SOURCE: Bartlett, Robert. “Medieval and Modern Concepts of Race and Ethnicity.” Journal of Medieval and Early Modern Studies 31, no. 1 (winter 2001): 39-46.

[In the following excerpt, Bartlett examines William's use of language to describe and differentiate people by race, nationality, and ethnicity.]

Historians working in the present day, just like their medieval and early modern predecessors, are confronted with difficult choices when they write of human population groups.¹ When, if at all, is it reasonable to employ the word race, the word nation, the word tribe? What collective term best describes, say, the Goths, the English, the Jews? What meaning does the concept “ethnic identity” have? It is hard to do without some collective terms, but neither the medieval nor the modern terminology of race and ethnicity is simple or uncomplicated. Even the distinction between those two central terms, race and ethnicity, is drawn in different ways by different people. In the United States both popular and official usage tends to associate race with the troubled history of white and black, while the term ethnicity summons up Italians, Irish, or Greeks, for example. Hence the former term suggests a distinction based on an inherited biological feature, skin color, while the latter points to cultural differences between groups. Recent large-scale immigration into the United States from Asia and Latin America has complicated the issue by posing the question of whether the categories Oriental and Hispanic belong to the question of “race” or the question of “ethnicity.” For the historian, such usage is to be regarded as an interesting fact about the intellectual and political history of our own times but cannot itself provide a tool of analysis. The expedients of the U.S. Census or Immigration administration are no starting point for scholarly inquiry.

Among social scientists of the present generation ethnicity has a different set of connotations. For them it serves as an acceptable alternative to race, a word that many consider permanently unusable because of its association with racism. In the discourse of the social sciences, the word ethnicity with this meaning is recent, the first occurrence recorded in the Oxford English Dictionary dating to 1953. (An earlier meaning was “paganism,” but this is not relevant here, although it is impossible not to cite this stunning instance from 1782: “From the curling spume of the celebrated Egean waves, fabulous ethnicity feigned Venus their idolatress conceived.”)

The first United Kingdom census to ask questions about ethnicity was that of 1991, and much interesting empirical data was obtained as a result. An attempt was also made to address the theoretical or conceptual issues involved in using the term. In the report Ethnicity in the 1991 Census published by the Office of National Statistics, the question is asked, “Do we really ‘know’ what ‘ethnic’ actually means?” The scare quotes around “know” and “ethnic” warn that the author wishes simultaneously to assert something and to retract it, and such hesitancy characterizes the discussion in this report. The author outlines three possible approaches to ethnicity: it is primordial, it is a constantly changing sense of group identity, it is situational. In the last case, the author writes, “there may be no single, unambiguous ‘true answer’ to a question about one's ethnic identity.”²

It is hard to countenance the idea that there could be a “true answer” to a question about one's ethnic identity. Ethnic identity results from a process of labelling (identification). This may be self-labelling, but labelling by others is also involved, since ethnic identity may be contested. In the 1930s many people who considered themselves Germans were told they were not; they were Jews instead. It made a good deal of difference which label stuck. This labelling and self-labelling is also strategic and situational. To identify oneself or others in this way is almost invariably to claim something or deny something. To call oneself, or be called, “black” or “British” or “Irish” or “Jewish” is not a neutral statement of the obvious but a political and historical assertion, with implications for one's rights and relationships. Different identities can be asserted in different situations. It is certainly not the case that a “true answer” could be found.

The author of the theoretical discussion in Ethnicity in the 1991 Census concludes:

Ethnicity … is clearly not amenable to static categorization … but despite its complexity it could at least be argued that it represents one possible way of conceptualizing social divisions and cleavages. The ontological status of “race” is, however, much less secure. There is now almost universal agreement in the literature that the biological notion of race should be rejected in favour of a view which sees the term as essentially merely a social construct.³

The “literature” referred to here is that of the current social sciences, where race is viewed as the bad old word and ethnicity as the acceptable new one. What is astonishing in the passage is the idea that race is “merely a social construct,” with the implication that ethnicity is not.

