Unruly

Intelligences in Music AI

Unruly Intelligences in Music AI is an interdisciplinary research project at UCLA, bringing together musicology, law, statistics, and creative media to ask how artificial intelligence hears — and mishears — the world's sound and music.

  • 8members
  • 4disciplines
  • 3questions
  • 1pilot

AI systems for classifying and recommending music often promise a borderless, universal listening experience. Yet they are built on data and assumptions drawn overwhelmingly from Western, English-language, commercially dominant traditions. When these systems encounter music that does not fit — a raga with no key signature, a tradition with no genre tag, a form of authorship that is communal rather than individual — they tend to treat the difference as error, noise, or lack. We take that friction as our starting point. Rather than smoothing it away, we treat the "unruly" — the outlier, the residual, the uncategorizable — as the thing most worth studying.

Across workshops, prototypes, a shared glossary of contested terms, the project examines what is lost when music becomes data, how cultural particularity hardens into technical default, and what more just alternatives might look like. We are less interested in making AI's map of music more complete than in asking what any such map necessarily leaves out — and who is left out with it.

This site is a working space, not a finished article. It gathers our strands of research as they develop.

Unruly Questions

  1. What gets lost when music becomes data, and how can it be recovered?

    From the moment a music is recorded, classified, and ingested into a model, layers of cultural meaning and context are stripped away. We ask what is lost in that translation, at whose cost, and whether it can ever be ethically traced.

  2. How do classification and recommendation systems turn Western musical norms into global defaults?

    Recommendation systems standardize and subsequently propagate assumptions about (usually Western) music theory to train their models. We examine how cultural particularity comes to masquerade as universality within technical infrastructures, and how/where that becomes visible.

  3. What would it mean to build systems that treat musical "unruliness" as a norm, rather than something to be flattened out and tamed?

    Rather than asking only how current systems fall short, we ask what alternatives might look like: different datasets, different interfaces, different metrics. And, by extension, whether "more data" is necessarily the right answer.

Unruly Terms

About this glossary In transdisciplinary work, a lot of time is spent on translation. As navigational guides to our joint exploration, we identified eight core concepts worthy of definition and debate. This glossary compiles our starting gambit for unruly definitions that articulate what we hold in common and where we differentiate.

Music

There is no such thing as music. Or, rather: there is no such thing as music in the universal, self-evident sense the…

An unruly definition, by Thomas Hodgson and Emmie Head

There is no such thing as music. Or, rather: there is no such thing as music in the universal, self-evident sense the word implies. The idea that music is a singular human practice — a language we all speak — is itself a culturally specific belief, and a relatively recent Western one. Ethnomusicology has wrangled with this for over a century: indeed, one of the discipline's first moves was to pluralize the term, to insist on musics. Its more provocative assertion, though, was to ask whether the word itself makes any sense at all in parts of the world distant from the European concert tradition. Many of the world's sonic practices are not thought of as "music" by those who perform them, but instead something more deeply intertwined with ritual, labour, devotion, or speech.

And so, when we say music, we are not naming a universal. In a historical sense, we are describing a (mostly Western) tradition organized around a set of discrete pitches of equal temperament, harmony, the work-concept, the author (see also Authorship below), etc. The unruliness of "music", then, happens when these particularities begin to masquerade as generalities, especially within technical infrastructures. When a streaming platform categorizes a monsoon raga as "E minor," for example, or a dataset recognizes only key signature and tempo, Western music theory is being invoked less as a way to describe music than to define it, globally, in epistemologically domineering ways. Music, in this sense, becomes a universal marker not of what we hear, but of a longstanding process of infrastructural universalization.

Intermediary (also Platform)

Intermediaries are ubiquitous and mercurial. Deriving from inter- (between or among) and medius (in the middle), but…

An unruly definition, by Julia Powles

Intermediaries are ubiquitous and mercurial. Deriving from inter- (between or among) and medius (in the middle), but distinct from the declarative mediator (to intervene or reconcile between two sides), the intermediary inhabits ambiguity. Its digital origin story is in the junctures of the material — nodes, routers, switches, proxies, gateways. Here intermediaries are both situational and propulsive, industriously shuttling suitcases of digital information across notched networks; an enterprise prided as mercantile, not mercenary. From this inheritance comes the core identity crisis of the modern intermediary. It aspires to clerical neutrality to the point of invisibility, but it does so with endemic self-importance. In between, everything is politics.

As intermediaries scaffold to platforms, the stickiest and best-known become monstrous assemblages of information exchange, attention-hacking, vanity, and profiteering. In the policy world, intermediaries are flashpoints. They entice and frustrate, thwarting all but the most determined with their aspirational camouflage, magnificent inscrutability, colossal power and, just as readily, wafer-thin responsibility. "Nothing to see here," the machine throbs.

