This work introduces a physics–legal admissibility framework for executable systems, defining a minimal condition under which computational events may be recognized as real within a shared domain. Execution is treated not as a logical outcome, but as a discrete physical–informational transition subject to boundary conditions. An event exists if and only if a valid Reality Token (RT) is present at the moment of activation:E ≠ ∅ ⇔ Valid(RT) = true.In the absence of a valid token, execution is not erroneous, unauthorized, or misaligned; it is ontologically null and non-addressable, regardless of observable effects. The Reality Token is defined as a non-persistent authorization primitive binding temporal validity and informational identity into a single structural, non-probabilistic constraint. It does not regulate behavior, optimize decisions, or enforce ethics. It establishes existence. Within prudential artificial intelligence systems, the Reality Token functions ex ante as a legitimacy boundary, preserving interruptibility, attribution, and refusal prior to action. Rather than correcting probabilistic models, the framework constrains the conditions under which outputs may enter reality. This protocol does not prescribe governance outcomes or alignment strategies. It defines a minimal physics–legal prerequisite under which responsibility, verification, and accountability remain computable at scale. In the absence of such a prerequisite, these notions do not degrade or fail — they collapse.
Pablo Octavio Feria Hernández (Wed,) studied this question.