Defining Manipulative Discourse: The Pragmatics of Cognitive Illusions

TL;DR: The authors proposed an alternative account of manipulation couched in the relevance-theoretic framework which treats manipulation as a two-step communicative attempt at misleading the context-selection process when interpreting a target utterance.

Abstract: Manipulative discourse has attracted a lot of attention in various adjacent domains of linguistic research, notably in rhetoric, argumentation theory, philosophy of language, discourse analysis, pragmatics, among others. We start with a review of the existing definitions provided in these fields and highlight some of the difficulties they encounter. In particular, we argue that there is still a need for an analytic model that makes predictions about manipulative discourse. We propose an alternative account of manipulation couched in the relevance-theoretic framework which treats manipulation as a two-step communicative attempt at misleading the context-selection process when interpreting a target utterance. We argue further that such attempts systematically exploit the inherent weaknesses or flaws of the human cognitive system that are amply discussed in cognitive psychology under the heading of “cognitive illusions”. We claim that such a model correctly captures classical instances of manipulative discourse which fall outside the scope of other accounts.

Summary

Introduction

• Scholars from disciplines in the Humanities as varied as social psychology, philosophy, anthropology, communication science, argumentation theory, linguistics, psychology and critical discourse analysis have tried to tackle the phenomenon or to explore some of its facets, but consensus, let alone a unifi ed account, has not been reached yet.
• The authors proposal aims at reviewing these criteria and explores a relatively new direction for research on manipulation, i.e. a cognitive one.
• Some of these criteria, the authors will also argue, fail to make the right predictions as they either over-generate or under-generate.

1. Th e Heterogeneous Nature of Manipulation: Available Clues

• Contemporary research on manipulation has usually been concerned with the nature of the manipulative message, its properties, and the social conditions in which manipulation is likely to occur.
• Th is section discusses the limits encountered by such accounts which try to isolate the necessary features that make up manipulative discourse.

1.1. Truth-Conditions and Felicity Conditions

• Truth and felicity conditions have been proposed as relevant criteria defi ning manipulative strategies in terms of falsity and insincerity (Rigotti, 2005 ); the main assumption behind this view roughly amounts to considering that some of the manipulative message’s features, in terms of the content it encodes, do not comply with the communicative standards of verbal interaction.
• From a speech-act-theoretical perspective, it could be envisaged to consider that manipulation violates felicity conditions, and among them, the sincerity condition in particular; the authors mentioned earlier in this respect the scenario of a politician uttering a commissive without actually being committed to the ulterior satisfaction of the propositional content conveyed.

1.2. Speaker Interest

• As van Dijk notes (2006: 360), “manipulators make others believe or do things that are in the interest of the manipulator and against the best interests of the manipulated”.
• Saussure also highlights the connexion between speaker interest and the manipulative attempt: […] communication is manipulative when the speaker retains some relevant information, or provides the correct information in order for the hearer to conclude that he should behave in a way which favours the speaker’s interests, without being aware of it.
• Now, several problems arise when examining the issue in further detail.
• It can hardly be claimed that this is an instance of hearer interest violation, to the extent that both participants in the exchange can reasonably be taken to pursue the satisfaction of the sick hearer’s interest (i.e. to get better).
• People do have confl icting interests, but this does not mean that manipulation is only about violating a hearer’s interest: some violations, which in this sense can appear to be incidental, may happen in the pursuit of a goal which, in the end, satisfi es one of the hearer’s main interests.

1.4. Social Inequality

• Research in Critical Discourse Analysis (e.g. van Dijk, 2006) and in psychosocial science (e.g. Milgram, 2004; Cialdini, 2009 ) stresses the importance of the social conditions manipulative communication builds on.
• While the authors do not deny that social parameters indeed contribute to successful manipulation, there are reasons to doubt that these are necessary conditions.
• Van Dijk is aware of the diffi - culty but dismisses these cases on the grounds that they hinge upon personal psychological factors: I limit my analysis to social criteria, and ignore the infl uence of psychological factors, such as character traits, intelligence, learning, etc. (van Dijk, 2006: 362).
• The authors propose that an appropriate pragmatic model of manipulation must shift its focus on and account for the cognitive processes which underlie the interpretation of manipulative discourse.
• (van Dijk 2006: 364) We will see in section 2, how a more comprehensive view of manipulative communication can extend van Dijk’s account in order to be able to deal with cases of manipulation that go beyond relations of social dominance.the authors.the authors.

