Ethnicity and Conflict in the Roman Conquest of Spain

Lyra D. Monteiro

People do not kill one another because their customs are different. Abner Cohen (1969: 200)

Introduction The integration of the Iberian Peninsula into the Roman Empire took an unusually long time, from the initial Roman engagement on the Mediterranean coast, which served as an excuse for the Second Punic War in 218 B.C.E., through to the final conquest of Galicia under Augustus, in 19 B.C.E. Modern scholarship on this period has been heavily based on the work of various ancient authors, most of whom wrote centuries after the events they describe. This creates familiar problems. Due to the nature of Greco-Roman historical writing, these sources reflect a palimpsest of the opinions, attitudes, and knowledge of the earlier writers on whom they base their work, all screened through the individual biases of the authors themselves. Thus, they cannot be taken as an accurate description of the state of things at any one time. If we want to assess the importance of different identities during this period of conquest and integration, we need to incorporate non-literary sources, including work on more recent empires, to help us consider the range of possibilities that could have shaped the context we are studying. In this paper, I explore the utility of concepts of cultural difference, ethnicity, and ethnic conflict for understanding the situation in early Roman Iberia.

The use of the concept of ethnicity in Romanization debates has opened up exciting new avenues in how we study the changes that occur with the creation of the Roman Empire. Such modern phenomena are ‘good to think with,’ and we should be careful to take advantage of scholarship from other disciplines, which study these phenomena directly in the present. At the same time, we must be sure not to extend our analyses in directions that they simply will not go. A further challenge is presented by areas such as Roman Spain, where questions about the ethnicity of various populations have been a standard aspect of scholarship for quite some time. This legacy requires that we critically re-examine our starting assumptions in light of new ways of defining and thinking about ethnicity. For example, scholars of the archaeology of the Roman period in Spain tend to assume that there are two options available to people, particularly in the pre-Augustan period: they can act like natives, or they can act like Romans (e.g. García y Bellido 1963; Mierse 1999). This presupposes the significance of ethnic identity in this context, something that instead needs to be interrogated and proven. Unpopular as it is in this era of post-colonial guilt, I would like to suggest that the key to understanding the Roman Empire lies in the ability of the Romans to absorb new populations. Such a position need not imply either coercion on the part of the Romans, or patronizing assumptions about the inherent superiority of the Roman way of life. Rather, I will suggest that for the Roman ‘colonial style’, ethnicity simply did not matter in the majority of cases, and as a result, people were able to let go of the characteristics that marked them as distinct (there are, of course, exceptions, such as the ‘ethnic soldiery’ described in Batavia by van Driel-Murray 2002). While some cultural differences remained, the use of these differences as markers of ethnic identity gradually disappears. Scholars who have argued that the Roman Empire lacked the power to coerce assimilation are correct – so are those who argue that there was something ‘in it’ for the native elites to Romanize themselves. What both of these points of view neglect are the specific political circumstances that would have allowed this kind of open and fluid adoption of cultural traits, without a need to maintain distinct ethnic identities.

The role of colonial style in shaping ethnic relations in populations under imperial control has usually been studied in the context of modern European colonialism. This topic has particular poignancy and urgency due to the prominence of ethnic conflict in recent years, which has been recognized as ‘the heritage of the great wave of decolonisation through the 1960s’ (Williams 1994: 51). One approach to addressing these issues is the comparison of British and French approaches in governing their African colonies, by Blanton, Mason and Athow. They find that ‘differences in the strategy of colonial administration contributed to differences in the systems of ethnic stratification and ethnic conflict that emerged in each system after independence’ (Blanton et al. 2001: 477). The British strategy involved playing pre-existing groups off of one another while keeping all at a more or less equal position, and led to a post-colonial situation in which all the ingredients for ethnic conflict were present (ibid.: 479–480). The effect of this unranked ethnic system is visible both in an increased frequency of ethnic clashes, as well as in clear geographical concentration of different ethnic groups. The French, on the other hand, attempted to thoroughly replace pre-existing authority structures, which had the effect of rendering impotent most ethnic groups, leading to a very different post-colonial situation (ibid.: 475, 478–9). Thus, it is not merely a matter of looking for ethnic differences to anticipate clashes. The structure of relations between ethnic groups varies according to the rules by which they are asked to play. These differences show the danger of assuming a uniform effect of empire upon ethnicity, even within the modern world.

