Generation & Composition
生成と作曲

As we enter the 2020s, artificial intelligence (AI) technology is becoming more common in music.
Now that music is being created without the need for people to create it, it is necessary to redefine rights, originality, and creativity, which have not been considered in the past, and musicians may be asked what they create and for whom.

The creation of music by machines is sometimes considered separately from traditional composition and called generation, since it is sometimes created by a program or algorithm that computes instead of a human brain.
However, what is produced is music in both cases.
Is there really a difference between generation and composition?
In this page, we will consider the fundamental question of what creativity means in music, looking back at the history of composition in Western music from the Middle Ages to see what is different and what is the same between traditional composition and machine generation.

The history of compositional techniques runs parallel to the history of human music.
Knowing the history of composition may be a guide to imagining where music is going today and how the future may change.

Western music can be traced back as far as the monophonic Gregorian chant of the 9th century, but at that time the author was unknown and the concept of a composer had not yet become independent.
Music meant singing a chant that already existed.
It is said that the first composer in history to consciously assign a value to his work by naming it as the author and releasing a manuscript of his work before his death was Guillaume de Machaut in 14th century France.
Since then, it has become common practice for composers to write their names on the scores of their works.

Since the Middle Ages, compositional techniques have evolved from monophonic to polyphonic music in which multiple melodies are played simultaneously, reaching its peak in the 16th century.
Then, from the 17th century, there was a shift to tonal harmony music, which is driven by the linking of chords that have a function.
From the 20th century onward, even a single composer wrote works in a variety of styles. It is no longer unusual to find music in which neither melody nor rhythm exists, and we have reached the present day.

In the first period when music became polyphonic, composition was considered to be the writing of melodies in the upper part of the tenor fixed melody, which was a long-note stretched version of the existing Gregorian chant.
Creating original music from scratch was not considered composition.
Cantus firmus is not audible in the original chant, but the values of not finding beauty only in what is audible, as is the case in atonal music of the 20th century, were already perfected in this period.
The range of harmonic intervals in music up to the Middle Ages was also different from that of today.
For example, the third degree interval, which is essential for creating beautiful harmonies in the modern age, was considered a dissonant interval in the past.
The support base of musicians of this period was the catholic church.
The definition of what is beautiful is not so much a universal aural truth as it is changing with the demands of the authorities of the time.

The Renaissance period marks the pinnacle of polyphonic musical technique. One of the most important concepts to emerge from this period is that of imitation. A single melodic theme is alternately passed on to multiple voices in different time periods, creating dialogue and interaction, with the effect of solidifying the unified musical idea as a whole.

In the 16th century, when the composer belonged to the church and was greatly influenced by it, the Reformation was taking place and the power of Catholicism was threatened, and clear, simple music that the whole congregation could sing was preferred over complex polyphony with many voices in a trained choir.
The composer Giovanni Pierluigi da Palestrina, fearing that church censorship would cause music to revert to monophony, began writing masses in four voices, with clearly audible lyrics, and this led to the tonal music of the Baroque and Classical periods that followed.

The Baroque period of the 17th century, when the weakened church was replaced by royalty and nobility who became patrons of composers in order to exert their power, replaced polyphonic music that continued until the Renaissance and evolved into a form of single melody plus chordal accompaniment in order to give it narrative and passionate expression.
As musical instruments developed, a style of instrumentation other than that of song accompaniment began to be explored. A style called "basso continuo," in which players improvise within the rules of the score by adding ornaments to the bass line, was born, and the concerto and chamber music forms that are used today developed.
The leading composer of this period was Claudio Monteverdi, who distinguished between the former and the latter, regarding the former as Prima pratica and the latter as Seconda pratica.
After the Baroque, music became more harmonic, and harmony, which until the Renaissance had arisen as a side effect of the layering of melodies, changed to the idea of progression through the concatenation of vertically stacked chords.

The era that most people associate with classical music is probably the Romantic period of the 19th century.
Composers became financially independent from both the church and the aristocracy, and were able to publish scores and perform them publicly according to their own plans, and harmony became more complex toward the 19th century, when it became possible to express oneself musically.
Romantic composers attempted to break the norms of traditional forms and structures. This led to extensions and transformations of traditional harmony and form.

