Corps de l’article

In "Economic Generalizations or Laws," an early chapter in his Principles of Economics (1890), Alfred Marshall paused to consider "the unity in difference which underlies various uses of... the term Normal." Noting that "a good jockey is not always of exceptional virtue," and that a strong judge and a strong rower had "seldom the same qualities," he concluded that normal things, like good and strong things, still shared something in common: "every use of the term normal implies the predominance of certain tendencies which appear likely to be more or less steadfast and persistent in their action over those which are relatively exceptional and intermittent." As a final example, he turned to the case of illness. Everyone agreed that illness was "an abnormal condition of man," but also that its absence was abnormal if it persisted throughout "a long life" (28). All this was preliminary to his introduction of the concept of a normal price, which was neither always dependent on "the undisturbed action of free competition" nor always "right morally," but rather that which persisted "in the long run... under given conditions" (29-30).

Marshall's use of the term ‘normal,’ bursting at its seams with qualifications, was itself normal (by his own definition) at the end of the nineteenth century. A common understanding of the normal joined together doctors, statisticians, and economists, among other late-Victorian experts, who staked their expertise in part on their ability to discern normal bodies, normal curves, or normal prices among a profusion of pathological specimens or accidental outliers. In this article, I explore what happened to the Victorian conception of the normal as it passed through bodies, curves, and prices—at the abstract planes of probability theory, medicine, and economics, but also in the more practical setting of life insurance, where it was harder to avoid or explain away disjunctions that emerged among these different disciplines. In the process, I illustrate interdisciplinarity as an historical practice as well as an approach to writing history.

Life insurance offices directly employed doctors and actuaries to assess the financial risks associated with human mortality. Doctors assessed risk at the level of the individual insurance candidate, with the goal of forming a pool that would approximate a normal standard of health. Actuaries, in their turn, evolved normal mortality curves out of the resulting aggregate of certifiably healthy lives, on which curves they based the prices that most customers paid in order to guard against the financial risk associated with premature death. Although life offices never employed economists during the nineteenth century (and only occasionally did so in the twentieth), the economic conception of a natural or normal price was far from irrelevant to insurance practice. In the abstract, the "natural price" of Adam Smith was derived from the same Newtonian logic that informed the actuaries' "mortality laws"; in practice, competition among life offices pushed insurance premiums toward equilibria that overlapped, but never wholly corresponded with, the prices that actuaries and doctors recommended. Although not economists themselves, insurance salesmen and branch managers continually kept their directors aware of the many "business" considerations that often qualified the findings of scientific risk management.

In all three cases, life insurance companies insistently focused on normal as opposed to pathological or accidental features of society. In the process, they engaged in what Michel Foucault called governmentality: the diffuse exercise of power by non-state institutions. Governmentality has emerged as a central legacy of Foucault, not least in studies of insurance and society (Heimer; Ericson et al.; Simon). Although the mere act of applying Foucault to another field of study has often been performed under the guise of interdisciplinarity, scholars have rarely applied the concept of governmentality directly to the historical problem of how different disciplines interact and produce results in practical situations.[1] In the case of life insurance, a common concern with the normal (whether it regarded bodies, curves, or prices) provided what Marshall called a "unity in difference" (29) that helped keep the division of labor within the industry from becoming dysfunctional. The differences mattered, though, since these required a constant negotiation of what qualified as normal in the first place. Interdisciplinarity consequently altered disciplines, at least to the extent that these descended from the level of theory to that of practice. At the same time, the dialogue and resulting compromises it engendered enabled insurance companies to wield sufficient power over their customers to succeed financially.

If forging a common conception of the normal enabled life insurance offices to find unity in difference, this process was further aided by the fact that their line of work did not require substantial agreement about that which was pathological. Once they had settled on who was insurable, they could exclude the rest—leaving the uninsured at the disposal of other instruments of governmentality. In large part, the people whom life insurers excluded (including consumptives, the insane, slum dwellers, and criminals) were the most interesting from the perspective of late-Victorian doctors and statisticians, who spent much of their time identifying pathologies and trying to do something about them. Although these experts still distinguished normal from pathological lives when they worked for an insurance company, they needed to put their therapeutic impulses on hold in that setting. The primary thrust of clinical medicine, with its ceaseless focus on deformity and disease, had little place in life insurance beyond identifying what was to be avoided. Neither did eugenics and social hygiene figure in any significant way in late-Victorian life insurance, much to the annoyance of would-be improvers who sought well-capitalized allies in the insurance industry.