An apparent parallel to the race/ethnicity tangle can be brought in at this point in the attempt to clarify issues. This is the sex/gender distinction. Those who use the terms sex and gender carefully are seeking to distinguish a chromosomal, biological distinction between people, something they are born with, and the forms of sexual identity they are socialized into. Breasts are one thing, lipstick another. This cannot be the distinction between race and ethnicity. As opponents of racism have repeatedly pointed out, there are no pure races; there are no clear-cut ways of grouping human beings into discrete biological populations. Ethnicity does not stand in the same relation to race as gender does to sex. Put another way, both race and ethnicity can only be at the gender end of the polarity.

This is not to deny that there is genetic and biological variation between human groups. Much of this is invisible, but some is highly visible: color of skin, shape of eye, type of hair. These biological differences do not themselves constitute race or ethnicity but are part of the raw materials from which race or ethnicity can be constructed—along with language, religion, political allegiance, economic position, and so on. The significance attached to visible genetic markers varies from society to society—the difference between the meaning of skin color in Brazil and in the Old South being a locus classicus in discussion of such issues.

Our own usage should be based on the intellectual value of the distinctions our terms embody. Dictates of fashion are obviously strong. Just as some people use gender simply as the acceptable modern version of sex and talk of gender ratios in the human population, so others feel, perhaps from noble motives, that the word race should be banned from our discourse. Neither group has enriched the terms of analysis. Revising our vocabulary is not in itself meritorious. To distinguish sex and gender is to observe the world in a more nuanced way. To relabel sex as gender has the intellectual value of preferring pine furniture to mahogany.

As regards the terms race and ethnicity, it is better to be both more radical and more reactionary. It must be possible to reclaim the word race from the racists. It is a short, everyday word. As long as it is made clear that race is not a biological category, then it is stylistically preferable to speak and write of races rather than of ethnic entities. Ethnicity and race both refer to the identifications made by individuals about the groups they belong to. If one word has a use, then the other does. One is not the dark side of the force in a Manichean dualism. For the rest of this discussion racial and ethnic will be treated as synonyms.

Before turning to the usage of medieval authors, there is one other large conceptual issue that must be addressed. This is the problem of distinguishing groups and identities of an ethnic or racial kind from religious ones. Especially in a period like the Middle Ages, when religion meant membership of a community much more than adherence to a set of principles or beliefs, there was a sense in which one was born a Christian, a Muslim, or a Jew, just as one was born English or Persian. The shape of ethnic identity and ethnic strife and that of religious identity and religious strife are identical, as some modern conflicts, such as those in Northern Ireland or Bosnia, show. Indeed, in Bosnia, the tendency has been to see a three-cornered fight between Croats, Serbs, and Muslims—two ethnic categories and one religious category. Noel Malcolm's superb Bosnia: A Short History refuses to be drawn by the idea that these ethnicities are primordial. Indeed, when talking of the period prior to the nineteenth century, he avoids the terms Croats and Serbs altogether, referring instead to Catholic Bosnians and Orthodox Bosnians, alongside the Muslim Bosnians. If anyone were to doubt that ethnicity were situational and strategic, his example of the twentieth-century Bosnian Muslim leader, who had two brothers, one classified as a Serb, the other as a Croat, might help to convince.⁴ We do not have a simple word to denote both ethnic and religious identities and groups, but it is hard to find good intellectual grounds for distinguishing them.

The medieval terminology of race and ethnicity was no more straightforward than our own. Some of the key terms of medieval Latin usage, such as gens and natio, imply, etymologically, a concept of races as descent groups. Others, such as populus, do not. The actual semantic field of such terms can only be mapped by detailed investigation of individual usage. As an example of one medieval writer's mental and verbal habits and also of the problem in rendering those medieval patterns accessible to modern readers of English, the following section looks at William of Malmesbury's Gesta regum Anglorum (Deeds of the Kings of the English) and its recent translation in the respected series Oxford Medieval Texts.⁵

William, who was a monk at the Benedictine abbey of Malmesbury in southern England, wrote his history, which is a source of primary importance for the Anglo-Saxon and Norman period, in the 1120s. The word gens and its grammatical variants occur about one hundred times in the Gesta regum. It is obviously a basic component of William of Malmesbury's view of the world. The range of the term can be suggested by looking at the various ways it has been translated in the modern edition and translation. Primary responsibility for the translation belongs to Sir Roger Mynors (1903-1989), whose work was done in the 1950s and subsequently lightly revised by scholars of a younger generation, so the preferences are those of a classically educated Oxford scholar whose own university education took place in the early 1920s. He was expressly committed to a method of translation that aimed at good literary English even at the cost of literalness.

