This paper advances a limited architectural claim about what any viable pre-physical generative theory must preserve, at minimum, once the entry conditions are already granted. It does not re-establish the minimal conditions under which structural existence can count as genuinely attained within process. It also does not offer a completed ontology, a physical theory, a formal dynamics, or a direct adjudication of competing scientific proposals. Instead, it takes as input the result of Author (2026a): that a viable theory must preserve genuine attainment, real internal failure, genuine status-difference, and some internal basis of processual determinability. The present question is what follows organizationally if those conditions are not to collapse into one another. The paper argues that collapse is the central architectural danger once the minimal conditions are in place. A theory may name ontology, generation, and readout while silently allowing one burden to absorb the work of another. Ontology may preload attainment by making the starting basis so rich that structural existence is already secured before generation begins. Generation may be reduced to neutral sequence or classificatory staging that bears no internal status-consequence. Readout may reverse explanatory order by allowing physical legibility to constitute what counts as attained rather than remaining downstream of it. In each case, the vocabulary of three layers is preserved while the functional distinctions those layers are meant to secure are evacuated. Against this pattern, the paper defends two claims. The stronger claim is functional: ontology, generation, and readout identify three irreducible explanatory burdens. Their irreducibility is established not by the observation that they answer different questions, but by showing that the characteristic failure of each burden cannot be redescribed without remainder as the successful operation of the others. Ontology cannot exhaust generation because even the most fully specified admissible characterization does not determine whether the accumulated differences among participating events will converge on the configuration required for closure. Generation cannot exhaust readout because even a fully borne generative trajectory does not determine how its attained outcome becomes physically legible. Readout cannot exhaust generation because even a fully specified account of physical legibility conditions does not bear the internal difference-accumulation that produces attained status. The weaker but still substantial claim is architectural: a three-layer structure of ontology, generation, and readout is the minimally viable organization under which this functional non-collapse can be preserved. The paper further introduces four metatheoretical diagnostics for identifying burden-collapse in practice: the Admissibility Compression Test, the Consequence Test, the Readout Reversal Test, and the Failure-Profile Test. These are not mechanical decision procedures but disciplined probes that require charges of collapse to be tied to determinate questions about failure-profile and explanatory remainder. The paper applies these diagnostics to Whitehead's framework of actual occasions and satisfaction, showing that it fails the Admissibility Compression Test because satisfaction is guaranteed by the nature of actual occasions, and fails the Consequence Test because the process does not bear attainment as an open question. The Failure-Profile Test confirms that the generative burden has no distinct explanatory work once the ontological burden is counted as successful. The present result is the middle term in a three-paper series. Author (2026a) establishes the necessary entry conditions for pre-physical generative theory. The present paper establishes the necessary architectural conditions: that the three burdens must remain functionally non-collapsed. Author (2026c) addresses the necessary bridge/interface conditions by asking what would additionally be required for attained status to stand prior to readout without being constituted by it. Satisfying the conditions of one level does not automatically satisfy those of the next; the levels are independent but cumulative. The present paper is offered as a clarification of the architectural threshold that must be crossed before that later bridge/interface work can be well-posed.
Li et al. (Mon,) studied this question.