In approaching ethnic conflict archaeologically, we should think carefully in terms of three different classes of phenomena: cultural difference, ethnicity, and ethnic conflict. Each term presupposes the one before it, but not the other way around – for example, the existence of cultural difference does not indicate ethnicity, nor does ethnicity imply violent ethnic conflict (Nagel and Olzak 1982: 130). For the sake of clarity, I will begin by differentiating cultural difference from ethnic identity, and finally consider the conditions necessary to have a state of ethnic conflict.

Cultural Differences Cultural differences, sometimes called ‘ethnic features’, represent basic differences in the materials and beliefs that shape one’s way of life. Clear cultural differences are to be expected, particularly in colonial situations, where groups of people arrive in a dramatically different cultural area (van den Berghe 1996). However, these are not inevitably transformed into ethnic groups (Barth 1969: 17–18). If external settlers move into a conquered territory, a degree of material culture differentiation may be visible at first, but such initial signs of distinct population groups should not be mistaken for signs of long-term ethnic differences. The first one hundred years, or around three generations, following incorporation into an empire seems to be the crucial length of time for things to ‘settle down’ into the pattern that a given colonial style favors. During this time, sharply defined cultural differences may gradually dull, as memories of being on opposing sides fade, and a new, more inclusive identity may be formed. Alternatively, these distinctions could be maintained, either by culturally defined group boundaries; legal measures imposed by the empire; or using phenotypical features, such as skin colour or height. In any event, ethnicity is signaled through material culture, so we would need to find, after that hundred-year transition period, a continued cultural division between groups in order to justify any discussion of the significance of ethnicity (cf. Barth 1969: 15–16).

No matter how much we feel the need to emphasize value of the unique cultures of the peoples whom the Romans encountered and often assimilated, this should not lead us to form analytical ethnic groups on which to base discussion of any form of ethnic conflict. Cultural markers can be important in the definition of ethnicity, but culture change can occur independent of changes in ethnic identity (Barth 1969: 12–13). The most widely noted cultural changes in Roman Spain – imported Roman goods, Roman-influenced houses and temples, adoption of Latin language and Roman practices of dedicatory inscriptions, use of Roman or latinized names, changes in burial rites, adoption of Roman coinage, etc–do not necessarily tell us anything about ethnic identity, without further analysis. Rather than diligently seek out evidence of the persistence of individual cultural elements of the pre-Roman material culture of an area, and then characterizing this as evidence of the survival of local ethnic identity (e.g. Mierse 1999; Webster and Cooper 1996), we ought to allow for the possibility that the social structure created by the Roman Empire was not one which allowed for, encouraged, or rewarded ethnic differentiation. Only where these cultural differences persist over time, and present clear patterns, can we argue for the development of ethnic differences as part of the colonial system.

Ethnicity Ethnicity, as opposed to cultural differences that vary more widely, is a social identity that is always developed through interaction with other groups under particular kinds of power relations. This sort of situation often occurs in imperial contexts, in which a group in power perceives itself as having a different origin, character, etc, from groups that it controls. This may seem like an inherent part of imperial ideology, but in the case of the Romans, we should be careful of assuming that it was. Roman imperial ideology contained within it the ability to incorporate almost any new population it encountered – while universal citizenship did not arrive until 212 C.E., provisions for foreigners and even slaves to become citizens were in place as early as the republican period (Borkowski and du Plessis 2005: 97–103). This may have had something of a ‘non-othering’ effect, which would have affected the development of ethnicity in places under Roman control.

Ethnicity is not a constant feature of human cultures, but neither is it an invention of the modern era. That is, there is a latent potential for ethnic sensibility in almost any population, but whether or not it is activated, and the level at which it is advantageous to activate it, vary with the context. This is not a reference to the multilayered nature of identity (Meskell 2001), but rather an argument that the combinations in which those layers appear and the importance which different aspects of those identities can have is not random, externally predictable, or entirely up to individual choice. That is, as logical as it seems to us that a particular group will have clearly defined gender roles, age grades, and a sense of ethnic identity, depending on the precise structure of their society, these things may matter to different degrees: ‘Which of these identities becomes salient at a particular moment depends upon the situational constraints and the strategic utility attached to the identity’ (Nagel and Olzak 1982: 129; c.f. Barth 1969: 18). One implication of this point of view is that loss of ethnic uniqueness or identity is not necessarily something to be lamented. It only matters if we think that ethnicity is something that ontologically exists and has inherent value, rather than something that is important under very specific circumstances, and is not under others. The higher the level of opportunity in an empire, the more likely it is that local ethnic identities will not be, as it were, required, as people can have opportunities beyond them, and would only be limiting themselves by arguing that they should be treated differently because of their origin.

