Toward a Heuristic Model of Scientific Invention Beyond the Explicit Hypothesis Space

抄録

The literature on scientific discovery generally describes theoretical invention as a process of generating, exploring, and selecting hypotheses within conceptual spaces that are already at least partially structured. Although this framework is useful, it has greater difficulty explaining the emergence of hypotheses that seem to arise before they become available as explicit possibilities for inquiry. This article proposes the concept of heuristic extraction of the unthought in order to name this problem and to offer a preliminary analytical model for its interpretation. The central argument is that part of scientific invention can be understood as the heuristic exploration of pre-articulated regions of conceptual space: zones in which there are not yet fully formulated hypotheses, but where constraints, structural traces, dispersed intuitions, or adjacent possibilities may already exist and become susceptible to future explication. To support this thesis, the paper articulates three lines of discussion: (i) debates on the internal limits of knowledge, with reference to Gödel and Foucault; (ii) models of heuristic search and problem solving, with emphasis on Herbert Simon and the notion of hypothesis space; and (iii) approaches to creativity and the transformation of conceptual spaces, especially in Margaret Boden and in the notion of the adjacent possible. The contribution of the article is neither empirical nor experimental. It is a conceptual and metatheoretical proposal. Its aim is to provide a more precise vocabulary for describing the passage between what is not yet formulated and what becomes explicitly hypothesized, while also suggesting why computational metaheuristics may be taken—at least in a strong analogical sense—as models for exploring this threshold. In doing so, the paper seeks to fill a gap between theories of discovery, philosophies of the limits of thought, and contemporary forms of computational search. By methodological delimitation, the article does not present a technical pipeline, a benchmark, or an experimental validation. These dimensions remain outside the scope of the present
paper and belong to a later development. The focus here falls exclusively on the philosophical and cognitive intelligibility of the problem. The paper concludes that the notion of heuristic extraction of the unthought is useful as a working hypothesis for rethinking scientific invention not merely as search within what is already thinkable, but also as a disciplined approach to
conceptual regions that are not yet fully available.