A second important strand in this 'sociology of repression' has derived from looking at the official rationales given (by legislators, policy-makers and influential figures of the time) for the purposes of punishment, and comparing these with what actually happens. For example, if the official rationale for punishment is reform of offenders, but the forms of punish¬ment which are developed most are those that seem least effective in reforming offenders, does this mean that the 'real' goal of punishment is something else, and reform is only a gloss adopted by politicians and penal professionals to make their policies and practices publicly acceptable? This dissonance between the supposed goal of reform and the actual outcome of punishments making people worse has been most frequently commented upon in relation to prisons. If prisons do not in fact reform people, it is sug¬gested, but nevertheless people continue to be imprisoned, then the 'real' goal of punishment must be something other than reform. While some sociologists have accordingly identified other 'real' functions of prisons in particular and punishment in general (e.g. Mathiesen 1974; 1990; Reiman 1979), others have objected to the 'functionalism' of this argument. Func- tionalism, in sociology, is a perspective which suggests that social forms and institutions do not survive if they do not perform some useful function for society, and it is a general criticism of this approach to prisons and punishment that it falls too easily into either functionalism or economic determinism. Critics of this 'hidden functions' approach to penality argue that prisons really do not work, but that we have failed to come up with more effective alternatives (e.g. Ignatieff 1983).