There are many ways to judge the impact of a book. Although awards and citations are common metrics, whether a book prompts a reader to discuss the work with a wide variety of scholars is another important marker of a book’s reach. In this regard, Avidit Acharya, Matthew Blackwell, and Maya Sen’s Deep Roots has provoked many discussions in my concentric intellectual circles. I probably have had more discrete conversations about this book with colleagues from more disciplines (political science, history, and sociology) than I have had about any other book. I appreciated the opportunity to engage my colleagues, and those discussions inform my comments here.

The main contribution of Deep Roots is its replication and extension of V.O. Key’s main findings in Southern Politics in State and Nation (1949), using causal inference. Building on Key’s work, Acharya, Blackwell, and Sen use counties in slaveholding states as their unit of analysis and find that counties with higher concentrations of enslaved people in 1860 still exhibit politically distinct behavior today. More of their white residents identify as conservative, and the counties boast a higher Republican voting rate in recent presidential elections. White survey respondents in these counties report higher levels of racial resentment and also tend to oppose affirmative action. To be sure, the effects of high concentrations of slavery on some types of black–white inequality lessened after the passage of civil rights legislation in the 1960s. However, the attitudinal measures are contemporary, suggesting that sentiments have been slow to change.

The causal models that the authors present are convincing. They comport with Key’s original findings and the baseline conclusions of numerous African American politics scholars who arrived at those conclusions via process-tracing methods. As such, my concern with this book is not with the findings. I do, however, want to push the authors to go further in their modeling, and I raise concerns about the book’s framing and overall narrative.

The authors frame 1860 as the genesis of their story, using the percentage enslaved in each county that year as the basis for their primary explanatory variable. This makes perfect sense, and the authors use it consistently throughout their various models. However, in chapter 6, they also frame the aftermath of Reconstruction as a critical juncture in Southern political history—another stop on the path dependence road to the South’s continued political distinctiveness. This, too, is a completely reasonable assertion. It raises the question, however, of whether the history of both slavery and the South’s attempt at “redemption” is even more complex than the way in which authors modeled it. It seems to me that the legacy of slavery, compounded by the end of Reconstruction and Southern efforts to codify white supremacy, could be reasonably hypothesized to jointly and independently affect Southern political attitudes today. As such, I suspect that a structural equation model would better capture the relationships the authors are presenting. This method would allow them to assess the direct and indirect effects of the concentration of enslaved persons in a county, in addition to capturing the cumulative effects of post-Reconstruction decisions such as support for state constitutional conventions that, among other things, functionally disenfranchised blacks until the passage of the Voting Rights Act. Despite this modeling concern, I do not question the authors’ findings.

There are places in the narrative, however, where the story they tell about the data muddles the statistical argument. On the first page of the book, they contrast Greenwood, Mississippi, and Asheville, North Carolina to introduce their analytical frame that counties with large and small proportions of enslaved persons in 1860 are different. These two cities are an odd pairing. Yes, the counties that included these cities in 1860 had different concentrations of enslaved persons. However, they are distinct in other important ways (namely, geographic terrain and land arability), which arguably make them an inappropriate matched pair for even a rhetorical comparison. (To be clear, I am not suggesting that these counties were matched in the statistical analysis.) Comparing the county that included Greenwood in 1860 to a county in inland Southeast Georgia, for instance—where the soil is (and presumably was) rich but where there were relatively few slaves in 1860—would have been a more convincing setup that would have burnished the authors’ authoritative voice.

As I mentioned earlier, I am persuaded by the findings, though they may seem intuitive to many expert readers. That the findings are intuitive is not a problem. Replication is a hallmark of science, and that the authors are able to build on Key’s methods and extend his argument is commendable. That said, the more challenging task for any author is answering the “So what?” question. This brings me back to my original concern: Do the authors clearly convey the purpose for replicating Key beyond the obvious methodological innovation? I am not sure they do. Although I value applying causal inference to this question, that should not be the point of this book. The larger purposes are to demonstrate the lingering impact of one of the United States’ original sins on political attitudes and behavior and to challenge skeptical readers who argue that slavery is ancient history and does not affect American politics today. The authors succeed in addressing the former purpose, but they fall short in addressing the latter goal. For instance, in their conclusion, the authors suggest that the takeaway of their findings is that civil rights legislation works, because the 1960s legislation (when it was not challenged by the courts) attenuated some of the structural effects of the legacy of slavery. That is a fair conclusion, but it is incomplete. The more relevant question is, How do we address the problem that attitudes have been slower to change? The authors neither adequately address that issue in their conclusion nor do they suggest prescriptions, which would have been helpful.

In addition, I think this book misses an important opportunity to bridge disciplinary boundaries to engage contemporary debates about race and Southern history and politics. The authors make a conscious and understandable decision not to replicate the work of historians. However, by declaring their work “theoretical and empirical” and rendering the history as “context,” as they do on p. 21, they ignore important historiographical debates with which they need to contend, even as they incorporate classic historical scholarship into their citations. For instance, how would the authors respond to potential pushback from contemporary scholars of Southern history who challenge the notion of Southern exceptionalism?

In short, although this book was stimulating, it raised more questions for me than it answered. That creates opportunities for future work, however. I have no doubt that the authors and their peers will continue to study this foundational problem of US democracy and provide new and important insights.