specifies that the
thing can be done, but the rules as a whole actually lack a mechanism.
We're now in exactly the situation that R2125 is supposed to prevent.
Restricting its application based on the specific structure of rules
involved would undermine its purpose.

So what exactly is a "method" as contemplated by R2125? Similar to the
above, consider if a rule stated that an action "CAN be performed by
Assembling a Contraption, where Assembling a Contraption is as defined
by other rules". Without such a definition provided by another rule,
it seems hard to consider this as much other than "CAN be performed"
without qualification. The rule asserts that it is possible but leaves
too many details unspecified. We know based on game custom and
precedent that R2125 functions to make sure that "CAN be performed",
without any qualification, does not alone make something possible. It
should equally apply to situations where a method exists but is left
undefined.

Thus, we can conclude that a "method explicitly specified in the
Rules" is a procedure by which the rules specify that an action
contemplated can actually be realized. It may be indirect, occurring
as a result of another game action, and it may be a wilful natural
action of a sentient being that starts things, or some other
occurrence such as the passage of time. But the process cannot be cut
from whole cloth. Nor can a significant omitted portion of the
procedure be inferred, because of the requirement that the method be
explicitly specified. Where only some details are missing, then a good
litmus test, in line with legal interpretation of contracts that need
to be explicit, is whether the correct missing details are fairly
immediately clear. If we have to engage in a detailed interpretive
exercise to discover the true mechanism, then it is most likely not
explicit enough to qualify.

Because R2125 is a general rule of interpretation, it also applies in
interpreting such a conditional: a rule that inquires into whether an
action is possible also inquires into whether or not a method exists.
If a method does not exist, the action is not possible. A rule wishing
to determine if the rules attempt to permit an action, including or
exclusively when no method is specified, must explicitly communicate
this intent. This is consistent with the plain language and historical
interpretation of such conditionals in the rules, such as for
deputisation or in acting on behalf. R2160 and R2466 both explicitly
examine the mechanism of performance authorised by the rules and
distinguish based on that.

These sorts of conditionals are often applied to the specification of
other methods for performing the same action, and it is often possible
to resolve such conditions into circularity so that they depend on
themselves. R2160 would be an example of such if it didn't have the
phrase "other than by deputisation": it would be possible for a player
to deputise to perform any action, provided e could deputise for it.
Sometimes, we may see the bootstrap argument that such a circular
conditional should be resolved by interpreting the action as being
possible. This argument begs the question. Such a circular conditional
is a tautology and is equally consistent if the action is not
possible. It can hardly be considered an explicit endorsement of the
action's possibility. Consequently, without needing to consider
generally how Agora handles question-begging, R2125 does not accept
it. Any such ambiguity must be resolved against the action being
possible.

## Application to Zombie Auctions

With this background out of the way, we can turn to zombie auctions,
and in particular whether or not they provide a mechanism to the
satisfaction of Rule 2125. The arguments and precedent suggest that
Rule 2545 provides this, by stating that an auction is a method of
transferring things away. But this is only a general statement, and
does not on its own provide any details on how to actually realize the
transfer. I agree with the argument that this means that Rule 1885, by
stating that an auction can be initiated, thereby implies that the
zombie can be transferred. This follows logically from those two
rules.

Rule 2532 defines the master switch as secured at power level 2,
meaning that, per Rules 2162 and 1688, master switches can be flipped
only as allowed by instruments of power 2 and greater. Rules 2545 and
1885 both have sufficient power to authorize changes to master, and
unlike with R2125 need not do so explicitly nor specify the methods,
so this could create an argument that they authorize lower-powered
rules which flesh out the definition of an auction to make a change to
the zombie switch despite its security. Indeed, if R1688 matched
R2125's requirements, the method would clearly not be explicitly
specified by rules of power 2 or greater and therefore we could stop
our considerations here. However, that is not the case, so I will
proceed under the assumption that, for R2125 purposes, we may
interpret even the lower-powered auction rules as part of the whole,
either because it is intrinsic to R2125's interpretation to do so, or
because R2545 and R1885 implicitly delegate the definition to the
auction rules. I will only revisit this assumption if needed.

Turning to the auction rules, we have two directly relevant clauses.
First, Rule 2549 states "An Auction also CANNOT be initiated unless
the Auctioneer is able to give away each item in each of the Auction's
lots." Second, Rule 2551 says "When [the winner pays the fee], if the
auctioneer CAN transfer the items in that lot to that winner at will,
e immediately does so; otherwise, e SHALL do so in a timely fashion."

R2459 is not helpful. It does not authorize anything, only prohibit.
It can't include an auction as the mechanism to give items away, as
that would create circular logic. Rule 2551, on the other hand, is
more helpful. It indicates the precise mechanism by which the actual
transfer of assets is effected: if the auctioneer can give the items
away at will, it happens automatically; otherwise e merely put under
an obligation to do so. Where the asset can already be transferred
away, the procedure is explicitly specified and there are no concerns
with respect to R2125. The obligation, on the other hand, clearly
contemplates the possibility that the Auctioneer is restricted in eir
ability to transfer the assets away. In such a case, no mechanism is
provided and the Auctioneer is left to do it on eir own.

This is exactly the bootstrapping situation I discussed above. In the
case of zombie auctions, where there are no other mechanisms for
transfer, the only way that the condition in R2551 could be met for a
zombie auction is with its own conclusion. And since R2551 implements
the auction mechanism considered by R2545, attempting to include the
latter rule into the picture does not resolve the circularity. It only
widens the circle slightly: R2545 provides a mechanism because R2551
implements an actual mechanism, therefore a lot up for auction can be
transferred because R2545 provides a mechanism to transfer lots up for
auction, therefore R2551 implements an actual mechanism because the
lot up for auction can be transferred.

## Conclusion

In conclusion, this is a typical example of the rules say I do without
saying how, therefore I do, which has plagued Agora for a long time
but possibly not for as long as I say I do, therefore I do has.

Were this CFJ not trivial, I would nonetheless judge it FALSE.

Suggested annotations:
- On R2125: "This rule is an interpretive principle applied to the
rules as a whole and not to any individual rule. For a method to be
explicitly specified, it must actually specify a way that the action
can be realized; it is not enough that a rule claims there is a
method. Circular reasoning cannot be used to permit an action or to
provide a method."
- On R217: "The 'text of the rules', for the purposes of Rule 217, is
the text of the rules as actually written, as augmented by the
semantic content of applicable definitions. Substituting definitions
into the text of a rule cannot eliminate unclarity caused by ambiguous
grammar or other linguistic aspects."

P.S. Here's a link to CFJ 3403, which I didn't use in this judgment,
but wasn't in the archive:
https://www.mail-archive.com/agora-business@agoranomic.org/msg26252.html

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