命名性失语的汉语普通话语料库构建

Construction of Chinese Mandarin corpus with anomic aphasia

目的:以命名性失语症为例,构建符合国际失语库标准的普通话失语症语料数据集。方法:根据国际失语库标准,采集9例命名性失语症患者自发性言语的多媒体样本,采用CLAN软件进行转录、编码和校验,并对所建立的样本数据集进行语言分析。结果:在失语库测试的整体水平,患者产生句子数96.89±22.45,耗时为709.1±283.78s,词类187.2±58.89个,词次563.67±286.15个,名词103±56.39个,动词99.67±44.65个。在单项语言任务水平,这些参数还支持语言任务间比较。在看图说话和命题任务中,名词和动词的产出量呈显著正相关(n=9,r=0.765,P<0.05;n=9,r=0.945,P<0.001);在命题任务中,动词的产出量与AQ值呈显著正相关(n=9,r=0.686,P<0.05)。结论:命名性失语症患者名词和动词的产出量相同,动词更能反映言语改善程度。国际失语库为汉语失语症提供了有效的语料分析工具。

Objective： With the example of anomic aphasia, we construct a set of data from Chinese Mandarin aphasia corpus which meets the criterion of International AphasiaBank. Method： According to the establishing standard of International AphasiaBank, nine persons with anomic aphasia’s spontaneous discourse samples were collected by multi-media. Language analysis was performed after the application of CLAN software with samples transcribed, coded and verified. Result： Exporting the patient’s parameters on the overall level of AphasiaBank tests, total utterances 96.89±22.45, during 709.1±283.78s, freq types187.2±58.89, freq tokens 563.67±286.15, nouns 103±56.39,verbs 99.67±44.65. In a single level of language task, these parameters also support the comparison between language tasks. In the tasks of picture description and story narrative, the number of the nouns and verbs is positively correlated （n=9,r=0.765,P〈0.05;n=9,r=0.945,P〈0.001）. In the task of story narrative, the number of the verbs and aphasia quotient is positively correlated（n=9,r=0.686,P〈0.05）. Conclusion： The number of outputs of nouns and verbs is consistent in anomic aphasia, and verb is better to reflect the degree of speech improvement. AphasiaBank provides an effective tool of corpus analysis for Chinese aphasia.