Figure 4.1 shows the percentage of Oxford and Cambridge members entering the universities between 1170 and 2012 who had artisan surnames. These surnames are very rare at the universities before 1350, but their frequency increases substantially in the late Middle Ages to achieve its modern level by 1500. Also shown is the percentage of probates in the PCC associated with artisan surnames, shifted back thirty years for comparability with university entrants.3 Finally, the figure shows “elite” PCC probates, cases in which the deceased person was referred to by a title indicating high status, such as Gentleman.
At this scale, medieval England looks like a world of astonishing mobility. Artisans in 1300 were mostly illiterate workers scattered across English villages, yet by 1500 their descendants were fully incorporated into the English universities. And by 1620 they were fully represented even among the gentry whose wills were proved in the PCC. Even before the Enlightenment proclaimed the idea of the fundamental equality of humanity in the abstract, the social and economic system of medieval England was delivering equality of opportunity in the concrete.
FIGURE 4.1. Percentage of artisan surnames among English elites, 1170–2012.
The intergenerational correlation implied by this pattern depends on two things, however. First, how elite was the population at Oxford and Cambridge in this period? One way to calculate its exclusivity would be to look at the share of males in each generation who attended the universities, which in the fifteenth century was 0.3–0.7 percent. This would make the universities potentially as exclusive as the top 0.5 percent of the medieval status distribution.
However, while the universities attracted those seeking a career in the church or administration, there were other career paths for medieval elites. Those seeking a legal career would enroll at one of the Inns of Court in London. Young men aspiring to a career in commerce might apprentice with a merchant or banker. Youths pursuing a military career would train at the tournament and the campaign. As a result, university attendees would have represented a larger share of the population, as much as the upper 2 percent.
The second factor that affects the calculation of mobility rates is the place of artisans in the status distribution of society in 1300. They ranked above unskilled laborers, who constituted a quarter or a third of the society, and above the semiskilled husbandmen of the farm sector. But they ranked below the many landowners, manorial officials, farmers, clerics, merchants, civil servants, and attorneys.
Here persistence rates are calculated assuming that artisans started between the fortieth and sixtieth percentile from the bottom of the socioeconomic distribution. The higher the starting position of artisans on the social ladder, the lower the estimated mobility rates. Assuming Oxford and Cambridge students represented the top 0.5–2 percent, and that artisans were at the median or the upper fortieth percentile of the status distribution, the persistence rate implied by figure 4.1 lies between 0.77 and 0.85. Figure 4.2 shows the best fit of 0.8 for the preferred assumption: Oxford and Cambridge represented an elite of 0.7 percent of the general population, and artisans started in the middle of the status distribution. The figure also shows that there was no possibility that the intergenerational correlation was as low as 0.7 or as high as 0.9.
This finding means that medieval England had mobility rates similar to, though perhaps modestly higher than, those of the modern United States and Sweden. In terms of social mobility, then, what did the Scientific Revolution, the Enlightenment, and the Industrial Revolution achieve? Very little. Social mobility existed long before people even thought of it as a feature of the good society. It was never fast, but over generations, all ranks of society could enter equally into its upper echelons. By implication, the early elites eventually saw substantial downward mobility.
FIGURE 4.2. Alternative persistence rates for medieval England versus data, Oxford and Cambridge students.
A good example of the operation of social mobility in these years is Geoffrey Chaucer (1343–1400), the celebrated author of The Canterbury Tales, itself a satirical commentary on the various strata of English social life (figure 4.3). The surname Chaucer is believed to derive from the French word for a shoemaker, chausseur. But the Chaucers had left shoemaking generations before Geoffrey. Both his father and grandfather were prosperous vintners, a higher-status occupation. Though he was of completely common background, Chaucer became, by dint of his abilities, a courtier, diplomat, and civil servant in the court of Richard II.4 By 1386 he was member of Parliament for Kent, and by 1389 clerk of the king’s works, an important administrative position. His son Thomas was later speaker of the House of Commons. Thomas’s daughter Alice married William de la Pole, Duke of Suffolk. Geoffrey’s great-great-grandson, John de la Pole, Earl of Lincoln, was named heir to the English throne by Richard III in 1485 (but executed after Richard’s overthrow). Thus during the medieval period, the Chaucer family went from shoemakers to claimants to the English throne, albeit under a different surname.5 Chaucer’s family history conveys the fluidity and flexibility of medieval English society, as opposed to the common stereotype many have of a rigid and static preindustrial social world.
