In order to improve our operationalization of class in sociolinguistic analysis, this paper draws on sociological theory as the foundation for a new approach to the conception and coding of occupation. The 162-speaker dataset is drawn from the larger corpus of sociolinguistic interviews conducted in Raleigh, NC. F1 and F2 measurements for the five front vowels of the SVS were extracted at 25% vowel duration and Lobanov-normalized (1971), and the vowel diagonal (Z2-Z1) was included as the dependent variable in regression analyses. To operationalize a sociologically-based theory of occupations, we implement a five-way distinction between industrial/occupational sectors (Law and Government, Technology and Finance, Interactive Service Work, Care Work, and Blue Collar) based on historical changes in Raleigh’s economy. Net of social and linguistic controls, models show significant differences between groups formerly grouped together as White Collar occupations, attributable to historical embeddedness in the greater Raleigh area.