[EM] Arrow's axioms & an alternative to elections

[Philippe Errembault]

Hi Ernest,



I hope we understand each other's. Since English is not my mother tongue, I could pass over some misunderstanding without realising
it. I will try to be clearer.



My point is that if you want to rank multi-dimensional information, you will have to project your space to a one-dimensional space.
This will be done using a function that, especially for human beings, will depends on your mood, on context, on... I do not know
what else... the fact is that this will not be stable data. Even more: the space of this information could even be non Euclidian,
because it's probably easy to determine preferences between two things, but if you try to establish the graph of all your
preferences, I'm not sure you will easily obtain an acyclic graph. Just try to rank all your friends on the basis: which you prefer
most... not quite easy, isn't it !?



So the problem is the same: If you ask people to rank a few quite simple choices, well, most of them will probably be able to do it.
However, if you then try consolidating the choices for a large group of people, the problem will become as complex as the one of
ranking all your friends. (In fact, this is not really a good example because here I have many dimensions instead of many points to
rank, but the idea is that human being is not a logical machine which easily sort things. Human brain works more like the
collaboration of conflicting tendencies; and a group works is, exactly in the same way, the collaboration of conflicting persons, so
you could not ask a group to have a simple decision process, when it is not even easy for one single person.)



This is why I came to think that a voting system is not the right way for implementing democracy, and why I tried to imagine a good
alternative process.







Ok, I will try to explain my idea: (I hoped I would try to write it down in French first... now it's to late ;-) )



First of all, about your non-deterministic argument:

I understand your point of view, and, especially for an election process, I agree that you should have a deterministic result, at
least for reasons of verification/validation, BUT this is only true if you go through an election process.



Since you agree to say that human being is *NOT* deterministic, I feel that a deterministic democratic process is quite an illusion,
which probably gives you a warm feeling of security, but that's all. The opposition between the need of deterministic results in an
election model, and the non-deterministic nature of human and human society, comforts me in the idea that the election model is a
wrong choice.



Finally, we do not really need a deterministic process. The important point is that we need to make something efficient; which tries
to take intelligently the best possible decisions; which make people happy to use it and respects the minorities; and which makes
the best use of every one without favouring someone for bad reasons. (I mean we could not favour any one more than "the number of
persons that would agree to favour him")







So, what else could we do, then?

My idea is a multiple steps proposal, which gradually goes from our current systems, to much more neural systems







First step: try to find a system that suppresses the inconveniences of both freedom to vote and obligation to vote

I do not know what happens in other countries, but here in Belgium, we are obliged to vote. In comparison, French, our neighbours,
are not obliged to vote. What are the consequences?



In Belgium, most people are obliged to vote, but are not interested. Therefore, they will vote easy. They will vote for well-known
people, they will vote for people the press presented, etc. So, supposing that only 20% (I have no idea how far this is from
reality) of population are really interested, the result of the vote will be 20% an interesting vote, 80% the vote of the press,
which is mostly the vote of the big influence groups already governing. Therefore, we have a system, which is quite stable, but
which is not very democratic, even if it looks it is. In reality, most votes have been manipulated (1).



In France, the last presidential elections have shown what happens when people are not obliged to vote. They have a two turns
election process, so at the first turn, many people were too lazy to vote, and one of the winners was extreme right. So, many people
went to vote to block him from becoming president at the second turn, which renewed president Chirac for a second mandate with an
exceptional score, which was absolutely not representative of what people really thought of him. As a conclusion, the French system
is probably not that easy to manipulate by press, but it is much more instable and random than the Belgian system.



So, what I suggest, is to have people chose one, let's say: "level one" representative, which will eventually be themselves, the
obligation they have is to give this representative, proxy to vote for them. "Level one" representatives and only them will have to
vote (and be obliged to) when there is something to vote for. Let us put restrictions on this: 1/It is strictly forbidden to chose
as representative, someone that could have a power on you. 2/ a representative can only represent, let us say, at most (e.g.) 19
persons. To be a representative, you at least need to represent yourself.



