In the context of the ‘knowledge society' paradigm, the paradoxes and tensions observed in the evolution of the French higher education and research system are instructive for an appreciation of the impact and limits of public action in this area. As we have seen, the rapid increase in enrolment rates during the 1988–98 period was accompanied by a very clear ‘vocationalization' of curricula. This was particularly true in higher education, which made a significant effort to develop a system of higher technological education recognized and appreciated by industry. These educational resources were mobilized for an in-depth productive reorganization, notably in the large industrial concerns but also in the financial sector. This reorganization was carried out by the elite traditionally coming from the Grandes Ecoles, especially those producing the senior branches of the civil service. What distinguishes this phase of the restructuring, as Hancke´ (2002) has clearly shown, is the managerial elite's growing autonomy in relation to the State. This has ensured the return of France's major firms to industrial competitiveness and high profitability. Nonetheless, such an ‘economy coordinated by an elite network' seems fragile in so far as it is challenged by the firms of other coordinated economies which can rely on compromises reached with labour. These questions are all the more vital in face of serious doubts about the ability of vast technological programmes to renew a capacity for innovation on the ‘technology frontier'. This was achieved in the past through a mission oriented innovation policy centred on hi-tech goods and largely realized within the framework of State markets. As the telecommunications example shows, this model relied heavily on the excellence of the elite engineering schools and specialized public research centres. This institutional configuration, however, does not respond as efficiently to the demands of the ‘diffusion- oriented' model. From this standpoint, the French experience, in spite of the considerable expansion of its higher education system, reveals an inadequacy and a tension. Entrepreneurship training, notably in order to develop knowledge and create hi-tech start-ups, seems largely inadequate; training through research is not really recognized in the private sector, which is reluctant to hire Ph.D.s to man R&D teams if the thesis is not combined with an engineering diploma from a Grande Ecole. In addition, masters level vocational training programmes tend to attract the best university students and thus to divert them from training through research. As a result, the base of knowledge and competences produced by the higher education and research system seems much too narrow for a diVusion-oriented innovation model. If the recruitment pool for the top levels of higher education at the highest levels is not considerably broadened, the tension between a policy of vocational oriented training and one valuing career paths based on training through research will only be reinforced. In some way, this might give rise to a tension between the ‘learning industry' a` la française', an unquestionable success of the vocationalization of studies as attested by the emergence of a new form of coordinated economy, and the ‘knowledge society' which is struggling to achieve social legitimacy. One irony of this history, however, is the emergence of highly dynamic technology parks which are spin-offs of the heyday of the great technological programmes and notably of one of the most flagrant industrial failures, the Plan-Calcul which was intended to stimulate a national computer industry. The new forms of public policy likely to support the growth of these technology parks and promote the creation of others remain to be invented.