“…While the agent exerts more effort in the longer push, the double push provides particularly strong cues about the agent's intention. Inferring intentions is a non-trivial computational task (Gao, Baker, Tang, Xu, & Tenenbaum, 2019) -a task that humans are extremely good at (McEllin, Sebanz, & Knoblich, 2018). Recent work has linked intentions to plans (Bratman, 2009;Shu, Kryven, Ullman, & Tenenbaum, 2020), and defined intended outcomes as those that made a difference to an agent's plan (Kleiman-Weiner et al, 2015).…”