Using Agglomerative Hierarchical Cluster analysis to validate Turkish perceptual attributes emerged by a corpus-driven data

Abstract

This study aims to provide validated Turkish perceptual attributes and ponders how a pool of affective quality attributes can be obtained and be representative of the language in the Republic of Türkiye. Based on the historical development of Türkiye, the study also presumes that various lexical words having the same meaning can be achieved when exposed to different socio-cultural contexts. Firstly, to consolidate a corpus, an online questionnaire was prepared and sent to 196 native Turkish bilingual speakers from all around Türkiye. Secondly, twenty-four binaural sound recordings were collected from seven public spaces to be performed in the listening test. In data analysis, bipolar adjective pairs were found by using Spearman's rank correlation coefficient. The result showed that the highest correlations among the sixty-four pairs are mainly on the pleasant-unpleasant continuum. Agglomerative hierarchical cluster analysis was used to cluster the collected adjectives. The result indicated four top-level categories. The terms grouped on the first cluster found their dichotomous on the fourth cluster while maintaining the same relationship in the pleasant-unpleasant continuum.