Classification

To classify music is in a sense to draw borders around sound — to decide what belongs with what, and, by implication,…

An unruly definition, by Thomas Hodgson

To classify music is in a sense to draw borders around sound — to decide what belongs with what, and, by implication, what does not belong at all. "Genre" is one of the most familiar of these epistemic borders, and one of the most contested. Musicology, and ethnomusicology in particular, have long been critical of the term: the names of genres are rarely neutral indicators of musical content, but rather carefully curated labels, motivated by corporate histories of marketing, taste, race, and geography. The World Music debate of the 1990s, in particular, showed how virtually an entire planet's worth of (non-Western) musical difference came to be lumped together under a single label — a label that revealed much more about the industry doing the sorting than the musics being sorted.

One of the big questions for our moment is whether streaming and AI dissolve these anxieties or magnify them. The corporate promise is one of dissolution: platforms speak of a "borderless" ecosystem, of infinite micro-genres, of recommendation unshackled from physical spaces (record stores, yes, but also local sites of music making). The reality, though, is that classification has not disappeared but rather become ever more automated, hidden within platform infrastructures. Genre tags are today assigned at scale — sometimes by experts (self-appointed or otherwise), increasingly by algorithms, and at least until recently, as one Spotify engineer admitted, simply invented in the name of algorithmic efficiency. Classification, then, collapses music's plural ontologies through technical infrastructures' requirement that everything be sortable.

Authorship

Authorship, sometimes correlated with "ownership," associates a creator or a group of creators with their creations,…

An unruly definition, by Emmie Head

Authorship, sometimes correlated with "ownership," associates a creator or a group of creators with their creations, a composer with their compositions. In the Western sense, authorship comprises an individual person (an "author") who originates or creates an intellectual or creative work. The humanities, including studies of Western music, romanticize the notion of the single author, establishing that the individual authorship of original ideas is accompanied by a level of genius or mastery over the subject matter or art form. This sense of authorship, while frequently referenced, is only a singular point in a web of diverse social and cultural understandings of authorship.

Authorship is culturally and socially dependent, not unlike music (see Music above). Beyond Western notions, authorship can be unknowable, unattributable, communal, collective, or unapplicable. Authorship does not always equate to ownership. Authors may choose to remain anonymous, develop pseudonyms, or even make their work freely accessible. Authors may create derivative works, transform pre-existing works, or parody. Provenance, or the lineage of custody of a product of someone's (or some people's) creativity is often associated with authorship should the author retain authority over their creative work or seek to assert it. Exchanges of intellectual property rights, works-for-hire, fair use, or Creative Commons may introduce novel complications to provenance. Yet, it is important to remember that even provenance can be falsified, debated, uncertain, or entirely unknown.

As music making artificial intelligence advances, authorship and provenance become increasingly complicated. Such Platforms (above) blur and obfuscate lineages of authors and creators by curating an anonymized front-end for users and a back-end black box within which answers to questions of authorship and provenance are housed but left invisible to users.

Dataset

Music, as a medium, is ephemeral, atmospheric, and embodied — all qualities that resist any neat translation into a…

An unruly definition, by Devon Baur

Music, as a medium, is ephemeral, atmospheric, and embodied — all qualities that resist any neat translation into a "dataset."

Broadly, a dataset is defined as a collection of materials used to train, test, or evaluate an AI model. In the case of music, this might include: recordings (such as MP3, WAV, video files); computational representations derived from those recordings (spectrograms); symbolic representations (MIDI files, scores); textual data (lyrics, reviews); descriptive metadata (genre, tempo, key, instrumentation); contextual metadata (release year, record label, language, country of origin); or listener behavior data (streams, skips, likes, etc). None of these categories capture music — they are merely attempts to find small footholds in an expansive atmosphere of flowing sonic mist.

Yet while these data points cannot wholly capture music, they frame what counts as music, especially in processes of classification and recommendation. In the dataset, presence is often viewed as power. Models are trained on a corpus of music and go on to prioritize and recommend those same types of music. Music datasets largely skew towards Western English-language songs, reflecting larger systemic inequities, including who has access to recording and digitization. In a pattern of algorithmic bias, this loop becomes recursive, as models trained on Western English-language music continue to elevate music marked as similar. The way forward is murky. A more diverse or intentionally counterhegemonic dataset could help decenter this privilege — but to say that more data and more extraction is the tidy solution would also miss the central heartbeat of music as a medium.

Instead, we need to also consider all that slips beyond the dataset. In performance studies, Diana Taylor distinguishes between the archive and the repertoire.1 The archive is made of the tangible artifacts or materials that can be housed by an institution, such as the text, scores, recordings, which is similar to the corpus of a dataset. By contrast, the repertoire is the corporeal knowledge or experience that transfers from body-to-body, which is resistant to the archive. In Taylor's model the archive is often curated and maintained by institutions of hegemonic power. And yet, the repertoire which dances around and between these artifacts can be counterhegemonic, passed below the eye (or ear) of the institution. From this perspective, we provoke questions about the dataset and its limits: What elements of music are beyond the reach of the dataset entirely? In what ways does eluding capture protect music from being reduced to data in dominating systems?

1 Taylor, Diana. The Archive and the Repertoire: Performing Cultural Memory in the Americas. Duke University Press, 2003.

