Word and pseudoword superiority effects: evidence from a shallow orthography language