Learning how to obtain rewards (e.g., food) is important for survival. Behavioural and neuroscience research have suggested that reward learning reflects the operation of two distinct neuro-cognitive systems: the goal-directed and habit systems (Balleine & O’Doherty, 2010). Recently, this dichotomy has been challenged by authors proposing that, what we thought were habitual responses, are better understood as goal-directed actions (Kruglanski & Szumowska, 2020). This letter is a critical assessment of this hypothesis, which in our opinion can hardly explain the results of recent studies. Also, we suggest that the debate about the goal-directedness of habits is probably insoluble from an experimental point of view and almost irrelevant from an applied/clinical perspective. A more fruitful approach might be to analyse to what extent a behaviour is controllable —regardless if it is purposeless or not.