People can code their actions on different levels of abstraction, with reference to meaningful goals or arbitrary SR-mappings. Does this coding determine how they are influenced by others’ actions? This question was addressed in a Simon task where participants reacted to the colour of circles with left or right keypresses, after which a partner moved another circle in the same or opposite direction. We varied the contextual embedding between three groups. The Goal group was instructed to sort dishes via keypresses, moving dirty (brown) plates to a sink and clean (white) plates to a cupboard. The SR-static group was simply told to perform keypresses depending on circle colour, but triggered no distal effects. The SR-dynamic group received the same instruction but triggered movements like the Goal group. We expected partner compatibility effects in both SR groups, specifically in Simon-incongruent trials, but no influence of the partner in the Goal group. The results revealed an interaction of partner compatibility and Simon congruence, but not in the expected way. Partner effects were absent regardless of Simon congruence. However, when the partner moved to the opposite direction, Simon interference was stronger, indicating increased susceptibility to irrelevant spatial information. There was no interaction with group. However, in the SR groups, individual participants’ Simon effects seemed affected by partner reactions in opposing ways, whereas in the Goal group, most participants’ Simon effects were completely indifferent to partner reactions. Thus, abstraction in action coding might affect susceptibility to distraction, but in more complex ways than expected.