From the first soothing lullabies babies hear to the funereal dirges that accompany the deceased, music is a fixture of the human experience. The latest findings on the cultural basis of singing have ascertained that certain song types—whether lullaby, dance, love or healing melodies—turn up in all cultures. And each kind of song has common cross-cultural acoustic features.
By analyzing a database of ethnographic descriptions and recordings of songs from around the world, a research team from multiple universities discovered commonalities in acoustic features such as tone, tempo and pitch. The organization of melodies around a hierarchical chord structure—tonality—appears as a common attribute across all cultures included in the database. Music varies more within than between cultures, indicating that each uses song in the same social context.
The study tries to come to grips with the debate among musicologists as to whether song is a universal human trait that may have deep evolutionary roots. Darwin thought early hominins sang before they developed language, and certain aspects of music appreciation appear to be hardwired in the human brain. “Music is a biological human function, like language is,” says Isabelle Peretz, a psychologist at the University of Montreal who was not involved in the study. “As a consequence, music is a universal trait and should exhibit universal properties.”
The new research, published Thursday in Science, suggests that humans indeed share a “musical grammar,” says psychologist Samuel Mehr, director of the Harvard Music Lab and the study’s lead author. “Music is built from similar, simple building blocks the world over,” he says.
Mehr and his colleagues compiled 4,709 written accounts of vocal music made by ethnographers and anthropologists over more than a century of research in 60 societies, and combined them with a discography of lullabies, love songs, healing songs and dances to create a cross-cultural database they dubbed the Natural History of Song (NHS), which is available online. They focused on song because unlike instrumental music, singing is independent of technology and has clear biological underpinnings. They chose four kinds of songs to analyze: lullabies, healing songs, love songs, and dances.
“Our goal was to representatively sample the music that people produce around the world,” says study co-author Manvir Singh, an anthropologist at Harvard. “The first thing we had to confront was: How do you represent this music? Ethnomusicologists have rightly critiqued projects that use a single scheme, like representing the world’s music in a Western staff because obviously [that] brings all kinds of biases.” To avoid this, Singh says, the team used a suite of analytical approaches. They transcribed the songs with Western notation, cataloged them with music information retrieval software and had 30 expert musicologists and hundreds of laypersons listen to and make annotations for excitement and pleasantness and other qualities. They then applied machine classifiers to test whether the identifiable acoustic features of each kind of song predict a song is a lullaby or some other in each of the 60 cultures surveyed.
The analyses revealed that music appears in every culture the researchers observed, and that songs vary primarily in three ways—their formality, religiosity and arousal (or level of excitement). Those common acoustic features predict the context in which a song is used in every culture they studied. Music also varies along the two dimensions of melodic and rhythmic complexity.
“Humans everywhere share some cognitive mechanism that may have evolved specifically for music, or they may have evolved for mundane purposes like analyzing an auditory environment,” Singh says. “On the basis of these cognitive mechanisms, certain songs sound appropriate in certain behavioral contexts.”
The researchers took pains to correct for potential biases that any Western ethnographers may have unwittingly brought to their study of non-Western cultures, and noted other possible constraints on their work: experts could have brought their own preconceptions to the annotation process, and the data do not reflect how people in the societies included in the study perceive tonality in their music.
Some musicologists criticized certain aspects of the methodology. “Using Western notation to notate examples and then drawing conclusions from those notated scores is a really problematic practice,” says Shannon Dudley, an ethnomusicologist at the University of Washington who was not involved in the study. “Subtleties of rhythm, subtleties of pitch differentiation, articulation and timbre—there are a lot of things that have a huge impact on the way people hear music that aren’t there in [Western] notation.” Ève Poudrier, a music theorist at the University of British Columbia who was not involved in the study, agrees. Poudrier notes that one of the musical features the researchers mapped to behavioral contexts was whether the song was in a major or minor key. “Major [and] minor is one system of classification of scales that actually is accurate only in Western music,” she says.
Mehr says that staff notation was just one of the five methods the researchers used to analyze the songs. “We find the same result each of the five ways—that form and function are linked worldwide,” he says. “This suggests that while indeed the transcriptions are missing out on some information related to the singing (like timbre, words, accompanying instruments, etc.), they are nonetheless capturing meaningful information about the vocalizations in the songs.” He also says the use of major or minor keys is “simply about the pattern of pitches that are audibly present in the recording. Either they’re present or they’re not, and it’s up to the data to show us whether these features are meaningful in terms of the relation between form and function.” Singh adds that no methodology is perfect, and that the NHS data is all open-access. “We’d be glad for anyone to test our conclusions using an alternative method,” he says.
W. Tecumseh Fitch, a cognitive biologist at the University of Vienna who was not involved in the study, co-authored an accompanying commentary in Science. He says the discovery that musical traits are widespread in humans is consistent with the idea that musicality is very old. “While this [study] certainly doesn’t confirm Darwin’s proto-language hypothesis, I think it’s consistent with the idea that music is very ancient indeed in our species,” he says. “It’s amazing how diverse it is, but at the same time … these commonalities are still there.”