Blurb:

There has been a lot of discussion recently in the Distributed Morphology literature as to how big the local domain for allomorphy is. I argue that an approach in terms of harmonic serialism that envisages word-internal movement triggered by alignment constraints makes it possible to maintain maximal locality for allomorph selection; the reason is that the overt position of some morphological exponent does not have to correspond to its base position (which is the position where it is selected).