Thesis
Reference
The flexibility of olfactory preferences : impact of decision-making processes
COPPIN, Géraldine
Abstract
Pleasantness evaluations and preferences related to odors have traditionally been thought of as rather inflexible. An increasing number of empirical studies reveal, however, some evidence of flexibility. This thesis sought to examine how and to what extent decision-making processes can shape olfactory preferences, as well as the level of processing required in such a choice-induced preference modulation (explicit vs. implicit) and how long the modulation persists. In the empirical part of this thesis, we present empirical evidence demonstrating that (i) olfactory preferences can indeed be shaped by decision-making processes, (ii) this shaping does not require explicit processes, and (iii) the change can be long-lasting. These results invite further consideration of the modulatory impact of decision-making processes on olfactory preferences and of the importance of implicit processing in olfactory preference modulation and stability across time. The theoretical implications of these results, in particular regarding cognitive dissonance theory, are discussed.
COPPIN, Géraldine. The flexibility of olfactory preferences : impact of decision-making processes. Thèse de doctorat : Univ. Genève, 2012, no. FPSE 504
URN : urn:nbn:ch:unige-218713
DOI : 10.13097/archive-ouverte/unige:21871
Available at:
http://archive-ouverte.unige.ch/unige:21871
Disclaimer: layout of this document may differ from the published version.
The flexibility of olfactory preferences:
impact of decision-making processes
Thèse
présentée à la
Faculté de Psychologie et des Sciences de l’Education de l’Université de Genève
pour obtenir le grade de
Docteur en Psychologie
par
Géraldine COPPIN
de Lille, France
Numéro d’étudiant : 06-311-500 Numéro de thèse : 504
sous la direction du
Professeur David SANDER
Genève, Juin 2012
ACKNOWLEDGEMENTS
The completion of this thesis would not have been possible without the advice, support and friendship of a great number of people. I can never fully express how grateful and lucky I have been to complete this thesis in such a professional, dynamic and caring environment.
La première personne que je voudrais remercier pour m’avoir permis de réaliser ce travail est mon directeur de thèse, le Professeur David Sander. Depuis ma première visite à Genève au printemps 2006, j’ai été impressionnée par ta connaissance de la littérature et ton aisance à la communiquer simplement. Les innombrables discussions que nous avons eu au cours de ces six années m’ont énormément apportées. J’espère que j’honorerai tes attentes de première thésarde officielle sous ta supervision en tant que Professeur.
Je voudrais remercier du fond du cœur le docteur Sylvain Delplanque, pour avoir toujours compris ce que je dis, même quand je m’exprime en ch’ti, et surtout pour avoir toujours une ou des explications qui répondent exactement à mes questions, même les plus farfelues.
Aucun test statistique n’est nécessaire pour savoir que vous êtes mes deux mentors. Je vous en suis, sincèrement, reconnaissante.
Je voudrais aussi remercier ainsi le Dr. Moustafa Bensafi, le Dr. Christian Margot et le Prof. Klaus Scherer pour avoir accepté d’être membres de mon jury de thèse. Moustafa, je voudrais sincérement te remercier pour m’avoir permis de faire un stage dans ton laboratoire et appris les secrets du sniff. Christian merci pour m’avoir fait partager une petite part de ta vaste connaissance et de ta vision du monde de l’olfaction, ainsi que ton courage face à la fabrication de l’androsténone. Merci beaucoup Klaus pour les discussions extrêmement enrichissantes que nous avons pu avoir sur les théories des émotions, et plus généralement, de l’affect.
Thank you to my friends and colleagues at the CISA and FPSE (y compris les personnes qui n’y font pas de recherche) : Andres, Andy, Ben B., Blandine, Camille, Carole, Céline, Dajana, Didier, Guido, Katia, Katja, Kerstin, Kim, Konni, Isabelle, Léo, Matthieu, Michael, Olivier, Sandra, Sascha, Sebastian, Sophie, Stéphane, Vera... Un grand merci à Alison et Christian pour avoir fait du bureau 5137 un endroit si chaleureux (et presque tropical avec le
papyrus) durant mes premières années de thèse (mais aussi après mon déménagement en 5139). Merci à Laure pour avoir si souvent pointé le bout de son nez dans mon bureau (quelqu’il soit). Thank you Tobias for being an awesome office-mate who taught and reminded me of very relevant things, like Käse in Kühlchrank. Thank you Ben M. for your help with missing data (I wish I will be able to use it) and your singular sense of humor. Even though this work is not included in this document, thank you Corrado for teaching Aline and I the secrets of MRI analysis in such a fruitful and patient way and tolerating our, sometimes, foolish behavior. Merci Pichon pour avoir mêlé l’acquisition et l’analyse des données IRM à d’innombrables fous rires et à une excellente vinaigrette. Merci Danny pour ton soutien et ta franchise – si tu n’existais pas, il faudrait se pencher sur la question de comment t’inventer.
Merci aux employés du service Recherche et Développement de Firmenich, en particulier Isabelle Cayeux, Christelle Porcherot, Bénédicte Le Calvé, Nadine Gaudreau, Christian Margot et Maria Inés Velazco. Ca été un grand plaisir d’interagir avec vous.