The music of Igor Stravinsky, with its multiple tonalities and complexities, has been the subject of riots between those who thought it was new music and those who felt that it was not. The expansion of music never stopped, leading to the atonal music of the 20th century, which removes the center of gravity from music by treating all 12 notes of an octave as equivalent.

Furthermore, after Thomas Edison's invention of recording technology in the 19th century, music became more accessible to the public, and popular music became the dominant form of music listening. Composers were able to modify their works by listening to the sound, and creativity and the spirit of experimentation expanded.
Since the democratization of computers, various electroacoustic experiments and works have become easily accessible to the public, and the evolution is ongoing.
The real history of popular music, which is slightly different from that of secular music in the pre-democratic era when no records were kept, and before the invention of recording technology and the Internet, may have just begun.
Understanding how classical music composition has evolved in relation to society may help us predict the future.

Music is an art completed not only through composition and performance, but also through listening. Next, we will look at how the environment for listening to music has evolved.

Music until the Renaissance was religious music for liturgical purposes in churches and music for social occasions of the nobility during the Baroque period, and it was not until the 17th century that theaters were born and public performances became available for the general public to enjoy.
Until the 18th century, when classical music was contemporaneous music, the subject of the repertoire of works performed did not extend back hundreds of years.
Johannes Tinctoris, a Renaissance music theorist who defined composition and improvisation separately and wrote numerous books on counterpoint and music theory, wrote in a 1477 treatise that there is no older work worth hearing than those created in the last 40 years.
https://earlymusictheory.org/Tinctoris/Tinctoris/BiographicalOutline/ This is an unthinkable value in today's classical music concert culture, where performances of works by composers who were active hundreds of years ago are taken for granted, and one can imagine that the definition of what is musical and what is creative has always changed with the times.
s Popular music, on the other hand, may not be as ingrained in the culture of recycling past works as classical music.

The 19th century composer Felix Mendelssohn was the first to rewrite the centuries-old music industry practice of not performing works from the past.
His arrangement and restaging of J.S.Bach's Matthew Passion for that era led to the rediscovery of Bach as a composer and his worldwide recognition, but before that, Bach was a forgotten figure who was only appreciated in a few regions.
Since then, classical music has become, and continues to be, a practice of re-performing works from the past over and above the works of living composers of the same era.
However, classical music has not been a long tradition in the long history of Western music since the Romanticism of the 19th century, when it became a culture of performing music from the past.

Since the invention of recording technology by Thomas Edison at the end of the 19th century, music has been recorded and sold, allowing more people to experience and enjoy the works of composers.
This has facilitated the spread of a composer's influence and fame throughout the world and has helped music to take root in the lives of more and more people.
Recording technology has also contributed as a means of recording and preserving the performances of past composers and performers.
Historical recordings have become a valuable resource for future generations of composers and play an important role in understanding musical traditions and evolution.

In the 21st century, streaming playback over the Internet has become widespread, and now that the option of music appreciation that does not depend on live performers has become commonplace, what kind of music experience are people looking for?
With such a vast amount of past works available and the ability to immediately play back recordings, why create new music?
Furthermore, are musicians really expected to create music in a situation where AI is creating works with a higher degree of perfection at a speed that surpasses that of humans?
It is hard to imagine the culture of music disappearing from the world, but there are so many recordings of musical works that even if we keep playing Spotify and Apple Music until we die, we will not be able to listen to them all.
In this age of music, what are the activities that should be undertaken by musicians who are experts in music?

The method of creating notes and musical structures logically and computationally using arithmetic rather than human thought is called algorithmic composition, and its practice began in the 1950s.
It may seem like the sudden appearance of a new concept that has never existed in the history of music, but an algorithm is a constraint, and in the world of music, it is a concept similar to the style that is brought about by prohibitions.

