I-Override

Canonical definition

I-Override is the cognitive operation by which an Individual redirects a Mimetic signal from automatic Decision Hub processing to deliberative Intrinsic Motivation scrutiny, thereby inferring the source-attribution that integration destroyed, which routes around Source Opacity rather than restoring the source-tags, and transforming felt necessity into an analysable sensation with identifiable Mimetic origins that can be evaluated rather than merely acted upon. I-Override is available only once the Thinker has been recruited (Chapter 4, Section 4.4.3): until then action-selection runs automatically on habit and on the priors already held, and an inference completed after selection revises the Beliefs for the next selection rather than redirecting the one already made. I-Override is Ground-Resolution run deliberately on the source-attribution special case: the reasoner maintains a non-zero interference term (Epistemic Suspension), preserves or expands the representable Hypothesis Space $\mathcal{H}$ through AIM Vocabulary Acquisition to admit H_M and other structural alternatives, and resists collapse onto a single Hypothesis pending further evidence, thereby enabling Belief revision. I-Override is not a separate faculty and not a state of the engine distinct from ordinary evaluation. I-Override is the evaluative operation under a configuration, and it runs Imagination, because expanding $\mathcal{H}$ to admit the Mimetic Hypothesis is what the generative operation performs. The operation is one instance of a general capacity. I-Override is not a separate faculty added to the architecture. It is the same inference engine, described in Chapter 13, running on the widest hypothesis space available to the Individual and maintaining a non-zero interference term against narrative-closure decoherence and M-contaminated evidence weighting. Run on a restricted hypothesis space, that same engine produces Confabulation. This is why the two are not opposites, and why the Knowledge Patch operates by widening the space rather than by installing a new capability.