See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/254116598

The Identity Trap: The Language of Genocide

Article  in  Journal of Language and Social Psychology · March 2012

DOI: 10.1177/0261927X11425033

CITATIONS READS

7 1,264

1 author:

William A Donohue
Michigan State University
73 PUBLICATIONS   1,566 CITATIONS   

SEE PROFILE

All content following this page was uploaded by William A Donohue on 26 August 2015.

The user has requested enhancement of the downloaded file.

 Journal of Language and Social
Psychology
http://jls.sagepub.com/

The Identity Trap: The Language of Genocide

William A. Donohue
Journal of Language and Social Psychology 2012 31: 13 originally published online 21
November 2011
DOI: 10.1177/0261927X11425033

The online version of this article can be found at:

http://jls.sagepub.com/content/31/1/13

Published by:

http://www.sagepublications.com

Additional services and information for Journal of Language and Social Psychology can be found
at:

Email Alerts: http://jls.sagepub.com/cgi/alerts

Subscriptions: http://jls.sagepub.com/subscriptions

Reprints: http://www.sagepub.com/journalsReprints.nav

Permissions: http://www.sagepub.com/journalsPermissions.nav

Citations: http://jls.sagepub.com/content/31/1/13.refs.html

>> Version of Record - Feb 6, 2012

OnlineFirst Version of Record - Nov 30, 2011

OnlineFirst Version of Record - Nov 21, 2011

What is This?

Downloaded from jls.sagepub.com by guest on October 11, 2013

425033
ueJournal of Language and Social Psychology
JLS31110.1177/0261927X11425033Donoh

Journal of Language and Social Psychology

The Identity Trap: The 31(1) 13­–29

© 2012 SAGE Publications
DOI: 10.1177/0261927X11425033
Language of Genocide http://jls.sagepub.com

William A. Donohue1

Abstract
The purpose of this article is to create a framework for thinking about how language
themes propagated in radio broadcasts start to degenerate into the reification of
a culture that tolerates and even encourages classification and dehumanization.
Understanding these language themes becomes an early warning system that begins
to signal the beginning of a genocidal spiral. This framework is termed the Identity
Trap, and it demonstrates how various linguistic conventions combine to establish
a social context that builds up the speaker’s social identity while denigrating the
“enemy’s” social identity, which provides the rationale for escalating conflict against
that enemy. Two broadcasts from the 1994 Rwandan genocide are used as examples.

Keywords
genocide, social identity, dehumanization, Rwanda

The hate speech inciting the 1994 Rwandan genocide is well documented by Simon
(2006). Extremist Hutu media outlets produced a title wave of hate speech advocating
the ethnic cleansing of the Tutsi minority. One radio station even identified targets for
Hutu militias, resulting in the executives from this station being convicted for their
role in the genocide. Although the airing of these messages alone did not cause the
genocide, it is clear that they created a social climate that legitimized tribal hatred that
ultimately eliminated any social sanctions preventing genocide.
In his article asking whether the Rwandan genocide could have been prevented,
Stanton (2004) argues that prevention means understanding the process of changing a
social climate within any given state and how that climate devolves into one group seek-
ing to exterminate another. He identifies eight stages of social change that each yields

1
Michigan State University, East Lansing, MI, USA

Corresponding Author:
William A. Donohue, Department of Communication, Michigan State University, East Lansing,
MI 48824, USA
Email: [email protected]
14 Journal of Language and Social Psychology 31(1)

markers that, if recognized, can serve as early warning systems for possible interven-
tion. The first three stages are Classification (“us versus them”), Symbolization
(groups are given names or symbols of their second-class citizenship such as ID cards),
and Dehumanization (groups are given names of bad things, e.g., cockroaches or can-
cer). These first three stages mark the onset of the death spiral toward genocide. The
final five stages include Organization (hate groups organize into armed forces),
Polarization (in which moderates are targeted and assassinated), Preparation (planning
and training for the final genocidal attacks), Extermination (beginning the genocidal
extermination of “less than human” enemies), and Denial (justifying their actions and
denying that a crime has been committed). Stanton argues that intervening in the social
climate at early stages provides a greater chance of success than waiting until extermi-
nation begins and then trying to stop that process.
These three early stages are important because they have the greatest potential for
prevention by first understanding how the public language displayed through the
media and other outlets is classifying, symbolizing, and dehumanizing groups within
that society. Based on that understanding, interventions can then be crafted to both
warn of impending violence and perhaps even turn that language in a more construc-
tive direction. Nevertheless, crafting an intervention system that might include a warn-
ing mechanism means first focusing on the themes displayed by the classification
language that sets the kind of social context within the society at risk for genocide. The
article applies a framework developed to understand the linguistic context of the Oslo
I accords (Donohue & Druckman, 2009) to an analysis of the language leading up to
the Rwandan crisis of 1994. The purpose is to determine how the media systematically
constructed a context of classification, symbolization, and dehumanization that made
the genocide possible. These language themes signal the beginning of the genocidal
process. The framework that will be used to understand this genocidal path is termed
the Identity Trap, and its development begins by better understanding how language
establishes social contexts in the course of ethnic disputes.