Merci à mes amis qui ne font pas de recherche (je pense notamment aux deux Ninis et à Sarah) pour toujours être là. Merci à ma famille pour savoir prendre avec humour les retombées des articles journalistiques (cf couverture de Psychologie le 1er avril 2010).
Last but not least, I would like to thank Bastiaan, for vigilantly enforcing proper English usage under all circumstances (including in this thesis). Much more importantly, thank you for enjoying your life with me.
PRELIMINARY NOTE
I wanted to emphasize that the work presented in this thesis a collaborative one. The credit of this work is really to be shared with all the persons involved in it, and listed as co-authors on the papers.
In order to clarify the key terms and concepts (followed by a *) used in this thesis, a glossary is provided at the end (p.177-178).
ABSTRACT
Pleasantness evaluations and preferences related to odors have traditionally been thought of as rather inflexible. An increasing number of empirical studies reveal, however, some evidence of flexibility. This thesis sought to examine how and to what extent decision- making processes can shape olfactory preferences, as well as the level of processing required in such a choice-induced preference modulation (explicit vs. implicit) and how long the modulation persists. In the empirical part of this thesis, we present empirical evidence demonstrating that (i) olfactory preferences can indeed be shaped by decision-making processes, (ii) this shaping does not require explicit processes, and (iii) the change can be long-lasting. These results invite further consideration of the modulatory impact of decision- making processes on olfactory preferences and of the importance of implicit processing in olfactory preference modulation and stability across time. The theoretical implications of these results, in particular regarding cognitive dissonance theory, are discussed.
TABLE OF CONTENTS
ACKNOWLEDGEMENTS... 2
PRELIMINARY NOTE ... 4
ABSTRACT... 5
1. INTRODUCTION AND OVERVIEW ... 10
2. THEORETICAL PART ... 15
2.1. Preferences: Definition and discussion of immutability vs. flexibility ... 16
2.1.2. Preferences as choices... 17
2.1.3. Preferences as attitudes ... 18
2.1.4. Preference as a behavioral predisposition ... 20
2.1.5. The concept of preferences in this thesis... 20
2.1.6. Immutability vs flexibility of preferences ... 22
2.1.6.1. Immutability of preferences... 22
2.1.6.2. Flexibility of preferences ... 23
2.1.7. Summary of this part... 25
2.2. Immutability vs. flexibility of chemosensory preferences ... 25
2.2.1. Immutability of chemosensory preferences... 25
2.2.2. Flexibility of chemosensory preferences ... 27
2.2.2.1. Needs, goals, and values in the flexibility of chemosensory preferences... 28
2.2.2.2. The role of learning and exposure factors in the flexibility of chemosensory preferences... 32
2.2.2.3. The role of other sensory modalities and cognitive factors in the flexibility of chemosensory preferences... 35
2.2.3. Summary of this part... 38
2.3. Investigating preference flexibility induced by decision-making processes ... 39
2.3.1. The free-choice paradigm... 39
2.3.2 Choice blindness... 40
2.3.3. Do choices truly affect preferences?... 41
2.3.4. The effort-justification paradigm ... 43
2.3.5. The induced-compliance paradigm ... 44
2.3.6. Summary of this part... 45
2.4. Theoretical interpretations of postchoice preference modulation and unanswered questions... 47
2.4.1. Classical interpretation: Cognitive dissonance theory ... 47
2.4.2. Explicit vs. implicit level of processing in cognitive dissonance theory ... 48
2.4.3. Stability of preference modulation following choice... 51
2.4.4. The first alternative to cognitive dissonance theory: Self-perception theory... 52
2.4.5. Self-anchoring and endowment effect ... 54
2.4.6. Summary of this part... 55
2.5. Thesis objectives... 55
3. EXPERIMENTAL PART ... 59
3.1. Methodological overview... 60
3.2. Experiment 1 - I’m no longer torn after choice: How explicit choices implicitly shape preferences of odors ... 63
3.2.1. Introduction... 63
3.2.2. Methods... 65
3.2.3. Results ... 67
3.2.4. Discussion... 70
3.3. Experiment 2 - Blind choices but not random preferences: Does the impact of blind choices on subsequent preferences depend on their congruency with pre-existing preferences?... 72
3.3.1. Introduction... 73
3.3.2. Methods... 75
3.3.3. Results ... 78
3.3.4. Discussion... 79
3.4. Experiments 3-4 - Choice both affects and reflects preferences... 81
3.4.1. Introduction... 82
3.4.2. Experiment 3 ... 85
3.4.2.1. Method ... 85
3.4.2.2. Results ... 87
3.4.2.3. Discussion... 88
3.4.3. Experiment 4 ... 90
3.4.3.1. Method ... 90
3.4.3.2. Results ... 93
3.4.4. General discussion... 95
3.5. Experiments 5-6 - Sniff of change: Sniff as a physiological correlate of pleasantness rating modulation following choice? ... 99
3.5.1. Introduction... 100
3.5.2. Experiment 5 ... 101
3.5.2.1. Method ... 101
3.5.2.2. Results ... 104
3.5.2.3. Discussion... 107
3.5.3. Experiment 6 ... 109
3.5.3.1. Method ... 109
3.5.3.2. Results ... 112
3.5.3.3. Discussion... 115
3.5.4. General discussion... 117
3.6. Experiments 7-9 - When flexibility is stable: Implicit long-term shaping of olfactory preferences... 119
3.6.1. Introduction... 120
3.6.2. Experiment 7 ... 126
3.6.2.1. Method ... 126
3.6.2.2. Results ... 130
3.6.2.3. Discussion... 132
3.6.3. Experiment 8 ... 133
3.6.3.1. Method ... 133
3.6.3.2. Results ... 136
3.6.3.3. Discussion... 139
3.6.4. Experiment 9 ... 141
