Thursday, June 08, 2006

Group Polarization and Extremism



One of the central concerns of the liberal/ communitarian debate of the 1980’s and 1990’s was the conception of the person (or self) that political theorists functioned with. Critics of liberalism, like Michael Sandel, argued that the conception of the self liberals espoused had important consequences for the public ethic liberalism inspired (e.g. state neutrality). Unlike the unencumbered conception of the self typical of contemporary liberalism, communitarians embraced a “social” conception of the person. A social conception of the self recognises the fact that humans are social beings and as such they are embedded in a web of social networks (e.g. family, neighbourhood, national identity, etc.) that shape their constitutive ends. This interconnected web of social relations is ignored by individualistic social theories that envision persons as rational agents operating behind a veil of ignorance (Rawls) or as rational utility maximizers who are placed in a prisoner’s dilemma (Gauthier).

The emphasis on our social nature enjoys a rich history in Western political thought, from Aristotle’s account of the political nature of man to Hegel’s ethical life thesis and Marx’s account of class. The social nature of human beings is an integral part of political theories that emphasize a thick communitarian ethos, whether that ethos is one that strives to foster democracy, socialism or cosmopolitanism.

I believe that we (i.e. political theorists) should function with a social conception of the person and that such a conception can inspire a viable liberal public ethic that helps us address a myriad of concerns that arise in real, unequal, multicultural, liberal democracies that exist in an era of rapid globalisation. A lot really depends on what we mean when we say that we need to function with a “social” conception of the person. Let me expand a bit on this here and draw out one of the implications this account of the self has for democratic theory.

As social beings we place significant weight on reputational considerations. Perhaps the obvious example is impressionable adolescents who, when seeking acceptance from their peers, will engage in behaviour that contravenes the moral values their parents and teachers have vigorously tried to engrain in them. Peer pressure might even lead people to engage in behaviour that endangers their own health and safety. But feeling the pressure to acquiesce to these reputational considerations is not limited to adolescents. Adults are also swayed by such factors. From the clothes we wear to the cars we drive, we (consciously or unconsciously) are influenced by reputational considerations of various sorts.

So what does this offer us in terms of useful insights for political theory, you might wonder? It has a number of important uses. These kinds of considerations have figured prominently into Cass Sunstein’s work on group polarization. See, for example, Sunstein’s article “Deliberative Trouble: Why Groups Go to Extremes” or his book Why Societies Need Dissent. “Group polarization means that members of a deliberating group predictably move toward a more extreme point in the direction indicated by the members’ predeliberation tendencies” (Sunstein, 2000, p. 74). So, for example, if a group of like-minded individuals (e.g. environmentalists, conservatives, feminists, etc.) get together to deliberate they are likely to become more extreme in the position. This results, argues Sunstein, from social influences on behaviour (e.g. our desire to maintain our reputation, pressures to conform, etc.) and the effect of limited “argument pools”. Suppose you believe that the American invasion of Iraq was unjustified. If you only discuss and debate the war with like-minded people you are likely to become more extreme in your conviction that the war is unjust. Why? Chances are the pool of arguments you have entertained will be narrower than the pool of arguments you might have entertained if you deliberated with a group of people who support the war in Iraq. Supporters of the war might bring different information to the debate, they might invoke different principles or concerns than those who oppose the war. So to exercise the deliberative virtues one must not limit the deliberative process to homogeneous groups.


Sunstein appeals to the experiments of Solomon Asch as evidence of the influence group influences can have on us. Sunstein’s work has important implications for democratic theory, especially deliberative democracy. The dominance of conformity is an excellent example of a non-ideal consideration that must be incorporated into a social theory. How do we do this? This is part of my concern in my new book Justice, Democracy and Reasonable Agreement in which I defend a “virtue oriented”’ account of justice entitled “civic liberalism”. Civic liberalism recognises that the virtue of civility is a multidimensional trait, and that the danger that civil aptitudes could give rise to vice (rather than virtue) is very real. This danger arises even in those associations that deliberative democrats might believe exemplify the deliberative virtues- associations where members are connected by bonds of affection, friendship, and solidarity. “In such groups, members are often less willing, or even unwilling, to state objections and counterarguments for fear that these will prove disruptive and violate the group’s internal norms. Families sometimes work this way” (Sunstein, 2003, p. 79). Virtuous families, friends, workplaces, legislatures, countries, etc. will recognise that dissenters often benefit others. And a non-ideal account of deliberative democracy will recognise that conformity does not necessarily promote societal interests.

The danger of group polarization is another reason why civic liberalism’s prescription that we avoid legislative or judicial supremacy (see my previous post) is judicious. Legislatures and courts are susceptible to the effects of limited argument pools. The dangers of group polarization are reduced when one endorses the dialogical model of judicial review. By dispersing power among independent deliberative bodies we ensure that opportunity exists for a diverse array of arguments and concerns to be raised. As a public ethic equipped to take non-ideal considerations seriously, civic liberalism prescribes that we guard against the potential dangers of group polarization. This has important consequences for the institutional design of a deliberative polity. Like Sunstein, I believe that deliberative democrats must take the empirical realities of non-ideal societies seriously. Otherwise we risk embracing a moral ideal that would lead to greater injustice (e.g. extremism) rather than a more humane society.

Cheers,
Colin