Since in the medieval period artisan surnames would have connoted low status, there is some chance that as people rose in social position, they adopted higher-status surnames. A man could do this by taking his wife’s surname at marriage or adopting the surname of a patron. English citizens may have lost the right of their American cousins to wield death-spitting weaponry, but they have long held the right to change their surnames to whatever they wish by personal fiat, as long as there is no intention to deceive. If this practice was common, the rate of mobility measured through surnames would be lower than the actual mobility rate, since there would be spurious continuation of the status of surnames. So for artisans, we can be confident that the mobility rate for the Middle Ages measured by surnames is the lower bound of actual mobility rates.
FIGURE 4.3. Portrait of Geoffrey Chaucer ca. 1415–20 by his friend, the poet Thomas Hoccleve.
The Decline of Elites: Locative Surnames
If medieval artisans enjoyed upward mobility, what signs are there of the concomitant downward mobility of thirteenth-century elites? A large elite group of surnames in medieval England are those surnames that came from town and village names: locative surnames.
In preindustrial England, where most people lived in one place all their lives, identifying the average person by the name of their town or village would make no sense. However, among the elite, who left their places of origin to go to court, to the universities, to the religious centers, and to the towns and cities to work as merchants, lawyers, and bankers, the most typical surname was one that identified their ancestral home or place of origin.
This locative naming practice started with the Norman conquerors of England. This new elite took surnames that linked them to their home villages in Normandy, such as Mandeville, Montgomery, Baskerville, Percy, Neville, and Beaumont.6 But as the Norman elite was gradually displaced by an indigenous English propertied class, new locative names associated with high status appeared: Berkeley, Hilton, Pakenham, and so on.
These surnames are prominent in the early records of Oxford and Cambridge: they account for nearly half of the names associated with the universities in the thirteenth century. But these surnames were a much smaller share of the overall population stock of surnames. The frequency of high-status surnames tended to increase in preindustrial England until 1800. Thus while the locative surnames used here account for 7.1 percent of all surnames among marriages from 1800 to 1829, they account for only 6.7 percent in 1650–79 and 6.1 percent in the period 1538–59. Projecting backward from the growth rate by generation between 1538 and 1800 gives an estimated share of 5 percent in 1250.
The advantage of using these locative surnames as a measure of mobility is that they represent a large share of the stock of all surnames, and most of them are not associated with any notable status or distinction. The most common locative sur
names, for example, are Barton, Bradley, Greenwood, Newton, Holland, and Walton. Such names would not themselves influence the status of the holders.
Figure 4.4 shows the relative representation of a sample of locative surnames at Oxford and Cambridge from 1200 to 2012, calculated, as before, as the ratio of their share in the universities to their share in the general population. Until 1350, the relative representation of these surnames remains close to four. The reason for the absence of any downward mobility for these surnames in these years is likely that locative surnames were still being adopted by people of higher status. So the initial period for measuring rates of social mobility is taken here as 1320–49. We see above that artisan surnames did not begin to displace other surnames at the universities on a significant scale until after 1350.
FIGURE 4.4. Locative surnames at Oxford and Cambridge, 1170–2012.
From 1350 all the way until the present, locative surnames decline steadily in relative representation at the universities. One persistence rate, 0.86, turns out to fit the data well for the period 1320–2012, a span of almost seven hundred years. But this is a high persistence rate. It implies that nearly three-quarters of the variation in general social status across families derives from inheritance in any generation. Thus mobility was consistent even in the Middle Ages, but very slow.
The estimated intergenerational correlation for the downward mobility of the locative surnames is quite consistent with the 0.75–0.85 range suggested for persistence in the upward mobility of the artisan surnames.
Thirteenth-Century Property Owners
An even more elite group of thirteenth-century surnames is the rarer surnames belonging to landowners appearing in the Inquisitions Post Mortem (IPM) for the years 1236–99. These were inquiries held at the death of tenants of the English crown to establish what lands they had held and who should succeed to those lands. The holders of these properties were typically members of the medieval upper classes. The surnames chosen for this sample were those of decedents who transmitted the largest average amounts of property to their heirs. They were also surnames for which the modern English form of the surname is known or easily inferred.
Table 4.1 lists some sample surnames as they appear in the IPM and in their modern forms. Many of these surnames, being derived from the names of French villages, lost their meaning to the later English-speaking bearers of the names and thus mutated into names with a similar sound that had a meaning in English. Taillebois, for example, is now sometimes Tallboys.
It is necessary to follow all possible derived variants of the initial surname, because the less well-connected variants are likely to be associated with lower-status descendants of the original surname holders. Baskerville, for example, began as the Norman surname de Basqueville, from a village in Normandy. By the thirteenth century it had become de Baskerville. In a further mutation, it became Baskervilde and then Baskerfield. The -field variant is lower in status on average then the -ville variant: this difference is predictable because the mutation to -field was much more likely to occur among lower-status and illiterate holders of the surname.