3.6.4.1. Method ... 141
3.6.4.2. Results ... 146
3.6.4.3. Discussion... 148
3.6.5. General discussion... 148
3.6.6. Conclusion... 151
4. GENERAL DISCUSSION AND CONCLUSIONS ... 152
4. 1. Summary of the experimental part ... 153
4.2. Theoretical relevance... 155
4.2.1. Modulation of olfactory preferences by decision-making processes ... 155
4.2.2. Implicit mechanisms in postchoice preference modulation ... 156
4.2.3. Long-term effect of decision-making processes... 159
4.2.4. Summary of this part... 163
4.3. Perspectives and limitations... 163
4.3.1. In the end, do choices truly affect preferences? ... 163
4.3.2. On using a-priori emotional odors ... 167
4.3.3. On using other measures of preferences ... 168
4.3.3.1. Explicit measures ... 168
4.3.3.2. Implicit measures ... 169
4.3.3.3. Psychophysiological and cerebral measures... 170
4.3.4. On using consequential choices... 171
4.3.5. Cross-modal choice-induced preference modulation... 173
4.3.6. Summary of this part... 175
5. GLOSSARY... 176
6. REFERENCES... 179
7. RESUME EN FRANCAIS ... 208
1. INTRODUCTION AND OVERVIEW
One hot summer’s day a Fox was strolling through an orchard till he came to a bunch of Grapes just ripening on a vine which had been trained over a lofty branch. “Just the thing
to quench my thirst,” said the Fox. Drawing a few paces, he took a run and jump, but just missed the bunch. Turning round again with a One, Two, Three, he jumped up, but with no
greater success. Again and again he tried after the tempting morsel, but at last had to give it up. As the Fox walked away with his nose in the air, he said: “I am sure they are sour”.
It is easy to despise what you cannot get.
(The Fox and the Grapes, Aesop’s Fable)
Making decisions is like speaking prose – people do it all the time, knowingly or unknowingly.
(Kahneman & Tversky, 1984, p. 341)
Let us begin by breaking the overall topic of “The flexibility of olfactory preferences:
impact of decision-making processes” down into a sequence of smaller questions.
“The flexibility of olfactory preferences – are olfactory preferences really flexible?”
The extent to which olfactory pleasantness evaluation and preferences are flexible has been under debate in the literature. Some authors have argued that olfactory pleasantness evaluation and preferences are to some extent predetermined. This theoretical position, as well as empirical evidence to its support, has been criticized by authors insisting on the intrinsic ambiguity of olfactory perception and its sensitivity to modulation. An increasing amount of evidence seems to weigh in favor of flexibility. This thesis focuses on one of the factors that modulate preference changes: decision-making processes. The aim of this thesis is thus to describe to what extent olfactory preferences are sensitive to decision-making processes, and to investigate the underlying processes of such an influence.
“Impact of decision-making processes” – Do preferences shape choices, or do choices shape preferences, or both? (This discussion is not limited to olfactory preferences). According to the most dominant approach in contemporay social sciences, human beings indeed have a given set of preferences, and their choices – the decision- making processes that lead to a behavioral output – are a consequence of these preferences.
Preferences shape choices, and not the other way around.
An increasing amount of experimental evidence however suggests that rather than being stable, preferences are context dependent (e.g., Warren, McGraw, & Van Boven, 2011). According to this view, preferences do not reflect a set of values already present ex ante in an individual, but instead are a contextualized mental construction generated on a situational basis. If this is true, then preferences are the result of a multidimensional integration between numerous physical and cognitive dimensions. One of the most well established of these cognitive dimensions is precisely the impact of decision-making processes on current perceptual constructs.
Like the Fox in Aseop’s Fable, we devaluate options that we cannot get. More precisely, after a choice between equally liked stimuli, we evaluate the one we chose more positively, while devaluating the one we did not pick. These three steps (evaluation – choice – evaluation) constitute the free-choice paradigm (Brehm, 1956), a paradigm that has been used in hundreds of studies (for an overview, see e.g. Harmon-Jones & Mills, 1999).
If it is clear that decision-making processes can impact preferences, why study the impact of decision-making?
Despite the fact that there is now a substantial amount of work that demonstrates preference modulation following choice using this paradigm, little is known about the nature of the mechanisms underlying this effect. Most work assumes that cognitive dissonance reduction is the best way to theoretically interpret preference modulation following choice. Cognitive dissonance theory (Festinger, 1957) will be discussed in detail across this thesis, but in a nutshell, it claims that the choice puts the desirable aspects of the rejected option into conflict with the undesirable aspects of the chosen, which elicits discomfort (i.e., cognitive dissonance). Such discomfort can be reduced by ex-post realigning one’s preferences with the choice made – by devaluating the rejected option and valuating the chosen one more positively. This mechanism is thought to rely on explicit, conscious, mechanisms (Festinger, 1964).
