Re-evaluating split-fovea processing in word recognition: A critical assessment of recent research