Summary
of Chapter 1 of Gerry Mackie’s Democracy
Defended, “A long, dark shadow over democratic politics”
·
Mackie starts the book by highlighting a minor ‘paradox’
relating to the role of democracy in the world as of 2003 (could be written
now, more or less): whilst ‘representative democracy’ seems to be the most
successful system of government in the world – on the march for decades, only
now experiencing a deceleration – it has, in this period of its greatest
triumph, come under considerable attack by liberal political theory produced by
academics holding citizenship in the exemplar democratic states.
·
“I do not know why, but from the beginning
academics have tended to be more disdainful of democracy than are, say, the demos” [2]. This is a very stupid
sentence. It’s a silly claim (sweeping vague claims like that are always silly
because their truth conditions aren’t
clear (people can reasonably disagree on whether datum x is a truthmaker for such claims), meaning that disagreement about
truth value is intractable), and,
even if you assume that the claim is ‘true’, there’s a very obvious candidate theory.
·
Mackie’s quick summary of 20th
Century academic attitudes to democracy:
“The US
had more of a democratic tradition, personified by Dewey. Dewey’s most
influential rival was Lippmann, who argued that the citizenry is ignorant and
that experts must rule in spite of the “democratic fallacy” (Wiebe 1995). In
Europe during the interwar period Lindsay (1935) and Barker (1951) were
virtually alone as academic defenders of democracy. In the period after World
War II, an exhausted conformism in American culture was accompanied by an
empirical democratic theory that apotheosized the “beneficial apathy” of the
citizenry, and by positivistic animosity to normative theory; Dahl (e.g., 1956)
was nevertheless a milestone in democratic theory. In this period, although
little good was said about democracy, not much bad was said about it either.
The revival of liberal political theory following Rawls (1971) was kinder to
democracy, but was much more liberal than democratic: for Rawls (1993,
231-240), the Supreme Court is the exemplar of public reason, not the
parliament, not the people. After Habermas (1984; 1987), an emphasis on the
transformation rather than the mere aggregation of preferences stimulated wider
academic interest in democracy (Elster 1986b; 1998). A robust normative
democratic theory, primarily but not exclusively on the theme of deliberation,
is beginning to appear.” [2]
·
Mackie introduces the primary subject matter of
the book:
“Although
democratization is the main trend in the world today, the main intellectual
trend in American political science is the view that democracy is chaotic,
arbitrary, meaningless, and impossible. This trend originated with economist
Kenneth Arrow’s impossibility theorem, which was applied to politics by the
late William Riker, political scientist at the University of Rochester. The
earlier academic attack on democracy by Mosca, Michels, and Pareto was revived
with fashionable new methods. Riker had great organizational resources, and
used them to promulgate a particular interpretation of Arrow’s theorem, to further
elaborate a doctrine he called “positive political theory” (“scientific,”
rather than “ethical”), and to recruit and place his students far and wide.
Riker calls populist
any democratic theory which depends on a systematic connection between the
opinion or will of the citizens and public policy, and liberalist any democratic theory which requires only that voting
result in the random removal of elected officials. Riker rejects populist
democracy as infeasible, and offers his liberalist democracy in its place. What
almost everyone means by democracy is what Riker calls populist democracy; and,
I shall argue, Riker’s liberalist alternative fails, descriptively and
normatively. […] Riker’s irrationalist doctrine emphasizes principled failings
of democracy and recommends a constitutionalist libertarianism and the
substitution of economic markets for much of political democracy (Riker and
Weingast 1988).
[…] The proposition that democratic voting is arbitrary
and meaningless can be used not only to justify a constitutional libertarianism
such as Riker’s, it can also be used to justify a dictatorship that appeals to
the values of stability and order. The irrationalist doctrine is taught in
America’s leading political science departments, law schools, and economics
departments. Students absorb these teachings, and then move on to join the political
and economic elites of the world. I shudder to think of the policies demanded
in the international consultancies and financial agencies and the national
treasury departments of the world by people who were taught the findings of
Arrow as interpreted and expanded by Riker’s school of thought. I worry that
authoritarian movements might find comfort in Riker’s (1982) irrationalist
credo, Liberalism against Populism.
