Rational Imitation or Relevance Guided Imitation?

Ildikó Király
Cognitive Psychology Department
Eotvos Lorand University, Budapest, Hungary

Seminal research with a simple paradigm (bending forward and 
lighting up a lamp with the forehead with the hands being free 
or occupied) has confirmed that early imitative learning is a 
selective, non-automatic, and inference-guided process (Gergely, 
Bekkering & Kiraly, 2002). According to the natural pedagogy 
theory, infants’ selection of what to learn from novel actions 
they observe is a) sensitive to their evaluation of the 
rationality of the action in relation to its outcome within 
the situational constraints, and b) sensitive to the presence 
of an ostensive communicative cueing context that serves to 
guide their relevance-reading process.

This view has been challenged by Paulus et al. (2011) who argue 
that imitative learning is non-inferential and mediated by 
motor resonance through automatic matching by the mirror 
neuron-system, since the original paradigm was confounded – the 
free hands served as support for the bending trunk, so this 
target action was easier to perform.

In this lecture I defend the natural pedagogy view of selective 
imitation against this challenge. I present new studies that 
modified Paulus et al.’s procedure in such a way that the motor 
resonance account and the natural pedagogy account generated 
contrastive predictions. Our results provide support for the 
inferential relevance and rationality-sensitive account of 
selective imitation.