Recommendation (also Recommender; see also Attention)

Recommendation is a process of deciding what should come next, what should be heard again, and what should hide from…

An unruly definition, by Sarah Min

Recommendation is a process of deciding what should come next, what should be heard again, and what should hide from a user's consciousness. In its most familiar human form, recommendation is intimate and situated. A friend might assert, "you have to listen to this." A DJ follows one song with a transition to another. A critic, elder, teacher, fan, or community member points toward something because it carries memory, taste, trust, mood, obligation, expertise, or care. The partiality of human recommendation is often visible. We know who is speaking, where from, and with what kind of authority.

Algorithmic recommendation borrows this social intimacy while reorganizing it through machine process. Rather than the simple human instinct to "like" music, it numerically calculates relations between users, tracks, skips, playlists, streams, tags, reviews, metadata, audio features, and behavioral patterns, turning listening patterns into data. Collaborative filtering gathers preferences from people similar to you. Natural language processing absorbs what has been written about artists and songs and turns that into binary form. Audio analysis extracts tempo, key, timbre, dynamics, and other measurable signals. Together, these systems produce the feeling of personal knowledge, to anticipate you, and, like a friend, suggest "you might like this too."

The unruliness of recommendation begins in this imitation of the human. Platforms often present recommendations as discovery, democratization, or connection: a way for any artist to find any listener across the globe. But the recommender is also an intermediary, and a market-making classifier. It does not merely reveal musical value, but by algorithmically consistently suggesting some songs over others, it helps produce that value. A song recommended at scale becomes more listenable because it has been made more visible. A song skipped too early, misclassified too crudely, or absent from the dataset may never enter the field of possibility at all. Recommendation therefore does not sit after music, as a neutral delivery mechanism. Although it presents as such, it does not hold the subjective consciousness of a human being.

This is why machine recommendation changes the human process as much as it automates it. Artists and labels learn to compose, release, title, tag, promote, and structure music in anticipation of the system. This may unravel in a hook earlier in the song or a shorter intro, possibly lyrics that become "playlist-friendly." Listeners, too, are folded into the loop: every skip, save, replay, or silence becomes a signal that trains the next suggestion.

To define recommendation unrulily is to refuse the fantasy that it is merely helpful. Recommendation is a social relation disguised as a technical one. It is where human trust, machine calculation, corporate scale, and musical creativity meet. It can connect but also isolate within niches. It can widen access, but it can also intensify the power of the already-heard and dilute exposure to less popular tracks. It can feel personal while operating through impersonal infrastructures like manufactured clicks and hooks to guide recommendations.

Recommendation names not only the act of pointing someone toward music, but the larger struggle over who gets to decide what music is listened to (see also Attention).

Attention (also Attention Conductor; see also Intermediary)

The attention of others is a potent thing. The more of it you command, the more influence you are in a position to…

An unruly definition, by Jenny Judge

The attention of others is a potent thing. The more of it you command, the more influence you are in a position to wield. But in our information-saturated world, the odds that you will actually attract the attention of anyone at all are miserably low, never mind the attention of the masses. In this context, a powerful figure emerges: the 'attention conductor' (see also Intermediary). An attention conductor is someone that has access to the attention of others, and is capable of directing that attention in whatever way they choose. The more people whose attention the conductor can access, and the more effectively they can direct that attention, the more the conductor stands to gain. Many of the wealthiest people in the world today are attention conductors: people (mostly men) who have developed effective and ubiquitous technologies aimed at harnessing and directing the attention of hundreds of millions: Meta, Google, Amazon, the media barons, and so on.

Attention is, or is what underlies, the selective directedness of our mental lives. It's that mental capacity that we have to focus on one thing, or one set of things, to the exclusion of others. Any act of attention can vary along a number of dimensions. It can be more or less focused, ranging from intense scrutiny to something more like dim awareness. It can be effortful, or effortless. It can be voluntarily deployed, or involuntarily harnessed by outside forces (and of course, the attention conductors know all about that). Attention has a deep connection to consciousness: what we attend to just is what we're aware of at all.

Attention is our most powerful resource. It's how we gain knowledge. It's how we develop virtues and skills. Mutual attention, and joint attention to a common object, are the building blocks with which we construct a shared world. But it's also our deepest vulnerability. Because when it's directed at the wrong things, or the right things in the wrong way, our attention is the means by which we can be misled, disempowered, and morally compromised.

Music

An unruly definition, by Thomas Hodgson and Emmie Head

There is no such thing as music. Or, rather: there is no such thing as music in the universal, self-evident sense the word implies. The idea that music is a singular human practice — a language we all speak — is itself a culturally specific belief, and a relatively recent Western one. Ethnomusicology has wrangled with this for over a century: indeed, one of the discipline's first moves was to pluralize the term, to insist on musics. Its more provocative assertion, though, was to ask whether the word itself makes any sense at all in parts of the world distant from the European concert tradition. Many of the world's sonic practices are not thought of as "music" by those who perform them, but instead something more deeply intertwined with ritual, labour, devotion, or speech.

