Formulating Manipulable Argumentation with Intra-/Inter-Agent Preferences