This paper assesses G. A. Cohen's critique of Rawlsian special incentives. Two arguments are identified and criticized: an argument that the difference principle does not justify incentives because of a limitation on an agent's prerogative to depart from a direct promotion of the interests of the worst off, and an argument that justice is limited in its scope. The first argument is evaluated and defended from the criticism that once Cohen has conceded some ethically grounded special incentives he cannot sustain his critique of special incentives. But it is finally rejected as a subtle form of an unreasonably demanding moral rigorism. The second argument is interpreted as the more plausible of Cohen's claims. It has, however, to be defended via two subsidiary theses: the claim that Rawls endorses a moral division of labour and that this in turn grounds a further commitment to moral dualism as opposed to moral monism. This argument is assessed and rejected. Neither the moral division of labour nor moral monism supports the claim that in applying the principles of justice to a basic structure one does not thereby apply them to the individuals constrained to act within that structure in the marketing of their labour. Nor is it plausible to identify local aspects of social relations where the principles of justice are suspended. Such principles are presupposed, for example in market relations or the family, but limitation in scope of direct application does not limit the scope of justification. That scope extends at least as far as individual decisions to market one's labour. The latter are made fair in the only possible way they could be made fair. Rawls's commitment to the revisionary socialism of James Meade illustrates this point. It is concluded that no version of Cohen's critique succeeds. However, Cohen's critique identifies the most plausible version of Rawls's egalitarianism.