The IA Model of Letter Perception

Questions

Can the model account for the word supiority effect?
Can the model account for the advantage of a pseudo-word context?
Can the model account for the finding that letters are sometimes misperceived so as to complete a word?