1.5. Intention

• Another speaker-oriented parameter which has been regarded as relevant to research on manipulation, though a highly problematic one from a descriptive point of view, is intentionality.
• Providing a negative answer to this question amounts to considering that people may accidentally manipulate others.
• Furthermore, the authors claim that the focus of the account should be on the cognitive processes of interpretation.
• It is precisely this assumption that will underlie their discussion of manipulative communication.

2. An Alternative Account

• In the remaining sections the authors will show how a new theoretical model of manipulation can be captured in relevance-theoretic terms, and how the latter can help us 8 See section 3 below.
• Crucially, whereas many accounts assume that manipulation is a form of inherently fl awed communicative activity, the authors argue that, from an interpretative perspective, manipulative discourse relies on the same context-construction or context selection procedure as in the case of a “normal” utterance.
• In an experimental design testing the impact that (1) and other similar illusions have on people, Erickson & Mattson ( 1981 ) found that up to two thirds of the subjects who were tested would miss the anomaly and be tricked into answering “two”.
• Allott & Rubio Fernández (2002) argue strongly for a pragmatic explanation of these phenomena and propose that they are traces of a form of shallow processing that ensues logically from the general principles governing human communication and in particular the principle of relevance.

Th eory

• As the authors saw, Relevance Th eory (see Sperber & Wilson, 1995 ) provides a model for standard inferential communication which draws on the assumption that speaker meaning is arrived at through the contextualisation of the utterance that is being interpreted.
• Th is process is governed by a principle of optimal relevance whereby the addressee tries to maximise the cognitive eff ects he can generate from the utterance while minimising the cognitive eff orts he has to put into the construction of the appropriate context of interpretation.
• Th ey write that [h]umans are not in the business of simply assessing the relevance of new information.
• The cognitive processing of Moses illusions is aff ected by the fact that the addressee/subject is busy performing a question-answering task and is therefore assessing relevance in connection with the task at hand.
• It is also equipped to handle recent infamous examples of manipulation that the authors are all familiar with.

4. A Concrete Example

• A striking case of manipulation that took place in the US in the aftermath of the 9/11 attacks against the World Trade Centre and the Pentagon conveniently attracted a lot of attention from the media.
• Wilson, a former US ambassador, was one of the persons hired to investigate a lead which linked Iraq with Niger in an attempt to buy uranium for nuclear weapons.
• As is obvious from the synopsis provided above, the authors see that the fi rst manipulative step consists in establishing a link between terrorist attacks and Iraq, thereby limiting the context selection process to C= {9/11}.
• More generally, this example shows that a key to context selection constraint, and therefore manipulation, relies on salience.
• 10 Finally, coming back to their earlier example of second-hand manipulation, the authors see how a credulous disciple can still be indirectly manipulative, in spite of his not having the manipulative intention which is part of the initial strategy.

5. Conclusion: Cognitive Optimism

• Sperber, Cara & Girotto (1995) propose an interesting view of the cognitive underpinnings of the quest for optimal relevance.
• More precisely, in order to increase the performance of the interpretative system, humans tend to take cognitive shortcuts.
• Sperber, Cara & Girotto write that […] people are nearly-incorrigible “cognitive optimists”.
• 90) The authors pragmatic approach of manipulation shows that it is precisely this kind of cognitive optimism which is exploited by manipulative strategies, also known as (1995.
• Also, this account seems capable of handling cases and manipulative scenarios which do not fi t into other models.

Full text

International Review of Pragmatics 1 (2009) 348–370
brill.nl/irp
© Koninklijke Brill NV, Leiden, 2009 DOI 10.1163/187730909X12535267111651
D e fi ning Manipulative Discourse:  e Pragmatics
of Cognitive Illusions
Didier Maillat and Steve Oswald
University of Fribourg and University of Neuchâtel, Switzerland
didier.maillat@unifr.ch and steve.oswald@unine.ch
Abstract
Manipulative discourse has attracted a lot of attention in various adjacent domains of linguistic
research, notably in rhetoric, argumentation theory, philosophy of language, discourse analysis,
pragmatics, among others. We start with a review of the existing defi nitions provided in these
fi elds and highlight some of the diffi culties they encounter. In particular, we argue that there is
still a need for an analytic model that makes predictions about manipulative discourse. We pro-
pose an alternative account of manipulation couched in the relevance-theoretic framework
which treats manipulation as a two-step communicative attempt at misleading the context-
selection process when interpreting a target utterance. We argue further that such attempts sys-
tematically exploit the inherent weaknesses or fl aws of the human cognitive system that are
amply discussed in cognitive psychology under the heading of “cognitive illusions”. We claim
that such a model correctly captures classical instances of manipulative discourse which fall
outside the scope of other accounts.
Keywords
argumentation , cognitive illusion , fallacy, manipulation , context selection process , Relevance
 eory
Introduction
Although the distinction between verbal manipulation and alternative non-
manipulative means by which speakers may get their addressees to engage in a
particular course of action or to believe the truth of certain propositions is
fairly intuitive, a consistent theoretical model of the former is not available
yet. Scholars from disciplines in the Humanities as varied as social psychology,
philosophy, anthropology, communication science, argumentation theory,
linguistics, psychology and critical discourse analysis have tried to tackle
the phenomenon or to explore some of its facets, but consensus, let alone a
unifi ed account, has not been reached yet.  e reason for this, in our opinion,
relates to the heterogeneous nature of manipulation; most of these approaches

