AbstractThis paper attempts a critique of the notion of ‘dialogue’ in dialogue theory as espoused by Linell, Markova, and others building on Bakhtin’s writings. According to them, human communication, culture, language, and even cognition are dialogical in nature. This implies that these domains work by principles of other-orientation and interaction.In our paper, we reject accepting other-orientation as an a priori condition of every semiotic action. Instead, we claim that in order to be an empirically useful concept for the social sciences, it must be shown if and how observable action is other-oriented. This leads us to the following questions: how can we methodically account for other-orientation of semiotic action? Does other-orientation always imply interaction? Is every human expression oriented towards others? How does the other, as s/he is represented in semiotic action, relate to the properties which the other can be seen to exhibit as indexed by their observable behavior?We study these questions by asking how the orientation towards others becomes evident in different forms of communication. For this concern, we introduce ‘recipient design’, ‘positioning’ and ‘intersubjectivity’ as concepts which allow us to inquire how semiotic action both takes the other into account and, reflexively, shapes him/her as an addressee having certain properties. We then specifically focus on actions and situations in which other-orientation is particularly problematic, such as interactions with children, animals, machines, or communication with unknown recipients via mass media.These borderline cases are scrutinized in order to delineate both limits and constitutive properties of other-orientation. We show that there are varieties of meaningful actions which do not exhibit an orientation towards the other, which do not rest on (the possibility of) interaction with the other or which even disregard what their producer can be taken to know about the other. Available knowledge about the other may be ignored in order to reach interactional goals, e. g. in strategical interactions or for concerns of socialization. If semiotic action is other-orientated, its design depends on how the other is available to and matters for their producer. Other-orientation may build on shared biographical experiences with the other, knowledge about the other as an individual and close attention to their situated conduct. However, other-orientation may also rest on (stereo-)typification with respect to institutional roles or group membership. In any case, others as they are represented in semiotic action can never be just others-as-such, but only others-as-perceived-by-the-actor.We conclude that the strong emphasis which dialogue theories put on other-orientation obscures that other-orientation is neither universal in semiotic action, that it must be distinguished from an interactive relationship, and that the ways in which the other figures in semiotic actions is not homogeneous in any of its most general properties. Instead, there is a huge variation in the ways in which the other can be taken into account. Therefore close scrutiny of how the other precisely figures in a certain kind of semiotic action is needed in order to lend the concept of ‘other-orientation’ empirical substance and a definite sense.