Functions via everyday actions: Support or obstacle?

In an earlier paper, we have documented a clash between intuitive and analytical thinking concerning functions, which we have termed the changing-the-input phenomenon. The discovery of the changing-the-input phenomenon, however, left us with a puzzle: Why has this phenomenon concerning functions - a purely mathematical concept - been observed in computer science classes but not in mathematics ones? The purpose of the present paper is to address this puzzle. More generally we ask, under what conditions the changing-the-input phenomenon will or will not be manifested? Still more generally, in learning about functions, when is the intuitive scaffolding of functions via actions-on-tangible-objects helpful, and when does it get in the way of deeper understanding?