New use of an old discourse marker: The rise of prefacing answers to questions with so.

This study investigates a previously unresearched use of the discourse marker so : prefacing answers to questions from an interlocutor, informally coined "backstory so " and found to signal that the answer necessitates background information, or more complexity or length than the asker assumes the questioner expects, a function known to be carried out by well. This investigation was motivated by (i) negative attitudes toward this use of so , describing the speakers with attributes like annoying, condescending, confusing, and wrong; (ii) layperson claims that it is new; and (iii) non-scholarly writings by linguists reporting controversy over whether answering questions with so is actually new or a Zwickian Recency Illusion. This paper draws on spoken data from the Corpus of Contemporary American English from 1990 to 2016, and presents findings of 774 target tokens by 544 unique speakers starting in 1992. Results of a logistic regression show a statistically significant increase in the rate of this form over that time. I tentatively suggest that previous constraints against the discourse marker so prefacing answers to mark added information seem to have undergone a rapid language shift, competing with the older use of well , and supporting the layperson intuitions of newness. • Prefacing the answer to a question from an interlocutor with so shows evidence of being an innovation on the rise. • This use of so marks added background information that the answerer assumes the questioner lacks, a function often carried out by well. • Grassroots prescriptivists in the online discourse criticize so -prefacing answers and claim that it has been increasing. • The rise of this form may have started in the early 1990s, which does not support a Zwickian Recency Illusion. • Those laypeople associate so -prefacing answers with expert interviews on NPR , which was also the most common source in the corpus. [ABSTRACT FROM AUTHOR]