It is frequently asserted that methodological individualism, the view in social science according to which all social phenomena must be accounted for in terms of what individuals think, choose, and do, is true but trivial and that the debate between individualists and nonindividualists is futile. By claiming that methodological individualism is neither trivial nor obviously true the present study challenges this assertion and seeks thereby to revive a flagging debate. The nontriviality of methodological individualism is shown by identifying its most plausible version and by unmasking its crucial, individuating but often concealed assumption. Its apparent unassailability is challenged by outlining an equally, if not more, credible non-individualist alternative.
Current interest in methodological individualism is largely due to analytical Marxists, notably Elster, who provocatively challenged Marxist orthodoxy which has traditionally shunned the individualist methods of economists in favour of an alliance with Hegelians, Functionalists, and Structuralists.1 But methodological individualism (hereafter referred to as MI) has a long and distinguished history. Hobbes can be credited with its first formulation, followed by thinkers of the Enlightenment, especially those who sought to provide a contractualist and therefore an individualist explanation of society. John Stuart Mill was its earliest systematic exponent.2 It was cast in its classic anti-naturalist mould by Dilthey. Weber, attempting to place himself delicately between Mill and Dilthey, emphasizing both causal generalization and autonomous human understanding, believed that MI was the surest way of exorcizing 'the spectre of collective conceptions that lingers among us'.3 It was deeply ingrained as one of the principal assumptions in the writings of Austrian economists such as Menger, von Mises, and Schumpeter. In fact, Schumpeter is generally acknowledged as having coined the term 'methodological individualism'.4 He claimed that MI is exclusively a scientific strategy according to which 'in the description of certain economic processes one had better begin with the actions of____________________