The Theory of Complex Predictability Accounts for Musical Pleasure

There is a recent literature suggesting that surprise, or deviations from expectations, leads to pleasure from listening to music. I disagree. It is the confirmation of predictions that leads to the greatest pleasure in music. This happens at several levels:

1. Rhythm is obviously predictable temporal sequences
2. Melody, with its predictable sequences of fifths, thirds and resolution, are predictable temporal sequences.
3. Harmony, with its predictable consonant combinations of nodes, are predictable.
4. Musical phrases and their repetitions are predictable
5. Systematic variations in volume, such as in Ravel's Bolero, are predictable
6. Interplay between different instruments are predictable
7. The chorus or hook of a song is the most repeated and thus predictable part
8. Familiarity of a song leads to increased pleasure through improved predictability for the corresponding brain

Several studies are consistent with this theory:

• Margulis (2014): This study investigated the effects of repetition on the pleasure experienced while listening to music. Repetition, in this case, can be considered as predictability. The results showed that repetition increased the liking for initially less familiar pieces, indicating that predictability could lead to pleasure. However, the effect was not observed for initially familiar pieces, suggesting that repetition leads to increased liking or pleasure only when it is novel enough to increase predictability.

• Pearce et al. (2010): This study explored the relationship between the predictability of melodic sequences and their perceived pleasantness. They found a U-shaped relationship, with both highly predictable and highly unpredictable sequences rated as less pleasant than moderately predictable sequences. Music that is too simple does not lead to predictability at multiple levels of the cognitive hierarchy.

• Cheung et al. (2019): This study investigated the role of surprise in music-induced pleasure using computational models and fMRI data. The results indicated that pleasure was associated with the degree to which listeners could predict upcoming events in a piece of music. Additionally, the researchers found that participants showed higher levels of pleasure when they were better able to predict musical events, but still experienced some level of surprise.

Since dopamine is believed to code for deviation from predictions, I predict that a different neurotransmitter will be found to increase in conjunction with musical pleasure.

In sum, the most pleasurable music is that which the brain is able to predict at many different levels.