Searching for Personality [Social Sciences]

In this column, the author discuss two approaches to automatic personality prediction. The first approach takes inspiration from human processes of personality attribution, whereby humans deploy knowledge about personality to attribute traits to other people, even those they never met before and even on the basis of very short sequences (down to a few seconds) of expressive behavior (so-called thin slices). The human attribution process can be described by means of Brunswick's lens model, as modified by Scherer. In this model, personality traits are externalized or made manifest in behavior by means of objectively measurable variables called distal cues, which the perceiver represents him/ herself as subjective/proximal percepts; the latter are then subjected to inferential processes leading to attribution. For instance, a distal cue (an externalization) of the extraversion trait can be the voice pitch, which the observer represents as loudness (the proximal percept) that he/she uses in the course of the inferential process.