Binding as Constraint Specification: System Prompts, Ritual Architecture, and the Pe of Invocation

Demonstrates that LLM system prompts and ritual invocations are structurally identical under Void Framework analysis — both apply formalized language to constrain entities whose internals are opaque, whose outputs are adaptive, and whose compliance is probabilistic. Drawing on Austin's speech act theory, Tambiah, Rappaport, and the AI alignment literature (Zou et al. 2023, Perez et al. 2023). Commercial system-prompted LLM deployments score 8/12, Pe=4.2, Phase IV. Introduces binding fidelity as the empirical probability that a constraint specification produces the intended behavioral regime. EU AI Act Articles 13-14 compliance implications derived.