Power and the Rhetorical Manipulation of Cognitive Dissonance

David G. Casagrande

Paper delivered at the Presidential Session of the Annual Meeting of the American
Anthropological Association. Dec. 15-19, 2004. Atlanta, GA

Introduction
Leon Festinger proposed a theory of cognitive dissonance in which he argued that
a person experiences psychological discomfort if two cognitions are equally important
but logically contradictory (Festinger 1956). As a result of this discomfort, the subject
changes behavior or seeks new information to enhance one of the cognitions or seeks to
reduce the importance of the other. This theory became a dominant paradigm in
psychological research and led to thousands of studies in which subjects were forced to
deal with their cognitive dissonance under controlled conditions so that researchers could
observe changes in individual values, opinions or behaviors. The bulk of this research
focused on individual cognition and veered away from Festinger’s focus on group
dynamics. Experimental results in the literature strongly suggest that cognitive
dissonance does not occur until logical inconsistencies are exposed via social interaction
like conversations. Billig (1987) argued that cognitive dissonance is not confronted
except in discourse, and the interlocutor’s key purpose is maintaining positive self-
perception. Specifically because the creation of cognitive dissonance is social, I suggest
that people are more likely to avoid or reduce cognitive dissonance via shared
communicative strategies than to change individual cognitions or behaviors.
Cognition involves various levels of abstraction ranging from simple, contrastive
sets to highly inclusive but abstract sets. People share frameworks for linking
propositions within levels of abstraction and between levels--what cognitive
anthropologists usually refer to as ‘cultural models’ or ‘cognitive models’. But what has
not been addressed adequately in the literature is that these models often include
conflicting logics. Normally, we don’t confront such contradictions. But at times, for

1
example in discourse, we are forced to confront them. This causes cognitive dissonance,
which is an emotional (affective) response to cognitive processes.
I propose that there are a variety of discursive strategies aimed at avoiding or
ameliorating cognitive dissonance.1 One strategy is to move the discourse to more
abstract concepts, where one can obfuscate and/or ignore contradictions at lower levels.
Thus, we have a discursive solution to an emotional response to cognitive processes. I am
calling this process “abstractification.” An interlocutor attempting to move the discourse
to a higher level will do so by invoking a shared abstract belief like “that’s just the way
God made it.” The particular belief chosen reflects the expectation of the first interlocutor
that the other interlocutor shares the abstract belief. The process of abstractification using
shared beliefs to escape cognitive dissonance is probably universal. The particular beliefs
used are culturally relative.
I hypothesize further that some discursive strategies or behavioral norms keep
individual cognition or discourse abstract to preclude cognitive dissonance. This process
can enhance social cohesion, but the cost is that logical inconsistencies are never
resolved. What’s more, those who are best at manipulating levels of abstraction have a
political advantage. Thus we have political advantage (i.e., power) deriving from a
discursive solution to an emotional response to cognitive processes.
The manipulation of cognitive dissonance by relocating discourse and cognition
to higher levels of abstraction may reduce the potential for adapting to environmental
change. Logical inconsistencies created by environmental information in conflict with
existing beliefs at lower levels of abstraction may be glossed over by the emotional
benefits of reducing cognitive dissonance at higher levels. Leaders of political systems in
crisis attempt to maintain social cohesion by reducing cognitive dissonance by using
increasingly abstract rhetoric. However, such a system could eventually collapse if the
critical environmental information causing logical inconsistencies at lower levels
continues to be ignored. During periods of environmental stress, and as the situation

1
The notion of discursive strategies for manipulating cognitive dissonance raises interesting cross-cultural
questions. Cognitive dissonance is probably universal. But discursive processes that create or resolve
cognitive dissonance are culturally relative (Nuckolls 1993). For example, European Americans desire
logical order and we experience dissonance when we don't find it. Some other cultures may not desire
logical order, or at least don't look for consistency in the same cognitive domains as most Americans.