This absence of a pervasive interest in the pathological is exactly why life insurance is such a useful case study of late-Victorian interdisciplinarity. As long as doctors or statisticians focused on the pathological, they were free to divide themselves into the proliferating range of specializations that marked the late-Victorian human sciences. Psychologists could argue with each other about what qualified as mental illness without having much to do with criminologists, who were busy arguing with each other about what constituted crime. Even projects that were avowedly interdisciplinary tended to reduce collections of pathologies to a single cause and argue with each other about its primacy: nature (for eugenicists) or nurture (for social hygienists). Because these projects made strong claims to discursive autonomy—identifying a perfectionist norm and seeking to alter society so as to attain it—they expanded their view of the pathological to include everything that did not accord with that norm (Rabinow 13). Many scholars who see the Victorian era through the lens of governmentality have focused on those parts of society where disciplines or interdisciplinary projects claimed to have the most autonomy, and have identified as a leading feature of modernity what Georges Canguilhem called the "enormous... problem of pathological structures and behaviors in man" (33). As a result, they come away from their research with the conviction that a primary legacy of modernity was fragmentation, which they present as an inescapable outcome of warring universalisms.[2]

Where theorists like Rabinow and Canguilhem focus on the discursive side of discipline formation, a more conventional style of sociology and social history viewed the Victorian era in the comparable terms of "professionalization." Here, the "discipline" in discipline-formation emerges as a form of border patrol, having as much to do with the self-regulation of experts as with the regulation of pathologies in nature or society. Although disciplinarity is a post-Victorian coinage, its basic tenets were far from foreign to Victorian doctors who guarded their professional status against the incursion of homeopaths and patent medicine salesmen (Peterson), or to the Cambridge wranglers whose rigorous training for the Tripos helped shape the scientific field of mathematical physics (Warwick). Representative of this way of thinking in the older sociology literature is Andrew Abbott's claim that "control of knowledge and its application means dominating outsiders who attack that control" (2). Timothy Lenoir, an historian of science who joins the sociological tradition to Foucault's focus on governmentality, defines "disciplinary programs" (mainly drawing on nineteenth-century examples) as "dynamic structures for assembling, channeling, and replicating the social and technical practices essential to the functioning of the political economy and the system of power relations that actualize it" (72). If this definition is taken to include a concern in disciplinary thinking with "facilitating links with other disciplines" (79), one channel for enabling these links has frequently been a common-enough conception of what counts as "normal" or "natural" in a field of knowledge.[3]

Without belittling the reality of the fragmentation that accompanied the rise of modernity, it is worth pointing out that most Victorians, by definition, did not inhabit bodies that their contemporaries defined as pathological. Nor, for that matter, did most Victorians (even self-avowed "experts") spend most of their time defining other Victorians in this way. Rather, most people thought of themselves, and were thought of by others, as normal. Discovering how they did so, and the impact this had on disciplinary thinking, requires looking in a different direction than most scholars of Victorian governmentality have hitherto done. To the Victorianist leafing through the pages of the British Medical Journal or the Transactions of the National Association for the Promotion of Social Sciences, or even (for that matter) a Dickens novel or an issue of the Times, what was deemed at the time to be pathological might seem to dominate. Leafing through an insurance company's policy register or a bank's deposit ledger tells a different story—less newsworthy at the time, perhaps, but also more revealing about life as most Victorians lived it.

I. Normal Bodies

The second chapter of Edward Sieveking's The Medical Adviser in Life Assurance (1874) was entitled "The Normal Man" (18). In it, Sieveking (who was the chief medical advisor for the Briton Medical and General Life Assurance Society) claimed that for an insurance candidate to qualify as being in "good health," it was necessary that he "possesses no hereditary taint, and that his organs and functions are in a condition to enable him to bear the wear and tear of life without unnecessary detriment"—that he should, in short, "approach as nearly as possible to the ideal or typical man" (19). Other insurance doctors fleshed out this prescription for a "normal man." For A.P. Stewart of the Scottish Widows' Fund, "a model life" (37) should possess "a capacious and well-arched chest... which sounds everywhere clear when struck with the finger, and expands equably in every part" (38). He went on to expand this description, equably in every part:

his looks should correspond with the age stated, his limbs and body should be plump, but firm and muscular, and free from palpable traces of injury or deformity, his figure should be erect, well-formed and well-proportioned, his gait steady and easy, his step elastic, his eye bright, his complexion clear and fresh—neither pale nor highly florid, above all neither sallow nor subject to deep and sudden flushings ... the head, which must not be prematurely bald or grey, may be conspicuously large, ... [if] it does not strike the eye as out of proportion to the neck and chest.

38

These perfectionist descriptions of health emerged alongside Victorian doctors' expansive definitions of disease, which encompassed "the systematic destruction of bodies, physically, socially, [and] spiritually" (Hamlin 41). For the middle- and upper-class patients who formed the customer base for insurance companies, warding off disease consequently required regimens of diet and exercise that included "how to govern one's passions, choose a career, or raise children" (Hamlin 68). More generally, health for Victorians resulted in bodies that were "purposively functional," that is, "not just working but working usefully" (Haley 20). What this meant in practice, as recorded in a mid-Victorian "Sanitary Catechism” was "being able to do a good day's work easily" (J. Bridges 142). Yet even this straightforward definition, which apparently accorded so well with the labor theory of value, went well beyond mere productive activity. It required, in the words of the Comtist medical inspector John Henry Bridges, "the greatest economy of vital energies... [and] the minimum of loss, resulting from antagonism of functions, and from degradation of the higher into the lower forms of force" (142).

