Non-Isochronous Meters: Why 7/8 Is Not Irregular
A meter whose beats are not all the same length is still a meter. The word for it is non-isochronous, and calling it irregular gets the music backwards.
Short answer: in most Western music the beats of a bar are isochronous, meaning equal in duration. In a large body of music from the Balkans, Turkey, the Middle East, South Asia and West Africa, they are not. A 7/8 grouped as 2+2+3 has two short beats and one long one, in that order, every bar. The pulse underneath stays perfectly even. What alternates is the beat layer built on top of it. That is a different regularity, not a missing one.
Isochronous, non-isochronous, and what stays even
It helps to separate two layers that everyday language runs together. Our guide to beat, rhythm and meter lays out the full hierarchy; the part that matters here is the distinction between the fast even pulse and the beats you actually tap.
In 4/4 at 120 BPM the beats are 500 ms apart and the eighth-note pulse is 250 ms. Every beat is the same length. In 7/8 at the same eighth-note rate, the pulse is still 250 ms, but the beats are 500, 500 and 750 ms. Two shorts and a long. The music has not lost its grid. It has grouped the grid unevenly, and it does so the same way in every bar.
This is why the older label "additive meter" is useful. The bar is not divided down from the top; it is added up from the bottom, out of twos and threes. It is also why the practical advice below turns out to be so simple: count the groups, not the units.
The common patterns
| Meter | Usual grouping | Beat pattern | Where you meet it |
|---|---|---|---|
| 5/8 | 2+3 or 3+2 | short-long or long-short | Balkan dance music, Greek and Turkish repertoire |
| 7/8 | 2+2+3 | short-short-long | Bulgarian růchenitsa, a great deal of Balkan dance |
| 7/8 | 3+2+2 | long-short-short | Turkish and Greek repertoire, and much progressive rock |
| 8/8 | 3+3+2 | long-long-short | Turkish and Middle Eastern practice, Latin American clave-adjacent patterns |
| 9/8 | 2+2+2+3 | three shorts and a long | Bulgarian daichovo, Turkish aksak forms |
| 11/8 | 2+2+3+2+2 | a long in the middle | Bulgarian kopanitsa |
Notice that in every row the numbers are twos and threes. That is the general rule for this family: the beat layer is assembled from short units of two pulses and long units of three, in a repeating order. The whole apparent complexity of an 11/8 is the order of five small numbers.
Where the word aksak comes from
The term most often used in the research literature is aksak, from a Turkish word meaning limping. It entered ethnomusicology through the Romanian scholar Constantin Brăiloiu, who wrote about aksak rhythm in the early 1950s and set out the arithmetic relationships between the patterns found in Turkish and Balkan music. His essays reached English readers only much later, in a translated collection published in 1984.
"Limping" is a vivid image and a slightly unfortunate one, because it implies a defect. To a dancer inside the tradition the long beat is not a stumble. It is the step the dance is built around, and it arrives on time every single bar.
Who can hear these meters, and when that gets decided
This is where the empirical work is genuinely striking. Hannon and Trehub ran two studies published in 2005 that together make a strong claim about exposure.
In Psychological Science, they compared listeners on disruptions inserted into simple Western metrical patterns and into complex Balkan ones. Bulgarian and Macedonian adults handled both categories about equally well. North American adults performed clearly better on the simple Western patterns. Infants, meanwhile, detected disruptions in both.
In PNAS, they narrowed down when that adult specialisation appears. Twelve-month-olds already showed an adult-like, culture-specific pattern of responding, in contrast to the culture-general responding evident at six months. Brief exposure to foreign music let the 12-month-olds, but not adults, perceive rhythmic distinctions in the foreign metrical context. The authors raised the possibility of a sensitive period early in life for acquiring rhythm.
The honest reading of that pair of results is not that adults cannot learn non-isochronous meters. It is that adults have to learn them, deliberately, against a set of expectations that formed before they could walk. Anyone who has spent a frustrating hour with a 9/8 has felt exactly this.
What performers actually do with the long beat
A reasonable suspicion about notated non-isochronous meter is that the printed 2+2+3 is an idealisation and real players just approximate it. Polak, London and Jacoby addressed that with a corpus study of Malian jembe drumming, published in Frontiers in Neuroscience in 2016. Their conclusion was that both isochronous and non-isochronous metrical subdivision afford precise and stable ensemble entrainment. The uneven grid was not a looser grid.
That result matters beyond ethnomusicology, because it undercuts a common assumption in timing research: that even spacing is a precondition for tight synchronisation. It is not. What synchronisation needs is a predictable pattern, and a repeating short-short-long is entirely predictable.
The cross-cultural picture
Jacoby and McDermott introduced a neat method for reading listeners' internal rhythmic categories directly, published in Current Biology in 2017. Participants heard a random rhythm and reproduced it; their reproduction became the next stimulus, like a game of telephone. Over iterations, reproductions drift towards whatever categories the listener already carries. Those categories turned out to sit at small integer ratios, and the specific set of ratios differed between listeners in the United States and Tsimane' participants in the Bolivian Amazon.
