Comments on "Social Choice and Chance in an Open Society"

What this Page Is About

Social Choice and Chance in an Open Society is a proposal for deliberately injecting a stochastic (chancy) element into electoral and legislative processes. The paper gives a simple, practical example of how to apply chance to a committee proceeding and presents the salient arguments for doing so. The conclusion recognizes that political science is an empirical discipline and calls for experimentation on a small scale before any such scheme is promoted in public political institutions.

The ideas inspiring this proposal have an origin in the political philosphy of Sir Karl R. Popper, who coined the phrase "open society" as better than the ambiguous "democracy." (There are, of course, other sources of the scheme as well.) Accordingly, at the suggestion of J. Peizer of the George Soros Foundation, I began to monitor The Karl Popper Web, a list server operated at the Dublin [Eire] City University, dedicated to examining critically, as well as promoting, the philosophy of the late Sir Karl. The main part of this page is devoted to a lightly-edited chain of quotations from The Karl Popper Web relevant to the scheme. At the end of this page, there are references to other comments on "Social Choice and Chance in an Open Society."

Comments from The Karl Popper Web:

Barry McMullin asks what KRP might have to say about the ethnic and religious conflicts that have been damaging the open society and eroding public confidence in political processes:

Date: Mon, 05 Feb 1996 14:12:50 +0100
From: Barry McMullin <>
Subject: CRITICAL-CAFE: When "democracy" fails? (long)
(Another week, another thread...)
I have been recently musing on some problems of Government, or,
more generally, of the "open society".  Perhaps some of you can
help me out.
Karl Popper consistently advocated that philosophy should grapple
with "real" problems - problems that have effects on people in
their ordinary lives.  With that in mind, I will dare to raise a
contemporary problem which is perhaps not quite the greatest
facing modern civilisation, but is certainly the greatest problem
facing my own small country of Ireland.
I am reasonably familiar with Popper's debunking of the notion
that "representative" government (or any other kind of
government) can be perfect, or reliably deliver an open society
(in the sense of a society which, inter alia, defends and protects
equitable and just treatment of all its citizens).  I believe
that Popper suggested, nonetheless, that some form of
representative government, ultimately relying on "free"
elections, with universal suffrage, is the best (or "least bad"?)
form of government we have yet been able to formulate.
My question is this: what can we then do when representative
government does, in fact, fail - drastically, and seemingly
By this I mean that there are situations in which representative
government fails to deliver or protect an open society; indeed,
in certain circumstances, representative government can lead
directly (and almost inevitably?) to sustained oppression and
I have the situation of Northern Ireland explicitly in mind here,
but the problem is, I think, fairly generic.
It arises fairly naturally wherever you have a single state with
two (or possibly slightly more) strongly distinguished ethnic
communities, with a history of mutual conflict, and with both
communities having roughly comparable population sizes.  In
such a situation, each community quite rationally fears the
formation of a government dominated by the other community; this,
in turn, reinforces and perpetuates the polarisation of the
communities, particularly in the context of elections, lest any
fragmentation or split might allow the "enemy" achieve power.
Unfortunately, in practise, the two communities will rarely be
close enough to exact numerical equality to place the overall
outcome of elections in doubt. Indeed, the closer the situation
is to such an instability, the greater the pressure on the
members of the two communities to vote on straight ethnic
So: under "conventional" systems of "representative" government,
some one of the two communities may, in fact, quite consistently
and reliably achieve dominence in government.  Once in power of
course, there will be a very strong pressure not to alienate
one's own community in any way; that, in itself, is not wrong -
but it can very easily (and without actual malice!) tip over into
positive discrimination against the minority community.  I
suggest that there can then quite easily arise a self-reinforcing
cycle of opposition and protest by the minority community, and
repression or persecution by the majority community (the latter
all defended as being necessary to protect the "state" against
subversion, or being in the "national interest").
A traditional (and sometimes successful!) response to this
problem is territorial partition of some sort - redrawing
boundaries of a state or states.  This can certainly
substantially perturb the situation, and make actually solve the
problem.  On the other hand, it may not.
In the Irish situation, for example, one can interpret the
establishment of Irish "independence" in 1921 as a partition of
the British Isles, designed specifically to solve the problem of
ethic British ("protestant") domination of an ethnic Irish
("catholic") minority in the Westminster Parliament. The original
suggestion (and the aspiration then - and since - of Irish
"nationalists") was a simple geographical partition. The
"British Isles" consists of two large islands - Great Britain and
Ireland. The "nationalist" idea was to that these should simply
be divided into two states with separate parliaments. This was
opposed (not surprisingly) by the substantial ethnic British 
(protestant) community in the northern part of the island of 
Ireland, on the basis that this would then leave them subject to
("democratic") oppression and persecution by an ethnic Irish
("catholic") majority in a Dublin parliament.  The final
compromise was a three way partition, with three parliaments, one
in Dublin, one in Westminster, and one in Belfast (Stormont).
(OK, of course I simplify more than somewhat: the state of
Northern Ireland remained - and remains - part of the "United
Kingdom", and its local parliament was subservient to
Westminster; but Stormont had sweeping control of affairs
within Northern Ireland, and was rarely interfered with by
Westminster - until the level of violent protest in Northern
Ireland finally forced its dissolution, and a reversion to direct
rule by Westminster.)
This partition worked (?) fairly well in achieving open societies
in the consitutional monarchy of Great Britain and what
eventually became the Republic of Ireland.  Neither was left with
a homogenous and substantial (in voting terms) ethnic minority, 
and political divisions were, or gradually became, oriented along 
policy rather than ethnic lines (which is, precisely, the situation 
under which representative government can work to protect and 
defend an open society?).
However: the statelet of Northern Ireland was left with a very
substantial (roughly one third) minority ethnic population - now
effectively disenfranchised.  During the following fifty
years, electoral divisions in Northern Ireland remained stubbornly
oriented along ethnic rather than policy lines. While there is
considerable dispute about the exact scale or degree of abuse or
oppression which was then inflicted on the minority community, it
was certainly significant, and certainly became self
reinforcing. And while one can regret (and even condemn) the
emergence of violent opposition to this "democratic" regime, I
think we should certainly regard it as very understandable, and
not "obviously" mistaken. (I can still hear Margaret Thatcher's
infamous "A criminal is a criminal is a criminal" speech...)
Incidentally: the original partition agreement was very strongly 
opposed by many in the new "Irish Free State". At least some of 
these were surely motivated precisely by the expectation that
the minority in Northern Ireland would suffer oppression and 
persecution. In any case, this dispute over the partition of the
island of Ireland led to a bitter, protracted, and extremely brutal 
cival war in the new state, memories of which are still very 
vivid in many Irish families, and which still underlies major 
party political divisions. In this sense, partition certainly cannot 
be ranked as a "success" even for the southern state!
Now, one still hears calls for another further attempt at
(re-)partition to fix the problem (most commonly from Irish
"nationalists" who cling to the touching - but extremely
dangerous - belief that, in a "united" Ireland, the majority
catholic community would not indulge in oppression and
presecution of the new minority protestant population). But it
seems to me that it may be time to admit that partition is at
best a very crude and imperfect instrument - and may well have
unexpected and unintended consequences - and that perhaps the
time has come to challenge the hallowed notion of representative
government itself.
But there's the rub: I have no very coherent alternative to
offer. So I throw the question open. But please, while I have,
for reasons of personal interest and emotion, phrased this in
terms of Northern Ireland, I'd like to hear whether you agree
that it is a generic problem, and, more importantly, whether you
can offer any generic solutions.

