Résumé
How should societies compare and choose among alternative futures under major transformation? Existing discussions of reform, institutional change, and long-run development often fail to distinguish clearly between what is preferable in terms of outcomes, what is practically achievable, and what is socially supportable. This paper develops a unified framework built around three sets: the outcome-layer choice set, the realizable set, and the acceptable set. The first captures social states that can be normatively compared at the level of outcomes; the second captures those that can actually be reached under technological, resource, institutional, and behavioral constraints; the third captures those states and paths that are not only realizable but also capable of securing sufficient social support. On this basis, the paper develops analytical tools for evaluating whether major reforms are trustworthy, identifying dominant social mechanisms and the compensatory mechanisms required when they change, and assessing civilizational progress over the long run. It also provides a framework for analyzing blocked transitions, including the conditions under which adjustment proceeds through negotiation, escalates into conflict, or culminates in war. The framework is applied to major technological transformation, especially artificial intelligence. The analysis suggests that large-scale AI substitution may shift scarcity toward physical and institutional control factors. With effective redistribution and compensation, technological change may generate highly concentrated yet sustainable structures; without them, markets may cease to function stably as society’s primary distributive mechanism.
Lire le preprint