The Son Also Rises Read online

  Supporting this view is the remarkable pervasiveness of New France disadvantage. Figure 3.11 shows the rate of occurrence of the most common New France surnames among physicians, compared to the most common Irish surnames.17 The New France surnames look as though they are drawn from a completely different distribution than the Irish surnames. There is something pervasively different about these two groups.

  FIGURE 3.12. Marital endogamy among New France descendants, 1950s.

  Interestingly, even going back to the 1950s and considering data from states with many people of New French descent, rates of intermarriage between those with New France surnames and those of surnames of other heritages have been substantial. This is not an isolated social group.

  Figure 3.12 shows the percentage of individuals of Franco-American heritage in four New England states and in Oregon, according to the 2000 census. Also shown is the percentage of those in the 1950s with common New France surnames who married a partner with any New France surname. By the 1950s, a large majority of New France descendants were marrying outside that community, even in Maine and Vermont, where they still constitute a quarter of the population. This has been a largely open community for generations. Interestingly, despite the evidence of persistently lower status, many of these exogamous marriages were with individuals bearing Irish and Italian surnames, who were coreligionists, and there is a substantial overlap between the New France and Irish populations in New England.

  The low average occupational status of New France surnames cannot be the result of recent assimilation into U.S. society. Some of the New France population was incorporated into the United States in the colonial era. The rest arrived as immigrants from Quebec and New Brunswick, mostly between 1870 and 1930.18 So the vast majority of those in the current stock of physicians and attorneys are from families who have been in the United States for three or more generations.

  An alternative explanation of the low socioeconomic status of this group is that it was drawn, for reason of accidents and history, from the lower end of the French status distribution. U.S. citizens of French colonial origin typically experienced selection through two migration experiences. First, the modern population of seven million people of New French descent derives from a small stock of migrants to France’s North American colonies in the seventeenth and early eighteenth centuries. It is estimated, for example, that the French population of Canada in the late nineteenth century derived from fewer than nine thousand original French settlers.19 Within this settler population, some people were much more reproductively successful than average, and they are the progenitors of a disproportionate share of the modern population. The high incidence of Mendelian (single-gene) disorders in the Province of Quebec, for example, has been argued to stem from a disproportionate contribution of a small share of the founder population to the modern genetic stock in French-speaking Canada: according to one researcher, “as few as 15% of the founders could account for 90% of the total genetic contribution from the founders.”20

  The surname Gagnon, as noted, is rare in France, with only about nine hundred holders now, and presumably less than three hundred possessing the surname in France in 1700. Yet there are 54,000 Gagnons now in North America, most of whom must have derived from a tiny group of Gagnon migrants to New France.

  Another distinctive feature of the demographics of Quebec in the seventeenth to nineteenth centuries was that the most reproductively successful group in the population was of lower socioeconomic status.21 So the founder population of the French population of North America could well have been a draw from the lower end of the French occupational distribution.

  In addition, there is evidence that French Canadian immigrants to the United States between 1860 and 1920 were negatively selected from the Francophone population of Canada. Byron Lew and Bruce Cater show that from 1900 to 1920, illiterate French Canadians were significantly more likely to emigrate to the United States than their literate counterparts.22

  It is possible, then, that the low status of modern Franco-Americans stems from the persistence of an ethnic culture maladapted to economic success. But it seems more plausible that their low status in the United States stems from the fact that they are a twice-selected low-status subgroup of the parent French population. Their continued low occupational status is witness to the slow processes of social mobility revealed once we turn to surnames as a measure.

  Japanese Surnames

  Today, people bearing Japanese surnames are an elite group in the United States with respect to occupational status. Unusually, Japanese Americans until recently showed no tendency to regress toward the mean. Instead, from 1940 to 2000, they became more distinctive as a population subgroup (see figure 3.13). Only since 2000 has there been regression to the mean among Japanese Americans of the kind seen among the descendants of Jewish immigrants from the 1980s onward.23 The Japanese experience also contrasts strongly with that of New France descendants, although these groups arrived in the United States at roughly the same time. Their experience in the medical profession reflects a general tendency for Japanese Americans to become a highly educated, high-income subgroup in America.

  We can attribute the delayed rise of Jewish Americans among physicians at least in part to the quota systems that operated increasingly in East Coast universities from the 1930s until the 1950s. The delayed rise of Japanese Americans as an elite group is less plausibly explained by such barriers. On the West Coast and in Hawaii there do not seem to have been such barriers. Unlike Jewish Americans, Japanese Americans were not graduating from colleges at a much higher rate than the rest of the local population, which was what led to quotas being placed on Jewish admissions in the East. Also, they represented a smaller share of the population in states like California than Jews did in New York.24

  FIGURE 3.13. Relative representation of Jewish, Japanese, and New France surnames among physicians, United States.