The most neutral possible translation of gens is “people,” and this occurs frequently in the English version. Perhaps more highly charged is “nation” and this too is a common choice. Thus William's gens Anglorum tends to appear as either “the English people” or “the English nation.” Sometimes the yet more specific “nationality” is preferred, as when William the Conqueror decides to promote no Englishman to high office in the church. The chronicler's nullum eius gentis becomes “no one of English nationality” (3.254; 470-71).

“Race” is not an especially favored translation, but it does occur. The languages of the Franks and of the Anglo-Saxons are similar because “both races” (ambae gentes) originally came from Germany (1.68; 98-99). Sometimes the context exercises a semantic pull. When we read of the depredations and attacks barbarorum paganorumque gentium, it is easy to see how the rendition “of barbarous and pagan tribes” could slip from the pen of an Oxford classicist. What is barbarous and pagan is tribal (2.114; 172-73).

The association of the word gens with biological descent is visible in more than one English counterpart. A barbarian is described as gente et animo barbarus—“a barbarian in blood and behaviour”; a noble Normannicae gentis is “of Norman stock”; while Bohemond, the Norman leader in southern Italy, who is loco Apulus, gente Normannus, is “Norman by family,” not, more mnemonically, “Apulian by place, Norman by race” (2.134, 145, 4.349; 212-13, 232-33, 608-9). “Blood,” “stock,” and “family” thus stress the breeding and pedigree connotations of gens.

William of Malmesbury's usage sometimes suggests that there may be gentes of more than one kind, specifically that one gens may be a subdivision of a larger gens. He is willing to call the Northumbrians, Mercians, East Anglians, and men of Kent gentes, but also refers continually to the gens Anglorum. A member of the “Kentish nation” (gens Cantuariorum) was also presumably a member of the “English nation,” hence having more than one ethnic identity simultaneously (1.88; 128-29). Nor do gentes have to be immutable. After the Viking conquests in northern and eastern England, the Northumbrians and the Angles coalesced with the Danes into one gens (2.125; 196-97), an example of what early medieval historians studying the Germanic and other peoples have termed ethnogenesis.

This brief survey of the use of one racial or ethnic term by one medieval author and its translation by one translator cannot, of course, enlighten more than a corner of the topic, but it highlights a few of the issues and may serve as a kind of bridge between discussion of modern and medieval terminology. The translator is confronted with a foreign language that has to be rendered into another tongue. The fundamental question concerns the closeness of the fit between the two sets of terms, foreign and native. Some might argue that a foreign word, like gens, should be translated by the same English word on each occasion, others, like Mynors, that it either cannot or need not be. If we do wish for word-for-word translation, there is the tricky issue of which English word to chose. As we have seen, in the space of one work by one author gens can be rendered “race,” “nation,” “people,” “tribe,” “stock,” or “family.” Mynors himself recognized that translation is not only an “enjoyable art” but also a “perilous” one.⁶

Perhaps this issue has been labored. Yet we must consider that it is very unlikely that William of Malmesbury's Gesta regum will be translated into English in its entirety again within the next century and, as Latinity becomes more and more a minority attainment, that translations of Latin texts will be the gateways to the culture of the Middle Ages much more than the texts in the original. For many generations Mynors's choices of words will dictate the impression the reader gains of this text. It may well be that there is no misrepresentation in imagining William of Malmesbury writing of “the English nation” or “barbarous tribes,” but for most readers there will be no way of checking. The complex and individual contours of William of Malmesbury's gens have been cloaked by the translator's choices of ethnic and racial terms from modern English.

