This chapter, which is of a theoretical-conceptual rather than an empirical nature, is concerned with the characterization of structural innovations in New English varieties and the question of whether they can be described as transfer phenomena and learner errors. It first gives a review of relevant empirical studies and presents the state of the art in research in second language acquisition on the nature of cross-linguistic influence and the factors that constrain it. Based on this, it subsequently proposes a model of how structural innovations in New English varieties might have emerged. In addition, previous methods of studying these innovations are discussed, and a best-practice methodology for future research is proposed. It is argued that the classification of a structure as either an error or as an innovation depends crucially on the speakers’ and speaker communities’ norm-orientation and attitudes.