Pseudo-morphemic structure inhibits, but morphemic structure facilitates, processing of a repeated free morpheme

Five experiments examined whether words with embedded morphemes are automatically morphologically parsed, even when doing so does not reflect the actual morphological structure. We found that the presence of an embedded morpheme in a word affects the subsequent processing of those embedded morphemes and that the effect depends on a mixture of facilitation due to the orthographic overlap and inhibition that depends on whether the target functions morphologically in the prime. Exposure to a word in which the target (e.g. car) does not function as a morpheme (e.g. carpet) made it more difficult (relative to an unrelated prime) to identify that target as a word, whereas exposure to a word in which the target was a productive morpheme (e.g. hogwash) made it easier to process the target (e.g. hog), and that these effects cannot be reduced to semantic, orthographic, phonological, or syllabic overlap.