Morphological analysis when morphemes are hard to detect: Evidence from infixation and reduplication.

There is substantial psycholinguistic evidence that skilled reading involves morphological analysis, whereby morphologically complex words like teacher are recognised via constituent morphemes {teach}+{-er}. The ability to recognise morphemes is vital as it allows readers to interpret both familiar and unfamiliar words. However, over 50% of the evidence for this claim comes from studies of suffixation in English, where morphemes typically have a relatively consistent surface form across exemplars (clean in cleaner, cleanliness; -ed in walked, jumped). It remains unclear whether morphological analysis is also robust when morphemes lack a consistent surface form. Using the morpheme interference paradigm in Tagalog, we test whether infixes and reduplicants are detected during prelexical processing. Infixes are inserted within a stem and therefore reduce the consistency of its surface realisation (e.g., tawag → t-in-awag 'called'), while reduplicants involve partial or full copying of a stem and therefore lack a consistent surface realisation (e.g., tawag and lundag become tatawag 'will call' and lulundag 'will jump', where ta- and lu- are instances of the same reduplicant morpheme). We found robust morpheme interference effects in both accuracy and RTs. Specifically, nonwords with morphological structure induced significantly lower accuracy and longer RTs than nonwords without. Our findings show skilled readers' acquired sensitivity to these sophisticated morphological structures and motivate refinements to theories of morpheme learning.