Not to lose “a physiognomy of a free scientific association”

In 1925, Mykhailo Ptukha (or: Mikhail Ptukha) travelled to Rome to take part in the meeting of the International Statistical Institute (ISI). Since 1919, Ptukha was the head of the Demographic Institute at the Ukrainian Academy of Sciences in Kyiv—the first such specialized research institute in demography in Europe and perhaps in the world. (NB: it does not mean that the demography practiced there was the one we now think of: no transition theory, a broader focus, rather sociological aspirations). Ptukha was trained internationally before the First World War: in St. Petersburg, Berlin, and London, specializing in mathematical statistics and drawing inspiration from the Italian school of statistics and demography, as personified by professor Rodolfo Benini. Visiting Rome in 1925 offered a rare opportunity not only to meet Benini, who was in the organizational committee, in person, but also to listen to a certain Benito Mussolini, since recently Italy’s head of government, deliver an inaugural speech. It was also one of Ptukha’s first, and, it seems, last visits abroad during the interwar period.

In his report about the meeting, before delivering the synopsis of what he noted as the most important debates at the Institute’s meeting, Ptukha took some time and space to elaborate on what the ISI and the new international situation in which it existed were. The need to organize the ISI, his narrative goes, came from the statisticians’ only very logical tendency to cooperate internationally. Since “in no other science are the elements of theory and practice so thoroughly intertwined as in statistics,” statistics must follow the practical life, and maybe even to surpass it and see ahead. And then, of course, statisticians of every country naturally want to know how statistical systems are functioning in other countries. The pre-war organization of statistics as based around national systems definitely had a strong element of competitiveness to it. But besides this, statisticians needed to coordinate their national systems to allow for the comparative perspective, for which they required the uniformity of methods and thus the unification of administrative statistics across countries.

Between the 1860s and 1880s, there were regular international statistical congresses, which faded with time and then were replaced by the ISI in the 1880s. The Institute met more or less regularly until its last meeting before the war, in Vienna in 1913. The next meeting would only happen in ten years, after the war, in Brussels in 1923. In the report, Ptukha stressed that the context was different this time. “Throughout this period (i.e., within the ten years between ISI meetings—D.T.), such tremendous changes in world economy and international relations occurred […] that we now have to deal with a different world social organism.” What was different, in short, was the First World War’s effects on the economy and international cooperation: state interventionism and the rising international governance within the framework of the League of Nations and satellite organizations on labour, migration, food and agriculture, and others.

For Ptukha, because of the rise of international governance which relied on statistical data, there was a temptation for the Statistical Institute to cooperate with them. The allure for the statisticians is clear: the cooperation with international organizations provides a unique opportunity for the statistical specialists to shape national statistics systems, aiming at their “unification and rationalization” (since it is good for comparison and quality data!—the words “unification and rationalization” are percolating this discourse). This is particularly so, Ptukha writes, since it is much more difficult to reform the already existing system than to intervene at the stage of its conception and planning. And so it is, of course, “desirable” for the statisticians “to come closer to the new [administrative] bodies.” This closeness, however, crucially bears a risk of making the statisticians and their professional organization lose “the physiognomy of a free scientific association,” since the international bodies will be prone to “use the respectable scientific organization for their own purposes.” “Having lost its independence and having become an administrative service body of international organizations, the Institute [ISI] will largely lose its authority within the statistical science.” The task for the Institute, says Ptukha, was to “preserve its independent standing, while at the same time secure the Institute’s influence on the particular/specific statistical interventions [meropriiatiia].”

Wait, what? Secure an independent standing but have enough say to persuade the authorities to influence particular statistical measures—but how? What would be the modality of this interaction and the role appropriated by the statisticians themselves?

I argue that this normative statement of Ptukha points to the very core of the dilemma that the statisticians had to face and deal with during that era of massive state expansion, the solution to which might have effectively transformed (read: obliterate) the statistician’s social and intellectual persona itself as it was known before the First World War. And that is: the need to rely on the state for the data collection and measure implementation, on the one had, but also suffer from the strong demand from the state for a particular kind of research.

Statistics as a social movement and a profession

What if while describing the Statistical Institute’ situation, Ptukha was in fact reflecting on (one could say—projecting) his own position as a statistician and demographer in the early decade of the Soviet rule? Indeed, parallels are not too difficult to find: the international situation with expanded governance was mirrored, if not also amplified at the domestic level in Soviet Union during its first decade already. This included the enormously growing bureaucratic structures and the nudge for the statisticians to cooperate with the government to be able to shape the administrative statistics in accordance with the needs of the statistical science.

These developments overlapped with the shifts in the community of statisticians themselves in the late Russian Empire. The old school of statistics, reinforced by the zemstvo movement, was strongly linked to activism and even, after 1905, to parliamentary politics. This model of statistics could be called “activist-associational”, to use Peter Wagner’s term. Prior to 1917, many statisticians were organized largely around the zemstvo movement or under the auspices of the Constitutional-Democratic party. They were part of the broader left-liberal opposition to autocracy. Some famous coordinators of statistics and statisticians themselves were even quite prominent members or even leaders of the Kadet party, such as Alexander Chuprov the younger or Aleksandr Kaufman or Nikolai Chernenkov.

In a parallel, in the early decades of the 20th century, the movement started to professionalize, forming its own society and founding its own professional journal—Statisticheskii vestnik. Part of this process had to do with the introduction of the increasingly esoteric knowledge—the adoption of mathematical statistics and appropriation of the probability theory into statistical operations, as Libby Schweber argued in her book Disciplining Statistics on this transformation in the British case. Statistics was increasingly taught at higher educational institutions, also outside the classical imperial university. The community would now have higher standards for the qualifications of the statisticians. However, one could say that there was still a thin reformist, ameliorative liberal ideology percolating the movement.

