Affirmative action policies — those that take race, caste, gender, or other identity characteristics into account in decisions about employment, education, or contracting — are among the most debated instruments of social policy. Their proponents argue that they correct for structural disadvantages that meritocratic selection alone cannot overcome: a system that selects on the basis of test scores will reproduce the inequalities embedded in the conditions of preparation for those tests. Their critics argue that they violate the principle of individual merit and introduce reverse discrimination that harms candidates from non-target groups.
A more sophisticated version of the proponent's argument holds that 'merit' is not a neutral measure but a socially constructed one: what counts as merit in a given selection process reflects the values and competencies of those who designed the process. If the design reflects the experiences and cultural assumptions of a dominant group, the test will systematically favour candidates from that group even when their superior performance reflects advantage rather than ability. Correcting for this requires adjusting not just the inputs (preparation) but the measurement itself.
The critics' most powerful response to this is that the argument proves too much: if merit is entirely socially constructed, there is no principled basis for any selection at all. The practical resolution most democratic societies have reached is that merit is imperfect but not meaningless, and that affirmative action should be time-limited and targeted at identifiable structural disadvantages rather than used as a permanent redistribution mechanism.