Language and Social Climate

In a recent article focusing on Israeli–Palestinian language leading up to the Oslo
Accords, Donohue and Druckman (2009) sought to understand how leaders from both
groups framed their public language to create the larger social context surrounding the
secret negotiations. Selecting a series of speeches and interviews in the 6 months lead-
ing up to the accords, they coded the extent to which the messages were displaying
power/affiliation, trust/mistrust, and forward-/backward-looking relational message
frames. The study found that Palestinian language consistently displayed more power
and backward-looking justice themes, and less trust. The primary explanation for this
“tougher” looking language was centered on the notion of outbidding. This is a strat-
egy that nonstate actors often adopt to appear “tougher” to show resolve to constituents
and bid for their support over rival factions. In contrast, Israeli language alternated
Donohue 15

between displaying high power and high affiliation; yet it was also more forward
looking while showing glimpses of trust at times. This vacillation reflected more of a
front-stage orientation, which is typical of state actors who must moderate their lin-
guistic positions to appeal to broader political audiences. Israeli politicians must
appeal to both conservative and liberal factions in order to win votes.
The study also explored the intersection of these three frames. When speakers
focused on the past (looking backward with an emphasis on justice concerns), these
issues were more likely to be accompanied with messages of high power and mistrust.
However, when they focused on the future and the formulation of peace issues, they
were more likely to infuse their language with trust and affiliation messages. In short,
the data suggest that these constructs are highly interdependent; the use of more col-
laborative or competitive relational messages appears to be a function of whether indi-
viduals adopt a justice or peace frame in their communication. These results build on
and extend the forward-backward, peace versus justice theme explored in Zartman and
Kremenyuk (2005) by adding relational features to their substantive distinction.
This finding is relevant to the issue of genocide because it reveals a linguistic profile
that typifies the “classification” language described by Stanton (2004). This classifica-
tion process works to create two polarized identities aimed at building separation
between in-groups and out-groups or separating one’s own belief system from the belief
systems of others (Rokeach, 1960). The process typically begins by classifying the out-
group as possessing a set of dangerous and morally bankrupt characteristics. The goal is
to create an out-group identity for the main enemy by proposing that this group is a
significant and credible threat to the well-being of the in-group. The main vehicle for
creating this identity is forging a simple, evil stereotype of the out-group, a “disbelief
system” (Rokeach, 1960) that stems from their past injustices. “Here’s the evil they have
perpetrated,” so “they can’t be trusted,” and as a result, “they must be controlled and
eliminated.” Notice the themes of looking backward, mistrust, and power in this lan-
guage and how this vilification process begins by looking backward and focusing on the
specific evil acts perpetrated by this out-group. After building the case that the out-
group is evil, the conclusion is that they cannot be trusted and must be dealt with sum-
marily. The steady drumbeat of injustices and mistrust aims at inciting emotional
reactions of fear and anger. In turn, this negative emotion motivates those who consume
these messages to take quick and decisive action that will eliminate the threat.
Part of this polarized identity for the out-group is forging an identity for the in-group
as the only viable counterforce to fight the threat. The foundation of this identity is the
in-group’s claim that only it has the courage and strength to identify the threat and
stand up to it. “We see what they’re doing and we’re not afraid to say it” is the general
theme of the language. The more they “call out” the out-group members as perpetuat-
ing outrageous injustices and forming unholy alliances with other recognized evildo-
ers, the more prominently the in-group displays its identity as called on to save the
day. Thus, the purpose of fighting the out-group is not just to eliminate the threat but
to bolster the ability of the in-group to provide justice and create a better society.
16 Journal of Language and Social Psychology 31(1)

The argument in this article is that creating these identity extremes through polar-
ized language forms an “Identity Trap” in the sense that it frames the conflict as being
about the clash of identities instead of focusing on substantive issues and the explora-
tion of a middle ground. The more extreme and polarized the identities become, the
bigger the trap that is created and the fewer choices people have to stay out of the
conflict. The idea of this process creating a “trap” stems from the extreme paradoxical
state this situation presents to individuals. The steady drumbeat of inflammatory,
extremist language simultaneously pulls enemies closer together (more talk about,
confrontation of, and more physical engagement with the enemy) in a struggle to push
(defeat, separate from, or in other ways eliminate) the enemy further away. This para-
doxical push–pull state traps people into a sense of confused outrage and becomes the
new “normal” for them. At its worst, this push–pull paradoxical identity entraps entire
societies, making individuals feel caged and backed into a corner.
As their frustration builds they look for authority figures that will provide a way out
of the paralyzing trap. These authority figures emerge from the in-group and become
the voice of action by first building an increasingly extreme identity of the targeted
out-group and then providing a path that leads to the elimination of the out-group. If
there are no moderating influences on this destructive path, individuals will face a dif-
ficult challenge in both recognizing their entrapment and escaping it constructively. If
the authoritarian leader is discredited, or some critical incident causes parties to dra-
matically shift their views of the out-group members, they may recognize and uncover
the paradox. In fact, the goal of any kind of political reconciliation process is essen-
tially a guided introspection about the out-group’s identity and an attempt to broaden
that identity to shift the conflict away from identities and toward a focus on the issues
(Moon, 2008). This shift eliminates the paradox and essentially closes the identity
trap. However, the longer the conflict focuses on identities by referencing past atrocities,
the deeper the paradox becomes and the bigger and more formidable is the identity trap.
Unless some kind of intervention occurs, the language shifts from grievances to
classification and can then move quickly toward symbolization and ultimately
dehumanization.
Thus, the building of an identity trap is best viewed as a process that begins with a
steady drumbeat of grievances or criticism against the out-group while simultaneously
forging the identity of and praising the in-group as a counter to the threat. The process
escalates with the use of symbolization language as a means of making the out-group
threat more concrete and significant. The symbolization then leads to dehumanization
and the call to action aimed at eliminating the out-group. Whillock (1995) provides
further insights about how this process evolves. He offers the idea of a “hate strata-
gem” to describe how groups build a social identity around hate themes. He argues
that groups first seek to inflame members’ emotional reactions to the targeted out-
group as a means of enhancing the in-group’s identity. By using hateful, verbally
aggressive discourse that vilifies the targeted individuals, the group creates a stronger
sense of purpose. They must stick together to eradicate the threat. To further solidify
this collective identity, the group pulls the enemy closer by expanding the list of
Donohue 17

