Psychometrics constructed its measurement instruments before it possessed an adequate theory of whatit intended to measure. The general factor of intelligence g, introduced by Spearman (1904) andconsolidated in the Cattell-Horn-Carroll model (McGrew 2009), is a construct inferred statisticallyfrom the structure of correlations among cognitive performances. It is sufficient as an explanation ofthe positive manifold (the structure of positive correlations) but not necessary, and it remainsontologically controversial. Intelligence tests currently in use measure six cognitive abilities understandardized conditions and produce a composite score that compresses a multidimensional structureinto a single scalar number. This paper argues that such compression is not a tolerable approximationbut a structural loss of relevant information.Contemporary neuroscience provides a theory of intelligence that psychometrics has neverdeveloped. The Network Neuroscience Theory (Barbey 2018) identifies g as an emergent property ofthe global topology of the brain connectome, not a localizable entity, but rather the result of the dynamicflexibility of neural networks and their capacity to transition between states that are easy or difficult toreach. The present work integrates this framework into a broader picture that includes axonalconduction velocity as the physical substrate of topology, the Free Energy Principle (Friston 2010) as amodel of the brain as an open system coupled to body and environment, and the emotional factor as astructural moderating variable of the cognitive process.On the basis of this integration, the paper proposes a three-layer model of cognitive assessment(neurological structural baseline, dynamics under naturalistic conditions, and coherence underperturbation) that repositions existing psychometric instruments as partial proxies of a richerconstruct. The IQ is not eliminated; it is placed in the position it deserves. The paper identifies an openresearch program and provides its conceptual foundation and operational structure.
Mirko Bradley (Tue,) studied this question.