The study of human craniofacial variation exemplifies general problems associated with the analysis of morphological plasticity that owe to the dependence of results on the methods by which phenotypic variation is quantified. We suggest a definition of plasticity that does not subordinate the developmental to the evolutionary: A process model in which changes are not a function of any mean or average, but only of the current state. Geometric morphometrics, a toolkit for assessing and visualizing biological form and its covariates, avoids some of the traditional pitfalls by focusing directly on the analysis of the two- and three-dimensional coordinates of anatomical landmarks. We discuss its potential relevance to phenotypic and developmental plasticity research, as well as some of its limitations, and demonstrate two useful analyses: assessment of asymmetry, and appraisal of integration. We itemize some of our previous studies on causes (inbreeding, environmental circumstances, etc.) and consequences (attractiveness perception) of asymmetry in humans, present some findings relating to the impact of sex on shape, and speculate about the adaptive relevance of one of these processes in particular. A closing argument points out that such considerations are possible only because of the careful separation of assumptions from empirical evidence entailed in the course of this type of data analysis.