undesirable qualities of these targeted individuals and making that list very public. It
becomes important to find more and more qualities that threaten the in-group’s iden-
tity. This continuous throng motivates the group to then inflict harm on the out-group
by verbally attacking its members, its perceived friends, and things they value. If suc-
cessful, the group will then begin physically attacking the out-group members directly
to eliminate them. That is, verbal aggression gives way to physical aggression. The
final part of the hate stratagem is to rhetorically conquer the enemy by glorifying the
killing and destruction of the out-group, its friends, and the things it values. This glo-
rification is needed to show that the group can successfully achieve its goals, which
further build its collective identity.
In other words, hate language that continuously performs the task of classifying the
targeted out-group from start to finish is an intrinsic part of the genocidal process. The
hate language serves to continuously reinforce the out-group’s negative identity and
the in-group’s positive identity further widening the identity trap and making it more
and more difficult to avoid succumbing to it. To understand the specific linguistic
structures that promotes this continuous vilification and classification process, it will
be useful to further describe the three-part conceptual framework described above that
forms the outline of identity trap and ultimately propels group members toward
violence.

Identity Trap Language

Forward- and Backward-Looking Language

One way to understand political language is in terms of forward- and backward-looking

statements (Zartman & Kremenyuk, 2005). Backward-looking statements focus on
justice themes such as ending the violence, accounting for past wrongs, and objections
to compromise outcomes that do not address the underlying sources of conflict. These
statements emphasize the symptoms of conflict, often implying that the other party is
responsible for the problem and that a resolution depends largely on their concessions.
They reflect and reinforce a competitive (or distributive) bargaining process. In con-
trast, forward-looking statements are characterized by an attempt to address the under-
lying causes of the conflict by creating a new framework intended to build a more
constructive future. They focus on similarities, the acknowledgement of mutual
responsibility, and an awareness of the underlying reasons for the conflict. The empha-
sis is on imagining a future in which peaceful relationships are sustained.

Power and Affiliation Language

An earlier conceptual scheme, devised by Donohue and Hoobler (2002), was used to
capture the relational features of communication. It distinguishes between messages
that emphasize power and those that promote affiliation between disputants. Power
includes (a) strong, forceful acts; (b) control, or attempts to regulate or manage the
18 Journal of Language and Social Psychology 31(1)

other; (c) attempts made to influence the other with direct or implied threats; (d) efforts
to impress the other with reputational tactics; (e) a strong positive or negative emo-
tional reaction; and (f) giving unsolicited advice or help. Affiliation messages consist
of (a) expressions of positive, friendly, or intimate feelings; (b) expressions of sadness
or regret for lost opportunities to restore relations; (c) statements of companionship or
camaraderie; and (d) nurturing statements or acts. In the analysis of the Oslo I data, it
was clear that power-oriented language was a standard tactic used to appear tough
toward the enemy, whereas affiliation language was aimed at trying to appear more
“reasonable” or “understanding.”

Trust and Mistrust Language

Trust has been conceptualized frequently as risking vulnerability in situations where
a person has little control over an outcome (Hoffman, 2002; Lewicki et al., 1998).
The key to a trusting relationship is a willingness to take a chance on being exploited
for the other’s gain (Lewicki & Wiethoff, 2000). Among the factors thought to con-
tribute to trust are (a) perceived similarity and positive regard, (b) an interdependent
relational structure in which the parties can help one another, (c) the initiation of trust-
ing behavior by the other party, and (d) making concessions on key issues. When one
or more of these factors are present, the parties are more likely to risk vulnerability
and thereby strengthen a sense of trust (Lewicki & Litterer, 1985). The contrasting
state of mistrust occurs when parties are unwilling to risk vulnerability or to acknowl-
edge their dependence on one another. Expressions of (a) separation, (b) denials of
similarity, (c) promoting suspicion of the other party’s intentions, and (d) an unwill-
ingness to offer or to reciprocate concessions are indicators of mistrust. In their study
of negotiations to end violent international conflicts, Irmer and Druckman (2007)
found that increasing levels of trust during the course of talks improved the prospects
for obtaining a comprehensive, rather than a partial, agreement—or no agreement at
all. Parties that made a transition from mistrust early in the talks to knowledge-based
and then identity-based trust achieved better outcomes than those who remained mis-
trusting or were willing only to evince a knowledge-based trust.

