Repetition-based rule learning: is it domain-specific, is it abstract?
The talk will present a series of near-infrared spectroscopy studies with newborns and 6-month-old infants investigating whether and if yes how these populations are able to learn simple identity/repetition-based structures (e.g. ABB: mubaba, penana etc.) from auditory and visual, linguistic and non-linguistic input. First, newborns will be shown to be able to distinguish ABB patterns from unstructured ABC ones, as well as from AAB ones, but not from ABA ones. On the basis of this evidence, it will be discussed whether this learning is based on low-level perceptual biases or more abstract representations. Then, 6-month-old hearing infants never exposed to sign language will be shown to be able to discriminate between AA and AB patterns presented in (nonsense) signs. The pattern of results obtained will be compared to responses to auditory AA vs. AB as well as to AA vs. AB in a non-linguistic visual control. The implications of these results for the domain-specificty of repetition-based rule learning will be discussed.