Ethnicity, as much as we have come to see it as a basic and inherently valuable kind of group identity, is not encouraged to the same extent by all imperial systems. Even when ethnicity matters a great deal, the freedom of choice which individuals have, whether or not they choose to identify with a particular ethnic group, varies greatly, since, as noted by Robin M. Williams ‘states are major actors in creating, accentuating, or diminishing ethnic identities’ (Williams 1994: 49). At one extreme, apartheid South Africa being an example, Thomas Eriksen sees that ethnic identity ‘becomes an imperative status, an ascribed aspect of their personhood from which they cannot escape entirely’ (1993: 5–6). At the other end of the spectrum, Juan Diez Medrano suggests that for modern Basque identity, ethnic belonging is as simple as living and working in the Basque country (1994: 877). These possibilities for the way in which ethnicity operates do not vary randomly, of course–an essential difference between these cases is that in the former, phenotypical features are key to the identification of ethnicity, while in the latter, they are not. When the features of ethnic identification are purely cultural, many scholars have found that, as summarized by Diez Medrano, ‘unless they face strong discrimination or prejudice, members of an economically disadvantaged group will tend to adopt the ethnic identity and political behavior of the economically advantaged ethnic group’ (1994: 876). Such ‘ethnic switching’ has been noted in many contexts (Nagel and Olzak 1982: 129–130), and this suggests that, under certain circumstances, we should not expect different origins to be maintained and manifested in the form of different ethnicities.

The Roman colonization of Spain had much more in common with the Basque example than with that of South Africa, and indeed it is difficult to suggest that phenotypically based ethnic categories (which can be described as ‘racial categories’) could have existed in this early phase of the empire. Since culturally based ethnic categories, even when they do exist, are much harder to maintain over time than those based on phenotype, it is hard to support the assumption by many scholars that categories of native and Roman still have relevance in the imperial period in Spain. We know that from the very beginning, in every province, native elites adopted Roman patterns of behavior (Woolf 1998; Keay and Terrenato 2001; Curchin 2004), illustrating the common adoption of the prestige culture or status on the part of those who are able to do so. This shows that there was no solid boundary between Iberians and Romans that could have persisted for centuries. Our attempts to make one are entirely inappropriate, as is the application of analogies derived from our experiences of racially based modern colonialism.

Ethnic Conflict Ethnic conflict is not an inevitable consequence of the existence of ethnic identity, even when that identity is fairly marked. Sociological studies of ethnic conflict, though they are, of course, almost exclusively concerned with modern cases, can give us an idea of the other characteristics of situations which lead to ethnic conflict. In an annual review article on the topic, Williams (1994) discussed the primary conditions which seem to facilitate ethnic conflict, which include: ethnic distinctiveness; competition for social goods, such as prestige, status, and economic or political power; grievances, unjust acts performed by one ethnic group against another; leaders who can mobilize effectively group action, and who tend to be among the elite or intelligentsia; and opportunity structures to allow groups to organize and carry out their actions. Thus, ethnic conflict is not simply a matter of intuitive, emotional reactions. Cultural context, and the structures of power both locally and imperially can have a huge impact on what this reaction would be, and beyond that, what form the reaction on the part of any potential ethnic group would take. Each of these five features should be present if we are to argue for the likelihood that a given conflict was ethnically based.

Attention to these aspects of cultural differences, ethnicity, and ethnic conflict can help us understand the different patterns of rebellion and peace across the Iberian peninsula, without having to resort to statements about the ‘primordial nature’ of different groups. I will discuss briefly a few of the places where ethnic identity may be an issue during the conquest period. This is only a preliminary list, and greater certainty about any of these issues requires additional research, particularly with the abundant archaeological material available from most contexts.