One purpose of my work here is to show that Riker’s irrationalist doctrine is
mistaken, and thereby to restore democracy as an intellectually respectable
method of human organization.” [2-4]
·
Mackie’s first introduction of social choice
theory, a sketch of the “problems of voting” (the text has tables to help the
reader but they are just the standard ones you can find on Wikipedia, so I feel
no contrition for excluding them (this section probably didn’t need summarising
precisely because of the abundant online resources on this simple maths but I’m
doing it anyway, in the interests of ‘comprehensiveness’)):
“Ordinary
majority rule seems to the most natural, or commonsensical, way of voting […]
When there are three or more alternatives there can be problems with majority
rule. If there are three candidates, and none receives a majority, then there
is no winner, and the method is incomplete. Perhaps without too much thought we
might turn to plurality rule as a simple extension of majority rule: whoever
gets the most votes, even if short of a majority, is the winner.
[…]
There can be a problem with simple plurality rule, however. Suppose that there
are three candidates A, B, and C in an election, and 100 voters. For
simplicity, everyone has strong preferences [here Mackie is stating a crucial formal
assumption of all this theoretical work in social choice theory, along with
decision theory (there can only be decisive preference or indifference, no
shades)]. Faction 1 is made up of 40 people, and ranks the candidates A > B
> C. Faction 2 is made up of 35 people and ranks the candidates C > B
> A. Faction 3 makes up 25 people and ranks the candidates B > C > A. With
plurality rule, everyone casts a vote for their first-ranked alternative. […] A
would win by plurality rule, even though 60 percent of the voters are against
A. […]
Borda wrote on the theory of elections in 1784 (see Black
1958; McLean and Urken 1995). Borda noticed this defect with plurality rule,
and proposed his method of marks, which we shall call the Borda count, to
remedy the defect. Borda thought we should count whether alternatives are
ranked first, second, third, and so forth. He proposed that [for n alternatives we should assign n – x points to each voter’s x-ranked preference, i.e. such that the first-ranked
preference always gets n – 1 points
and the last-ranked preference always gets 0 points (n – n). A quick
calculation shows that the ‘Borda method’ outputs candidate B as the winner in
our example set up above, which seems to most people like the ‘right result’
(seeing as no faction places B last).] [What you just read is not really a
paraphrase of Mackie; he takes several more sentences to explain the method
since he avoids algebra (and he shows the Borda count calculations on the page)].
[…]
Condorcet, another French thinker, wrote on the theory of
elections in 1785 (see also McLean and Hewitt 1994; McLean 1995). Condorcet
proposed as a criterion that the alternative that beats all other alternatives
in pairwise comparison should be the winner. In our example, examining the
italicized cells in the matrix, B > A, B > C, and C > A, or B > C >
A. In this example (and in most practical circumstances) the Condorcet winner
and the Borda winner coincide. They need not, however. Condorcet objected to
the Borda method on the ground that it is possible for it to violate a
condition that later came to be called the independence
of irrelevant alternatives [my italics (assume in future that they’re his
italics unless mentioned like this)]. [Imagine another three-candidate scenario
where the first faction of 51 ranks A > B > C, the second faction of 35
ranks C > B > A, and the third faction of 14 ranks B > C > A.] By
the Condorcet method, the social ranking [in the scenario you just imagined] is
A > B > C, the same as the ranking of the faction with the slender
majority of 51. Observe, however, that A is the last choice of 49 voters. The
Borda method takes that into account and reports a social ranking of B > A
> C. The dispute is this: Condorcet insists that in pairwise comparison A
beats every other alternative, Borda insists that B gets more votes over every
other alternative than does any other alternative. The Borda method violates
the independence condition because in deciding the social ranking between two
alternatives X and Y it takes into account individual rankings of alternatives