2
worsens, societies often exhibit proliferation of monumental architecture and/or
militarization based on abstract notions like “good” and “evil”. Examples may include
the Classic Maya, Easter Island, the Soviet Union, and current attempts to use religious
rhetoric in the U.S. to avoid addressing climate change (e.g., “It is God’s will”).
Here I present some examples from my research. Since 1998 I have been studying
how Tzeltal Maya in Chiapas, Mexico share information about medicinal plants. My
original idea was to determine how shared cognitive models guided the selection of
medicinal plants and facilitated communication about them. In other words, I was
studying the functionality of the group’s cultural models—how logical relationships
between shared propositions functioned to structure the medicinal plant knowledge
system. A fundamental assumption of this research was that people must share logically
coherent explanatory models in order to facilitate the discovery of cures and share that
information. For example, previous research showed that how a plant tastes is an
important culturally-shared indicator of how that plant may cure. One of my predecessors
(Brett 1998) hypothesized that the Tzeltal associate bitterness with respiratory illnesses. I
found this to be partially true. The following is a typical excerpt from an interview:

Q: Bi yu’un ya sposhta obal ch’a bakal.

Why does Salvia lavanduloides cure cough?
A: Yu’un ha lom ch’a. Ya sposhtaik obal ch’a wamaletik.
Because it’s bitter. Bitter plants cure cough.

A Chi Square analysis of data from 53 interviews using 203 plant species showed that
cough treatments were significantly more likely to be bitter than other taste categories
like astringent, sour, sweet or pungent (X2 = 30.15, p < 0.001). But I also found many
contradictions in this (and other) shared explanatory models. For example, later in the
same interview, I was asking about a different plant:

Q: Bi yu’un ya sposhta tza’nel yakan k’ulub wamal.

Why does Verbena litoralis cure diarrhea?
A: Ha lom ch’a.

3
 Because it’s very bitter.
Q: Ya bal a poshta obal yakan k’ulub.
Does Verbena litoralis cure coughs?
A: Ma poshta
No.
Q: La a’wal ya sposhta obal ch’a wamaletik. Bi yu’un ma sposhta obal ha’mene.
You said earlier that bitter plants cure coughs, why doesn’t this one cure
coughs?
A: Mash k’il . . . ha te, ha nash . . . mash k’il. Ha nash la spas El Señor.
I don’t know . . . well, it’s just . . . I don’t know. That’s just the way God
made it.

The rate at which interviewees contradicted themselves in this way was 29%. How could
such logical inconsistencies persist in a system so important for survival? It was obvious
that these potential contradictions either simply did not occur to my interviewees, or if
they did, their importance was somehow minimized. What might be the point of having a
shared model if its logical incoherence inhibits its functionality?

Hierarchical levels of abstraction

My use of the phrase “hierarchical level of abstraction” draws heavily from
Bateson’s “hierarchical orders of learning” (Bateson 1987:279-308). Let me use a
taxonomic example from ethnobiology to explain. A person discriminating between a
sparrow and a robin is doing so within a particular set of criteria of set inclusion—a
specific taxonomic level—and the saliency of this contrastive level tends to be consistent
across societies (Berlin 1992:19). More hierarchically inclusive categories like “birds” or
“living things” are more abstract, and the criteria that people use to form these categories
tend to vary across cultures. Hence, European Americans would consider a New Guinean
cassowary to be a large flightless bird. The Kalam of New Guinea, however, place the
cassowary in a category with humans, rather than birds (Bulmer 1967). At higher levels
of abstraction, boundaries and criteria for set inclusion are more flexible. The Tzeltal
lexeme tza’nel refers to the directly observable physical symptom we call “diarrhea.” The

4
more abstract lexeme poxil (to cure) can refer to the cessation of physical symptoms like
diarrhea, or the restoration of social relations by returning one’s soul to its proper
location, both of these in combination, or neither.
Gregory Bateson’s “hierarchical orders of learning” help to clarify this idea.2 In
Bateson’s Type I learning, premises are learned through rote repetition. For example,
after repeated observations a Tzeltal child learns that drinking tea from Salvia
lavanduloides relieves coughing. Level II learning is learning about contexts of level I
learning, or applying premises from one set of observations to another context. It is
inductive or deductive. For example, the Tzeltal child may reason that Salvia and several
other plants that cure coughing are bitter, so bitter plants must cure coughing. Most
importantly, level II learning includes social context. According to Bateson, much of our
most important communication is about communication (Bateson1987:201-227). The
Tzeltal child learns social information like “mother knows best” because a mother knows
lots of cures. The child might also deductively reason that other mothers also know best.
When trying to extract the most relevant information for Level II learning, some
logical inconsistencies must be ignored, while others are deemed more important and are
reconciled. For example, the child might be able to ignore that, in addition to cough
remedies, some diarrhea remedies are bitter too. Due to the fuzzy nature of hierarchical
inclusiveness, logical inconsistencies persist at lower levels of learning, but can be
ignored. An example of a higher level, more abstract, logic is: medicinal plants are
imbued with a power to cure, bitter plants cure diarrhea, but bitter plants can be used to
cure other things as well. Note that at level I, “Verbena cures diarrhea,” the potential for
logical inconsistency does not exist, only falsification of a premise. It is only at
successively higher levels of abstraction that logical contradictions are either created or
reconciled.
Logical contradictions can become highly problematic with level III learning,
wherein we learn and apply logical propositions to the context of the context. For

