Bradley, L.
University of Freiburg, Germany
Speakers of morphologically rich languages such as Hebrew, Arabic and German have reliably shown priming between both semantically opaque and transparent morphological pairs relative to semantic and form conditions. However, the question of a processing advantage for transparent primes at the earliest stages of processing is still cross-linguistically controversial. In German, a language in which semantic transparency advantages are not typically observed in long-latency overt priming, we instead tested for such advantages in masked priming, following Feldman et al (2015), whose results challenge a semantically blind early segmentation stage with morphologically complex primes.
To answer the question of whether 'form-then-meaning' accounts of lexical processing hold in German, we describe a lexical decision experiment with prime-target pairings at three short SOAs arranged into inflectional (gesagt-SAGEN; genannt-NENNEN), derivational (abholen-HOLEN; verbrechen-BRECHEN) and compound (Notlösung-LÖSEN; Buchhalter-HALTEN) blocks, with SOAs randomized within each block (34 ms, 50 ms, 67 ms). This allows us to determine not only the time course of possible early semantic modulation of priming, but also whether a putative need for stem-based representations in German is motivated by priming of derivational morphology alone, or whether it is supported by analogous effects at the compound level.