Music theory, the study of finding and systematically summarizing the rules inherent in existing musical compositions, has been the result of various academic studies for centuries of humanity.
On the other hand, composers are conscious of creating authorship and style through the rules and restrictions they plan to impose on their own compositions.
For example, part-writing in counterpoint and harmony also has a great many restrictions, and it is possible to compose music that fulfills one particular period style just by adhering to its rules, but there are still too many random elements that may result in tasteless, textbook music unless more restrictions are set by the composer.
If the composer adheres to most of the prohibitive rules with reference to traditional theory, and then imposes certain additional detailed restrictions, the work will take on the distinctive style of the composer's individual style.
On the other hand, it is also creative to design and produce music with sound composition rules that are completely different from traditional theory.
When taking this approach, it is inevitable to consider the fundamental question of what music is, as one unfamiliar piece of music after another is created.
John Cage is a prime example of a composer who conceived of the generation of music from rules unknown in the history of music.
Even though composers are not conscious of the term algorithm, there is a history of using, inventing, and transforming it over the centuries to create works as a concept of style, which has a close meaning.
Next, we will look at specific musical techniques from a time when computers did not exist, keeping in mind that algorithms approximate style.

The mechanism of generatively creating music can be traced back in time as far as ancient Greece to systems such as the Aeolian harp, which used the force of wind to vibrate the strings.
The sense of scale in which the musical structure is created using only the forces of nature, avoiding artificial calculations, is similar to that of contemporary art.
Ancient music was an important discipline along with arithmetic, geometry, and astronomy, Not a few theories, such as musical scales and intervals, have been handed down from the ancient Greeks and remain in use today.

The various achievements of Guido d'Arezzo (992 - 1050), an important Italian medieval music theorist, have left their imprint far into the present day.
The solmization, which represents the height of a note in Do Re Mi, which is taken for granted not only by musicians but also by many people around the world, was invented by him and takes its name from the initial letters (Ut, Re, Mi, Fa, Sol, La) of the punctuation in the lyrics of the chant "Ut queant laxis".
Ut was changed by musicologist Giovanni Battista Doni in the 17th century to Do, meaning Dominus and also the first letter of his name, to facilitate pronunciation.
The arithmetical composition process presented by Guido in his eleventh-century treatise Micrologus involves the transformation of a Latin text into a melody under certain rules.
This is a rule-based composition technique in which the composer is not free to create music, but rather the music is automatically determined depending on the text and lyrics, and is like data mapping in media art, and is now considered one of the oldest algorithmic composition techniques.

Invented by 14th century composer Philippe de Vitry, isorhythm later reached its peak with Guillaume de Machaut.
This is a compositional technique that combines a set of pitches called colore and a set of rhythmic patterns called talea to produce complex hochetus and other pieces.
It is a kind of repetition technique, but since the number of notes in the talea differs from the number of notes in the corolla, it avoids simple repetition of patterns that could be made close together.
Compositions at that time were required to quote Gregorian chant as Cantus Firmus. While using this sequence of notes as a corollary, a new rhythmic pattern was produced as a talea, and the two were combined to create a melody.
The minimalist approach has attracted attention from minimalist composers, and even in the 20th century, composers such as Reich, Pärt, Messiaen, and Cage have applied isorhythms to their works.

The Renaissance canon is a technique in which the singer begins to sing the theme of a fixed melody at different times, and the theory is that the harmony is determined by the timing at which the singer begins.
The algorithm is what controls the timing. Unlike fugues, which are relatively free, canons require strict imitation of the melody based on a set of rules.

Mozart, the master of classical music, was another composer who composed music using automatic generative processes.
This was a bit of a fad at the time, as several composers, not just Mozart, were experimenting with this kind of thing.
The algorithm inherent in Musikalisches Wurfelspiel is to number the myriad of several bars of fragmented music and randomly select the order of each fragment using dice.

