The multiplicity of emotions: A framework of emotional functions in decision making

Judgment and Decision Making, Vol. 3, No. 1, January 2008, pp. 5–17
The multiplicity of emotions: A framework of emotional functions
in decision making
Hans-Rüdiger P?ster ?
Department of Business Psychology, University of Lüneburg
Faculty of Psychology, University of Bergen
Gisela Böhm
Faculty of Psychology, University of Bergen
Abstract
A four-fold classi?cation of emotions with respect to their functions in decision making is proposed. It is argued
that emotions are not homogenous concerning their role in decision making, but that four distinct functions can be
distinguished concerning emotional phenomena. One function is to provide information about pleasure and pain for
preference construction, a second function is to enable rapid choices under time pressure, a third function is to focus
attention on relevant aspects of a decision problem, and a fourth function is to generate commitment concerning morally
and socially signi?cant decisions. The pertinent literature on the relationship between emotion and decision making is
reviewed, and it is concluded that most approaches ?t into the proposed framework. We argue that a precise conceptu-
alization of emotional phenomena is required to advance our understanding of the complex role of emotions in decision
making.
Keywords: emotion, affect, decision making.
1 Introduction
cision making. In particular, we argue that emotion(s)
should not be construed as a homogenous category, that
In this paper, we address the question of how to conceptu-
the positive-negative valence dimension is not the most
alize emotions concerning their role in decision making.
important aspect of emotions in decision making, and that
The study of emotions in the context of decision mak-
emotions do not imply irrationality.
ing, beginning more than twenty years ago (Baron, 1992;
We argue that it is more useful to think of emotional
Bell, 1982; Elster, 1985; Frank, 1988; Loomes & Sugden,
phenomena as implementing speci?c mechanisms to ac-
1982; P?ster & Böhm, 1992; Toda, 1980), has received
count for different functions that arise in decision mak-
increasing attention over the past decade (Loewenstein &
ing. Four functions are proposed which are conceptu-
Lerner, 2003; Mellers, 2000; Naqvi, Shiv, & Bechara,
ally independent, though empirically correlated. The
2006; Peters, 2006). There is, however, little consensus
four-fold classi?cation of functions which we propose in-
in the literature on what is actually meant by emotion or
cludes information, speed, relevance, and commitment as
affect. This paper tries to contribute to a more precise
the basic aspects.
and useful conceptualization of emotion concerning the
emotion-decision making relationship.
First, we brie?y sketch two approaches with a simi-
2 Related work
lar objective, the classi?cation of emotions by Loewen-
stein and Lerner (2003), and the functional typology pro-
2.1 The Loewenstein-Lerner classi?cation
posed by Peters (2006). We will then discuss some com-
Loewenstein and Lerner (2003) construe emotions ac-
mon assumptions about emotions which we believe have
cording to their place along the time course of a decision
confused the understanding of the role of emotion in de-
process, beginning with a deliberation phase leading to
a choice, then implementing the choice, and, eventually,
?Address correspondence to: Hans-Rüdiger P?ster, University of
Lüneburg, Wilschenbrucher Weg 84, 21335 Lüneburg, Germany, email:
experiencing the outcomes. They distinguish between an-
p?ster@uni-lueneburg.de
ticipated emotions and immediate emotions, with imme-
5

---

Judgment and Decision Making, Vol. 3, No. 1, January 2008
Multiplicity of emotions
6
diate emotions further classi?ed into incidental and antic-
ferences concerning the Peters taxonomy as we present
ipatory emotions. Anticipated emotions are beliefs about
our four-fold classi?cation of emotional mechanisms, and
one’s future emotional states that might ensue when the
take it up again in the discussion section.
outcomes are obtained. Immediate emotions, in contrast,
are actually experienced when making a decision, thereby
exerting an effect on the mental processes involved in
3 Contentious issues in emotion re-
making a choice; for similar distinctions see Kahneman
(2000). Immediate emotions come in two variants, either
search
as incidental emotions caused by factors which are not re-
lated to the decision problem at hand, and as anticipatory
Before presenting a framework of emotional phenom-
or integral emotions, which are caused by the decision
ena in decision making based on functional considera-
problem itself.
tions, we brie?y discuss a number of common assump-
There is ample evidence that these kinds of emotion
tions which, as we argue, have hindered a consensual
frequently do in?uence the judgments and choices people
conceptualization of the emotion-decision making rela-
make. Lerner and Keltner (2000) demonstrated the ef-
tionship. Though our framework does not logically de-
fects of incidental fear and anger on risk judgments. The
pend on these points, we consider it helpful to clarify
in?uence of immediate anticipatory emotions in intertem-
these contentious issues in advance.
poral choice has been examined by Loewenstein (1996).
The importance of anticipated emotions such as antici-
pated regret and disappointment in decision making has
3.1 The in?uence-on metaphor
been demonstrated by Zeelenberg, van Dijk, Manstead,
and van der Pligt (2000).
Both Peters (2006) and Loewenstein and Lerner (2003),
and, arguably, a vast majority of other researchers in
2.2 Peters’ functional roles of affect
the ?eld of decision making, adhere to what we call the
in?uence-on metaphor. Emotions - or affect, or feelings -
Peters (2006) recently proposed a classi?cation of the
are portrayed as external forces in?uencing an otherwise
roles that affect plays in decision making. Affect is
non-emotional process. It is assumed that the domain of
loosely de?ned as experienced feelings about a stimulus,
emotion is qualitatively different and functionally sepa-
either integral or incidental (Peters, Västfjäll, Gärling,
rate from the domain of cognition. Decision making is
& Slovic, 2006). Four roles are identi?ed: First, affect
then seen as an essentially cognitive process, which does
plays a role as information, especially via the affect-as-
not necessarily entail emotions. Emotions may have an
information mechanism (Schwarz & Clore, 1988). These
in?uence on decision making, but decision making per
feelings, possibly misattributed to the stimulus, act as
se might as well proceed without emotion. This is the
good-versus-bad information to guide choices, accord-
premise of traditional approaches of behavioral decision
ing to the affect heuristic proposed by Slovic, Finu-
making (Slovic, Lichtenstein, & Fischhoff, 1988), but is
cane, Peters, and MacGregor (2002). The second role
also re?ected in current dual-system theories (Kahneman,
played by affect is as a spotlight, focusing the decision
2003; Sloman, 1996; for a critical discussion see Price &
maker’s attention on certain kinds of new information
Norman, this issue).
and making certain kinds of knowledge more accessible
This antagonism of emotion and decision making is
for further information processing. This role is reminis-
commonly accompanied by further dichotomies: Irra-
cent of mood-congruent memory as studied by Bower
tional emotions disturb rational cognitions, intuitive feel-
(1991). Third, affect operates as a motivator, in?uenc-
ings outsmart deliberate thinking, and hot affect over-
ing approach-avoidance tendencies as well as efforts to
whelms cold logic.
process information (Frijda, 1986; Zeelenberg & Pieters,
We believe that the in?uence-on metaphor is mislead-
2006; Zeelenberg, Nelissen, Breugelmans, & Pieters, this
ing in several respects. In particular, we argue (1) that the
issue). Finally, a fourth role of affect is to serve as a
class of emotional phenomena does not form a homoge-
common currency in judgments and decisions (Cabanac,
neous category, but breaks down into qualitatively differ-
1992). Just as money does for goods, affect provides a
ent categories, (2) that many emotions are not unambigu-
common currency for experiences. Following Cabanac
ously mapped onto a simple positive-negative valence di-
(1992), Peters claims that affective reactions enable peo-
mension, and (3) that emotional mechanisms are ubiq-
ple to compare disparate events and complex arguments
uitous in decision making and do not constitute an ex-
on a common underlying dimension.
ternal irrational force which interrupts an allegedly non-
The Peters approach is similar in some respects to
emotional rational process. In the following, each of
our proposal. We will point out commonalities and dif-
these claims will be discussed brie?y.

