Date of Degree


Document Type


Degree Name





Cecelia Cutler

Committee Members

Matthew Garley

William Haddican

William McClure

Subject Categories

Anthropological Linguistics and Sociolinguistics | Discourse and Text Linguistics | Semantics and Pragmatics


discourse markers, language attitudes, rapid language change, corpus analysis, societal treatment study, matched guise survey


This dissertation investigates a linguistic feature called “backstory so,” defined as discourse marker so when it prefaces the answer to a question or request for information from an interlocutor. The motivation for its investigation is a collection of highly negative internet comments expressing irritation and insulting attitudes toward this use of so and the people who say it, calling them annoying, inarticulate, and condescending, for example. I also examine controversy in the (limited) literature about whether or not this language feature is new.

I therefore first present findings that this use of so is an instance of rapid language change that seemingly did not exist in English before the mid-1990s. This diachronic study also suggests that actual rates of backstory so may be the same across women and men. I further present results of a societal treatment study of the nature of the attitudes that backstory so invokes when collected through un-elicited means, which express explicit attitudes in people’s minds. This societal treatment study explores the layers of stereotypes that backstory so invokes. For example, while many people complained that it is frequently said by scientists and experts, others complained that it is frequently said by women who are young and lacking education and/or intelligence. The comments that express dislike toward it yet attribute it to members of society who hold high levels of social prestige are analyzed within the framework of a lesser-studied area of prescriptivism: People who mock “educated” speech characteristics by way of establishing solidarity and preserving their vernacular. This analysis also uncovers one of the more tangible reasons that people dislike backstory so: a shift from so as a discourse marker that changes the topic to so as a discourse marker that prefaces the response to a question—a speech act that requires (by definition) keeping the topic the same. This shift is disorienting in a way that suggests the experience of a violation of Grice’s Maxims of Relevance and Quantity—a violation that is resolved somewhat when the answerer gets back on track after providing some amount of unrequested backstory.

Finally, I present findings of a matched guise survey that tested these attributes by eliciting them in a way that collected implicit attitudes toward backstory so. These findings suggest that backstory so is perceived negatively across the population, on average, when compared to a control group. This survey analysis presents statistical analyses, including t tests, bivariate correlations, factor reductions, and principal component analyses to confirm that findings are statistically significant. The results suggest that backstory so is perceived negatively in both men and women, particularly when it comes to perceived levels of education and intelligence, but that in men, backstory so is only indexed with masculinity when being masculine is grouped with being condescending, while in women, backstory so is only indexed with femininity when being feminine is grouped with trendiness and with talking “like a Valley Girl.”

This dissertation also includes a review of the literature on discourse markers and on language attitudes, and a brief proposal for future work to (1) collect actual rates of backstory so across demographic categories and (2) investigate whether perceived rates of backstory so differ from actual rates.