The Measured Listener | Los Angeles Review of Books

This is the 15th installment in the Legacies of Eugenics series, which features essays by leading thinkers devoted to exploring the history of eugenics and the ways it shapes our present. You can read the first part here. The series is organized by Osagie K. Obasogie in collaboration with the Los Angeles Review of Books, and supported by the Center for Genetics and Society, the Othering & Belonging Institute, and Berkeley Public Health.

¤

A CENTRAL PREMISE of High Fidelity—Nick Hornby’s 1995 novel, which was turned into a film in 2000 and then a Zoë Kravitz–starring TV series in 2020—is that its main character, Rob Fleming, owner of a failing record store, cannot understand himself without first measuring and ranking his experiences. “You are what you like,” as the line goes. For this protagonist, the act of ranking every aspect of his life is both a coping mechanism and a pathology. Rob Fleming can only discuss his feelings and emotions in list form—“they have opinions and I have lists,” he says.

In the last several years, models of algorithmic recommendation have one-upped Rob’s lists and human curators in general. In part, they have done so by hitting the sweet spot between surprise and not too much of it—by going beyond the titular list. In my case, they have learned what kind of listener I am. Beyond knowing my favorite artists and genres (Pavement, Courtney Barnett, Lucy Dacus, and other artists falling under the broad umbrella of “indie rock”), they know which music makes me feel most nostalgic (Elliott Smith, Aimee Mann, Wilco, and other “indie folk” artists) and which songs I listen to for myself (these days, Phoebe Bridgers), for my teaching at a school of music (everything from Chopin’s Mazurkas to J Dilla to Katy Perry), and for my kids (the KPop Demon Hunters and Percy Jackson soundtracks). They know I prefer certain kinds of music only at certain hours, know that I prefer my favorites during the day but am more likely to listen to news podcasts in the morning. Their recommendations invariably feel intimate, even inevitable.

They are neither. The algorithms are built upon decades of research that asked not what music is, nor what it means to listen to music, but how a human might be measured based on their musical preferences. Like IQ tests, this research has a dark history. We’ve inherited methodologies born of a desire to associate musical preference with inherited “fitness” in a eugenicist sense. The original goal was diagnostic: to use what people listen to as a window into their selves, including their personality disorders and any mental illnesses they might have. Algorithmically mediated music has absorbed these methodologies.

¤

In a January 2024 interview, Spotify co-founder Daniel Ek announced what music platforms would do with the multitude of tunes now at people’s fingertips: “[W]e’re going to become that trusted friend where we’re going to introduce you to things that you probably thought, ‘No way in hell am I going to be interested in this,’ and you’re going to be totally open to it.” He went on to claim that the company would “do a better job” choosing your music than you possibly could. It would transcend any list you could make yourself. “[E]ven if you spent a whole working day trying to figure out what you wanted to listen to,” he bragged, “we will be able to create a playlist that is so much better than any of that.” Spotify alone has more than 700 million users, and by the company’s own accounting, a third of all new artist discoveries happen through algorithmic recommendations rather than active search—a measure, in every sense, of how thoroughly the listening experience has been quantified.

The algorithms are based on certain assumptions about music and the listener. The first is that music can be understood as discrete streams of information—made up of tempo, key, and lyrics—which become identifying parts of a larger musical style or genre. This approach descends from a long history of stylometrics (or stylistics)—the use of quantitative methods to ascribe authorship and style markers. Pandora’s Music Genome Project, for example, looks at 450 musical features, which the company regularly refers to with terms such as “genes” or “DNA.” This project became pivotal to how, in the 2010s, streaming platforms grouped artists and songs into genres and microgenres, a prerequisite for curation. The website Every Noise at Once, created by former Spotify engineer Glenn McDonald, shows how this data can then be used to cluster genres into relationships that are at once meaningful and surprising—it makes sense that K-pop would show up close to Japanese teen pop, Swedish idol pop, and Chinese electropop, but what is it about Polish reggae that leads to its proximity to these other genres?