D. Maillat and S. Oswald / International Review of Pragmatics 1 (2009) 348–370 349
indeed focus on some of its particular aspects, but do not – and sometimes
cannot – take a step further to consider the phenomenon comprehensively.
Answering the “what” question (“what is manipulative communication?”) is a
very diffi cult task, in so far as the identifi cation of manipulation as a commu-
nicative phenomenon proves to be problematic with regard to a goal of
descriptive adequacy. Contemporary approaches, in this line of work, have
tried to bring forth several criteria in order to assess what it is that makes a
discourse manipulative; yet the question is far from being settled, as we will
see further along.
Our proposal aims at reviewing these criteria and explores a relatively new
direction for research on manipulation, i.e. a cognitive one. In section 1 we
will discuss some of the criteria that have been proposed to characterise
manipulation.  ese are truth-conditionality violation, speaker interest, co -
vertness, social inequality and intention. We will show how none of these
may constitute a necessary condition for manipulation, as they might consti-
tute an obstacle to providing an operative criterion in order for a theore -
tical model to be adequate from a descriptive point of view. Some of these
criteria, we will also argue, fail to make the right predictions as they either
over-generate or under-generate. Consequently we will suggest that studying
this phenomenon may benefi t from a shift of perspective. Building on the
assumption that manipulation gains its eff ectiveness through the manipulated
individual’s processing of the manipulative input, section 2 will review litera-
ture on cognitive processing of linguistic information, in order to establish the
inherent risk associated to meaning derivation and to defend the idea that
manipulation may actually exploit the fact that “errors” are likely to occur in
speech processing. In this respect, we take on board Rigotti’s idea according to
which “the dynamics of manipulation are very close to the dynamics of human
error” (2005: 69).  is hypothesis will draw on contemporary research in
psychology and cognitive pragmatics (e.g. Pohl, 2004 ; Sperber & Wilson,
1995 ). Section 3 will in turn develop the idea and sketch out a pragmatic
model which accounts for the way manipulators may actually induce such
“undue” processing.  is alternative analytical approach will allow us to recon-
sider the nature of manipulative discourse and to provide a new defi nitional
framework to the processes involved in manipulation. In order to do so we
propose to integrate the complexity inherently associated with the heteroge-
neous nature of manipulation by focusing our attention on the processes
involved in the actual success of the manipulative attempt (i.e. the cognitive
mechanisms involved in speech processing). In other words, we will try to
understand the “whatness” of manipulation by answering the “how does it
work” question from a cognitive point of view. In doing so we will address

