Just a mark: Diacritic function does not play a role in the early stages of visual word recognition

A very common feature in most writing systems is the presence of diacritics: distinguishing marks that are added for various linguistic reasons. Most models of reading, however, have not yet captured the nature of these marks. Recent priming experiments in several languages have attempted to resolve how diacritical letters are represented in the visual word recognition system. Since the function and appearance of diacritics can change from one language to the other, it is hard to interpret the accumulated evidence. With this in mind, we conducted two masked priming lexical decision experiments in Hungarian, a transparent orthography with a wide use of diacritic vowels that allows for clear-cut manipulations. In the two experiments, we manipulated the presence or absence of the same diacritic (i.e., the acute accent) on two specific sets of letters that behave differently. In Experiment 1, the manipulation changed only the length of vowels, whereas in Experiment 2, it also changed the quality (e.g., a↝/ɒ/ vs. á↝/aː/). In both experiments, we found that primes with an omitted diacritic work just as good as the identity primes (nema→NÉMA = néma→NÉMA [mute]), whereas the addition of a diacritic comes with a cost (mése→MESE > mese→MESE [tale]). This asymmetry favors a purely perceptual account of the very early stages of word recognition, making it blind to the function of diacritics. We suggest that the linguistic functions of diacritics originate at later processing stages.