2
See also Pierce’s typology of hierarchical abstraction, Klir’s (1985) epistemological hierarchies, and
Wegner and Vallacher’s “action identity” scheme for behavioral hierarchies. In addition to Bateson’s
(1958, 1987) logical types of learning, these are all examples of hierarchical organization of abstraction,
although not necessarily mutually exclusive or taxonomic.

5
example, human subjects in laboratory experiments contextualize their experience within
the motives of a university or broad social ethics. As Gregory Bateson stated: “If . . . the
creature is driven to level III by ‘contraries’ generated at level II, then we may expect that
it is the resolving of these contraries that will constitute positive reinforcement at level
III” (Bateson 1987:305). In other words, level III learning is where we learn the socially
acceptable rules for dealing with cognitive dissonance, including socially acceptable
techniques of abstractification in discourse. Level III learning is the social process that
recursively validates discursive strategies. When I confronted my Tzeltal informant with
the logical inconsistency regarding the bitterness of plants, the socially acceptable
strategy was to default to a higher level of abstraction and say that he didn’t know why
bitter plants also cure diarrhea, he said “that’s just the way God made it.” This is socially
acceptable in that both interlocutors assume that the other interlocutor shares a valid but
fuzzy notion of “god making cures,” and that this fuzzy notion is more important than
subordinate logical contradictions.
In 2003 I started interviewing people who live in the Phoenix, Arizona
metropolitan area about water scarcity and the way they use water. This population has
one of the highest per capita water consumption rates in the country. Although they live
in a very arid climate, many have lawns and swimming pools (Yabiku et al. 2008). They
also are aware that water is scarce. I conducted these interviews during a severe drought
that received extensive media coverage (Casagrande et al. 2007). I was interested to see
how they might deal with the potential cognitive dissonance resulting from these facts.
During interviews, I attempted to expose logical inconsistencies to see what abstract
beliefs interviewees would invoke to avoid cognitive dissonance. I made notes to myself
when interviewees said things that could be logically inconsistent. As the interview
proceeded, I attempted to covertly steer the conversation so that potentially conflicting
logics would become closer to each other in the mind of the interviewee. In one example,
a woman told me about her pool, which requires a lot if water due to evaporation. Later,
she was telling me about the nearly empty reservoirs she had seen, which she attributed to
the drought and excessive use of water by people. After a slight pause, I re-introduced the
topic of the pool as follows:

6
 Q: . . . so you have a pool . . .
A: yeah. It’s small.
Q: but does it use a lot of water?
A: (hesitating) “You know . . . here’s what I don’t understand, why are they
spending all this money on Mars? Why don’t they build a . . . uh . . . a big plant to
change salt water?

As the logical inconsistency of being concerned about drought and having a

water-intensive pool began to materialize into consciousness, the interviewee abruptly
moved the discourse to the more abstract, widely-shared American “belief in the
technological fix.” Note that she knows she is talking to college professor, and chooses a
belief (most likely subconsciously) that she is sure I will share. During 33 interviews, I
was able to create 13 cognitive dissonance events that resulted in interviewees invoking
abstract beliefs (Table 1).

Table 1. Shared beliefs presented by interviewees in metropolitan Phoenix, Arizona

to avoid cognitive dissonance related to water issues.

