This paper demonstrates that the three movements of Beethoven's "Moonlight Sonata" (Op. 27 No. 2) instantiate three distinct machine learning architectures through structural correspondence rather than mere analogy. By computationally analyzing the score through entropy, Jensen-Shannon divergence, dissonance, hand distributional overlap, self-similarity matrices, temporal memory decay, and contextual pitch embeddings, we establish four counterintuitive findings:
- The perceived musical "temperature" is governed by throughput, not distributional width;
- The lightest movement carries the highest dissonance;
- The movements implement streaming, recurrent, and periodic positional encoding memory architectures;
- The same pitch class acquires different contextual identities across movements, analogous to contextual vs. static embeddings in NLP—unsupervised clustering recovers the tonal structure without music-theoretic input.
We construct a reverse sonification (decoding analytical features back into MIDI) and quantify the chirality of the encode-decode cycle: what distributions preserve and sequential ordering destroys. Prompted by a listener's observation that the decoded piece sounds like "mirror isomers that can't be superimposed," the chirality measurement reveals reconstruction loss increasing monotonically with n-gram order. Bootstrap baselines and subsample checks confirm all movements carry sequential information above noise, though raw values are confounded by sample size. Cross-domain comparison shows natural language has higher chirality than music, reflecting stronger sequential constraints.
Blogger's Review: This article explores the relationship between music and machine learning from a unique interdisciplinary perspective, revealing how computational analysis can uncover the underlying structure of musical works. The discussion on chirality and its implications in audio and language provides new avenues for thought and warrants further investigation.