The Paradox of the Identity Trap

At this point, it is useful to understand how these three pairs of linguistic markers
function collectively to form an identity trap. To summarize the issue of paradox in
the context of crisis communication, it appears that parties create and manage rela-
tionships by negotiating their identity as they exchange information. As they become
more interdependent (pulling each other closer and being more direct) and less affili-
ative (pushing each other away through expressions of mistrust and dislike) they find
themselves immersed in an identity trap. Individuals’ identities are at risk in a conflict
as they seek to support both their positive and negative face needs (Brown &
Levinson, 1978). Both parties place a great priority on the need to protect how others
Donohue 19

see them (positive face) while resisting others’ attempts to constrain their behavior
(negative face). When these identity needs drive the paradoxical context of conflict
interaction, parties risk falling into the trap of needing to use more extreme behavior
(both linguistic and physical) to repel ever-expanding identity threats. Falling into the
trap is easy since individuals are well practiced at protecting identity as it is essential
to maintaining one’s social standing. Unfortunately, staying out of the trap is often
difficult since that requires not reciprocating identity attacks while also understanding
the conflict from the other’s perspective.
As the identity trap deepens and language escalates in response to fulfilling identity
needs, the language can signal what Zartman (2000) argues is a mutually hurting stale-
mate. In other words, both sides are involved in an escalating spiral in which both
sides continue to endure the pain of conflict until that conflict becomes “ripe” or too
painful to continue. This ripeness becomes manifest as an expressed willingness to
move out of an identity-focused conflict trap and into a more constructive frame. This
movement, or ripening, requires that parties focus quickly on two separate issues
before they can begin dealing with the substantive or material concerns in the dispute:
their relationship to each other and the context that has emerged in the course of the
dispute. The challenge with the first problem is building a consensus about the funda-
mental parameters of their relationship simply as communicators. Regarding affilia-
tion, how are they prepared to manage the issues of trust/mistrust, openness/closedness,
public/private? And how can they also manage their interdependence issues focusing
on connection/ autonomy or control/yielding? They might engage in a variety of face-
saving and face-maintaining strategies all aimed at finding new levels of affiliation
and interdependence that will free them from the trap and simply allow the parties to
communicate again more constructively, or disengage for some period to cool off.
For example, in response to a mutually hurting stalemate fed by one identity-riddled
crisis after another, one side may choose to renegotiate affiliation by being more open
and revealing some sensitive information to show that they are being more trusting. Or
they might yield control over some critical negotiation procedure that they used to
stonewall the other. They might also agree to work with a third party that was previ-
ously not acceptable. The key here is that once parties feel trapped and see themselves
imprisoned by it, the conflict is ripe for renegotiation.
Unfortunately, genocide occurs when one side chooses to resolve the stalemate by
physically eliminating the enemy instead of promoting negotiation. If the individuals
from one of the groups are nonstate actors, the Donohue and Druckman (2009)
research indicates that they are more likely to use their linguistic opportunities to prop-
agate extremist language as an outbidding strategy to gain legitimacy among their own
constituents. This outbidding, extremist language began proliferating in the Rwandan
conflict in late 1993 and escalated throughout the genocide becoming more extreme
by the day. The key question becomes “what is that tipping point?” Is there a way to
chart the course of the language using the Identity Trap framework to determine how
close individuals are to experiencing a sense of ripeness and conflict resolution or to
sliding toward a genocidal solution to the issues? The first step in answering this
20 Journal of Language and Social Psychology 31(1)

question is conducting a qualitative analysis of some of the radio broadcasts leading

up to the Rwandan genocide of 1994.

Rwandan Genocide
According to the British Broadcasting Corporation’s time line (http://news.bbc.co.
uk/2/hi/africa/3580247.stm), the 100 days of genocide began on April 6, 1994, when
President Habyarimana and Burundian President Cyprien Ntaryamira were killed
when their plane was shot down as it was about to land at the Kigali, Rwanda, airport.
That event triggered the systematic killing of Tutsis and moderate Hutus by the
Rwandan armed forces and the Interahamwe militia, which was dominated by Hutu
extremists. This killing continued throughout April, so that by mid-May about
500,000 Rwandans had been killed. The conflict ended on July 18, 1994, when the
RTF (the Tutsi Rwandan Patriotic Front) seized control of the country and announced
that the war was over. During the conflict, there were various attempts by the
International Red Cross, French, and Belgian forces to intervene, but their troop levels
were ineffective, rendering them unable to end the hostilities.
From mid-1993 to the end of July 1994, an extremist Hutu group controlled a popu-
lar radio station in Kigali, the capital of Rwanda. The station sought to stir up ethnic
hatred for the Tutsi minority (about 9% of the population). Here is an excerpt from a
November 24, 1993, broadcast (the Inkotanyi is a nickname for Tutsis). These tapes
are available at http://surplusknowledge.com/index.php?option=com_content&view=
article&id=55&Itemid=64.

However, what has been found out, what is well-known is that this present
government which spent much time in an impasse is a powerless one. Because
its members are trying to work with Inkotanyi if they were ever to join that
government. Because they are trying to get posts in that government. So, they
can never say anything that should oppose them to the Inkotanyi. That is the
reason people are being killed and they can never talk of “Inkotanyi.” They are
really afraid of that word so that they (Inkotanyi) will not hear them and get
angry with them and scold them. That is why in the speech that Mrs.
Uwilingiyimana Agathe delivered yesterday, people have always been telling
her, “You are saying that soldiers decided that they will no longer negotiate
because people died. Why did RPF make a great deal of fuss when thirteen
people were lost? They even ordered you to rush and look for them in Ndusu
until you handed them over to them. Even radio and television stations attended
that. . . . So, why had they stopped the negotiations and instead of condemning
them, you rather got afraid, you fussed about and you looked for those young
men whom you handed over to the Inkotanyi? Maybe if you do not agree with
me. . . . Furthermore, now that people have been killed and there is no doubt
that it is RPF who killed them. . . . Now we are stopping the negotiations. I think
Donohue 21

that if agreements are signed, I think that they have to be respected by two sig-
natory parties.

