The project will construct an internal, length-dependent plastic theory for granular materials to be applied to the study of instability phenomena. Analytical, numerical and experimental evaluation of this concept will be conducted to determine the ability of this approach to provide information on the growth and wave length selection of the associated deformation inhomogeneities. The internal length scale will be introduced either through a Cosserat- type modification characterizing the assymetry of stress, or a gradient-dependent term characterizing the heterogeneity of the plastic flow-rule. The merits and drawbacks of both approaches will be examined and the simplest possible form will be used to study three classes of hazard-related problems: shear banding, surface instabilities, and liquefaction. This work will provide a better understanding of the physical failure processes which are fundamental for the development of appropriate mitigation measures.