Distal and non-distal NIP theories

We study one way in which stable phenomena can exist in an NIP theory. We start by defining a notion of 'pure instability' that we call 'distality' in which no such phenomenon occurs. O-minimal theories and the p-adics for example are distal. Next, we try to understand what happens when distality fails. Given a type p over a sufficiently saturated model, we extract, in some sense, the stable part of p and define a notion of stable-independence which is implied by non-forking and has bounded weight. As an application, we show that the expansion of a model by traces of externally definable sets from some adequate indiscernible sequence eliminates quantifiers.