EMI

Experiments in Musical Intelligence

Experiments in Musical Intelligence is a software developed in the 1980s by David Cope, designed to generate stylistically accurate imitations of composers' works by analyzing their existing music as input. The software was never released in its full form to the public and most of its databases were deleted by Cope himself in 2003.

From its first appearance, EMI's deep nature was highly controversial and sparked a wide range of conflicting reactions: some despised it as a threat to creativity and nothing more than a pale imitator of the great Masters, others hailed it as a groundbreaking artistic revolution.

There are several uploads on David Cope's YouTube channel featuring EMI’s music. However, most of the recordings are done with digital MIDI instruments and without any phrasing at all. Machine-like, indeed. It's interesting to open some of his videos and read the comments: some people are fascinated, some disturbed, others are barely touched by the subject.

Without a doubt, EMI was a groundbreaking experimental endeavor that anticipated much of the neural-network technology now widely accessible to the public.

A very sweet moment in the "2-part Invention 7 after J.S.Bach" by EMI. Was it luck?

Similarities -part 0

It's possible to notice strict similarities with the 2 part inventions by J.S.Bach every once in a while, some of which were already noted by audiences in the 1980s during one of Cope’s Conferences.

It's easy to argue "the computer merely copied" if not "plagiarized", but I would be more careful examining the output. While it is true that entire bars can be identified -still, very rarely I must add- most of the time, similar ideas are put into diverse contexts from the originals thus requiring some form of intuitive reasoning (or surrogate) to apply the necessary changes for the adaptation.

 

Could it be that we tend to overinterpret patterns and similarities with Bach simply because we know as a fact that 1) Bach came earlier and 2) The computer is taking inspiration from it?

Similarities -part 1

 

EMI's Invention 2 clearly uses the same chromatic/contrary-motion-like figure as Bach's invention 6, but because the piece is set in a different tempo and key, the computer needs to figure out a way to "adapt" a similar idea to a slightly different scenario. Unlike in Bach's invention, EMI's piece does not include 32nd notes, making it perhaps exhibit its own darker and more serious atmosphere.

Another great similarity I found is this use of up-down scale motif, present in Bach's Invention 13. It can be found in multiple pieces. Here is one example:

 

Similarities -part 2

 

Despite my gut feelings tell me to mostly credit Bach while listening to EMI, I find hard to argue there is a difference in the way music most likely might have been taught to pupils. It is widely accepted one of the main ways to learn the skill of music was indeed to copy the teacher's ideas and try to re-formulate them or put them into different contexts.

Here is the opening of Toccata and Fugue in A minor KrebsWV 411 by Johann Ludwig Krebs, one of best J.S.Bach's scholars.

Source: (https://imslp.org/wiki/Toccata_and_Fugue_in_A_minor%2C_Krebs-WV_411_(Krebs%2C_Johann_Ludwig))

Now take the F Major Toccata by J.S.Bach, BWV 540

Source (https://digitalesammlungen.bach-leipzig.de/main/thumbnailview/qsr=bwv%20540)

It's not hard to spot similarities. Despite not being exact copies, both pieces share the same tempo of 3/8, both start with a steady pedal note on the fundamental, both begin with a 2 voice canon-like passage going on for a while on the manuals.

What follows in both cases, is a pedal solo:

Even when looking closer at the shape of the 16th notes in the pedal solo, it's hard to see changes:

Another passage harmonically almost identical appears later in the same piece:

 

 

I wouldn't say the mentioned stylistic features of KWV 411 and BWV 540 are a standard in organ music all-together. Only when taken singularly ("pedal notes", "canons", "pedal solos", etc) yes, they are indeed Common Practice of baroque organ music.

Krebs's Toccata eventually goes on another route and the Fugue has little to do with Bach's.

Similarities -part 3

I personally find both Toccatas and Fugues to be absolute fine treasures of contrapuntal interweaving: the voice texture is rich and dense but never overwhelming. I find it hard to rate one piece as better than the other based solely on the notes.

J.L.Krebs signature on the front page of KrebsWV 411.

As none of the available Krebs WV 411 manuscripts online contains even the smallest credit to Bach from the composer, our current view of copyright is challenged.

Perhaps the piece was meant for private use only?

Perhaps the F major toccata was such a popular piece that everyone knew Bach was the original author?

Perhaps I am overseeing similarities and being superficial in assuming one composer was inspired by the previous.

Perhaps who made the art -the composer as individual- wasn't that relevant afterall because art itself was what mattered?

 

For a more in-depth discussion of this topic it is recommended the reading of "Musical Production Without the Work-Concept" from The Imaginary Museum of Musical Works: An Essay in the Philosophy of Music by Lydia Goehr.


 

Click on the archer to

go back to the Home Page