---

Judgment and Decision Making, Vol. 3, No. 1, January 2008
Multiplicity of emotions
7
3.2 Is emotion a homogenous category?
may be called the coarser emotions, grief, fear, rage, love,
in which every one recognizes a strong organic reverber-
Most researchers would agree with the common intuition
ation, and afterwards speak of the subtler emotions, or
that the variety of particular emotions – such as anger,
of those whose organic reverberation is less obvious and
joy, or envy – are all instances of a general category
strong " (James, 1890/1952, p. 743).
called emotion (Charland, 2002). The hypothesis is that
all individual emotions share a few essential characteris-
tics, which permit to subsume particular emotions under
3.3 Positive and negative emotions
a single category of a more abstract type. The emotion
Another assumption shared by the vast majority of re-
category is perceived as referring to something real and
searchers is the idea that all emotions are naturally clas-
natural, and is not to be taken as an arbitrary conceptual
si?ed as either positive or negative. More precisely, it is
construction.
assumed that all emotional states can be mapped onto a
The view of emotion as a homogenous category has
one-dimensional scale of valence, characterized by con-
been discussed under the caption of emotion as a natu-
trasting labels such as positive versus negative, plea-
ral kind. A natural kind can be de?ned as exhibiting a
surable versus painful, or helpful versus harmful (Bar-
so-called homeostatic property cluster, that is, an interre-
rett, 2006b; Russell, 2003). This assumption of one-
lated set of properties causing a dynamic but stable con-
dimensional scalability corresponds to the economic no-
dition, which allows reliable inductions and generaliza-
tion of utility, which takes for granted that choice reveals
tions (Boyd, 1999; Grif?ths, 2004). The notion of an af-
an underlying one-dimensional utility scale. In a parallel
fect program as proposed by Panksepp (2000) or Ekman
manner, research on hedonic feelings and happiness pos-
(1999) is consistent with that view. Neuroscientists claim
tulates a general dimension of pleasant versus unpleasant
to have identi?ed particular areas of the brain which are in
feelings on which all experiences can be evaluated (Ca-
charge of emotional processes, whereas other parts are in
banac, 1992). Empirically, however, this view just does
charge of cognitive processes (Panksepp, 2000; LeDoux,
not hold, and ample evidence demonstrates that human
1996).
preferences do not conform to simple scalability (Licht-
However, the conception of emotion as a natural kind
enstein & Slovic, 2006; Tversky & Thaler, 1990).
has come under severe critique, mainly from a philosoph-
Such an unambiguous classi?cation of all emotional
ical perspective (Grif?ths, 1997, 2004). Following Grif-
states as positive or negative, though one of the most
?ths (1997), the main counter-arguments are: (i) There is
unanimous beliefs, may be impossible.
Following
no cluster of properties, which is common to all instances
Solomon and Stone (2002), we agree with the view
of the emotion category, and which allows for lawlike
that for many emotions a unique mapping as positive
generalizations; (ii) most generalizations are based on
or negative is impossible. There are multiple meanings
similarity by analogy, but not on homology, that is, not on
of this underlying dimension, and Solomon and Stone
a common evolutionary origin. Grif?ths argues that no
(2002) point out that good/bad, pleasurable/painful, use-
general regularities, physiological, neurological, or be-
ful/useless and similar contrasts have different origins
havioral, can be reliably identi?ed that are common to
and different meanings, and actually represent qualita-
and essential for all emotions.
tively different dimensions. What is good or bene?cial
Along similar lines, the status of particular emotions
need not be pleasurable, and what is harmful might nev-
such as anger and sadness as natural kinds has been ques-
ertheless be satisfying.
tioned by Barrett (2006a). Barrett (2006a) summarizes
Furthermore, many emotions constitute complex ap-
a vast array of evidence discon?rming the view that par-
praisals, consisting of mixtures of pleasurable and un-
ticular emotions exhibit unique response patterns (phys-
pleasurable aspects. Taking anger as an example, the
iological, facial, behavioral, etc.). Furthermore, Barrett
emotion’s object (another person) may be judged as hos-
(2006a) argues that empirical evidence supporting the
tile, the situational context as undesirable, but the arousal
claim of a unique causal mechanism, for example, partic-
(e.g., feeling strong) may be experienced as pleasurable,
ular neural circuitry in the brain for particular emotions,
and the consequences of expressing one’s anger (putting
is far from being consistent.
the other person in his place) may be quite enjoyable. The
We endorse the view that emotion is not a homoge-
pleasurable and the painful, the positive and the negative
nous category. Emotion is a word used in the vernacular
are not exclusive in emotional experiences; ambivalence
to refer to loosely related phenomena. The conceptual
might be the norm and not the exception.
confusion to be observed in the literature about de?ni-
Complex emotions are contextual states of mind, and
tions of emotion manifests that fact. Ever since William
depending on context might be positive or negative or
James, there have been doubts if all emotions are to be
both. Fear, commonly quali?ed as a negative emotion, is
treated in the same way: "I shall limit myself ... to what
experienced as positive and joyful in the movies when the

