That human pragmatic action is heavily dependent on, and interacts with, the situation of utterance, is illustrated by Jacob Mey (J Pragmat 42:2882–2888, 2010) by contrasting human conversation with interactions with artificial intelligence, where the utterances produced are characterized by a lack of situatedness and are thus excellent examples of speech acts that are not pragmatic acts. In the current paper, I also use interactions with artificial communication partners to investigate the effects of different degrees of situatedness on pragmatic actions. In particular, I present an experiment in which people explain the use of an ordinary household item, a table-top lamp, either to their 10-month old children, or to a robot which resembles a young child. The robot’s behavior differs with respect to the degree to which the robot takes the joint interaction situation into account. The qualitative and quantitative analysis of tutors’ utterances shows that the communication partner, but also the degree of the robot’s situatedness, have a considerable impact on people’s verbal behavior. At the same time, tutors also bring in certain preconceptions about human-robot situations. Together, the results support a situated notion of the pragmatic act, the pragmeme.