There is, however, a surprising dearth of studies that directly examine whether the mechanisms at play are purely explicit, particularly since recent evidence suggests that implicit mechanisms may also be at play. Similarly, little is known regarding how long the preference modulation following choice lasts. Clarifying these two aspects would thus have important implications for cognitive dissonance theory, and more generally for the study of the acquisition and flexibility of preferences.
Why do we concentrate on the domain of olfaction?
Moreover, this thesis focuses on preferences in the olfactory domain. To the best of our knowledge, this work constitutes the first attempt to systematically study the impact of decision-making in this sensory modality. There are two important reasons which motivated our concentration on the domain of olfaction. First, we believe that the use of olfactory stimuli is ideally suited for studying implicit processes in preference modulation induced by choice because olfaction might rely more on implicit processes than vision or audition (Köster, 2005). Second, there is an ongoing discussion in the domain of olfaction about the flexibility of the evaluation of olfactory pleasantness. Our work contributes to this by investigating potential modulations caused by decision-making processes. We will discuss at length why we believe that studying this empirical question is important in the thesis objectives (section 2.5). All reported studies (apart from one) hence employed olfactory stimuli.
An overview of the structure of this thesis
This thesis is structured as follows: In chapter 2, we first define the concept of preferences, as used in this thesis, and discuss its assumed stability. We then review literature on the flexibility olfactory preference. Next, we present the main paradigms that have been used to investigate the way that decision-making influences preference flexibility driven. The main theoretical interpretation to such a modulation is discussed.
This theoretical part ends with the description of the three main thesis objectives. In summary, these are (i) investigating whether olfactory preferences can be modulated by decision-making processes; (ii) studying the level of processing (i.e., explicit vs. implicit) required for such a modulation to occur and (iii) testing the stability of this modulation.
In the empirical part (chapter 3), we present the nine different experiments we ran to fulfill these objectives. Eight of the nine reported experimental studies used variants of the free-choice paradigm, which was adapted to the goals of each particular study.
The first study investigates the modulation of olfactory preferences by explicit choices, and the extent to which such modulation could be present at an implicit level of processing (section 3.2). The second study investigates the extent to which blind choices can modulate olfactory preferences. In this study, participants were led to believe they were making certain choices, when actually they were predetermined by the experimenter.
We then tracked whether olfactory pleasantness changes, and whether this change depends on participants pre-existing pleasantness ratings or not (section 3.3). The third and fourth experiments address a methodological critique of the traditional free choice paradigm (Chen & Risen, 2010). The fifth and sixth studies investigate sniff patterns as a potential implicit behavioral correlate of preference modulation following choice (section 3.5). The seventh to ninth experiments explore the consolidation of implicitly shaped choice-induced preferences in long-term memory (section 3.6).
To conclude, we combine and integrate the theoretical and experimental parts of the thesis with a discussion of perspectives and limitations (chapter 4). We emphasize the importance of considering the role of implicit processing in olfactory preference flexibility and classical theories that explain postchoice preference modulation. We also develop future perspectives.
2. THEORETICAL PART
The theoretical part of this thesis is divided into four sections.
After making explicit how the term “preference” will be used in this work, we will focus our attention on one particular property typically assumed of preferences: their stability.
We will also briefly present empirical evidence arguing for the immutability vs. the flexibility of preference. These two points alone could of course fill many volumes, so we will restrict the discussion to the particular aspects that are essential for the present investigation (section 2.1).
We then proceed to the exposition of the theoretical backbone of the thesis, which is developed in sections 2.2-2.4. We first review empirical evidence regarding the immutability of olfactory preferences, as well as regarding the flexibility of them (section 2.2). This is followed by a review of studies that investigate preference flexibility induced by decision- making processes. An overview of the main paradigms used to do so is given (section 2.3).
Section 2.4 follows this up with an overview and discussion of theoretical interpretations of preference modulation following decision-making processes. Finally, in section 2.5, we derive from this a motivation to study olfactory preference flexibility, as induced by decision- making processes – which is the core topic of the present thesis.
2.1. Preferences: Definition and discussion of immutability vs. flexibility
The concept of preferences, like most concepts in psychology, is a hypothetical construct, i.e. a concept that is not directly observable, but which is assumed to underlie observable behaviors. This makes a precise definition and operationalization of the concept of preference particularly relevant.
In the social sciences literature, the term “preferences” is used in a variety of related, but not identical, ways. As noted by Furrer (2012), even though various disciplines make extensive use of notions labeled preferences, attempts to unify them into a single concept are few. After presenting a collection of common definitions of preferences, we will make explicit how we will use the term in what follows.
2.1.1. Preferences as comparative evaluations
Etymologically, the word preference comes from Latin praeferre – to carry in front.
Current definitions of the term of « preference » given by dictionnaries are very close to this etymological meaning. For instance, the Oxford English dictionnary (http://oxforddictionaries.com/definition/preference) defines a preference as « a greater liking for one alternative over another or others ».
Some authors, notably philosophers, have defended a similar position. For instance, Hausman (2012) has argued that preferences are best described as comparative evaluations.
According to him, “preferences and desires are not the same. The most important difference is that preferences, unlike desires, are comparative. To prefer something is always to prefer it to something else. If there are only two alternatives, one can desire both, but one cannot prefer both. Because they are comparative, preferences, unlike desires, require that one weight alternatives. They are thus more cognitive, more like judgments, than desires.” (p.X).