350 D. Maillat and S. Oswald / International Review of Pragmatics 1 (2009) 348–370
a point recently raised by Cummings (2004: 178ff ) who emphasises that while
there exists a clear potential for a pragmatic analysis of argumentation phe-
nomena and their governing principles, (neo-)Gricean pragmatic models have
not been applied to this domain, which renders “[p]ragmatic studies of argu-
mentative principles […] all the more urgent […]” (Cummings, 2004: 185).
1.  e Heterogeneous Nature of Manipulation: Available Clues
Contemporary research on manipulation has usually been concerned with the
nature of the manipulative message, its properties, and the social conditions in
which manipulation is likely to occur.  is section discusses the limits encoun-
tered by such accounts which try to isolate the necessary features that make
up manipulative discourse. Ultimately, we will propose that the identifi cation
of manipulation may not be as informative, from a scientifi c point of view, as
an account of the mechanisms by which manipulation may be successful.
1.1. Truth-Conditions and Felicity Conditions
Truth and felicity conditions have been proposed as relevant criteria defi ning
manipulative strategies in terms of falsity and insincerity (Rigotti, 2005 ); the
main assumption behind this view roughly amounts to considering that some
of the manipulative message’s features, in terms of the content it encodes, do
not comply with the communicative standards of verbal interaction.
1
 is is the
point made by Rigotti, who suggests that in manipulation, “what is negative has
to be somehow disguised as something positive”, and that manipulation “twists
the vision of the world […] in the mind of the addressee” (2005: 68). Typically,
lies would fall under this category, since it is crucial for a lie’s success to be taken
as a truthful statement. If a child lies to his parents by telling them that he did
not break a valuable vase,
2
it might be argued that he has tried to be manipula-
tive. Similarly, insincere promises, such as those politicians might utter in
1
In terms of speech act theory, this amounts to saying that a manipulative utterance generally
violates some felicity conditions of the speech act at stake, or, from a Gricean perspective, that it
covertly violates the Cooperative Principle and possibly the conversational maxims.
2
 e lie itself is merely about providing a false statement while asserting it as a truth; more-
over, if the child provides an alternative explanation, by for instance convincingly blaming the
dog instead and thus avoiding potential punishment, we could say the child’s contribution was
manipulative, in so far as it is further constraining the parents’ processing of information by
providing a plausible context to explain the situation.

D. Maillat and S. Oswald / International Review of Pragmatics 1 (2009) 348–370 351
3
In this example, the felicity (and by “unorthodox” extension, its truth) of the statement is a
function of the sincerity with which it is uttered, hence the truth-functional violation.
support of their future election, can also turn out to be manipulative, as they
may break the sincerity conditions which defi ne a promise.
3
However, it seems diffi cult to regard a false statement as a necessary or even
as a suffi cient condition for manipulation to take place, as Rigotti himself
(2005) acknowledges. First of all, there is strong evidence in favour of alterna-
tive manipulative mechanisms which do not involve lying. Such is the case, for
instance, when a speaker provides a plausible justifi cation for a proposition
which complies with an actual state of aff airs, while the real motivation for her
statement is not the one she makes manifest, but one which she wants to keep
concealed, while at the same time obtaining the desired eff ect on the addressee.
In this fi rst type, the manipulative strategy relies on shifting the addressee’s
attention as it were, stating A (which is true) so as to hide the truth of a propo-
sition B which would weaken or contradict the view entertained by the
speaker.
Along this line of thought, Parret ( 1978 ) argues that lying is limited to
cases where a speaker communicates a false proposition which misleads the
hearer into entertaining a new belief, while manipulation may function in
several diff erent ways: for example, the communication of a proposition may
mislead the addressee into giving up a legitimate belief, into reinforcing a
problematic belief or into preventing him from entertaining a legitimate
belief, all of which can be achieved without resorting to lying on the speaker’s
part, which would, therefore, be diffi cult to capture either by the restrictions
imposed by the very notion of lying or through an account based solely on the
notion of truth. What this shows, beyond Parret’s analysis, is that truth turns
out to be an insuffi cient criterion when it comes to characterising manipula-
tive communication.
Secondly, some researchers point out that the manipulative status of a lie
may even be questioned in a few cases: the case of parents telling their children
about Santa Claus raises this particular issue, as highlighted by Saussure &
Schulz when they argue that “it would be counter-intuitive that the parent
telling the child about Santa Claus is performing manipulation” (2005: 2). If
we are simply looking at the nature of the message parents make manifest to
their child by talking about Santa Claus, truth-conditions are indeed violated,
and, specifi cally, the existential presupposition. However, existential presup-
position is also violated when parents tell tales about unicorns, and more gen-
erally, when a speaker’s utterance is about fi ction. Although it could be said
that there is a diff erence between Santa Claus and unicorns in that children are