Assumptions that focus on the listener are more difficult to pin down. Making a recommendation based on what type of listener you are requires a calculation of what it means for you to listen in a certain way and then basing predictions on your perceived listening behavior. But what does it mean to measure you as a listener? Or to measure musical taste? Spotify answers these questions by classifying listeners into one of 16 “listening personalities,” including “Early Adopters,” who rate highly on exploration and variety; “Musicologists,” who value “timelessness”; and “Specialists,” who value “uniqueness.”

A paradox is at play here: streaming services promise their users agency—namely, the ability to be curious, exploratory, and adventurous. Under the hood, though, they assume something else entirely—that, just as High Fidelity’s Rob Fleming would argue, you are what you like, and certain types of listeners will prefer certain types of music. The opposite is also thought to be true: those who prefer certain types of music are a certain type of listener. This paradox—perhaps unsurprisingly—is present throughout the history of science and technology. As Jessica Riskin argues in her 2016 book The Restless Clock: A History of the Centuries-Long Argument over What Makes Living Things Tick, technology has always reflected preconceptions of what it means to be alive, blurring the distinction between a passive mechanism (a living being modeled as clockwork) and an active, restless being with its own agency. Similarly, algorithmic recommendation systems in music are built upon a long history of models that treat musical preference as indicators—symptoms, in effect—of listeners’ underlying traits. Listeners, in these systems, are not restless agents capable of free will.

¤

The notion of measuring what it means to be a “good listener” originated in the early 20th century, as part of Carl Seashore’s “measures of musical talents,” first released in 1919 by the Columbia Graphophone Company. Born in Sweden and raised in Iowa, Seashore received Yale’s first PhD in the then-nascent discipline of psychology in 1895. He then returned to his home state to teach at the University of Iowa. His standardized test was designed to uncover innate musical talent by asking participants whether two musical ideas—for example, tones, pitches, rhythms, or timbres—were the same or different. In the case of tones, these would get closer together as the test went on, and they would even begin to warble. In another part of the test, participants were asked if the durations of the tones, including two artificially generated tones, were also the same or different. In the “tonal memory” part, students would be asked the same question of melodies of increasing length.

In her 2021 book “Destined to Fail”: Carl Seashore’s World of Eugenics, Psychology, Education, and Music, Julia Koza highlights Seashore’s avid and highly successful commercialization of these tests, which sold for a profit of about $7.50 each (roughly $122 in 2026 dollars). Advertised to both music educators and eugenicists—in, for instance, the Music Supervisors’ Journal and Eugenical News—they won over both constituencies. This was no accident: diagnosing inherited musical talent was, Seashore believed, a window into someone’s overall “fitness.” Starting in the 1920s, Seashore corresponded on the subject with Charles Davenport, a central figure in the American eugenicist movement, with the latter declaring Seashore’s work “of the greatest possible importance.”

During World War II, Seashore’s test was used to select submariners, and in 1947, Seashore, by this point a dean, proposed a course on “the nature of euthenics,” which he argued was “the logical sequel to Eugenics which now has scientific and practical status as a normative science.” Whereas eugenics was “the science of being born well,” euthenics, he argued, was “the science and art of living well.” His tests, which incidentally emerged in tandem with the phonograph, normalized the idea of being a “good listener,” and they assumed the trait’s heritability and measurability. The tests became the standard in elementary music education until the 1970s. If students were interested in participating in middle school orchestras, then they would often have to take the Seashore test first. Those who passed would be allowed to play, sometimes receiving free instruments and lessons. Those who failed were summarily classified as unmusical. They essentially had bad musical genes, and there wasn’t much that could be done about that.

¤

While Seashore sought to define what it means to be musical, and to listen well, others were seeking to correlate listening with personality, mental health, and overall fitness by capitalizing on new statistical techniques such as factor analysis. A notoriously difficult activity to measure, music is a complex and multidimensional experience. Musical recommendation systems use statistical techniques to reduce these dimensions into fewer underlying traits. For example, songs with an upbeat tempo, a steady beat, and a strong backbeat all contribute to what Spotify calls “danceability.” Songs that cluster together can be reduced into broader genre factors—“escape room,” “easycore,” and “K-pop reality show” are notable genres found on Every Noise at Once.