In the same vein, « a preference expresses a relational value judgment. It is relational in the sense that it connects two or more relata. These relata may be propositions expressing states of affairs, events, etc. or they may be bundles of goods. Preference is a value judgment in the sense that it compares relata with respect to (some aspect of) their value » (Grüne-Yanoff &
Hansson, 2009, p.8).
Preferences could consequently be understood as a comparative evaluation between two or more alternatives. Economics, which makes an extensive use of the term of
« preference », has however a very different notion of the concept.
2.1.2. Preferences as choices
In economics, preferences usually refer to choices. In the « revealed preference » framework (Samuelson, 1938), choice and preference are actually tautologically linked concepts. Indeed, choices, which are measurable, would reveal unmeasurable preferences. To illustrate this definition, let us consider how Kahneman, Ritov & Schkade (1999) contrast preference (used here in the economical sense) and attitude: « attitudes are defined by the affective value of objects considered one at a time, not by choices. […] The same object may
evoke different valuations depending on its description and on the context in which it is evaluated. […] The separate attitudes to two objects do not necessarily predict the outcome of a choice or direct comparison between them [i.e. preference] » (p. 568).
Although choice is the central notion of preference, in empirical work, preference is also frequently measured by willingness to pay (e.g., Farquhar, 1984), i.e. the maximum amount one is willing to pay to obtain something desired or to avoid something undesired.
The idea is that people exhibit a higher willingness to pay for preferred objects. When measuring willingness to pay, one is essentially measuring the indifference point for a set of choices between money and the choice object at hand. As long as people would choose more money over less money, one can then simply rank the willingness to pay for the different choice objects by size, in order to predict what people would have chosen if they had instead directly chosen between these choice objects. In this sense, willingness to pay is also fundamentally a choice-based approach to preferences.
To summarize, in the economic approach of revealed preference theory, preferences are linked to choice by definition. Choices are not caused by preferences, they are the preferences. According to Warren et al. (2011), the notion of revealed preference most closely corresponds to what psychologists would call expressed preference. In psychology, preference is more typically thought of as reflecting psychological (internal) processes.
2.1.3. Preferences as attitudes
According to Allport (1935, p. 798), “the concept of attitude is probably the most dinctinctive and indispensable concept” in social psychology. As emphasized by Maio and Haddock (2010, p.4), although different definitions of the term of attitude have been proposed in the psychology literature, all of them conceptualize attitude as the expression of an evaluative judgment. More precisely, attitude has been defined as a « psychological tendency that is expressed by evaluating a particular entity with some degree of favor or disfavor » (Eagly & Chaiken, 1993, p.1). Attitudes consequently vary in both valence and strength.
On the other hand, psychologists use the term preference to refer to “a latent tendency to consider something desirable or undersirable” (Warren et al., 2011, p. 194). This is the most typical definition employed in psychology. In psychology, preference is indeed often used to mean attitude (Phillips, Reed Johnson, & Maddala, 2002). According to this view, a
preference is a “relatively stable evaluative judgment in the sense of liking or disliking an object” (Scherer, 2005, p.703). Preferences have further been conceived of as an individual’s attitude towards a set of objects, typically reflected in an explicit decision-making process (Lichtenstein & Slovic, 2006). In Juslin & Scherer (2003), preferences and attitudes are conceptually merged together, and as such, share the same properties. For the purposes of our point here, the important thing to note is that according to the view just described, the affective states produced by the evaluated stimulus are of low intensity, stable, and have low behavioral impact (see Figure 1).
Figure 1. Preferences/attitudes properties (taken from Juslin & Scherer, 2005).
Thus, the terms “preference” and “attitude” are often used interchangably (Phillips et al., 2002). In studies using the free-choice paradigm, the main paradigm used in the framework of this thesis, authors have also employed both terms. For instance, Lieberman, Ochsner, Gilbert, & Schacter (2001) used the term “attitude” while Egan, Santos, & Bloom (2007) and Sharot, De Martino, & Dolan (2009) used the term “preference” – for what is in both cases an evaluative, explicit judgment that is related to hedonic experience. Such a measurement is an assessment “that taps a person’s more deliberative and acknowledged evaluations” (Briñol, Petty, & McCaslin, 2009, p. 285), i.e. an explicit measure.
Let us get back to the core definition of preference as an evaluative judgment: what does that mean exactly? While a judgment could be defined as “the assessment of an external or internal stimulus on an internal scale”, an evaluative judgment would be “a special type of
judgment, in which the internal scale is related to the person’s value system” (Zysset, Huber, Ferstl, & von Cramon, 2002).
To sum up, preferences, as attitudes, can be conceptualized as evaluative judgments.
This notion of preferences conceives of them as underyling preferences (Warren et al., 2011).
But psychologists, who are, by definition, interested in behavior, have also considered this facet.
2.1.4. Preference as a behavioral predisposition
Some definitions of preference highlight the goal-directed behavioral tendency component of this concept. For example, Zajonc and Markus (1982) have defined preference as “a behavioral tendency that exhibits itself not so much in what the individual thinks or says about the object, but how s/he acts towards it” (p.128). Preference towards X would be an
“indicator of the subjective expected value of engaging in goal-directed behavior towards object X” (Changizi, & Shimojo, 2008). Although Scherer considers preferences to have low behavioral impact, he still sees a role for them to generate tendencies to approach or to avoid (see Scherer, 2000).