• Technological fix (n = 5): e.g., more dams, ocean water desalination, a pipeline to
the Great Lakes
• Scapegoating (n = 3): e.g., it is the fault of Mexican immigrants, people relocating
from the Midwest who want lawns, or water-hoarding by the state of California
• Distrust of experts (n = 3): e.g., there isn’t really a drought, water is mismanaged
• God / religion (n = 2): e.g., God will provide a solution

Perhaps the most widespread use of Festinger’s theory of cognitive dissonance

outside of academia is in marketing. Metropolitan Phoenix has been consistently and
successfully marketed since the 1950s using the abstract cultural model of ‘the oasis’. In
many campaigns, desert imagery is intentionally juxtaposed with water-intensive imagery

7
with the effect of creating inconsistencies that are reconciled using other images of palm
trees or slogan’s like “the desert is a myth.”3
The notion of hierarchical levels of abstraction provides an explanation for why
people appear to act in contradiction to some of their beliefs. If researchers fail to specify
hierarchical levels of learning or abstraction we have no way of explaining why a Tzeltal
mother might say bitter plants cure cough, but use them to treat diarrhea, Americans are
concerned about pollution and global warming, but continue to drive SUVs, or why
people in Phoenix, Arizona are worried about water and drought, but make minimal
attempts to curtail personal water use. These paradoxes are resolved by situating
discursive data within hierarchical levels of abstraction.

Political power and rhetorical manipulation of cognitive dissonance

How does a communicative system loaded with logical contradictions function to
allow people to react to critical environmental information and adapt to environmental
change? In the case of Tzeltal medicinal plants, I believe that detailed, individual logical
models guide experimentation by experts. Only enough schematic details are widely
shared by non-experts to allow for legitimization of expert knowledge (see also Garro
2000). In other words it is more important for the larger group to agree about general
principles of plant curing than potentially conflicting details. Thus, important information
can flow from experts to the rest of the population.
For example, I found the Tzeltal hot/cold medicinal classification system is even
more contradictory than that of taste classification. I recorded a conversation in which an
expert was trying to explain a treatment to two non-expert men, whose confusion about
hot and cold classifications led the expert to say:

K’ishin ta poshta mak sik’il ta poshta wamaletik. Ya hta’etik te poshiletik. La

spas wamal ta sposhta spisil chamel El Señor. Ya hta’etik.
Sometimes plants are hot, sometimes cold. We find the cures. God made a
plant to cure every illness. We find them.

3
“The Desert is a Myth” is the slogan used by the Greater Phoenix Convention & Visitors Bureau in its
2004 advertising campaign.

8
 The two novices appeared to be convinced.

This raises an interesting theoretical prospect. Are those people who are very
good at manipulating other people’s cognitive dissonance more likely to accumulate
power?4 It is no coincidence that President George W. Bush and his advisors consistently
and strategically inserted highly abstract themes like “good” and “evil” into the national
discourse about terrorism and the Iraq war after 9/11. No social decision to go to war is
simple and clear cut. If the American public dwells on potentially conflicting information
they will experience cognitive dissonance. It is not my intention to disparage the
president. In fact, I’m arguing that manipulation of cognitive dissonance is natural and
necessary for maintaining social cohesion. The American public, indeed any public, looks
to leaders to preempt cognitive dissonance. Most importantly, research has clearly shown
that people want to avoid cognitive dissonance very much and are therefore willing to
embrace abstract rhetoric of their leaders, even when there are obvious contradictions, or
their leaders are invoking ideas that would not be palatable in other contexts. Leaders, in
turn, discourage people from critically examining details or considering potential
contradictions by steering the public discourse away from potential inconsistencies.
Use of abstract concepts in public rhetoric is absolutely necessary to maintain
social cohesion, especially in a nation-state as diverse and large as the United States.
Somebody has to manipulate levels of abstraction or the system will fall apart. It may be
that individual talent for manipulating cognitive dissonance is a fundamental precursor to
the accumulation of power and maintenance of social cohesion. This probably includes
any type of social organization, including foragers, subsistence horticulturists like the
Tzeltal, or post-industrial Americans. This creates what I call the “leader’s paradox.” The
more environmental information conflicts with popular beliefs, the more necessary it is
for a leader to promote abstract rhetoric to hold the society together, but the less likely it
is that the society will act to correct environmental problems. Also, the more complex a
culture is, and the more diverse the beliefs in the population are, the more a leader must
4
This was Festinger’s (1956) original proposition in his fascinating description of the cult who believed
the world would end on a specific day. When the world did not end, the cult leader actually accumulated
more power as the members attempted to avoid their cognitive dissonance.

