The modality switching costs of Chinese–English bilinguals in the processing of L1 and L2

Abstract

Modality switching cost indicates that people’s performance becomes worse when they judge sequential information that is related to different sensory modalities than judging information that is related to the same modality. In this study, we conducted three experiments on proficient and non-proficient bilingual individuals to investigate the modality switching costs in L1 and L2 processing separately. In Experiment 1, materials were L1 and L2 words that were either conceptually related to a visual modality (e.g., light) or related to an auditory modality (e.g., song). The modality switching costs were investigated in a lexical decision task in both L1 and L2. Experiment 2 further explored the modality switching costs while weakening the activation level of the perceptual modality by adding a set of fillers. Experiment 3 used a word-naming task to explore the modality switching effect in language production in L1 and L2. Results of these experiments showed that the modality switching costs appeared in both language comprehension and production in L1 and L2 conditions. The magnitude of the modality switching costs was conditionally modulated by the L2 proficiency level, such as in the L2 condition in Experiment 1 and in both L1 and L2 conditions in Experiment 3. These results suggest that sensorimotor simulation is involved in not only language comprehension but also language production. The sensorimotor simulation that is acquired in L1 can be transferred to L2.

Motor features of abstract verbs determine their representations in the motor system

“…Only the RT for the correct target trials was analyzed. The RT that exceeded the mean by more than three standard deviations was treated as an outlier to ensure the lowest proportion of elimination data ( Marmolejo-Ramos et al, 2015 ) and to be consistent with recent studies on the embodiment of language (e.g., Muraki et al, 2020a ; Zhao et al, 2020 ). The significance criterion for all analyses was set at α = 0.05.…”

“…Embodied cognition theory states that human concepts are represented in the sensorimotor and emotional systems that are engaged by corresponding experiences during concept acquisition, and comprehending words comprises reactivating the same experiences that are responsible for perceiving and acting upon these concepts ( Barsalou, 1999 , 2008 ; Gallese and Lakoff, 2005 ; Pulvermüller et al, 2005 ; Jirak et al, 2010 ; Kiefer and Pulvermuüller, 2012 ; Borghi et al, 2017 ; Buccino et al, 2019 ). While the involvement of motor and somatosensory systems in concrete concept processing is well established in both behavioral and neuroimaging literature ( Hauk et al, 2004 ; Marino et al, 2013 ; Su et al, 2013 ; García et al, 2019 ; Klepp et al, 2019 ; Zappa et al, 2019 ; Dreyer et al, 2020 ; Zhao et al, 2020 ; Togato et al, 2021 ), it is still debated whether abstract concepts are represented in the sensorimotor system. The main challenge for embodiment is to provide convincing explanations and evidence of the embodiment of abstract concepts ( Barsalou, 2008 ; Buccino et al, 2019 ; Ostarek and Huettig, 2019 ).…”

The organization of semantic associations between senses in language

“…Neuroimaging results also support the connections between perception, motor skills, specific sensory brain regions, and linguistic concepts 2,3 . Not only language understanding, but also production, might be affected by sensorimotor stimulation 4 . Likewise, the use of language affects performance on sensorimotor 5 , and even judgment/choice tasks 6 .…”

“…Bold semantic domains had related Wikipedia topics. Semantic domain names and keywords were assigned not only based on central words and related SIL semantic fields, but also in emotional scores and the graphs of domains (figures[2][3][4][5][6][7][8][9]…”

The Mental Lexicon Features of the Hakka-Mandarin Dialect Bilingual

“…We employed linear mixed-effect models to analyze the data because of the strengths of a more reliable analysis of the original dataset and more generalizable findings [ 45 , 46 ]. The analysis was conducted with the lme4 package (version 1.1.27.1, Bates et al) [ 47 ] of R software (version 4.1.2, R core team, Vienna, Austria) [ 39 ].…”