Note that early definitions of attitudes have also emphasized the behavioral predisposition.
For instance, Allport (1935) has defined attitudes as a state of preparedness, and Campbell (1963) as a tendency to behave in certains ways towards the attitude object.
In this context, preferences are conceptualized as a behavioral predisposition, rather than as an evaluative judgment. If this is the notion one has in mind, one may thus require different measures that are more tailored towards actual behavior.
2.1.5. The concept of preferences in this thesis
As we have seen, there is a variety of definitions for the term “preference” – which points to the importance of specifying how the term will be used in this thesis.
The definition used in the context of this thesis takes the term “preference”* to mean a hedonic evaluation, i.e. an evaluation or appraisal of the (un)pleasant character of a stimulus or a situation. Such a definition relates to, at least, four different concepts in the literature.
First, it relates to liking (e.g., Berridge & Kringelbach, 2008). Liking refers to subjective hedonic reactions (Berridge & Kringelbach, 2008). As described by these authors,
“hedonic valence is neither an inherent property of a stimulus nor a mere epiphenomenon but results as an objective neural-psychological process from the interactive ability of the stimulus to engage hedonic brain systems” (p. 472). This conceptualization completes the notion of intrinsic pleasantness (Scherer, 2005), which are stable associations between a given stimulus (see also section 2.1.6.1.) and, to take Berridge’s terms, the ability to engage hedonic brain systems. Much like our definition, this conception emphasizes hedonic experience. It is, however, not fully clear what scope the term “reaction” encompasses. In our definition, we therefore decided to use the term evaluation, used in the sense of appraisal theories of emotion (e.g., Sander et al., 2005), due to the comparative clarity of its definition.
Second, this definition of preference refers to experienced utility. Experienced utility refers to the subjective pleasure experienced with a stimulus, while predicted utility refers to the beliefs, prior to consumption, about the subjective pleasure that will be experienced with a particular stimulus (Kahneman, Wakker, & Sarin, 1997). Here, we presented smells directly to the participants, rather than asked them to anticipate the pleasure they would have from a given smell, which allowed an assessment of experienced utility (see section 2.5).
Anecdotally, smells were already recognized and used for this purpose by Martin (1922). The notion of experienced utility is essentially identical with our conception of preference.
Third, this definition of preferences relates to the recently presented concept of a macro-valence (Sacharin, Sander, & Scherer, 2012). Macro-valence is assumed to be one- dimensional, an integrative “common currency” (Montague & Berns, 2002) of different qualitatively different types of evaluations, or micro-valences. In this respect, macro-valence is assumed to be the primary predictor of choice (Sacharin et al., 2012). Note that this concept is more akin to what Kahneman et al. (1999) label decision utility, than experienced utility – it is essentially a value assigned to an option before choice, as opposed to pleasure experienced with a stimulus. While these two relate, the relation is complex. For instance, the utility experienced with a stimulus in the past may serve as an input for computing decision utility.
Influences may hypothetically also go in the other direction, if chosen objects (i.e. the ones with the highest decision utility) are later experienced differently from unchosen objects – a thought that lies at the heart of this thesis.
Thus, in the framework of this thesis, the term preference will refer to hedonic evaluations, in other words, an evaluation of the degree of pleasantness of a stimulus. This term is most closely related to the concepts of experienced utility (Kahneman et al., 1997) and liking (e.g., Berridge & Kringelbach, 2008). There is also a relation to the concept of macro- valence (Sacharin et al., 2012), but this is more indirect and complex.
After having defined what we mean by preference, we will dedicate the next section to a discussion of the stability of preferences, or in opposite terms, their flexible character.
2.1.6. Immutability vs flexibility of preferences
2.1.6.1. Immutability of preferences
Some authors have emphasized the immutability of preferences, claiming that “many judgments and choices are based on stable, revealed, or inherent preferences” (Warren, McGraw, & Van Boven, 2011, p.193). The definition of what constitutes an inherent preference has been the object of discussions (Dhar & Novemsky, 2008; Kivetz, Netzer, &
Schrift, 2008; Simonson, 2008). Other authors like Scherer (2005) have linked preferences, conceived of as stable evaluative judgments, and intrinsic pleasantness. Thus, “preference should generate intrinsic apparaisal (intrinsic pleasantness check), independently of current needs or goals, although the latter might modulate the appraisal” (Scherer, 2005, p. 703).
Intrinsic pleasantness checks are presented as follows: “on a very low level of processing, the organism evaluates, with the help of genetically fixed schema or overlearned associations, whether a stimulus event is likely to result in pleasure or pain (in the widest sense). The term intrinsic refers to the fact that the evaluation is produced by a feature of the stimulus and is thus independent of the momentary preferences or goal states of the organism” (Sander, Grandjean and Scherer, 2005, p. 319). Both inherent and intrinsic aspects emphasize the stability of preferences. Here, rather than using the terms of inherent or intrinsic, we will talk about immutable preferences, which we feel is more tractable. By immutable, we mean preferences that remain stable during a prolonged period of time.
