Second-grade elasticity revisited