Author: Noah Stern Weber

Your Computer is Listening. Are you?

Six years ago, I wrote an article stemming from a lively discussion that I had with a few friends on the work of David Cope’s artificial intelligence compositional program “Emily Howell.” My intention had been two-fold: to approach the philosophical challenges of our society accepting music originating from an extra-human source, while also attempting to discuss whether “Emily Howell’s work” met the definition of a composed piece—or if extraordinary human effort was involved in the final product.

This inquiry will take a very different approach.

We begin with the hypothesis that, due to the rate of growth and development of A.I. technology, #resistanceisfutile. Which is to say that computer-composed music is here, and the conversation needs to change.

Need proof? When I wrote the article six years ago, there were roughly two or three A.I. programs, mostly theoretical and almost exclusively confined to academic institutions. In the two weeks between agreeing to write this article and sitting at down to flesh out my notes, a new program using Google’s A.I. open platform was released. In the week and a half between writing my first draft and coming back for serious revisions, another A.I. music system was publicly announced with venture capital funding of $4 million.  The speed at which new technology in this field is developed and released is staggering, and we cannot discuss if it might change the musical landscape, but rather how we will adapt to it.

Advances in the capacity and ease of use in digitally based media have fundamentally changed the ways that creators and producers interact with audiences and each other and—in many ways—they have bridged some of the gaps between “classical” and “popular” music.

Ted Hearne introduced me to the beauty and artistic possibilities of Auto-Tune in The Source (digital processing design by Philip White). After seeing a demo of Kamala Sankaram’s virtual reality operetta The Parksville Murders, I programmed a session at OPERA America’s New Works Forum, bringing in the composer, producers (Opera on Tap), and director (Carri Ann Shim Sham) to introduce their work to presenters and producers of opera from around the country. While still a beta product, it led to a serious discussion about the capacity of new technologies to engage audiences outside of a more traditional performance space.

The Transactional Relationship 

In the tech world, A.I. is equated to the Holy Grail, “poised to reinvent computing itself.” It will not just automate processes, but continually improve upon itself, freeing the programmer and the consumer from constantly working out idiosyncrasies or bugs. It is already a part of our daily lives—including Google’s search function, Siri, and fraud detection on credit cards. The intuitive learning will be essential to mass-acceptance of self-driving cars, which will save tens of thousands of lives annually.

So why is A.I. composition not the next great innovation to revolutionize the music industry? Let’s return to the “Prostitute Metaphor” from my original article. To summarize, I argued that emotional interactions are based on a perceived understanding of shared reality, and if one side is disingenuous or misrepresenting the situation, the entire interaction has changed ex post facto. The value we give to art is mutable.

A.I.’s potential to replace human function has become a recurring theme in our culture. In the last 18 months, Westworld and Humans have each challenged their viewers to ask how comfortable they are with autonomous, human-esque machines (while Lars and the Real Girl explores the artificial constructs of relationships with people who may or may not ever have lived).

I’ll conclude this section with a point about how we want to feel a connection to people that move us, as partners and as musicians. Can A.I. do this? Should A.I. do this? And (as a segue to the next section), what does it mean when the thing that affects us—the perfectly created partner, the song or symphony that hits you a certain way—can be endlessly replicated?

Audiences are interested in a relationship with the artist, living or dead, to the point that the composer’s “brand” determines the majority of the value of the work (commissioning fees, recording deals, royalty percentages, etc.), and the “pre-discovery” work of famous creators have been sought after as important links to the creation of the magnum opus.

Supply and Demand

What can we learn about product and consumption (supply and demand) as we relate this back to composition in the 21st century?

If you don’t know JukeDeck, it’s worth checking out. It was the focal point of Alex Marshall’s January 22, 2017, New York Times article “From Jingles to Pop Hits, A.I. Is Music to Some Ears.” Start with the interface:

 Two JukeDeck screenshots--the first shows the following list of genres: piano, folk, rock, ambient, cinematic, pop, chillout, corporate, drum and bass, and synth pop; and the second shows the following list of moods: uplifting, melancholic, dark, angry, sparse, meditative, sci-fi, action, emotive, easy listening, tech, aggressive, and tropical

Doesn’t it seem like an earlier version of Spotify?

Two smartphone screenshots from an earlier version of Spotify, the first one features an album called Swagger with a shuffle play option and a list of four of the songs: "Ain't No Rest for the Wicked," "Beat The Devil's Tattoo," "No Good," and "Wicked Ones"; the second one features an album called Punk Unleashed with a shuffle play option and a list of five of the songs: "Limelight," "Near to the Wild Heart of Life," "Buddy," "Not Happy," and "Sixes and Sevens."

“Spotify is a new way of listening to music.” This was their catchphrase (see way-back machine to 6/15/11). They dropped that phrase once it became the primary way that people consume music. The curation can be taken out of the consumer’s hands—not only is it easier, but also smarter. The consumer should feel worldlier for learning about new groups and hearing new music.

The problem, at least in practice, is that this was not the outcome. The same songs keep coming up, and with prepackaged playlists for “gym,” “study,” “dim the lights,” etc., the listener does not need to engage as the music becomes a background soundtrack instead of a product to focus on.

