Based on Moscovici’s classical study on the cultivation of psychoanalytic ideas in France in the 1950’s and our own research on modern biotechnology, we propose a paradigm for researching social representations. Following a consideration of the nature of representations and of the ‘iconoclastic suspicion’ that haunts them, we propose a model of the emergence of meaning relating three elements: subjects, objects, and projects. The basic unit of analysis is the elongated triangle of mediation : subject 1, object, project, and subject (...) 2, captured in the image of a ‘Toblerone’. Such social units cultivate, that is produce, circulate and receive representation which may be embodied in four modes–habitual behaviour, individual cognition, informal communication and formal communication–and in three mediums–words, visual images or non-linguistic sounds. We propose an operational definition of a ‘social representation’ as the comparison of four characteristics of communication systems: the content structures , the typified processes , and their functions , within the context of segmented social milieus. Seven implications for research on social representations are outlined: content and process; segmentation by social milieus rather than taxonomies; cultivation studies within social milieus; multi-method analysis; time structures and longitudinal data; the crossover of cultural projects and trajectories; the disinterested research attitude. This ideal type paradigm leads to an operational clarification to identify new research questions, and to guide the design and evaluation of studies on social representations. (shrink)