Jeff Russell

Regarding procedural knowledge, it makes me think of Michael Polanyi (Karl's brother) and his work on tacit knowledge (Practical Knowledge is a good starting point) - and his famous statement: we know more than we can explain.

Regarding agile, most large established enterprises struggle with it because it is oriented towards building new things from scratch, whereas their essential issue is with modifying existing things (for which existing deep tacit knowledge is needed) and in some cases implementing known requirements with fixed deadlines in the form of government regulations. That and agile has yet to adequately address the accounting aspects - new code is capital expenditure, not revenue expenditure, and so the taxation treatment is different and this affects budgeting significantly for large enterprises that can't just pursue a burn rate on capital expenditure.
(I always find it funny that the original project that gave rise to agile methods (with GM or Ford I think) was a failure.)

One other thing that you might want to look at is Christopher Alexander's work "The Timeless Way of Building" which outlines the thinking which led to his "Pattern Languages". He addresses the observable fact that in Europe barns in a region are all different at the level of detail yet follow a set pattern, despite them traditionally being built by the locals and not professionals. The answer to why that is so probably bears upon what you are angling in on, and has a lot to do with tacit knowledge IMHO.