In the last several decades a number of studies on perceptual learning in
early infancy have suggested that even infants seem to be sensitive to the
way objects move and interact in the world. In order to explain the early
emergence of infants? sensitivity to causal patterns in the world some
psychologists have proposed that core knowledge of objects and causal
relations is innate (Leslie & Keeble 1987, Carey & Spelke, 1994; Keil, 1995;
Spelke et al., 1994). The goal of this paper is to examine the nativist
developmental model by investigating the criteria that a mechanistic model
needs to fulfill if it is to be explanatory. Craver (2006) put forth a number
of such criteria and developed a few very useful distinctions between
explanation sketches and proper mechanistic explanations. By applying these
criteria to the nativist developmental model I aim to show, firstly, that
nativists only partially characterize the phenomenon at stake without giving
us the details of when and under which conditions perception and attention in
early infancy take place. Secondly, nativist start off with a description of
the phenomena to be explained (even if it is only a partial description) but
import into it a particular theory of perception that requires further
empirical evidence and further defense on its own. Furthermore, I argue that
innate knowledge is a good candidate for a filler term (a term that is used
to name the still unknown processes and parts of the mechanism) and is likely
to become redundant. Recent extensive research on early intermodal perception
indicates that the mechanism enabling the perception of regularities and
causal patterns in early infancy is grounded in our neurophysiology. However,
this mechanism is fairly basic and does not involve highly sophisticated
cognitive structures or innate core knowledge. I conclude with a remark that
a closer examination of the mechanisms involved in early perceptual learning
indicates that the nativism/empiricism debate (as usually construed in
developmental literature) is wrong headed and should be closed.
A computational developmental model for specificity and transfer in perceptual learning