Differences of split and non-split architectures emerged from modelling Chinese character pronunciation

The split fovea model, which reflects some aspects of the anatomy of the visual pathways, has successfully addressed several phenomena in visual word recognition (e.g., Shillcock, Ellison & Monaghan, 2000). However, it is still unclear what qualitative processing differences exist between a split architecture and a non-split counterpart. In the current study, we compare the performance of split and non-split architectures in modelling Chinese character pronunciation and show that Chinese left-right structured phonetic compounds create a unique opportunity for understanding the qualitative processing differences between the two possible versions of cognitive architectures.