Recent works indicated that performing a joint spatial compatibility task with an incompatible stimulus-response mapping affects subsequent joint Simon task performance, eliminating the social Simon effect (social transfer of learning effect or SToL effect). Crucially, the SToL effect was not tuned to the specific identity of the co-actor, and depended on the overlap between the spatial relations of the practice and transfer tasks. Starting from these findings, this study aimed at investigating which spatial relations between stimulus (S), response (R) or participant (P) positions are relevant for the SToL effect to occur. Two experiments were run in which the participant-response associations were incompatible (participants were required to respond with crossed arms), whereas the stimulus-response and stimulus-participant associations were manipulated. We found that learning derived from the practice task did not transfer to the subsequent task when stimulus-response associations were spatially incompatible and stimulus-participant associations were compatible (Experiment 1). However, a SToL effect was evident when stimulus-participant associations were spatially incompatible and stimulus-response associations were compatible (Experiment 2), hence suggesting that the spatial relation between stimulus and participant positions is crucial for the SToL effect to occur.