A UNESCO study (UNESCO, 1979) has stated that the term 'culture' eludes definition and that there is no universally accepted description of its contents; it therefore recommended an empirical approach for the Framework of Cultural Statistics (FCS) which uses existing classifications and practices to infer its statistical contents. There are, of course, diverse views of what constitutes 'culture'. The original Latin 'cultura' was associated with growing and improvement and since then the term has become more narrowly associated with a certain value system that varies as between countries, with different connotations of 'culture' in the English/American region, 'culture' in France or 'kultur' in Germany. But this ambiguity also applies to the definition of other human and societal activities, and if we want to describe them we need at least some general typology, as provided in the FCS (see Horowitz, 1981). In such a framework, the constituent elements must be clearly specified so that they can be ordered into a coherent system. This is in contrast to a priori economic models, such as the National Accounts (SNA), where popular terms are used in a well-defined and articulated sense. Some cultural activities can be fitted into an economic schema of provision of goods and services, but the scope of culture obviously goes beyond the economic aspects. Social scientists from other fields, e.g. sociology or political science, have looked with some envy at the seeming certainty of economic systems, and the aggregates derived from them, because they can use a standard terminology and a common monetary numeraire. They then lend themselves to both empirical induction and a priori deduction, and can be applied to projections and planning. By contrast, (other) social systems lack a common unit and rest on a diverse and abstract terminology which leads to a complex structure whose evaluation and application to