Re: [Ag Meinungsfindungstool] [MG] Decision Rule

["marc" <marc AT merkstduwas.de>]

Hi Ned,

Thank you so much for red flagging this!

This is a very important point: currently the AG-MFT model locates (NUO) 
within the Discussion System area.

That means the "vote arrow" is NOT valid for (NUO) driven decisions! 
Currently for (NUO) the *participation* is *implicit*, as you already 
pointed out earlier.


As far as I can say, there are currently two possible workflows of making a 
decision using the AG-MFT model:

(1) The (CV) run of things:
   (a) In the Discussion System a discourse is going on (, and interrupted 
by some trigger).
   (b) The Resolution System is notified about the decision to be made.
   (c) The Resolution System executes the decisio-making process using the 
(CV) methodology.
   (d) If there is a decision, the Information System is notified about the 
resolution, else the resolution is rejected and send back to the Discussion 
System.

(2) The (NUO) run of things:
   (a) In the Discussion System a discourse is going on (, and on, and 
on..).
   (b) The Resolution System is notified about the decision to be made.
   (c) The Resolution System executes the decisio-making process using the 
(NUO) methodology.
   (d) If there is a decision, the Information System is notified about the 
resolution, else the resolution is rejected and send back to the Discussion

IMHO the relevant difference between both is, that with (2) the execution of 
(a) and (c) is definitely in parallel, while (a) might be on hold with (1).

But I am going to discuss this within the AG-MFT weekly mumble meeting on 
Monday.

In fact the domain of AG-MFT is currently the Position Forming (Discussion 
Systems). Our main focus is not on Decision-Making (Resolution Systems) 
right now.

Even if we already know that there will be an extensive interaction between 
all systems taking part in the decision-making process, we did not specify 
the interfaces between them yet.

Cheers
marc


-----Original Message----- 
From: Ned Conner
Sent: Thursday, January 16, 2014 10:14 PM
To: Metagovernment Project
Subject: Re: [MG] Decision Rule

Hi Marc -- Thank you for that excellent response! (more flowers for you!)

I will respond to several of your points, each in a separate email in
this thread. But, they are each major points, so that we should probably
spin each of them into its own separate thread pretty quick.

------------------------------------------------------------------------

   Marc> I thought the thread is about choosing either CV OR NUO and
   then apply that one anytime and everywhere for any single decision
   from here on, e.g. to develop a software system for decision-making
   that implements solely this single decision rule, but not both.

Ned> Actually (up until your excellent response) I had taken the AG_MFT
"Big Picture" to be doing exactly that: developing a software system for
decision-making that implements solely (CV) at the end of the decision
process, for every decision.

The Big Picture arrow to the Resolution System is quite clearly labeled
"Vote", which had suggested to me that labeling that arrow "Propose"
instead of "Vote" was not being considered, for any decision.

I had taken the "Big Picture" to indicate that, at the end of the whole
process, an instantiation of the (CV) rule would be used to actually
make the decisions, without exploring the possibility that it might
actually work better to use an instantiation of the (NUO) rule to
actually make the decisions at the end of the process. That is why I was
shouting: Wait! First we need to experiment to see which sort of rule
will give us better decisions at the end of the process!

To avoid that confusion, you might want to change the label for the
arrow that points to the Resolution System from "Vote" to something more
neutral, such as "Participate", and maybe remove the "Individual Voting:
Direct/Delegation" label. Otherwise, AG_MFT is telling the world that it
is going to be making the final decisions using (CV), and is not the
least bit open to using (NUO) instead, even if the experiments show that
it works better.