other than X and Y, such as between X and Z and between Y and Z [as soon as I
learned about this condition, I thought that it was silly and that its
violation by the Borda method was a non-issue; as we will see, Mackie 100%
agrees with me on this and uses it as his most serious objection to Arrow
himself (as opposed to Riker’s interpretation of Arrow), since, of the ‘crucial’
conditions for a rational voting system set up by Arrow, the Borda count only
violates IIA. Unfortunately, the reason Mackie wrote his book is that most in
the field don’t think IIA is silly (probably there’s a big selection effect at
play, since the Impossibility Theorem is virtually the central monument (‘crowning
glory’) of political ‘science’ and the person who thinks it’s of very little
significance is much less likely to go into this field)]. […]
There’s also a problem with the Condorcet method, however,
known as Condorcet’s paradox of voting. Suppose there are three (or more)
alternatives and two (or more voters). Given three alternatives, there are six
possible strong preference rankings […]. Given three voters, one each with cyclical
rankings [A > B > C, C > A > B, and B > C > A], the result of
voting by the Condorcet method over three alternatives is inconsistent, that
is, A beats B, B beats C, and C beats A [this is called intransitivity]. […] Arrow’s possibility theorem can be understood
as a generalization of Condorcet’s paradox, applying not just to simple voting
but to any social welfare function that aggregates individual orderings over
alternative social states. The Arrow theorem requires that the social ranking
be transitive, not intransitive as is the cycle. The Borda method would count
the cyclical profile in this paradox example as a tie, A ~ B ~ C, and thus
would not report an intransitive social ranking, but the Arrow theorem also
requires that a voting rule not violate the independence of irrelevant
alternative condition, thus disqualifying rules such as the Borda count [I just
said this]. Historically, Arrow’s theorem is the consequence of
noncomparabilist dogma in the discipline of economics, that it is meaningless
to compare one person’s welfare to another’s, that interpersonal utility
comparisons are impossible [this is an explanation of the one consideration used
to defend the non-silliness of IIA].
Cycling is one problem with Condorcet voting. A second,
and related problem, could be labeled path
dependence. What if there were first a vote between A and B, which A wins,
and second a vote between A and C, which C wins? It seems that we have voted
over all three alternatives and that we have a winner, C. We neglected,
however, to vote between C and B, which B would win, and which would have
disclosed the cycle to us. Unless we take pairwise votes over all alternatives
we might not notice the cycle, and normally we don’t take all pairwise votes.
To make things worse, what if Louis controlled the agenda, and arranged for
that order of voting, A against B, and then the winner against C? Then Louis
would have manipulatively brought it about that his first-ranked alternative,
C, won, arbitrarily, and voters
Huebert and Deuteronomy might even not have noticed.
A third problem is strategic
voting. Suppose again that we have a cycle as above, and an agenda as
above, A against B and then the winner against C. Then Huebert would have an
incentive to vote strategically in the first round: rather than sincerely
voting for A over B, Huebert strategically votes for B over A. B wins the
contest in the first round, and beats C in the second round. By voting strategically,
Huebert has avoided the victory of his third-ranked alternative C and brought
about the victory of his second-ranked alternative B. Inaccuracy is a fourth problem. I showed already that the Borda and
Condorcet procedures can select different social outcomes from the same profile
of individuals’ preferences. If apparently fair voting rules each select a
different public good from the same voter profile, then arguably the public
good is arbitrary. Inaccuracy, agenda control, and strategic voting also raise
the possibility that a social outcome might tell us nothing about the sincere
individual preferences underlying the outcome. Based on these and further
considerations, Riker’s hypothesis is that democratic politics is in pervasive
political disequilibrium.”
