Abstract
It has long been debated whether words are processed serially or in parallel. Recent studies provided supporting evidence for serial processing of English words (White et al., 2018, 2019, 2020). Here we question whether this result can be generalized to the other languages, and whether semantic relation between the words would lead to different findings. Using a similar experimental design in White et al., two semantically related or unrelated Korean words were simultaneously presented in the left and right visual fields. Whereas a word either in the left or right visual field was cued in the single task, both words were cued in the dual task. Participants (n = 12) judged whether the cued words were surely or maybe living, or surely or maybe non-living. Word presentation duration was tailored for each participant to keep all participants’ performance in the single task roughly the same. The area under the receive operating characteristic curve was calculated as a measure of accuracy. Dual task accuracy was lower than single task accuracy, replicating the dual task deficit observed in the previous studies. Further, we used the attention operating characteristics to compare dual task accuracy with specific model predictions (Sperling & Melchner, 1978). If the two words were semantically related, capacity-limited parallel-processing model accounted dual task accuracy better. On the other hand, if the words were semantically irrelevant, accuracy was closer to the prediction of the all-or-none serial-processing model. In addition, accuracy of one word was not associated with accuracy of the other word, meaning that there was no trade-off which probably result from all-or-none serial processing. Our results suggest that semantic relation between the words is the critical factor that determines how (Korean) words are processed.