------------------------------------------------------------------------


On 1/16/2014 8:07 AM, marc wrote:
> Hi Ned,
>
> "Thank you for the flowers" as we use to say in Germany (which indeed mean
> "Thank you for the compliment", but with a twinkle).
>
>
> Ned Conner wrote:
> >    Ned> Which raises the question: Why would we put time and effort
> >    into discussing all of these topics?
> >
> >        My answer: To choose between the abstract "No Unresolved
> >        Objections" decision rule and the abstract "Count Votes"
> >        decision rule is (in Thomas Kuhn terms) to choose between two
> >        entirely distinct decision making paradigms; two distinct ways
> >        to do democracy. In my view, there is no more fundamental choice
> >        that we Metagovernment Mavens could be discussing.
> >
> >    Marc> Well, is this really the one-million-dollar-question: Which
> >    decision rule to choose?
> >
> >        * (NUO) "No Unresolved Objections"
> >        * (CV) "Count Votes"
> >
> >        IMHO there is no need to choose one rule over the other!
> >
> > Ned> I am not sure that I am following you here. It appears to me that in
> > every particular case that an actual decision is actually made, a 
> > decision
> > rule has been used in making that decision. The only choice is whether 
> > the
> > decision rule is explicit or (wink wink) implicit. We always in fact do,
> > in every particular case, choose one rule over others.
>
> That's exactly what I meant when I said "[...] that we should implement 
> *both* rules and choose which one to use depending on the current context 
> of a decision to be  made on demand".
>
> But maybe I didn't got the point of this thread correctly (that's the 
> danger on *just* jumping in ;o).
>
> I thought the thread is about choosing either CV OR NUO and then apply 
> that one anytime and everywhere for any single decision from here on, e.g. 
> to develop a software system for decision-making that implements solely 
> this single decision rule, but not both.
>
> If this is not the case here, sorry for the inconvenience and please skip 
> the rest of this email!
>
>
> >     Marc> Because both rules not only typify different paradigms for
> >     decision-making, but they focus on entirely different aspects of the
> >     democratic process also! Hence in the end they are not comparable to
> >     each other. Thus no decision possible. (This reminds me of the
> >     Mission Impossible tune - really no clue why! ;o)
> > Ned> Again, I am not quite following you. The claim that the two abstract
> > rules "focus on entirely different aspects of the democratic process" is
> > true in the AG_MFT model because the model declares it to be true. But in
> > other equally democratic models (such as the Blinap model), the claim is
> > not true, because those other models declare it to not be true.
>
> IMHO the different aspects are not only different in the context of the 
> AG-MFT model. I think they are universal different in general. It is the 
> scalability of the decision rule that makes it act as a different aspect 
> on the democratic process, which might be encountered by the emergence of 
> the decision rule and the decision-making process itself.
>
> > The AG_MFT model separates the democratic process into:
> >
> >     * opinion forming
> >     * position forming
> >     * decision making
> >
> > The Blinap model contains no such separation of the democratic process
> > into three parts. Instead, decisions are arrived at through successive
> > recursive proposing. (If you think that Blinap actually does contain the
> > three parts as separate parts, show me.)
>
> At the beginning of this week the AG-MFT extended the model by a 
> "Coordination" system. This is the fourth aspect of the 
> decision-making-process, which indeed can be seen as a cross cutting 
> concern of the other three aspects 
> (http://wiki.piratenpartei.de/wiki/images/7/72/MFT_BigPicture_v01.jpg).
>
> Just for clarification: The AG-MFT model just states that there are 
> different aspects of the decision-making process that a system *could* 
> focus on in general. In an ideal world those different aspects might be 
> represented by different and seperate systems that interact with each 
> other. But in reality each existing theory or system will be part of at 
> least one of those aspects.
>
> Unfortunately I don't know much about Blinap in this regard, so I can't 
> say which aspects of the AG-MFT model it really covers. But IMHO the 
> seperation just exists; independently from any existing theory or system.
>
