I will make no claims as to the accuracy of the drafts, but you did forget
a "what" in the wording "D. Margaux calls is later named CFJ 3727." :)
Jason Cobb
On Sat, Jun 1, 2019 at 11:59 PM James Cook <jc...@cs.berkeley.edu> wrote:
> Comments welcome. Sorry that it's so long. I went back and forth on
> 3726 a couple of times.
>
> I believe this is due on June 4 at 21:53 UTC. I plan to send it out
> the next couple of days.
>
> --------
>
> This is my judgement of CFJs 3726 and 3727.
>
> CFJ 3726 was called by Aris, with the statement: "The most recent
> attempted imposition of the Cold Hand of Justice by Aris was effective."
>
> CFJ 3727 was called by D. Margaux, with the statement: "D. Margaux has
> more than 0 blots."
>
> 1. Arguments
> ============
>
> There was a long conversation on the discussion list, starting around
> when D. Margaux called a CFJ (later withdrawn) on the thread "[Referee]
> Recusal (attn H. Arbitor)" in May 2019. I will not try to repeat
> everything here.
>
> 2. Sequence of events (all times UTC)
> =====================================
>
> 2019-05-20 01:25
>
> The Referee publishes a weekly report specifying that D. Margaux has 0
> blots.
>
> 2019-05-20 20:32
>
> D. Margaux publishes the below document and announces intent to ratify
> it "as true at the time 00:00 GMT on 20 May 2019":
>
> { For purposes of this document, “Politics Rules” and “Spaaace Rules”
> have the meaning ascribed to those terms in Proposal 8177.
>
> Any switch created directly by any of the Politics Rules or the
> Spaaace Rules has its default value.
>
> There are no currently existing entities or switches created by the
> Clork pursuant to the Politics Rules or by the Astronomor pursuant to
> the Spaaace Rules. }
>
> 2019-05-21 10:20
>
> D. Margaux deputises as Astronomor and Clork to publish the following
> weekly reports:
>
> {there are no entities in existence for which the Astronomor is the
> recordkeepor other than those created directly by the Rules. All
> switches for which the Astronomor is recordkeepor have their default
> value.}
>
> {there are no entities in existence for which the Clork is the
> recordkeepor other than those directly created by the Rules. All
> switches for which the Clork is recordkeepor have their default value.}
>
> 2019-05-25 22:02
>
> omd Points eir Finger at D. Margaux for publishing inaccurate
> information in the above reports.
>
> 2019-05-25 22:54
>
> D. Margaux, the Referee, authorizes the Arbitor, Aris, to act on eir
> behalf to "investigate and conclude the investigation of the finger
> pointed".
>
> 2019-05-26 22:43
>
> Aris attempts to act on D. Margaux's behalf to impose the Cold Hand of
> Justice on D. Margaux and fine em 2 blots, with the following message:
>
> > Alright. There was a clear rule violation here, as the information in
> the
> > report was inaccurate. The violative conduct was undertaken for the
> good of
> > the game, but there were also other options available (proposal, or
> > ratification without objection, which would have been unlikely to
> cause any
> > problems done correctly). Ordinarily, a rule violation for the good of
> the
> > game would be a forgiveable one blot fine. Under the circumstances
> though,
> > some additional penalty is warranted for failing to adequately
> consider and
> > discuss options that would have avoided violating the rules.
> >
> > I act on behalf of D. Margaux to impose the Cold Hand of Justice on D.
> > Margaux, penalizing em with a forgiveable fine of 2 blots. The required
> > words are {optimize, preferentially, consider, supersubtilize,
> > adjudication, law, good, bad, future, duty}.
>
> 2019-05-26 22:50
>
> D. Margaux ratifies the document they earlier announced intent to
> ratify.
>
> 2019-05-27 14:11
>
> D. Margaux calls is later named CFJ 3727.
>
> 2019-05-27 19:58
>
> Aris calls what is later named CFJ 3726.
>
> 3. Effectiveness of the fine ignoring ratification
> ==================================================
>
> It is helpful to first consider whether the attempt to levy a fine would
> have been effective if no ratifications had taken place.
>
> I believe that Aris's attempted imposition of the Cold Hand of Justice
> by levying a fine (2019-05-26 22:43 message) met the requirements of
> Rule 2557, so it remains only to check that it does not run afoul of any
> of the conditions in Rule 2531 ("Any attempt to levy a fine is
> INEFFECTIVE if...").
>
> Condition 1 of Rule 2531:
>
> The attempt in Aris's message included the value the fine (2 blots) and
> the name of the person being fined (D. Margaux). The sentence performing
> the action did not specify the specific reason for the fine (message is
> copied earlier: "I act on behalf of D. Margaux to impose..."). However,
> Aris states the message earlier in the same message "...the information
> in the report was inaccurate" which is part of eir attempt, so Condition
> 1 does not trigger.
