Introduction

1. Hemispheric Lateralisation of the Word Length Effect in Chinese Character Recognition Yu-Ju Chou (sallyc@cogsci.ed.ac.uk), Richard Shillcock (rcs@cogsci.ed.ac.uk)IANC, Division of Informatics, University of Edinburgh, 2 Buccleuch Place, Edinburgh, EH8 9LW UK Abstract:In the last decade, researchers of hemispheric superiority have become increasingly interested in the length effect in word recognition in alphabetic languages. But little has been known about ideographic languages like traditional Chinese. The primary aim of this study is to investigate hemispheric laterality and the word length effect in Chinese script recognition. Different-length words consisting of two-, three- and five-characters were presented unilaterally in a lexical decision task. The results, from 23 Taiwanese subjects supported the word-length effect showing significantly different recognition latencies for the multi-character words of different length, but no significant hemispheric lateralisation. There was a significant interaction between gender and visual field, with males tending to show a right visual field advantage. Introduction Previous studies have demonstrated hemispheric lateralisation of a right visual field (RVF) superiority in recognizing English words (Ellis & Young, 1985).The principle finding was: increasing word length affects the left visual field (LVF) but not the RVF presentations, resulting in a RVF superiority (Ellis & Young, 1985). Chinese, the so-called ideogram, is distinctive from the English writing system and is supposed to present more pictorial characteristics, involving an LVF superiority in recognition tasks, because the right hemisphere is dominant in processing pictorial images. Fang (1994) conducted experiments with different-length Chinese words but failed to find significant interactions between Visual Field and Word Length. This failure is important given the robustness of the effect in English. The methodology of this experiment was taken from Fang (1994), but with an added manipulation of gender, to investigate word recognition in Chinese. • Discussion • In conclusion, hemispheric superiority for word length, as found in English, does not appear to exist in Chinese, from the results of this experiment. • There was, however, a significant interaction between Gender and Visual Field, the slowest reaction latencies coming from males responding to the LVF, supporting the idea that there are indeed visual field differences to be found in the processing of Chinese script, and comparable to those found in reading English orthography. • Complexity of single Chinese characters • Single Chinese characters have richer information than alphabetises because large portion of Chinese characters contain phonetic and semantic radicals, and needs more fixation time before recognition. • Because of their interaction with gender, these visual field effects would seem to involve the lateralisation of phonological versus spatial processing. • The vertical orientation • Speculatively, the principal difference presented in the current experiment is the vertical presentation of the words, compared with the exclusively horizontal presentation found in the relevant English experiments. • Shillcock and Monaghan (1999) used a statistical analysis of the English lexicon to explain the lateralisation found in English (see Figure 4 for details). 荷 包 蛋 The central fixation point, a 4 mm × 4 mm cross, presented for 1000 msec. An unilaterally presented vertical Chinese stimulus, ended by the critical response or ended automatically after 2000 msec. The masking pattern, overlapped Chinese characters, presented for 1000 msec. Figure 1.The sequence of presentation. • Experiment • This study aims to investigate hemispheric lateralisation and the length effect in Chinese script recognition. Various words consisting of two, three and five characters were presented unilaterally and vertically in a lexical decision task. • Subjects • Twelve male and eleven femalenative Mandarin speakers with normal or corrected-to-normal vision participated the study. Only one subject was ambidextrous, the rest were right-handed according to self-report. • Stimuli • The stimuli were different-length vertically displayed Chinese wordsand non-wordsconsisting of2, 3 and 5 Chinese characters, chosen from the Corpus of Journal Chinese (1993). • 20 real-words and 20 non-words in each length category, 120 words in total. • The non-words were made by switching two characters within one word. For example, a three-character word: • 荷包蛋 (he bao dan) was changed to 荷蛋包 (he dan bao). • Each character was13 mm × 13 mm and the inter-character space was 9 mm, adjusted by PhotoShop and presented by Psyscope Version 1.2b5 (1994) and a Macintosh computer. • Each subject saw half of the stimuli presented on the screen 2 mm to the right hand side of the fixation point, and the other half presented to the left hand side of the fixation point. The smallest visual angle was equivalent to 0.25 degree from the fixation point. • Procedure • Subjects were instructed to press the the right button on the response box with their right index finger for real-words, and press the left button with their left index finger for non-words. • Psyscope software recorded response latencies with millisecond precision. The sequence of presentation was as Figure 1. • Analysis and results • An analysis of variance of response latencies was carried out with Visual Field and Word Length as within-subject factors and Gender as a between-subject factor. • A significant main effect was found for Word Length see Figure 2 for details), but was not for Genderor Visual Field (Length F(2,42)=270.832, p< .001; Gender F(1,21)=.429, p> .05; Visual Field F(1,21)=.423, p> .05). • The two-way interaction between Gender and Visual Field was significant (F(1,21)=6.014, p< .05, see Figure 3 for details). Gender differences were not significant either in the RVF or in the LVF. On the other hand, Visual Field difference was marginally significant for Males (p= .086) but not for Females (p= .121). No significant three-way interaction was found either. Figure 4.The information is asymmetrically distributed either side of the splitting points, through which the words are initially divided and hence, the information is transferred and processed through two visual fields ( Shillcock, et al, 2000). • When words are presented vertically, no such straightforward splitting may occur and the lateralisation of the different information is minimised. • We thereafter conducted a survey for the distribution of information in Chinese by using639 2-character and 650 4-character Chinese words and counted the numbers of variety of the first halves and second halves split from the mid point of the words. • The results supported that the information of Chinese was also asymmetrically distributed, the varieties in the first half of four-character words were larger than the second half (p< .05 for 4-character words and p= .104 for two-character words) by Binomial Test. Initial splitting and lateralisation should also occur in the reading process. The lack of significant Visual Field effect may result from the vertical presentation of the stimuli. • We may conclude that there is a qualitative difference between text processing in English and Chinese that prevents the emergence of a significant interaction between word length and hemifield in Chinese. RVF LVF RVF LVF Figure 2. A significant main effect was found only for Word Length, not for Visual Field or their interaction. Response latencies to 2-character words were shorter than those to 3-character words which were in turn shorter than those to 5 character words. Figure 3. The interaction of Visual Field and Gender was significant. The Visual Field difference was marginally significant for Males: words presented in the RVF tended to be better recognized than those in the LVF. • References • Chinese knowledge information processing group (1993). The most frequent nouns in Journal Chinese and their classification: Corpus-based research series no. 1~ 4. Taiwan: Institute of information science Academia Sinica and Institute of History & Philology Academia Sinica. • Fang, S. P., (1994).English word length effects and the Chinese character-word difference: Turth or Myth? Chinese Journal of Psychology, 36(1), 59-80. Acknowledgments I would like to thank my supervisor Dr. Richard Shillcock for his guidance and thank Dr. Sheng-Ping Fang for her generosity in providing the corpus and essential research papers. • Shillcock, R., Ellison, M. and Monaghan, P. (2000). Eye-Fixation Behavior, Lexical Storage, and Visual Word Recognition in a Split Processing Model. Psychological Review, 107(4), 824-851. • Young A.W. & Ellis, A. W. (1985). Different Methods of Lexical access for words presented in the left and right visual hemifields. Brain and Language, 24, 326-358.