Linguists have long observed that the sounds of a language change over time. One such sound change involves a process in which two distinct sounds disappear in a language--a process called "merger". For example, the "low back merger" describes the loss of contrast between the vowels in "cot" and "caught" in American English. Some speakers maintain a contrast in these two vowels, while other speakers do not. A standard claim is that mergers can occur within an individual over time. However, while numerous published studies have documented the occurrence of specific mergers and theorized on the causes and course of merger, there have been no published accounts testing this claim and clearly documenting the course of a merger within an individual. This project is a first-ever longitudinal study designed to determine whether "unmerged" speakers show signs of merger over time. The project takes place in suburban Minneapolis/St. Paul, Minnesota, and tests the claim about merger using recorded speech and other data obtained from subjects exposed to two ongoing mergers in that area, the "low back merger" (i.e. "cot" vs. "caught") and "pre-velar raising", where the stressed vowel in a word like "haggle" sounds similar to the one in "Hegel". Both mergers are changes in progress but are not completely established in the Minneapolis/St. Paul metro area. Participants include 13- to 15-year old students at a suburban junior high school, a population beyond the critical acquisition period but also of an age likely to be vulnerable to change. This group has received little attention in the area of sound change. The longitudinal study consists of collecting data at three times during the project period. The data to be collected consists of speakers being recorded saying a list of words, taking a listening task based on artificial vowel stimuli, and being recorded in a 10- to 20-minute semi-structured conversation.

The current study is intended to establish a baseline for understanding early adult speech patterns. It will shed light on the process of sound change and advance our understanding of an important type of sound change in a way that integrates sociolinguistics, historical linguistics, and dialectology. The project is complemented by an outreach component in which university students will construct audiovisual and electronic media displays to introduce high school and junior high students to basic linguistic concepts. Other graduate students will be trained in sociolinguistic research methodologies. The project will add a large volume of publicly-available data to the discussion about how sound change works.

Project Report

The Twin Cities of Minneapolis/St. Paul has been thought to exhibit characteristic features of the Northern Cities Shift: the presence of general /æ/-raising and the relative absence of low back merger (Labov et al. 2006). Recent reports of the greater Twin Cites area, however, have shown an increased incidence of low back merger as well as a higher incidence of raising of /æ/ in conditioned environments, particularly before voiced velars (so that bag sounds more like beg) (Bauer and Parker 2008, Ito 2010, Benson et al. 2011). The presence of the low back merger in the area is thought to result from pressure from dialect regions west of the Twin Cities, and pre-velar raising of the low front vowel is thought to result from pressure from dialect regions to the east, in Western Wisconsin (Benson et al. 2011). Despite observation of the fact of ongoing change in the region, the mechanism of how these apparent changes emerge within an individual remains an ongoing matter of debate. One often-cited proposal predicts that individuals are likely to become merged over time as a result of contact with other merged individuals (Herold 1990). A further prediction states that unconditioned mergers, such as the low back merger, are particularly vulnerable to change, whereas conditioned mergers, such as pre-velar raising, are more resistant to change within the individual (Labov 2007). Neither prediction has been tested against longitudinal data from the same speakers over time. The current study was a longitudinal study of 45 adolescent speakers in suburban Minneapolis, sampled each year over a period of three years, resulting in a dataset of 16,080 vowel tokens. Overall, there is some support for the predictions of change over time: individuals are more likely to show change related to unconditioned mergers than conditioned ones. Our results show a few speakers who had no low back merger at the start of the study did in fact lose the vowel height distinction between the two vowels over time. However, all such speakers preserved the acoustic distinctions between the vowels often associated with lip-rounding, and a social network analysis of all speakers revealed no relationship between friend networks and the likelihood of change over time. In addition, while a number of speakers showed evidence or pre-velar raising at the onset of the study, there was limited evidence for any change in the number of speakers to show pre-velar raising over time, as predicted. We conclude that the height distinction of unconditioned low back vowels in the Twin Cities is corruptible over time, but the low front space, particularly conditioned before velars, is more or less stable. Moreover, if social networks do play a role in change within the individual, simple friend networks do not appear to contribute to such change.

National Science Foundation (NSF)
Division of Behavioral and Cognitive Sciences (BCS)
Standard Grant (Standard)
Application #
Program Officer
William J. Badecker
Project Start
Project End
Budget Start
Budget End
Support Year
Fiscal Year
Total Cost
Indirect Cost
Illinois Institute of Technology
United States
Zip Code