Some preferences appear to be present at birth and stay all life; the most obvious ones are probably the pursuit of pleasurable stimuli or situations and the avoidance of painful stimuli or situations, as mentioned in the previous quotation by Sander et al. (2005) (see also Elliott, 2006, for a historical overview of this notion). But they are not the only preferences present at birth that last across life. For instance, newborn humans preferentially look at facelike stimuli within minutes or hours after birth (e.g., Goren, Sarty, & Wu, 1975), thereby expressing a preference underpinned by a perceptual mechanism that ought still to be present in adults (Stein, Peelen, & Sterzer, 2011).
Stability may also be created by predispositions. In this respect, preparedness may provide a facilitated or impeded basis for acquiring particular preferences across life.
Seligman (1971) has proposed that the ease with which learned associations can be formed depends on biological predispositions. These predispositions in learning an association are called “preparedness”. According to this hypothesis, an organism is prepared, nonprepared or even contraprepared for learning particular associations. Twins studies have been used extensively to investigate biological predispositions towards given preferences and behaviors.
Those studies tend to show that certain preferences are more shared across monozygotic twins rather than dizygotic twins (e.g., Simonson & Sela, 2011).
Some preferences may thus be quite stable through time and/or predisposed. However,
“in many situations, we do not really know what we prefer; we must construct our preferences as the situation arises” (Lichtenstein, & Slovic, 2006, p.1). According to Warren et al. (2011), almost always, preferences are constructed. At least two definitions exist of constructed preferences – preferences that change across contexts and preferences that are computed during the judgment and choice process (Warren et al., 2011). Importantly, both definitions emphasize the flexibility of preferences (see Warren et al., 2011).
2.1.6.2. Flexibility of preferences
“We make choices and set preferences with respect to our construals of objects rather than the objects themselves”. (Trope & Liberman, 2010, p. 451). Factors able to influence this construction may consequently lead to preference modulation. The next section will enumerate some of these factors.
The most famous effect illustrating the flexibility of preferences – the mere exposure effect – has been theorized by Zajonc (1968). The mere exposure effect refers to the increased preference for a stimulus following repetead exposures to it. Replicated over and over, it has been demonstrated using visual (e.g., Zajonc, 1968), auditory (e.g., Bornstein, 1989), haptic (Jackesch & Carbon, 2012), olfactory (e.g., Delplanque et al., 2008; Delplanque et al., 2012) and gustatory (e.g., Pliner, 1982) stimuli.
The idea is that repeated exposures increase perceptual fluency, i.e. the ease of processing an object or a situation, which increases liking (e.g., Reber, Schwarz, &
Winkielman, 2004; Reber, Winkielman, & Schwarz, 1998). Consistently with this hypothesis, people favor their mirror image over their their non-inverted actual image, while they favor their partner’s regular image to his/her mirror image (Mita, Demer, &
Knight, 1977). Note that preference fluency, i.e. the subjective feeling that the preference formation is easy or difficult, may impact choices too (e.g., Novemsky, Dhar, Schwarz, &
Simonson, 2007).
Context has been shown to importantly impact choice and preferences in many experimental contexts (for a review, see Lichtenstein, & Slovic, 2006).
Preferences can also be modulated by psychological distance, i.e. « the subjective experience that something is close or far away from the self, here, now » (Trope & Liberman, 2010, p.440). There has been research on the impact of probability on preferences – the rationale being that outcomes with a low probability seem more psychologically distant than likely outcomes (Todorov, Goren, & Trope, 2007). Results show that the willingess to sign up for a given outcome, as well as the attractiveness of outcomes (the two measures of preference used in those experiments) depends on the probability of obtaining the outcome.
Values are broad motivational constructs that determine what we consider important and we goals we choose to pursue (Rohan, 2000). In addition to impacting choices (Brosch, Coppin, Scherer, Schwartz, & Sander, 2011; Brosch, Coppin, Schwartz, & Sander, in press), values can also impact preferences (e.g., Goodwin & Tinker, 2002).
2.1.7. Summary of this part
The aim of this part was two-fold: to define preferences and to discuss the assumption of their stability.
After discussing different notions, which take the meaning of the term
“preferences” as comparative evaluations (as some philosophers tend to), or choices (in the vein of economics), we presented psychological definitions of preferences. In this context, we discussed, first, preferences, as attitudes, i.e hedonic evaluative judgments, and second, preferences as behavioral predispositions. We specified that we will use the term
“preferences” in this thesis to designate hedonic evaluations.
We then mentioned that some authors have conceptualized preferences as immutable, i.e. asserted the stability of preferences over a long period of time. However, empirical evidence tends to show that preferences are rather flexible.
We will now review empirical evidence for the immutability vs. flexibility properties of olfactory preferences. As we will try to illustrate, the question of immutability and flexibility in the olfactory domain has received much attention.
2.2. Immutability vs. flexibility of chemosensory preferences1
2.2.1. Immutability of chemosensory preferences
We will focus here on the particular case of chemosensory preferences, i.e.
preferences regarding odors*, flavors* and tastes*. Chemosensory preferences can be argued to be, at least partially, innate and hard wired (Steiner, 1979), and as such, rather inflexible.
We will now review three main lines of evidence supporting this idea.
1This thesis section is adapted from Coppin, G., & Sander, D. (2011). The flexibility of chemosensory preferences. In R. J. Dolan and T. Sharot (Eds.). The neuroscience of preference and choice, pp. 257-275. Elsevier Publishing.