My contention is not that the quality of music decreased, but that the changing consumption method devalues each moment of recorded sound. The immense quantity of music now available makes the pool larger, and thus the individuals (songs/tracks/works) inherently have less value.

We can’t erase the Pandora’s Box of Spotify, so it is important to focus on how consumption is changing.

A.I. Composition Commercial Pioneers

Returning to JukeDeck: what exactly are they doing and how does it compare to our old model of Emily Howell?

Emily Howell was limited (as of 2011) to the export of the melodic, harmonic, and rhythmic ideas, requiring someone to ultimately render it playable by musicians. JukeDeck is more of a full-stack service. The company has looked at the monetization and has determined that creating digital-instrument outputs in lieu of any notated music offers the immediate gratification that audiences are increasingly looking for.

I encourage you to take a look at the program and see how it creates music in different genres. Through my own exploration of the JukeDeck, I felt that the final product was something between cliché spa music and your grandparent’s attempt at dubstep, yet JukeDeck is signing on major clients (the Times article mentions Coca-Cola). While a composer might argue that the music lacks any artistic merit, at least one company with a large marketing budget has determined that they get more value out of this than they do from a living composer (acknowledging that a composer will most likely charge more than $21.99 for a lump-sum royalty buyout). So in this situation, the ease of use and cost outweigh the creative input.

The other company mentioned in the article that hopes to (eventually) monetize A.I. composition is Flow Machines, funded by the European Research Council (ERC) and coordinated by François Pachet (Sony CSL Paris – UMPC).

Flow Machines is remarkably different. Instead of creating a finished product, its intention is to be a musical contributor, generating ideas that others will then expand upon and make their own. Pachet told the Times, “Most people working on A.I. have focused on classical music, but I’ve always been convinced that composing a short, catchy melody is probably the most difficult task.” His intention seems to be to draw on the current pop music model of multiple collaborators/producers offering input on a song that often will be performed by a third party.

While that may be true, I think that the core concept might be closer to “classical music” than he thinks.

While studying at École D’Arts Americaines de Fontainebleau, I took classes in the pedagogy of Nadia Boulanger. Each week would focus on the composition of a different canonical composer. We would study each composer’s tendencies, idiosyncrasies, and quirks through a series of pieces, and were then required to write something in their style. The intention was to internalize what made them unique and inform some of our own writing, if only through expanding our musical language. As Stravinsky said, “Lesser artists borrow, greater artists steal.”

What makes Flow Machine or JukeDeck (or Emily Howell?) different from Boulanger’s methodology? Idiosyncrasies. Each student took something different from that class. They would remember, internalize, and reflect different aspects of what was taught. The intention was never to compose the next Beethoven sonata or Mahler symphony, but to allow for the opportunity to incorporate the compositional tools and techniques into a palate as the student developed. While JukeDeck excludes the human component entirely, Flow Machine removes the learning process that is fundamental to the development of a composer. In creating a shortcut for the origination of new, yet ultimately derivative ideas or idioms, composers may become less capable of making those decisions themselves. The long-term effect could be a generation of composers who cannot create – only expand upon an existing idea.

What would happen if two A.I. programs analyzed the same ten pieces with their unique neural networks and were asked to export a composite? Their output would be different, but likely more closely related than if the same were asked of two human composers. As a follow up, if the same ten pieces were run through the same program on the same day, would they export the same product? What about a week later, after the programs had internalized other materials and connections in their neural networks?

What makes Flow Machine unique is the acknowledgment of its limitations. It is the Trojan Horse of A.I. music. It argues that it won’t replace composition, but help facilitate it with big data strategies. If we were discussing any non-arts industry, it might be championed as a “disruptive innovator.” Yet this becomes a slippery slope. Once we can accept that a program can provide an artistic contribution instead of facilitating the production of an existing work, the precedent has been set. At what point might presenters begin to hire arrangers and editors in lieu of composers?

No one can effectively predict whether systems like Flow Machine will be used by classical composers to supplement their own creativity. Both recording and computer notation programs changed the way that composers compose and engage – each offering accessibility as a trade-off for some other technical element of composition.

I could foresee a future when multiple famous “collaborators” might input a series of musical ideas or suggestions into a program (i.e. playlist of favorite works), and the musically literate person becomes an editor or copyist, working in the background to make it cohesive. Does that sound far-fetched? Imagine the potential for a #SupremeCourtSymphony or #DenzelWashingtonSoundtrack. They could come on stage after the performance and discuss their “musical influences” as one might expect from any post-premiere talkback.

So what does it all mean?

In the short term, the people who make their living creating the work that is already uncredited and replicable by these programs may be in a difficult situation.

A classically trained composer who writes for standard classical outlets (symphony, opera, chamber music, etc.) will not be disadvantaged any further than they already are. Since Beethoven’s death in 1827 and the deification/canonization/historical reflection that followed, living composers have been in constant competition with their non-living counterparts, and even occasionally with their own earlier works. It will (almost) always be less expensive to perform something known than to take the risk to invest in something new. There may be situations where A.I.-composed music is ultimately used in lieu of a contemporary human creation, if only because the cost is more closely comparable to utilization of existing work, but I suspect that the priorities of audiences will not change quite as quickly in situations where music is considered a form of art.

