A Reinforcement-learning Framework for Interpreting Trial-by-trial Motor Adaptation to Novel Haptic Environments