The name evidence thus suggests that many of the modern-day Gypsies and Travellers of England are not descended from those of the seventeenth century. The normal process of social mobility should have brought those descendants closer to the social mean. Instead there is a steady flow of people into and out of Gypsy and Traveller communities. The more economically successful members of these communities acquire permanent homes and occupations more typically associated with the majority population. Because they are in no way racially distinct from the rest of the indigenous English population, and because their surnames do not reveal much about their background, they can at any time blend into the larger society. But at the same time that people leave the group, others from the margins of society flow in. These entrants adopt the lifestyle and mores of the Gypsies and Travellers. For example, the recent addition to the traveling community, the New Age Travellers who took to the road in the past generation, will likely merge into future generations of Gypsies and Travellers.
This hypothesis of an open Gypsy and Traveller community predicts that surnames concentrated in the Traveller community in 1891 or earlier, such as Boswell, Penfold, Loveridge, Brazil, and Beeney should trend upward toward the mean in social status.27 Each generation, some of the Travellers with these surnames will move up and be incorporated into the settled society, so that the average surname status will rise. But only those who do not experience this upward mobility will continue to identify as Travellers. Thus self-identified Travellers would seem to be a minority not subject to social mobility.
However, testing this hypothesis by looking at the social status of a Traveller surname, such as Loveridge, from 1858 to the present does not yield the expected result. Figure 13.8 shows a measure of the social status of the surname Loveridge from 1858 to 2012, by decade. The status is measured by comparing the fraction of people called Loveridge whose estates were probated to the fraction of the general population whose estates were probated. For a surname of average status, this fraction is one. For names of above-average status, the figure is greater than one, and for names of below-average status, it is less than one. Interestingly, until 1910, Loveridge was a relatively high-status surname. But since then its average status has declined steadily, so that by 2000 the probate rate of Loveridges was only about 60 percent of the average.28
FIGURE 13.8. Social status of Loveridge by decade, 1858–2012.
What is happening here? Although rare surnames can move away from the social mean of status as a result of random forces, Loveridge is so common as to make such a random movement wildly improbable. By 2002 there were more than five thousand Loveridges in England and Wales. Does the decline of this name suggest that the law of mobility sometimes does not predict social outcomes? Can social groups systematically move downward in status?
The likely solution to this puzzle, which does not violate the law of mobility, is the following. Loveridge had exceptional growth in frequency for a common surname in England in the years 1881–2002. In that interval, the stock of the average indigenous surname did not quite double. Yet in the same interval, Loveridges increased nearly fourfold. This disproportionate increase may be attributed to the very high fertility rates among Gypsy and Traveller families in modern times, illustrated in table 13.2 above. Such fertility rates would double the population of Travellers in each generation and could explain why, even though a substantial fraction of Loveridges are not Travellers, the overall stock of the name could increase so much over time. If fertility is much higher for low-status families with a given surname, then even if every family conforms to the law of mobility, the average social status of the surname can move downward from the mean over time.
The implication here is that the children of the low-status Loveridges are indeed regressing to the mean over generations, and they have so many more children than the high-status families that the surname group is diverging toward the bottom end of the status distribution. Figure 13.9 simulates this effect for a population that starts in the first generation with average social status, and in which status regresses to the mean with a persistence rate of 0.7. Fertility, however, is at twice the average rate at the bottom of the distribution and half the average rate at the top. In this case, mean status moves downward from the mean, despite every family’s regressing to the mean.
FIGURE 13.9. Simulated downward movement of surname-group social status resulting from fertility effects.
The downward mobility in this example above will continue until the average status reaches an equilibrium at which the mean for the group is sufficiently low that regression to the mean can balance the excess fertility at the bottom end of the status distribution. Thus another explanation for long-lasting underclasses in a society, even with intermarriage between the underclass and the rest of the population, is that the underclass has much higher fertility rates than the society as a whole. The effects of marital exogamy in pulling the group toward the mean are offset by the higher fertility of poorer members of the group. However, in preindustrial society, poor groups typically had lower fertility rates than richer groups, so this effect could operate only in the modern world (in this case, since 1880).
The example of the Loveridges does suggest, however, that under current conditions, no matter how many campsites and social services are provided by local authorities in Britain, a distinct and poor Gypsy and Traveller population is likely to persist in England and Ireland for the foreseeable future, beyond the reach of the normal processes of social mobility.
Elites and Underclasses in the Modern United States
The proposition that elites and underclasses are not created by religion, culture, or race is supported by evidence from the United States on current elite and underclass populations. A quick confirmation of this proposition can be obtained by looking at surnames identified with particular ethnic or national groups and counting the numbers of registered physicians per thousand of each surname type in 2000. We can divide this number by the average number of physicians registered per person in the United States in 2000. For the population as a whole, this number will be one.