And so, when we say music, we are not naming a universal. In a historical sense, we are describing a (mostly Western) tradition organized around a set of discrete pitches of equal temperament, harmony, the work-concept, the author (see also Authorship below), etc. The unruliness of "music", then, happens when these particularities begin to masquerade as generalities, especially within technical infrastructures. When a streaming platform categorizes a monsoon raga as "E minor," for example, or a dataset recognizes only key signature and tempo, Western music theory is being invoked less as a way to describe music than to define it, globally, in epistemologically domineering ways. Music, in this sense, becomes a universal marker not of what we hear, but of a longstanding process of infrastructural universalization.

Intermediary (also Platform)

An unruly definition, by Julia Powles

Intermediaries are ubiquitous and mercurial. Deriving from inter- (between or among) and medius (in the middle), but distinct from the declarative mediator (to intervene or reconcile between two sides), the intermediary inhabits ambiguity. Its digital origin story is in the junctures of the material — nodes, routers, switches, proxies, gateways. Here intermediaries are both situational and propulsive, industriously shuttling suitcases of digital information across notched networks; an enterprise prided as mercantile, not mercenary. From this inheritance comes the core identity crisis of the modern intermediary. It aspires to clerical neutrality to the point of invisibility, but it does so with endemic self-importance. In between, everything is politics.

As intermediaries scaffold to platforms, the stickiest and best-known become monstrous assemblages of information exchange, attention-hacking, vanity, and profiteering. In the policy world, intermediaries are flashpoints. They entice and frustrate, thwarting all but the most determined with their aspirational camouflage, magnificent inscrutability, colossal power and, just as readily, wafer-thin responsibility. "Nothing to see here," the machine throbs.

Classification

An unruly definition, by Thomas Hodgson

To classify music is in a sense to draw borders around sound — to decide what belongs with what, and, by implication, what does not belong at all. "Genre" is one of the most familiar of these epistemic borders, and one of the most contested. Musicology, and ethnomusicology in particular, have long been critical of the term: the names of genres are rarely neutral indicators of musical content, but rather carefully curated labels, motivated by corporate histories of marketing, taste, race, and geography. The World Music debate of the 1990s, in particular, showed how virtually an entire planet's worth of (non-Western) musical difference came to be lumped together under a single label — a label that revealed much more about the industry doing the sorting than the musics being sorted.

One of the big questions for our moment is whether streaming and AI dissolve these anxieties or magnify them. The corporate promise is one of dissolution: platforms speak of a "borderless" ecosystem, of infinite micro-genres, of recommendation unshackled from physical spaces (record stores, yes, but also local sites of music making). The reality, though, is that classification has not disappeared but rather become ever more automated, hidden within platform infrastructures. Genre tags are today assigned at scale — sometimes by experts (self-appointed or otherwise), increasingly by algorithms, and at least until recently, as one Spotify engineer admitted, simply invented in the name of algorithmic efficiency. Classification, then, collapses music's plural ontologies through technical infrastructures' requirement that everything be sortable.

Authorship

An unruly definition, by Emmie Head

Authorship, sometimes correlated with "ownership," associates a creator or a group of creators with their creations, a composer with their compositions. In the Western sense, authorship comprises an individual person (an "author") who originates or creates an intellectual or creative work. The humanities, including studies of Western music, romanticize the notion of the single author, establishing that the individual authorship of original ideas is accompanied by a level of genius or mastery over the subject matter or art form. This sense of authorship, while frequently referenced, is only a singular point in a web of diverse social and cultural understandings of authorship.

Authorship is culturally and socially dependent, not unlike music (see Music above). Beyond Western notions, authorship can be unknowable, unattributable, communal, collective, or unapplicable. Authorship does not always equate to ownership. Authors may choose to remain anonymous, develop pseudonyms, or even make their work freely accessible. Authors may create derivative works, transform pre-existing works, or parody. Provenance, or the lineage of custody of a product of someone's (or some people's) creativity is often associated with authorship should the author retain authority over their creative work or seek to assert it. Exchanges of intellectual property rights, works-for-hire, fair use, or Creative Commons may introduce novel complications to provenance. Yet, it is important to remember that even provenance can be falsified, debated, uncertain, or entirely unknown.

As music making artificial intelligence advances, authorship and provenance become increasingly complicated. Such Platforms (above) blur and obfuscate lineages of authors and creators by curating an anonymized front-end for users and a back-end black box within which answers to questions of authorship and provenance are housed but left invisible to users.

Dataset

An unruly definition, by Devon Baur

Music, as a medium, is ephemeral, atmospheric, and embodied — all qualities that resist any neat translation into a "dataset."

Broadly, a dataset is defined as a collection of materials used to train, test, or evaluate an AI model. In the case of music, this might include: recordings (such as MP3, WAV, video files); computational representations derived from those recordings (spectrograms); symbolic representations (MIDI files, scores); textual data (lyrics, reviews); descriptive metadata (genre, tempo, key, instrumentation); contextual metadata (release year, record label, language, country of origin); or listener behavior data (streams, skips, likes, etc). None of these categories capture music — they are merely attempts to find small footholds in an expansive atmosphere of flowing sonic mist.

