10.1 Introduction

Although the Prisoner's Dilemma was originally developed and analyzed as a two-person interaction, many of the most important applications of what we might loosely call “Prisoner's Dilemma thinking” involve issues in the social sciences that are concerned with much larger numbers. This fact immediately poses a question: How does the two-person version differ from the large number Prisoner's Dilemma? Do the lessons of (and intuitions arising from) the two-person case carry over to larger scale social applications?

The general consensus in the economics literature is that the differences are very considerable – amounting to something like a qualitative difference between small-number and large-number situations. Consider, for example, the case of market provision of so-called “public goods.” As Richard Tuck observes in the epigraph, virtually all the classic writers on public goods provision make two points: first, that the public goods problem is very like the Prisoner's Dilemma problem in certain critical respects; and second, that small-number cases are unlike large-number cases in that voluntary action is much more likely to secure satisfactory levels of public goods provision in the small-number setting. Buchanan, for example, observes:

There is in short consensus that numbers make a difference; but much less consensus concerning exactly how and why they do.

ESTEEM, REPUTATION, AND THE ‘COMPOUNDING EFFECT’

We assume in this chapter, in line with what we have argued elsewhere (Brennan and Pettit 2004), that people desire the esteem of others and shrink from their disesteem. In making this assumption, we are deliberately associating ourselves with an intellectual tradition that dominated social theorizing until the nineteenth century, and specifically until the emergence of modern economics. That tradition includes Adam Smith, Thomas Hobbes, John Locke, the Baron de Montesquieu, David Hume – indeed, just about everyone who is recognized as a forebear of modern social and political theory, whether specifically in the economistic style or not. There is scarcely a social theorist up to the nineteenth century who does not regard the desire for esteem as among the most ubiquitous and powerful motives of human action (Lovejoy 1961). Smith's elegantly forthright formulation, offered as the epigraph to this section, simply exemplifies the wider tradition.

We can think of a minimalist version of the basic esteem relationship as involving just two individuals – actor A and an observer, B. The actor undertakes some action, or exhibits some disposition, that is observed by B. The observation of this action/disposition induces in B an immediate and spontaneous evaluative attitude.

PRELUDE

There are at least two possible approaches to the issue of political extremism. One is to study the operation of groups that are considered on some independent grounds to be extremist – perhaps in terms of the methods they use – and that see themselves and/or are seen by others to be broadly political in some meaningful sense. Thus, we might study the behaviour of the Red Guards or various of the nineteenth-century Russian anarchist groups for whom violence was an explicit tool in the pursuit of their political agenda. And we would distinguish such political terrorism from violence used by organized groups for merely criminal purposes. All such groups are extremist in method. Only those who have an explicit political agenda would be classified as relevant to political extremism. The reference in such a classification would be to the methods such groups use. Whether their political ends are extreme, and how indeed we would identify extremism of political ends, are separable questions. It is these latter questions with which I shall be concerned here. And I shall be concerned about them in a narrow, though not unfamiliar, context – the context of equilibrium outcomes in standard rational actor models of democratic political determination. Part of my reason for taking this approach is that this is where my expertise, such as it is, lies. But there is another reason, one perhaps more defensible.

Representation and political agency

Ideas of representation in political theory are notoriously diffuse and recalcitrant. We shall not here be concerned with the full array of these ideas. Our attention in this chapter will be focused on the issue of political agency – on the simple fact of representation, rather than its detailed form. The essential feature of representation, as we shall understand it, is that a mediating assembly of some kind is set between the citizenry and political decision making.

A simple model

In this chapter we take up the question – first broached in chapter 4 – of the relationship between the operation of institutions that seek to economise on virtue in use or in allocation, and the dispositional choices of individuals. The basic question is whether – or in what circumstances – institutions that economise on virtue in either of these senses may undermine or erode the virtue that they economise on. We will begin by sketching what we consider to be the simplest possible version of a model that is capable of capturing the feedback effect from institutions to dispositions; that is, a model that incorporates both dispositional choice and a structure of political institutions that operate as both sanctioning and screening mechanisms. In this context we will investigate the question of the conditions under which such an institutional structure has virtue producing properties, and the conditions under which the institutional structure may act to destroy virtue. Some of the limitations of this simple model will be addressed in the following section, where we will also outline some generalisations.

The basic model is organised around the choice between dispositions in the face of an imperfect screening device and an imperfect sanctioning device.

Motivation in politics

It seems self-evident that any account of the operation of democratic political institutions must depend on assumptions made about human nature – and specifically about human motivation. After all, within the analytic tradition that we will be working in – the rational actor tradition – human behaviour is understood as the outcome of rational choices, and rational choice is understood in terms of agents' beliefs and desires. The rational option, in the standard Humean/Davidsonian account, is just that option that maximises the agent's desire satisfaction, given the agent's beliefs (beliefs, say, about the consequences of alternative actions). Economists may, in most settings, talk about preferences rather than desires, but the Humean story – or something very like it – underlies virtually all modern economics and correspondingly all rational actor political theory.

Not all scholars admire the rational actor approach to politics. Their criticisms are varied and we will not try to address them all in this book.