Emotional Manipulative Language (EML) plays a significant role in exerting influence in both online and offline communication. While recent advances in natural language processing have enabled the automatic detection of manipulation in text (Goglia et al., 2025), less attention has been paid to understanding how manipulation manifests linguistically. In this project we aim at *extracting interpretable linguistic markers and features of EML*, moving beyond detection and toward interpretation. Our goal is to develop a methodological framework for EML interpretation and to provide theoretical insights into the linguistic mechanics of emotional manipulation through a data-driven linguistic analysis. We will extract human-interpretable linguistic markers and features indicative of EML, and categorize them into a taxonomy, useful for practical applications such as the early detection of online harmful behaviors.