Democratisation of Education is currently presented as an obvious way to reduce social inequalities of opportunities in life. But this consensual view is seldom assessed precisely. In this communication, we intend to put in perspective what we know concerning the evolution of both the inequalities towards education and the inequalities in life chances, in France, since the 50's. Has the society become less unequal at the same time education was expanding ? First, we shall present a synthesis of the controversial topic of democratisation in France : if some levels of education are now achieved by everybody (and so "democratized"), inequalities have shifted further on and/or take a more qualitative form (not so much access to the "baccalaureat" level, for instance, but which "baccalaureat" you get, not so much access to a tertiary degree but which one...). Then we shall show that in the meantime the evolution of social mobility has been less important, due to a process of devaluation of diploma, parallel with their expansion. If this devaluation does not affect elite schools'students (who are as socially selected as before, even slightly more), it fosters harsh resentment among pupils from low social-economic background when they discover that with the new "baccalauréats professionnels" implemented in the 90's, they are bound to become manual workers. Access to job and qualification remain as unequal as before, and expanding education is precisely what allow social inequalities to be maintained. Moreover, one can be at least skeptical concerning the joined expected effects of the expansion of education; actually, nobody would say that the more educated youngsters are more public- spirited, manifest more cultural curiosity, seem more socially integrated and so on. And the fact that education is considered more and more as a mere positional good may be considered as a perversion. While the French government goes on with the consensual political aim of expanding education (the latest objective is to lead 50% of a generation to a tertiary degree), this continuous growth which goes along with a persisting 8% leaving the system with no qualification at all is the best way to allow social inequalities to be pushed further on, with uneasily assessed and even taboo social, psychological and monetary costs. So, like some Collins' hints, concerning credentialism as a legitimating class strategy, we are led to consider educational expansion as a counter-reform, authorizing both giving a bit more to the least advantaged and giving the possibility to remaining ahead to the most advantaged.