This technique of reduction—factor analysis—was invented by Charles Spearman, but it was his student Raymond Cattell who, over a career spanning seven decades, would deploy it as a method for measuring human nature. Perhaps second only to Seashore, Cattell was the most consequential figure in creating a listener who could be captured by a fixed set of measurable traits. Cattell used factor analysis to argue that musical preference could reveal not only the underlying personality of listeners but also their fitness for contributing to society (“musical creativity” as an important factor for social good). Born in 1905, Cattell had earned a degree in chemistry and physics before turning to psychology for his graduate work. He published around 56 books and more than 500 articles, and trained dozens of graduate students, postdoctoral fellows, and visiting researchers.

In 1997, the American Psychological Foundation selected Cattell to receive the Gold Medal Award for Lifetime Achievement. Several scientists raised objections, however, based on his long history of eugenicist and racist beliefs. In an open letter to the American Psychological Association, Cattell insisted that the accusations were false—he was never a “racial supremacist” and the use of eugenics should always be “voluntary.” His more recently published book, Beyondism: Religion from Science (1987; reworked from parts of his 1972 book A New Morality from Science: Beyondism), was, he insisted, a more accurate depiction of his views. He withdrew his name from consideration for the award, died shortly thereafter at the age of 92, and the debate went quiet.

Cattell’s early work, however, was not simply the source of an occasional quote taken out of context; it consisted of sustained Nazi talking points buttressed by the heft of statistics. For example, his 1931 book Psychology and Social Progress: Mankind and Destiny from the Standpoint of a Scientist (note the intended gravitas of “the standpoint of a scientist”) includes a chapter titled “Nation and Race: Their Significance for Human Progress,” in which he declares that “human improvement” over time has been the result of group competition. In less than a paragraph, he argues not only for racialized warfare but also for the extraction of mineral resources from colonized peoples, writing:

The members of one competing group must be of a different inborn racial type from the other, else there is no selective action resulting in the greater prevalence of a fitter type. […]

Success in the struggle must be determined by these inborn abilities, not by biologically irrelevant, acquired factors such as the possession of greater natural resources or the chance legacy of greater knowledge and technical skill handed on from some previous civilisation.

This assertion is then followed by a pullout map labeled “Distribution of Races of Europe,” which is a simplified version of the map in Hans Günther’s publication The Racial Elements of European History (1927). Cattell argued that “innate racial differences in emotional endowment and mental power are already recognisable and clearly account for many historical developments,” and he provided his own ranking of various races. The “Mediterranean” race, on the other hand, was “highly gregarious, lacking in self-assertion,” with a tendency for “crimes of violence and passion.”

According to Cattell’s open letter to the APA, these were merely the writings of his blinkered youth in the 1930s, “later amended based on subsequent observations.” The books he did stand behind are, however, hardly less offensive. In Beyondism: Religion from Science, Cattell gives a seemingly L. Ron Hubbard–inspired name to the thoughts that had been percolating for decades—that statistics can be used to determine which cultures should survive and which should not. To determine relative fitness, Cattell argues that nations and different cultural groups should compete, with a world council setting the rules. War was to be expected, and indeed was dictated by evolutionary law.

Interestingly, many of the central tenets of Beyondism are echoed in recent movements arguing for the broader advancement of civilization at the expense of global communities and environmental concerns in the present (Timnit Gebru and Émile P. Torres call this collection of ideas “TESCREAL,” for transhumanism, extropianism, singularitarianism, cosmism, rationalists, effective altruism, and longtermism). For those committed to TESCREAL-centered ideas, one of the quickest paths to a promised technotopia is to adhere to a metrics of worth. Clear ideological connections link Marc Andreessen’s “Techno-Optimist Manifesto,” which states that “Techno-Optimists believe that societies, like sharks, grow or die,” to Cattell’s Beyondism movement. History, it would seem, is recursive.