Classification Linguistic Analysis

Using the Identity Trap framework outlined above, the classification themes dis-
played in this broadcast begin to emerge. Notice first that this broadcast was made in
late 1993, about 6 months prior to the genocidal attacks. This excerpt is just one of
the daily broadcasts by this extremist Hutu group, and their goal was to vilify both the
government controlled by moderate Hutus and the Inkotanyi, or the minority Tutsi
population in the country. As indicated above, these broadcasts were made by non-
state actors looking to gain support for their extremist cause. This is a typical outbid-
ding goal that is achieved by using the kind of extreme language that appeals to
extremists.
Before examining the specific themes used in the broadcasts, it is useful to examine
the general approach the speaker is using to spring the identity trap. The first several
lines are filled with accusations aimed at discrediting two out-groups: the moderate
Hutu-run government led by Mrs. Agathe and the Inkotanyi, a pejorative nickname for
the Tutsi minority. The accusations stem from corruption, improper negotiation strate-
gies, and, most important, murder. On one hand, the speaker is working hard to create
out-groups for both the government and the Inkotanyi. On the other hand, the speaker
is creating an in-group identity for those who oppose these injustices. The focus then
moves toward making sense of these incidents by building negative identities for the
out-groups and a positive identity for the in-groups who would oppose such practices.
To understand this extreme language more precisely, it might be useful to focus first
on how power and affiliation messages were displayed in the broadcast.

Power/Affiliation Classification Themes

Recall that power themes focus on control, influence, emotional reactions, and the
like, whereas affiliation themes emphasize camaraderie, friendliness, and other nur-
turing acts. In the first couple of sentences, the speaker is seeking to influence public
opinion. This power move criticizes the government for being powerless because they
are trying to work with the Tutsis. Notice first that the speaker is using the pejorative
nickname for the Tutsis, calling them the Inkotanyi. The speaker then claims that the
Inkotanyi are collaborators with the RTF (Rwandan government) and that both groups
are involved in various killings. In fact, throughout this broadcast the speaker is trying
to link the RTF and the Tutsis together in a conspiracy of some sort against the people.
In essence, the speaker is trying to classify the Tutsis as another enemy of the people;
the objective is to justify later dehumanization that occurs in subsequent broadcasts in
April and May 1994. The dominance of power themes and the absence of affiliation
themes in the language begin to set the context for later events.
22 Journal of Language and Social Psychology 31(1)

The power themes are also propelled forward by the use of emotional language, the
first component of Whillock’s (1995) hate stratagem. The accusation that the Inkotanyi
are working with the government to acquire posts strikes at the issue of fairness and
works to forge the identity of the Tutsis as an entitled class who see themselves as
superior to “the people.” The Inkotanyi are then portrayed as people who must be
feared because they easily angered and should not be trusted. The broadcast ultimately
accuses them of working with the government to kill 13 people. These accusations are
aimed at inciting the listeners’ emotional reactions against the Tutsis while building
the counteridentity of the group watching over the interests of “the people”—the Hutu
majority.

Trust/Mistrust Classification Themes

The broadcast also seeks to establish listener mistrust of both the RTF and the Tutsis.
Recall that the mistrust themes focus on separation, denials of similarity, fomenting
suspicion, and an unwillingness to reciprocate concessions. Looking at each sentence
in the broadcast, it is clear that the intent of each comment is to build suspicion about
the RTF as well as the Tutsis. The speaker accuses the government of giving posts to
the Tutsis, unfairly stopping negotiations after the 13 people had been killed, giving the
killers over to the Tutsis, and accusing the RTF of committing the killings. The
speaker’s goal also appears to separate listeners from the RTF and Tutsis. The speaker
is essentially accusing these two groups of doing bad things that are wrong and unfair
for the Rwandan people. The mistrust theme is very strong in this broadcast and
serves to again classify the RTF and the Tutsis as bad people.

Forward-/Backward-Looking Classification Themes

Finally, the hallmark of backward-looking language is the idea of seeking justice for
past wrongs and not addressing the underlying sources of the conflict. The goal is to
lay blame on someone for past wrongs such that if these individuals were eliminated,
justice would be served in some way. Again, the goal of the broadcast is to lay blame
for various injustices (e.g., killings, stopping negotiations) at the feet of the RTF and
the Tutsis. The language clearly looks backward with no attempt to propose any alter-
native strategies for either understanding the current problems or negotiating a new
set of arrangements. There is also no attempt to provide others’ opinions or to get
reaction from any other source. In other words, it is not about debate but rather stern
accusation to incite hate.

