Abstract
<title>Abstract</title> <p>Emotion analysis is a popular sub-field in Natural Language Processing (NLP), and its application to literary narratives has gained increasing attention in recent years. This paper introduces BhaRaTa (Bh\={a}va \& Rasa Tagging), a novel framework for emotion analysis of Sanskrit literary narratives based on the classical Indian aesthetic theory described in the N\={a}\d{t}ya\'{s}\={a}stra. Focusing on Sanskrit dramas, we explore the applicability of this framework through the annotation of the different bh\={a}vas (physical, mental, emotional, and physiological states) and rasas (aesthetic experience). We conduct a series of experiments and analyses to examine the characteristics of the annotated data and evaluate the effectiveness of the proposed framework for computational emotion analysis. By introducing a theory-driven approach grounded in Indian literary aesthetics, this work lays the foundation for computational emotion analysis of Sanskrit narratives and provides a basis for extending such approaches to other languages.</p>