A recent study published in Psychological Science provides evidence that people naturally absorb the underlying rules of music just by listening to it over their lifetime. The findings suggest that both trained musicians and people with no musical background use harmonic context in remarkably similar ways to predict and remember musical patterns.
Scholars have long debated whether formal training is required to understand the deeper harmonic frameworks of music. Music is organized into layers of notes, phrases, and sections, similar to how language is structured into words and sentences. Some experts believe that understanding this organization requires explicit instruction in music theory.
Other scientists argue that mere passive exposure to music allows the brain to implicitly learn these rules. Previous studies have yielded mixed results regarding how formal training impacts a listener’s ability to process tonal context. Tonal context refers to the overarching harmonic organization or key of a piece of music.
“There has been quite a bit of work looking at how listeners build up musical context, but the open question was how much context is actually used,” said corresponding author Riesa Y. Cassano-Coleman, a PhD candidate at the University of Rochester and member of the SoNIC (“Science of Neural, Interpersonal Communication”) lab.
“This study was also motivated by an analogous line of research in language/narrative: different areas of the brain respond to different amounts of coherent context in narratives (Lerner et al., 2011 J. Neurosci.). Basically everyone is an “expert” in narratives (at least in the sense that people use language and stories to communicate in everyday life) but not everyone is an expert in music. So music provides an interesting test case: do we need formal training in music to understand musical structure?”
To resolve this debate, the researchers designed a systematic way to test listeners. They wanted to see how the amount of coherent musical information affects a person’s ability to encode, predict, and segment music. By scrambling musical pieces at different time intervals, the scientists manipulated the amount of musical context available to the listener.
The researchers conducted four separate experiments using piano pieces from Pyotr Ilyich Tchaikovsky’s Album for the Young. They created different versions of the music by scrambling the pieces at various timescales. The conditions included one-bar scrambles, two-bar scrambles, eight-bar scrambles, and completely intact music.
A musical bar, or measure, is a small segment of time containing a specific number of beats. By keeping features like volume, instrument type, and speed constant, the scientists ensured that participants were reacting only to changes in the harmonic structure.
“On the surface, our different musical stimuli sound pretty uniform: same piano timbre, same tempo, no change in dynamics (volume),” Cassano-Coleman explained. “So the only thing that changes across conditions is the underlying structure. What we wanted to test is to what extent listeners used that structure, specifically to remember and predict in the music.”
In the first experiment, the researchers tested musical memory. They recruited 108 adults between the ages of 19 and 41, splitting them evenly into musicians with at least five years of training and nonmusicians with no training.
Participants listened to a sixteen-second prompt from one of the scrambled conditions. After a brief delay, they had one and a half seconds to identify which of two short musical clips had appeared in the prompt. The results showed that memory improved for both groups as the music became less scrambled.
Musicians performed better overall on the memory task, but both groups benefited from longer stretches of intact music at the exact same rate. Even within the musician group, having more years of practice did not predict better memory performance. This suggests a shared underlying mechanism for memory encoding.
The second experiment tested musical prediction using a distinct sample of 108 adults, again divided evenly between musicians and nonmusicians. Participants listened to fourteen-second prompts and then chose which of two short clips best completed the musical sequence. The data revealed that prediction accuracy increased as the amount of intact musical context increased.
In this prediction task, musicians did not perform better than nonmusicians. Both groups utilized the available harmonic information equally well to guess the upcoming notes. This provides evidence that people unconsciously apply the rules of music theory to anticipate what comes next, regardless of their formal education.
In the third experiment, the scientists explored event segmentation, which is how people mentally divide continuous sounds into meaningful chunks. A sample of 95 adults listened to longer, one-minute musical pieces. They were instructed to press the spacebar on their keyboard whenever they heard a meaningful change in the music.
“The event segmentation task requires context integration in real time: in order to segment the music into meaningful events, as you’re listening, you have to remember what you just heard, predict what’s coming next, and decide if it’s enough of a meaningful change to mark an event boundary,” Cassano-Coleman told PsyPost. “This gives us some insight into how these processes unfold under more natural listening conditions over tens of seconds or minutes, rather than just the 15 or so seconds in the memory and prediction tasks.”
As the music became more heavily scrambled, all participants pressed the button more frequently. Their responses naturally aligned with the new boundaries created by the scrambling process. When the music was left intact, both groups successfully identified standard eight-bar phrases as meaningful events.
A difference between the groups emerged when looking at longer musical structures. Musicians tended to align their button presses with sixteen-bar hyperphrases, which are larger musical sections made up of multiple smaller phrases. Nonmusicians tended to stick to identifying the shorter eight-bar phrases.
The researchers conducted additional checks to ensure that simple changes in pitch height or rhythmic speed were not guiding these responses. This provides evidence that listeners were genuinely tracking the underlying harmonic rules.
The fourth experiment tested explicit awareness of the structural disruptions. The researchers asked 108 participants from the first two experiments to listen to the one-minute pieces and explicitly identify the level of scrambling. Participants had to choose whether the music was scrambled every bar, every two bars, every eight bars, or left intact.
Musicians performed better on this categorization task than nonmusicians. This suggests that explicit theory training helps people consciously reason about musical structure. Yet both groups struggled most with identifying the completely intact music and the highly chaotic one-bar scrambles, achieving their highest accuracy on the midlevel scrambles.
“What we found is that listeners do integrate musical context over time, and that you don’t need formal training to make use of it,” Cassano-Coleman summarized. “In other words, disrupting structure (via scrambling) disrupted listeners’ ability to remember and accurately predict what comes next. The biggest thing that surprised us was just how similarly musicians and non-musicians perform in these tasks – musicians did seem to have an advantage in explicit labeling (experiment 4), but otherwise both groups performed better (at similar rates) with more intact context.”
While this study offers a detailed look at music cognition, it does have a few limitations. The researchers utilized pieces from Western classical music, which follows a very specific set of harmonic rules. It remains unclear if listeners would show the exact same patterns of context integration when listening to diverse genres or music from other cultures.
Future research could explore how different musical features, such as changing rhythms or different combinations of instruments, drive event segmentation. Scientists also plan to investigate how highly trained musicians use this type of context while actively performing. Combining behavioral tests with brain imaging could help pinpoint exactly how the mind merges multiple streams of auditory information.
“In terms of future directions, we’re interested in what’s happening in the brain as people listen to or as expert pianists play these scrambled stimuli in the fMRI scanner,” Cassano-Coleman said. “Stay tuned for that!”
The study, “Listeners Systematically Integrate Hierarchical Tonal Context, Regardless of Musical Training,” was authored by Riesa Y. Cassano-Coleman, Sarah C. Izen, and Elise A. Piazza.