---

Judgment and Decision Making, Vol. 3, No. 1, January 2008
Multiplicity of emotions
8
context of a factual threat is stripped away. The notorious
anatomical grounds.
’Schadenfreude’ (gloating) represents a complex mixture
We agree with that view and believe that the issue of
of positive (subjective feeling) and negative (moral im-
rationality should be based on the validity of emotional
plications) facets, and might be painful or enjoyable de-
evaluations rather than on formal coherence. If our emo-
pending on the context and attentional focus.
tional appraisals are appropriate, that is, if we fear what
To be sure, we hold that some type of implicit or ex-
objectively is to be feared, and if we hopefully anticipate
plicit appraisal is at the core of emotional states, repre-
what will actually make us happy, then these emotions
senting an evaluation of the relationship between the self
might be called rational. Ample evidence demonstrates,
and the situation. For some emotions, this evaluation can
unfortunately, that people are not exceptionally good in
be reduced to a simple dimension of pleasure and pain,
making appropriate judgments about what makes them
as we will discuss shortly. For others, however, we en-
happy (Gilbert, 2006; Hsee & Hastie, 2006). Emotions,
dorse the argument of Solomon and Stone (2002), that
thus, may be appropriate, hence rational, or inappropri-
it is impossible to reduce the qualitatively complex and
ate, hence irrational.
multidimensional appraisals to a simple valence dimen-
We view emotions not as threats to rationality. The
sion.
rationality of decision making might actually depend
With respect to decision making, this implies that sub-
on people’s capacity to form appropriate emotions (de
stituting utility with valence does not solve the ques-
Sousa, 1987). Again, the in?uence-on metaphor is mis-
tion of the emotion-decision making relationship; there
leading here, assuming a detrimental in?uence of irra-
is more to the function of emotions than supplying a lo-
tional forces on an otherwise rational process.
cation on the valence dimension. More than two-hundred
The in?uence-on metaphor, we suppose, partly re-
years ago, already, Bernoulli (1954/1738) and Bentham
sults from these assumptions: Homogeneity of emotions,
(1948/1789) interpreted utility not as a formal measure,
unique valence, and opposition to rationality. If it is ac-
but as the subjective sensation of objective value, for ex-
cepted that emotion is not a homogenous category, and
ample, as the pleasure associated with receiving some
that valence cannot serve as a unifying aspect of emo-
amount of money. By maximizing utilities, decision mak-
tion, it follows that emotional phenomena should be clas-
ers might in fact maximize pleasure and minimize pain
si?ed in a more heterogenous way. Thus, we ask what
(Mellers, 2000). However, to assign more utility, that
functional requirements need to be solved in decision
is, more pleasure, to more money, might not be an emo-
making, and propose to classify emotional mechanisms
tional process at all, and to the extent that emotions are
along these functions. It will turn out that what is usually
involved, the one-dimensional mapping might soon break
conceptualized as simply ’emotional’ actually consists of
down.
four separate kinds of emotional mechanisms.
3.4 Emotion and rationality
4 A four-function framework of
A further common view is that decision making is a ra-
emotional mechanisms in decision
tional mental process without emotion, and that emotions
disrupt and jeopardize the rational process. In decision
making
research, rationality is mostly understood as formal con-
sistency, that is, conforming to the laws of probability and
The functional approach takes as its starting point po-
the axioms of utility theory. If people behave rationally
tential requirements of decision tasks, and potential at-
in that sense, they will make optimal choices. Emotions,
tempts to meet these requirements by instantiating partic-
then, can only interrupt and impede the process of achiev-
ular emotional mechanisms. At present, we identify four
ing an optimal decision.
functional requirements.
However, evidence is accumulating that this concep-
First, any decision requires some kind of information,
tion might be false (Bechara & Damasio, 2005; Bechara,
and with respect to individual decisions which are made
Damasio, Tranel, & Damasio, 1997). Without emotional
to promote the well-being of the decision maker, this in-
involvement, decision making might not even be possi-
formation needs to be personally relevant. Hence, the ?rst
ble or might be far from optimal (Damasio, 1994). Fur-
requirement and function is to provide information which
thermore, evidence from neuropsychological studies sug-
is useful for evaluation; consequently, we suggest that a
gests that at the level of brain structure and functioning,
particular class of emotions serves that purpose.
a clear-cut topological distinction between cognition and
Second, decisions in real life are an integral part of the
emotion might not be feasible (Phelps, 2006). Hence,
?ow of human activities, and thus subject to many situa-
the opposition of irrational emotion and rational cogni-
tional constraints. One of the most signi?cant constraints
tion turns out to be dubious on behavioral and neuro-
is time and time pressure. Making decisions requires to