Such notions of health had little place in the sort of work that doctors were paid to do for insurance companies, which required them to walk a line between protecting their employers from high-risk lives and turning away customers that another company would gladly insure. Head-office medical advisors like Sieveking, who typically combined a busy hospital practice with a wealth of high-paying aristocratic patients, were secure enough in their professional status to bend their perfectionist tendencies to accommodate companies that might pay them up to £400 a year to screen lives (Dupree). It was harder for these doctors to convince their junior colleagues in the provinces, who examined lives at a guinea per candidate, that (as one medical advisor urged in 1905) "every person should be allowed the privilege of insurance unless good cause to the contrary can be shown" (R. Smith, 14). The actuary Arthur Bailey observed that "the younger practitioners exercise all their skill—their knowledge recently brought from the college and hospital—to write elaborate reports, showing why you should reject this case on rather trivial grounds" (Walford, "Position" 170). The Law Union and Crown's medical advisor agreed, noting that "[y]oung pathologists are apt to be too fanciful in excluding men—they do not take fair risks" (Poore 46).

When these doctors recognized the need to descend from the "normal" life (as defined by the clinician or health inspector) to the insurable life, they indicated their awareness that life insurance as a business hinged on accepting risks—and that their job was to distinguish between good and bad risks, not to eliminate risk altogether. A.P. Stewart was aware that his list of attributes for the "model life" might be a good target for the aspiring clinician, but was overkill from a gate-keeping perspective. He followed his list with a rhetorical question and answer: "If this be a sketch of your model life, many will be apt to exclaim, who can expect to be placed in the first class? Few indeed, if this standard were rigorously adhered to by Life Offices" (39). He added that "the allowances made for deviations from the strict standard of health are not only ample, but very liberal"—including, by his reckoning, a pulse five to ten beats higher than normal, childhood pleurisy, and occasional "acidity and flatulence" (40). Later insurance guides noted the need "in practice to make certain recognised exceptions... to the necessary standard" (Pollock and Chisholm 15) and deemed lives with "some slight departure from the normal in the physical condition" to be "acceptable at ordinary rates as a fair business risk" (Brockbank 8).

It was one thing for doctors to shift their assumptions from "normal" to "normal enough" in order to accommodate life insurance companies' conceptions of acceptable risk. It was another thing for them to embrace the classificatory thinking that deeply informed the actuaries who managed these companies.[4] A central basis of life insurance (as will be discussed in more detail below) rested on the law of large numbers, which asserted a great deal of confidence about predictions concerning people en masse but remained staunchly agnostic about predicting outcomes at the level of the individual. In one sense, insurance companies made a major exception to this rule when they invited doctors to examine individual lives and predict whether or not each candidate qualified as a good or bad risk. But as insurance doctors quickly discovered, they could not opt out of thinking in terms of classes of people just because they examined lives one candidate at a time. It was in this sense that their disciplinary tendency to focus on singularity—whether defined as a pathological condition that was not exactly like any other, or a "normal" standard of health that no human being precisely achieved—was ill-suited to their insurance work.

This was less of a problem before 1850, when most insurance companies either accepted lives at ordinary rates or rejected them outright. After that time, owing to competition from a handful of firms that specialized in insuring higher-risk lives, most companies increased the proportion of candidates who qualified as "under-average" but still insurable at higher premiums.[5] Once this occurred, many life offices began providing doctors with a list of three or four classes into which to place each candidate, always including "unexceptionable" or "first-class" for completely healthy proposers and "inferior" for rejected cases (Chisholm 411; Low 117; Purdon 3). Between these extremes, companies distinguished between "Average" and "Doubtful" lives, with the latter warranting an extra premium (Lyon 140).[6] Although doctors complied with these new instructions, they did so reluctantly. The Equity and Law's medical advisor argued in 1899 that "we cannot classify": it was "in the estimation we form of the individual before us that we can best do our work" (Symes Thompson 160). The Aberdeen insurance doctor George Williamson similarly complained in 1912 that "in most cases of a defect there is such a variety of influences at work, and so interwoven, that it would be practically impossible ... to make any sort of classification that would be at all satisfactory" (4).

Behind this suspicion of classificatory thinking was a holistic view of health and disease, a continuation of earlier views that wrapped health in a person's spirituality and morality (Lawrence 503-505). Andrea Rabagliati, who examined insurance candidates for the Friends' Provident Institution, told the Yorkshire Insurance Institute in 1908 that nature was "not made in aether-tight or energy-tight compartments, each demarcated from the other by well-marked boundaries" (31). Instead, he urged, there was "one continual heave and throb, and change and interchange going on among all the innumerable parts of her harmonious whole" (31). He concluded that "health and disease... shade off into one another by gradations so insensible that it is very difficult to say where the one leaves off and where the other begins" (31). Although this way of thinking sometimes made it harder for doctors to fit their conception of a normal life within the parameters laid out by actuaries, it was not necessarily as much of a problem for life insurance companies. When their medical advisors refused to draw a clear line between normal and pathological bodies, life offices could still appeal to the market to determine whether or not a borderline life was insurable. As will be discussed in the final section, this appeal was often more popular among customers, but invariably less uniform than a recourse to medical or actuarial expertise.

II. Normal Curves

If insurance doctors only grudgingly classified, actuaries classified as a matter of course. The philosopher Henry Mansel, writing in 1860, stated what had by that time become a staple of life-insurance marketing:

The actuary of an insurance company, if he were to predict the duration of life of any one individual on the books of his office, would in all probability guess wrong... But if the same experiment is tried on a sufficiently large scale, opposite errors will counteract each other, and the general approximate result attains almost to a moral certainty.

