Roger Martin recently diagnosed a kind of complexity that is manufactured by us and largely unaddressed: inter-domain complexity. It comes about as fields of knowledge are segmented into multiple domains, and each domain develops deep algorithmic knowledge and specialized tools that work by ignoring many of the variables actually in play. Martin notes that the difficulty of reintegrating such simplified and divided disciplines is what gives us the feeling, when we look at any large, adaptive system, of being overwhelmed by massive, un-addressable complexity.