A Study on Fuzzy Linguistic Scale

[[abstract]]本文主要目的在發展模糊語意尺度的建構程序，不但可以有效地解決傳統問卷語意轉換值，假設為等距的量化方式所產生的誤差；並可降低模糊問卷填答時的困難度。經由實證研究顯示：本文所發展的模糊語意尺度轉換方式，能反映出人類思維的複雜性及不確定性的現象。其次，由實證發現不同的受訪者，在不同的知覺型態下，會產生不同的語意轉換值，充份表現出人類語言表達的模糊性，而且顯示存在著語意膨脹與貶值的現象，可反映出受訪者內心「一把尺」的寬緊狀態。此外本文亦發現受訪者對於語意措辭的兩極語意值有較趨於一致性的看法，但當語意值愈趨中間其看法就愈紛歧，此特性可有效改善傳統問卷，語意值轉換所造成的衡量結果稀釋問題。[[abstract]]The purpose of this paper is to develop the process of a “fuzzy linguistic scale” to solve the linguistic scale transformation problems generated by the traditional quantitative methods based on Likert scale and semantic scale, and to reduce the difficulties of answering the fuzzy questionnaire. We have the following findings: First, the transformation method of fuzzy linguistic scale we develop in this paper can reflect the complexity and uncertainity of human thinking. Second, we show the phenomenon of semantic inflation and devaluation by providing the subjects’ different cognitive styles and different linguistic variable combinations. This shows the fasten and widen conditions of the scale in the questionnaire takers’ mind. Finally, in answering questionnaires we find something about the most positive scale and the most negative one. The closer to the middle scales the linguistic terms are, the fuzzier they are. Therefore, the problem of dilution of measuring results due to the traditional linguistic value in which every scale interval is equal, can be solved by using the fuzzy linguistic scale