This paper shows that, Rueff (1925, 1931) distinguished [a] a « permanent » unemployment due to excessive real wages relative to the labor productivity, [b] a “temporary” unemployment due to a decline in the economic activity resulting from a cyclic decrease of the price level, and [c] a « minimum » frictional unemployment prevailing in the normal functioning of the economy. Using empirical data, Rueff suggested that the unemployment of type [a] was largely dominant in England during the 1920’s (i.e. the socalled « law of Rueff »). The confrontation between this analysis and the subsequent analysis of unemployment in the literature reveals that : (i) the Phillips curve and its extension with the NAIRU appears as a non-legacy ; (ii) the wage curve is in accordance with the « law of Rueff » and provides an interesting complement to it; (iii) the equation proposed by Allais to explain the french unemployment rate includes the three types of unemployment pointed out by Rueff; (iv) although now abandoned, the fixed price temporary equilibria theory includes the unemployment of type [a] with both the classical regime and the keynesian regime of unemployment ; (v) the new keynesian microeconomy of the labor market shows that unemployment of type [a] can be explained by the behavior of rational agents without involving rigidities imposed by the Government ; this result generalizes the concept of unemployment of type [a] but is a refutation of the possibility accepted by Rueff to get a competitive equilibrium in a labor market without exogenous rigidities; (vi) the imperfect competition WS-PS model based on negotiation between employees and employers appears in accordance with the three kinds of unemployment [a], [b] et [c] so that it can be seen in this model a synthesis joining Rueff and Allais.