A West African master drummer in a Ewe ensemble, a tabla player in a Hindustani recital, a clave-pattern percussionist in a Cuban son band, and a Bulgarian wedding band playing in 7/8 are all doing something a 4/4-trained Western ear initially cannot follow. Each is placing their pattern against an underlying meter the listener is expected to hold internally — a meter that may not be played by any single instrument, that may run for seven, twelve, or sixteen beats before repeating, that may consist of several rhythmic strands simultaneously. To follow such music is to maintain more than one prediction at the same time, a cognitive demand that with practice becomes effortless to native listeners and feels essentially impossible to non-native ones.
West African polyrhythm (Ghana, Nigeria, Mali) is the deepest root: ensembles of three to six drummers play interlocking patterns, each with its own felt downbeat, and the aggregate texture is not a single meter but a crossing of several. The tradition shaped the rhythmic complexity of jazz, funk, Afro-Cuban music, and much of contemporary popular music. Indian tala organizes time into cycles of 7, 9, 12, 16, or sometimes 32 beats, with named vibhags (sub-groupings) and characteristic theka (drum patterns); the listener learns to hear the cycle's sam — its downbeat-of-downbeats — and the entire performance breathes around it. Cuban son clave — the 3-2 or 2-3 clave pattern — is the rhythmic backbone of son montuno, salsa, mambo, and most descendants; a band falling out of clave is, in Cuban music criticism, a serious failure. Bulgarian aksak (limping) meters use uneven groupings (7/8 as 2+2+3, 9/8 as 2+2+2+3); folk traditions that influenced Béla Bartók and re-entered Western art music through him.
What makes these traditions hard for outsiders is not abstract difficulty — children raised inside them produce them effortlessly — but enculturation. The brain learns rhythmic templates in early childhood much as it learns the phonemes of a native language; templates not encountered young become difficult to acquire later, in roughly the same way a phonemic distinction never heard becomes difficult to hear. Patel and Iversen's rhythm-perception research has documented this rhythmic-mother-tongue effect across cultures and age groups. The implication is concrete: hours spent with an unfamiliar tradition build the prediction templates that make its music intelligible. There is no real ceiling; adult learners can become competent with sustained exposure.
AI music generators trained on Western pop default to 4/4 because their data does, and one of the more interesting open problems in algorithmic music is the cross-cultural training-data correction. Sonic Pi and TidalCycles — domain-specific languages for live coding music — are designed in part to make polyrhythmic and asymmetric structures expressible without first translating them into Western notation. Algoraves, live-coded electronic dance events, have made cross-cultural rhythmic vocabulary newly audible to a generation that grew up on common-meter pop. Western popular music in 2025 is rhythmically more cosmopolitan than at any point since recording began.