352 D. Maillat and S. Oswald / International Review of Pragmatics 1 (2009) 348–370
4
In fact, we believe that making the manipulative intention explicit would make the manipu-
lative attempt collapse altogether. We take this to be in the very nature of manipulation: a
manipulative intention is not meant to be communicated nor recognised (see sections 1.3 and
1.5 below).
aware that unicorns are to be found in fairy tales only, while they can reason-
ably be taken to believe in the actual existence of Santa Claus in the real world,
to the extent that it might even have a tangible eff ect on their behaviour (being
good all year in view of getting presents at Christmas), there is a priori no
reason to assume that truth-conditions are responsible for that diff erence. Be
that as it may, it could still be argued that the deliberate use of a false state-
ment is manipulative, as long as the addressee is not aware of it; that is, if we
consider that the hearer’s processing of information turns out to be biased by
virtue of the constraints the speaker imposes on the message. Of course, from
a somewhat more ethical and less technical point of view, we might express
reservations to call the Santa Claus lie manipulation, as Saussure & Schulz
(2005) do, simply because the example seems radically diff erent from one
where a cult guru manages to manipulate his followers and gets them to will-
ingly give him large amounts of money, or, more dramatically, to give up their
own lives, as was the case a few years ago with the sadly famous Order of the
Solar Temple collective suicide. But this is a separate question, and defi ning
manipulation in terms of moral or ethical acceptability would call for discus-
sions that exceed the scope of the present paper, although it does suggest
that a discussion about the manipulative status of lies bears technical as well
as ethical implications, since it involves far more than unsatisfi ed truth-
conditions.
From a speech-act-theoretical perspective, it could be envisaged to consider
that manipulation violates felicity conditions, and among them, the sincerity
condition in particular; we mentioned earlier in this respect the scenario of
a politician uttering a commissive without actually being committed to the
ulterior satisfaction of the propositional content conveyed. However, if one
were to defi ne manipulation within a speech-act-theoretic framework, a major
problem, discussed by Parret ( 1978 ), would emerge: it is the idea that the
speech act of manipulation cannot comply with the principle of expressibility
(cf. Searle, 1969) because it cannot be translated with an explicit performative
such as “I manipulate you + propositional content”.
4
 erefore, trying to
couch a defi nition of manipulation in terms of felicity conditions would not
take us very far; as Parret ( 1978 ) noted: the notion of manipulation as a speech
act is either self-destructive or it destroys the conceptual frame itself, were it to
be elaborated in terms of speech act theory.  at is to say that manipulation is

Frequently asked questions

Q1. What contributions have the authors mentioned in the paper "Defi ning manipulative discourse: th e pragmatics of cognitive illusions" ?

The authors start with a review of the existing defi nitions provided in these fi elds and highlight some of the diffi culties they encounter. The authors propose an alternative account of manipulation couched in the relevance-theoretic framework which treats manipulation as a two-step communicative attempt at misleading the contextselection process when interpreting a target utterance. The authors argue further that such attempts systematically exploit the inherent weaknesses or fl aws of the human cognitive system that are amply discussed in cognitive psychology under the heading of “ cognitive illusions ”. The authors claim that such a model correctly captures classical instances of manipulative discourse which fall outside the scope of other accounts. 

Q2. What are the future works in "Defi ning manipulative discourse: th e pragmatics of cognitive illusions" ?

To conclude, the authors hope to have shown that such an approach off ers a potentially greater descriptive and explanatory power than other traditional accounts in capturing manipulative strategies. 

Q3. What is the key element in the analysis of manipulative discourse?

Th e crucial element in this analysis of manipulative discourse lies in the manipulator’s attempt to prevent the addressee from accessing some other less accessible contextual assumptions. 

Q4. What is the principle of optimal relevance?

Th is process is governed by a principle of optimal relevance whereby the addressee tries to maximise the cognitive eff ects he can generate from the utterance while minimising the cognitive eff orts he has to put into the construction of the appropriate context of interpretation. 

Q5. What is the purpose of the assessment of relevance?

Th e assessment of relevance is not the goal of the comprehension process, but only a means to an end, the end being to maximise the relevance of any information being processed. 

Q6. What is the idea behind the manipulation view?

Underlying this view is the idea that manipulation exploits these forms of social asymmetry and the associated imbalance in the relations of power and domination in favour of the manipulator. 

Q7. What is the real motivation of the manipulative attempt?

Here the manipulative goal is foremost to improve the person’s health, not to trick him into getting stung in spite of his phobia, and so the authors would say that in this particular case, the real motivation of the manipulative attempt translates into a convergence of speaker and hearer interest. 

Q8. What is the definition of manipulative discourse?

In other words, manipulative discourse is a form of communication that puts the addressee in a situation where s/he will be led to shallow-process contextual assumptions. 

Q9. What is the main assumption behind this view?

Truth and felicity conditions have been proposed as relevant criteria defi ning manipulative strategies in terms of falsity and insincerity (Rigotti, 2005 ); the main assumption behind this view roughly amounts to considering that some of the manipulative message’s features, in terms of the content it encodes, do not comply with the communicative standards of verbal interaction. 

Q10. Who was hired to investigate the alleged link between Iraq and Niger?

a former US ambassador, was one of the persons hired to investigate a lead which linked Iraq with Niger in an attempt to buy uranium for nuclear weapons.