2.1 Method
Participants. Twenty-five Mandarin native speakers (12 women) residing in Aix-Marseille-Provence metropolis, aged 19 to 34 years (M = 25.8, SD = 4.2), were recruited. They reported living in a Mandarin-speaking country until the age of 18 with an average length of stay in France of 2.5 years (range: 0.5-7 years, SD=2.0). All participants had normal or corrected-to-normal vision, were right-handed (Oldfield, 1971), and none presented hearing impairment or psychiatric disorders. All participants gave written consent prior to participation and received a gift card after the experiment.
Materials. Target words were chosen from Affective Norms for English Words (ANEW) and (Bradly & Lang, 1999), based on mean valence from 1 (negative) to 9 (positive), and arousal scores from 1 (low) to 9 (high). The pool of items was initially selected following the criteria: (1) valence values were greater than 6.5 for positive words, less than 3.5 for negative words and between 4.5-5.5 for neutral words; (2) arousal values were greater than 5 for both positive and negative items and less than 4.5 for neutral words. The set of words did not contain taboo words. All selected items were translated into Mandarin and French, and the printed frequency was obtained from Sinica Corpus for Mandarin and from Lexique 3 for French. A further selection was performed based on the printed frequency of selected items in both languages. Twenty-five items were retained for each of the three emotion categories. All Mandarin words were disyllabic and contained 2 characters. An overlap of character was avoided so that no minimal pairs were included in the pool of words. LME models with Item as a random effect and the fixed effect of Emotion Category (CAT) were performed to examine differences in lexical properties, arousal and valence (cf. Table 1 ). None of the differences in either printed frequency or number of phonemes was significant across categories. Number of strokes did not differ between negative and positive items, whereas both tended to differ from neutral items. Arousal ratings did not differ between negative and positive target words, whereas both differed from neutral words. Valence ratings differed significantly between all 3 categories; negative words were rated lower than both positive and neutral words, and the latter were rated lower than positive words.
Target items were recorded in a professional sound booth and sampled at 48 kHz with a format of 32-bit float during a single session by a female Mandarin native speaker who did not participate in the main experiment. They were subsequently segmented and annotated automatically into individual tracks with SPASS(Bigi, 2021) and manually verified in PRAAT(Boersma et al.). LME models revealed no significant difference in auditory word duration between any of two Emotion categories (cf.Table 1 ). Altogether, there were 75 trials divided into 3 blocks, with an even distribution of Valence category and Mandarin printed frequency across blocks; block order was counterbalanced across participants.
Procedure. Participants were seated 70 cm from the computer monitor in a dimly lit, sound-attenuated room. Each trial began with a fixation cross, concurrent with an auditory target word in Mandarin. Participants were asked to rate each word according to its valence from 1 to 5 on a Likert scale (1: very negative, 2: somewhat negative, 3: neutral, 4: somewhat positive, 5: very positive) manually on an external response box upon the presentation of a visual response prompt. Subsequent to the response, a blink prompt was displayed for 1 second prior to the onset of the next trial. A practice block of 5 trials initiated participants to the task requirements. Participants could take a short break between the blocks.
Electroencephalographic (EEG) activity was recorded continuously from 64 scalp locations using BioSemi Active Two system AD box over frontal, temporal, central, posterior temporal, parietal and occipital areas of the left and right hemispheres and midline. EEG data were sampled online at 512 Hz and impedance for individual electrodes was kept below 20µV. Electrodes were placed beneath and at the outer canthus of the left eye to monitor blinks and horizontal eye-movements, and over the left and right mastoids. Offline analysis was referenced to the electrode placed on the left mastoid. EEG continuous signal was segmented offline starting from 100 msec prior to stimulus onset, which served as reference point for the pre-stimulus baseline correction, to 1100 msec post stimulus onset All channels were filtered offline with a high frequency filter of 30 Hz. Automated routines were applied to exclude trials contaminated by ocular-motor or muscular artifacts.