Figure 13.10 shows the implied elite populations in descending order. Topping this list, surprisingly, are names of Egyptian Coptic origin. For such surnames, there are a remarkable forty-eight physicians per thousand holders of the surname in 2000. This rate of representation among physicians is thirteen times that of the average surname in the United States. Next come Hindu surnames, then Indian Christian surnames, then Iranian Muslim surnames, then Maronite Christian surnames, originally from Lebanon. All these groups are represented among physicians at a frequency six or more times that of the general population. All are relatively more frequent among physicians than even Ashkenazi and Sephardic Jewish surnames. These new elites in the United States span three major religious traditions: Christian, Muslim, and Hindu. But the Christian groups here represent mostly a very traditional, unreformed Christianity. The Coptic Church, for example, has a liturgy in Coptic, a language no longer spoken by church members. Part of the Maronite liturgy is in Aramaic, again not spoken by the congregants. Many Indian Christians are of Portuguese descent or from families converted by the Portuguese, or are from the even older Syriac tradition. U.S. elites exhibit an astonishing diversity now of backgrounds and cultural heritage.29
FIGURE 13.10. Elites in the modern United States, measured by relative representation among physicians.
The remaining elite groups are Koreans and Chinese, followed by Filipinos, black Africans, and Greek surnames, and finally Armenians, Japanese, Vietnamese, and black Haitians. Almost every major race and religious tradition is represented—except for European Protestants.
Many of these modern elites are creations of U.S. immigration policy, which for countries far from the United States is biased strongly in favor of skilled immigrants. The elite status of Hindu, Indian Christian, and black African surnames can be attributed almost entirely to this factor. But other elite groups are the product of events in oth
er countries that led to selective migration by groups within these countries. During the rule of the Shah of Iran, many Iranian students attended U.S. universities. With the revolution of 1979, many of these students chose to stay in the United States, and many other highly educated Iranians fled the new Islamic republic to take up residence in the United States. Similarly, before the fall of South Vietnam to the Communists in 1975, Vietnamese immigration to the United States was negligible. But in the initial years of the new regime, many Vietnamese associated with the previous regime fled, including many skilled and educated families. Other groups, such as the Copts, the Maronites, and Jews, already constituted elites in their home societies. But for the Copts, immigration to the United States has because of immigration policy attracted higher-status Copts, which has made this group even more elite in the United States.
In a recent study, Cynthia Feliciano examines the educational attainment of migrants to the United States relative to the average for their home country. She creates an index of selectivity in migration for each sending country, which is based on the relative education level of migrants to the United States compared to nonmigrants. The correlation of physician frequency in the United States with this measure of migration selectivity is 0.75 for the eleven countries where there is national-level information, as shown in figures 13.10 and 13.11. The countries with the greatest educational disparities for migrants to the United States include Iran and India, which also have among the greatest disproportions in physician frequencies.30
Figure 13.11 shows, in contrast, surname groups underrepresented among physicians in the United States. An interesting variation among the long-established, largely white population is the underrepresentation of Cajun and New France surnames relative to Dutch and English surnames.31 This finding, as discussed in chapter 3, seems to relate to the history of French settlement in North America and negative selection in migration of French Canadians to the United States.
FIGURE 13.11. Underclasses in the modern United States, measured by relative representation among physicians.
The other surnames heavily underrepresented are Cambodian, Latino, black American, Hmong, Mayan, and Native American. The Hmong engaged mainly in subsistence farming in the hills of Laos before coming to the United States. The U.S. Hmong community seems to represent a broad cross section of the Hmong population of Laos, entire communities having moved to refugee camps in Thailand out of fear of the Communist Laotian government and then been admitted en masse to the United States. Thus an entire disadvantaged refugee community in Laos has been transplanted to the United States. As a consequence of these processes, there are eighty times as many physicians per capita with Hindu Indian surnames such as Banerjee or Ganguly than there are with Hmong surnames such as Her, Lor, or Vang.32
For Latino surname groups, selective immigration again seems to be a powerful force shaping social status in the United States. Much of this population originates from Mexico and Central America, and a considerable proportion of the original migrants entered or remained in the United States illegally. Illegal migration should not be an attractive option for educated populations, given the manifest disadvantages that illegal status imposes in the United States. There has been considerable debate about whether Mexican migrants to the United States are negatively selected. Recent research suggests convincingly that Mexican migrants to the United States in recent years have substantially less schooling than nonmigrants and earned considerably less in Mexico before migrating than did nonmigrants.33
We saw above in figure 3.3 that long-established populations descended from European origins, aside from the French, tend to have social status close to the average. These populations mainly arrived in the era of open immigration before 1914. A recent study of Norwegian migration in this period finds that, consistent with this, the forces of selection for emigrants were not strong. There was definite negative selection for migrants from urban areas in Norway. But for the more numerically important rural areas the evidence was ambiguous as to whether selection was positive or negative.34
Thus the accidents of geography, immigration policy, and social and political events in countries around the world are even now creating new upper and lower social classes in the United States. These classes will be a feature of U.S. society for many generations to come, until the processes of intermarriage eventually eliminate these distinctions.
