I would like to show that the problem of logical constants can be helped by treating the problem of relationship between logic and human reasoning. Thus I will present some parallels between the respective dilemmas and show that choice of a proof-theoretic answer in one case induces an expressivist choice in the other and the other way round, as well. This does not mean that other options are closed, though the two selected ones are thus given a new plausibility. Furthermore, the proof-theoretical demarcations of logical constants can provide missing details into the expressivist story, as they say which constants and why can actually perform the expres- sivist job.