Yet while these data points cannot wholly capture music, they frame what counts as music, especially in processes of classification and recommendation. In the dataset, presence is often viewed as power. Models are trained on a corpus of music and go on to prioritize and recommend those same types of music. Music datasets largely skew towards Western English-language songs, reflecting larger systemic inequities, including who has access to recording and digitization. In a pattern of algorithmic bias, this loop becomes recursive, as models trained on Western English-language music continue to elevate music marked as similar. The way forward is murky. A more diverse or intentionally counterhegemonic dataset could help decenter this privilege — but to say that more data and more extraction is the tidy solution would also miss the central heartbeat of music as a medium.

Instead, we need to also consider all that slips beyond the dataset. In performance studies, Diana Taylor distinguishes between the archive and the repertoire.1 The archive is made of the tangible artifacts or materials that can be housed by an institution, such as the text, scores, recordings, which is similar to the corpus of a dataset. By contrast, the repertoire is the corporeal knowledge or experience that transfers from body-to-body, which is resistant to the archive. In Taylor's model the archive is often curated and maintained by institutions of hegemonic power. And yet, the repertoire which dances around and between these artifacts can be counterhegemonic, passed below the eye (or ear) of the institution. From this perspective, we provoke questions about the dataset and its limits: What elements of music are beyond the reach of the dataset entirely? In what ways does eluding capture protect music from being reduced to data in dominating systems?

1 Taylor, Diana. The Archive and the Repertoire: Performing Cultural Memory in the Americas. Duke University Press, 2003.

Recommendation (also Recommender; see also Attention)

An unruly definition, by Sarah Min

Recommendation is a process of deciding what should come next, what should be heard again, and what should hide from a user's consciousness. In its most familiar human form, recommendation is intimate and situated. A friend might assert, "you have to listen to this." A DJ follows one song with a transition to another. A critic, elder, teacher, fan, or community member points toward something because it carries memory, taste, trust, mood, obligation, expertise, or care. The partiality of human recommendation is often visible. We know who is speaking, where from, and with what kind of authority.

Algorithmic recommendation borrows this social intimacy while reorganizing it through machine process. Rather than the simple human instinct to "like" music, it numerically calculates relations between users, tracks, skips, playlists, streams, tags, reviews, metadata, audio features, and behavioral patterns, turning listening patterns into data. Collaborative filtering gathers preferences from people similar to you. Natural language processing absorbs what has been written about artists and songs and turns that into binary form. Audio analysis extracts tempo, key, timbre, dynamics, and other measurable signals. Together, these systems produce the feeling of personal knowledge, to anticipate you, and, like a friend, suggest "you might like this too."

The unruliness of recommendation begins in this imitation of the human. Platforms often present recommendations as discovery, democratization, or connection: a way for any artist to find any listener across the globe. But the recommender is also an intermediary, and a market-making classifier. It does not merely reveal musical value, but by algorithmically consistently suggesting some songs over others, it helps produce that value. A song recommended at scale becomes more listenable because it has been made more visible. A song skipped too early, misclassified too crudely, or absent from the dataset may never enter the field of possibility at all. Recommendation therefore does not sit after music, as a neutral delivery mechanism. Although it presents as such, it does not hold the subjective consciousness of a human being.

This is why machine recommendation changes the human process as much as it automates it. Artists and labels learn to compose, release, title, tag, promote, and structure music in anticipation of the system. This may unravel in a hook earlier in the song or a shorter intro, possibly lyrics that become "playlist-friendly." Listeners, too, are folded into the loop: every skip, save, replay, or silence becomes a signal that trains the next suggestion.

To define recommendation unrulily is to refuse the fantasy that it is merely helpful. Recommendation is a social relation disguised as a technical one. It is where human trust, machine calculation, corporate scale, and musical creativity meet. It can connect but also isolate within niches. It can widen access, but it can also intensify the power of the already-heard and dilute exposure to less popular tracks. It can feel personal while operating through impersonal infrastructures like manufactured clicks and hooks to guide recommendations.

Recommendation names not only the act of pointing someone toward music, but the larger struggle over who gets to decide what music is listened to (see also Attention).

Attention (also Attention Conductor; see also Intermediary)

An unruly definition, by Jenny Judge

The attention of others is a potent thing. The more of it you command, the more influence you are in a position to wield. But in our information-saturated world, the odds that you will actually attract the attention of anyone at all are miserably low, never mind the attention of the masses. In this context, a powerful figure emerges: the 'attention conductor' (see also Intermediary). An attention conductor is someone that has access to the attention of others, and is capable of directing that attention in whatever way they choose. The more people whose attention the conductor can access, and the more effectively they can direct that attention, the more the conductor stands to gain. Many of the wealthiest people in the world today are attention conductors: people (mostly men) who have developed effective and ubiquitous technologies aimed at harnessing and directing the attention of hundreds of millions: Meta, Google, Amazon, the media barons, and so on.

