Index ← 2376a CFJ 2376 2377 → text
==============================  CFJ 2376  ==============================

    It is possible for a Rule to change a proposal's text.


Caller:                                 omd

Judge:                                  Murphy
Judgement:                              FALSE

Appeal:                                 2376a
Decision:                               REASSIGN

Judge:                                  G.
Judgement:                              TRUE



Called by omd:                          11 Feb 2009 02:20:11 GMT
Assigned to Murphy:                     14 Feb 2009 17:50:43 GMT
Judged FALSE by Murphy:                 14 Feb 2009 18:31:10 GMT
Appealed by omd:                        14 Feb 2009 21:07:32 GMT
Appealed by ais523:                     14 Feb 2009 21:20:42 GMT
Appealed by Taral:                      14 Feb 2009 22:05:25 GMT
Appeal 2376a:                           14 Feb 2009 22:23:18 GMT
REASSIGNED on Appeal:                   24 Feb 2009 00:07:36 GMT
Assigned to G.:                         24 Feb 2009 19:32:32 GMT
Judged TRUE by G.:                      24 Feb 2009 20:34:53 GMT


Caller's Arguments:

Suber's rule in the subject is no longer in the Agoran ruleset, but I
forsee arguments that this doesn't work because a proposal, as a 'document
outlining changes to be made to Agora' (R106), is necessarily immutable
(corresponding to the standard requirement that amendments to proposals
must be performed by retraction and resubmission).  Although some types of
documents (such as public documents and official documents) are immutable,
I don't think 'document' implies immutability.  For example, Rule 754
implies that a contract is a legal document; Rule 2166 refers to a rule or
contract as a 'backing document'; and Rule 1586 temporarily defines
'document' as 'rule or contract'; rules and contracts can be amended.
Alternately, in ordinary language, amendable texts such as the United
States Constitution are referred to as 'living documents'.  Therefore,
there is nothing implicit in the definition of a proposal to prevent one
from being amended; as the attributes of proposals are not secured, it is
possible for a Rule to (grant permission to) amend them.


Gratuitous Arguments by G.:

A rule specifically allowing a proposal text to change is no more
difficult to accept than one changing chamber, AI, etc.  But what's
important is what is needed to actually *adopt* a Proposal.  That is
defined under Agoran Decision rules, not proposal rules.

Specifically, in R107 and elsewhere, an Agoran Decision is a decision on
a "matter" to be decided on, which I argue could be defined as referring
to a particular and specific version of a text, not a different version
which later bears the same ID number.  The specific matter must be in the
initiation notice for the decision[*].

I fully admit that's a "for the good of the game" style argument relying
on comparing a general definition of a "matter" against a general
definition of "document", "text", etc. provided by comex.

[*] Note that the example in R107(a) is shorthand as it doesn't include
the full text of the proposal and thus doesn't show the "matter" -
therefore it is an incomplete example - ID numbers themselves are not a
"matter" but a handy method for referring to a "matter" as long as there
is no reasonable ambiguity (the example in R107(a) predates ID numbers).
It would be keeping with very long Agoran custom and English definitions
to rule that a "matter" in R107a consists of an entire proposal text of
a specific version published at the initiation of the decision.


Gratuitous Arguments by G.:

1.   The consideration of what an entity "is" (e.g. is it its text?
its name?  its form?  the sum of its changing properties? ) is as
unclear in ancient philosophy as it is in Agoran proposals (cf.
Ship of Theseus) so there is considerable room for interpretation
and debate for any particular type of entity.   I would argue that,
unlike documents that report other values (e.g. a report identified
by its changing currency holdings), or a "living" contract (that is
identified by its parties and their fact of agreement as well as its
text) that the caller cites, there is sufficient linguistic,
customary, procedural, and Agoran specific and nomic-in-general
historical evidence to strongly support (and perhaps make a "good of
the game" argument) that a proposal *is*, by its very nature and
definition, a precise text in a precise and immutable version, that
other properties; AI, ID number, etc. are secondary, and that a
different textual version is a different text and therefore a
different proposal (or not a proposal).

