Measurements of seismic anisotropy in continental regions are frequently interpreted with respect to past tectonic processes, preserved in the lithosphere as “fossil” fabrics. Models of the present-day sublithospheric flow (often using absolute plate motion as a proxy) are also used to explain the observations. Discriminating between these different sources of seismic anisotropy is particularly challenging beneath shields, whose thick (≥200 km) lithospheric roots may record a protracted history of deformation and strongly influence underlying mantle flow. Eastern Canada, where the geological record spans ∼3 Ga of Earth history, is an ideal region to address this issue. We use shear wave splitting measurements of core phases such as SKS to define upper mantle anisotropy using the orientation of the fast-polarization direction ϕ and delay time δt between fast and slow shear wave arrivals. Comparison with structural trends in surface geology and aeromagnetic data helps to determine the contribution of fossil lithospheric fabrics to the anisotropy. We also assess the influence of sublithospheric mantle flow via flow directions derived from global geodynamic models. Fast-polarization orientations are generally ENE-WSW to ESE-WNW across the region, but significant lateral variability in splitting parameters on a ≤100 km scale implies a lithospheric contribution to the results. Correlations with structural geologic and magnetic trends are not ubiquitous, however, nor are correlations with geodynamically predicted mantle flow directions. We therefore consider that the splitting parameters likely record a combination of the present-day mantle flow and older lithospheric fabrics. Consideration of both sources of anisotropy is critical in shield regions when interpreting splitting observations.