Norwegian stev, short 4-line songs, are performed in southeastern Norway in a reciting-singing style in Telemark and Setesdal that may stem from a millenium ago. The reciting-singing style, called kveding, has been thought to be an unsystematic “free rhythm”. However, solo reciter-singers called kvedarar (pl.; kvedar, s.) often foot-tap the irregular rhythm. Groups of listening kvedarar can foot-tap the irregular rhythms simultaneously, independent of whether the solo kvedar foot-taps or not. This raises the question: how can these irregular rhythms be predictable? Analysing foot-tapping in stev performance documented by film reveals a short-long two-pulse rhythm that can vary considerably. This two-pulse may be called a “dipod”, meaning that 2 poetic (metrical) accents establish the backbone or core of each two-pulse phrase. It seems that poetic accents in stev are always word accents, bringing sung stev close to “speech” and in that sense “reciting”. The reciter-singers (kvedarar) traditionally foot-tap these poetic accents and only these; they never foot-tap linguistically unaccented syllables (such as “ble” in “syllable”). This then is accentual poetry, literally. The poetic dipod and its two-pulse in performance is old, reminiscent of two-accent lines in Old Norse folk poetry, yet still emerging in popular and classical lyrics, which will be demonstrated.