Conclusion
The persistence of high and low status for some groups in various societies would seem to contradict the simple law of mobility for social status. However, in each of the anomalous cases discussed above, there are factors at play that can make even extreme persistence consistent with the same universal tendency for families to regress toward the mean over time. Elites and underclasses seem to be created by mechanisms that select them from the top or bottom of the established status distribution. They can also be created, as in the case of the Gypsy/Traveller community in England, by differential fertility between higher- and lower-status members of a group.
Once established, these differences in social status can be maintained by marital endogamy, as seems to have happened with Christian and Jewish minorities in the Muslim world, or they can be maintained by selective movement between social groups, as in the case of Catholics and Protestants in Ireland.
1 Botticini and Eckstein 2012, 71.
2 Botticini and Eckstein 2012, 80–94.
3 “Families with low-ability sons or with sons who do not like studying … will be less likely to invest in children’s literacy” (Botticini and Eckstein 2012, 93).
4 Botticini and Eckstein 2012, 120. However, a model of selective survival of Judaism among elite Jewish families would show a relatively uniform decline of the Jewish population across different geographic areas. Botticini and Eckstein also emphasize that Jewish populations disappeared from much of the Middle East and North Africa by 650 CE, the surviving populations being concentrated in Mesopotamia and Persia.
5 See Kennedy, Gurrin, and Miller 2012. The discussion here just amplifies their observations.
6 The following surnames were chosen as exclusively Scottish in origin: Bothwell, Buchanan, Cathcart, Fullerton, Girvan, Hamilton, Laird, McGregor, Orr, and Sproule.
7 These Irish surnames are: Boyle/O’Boyle, Doherty/O’Doherty, Grady/O’Grady, Han(n)away, and McBride. Hanaway was included because it is the surname of my maternal grandfather, who himself appears in the census.
8 Though, as Kennedy, Gurrin, and Miller (2012) note, there is a question of causation here: “Whether lower socio-economic status preceded or coincided with absorption into the Catholic community, or gave rise to this outcome in the form of downward social mobility, opens a further intriguing set of possibilities” (p. 21).
9 Botticini and Eckstein 2012, 40.
10 “Scholars Debate the Roots of Yiddish, Migration of Jews” 1996.
11 Clark 2007, 116–121.
12 Van Straten and Snel 2006; Van Straten 2007, 43.
13 There is dissent on this point. Elhaik (2013) reports genetic evidence favoring the Khazarian hypothesis.
14 Ostrer and Skorecki 2013, 123.
15 Gregory Cochran, Jason Hardy, and Henry Harpending (2006) have argued that the elite status of the modern Ashkenazim stemmed from the greater reproductive success of more-intelligent members of the community in Eastern Europe because of their occupational concentration in finance and trade. This is another potential mechanism consistent with the law of mobility. However, greater reproductive success in the preindustrial era by the elites in society is not a pattern unique to Ashkenazi Jews: in preindustrial England, for example, economic success was associated with reproductive success (Clark 2007).
16 Saleh 2013.
17 Once someone converted to Islam, or a child was born Muslim, however, conversion to another religion was punishable by death.
18 Muslims had their own taxes to pay under Islamic law, though these were generally, by design, less burdensome than the jizya.
>
19 Firoozi 1974, 65.
20 Walter, Danker-Hopfe, and Amirshahi 1991.
21 It is not clear whether this same mechanism accounts for the low socioeconomic status of Muslims in India by the time of the British Raj. See Eaton 1993.
22 U.K., Equality and Human Rights Commission 2009, 5.
23 Since I am Scottish and Irish in origin, I know at first hand how one’s supposed cultural heritage can turn out to be largely a modern invention.
24 In line with this hypothesis, genetic testing suggests that the Irish Traveller community is entirely of Irish origin (North, Martin, and Crawford 2000). This article concludes that “these data support that the origin of the Travellers was not a sudden event; rather a gradual formation of populations” (p. 463). There are no equivalent genetic studies of the origins of English Gypsies and Travellers.
25 These first names are from the Romany and Traveller Family History Society, n.d. In later periods, these names are associated with surnames held by many Traveller families.
26 FamilySearch, n.d.
27 Modern surnames associated with the Traveller community in England can be identified by their unusually fast population growth and their high rate of intermarriage with Smiths, since Smith is unusually prevalent among Travellers.
28 There are other signs of the low social status of the Loveridges in recent years. A search on the Internet for recent arrests and convictions in England and Wales showed eight times as many Loveridges as Barclays, even though the name Barclay has about 20 percent more holders than Loveridge.
The Son Also Rises Page 25