Abstract
Recent evidence has revealed conflicting results regarding the influence of letter transpositions during the recognition of morphologically complex words. While some studies suggest that the disruption of the morpheme boundary through across-boundary transpositions (e.g., darnkess) leads to the absence of masked transposed-letter (TL) priming, other studies have found that TL priming occurs independently of whether or not letters have been transposed across the boundary. We conducted three experiments to test whether the difference between TL- within and TL-across priming is modulated by (a) the transposition of internal versus external letters of the stem (Experiment 1), (b) the overall proportion of affixed trials (Experiment 2), or (c) the relative frequency between prime and target (Experiment 3). The results revealed equal TL-within and TL-across boundary priming across all three experiments, which adds to previous findings suggesting that across-boundary transpositions do not interfere with the recognition of morphologically complex words.
Notes
1 We also included by-subject or by-item random slopes in the effects of stimulus variables in the random structure of the model. However, likelihood ratio tests revealed that the inclusion of random slopes did not significantly improve the model's fit and were excluded from the analysis.
2 As in Experiment 1, the inclusion of by-subject or by-item random slopes did not significantly improve the model's fit and were excluded.
3 By-subject and by-item random slopes were included in the effects of stimulus variables in the random structure of the model. While by-item slopes increased the model's fit, the overall model was very similar to the model without random slopes and did not affect the significance or direction of our results.