  Masao Suzuki argues that one factor that explains the high status of the Japanese is that emigrants from Japan were always a relatively elite group and became more so as barriers to Japanese immigration were set in place. Table 3.4 shows the occupational distribution of Japanese immigrants entering the United States from 1899 to 1931, compared to the occupational distribution in Japan as a whole in 1920. Even in the period before the Gentleman’s Agreement of 1907 (a tacit agreement between the governments of the United States and Japan that placed informal limits on Japanese immigration), the immigrants of 1899–1907 would have likely been more skilled than the Japanese home population in those years. Because the Japanese economy was rapidly modernizing, the occupational distribution in 1920 would have been more skilled than twenty years earlier. After the Gentleman’s Agreement, Japanese immigrants to the United States became distinctly more skilled.

  TABLE 3.4. Occupational distribution of Japanese immigrants to the United States and Japanese domestic population, 1899–1931

  The rise in status of the Japanese American community until the 1980s was substantially driven by the high skills that Japanese immigrants brought with them in the years 1908–70, when Japan was a substantially poorer economy than the United States. According to the 1960 census, of Japanese Americans born in the 1920s, 16 percent were born in Japan, and for those born in the 1930s the figure is 27 percent. The high skill level of Japanese immigrants in the early to mid-twentieth century is evident from the AMA register of physicians. Of those physicians with Japanese surnames completing medical school in the 1940s, 69 percent were trained in Japan. In the 1950s, 52 percent were still Japanese trained, and in the 1960s 44 percent.

  Conclusions

  This chapter establishes through analysis of surname distributions that the underlying social mobility rates in the United States since 1920 are much lower than conventional estimates would suggest. Although surname groups tend to regress to the mean in occupational status, they do so far more slowly than conventional estimates imply.

  Looking at ethnic groups such as Jews, blacks, J
apanese Americans, and Franco-Americans, it might seem that this slow social mobility is connected to some shared social capital, or lack thereof. But we see the same slow rates of mobility within groups of surnames that are not ethnically or culturally homogeneous, such as the bearers of the rare surnames of the rich of 1923–24. Chapter 6 offers explanations of why these surname estimates of social mobility rates reveal the true underlying rates of social mobility in a society such as the United States.

  We do not, however, find any evidence for the dystopia that Charles Murray fears in his recent book, Coming Apart: The State of White America, 1960–2000. Murray argues that there is an increasing disparity between the values of the white upper class and the white lower class, and increasing geographic, educational, and social isolation of that lower class. The data presented and discussed in this chapter give no indication that social mobility rates have declined in the past few decades in the United States. They are slow, but not any slower than in the previous forty or fifty years.

  1 See Ward et al. 2012.

  2 Social Security Death Index, n.d.

  3 The average surname incidence for the 2000 population for domestically trained physicians is 2.85 per thousand. We show below that some recent immigrant groups are even more elite according to this measure than the Jewish population, especially once foreign-trained physicians are included. The Jewish population is losing its distinction as the highest-status ethnic group in the United States to such newcomers as Egyptian Copts, Hindus, and Iranian Muslims.

  4 Rutgers and William and Mary are, of course, not members of the Ivy League but are of similar antiquity.

  5 It was also required for the chosen surnames with a hundred or more holders in the 2000 census that the holders of the name be at least 80 percent white and less than 10 percent black. This restriction was specified because of the aim of identifying just elite surnames from the early history of the United States.

  6 Barack Obama is the most visible member of this elite. Chapter 13 shows that black Africans, for example, have substantially more physicians per 1,000 members than the general white population in the United States.

  7 Jefferson is another surname that is predominantly black. It presumably arose in the same way as Washington. But only about two-thirds of Jeffersons are black.

  8 Because the numbers of Native American physicians are so small, their intergenerational status correlation cannot be meaningfully estimated, and this group is therefore omitted from the discussion below.

  9 Using the method adopted here, this would imply a persistence parameter for these groups greater than one. In this case, such a parameter cannot be an intergenerational correlation, since it would imply that the distribution of status is not constant over time.

  10 Borst 2002, 210. These quotas were progressively tightened during the 1920s and 1930s. Thus, in the most dramatic case, Jewish enrollment at Boston University Medical School was cut from 48 percent in 1929 to 13 percent by 1934 (Borst 2002, 208).

  11 Estimated persistence rates change little if this assumed cutoff is changed.

  12 We define convergence as being within 10 percent of the expected representation.

  13 Olson/Olsen also shows an average representation among physicians.

  14 This assumes that attorneys represent the top 1 percent of the occupational status distribution, whereas physicians were assumed to represent the top 0.5 percent.

  15 Borjas 1995.

  16 MacKinnon and Parent 2005, table 1.

  17 New France surnames were included only if fewer than 5 percent of the holders were black. The figure excludes the three most common Irish surnames, O’Brien, Gallagher, and Brennan, which each had more than forty-five thousand holders.

  18 MacKinnon and Parent 2005, appendix, table 1.

  19 Scriver 2001, 76.

  20 Scriver 2001, 78. This is not to imply that the source of Franco-American disadvantage in the modern United States is genetic in nature. The possible mechanisms of status transmission are discussed in chapter 7.