Things changed more significantly after the revolutions. With the full-fledged onset of the Soviet rule in most of the post-imperial space, legal political parties and movements were prohibited and gone. The statisticians, at least those who did not emigrate, would still be working in the academia or for governmental bodies at different levels. There would be still some professional self-organization, such as in all-Ukrainian or all-Soviet statistical congresses. Also, some informal peer networks in and around the Central Statistical Offices remained still throughout the 1920s—ruined later during the Stalinist onslaught, as described by Alain Blum and Martine Mespoulet. Professionalization was still underway, and so was also the transfer of the statisticians’ activity from the public space and broader left-liberal coalitions to the closed-door expert service done for the government offices. Now, in order to impact policies, you would not have to publicly campaign but to gain the party officials’ favour.

One needs to say that even those who preferred to stay in the academia must have felt pressure to turn to the government structures for part- or full-time work or even for irregular assignments. At this point, standing closer to the “new bodies” was desirable not only from the administrative perspective, but also from the point of view earning enough and ensuring sufficient quality of life. The very structural socio-economic conditions in academia in the 1920s pushed and nudged social science experts ever closer to the government structures. The academic pay for teaching positions was not high in the first place. An economy with nationalized “commanding heights” and only widely dispersed private capital during the New Economic Policy meant that the government was the sole possible source of funding for these academic institutions. Besides, if someone wanted to get a better pay and provide for the maintenance of their life and that of their family, getting a job in or around the government was the way to do it. Throughout the mid-1920s, after his colleague Aleksandr Volkov became head of the Ukrainian Central Statistical Office, Ptukha started frequenting the governmental meetings of the statistical office and Gosplan in Kharkiv. In his letters to his Berlin-based former informal supervisor Ladislaus von Bortkiewicz, Ptukha admitted (or complained?), for instance, that he wrote his 1925 publication Naseleniie Kievskoi gubernii [The population of Kyiv province] for a living. Same was true regarding the material on the population movement in the nine Ukrainian provinces. In late 1922, Ptukha was also hoping at some point to obtain a consultant position at the Kyiv provincial statistical bureau, where “the pay was not so bad.”1 At the same time, he struggled finding spare time for what he called “purely scientific” work, due to all the teaching and smaller research assignments. This situation, he confessed, depressed him.2

But what was this “pure science” of statistics which Ptukha would allegedly practice if undisturbed by the administrative concerns of the Soviet state?

Statistics, a science pure and applied

This distinction between the “pure sciences” and “practical” or “administrative” purposes (or those serving “the needs of life”) was a salient one in Ptukha’s writings. When construing the first life table for Ukraine in 1921—the model representing the mortality by age throughout society—he defined it as a “purely scientific” task (of studying mortality as such), which, at the same time, could have practical applications in insurance or in the development of measures to fight mortality. In 1927, when Ptukha was invited to become head of the demographic department in the Union’s Central Statistical Office in Moscow, he rejected the offer, commenting this privately to Bortkiewicz that he saw no purpose in becoming “an administrative statistician of a pure type.”

It is not hard to see how the dichotomy between the “purely scientific” and the “practical needs” overlaps with Ptukha’s concerns about the International Statistical Institute’s possible loss of its “physiognomy of a free scientific association.” Behind these two is the model of scientific autonomy, a discourse about scientists—in this case, statisticians—pursuing the science for the science’s sake, the interests of their own discipline, driven by the will to know and discover, which is not (or should not, since it is a normative discourse) be impacted by any political and “purely applied” needs. The term “pure science,” of course, has its own history. However, this is not a time and place to recapitulate the decades-old discussions about the possibility or impossibility of the apolitical science in general. But in the case of statistics—a social science with its instruments intervening directly into the social “fabric”—its political potential appears clear. Moreover, statistics in the Russian Empire, not to say elsewhere, was still quite embedded into left-liberal ameliorative (and largely not radical Bolshevik) political projects and ideologies.

I would posit that Ptukha’s “pure science” was not an inherently apolitical project but a leftover from his pre-Soviet quasi-sociological project of demography that shared many “post-liberal” (in terms of Peter Wagner) concerns with the first sociological projects, particularly those in Italy. The “pure science” discourse was thus functioning as a protective rhetorical shield, which helped Ptukha to function within the early Soviet (pre-Stalinist) academia in Kyiv and cooperate with the government in Kharkiv while keeping intellectual distance from the Bolshevik political project. In the words of Robert N. Proctor, “[v]alue-freedom, i.e., the discourse of the science’s purity is an ideology of science under siege—a defensive reaction to threats to the autonomy of science from political tyrants, religious zealots, secular moralists, government bureaucrats, methodological imperialists, or industrial pragmatists asking that science be servile or righteous or politically correct or practical or profitable. This may help to explain why it is that the ideal of pure science arises, curiously enough, at the very time that science is perceived as vital for the realization of political, military, and industrial objectives.” Part of its “purely scientific” autonomy relied on the increasingly complicated mathematical language which accompanied the construction of life tables and other statistical operations. Claiming some semi-autonomous logic of his science, as well as its alleged neutrality covered under the “purely scientific” language, Ptukha could always wear an independent “expert hat” when facing the government requests. And not, as it could have been a generation before, participate in a debate as a member of a political party or social movement at some parliamentary committee.

To recapitulate, my hypothesis is that, besides disciplinary claims and aspiration to establish statistics / demography as a social science on par with others in academia, the call for “pure science” signalled also a claim to a space for research that the statisticians themselves deemed important—even if those were political themselves.