First of all we would like to thank all the commentators for their valuable contributions. They provide additional evidence for our analysis, but they also challenge aspects of it in interesting ways. Unfortunately, space prevents us from addressing most of the points raised here. As Hodge and Cormier demonstrate, discussions about reported speech quickly touch upon fundamental questions about the role of convention in linguistic description and the contribution of multimodality. Dancygier highlights the importance of finding specific theoretical answers for data we encounter in the realm of reported speech, which cannot be easily treated with familiar notions such as 'family resemblances'. And McGregor shows the exciting prospect that our observations can be taken even further than we had anticipated. We are also grateful for the avenues of debate Maier has opened up between a typological approach and the rich literature in formal semantics on quotation, and we thank Quer for doing the same with the more recent but equally rich literature on sign languages. Much to our delight, there seems to be a rather general (though not complete) agreement among the commentators that reported speech is a relatively coherent phenomenon with distinct properties. However, there is variation in their interpretation of the level of analysis at which this phenomenon should be treated. In this response we would like to briefly address one question that emerged in several commentaries, particularly that of Rumsey, Reesink, Goddard and Wierzbicka, and Güldemann (who even chooses it as the title of his commentary): What, exactly, is syntactic about reported speech? After outlining our answer to this question in Section 1, we sketch how this approach fits within a broader research programme (Section 2), and wrap up with some concluding remarks (Section 3).