The Caller might argue that eir added text to R2238, "comex CAN
modify any attribute of a proposal, including AI and text, by
announcement" does indirectly define a proposal's text as mutable.
However, at the time this definition is needed, R2238 is at power-1,
and the current R754/9 [(2) and (4)] states that so long as the
definitions (even implicit definitions with historical context) of
immutability are contained or strongly supported in rules>1, R2238
SHOULD be used for guidance, but "guidance" not strong enough to
overrule the immutability.  (R754/9 makes the "guidance" of low-power
rules purposefully weak; the expressed legislative purpose was to
prevent low-power definitions, especially implicit ones, from
overruling common usage or strong contextual definitions of terms
contained in high-power rules, to prevent just this sort of scam).

2. *If* the judge is to accept that one proposal can have two
versions of its text and be the same proposal, the judge should
first consider the impacts on the Agoran Decision process for
adopting proposals as well as the definition of proposal itself,
with specific consideration of what is meant by the "matter" to be

For example, in order to "clearly identify" a text with two versions
(as required by R683 and R208 for voting and resolution), I would
argue that, if two versions exist, the version must be identified at
each vote (e.g. it is not enough to identify a "report" in general
to ratify, one must identify a specific copy of the report).  This
might lead to various parts of the process (but not all of it)
breaking depending on how and when identifications are made, or it
might simply reinforce the procedural evidence for accepting
argument #1.


Judge Murphy's Arguments:

I accept Goethe's gratuitous arguments regarding Rule 754.  comex's
scam legislation also fails to address the obvious and significant
ramifications of its claim (particularly the impact on an in-process
Agoran decision on whether to adopt the proposal), and is too low-power
to impose brokenness rather than merely being ineffective (cf. Zefram's
arguments against "the rules are a contract" at Power=1).


Appellant omd's Arguments:

I intend to appeal this with two support.  The judge did not address
the debate about a proposal's document-ness in any way, but assumed
that a proposal is immutable without addressing any of the arguments
about the matter.


Appellant ais523's Arguments:

I support. This judgement is completely missing the point. Goethe's arguments
about rule 754 are about contracts, not proposals, and rather tangential to
the point (they imply that proposals are either natural-language proposals
per rule 754 or natural-language documents per rule 106); although Goethe's
argument is relevant, it is only relevant in establishing what the judge must
look at, not an argument for FALSE in itself! (Imagine the following
hypothetical: CFJ "ais523 is a first-class person", gratuitous argument "a
first-class person is a person of a biological nature", Judge "I accept the
gratuitous argument, FALSE"; that would only be appropriate if it was very
obvious as to my nature. This situation is similar; it is not at all obvious
whether a natural-language proposal is modifiable in general, or Goethe and
I wouldn't have been debating it for several days with neither of us ending
up entirely sure at the end of it.)

Also, why should comex's scam legislation address the ramifications of its
claim? That's for a judge to do, not for the scam rule itself. (If the scam
rule did state a judgement to be used in any CFJs regarding it, I rather
suspect that would either be considered judicial corruption, or ignored,
rather than something that the rule is required to do!) If a rule says, or
implies, that a proposal can be modified, it is not ineffective merely
because it does not consider every possible resulting corner case.