274

This formulation was central to insurance marketing because it joined together a reason why people should desire to insure their lives (no individual, not even an actuary, could be certain when he or she would die) with a reason why they should trust insurance companies to pay a full claim when that uncertain hour arrived (premiums were based on the law of large numbers). For the large majority of customers who passed their medical examination as "first-class" or "average" lives, finding a price to cover the risk of premature death was a simple matter of looking up their age on premium tables, which actuaries refined by compiling new data sets throughout the nineteenth century. The fact that actuaries spent much less time generating relevant statistics for doubtful and inferior lives, or singling out superior lives for special treatment, set them apart from most other Victorians who made a career out of appealing to statistical laws.

The large majority of the time and money life insurers spent on gathering vital statistics concerned the class of people whom they deemed to be insurable at standard rates. On the basis of medical guesswork, they turned away a small portion of applicants, ushered a second portion (around 7%, as of 1863) into a surcharged "under-average" class, and often also treated women (a little over 10% of policyholders in 1863) as a separate class of higher-risk lives (Aldcroft 370; Meikle 5; Metcalf 86). To find a price that would cover the risk of the majority who remained, actuaries sought data on the mortality of a healthy subset of humanity. Initially, they collected this data from birth and burial records of towns that they deemed to be healthier than the norm: first Northampton, which (in one actuary's words) was "a small central and healthy borough town, which in itself combines many of the advantages of both town and country" (Hardy 229); then, from the 1820s, Carlisle, which yielded mortality data that Augustus De Morgan deemed to be "the best existing tables of healthy life which have been constructed in England" (167).

By the 1850s, actuaries turned from using healthy towns as proxies for their medically-certified policyholders to mining their own policy registers for mortality data. Samuel Brown, a leading member of the recently-established Institute of Actuaries, urged insurance companies in 1854 to "unite their experience, and ascertain in what respect different classes of assured lives vary in mortality from the average of the country at large"; with this, he concluded, "there would be no need of an elaborate research into old books framed for a totally different purpose" ("Simple Plan" 290). A decade later, the Institute started this process in motion, in conjunction with the Faculty of Actuaries in Scotland (Alborn, "Calculating" 108-109). By 1869, twenty firms had combined to produce a "Healthy Males" table that charted the age-specific mortality of more than 130,000 men whom doctors had deemed to be sufficiently average to pay standard premiums for their life insurance (Aldcroft 370). A "British Offices" table followed in 1901, which was also limited to male candidates of average or better than average health but this time encompassed nearly a million lives from sixty insurance companies (Newman 271).[7]

When actuaries began processing the data they had collected on insured lives, they looked for regular distributions and interpreted these as naturally-occurring statistical laws. One of the first to do so was Benjamin Gompertz, a stockbroker and amateur mathematician who took over in 1825 as the first actuary of the Alliance, a heavily-capitalized fire and life office (Cottrell 31). The same year, he presented a formula for "the law of human mortality" (Gompertz 513) to the Royal Society, which exponentially related age with the random chance of dying then multiplied that function by the initial "power to avoid death" (517-19). The main practical use of such a law was to help actuaries smooth their raw data into curves, which could then be mapped onto premiums that increased with age by even increments. Although such formulae coexisted with more labor-intensive summation techniques that could be used to smooth curves, efforts to refine "an algebraic or analytic law among the numbers" that would let actuaries "sum the progression at once" persisted into the 1870s (Sang 6).

With such formulae in hand, British actuaries repeatedly claimed that their life tables displayed "the presence of a just order and law" (Hodgson 19) or traced "great nature's law... within the narrowest limits" (Farren 5). These actuaries shared an "a priori belief that there are Newtonian laws about people" with statisticians like Adolphe Quetelet, William Farr, and Louis Villermé (Hacking 48, 110-21). Victorian insurance promoters amplified this belief in their effort to convince prospective customers of the safety of their undertakings. William Bridges of the Mitre General Life Assurance Company professed in 1854 that "in all generations death and disease pursue their mission according to a law as rigid and mathematical, amid a vast population, as that of the planetary orbits" (29). The insurance lecturer George Sexton offered a different celestial analogy to make the same point: "As the astronomer knows that an eclipse will be the result of the sun, moon, and earth being brought into a certain position.... so does the Actuary of the Life Office know that the deaths in a given population will be, ceteris paribus, always the same" (11).

In their Newtonian pretensions, the mortality curves produced by life insurance companies were similar to the "normal curves" that contemporary statisticians were starting to generate—most famously the curve that gave birth to Adolphe Quetelet's l'homme moyen. Quetelet, who was trained as an astronomer, extended Pierre Simon de Laplace's probability theories (which were mainly used to confirm a "law of errors" for celestial observations) to human physical and moral characteristics (Hacking 113). Samuel Brown, a leading British advocate of Quetelet's project, defined and defended it in 1867 as "the collection of [statistics] in such a form as will enable the true law of their recurrence to be discovered" ("Report" 11; see also Brown, "Uniform"). Life insurance companies directly applied variants of Quetelet's approach in assessing risk, starting with his favorite example of the normal distribution of height among a random selection of men (Hacking 109-10). A chart of average heights and weights, encompassing 2650 "healthy males," formed the basis for a standard question on all insurance forms starting in the 1850s (Hutchinson 57) and later tests referred to mean values for the specific gravity of urine and for the pulse of insurance candidates (Bedford 21; Hall 60).