Show me the money

I focused on JukeDeck and Flow Machine over the many other contributors to this field because they are the two with the greatest potential for monetization. (Google’s Magenta is a free-form “let’s make something great together” venture only possible with the funding of Google’s parent company Alphabet behind it, and various other smaller programs are working off of this open-source system.)

Acknowledging monetization is the key question when considering a future outside of academia. The supposed threat of A.I. music is that it might eliminate the (compensated) roles that composers play in the 21st century, and the counter-perspective is how to create more paying work for these artists.

Whether it is a performing arts organization looking to strengthen its bottom line or composers trying to support themselves through their work, acknowledging shifts in consumer priorities is essential to ensuring long-term success. We need to consider that many consumers are seeking a specific kind of experience in both their recorded and live performance that has diverged more in the last 15 years than in the preceding 50.

It is cliché, but we need more disruptive innovations in the field. Until we reach the singularity, A.I. systems will always be aggregators, culling vast quantities of existing data but limited in their ability to create anything fundamentally new.

Some of the most successful examples of projects that have tried to break out of the confines of how we traditionally perceive performance (in no particular order):

  • Hopscotch, with a group of six composers, featuring multiple storylines presented in segments via limousines, developed and produced by The Industry.
  • Ghosts of Crosstown, a site-specific collaboration between six composers, focusing on the rise and fall of an urban center, developed and produced by Opera Memphis.
  • As previously mentioned, Ted Hearne’s The Source, a searing work about Chelsea Manning and her WikiLeaks contributions, with a compiled libretto by Mark Doten. Developed and produced by Beth Morrison Projects (obligatory disclaimer – I worked on this show).
  • David Lang’s anatomy theater—an immersive experience (at the L.A. premiere, the audience ate sausages while a woman was hanged and dissected)—attempting to delve not just into a historical game of grotesque theater, but also creating the mass hysteria that surrounded it (the sheer number of people who were “unsettled” by this work seems to be an accomplishment – and once again, while I did not fully develop this show, I was a part of the initial planning at Beth Morrison Projects).

Craft is not enough. Quoting Debussy, “Works of art make rules but rules do not make works of art.” As we enter this brave new world of man versus machine, competing for revenue derived not just of brawn but increasingly of intellect, composers will ultimately be confronted—either directly or indirectly—with the need to validate their creations as something beyond that of an aggregate.

I am optimistic about the recent trend of deep discussion about who our audiences are and how we can engage them more thoroughly. My sincere hope is that we can continue to move the field forward, embracing technologies that allow creators to grow and develop new work, while finding ways to contextualize the truly magnificent history that extends back to the origins of polyphony. While I am doubtful about the reality of computer origination of ideas upending the system, I’m confident that we can learn from these technological innovations and their incorporation in our lives to understand the changes that need to be made to secure the role of contemporary classical music in the 21st century.

 

Emily Doesn’t Care If You Listen

name

“Cope remembered that Bach wasn’t a machine—once in a while, he broke his rules for the sake of aesthetics. The program didn’t break any rules; Cope hadn’t asked it to.”

—Ryan Blitstein, “Triumph of the Cyborg Composer,” Miller-McCune, February 22, 2010.

Technology evolves at such a fast pace that it is often difficult to discuss a new innovation critically without fearing that the argument will be anachronistic before the discussion is even complete. Yet we must consider the software and hardware around us as they relate directly to the moment we live in, for any attempts at trying to write about the evolution or limitations of the technology of the future is a folly. To date, there is very little by way of critical analysis of David Cope’s computer programs “Emmy” and “Emily Howell” outside of the relatively esoteric field of research into artificial intelligence, and thus I hope to offer one as best as 2011 will allow.

Emmy (a play on EMI—Experiments in Musical Intelligence) is software, capable of creating original music when given certain parameters. If Bach’s six Brandenburg Concerti are uploaded into the program, a new work “in the style” of the Concerti will appear. If Mahler’s Ninth Symphony, Schoenberg’s Verklärte Nacht, and Zemlinsky’s Lyric Symphony are synthesized by Emmy, a work of distinctly late 19th-century Vienna will be produced. As a strong proponent of computer sciences as they relate to music, Cope felt the desire to go further with artificially intelligent composition and put aside the program that merely synthesized the works of previous composers to create a new, dynamic platform, “Emily Howell.” Emily only uses musical sources produced by Emmy, which include everything from Navajo Songs to Mahler, and (officially) the only living composer whose work is present in the database is Cope’s.

The implications of these programs are tremendous. For musicians and audiences alike, the progress and even nature of music could be fundamentally changed. If a computer program could create new music with the push of a button, customizable to anyone’s taste and without distinct intellectual copyright issues, it would revolutionize the way music is produced and consumed in everyday life.

Yet there are technical, philosophical, and practical issues which must be considered when weighing the value of Emmy and Emily Howell in society. How does Emily potentially affect music from the perspectives of composers, performers, computer programmers, classical music fans, and the general public? Does Emily live up to Cope’s claims regarding Howell’s capacity?

