A learning model can explain both shared and idiosyncratic first impressions from faces

Why is the literature on first impressions so focused on White faces?

“…Similarly, high levels of consensus, where observed, have been cited as evidence for 'universality' and an innate explanation of first impressions [30,108,110]. By contrast, evidence that first impressions lack accuracy and vary between cultural groups accords well with the view that first impressions are culturally acquired [43,100,104]. As described above, however, the focus on White faces and the use of predominantly White participants has artificially exaggerated evidence of accuracy and consistency, thereby 'stacking the deck' in favour of nativist accounts of first impressions.…”

“…In recent years, we have argued that first impressions are the products of culturally acquired mappings between visual representations of different face shapes and representations of the various trait profiles that other people may possess [ 19 , 42 , 43 , 100 , 104 , 105 ] ( figure 2 ). According to this account, trait inferences from cues to ethnicity are closely comparable to trait inferences from other facial cues.…”

The cultural learning account of first impressions

“…Contrary to the prevailing view at the time, TIM [21][22][23][24] proposed that automatic first impressions are the result of associative mappings acquired by individuals during the course of their lifetime that connect perceptual descriptions of facial appearance (points or regions in face space [ Box 1]) with representations of the potential trait profiles that others may possess (locations in trait space [25][26][27][28]). These mappings are acquired via domain-general associative processes following exposure to correlated face-trait experience; i.e., learning episodes where certain facial features or feature configurations are predictive of particular trait profiles.…”

Reply to Cook and Over: Social learning and evolutionary mechanisms are not mutually exclusive

“…We thank Cook and Over (1) for their response to our recent article, where we find that unique experiences drive individual differences in impression formation (2). Overall, our theoretical positions are similar: We suggest that social associative learning underlies this core finding (2), and Cook and Over (1) clearly agree.…”

“…Social learning cannot easily account for this example nor other findings in facial impression research. For instance, Cook and Over (1) suggest that impressions should be measured in samples that are non-Western, educated, industrialized, rich, and democratic: Indeed, Zebrowitz et al (9) carried out an elegant study with the Tsimane' people, a culturally isolated group in Bolivia. Although there were cultural dialects, overall, these remarkably different cultures formed highly similar impressions, suggesting that at least some contingencies between cues and impressions are evolutionarily predisposed.…”