status: https://faculty.washington.edu/kerim/nomic/cases/#3745
(This document is informational only and contains no game actions).
=============================== CFJ 3745 ===============================
There exists an Agoran Decision to adopt a proposal with the title
'It's caused enough trouble already' and with a valid adoption
index.
==========================================================================
Caller: Jason Cobb
Judge: D. Margaux
Judgement: FALSE
==========================================================================
History:
Called by Jason Cobb: 23 Jun 2019 22:07:24
Assigned to D. Margaux: 30 Jun 2019 20:02:27
Judged TRUE by D. Margaux: 01 Jul 2019 00:47:15
Motion to reconsider filed by D. Margaux: 01 Jul 2019 06:16:07
Judged FALSE by D. Margaux: 01 Jul 2019 06:16:07
Motion for D. Margaux to reconsider: 06 Jul 2019 08:23:08
==========================================================================
Gratuitous Arguments by G.:
When you submit a proposal, it is "optional" to include an adoption index
(R2350). The default value in R1950 is "none" so that likely means the
result (if you submit without specifying at all) is a proposal with "AI =
none".
If you submit with an invalid (but optional) AI, I'm not at all sure
whether it invalidates the proposal creation or sets it at default.
For example, in the case of Rule Changes, if you say "Amend Rule XXXX
([Title])", including the [Title] is optional, but if you specify the
title incorrectly, precedent holds that it invalidates the whole rule
change as overly ambiguous, though that relies on the "Any ambiguity..."
clause in R105 specific to Rule Changes.
Falsifian specified an (invalid) AI of 0.5 when submitting the proposal
in question. So for more general by-announcement actions, does
specifying an invalid but optional parameter invalidate the whole process?
--------------------------------------------------------------------------
Gratuitous Evidence by G.:
On 6/23/2019 9:08 AM, James Cook wrote to agora-business:
I create a proposal with the following attributes and text:
Title: It's caused enough trouble already
Adoption index: 0.5
Co-authors: none (empty list)
Text: Repeal Rule 2596 (The Ritual).
Rule 2350/11 (Power=3)
Proposals
A proposal is a type of entity consisting of a body of text and
other attributes. A player CAN create a proposal by announcement,
specifying its text and optionally specifying any of the following
attributes:
* An associated title.
* A list of co-authors (which must be persons other than the
author).
* An adoption index.
Creating a proposal adds it to the Proposal Pool. Once a proposal
is created, neither its text nor any of the aforementioned
attributes can be changed. [...]
--------------------------------------------------------------------------
Judge D. Margaux's Arguments:
I judge it FALSE for the reasons given by g.
On Jul 1, 2019, at 12:30 AM, Kerim Aydin <ke...@uw.edu> wrote:
The Proposal Distribution (not the Proposal) was CoE'd on the AI (the
Distribution listed the AI as 0.5, which is wrong regardless). Since
AI is an essential parameter, that means the attempt to distribute the
proposal and create a decision failed, by R107. (using R107 language,
the CoE "correctly identified the lack" of a valid AI).
Therefore the decision was never initiated - it was invalid. (This
all happened before the CFJs were called).
--------------------------------------------------------------------------
Arguments filed with motion to reconsider, by Aris:
I’m strongly tempted to move to reconsider this, and apologize for
failing to provide arguments earlier (honestly, I totally forgot about
this case). I really don’t think this opinion adequately considers the
other sensible possibility: that the proposal fails entirely.
To begin with, when someone submits a proposal, they’re performing the
action of creating an entity that has certain properties. If creating an
entity with the properties specified is impossible, there’s no way for
them to do that (the speech act cannot make itself true). This is a very
different situation from the one where they leave something blank, and
we fill it in with a default value. There, the person is still
performing the speech act they wrote out, it’s just that the entity they
create also has additional properties. Here, we’re actually rewriting an
ineffective speech act to turn it into an effective one, which feels
deeply disturbing and totally unlike the way Agora speech acts normally
work.
Another problem is that, since we’re rewriting what someone said, we
break referential transparency and bump into other counter-intuitive
behavior. If someone says “I submit a proposal with properties X, Y, and
Z, and AI 0.5. I retract that proposal.”, no proposal would result under
this theory. However, if we substituted the variable “that”, we get “I
submit a proposal with properties X, Y, and Z, and AI 0.5. I retract the
proposal with properties X, Y, and Z, and AI 0.5.” However, this leads
to a proposal still existing, since there is no proposal with AI 0.5,
unless it’s proposed that we rewrite that sentence too, which makes even
less sense. To be clear, I’m not saying that Agora always follows
referential transparency, just that we need a principled reason to break
it and this really doesn’t feel like one.
In short, I think it’s by far the cleanest if we read statements
literally and don’t allow the defaulting of invalid values. I’d like to
know if others think I could have a point, and if they do I’ll file a
motion (unless the judge cares to self-file one). I really think their
are some counter-arguments here that deserve consideration, and would go
so far as to suggest that if the current interpretation is sustained we
might want to legislate to make sure that announcements are read literally.
==========================================================================
