The present paper investigates whether lexicalized visuo-spatial configurations and/or an explicit reference to a human body part can facilitate the interpretation of unknown idioms by comparing the levels of correct interpretation of unknown English and Serbian idiomatic expressions. Two groups of respondents, American and Serbian engineering students, had the task to interpret literally translated idiomatic expressions from a target language they were not familiar with (Serbian or English). The idioms were divided into three groups: (1) visuo-spatial bodily idioms, (2) bodily idioms only, and (3) non-bodily idioms. The goal was to test whether references to visuo-spatial information and/or parts of the body would help respondents interpret the expressions correctly. Our results, compiled from both groups of respondents, suggest that there are significant differences for the three scores, with the combined visuo-spatial and bodily components taking the lead, the bodily component only coming second, and the non-bodily idioms falling strikingly far behind. This could provide support to the well-known assumptions of cognitive linguistics that visuo-spatial configurations and embodiment play a major role in the construction of abstract concepts.