汉字手写轨迹及特征的自检效应研究

【作者】 吴喜娜；

【导师】 乐竟泓；

【作者基本信息】 华东师范大学 ， 基础心理学， 2004， 硕士

【摘要】 汉字的认知研究是认知心理学、尤其是语言认知研究中的一个热点，已取得了丰硕的成果，但以往汉字认知的研究只是对标准化汉字认知的探讨，而很少或几乎没有任何研究涉及到手写汉字的认知特点，而在笔迹学领域，多数研究志在分析笔迹与人格的关系等应用型研究，也没有作过这一基础的理论工作，这正是本文的一个切入点。 鉴于研究的立意，本研究通过两个实验，就个体如何检测自身手写汉字笔迹的认知特点作一些探索，并将在方法上进行一些新的尝试。实验一是从整字的角度，刺激字在快速呈现的条件下考察了个体对自身手写汉字检测的SOA(刺激呈现时间)效应。本实验采用2(性别)×5的两因素实验设计，自变量为：SOA(50ms、100ms、150ms、200ms、250ms)，因变量为个体对刺激字的判断反应类型。结果表明：在不同SOA条件下，个体的判断没有显著性差异。证实了手写汉字检测的整体效应；但男女的检测率有显著的差异，女性的成绩显著的高于男性。证实了手写汉字检测存在性别差异；不同SOA条件下较高的检测率证实了手写汉字的可识别性，这为笔迹鉴定搭建了理论基石。实验二采用新的研究方法一笔画拆除法，考察了手写汉字检测的基本笔画效应。本实验采用2(性别)×7(笔画拆除类型)的两因素实验设计，自变量为7种笔画拆除类型，除了对基本笔画的拆除之外(六个基本笔画即六种拆除方式)，还有一种就是去掉笔画中的一笔，仍然能成为一个字的这类汉字，称为准原型汉字，将去掉基本笔画之后的汉字称为破坏原型结构的汉字。因变量为个体对刺激字的判断反应类型。结果表明：原型汉字检测率要高于准原型汉字和破坏原型结构的汉字，而准原型汉字和破坏原型结构的汉字检测率没有显著差异。再次证实了汉字检测的整体效应；此外缺少基本笔画对手写汉字的检测在击中率上都有显著性差异，这说明基本笔画对手写汉字检测的贡献率不同。以击中率作为贡献率的指标，基本笔画的检测贡献率分别为：横—0.209，撇—0.185，竖—0.168，捺—0.125，点—0.069，提—0.032。这点证实了手写汉字检测的笔画效应；与实验一中同样发现了性别效应，女性击中率的成绩还是要高于男性。 综合本研究结果，对手写汉字的检测，发现了整体效应，笔画效应和性别效应。手写汉字的高检测率证实了手写汉字的可识别性，且提出来基本笔画对手写汉字检测的贡献率。对笔迹鉴定的理论、实践工作具有指导作用。更多还原

【Abstract】 In this paper we discuss the recognition of Chinese characters handwriting in the light ofcognitive psychology, which is the area almost not concerned in earlier literature. Although cognitive research on Chinese character is one of the focuses in Cognitive Psychology, especially in psycholinguistics.In the first experiment, Chinese high frequency characters handwriting recognition was investigated at 5 different SOAs (the time of stimulate appearance), that is 50ms,100ms,150ms,200ms,250ms. Gender is another variable, response type is dependent variable. The result indicated that (1) there was no significant difference among different SOA levels. It is concluded that gestalt effect was revealed; (2) The correct response rate of girls is significant higher than that of boys, which illustrate gender effect exists in Chinese character handwriting recognition; (3) High correct response rate under different SOA level testify the handwriting can be recognized that is base theory of handwriting identification.New paradigm was adopted to exam stroke effect of Chinese character handwriting recognition in the second experiment, which is called stroke-leaving. The basic stroke is the least structural units of Chinese characters, so basic strokes were leaved to examine stroke effect on Chinese character recognition. There should be six kinds of leaving paradigm, which is heng-leaving, shu-leaving, pei-leaving, na-leaving, dian-leaving, ti-leaving. Besides these, another leaving paradigm was considered, still a word without one stroke, we called it quasi-antitype. It was compared with antitype and leaving-type Handwriting. Gender is also another variable, response type is dependent variable. The experimental results showed that the right recognition rates of antitype character is significantly higher than those of quasi-antitype and leaving-type, and no significant difference between quasi-antitype and leaving-type, which is testified gestalt effect again. We choose the hit rate as recognition contribute rate of basic stoke which is showed in order: heng-0.209, pie-0.185, shu-0.168, na-0.125, dian-0.069, ti-0.032. They will benefit handwriting identification in practice. Gender effect is also found in this experiment.On the whole, the result of both two experiments showed not only gestalt effect but also stroke effect in Chinese characters handwriting recognition. The high correct rate under different SOA levels in first experiment and basic stroke of Chinese character in second experiment testify the theory of handwriting identification, that is handwriting can be recognition and basic strokes contribute to it.更多还原