1. ‘Spanish’ troops in ‘Roman’ armies In an article from 1963, Antonio García y Bellido considered the pattern of use of native auxiliaries in the armies of the various Roman commanders in the peninsula. In particular, he notes changes between 205 B.C.E., when, at the conclusion of the Second Punic War, the only external power in the peninsula was the Romans; and 133 B.C.E., an important date, which represents the end of what he characterizes as ‘native rebellions’, with the conclusion of the Celtiberian wars. He argues that the scarce references to native auxiliaries in the literary sources is evidence of the ‘open rebellion’ of ‘all of Spain’ against the Romans (1963: 213). In keeping with his own nationalist sentiments, this apparent reluctance of ‘Spanish’ troops to fight for ‘foreigners’ was easy to explain, and helped to support his case for the deep-rooted unity of the Spanish people (Jiménez Diez 2005: 24–5). However, a closer look at the sources he cites suggests that they only mention native troops under exceptional circumstances, such as when Roman generals were unsuccessful in their efforts at recruitment (e.g. Livy 34.19; 40.30), when they acquired particularly high-status native fighters (e.g. Livy 40.47), or when they were aided in their recruiting efforts by pre-existing tribal rivalries (e.g. Livy 34.20). Thus, it seems equally possible that the silence of the sources reflects only how unremarkable it was that the Romans continued to recruit local troops, as they had done throughout the Second Punic War (Richardson 1986: 35ff). It is unrealistic to suppose that all the indigenous inhabitants of Spain would have acted as one against the Romans, or vice versa. As Simon Keay has pointed out, the wars that were fought on the peninsula had a relatively restricted geographical impact (Keay 2001: 126–127), and their ultimately limited impact is further evidenced by the fact that there are fewer of them after 133. Importantly, the late second century sees other changes aside from the end of the wars, including the beginning of Roman exploitation of the mines in the Sierra Morena, and changes in agricultural practices in Baetica, which had previously continued more or less along pre-Roman patterns (Keay 2003: 161–162). In other parts of Spanish provinces, crucial changes around this time have also been noted (García y Bellido 1963: 220; Curchin 2004: 94; Richardson 1986: 13–14). The changes at this date seem likely due both to a change of policy on the part of the Romans – deciding not to continue harassing and provoking Spanish tribes for the sake of glory – as well as to the gradual adjustment of all concerned to the new, provincial state of affairs, which was a process that took many decades after the conquest. The absence of further rebellions is therefore a symptom, rather than a cause, of a shift in interactions, as the descendents of people who were considered and considered themselves non-Roman evolved to consider themselves, and be considered, Roman. The adults of 133 were the great-grandchildren or great-great-grandchildren of those who had initially taken sides in the conflict between Romans and Carthaginians. When we allow for the changes in culture and identity with the passing of generations, we can observe more clearly the impact of Roman colonial style upon ethnicity.

2. Hybridae According to Livy (43.3), the city of Carteia, near Cadiz, was founded in the 170s B.C.E., as a special settlement for the children of Roman soldiers and native women. It seems to be unique in the literary sources, and its date may help explain this. That is, it may be a symptom of the ethnic situation of the period in which it was founded. This was the first or second adult generation since the Second Punic War brought Romans to the peninsula, and the hybridae’s difficulty in ‘fitting in’ on either side suggests that ethnic categories along the lines of ‘Roman’ and ‘Iberian’ were indeed important at that time. We should not be too excited by this information, though – despite its vehement claims of immutability and permanence, ethnicity is often a very transient phenomenon. Soon after this first generation had worked its way through the cultural and ethnic negotiations of their new colonial situation, the categories of Roman, native, and hybrid lost some or all of their importance. Though many authors recognize, in theory, that intermarriage must have been very common (e.g. Curchin 1990: 87–88), they neglect the implication of this for the maintenance of opposing ethnic groups (e.g. Curchin 2004: 3). That is, it may be that Mary Downs’ question of ‘how many hybridae were there a century later?’ (2000: 204) is left answered by our sources because the category lacked meaning in a context that had succeeded, over the course of a century, in downplaying ethnic distinctions among the inhabitants of this part of the Roman Empire.

