Why transactions matter in thinking about the language environment

In the past decades, psycholinguistics has convincingly demonstrated that the way humans use and process language is incredibly sensitive to the distribution of language in the environment. Since researchers do not have access to language users’ true environment, they have used large text corpora as a proxy to build measures of the environment. Although the effects of these corpus-based measures appear to be very large and robust, there is empirical and analytical evidence that corpora do not offer a faithful representation of the language environment. I will argue that psycholinguistic research requires a framework for understanding the language environment as a transaction record and I will discuss a set of research questions that arise from this framework.