The Illiac Suite: The 1957 Computer Score That Started AI Music
The Illiac Suite, also called String Quartet No. 4, is a piece for two violins, viola and cello created in 1957 when two Illinois researchers, Lejaren Hiller and Leonard Isaacson, programmed the university's ILLIAC I machine to help write it. It is widely regarded as one of the earliest substantial musical scores generated with a computer. The researchers programmed rules of counterpoint and probability, let the machine generate many candidate notes, kept those that obeyed the rules and transcribed the results for human players. The work is a direct ancestor of today's AI music systems.
Here is the story behind it and the surprisingly practical lessons it offers anyone who studies harmony and composition.
A room-sized machine and a musical question
The ILLIAC I was one of the early university computers, large enough to fill a room and programmed with punched tape. Hiller was a chemist with serious musical training, and Isaacson was a fellow researcher. They asked whether the computer could imitate the decisions a composer makes, not by understanding music, but by following explicit instructions.
Generate and test
The method behind the suite is simple to describe.
Generate
The computer produced random candidate notes.
Test
Each candidate was checked against a list of rules. If it broke a rule, it was rejected and a new one was generated.
Keep and continue
Accepted notes were added to the music, and the process repeated.
Today this approach would be called rule-based generation combined with random sampling. It is slow and wasteful, because many candidates are thrown away, but it shows exactly which rules shape the final music.
Four experiments, four movements
Each of the four movements tested a different idea.
- The first explored simple melody writing and then two- and four-voice counterpoint using strict rules similar to those taught in traditional harmony classes.
- The second applied more varied counterpoint rules to create contrasting textures.
- The third experimented with rhythm, dynamics and playing techniques, and included more dissonant, modern sounds.
- The fourth used probability tables, called Markov chains, where the choice of each note depended on the notes before it.
How the world reacted
The premiere and the publication of the score sparked debate. Some musicians found the idea exciting; others felt a computer could never truly compose. The authors themselves described the work as an experiment rather than a masterpiece. Yet the questions they raised, about creativity, rules and randomness, are the same questions people ask about generative AI today.
What pianists can learn from it
The suite is a reminder that much of musical grammar can be written as rules. Exploring those rules at the keyboard builds understanding.
Try a rule game
Choose a key and write a short melody by rolling through the scale step by step, then apply three rules: begin and end on the tonic, avoid leaps larger than a fifth, and follow a leap with a step in the opposite direction. Play the result and listen to how much the rules shape the character.
Try a probability game
Write the notes of a simple folk tune and count which note tends to follow which. Use those tendencies to create a new tune. You have just built a tiny Markov chain by hand.
Notice what rules cannot do
Your rule-based melody may be correct but bland, a little like a sentence with perfect grammar and nothing interesting to say. Expressive phrasing, surprise and emotion usually come from choices that go beyond the rules.
From 1957 to modern AI
Modern systems learn patterns from large collections of music instead of relying only on hand-written rules, but the core idea remains: generate possibilities and evaluate them. Practice tools that analyse your playing work in a related way, comparing what you did with what the score expects and highlighting differences.
FAQ
Q: Who composed the Illiac Suite?
A: Two researchers, Hiller and Isaacson, who worked with the ILLIAC I machine in Illinois in 1957.
Q: What instruments play the piece?
A: A string quartet: two violins, viola and cello.
Q: Did the computer understand music?
A: No. It followed programmed rules and probabilities, and humans chose and arranged the results.
Q: Is the Illiac Suite real AI?
A: It predates modern machine learning but is considered an important early step toward AI composition.
Learn the rules, then make music with TIGA ALPHA
TIGA ALPHA is an AI piano teaching app built by a Thai team, with lessons in Thai, English and Mandarin. Its CEO, Teacher Aum, has 20 years of experience in piano teaching, and that music teaching background shapes every lesson, while piano teaching with AI gives you feedback each time you play piano. Use it as an online piano lessons app for short daily sessions, follow it as a complete AI online piano course, or combine it with a piano course or piano learning course led by a teacher. Whether you want to learn piano from zero, learn piano online after a long break, learn music alongside your children or simply play music for fun, this online piano teaching application keeps your next step clear. New users get a 7-day free Premium trial, then 799 THB/month. Open the piano teaching app below and pick a first lesson that fits your week.
- Download and start: TIGA ALPHA landing page
- In-person lessons in Thailand: TIGA Studio
- About the company: tigaalpha.com
Knowing the rules of music is the starting point; deciding when to bend them is the art.