The reconstruction of the Proto-Indo-European system of personal pronouns is highly controversial. This dissertation puts forth a scenario that has internal coherence on the level of the proto-language and explains the whole range of attested forms in the individual branches. It is principally concerned with the tonic oblique plural and dual pronouns, which I reconstruct as 1pl. *usd\rm n\!\!\!{\sb{\sb{\sb{\sb\circ}}}}\!susd-me, 2pl. *us-usd\rm u\!\!\!{\sb{\sb{\sb{\sb\\enspace}}}}\!usde and 1du. **usd\rm n\!\!\!{\sb{\sb{\sb{\sb\circ}}}}\!h\sb3usd-me, 2du. *uh3-usd\rm u\!\!\!{\sb{\sb{\sb{\sb\\enspace}}}}\!usde and whose development in Celtic, Germanic, Anatolian, Tocharian, Armenian, Greek, and Indo-Iranian I trace in detail. My major conclusions are that the second-person plural pronoun is *us-usd\rm u\!\!\!{\sb{\sb{\sb{\sb\\enspace}}}}\!usde (and not *us-me vel sim.), that aphaeresis has obscured many otherwise straightforward derivations, and that the dative and instrumental cases play an important role in the creation of pronominal paradigms.