Attention is, or is what underlies, the selective directedness of our mental lives. It's that mental capacity that we have to focus on one thing, or one set of things, to the exclusion of others. Any act of attention can vary along a number of dimensions. It can be more or less focused, ranging from intense scrutiny to something more like dim awareness. It can be effortful, or effortless. It can be voluntarily deployed, or involuntarily harnessed by outside forces (and of course, the attention conductors know all about that). Attention has a deep connection to consciousness: what we attend to just is what we're aware of at all.

Attention is our most powerful resource. It's how we gain knowledge. It's how we develop virtues and skills. Mutual attention, and joint attention to a common object, are the building blocks with which we construct a shared world. But it's also our deepest vulnerability. Because when it's directed at the wrong things, or the right things in the wrong way, our attention is the means by which we can be misled, disempowered, and morally compromised.

To come — Algorithmic Systems

Datasets · Interfaces · Wisdom

Three strands branch from one unruly center. Pick a strand to read it here.

Unruly Datasets The datasets and audio models this project examines, and how each one logs (or doesn't log) musical difference. Unruly Interfaces How do we see music data? Unruly Wisdom The shared reading list the project is built on.

Unruly Datasets

The datasets and audio models this project examines, and how each one logs (or doesn't log) musical difference.

Dataset audit table — coming soon.

Unruly Interface

How do we see music data? The interface through which we encounter a dataset quietly shapes how we think with and through that data. It frames what appears central, legible, and prioritized as clickable, sortable, searchable, visible, or absent. As visual theorist Johanna Drucker reminds us in Graphesis, "data does not have an inherent visual form" and so all visualizations are expressive interpretations.

To observe this process is to notice the dramaturgy of the interface. Graphic representation is a theatrical problem, much the same as the ways that blocking, set design, choreography, and point of view shape the meaning of what appears before us on stage. In this sense, an interface is not a neutral tool for accessing information, but a creative structure that arranges the data and directs our attention in a way that teaches us how to look. If we stage data differently, particularly in ways that attend to its unruliness, we might also come to think about its possibilities and limitations differently.

This type of question is central to (and reliant on) interdisciplinary collaboration. As Drucker continues, "The majority of information graphics, for instance, are shaped by the disciplines from which they have sprung: statistics, empirical sciences, and business. Can these graphic languages serve humanistic fields where interpretation, ambiguity, inference, and qualitative judgment take priority over quantitative statements and presentations of 'facts'?"

We are interested in exploring ways that interfaces can become unruly sites of inquiry, reframing ways to interpret what is present and absent in our datasets.

In this section, we build an alternative interface for viewing the Million Song Dataset, a large-scale collection of metadata and audio features for popular music. At first glance, the dataset's scale suggests abundance with one million songs, one million entries, one million possible pathways through musical culture. Yet such a scale can also conceal imbalance. When the dataset is mapped geographically, its uneven distribution becomes more immediately legible. The globe reveals dense clusters of data across North America and Europe, while other regions appear sparsely represented or nearly blank. In this sense, the interface stages not only the information that is present, but also the information that is glaringly absent.

This act of mapping reveals the expressive potential of the interface but it does not aim to be a perfect solution. The globe introduces its own problems. It was difficult to decide which version of the globe we should use; we used a current global map API, but with equal justification, we could have used a map of the world in 1966 with different geopolitical mappings, or a speculative map of 2066 where rising sea levels and climate change threaten the very geographies we are trying to represent. Each choice produces a different understanding of the dataset and its relationship to history, territory, and power.

The question of origin is equally unstable. The Million Song Dataset tags songs with a latitude and a longitude. How can we pin a fixed geography to a song? It is unclear if this geotag is an attempt to represent location determined by the locale of the artist, the site of recording, the record label, the language of the lyrics, the market where it circulated, or the communities that gave it meaning. A tidy geographical interface may reveal aggregate bias, but it also fails to represent migration, diaspora, remix, sampling, and international collaboration. To pin a song to a single point on a globe is already to make an argument.†

Our alternative interface always highlights the disorganized components of the data. Where and how do we map data that is incomplete? Songs with missing geographic tags are not simply neutral unknowns. They raise questions about what kinds of music become richly described and what kinds of music remain under-documented. Does the lack of a geographic tag privilege or penalize the song in processes of categorization and recommendation? Illegibility in these systems remains political.

By staging the Million Song Dataset on a globe, we do not claim to show the dataset as it truly is. Instead, we use the interface as a critical instrument. The map makes imbalance visible, but it also exposes the impossibility of any single, stable view. An unruly interface does not tidy the data into certainty. It asks us to see data (which, again, has no inherent visual form) as a staged entity that is always partial, situated, historical, and incomplete.

Interactive globe — coming soon The Million Song Dataset staged on a globe. The visualization is not yet part of this delivery; the text above describes it.