Gratuitous Arguments by ais523:

On Wed, 2009-02-11 at 11:21 -0800, Kerim Aydin wrote:
> On Wed, 11 Feb 2009, comex wrote:
> > The only way around this is to argue that a proposal also includes its
> > AI and II (which is dubious given a reading of the relevant rules);
> No, it's trivial.  Think databases.  The relation "One text per
> proposal" does not imply "one proposal per text."  All you're raising
> is that there's some ambiguity in referring to two proposals in the
> latter case.  Various aspects (in particular ID numbers) make
> referring to multiple proposals with the same text easier, but if
> those all break down due to ambiguity, you can always refer to
> "I vote for the proposal submitted with the following text in the
> following message" as a trivial though annoying solution.  But
> you're right, text alone is not enough.  See my last paragraph
> below.
I agree that there's only one text per proposal.
      A proposal is a document outlining changes to be made to Agora,
      including enacting, repealing, or amending rules, or making
      other explicit changes to the gamestate.
It's pretty clear that a proposal can have an AI, II, title, etc, as per
the rules. Text is a feature of a document, just as those other things
are features of documents which happen to be proposals.
> You've made various analogies "proposals may be like contracts" or
> "proposals may be like reports".  I wonder how you get around the
> English definition we're required to use; e.g "proposals" (Agora)
> are like "proposals" (dictionary):
> 1 an act of putting forward or stating something for consideration
> 2 a: something proposed : suggestion
> By this definition, a proposal is tied to "something" proposed.  The
> only "something" that a person is required to submit for our
> proposals is published, marked block of text.  By common definition,
> a different block of text would a different "something" put forward
> and a different proposal.  Or are you saying that two different
> texts are somehow the same "suggestion"?
If I remember correctly, B Nomic used to allow players to amend
proposals, before they were voted on. PerlNomic allows players to amend
proposals even /after/ they've been voted on, if they had more abstain
than no votes (which resets the vote count). Certain real-life legal
systems also allow amendments of their equivalents to proposals; it is
common for proposed laws to change several times in both the US and UK
parliaments, for instance, before being adopted or rejected. Whether
common-language suggestions can be modified without retracting or
remaking them is, I suspect, normally irrelevant in conversation; I
expect that altering a suggestion would be treated as an entirely
possible action in real-life, and whether that is just an abbreviation
for making a different similar suggestion is nearly always irrelevant,
confined just to the outer realms of philosophical argument (and nomic,
of course). That makes it not particularly helpful...
> Actually, by this definition, if a second person puts forward the
> same text, it *is* the same proposal.  There used to be a specific
> clause preventing that:
>       The delivery of text identical to another Proposal which was
>       Proposed less than three weeks previous does not create a
>       new Proposal unless accompanied by a clear and explicit
>       indication that the intent was to Propose a duplicate Proposal.
> This is part of the historical record I was thinking of.  But I
> wonder if its removal was a mistake?  That some of those case of
> the same text were, by definition, the same proposal?
That rule looks to me like it's an attempt to prevent duplicate
proposals being submitted by mistake; an anti-typo rule, so to speak,
rather than anything inherent in the nature of proposals. (Remember all
the accidental double-distributions from the PNP? Presumably, something
similar happened on a whole-Agora scale back then.)

Actually, the PNP double-distributions are interesting, as I think we've
ruled that they create two different proposals. Which one is the
originally proposed one, I wonder?
> A way out of this is to say that text alone is not enough.
> "Text submitted by a certain person in a certain message" is
> needed to identify a proposal (R106). (evidence: what happens
> when a copy/paste error is made by the distributor?  A new text
> therefore a new proposal).  On the other hand, AI, etc do
> not make a new proposal.
I'd say a copy/paste error creates a new proposal because it could have
been created. Clearly, typos shouldn't be capable of editing the text of
a proposal! (Also, see CFJ 1669. That CFJ was back when attributes of
proposals needed to have particular values for them to be distributed.)
The important thing here is that the Promotor is assigning an ID number
to the proposal, and so has to identify it by some means other than
number. In this case, the text is used to identify the proposal. If the
text is incorrect, therefore, the proposal isn't identified.

However, given that the Promotor is entirely capable of creating a
proposal on eir own accord, I expected that such a proposal would count
as submitted by the Promotor if submitted. (See, however, CFJ 1780,
which seems inconsistent with both of our views; I wonder what the
result would be if it were made again nowadays?)

