7. Balkan: Aksak and Asymmetric Groove

In the villages of Bulgaria, North Macedonia, and Serbia, people dance in metres that Western music theory calls irregular. 7/8, 9/8, 11/8, 13/8 — time signatures that look uncomfortable on paper but feel entirely natural when your feet are moving. The secret is that these are not subdivisions of a regular pulse. They are additive rhythms: groups of twos and threes assembled into cycles that are asymmetric by design.

Theory deep dive

For the rules of additive counterpoint stated explicitly — inviolable cell sequences, cell-head accents, the tupan’s stroke grammar, and why swing is off-limits — see Additive Counterpoint: Balkan.

Aksak: The Limping Rhythm

The Turkish word aksak means “limping” — an apt description for rhythms that alternate between short (2) and long (3) groupings (Brăiloiu 1951). A rachenitsa in 7/8 might be felt as 2+2+3, or equivalently as “short-short-long.” Daichovo horo is 9/8 as 2+2+2+3. Kopanitsa is 11/8, typically 2+2+3+2+2.

These are literally Euclidean rhythms. E(3,7) distributes three onsets across seven positions, producing the inter-onset pattern 2+2+3 — exactly the rachenitsa.

One caveat before the diagram: written aksak notation makes the long beat look like exactly 1.5× the short beat, but performed aksak long beats are not exactly 1.5× short beats. Goldberg’s field measurements (Goldberg 2015) show that the long-to-short ratio deviates systematically from 3:2, varying by tradition, tempo, and player. The additive-cell notation is a good structural abstraction, not a claim about the actual millisecond ratio.

E(3,7)

This is not a coincidence. The correspondence is mathematical, not historical: Bjorklund’s algorithm was originally derived for pulse distribution in neutron-beam timing at the Spallation Neutron Source⁴⁶, and Toussaint¹ later identified that the same even distribution appears in additive dance meters that Balkan folk musicians have long favored. The algorithm did not discover the tradition and the tradition did not discover the algorithm — the same mathematical distribution describes both.

Rotation Selects the Variant

The rachenitsa exists in multiple rotational forms: 2+2+3, 2+3+2, and 3+2+2. Each variant has a different character. The first pushes forward, the second sways, the third plants the long group on the downbeat and rushes through the short groups.

In Poly, these three variants are one necklace of gaps read from three different starting points; the rotation parameter selects which onset lands on the downbeat. E(3,7) at rotation 0 gives 2+2+3. At rotation 3, the long gap falls first, giving 3+2+2. At rotation 5, the long gap moves to the middle, giving 2+3+2. The rotation parameter is not an abstract offset — it is the selector for which aksak variant you are playing.

E(3,7) r5

E(3,7) at rotation 5 shifts the long group to the middle of the cycle, producing the swaying 2+3+2 variant. Try all three rotations to feel how the same three accents in seven positions create distinctly different dance grooves.

Daichovo and Kopanitsa

The same principle extends to longer cycles. Daichovo horo in 9/8 uses four accents across nine positions:

E(4,9)

E(4,9) at the standard rotation used here produces the grouping 2+2+2+3 — which is the daichovo horo grouping. This is one of the tightest matches between Euclidean distribution and traditional aksak in the entire repertoire: the algorithm and the tradition agree exactly. See E(4,9) in the Euclidean reference appendix for the pattern in full.

Kopanitsa in 11/8 is a different case. Its primary form is 2+2+3+2+2 — five cells, five accents across eleven positions, not four. That five-cell grouping is not the Euclidean distribution of any 4-hit pattern in 11 (it has five onsets, not four); it belongs to the five-hit family, and its canonical Bjorklund spelling is E(5,11). E(4,11), which distributes four accents across eleven, is a related four-accent variant rendering rather than the canonical kopanitsa grouping:

E(4,11)

E(4,11) gives 3+3+3+2; rotation moves through its variants. The point is that Euclidean distribution naturally generates a family of additive groupings, several of which correspond to Balkan dance metres — but the canonical five-cell kopanitsa 2+2+3+2+2 is a five-accent pattern, not the four-accent E(4,11) row.

Rachenitsa GrooveRachenitsa 7/8

Lane	Role	Steps	Hits	Rotation	Subdivision	Note	Velocity	Swing
1	Tupan bass	7	3	0	1/8	36	110	0
2	Tupan rim	7	4	2	1/8	37	85	0
3	Kaval accent	7	2	1	1/8	76	75	0
4	Gadulka pulse	7	5	0	1/16	42	65	0

Two of those lanes are honest mappings and two are stand-ins. The tupan is a double-headed bass drum, so lanes 1 and 2 land on a GM kick (36) and side stick (37) — a drum standing in for a drum. The kaval is an end-blown flute and the gadulka a bowed fiddle, and there is no GM drum equivalent for either: lane 3 is a woodblock (76) and lane 4 a hat (42), holding the rhythmic position while the melodic instrument is absent. Route those two lanes to a pitched instrument if you want the timbre as well as the rhythm.

Rachenitsa 7/87/8 aksak groove — tupan, rim, kaval, and gadulka in additive meter

PlayTry it

Notice that Swing is set to 0 on every lane. Aksak rhythms are not swung — their asymmetry comes from the additive grouping structure, not from micro-timing displacement. Applying swing to a 7/8 pattern would fight the natural limping quality rather than enhance it.

Bulgarian Wedding Music: Dense and Fast

The Bulgarian wedding music tradition (svatbarska muzika) pushes aksak metres to extraordinary tempos (Rice 1994). The tradition was condemned by the socialist state and adored anyway, then declined after 1989 as chalga rose (Silverman 2007) — the virtuosity below is a competitive scene’s, not a museum’s. Kopanitsa at 200+ BPM becomes a blur of interlocking accents where individual onsets merge into a continuous textured flow.

Kopanitsa GrooveKopanitsa 11/8

Lane	Role	Steps	Hits	Rotation	Subdivision	Velocity	Humanize
1	Kick accent	11	4	0	1/8	115	0.1
2	Snare pattern	11	5	3	1/8	90	0.15
3	Hi-hat pulse	11	7	0	1/16	70	0.1
4	Bell accent	11	3	2	1/8	80	0.1

Kopanitsa 11/8Fast Bulgarian kopanitsa — kick, snare, hat, and bell in eleven-beat cycle

PlayTry it

Humanize is kept low — 0.1 to 0.15 — because Balkan folk ensembles play with remarkable tightness. The rhythmic interest comes from the additive structure, not from loose timing. Even at extreme tempos, the players lock together with mechanical precision that would impress a drum machine.

Asymmetry as Natural Order

Western music theory treats these metres as exotic exceptions. But in Southeastern Europe, they are the default. Children learn to dance in 7, 9, and 11 before they encounter 4/4. The asymmetry is not heard as irregular — it is simply how music moves.

Poly’s engine is indifferent to whether a step count is even or odd, divisible or prime. A 7-step cycle is no more complex for the algorithm than a 16-step cycle. This neutrality means that aksak rhythms are not a special mode or an advanced feature — they are just another set of valid parameters.

What to Listen For

Play the rachenitsa patch at a moderate tempo and try counting along. At first, the 7/8 feel will seem to stutter against your internal 4/4 expectation. After a few cycles, the asymmetric pulse becomes natural — you stop trying to fit it into even groups and start feeling the 2+2+3 lilt as its own kind of groove. That shift in perception is the point: aksak rhythms are not broken symmetries, they are complete asymmetries.