Lexical selectivity in visual morphological decomposition

In some preliminary results reported in my dissertation, priming effects are shown to arise for specific affixes: i.e., -able, -ful, im-, -ity, -ment, -ness, -s; but not for others: dis-, in-, -er. These unexpected results suggest a lexically selective procedure of visual morphological decomposition, which is contingent on some property (let it be purely orthographical or, rather, more linguistically abstract) associated to the affix being involved.

Such a claim, if true, would challenge all existing theories of early decomposition, in which all morphemes are predicted to systematically undergo decomposition (for a review, Rastle & Davis 2008). At this stage of research, there seems to be no theoretically-relevant, testable hypothesis behind the distinction between decomposable and non-decomposable morphemes. It is also important to acknowledge that the experiments reported in the dissertation were run online, which have brought about technical issues during data collection (prime duration fluctuations, suboptimal RT reliability) and, as a result, a high amount of noise in the data. The lack of priming for some affixes might have therefore arisen because of technical and methodological issues, rather than theoretically interesting ones. As preliminary step towards further understanding the results reported in the dissertation, we expect to re-run the same experiments in an in-lab environment, which guarantees more control over sources of noise.