Tropical cyclone track forecasts are very sensitive to the initial state. This sensitivity can occur in regions both local and remote to the initial tropical cyclone. In this study, linear sensitivity analysis in the form of singular vectors is applied to understand the relationship between tropical cyclone structure and motion, and to examine the influences of the background flow on tropical cyclone movement. Singular vectors represent the fastest growing perturbations to a given trajectory in a linear sense. In this study, the SVs are calculated using the dry tangent and adjoint models of the Navy Operational Global Atmospheric Prediction System (NOGAPS) that have been linearized about the full-physics high-resolution operational forecasts. We use the SV tools to study the large-scale dynamical processes that influence storm motion. It is found that straight-moving and recurving storms exhibit systematic differences in initial sensitivity patterns. The straight moving storms are particularly sensitive to the analyses of the rear quadrants of the storm, about 5 degrees from the storm center, where the background vorticity gradient changes sign. In addition to this sensitivity, recurving storms also exhibit strong sensitivity to the upper-level potential vorticity field in regions of confluent flow. An examination of the appropriateness of dry linear diagnostics for these tropical cyclone motion studies is also included.