This paper addresses a seemingly paradoxical phenomenon: thought processes are often initiated not by large-scale inputs, but by small, unexpected micro‑triggers that shift the current frame of perception. The strength of a trigger lies not in its magnitude, but in its informational surprise against the background of existing knowledge. We propose a conceptual model at the intersection of associative psychology, predictive coding, and signal-to-noise theory: (i) the knowledge base creates the context and the 'energy landscape' for search; (ii) a micro‑trigger acts as a localized perturbation of high surprisal, lowering thresholds between representational states; (iii) weak noise can enhance signal detectability (stochastic resonance). Predictions, methodological implications, and limitations are discussed.