What Quetelet actually said about the average man, however—that his "qualities were developed in due proportion, in perfect harmony, alike removed from excess or defect of every kind" (qtd. in Cooper 175)—was much closer to doctors' conceptions of "perfect health" than to actuaries' working definitions of "healthy enough." The crucial difference was that while Quetelet valorized the mean (T. Porter 102), insurance companies worked outward in both directions from the mean until (in the direction of poor health) they finally reached a class of lives that they deemed to be uninsurable. A similar gap existed between what counted as normal for insurance companies and for later adaptations of Quetelet's average man, which appeared both in the realm of social hygiene and eugenics. As Ian Hacking has observed, these late-Victorian exercises in "taming chance" avoided the statistical fatalism that was inherent in Quetelet's conception in one of two ways. In the first case, statisticians identified "a governed class whose comportment is offensive" and tried, by improving both their behavior and their surroundings, to "change the laws of statistics" that such people obeyed (119).[8] In the second, they focused on the "tails of the distribution" (185)—the governing as well as ungovernable classes—and sought, by manipulating their relative success at breeding, to shift the mean toward the former group.

Each of these projects viewed the life insurance industry as a potential ally in their respective efforts to apply statistics to the improvement of mankind. The social hygienist A.J. Hume, for instance, called on insurance companies to "have a voice in county, municipal, and legislative councils in so far … as health is concerned" (316), and G.W. Hambleton, the president of the Polytechnic Physical Development Society, urged them to "become powerful agents in the promotion of national physique and public health" (36). The eugenics pioneer Francis Galton, for his part, asked insurers in 1887 "to combine in order to obtain a collection of completed cases for at least two generations, or better still for three" to help determine the effect of heredity on disease (cited in Pearson III: 72). When the actuary William Palin Elderton suggested in 1902 that such an effort was feasible, Galton drew up a circular citing "a serious actuarial need, namely of better data than are now available for computing the influence of family and personal antecedents on the longevity and health of individuals." Such a study, he hoped, "would be especially serviceable for my own inquiries into what the University of London has now recognised under the title of 'National Eugenics'" (cited in Ibid. 538).

Besides looking to insurance companies' resources (whether financial or statistical), social hygienists and eugenicists hoped to tap into their potential to regulate their customers' lives. The Manchester doctor G.H. Darwin proposed that insurers could give all their customers "a small pamphlet" containing rules regarding "the upkeep of the body" and "sanitary regulations of a simple character, according to which the home should be managed" (49-50). Hume similarly hoped that insurance companies would issue "a short Code of Rules by means of which longevity is to be attained... to every accepted life along with the policy" and would exact "some pledge for future conduct" upon admission (316). Defenders of eugenics also hoped to enlist life insurers as allies, this time by amending their system of classifying lives to make room for those who exceeded the merely average. James Barr, a Liverpool doctor and Eugenics Society Vice President, called for special treatment of "unexceptionable lives" who were "muscular and proportionally developed" and of "long-lived stock" (242-43).

It is possible to find some fellow travelers in the insurance industry who recognized its potential to assist in the goals of either social hygiene or eugenics. Henry Porter, an assistant actuary at the Alliance, urged life offices to "take their share" in "the work of sanitary, political, and moral improvement" that was underway in the 1850s (111-12). Two generations later, the Eagle Star and British Dominions Insurance Company issued a Guide to Health as part of a "Household Series" of pamphlets, which taught (among other things) that although "disease germs are always floating about in the atmosphere, " it was "possible, by means of disinfectants, to destroy them while they are still outside the body" (23). Eugenic ideas, at least in the most general sense of favoring hereditary over environmental aspects of disease, had a wide currency among most actuaries, who were convinced that family history was an important indicator of risk for tuberculosis (Alborn, "Insurance"). A few actuaries went further than this, and personally participated in the eugenics movement. Besides Elderton, whose sister Ethel worked at Galton’s Eugenics Record Office (Love 147-48), another sympathizer was the Scottish Life secretary Lewis Orr, who spoke to the Eugenics Society in 1913 regarding the "identity of interest between Eugenic and Actuarial Science" (331).

Yet such support among life insurance actuaries for social hygiene and eugenics had relatively narrow limits. In the case of social hygiene, this was because the target population of most public health reforms was a class of people that insurance companies already excluded, and hence had little interest in improving. The Scottish Amicable’s manager argued that sanitary reform only meant "prolonging weakly lives a short time longer, and in the process of nursing them deteriorating the lives of the mothers" (Marr 48). As for eugenics, even Orr admitted that the actuary was "at once a cold critic and a warm friend" to that movement, since he "looks upon human life... with a more restricted range of vision" (333). Here, the problem was that eugenics aspired to levels of specificity in classifying risk that ran the danger of failing to include enough policyholders to form a safe average. The Law Union’s actuary, Samuel Warner, concluded that performing "elaborate, difficult and laborious" investigations into more specific risk groups would not be worth the trouble, since "actuaries could seldom hope, whatever the general experience might be, to get enough cases to form anything like an average for their own practice" (Lutt 524).[9]

