I thank Alessandro Stanziani and the editors of Slavic Review for bringing the problem of quantitative economic history to a diverse readership. Colleagues divided between departments (economics and history) and by method (positivist and critical) tend not to publish in the same journals or attend the same panels. This is a rare opportunity to share in a discussion. Allow me to use Stanziani’s points as an occasion to go further and to discuss how I have appreciated numbers in some of my research because I am, as he puts it and as I agree, radical.

At issue is a practice of economic history that is skillful but not considered. In a body of work cited by the author, statistics are “mined” with the assumption that they describe one or another reality. The conclusions of these studies are meant to settle one or another argument—ultimately, about the material condition of the peasantry—in a subfield of economic history that is satisfied by the good and abundant statistics.

Good for what and abundant in what? Are numbers a form of truth, which we study in order to answer a question that is self-evident—say, that the measure of social stability is material well-being and rationality, quantified by the expert use of the numbers we have excavated? I agree with Stanziani that this is not the case: numbers were produced with different questions in mind. I add that material well-being does not translate into social stability; Tocqueville taught us this 160 years ago.Footnote ¹ Apply this to the topic at hand. There is greater consensus that peasant material well-being was improving as 1914 approached, at least in the aggregate, but without the insight of a political philosopher or even a thinker, the Russian Revolution that follows is a mystery. Was the Revolution a terrible misunderstanding of the quantifiable reality on the part of peasants who should have read the statistics rather than rebel? Should the village gathering have used its time to find a bigger abacus, adjust the numbers, infer from the available variables, and subject them to regression analysis, at which point peasants would have been convinced that they were, after all, happy? And then there is the matter of the categories being quantified. Did peasants march on noble estates because the Empire had not gone far enough with “economic liberalization”? Did peasants fail to make “rational choices”? Did peasants in Tambov engage in pitched battles with the Cheka in 1920 because they had a burning desire for better “factor markets”? Que dieu nous sauve.

Used in this manner, the numbers in question may be old, but the analysis is not historical. To use Stanziani’s distinction, numbers are not only data but also sources. The numbers were so deeply informed by the political and professional context and the motivations of the statisticians and funding agencies, that they should be appreciated not only with the expertise of a quantitative analyst, but with a historian’s inquisition into the politics of numbers. I stress that this is not a call to disregard numbers, but to use them in savvy and intelligent ways. We should be deeply educated: what did the statistician count and why, which requires reading the surrounding texts rather than uprooting the numbers from the text? We should also be broadly educated in Russia—its literature, its political debate, its philosophy, its painting—in order to read the numbers as the texts that they are, as the aesthetic movements and stabilities that they were meant to express.Footnote ²

There is something annoying about this debate because we were supposed to have learned in secondary school what we are repeating today: numbers are political at their inception, and they are useful because of this, not despite this. High school diploma in hand, one may look at some general critical analyses of statistical thinking that are now twenty years old and available in paperback.Footnote ³ This is not a Russian problem, nor a nineteenth century problem. When the US Census Bureau asks about the racial composition of the US population, one might be tempted to parse those numbers and ensure they are accurate. How many respondents left it blank, and who were they, really, in the racial terms they have avoided? What does multiracial really mean? Should we take each multiracial person and derive fractions of a person? We may adjust the figure of 13.2 percent Black or African-American upward or downward, but we will miss the first and more profound point: the Bureau is proposing that race is a fundamental criterion for identifying that population. It is the statistical equivalent of push-polling, as any statistical investigation will be, and the data invite us to look for ourselves in those numbers. That is why they are published. I learned that I was “white” when I moved to the USA in 1986; before that I thought I was complicated. I had to replace one series of binaries (totally natural ones, like Hellene-barbarian, Orthodox-Frank) with another (white and non-white). The census produces and reproduces a category, and we give life to the category when we read and cite the census and locate our very selves in it. The numbers are an expert counting of what the Bureau counted. Think about it.

By the same token, when Russian statisticians began in the 1870s to quantify the condition of the peasantry, they were part of a recent process that brought a single peasantry into being. It had not always been thus. As late as 1886, the peasant estate was divided into those who were being freed from the nobility, those who were being freed from the crown, and those termed state peasants who were not quite freed from anyone but merged into the larger estate of peasants. Until 1886, each group was counted differently and had different legal obligations and statuses. It was only in 1886 that all peasants were placed in the same category with more or less the same rules of obligation and self-government, owing the same land tax (rather than obrok, barshchina, or poll tax) and the same redemption payments. An entire statistical industry was built on those land holdings and those payments.

