Given a large model M of some theory T, I will describe a method for lifting well-behaved notions of independence from the theories of substructures of M to a reasonably well-behaved notion of independence in M. (In essence, we take the limit of the independence relations in the substructures.) The motivating example – two-sorted theories of infinite-dimensional vector spaces over an algebraically closed field and with a bilinear form – was worked out by N. Granger in his thesis. I will outline this example before launching into the more general framework.