In both these cases, the more basic issue separating life insurance from social hygiene or eugenics was that it was a profitable business that focused on insuring normal lives, as opposed to a (usually unprofitable) project seeking to normalize pathological lives. As Warner argued in a different setting, "the backbone of life assurance companies' business did not consist of impaired lives, but of lives which were considered to be in average health" (Carruthers 322). Although this translated into what Warner called "a very rough and ready… system" for classifying risks (Ibid.), it was a system that would remain in place as long it made money for insurance companies. Any hoped-for partnership between insurance and eugenics died on the vine, leaving Galton's disciple Karl Pearson to complain in 1930: "Fifteen years passed after Galton threw out the suggestion that material might be available in assurance offices, before an actuary told us it did actually exist. Twenty-five further years have rolled by and still nothing has been done!" (III: 73).

III. Normal Prices

For most of the nineteenth century and into the twentieth, most British life insurance premiums came closer than most other services and commodities to approximating what economists would have claimed to be a "natural" or "normal" price. In 1842, more than 70% of life offices charged within 5% of the mean price, despite the fact that many were still basing their rates on the higher mortality of the Northampton table (Pocock 158-63). By 1910, when the Healthy Males table had taken hold as the industry standard, the proportion of companies with published rates that clustered within 5% of the mean had increased to more than 90% at most ages (Life Assurance Premiums 6-7). This uniformity in price could be interpreted as the result of supply and demand, since competition forced companies to keep their premiums as close as safety permitted to the cost of covering their risks. At the same time, the precise level to which prices fell had more to do with the professional judgment of the actuaries who translated mortality statistics into premiums than with the direct agency of customers. This latter form of bargaining, which bore more resemblance to a Smithian model of an untrammeled market, only took place in that realm of life insurance where the least uniformity prevailed: the "under-average" risks, for which actuaries ceded their price-setting powers to a market mediated only by salesmen, branch managers, and company directors.

The concept of a natural price dictated by supply and demand dates back at least to the 1730s, when Richard Cantillon wrote his Analysis of Trade, Commerce, Coin, Bullion, Banks, and Foreign Exchanges.[10] Not published in English until 1759, this book featured as its centerpiece a theory of value that would persist well into the twentieth century:

The intrinsic Worth of every Thing is proportioned to the Value of the Land, Labour, Risk and Time necessarily had in producing it into Use and Form; and the several Variations which happen in the Prices of every Species of Goods and Merchandize, depend on the Demand there is held for them, to supply the Delicacy, Taste and Luxury of the Rich.

34

With the exception of the last clause, which soon gave way to a more populist rendering of the market, the basics of this definition remained unchanged for a century and only slightly modified for seventy years after that (Milgate). Adam Smith assumed that a commodity's market price approached its natural price when "the quantity brought to market is just sufficient to supply the effectual demand and no more" (69), and Malthus observed that prices were "regulated by the ordinary and average relation of supply to the demand" (78).

Through the 1860s, a primary debate in political economy regarded the relative contributions of Cantillon's component parts of land, labor, risk and time to a commodity's value (Schumpeter 188-89, 588-605). One new wrinkle, starting with Adam Smith, was the specific identification of this price as "natural" and the association of this adjective with Newtonian metaphors to illustrate the movement of prices.[11] Smith's natural price was that "to which the prices of all commodities are continually gravitating" (70); "any deviation" from that price, according to John Stuart Mill, was "but a temporary irregularity, which, the moment it exists, sets forces in motion tending to correct it" (535). A favorite analogy, which continued to be used by neoclassical economists, was that of a body of water. Mill compared the economy to an ocean that "preserves its level" despite always being "ruffled by waves, and often agitated by storms" (Ibid.), and the French economist Léon Walras compared the market to "a lake agitated by the wind, where the water is incessantly seeking its level" (cited in Milgate 2).

Although neoclassical economics retained this earlier tendency of viewing prices as inevitably finding their level, it departed from the economics of Smith and Ricardo in other regards. One of Alfred Marshall's accomplishments in turning economics into a profession was his substitution of a "normal" for a "natural" price as the field's focal point. He faulted Smith's "doctrine of natural organization" because it prevented people from "inquiring whether many even of the broader features of modern industry might not be transitional" (205). What Marshall called "normal action in economics," in contrast, encompassed "much action which we should use our utmost efforts to stop" (29). In other words, his "normal" was the same as that employed by Hacking's tamers of chance: for him, economics shared a "peculiarity... with some other sciences" since its "material... can be modified by human effort" (Ibid.).