>
> Conditions 2 and 3 of Rule 2531:
>
> The officer reports were false at the time they were published. For
> example, The Astronomor's 2019-03-05 weekly report states that many
> players own Spaceships, and that part of the report has long since
> self-ratified (under Rule 2166), but D. Margaux's report claims no
> Spaceships exist.
>
> Therefore, ignoring the ratification of the more recent reports,
> D. Margaux did commit a rule-breaking action, so these two conditions do
> not apply.
>
> Conditions 4 through 8 are straightforward.
>
> Therefore, if D. Margaux's document and reports were never ratified, the
> attempted imposition would have been EFFECTIVE, and D. Margaux would
> have blots.
>
> 4. The timeline hypothesized by rule 1551
> =========================================
>
> When D. Margaux ratified eir document on May 26, according to Rule 1551,
> the gamestate was "modified to what it would be if, at the time the
> ratified document was published, the gamestate had been minimally
> modified to make the ratified document as true and accurate as
> possible".
>
> (D. Margaux specified 2019-05-20 00:00 GMT as being the time the
> document was true, but since the time was not part of the document
> itself, it does trigger Rule 1551's clause "if the document explicitly
> specifies a different past time...". Anyway, I think the document was
> published close enough to that time that the effect is the same.)
>
> To understand how Rule 1551 modified the gamestate, the first step is to
> understand the minimal modification that is hypothetically applied at
> the time the document was published. This minimal modification is clear:
> all switches created directly by the Politics and Spaaace Rules are
> modified to be at their default values, and all entities and switches
> created by the Clork or Astronomor according to those respective Rules
> are removed so that they do not exist in the minimally modified
> gamestate.
>
> Having understood the hypothetical minimal modification, the next step
> is to understand what would have happened from that point in time
> onward. We end up with the following timeline:
>
> * D. Margaux publishes correct Astronomor and Clork reports.
> * omd Points eir Finger at D. Margaux for publishing inaccurate
> information in those reports.
> * D. Margaux authorizes Aris to act on eir behalf for the investigation.
> * Aris attempts to impose the Cold Hand of Justice, but the attempt is
> INEFFECTIVE because D. Margaux did not violate the rules.
>
> In this hypothetical timeline, the appropriate judgement of CFJ 3726
> would be FALSE, since Aris's attempted imposition was INEFFECTIVE, and
> the appropriate judgement of CFJ 3727 would be FALSE as well, since D.
> Margaux did not gain any blots after the 2019-05-20 referee report,
> which self-ratified before CFJ 3727 was called.
>
> 5. The gamestate after ratification
> ===================================
>
> After D. Margaux ratified eir document, the gamestate was modified to
> what it would be in the hypothetical timeline above. Our task now is to
> determine how that modification affects the value of the two CFJs.
>
> I think it is clear that blot holdings are part of the gamestate. For
> one thing, they only exist because the game defines em, and are part of
> the state that we keep track of, so the terms "game" and "state" fit.
> For another, blots are an asset tracked by some recordkeepor (the
> Referee), and Rule 2166 states that the portion of a recordkeepor's
> report that lists assets is self-ratifying, which wouldn't mean much if
> such assets weren't part of the gamestate. Therefore at the end of this
> message I will judge 3727 FALSE.
>
> The harder question is whether the change to the gamestate affects CFJ
> 3726. I believe it comes down to this question:
>
> Does the gamestate include information about past events, such as
> whether the attempted imposition of the Cold Hand of Justice was
> effective, and is CFJ 3726 referring to the information contained in
> the gamestate?
>
> If the answers are no, then CFJ 3726 should be judged TRUE, as argued in
> Section 3 above. If the answers are yes, then CFJ 3726 should be judged
> FALSE, according to the timeline described in Section 4.
>
> I believe the answers are yes, and so at the end of this message I will
> judge CFJ 3726 TRUE. Before I say why, I'd like explain why there could
> be doubt about this.
>
> 6. An interpretation causing CFJ 3726 to be FALSE
> =================================================
>
> When the rules talk about a concept that isn't defined by the rules,
> we generally assume they're referring to the concept in reality. For
> example, the rules don't define what a message is, and they don't define
> the meaning of fundamental words like "is", "for", etc.
>
> Therefore it's natural to assume that when the rules discuss what "was",
> they are referring to past events. I think it's defensible to assume
> that even if the past event being referred to is a legal fiction, the
> rules are referring to what the Rules had to say about the legal fiction
> in the past, rather than some new legal fiction the Rules currently
> define about the past.
>
> To instead say that whenever the Rules refer to the past, they are
> referring to some fictitious past defined by the rules, could be said to
> be a very drastic statement, requiring strong evidence to back it up.
