Re: What can a Society Do?

From: Anders Sandberg (
Date: Wed Sep 12 2001 - 05:22:52 MDT

On Wed, Sep 12, 2001 at 11:17:01AM +0800, Chen Yixiong, Eric wrote:
> From: Extropians List
> Originally: RE: TERRORISM: looking for solutions
> The concept of trustability has its roots in Game Theory.
> >From what I know of Game Theory, I can draw some conclusions:
> 1) Without trust, a society will function only on its lowest
>denominator, if we consider it a society at all. People will only
>look out for their self-interests and would not perceive to
>preemptively strike those they perceive as a threat, to ensure
>their own survival.
Yes. This seems to be in line with the work of Francis Fukuyama.

> 2) To create trust without a free information flow, the most
>common solution would require establishing a hierarchy. In such an
>organization (e.g. Governments), by virtue of the inequality of
>power, can track and punish individuals who infringe on trust and
>thus maintain a working level of trustfulness (and also a related
>concept called security).
The drawback is of course that even benign hierarchies have a
tendency to become seriously self-serving, as described by the
public choice school in economics. And due to their basis in power
inequalities rather than respect for rights there are few factors to
correct for this.

> 3) A society without any structure, if it has to function
> efficiently, would have to self-create a Government. thus,
> anarchismw would not work because it would sow its seeds for its
> own destruction. Humans have an instinct to form groups (of
> whose's members they can trust much more than others) thanks to
> evolution selecting out anti-social humans.
To my knowledge, anarchists are seldom atomists. All serious
anarchist social ideas (be they anarchocapitalist or left
anarchists) include various forms of *voluntary* social
organisations (companies, communes) that would help society
function. Anarchocapitalists expect trust through contract
and legal systems, leftist anarchists expect trust to develop due to
self interest and individual social maturation.

Note that you have so far ignored the spread of information and the
effect of including individual rights into this framework.

> Notes: Godel's Incompleteness Theorem forbids any system that
> claims to cater for all people in all situations.

How do you come to this conclusion? It just claims that a formal
system powerful enough to handle arithmethic is either inconsistent
or incomplete. Human societies are not formal systems in the first
place (despite the efforts of constitutional lawyers), and obviously
inconsistent and random to some extent.

You might be thinking of the Arrow theorem?

> Surely, the claims
>by capitalist-anarchists do not work because of such. While the
>Theory of Evolution may *appear* to apply to human society, it
>(as the Theorm predicts) to apply to human societies because humans
>can and will think out of the system. Placed in simple terms,
>can and do exploit the social systems they encounter. This ability
>confers the ability to lie, to see through paradoxes and to
>"make-believe", and confers the greatest difference between a human
>and a Turing machine. Evolution Theory applies well only to beings
>simple enough to remain within the system rules.

Why wouldn't a Turing machine be able to lie, make use of the system
to suit its purposes and handle paradoxes? My artificial life
simulations and neural networks do that every day (in a limited
form), while running on a finite Turing machine. Look out for the
Penrosian "we are better than machines" meme!

You are also wrong in claiming it is the theory of evolution is
inapplicable because humans are so clever and unpredictable agents.
While there are valid arguments against memetics as a good model of
human behavior (such as Torsten Nahm's speech at TransVision 01),
human behavior does appear to fit a Popperian evolutionary
epistemology, where we learn and act in a feedback loop. It is very
adaptive, leads to evolutionary change but can be remarkably
Lamarkian if needed.

> 4) The fastest way to create trust would require the removal of
> the ability to make deceit in all of us, or our disire to do so.
> While we know this has high impossibility, we can approximate
> something like that by:
> a. Having a free flow of *all* information, including what we
consider as private
> b. Having a group of people capable of handling such free flow of
information (such as those not prone to gossip)
> c. the exclusion of people who refuse, or has insufficient ability
to participate in such open-ness

I.e. what you suggest is a voluntary transparent society? Halperin's
_The Truth Machine_ might also be approaching, given the advances in
brain fingerprinting.

However, this might not be the fastest way of creating trust, since
it would require a serious re-training of many cultural aspects of
the participants, and if it is to be relevant has to encompass a
sizeable number of people.

> Project Sociologistics aims to create a social system (in theory)
>that supports such a fundamental "right", as long as such pursuit
>self-actualisation aims to understand the universe and conducted in
>an ethical manner.
You run into the usual problem of positive rights here, of course.
Should society support me if I claim my personal growth hinges on
counting grains of sand? It might even be true, but it is clearly
unproductive for everybody else.

> 2) Freedom from Mundane Issues

Note that the problems you name in this section are those which are
commonly dealt with in liberal studies through the open society. By
having a society with strong member feedback, enforced
accountability and the freedom to adapt to meet new challenges, most
of these problems can be ameliorated.

> 4) Freedom to Trust
> What good would we, to live in a society of whom members we cannot
>trust. This freedom provides for high levels of trust among a
>society's own members, such as by disclosure of all data each one
>has on everything. It facilitates, via unity, the ability to
>projects that require many people. If everyone will trust each
>like a mother with a child, our societies will have much lesser
But is this also a freedom not to trust? If I know enough about a
certain person, I might decide not to deal with him since I consider
him untrustworthy or his interests counter to mine.

> 5) Freedom of Non-Interference
> This applies to the society both ways, in that it will not
interfere in other's affairs if they don't interfere with it, and
the hope that others will also comply with this rule. This also
applies on the microscopic manner, where one seeks not to interfere
with each other's activity unless one has his or her interests

I.e. the standard liberal dictum of "my rights end where yours

> Clearly, creating an independent rational society by taking over
>another nation would just spur military action, but also face
>uncooperation among the irrational people who would break the
>of trust. Hence, I recommended, and continue to recommend, the
>setting up of a colony in unclaimed or purchased land or space.
>Making a colony work gets difficult enough, but having to fend off
>military attacks while doing so, will make it almost impossible to
>work. Converting an irrational to an rational one would sounds so
>impossible that I think this problem does not have enough worth to
>justify assigning our computing resources to solve it.
As I have previously pointed out, this kind of emigration strategy
is unlikely to be as efficient as actually changing existing
societies. You don't need to make everybody and the entire society
rational in order to reap substantial benefits.

It might be worth pondering why democracy has spread worldwide.

> > hence, I hope with this posting, to provide some hints on how
we can construct a society based on trust. I hope it helps. Please
read the paper, or part of it, before formulating the reply, as you
could contribute redundant points that I would find reluctant to
reply due to insufficient time.

A bit too late for that now, isn't it?

Anders Sandberg                                      Towards Ascension!                  
GCS/M/S/O d++ -p+ c++++ !l u+ e++ m++ s+/+ n--- h+/* f+ g+ w++ t+ r+ !y

This archive was generated by hypermail 2b30 : Fri Oct 12 2001 - 14:40:29 MDT