Do the resulting data about land holdings and fiscal dues tell us about the condition of the peasantry? No, they tell us how the peasantry was being conceived, as a relationship to land as such, and as a relationship with the state that was held responsible for the material condition of the peasantry. It was a political argument that was carried out with numbers, and the numbers do not make sense without the political argument. Surely, before we quantify the condition of the peasantry, we should ask what a peasant was.

I know, this is the dreaded deconstruction that delays our arrival at “the real point,” which concerns the levels of material abundance and rationality, and which dampens the glee of locating just about anything that was counted, only because it was counted. Well, the statistical creation of a peasantry is “a real point” because the overwhelming weight of our data relates not to the condition of the peasantry as such, but to the creation of the post-Emancipation peasantry and its relationship to power. The bulk of those numbers relate to the practical problem of taxation at the state and zemstvo levels, because both levels of government were mandated to count acres in order to tax those acres. It was the origin of the peasantry in that data that gave us allotment land as the one consistent measurement of tax liability over time and space, rather than the non-allotment land that peasants were buying up and renting all the way to 1917, and rather than the non-farming occupations that were important everywhere to one degree or another. Statisticians were in the process not of statistically observing the peasantry, but of bringing into being this simplified peasantry rather than some other peasantry. The numbers are more interesting and revealing, not less useful, if we keep them rooted in their time and place.

The use of a single criterion over time and space—allotment land—is understandable: Russia did not have enough statisticians to really measure well-being so they turned to readily available stand-ins. The first zemstvo statistician was hired in 1870, but things were so bad that no one even counted the statisticians—until the Interior Ministry launched a search for subversives in 1882 and counted 227 professional and amateur statisticians.Footnote ⁴ Not a good start for quantifying the largest country in the world. Some problems got worse, the more statisticians were hired and got to work. Each unit and level of government followed different methods and addressed different questions, depending on their assessment of what was important and what a given zemstvo mandated. As statisticians added new variables, like forests, pasture, crafts, and wage-work, the data for each territory became that much less comparable with any other territory. How does one compare a birch-bark shoe produced in Tula with a bundle of hay in Penza and a day of work in Kherson? How does one compare an acre of agricultural land in Kursk with an acre of pasture in Vologda?

There is more. Long before the statisticians counted peasants as such, they produced an aggregate peasantry as such. Small samples became the case studies in all the peasantry, and aggregate land funds were correlated with aggregate populations. This was in part for lack of statisticians but also because of the belief that peasants were indeed a collective rather than a sum of individuals or even individual households.Footnote ⁵ Virtue and necessity merged quite easily into a cultural whole, of which statistics were a part. The entire undertaking had a whiff of Physiocratic assumption that wealth derives from land and tangible goods. Stephen Hoch calls this Malthusian, and he is right.Footnote ⁶ The rest of the population was being measured by the new standard of money and income (what derived from property rather than the property itself, what was monetized, and what was in motion and not fixed—in short, a commercial capitalist economy).Footnote ⁷ Peasants had a different statistics that kept them separate and focused on land rather than income, and were collective rather than individualized. Ultimately, peasants became the subject of a different branch of economics altogether, the labor-production school associated with Aleksandr Chaianov. David Darrow tells us that peasants were measured by a standard of “sufficiency,” as per capita grain or calorie needs, meaning that survival was the central question when it came to a peasant.Footnote ⁸ No other population was asked whether it was merely surviving, as if no other question mattered. This is why we do not have a statistical series on the well being of “Russians.”

This is not beside the point; it is the point. We can try to compensate, to be sure. We can mobilize our surrogate criteria, infer, and guess in an educated manner, but we should not rush past the historical point: statistics grew out of social separateness and reproduced social separateness. We also need humility: our guesses will always be poor because the Old Regime was not equipped to count what we think it should have been counting. The entire tax system, which has given us the bulk of our data on peasant well-being, was built on a caste system. This applies to the zemstvo that produced those abundant numbers, because the zemstvo was itself socially segregated into different electoral curiae, with an unelected third element mediating. No amount of recalculation can or should overcome the image of Grigorii Miasoedov’s “Lunchtime at the Zemstvo” (1872), in which peasant zemsvto deputies took their meal outside the zemstvo building alongside the chickens, while the nobility dined inside with silverware on table cloths.

The most common statistical mechanism used at the time to transcend difference was monetary measurement and income. It was that era’s great leveler in that it dissolved all persons, regardless of differences, into rubles. But when the Russian government first estimated “national income” in 1906, it excluded peasants from the exercise.Footnote ⁹ Quantify this: when the personal income tax was introduced in 1916, the provincial offices that were directed to implement it were not sure whether peasants were also “persons” because they had never been studied systematically as incomes, let alone as personal incomes. If we really want to know about the peasant revolutions of 1905–7 and 1917, we would do much better to look at matters of social estrangement and sheer ignorance—reproduced in the data, not overcome by the data.

