ABSTRACT
Recent data from infant-mounted webcams (Sugden & Moulson, 2012) may
be especially useful in this regard, offering quantitative estimates of how
faces are tracked by infants.
The current model offers one account of how face recognition changes
during infancy as a function of exposure to a predominance of faces
belonging to one race. Besides considering the development of the other-race
effect during this time period, the model may also yield insights into how
own-and other-race face recognition change if experience changes sig-
nificantly over the life-span. The other-race effect is known to be malleable in
children (Anzures et al., 2012; Bar-Haim, Ziv, Lamy, & Hodes, 2006; de
Heering, de Liedekerke, Deboni, & Rossion, 2010; Sangrigoli, Pallier,
Argenti, Ventureyra, & de Schonen, 2005) and adults (Degutis, DeNicola,
Zink, McGlinchey, & Milberg, 2011), and recent results indicate that
chimpanzees’ face recognition abilities for human and chimpanzee faces
undergo an extended developmental trajectory (Dahl, Rasch, Tomonaga, &
Adachi, 2013)*at early ages chimpanzee faces are recognized more effectively than human faces (assuming that this is the dominant category
in the environment; Sugita, 2008), but in later life (and with extended
experience), human faces are ultimately favoured. I do not know of any data
from human observers describing the time course of changes in the other-
race effect over long periods of time, but the data from chimpanzees
indicates that recognition accuracy obeys a fairly simple function. Is this also
an emergent property of learning intra-and extrapersonal variation in the
manner described here? The current model makes it possible to at least
crudely quantify experience by manipulating the amount and composition of
training data, meaning we may examine how it responds to both acute
changes in the visual environment and gradual changes in the diversity of
experience available to the observer.