Symbolizing and Dehumanizing Linguistic Analysis

It is clear that the radio broadcasts leading up to the genocidal acts were aimed at
accomplishing two key goals. The first goal was to focus listener attention on iden-
tities rather than on specific issues. Creating an extremist identity for the out-group
Donohue 23

Tutsis on the one hand and building a rationale for viewing the extremist Hutu group
as the salvation of the country on the other hand was a necessary first move. The
second goal was to refine these identities by classifying the Tutsis as dangerous
criminals causing problems for the country while offering the Hutu group as the voice
of salvation. Having pounded away at these goals for months, subsequent broadcasts
moved systematically toward genocide by concentrating on language that promoted
symbolization and ultimately dehumanization. Below is a broadcast from May 16,
1994, about 7 months after the first transcript and about 1 month into the genocide:

Those who are at the market you will be told things about those children you
call street children. You will hear that many of them are Inkotanyi. Look at all
of them and then touch on their heads, you will see a sign showing you that
there are Inkotanyi. There is a zero sign shaved at the back side of their heads;
touch and see. You will see RPF children who, in daytime, make themselves
street children but who, in the night, disguise themselves and go to inform RPF
on the situation.
And you people who live down there near Rugunga, even though it is rain-
ing, go out. You will see Inkotanyi’s straw-huts in the marsh where horses are
kept. It is clear then that this place shelters Inkotanyi. I think that those who
have guns should immediately go to these Inkotanyi before they listen to Radio
RTLM and flee. Stand near this place and encircle them and kill them because
they are there.
You are then going to listen to this Inkotanyi child we have captured this
morning. You will know the course of the situation and how Inkotanyi are dare-
devils. You will learn that, our victory against them is certain. It is not even
necessary to negotiate with them. We need ammunitions and enough weapons
so that we can fight against them and no doubt we will have the victory we will
win since Inkotanyi are desperate. People who bring sorcerers, women who
suck and. . . . In a short moment you are going to listen to the interview I have
had with that Inkotanyi but you should put into action what he tells you, arrest
those street children. There are three at Kimisagara near Rose’s. There is a team
of 30 street children which informs Inkotanyi on the number of people who are
on barriers, the kind of weapons they have, how they sleep. I don’t want to take
your time. Listen to this child.

Although this transcript is very difficult to read, it clearly demonstrates an important

transition from the previous example. The first broadcast aired some months before the
genocide began focused more on the classification process. The goal was to portray the
Tutsis as dangerous and dishonest. However, this subsequent broadcast aired after
the genocide began moves past this classification and works to achieve the goals of
symbolization and dehumanization. Notice first that the broadcast lists in great detail
how to identify members of the Tutsi Inkotanyi group. Just as with past genocides in which
members of the majority population were given visual characteristics of the targeted
24 Journal of Language and Social Psychology 31(1)

persons, the broadcasts revealed ways of physically differentiating Tutsi children from
others. The “zero sign” shaved on their heads became a symbol of differentiation. That
symbol enabled the classification of Tutsis as deceitful and dangerous.
The broadcasts also reveal many attempts at dehumanizing the Tutsis. This lan-
guage labels them as daredevils, street children, sorcerers, and “women who suck.”
Other broadcasts labeled these Tutsis as cockroaches needing to be exterminated. The
goal is therefore to hunt them down and totally eliminate them. No attempts at negotia-
tion or problem solving of any kind are suggested. Victory is the main objective. These
kinds of explicit instructions on how to hunt down and find these enemies are apparent
in broadcasts after the April 6, 1994, plane crash that triggered the genocide. After
months of classification, the broadcasts only needed to provide the instructions and the
genocide was in full swing.
Notice that the language is oriented toward power, justice, and mistrust. The power
language commands listeners to identify the Tutsis and hunt them down. The language
is directive and forceful with few qualifications. Consistent with this language are
attempts to show how the Tutsis cannot be trusted. The broadcast calls them daredevils
and sorcerers who pose a significant threat.
The overriding theme of the broadcast is the need to seek justice for these problems.
The only way the broadcast advocates for addressing the problem is to take up arms
and exterminate the enemy as they sleep. The concept of victory is offered as the moral
imperative for achieving this justice and for ultimately escaping the identity trap. The
broadcasts make the case that as long as the Tutsis are allowed to live, they offer noth-
ing but problems. The people can crawl out of the trap and find a certain resolution by
pushing away the enemy completely and eliminating the paradox. The broadcasters
established themselves as the saviors of the people with a unique and forceful under-
standing of the threat posed by the Tutsi minority. They then used powerful language
to make sure the threat was addressed decisively.

Avoiding the Identity Trap

The United Nations passed only 2 resolutions during the genocide in an attempt to
separate the parties and avoid further violence but then another 16 after the genocide
ended. A more preventive approach to these kinds of tragedies is to understand the
clear signals leading up to such an event. The argument forwarded by this article is
that this prevention effort can begin by concentrating on how language promotes
extremist identities and ultimately traps populations into making horrendous choices.
The key to avoiding identity traps is first to recognize that one is present or looming
in the language. Extreme responses to problems such as genocide are not sudden,
unexpected events. They are telegraphed. Left to fester, such language catches people
in an identity trap. As the trap begins to close around the population, it is important for
dissenting voices or external forces to intervene and figure a way out of the trap that
can transform the culture away from one that tolerates or even encourages the transi-
tion to genocide. To achieve this goal, it might be useful to highlight strategies for
Donohue 25