Perhaps no compositional technique is more infamous in music history than the twelve-tone technique. Atonal music, as represented by Arnold Schoenberg and his students Anton Webern and Alban Berg of the Second Viennese School, eliminated all of the gravity and centrality of sound that was typical of it in order to explore the next era of tonal music.
In order to make the value of all 12 semitones equal, they revived the concept of note rows and canons that had existed since the Middle Ages, and succeeded in creating music that kept the probability of occurrence of all notes equal by performing mathematical operations such as antigravity, retrograde, and inverse retrograde on the sequence of notes determined by the composer in the matrix.
However, the dissonant music produced by this overly systematic compositional technique continues to be one of the most controversial sounds in history, often isolated from understanding even by performers who are supposed to be musicians like the composer.
Webern studied musicology under Guido Adler, and received a degree in musicology from the University of California, Berkeley, with a thesis on Heinrich Isaak, His interest in the compositional techniques of polyphony during the Renaissance influenced his structuring of atonal music.

Continuing to confront the question of what composition is, John Cage used entirely new compositional techniques based on various concepts, such as chance music and indeterminate music, to escape the old stereotypes. One of the earliest examples of this is the 20th century application of a concept called gamut, which has been around since the time of d'Arezzo.
The music was created by sensuously picking up fragments of diverse sound materials, including dozens of arbitrarily prepared single notes, heavy notes, and chords, and piecing them together in a patchwork-like fashion.
There are no conventional classical musical conventions.
The next step was to prescribe a strict arrangement for the gamut, using chance operations to move up, down, left, and right, and finally to choose at random.
Cage's goal was to create a world where music is automatically generated according to rules that do not easily reflect the author's intentions.
Although Cage initially interacted with and influenced Boulez, who led the European avant-garde movement, he gradually became more and more devoted to Eastern philosophy and expanded his activities into a far wider range of fields, including contemporary art, and his unique view of music parted company with the avant-garde.

The works of Estonian composer Arvo Pärt are also very often systematically generated in an arithmetic manner.
His works in the tintinnabuli style, such as Für Alina and Berliner Messe, which are based on two-voice counterpoint consisting of M-Voice, the sequential progression of diatonic scales, and T-Voice, the main triadic arpeggiation, in order to redefine tonal music, are like Renaissance music, and the 12 The systematic technique after the 12-tone technique further pushed minimalism in some aspects.
One of his masterpieces, Cantus in memoriam Benjamin Britten, is a very experimental approach in which the first violin melody begins with a single note, then expands to two notes, then three, and finally a complete melody appears.
Again, medieval isorhythms are applied here, and for the fixed two-note tarea rhythmic pattern of one half-note and one quarter-note in less than one measure, the colore continues to extend to the end.
In this way, the repetitions and variations are completely different from the music that loops bar by bar.
Talea is expanded to double, quadruple, etc. its size in the voices of the 2nd violin, viola, and cello, which is also an application of the medieval canon technique.

Through programmatic and automatic composition and generation of compositional algorithms, it is easier to evaluate the musical style itself that the algorithm encompasses.
Normally, a style cannot be seen unless a number of musical works are created through the act of composition, but with algorithmic composition, an infinite number of pieces within the same style can be generated instantly, making it easy to identify the characteristics of a style (algorithm).
Next, we will look at examples of music composed by computer arithmetic since the invention of the computer.

The first example of the use of computer algorithms in composition is Hiller's String Quartet, published in 1956 by Lejaren Hiller and Leonard Issacson of the University of Illinois.
In this piece, the Iliac computer at the University of Illinois generates pitch, rhythm, and other musical symbols and outputs them as alphabetical and numeric combinations. This is then translated by human hands to complete the score for a traditional string quartet. The music for this piece is constructed based on the Markov chain algorithm.

Xenakis is famous as a composer who used various mathematical logics and probabilistic methods of composing with computers.
The rest of the piece was composed by algorithmic composition using IBM's 7090 computer.
Other instrumental pieces ST/10, 080262, ST/48-1, 240162, Atrees, and Morsima-Amorsima were written using probabilistic methods.

David Cope is an American musicologist/composer and the developer of Experiments in Musical Intelligence (EMI).
EMI searches its internal database for appropriate melodic fragments, and then connects them according to a certain musical grammar to generate music.
Furthermore, he has succeeded in analyzing his own past works and automatically generating music in his own style.
This has given rise to the value that a composer is no longer engaged in a creative act if he does not create a new style himself.