RANDOM WALK

February 2023 - Present

RANDOM WALK is an ongoing composition that uses the mathematical process of a "random walk" to generate new piano works from existing Western-classical piano pieces. It takes apart the canon and strings it back together in a "random" process, whose possibilities are determined entirely by my memories of finding joy in the physicality of playing piano. It thus reflects my complicated relationship with learning piano as a child, which helped me find delight in my body through the power of my hands while simultaneously imposing gendered and racialized expectations through its traditional conservatory environment. RANDOM WALK is a love letter to Western classical piano; it is a denunciation and destruction of Western classical piano; it is an exploration and reclamation of the euphoric-dysphoric physicality I experienced through Western classical piano.

In-Progress Video: Composition of Cycle 1, Movements 1-4

Performance Instructions

RANDOM WALK will first be performed by myself. Every note of every measure will be played, as much as possible, with the same tempo, dynamics, and musicality that I use when playing the original piece, but I will follow my instincts to change phrasing where necessary. Measures will be played smoothly from one to the next without pausing. There will be short pauses between movements of the same cycle, and longer pauses between cycles. RANDOM WALK will always be performed in its entirety - never in isolated cycles or measures.

Other performers must first demonstrate the current or prior ability to confidently perform all 19 source pieces. Then, in collaboration with myself, they will undergo the same rehearsal process of learning each measure of RANDOM WALK in accordance with how they play the original pieces.

Composition Process

Math

A random walk is a mathematical process for randomization. Given a graph of nodes connected by lines, first choose a starting node. Then, randomly choose one of its outgoing lines to follow to the next node, and repeat until the desired "walk" length is reached. Thus, the possibilities for the next node are always constrained by the initial graph, which dictates the possible next nodes connected to the current one.

Prep

For this composition, I started by choosing piano pieces from my time in conservatory that have stuck with me throughout the years. I ended up with 19 pieces that I played from ages 9-17, which are all significant to me due to the joy I found in the sheer physicality of playing them.

In the chronological order that I learned them, they are as follows:

1. Dmitry Kabalevsky: Fairy Tale, Op. 27 No. 20

2. Ludwig van Beethoven: Six Variations on "Nel cor piu non mi sento", WoO 70

3. Dmitry Kabalevsky: Sonatina Op. 13 No. 1, I - Allegro assai e lusingando

4. Aram Khachaturian: Sonatina in C, III - Allegro mosso

5. Moritz Moszkowski: Etude in g minor, Op. 72 No. 2

6. Sergei Rachmaninoff: Prelude in g minor, Op. 23 No. 5

7. Ludwig van Beethoven: Sonata Pathétique in c minor, Op. 13 No. 8, I - Grave - Allegro di molto e con brio

8. Frédéric Chopin: Revolutionary Etude, Op. 10 No. 12

9. Robert Muczynski: Six Preludes, Op. 6, I

10. Robert Muczynski: Six Preludes, Op. 6, II

11. Robert Muczynski: Six Preludes, Op. 6, III

12. Robert Muczynski: Six Preludes, Op. 6, IV

13. Robert Muczynski: Six Preludes, Op. 6, V

14. Robert Muczynski: Six Preludes, Op. 6, VI

15. Béla Bartók: Suite, Op. 14, I

16. Béla Bartók: Suite, Op. 14, III

17. Béla Bartók: Suite, Op. 14, IV

18. Sergei Rachmaninoff: Étude-Tableau in d minor, Op. 33 No. 4

19. Johann Sebastian Bach and Sergei Rachmaninoff: Violin Partita No. 3 in E Major BWV 1006, Prelude

I first made a graph with each piece as a node and then connected them to other pieces that I find similar in physicality or experience. To get a "nicely-random" random walk, your graph of N nodes should have each node connected to sqrt(N) other nodes; therefore, for my piece-graph, I gave each piece-node 3-5 connections.

Then, for each piece, I chose 6-13 of what I felt to be the most significant measures and made a measure-graph of those measures, again connecting them to measures I find similar in tactility. Every measure of every measure-graph has 3 connections.

Finally, I split each measure into right-hand and left-hand lines and then split each line apart, choosing based on what notes I felt should be grouped together or stand apart. These deconstructed measures are organized into measure-pouches holding their constituent parts.

Process

RANDOM WALK has 19 movements, split into Cycles of four movements each (with one last Cycle of three movements). Each of the 19 movements has 19 measures and is composed using a different color of yarn.

I started by choosing a beginning node for the piece-graph and the measure-graphs, and I marked the measure-graph beginnings with black thread.

For the first measure of Movement I, the Source Piece is the beginning node of the piece-graph (Muczynski Six Preludes III) and the Source Measure is the corresponding measure-graph's beginning node. To arrange the measure, I use a physical randomizing process of shaking the measure-pouches and laying the parts in the order that they fall out.

To make the second measure, I take one "step" in the piece-graph by choosing one of the possible paths from the previous Source Piece. The new Source Measure is then the corresponding measure-graph's beginning node, and the measure is arranged similarly by shaking the measure-pouches. All following measures are composed in the same way: stepping forward in the piece-graph to find the new Source Piece, and then using the corresponding measure-graph to find the Source Measure. When measure-graphs are revisited, they step forward to proceed to a new measure; the black thread is used for place-keeping in the measure-graphs.

The path through Source Pieces of a movement is marked on the piece-graph using colored yarn, and the path through Source Measures of a movement is marked between all the measure-graphs using colored yarn.

At the end of a movement, the colored yarn on the piece-graph is snipped and the new color begins on the same Source Piece. The corresponding measure-graph steps forward to find the first Source Measure of the next movement.

Between cycles, the built-up yarn on the piece-graph is removed, reflecting how I would discard pieces after conservatory exams to start the next set of pieces for the next year's exam. The built-up yarn on the measure-graphs remain throughout all cycles, reflecting how each piece stayed in my memory and subconsciously became part of an internal web of referential physicality.