3. The Lusitanian Wars Of all the conflicts described in the literary sources, this long struggle, which lasted in two phases from 155–139 B.C.E., would seem to have most of the signs of ethnic conflict. The famous leader Viriathus, responding to grievances such as the betrayal of a truce by a Roman general who treacherously massacred his countrymen, led a long war for land rights (Richardson 1986: 126ff). Territory is closely tied to concepts of ethnicity, and is frequently a driving factor in ethnic conflict (Williams 1994: 59). But if this was an ethnic concern, the ethnic categories seem not to have maintained their significance after the conflict itself was over. Stephen Dyson found evidence that the name of the Roman governor who settled natives after this conflict, Decimus Iunius Brutus Callaicus, was adopted by those he settled (1980/1981: 276–79), which suggests that the importance of this transiently adopted ethnic identity began to fade as soon as the conflict was over.

4. Urbanization Cities are frequently important sites of imperial intersection with, and encouragement of, ethnic identities. Comparative examples from contexts as diverse as the Uruk period in Anatolia (Stein 1999: 117–169), Cuzco in Peru (D’Altroy 2002: 119), and modern African colonial centres (Nagel and Olzak 1982: 131–133) show how imperial cities tend to encourage the formation of neighborhoods segregated by ethnicity. Nagel and Olzak suggest the reasons for the increase in ethnic salience that often accompanies the growth of cities: ‘Urbanization promotes ethnic mobilization to the extent that: (1) ethnic affiliations ease the urban transition of rural emigrants; (2) urban settings facilitate the development of ethnic organizations; and (3) urban economic competition is organized along ethnic lines’ (1982: 131). Thus, it seems worthwhile to see if similar processes occurred in the Roman Empire, where cities are acknowledged to have played a large role in the process of Romanization (Fentress 2000). In Baetica, two important examples of this, where we have literary testimony of the presence of a first-generation Roman or Italian element in the founding population, are Itálica and Córdoba. Appian’s Iberike, which is the only literary source for the foundation of Itálica, calls it a Roman colony (38.115), while Strabo explicitly refers to Córdoba as a joint foundation of Romans and natives (Geography 3.2.1). In both cases, scholars have traditionally spoken in terms of the different degrees to which the Romans or Italians among the population ‘won out’ culturally against the native elements. We see here, again, a problematic assumption that the two groups were kept separate by Roman colonial policy, or that underlying ethnic differences were so universally powerful as to maintain this distinction. However, the evidence we have for settlement at both places does not support a picture of segregated neighborhoods of Italians, next to segregated neighborhoods of natives. Cordoba is particularly interesting, because the Roman settlement was founded next to a preexisting native settlement (Ventura et al. 1998: 87). By the end of the second century, the same time as other major changes were occurring in Baetica, this native settlement disappeared, presumably as its former inhabitants moved into the Roman city (Ventura et al.1998: 91). As with the issue of the vanishing hybridae, as time passes the importance of cultural differences fades.

Conclusion These four examples from the conquest period in Spain have emphasized the importance of carefully distinguishing cultural differences from ethnicity, and furthermore considering the contributing factors to ethnic conflict. A major barrier to the direct comparison of ethnic circumstances in different empires is the impact of colonial style upon ethnicity–where ethnicity is emphasized as part of the workings of the empire, features such as ethnic neighborhoods and ethnic conflict are to be expected. In Roman Spain, however, the apparent absence of ethnic distinctions, which is most evident after the gradual cultural changes in the century following conquest, helps to explain the position of these provinces in the empire. While awareness of resistance to Roman dominance is important, we ought not to lose sight of the ways in which the Roman Empire seemed to produce a decrease in the importance of ethnic identity in Spain, particularly a few generations after conquest.

Joukowsky Institute for Archaeology and the Ancient World, Brown University

Acknowledgements This paper has benefited from the ideas and writing of many talented people. In addition to those cited in the bibliography, I am grateful to Dr. Irene Schrüfer-Kolb, the organizer of the TRAC session on the archaeology of ethnic conflict, which inspired this paper; Dr. Alicia Jiménez Diez who has helped me throughout my exploration of the archaeology of ethnicity in Roman Spain, and kindly commented on an early draft of this paper; my advisor, Professor Susan Alcock, for her comments and encouragement with this paper; Kenny Sims, whose challenging debates have contributed much to my thinking on the archaeology of ethnicity, and whose critical comments have vastly improved the still-developing ideas presented here; and Andrew Losowsky, whose patience, support, and editor’s eye remain invaluable.