Right-edge lapse has prominence alternation

*Lapse and *Clash have previously been seen to go a long way toward capturing the alternating pattern of stress systems, without reference to feet (e.g. Gordon 2002). We propose a perceptual reason for the fact that many languages allow a lapse to occur specifically at the right edge, arguing that final lengthening can be perceived as a type of prominence and showing a correlation between languages that allow final lapse and having duration as a stress correlate. We show how a typical final-lapse-allowed pattern can be captured under this assumption.