What is Music? The answer might seem obvious, but it is important to define what music is before it can properly be discussed. Leonard Bernstein argued in his 1973 Norton Lectures (later titled The Unanswered Question) that based upon Noam Chomsky’s analysis of languages, syntax, and grammar, music functions as a language. By extension, it suggests that music is just a root—as the Phoenician and Cyrillic alphabets offer basic structures for languages to be built, so can music be parsed in various ways to offer many languages. Thus music is not a universal language. Contrary to romantic notions that music transcends nation, culture, and status, any Western listener who has ever sat down to traditional Tibetan chants can attest that what one culture considers sacred, another would consider cacophonous. Tritones (and their surrounding microtones) are considered harmonious in many of these musical languages, while Western classical music referred to the tritone as “the Devil’s Interval” for a significant period of time—not only unpleasant but, by association, immoral.

Languages also inherently evolve over time; while Chaucer wrote in “English,” it is virtually unintelligible to the modern speaker. A mid-point can be seen in Shakespeare—certainly understandable to an English speaker, but syntactically far removed from our present language. Each musical language evolves as well. During the Renaissance, Western classical music revolved around modal harmonies, with the Ionian and Aeolian (major and minor) scales rarely used. From the Baroque until midway through the Romantic Period, Ionian and Aeolian (tonal) harmonies were the principal musical dialects. Toward the end of the Romantic period, tonal harmony began to disintegrate and a slew of new harmonic structures were used. Music then can be said not only to function as a language, but to evolve as one as well. While not immediately capable of relaying concrete ideas (i.e. “Go take out the garbage!”), it is nonetheless a form of communication: A poetic communication, with grammatical and syntactical rules guiding its usage and comprehensibility as the language evolves.

Ultimately, the discussion involves intention. A language is a collection of sounds, organized into roots, expanded into words, and then governed by grammar to take otherwise incoherent tones and give them meaning. Neither “hjew wejklf jjeih” nor “anatomical shelf run quarry” conveys anything, because they are not following the basic rules of the language used to write this essay. Is a birdcall music or literal language? To a bird, it communicates a specific idea (danger, mating, etc.) and thus should be considered a literal language. The average human lacks the ability to understand the intended message, yet we can reconceive it and render it into a musical concept. A powerful and congruent example of this can be seen in Olivier Messiaen’s fascination with birds and his usage of their songs in music and ornithography. He spent a significant portion of his life cataloging birdcalls, using musical notation only as a medium for documentation. Yet in the composition of his Quartet for the End of Time, he utilizes birdcalls interpretively, using the free harmonic language of the nightingale to conjure an image of an Eden-like heaven. In doing so, he removes the literal meaning of the birdcalls for the picture he is painting in tones—and it becomes music.

As we relate this back to computers attempting to communicate, why aren’t there any computers creating works of literature, either in the style of a specific author or in their own, unique style? Would the general public want to consume a book written by a computer? Would it have any substance?

At the Rensselaer Polytechnic Institute, scientists have made a noble attempt. Their program, “Brutus,” can produce simple stories, based upon certain paradigms of contemporary struggle and conflict. The researchers state:

We use Brutus.1 in support of our philosophy of Weak Artificial Intelligence—basically, the view that computers will never be genuinely conscious, but computers can be cleverly programmed to “appear” to be, in this case, literarily creative.

(—From the website for Brutus;
examples of the literary output can also be read there.)

While it is fascinating to observe a computer able to use algorithms to analyze the basic elements of human interaction and mimic them, the end result is, at least at this time in computer development, lacking in the originality of insight into a human issue. A computer can only generate responses based upon previously programmed ideas. This “Weak Artificial Intelligence” can only mimic the process of learning—its ability to learn is confined to the parameters it already understands. Profound literature is created through a unique perspective, conveyed through language in a way that resonates with readers. By definition, a computer cannot have a unique perspective or original idea until it is able to achieve true artificial intelligence—specifically, self-awareness.

name

So why then has Emily Howell gained a certain amount of popular acceptance through live performances and record sales while her literary counterpart remains an academic novelty?

Music is mystical. Or more appropriately, its composition is romanticized. Most people “compose” text on a daily basis, whether for mundane or creative purposes. Comparatively few compose music so prolifically. Improvisational music can be viewed as the one field of composition that captures the same freedom and immediacy in the use of a musical language, but it requires a certain amount of training, both in the grammar of a musical language and in the technical execution required to perform it on an instrument. So the composition of music is left to “talented” or “gifted” individuals, while the majority of society remains content simply to consume the products. This is where the Turing Test becomes a significant factor.

In 1950, the English mathematician and computer scientist Alan Turing proposed a test for determining true artificial intelligence. It proceeds as follows: a human judge engages in a natural language conversation with one human and one machine, each of which tries to appear human. All participants are placed in isolated locations. If the judge cannot reliably tell the machine from the human, the machine is said to have passed the test. In order to test the machine’s intelligence rather than its ability to render words into audio, the conversation is limited to a text-only channel such as a computer keyboard and screen. To date, no computer has won the Loebner Grand Prize, an ongoing challenge for a computer to pass the Turing Test. (There is an annual prize for the computer “which is most indistinguishable from a human,” thus attempts are regularly made.)

