In a first-order Markov chain model for melody generation, what determines the next note?
AThe key and tempo of the piece
BA probability distribution conditioned only on the current note
CThe last three notes and the current beat position
DA random selection from all available pitches with equal probability
A first-order Markov chain has the Markov property: the next state depends only on the current state (the current note), not on the history before it. This produces statistically plausible melodic steps but lacks long-range coherence.
Question 2 True / False
True or false: Algorithmic composition necessarily produces music with no human involvement or artistic intent.
TTrue
FFalse
Answer: False
Algorithmic composition is a compositional technique, not a replacement for human authorship. The composer designs the system, selects parameters, curates outputs, and makes all the significant aesthetic decisions — the algorithm extends and executes their intent.
Question 3 Short Answer
What is an L-system in the context of algorithmic composition, and what type of musical structures does it produce?
Think about your answer, then reveal below.
Model answer: An L-system (Lindenmayer system) is a parallel rewriting system that recursively replaces symbols in a string with more complex substitutions according to production rules. Applied to music, it generates self-similar, fractal-like structures — nested rhythmic or melodic patterns that exhibit similar structure at multiple scales.
The recursive nature of L-systems produces structural self-similarity: a theme that appears at multiple hierarchical levels simultaneously, similar to fractal geometry. Bach's counterpoint and natural forms like branching trees exhibit similar nested self-similarity.
Question 4 Multiple Choice
A composer wants to generate music that statistically resembles Bach chorales — using the same harmonic progressions and voice-leading tendencies — without directly copying any existing chorale. Which algorithmic approach is most appropriate?
ARandomly select notes from Bach's scale choices
BTrain a Markov chain or neural network on a corpus of Bach chorales, then generate new sequences by sampling from the learned distributions
CApply the same rules to a different tone row
DUse an L-system with Bach's melodic intervals as production rules
Markov chains or neural networks trained on Bach chorale data learn the statistical regularities of his harmonic and voice-leading choices — transition probabilities between chords, melodic interval tendencies, cadence patterns. Generating from these models produces new sequences with similar stylistic properties.