To move away from the individual case to a wider consideration of ethnic and racial terms among medieval writers, one can clearly see how such words as gens and natio fit in well with what can be called the genealogical idiom of much medieval thinking. This was a world in which blood and descent were seen as fundamental. A noble was generosus or gentle or gentil—“well born.” A serf was a nativus—“born unfree.” Kin solidarities were central in shaping patterns of property, power, and violence. To people who constantly saw the fate and fortunes of individuals determined by their birth and descent, it was natural to conceptualize humanity in similarly genealogical terms.

One frequently held theory proposed that the different gentes descended from the three sons of Noah, with Shem's descendants receiving Asia, Ham's Africa, and Japheth's Europe. A further theory postulated that there were a total of 72 races, paralleling the 72 apostles sent out by Jesus according to chapter 10 of the Gospel of Luke in the Vulgate version. Moreover, most races had etymological founders, like Scota, daughter of Pharoah, ancestress of the Scots, or Brutus the Trojan, who gave his name to Britain and from whom the Britons descend. These various strands—the three sons of Noah, the 72 races, and etymological ancestors—could be combined into a complete family tree of all humanity.

This immediately points to a latent idiosyncrasy in the medieval genealogical idiom. For while the starting point or premise of such genealogical-ethnic thinking was that each race was a group of human beings of common biological descent, the specific biblically inspired form it took also maintained the common biological descent of all human beings, from Adam and Eve and from Noah and his wife. In Augustine's words, “Whoever is anywhere born a man, that is, a rational mortal animal, no matter what unusual appearance he presents … no Christian can doubt that he springs from that one protoplast … if they are human, they are descended from Adam.”⁷ Hence, although it may seem that medieval racial thinking was just as biological as modern pseudoscientific racism, it had as inbuilt components the common descent of all human beings with, as a necessary corollary, the implication that races had developed over the course of time.

In fact, while the language of race in the Middle Ages may often seem primarily concerned with descent groups, a closer look shows that this genetic component was often overshadowed by considerations of a different order. Two of the most significant are, at the level of learned theory, the importance of ideas of environmental influence and, more generally, the consistent emphasis on the cultural and social component of ethnic identity.

The learned theory of the Middle Ages drew on an ancient tradition of geographical determinism. As one of the founding texts, the Hippocratic Airs, Waters, and Places, put it, “In general you will find assimilated to the nature of the land both the physique and the characteristics of the inhabitants.”⁸ This belief, that climate and geography shape the bodies and characters of the different peoples, is obviously not logically incompatible with the theory that races are descent groups, but it stresses quite different things and contradicts the idea of a constant national character.

Notes

1. Some of the issues discussed in this essay are addressed in Robert Bartlett, The Making of Europe: Conquest, Colonization, and Cultural Change, 950-1350 (Princeton: Princeton University Press, 1993), 197-242.

2. Ethnicity in the 1991 Census, 4 vols. (London: Office of National Statistics, 1996), 3:2-3.

3. Ibid., 3.

4. See Noel Malcolm, Bosnia: A Short History (New York: New York University Press, 1994), xi, 166.

5. William of Malmesbury, Gesta regum Anglorum, ed. and trans. R. A. B. Mynors, R. M. Thomson, and M. Winterbottom, vol. 1, Oxford Medieval Texts (Oxford: Clarendon Press, 1998). Book and section are given first, followed by page numbers.

6. Quoted in Michael Winterbottom's obituary of Mynors, Proceedings of the British Academy, “1991 Lectures and Memoirs,” 80 (1991): 371-401, at 394.

7. Augustine, De civitate Dei 16.8; ed. B. Dombart and A. Kalb, Corpus Christianorum, Series Latina 47-48, 2 vols. (Turnhout: Brepols, 1950), 2:508-10. Translations here and following are my own unless otherwise noted.

8. Airs, Waters, Places, in Hippocrates, trans. W. H. S. Jones, Loeb Classical Library (London: Heinemann, 1923), 24.