The ENSO framework has maintained that physical time is not primitive. A conceptual implication is immediate: causation, when understood in the ordinary physical grammar of production, change, propagation, or cause followed by effect, cannot be primitive either. Earlier ENSO work nevertheless used the term "atemporal causation" to describe constraint-bearing dependence in an atemporal regime. That phrase was useful because it forced the problem into view, but it retained too much of the temporal grammar it was intended to remove. The present paper therefore makes an explicit ontological and terminological correction. The replacement term is "atemporal instantiation." It describes a physically realized state of relational closure constituted by constraint dependence in a regime in which temporal succession is not an intrinsic physical relation. The primitive object is not a process C -> R in which a constraint first exists and subsequently produces a result. It is one closure, written here as Inst(C,R), in which constraint-character and result-character are distinguishable aspects of the same relational state. In deterministic cases the closure may select a unique result; in probabilistic quantum cases what is fixed is the admissible probability structure rather than an individually predetermined detector outcome. Result is therefore treated as an inevitability of the relational specification, not as a temporal consequence. The paper distinguishes the ontology of a primitive relation from the way such a relation is empirically accessed. Laboratory evidence is necessarily acquired through temporal preparation, intervention, detection, recording, and comparison; this does not by itself imply that the primitive relation disclosed by those records is a temporal production process. A concrete self-embedding example makes the distinction explicit: the static closure conditions W = L + S and (W,L) ~ (L,S) uniquely fix the positive ratio L/S = phi without a temporal variable or iterative production step. Iteration may reveal the fixed point, but it is not required by the closure proof. Operationally, temporal structure is treated as an augmentation proposed by a process account. Such an augmentation earns explanatory status only if it introduces a reproducible result distinction that the constraint-defined closure cannot represent. Entanglement supplies a positive physical application, while Bell correlations and moving-analyser experiments provide a bounded stress test: the joint distribution depends on the state and measurement settings, opposite precedence labels for commuting separated measurements do not alter that distribution, and a concrete model that made such temporal ordering result-bearing predicted a loss of correlation that was not observed. These facts do not prove that all quantum physics is timeless or exclude every possible temporal ontology. They support a more precise methodological claim: temporal process should not be assigned to a primitive relation without an operational reason. The revised ENSO ontology is therefore: constraint dependence is the structure; atemporal instantiation is its physically realized relational closure; temporal causation belongs to the temporal regime in which temporality itself is physically available.