The music is both beautiful and complex, and it was written by a machine.

David Cope, a professor at UC Santa Cruz, has spent the last thirty years wrangling with the duality of music composition and computer intelligence. Namely, can a software algorithm truly compose original music, or will it just emulate the preferences and predilections of its programmer?

Cope, who is both a composer and a programmer, originally sought to construct algorithms capable of imitating the great composers (Bach, Brahms, Chopin, etc.). He dubbed his first such foray EMI, or Experiments in Musical Intelligence (aka, Emmy). Classical aficionados reluctantly gave it a nod, but not without condescension. Cope went on to publish a book, Computers Models of Musical Creativity, detailing his concept of Inductive Association and how it relates to synthesizing creativity.

Five years later, the next evolution of his music engine is also the latest artist found in the classical section of the iTunes music store. Welcome, Emily Howell. She/It takes the process of computer-generated music even further in that it is not driven by databases of existing compositions. Free associations perhaps? The music lives somewhere between Ruth Crawford, Morton Feldman, and Scriabin performed not by a computer but by established classical artists on real instruments. It seems then that Ms. Howell, the computer, has produced something innovative and original. This leads us to conclude that Cope is either a clever magician, one who infuses his own tastes into the output of a machine, or a pioneer who’s pushing computing one-step closer to the Turing ideal.

Futurist Ray Kurzweil along with noted Philosopher Douglas Hofstadter assert that what Cope has achieved is indeed the dawn of the intelligent machine. Then again, if a mannequin draws you into its character, don’t we credit the puppeteer? You be the judge. The music is on iTunes:

Centaur Records presents the music of “Emily Howell”, From Darkness, Light.

Cope’s book Computers Models of Musical Creativity on MIT Press: