Purpose ‐ When coping with complex, but also possibly disruptive and open-ended social dynamics, the anticipatory system idea, which was developed by Rosen in the realm of physical and biological system observation, remains a reference framework, but one that may need to be reinforced by other theoretical considerations. This paper aims at using a debate that took place in a specific foresight discussion arena on early detection and weak signal analysis, as a constructive epistemic detour to eventually contribute to such a reinforcement of Rosen's anticipatory system proposal.Design/methodology/approach ‐ The author aims at revisiting Rosen's framework with stimulating inputs drawing upon the early detection debates, by first assessing the original concepts brought up by Ansoff in the 1970s and 1980s and its further enhancements by contemporary scholars. A rather constructivist approach is then developed to weak signal analysis, aiming at emphasising the need, in analytical situations involving social system features, for reflexive stages and capacities. Bearing this requirement in mind, the productive value of the "framing" and "meta-framing" notions is explored, in order to apply them to Rosen's anticipatory systems and possibly contribute to enriching his original concept.Findings ‐ How effective the framing and meta-framing couple can be for a series of anticipatory issues is described in a detailed manner and, then more specifically, Rosen's anticipatory system concept is revisited in the light of those inputs, aiming at putting into perspective new options for research and anticipation activities in general.Research limitations/implications ‐ The paper is essentially conceptual and based on a rich but disputable detour by early detection and weak analysis issues so as to emphasise key reflexive references and method. However, most of this material is taken from domains rather untypical of Rosennean debates and in addition would need to be completed by a series of supportive cases, but that is beyond the scope and scale of this paper.Practical implications ‐ The paper sets clear distinctions and boundaries for when and when not to apply reflexive steps in a foresight exercise, including in the context of rolling out a Rosen type of approach. Research decision making both in the corporate and policy-making contexts can benefit from such clues and supportive framework conditions.Social implications ‐ Social systems are typically complex and involve multiple perspectives and viewpoints; they concern a series of major challenges to be coped with locally or more globally, at environmental, political, cultural or technological level, and in that category of anticipatory endeavor, the framing/meta-framing epistemic couple may be of great usefulness.Originality/value ‐ Although rather conceptual, the detour proposed by the paper aims at creating a reflexive distance and enriched capability to evaluate one's potential biases and blind spots in anticipatory modelling activities.