> One of our ambitious aims is to cooperate with (nearly) any other existing 
> system or theory in the area of decision-making. Therefore we are very 
> keen to learn how Blinap fits into our model.
>
>
> > The argument that you need to make is that the AG_MFT model of the
> > democratic process is "better" than other models such as Blinap, not that
> > the democratic process can only be what the AG_MFT model declares it to
> > be.
>
> Interesting point! The AG-MFT does not try to define a *better* model - 
> quite the contrary we try to define a *common* model. Another ambitious 
> aim of ours is to enable cooperation between all those competing systems 
> and theories around.
>
>
> >     Marc> From my point of view there are two serious differences; a
> >     qualitative and a quantitative one:
> >
> >         * (1) While (NUO) focuses on equity, (CV) strives for majority.
> >         * (2) While (NUO) solely suits well for small groups, (CV) can
> >           also handle large scale effectively.
> >
> > Ned> I think I may agree that "(NUO) focuses on equity", but I would need
> > to know more about what you mean by "equity" to know for sure.
>
> Bingo! IMHO this is one of the elementary questions! What is "equity"? To 
> be honest; I can't tell you! But, I believe that we all together as a 
> society need to define what we would like the answer should be...
>
> ... for me this includes ethical and morally guidelines of how to reach a 
> *wise decision* that respects (nearly) all facets of the domain where the 
> decision is to be made.
>
>
> > I would agree that (CV) does strive for some measure of vote-count
> > popularity, but "majority" is only such measure among many.
>
> But at the end any "vote-count popularity" is a kind of measurement, isn't 
> it? That's my point here:
>
>    Thesis I: CV is about quantity.
>    Thesis II: NUO is about quality.
>
>
> > I agree that (NUO) can suit well for small groups.
>
> Great! We have consensus ;o)
>
>    Thesis III: NUO can suit well for small groups (due to Thesis II).
>
>
> > I *STRONGLY* dispute that (CV) can successfully scale to large
> > populations.
>
> Great also! Here we have dissent ;o)
>
>
> > (I keep posting this quote only because no one has so far even tried to
> > counter it. We really need to "grok in fullness" that "count votes"
> > systems have been around for thousands of years, and have *NEVER* shown
> > themselves to be sustainable at the scale of large populations -- for
> > reasons that are intrinsic to voting, irrespective of the technical
> > particulars of the vote-count system.)
>
> IMHO for now just the following fact seems to be relevant:
>
>    Thesis IV: CV might scale better on large groups than NUO do (due to 
> Thesis I).
>
>
> >     Marc> This leads me to the assumption that we should implement
> >     *both* rules and choose which one to use depending on the current
> >     context of a decision to be made on demand.
> >
> > Ned> I totally agree with you that we need to run the experiments,
> > preferably in parallel (resources permitting).
> >
> > But, you appear to be asserting that you already know which particular
> > decision rule will work best in which sort of context. I don't think that
> > you do. Show me the data. (I state this provocatively to emphasize that I
> > really think that WE NEED TO RUN THE EXPERIMENTS in order to have ACTUAL
> > EMPIRICAL SUPPORT for any claims regarding which particular decision 
> > rules
> > will work best in various particular contexts.)
>
> No, I don't know! And I guess nobody does! This is why the AG-MFT has 
> started last year to prototype several different systems as proof of 
> concept implementations. We don't know if our model holds or not. So we 
> need to eat the pudding!
>
>
> >     Will Rogers> It isn't what we don't know that gives us trouble, it's
> >     what we know that ain't so.
>
> and even more enlightening: as soon as man thinks, he is wrong already!
>
> To cut a long story short: IMHO it is all about quality and quantity where 
> NUO is towards quality while CV is towards quantity.
>
> But to get to a certain level of quality using NUO will take its time. 
> Hence I think the following might be true:
>
>    Thesis V: CV seems to increase the speed of decision-making.
>
>
> IMHO given these five theses we can start our experiments...
>
> Cheers
> marc

_______________________________________________
Start : a mailing list of the Metagovernment project
http://www.metagovernment.org/
Post to the list: Start AT metagovernment.org
Manage subscription: 
http://metagovernment.org/mailman/listinfo/start_metagovernment.org