References

  1. Johanna Drucker, Graphesis: Visual Forms of Knowledge Production.
  2. journals.sagepub.com/doi/full/10.1177/2057047319850192
  3. upress.umn.edu/9780972969628/elsewhere-mapping

Unruly Wisdom

reading list A reading list is a kind of argument. These are the works the project is thinking with — across musicology, law, statistics, and media — and a note on why each one earned its place. The list is maintained collectively in Zotero and updated as the project grows.

filter by topic

25 works

Bonadio, Enrico, and Chen Wei Zhu. Music Borrowing and Copyright Law: A Genre-by-Genre Analysis. Bloomsbury Publishing, 2023.

why it's here

  1. Edited collection of short essays (approx. 20 pages) which parse out the challenging relationships between music law and different genres.
  2. Includes multiple sections, navigating different music in American but also  across the world. Sections include: Americas, Europe, Africa and Middle East, and Asia and Oceania.
  3. Gives the largest breadth of literature of music and its relationship to copyright law in different parts of the world.
Burkart, Patrick, and Tom McCourt. Digital Music Wars: Ownership and Control of the Celestial Jukebox. 2006. Reprint, Rowan & Littlefield, 2006.
Burkart, Patrick, and Tom McCourt. Digital Music Wars: Ownership and Control of the Celestial Jukebox. Bloomsbury Publishing PLC, 2006.

why it's here

  1. Evaluates transition from analog to digital modes of listening.
  2. Traces current-to-the-early-2000s developments in law, technology, and corporate to evaluate how the digital music industry will work with distributiuon, access, and consumer privacy.
Cummings, Alex Sayf. Democracy of Sound: Music Piracy and the Remaking of American Copyright in the Twentieth Century. Oxford University Press, 2013.

why it's here

  1. Detailed discussion of Napster’s legal situation and role of piracy of copyrighted music in the advent of digital streaming
Drott, Eric. “Copyright, Compensation, and Commons in the Music AI Industry.” Creative Industries Journal 14, no. 2 (August 4, 2021): 190–207. https://doi.org/10.1080/17510694.2020.1839702.

why it's here

  1. Survey of the relationship between copyright and compensation in the ecosystem of music streaming, focus on introduction of AI-based systems onto those platforms.
  2. Brings into discussion the role of the “commons” as a potential soulution to distributive justice in the face of commercial misoc and AI’s integration into user platform.
Feld, Steven. “A Sweet Lullaby for World Music.” Public Culture 12, no. 1 (January 1, 2000): 145–71. https://doi.org/10.1215/08992363-12-1-145.

why it's here

  1. response to world music debate - questions whether or not the use of Afunakwa’s voice in Deep Forest should be justified/what rights and affordances singers who are the subject of ethnographic recordings should have over their works.
Fishman, Joseph. “Music as a Matter of Law.” Harvard Law Review 131, no. 7 (May 1, 2018): 1861. https://scholarship.law.vanderbilt.edu/faculty-publications/578.

why it's here

  1. Seminal piece of legal writing, questioning legal ontologies of music and how “Blurred Lines” decision (2018) brought into question whether or not the law should have as much oversight as it does on music making and regulating copyrights
  2. Surveys historical approach to legal framing of music with particular attention to the United States.
Hesmondhalgh, David. Music Streaming Around the World. Univ of California Press, 2025.

why it's here

  1. Surveys role streaming platforms play in 12 different global music economies.
  2. Considers implications on contemporary cultures of music making.
Hodgson, Thomas. “Quantifying Music: Imagined Metrics in Digital Startup Culture.” Culture, Theory and Critique 61, no. 4 (October 1, 2020): 424–39. https://doi.org/10.1080/14735784.2021.1894961.
Hodgson, Thomas. “Spotify and the Democratisation of Music.” Popular Music 40, no. 1 (February 2021): 1–17. https://doi.org/10.1017/S0261143021000064.
Jaakola, Lyz, and Timothy B. Powell. “‘The Songs Are Alive’: Bringing Frances Densmore’s Recordings Back Home to Ojibwe Country.” In The Oxford Handbook of Musical Repatriation, edited by Frank Gunderson, Robert C. Lancefield, and Bret Woods, 575–90. Oxford University Press, 2019. https://doi.org/10.1093/oxfordhb/9780190659806.013.32.

why it's here

  1. Recounts the digital repatriation efforts of Frances Densmore’s Anishinaabe and Ojibwe song collections.
  2. Poses that we need to allow Indigenous peoples in the United States, in particular, the chance to reclaim ownership over their music and their voices, despite legal challenges they might face as ethnographic recordings often attributed to individuals who take the recordings.
  3. Discuss challeges of repatriating digitized recordings.
Loughridge, Deirdre. Sounding Human: Music and Machines, 1740/2020. New Material Histories of Music. Chicago, IL: University of Chicago Press, 2024. https://press.uchicago.edu/ucp/books/book/chicago/S/bo208042715.html.

why it's here

  1. Draws on posthumanist perspective to analyze relationships between music making humans and assistive machines.
  2. Provides a concise history of different technologically-informed musical products, explaining how music and technology together can explain and challenge what it means to be human.
Mann, Larisa Kingston. Rude Citizenship: Jamaican Popular Music, Copyright, and the Reverberations of Colonial Power. UNC Press Books, 2022.