recognizing when identity traps are emerging. First, it is important to note that traps
generally emerge when extremist groups gain access to popular media and begin the
outbidding process that consists of attempts to classify. Often, these classification
attempts can be very subtle. Showing disrespect by dismissing the other’s issues, dem-
onstrating a lack of interest in listening to the other, or labeling the other as unjust or
illegitimate are active attempts to build a climate of mistrust, power, and looking back-
ward. Repeated over time with more intense language or sensitive topics can further
escalate the move from classification to symbolization and dehumanization. These
strategies are common in most international disputes, such as the Palestinian–Israeli
conflict, in which both sides seek to paint a black-and-white picture of the other side’s
words and actions.
Another side of the identity trap that emerges shows a fixation on a linguistic trope.
Tropes represent frames for shaping the rhetorical frame. Putnam (2004) examines the
concept of tropes (the use of metaphors, irony, metonymy, and synecdoche) to under-
stand how individuals reference and manage relational tensions. She focuses specifi-
cally on the use of metonymy (a figure of speech in which a word stands for its
constituent parts, e.g., using the word culture to stand for an organization’s values,
rituals, and myths) and synecdoche (a figure of speech in which the part stands for, or
symbolizes, the whole, e.g., substituting the word crown to stand for the role of the
king or queen). In other words, how are people using words to conceptualize complex
relational dialectics that define the nature of the parties’ interdependencies? In her
research, Putnam (2004) focuses on how tropes are used to negotiate autonomy and
interdependence in a teacher bargaining context. She found that the term language
referred to the creation and discussion of policy issues and the term money referred to
any budget-related item. The bargaining often revolved around who owned the money
and the ability to craft the language that would ultimately guide the negotiation. Thus,
these concepts were used to negotiate the dialectical tensions of autonomy and inter-
dependence between the parties.
In classification language, similar kinds of tropes often emerge that serve to build
identity traps. Several tropes are featured in the November 24 broadcast. The most
obvious is the use of the pejorative nickname “Inkotanyi” to refer to the Tutsis. The
speakers made several references to this nickname and often formed the broadcast
around it to ensure that the classification objectives were reached. The other trope that
appears is use of the word negotiation as a way of denigrating what would be a more
reasonable and rational way of resolving the conflict. The repeated references to nego-
tiation seek to justify the preferred strategy of the speaker’s extremist Hutu supporters
and characterize it as a symbol of the moral high ground in the dispute. Such tropes are
commonly used by groups seeking to build a case for their side, or in this case the
listeners, to adopt a particular frame for the dispute and ultimately a course of action
to resolve the dispute.
These tropes often serve as windows for disclosing paradoxes. The label “Inkotanyi”
is not only used to classify all the individuals who identify with the Tutsi tribe, but it
also serves as a linguistic tool to justify actions. For example, the Hutu fixation in
26 Journal of Language and Social Psychology 31(1)

calling all of the people who oppose their extremist vision as “Inkotanyi” represents a
classic identity trap. The Hutu extremists use this term in their language to express
their fears and press their demands with the listeners and other supporters while also
using it to resist demands and assert rights. Thus, whatever trope either side uses rep-
resents a bid for the support of a specific identity while also signaling the entrance into
the identity trap. The fight between the Hutus and Tutsis was not about the appropriate
form of government or specific grievances. Rather, it was about a lack of trust and
affiliation, and a refusal to work together for a better future. Listeners were led into an
identity trap that deepened dramatically as the broadcasts progressed.
Once caught within a trap, which is often very difficult to avoid, it is important to
understand how the interaction can be transformed to escape it. The most obvious
strategy is to recognize first that the other is setting a trap with some kind of identity
attack. Reciprocating the attack springs the trap and captures everyone. Avoiding the
attack by reestablishing another topic or focusing on the substantive issue begins to
circumvent the trap. Another strategy for avoiding the trap is to simply decrease inter-
dependence. Sometimes it is useful to separate for a while and not continue to interact.
Or one can bring in a mediator to create and enforce interaction rules.
The Rwandan situation is typical of many genocidal situations in which a power-
less group finds itself in a death spiral at the hands of the more powerful party. Caught
in an existential struggle, the powerless group must pursue a unilateral strategy to
avoid the identity trap and deescalate the conflict. History tells us that outside inter-
vention, if it ever materializes, comes only after significant atrocities have already
taken place. And often the interventions, as in the case of Rwanda, are ineffective at
stopping the violence. Recent events in Libya suggest that Arab League and United
Nations interventions may have prevented a humanitarian crisis in that country. Far
too often genocide continues unabated until a major event blunts the attacks. Within
their existential struggle, what could the Tutsis have done as a unilateral strategy?
Given their limited resources, their last hope would have been outside intervention
when it was clear that a climate of hate was emerging in Rwanda. Perhaps there is a
case for a genocide that might alert the United Nations or the International Criminal
Court (ICC) to take action.

Establishing an Early Warning

System for Genocide Avoidance
Language that classifies signals the slippery slope toward genocide. What kinds of
systems ought to be in place to both detect such language and then to mitigate its
effects? Creating a linguistic early warning system would certainly be easy to imple-
ment. One could imagine a system in which a computer program sponsored by, for
example, the ICC could analyze the verbally aggressive language coming from vari-
ous media broadcasts or computer websites and then look for trends over time to
determine the appropriateness and timing of an intervention. That system could focus
on the three constructs proposed here including power/affiliation, trust/mistrust, and
Donohue 27