The response to Barry's question:

[Here you may pick up the thread "When Democracy Fails" in The Karl Popper Web's Critical-Cafe Archive. Since the archive does not have a back-link to this page, you may wish to bookmark this location.]
Go to C-C Archive.

Other Comments (via e-mail):

Tom Deering of Mensa raises the problem of complexity relative to binary choices:

From: Sander Rubin <>
Subject: Voting Procedures
At 09:08 8/19/96 +0000, Tom Deering wrote:
>Sander Rubin wrote:
>> For thoughts on voting procedures, have a look at:
>I read this page, and a couple of the related ones. (Now you know why it takes me an hour 
>to go through my email each morning.)  I enjoyed the piece, but you missed one problem.
>You assume that people will reflect on issues.  I think it's clear that people do very 
>little thinking in general, never mind in the heat of an election.  Sound bites, dirty 
>campaigns and empty promises are proof.  A thinking populace would demand real quotes, 
>issue-based campaigns; and they would check if past promises were kept.  We do not.
>We use a simple binary choice because: it's simple.  A chicken can make a binary choice.  
>The math is simple, the method is simple.  Simple is the key, not fairness or accurate 
>determinism.  We demand a system that can be operated with a simple coin flip.
>Thanks for making me think,
I didn't miss that problem; it just isn't the problem that I need to address at
the outset.  I'm well aware, of course, of inertia and laziness.  And as they
pervade society, we get consequences that do damage.
The problem is: given (many but not all) people's laziness and inertia, can we 
devise rules that will favor the thoughtful and informed among them?  Who, after
all, is the "We" who are "demanding" binary choices?  It isn't unanimous, and 
those who do demand are free to put all their votes on the one choice they favor.
Turn your statement around.  The sound bites, dirty campaigns, and empty promises 
are pervasive because the system of binary voting favors that kind of political 
activity.  Would a different system produce a more rational political process
and better social choices?
Mensa is the ideal test bed for the scheme for two reasons.  1) The decisions
to be made are of no consequence in the "real world".  2) The population has
the ablitity to deal with (relatively) complex rules (though they are not
really that complex).

On the other hand, if you want to give chickens the franchise for some ulterior
purpose, think about the consequences for us all and recoil.

Return to "Choice & Chance in an Open Society".

Created: 19 Aug 96