Now, suppose the /Assessor/ makes a typo in the text of a proposal
Assessing a proposal. Does that cause the assessment to be ineffective?
No, because the the text of the proposal need not be published during
the assessment of the proposal. The proposal is now being identified by
ID number, and there can only be one proposal with any particular ID
number (rule 2161). It's possible to amend the text of a proposal (a
document), and the ID number continues to refer to the proposal, just as
a rule (an instrument whose content takes the form of text) can be
amended and keep the same ID number, and a contract (a binding
agreement) can be amended and still keep the same ID number. (I wonder:
if proposals didn't have ID numbers and had to be identified by their
text during assessment, and the Assessor made a typo that fundamentally
changed the meaning during assessment, would it adopt that proposal with
a different text? Would it adopt a different, newly created, proposal
with a different text?)


Judge G.'s Arguments:

First, it is trivially TRUE that the Rules could be altered at high level
to allow this.  However, in the spirit of dealing with the Rules at the
time of the CFJ (rather than dealing with all hypothetical rulesets),
I'll write the opinion towards a narrower question that I suspect was
intended by the Caller:   "It is possible for a clause to be added to a
Rule (and no other changes be made) that would allow a Rule to change a
proposal's text."

In the Ruleset at the time of the CFJ, the definition of a proposal and
how it is to be submitted was contained in R106/15; the question is whether
this definition or submission process makes the text inherently immutable.
The previous arguments (gratuitous and otherwise) for this case and for
Appeal 2376a have made it clear that in common or legislative processes,
proposals can be mutable or immutable; in legislative processes in
particular, mutability is defined by the existence of a mutation process.

In a gratuitous argument, I stated:
  - The process of R106, and of Agoran Decisions, does suggest, rather
    strongly, immutability;
- The history of the process also strongly suggests immutability;

However, neither the general R106/15 description of what a proposal "is",
nor the explicit submission process in R106, nor the historical context,
contain explicit Rules-based definitions of immutability.  Therefore, using
the guidance of R754/9(2), if a clause were added to a rule *of the same
power as R106/15* that declared a proposal mutable, *or* contained an
explicit process of mutability, it would define the mutability without
contradicting R106.  This, again leads to TRUE.

So what if a lower-powered rule defined a mutability process?  By
Rule 754/9(4), "In determining the ordinary-language meaning of a term,
definitions contained in lower-powered Rules, followed by definitions
used in contracts or other Agoran legal documents, SHOULD be used for

The question is, how strong is this guidance?  Rule 754 has been
specifically re-written to allow common definitions of words in higher-
powered rules (e.g. "proposal") to have higher authority than lower-
powered definitions.  If common dictionaries clearly and directly linked
"proposals" to "immutable text", that definition would win.  However,
in this case, as discussed at length, the common definition of "proposal"
does not forbid mutability and some legislative systems allow it.

Further, while the "weak suggestions" and "historical context" do in
fact suggest immutability, it is not at this point for the good of
the game to disregard the guidance of the SHOULD in R754; rather, in
the name of respecting process, a well-specified power-1 mechanism for
altering a proposal should not be taken as conflicting with R106/15.
Again, this leads to a judgement of TRUE.

Therefore, with a finding of TRUE on any reading of the CFJ in question,
this Court finds TRUE.  HOWEVER:  it is important to note that this
decision does *not* address, nor is it meant to imply an opinion on, what
happens to an Agoran Decision to *adopt* a proposal if the proposal
text changes after the initiation of a Decision to adopt it.

Also note, that if we accept that a Proposal text can be changed, it goes
against the precedent that "any body of text (such as a rule) conforming
to the first paragraph of R106 is a proposal" because this latter
precedent coupled with mutability would imply that a proposal, when it
changes it text, actually turns into a copy of a previously-existing
platonic proposal and so is no longer its own proposal.  But that's
another matter for philosophers to moot on another day.