"Democrats and Republicans" or "Republicans and Democrats"? Ask Anastasia Smirnova if you want to know
A paper co-authored by Anastasia Smirnova and Rumen Iliev (Ford School of Public Policy) has just appeared in the Journal of Psycholinguistic Research. This innovative paper explores how the syntactic properties of English binomial expressions can be used to probe the relationship between linguistic and social cognition. Below is how Anastasia describes the content of the paper:
"In this paper my colleague and I study how syntax interacts with extra-linguistic factors, such as psychological properties of the speakers. While word order in English is fixed for the most part – we can say "John read the book" but not "Read the book John" – in binomials, i.e. constructions with two conjoined nouns, the order is flexible. Thus, we can say "Democrats and Republicans" as well as "Republicans and Democrats". What principles determine the choice of word order in these constructions? To answer this question, we examined the behavior of binomials in blogs, websites, a corpus of US Senate speeches, and Project Gutenberg. Our results provide strong support for the "Me first" principle: properties that are psychologically closer to the speaker are mentioned first. The data show that the principle holds not only for political orientation, but also for gender, place of origin, religion, race, and consumer preferences."
The bibliographical information of the paper, together with an abstract, is given below.
Iliev, Rumen & Anastasia Smirnova. (2014) Revealing word order: using serial position in binomials to predict properties of the speaker. Journal of Psycholinguistic Research. DOI 10.1007/s10936-014-9341-3.
Three studies test the link between word order in binomials and psychological and demographic characteristics of a speaker. While linguists have already suggested that psychological, cultural and societal factors are important in choosing word order in binomials, the vast majority of relevant research was focused on general factors and on broadly shared cultural conventions. In contrast, in this work we are interested in what word order can tell us about the particular speaker. More specifically, we test the degree to which word order is affected by factors such as gender, race, geographic location, religion, political orientation, and consumer preferences. Using a variety of methodologies and different data sources, we find converging evidence that word order is linked to a broad set of features associated with the speaker. We discuss the theoretical implications of these findings and the potential to use word order as a tool for analyzing large text corpora and data on the web.