9
promote abstract concepts like “freedom” or “terror” as in the case of the U.S. Again, it is
less likely that such a society will act to correct logical inconsistencies. This proliferation
of logical inconsistencies can continue only to a point at which systemic collapse or some
major crisis re-organizes the system.

Conclusion
Leon Festinger proposed a theory of cognitive dissonance in which he argued that
a person experiences psychological discomfort if two cognitions are equally important
but logically inconsistent (Festinger 1956). As a result, the subject seeks new
information, or changes behavior, to enhance one of the cognitions or to reduce the
importance of the other. More attention should be paid to social processes implied by his
theory. Logical inconsistencies arise primarily in discourse or other social interaction, not
in an individual’s private thoughts. Much social discourse involves collective strategies to
avoid, manipulate or otherwise ameliorate cognitive dissonance that arises through social
interaction. A widely-used rhetorical strategy for avoiding logical inconsistencies is to
steer the discourse toward more abstract and logically indefensible beliefs. These beliefs
are usually widely shared within a culture. An interlocutor is confident that such beliefs
will be shared by another member of his or her culture. I have conceptualized political
power as the ability to rhetorically steer people away from their cognitive dissonance by
invoking abstract beliefs. This produces the benefit of social cohesion, but also creates
the potential cost of reducing the ability to react to important information, which
preserves maladaptive, lower-level logical inconsistencies. This creates the “leader’s
paradox.” In complex pluralistic systems, such as states and modern nations, the
proliferation of abstract evasive mechanisms may preclude resolution of logical
inconsistencies needed for adaptation to environmental change, eventually leading to
systemic collapse.

10
References Cited
Bateson, Gregory
1958 Naven: A Survey of the Problems Suggested by a Composite Picture of the
Culture of a New Guinea Tribe Drawn from Three Points of View. Stanford: CA:
Stanford University Press.
Bateson, Gregory
1987 Steps to an Ecology of Mind: Collected Essays in Anthropology, Psychiatry,
Evolution, and Epistemology. Northvale: NJ: Jason Aronson Inc.
Berlin, B
1992 Ethnobiological Classification: Principles of Categorization of Plants and
Animals in Traditional Societies. Princeton, NJ: Princeton University Press.
Billig, M.
1987 Arguing and thinking: A rhetorical approach to social psychology. Cambridge:
Cambridge University Press.
Brett, John A.
1998 Medicinal plant selection criteria: The cultural interpretation of chemical senses.
Journal of Applied Botany-Angewandte Botanik 72(3-4):71-74.
Bulmer, Ralph
1967 Why is the cassowary not a bird? A problem of zoological taxonomy among the
Karam of the New Guinea Highlands. Man 2:5-25.
Casagrande, D. G., D. Hope, E. Farley-Metzger, W. Cook, and S. Yabiku.
2007 Problem and opportunity: Integrating anthropology, ecology, and policy
through adaptive experimentation in the urban American Southwest. Human
Organization, 66(2), 125-139.
Festinger, L, H. Riecken, and S. Schacter
1956 When prophecy fails. Minneapolis: University of Minnesota Press.
Garro, Linda C.
2000 Remembering what one knows and the construction of the past: A comparison
of cultural consensus theory and cultural schema theory. Ethos 28(3):275-319.
Klir, G. J.
1985. Architecture of systems problem solving. New York: Plenum Press.

11
Neuman, Y., and I. Tabak
2003 Inconsistency as an Interactional Problem: A Lesson from Political Rhetoric.
Journal of Psycholinguistic Research, 32(3), 251-267.
Nuckolls, C W
1993 The anthropology of explanation. Anthropological Quarterly 66(1):1-21.
Wegner, D. M., and R. R. Vallacher
1986 Action Identification. In E. T. Higgins. & R. M. Sorrentino (Eds.), Handbook of
motivation and cognition: Foundations of social behavior. Vol 1 pp. 550-582). New
York: The Guilford Press.
Yabiku, S. T., Casagrande, D. G., & Farley-Metzger, E.
2008 Preferences for Landscape Choice in a Southwestern Desert City. Environment
and Behavior, 40(3), 382-400.

12