The Metacognitive Attractor: A Structural Account of IQ Score Convergence in the Upper Tail of the Ability Distribution Franny Philos Sophia

Standard intelligence tests yield scores in the 120–130 range for individuals whose real-worldaccomplishments suggest far greater cognitive capacity—a pattern exemplified by RichardFeynman's recorded IQ of 125. This convergence is conventionally attributed to ceiling effects ormeasurement error. Here I argue that it reflects a structural attractor arising from metacognitiveoverhead: the reflective processing that accompanies high cognitive ability imposes superlinear costsunder timed testing conditions, producing a scoring function with a maximum in the IQ 120–130range. Above this critical threshold, additional ability decreases observed scores.Three independent lines of evidence converge on this account. First, gifted populations exhibit largeVerbal Comprehension–Processing Speed discrepancies (partial η² = .52), indicating that full-scalecomposites systematically misrepresent their ability profiles. Second, experimental studiesdemonstrate that concurrent metacognitive monitoring impairs timed cognitive performance, withverbal overshadowing selectively disrupting the spatial reasoning processes central to matrix-basedintelligence tests. Third, neurobiological research reveals qualitatively distinct cortical developmenttrajectories—prolonged synaptic proliferation followed by intensive pruning—in the highest-abilitychildren, consistent with a late-maturing cognitive architecture that standard tests are not designed tocapture.Integrating these findings, I propose a two-type model distinguishing rapid pattern recognitionintelligence (Type A: early-maturing, accurately captured by standard tests) frommetacognitive-generative intelligence (Type B: late-maturing, systematically underestimated). Thistypology unifies previously disconnected frameworks from creativity research, career trajectoryanalysis, and developmental neuroscience. The resulting evaluation asymmetry—standardinstruments are causally valid for Type A but progressively invalid for Type B—has directimplications for gifted identification, educational policy, and the design of assessment systems. Fourtestable predictions are offered to distinguish the attractor hypothesis from conventionalceiling-effect accounts.