Language and Conflict
| Published date | 01 March 2011 |
| Date | 01 March 2011 |
| Author | Tristan James Mabry |
| DOI | 10.1177/0192512110379489 |
| Subject Matter | Articles |
Corresponding author:
Tristan James Mabry, Naval Postgraduate School, 1 University Circle, Herrmann Hall #312, Monterey, CA 93943, USA
[email: tjmabry@nps.edu]
Language and Conflict
Tristan James Mabry
Abstract
Language divides are common components of group conflict, a phenomenon reflected widely in theories of
nationalism. This article evaluates measures developed by David Laitin and James Fearon in the minorities at
risk dataset claiming to quantify language difference and concludes they are deeply flawed. The introduction
outlines language divides vis-a-vis conflict. A theoretical analysis in the second section argues against rational
choice analyses of language politics; in the third section a sociolinguistic matrix shows that these fractional
measures represent language ancestry but nothing else (morphology, syntax, lexicon, orthography, status).
Theoretical implications and alternative methods are considered in the fourth section followed by a summary
conclusion.
Keywords
ethnic conflict, language, methodology, nationalism, rational choice
Introduction
The relationship between language and conflict parallels the relationship between ethnicity and
conflict. To the extent that an ethnolinguistic population is synonymous with an ethnocultural
group, general models of modernity and nationalism typically parse cultures according to lan-
guages. Hence, the Russian culture and the Russian language, for example, are considered inextri-
cable. This model is framed theoretically within the modernist paradigm of nations and nationalism
(Smith, 1998; cf. Roshwald, 2006). Credited most often to the works of Ernest Gellner (1964,
1983, 1987) and Benedict Anderson (1983, 1998), this model includes structural and constructivist
arguments framed against longue durée processes of urbanization and mass communication. Thus,
modernization homogenizes low cultures into high cultures that are institutionalized in states as
national cultures. If rival cultures clash, however, then ethnic groups may mobilize members in a
group conflict.
David Laitin (2000a: 102), one of the very few American political scientists persistently asking
penetrating questions about language, considers the modernist paradigm of nationalism a ‘blunt
theory’. It is his position that the power of language (along with most other markers of ethnicity
International Political Science Review
32(2) 189–207
© The Author(s) 2010
Reprints and permission:
sagepub.co.uk/journalsPermissions.nav
DOI: 10.1177/0192512110379489
ips.sagepub.com
190 International Political Science Review 32(2)
and culture) to engender conflict is greatly over-estimated, especially in regard to the mobilization
of groups to commit large-scale violence, such as insurgency or civil war (Fearon and Laitin,
2003). He is therefore critical of authors and analysts who routinely connect language division to
violent ethnic conflict. Yet at the same time, Laitin (like Gellner, Anderson, and Adrian Hastings)
endorses language as a proxy for culture, and by extension a proxy for national membership. He
(2007, p. 59) argues this is sensible because ‘language has special attributes that make it especially
amenable’ to his model of nation formation, a phenomenon he attributes to rational individuals
calculating whether or not to join a mass movement ‘for purely instrumental reasons’ that need not
change one’s ‘deep-seated identity.’ This is where Laitin diverges sharply from the canon of nation-
alism theory, which generally posits that the deepest identity – an identity so inextricable with the
self that its protection is worth dying for – is, in fact, a national identity.
To falsify a link between language and conflict, Laitin sought data that would enable a large-
scale statistical analysis, a method he views as a critical component of the best political science.
This, of course, is a debatable position (Green and Shapiro, 1994; Monroe, 2005), especially in
regard to the role of language, which, as argued below, is exceptionally resistant to categorization
and quantification. Moreover, if large-N analysis of language community relations were useful to
disciplines that routinely research language, such as sociolinguistics or anthropology, it seems
more than a little surprising that nobody before Laitin came up with the idea. It is suggested here
that the method was not applied elsewhere because it is incapable of uncovering anything of value,
particularly when compared with the abundance of research from other disciplines that already
tells us what we need to know about language divides.
Thus, this article does not endorse quantitative analysis of language politics, but rather evalu-
ates quantitative variables ascribed to characteristics of language communities, variables devel-
oped by David Laitin and James Fearon and eventually incorporated into the minorities at risk
(MAR) dataset as ancestral language scores (ALS). In sum, such variables are rejected as incapable
of representing that which is nominally under consideration: the relative importance of language
difference in relation to cultural-cum-national conflict. Therefore, any conclusions drawn from
analyses that employ such data are questionable.
What then is language difference? First, language difference may be disaggregated across a very
broad spectrum of characteristics. In the analog of comparing two people, there may be similarities
or differences according to physical traits such as height, weight, coloring, gender, allergies, immu-
nities and so on, and many more conceptual distinctions such as ‘personality’ or ‘beliefs.’ In the
case of languages, it is possible to observe similarities or differences according to structural traits
such as whether a subject precedes a verb and object (or some other ordering), but also such dis-
tinctions as the range of phonemes, patterns of intonation, number of words (and for what things
and ideas), the presence or absence of writing and the many variations of alphabets or scripts, and
whether a language pair is more or less related to a common ancestral language. It is important to
acknowledge that a measure of language distance is reasonably objective, though as an instrument
of measuring language difference the measure appears insignificant. Imagine a meteorologist who
produces a weather forecast based on quantifiable and objective data: so far, so good. On closer
inspection, his or her data includes points measuring air temperature, but does not include mea-
surements of barometric pressure, precipitation, relative humidity, cloud cover, wind speed, wind
direction or dew point. Hence, the data is an objective measure of temperature, but it does not
describe the weather, let alone a climate.
The second degree of language ‘difference’ is socially and politically subjective: what matters
in the course of ethnic mobilization or mass behavior is the perception of inter-group commonality
To continue reading
Request your trial