From Pluribus to Unum? The Civil War and Imagined Sovereignty in 19th Century America
Research Question: Contestation over the structure and location of final sovereign authority -- the right to make and enforce binding rules -- occupies a central role in political development. Historically, war often settled these debates and resulted in the institutionalization of the victor's vision of sovereignty. Yet sovereign authority requires more than a set of institutions; it ultimately rests on the recognition and acceptance of the governed. How does war shape the popular imagination of sovereignty? Does war promote ideational convergence around the victor's ideals, or does it polarize and harden attachments to competing visions of sovereignty? We examine this question in the context of the United States, a case where the debate over two competing visions of sovereignty culminated in violence during the American Civil War.
Data: We measure imagined sovereignty in the U.S. context by exploiting the grammatical change in which the “United States” shifts from a plural noun to a singular noun. We treat this change in speech as indicative of a change in how Americans understand U.S. sovereignty, from the multiple and equal sovereignties embedded in the several states to the single final sovereignty of the United States as a national entity. We capture this shift using four sources of textual data: two proprietary newspaper datasets, Nineteenth Century Newspapers and America’s Historical Newspapers, and two journals of Congressional speech, the Congressional Globe and its successor the Congressional Record.
Methods: We analyze our data using statistical methods. The analyses using the newspaper data cover the entire 19th century and allow us to observe where and when the singular or plural form of the “United States” is used. The analyses using the Congressional data are limited to the second half of the century, but offer greater analytical depth because we can link singular usage to specific individuals. In turn, that allows us to conduct tests to probe the mechanisms underlying the shift to an imagined national sovereignty.
Challenges: Although the four data sources provide OCR text, we face two challenges. First, our method of identifying the plural or singular form of the “United States” relies on subject-verb agreement. This means that the “United States” must appear as a grammatical subject of a sentence, rather than a grammatical object. Simple keyword searches cannot distinguish between the subject or object forms. Second, because the OCR text is raw (i.e., unprocessed), automated searching must allow for errors such as “United Steves has} not ovvp,e(Lany land.” We developed a set of rules to facilitate automatic searching, in combination with the use of regular expressions, to identify grammatical subject mentions. We also manually examined every statement in the dataset to confirm that it is a grammatical subject mention.
Findings: Our analysis reveals a powerful effect of the Civil War on the popular imagination of sovereignty, but among winning partisans only. Evidence from our newspaper corpus demonstrates that the experience of the war accelerated adoption of the grammatical singular in the North but had no effect on the South. Evidence from our Congressional corpus points to a rapid, discontinuous increase in singular usage among Northern Congressmen. Our ability to link speech to specific individuals and their party identification allows us to tease apart the pathways at work in shaping sovereignty. We find that Northern Republicans are responsible for driving the increase in singular usage; we find no evidence that Northern Democrats, who fought with the North but had a much weaker ideological commitment to the vision of national sovereignty, thought about U.S. sovereignty differently after the war -- a result that suggests support for the norm promotion mechanism.