This thesis examines the nature and scope of linguistic variation within and beyond the Bantawa language and uses it to investigate Bantawa history. Bantawa-speaking communities are spread west-to-east across 200 km of the Himalayan hills of eastern Nepal and into neighbouring north-eastern India. The patterns of linguistic similarity and difference, and the ways in which they demarcate boundaries between Bantawa communities have been used in the thesis to explore Bantawa linguistic history and to connect this linguistic history to the broader social and political history of the Bantawa speakers. The core of this thesis is the detailed comparison of verbal agreement systems and honorific pronoun systems across the different Bantawa speech varieties, and the reconstruction of ancestral systems argued to have been present in proto-Bantawa. This comparison not only showcases the variation in Bantawa today but also demonstrates that the current linguistic landscape of Bantawa reflects complex patterns of linguistic continuity and change, which have been influenced by politico-linguistic dynamics in the region. The primary methodological framework used in this thesis is the Comparative Method. The method begins with a comparison of synchronic data and then reconstructs the proto-forms, which offers starting points to hypothesise most plausible processes and directions of change. Reconstructing complex systems of the verbal agreement for proto-Bantawa was possible by comparing the forms and discerning which form is an innovation and which is archaic. Some features are specific to certain dialects, which are apparently shared innovations within Bantawa; some other features are shared in higher levels with other sister languages, in which case they are treated as shared innovations at a higher level. Other features appear conservative and they are considered inheritances from early ancestor languages such as proto-Kiranti and perhaps even as far as Proto-Trans-Himalayan. One of the examples is the reconstructed 2nd person prefix *ta-, the affixal position and paradigmatic configuration of which is quite anomalous. What is astonishing is that a very similar form, function, and paradigmatic configuration was attested in the languages within the same language phylum but still distant both in linguistic and geographical respect. This fact suggests that this feature is inherited from the same ancestor language at a very early period. The honorific pronoun system is established by employing plural forms regardless of the actual grammatical number of the person that the speaker refers to. Probably this strategy is motivated by a sociolinguistic reason, i.e. to make the speech polite or to avoid potential offence by singling out the addressee. There is good evidential support that this innovation was induced by the contact with Indo-Aryan speakers who had a rigid caste system intertwined with Hinduism culture. The broad diffusion of this strategy indicates the wide impact of the contact that goes much beyond Bantawa, i.e. many Kiranti languages and even to a higher level. This study discovered that the most plausible motivation for the innovations both in the verbal agreement system and the pronoun system was a sociolinguistic reason, i.e. to make the speech polite or to avoid potential offence. It is also noted further that the majority variation is paradigmatic syncretism and it is the plural form that survives in the process of syncretism. This phenomenon suggests a close cause-effect tie between paradigmatic changes and politeness encoding. Besides the Comparative Method, this study looked at non-linguistic histories as well. In that respect, this thesis is interdisciplinary by nature. It turned out that these non-linguistic histories accord with linguistic history and eventually richly fleshed out our understanding of the history of the Bantawa language.