View this PageEdit this Page (locked)Uploads to this Page (locked)Versions of this Page over TimePrintable Version of this PageHome PageRecent ChangesSearchSign In


Project Description

The Artificial Intelligence researcher Marvin Minsky made the suggestion that a good way to learn about how music ‘worked’ was to represent masterworks as a constraint networks, then play around with the constraints, and study how this affected the music. The field of musical constraints has since explored this essential idea in numerous different ways. PlanC (pronounced "Plan C") was an early musical constraint system, but remains uniquely well-suited for manipulating some aspects of tonal harmony. While most musical constraint systems work with descriptions of musical materials such as individual notes and chords; PlanC draws power from a representation called Harmony Space that encodes high level information about tonal harmony in a three dimensional spatial form as shapes, paths, home locations, forbidden areas and trajectories. This representation builds on Balzano’s group theoretic cognitive theory of harmonic perception, allowing PlanC to represent and manipulate harmonic structures and processes in an explicit flexible and powerful manner.

PlanC (implemented in Prolog) can be used as an analysis tool to test theories of how particular pieces “work”, or can be used for learning and teaching about harmony, or can be used to compose new pieces of music. The system was implemented as a prototype, rather than a production quality system, but illustrated all of the key principles for a powerful new set of constraint based tools for manipulating tonal harmony.


  • A short piece composed using PlanC was transmitted on BBC Radio 4 as part of as part of thirty minute OU Mathematics programme on BBC Radio 4 transmitted 22.30 February 15th 1998.
  • PlanC is expected to play a role in future projects for tools to allow learners to compose music using constraint-based systems.

Selected Publications

Holland, S. (2000) Artificial Intelligence in Music Education: a critical review. In Miranda, E. (ed.) Readings in Music and Artificial Intelligence, Contemporary Music Studies Vol. 20. Pages 239-274, ISBN 90-5755-094-6, ISSN 0891-5415. Harwood Academic Publishers, Amsterdam. Artificial Intelligence MusicEd.pdf.

Holland, S. (1991) Preliminary Report on the Design of a Constraint-Based Musical Planner. Technical Report AUCS/TR9113, Dept of Computing Science, University of Aberdeen. (planc.pdf).

constraint diagram small.jpg