·
The next section of the chapter is Mackie’s “hall
of quotations”. In order to establish that there really is a serious “trend to
democratic irrationalism in academic opinion”, he includes a long series of supporting
quotations from texts written since the 1960s by economists, sociologists,
historians, legal theorists, political scientists and philosophers. I’ll take
the top quote from each of the six pages, to give you a taste of the taste that
Mackie gave me (to prepare yourself, you should know that all these quotes are
very stupid and often vaguely insane-sounding, and it annoys me that so many
academics are this dumb)[1]:
“The
fall of the Weimar Republic and, more broadly, the collapse of many other
constitutional democracies with the rise of fascism and bolshevism in the interwar
period alerted the [political science {Mackie’s brackets; mine will be curly
for the moment}] discipline to the terrible consequences of unstable
democracies. Later, Arrow’s impossibility theorem, a key instance of incisive
analytical work on the core problems of liberal regimes {????}, set forth the
theoretical challenge in stark terms {no it didn’t, it’s maths}. Instability is
an immanent feature of liberal democracy. Under broad conditions, majority rule
leads to the cycling of coalitions and policy; only nondemocratic practices can
alleviate this deep tendency, convoking a tradeoff between stability and
democracy.” [Katznelson and Milner 2002, 17-18]
“How can
we define and give expression to the collective wishes of a community? Arrow’s
argument shows that our intuitive criteria for democratic decision cannot in
fact be satisfied… Put crudely, what Arrow has done is to show that strict
democracy is impossible {very crude!}.” [Runciman 1963, 133]
“Arrow’s
contribution provides incontrovertible support for market process and
encouragement for those who seek to constrain the range of collective choice to
the limited functions of the minimal state {yes, that comes straight out of the
maths!}.” [Rowley 1993, xiii]
“Accurate
preference aggregation through politics is unlikely to be accomplished in the light
of the conundrums in developing a social welfare function.” [Riker 1982; Arrow
1963/1951]
“The
rhetorical convention of discussing “the majority” makes no sense {ummm….}.
When there exists a modest diversity of preference, which is, after all, the
bare necessity for political controversy, then cycles are ubiquitous – there
are “too many majorities.” {As we’ll see, this is actually a total falsehood.}
The actual social state chosen by the legislature is determined, not by some
process that yields an alternative presumably better than all the rest, but by
the order in which the alternatives arise for a vote. The absence of an
equilibrium implies that the person in control of the agenda (e.g., a committee
leader) can bias legislative choice in favour of his or her most preferred
alternative. Thus, there is a fundamental arbitrariness to social choice under
majority rule… Similarly, strategic voting, typically secret, is always
possible… Although strategic voting occurs often, it is hard to discover… All
of this shows that the notion of a “will of the people” has no meaning {What’s
your philosophy of language?}… In modern political science… electoral
majorities are seen as evanescent, and the legislator himself as a placeholder
opportunistically building up an ad hoc majority for the next election… Knowing
as we do that decision are often, even typically manipulated {stop lying}, but
being unsure just when manipulation occurs, we are forced to suspect that every
outcome is manipulated… Our examples show that this problem actually arises in
practice.” [Riker and Weingast 1988, 393-396, 399]
“The
various paradoxes of collective decision making seriously challenge the presumption
that legislative changes generally represent welfare improvements, even in the de gustibus sense of reflecting changes
in public taste {completely redundant Latin?}. Enactments that instead reflect
mere cycling, or changes in the agenda setter or in political tactics, may
better be viewed as random and purposeless from the social welfare perspective.”
[Shaviro 2000, 68]
·
Mackie’s comments on the quotations: “Notice
that people seize on the disequilibrium results in order to promote their more
favored and demote their less favored institutions. Tribe uses the results to
elevate the judicial over the other branches of government; Tushnet observes
that the judiciary is just as tained. Rowley, and Shepsle and Weingast, upgrade
the market by downgrading the government; Wolff would abolish government
altogether. Arrow (1997) has recently gone on record that his theorem does not
show that democracy is impossible, since it applies to all aggregations of
individuals’ preferences, whether by one branch of government or another, and,
I wold make clear, whether by government or market. The irrationalist doctrines
I criticize are not Arrow’s, they are
based on interpretations by others of
Arrow’s theorem.
Many influential people suggest that democracy is
impossible. The main purpose of this book is to argue against that view."
·
The last section of the first chapter is Mackie’s
outline of the book’s contents, chapter by chapter. I will not summarise this
summary, for obvious reasons. Somewhat strangely, Mackie includes perhaps the
most important table in the book in this final section of the first chapter.
This table is basically a summary of Mackie’s key anti-Rikerian empirical
findings: a long list of alleged examples of political decisions throughout
time and space which instantiate cycling or agenda control, and his own
analysis, contradicting this allegation. This table is included below, using
images taken from my iPhone.
[1] It’s
a shit species we’ve got here. I find it hard to summon any positivity about
its instances.
No comments:
Post a Comment