Jacoby and colleagues then scaled the method up, reporting in Nature Human Behaviour in 2024 on 39 participant groups across 15 countries. Every tested group showed a sparse prior with peaks at integer-ratio rhythms. The weight given to different ratios varied between groups, often in ways that reflected local musical practice.
So there does appear to be something general here, namely the pull towards simple whole-number relationships. Which whole numbers count as musical is learned. Our guide to time signatures covers the notation side of the same territory.
How to actually learn one
Count the beats, not the eighths. Counting to seven each bar keeps your attention on the subdivision layer, which is the layer that takes care of itself. Say a repeating short-short-long instead, and mark only those three arrivals with your body.
Use words. Dance traditions that use these meters teach them with spoken patterns, and there is no reason to invent something more clever. Two even syllables plus one longer one will do it.
Move. These are dance meters. Sitting still while counting is the least efficient possible route into them.
Get the pulse stable first. If the eighth-note layer wobbles, the long beat will feel arbitrary rather than long. Our guide to subdivision is the prerequisite here.
A note on this site's games. Chronkle's rhythm tasks are built on an even grid, so they will not teach you a 9/8 directly. What they train is the layer underneath: a stable pulse and an accurate sense of how long a given interval is. That is the part that has to be solid before an uneven grouping feels like anything at all.
Get the pattern into your hands first. Chronmory plays a short rhythm and asks you to repeat it back. It is even-grid practice for the exact skill an uneven meter leans on: holding a pattern accurately enough that the grouping, and not your memory, is what you are hearing.
Frequently asked questions
What does non-isochronous meter mean?
It means the beats of the cycle are not all the same length. In a 7/8 counted 2+2+3, two short beats are followed by a longer one, and that pattern repeats. The underlying subdivision pulse stays even, so the music is fully regular. It is the beat layer, not the pulse layer, that alternates.
Is 7/8 harder than 4/4?
Harder for whom is the right question. Hannon and Trehub reported in Psychological Science in 2005 that Bulgarian and Macedonian adults handled Balkan and Western metrical patterns about equally well, while North American adults did clearly better on the simple Western patterns. The difficulty tracks exposure rather than the arithmetic.
Are children better at learning these meters than adults?
Younger infants appear more open to them. Hannon and Trehub reported in PNAS in 2005 that 12-month-olds already showed a culture-specific pattern of responding, unlike the culture-general responding at six months, and that brief exposure to foreign music let the 12-month-olds, but not adults, pick up rhythmic distinctions in an unfamiliar metrical context.
Do performers of these meters actually play exact ratios?
Not exactly, and the deviations are systematic rather than sloppy. Polak, London and Jacoby analysed a corpus of Malian jembe performance and reported in Frontiers in Neuroscience in 2016 that non-isochronous subdivision supported entrainment that was as precise and stable as the isochronous case.
Is the preference for simple whole-number ratios universal?
The tendency appears to be widespread, but the specific ratios are not. Jacoby and colleagues measured rhythm priors in 39 participant groups across 15 countries and reported in Nature Human Behaviour in 2024 that every group showed peaks at integer-ratio rhythms, while which ratios mattered varied with local musical practice.
What is the quickest way to feel a 7/8?
Count beats, not eighth notes. Say a repeating long-short pattern out loud, such as short-short-long, and let your body mark only those three arrivals. Counting to seven every bar keeps you at the subdivision layer, which is exactly the layer that does not need your attention.
Keep reading
- Time signatures explained: what the two numbers do, and what they leave out.
- Subdivision: the even pulse layer these meters are assembled from.
- Polyrhythms: the other way music runs two groupings at once.
Sources: Hannon, E. E. and Trehub, S. E. (2005), Metrical categories in infancy and adulthood, Psychological Science; Hannon, E. E. and Trehub, S. E. (2005), Tuning in to musical rhythms: infants learn more readily than adults, PNAS; Polak, R., London, J. and Jacoby, N. (2016), Both isochronous and non-isochronous metrical subdivision afford precise and stable ensemble entrainment: a corpus study of Malian jembe drumming, Frontiers in Neuroscience; Jacoby, N. and McDermott, J. H. (2017), Integer ratio priors on musical rhythm revealed cross-culturally by iterated reproduction, Current Biology; Jacoby, N. et al. (2024), Commonality and variation in mental representations of music revealed by a cross-cultural comparison of rhythm priors in 15 countries, Nature Human Behaviour; Hannon, E. E., Soley, G. and Ullal, S. (2012), Familiarity overrides complexity in rhythm perception: a cross-cultural comparison of American and Turkish listeners, Journal of Experimental Psychology: Human Perception and Performance.