The most popular story about Emily Howell suggests that she does indeed pass the Turing Test. A 2009 article by Murad Ahmed published in The Times (London), offers an un-cited and un-attributed anecdote from Cope about a professor who came to him a break in the middle of a lecture performance at UC Santa Cruz, where both professors teach. After the piece was played, the “professor came to me and said this was one of the most beautiful pieces he’d heard in a long time.” Then, after listening to the lecture and learning that it was written by a computer, he was reported to have said, “From the minute it started I could tell it was computer-composed. It has no guts, no emotion, no soul.”

Cope’s anecdote extends the parameters of the Turing Test to music, offers that Emily has passed it, and implies that academia is reluctant to acknowledge his brilliant discovery. Yet looking more carefully at his statement, a serious question of judgment arises. Who was this professor? The reader is lead to believe that he is a professor of music—otherwise why mention his status within the school. More importantly, why would this “professor,” who presumably knew Cope’s work, go to a lecture recital, feel inclined to make such strong statements in support of the work, only to make a contradictory statement, insulting Cope’s achievement, shortly after?

There is no question that Cope, in devoting himself to his science, has anthropomorphized Emily and is fighting to gain recognition for her1, but to what extent? Music is a poetic, interpretive language, not a literal one, thus the standards for the Turing Test cannot be directly applied. Many works by living composers, such as Pierre Boulez’s Structures 1a, would more readily be accepted as the music of computers than Howell’s, but the intention behind Structures 1a was to give way to an intricately conceived, emotionless canvas. As an academic deeply involved in the world of artificial intelligence (he has written seven books and twenty articles on the subject of music as it relates to computers, with three of the books and nine of the articles dealing specifically with AI, as well as a creator of Artificial Intelligence Poetry Program), Cope’s attempts to suggest that computer music can be judged by the Turing Test don’t adequately consider the basic elements of the test as they relate to musical languages, as well as the limitations of the linguistic test itself.2

In many ways, this conversation is directed towards the producers of music—both the composers and the performers who devote their lives to the creation of music. Emily Howell is an extension of the iTunes “Genius Function,” and the “Personalized Recommendations” function utilized by Netflix, Amazon.com, Pandora, and a host of other media outlets that try to predict human desires to increase sales. Most consumers of this media appreciate the “Personalized” results, as it offers them suggestions as to what they will want to purchase next, giving them something new without straying too far from what they already know. The popular dating site eHarmony.com acknowledges the limitations of its algorithm, and after a 30-minute personality test, informs as many as twenty percent of its potential clients that they do not fall into a specific category, and are thus “unmatchable.”3 While this might be disheartening to a single individual in search of love, it suggests that human preferences cannot be simply reduced to a series of parameters.

As producers of music, it is important to understand the limitations of assuming that groupings or categories of sounds will elicit specific emotions. While a minor key often implies a more somber mood in Western classical music (obviously with exceptions), the effective usage of a key is different from composer to composer and piece to piece. At this moment, artificial intelligence cannot understand the subtle gradients, or their reception without a human intermediary.

Emily Howell has proved effective at mimicking a composer fairly well. The entire project was created when David Cope, struggling to complete an opera commission, designed a program that understood his music idioms, tendencies, and syntax, and then produced new material in his style. He used this basic material as a guideline to complete the opera. After this, Cope began to expand the parameters, inputting works of other composers into Emmy, trying to create a program that would offer something as authentic as the original. The idea itself was not new—Fritz Kreisler performed many “lost” works of the baroque era, only to later reveal that they were of his own pen, and Remo Giazotto used the bombing of Dresden to “uncover” several works “by Albinoni,” most famously the Adagio in G Minor. Yet there is a notable difference: while Kreisler and Giazotto were able to pass off the works as originals because of their fame, closer analysis of the works show distinct signs of musical influences other than the supposed original composer. Emmy’s output proved to be mediocre examples of the original composer, with no uncharacteristic elements, but also none of the “brilliance.” In the section of his website devoted to “Experiments in Musical Intelligence,” Cope describes Emmy’s process as:

(1) deconstruction (analyze and separate into parts)
(2) signatures (commonality – retain that which signifies style)
(3) compatibility (recombinancy – recombine into new works)

This process created a composite of the functional norms, and by his own admission, the results were rather uninspiring.

In “Creative Writers and Daydreaming,” Freud described the creative process as “the suspension of rational principles.” A computer can create an infinite number of random possibilities, but the concept of suspended rationality has yet to be attainted by a computer. As Cope explains in “Experiments in Musical Intelligence,” it was the lack of “errors,” or intentional breaking of the rules, that made Emmy’s work plastic. As nuanced as the programming can be, the difference between irrational and random (or intentional breaking of the rules for an aesthetic reason versus arbitrary straying because a certain amount of unpredictability is required) is what separates creation from amalgamation.