This first step would solve the problems presented above. 1/ you are sure that most representative know what the do better than
persons not interested in the voting process, 2/ the will have a weight according to the number of persons which gave them their
confidence. (First reason to have this be kind of neural), 3/ the system stability will be guaranteed by the fact that this process
keeps the relation one citizen = one vote







Second step: multi-level representation

Now that we have our "level one" representatives, we could imagine that they would themselves choose to give their representation to
"level two" representative, which would give their representation to "level three" etc. until one last level where people would have
more or less the same number as in a parliament. Let us say that a "level n" representative can represent at most 19^n peoples, and
probably there could also be a minimum number, but I'm not sure this would be a need.



The point is that now, you have multiple tiers of representation, the higher are at the level of a parliament, the lower are at the
level of a referendum, but at all levels, people are interested in the decision process. At higher level, you could ask for frequent
questions like in the actual parliament, at lower level, you could ask questions to people like in a referendum. By the way, you
would also have the possibility to ask a question at any intermediate level, which is a great improvement in the tuning of the
decision process.

One of the most important advantages of such a process, is that having more intermediate levels, gives more hindsight (not sure
about the use of this word...) to the decision process, which could make it smarter. We could also compare this principle to a multi
layer neural network.



On this should be added good criterions for choosing the level at which a question should be asked. This might be done by having the
level n decide that they want to ask one to level n-1.



I imagine we could have representatives be paid according to a logarithm of the consolidated number of people they would
represent... this is just an idea. I don't know if it's good.



Up to here, things could be done with the classical means.







Step three: where things got complicated and really neural. For this we will need the support of communication tools, like a
computer network, so then we need a secured system to avoid fraud.



Now that we have our multi layers system, we could decide to have various decision paths according to the subject. Let us say that
we would have the political topics divided like an outlined document. At the higher level you would have the whole decision process,
then you would have different ways of having it divided, by topics (education, economy, security, etc.) and subtopics, and/or by
zone (your town, your region, your community, your country, etc.). (At the finest level, you would find the laws propositions). Then
you would have to select for each topic or subtopic, to whom you would give your representation. (E.g. you could decide that you
would give your representation to someone, except for one specific subject you know well, for which you would like to represent
yourself. So then, you would have different persons/specialists/neurons implicated for each question and we could hope to have the
human society work like a huge brain. Like in a true brain where some neurons have an inhibitor role, it would be necessary to have
moderators, which would be non-specialists of a specific subject, participating with the specialists, to make the side implications
appear, and avoid decisions taken with bad side effects.



It could happen that at that level, that it would be difficult to have assemblies talking like now, just because of the difficulty
to have people, all be available at the same time, and it could happen that many discussions would go through forum tools or things
like that...



I hope this answered to your questions



Philippe







(1) In Belgium, this is especially true, since:

- Government have decided to put an electoral threshold of 5% under which a list cannot have representatives, which is a very
efficient way to block new lists and new ideas from growing.

- They installed an electronic voting system which, despite the fact that it could more easily be invisibly manipulated, also make
people feel less secure in front of a machine they do not know, so they tend to make themselves feel more secure by voting for
someone they know, which reinforce the power of press manipulation. It happened that really strange behaviours where detected, and
that the parliament decided according to the official experts report, that since the problem had a probability of occurring near to
zero it should not have happened anywhere else, and since it had been detected and corrected, there was no problem.

- They obliged the lists to have the same number of women and men, which for small lists makes quite difficult to find the right
amount of women, who are quite often not interested or do not dare to present themselves for an election. And by the way, if a small
list managed to put enough women on the list, they will probably be younger and with less experience (since the first they find will
be put on the list), so, if they manage to be elected, they will eventually be easier to manipulate by the other parties.



Therefore, my country is quite a good example of the democratic power to citizen ;-)