---

Judgment and Decision Making, Vol. 3, No. 1, January 2008
Multiplicity of emotions
9
Table 1: Emotional functions in decision making.
Function
Emotion type
Prototypes
Mechanisms
information
reducible emotions
joy, (dis)liking
integration, trade-offs
speed
affect-programs, drives
fear, disgust, sexual lust
stimulus-speci?c response
relevance
complex discrete emotions
regret, disappointment, envy
selective attention, appraisal
commitment
moral sentiments
guilt, love, anger
social coordination, perseverance
choose and act within a temporal window of opportunity,
tion about evaluative judgments, for example, about one’s
ranging from seconds to years. We suggest that a second
life satisfaction. This is particularly prominent when the
function, hence a second kind of emotional mechanism,
mood state cannot be attributed to an unrelated causal
is concerned with speed, enabling the decision maker to
event.
make rapid decisions under time constraints.
In contrast to the affect-as-information framework,
Third, when making a decision, the decision maker se-
which focusses mainly on incidental affect, decision af-
lects a subset of particular aspects of the situation un-
fect theory as advanced by Mellers (2000) is based on
der consideration, which consists, in principle, of an
integral emotion, that is, feelings of pleasure or displea-
uncountable number of aspects. This selection is con-
sure that originate directly from the choice consequences
trolled by relevance, and we propose that a third func-
under consideration. Decision affect theory assumes that
tion, hence a third kind of mechanisms, is to direct the
decision makers compute a weighted sum of anticipated
decision maker’s attention to relevant aspects of the sit-
pleasures which they believe to obtain from outcomes of
uation. What is relevant depends on how the situation is
risky choices, and then choose the option they believe to
appraised.
yield the greatest amount of potential pleasure. Using
Finally, a fourth requirement of decision making is to
pleasure as a substitute for utility, anticipated pleasure in-
adhere to decisions once made, that is, to implement de-
forms about the utility of a consequence, and ultimately
cisions in the long run. Most people show commitment
about the expected utility of a choice option. This inter-
in particularly complex decision situations, which, as we
pretation of utility as pleasure is in line with the concep-
will demonstrate below, are mainly concerned with social
tion of Bernoulli (1954/1738), which has been revived
and moral decisions. Hence, the fourth kind of mecha-
in a number of modi?cations of the Subjective-Expected-
nism will generate commitment in social decision mak-
Utility model (Elster & Loewenstein, 1992; Kahneman &
ing.
Tversky, 1979).
In sum, information, speed, relevance, and commit-
The affect-heuristic, proposed by Slovic et al. (2002),
ment are four requirements when making decisions, and
is a related approach, though it is somewhat equivocal
distinct types of emotional mechanisms serve to meet
with respect to the issue of incidental or integral affect.
these requirements. In the following sections, we will
The affect-heuristic - a quick and simpli?ed process of
elaborate on each function (Table 1).
evaluating a risky option by relying on one’s immediate
feelings of liking or disliking - refers to affect elicited by
4.1 The information function: Pleasures
the options under consideration, but the affective reaction
and pains
might as well be caused by undetected intrusions from
unrelated events or memories.
The information function of emotion has been acknowl-
In any case, emotion or affect as understood by these
edged by many researchers, albeit under different frame-
theories is essential in providing the decision maker with
works and with varying emphasis (Clore, Gasper, &
evaluative information about the target, be it attributed
Garvin, 2001; Mellers, 2000; Peters, 2006; Schwarz
correctly or incorrectly to the target itself. The most im-
& Clore, 1988; Slovic, Finucane, Peters, & McGregor,
portant feature of that process is its one-dimensionality:
2002). The information involved is information which is
Whatever the origin and whatever its qualitative speci?cs,
useful for evaluation and preference construction, that is,
the multitude of affective states is projected onto one sin-
for making a decision.
gle dimension of pleasure and pain. In theoretical terms,
For example, in Schwarz and Clore’s (1988) affect-
this is the valence dimension, commonly assumed to be
as-information framework, affective states such as pos-
bipolar, and assumed to constitute a core characteristic of
itive or negative mood are assumed to provide informa-
an emotional experience (Barrett, 2006b).

