When Prediction Errors Shape Perceived Truth: Temporal Alignment and Affective Appraisal in Misinformation

Why do transient changes in processing sometimes influence perceived truth, whereas similar processing experiences have little or even opposite effects? We propose a theoretical framework specifying when prediction-error-related changes in processing may become epistemically relevant to the evaluation of information, with particular implications for misinformation. The framework distinguishes between the occurrence of a prediction error and its use as a cue for judgment. We propose that temporal alignment with target evaluation acts as a gating condition: a processing signal is more likely to contribute to perceived truth when it occurs while the target information is being evaluated and can therefore be attributed to that target. Temporal alignment does not, however, determine the direction of its epistemic effect. Once subjectively accessible, the processing experience is interpreted through attributional and affective appraisal processes. Unexpected facilitation or coherence may be positively appraised and increase subjective plausibility, whereas processing difficulty, incoherence, conflict with prior knowledge, or source-related suspicion may be negatively appraised and decrease it. The framework therefore predicts both plausibility-enhancing and plausibility-reducing consequences of prediction error. By integrating predictive processing, processing fluency, metacognitive inference, and affective appraisal, we propose that temporal alignment constrains whether a transient processing signal enters an epistemic judgment, whereas attribution and appraisal contribute to determining how that signal is interpreted. This account generates testable predictions concerning when processing experiences should increase, decrease, or fail to influence the perceived truth of misinformation and invites direct empirical tests capable of corroborating, refining, or falsifying the temporal-alignment hypothesis.