Although Dr. Devereaux is taking a bit of time away from the more typical blogging topics he usually covers on A Collection of Unmitigated Pedantry, he still discusses books related to his area of specialty:

For this week’s book recommendation, I want to recommend N. Rosenstein, Rome at War: Farms, Families and Death in the Middle Republic (2004). This is something of a variation from my normal recommendations, so I want to lead with a necessary caveat: this book is not a light or easy read. It was written for specialists and expects the reader to do some work to fully understand its arguments. That said, it isn’t written in impenetrable “academese” – indeed, the ideas here are very concrete, dealing with food production, family formation, mortality and military service. But they’re also fairly technical and Rosenstein doesn’t always stop to recap what he has said and draw fully the conclusions he has reached and so a bit of that work is left to the reader.

That said, this is probably in the top ten or so books that have shaped me as a scholar and influenced my own thinking – as attentive readers can no doubt recall seeing this book show up a lot in my footnotes and citations. And much like another book I’ve recommended, Landers, The Field and the Forge: Population, Production and Power in the Pre-Industrial West (2003), this is the sort of book that moves you beyond the generalizations about ancient societies you might get in a more general treatment (“low productivity, high mortality, youth-shifted age profile, etc.”) down to the actual evidence and methods we have to estimate and understand that.

Fundamentally, Rome at War is an exercise in “modeling” – creating (fairly simple) statistical models to simulate things for which we do not have vast amounts of hard data, but for which we can more or less estimate. For instance, we do not have the complete financial records for a statistically significant sample of Roman small farmers; indeed, we do not have such for any Roman small farmers. So instead, Rosenstein begins with some evidence-informed estimates about typical family size and construction and combines them with some equally evidence-informed estimates about the productivity of ancient farms and their size and then “simulates” that household. That sort of approach informs the entire book.

Fundamentally, Rosenstein is seeking to examine the causes of a key Roman political event: the agrarian land-reform program of Tiberius Gracchus in 133, but the road he takes getting there is equally interesting. He begins by demonstrating that based on what we know the issue with the structure of agriculture in Roman Italy was not, strictly speaking “low productivity” so much as inefficient labor allocation (a note you will have seen me come back to a lot): farms too small for the families – as units of labor – which farmed them. That is a very interesting observation generally, but his point in reaching it is to show that this is why Roman can conscript these fellows so aggressively: this is mostly surplus labor so pulling it out of the countryside does not undermine these households (usually). But that pulls a major pillar – that heavy Roman conscription undermined small freeholders in Italy in the Second Century – out of the traditional reading of the land reforms.

Instead, Rosenstein then moves on to modeling Roman military mortality, arguing that, based on what we know, the real problem is that Rome spends the second century winning a lot. As a result, lots of young men who normally might have died in war – certainly in the massive wars of the third century (Pyrrhic and Punic) – survived their military service, but remained surplus to the labor needs of the countryside and thus a strain on their small households. These fellows then started to accumulate. Meanwhile, the nature of the Roman census (self-reported on the honor system) and late second century Roman military service (often unprofitable and dangerous in Spain, but not with the sort of massive armies of the previous centuries which might cause demographically significant losses) meant that more Romans might have been dodging the draft by under-reporting in the census. Which leads to his conclusion: when Tiberius Gracchus looks out, he sees both large numbers of landless Romans accumulating in Rome (and angry) and also falling census rolls for the Roman smallholder class and assumes that the Roman peasantry is being economically devastated by expanding slave estates and his solution is land reform. But what is actually happening is population growth combined with falling census registration, which in turn explains why the land reform program doesn’t produce nearly as much change as you’d expect, despite being more or less implemented.

Those conclusions remain both important and contested. What I think will be more valuable for most readers is instead the path Rosenstein takes to reach them, which walks through so much of the nuts-and-bolts of Roman life: marriage patterns, childbearing patterns, agricultural productivity, military service rates, mortality rates and so on. These are, invariably, estimates built on estimates of estimates and so exist with fairly large “error bars” and uncertainty, but they are, for the most part, the best the evidence will support and serve to put meat on the bones of those standard generalizing descriptions of ancient society.