The observation that inclusion trails in porphyroblasts commonly exhibit a more or less constant orientation over a large area even though the enclosing rocks have been folded has been considered puzzling by geologists and has generated considerable debate. An analysis of vorticity during multilayer folding in response to layer-parallel shortening shows that the observation is in fact theoretically expected. Folding of multilayers in response to layer-parallel shortening (bulk folding) can be considered kinematically as a time-variable combination of flexural flow and pure shear components. The rotation of a spherical porphyroblast during folding is related to the vorticity history in the vicinity of the porphyroblast by a set of differential equations. For likely folding histories, the solutions to these equations are in agreement with naturally observed patterns. Therefore it is unnecessary to invoke other mechanisms, such as assuming porphyroblasts to be in ‘islands' irrotational with respect to the fold axial plane. More importantly, the theory allows constraints to be placed, from current data set of porphyroblast inclusion trail geometry, on the history of folding and the competence properties of rocks at the time of folding.