This paper re-examines theories of linguistic politeness in Japanese, and holds that linguistic politeness is a very complicated issue influenced by multiple factors in different layers including general face wants of participants, the participants’ societal positions and social relationships, social norm that the interactants share, the interactants’ discernment or interpretation of the social rules, immediate context of the interaction, and possible strategies for the interactants to choose under the constrains of the other simultaneously functioning factors. Based on the data collected from recent Japanese TV dramas, this study maintains that, as a general principle, Brown and Levinson’s (1978, 1987) theory of face does apply to Japanese language and culture and forms the base of politeness. Similarly as in any other culture, facework in successful communication in Japanese is a result of choice by an interlocutor in accordance with normative polite practices. What makes linguistic politeness in Japanese unique is not that Japanese speakers need to act appropriately according to their social norm, but that their discernment (wakimae) and recognition of the social position and relationship (tachiba) of the participants, which form the second layer of the determining factors of politeness, make speakers of Japanese always attend to and try to fulfil the other participant’s face want including both positive and negative face, and, at the same time, maintain their own positive face but rarely claim their own negative face especially when an interactant has less power and in a lower social position in an interaction. The data also suggest a model of face-redressing strategies co-occurring with face threatening acts (FTA) in Japanese.