why it's here

  1. Identifies how cultures of collaboration are central to Jamaican creative practices.
  2. With ethnographic research, Kingston demonstrates how, as a result, music producers, musicians and audiences are resistant to the claims of copyright law and as an act of social/political commentary.
Michaud, Alyssa. Automatic Artistry: Negotiating Musical Creativity in a Technological Age. University of California Press, 2026.

why it's here

  1. Traces legacy of technological innovation and impact on music making

  2. Emphasis on interrelationship between performers and the technologies they use to make music.

Morrison, Matthew D. Blacksound: Making Race and Popular Music in the United States. Univ of California Press, 2024.

why it's here

  1. Navigates history of blackface minstrelsy and racialized foundations of American music.
  2. Unpacks relationship between performance, race, and intellectual property, arguing that blackface performance materials would be integrated into commercial entertainment through the privileges, affordances, and restrictions of copyright.
Parks, Gregory S., and Frank Rudy Cooper, eds. Fight the Power: Law and Policy through Hip-Hop Songs. Cambridge: Cambridge University Press, 2022. https://doi.org/10.1017/9781009019804.

why it's here

  1. Collection of legal commentaries, considering how current policy developments exist within the context of hip-hop performance and the fight for Black equality.
  2. Emphasis on the development of criminal proceedings, using rap lyrics as evidence of admission of crimes.
  3. Book’s chapters consider other roles of law in hip-hop such as: gendered relations, stylistic specificites, and the significant role of sampling in music of Black American musical practice.
Perullo, Alex. “Digital Repatriation: Copyright Policies, Fair Use, and Ethics.” In The Oxford Handbook of Musical Repatriation, edited by Frank Gunderson, Robert C. Lancefield, and Bret Woods, 0. Oxford University Press, 2019. https://doi.org/10.1093/oxfordhb/9780190659806.013.30.

why it's here

  1. Navigates the challenges of legal policy and the digitization of music and attempts at repatriation.
  2. Argues we should still create digital collections for certain audiences (often composed of individuals whose music is being repatriated) as a way to “pay back” the damages of acquiring that music in the first place.
Peterson, Richard A. Creating Country Music: Fabricating Authenticity. University of Chicago Press, 2013.

why it's here

  1. Traces origination of country music in the United States
  2. Explores constructions of archetypes within the genre, arguing that authenticity is manufactured and created through the negotation of space between new and old materials.
Reece, Frederick. Forgery in Musical Composition: Aesthetics, History, and the Canon. Oxford, New York: Oxford University Press, 2025.

why it's here

  1. Explores motivation behind the creation of “fake” classical music, written by other but attributed to compsores like Haydn, Mozart, and Schubert.
  2. Considers motivations of writers/compsoers to falsely attribute works, questioning the value of authorship and integrity of the single-author concept.
Reed, Trevor. “Fair Use As Cultural Appropriation,” April 4, 2019. https://doi.org/10.2139/ssrn.3456164.

why it's here

  1. Argues that fair use aces as a gatekeeping mechanism for unauthorized uses of copyrighted culture. This mechanism empowers courts to sanction or disapprove of cultural appropriations to futher copyright’s goal of promoting creativity.
  2. Claims the “forgotten” factor of fair use is how unauthorized appropriations impact culturally diverse forms of creativity.
Savage, Steve. Bytes and Backbeats: Repurposing Music in the Digital Age. University of Michigan Press, 2013. https://doi.org/https://doi.org/10.3998/mpub.3432847.

why it's here

  1. Lays out the relationship betwen digital technology and music creation, production, and consumption.
  2. Challenges the relationship between artist and engineer, putting into question traditional musicological notions of authenticity and authorship.
Seeger, Anthony. “Ethnomusicology and Music Law.” Ethnomusicology 36, no. 3 (1992): 345–59. https://doi.org/10.2307/851868.

why it's here

  1. Discusses the relationship between the United States’ intellectual property and entertainment law and those who do not ascribe to the United States’ specific definitions of originality and ownership.
Stanfill, Mel. Rock This Way. University of Michigan Press, 2023. https://press.umich.edu/Books/R/Rock-This-Way2.

why it's here

  1. Analyzes how American culture defines “legitimate” musuc.
  2. Considers what role covers, remixes, and mashups play in this definition.
  3. Argues that race, the law, and creativity shape perceptions of borrowing arguing for their role in white supremacy and questioing what makes something original in the first place.
Zemp, Hugo. “The/An Ethnomusicologist and the Record Business.” Yearbook for Traditional Music 28 (1996): 36–56. https://doi.org/10.2307/767806.

why it's here

  1. first hand notes of ethnomusicologists rights negotiations.

  2. pertaining world music debate and Afunakwa’s rights over recorded voice in Smithsonian collection

Unruly Team

Unruly Hopes

placeholder

Next steps, background, where this goes from here.