A function of government under this system, writes Cattell, should be “eugenic control” to maximize national fitness. Societies experiencing famine or poverty at the hands of colonizers and oppressors were demonstrably unfit and should not be allowed to continue. In a later book titled How Good Is Your Country? What You Should Know (1994), he would write that a “control of the instinct of compassion is called for.” It was published as part of the Mankind Quarterly Monograph Series by the Institute for the Study of Man, composed of members known for their opposition to civil rights and fondness for neo-Nazis. Discussing Mankind Quarterly in her 2019 book Superior: The Return of Race Science, Angela Saini notes that “the target audience didn’t appear to be the academic community at all, but racist movements searching for evidence that their prejudices might be rooted in scientific fact.”

To determine what was “good” in a culture, Cattell in How Good Is Your Country? concocted “a factor” that he called “intelligent affluence,” constructed by measuring subfactors such as expenditure on education, sterilization of the unfit, and musical creativity. The “morale” factor measured the number of deaths from syphilis (ideally, low), the birth rate (ideally, low), and the number of people per household (ideally, low). Another factor, “vigorous development,” measured income per capita, as well as percentages of the “Nordic race” and of Protestants (ideally, high).

Cattell considered none of this to be racist. A racist, he declared, believed without evidence that his own race was superior to others, whereas he, by contrast, relied only on hard evidence. It wasn’t his fault that his own race came out on top. In Beyondism, he constructed the portmanteau term “ignoracism” to describe those other “bigoted individuals” who have “totally refused to consider the scientific possibility that races may show statistically significant differences.”

¤

The Cattell papers, held at the University of Illinois Urbana-Champaign, aren’t like most other academic archives. There’s no correspondence, no drafts or mementos, no evidence of a love of music or the arts. It consists only of his 50 or so books, most of them out of print. The sheer count of his publication record clearly mattered enormously to his impact, even as he systematically sacrificed clarity for productivity, and never reread or revised. It seems strange—and uncanny—that this is the man who has so shaped our current musical environment. A pair of papers from 1954 and 1960 were especially notable in this regard, and the first of their kind as far as I know. In these essays, he sought to uncover whether musical preference can be condensed into measurable dimensions, and whether those dimensions can then be used to predict behavior and the psychological traits of listeners.

Co-authored with David R. Saunders, the 1954 study, entitled “Musical Preferences and Personality Diagnosis: I. A Factorization of One Hundred and Twenty Themes” and published in The Journal of Social Psychology, asked 196 participants (mostly university students) whether they liked or disliked musical excerpts from piano recordings. Of the 120 themes, 55 of them fall under the broader umbrella of “classical” (Bach appears five times, Chopin seven times, Debussy five times, Schumann five times, and Beethoven four times). Many themes are simply referred to as “bebop”—Black musicians are far likelier to be unnamed. Repeatedly lauding the neutrality of science, Cattell simultaneously ignores the glaring unneutrality of his design. For example, the excerpts should be “catholic,” he writes, and should represent all personality types, but none of the examples were written by women. A strong Western European and American bias prevails as well, with only a handful of token samples, unnamed and unattributed (“Yiddish Lullaby,” “Folk Song from India”), from other cultures.

“Is there a tendency for preferences for certain kinds of music to be systematically related to the kinds of personality structure?” Cattell asks in the paper. “What type of music can in fact be regarded as an adjusting or therapeutic agent for this particular personality[?]” And then: “[T]he investigation of these differences may be the best starting point for later work on psychotherapy itself.” The goal, it’s safe to conclude, was never to understand music, or to understand the listener, but to use music as a diagnostic tool.

Those who preferred bebop and Art Tatum were characterized as high in “surgency” (Cattell’s “factor” for something akin to extroversion, enthusiasm, and impulsiveness—he frequently came up with his own terms); those who disliked Schoenberg’s Op. 11 and Schubert’s “Death and the Maiden” were rated high on the “conventionalism and conservatism” factor. Some groups were trickier. What did it mean to dislike Brahms’s “Hungarian Dance” but appreciate Hanson’s “Romantic Symphony”? Unknown. Nevertheless, the foundation was laid: a factor analysis could determine 12 dimensions of musical preference. It’s worth highlighting the implicit hierarchies at work here: the preference for musical styles performed mostly by Black musicians mapped onto poor impulse control and, more broadly, unfitness. The preference for European classical music was related to what Cattell called the “I factor”—imagination, sensitivity, and cultural sophistication.

