Forming social impressions from voices in native and foreign languages

Abstract

We form very rapid personality impressions about speakers on hearing a single word. This implies that the acoustical properties of the voice (e.g., pitch) are very powerful cues when forming social impressions. Here, we aimed to explore how personality impressions for brief social utterances transfer across languages and whether acoustical properties play a similar role in driving personality impressions. Additionally, we examined whether evaluations are similar in the native and a foreign language of the listener. In two experiments we asked Spanish listeners to evaluate personality traits from different instances of the Spanish word "Hola" (Experiment 1) and the English word "Hello" (Experiment 2), native and foreign language respectively. The results revealed that listeners across languages form very similar personality impressions irrespective of whether the voices belong to the native or the foreign language of the listener. A social voice space was summarized by two main personality traits, one emphasizing valence (e.g., trust) and the other strength (e.g., dominance). Conversely, the acoustical properties that listeners pay attention to when judging other's personality vary across languages. These results provide evidence that social voice perception contains certain elements invariant across cultures/languages, while others are modulated by the cultural/linguistic background of the listener.

Figure 1

Principal Component Analysis solutions and main correlates of the Social Voice Space for “Hola” voices. (A) The two-dimensional solution of the PCA for male (left) and female (right) voices (black dots). Labels equate to: Agg – Aggressiveness; Att-Attractiveness; Comp- Competence; Con – Confidence; Dom – Dominance; Lik- Likeability; Tru – Trustworthiness; War – Warmth.

Figure 2

Upper panel. Correlation matrix and distribution density distribution (diagonal) of PC1 and PC2 of personality traits and voice acoustics for male voices. Black points represent Scottish voices in the original experiment (ScN: Scottish native), Gray points represent Scottish voices in Experiment 2 (ScF: Scottish foreign) and Red points represent Spanish voices in Experiment 1 (SpN; Spanish native). The lower panel is a zoomed in view of the correlation between PCs and F0 (Package GGally in R).

Figure 3

Upper panel. Correlation matrix and distribution density distribution (diagonal) of PC1 and PC2 of personality traits and voice acoustics for female voices. Black points represent Scottish voices in the original experiment (ScN: Scottish native), Gray points represent Scottish voices in Experiment 2 (ScF: Scottish foreign) and Red points represent Spanish voices in Experiment 1 (SpN; Spanish native). The lower panel is a zoomed in view of the correlation between PCs and F0 (Package GGally in R).

Figure 4

Spearman correlation matrix and hierarchical clustering of the voice acoustics in Experiment 1 and 2 (package ggplot2 in R). Black color represents one-to one correlations. The darker the grey, the stronger the correlation between acoustics. The order in which the acoustics are placed determines the ordered clusters.

Figure 5

Principal Component Analysis solutions and main correlates of the Social Voice Space for “Hola” voices. (A) The two-dimensional solution of the PCA for male (left) and female (right) voices (black dots). Labels equate to: Agg – Aggressiveness; Att-Attractiveness; Comp- Competence; Con – Confidence; Dom – Dominance; Lik- Likeability; Tru – Trustworthiness; War – Warmth.