---

Judgment and Decision Making, Vol. 3, No. 1, January 2008
Multiplicity of emotions
10
To the extent to which decision options can be mapped
not be extended to organizational decision making, since
onto the pleasure dimension, this provides a simple mech-
an aggregated unit such as a ?rm surely does not maxi-
anism of making a choice by integration and maximiza-
mize pleasure, it rather maximizes pro?t.
tion. Most models imply an implicit weighing process,
As discussed above, many emotions cannot be charac-
for example, weighing the pleasures of future outcomes
terized unequivocally as positive or negative (Solomon &
by degree of delay and by probability of occurrence.
Stone, 2002). Hence, information about valence is a spe-
In economic regret and disappointment models, ex-
ci?c property restricted to only a few emotional states.
pected utility is modi?ed by integrating deviations from
Peters’ (2006) notion of information is related to our con-
non-obtained but hoped-for outcomes into the overall
ceptualization, but extends the information function to all
utility equation. For example, the basic utility of an out-
emotions, and also includes incidental sources of affect.
come might be enriched by a function of the difference
between the obtained outcome and the expected outcome,
4.2 The speed function: Affect programs
or by a function of the difference between the obtained
outcome and a missed outcome (Bell, 1982; Loomes &
and somatic markers
Sugden, 1982). Many similar models have been proposed
Bechara, Damasio, Tranel and Damasio (1997) demon-
that try to overcome the limits of traditional expected util-
strated, using the so-called Iowa Gambling Task, that
ity models (Wu, Zhang, & Gonzales, 2004). However,
anticipatory affective reactions, measured as changes in
empirical evidence supporting these approaches is mixed,
skin conductance responses, towards risky and disad-
at best.
vantageous stimuli precede in time conscious knowledge
We suppose that the limits of pleasure-as-utility mod-
about the disadvantageous nature of the stimulus. The
els are demarcated by the one-dimensionality of the as-
somatic-marker hypothesis (Bechara & Damasio, 2005;
sumed emotional experience. To the extent that emo-
Damasio, 1994) maintains that these kinds of affective
tional experiences can be mapped unambiguously onto
signals, originating in bodily states and acquired by learn-
the pleasure-displeasure scale without loss of meaning,
ing from previous experiences, act as markers about the
they can serve as informative signals for the decision
positivity or negativity of current experiences. Somatic
maker: We call these emotions reducible emotions. Re-
markers operate automatically and obligatorily, in?uenc-
ducible emotions are emotions which are essentially car-
ing behavior even before a deliberate intention is gener-
riers of valence without representing a complex appraisal
ated.
of the situation, and are hence easily reducible onto a
Evidently, there is a class of bodily states that has
scale of pleasure and pain. Examples are joy (or distress),
the potential to guide behavior without cognitive control.
liking (or disliking), and attraction (or repulsion). It is
This ?nding is not totally new. Being in a state of in-
only a subset of emotional states that are characterized
tense hunger or strong sexual arousal drives behavior in a
by a unique valence, and that do not imply more complex
certain direction. Loewenstein (1996; Loewenstein, We-
appraisals (Ortony, Clore, & Collins, 1988; Solomon &
ber, Hsee, & Welch, 2001) has emphasized the role of
Stone, 2002; Zajonc, 2000).
visceral states in determining human choice and behav-
When making a choice, reducible emotions are men-
ior. The force of the body is, not surprisingly, especially
tally represented as beliefs about the decision’s conse-
irresistible when basic drives or addictive desires are in-
quences, not as momentary feelings. When pondering the
volved: An overwhelming craving for a drug or even for
question whether you should rather choose a risky gam-
a chocolate bar resolves the choice for us, leaving only
ble with the risk of loosing amount X, otherwise winning
a minor role to our cognitive re?ections concerning that
Y, or a safe option gaining Z (with X < Z < Y), you are
choice (Baumeister & Vohs, 2003). The ?nding of an hy-
consulting your beliefs about the expected pleasures of
perbolic discount function in intertemporal choice might
winning Y, and the expected pains of loosing X, respec-
be partially accounted for by the increasing intensity of
tively. From memory or from imagination you infer how
bodily states of deprivation when the object of one’s crav-
you might feel, but you do not necessarily feel it at that
ings nears in temporal or physical proximity (Frederick,
moment. More generally, these beliefs provide orienta-
Loewenstein, & O’Donoghue, 2003).
tion concerning a preferential problem. Reducible emo-
Somatic markers, visceral states, and (quasi-)addictive
tions with respect to preferences can be considered as a
cravings all show a common characteristic: They speed
special case of the general class of orientation feelings
up behavior, in contrast to the slow machinery of deliber-
(Norman, Price, & Duff, 2006; Price & Norman, this is-
ate choice. This can be viewed as adaptive from an evo-
sue).
lutionary perspective. Observations from patients with a
To orient one’s choices towards the amount of plea-
damage to the ventromedial prefrontal cortex, considered
sure provided applies to personal decision making, that is,
to be the brain structure which triggers relevant somatic
when individual preferences are concerned. This should
markers, suggest that these patients not only tend to make

