Abstract. Initial-and final-state rescattering, neglected in the parton model, have a profound effect in QCD hard-scattering reactions, predicting single-spin asymmetries, diffractive deep inelastic scattering, diffractive hard hadronic reactions, the breakdown of the Lam Tung relation in DrellYan reactions, and nuclear shadowing and non-universal antishadowing-leading-twist physics not incorporated in the light-front wavefunctions of the target computed in isolation. I also discuss the use of diffraction to materialize the Fock states of a hadronic projectile and test QCD color transparency, and anomalous heavy quark effects. The presence of direct higher-twist processes where a proton is produced in the hard subprocess can explain the large proton-to-pion ratio seen in high centrality heavy ion collisions. I emphasize the importance of distinguishing between static observables such as the probability distributions computed from the square of the light-front wavefunctions versus dynamical observables which include the effects of rescattering.
NOVEL FEATURES OF DIFFRACTIVE DEEP INELASTIC SCATTERINGA remarkable feature of deep inelastic lepton-proton scattering at HERA is that approximately 10% events are diffractive [1,2]: the target proton remains intact, and there is a large rapidity gap between the proton and the other hadrons in the final state. This observation presents a paradox: if one chooses the conventional parton model frame, the virtual photon interacts with a quark constituent with light-cone momentum fraction x = k + /p + = x b j . Furthermore, the gauge link associated with the struck quark (the Wilson line) becomes unity in light-cone gauge A + = 0. Thus the struck "current" quark apparently experiences no final-state interactions. Since the light-front wavefunctions ψ n (x i , k ⊥i ) of a stable hadron are real, it appears impossible to generate the required imaginary phase associated with pomeron exchange, let alone large rapidity gaps. This paradox was resolved by Hoyer, Marchal, Peigne, Sannino and myself [3]. Consider the case where the virtual photon interacts with a strange quark-the ss pair is assumed to be produced in the target by gluon splitting. In the case of Feynman gauge, the struck s quark continues to interact in the final state via gluon exchange as described by the Wilson line. The final-state interactions occur at a light-cone time ∆τ 1/ν shortly after the virtual photon interacts with the struck quark. When one integrates over the nearly-on-shell intermediate state, the amplitude acquires an imaginary part. Thus the rescattering of the quark produces a separated color-singlet ss and an imaginary phase.