Killing in the name of …? Types of ethnic groups and armed conflict
| DOI | 10.1177/0010836717705413 |
| Author | Lasse Lykke Rørbæk |
| Published date | 01 December 2017 |
| Date | 01 December 2017 |
| Subject Matter | Articles |
https://doi.org/10.1177/0010836717705413
Cooperation and Conflict
2017, Vol. 52(4) 537 –552
© The Author(s) 2017
Reprints and permissions:
sagepub.co.uk/journalsPermissions.nav
DOI: 10.1177/0010836717705413
journals.sagepub.com/home/cac
Killing in the name of …?
Types of ethnic groups
and armed conflict
Lasse Lykke Rørbæk
Abstract
This paper asks whether certain types of ethnic groups are particularly likely to take part
in armed conflict. Several theoretical arguments indicate that this should be the case, often
highlighting religious and racial boundaries as being more conflict-prone than, for instance,
linguistic boundaries. However, the potential effects of groups being mobilized around these
different boundary markers remain largely untested. The paper helps to fill this gap by analyzing
conflict propensity across types of ethnic groups in a global sample for the period 1946–2009.
At odds with common perceptions, the results show that the probability of armed conflict onset
is not affected by whether ethnic groups are mobilized around religious, linguistic, racial, or
regional markers. The effect of political discrimination on armed conflict is also not conditioned
by these different boundary markers. The paper thus lends support to an inclusive conception of
ethnicity and suggests that we need to focus on the social and political context rather than the
specific cultural content of ethnic boundaries if we want to identify the conflicts that are most
likely to escalate and turn violent.
Keywords
Armed conflict, boundary markers, ethnic groups, ethnic mobilization, religion
Introduction
Since World War 2 approximately two-thirds of all civil wars have been fought along eth-
nic lines (Denny and Walter, 2014). The conflict literature explains this prevalence by eth-
nic groups’ distinctive incentives and opportunities to mobilize for violent collective action
(see also Cederman et al., 2010). Ethnic groups can be defined as social groups for which
membership eligibility is determined by descent-based attributes rather than, for example,
political ideology, socioeconomic class, or gender (Chandra, 2006; Fearon, 2006). Most
Corresponding author:
Lasse Lykke Rørbæk, Department of Political Science, Aarhus University, Bartholins Allé 7, Aarhus 8000-
DK, Denmark.
Email: LLykke@ps.au.dk
705413CAC0010.1177/0010836717705413Cooperation and ConflictRørbæk
research-article2017
Article
538 Cooperation and Conflict 52(4)
commonly, ethnic groups are mobilized around religious, linguistic, racial, or regional
markers (Wimmer, 2013: 8). Although the ethnic conflict literature is rich, our knowledge
is still sparse when it comes to the potential effect of these different boundary markers. Are
certain types of ethnic groups particularly likely to take part in armed conflict?
To answer this question, I assess theoretical arguments that stress how the cultural content
of ethnic boundaries may affect behavioral differences. Religious boundaries are often argued
to incite violence (e.g., Reynal-Querol, 2002; Smith, 1996); a notion that resonates with the
public perception as well as current armed conflicts in, for instance, Iraq, Syria, and Yemen.
In somewhat similar fashion, racial boundaries have been described as inherently conflictual
(e.g., Cornell and Hartmann, 1998). Language differences have, by contrast, been highlighted
as a factor that can hinder outbreaks of violence (Laitin, 2000). Despite arguments like these,
recent conflict studies tend to pool all ethnic groups together (e.g., Cederman et al., 2013).
However, conclusions drawn from these studies might be misleading if the propensity for
armed conflict is systematically associated with types of ethnic groups.
As I will argue below, three potential effects of ethnic boundary markers can be hypothe-
sized. First, if certain types of ethnic groups on average are more likely to take part in armed
conflict, all other things equal, there would be a direct effect. Second, if the variation in con-
flict propensity is driven by certain groups’ increased likelihood of experiencing political
discrimination or other factors known to increase the risk of violence, then the effect would
be indirect. And, finally, if certain groups are particularly likely to engage in armed conflict
when they experience these conflict-enhancing factors, then the effect would be moderating.
I test these propositions by combining data on ethnic power relations and armed con-
flict onset (EPR-ETH v.2; Cederman et al., 2010) with coding of ethnic boundary mark-
ers (Wimmer, 2015) in a global sample of politically relevant ethnic groups for the period
1946–2009. In contrast to theoretical arguments as well as empirical findings showing
variations in conflict propensity depending on the type of ethnic boundary, I find no
statistical evidence of either a direct or indirect effect. Also, the effect of one of the most
important predictors of armed conflict, political discrimination, does not vary signifi-
cantly across types of ethnic groups when proxied as either political exclusion or power
loss (indicating absence of a moderating effect).
What these findings suggest is that religious, racial, linguistic and other ethnic bound-
ary markers are not sui generis when it comes to outbreaks of armed conflict. Accordingly,
the paper supports more generalizing or inclusive conceptions of ethnicity that treat vari-
ous ascriptive markers as functionally equivalent (cf. Brubaker, 2015; see also Horowitz,
1985; Olzak, 2006). To quote Rothschild (1981: 86), ‘the possible ethnic significance of
these marker-criteria is given, or withheld, or withdrawn, not by their [cultural] content,
but by their social and political context.’ An important implication of the paper is thus
that the intensive focus on religious violence in the public as well as the academic debate
is at least somewhat unjustified. Armed conflicts along religious boundaries as we are
currently witnessing in several Middle Eastern countries are likely more about political
dominance than about religious differences. However, as I discuss in the conclusion,
although types of ethnic groups cannot explain escalation into armed conflict, more
research is needed to examine whether ethnic boundary markers matter in other stages of
the conflict process. In this regard, the analysis is limited by only including politically
relevant ethnic groups because it cannot be established whether certain types of ethnic
groups are more likely to become politically relevant in the first place.
To continue reading
Request your trial