The Impact of Phonological Co-Activation on Written Language Switching

Language switching has mostly been investigated when switching while speaking and not while writing. As a result, written language switching and the factors that may impact it are not well-understood. In a previous study (Roembke et al., 2024), it was shown that written language switching is highly facilitated for translation-equivalent word pairs that are identical orthographically (i.e., homographs: TIGER/TIGER [English/German]), even though they mismatched in phonology. Thus, switching facilitation might be the result of limited phonological co-activation when writing homographs, since phonology constitutes the only difference between the translations. In a first experiment (preregistration: https://doi.org/10.23668/psycharchives.12509), we investigated this hypothesis more directly by manipulating the extent to which a word’s phonology had to be activated during written picture naming. Unbalanced German-English bilinguals switched between naming pictures of homographs and quasi-homographs in their dominant versus secondary language. Participants responded by typing the word, and simultaneously speaking the same word in the corresponding language (type-and-speak), tapping their tongue (type-and-tongue-tap) or doing neither (type-only). As predicted, we found larger switch costs in type-and-speak compared to the other conditions, suggesting that bilingual writers can avoid recruiting language-specific phonology when this interferes with their response and producing it is not requested (i.e., when typing without speaking). In this experiment, we aim to replicate this result when non-cognates are included in the stimuli list, reflecting a more standard setting wherein between-language conflict involves also orthographic representations, and not only phonological ones.