You are viewing the SNL 2017 Archive Website. For the latest information, see the Current Website.

Poster B45, Wednesday, November 8, 3:00 – 4:15 pm, Harborview and Loch Raven Ballrooms

Morphological processing in Chinese: An ERP study

Lin Chen1, You Li2, Charles Perfetti3;1Sun Yat-sen University, 2South China Normal University, 3University of Pittsburgh

There is a relatively broad consensus that morphologically complex words, such as playground, are decomposed into their constituent morphemes during word recognition. However, researchers disagree about how the morphological processing happens. Some models advocate that morphological processing is initially based on purely orthographic analysis (Lavric et al., 2007, 2012). The alternative hypothesis suggests that morphemic-meaning can be accessed rapidly and constrains initial morphological decomposition (Diependaele et al., 2011; Feldman et al., 2012). The present study addresses this decomposition issue, using ERPs to provide new evidence for morphological processing in Chinese. Twenty-five (13 male) undergraduate students participated the experiment. The stimuli were 116 sets of prime-target pairs. For each set, there are two versions, one for each of two conditions: transparent and opaque. In the transparent condition, the two-character word prime and the one-character target shared the first morpheme and had a semantically transparent relationship (公开-公 [make public-public]). The prime-target pairs in the opaque condition shared the first morpheme but had an semantically opaque relationship (公主-公[princess-public]). There were the two sets of control prime-target pairs in which the two-character word primes were transparent words and opaque words, respectively. The two-word prime was orthographically, morphologically and semantically unrelated to the one-character target. There were 116 filler pairs of which the targets were non-characters to counterbalance the responses. Each trial began with a central 500ms fixation (+). Primes were presented for 40ms followed by a backward mask (#####) for 500ms. The target appeared with offset of the mask, remaining until a response was made. Participants were instructed to make a character or non-character response to the target. There was a main effect of prime type for RTs, F(1,24)=23.21, p<.001. Responses in the morpheme-shared condition was significantly faster than those in the control condition. Paired t test was used to compare the magnitude of priming for transparent items and opaque items. No differences were found t(24)=1.19, p=.24. ERP recordings and data analysis The ERP amplitude analyses focused on time windows centered on P200, N400 and P600. Repeated measures ANOVAS with the factors word transparency (transparent, opaque), region (anterior [FZ, FCZ], central [CZ, CPZ], posterior [PZ, POZ]), and prime type (morpheme-shared, control) were carried out. There were no effects of experimental variables in the 200-300ms time window. There was a main effect of prime type in the 300-500ms time window, F(1,24)=18.75, p<.001. Morpheme-shared condition (4.47) produced a reduced negativity relative to non-repeated condition (2.94). There was an significant interaction between regions and transparency in the 500-800ms time window, F(2,48)=6.51, p<.05. Further analysis showed that in the posterior, the mean amplitude of opaque words (3.55) was significantly more positive than transparent words (2.95), F(1,24)=6.48, p<.05. The results support that, morphological processing in Chinese depends more on orthographic analysis.

Topic Area: Grammar: Morphology

Back to Poster Schedule