status: https://faculty.washington.edu/kerim/nomic/cases/#3817
(This document is informational only and contains no game actions).
=============================== CFJ 3817 ===============================
Falsifian initiated a zombie auction this month.
==========================================================================
Caller: Falsifian
Judge: G.
Judgement: FALSE
==========================================================================
History:
Called by Falsifian: 18 Feb 2020 20:35:45
Assigned to G.: 19 Feb 2020 15:23:56
Judged FALSE by G.: 26 Feb 2020 22:56:30
==========================================================================
Caller's Arguments:
I believe this is TRUE.
The second-last paragraph of R2549 would seem to make this FALSE:
since it was found in CFJ 3793 that Agora cannot transfer zombies,
R2549 therefore says that I could not have initiated a zombie auction.
However, R1885, which is higher-powered than R2549, says that under
certain conditions, "the Registrar CAN put that zombie (along with any
other zombies that fulfill the same conditions) up for auction.".
Those conditions were met for each zombie I listed.
So, I believe the rules authorized me to initiate an auction, and
therefore I did so by announcement (mechanism provided by R2549).
Caller's Evidence:
My message allegedly initiating the auction is at
https://mailman.agoranomic.org/cgi-bin/mailman/private/agora-
official/2020-February/013402.html, and is quoted here:
> I initiate a zombie auction, with the following lots (each zombie a
> separate lot) ordered as follows (highest-bid first):
>
> 1. Trigon
> 2. Nch
> 3. Cuddle Beam
> 4. Walker
> 5. ATMunn
>
> Agora is the Auctioneer, and the Registrar is the Announcer. The
> currency is Coins with a minimum bid of 1.
--------------------------------------------------------------------------
Judge G.'s Arguments:
The authorizing text for zombie auctions is in R1885/9 (Power=2):
> Whenever [conditions for auctioning a zombie are true] then the
> Registrar CAN put that zombie (along with any other zombies that
> fulfill the same conditions) up for auction.
I’ve confirmed from Registrar’s Reports that the [conditions]
referenced above were true for the zombies that were part of the
alleged auction announcement.
>From CFJ 3793, we know that the R2125/10 (Power=3) requires that the
R1885 “CAN put that zombie...” be linked to a method definition, by
which the action (quoting Judge Alexis) “can actually be realized”
e.g. a link to a “wilful natural action of a sentient being that
starts things”.
There is a little tension here with the definition of CAN itself in
R2152/8 (Power=3):
> 5. CAN, POSSIBLE: Attempts to perform the described action are
> successful.
which implies that “any” attempt to perform an action is successful,
however this tension can be greatly lessened by adopting a strict and
narrow definition of the term “attempt” - that is, to qualify as
“attempt”, the actor must meet all of the technical requirements of
the defined method that links the CAN to a realizable action. In
common language, we might define an attempt more broadly, but in
Agora, an “attempt” that does not follow the needs and restrictions of
the given method is ISID.
However, if an action has rules restrictions on its performance, it’s
important to distinguish whether the restrictions are an intrinsic
part of the method definition (part of the definition of “attempting”
the action) versus being an external restriction on the authorizing
CAN. Consider the following example rule texts:
(a)
> An entity authorized by a rule to initiate an Auction does so by
> announcement, provided that the announcement includes [necessary
> parameters] and provided that [method conditions] are true.
This is the way most methods in the rules (e.g. dependent actions) are
written, and I don’t think there’d be much argument that [necessary
parameters] and [method conditions] are part of the method definition
- the conditions must be fulfilled in order to make a successful
action “attempt” even if the text of (a) is at a lower power than the
authorizing CAN.
(b)
> An entity authorized by a rule to initiate an Auction CAN do so by
> announcement, provided that the announcement includes [necessary
> parameters] and provided that [method conditions] are true.
In (b), the only difference from (a) is the “does so by” is replaced
by “CAN do so”. I don’t think, in this context, that adding a CAN
changes this from a “definition” to an “added authorization” - it
amounts to “if e CAN, e CAN do so by an announcement with these
restrictions” which is essentially a method definition - and the
restrictions are still clearly part of the method.