Risk and insurance entered into Smith's analysis only in passing, usually alongside the "trouble and expense" involved in bringing a good to the market (10, 63), and Marshall similarly included insurance among "the general expenses" that manufacturers and traders "added to the prime cost" of their goods (330). In life insurance, what modern economists would later call a transaction cost became the primary cost associated with the service: around two-thirds of the basic premium, for most companies. The rest consisted in "loading," which accorded more closely with the costs of production that economists spent most of their time debating: agents' commissions plus "rent, taxes, salaries, &c," as one actuary put it in 1850 (Jellicoe 336). What this meant in practice was that the "level" to which insurance premiums "gravitated" had more to do with actuaries' ability to calculate the risk of premature death than with anything else. This was the main reason why most companies before 1820, using the less "healthy" Northampton table, charged between 20% and 30% more than most companies after 1840, when the Carlisle and later the Healthy Males tables came into general use (Alborn, "Fund Managers" 72).[12]

Although the concept of a natural price was not foreign to life insurance officials, its signification could not have been more different from the way most economists used the term. Actuaries typically referred to a natural price and related terms in order to oppose increased competition, instead of welcoming competition (as economists were likelier to do) as a means of reducing prices to their natural level. Once actuaries were convinced that they had pinned their prices to a timelessly Newtonian law of mortality, they were equally convinced that any effort to undercut that price must be an unnatural act. Samuel Brown made this argument when he invoked the law of diminishing returns, which David Ricardo had originally proposed to show that new investment in land was bound to lose money under the monopoly conditions of the Corn Law. Ricardo's point had been that free trade would change that, by opening up access to foreign grain and freeing up capital to get more out of existing land (Berg 43-72). Brown acknowledged this point, but insisted that it did not apply to life insurance:

Competition must always be expected in every branch of business where profits can be obtained; but assurance is not like an article of manufacture, of which, by the aid of capital and machinery, we can scarcely see the limit to which the cost of production can be reduced. The rate of mortality is a law of nature, and were the Companies, by laying out more capital in their expenses, to increase their business to any extent, they would not be able to vary the consequences of this law one iota. Let once the premiums come below the actual risks and the expenses combined, and every addition to the business is a fatal and irretrievable loss,—every new policy a fresh step to ruin.

"Summary" 208

The anonymous author of Free Trade versus Life Assurance (originally a letter to the editor of the Morning Advertiser) took this stance to its logical extreme, arguing that life insurance

transactions do not bear any analogy to the ordinary operations of trade, for you cannot buy, so to speak, the expectation of life at one price, and sell it at another... There are not those fluctuations in the average mortality of man, as there are in the various market prices of coals, cotton, tea, coffee, sugar, rice, and the numerous other articles of consumption.

5

If life insurance displayed its conservative face in such arguments, there was also room for a more progressive stance on determining a natural price for the risk of premature death. This stance approached the more flexible neoclassical definition of "normal price" that allowed for "transitional... features of modern industry" (Marshall 205). Brown followed his tirade against ruinous competition with an equally energetic call for life offices to determine, "not from conjecture, but from actual facts, how far we can go, and where our safety ends"; this, he concluded, might yield "the discovery of new laws or the correction of former theories," which might in turn open up a "field... for new and extensive operations all over the world" ("Summary" 209).

As discussed above, although such calls did lead to further refinements in statistics on the healthy males who formed the bulk of insurance customers, they did not open up new fields of business subject to "scientific" prices. Frank Knight, an American economist who wrote widely on both value and risk, was quick to identify this dichotomy. Life insurance, he wrote in 1921, qualified as the "most highly developed" branch of insurance because "its contingencies are most accurately measured [and] its classifications are most perfect" (Knight, Risk 247). Any such claims to accuracy evaporated, however, once one left "the relatively narrow circle of 'normal' individuals." Beyond this circle, life insurance faced no end of "difficult" cases, occupying categories about which actuaries possessed little or no statistical knowledge, and which in any case formed such small groups as to be impervious to the law of large numbers (248).

When actuaries moved from the relative certainty of normal lives to this more treacherous statistical realm, they ironically determined prices in a way that corresponded more closely to the original Smithian conception of supply and demand. Instead of scientifically determining standard premiums in these cases, actuaries left the details of setting prices to directors and business managers, who were as likely to do so on "general business grounds" (as the English and Scottish Law's board decreed) as on their medical advisor's prognosis (XIV, 23 March 1899). They did so on the assumption that as long as customers were free to compare offers among different companies, a sort of rough justice would condemn the least-insurable lives to higher premiums and provide normal rates to those who could convince at least one office that they posed no special risk. Such cases also engendered a new definition of risk, which extended beyond the simple chance of premature death to the chance (for instance) of losing a big account by failing to come down in price for a disgruntled candidate.

An example from the Norwich Union's archives is indicative of the process that many insurance offices used to determine premiums for such customers. It concerns a Liverpool cotton broker named John Lightbound, who applied in 1880 to add £1250 to a policy he had taken out three years earlier, and was charged £5 extra (£39 instead of £34) on the grounds that two of his sisters had died of consumption since the original policy had been issued. The agent, W.H. Andersson, first claimed that the sisters' deaths should not count as evidence of poor family history, arguing that they "were nuns (the family being very strict Catholics) & had to undergo very hard privations—the male side of the family all being good." When that failed, he added that he saw "most of the family daily all of whom I consider good insurable lives & ... very steady & regular," and implied that since "we have the whole of the family" insured, losing this policy would hurt this connection. All this was enough to reduce the extra by £2, and even then Lightbound only assented because Andersson agreed to cover it out of his commission. Three years later, Lightbound was still complaining "most bitterly at the Extra … charged him," and Andersson again reminded the head office that "we have all the family in our Books." The board agreed to re-examine the man, and eliminate the extra if nothing showed up (Norwich Union 35/2).

