Historically, transverse momentum dependent (TMD) distribution or fragmentation functions have first been suggested to explain the surprisingly large and otherwise puzzling single-spin asymmetries observed in hadronic reactions with transversely polarized protons. These measurements revealed up to 40% asymmetries in proton-proton collisions which persist to high energies. Such single-spin asymmetries are commonly expected to be suppressed at high energy. They indicate new, not yet explored features in nucleon structure: significant spin-orbit coupling in the nucleon associated with quark transverse momentum and the bound state structure of the nucleon. The unexpected experimental findings provoked the development of new concepts for parton distributions over the last two decades with the introduction of TMD distribution and fragmentation functions, which break with the common collinear approximation for the description of hard processes in QCD. The novel TMD distributions, including an additional degree of freedom the transverse momentum of the parton are a key to unravel the intricacies of the intrinsic motion of partons and the possible connection between their orbital motion, their spin and the spin of the nucleon, which cannot be described with standard PDFs. Observables of TMD distributions are spin and/or azimuthal asymmetries measured in proton-proton collisions as well as in semi-inclusive meson production in deep-inelastic lepton-nucleon scattering, where in addition to the scattered lepton at least one hadron is observed in the final state.