汉字通透性算法以及对结构方式效应的解释

在先前提出的汉字通透性算法中 ,包括了水平通透性和垂直通透性 ,而汉字作为整体 ,其通透性也应该是一个整体 ,所以 ,当前研究对这一算法进行了更正 ,然后运用新的算法对先前实验材料和基本汉字集 (6 72 4 )中的配对上下字和左右字进行通透性分析 ,证明了它的可行性。在实验研究中 ,采用新算法选择实验用字 ,以达到使上下字的通透性大于左右字的通透性或者使它们相等的目的 ,以此为前提观察到了逆转的结构方式效应以及结构方式效应消失。这一实验结果表明 ,结构方式本身并不影响汉字识别 ,影响汉字识别的直接因素是隐藏在结构方式之下的汉字通透性

MEASURE OF STROKE DISTANCE IN CHINESE CHARACTERS AND ITS APPLICATION TO EXPLAIN STRUCTURE EFFECT

In previous studies, we proposed hypothesis that stroke distance accounted for the influences of various Chinese character forms on visual perception of Chinese characters. Stroke distance was measured in horizontal and vertical dimensions and represented with the two quantities: horizontal and vertical distance. In fact, the two quantities might be integrated in some way. A new algorithm aiming to integrate horizontal and vertical distance was proposed in this present paper. When a character was presented on a monitor controlled by a computer, the stroke distance was measured in pixel. The new algorithm could be represented with the following formula: D=(∑d i )/n,in which n referred to the number of two pixels that could be directly connected with a line in horizontal or vertical dimension and d referred to the distance between the two pixels. The new algorithm was advisable for analyzing characters used in previous experiments. It was also applied to analyze all the paired horizontally and vertically structured characters that were selected from the Basic Chinese Character Set (6724 characters). The results showed that, generally, the horizontally structured characters had larger stroke distance than the vertically structured characters regardless of the number of radical. Structure effect refers that people responded to the horizontally structured characters with less time and greater accuracy than the vertically structured characters. In previous studies, we presented the explanation based on stroke distance. The Present study further tested the hypothesis. Two experiments were conducted. In experiment 1, a group of paired horizontally and vertically structured characters with statistically equal stroke distance were selected by applying the new algorithm. Participants in a dimmed room were asked to identify the characters presented on the computer monitor as quickly and accurately as possible. The results showed that the influence of Chinese character structure on visual identification disappeared. In experiment 2, the algorithm was applied to a selected group of paired characters in which the distance of vertically structured characters was larger than that of horizontally structured characters. Under this condition, we found that the structure effect was reversed. Vertically structured characters were easier to identify than horizontally structured characters. The results obtained from these two experiments were compatible with the hypothesis that stroke distance contributed to the structure effect. Meanwhile, the new algorithm for measuring stroke distance of Chinese characters was found to be helpful.