Our most basic and foundational categories have a history, too, which we should know before counting. In 1900, “national economy” was still a neologism, which is good to know. Should we really write histories of the national economy in, say, 1850, when there was no such concept in Russia and it was only just lifting its head in France (économie nationale) and Germany (Volkswithschaft)? Could Russians imagine a national economy when they assumed the divisions were too great to imagine a nation?Footnote ¹⁰ Or, to use the term that became current in Russia in the 1890s, it is hard to measure “the popular economy” (narodnoe khoziaistvo) when few believed there was a single “people.” “Property” is another example. Statisticians were unsure how to count all “property” and therefore did not, since different populations owned different kinds of property (communal, household, individual, private, state) with vastly different implications (could or could not be sold or mortgaged or redistributed). Each related to a different electoral franchise for the volost', zemstvo, municipal duma, and State Duma, and therefore to a distinct form of political participation.Footnote ¹¹

It is also worth considering what it meant to be an economist in 1900. These were thinkers sooner than mathematicians, and they counted in order to make one or another political argument. They wrote in good prose. An economic proposition for them was a political proposition, which is why the best-known economists of the day were better known as political activists: Petr Struve, Georgii Plekhanov, Vasilii Vorontsov, Mikhail Tugan-Baranovskii, and Sergei Prokopovich, to mention a few. Would their data make sense if we did not appreciate the fact that one was a moderate Marxist turned militant liberal and tended to count individuals, another the founder of the Marxist movement who was looking for classes, the third a populist who believed “the people” were laborers of any kind but not all persons, and so on? They knew full well that the question of the economy was a matter of political morality, so rather than clear away their subjective assessment in order to get at their numbers, we should listen to what they told us: economics were politics through and through.

We do need numbers because they are a language to express ideas and make arguments; read alone they are deracinated. We are better-equipped thanks to the quantitative studies of our colleagues to propose that peasant living standards did rise, as an aggregate. We can be more confident that taxes did not contribute to peasant immiseration, in the aggregate. But the statistical machinery of the Old Regime did not ask the main question about peasant income (and therefore the burden of taxes on peasant households could not be estimated), and the “rich” troves and “veins” of zemstvo statistics are poor in what concerns us. This is not an obstacle: it gives rise to serious interpretation about what was and was not imaginable under the Old Regime, and it tells us what was known and what was not even asked as officials and revolutionaries formulated policy. The implications were not “theoretical” but urgently practical: come 1917, neither tsarist officials nor Bolshevik commissars knew how to tax peasants in the midst of the food crisis, and instead they turned to guesswork, requisition, and collective violence.Footnote ¹²

I like numbers. It is not the main thing I do, but when I study the numbers in context (meaning by reading the surrounding texts), I learn a lot. By counting, but also reading around the numbers, I was able to trace what the Stolypin reforms did accomplish -- basically, spend a lot of money, which we knew, but on things other than the Stolypin land reforms, which we did not know.Footnote ¹³ I learned what it meant to be a homesteader and own “individual” property but not private property (still could not borrow money against the land).Footnote ¹⁴ This property and its owners were for many practical purposes merged back into the peasant estate for administrative, electoral, and tax purposes; so much for their individualism.Footnote ¹⁵ Research into context led to an understanding of what peasants did and did not pay in land taxes by 1913 (a lot less than anyone assumed, myself included).Footnote ¹⁶ Finally, I learned where Russia stood in relation to fiscal reform compared with other countries (not far at all, which required reading about the Britain, France, Germany, Italy, Austria-Hungary, the USA, and Canada).Footnote ¹⁷ To make sense of the numbers and draw these conclusions one needs to be educated in Russia as much as in numbers; to read texts of prose and parliamentary debate, minority opinions, preambles to draft laws, instructions, supplements, and also to read statistical tables as the texts that they are. Those numbers support the text, not vice versa. Numbers are useful when they have historical and cultural significance, beginning with the recent (and still, on some level I hope, absurd) notion that the human condition can be expressed in numbers alone.

Seriously, have some fun and read Evgenii Zamiatin, not to question his abundant math, and not to pose the wrong questions: “That’s very interesting but can you prove that odd numbers are men?” or “Is it just me or didn’t we already know that there’s no final number?” He understood that numbers are an idiom, not a truth, and he confronted his era’s confidence in numbers. A hint: it’s a dystopia.