There is no question that musical borrowing and referential passages play a significant role in composition. The Western classical tradition teaches composition through the study of the canon. Nadia Boulanger, the famous compositional pedagogue, explored a new composer each week with her class, analyzing pieces of the composer and then requiring her students to write a piece in that composer’s style. Yet, as Igor Stravinsky famously remarked, “A bad composer borrows, a good composer steals.” This suggests that to be influenced by another, one cannot simply break down the process and imitate, but must incorporate it into their own personal language. Boulanger’s methodology taught her students to understand the creative process of historical figures to enrich their vocabulary—teaching them grammar that they would then mold into their own voices—as seen in the multifarious careers of her pupils (Aaron Copland, Roy Harris, Virgil Thomson, Elliott Carter, Quincy Jones, Philip Glass, etc.). But Emily cannot generate her own style, because, as limited by a lack of self, she cannot interpret the works of others; she can only analyze them.4

There is something to be said for the sheer technical achievement that is Emily Howell. In a way, she is a work of art. From the technical perspective, she is on par with the most advanced forms of artificial intelligence currently in production, and more so because her product can mimic “human” music to the extent that listeners don’t assume that it is computer music. Yet having had the opportunity to review two scores (of six total works) by Emily Howell and compare them to the official recordings supervised by David Cope, I stumbled upon a fascinating issue: the music used for the recording has been edited from that which was originally produced by Emily Howell. Specifically in Land of Stone, Opus 3, passages are modified, instruments and parts are added and rhythms, octaves, and dynamics are changed. Also, comparing the score markings in Land of Stone to From Darkness, Light, Opus 1, there are significant differences in the quantity of markings. While From Darkness, Light is very loosely annotated (in the style of Bach, as seen below), the later work, for a chamber orchestra, is mired by articulations, stylistic directions and various dynamic and tempo markings.

name
From Darkness, Light, ii. Fugue, measures 7-9

I have broken down the changes into four categories: additions or subtractions; modifications (excluding passages that could conceivably be attributed to player error—in the last major section, from measures 165-263, the rhythms become complex and the accuracy of the recording suffers in this section, thus only pitch, register, and additions are considered there); errors (i.e. passages that are impossible or that contain incoherent markings); and passages that can reasonably be assumed to have stylistic markings not originating from the computer. There were at least 12 additions or subtractions, 29 modifications, 68 edits, and 17 errors in the 275 measures that comprise this piece.5

This lead me to contact Nicole Paiement, the music director of Ensemble Parallèle, the group that recorded Land of Stone for Cope’s commercial recording. She confirmed much of what I had inferred regarding the active editing of the work. When asked if and how the piece changed, she said:

There was more of an evolution of the piece prior to the performance. As we started to work on the piece we adapted things that were not working as well… So it was as if [Cope] sort of became the computer-composer. So he would listen, and most times he agreed, and at times he said, “we would keep that,” but generally speaking David was very agreeable to doing the changes that would make the piece a better piece.

Where then does this leave Emily Howell’s work? Aside from its inability to properly grasp the capacity of the instruments in the orchestra (i.e. asking a harp to crescendo and decrescendo with a single struck note, as in measures 12-13, writing out of the range of an instrument, such as the D5 written for the contrabass in measure 185), it is clear that Cope felt free to edit and modify “Emily’s” work as he saw it fit. In most every case of an addition (i.e. mm. 65-85, the addition of the harp and glockenspiel outlining the suggested harmonies in the bassoon and clarinet, various examples of instruments moved 8va or 8vb because the range it was written in was not conducive to the mood, the flute introduced a measure before each of the “bell-like” sections, measures 138, 149, 154), there is a good musical reasoning for it. Furthermore, specialized stylistic additions turn theoretical structures into quasi-lyrical passages, such as the addition of Senza Tempo at measure 57, allowing the pianist to freely explore a somewhat random twelve-tone quasi-aggregate6 phrase:

name
Land of Stone, Measures 57-58

So while Cope states in his program notes for Land of Stone that:

My role with Emily is to provoke her to compose good music, and once composed, act as her impresario—her agent —to obtain performances, recordings, and so on…

It is clear that he is integral to the compositional process. Despite his assertion that Emily Howell produces “good music” on her own, the interference with the program’s output, both through his editing and Paiement’s critiques, suggests that his claim of computer-composed music is still a distant vision.

Emily is a testament to the nature of music: by parsing compositions down to their bare elements, she offers definitive proof of music as a language. She can rapidly produce composite works, changing the way à la carte music is created for daily consumption, but let’s not confuse these experiments in mass-produced Muzak with human attempts to create art.

As this discussion moves toward the perspective of the audience member, an important question arises: Do people need a relationship with a composer?

To illustrate this question regarding audiences and their expectations, I offer the prostitute metaphor. A prostitute is paid for by friends of a particular person without his knowledge. She is instructed to approach him at a party as if she were genuinely interested in him. The man cannot believe his good fortune—she is the paradigm of what a woman should be. She is beautiful, intelligent, and finds him witty and humorous. The man brings her home and wakes in the morning to find her gone. Disheartened, he calls his closest friend to recount his brief encounter with love, only to be told that all of his perceptions were ill-conceived. He made the assumption that when she approached him at the party, she found him attractive. He then further assumed that she was enjoying the time spent with him as they conversed and flirted. As he went to sleep that night, he most certainly assumed that she would at least be there the next morning.