  21 Clark and Hamilton 2006.

  22 Lew and Cater 2012, table 2. In contrast, literate English Canadians were more likely than illiterate English Canadians to emigrate to the United States.

  23 The decline in representation among physicians from 2000 to 2012 is statistically significant, indicating that regression has begun.

  24 The internments of 1942–45 applied to only a minority of the total Japanese population, and this seems too brief an episode to explain the long-delayed rise in their representation among physicians.

  FOUR

  Medieval England

  Mobility in the Feudal Age

  MEASURED USING SURNAMES, social mobility is surprisingly low in modern Sweden and in the United States. How do mobility rates compare for the preindustrial era, before the whole panoply of public education and fair-employment laws of the modern state? Using surname evidence, we can estimate social mobility rates in England back to 1300 and the feudal world of lord and bishop, serf and slave.

  What most people would expect from this study is obvious: the illustration of a class-ridden past, with the majority of the population trapped under the feudal yoke, condemned to a brutish existence cultivating the heavy sod of the medieval fields. This unfortunate majority would have been supporting an entrenched and tyrannical elite who sustained their position through control of land, politics, and violence. We would then expect to see the liberation of the Industrial Revolution, which eventually freed the population from the constraints of the past, followed by a series of political reforms that enfranchised ever-larger shares of the population. After World War II we would note the introduction of redistributive taxation, mass higher education, and a modern world of social mobility. The stately ossification of the world of Downton Abbey would be replaced by the hectic social turbulence of the rude boys of the City and finance.

  We have seen that modernity did not bring rapid social mobility in Sweden and the United States. But was there gain at least relative to the bad old days of the Middle Ages? Do the modern English at least live in a society of greater social mobility than in medieval times?

  The Rise of the Artisans

  In using surnames to measure mobility in medieval England, the first set of surnames employed are those of medieval artisans: workers with some skills, located at the middle or lower middle of the social scale. More than one in ten of all modern surnames in England derive from the occupation of some medieval ancestor. Smith, Baker, Clark, Cook, Carter, Wright, Shepherd, Stewart, Chamberlain, and Butler are all easily identified occupational surnames, but there are plenty of others whose etymological origin is more obscure, such as Webb or Webber (weaver), Coward (cowherd), Walker (fuller), Coulthard (colt herd), Baxter (baker), and Dexter (dyer).

  In addition to Smith (the most common surname in England, Australia, and the United States), the whole range of the building trades is represented in surnames: Carpenter or Wright, Mason, Thatcher, Plumber, Glazier, Painter, Sawyer, Slater, and Tyler. The one trade that is not well represented is bricklaying. Surnames were well established in England by the fourteenth century, and brick became an important building material in England only after 1500, too late to leave a mark on surnames.

  Farming contributes such occupational surnames as Carter, Shepherd, Coward, Plowman, and Thresher. Textile and clothing production, important occupations in medieval England, generate Taylor, Webb, Webber, Webster and Weaver, Walker and Fuller, Barker and Tanner, Lister, Dyer and Dexter, Skinner, and Glover. The female versions of occupational surnames in medieval England were derived from the male forms by changing the ending to -ster or -xter. There are however, few Spinners, despite the importance of this occupation, because spinning was exclusively female work and would not become the stuff of an inherited surname.1 Food production gave us Baker and Baxter, Butcher, Coke and Cook, Brewer and Brewster, Salter, Miller and Milner, and Spicer. All these artisan occupational surnames connote the medieval origin of a family not at t
he bottom of society, among the laborers in town and field, but in the middle, below the landowners, the manorial officials, the clerics, the merchant class, and the attorneys.

  When were these surnames first affixed as heritable surnames? The exact period is unknown, but by 1381 these surnames were largely transmitted through inheritance. We know this from the surviving returns of the 1381 poll tax. Some of these returns show not just the surnames of taxpayers but also their current occupations. An occupational surname that differs from its holder’s occupation must be an inherited surname. By 1381, only 38 percent of people with artisan surnames were doing the job their name described, so a minimum of 62 percent of people with artisan surnames had inherited them.2 But given that many sons would follow their father’s or grandfather’s profession, by 1381 most surnames must have been inherited. The fact, however, that 38 percent of artisan surnames still described the occupation of the bearer suggests that these surnames could not have existed as hereditary surnames for more than three or four generations. Thus it seems likely that artisan surnames became hereditary around 1250–1300.

  There are four major sources that identify the elite in medieval England. The first is people associated with the ancient universities of Oxford and Cambridge, whose membership records start around 1170. The second is those whose wills were proved in the highest will court in the land, the Prerogative Court of the Archbishop of Canterbury (PCC), whose records span the period 1384–1858. Because of its location in London, this was the court for the elites of English society until 1858. The third is those whose wills were proved in the PCC who are referred to as Sir or Gentleman, suggesting even higher status. The fourth is members of Parliament, the House of Commons, from 1295 onward. What do these sources suggest about the social mobility rates implied by the distribution of artisan surnames over time?