>
> 7. The past is part of the gamestate
> ====================================
>
> However, I think sufficient evidence does exist to say the gamestate
> contains information about past actions, and that the Rules refer to the
> gamestate when they talk about the past.
>
> 7A. The text of the rules
> =========================
>
> In two places, the text of the Rules seem to imply that the past is part
> of the gamestate:
>
> Rule 1551 says:
>
> > Text purportedly about previous instances of ratification (e.g. a
> > report's date of last ratification) is excluded from ratification.
>
> Rule 2034 says that certain messages constitute self-ratifying claims of
> several facts about the past: "such a decision existed", "it had the
> number of voters indicated", etc.
>
> In both cases, if the gamestate did not include information about the
> past, or the Rules did not refer to that information when referring to
> the past, then these parts of the Rules wouldn't make sense. The
> gamestate couldn't be "minimally modified" to make these statements
> true in the hypothetical timeline described by Rule 1551, and after the
> ratification, the modified facts about the past, not being part of the
> gamestate, would not be.
>
> 7B. Past judgements about judgements about the past
> ===================================================
>
> Rule 217 instructs us that when the text of the rules is silent,
> inconsistent or unclear, we should augment it with past judgements
> (among other things).
>
> omd pointed out that the judgement CFJ 3337 [0] relied on
> past events being part of the gamestate. Judge G.'s judgement of that CFJ
> relied on the assertion that after a document stating the past date of a
> player's registration was ratified, that player became eligible to vote
> in a decision they otherwise would not have been eligible for, because
> the text of the ratified document implied that they were registered at
> the start of the voting period.
>
> The judgement of CFJ 2491 [1] also relied on precedent being part of the
> game state: Judge Yally's judgement was that since a document was
> ratified implying that Rodlen became an officer sufficiently long ago
> for em to be awarded a green ribbon, e should be awarded a green ribbon.
>
> Both of these judgements support the conclusion that the past is part of
> the gamestate.
>
> 7C. The best interests of the game
> ==================================
>
> Rule 217 also instructs us to consider the best interests of the game.
>
> Generally, the purpose of gamestate changes such as the effect of
> ratification is to clarify the current situation in the game in order to
> allow play to continue smoothly. Consider the following examples where
> rules make actions possible or required depending on what happened in
> the past:
>
> * Rule 2555: "If a player has neither gained blots nor expunged any
> blots from emself in the current Agoran week, e CAN expunge 1 blot
> from emself by announcement."
>
> * Rule 2496: Whether a player can earn a reward depends on whether they
> have fulfilled the condition in the past 7 days, and also on whether e
> has already claimed that reward.
>
> * Many rules allow or require a player to take an action "in a timely
> fashion" after a past event.
>
> When a ratification event clarifies the current situation in the game,
> it would be helpful if the above examples are among the things that are
> clarified. For example, after ratification of a document implying
> whether or not a player has gained blots in the current week, that
> player can confidently know whether they are able to expunge a blot. All
> other things being equal, I think it is in the best interests of the
> game if ratification clarifies these examples.
>
> (There may be best-interests-of-the-game arguments going the other way,
> e.g. maybe it's easier to untangle some situations if ratification isn't
> mucking around with the past. But 7A and 7B still apply.)
>
> We could at this point try to pick and choose just those elements of the
> past are directly relevant: e.g. since D. Margaux didn't have any blots
> to begin with, CFJ 3726 is not relevant to whether e can expunge any
> blots, so CFJ 3726 may not matter to the best interests of the game. But
> I think that would be overcomplicating things. If the gamestate includes
> the effectiveness of some past actions, it is natural to assume it
> includes the effectiveness of all of them.
>
> 8. It depends on what the meaning of the word "was" was
> =======================================================
>
> There's still a question of interpreting CFJ 3726 itself. I have
> concluded that when the Rules refer to the effectiveness of past
> actions, they are referring to information in the gamestate about the
> past. But CFJ 3726 is not a rule. What does "was effective" mean in the
> statement in CFJ 3726?
>
> Rule 591 has guidance on delivering judgements:
>
> > The valid judgements for an inquiry case are as follows, based on
> > the facts and legal situation at the time the inquiry case was
> > initiated, not taking into account any events since that time:
> >
> > * FALSE, appropriate if the statement was factually and
> > logically false
> >
> > * TRUE, appropriate if the statement was factually and
> > logically true
>
> I believe "legal situation" means I should base my decision on what the
> rules say. Moreover, the CFJ process is more useful if judgements on
> statements agree with what the rules say about those statements.
>
> So, I will interpret "was effective" as referring to information about
> the past contained in the gamestate.
>
> 9. Judgement
> ============
>
> I judge CFJ 3726 FALSE.
> I judge CFJ 3727 FALSE.
>
> [0] https://faculty.washington.edu/kerim/nomic/cases/?3337
> [1] https://faculty.washington.edu/kerim/nomic/cases/?2491
>