In 1960, Cattell and another co-author, Robert McMichael, drove to Manteno State Hospital in Kankakee County, Illinois, about 40 miles south of Chicago, to work with 230 patients. The earlier study had also done some work at Manteno but was thwarted when Cattell and Saunders were unable to hold the attention of any of the 188 patients for an additional two hours for a second test. With a reputation for treating patients poorly, Manteno was one of the largest psychiatric hospitals in the world, its patient population peaking at 8,195 in 1954. Forty-seven patients and staff had died there in 1939 of typhoid fever in what Time magazine would call “Manteno Madness.” During World War II, 60 beds were set aside for a US government project that injected unknowing patients with malaria. Various reports thereafter stated that patients were subjected to experimental and nonconsensual surgeries. Unsurprisingly, the hospital increasingly struggled to hire and retain staff. Electroshock therapy was introduced there in 1936 and then became a staple. A 1953 report from the Central Inspection Board of the American Psychiatric Association states that, in that year alone, 1,465 patients received a total of 7,379 electroshock treatments. The hospital was also overcrowded, at nearly twice the capacity of the APA’s standard.

It was in this environment that Raymond Cattell conducted the first studies exploring the relationship between personality and musical preference. In the 1960 study, 230 patients submitted complete answer sheets, but only 190 were usable after the authors discarded the invalid records, including those in which the participants filled in all the blanks before the test started, copied from their neighbors, or did not listen to the music at all. The remaining answer sheets were then compared with a control group—60 people from a church social group, who were allowed to take the test together in church. Those at Manteno were placed in groups by ward, with eight to 10 staff members present and “encouraging” them to complete the test. The participants had spent no fewer than three years at the hospital; those diagnosed with some form of psychoneurosis spent an average of more than seven years there.

Cattell and McMichael found that the Music Preference Test was able to distinguish between the patients at Manteno and the control group from the church, and that musical preference could be used to distinguish between types of psychiatric conditions. Schizophrenics, for example, scored high on the factor that Cattell referred to as “introspectiveness [versus] social contact”—they preferred music that the authors associated with withdrawal and inward focus. Those diagnosed with paranoia were distinguished by their low scores on “tenacity [versus] cyclothyme relaxedness”—the latter a trait associated with emotional warmth and sociability—suggesting that their preferences were rigid and immovable. Those diagnosed as psychoneurotics—the archaic umbrella term for what would now be thought of as anxiety disorders, mood disorders, or obsessive-compulsive disorder—were notable with respect to the “adjustment [versus] frustrated emotionality” factor, with their preferences clustering toward the frustrated end of the scale.

Ultimately, Cattell and McMichael hoped they might someday deploy dimensions of musical preference to diagnose any listener, and to predict who they might become and what pathologies they might inhabit. For example, those who preferred music associated with withdrawal and inward focus were more likely to be diagnosed as schizophrenic, while those who were unable to be moved by music were more likely to be diagnosed as paranoid. According to Cattell and McMichael, the listeners in the church group—referred to as “normals”—were more socially sensitive, more driven, and more emotionally open than the patient population. The church group defined the norm, and any deviations were symptoms.

¤

Cattell’s work is still systematically taught, often uncritically, in psychology classrooms. His work applying factor analysis to human personality traits is an early precursor of how we discuss the measurement of personality. Although he argued for a 16-factor model, psychologists today typically refer to a model that incorporates five factors, with the most popular being the “Big Five” or “OCEAN” model (for openness, conscientiousness, extroversion, agreeableness, and neuroticism). In his 2009 book The Cattell Controversy: Race, Science, and Ideology, William H. Tucker writes that “even though [Cattell] refused to acknowledge paternity, the Big Five was indisputably his illegitimate child.” Interestingly, Spotify’s “listening personality” model still uses a 16-factor approach akin to Cattell’s.

