SACHHIDANAND SIDRAMAPPA , MALLAMMA V. REDDY,MALLAMMA V. REDDY
DOI: https://doi.org/10.5281/zenodo.17557540The study of Kannada inscriptions has long focused on historical, linguistic, and cultural interpretations. However, the psychological dimensions embedded within these inscriptions have remained largely unexplored. In recent years, the integration of computational techniques, especially sentiment analysis has opened new possibilities for understanding the emotional and psychological undertones reflected in ancient writings. This approach forms a bridge between history, psychology, and digital linguistics, allowing researchers to interpret how rulers, poets, and scribes expressed emotions, intentions, and authority through language. Traditional epigraphy focuses on translation and interpretation, while sentiment analysis enables a deeper layer of psychological exploration. It helps to identify recurring emotional tones and assess how rulers used language to legitimize power, express gratitude, or evoke fear and loyalty among their subjects. This study introduces a computational approach to examine the emotional and psychological expressions found in Kannada inscriptions. To develop a methodological framework that integrates computational analysis with cultural and historical interpretation. The method applies a structured text processing pipeline, including cleaning, tokenization, and lemmatization, followed by TF–IDF feature extraction to represent linguistic patterns numerically. Multiple machine learning algorithms—such as Linear SVC, Logistic Regression, SGD Classifier, K-Nearest Neighbours, Multinomial Naive Bayes, and Random Forest were used to categorize the texts based on sentiment patterns. Among these, linear and ensemble models delivered the most consistent and accurate outcomes. Findings highlight that inscriptional language not only functioned as an administrative or religious tool but also as a vehicle for emotional expression, moral persuasion, and psychological influence in early South Indian society.