---

Judgment and Decision Making, Vol. 3, No. 1, January 2008
Multiplicity of emotions
11
disadvantageous choices, but that it takes them a tremen-
stein and Lerner’s (2003) approach. In the taxonomy of
dous amount of time to reach a decision at all (Bechara
Peters (2006), it is implicit in the motivation function.
& Damasio, 2005). These patients endlessly ponder the
However, the motivating function of emotions can take
pros and cons of options, continually procrastinating a ?-
different forms. The direct link between affect and a par-
nal resolution. They appear to be unable to make up their
ticular behavior in affect programs supports rapid imple-
minds to execute the vital step from choosing to acting,
mentation of actions. This is, however, not the case for
exhibiting a lack of ability to cross the Rubicon gap from
broad action tendencies (Frijda, 1986), referred to by Pe-
thinking to doing (Heckhausen, 1991). Generally speak-
ters (2006). Action tendencies may or may not be en-
ing, all decisions are constrained by temporal limits, and
acted, and might even slow down behavior when contra-
any choice is, at some point in time, preceded by a choice
dicting tendencies paralyze each other.
to choose at all.
The function to speed up choices is especially obvious
4.3 The relevance function: Discrete emo-
for a few affective states for which some authors have
tions
claimed that so-called affect programs (Panksepp, 2000;
Tomkins, 1984) can be identi?ed as specialized brain cir-
Many emotional states are comprehensible only with ref-
cuitries. Based on a meta-analysis of functional neu-
erence to their cognitive content, examples are regret,
roimaging studies of emotions, Murphy, Nimmo-Smith
envy, gloating, pride, guilt, shame, contempt, and many
and Lawrence (2003) conclude that partially separate
others. Note that this does not apply to simple pleasure or
neural systems can be identi?ed for fear, disgust, and
liking: An utterance such as "I like my spouse" is a legit-
anger. These emotions also show a typical structural dis-
imate statement without providing any knowledge about
tinctiveness, with fear associated with the amygdala, dis-
the spouse. In contrast, an assertion such as "I regret that
gust with the insula-operculum and the globus pallidus,
I chose to marry X" needs a meaningful justi?cation to
and anger with the lateral orbitofrontal cortex. On the
be acceptable in normal conversation. According to ap-
other hand, no particular brain areas can be cohesively
praisal theories, these emotions are the consequence of
ascribed to emotional activity in general, or to positive
the particular way people construe the situation or the
versus negative emotions. Though at present neuroimag-
event (Ortony et al., 1988; Smith & Lazarus, 1993). If
ing ?ndings should be interpreted with caution, this sug-
we are aware of the particular appraisals concerning di-
gests that fear, disgust, and anger might constitute a spe-
mensions such as valence, certainty, control, and respon-
cial kind of process, which is hard-wired in the brain, and
sibility, we can infer what emotion a person is experienc-
which enables humans to act quickly without delibera-
ing. And, vice versa, knowing the particular emotion a
tion.
person has with respect to an event provides us with a
This makes sense also from an evolutionary perspec-
justi?ed conjecture about how this person construes that
tive, assuming that a ?nite set of affect programs evolved,
event (Siemer & Reisenzein, 2007).
which are tied to stimuli that are exceptionally threaten-
Emotional construals are idiosyncratic, representing
ing, and associated with immediate withdrawal behavior.
the event from a subjective point of view. Yet, there is one
Interestingly, Murphy et al.’s (2003) meta-analysis pro-
invariant feature: All emotional construals focus on the
vides some evidence for a discrete neural system for with-
fundamental relationship between the self and the event,
drawal, but not for approach, restricting the set of quick
that is, on the relevance a particular event has on a per-
affective responses to behaviors that move the person out
son’s vital concerns and interests. We suggest that this
of a dangerous situation. Approach behavior might then
is the characteristic function: To focus people’s attention
be controlled by more deliberate functions, increasing the
on the particular aspects of an event that are appraised as
?exibility of approach behavior and enabling the explo-
relevant. Once this focus is established, further emotions,
ration of genuinely new situations.
or motivations, or actions concerning that relevant aspect
Not all fast decisions need to be mediated by affective
may ensue.
responses. An experienced chess player is able to make
Take as an example the event of a colleague who
rapid moves by simple pattern recognition. But these de-
proudly brags about his paper which has just been ac-
cisions are not vital for the chess player, whereas affect
cepted by the most prestigious journal. As a result of
programs address concerns of vital importance for the or-
appraising that event you might feel envy, somebody else
ganism, which is presumably why they have evolved in
might, however, feel admiration. In either case, the very
the ?rst place. A mere cognitive response is easily con-
same event tells a quite different story about the relevant
trolled by deliberation, but an affect program claims con-
concerns of the two characters experiencing envy or ad-
trol precedence, and enormous effort is needed to sup-
miration. This, in turn, entails different behaviors if a
press the response, if this is possible at all.
choice is involved (Böhm & P?ster, 2000, 2005; Zeelen-
The speed function plays no prominent role in Loewen-
berg et al., this issue).

---

Judgment and Decision Making, Vol. 3, No. 1, January 2008
Multiplicity of emotions
12
In decision research, the two most extensively studied
outcomes obtained and not obtained, irrespective of them
emotions that serve the relevance function are regret and
being positive or negative.
disappointment (Connolly & Zeelenberg, 2002; Zeelen-
berg, van Dijk, Manstead, & van der Pligt, 2000). Regret
and disappointment both result from counterfactual com-
4.4 The commitment function: Moral sen-
parisons of what one has obtained with what one could
timents
have obtained, signalling that we have made a bad de-
cision. Regret and disappointment draw people’s atten-
We all have strong intuitions about what is morally right
tion to different potential causes of the bad outcomes, and
and what is wrong. The morally right choice, however,
trigger different behavioral tendencies (Zeelenberg et al.,
is frequently opposed to what is in our best self-interest,
2000). With respect to regret, one’s own previous deci-
narrowly construed as the immediate maximization of
sion is highlighted as the relevant cause, and tendencies to
material wealth. We might make a steeper career if we
undo one’s decision and attempts to get a second chance
were ruthlessly competitive instead of collaborating with
will result. In contrast, with respect to disappointment,
others, we might live a less stressful life if we lied and
tendencies to get away from the situation and attempts to
cheated from time to time, or we might end up with more
blame one’s bad fortune will result.
money in our pockets if we invested in military stocks
Similarly, Böhm and P?ster (2000, 2005) have shown
instead of contributing to medecins sans frontieres.
that in the domain of environmental risks people tend to
Why, then, do most people act morally most of the
appraise a risk either with respect to its consequences,
time? Frank (1988, 2004) argues that some emotional
or with respect to its moral implications, depending on
states mainly operate as commitment devices, leverag-
how the risk is mentally represented. A consequentialist
ing moral choices and preventing people from pursuing
appraisal causes emotions such as fear or worry, which
hardnosed self-interest. From the point of view of self-
in turn trigger helpful behavior. In contrast, a moral ap-
interest, emotions such as guilt or shame are senseless.
praisal causes emotions such as anger or indignation, trig-
From a strict consequentialist and utilitarian perspective,
gering aggressive behaviors.
they might even prevent optimal decisions (Baron, 1994).
There are, of course, other processes that direct peo-
Ever since Adam Smith (1759), however, it is conceded
ple’s attention on speci?c features of the world, such as
that people are in?uenced by motives that go beyond, and
the startle response, or just the novelty of particular fea-
contradict, pure self-interest, understood as a narrow fo-
tures. These mechanisms, however, yield no meaningful
cus on maximizing one’s own material outcomes.
interpretation of the situation, they are, in a sense, neutral.
In the familiar prisoners’ dilemma, pure self-interest
To perceive an object as novel is neutral, but to perceive
dictates to choose the defective option, but with the con-
an action as regrettable is personally relevant.
sequence that both players are worse off than if they had
What people consider relevant, guided by their emo-
chosen the cooperative option. In fact, people frequently
tions, depends on their very personal histories and mo-
choose the cooperative option, even when playing the
tives. More important, emotional relevance does not nec-
game only once with a stranger, let alone when playing an
essarily imply good and wise judgment. To regret and
iterated version with repeated interactions with the same
grieve over consequences which cannot be undone and
player. People also contribute to public goods, and punish
could not have possibly been foreseen, is clearly irra-
others who behave uncooperatively, even if that entails a
tional. To ask for advice from other people might be
cost (Gintis, Bowles, Boyd, & Fehr, 2005).
a good strategy in these situations. For non-involved
The problem with that kind of moral or altruistic deci-
strangers, the decision problem under consideration is not
sion is not to ?nd out what the right choice is, but how
personally relevant, and they might grasp the problem in a
to enact it and stick to it when confronted with opposing
more comprehensive, more balanced way, and might pos-
motives to pursue one’s pure self-interest. To summarize
sibly provide advice that improves our decisions (Yaniv,
Frank’s basic argument by example (1988, 2004, 2006):
2004).
Suppose you are the owner of a business and have the op-
The spotlight function of Peters (2006) closely paral-
portunity to open a satellite of?ce elsewhere. Since you
lels the relevance function. However, whereas Peters is
cannot directly survey the manager of the new of?ce far
more concerned with the valence of affect, which directs
away, he might cheat you. If he cheats, he obtains, say, $
attention on either positive or negative aspects, we em-
1.500, and you loose $ 500. If he is honest, both of you
phasize the particular semantics of emotions. For exam-
obtain $ 1.000. Following pure self-interest, the manager
ple, regret, interpreted as a negative emotion, not simply
will cheat. Knowing this, you will not open the new of-
focuses attention on unpleasant aspects, it actually con-
?ce, so that each of you obtains nothing. Now you and
strues the situation as a regrettable one. This implies
the manager are worse off, relative to what you would
a focus on non-chosen alternatives, and comparisons of
have obtained with the new of?ce opened and an honest