Does this man’s discovery that his conception of the relationship was based on a lie change his appreciation of the event? I believe he might be forced to see it as nothing more than a mechanical act, destroying his concept of the presumed relationship.
There is an argument that music is nothing more than notes on a page. In theory, it is. It can sit in a cabinet for centuries and be as irrelevant as the dust that collects upon it. Yet once it is performed, there are inherent implications; including the assumption that this is an interaction between a composer and a listener. Only Emmy and Emily Howell can truly be impartial to music, breaking it down to its elemental roots without any response to the composite sound. As human beings, we have been exposed to music, and “understand” the language (although different dialects and varying degrees of “literacy” and natural sensitivity exist).

Another issue to take into account when considering the public: Who owns Emily Howell’s intellectual property rights? First, the issue of her sources needs to be discussed. Cope says of her musical sources:

In January of 2003, I retired my Experiments in Musical Intelligence program (Emmy).

My reasons were many, but mostly this retirement resulted from my desire to get on with my life. I had for years envisioned a new program which, unlike Emmy, included many aspects of computational creativity that Emmy did not include (I discuss this in depth in my book Computer Models of Musical Creativity, MIT Press, 2005). I call this new program Emily Howell.

Interestingly, the database for Emily Howell consists mainly of music composed by Emmy, thus linking this new program with my previous work in rather explicit and intimate ways. Emily’s style, a hybrid of the many styles which Emmy emulated during her twenty-year career, has fits and starts of Ivesian-sounding pastiches. My goal involves her creating a completely new style — a composite of those that most effectively complement one another.7

Emmy was intended solely to create additional material in the style of a previous composer, so her musical sources were (officially) the great composers of the past, as well as works by Cope. Yet theoretically, any composer could be plugged into the program to produce something similar. Assuming an orchestra wanted to perform a piece by Copland, but didn’t want to pay the rental fees, they could plug several pieces of Copland into the program and title the work Hommage à Copland. Unscrupulous composers could freely exploit works by their colleagues, using Emily Howell to create a work using sources from several pieces still under copyright. Or it could cheapen contemporary composition altogether, as Cope freely suggests in his original reasoning for creating Emmy—writer’s block.

These aspects should be troubling to anyone in the field of music. If the cost of producing “new” music through Emily is exponentially lower and the product created faster than through traditional composition, there is no question that it would be widely used. Yet the previously established inability of Emily to create truly original work will prevent music from evolving. Living composers, already struggling to exist outside of academia, simply won’t have the ability to compete, and Milton Babbitt’s vision of composition as a science completely removed from the real world, summarized in his article “Who Cares if You Listen?,”8 will be a reality.

Something which I had initially hope to explore further was Emily Howell’s ability to set text. Specifically because Cope said he originally used Emmy to complete an opera, and one of Emily Howell’s five works is From the Willow’s Keep, Op. 4, I requested this score from Cope with the hopes of offering a critical analysis, but was only provided with Land of Stone and From Darkness, Light. At the same time, there is no documented text in the English language with the phrase “From the Willow’s Keep,” so the text for his piece may have been produced by Cope’s “Poet Program.”9 As it relates to contemporary popular composition, the setting of text would be immensely important. Cope alluded to a “well-known pop group” inquiring about utilizing Emily Howell,10 thus the practical applications in popular music must be considered. “Emily Howell is adaptable and egolessly self-modifying in her ability to respond to audience criticism.”11 This is an audacious claim, both as it relates to classical and popular music. First, how is Emily able to evaluate and respond to audience criticism? Is she evaluating based upon “hoots and hollers?” Can she distinguish between a hoot and a holler? Presumably, she is responding using a decibel meter—but this cannot really compare a room full of jeers to a room full of cheers (or more likely a combination of both). Possibly in the future, each member of the audience will be connected to a device measuring serotonin levels, and a “Matrix”-esque construction of the auditory experience will be a reality. Furthermore, since standard concert etiquette suggests that audiences only show signs of approval or disapproval at the end of a piece, how exactly does this egoless self-modification occur? How does she “realize” that a certain note, passage, or even section was disliked? This again is a romanticized notion of Emily functioning without the help of Cope. Yet going further, her egoless state means she fundamentally has no artistic integrity. Right now she produces “pretty” music. Warren Riddle, editor for the popular tech magazine Switched, ended his optimistic article on Cope’s work with the bombastic tirade:

The current offerings (of Emily Howell) may be limited, but they’re definitely more appealing than the soulless, conformist tripe that so-called “humans” are currently regurgitating at an alarming rate.

Is there nothing to stop Emily Howell from churning out “soulless, conformist tripe”? She is nothing but a computer program, and can be utilized in a myriad of ways.

Regarding the practical application of Emily Howell to popular music, it could be the device that ultimately ends professional music on any level. Guitar Hero, Rock Band, and any number of other music simulation games offer people with no prerequisite capacity for music the opportunity to feel like they are creating it. If Emily Howell were widely distributed, anyone could mix their favorite five songs together to create a composite they would feel is theirs. In theory it is “music for the masses,” no holds barred, but in reality, once the computer becomes the originator instead of facilitator, communication ceases to happen, and the product reverts back to irrelevant sounds.

