Your proto clarified the underlying issue for me. For the most part, we
treat in-proposal textual substitutions (e.g. your "inserts") as part of
the sequence of the proposal "taking effect". In other words, the
following would work:
////
1. Every player with property X hereby wins the game.
2. If any player has won the game as a result of this proposal, amend
Rule Y to read as follows, substituting the player's name for [name]:
{ [name] is the ruler of Agora}.
////
I think no one would have trouble saying that the substitution happens
*at* step 2, after the Win, as part of the proposal taking effect.
However, the new comments clause, by using "before", has added some kind
of "pre-processing" step to R106:
> Clearly marked comments are considered
> to be removed from the proposal before it takes effect, unless
> otherwise stated by the proposal.
This introduces the concept of "marking" (textual/formatting clues that
may or may not be part of the body of the proposal), but also a new step
of "scanning the proposal for comments and removing them" before the
proposal takes effect.
And we don't know a lot about this procedure. Are substitutions allowed?
Are they simultaneous or applied in sequence? Do you have to specify
"pre-processor" directives separate from main effects? If the proposal
"otherwise states" something about comments, when are those statements
"executed"? Can "main effects" work retroactively and change things into
comments that are then (retroactively) removed? (The words "clearly" are
not really useful here - as discussed during your win-by-apathy thing,
"clear" allows for all sorts of substitutions in other contexts.
Depending on how these questions are answered there can be all kind of
paradoxes and other textual play introduced in some way...
On Tue, 18 Sep 2018, D Margaux wrote:
> For example, this protoproposal would
> resolve the bleach issue:
>
> ////
>
> 1. For the purposes of paragraph 2 of this proposal, let the following
> terms have the following definitions:
>
> “Execute X” means to insert into the proposal the value of X, where X
> is a text variable with a value set in this proposal.
>
> “L” is a text variable with the following value: “All text to the
> right of the double pipe marks is a comment. The following players win the
> game: [names].”
>
> “R” is a text variable with the following value: “All text to the left
> of the double pipe marks and to the right of the paragraph number is a
> comment. The following players win the game: [names].”
>
> “double pipe marks” is the name given to two back-to-back instances of
> the symbol “|”.
>
> 2. Execute L || Execute R
>
> On Tue, Sep 18, 2018 at 8:32 AM D Margaux <[email protected]> wrote:
>
> > One way to make it more airtight is to define variables for each string,
> > and then have the proposal execute them.
> >
> > On further consideration, I think the whitespace/bleach solution has a
> > flaw. The rule says that whitespace “generally” doesn’t matter. This
> > situation is one where that “general” rule probably doesn’t apply.
> >
> > On Tue, Sep 18, 2018 at 4:51 AM Cuddle Beam <[email protected]> wrote:
> >
> >> I love the proposal, I just I dislike the ambiguity because it might
> >> defuse
> >> it and then there's no fun lol
> >>
> >> I like the whitespace angle too, it opens the doors to another style of
> >> "paradox" (and scam, quite likely).
> >>
> >> Maybe we could figure out a way to make L||R more watertight and resubmit
> >> it.
> >>
> >> On Tue, Sep 18, 2018 at 2:44 AM Kerim Aydin <[email protected]>
> >> wrote:
> >>
> >> >
> >> >
> >> > On Tue, 18 Sep 2018, Alex Smith wrote:
> >> > > This may actually suggest a general principle for avoiding ambiguities
> >> > > like this one: as the formatting of text is irrelevant, an attempt to
> >> > > specify something in a way that doesn't have a clear "start-to-finish"
> >> > > order is failing to uniquely specify the text in question, and thus is
> >> > > too ambiguous to work.
> >> >
> >> > First, non-"start-to-finish" text has been allowed to have effects
> >> > before: https://faculty.washington.edu/kerim/nomic/cases/?1267
> >> >
> >> > But here's the new bit! The following text that was added to R106 in
> >> > July this year:
> >> >
> >> > > Clearly marked comments are considered
> >> > > to be removed from the proposal before it takes effect, unless
> >> > > otherwise stated by the proposal.
> >> >
> >> > (at least that's the text I was trying to get at, D. Margaux's Bleach
> >> > argument is a clever alternative).
> >> >
> >> > The thing is, marking comments often relies on textual formatting, at
> >> > least for end-of-line comments (not so much for parenthetical comments,
> >> > perhaps). So what does the "marking" in "clearly marking" mean? Can
> >> > we use formatting clues? (formatting clues *do* "mark" things). And
> >> > if a comment marks itself as a comment (but it is otherwise "considered
> >> > removed") it might still count as a comment "before" it is removed.
> >> >
> >> > Ignoring the rest of the rules and focusing on this clause, I could
> >> > imagine at least four outcomes:
> >> >
> >> > 1. The comments are ambiguous, therefore not "clearly marked",
> >> > therefore nothing in the proposal is removed, and there's enough
> >> > textual clarity (e.g. the marks demonstrated intended readings) in
> >> > both halves that both halves are effective.
> >> >
> >> > 2. As #1, but you're forced to read L to R which leads to
> >> > ineffective nonsense.
> >> >
> >> > 3. The comments mark each other, and then are considered removed,
> >> > so the whole thing is a big comment and does nothing.
> >> >
> >> > 4. Similar to CFJ 1267, we decide that one portion is "naturally in
> >> > front" of the other (probably the left portion, given CFJs 3534-3535)
> >> > and that portion succeeds in removing the other portion as a comment.
> >> >
> >> >
> >> >
> >> >
> >>
> > --
> > D. Margaux
> >
> --
> D. Margaux
>
