The Plurality Ethnic Group Founding Dates Dataset

Thanks to funding from the British Academy

Award ref#: SG091163 "In Search of Ethnic Diversity: Dating the World's Dominant Ethnic Groups"

Principal Investigator: Eric Kaufmann, Department of Politics, Birkbeck College, University of London

Data used in the article, Kaufmann, Eric P. 2014. 'Land, history or modernization? Explaining ethnic fractionalization,' Ethnic and Racial Studies, forthcoming

             The Plurality Ethnic Group Founding Dates dataset was compiled by PhD student researchers with the assistance of a grant from the British Academy. Work was completed by six PhD students who are/were working on issues of nationalism and ethnic conflict. All are members of, and have access to, the expertise base of the Association for the Study of Ethnicity and Nationalism (ASEN) at LSE in London. This is a 20-year old student-run organisation which publishes Nations and Nationalism, the leading nationalism journal, and Studies in Ethnicity and Nationalism, and has several hundred members at any one time. Directories of members and their research interests are one resource, alongside more conventional forms of bibliographic inquiry.

Experts were contacted and asked to reply to the statement in Appendix I. Experts were directed to address the question of 'first imagining' by an in-group member rather than the formation of fully-fledged ethnicity, as this involves making determinations about the degree to which a set of myths of origin must penetrate down the social scale before a group can be said to be in existence. Though the label 'ethnic group foundation date' appears to make an ethnosymbolist presumption, this is not the case. A modernist interpretation would see an ethnic founding date as an ideational curiosity whose significance and continuity has been retrospectively constructed by political entrepreneurs. This is precisely how Eric Hobsbawm, for instance, treats the 'cultural nationalist' activities of romantic intellectuals, i.e. philology, folk anthropology, festivals, archaeology (Hobsbawm 1990).

The dating of a subjective phenomenon like the first imagining of a group demands specialist historical knowledge and thus occludes attempts to use standard techniques of inter-coder reliability. In the event, the initial 'scientific' aim of achieving inter-coder reliability had to be replaced by a more historical approach in which the accounts of scholars were triangulated because there was no simple process by which coders could 'code' data. This is the method, for instance, used in the State Antiquity Data project dataset (Bockstette et al. 2002; Chanda and Putterman 2007). All dates were reviewed for consistency by the author and by project advisor and well-known subject expert Anthony Smith, to whom the author is especially grateful.

The majority of experts contacted did not respond. Even among those who did, some found the question difficult to answer while others thought the matter controversial given the theoretical debate over the question of the modernity of ethnic groups. We had, however, anticipated such controversies and were able to offer guidance (see Appendix I) regarding which changes we felt represented continuity of historical consciousness and which signified rupture. We also advised upon when consciousness was deemed to exist and when the evidence was insufficient to support such a claim. Based on this criteria, Sunni Arab Egyptians and Iraqis are deemed to have lost contact with the narrative created by ancient progenitors while Greeks, Persians and the Jews are seen to have maintained it, however minimally.

The great shifts in consciousness, i.e. from ancient Hellenes to Byzantines to modern Greeks, are viewed as insufficient to constitute a full rupture. By contrast, the collapse of Sumer, Babylon or Assyria clearly sundered any continuity of Iraqi consciousness. It is a difficult exercise to erect a pure standard for borderline cases to decide when a historic consciousness has been broken or not. This has been attempted here, however, and distinguishes this dataset from the State Antiquity dataset which only seeks to establish whether an 'indigenous' dynasty was in control in a portion of the geographic area of the present-day state in a given half-century. This is an important difference, but one which may account for the stronger connection between ethnic group founding date and current group size as compared with the State Antiquity measure.

We were able to compile a set of dates for 129 of the 156 countries in the Vanhanen (1999) dataset. Crucial to bear in mind is that there is a considerable distance to travel between the first imagining of a group by a putative insider and the fully-fledged emergence of mass ethnic consciousness. Walker Connor (1994) argues that a majority of members must be conscious of their ethnic status in order for a group to be considered to be in existence. By this stringent criteria, there would be few, if any, groups in existence prior to 1800. This may be the reality - but the astounding relationship between ethnic first imagining and modern-day ethnic group morphology suggests that the date of origin is far from meaningless. Naturally the path from initial imagining to active ethnic behaviour is episodic. In some cases, the time between first imagining and collective action is brief. In others, it is protracted. For example, kingdoms which formed the ultimate origin of the largest ethnic groups in the Democratic Republic of Congo, Nigeria, Sierra Leone and several West African states date to the 14th and 15th centuries. Yet it is unclear how much this gives these groups an advantage over their twentieth century equivalents in neighbouring Liberia or Cameroon. In conjunction with ethnic date, a measure of both continuity and social (vertical) penetration would most likely improve the predictive power of how well an ethnic group is able to expand geographically (horizontally) into neighbouring populations. These offer avenues for future research. Another possibility is that even if their early kingdoms achieved low social penetration, older ethnic groups (like the Wolof of Senegal or Kikuyu of Kenya, both successful assimilators) can point to these early antecedents as a form of cultural capital which endows them with a patina of prestige (Johnson 2004). This in turn attracts power and adherents from other groups, increasing the group's share of the population.

In many cases, historians proved to be the best sources, though scholars from a wide range of fields contributed (see downloadable Data Sources file). In most cases, scholars generally agreed on a ballpark founding date. However, it must be emphasised that in some cases there was significant disagreement on dates between scholars, and in these cases we were forced to compile an 'early' and 'late' founding year for the group. Arriving at an intermediate estimate in these cases was a detailed exercise which involved a great deal of trial and error, with discussions between researchers, scholars, outside experts and myself. The triangulation of dates received from scholars was not always unproblematic. In general, the older dates were selected unless there was reason to suspect these were mythical, post hoc inventions, or described a kingdom with no connection to the present ethnic group. We used textual sources to make estimates of the remaining 27 countries. Tests show that these did not differ significantly from other dates in the data.[1] The dates were also subject to consistency across the dataset and checked by the author and by Anthony Smith, one of the world's leading scholars of comparative ethnic history.[2]