---

Judgment and Decision Making, Vol. 3, No. 1, January 2008
Multiplicity of emotions
13
manager running it. However, if the manager feels guilty
with strategic interaction and coordination, has, like be-
when cheating, this feeling might incur a cost on cheat-
havioral decision research, increasingly focused on the
ing. Suppose the cost is as large as $ 10.000, then cheat-
study of emotional factors (Geanakoplos, Pearce, & Stac-
ing plus feeling guilt would amount to a loss of - $ 8.500.
chetti, 1989; Rabin, 1993). In line with the assumptions
Thus, the moral emotion of guilt causes the manager to
of Frank (1988), Fehr and Gächter (2002) demonstrate
be honest, for your bene?t as well as his.
that people cooperate with strangers in one-shot inter-
Note that the generation of guilt, in that framework, is
actions even when it is costly for them, showing strong
automatic and involuntary, though the decision which fol-
reciprocity (Gintis et al., 2005). Using a public good
lows is a deliberate act. The story could go on: Suppose
paradigm, they showed that a majority of participants was
the manager, in a moment of moral blindness, cheats, and
willing to punish violators who did not contribute and
feels guilt as a consequence. To remove that niggling
tried to get a free ride. Punishing, however, was costly for
feeling, he silently returns the money, undoing his mis-
those who punished without delivering a balancing bene-
demeanour, again to his and your advantage.
?t. This constitutes a social dilemma in its own right: If
Guilt, on that account, serves as the proximal cause of
everybody punishes the violators, then in the long run all
the manager’s commitment to be honest. Being honest
are better off, assuming that violators learn from punish-
is not an abstract moral act, it is advantageous to both
ment; but the one who does not punish, while others do,
agents, and without that mechanism, everybody would be
will be even better off. Fehr and Gächter’s (2002) ?nd-
worse off. There is an important corollary to this account:
ings suggest that this cooperative behavior is emotionally
The owner should somehow recognize managers who do
mediated by anger. Anger triggers behavior which from a
not cheat. Obviously, to trust a manager who simply tells
pure self-interest perspective is costly for the individual.
you that he is honest is not an advisable strategy. Moral
Note that other kinds of anger, such as road rage (Joint,
sentiments, following Frank (1988), also provide a solu-
1996), do not implement the commitment function. The
tion for the skeptical owner, since they generate observ-
word anger, thus, denotes quite different emotional phe-
able signals such as blushing which indicate the likeli-
nomena: In the public good situation, anger is close to
hood that a person is honest.
what might be called moral indignation, whereas in a traf-
This suggests that evolution has equipped humans with
?c jam situation, anger is close to rage and fury.
the capacity to produce reliable signals informing others
The role of moral sentiments and the commitment
that they are honest and trustworthy, which in turn leads
function is not addressed by Peters’ (2006) taxonomy, nor
to advantageous cooperation. These signals must be au-
by Loewenstein and Lerner (2003). The particular role of
tomatic and uncontrollable to be credible. On the other
commitment and moral emotions will be taken up in the
hand, they are not totally reliable, some cheaters survive
discussion.
and bene?t, leading to an evolutionary stable equilibrium
with a majority of trustworthy people and a minority of
cheaters in coexistence (Frank, 1988).
5 Discussion
As these examples demonstrate, moral sentiments
are closely related to social relationships.
Evidence
from neurobiology suggests that moral judgments involve
We propose a new look on the relationship between
brain areas which are associated with cognitive as well as
emotion and decision making. Emotions do not merely
with social-emotional processing, forming a specialized
in?uence an otherwise non-emotional process, as the
neural circuitry activated when making moral judgments
in?uence-on metaphor holds, but are part of virtually any
(Greene, Nystrom, Engell, Darley, & Cohen, 2004). Just
decision making process. Because, as we argue, emo-
as guilt and shame provide signals about trustworthiness,
tion itself is not a homogeneous category, the emotional
so do emotions such as sympathy and love signal so-
functions within decision making are multifaceted. Fol-
cial support and reciprocity. They indicate with whom
lowing functional considerations, we propose a four-fold
it is bene?cial and wise to form a bond, just as guilt
classi?cation on how emotional mechanisms shape deci-
and shame they serve also as commitment devices. The
sion making.
emotion of love, for example, helps you to be faithful to
The information function provides evaluative informa-
your partner, even when at times another person might be
tion which feeds into preference construction. Emo-
more alluring. In sum, moral sentiments such as guilt,
tional states such as joy or distress inform about the de-
shame, love, hate, or sympathy guide decisions in two
gree of (un)pleasantness of actions and consequences.
ways: They support people to stick to long-term commit-
They allow to map a diversity of experiences on a one-
ments, and they support decisions about whom to select
dimensional scale of pleasure and pain.
for cooperation.
The speed function enables rapid choice and action un-
The ?eld of game theory which is genuinely concerned
der time pressure. Affect programs for fear and disgust