Figure 1. Identity Trap framework

forward-/backward-looking language. If that system were fed into a central ICC site
that monitors such language, it would be possible for those analyzing the information
to create a list of countries/areas on a “watch list” of some kind. As the language
intensified, the more vulnerable places would be targeted for some kind of action.
Figure 1 presents an Identity Trap framework that provides a useful way of concep-
tualizing how a genocide alert system might function. Using Stanton’s (2004) frame-
work, it would be possible to adopt the seven power themes he describes in his
genocidal stages. If the rhetoric were analyzed and the verbally aggressive language
began to move into the red power theme zone, and away from the green affiliation
theme zone, then a body such as the ICC could begin to flag that nation or region for
greater scrutiny. If verbally aggressive language progressed past the first stage or two
into more dehumanizing themes, then it could be expected that the aggressors were
moving beyond language-based actions and more toward the implementation of physi-
cal violence in the form of militias that target victims.
What action would be appropriate to mitigate classification language? Aside from
military intervention, there are various aid programs that seek linguistic moderation
through economic development. Factions from different groups could learn to work
together and benefit from jobs and increased prosperity. Underlying these efforts
could also be attempts to initiate dialogue groups that allow individuals from different
sides to simply become more comfortable with one another. Would that have been
possible in preventing the Rwandan genocide? Certainly the government was weak
and vulnerable to an onslaught of classification language. Yet stronger interventions
aimed at addressing grievances might have created a different climate, one capable of
heading off the horrific escalation. Clearly, public language matters; it creates a
28 Journal of Language and Social Psychology 31(1)

context for how people can interact with one another. Will they choose a productive
path that emphasizes understanding and negotiation or will they choose a destructive
path that emphasizes vilification through classification and violence? The point of this
article is that language is real in the sense that it really matters. Language creates a
context that moves people to action, particularly when access to multiple communica-
tion sources is fairly limited, as it was in Rwanda. Events taking hold now in the Arab
world are another example of how language matters. As protestors from Tunisia to
Egypt and Syria to Libya find the communication tools and form the language needed
to cast leaders in an oppressive light, they are practicing identity trap politics. Certainly
as these events illustrate, there is no shortage of conflicts around which to devote pre-
ventive resources. However, the more we come to grips with the reality that language
may be the key to quick and effective conflict resolution, we can create interventions
to detect and avoid genocide.

Declaration of Conflicting Interests

The author declared no potential conflicts of interest with respect to the research, authorship,
and/or publication of this article.

Funding
The author received no financial support for the research, authorship, and/or publication of this
article.

References
Brown, P., & Levinson, S. (1978). Universals in language usage: Politeness phenomena. In E. Goody
(Ed.), Questions and politeness: Strategies in social interaction (pp. 56-289). Cambridge,
England: Cambridge University Press.
Donohue, W. A., & Druckman, D. (2009). Message framing surrounding the Oslo I accords.
Journal of Conflict Resolution, 53, 119-145.
Donohue, W. A., & Hoobler, G. D. (2002). Relational frames and their ethical implications
in international negotiation: An analysis based on the Oslo II negotiations. International
Negotiation, 7, 143-167.
Hoffman, A. M. (2002). A conceptualization of trust in international relations. European Jour-
nal of International Relations, 8, 375-401.
Irmer, C. G., & Druckman, D. (2007). Process and context in international negotiation. In
C. Dupont (Ed.), Négociation et transformations du monde [Negotiation and world transfor-
mations] (pp. 117-136). Paris, France: Publibook.
Lewicki, R. J., & Litterer, J. A. (1985). Negotiation. Homewood, IL: Irwin.
Lewicki, R.J., McAllister, D.J., & Bies, R.J. (1998). Trust and distrust: New relationships and
realities. Academy of Management Review, 23, 438-458.
Lewicki, R. J., & Wiethoff, C. (2000). Trust, trust development, and trust repair. In M. Deutsch
& P. Coleman (Eds.), The handbook of conflict resolution: Theory and practice (pp. 86-107).
San Francisco, CA: Jossey-Bass.
Donohue 29

Moon, C. (2008). Narrating political reconciliation: South Africa’s Truth and Reconciliation
Commission. London, England: Centre for the Study of Human Rights.
Putnam, L. L. (2004). Dialectical tensions and rhetorical tropes in negotiations. Organization
Studies, 25, 35-53.
Rokeach, M. (1960). The open and closed mind. Oxford, England: Basic Books
Simon, J. (2006). Of hate and genocide. Columbia Journalism Review, 34, 9.
Stanton, G. H. (2004). Could the Rwandan genocide have been prevented? Journal of Genocide
Research, 6, 211-228.
Whillock, R. K. (1995). The use of hate as a stratagem for achieving political and social goals.
In R. K. Whillock & D. Slayden (Eds.), Hate speech (pp. 28-54). Thousand Oaks, CA:
Sage.
Zartman, I. W. (2000). Ripeness: The hurting stalemate and beyond. In P. Stern & D. Druckman
(Eds.), Conflict resolution after the cold war (pp. 225-250). Washington, DC: National Research
Council.
Zartman, I. W., & Kremenyuk, V. (2005). Peace versus justice: Negotiating forward and back-
ward-looking outcomes. Lanham, MD: Rowman & Littlefield.

Bio
William A. Donohue (PhD, Ohio State University) is a Distinguished Professor of
Communication at Michigan State University. He has published extensively in the areas of
conflict, communication, negotiation, and mediation while also conducting workshops and
other intervention activities focusing on communication, leadership development, and conflict
management. He has coauthored Framing Matters: Perspectives on Negotiation Research and
Practice in Communication. He is a recent past president of the International Association for
Conflict Management and is on the editorial board of several major journals.

View publication stats