A large amount of research demonstrated that adults are able to process numbers in the symbolic as well as in the non-symbolic code. According to Verguts and Fias (2004), there are qualitative differences in the coding of quantity as conveyed by symbolic and non-symbolic formats: symbolic numbers are processed with a place-coded representation whereas non-symbolic numerosities are processed with a summation-coded representation (see also Roggeman, Verguts & Fias, 2006). Moreover, besides their differential coding, symbolic and non-symbolic representations also present differential scaling; the symbolic representation being linearly scaled (Verguts, Fias & Stevens, 2005) and the non-symbolic representation being logarithmically compressed (Dehaene, 1992; Verguts & Fias, 2004).
The aim of our work is to examine how these different features influence numerosity estimation performances.
Schematic representation of (A) the place-coded and linear symbolic representation and (B) the summation-coded and logarithmic non-symbolic representation. The red segment corresponds to the portion of the representation which is activated when the magnitude 40 is processed.
|Study of the under- and over-estimation process|