Edge-alignment is useful but not necessary for affix activation in visual word recognition: Evidence from infixation

Previous experiments have found that an early stage of visual word recognition involves aform-based decomposition process, where morphologically complex words like teacher arebroken down into morphemes {teach}+{-er}. However, recent models propose that thisdecomposition is constrained by edge-alignment—only stems and affixes at word edges aresubject to decomposition and subsequent activation. In this study, we conducted two maskedaffix priming experiments to examine whether edge-alignment affects morphologicalprocessing, particularly affix activation, using infixation where the affix is inserted in themiddle of the letter-strings. In Experiment 1, we found that having an infix at the left-edge of the prime (inahit ‘shaved’) significantly facilitated recognition of an infix in thetarget, even though it is not edge-aligned (TINAHI ‘sewn’). No reliable priming was observedin the reverse condition, where the prime contained a non-edge-aligned infix and the targetcontained a left-edge infix (tinahi-INAHIT ‘sewn-SHAVED’). These findings suggest thatthe infix may be more easily activated when it appears in the left-edge position in theprime. In Experiment 2, however, we observed significant priming effects of similarmagnitude for word pairs sharing the same infix in both non-edge-aligned positions (i.e., afterthe first consonant in the stem, tinalo-SINALI ‘defeated-JOINED’) and left-edge positions(inamoy-INIPIT ‘smelled-STUCK’). Findings from both experiments indicate that whileedge-alignment may provide a small advantage, it is not a critical factor that determines affixactivation, thereby challenging existing theories that require strict edge-alignment in visualmorphological processing.