Ian Hacking was one of the more innovative and adventurous philosophers to take up the philosophy of science as their field of inquiry. The Taming of Chance (1990) is a genuinely fascinating treatment of the subject of the emergence of the idea of populations of events rather than discrete individuals. Together with The Emergence of Probability: A Philosophical Study of Early Ideas about Probability, Induction and Statistical Inference (1975; 2nd ed. 2006), the two books represent a very original contribution to an important aspect of modern ways of thinking: the ways in which the human sciences and the public came to think differently about the nature of social and biological reality.
Hacking's contributions to the history of statistical and probabilistic thinking are particularly valuable for the light they shed light on a crucial moment during which fundamental change in the largest gauge intellectual framework took place -- the shift away from deterministic causation to the idea that phenomena present themselves with a distribution of characteristics.
Determinism was eroded during the nineteenth century and a space was cleared for autonomous laws of chance. The idea of human nature was displaced by a model of normal people with laws of dispersion. These two transformations were parallel and fed into each other. Chance made the world seem less capricious: it was legitimated because it brought order out of chaos. The greater the level of indeterminism in our conception of the world and of people, the higher the expected level of control. (TC, vii)Hacking compares his approach to the emergence of probabilistic thinking to that of Foucault in The Archaeology of Knowledge: as a genealogy of a new intellectual framework. (A mark of Hacking's originality as an analytic philosopher is exactly his readiness to find sources of inspiration in Foucault.) His effort is to capture and document the series of shifts in conceptual system and language through which scientists, philosophers, and ordinary people talked about such things as suicide, criminality, and disease.
This book is a piece of philosophical analysis. Philosophical analysis is the investigation of concepts. Concepts are words in their sites. Their sites are sentences and institutions. I regret that I have said too little about institutions, and too much about sentences and how they are arranged. (7)But the crucial point that Hacking is making is that the concepts and assertions that he studies are drawn from a scientific culture that was in a period of flux; so this kind of philosophical, conceptual, and linguistic analysis can document the shifting of a framework of ideas and ways of thinking.
One important aspect of Hacking's arguments in both books is the point that the emergence of probability and statistics was not solely a development of a field of mathematics; it was a new intellectual creation that involved parsing the social and natural worlds in ways that were very different from medieval and modern frameworks. It involved a shift from thinking of events in the world as being causally determined to thinking of them as emerging from a set of probabilistic laws or regularities. It is at bottom a question of metaphysics not mathematics (4).
Hacking also establishes a point about social constructivism that is fundamental to other parts of his work as well (including The Social Construction of What?): the idea that the act of conceptualizing and measuring is also often the act of constituting a particular slice of human reality. The definitions of mental disorder included in the Diagnostic and Statistical Manual of Mental Disorders, 5th Edition: DSM-5 do not merely describe mental realities; they serve to constitute delineations of populations of people with disorders. Prior to these stipulations, we can make the case that the disorder did not exist. (Human beings were troubled in various ways; but they were not "paranoid schizophrenics" until a set of symptoms and behaviors were parsed in such a way by the profession of psychiatry.) Hacking's account of the emergence of official statistics in the eighteenth and nineteenth centuries emphasizes that this kind of social construction was equally in play in the decision to classify certain deaths as "suicide".
Hacking refers to the object of his study as the emergence of a new "style of reasoning," explicitly paired with the ideas of paradigm, research programme, and themata (6). Here he tips his hat to A. C. Crombie. What he seems to mean by this is a congeries of assumptions about nature and society, system of conceptualization, modes of observation and measurement, and rough expectations about outcomes. (These ideas are sketched in the first chapter of The Taming of Chance.)
One part of the emergence of the new way of thinking derived from the consequences of the counting and measurement that came to be a part of government activity by the eighteenth century. Population size, trade data, mortality rates, disease rates ... all of these topics came in for extensive state scrutiny and investigation. And these sources of quantitative data permitted the formulation of new questions: were city dwellers more or less prone to suicide? Did Protestants or Catholics have higher birth rates? Hacking notes that Leibniz played an important role in trying to make sense of new statistical knowledge about one's society:
Leibniz had a lively interest in statistical questions of all sorts, and pursued an active correspondence on issues of disease, death and population. (18)And Leibniz believed that these questions were crucial to the health of the modern state; in fact, he offered a "white paper" to this effect to Prince Frederick of Prussia in 1700.
It was found that at least some statistical information about human processes -- birth, death, disease -- might reveal statistical laws; so the individual events might reflect chance, but they were subsumed under stable and enduring statistical laws like life tables. Hacking refers to a kind of "statistical fatalism" (126), which maintained that the apparent contingency of individual events conformed to an underlying causal necessity in the aggregate.
By 1830 innumerable regularities about crime and suicide seemed visible to the naked eye. There were 'invariable' laws about their relative frequency by month, by method, by sex, by region, by nation. No one would have imagined such statistical stabilities had it not been for an avalanche of printed and public tables. (73)This is a remarkable work of scholarship -- Hacking has pulled together a very detailed collation of the efforts at statistical measurement of national populations and the scientific reflections that these elicited. It crosses over histories of public data collection, epidemiology, theories of suicide, and mathematical representations of new concepts of the statistics of populations. The scholarship by itself is daunting. But even more important is Hacking's ability to place these developments into an intellectual and philosophical context. The book is a tour-de-force.
(Here is an earlier post on Hacking's appreciation of Thomas Kuhn; link.)