---

Judgment and Decision Making, Vol. 3, No. 1, January 2008
Multiplicity of emotions
14
trigger immediate avoidance responses. These mecha-
1999), is somewhat related to the relevance function. The
nisms are highly stimulus-speci?c and presumably have
spotlight function posits that affect directs attention to
evolved under evolutionary selection pressure.
particular kinds of information, just as selective atten-
The relevance function focuses attention on particular
tion directs visual search. The speci?c valence, positive
aspects that are of potential relevance for the decision
or negative, determines the kind of information selected,
maker. A discrete emotion such as regret or envy consti-
which is a well-known effect (Bower, 1991). In contrast,
tutes a particular appraisal, which implies particular eval-
the relevance function is about semantics, that is, how a
uations as well as particular action tendencies.
situation is construed. To feel regret entails that attention
The commitment function enables social coordination
is focused on alternative options one could have but has
by committing people to stick to decisions, even against
not chosen. Relevance is not restricted to valence, quite
their short-term self-interest. Guilt, for example, prevents
to the contrary, relevant emotions such as regret or pride
defection in social dilemmas, and thus guides decision
focus attention on a meaningful con?guration of options,
making in strategic choice situations.
outcomes, and their relationships.
We claim that there is no uniform in?uence of emo-
To view affect as a motivator (Peters, 2006) for action
tion or affect per se on decision making. In particular, the
has been a recurrent issue in emotion research (Frijda,
view of emotions as providing valenced evaluations, and
1986), and has recently been emphasized by Zeelenberg
substituting utility with valence, is just one of four qual-
and Pieters (2006; Zeelenberg et al., this issue). We fully
itatively different mechanisms. Contextual requirements
endorse this assumption, since it is of paramount im-
primarily determine which function dominates: Lack of
portance in decision making. Beyond simple approach-
information, time pressure, relevance ambiguity, or need
avoidance behavior, action tendencies associated with
of social coordination. Of course, no real decision situa-
discrete emotions trigger preferences for particular activ-
tion is an uncontaminated instance of just one functional
ities (Zeelenberg & Pieters, 2006). In our framework, the
requirement. We might, simultaneously, be pressed by
motivational function is implied by the speed function,
some affect program to escape, and at the same time be
by the relevance function, as well as by the commitment
committed to continue a morally endorsed course of ac-
function, differing in the granularity of the action urged
tion.
to implement.
The issue of rationality in decision making, thus, turns
Finally, the assumption of affect serving as a common
out to be one of appropriateness of emotions, not of for-
currency as has been advocated by Cabanac (1992), par-
mal consistency of preferences. Whenever several emo-
allels the information function and what we call reducible
tional functions generate antagonistic preferences, an in-
emotions. Cabanac (1992) showed that pleasure indeed
tense state of ambivalence occurs. Ambivalence might
serves as a common currency for different sensory ex-
be considered as an important cause that makes decisions
periences. We doubt, however, that the totality of expe-
dif?cult (Greenspan, 1980).
riences, including complex cognitions, can be collapsed
The approach most akin to our proposal is the func-
onto one hedonic scale of simple affect (Peters, 2006).
tional taxonomy of Peters (2006; Peters et al., 2006), and
As Solomon and Stone (2002) argued, many emotions
we will brie?y discuss central similarities and emphasize
cannot be mapped uniquely onto the positive-negative di-
differences. Though it seems that Peters (2006) is more
mension. Ambivalence, decision aversion and trade-off
concerned with affect in contrast to discrete emotional
dif?culties (Beattie & Barlas, 2001; Hanselmann & Tan-
states, the terminology is somewhat unclear and we will
ner, this issue) indicate the ubiquity of different curren-
neglect de?nitional issues for the moment.
cies with frequently unknown exchange rates.
The information function proposed by Peters (2006)
In sum, we see several similarities with Peters’ (2006)
is closely related to the affect-as-information theory
approach, and take this convergence as evidence of the
(Schwarz & Clore, 1988). People use immediate and
need to develop theoretically clear and comprehensive ac-
holistic affective responses as a substitute of deliber-
counts of the emotion-decision making relationship. We
ate reasoning (Kahneman, 2003); the affective substitute
also think that our framework extends the common ideas
might be completely incidental. In contrast, the infor-
in important ways. Peters’ information, spotlight, and
mation function in our framework de?nes evaluative in-
common currency functions all hinge on the existence
formation as the essence of a preferential judgment. Re-
of a one-dimensional valence scale. Affective substitu-
ducible emotions actually generate the information nec-
tion, affective selective attention, and the common cur-
essary to compare different options; what they accom-
rency proposal are all variants of one basic mechanism
plish is the very process of projecting diverse experiences
of valence ascription. This is captured in our notion of
on a common pleasure-pain dimension.
reducible emotions. The relevance and the commitment
Peters’ (2006) spotlight metaphor, taken from theo-
function, however, go beyond mere valence. Relevance
ries on visual perception (Fernandez-Duque & Johnson,
entails preferences for highly particular actions, and com-

---