Confabulation

Canonical definition

A sincere but wrong Hypothesis or Belief, where sincerity is guaranteed by the architecture (the Individual has no introspective access to the wrongness) and wrongness is the failure of the Hypothesis to correspond to the state of affairs it concerns. Confabulation is a classification on the Truth axis alone: a Belief is a Confabulation if and only if it is sincere and wrong, irrespective of its content, its motivational source, or the cognitive operation that produced it. Truth-value is therefore constitutive of Confabulation-status, while remaining orthogonal to a Confabulation's downstream consequences, which run identically whether or not the Belief is wrong (Truth-Value Orthogonality). The two claims are compatible because the first governs classification and the second governs effects. Two sub-cases are distinguished by the Referent of the Hypothesis. Source-Attribution Confabulation is the case in which the Hypothesis concerns the motivational source of a present desire and is wrong because the M-Hypothesis was unrepresentable or under-weighted in I's Hypothesis Space (the Vocabulary Gap or Hilbert Collapse); this is the output of I's Bayesian ground-resolution function under Source Opacity, and is the proximate mechanism of Object Parasitism, scapegoating, and the Confabulation Cascade. In that source-attribution case, completions preferentially attribute to A or I and systematically avoid M-attribution (Disaggregation Convention 2); the genus definition itself remains truth-axis-only. Predictive Confabulation is the case in which the Hypothesis concerns a future state of affairs and is classified as a Confabulation only once that state arrives and resolves the Hypothesis false (see Predictive Superposition Principle). In every sub-case the Bayesian operation that produces the Belief is rational and content-agnostic; the wrongness, not the operation, is what makes the Belief a Confabulation.