There may come a time when computers achieve true artificial intelligence, at which point, much of this argument will be moot. I am hesitant to stand in the way of progress, and historically some of the greatest inventions of our time were eviscerated by critics of their time for being the downfall of some great legacy. I believe, however, that the benefits of Emily Howell and other computer-composed technologies are currently exaggerated and little more than novelties, and as we explore this new frontier of computer-assisted composition, we should be wary of potential negative consequences.

***

Notes

[Ed. Note: In addition to the comments and references cited below, Noah Stern Weber, the author of the present article, wanted to cite Andrew Aziz’s article, “Algorithmic Style Analysis,” a history and evolution of artificial intelligence as it relates to music analysis and composition, which focuses on Cope quite a bit. According to Weber, “The article is fairly technical, but is the best discussion of the actual compositional process that is shorter than a book. It may be too esoteric for a general audience, but I found it very helpful as I began my research for this article. I believe it would be of interest to those who want to learn more on the subject.”—FJO]

1.“Unfortunately, I fear that she, like all of us involved with the making of new art, faces a hellish uphill battle, for all too often, performers, reviewers, and audiences rely on the proven rather than on the risky, and choose to repeat rather than initiate our creative experiences. What a shame that Emily’s music, like the music of her human counterparts, will most likely find itself buried in complacency, never having the opportunity to be heard and appreciated in her time.”—Excerpted from the Introductory Notes to the Emily Howell CD From Darkness, Light, Centaur 3023 (2010))

2. The Loebner Competition, an annual test of a computer’s capacity to pass the Turing Test is limited to a five-minute discussion. The computer’s best course of action was summarized by three time “Most Human Computer Award” Winner Richard Wallace as engaging in “stateless” conversation. “Experience with [Wallace’s chatbot] ALICE indicates that most casual conversation is ‘stateless,’ that is, each reply depends only on the current query, without any knowledge of the history of the conversation required to formulate the reply.” The programs that do best in the Loebner Competition tend to speak like a teenager through text-message, preprogrammed with stock responses, pop-culture references, and innocuous jokes. When it does not know how to respond, it abruptly changes the subject. (For a discussion on the Loebner Competition and AI, see: Brian Christian, “Mind vs. Machine,” The Atlantic, March 2011.)

3.A standard unmatcheable message from eHarmony: eHarmony is based upon a complex matching system developed through extensive testing of married individuals. One of the requirements for it to work successfully is for participants to fall into our rigorously defined profiles. If we aren’t able to match a user well using these profiles, the most considerate approach is to inform them early in the process… Our matching system is not suitable for about 20% of potential users, so 1 in 5 people simply would not benefit from our service.

4. This returns to the idea of suspended rationality. Howell’s algorithms only allow for an averaging of the discrepancies between a work and her core knowledge. If ten Emily Howells were given the task of analyzing the same series of works and producing something with that composite, the output would be very similar. Humans will be drawn to different elements of the series and will produce markedly different works. Our learning is idiosyncratic; it is what differentiates personalities and dispositions. We valuate new ideas by combining our past knowledge with our personal aesthetics and tastes. A person can be fundamentally changed by a concept that challenges their preconceived notions of a subject—Emily Howell’s algorithm can only weigh it against what it already knows.

5.Because of size constraints, I could not highlight the vast majority of these issues, but a copy of the score with each discrepancy can be downloaded at noahsweber.com/supplement.htm.

6.This cycle repeats every two measures: Nine of the twelve tones are used repeatedly, avoiding the last three. The missing tones are played simultaneously by the accented piano and the clarinet (later the flute), completing the aggregate, at which point the sequence begins again.

7.Excerpted from the introductory notes included in the score of Land of Stone.

8.“The attitude towards the indisputable facts of the status and condition of the composer of what we will, for the moment, designate as ‘serious,’ ‘advanced,’ contemporary music. This composer expends an enormous amount of time and energy—and, usually, considerable money—on the creation of a commodity which has little, no, or negative commodity value. He is, in essence, a “vanity” composer. The general public is largely unaware of and uninterested in his music. The majority of performers shun it and resent it. Consequently, the music is little performed, and then primarily at poorly attended concerts before an audience consisting in the main of fellow ‘professionals’. At best, the music would appear to be for, of, and by specialists.” (“Who Cares if You Listen?” High Fidelity, VIII/2 [February 1958] pp38-40, 126-7.)

9.Cope has written a program similar to Brutus.1 which composes poems.

10.Warren Riddle, ” A Robo-Symphony: David Cope Composes ‘Human’ Music With A.I.Switched, February 27, 2010.

11.Surfdaddy Orca, “Has Emily Howell Passed the Turing Test?,” H+ Magazine, March 22, 2010

***

name
Noah Stern Weber

Noah Stern Weber is the Founder and President of Burning Bayreuth, a concert series devoted to innovative presentation of classical and contemporary music. He has degrees in clarinet and conducting from the Peabody Conservatory of Music and the Bard College-Conservatory of Music, and has served as assistant conductor for the Center City Opera Theater, the Luzerne Music Center, the Gulf Coast Symphony and École d’Art Americaines de Fontainebleau.