Time-course variation of statistics embedded in music: Corpus study on implicit learning and knowledge

Area of Science:

• Cognitive Science
• Musicology
• Computational Musicology

Background:

• Statistical learning, an implicit cognitive process, underlies understanding sequential patterns.
• This implicit knowledge can be expressed through abstract mediums like musical melodies.
• Musical compositions may reflect a composer's evolving statistical knowledge over time.

Purpose of the Study:

• To investigate how musical statistics, specifically transitional probabilities, change throughout a composer's career.
• To determine if these statistical variations correlate with a composer's implicit knowledge development.
• To explore novel methods for evaluating temporal changes in a composer's implicit cognitive processes.

Main Methods:

• Analysis of transitional probabilities in the highest-pitch sequences of Beethoven's Piano Sonatas.
• Application of hierarchical Markov models to quantify sequential patterns.
• Ordering musical pieces chronologically by opus number to track lifetime variations.

Main Results:

• A gradual decrease in transitional probabilities of musically universal sequential patterns was observed over Beethoven's lifetime.
• These findings suggest a correlation between the time-course of musical statistics and a composer's statistical knowledge.

Conclusions:

• Musical scores can serve as a medium to evaluate the temporal dynamics of a composer's implicit statistical knowledge.
• This research introduces a novel computational approach to understanding cognitive evolution in music composition.