In 2018, Spotify applied for a patent entitled “Methods and Systems for Personalizing User Experience Based on Personality Traits.” Using the Big Five, the new model “assigns a first personality trait of a plurality of personality traits to the user. The electronic device provides personalized content to the user based on the first personality trait.” The personality trait might be, say, extroversion. Spotify checks the user’s listening history, assigns one or more characteristics of songs to “extroversion,” tracks the listener’s behavior to see how well this personality trait fits (do they skip these songs? do they share them?), and then provides personalized content. The process would then move to a second personality trait.

A separate Spotify patent filed in 2018 sought to analyze sound from the user’s device (speech content, environmental metadata, etc.) to analyze the listener’s gender, age, emotional state, and speech accents, and whether the listener was on a bus or subway, or alone or at a party. The patent argued that recommendation required users to “tediously input answers to multiple queries in order for the system to identify the user’s tastes.” Consensually reporting on your own musical preferences was a chore; let Spotify take care of it for you. In 2023, a company called Chime Health AI filed a patent for a method of assessing health based on listening preferences. The authors of the patent argued that “music and media preferences […] may be a good indication of a user’s health status.” Specifically, they analyze how a listener responds to music to provide “rapid, science-driven diagnostics for conditions like ADHD, Autism, Depression, PTSD, and 280 score reports.” Streaming services, once simply large databases of music, and eventually personalized recommendation machines in which listeners are treated as vectors of measurable preference dimensions, are thus once again on the verge of being used to diagnose well-being.

The origins of these methods have gradually been erased. Studies in the 1990s and early 2000s cited Cattell’s work, but the patents that followed did not. An epistemic and methodological laundering is at play. We’ve retained the methods that treat listeners as fixed vectors of preferences, but without recognizing that such factors are derived from an unethical study experimenting with patients at an overcrowded hospital who were repeatedly subjected to electroshock therapy. That study was conducted by eugenicists who sought to use these techniques to determine which cultures deserved to live on.

Whereas the late 20th century saw the rise of algorithmic recommendation as a technology, the 2010s saw a return to Cattell’s initial formulation that essential aspects of the listener could be inferred through musical choice. We can trace an arc of decreasing agency. For Seashore, the listener was a bundle of biologically inherited sensory organs; individual preferences were of no concern. For Cattell, the listener was a personality to be measured, with the measurement then signifying broader notions of “fitness.” Spotify’s listener is a behavioral trace—a vector extracted from conscious and unconscious actions, tracked through the unknowing communication of biomarkers and environmental cues.

The arc from Cattell to Spotify is not so much a story of influence as one of inheritance. Studies in the early 2000s on music and personality cite Cattell as the origin of the music-preference research program, but with no mention of Manteno and its electroshock wards, and certainly not the eugenicist religion he promulgated. The patents that followed would cite subsequent approaches to personality (the Big Five) as established scientific principles, without many supporting citations. Spotify’s listening personalities reverted to the same number of personality traits—16—as Cattell. Each layer of transmission drops another layer of context until what is left is structurally identical to Cattell’s diagnostic approach: the listener is an immovable vector of measurable preference dimensions whose choices reveal who they are and inform a prediction about who they will be.

Daniel Ek’s promise—that Spotify would become the trusted friend who knows what you want better than you do, who will introduce you to things you’d never have chosen yourself—is not a new idea dressed up in the language of technology. It is the same claim Cattell made at Manteno in 1954—that the right measurement apparatus understands subjects better than subjects understand themselves, and that this understanding can be used to predict, categorize, and ultimately manage them. Cattell thought the endpoint was clinical diagnosis. Ek thinks it’s a better playlist. While the two ambitions are different, they share the model of a listener who is passive, typological, and best understood by the system.

¤

Featured image: Underwood & Underwood, [Woman with headphones listening to radio], ca. 1920–30. Library of Congress, Prints & Photographs Division [LC-USZ62-23645]. CC0, loc.gov. Accessed May 26, 2026. Image has been cropped.