As this case suggests, life insurance premiums deviated most widely from what economists would recognize as their "normal" level when supply and demand had the most to do with their determination. The Norwich Union's archives, and those of many other firms, are full of examples where directors did not bend to their branch managers’ request, or acceded more fully than in the Lightbound case to the customer's demand; and these do not include the many instances where customers either accepted a surcharge without complaining or tried other companies until they could find one that charged less. As any modern economist would confirm, the reason for variation in these cases was that there was imperfect information on both sides: a customer suspected (but could not prove) that he or she posed less of a risk than the doctor had diagnosed, and the insurance company lacked sufficient data to be able refer complaints to a standard "under-average" mortality table. As discussed above, one reason insurance companies lacked information was because their managers had made a conscious effort not to seek it out—on the grounds that they were mainly in the business of assessing the risk, and generating prices, for "normal" lives.

Exactly how far this way of thinking about price departed from common economic assumptions is clear from a different quotation by Frank Knight, this time from an article he wrote in 1917 on "The Concept of Normal Price in Value and Distribution." Knight agreed with his fellow neoclassical economists that the normal price was "that at which production would exactly equal consumption," but he went on to insist that such a price "only rarely and accidentally does in fact"—concluding that "one price cannot be normal unless all other prices are so, and normality is really a condition of the whole system" (71). In contesting earlier economists' claims that market prices only accidentally and temporarily deviated from the norm, Knight revealed their "normal" price to be no less exceptional than the doctors' perfectionist conception of health or Quetelet's average man. Life insurers, in contrast, started with the assumption that normality was a condition of the whole system—in their case, the customers whom they decided to accept at standard rates. They were able to accomplish what for economists was an exceptional event by not professing to regulate society as a whole, but rather only their pre-screened part of it.

IV. Conclusion

In The Taming of Chance, Ian Hacking contrasts the case of illiberal Prussia, which exhibited "exquisite statistics [but] a resistance to the idea of statistical law," with that of liberal Britain and France, where such ideas bore much more fruit. The reason for this difference, he claims, is that in Prussia "the group... conferred identity upon the individuals who comprised it" (36), whereas "if you are a liberal... social laws are constituted by the acts of individuals" (37). His point in making this contrast parallels Foucault's aim in introducing the concept of governmentality: to argue that the human sciences, in a liberal state, can be (or at least can aspire to be) effective surrogates for more direct techniques of governance exercised by such a state’s less liberal counterparts. The case of British life insurance suggests that dichotomies like this require rethinking, since not all wielders of governmentality appeal to the human sciences in the way that Hacking (or, by extension, Foucault) implies. Although life insurance started with the assumption that individual acts constituted social laws, they ended up much closer to the "Prussian" model of dispensing with these laws once they discovered that they were able to confer identity directly upon the healthy men who qualified as their customers. They could do so without exercising anything close to "Prussian" bureaucratic powers, because they had the luxury of refusing to insure people who fell too far outside their definition of the normal—or, alternatively, of not bothering to regulate the price they charged those people through recourse to their "exquisite statistics."

That this was the case can be seen from the departure British actuaries made, beginning in the 1870s, from claiming that their premium tables were based on Newtonian laws of mortality. That claim, which had once been crucial for creating trust among their customers and smoothing mortality curves, faded with the accomplishment of greater levels of trust (through a combination of new legislation and decades of relative financial stability), and with the appearance of alternative smoothing techniques. As early as 1876, the manager of the Scottish Union Fire and Life commented that "[w]hatever abstract 'law' of mortality may exist in the nature of things, it is obviously liable to so many exceptional influences, so many caprices even, ... that although the law were known to us, we could not rely on it as an invariable factor" (McCandlish 16). When the British Offices table appeared at the turn of the century, the actuary Ralph Price Hardy did not bother to graduate it by recourse to an algebraic "mortality law"; instead he used one of "the new excellent and short formulas of approximate summation" that were available. A trade journal commented, with more than a hint of regret, that

it would appear that the search for a law of human mortality is now more or less officially abandoned, a course which, though from some points of view regrettable, is no doubt justified. The imaginary Law must be regarded as a 'Will o' the Wisp' and relegated to the society of the philosopher's stone and the elixir of life.

Insurance Record 40 [1902]: 374

What was beginning to be true for life insurance in 1902 did not extend very far into the human sciences of medicine, statistics, and economics. All of these continued to embrace a perfectionist conception of the normal—and, at least in the latter two cases, an accompanying faith in laws that could be derived from individual acts. Life insurance companies, in contrast, took the normal as they found it, because it was easier to make money that way than to try and convince people to aspire to a norm that did not yet exist. Far from accomplishing this in isolation from the human sciences, they accomplished it by diverting medicine, statistics, and "general business grounds" to their own ends. Doctors and actuaries could choose whether or not to bend their disciplinary commitments to this end, and enough of them did so to provide insurance companies with ample levels of expertise. As subjects of a history of interdisciplinarity, these actors remind us that the disciplines that shaped their expertise—with their tendency to focus on pathological behaviors in order to regulate and reform the human or social body—did not necessarily signify the last word on the shaping of late-Victorian society.