Fig. 4

Co-occurrence matrices and dendrograms illustrating emotion stimuli groupings based on the naming data in Maniq and Polish. The values displayed in the co-occurrence matrices are scaled by rows. On each dendrogram, the length of branches represents similarity, i.e., the shorter the branches the more similar the stimuli. The dendrograms were calculated with an average-link agglomerative clustering algorithm using cosine distances. The letter F in the stimulus name stands for ‘female’ and M for ‘male’ (actor). For translations of the Maniq and Polish responses, see Supplementary Tables S12 and S13.