Over the last two decades statistical learning (SL) has evolved into a key explanatory mechanism underlying the incidental learning of regularities across different domains of cognition, such as language, visual and auditory perception, and memory. Yet SL has mainly been investigated as an independent research area, separated from the primary study of the relevant cognitive domains. The aim of this special issue is to foster a bilateral integration of SL research with cognitive science: not only should domain-relevant evidence about the complexity of real-world input become more tightly integrated into SL research, but non-SL studies should also carefully consider the nature and range of statistical regularities that may affect learning and processing in a given domain. Four papers on reading in this volume demonstrate that such integration can lead to a better understanding of reading, while also revealing the complexity and abundance of different statistical patterns present in printed text. Moving beyond disciplinary boundaries has the promise to broaden the focus of SL research beyond simple artificial patterns, to examine the rich and subtle intricacies of real-world cognition. A final paper on the neurobiological underpinnings of SL and the consolidation of learned statistical regularities further illustrates what might be gained from a better integration of SL and memory research. We conclude by discussing possible directions for taking an integrative approach to SL forward.
- Statistical learning