In my last post, I discussed what I referred to as the endogeneity of historical data. The basic idea here is that the very existence of the historical records that researchers use to generate quantitative data is often a byproduct of the process being studied. To illustrate this concept, I described how the efforts to expand the regulatory power of the South Dakota Board of Railroad Commissioners led to the creation of the licensing that I used to reconstruct the market network defined by the relationship between rail lines, towns, and grain elevator companies. Perhaps a more familiar example, which is no less central to my work on the origins of electoral Populism, is the case of county-level census and voting data.
Focusing on the case of South Dakota, I use county-level census and voting data to examine changes in the correlates of third-party vote share between 1890 and 1896. These data speak not only to electoral politics, but to the expanding reach of the government and the organization of political space, which led to the creation of areal units such as territories, states, counties, minor civil divisions, and so on. It is easy to take this type of historical data for granted. This is no less true for traditional comparative-historical researchers than it is for their quantitative counterparts, both of whom routinely treat areal units as cases. The question is how these cases came to be. Neither their existence nor their physical form could be taken for granted. In the context of western settlement, the creation of states and counties had the potential to generate resources and shape the dynamics of contention. Far from arising naturally, these entities were a product of concreted human action.
This process was driven by the acquisition and distribution of public land by the federal government. To be clear, this land was anything but free. The acquisition of public land went hand in hand with state-sanctioned violence and the forced relocation of the country’s indigenous population. The result was a stratified system of governance in which members of indigenous tribes were effectively organized out of the polity through the creation of the reservation system, while white settlers were incorporated into the polity through the creation of territories and states. The rules governing the political organization of public land can be traced back to the Northwest Ordinance of 1787, which explicitly linked the creation of territories and states to the size of the population in the region. Yet population alone was not enough. The creation of territories and states required congressional action, prompting residents to proactively lobby for political organization.
While movements for political organization had the potential to generate mass support, they were often set in motion by local elites hoping to capitalized on the various resources that came with new territories and states. The most obvious benefit were the political offices. When a territory made the transition to statehood, for example, it went from having a single elected representative in Congress to having one or more congressional representatives, two senators, and an elected governor. In addition to generating political vacancies out of thin air, the creation of territories and states was typically accompanied by the creation of public institutions such as prisons and universities. The ability to create and distribute these types of spatially-specific resources dovetailed naturally with the logic of territorial politics in the American West, which tended to be organized around geographically-based factions. Without a doubt, the biggest of these spatially-specific prizes was the capital. While the capital could be located anywhere, there was a general sense that it should be located in the middle of the territory or state to which it belonged. What this meant was that there was an incentive to not only divide territories, but to divide them in very particular ways.
This was evident in the long of chain of events leading up to the admission of North Dakota and South Dakota as separate states on November 2, 1889. Prior to that point, the area that would eventually become North Dakota and South Dakota was part of a single political entity known simply as Dakota Territory. The push to divide Dakota Territory took different forms at different times, as shown by the figure below, which depicts three of the boundary proposals that emerged in the period between 1871 and 1889. While these were by no means the only proposals put forward during this period, this particular trio of maps succinctly captures the evolution of the division movement over time. These changes were a reflection of the changes in economic and political geography that came with the development of the territory. Development was initially concentrated in southern Dakota, with the town of Yankton serving as the capital, making it the center of political power. Though ultimately unsuccessful, the movement to create Pembina Territory in 1871 began in anticipation of the arrival of the Northern Pacific railroad, which would run westward to Bismarck—the capital of present-day North Dakota. The division movement was born anew a few years later following the discovery of gold in the Black Hills to the west. Faring no better than its northern counterpart, the western incarnation of the Dakota division movement is notable insofar as it introduced the prospect of an east-west divide. This is only hinted at when looking at the boundaries proposed for the Territory of the Black Hills; subsequent proposals in this vein would, however, call for a complete divide between eastern and western Dakota Territory.
In the end, the movement for division and statehood finally went forward under southern leadership. As historian Howard Lamar brilliantly describes, the movement to admit North Dakota and South Dakota as separate states was a defensive maneuver initiated by the Yankton elite in an effort to insulate themselves from the challenges of an increasingly dominant Bismarck faction. Backed by the Northern Pacific and territorial governor Nehemiah Ordway, the Bismarck faction succeeded in relocating the territorial capital from Yankton to Bismarck in 1883. While the capital removal project would eventually hold up in court, its legality of the scheme was questionable at best. The Yankton elite benefitted from the rise of the Populist movement, which took aim at the monopolization of political power and the growing influence of railroad corporations in party politics. This led to a situation in which a movement on the part of local elites could be productively coupled with a mass movement on the part of agricultural producers, in spite of the fact that the two groups were ultimately competing for power within what was, in effect, a one-party system. Once southern Dakota made the transition to statehood in 1889, Populist leaders in the new state were denied access to major political offices, which led to the creation of an independent party in 1890.
Looking at this example, we can see how expansion of the state and fights over the organization of political space contributed to the generation of historical data. The data generating process in question was endogenous in the sense that South Dakota was a product of Populist mobilization. To put it another way, the case of South Dakota was not so much a neutral site of inquiry as it was an actively contested space. The boundaries of the resulting state were consciously manipulated in order to influence the dynamics of political competition going forward. These types of contests were inherent to the expansion of centralized authority, in that the organization of political space was synonymous with the production of political and economic resources. This was true for the creation of counties as well as the creation of territories and states. In the case of counties, the main prize was the county seat. Today, the county seat may not seem like much. For western settlers, on the other hand, the county seat was a source of contention and even violence, as evidenced by the proliferation of “county seat wars” throughout the Midwest and Great Plains. In Dakota Territory, the location of the county seat was determined by county commissioners who were appointed by the governor. The fact the county commissioners were appointed meant that the governor had considerable control over who got the county seat, transforming the organization of counties into a source of patronage.
This style of politics reached new heights under Nehemiah Ordway, who was accused of working with his son to sell county commissionerships to the highest bidder, raising questions about the slew of counties that were created and organized during Ordway’s tenure, which lasted from 1880 to 1884. This was a period of rapid growth, so it is not entirely surprising that the number of counties being organized would increase. There was, however, clear evidence of nefarious dealings in more than one case. It would be interesting to try and quantify the extent to which changes in county organization were attributable to corruption as opposed to growth. Unfortunately, the population boom of early 1880s was followed by an equally swift bust, making it difficult to accurately estimate the interdecennial population using decennial census data. This all having been said, it worth noting that Dakota Territory is singled out in the Atlas County of Historical County Boundaries for the complexity of its counties’ boundaries.
The boundary problem in the Dakotas is an extreme example of a more general challenge in quantitative history, where much of the data refers not to individuals, but to areal units whose boundaries are subject to change over time. These changes are not always apparent when looking at tabular representations of data, but appear in sharp relief once you begin looking at maps of the underlying geographies. I quickly discovered this when looking at data on Minnesota and the Dakotas in the 1890s. In this instance, the number of observations was sufficiently small that the problem could be fixed by hand if needed. But the problem kept coming up, if not in my own work, then in the work of friends and colleagues, many of whom were working with data sets that were, for whatever reason, not so easily fixed. This led to me to look into more general solutions, as I will discuss in my next post. To make things interesting, I’ll turn my attention from the West to the South, where the legacy of slavery continues to have a lasting effect on contemporary politics.