基于标准分的跨方言元音系统比较数据规整法  

A Revised Z-score Normalization for Vowel System Comparison

作　　者：凌锋[1] 

机构地区：[1]上海大学

出　　处：《方言》2017年第3期290-301,共12页Dialect

基　　金：国家社科基金项目"汉语方言介音类型实验研究(16BYY150)"的资助

摘　　要：元音数据规整是跨方言元音系统比较的基础。由于前人提出的元音数据规整方案都没有考虑跨系统比较问题,所以无法直接在这一研究领域中使用。我们的研究在学界公认比较理想的Lobanov规整法基础上结合了均值比较法的长处,把标准分值重新转化成实际物理值,有效的避免前人规整法存在的问题。经过实际元音系统数据检验也证明我们的这种改进方法规整效果比较理想。Vowel normalization is elementary for cross-dialectal vowel system comparisons. However, the previous vowel normalization methods were not designed for trans-systemic comparisons and thus could not be applied directly. The paper introduces a new normalization method based on the z-score transformation and mean comparison and converts the standard value to physical one to avoid the problems of previous methods. Our revised method is proved more satisfactory by the very data gathered from the vowel system comparisons.

关 键 词：元音系统 数据规整 方言比较 

分 类 号：H17[语言文字—汉语]