(c)
> An entity authorized by a rule to initiate an Auction CAN do so by
> announcement; the announcement CANNOT be missing [necessary
> parameters] and [method conditions] CANNOT be false.
If we accept CANs as part of the method in (b), then using CANNOTS to
modify the “CAN do so by announcement” in (c) makes sense given the
sentence structure. In particular, it *doesn’t* make sense to
consider the “CAN do so by announcement” as a method definition and
then say “but those CANNOTS apply to the power-2 initial
authorization”. So (c) should work like (b) and (a).
(d)
> An entity authorized by a rule to initiate an Auction CAN do so by
> announcement; the announcement CANNOT be missing [necessary
> parameters].
>
> SEPARATE power=1 rule: An entity CANNOT initiate an auction if
> [method conditions] are FALSE.
Here, the final restriction, by being in a separate rule, is clearly a
restriction on the initial authorizing CAN, not part of the method
definition. The separate rule’s CANNOT would apply to the R1885
authorizing CAN, and if the separate rule was power=1, R1885’s power=2
CAN would overrule the CANNOT.
Unfortunately, the actual text for Auction Initiation (R2549/4
Power=1) is midway between (c) and (d):
> An entity authorized by a rule to initiate an Auction CAN do so by
> announcement.
> [...]
> An Auction CANNOT be initiated unless the announcer specifies all
> of [essential parameters]. An auction also CANNOT be
> initiated unless [conditions].
In favor of the CANNOTS being intrinsic to the method: They are in the
same rule, one CANNOT applies directly to the announcement (part of
the method) and the second CANNOT is linked to the first with an
“also” implying it modifies the ability to make the announcement.
In favor of the CANNOTS being standalone modifications to the R1885
CAN: it’s a separate paragraph, separated by the [...] text, and each
sentence starts “An Auction CANNOT be initiated” implying the
restriction applies to any statement in the Rules that says it CAN.
This is a close call - what is the best way to parse?
Well, importantly, any “method definition” will be interpreted under
R217 with respect to power and precedence, that is, from R217/12
(Power=3):
> Definitions in lower-powered Rules do not overrule
> common-sense interpretations or common definitions of terms in
> higher-powered rules, but may constructively make reasonable
> clarifications to those definitions. For this purpose, a
> clarification is reasonable if and only if it adds detail without
> changing the underlying general meaning of the term and without
> causing the higher powered rule to be read in a way inconsistent
> with its text.
Rule 2549/4 (Power=1) talks about “auction initiation”, but the CAN in
R1885 refers to “putting a zombie up for auction” and not “auction
initiation” directly. Therefore, the definition of “auction” as a
whole is relevant. This is found in Rule 2545/2 (Power=2):
> An Auction is a way for entities to give away items in exchange
> for a currency. Any rule CAN permit or require Auctions to be
> initiated.
The remainder of R2545 makes it clear that the “entity” in question in
this sentence is the Auctioneer of the particular auction, and the
“items” to be given away are the lots. Therefore, if a lower-powered
Rule defines auction such that an auctioneer need not be able to give
away the items in question, it is inconsistent with the definition of
“auction”.
This tips the decision. The particular CANNOT is:
> An Auction also CANNOT be
> initiated unless the Auctioneer is able to give away each item in
> each of the Auction's lots.
This is absolutely consistent with the definition of auction, as a
means for the auctioneer to “give away” items - if e can’t give an
item away, e doesn’t have an Auction. Therefore, the CANNOT is a
“reasonable clarification” of the Power=2 definition of “auction” and
part of the method definition, not a stand-alone conflict with R1885.
Overall, the power=1 method definition of “auction initiation”, which
is required to be followed to count as an “initiation attempt”,
includes (by definition) that the auction CANNOT be started in the
circumstances given.
I judge FALSE.
==========================================================================
