The Hierarchical Construction of Affect in Music

Valence and Arousal as Residual Quality versus Intensity

Abstract

Research Question/Problem

The study of musical emotions has historically been informed by Russell's Circumplex Model (1980), which organizes affect into a two-dimensional map of Valence and Arousal. However, this view assumes a reflexive nature, in which emotion causes the observed response and neglects signal magnitude as an organizing dimension. In contrast, the Theory of Constructed Emotion (Barrett, 2017) suggests that the brain constructs affective experience by integrating interoceptive and external signals. Based on the evidence of a General Factor as the primary structural base, this work examines how the hierarchy of affect, with the primacy of Intensity, redefines musical cognition and explains the limits of aesthetic perception.

Objectives

This work proposes a Hierarchical Model of Musical Affect. The objective is to demonstrate that Intensity is not a geometric byproduct of Valence and Arousal, but the primary and organizing dimension of perception, upon which qualitative distinctions of pleasure and energy are constructed.

Method

An integrated theoretical approach was adopted, supported by secondary data analysis from two large-scale studies: the validation of the Musical Emotion Evaluation Test (MEET; N=200) and the AMG1608 corpus (N=1608). Principal Component Analysis (PCA) and Partial Least Squares Structural Equation Modeling (PLS-SEM) were used to map the variance of human reports, and Symmetric Residual Mapping (SRM) was employed to isolate affective quality from sonic magnitude in audio recordings.

Results

The evidence confirms the hierarchical hypothesis that Intensity (PC1) explains between 24% and 70% of the total variance of affect, acting as the foundation of perception. By isolating this dimension, we reveal that the cognition of three of the four quadrants (Joy, Tension, and Sadness) is acoustically residual and decodable. However, the “Serenity Paradox” is identified: in the high valence and low arousal quadrant, the human system inverts physical logic, suppressing magnitude to preserve relaxation. This finding demonstrates that musical affect is constructed through modular cognitive filters, explaining why purely reactive computational models fail in high-order aesthetic states.

Conclusions

The transition from a two-dimensional to a formative hierarchical model allows for a more accurate understanding of the listener's trajectory. It is concluded that musical cognition must consider Intensity as the basis of affective processing. The inability of AI mechanisms to mimic Serenity validates Barrett's (2017) thesis that affect is a sovereign construction of the nervous system, mediated by aesthetic filters that transcend raw acoustics.