Mirjam de Jonge, postdoc at Université Côte D’Azur and PhD candidate at the UvA, is the guest speaker at this ACLC seminar on 10 January 2020 from 16:15 until 17:30 (followed by drinks). She will talk about ‘Inflection and derivation in silent word production: neurophysiological evidence’.
Inflection and derivation in silent word production: neurophysiological evidence
Sahin et al. (2009) presented evidence from patients with intra-cranial electrodes showing that lexical access, morpho-syntactic processing, and phonological processing can in principle be separated in time and space based on characteristic electrophysiological responses at 200 ms, 320 ms, and 450 ms respectively. These findings were elicited in a silent pronunciation paradigm where participants were presented visually with a cue followed by a single word, and were instructed to pronounce the word they saw silently in their head according to the preceding cue. The cue instructed the participants to simply read the word they saw, induced a null inflection, or induced an overt inflection of the word.
In a new set of experiments, we aimed to replicate Sahin et al.’s findings in surface EEG with healthy adults producing English plurals and extend the paradigm to English velar softening, a derivational process. The EEG from 20 participants was recorded from 64 scalp electrodes during silent pronunciation tasks involving either pluralisation (as in rock-rocks) or velar softening (as in critic-criticise) in the Overt conditions. In the velar softening experiment, alternations were induced in both real words and nonwords.
Preliminary analyses of the pluralisation study suggest that the basic patterns reported by Sahin et al. appear in surface EEG as well, with different scalp distributions for the effects in the different time windows, supporting the idea that these components indicate different processes. In the velar softening study, an overt derivational change in nonwords elicited stronger responses in the time window associated with phonological processing than real words, suggesting that participants are using a lexical route rather than apply online derivation for at least a subset of the real words.