Starting from R. K. Merton's now classic criticism of 'holistic' functionalism, i.e. of a functionalism which postulates social unity, universality and functional in-dispensability, the author stresses certain implications of this criticism more than they have been stressed hitherto. Classical and holistic functionalism) from H. Spencer, B. Malinowski, A. R. Radcliffe-Brown, etc to T. Parsons, postulates certain total unities (a global culture, an integrated system, etc.) in which each item (existence, actions, structures, etc.) is considered and defined on the grounds of (...) its consequences for the maintenance of the system as a whole; therefore holistic functionalism as a method is, in effect, the study of the consequences of the system on the items that compose it, since each of these items is defined within the sphere of the system and of its integrative functions. Merton's 'neo-functionalism', on the other hand, is remarkable not only in that it takes into account the 'dysfunctional' and 'nonfunctional' consequences of certain items on the system, but more especially because, within the context of functional analysis, it stresses the possible existence of structural substitutes and alternatives of functions, and therefore of latent structures which are foreign to objective functional consequences, as well as being able to deal with unanticipated and unexpected items and their consequences on the system. 'Neo-functionalism', which is susceptible of further development, is not limited to the study of the consequences of the system on its items: it can also reverse this scheme and study the consequences of certain items on the system. Merton's criticism of holistic functionalism therefore implies a broadening of the scientific resources of this method and a renewal of its interpretative scheme, thanks to which functional analysis ceases to appear as 'the* method of explaining sociology as a science, and becomes an interpretative method which complements the analysis of social structures and relations. Seen in this light the concept of structure becomes emancipated and independent of the concept of system and function; whereas, within the framework of universal functionalism, it was ancillary to the concept of function. Finally, latent structures and unconscious structures, conditions of possibility and subjective dispositions are favourable to social structures and social relations, not excluding those that are neither visible nor observable. This analysis, the author notes, is extremely meaningful and has great possibilities of development, especially in view of the structuralism recently to be noted in the human and social sciences: anthropology, history, linguistics, etc. (shrink)