TABLE 4.1. Some medieval surnames from the IPM and their modern English equivalents
IPM
Modern
De Bello Campo, De Beauchamp
Beauchamp, Beaucamp, Beacham
De Berkele, De Berkelegh, De Berkeley
Berkeley, Barclay
De Kaygnes, De Kaynes, De Caynes, De Keynnes, De Kahanes, De Keines
Keynes, Kaynes
De Menwarin, De Meynwaring, De Meynwaryn
Mainwaring, Manwaring
De Mortuo Mari, De Mortymer, De Mortimer
Mortimer, Mortimor
Taillebois, Tayleboys, Talebot, Talbot
Talboys, Talbot, Talbott, Tallboy
Many of the surnames of the English elite in the thirteenth century originated as surnames of the Norman conquerors of 1066. But in the two intervening centuries, a new class of English property owners had also emerged, such as the rich and influential Berkeley family.7
Figure 4.5 shows the status over time of this group of surnames as represented by their incidence among students at Oxford and Cambridge. As expected, this is an even more elite group of surnames than those shown in figure 4.4, which are simply associated with places. The IPM surnames peaked in status in 1230–59, when they were thirty times more common at the universities than in the general population. After that, they immediately began regressing to the mean. They show a persistence rate very similar to that of the locative surnames until about 1500.
Had that rate of regression to the mean been maintained, then by 2012 these surnames would have been only 14 percent more frequent in the top 1 percent of the status distribution than an average surname. But after 1500 the rate of regression to the mean slowed down further, and from 1500 to 2012 the persistence rate that fits the data throughout these years is 0.93, an extraordinarily high number. This implies that modern England actually has lower rates of social mobility than medieval England did. Surnames loosely associated with the rich of the thirteenth century still appear among Oxford and Cambridge students at a rate 25 percent more than expected 1980–2009. Since these results differ from those for the locative surnames, we have to consider other possible explanations of this outcome.
FIGURE 4.5. Incidence of surnames from the Inquisitions Post Mortem (IPM) at Oxford and Cambridge, 1170–2012.
One important possibility is that elites deliberately adopted these high-status names in recent centuries. When a high-status man with a common surname such as Smith married a woman with a high-status surname such as Darcy, he might choose to adopt her surname after the marriage rather than follow the convention of the woman adopting the husband’s surname.
Consider, for example, the Stanley family, the earls of Derby and descendants of an original medieval Stanley. In the eighteenth century, seeking cash through a matrimonial alliance with a rich heiress, Lucy Smith, Stanley became Smith-Stanley. In the later nineteenth century, the déclassé Smith was again dropped from the family name. Such selective name changing could give an artificially low impression of downward mobility by holders of high-status medieval surnames.
The Norman Conquerors
We can follow one group of surnames even further back than the propertied elite of the thirteenth century. These are the names of the Norman conquerors of England in 1066 who are recorded as property holders in the Domesday Book of 1086. What happened to this medieval super-elite?
Once he had vanquished Harold and his followers in battle, Duke William of Normandy’s first order of business was to secure his hold on the English throne by granting lands and positions to the adventurers from Normandy, Brittany, and Flanders who had won it for him. There was thus a wholesale replacement of the former Saxon upper class with a new Continental elite.
The Domesday Book, the record of property holders in England under William, records for the first time the names of the Norman adventurers. Detailed work in the arcane historical field of prosopography by the splendidly named Katharine Stephanie Benedicta Keats-Rohan has established the origins of many landholders in the Domesday Book.8
It is not clear how heritable surnames were in the eleventh century. However, of the nearly five hundred surnames identified in the Domesday Book, many seem likely to be unique to the Norman, Breton, and Flemish upper classes that dominated England after the conquest. Many of these surnames, which were associated with the holders’ villages of origin, have disappeared, but some have survived into modern England. Table 4.2 shows a sample of these names in the form that appears in the Domesday Book and in their modern form.
The table shows that one of these surnames, Sinclair, had more than seventeen thousand bearers in England and Wales in 2002. Could all of these people have descended from one family, or a small group of families? Once again we witness the amazing power of exponentiation. Thirty-one generations elapsed between 1086 and 2002. A population of seventeen thousand people with a g
iven surname in 2002 implies 8,500 males. For one forefather to produce 8,500 descendants in the course of thirty-one generations would require that each generation produce an average of only 1.34 surviving sons per family. There is evidence that the upper classes of preindustrial England easily achieved such fertility levels.9 So even these outlier surnames with large numbers of holders could have descended from one family, if it was consistently reproductively successful over the generations.
TABLE 4.2. Some Norman surnames, 1086 and 2002
Original
Modern
Number in 2002
Baignard
Baynard
54
De Belcamp
Beauchamp, Beacham
3,252
De Berneres
Berners
49
Burdet
Burdett
3,973
De Busli
Busly
52
De Cailly
Cailey
32
De Caron
The Son Also Rises Page 8
