Categorical perception of newly learned faces

Face Familiarity, Distinctiveness, and Categorical Perception

“…Indeed, in terms of CP, it would seem that the representations for a novel human face do not differ from those for a monkey face (Campbell et al, 1997). How familiarity, and how much familiarity, with a face achieves CP are still the critical questions to be answered in future research (see Viviani et al, 2007). One possibility from the current data is that familiarity makes a face more distinctive.…”

“…Psychometric sigmoid identification functions are found here for unfamiliar face continua but these should not be taken as symptomatic of a true underlying categorical process. As Viviani et al (2007) argue, it is only designs like the X-AX paradigm that can give unambiguous evidence in favour of CP. Thus, for the present data, one might even be reluctant to interpret the modest (but nonsignificant correlations) between the within-pair differences in classification and the performance in the better likeness task as reflecting a tendency towards CP for unfamiliar faces.…”

“…CP has been demonstrated for initially unfamiliar faces (Goldstone, Lippa, & Shiffrin, 2001;McKone, Martini, & Nakayama, 2001;Viviani, Binda, & Borsato, 2007) but training in McKone et al (2001) was extensive (between 1,000 -10,000 trials; 2 -14 hr /participant/face pair) and considerable (15 min of similarity ratings plus 54 trials for each of four faces in a category learning task in Goldstone et al, 2001; thousands of preliminary identification trials on the morphed continua in Viviani et al, 2007). However, there are reports that brief exposure to unfamiliar faces can result in CP (Campanella, Hanoteau, Seron, Joassin, & Bruyer, 2003;Levin & Beale, 2000).…”

What’s in the name? Categorical perception for unfamiliar faces can occur through labeling

“…In both Stevenage (1998) and Levin and Beale (2000), names were allocated to the faces during familiarization. In a more recent study by Viviani et al (2007), endpoint faces were labeled as "Face A" and "Face B" in a study phase that preceded a categorization task. Thus, it remains unclear whether CP resulted from the establishment of a facial representation that is activated independently of access to a person's name (or category label), or whether CP requires a learned association between the facial representation and the label.…”

“…In Experiment 3, the endpoint faces were presented with associated names ("George," "Luke") for participants to study for 1 min before the discrimination task. If CP for faces requires long-term perceptual learning (Beale & Keil, 1995;Stevenage, 1998;Viviani et al, 2007), better discrimination of between-category than of within-category pairs should occur only for familiar face continua in all three experiments. If CP for faces can be established with no preexposure (Campanella et al, 2003), then it should be observed for unfamiliar faces in all three experiments.…”