First, newborns display characteristic facial patterns as a function of hedonic variations of odors and tastes (Steiner, 1974, 1979). For example, in three-day-old infants, butyric acid, which is a rather unpleasant smell, elicits significantly more facial markers of disgust than vanillin, which is on average considered a pleasant smell (Soussignan, Schaal, Marlier & Jiang, 1997). Newborns also behave differently to different tastes: in comparison to regular water, they drink less water when citrid acid has been added to it (Desor, Maller, &
Andrews, 1975) while they drink more when sugar has been added (Desor, Maller, & Turner, 1973). When the sucking to access a sweet drink was made difficult, female newborns, although more responsive to it, were less willing to exert effort for it (see Nisbett & Gurwitz, 1970). Such innate preferences have the potential to permit the detection and selective consumption of sweet-tasting calorie-rich food, and the avoidance of potentially toxic food, which often tastes bitter (e.g., Ventura & Mennella, 2011).
Second, the innate facial patterns in response to particular odors and tastes are strongly homologous to those of other primate species, especially the species that are phylogenetically our closest relatives (Steiner, Glaser, Hawilo & Berridge, 2001). Chemosensory preferences could be shared across species and be partially genetically determined. For instance, the lifelong taste preference for sweetness might have a genetically coded component (Keskitalo, Knaapila, Kallela et al., 2007). Olfactory preferences are partially shared across mice and humans (Mandairon, Poncelet, Bensafi & Didier, 2009). Preferences for perfumes are correlated with the major histocompatibility complex (MHC) (Milinski & Wedekind, 2001).
Taken together, this evidence suggests that there is a genetically based and predetermined component of olfactory preferences.
Third, some physicochemical properties of odorant molecules can predict how humans will perceive their hedonic character (Khan et al., 2007). Those properties also share a direct relation with the perceived number of odor notes (Kermen et al., 2011). Moreover, such properties can allow an artificial nose to categorize the odors according to their pleasantness with high accuracy (Haddad, Medhanie, Roth et al., 2010). These results suggest that there is a predictable link between odorant structures and stimulus pleasantness.
However, as nicely put by Birnbaum (2011), “there is something about smell that defies structure” (p. 280). Olfactory pleasantness and preferences are more than just a predictable consequence of genetics and the molecular structure of a smell (see e.g., Coppin, Delplanque, & Sander, 2011).
2.2.2. Flexibility of chemosensory preferences
While there is thus evidence for some comparatively stable aspects of olfactory perception, other authors have emphasized that olfactory evaluation has a high degree of flexibility, and documented high plasticity in responsiveness to odors (e.g., Engen, 1979, 1988). A significant body of evidence substantiates the claim of flexibility. This goes so far that some authors claim that today, “there is no evidence for any innate odor preferences”
(Yeomans, 2006, p. 800).
Chemosensory preferences develop very early, as early as when a child is still in the womb, influenced by the mother’s diet (e.g., Beauchamp & Mennella, 2009; Schaal, Marlier
& Soussignan, 2000; Mennella, Jagnow, & Beauchamp, 2001; Schaal, Soussignan & Marlier, 2002). Olfactory preferences at birth change after few days (Varendi, Porter, & Winberg, 2008). Early and regular exposure at the mother’s breast to a given smell has an impact on a child’s preference for the smell itself. There is also evidence that such exposure has an impact on a child’s behavior at 7 and 21 months of age toward objects with the same smell (Delaunay-El Allam, Soussignan, Patris et al., 2010). Moreover, early exposure to a smell (Poncelet, Rinck, Bourgeat et al., 2010), such as first associations between an odor and an object, may have a “privileged brain representation”, in particular in terms of hippocampus activity (see Yeshurun, Lapid, Dudai & Sobel, 2009).
Cultural background and experience have been shown to matter more than genetics for preferences regarding sweet tastes (Mennella, Pepino & Reed, 2005). An appreciation of bitter tastes can be acquired across life, despite a newborn’s reaction of disgust to these tastes (Mennella et al., 2005). Harris (2008) has shown that neophobia, i.e. the reluctance to try new foods, is determined both by genetics and environmental influences. In species like Drosophilas, most odorants elicit attraction or avoidance behaviors. But even there, it
depends on their concentration: while Drosophilas are attracted to many odorants while at low concentration, they are repulsed at high concentration (see e.g., Wang et al., 2003).
Here, we will review empirical results regarding the flexibility of chemosensory preferences in restricting our attention to the factors that are most strongly linked with chemosensory processing. In particular, we will discuss modulatory factors as they relate to the three basic functions of the chemosensory system. These functions can be classified into three main categories: ingestion (e.g., appetite regulation); avoidance of environmental hazards (e.g., detection of microbial threats); and social communication (e.g., emotional contagion) (Stevenson, 2010).
This thesis’ part is structured as follows. First, we briefly present how needs, goals and values relate to chemosensory preferences. Second, we emphasize the role of learning and exposure in the flexibility of chemosensory preferences. Third, we discuss the importance of other sensory (e.g., visual inputs) and cognitive (e.g., verbal labels) information in the flexibility of chemosensory preferences. Pertinent data on the neural underpinnings of the described phenomena are presented throughout the section, where applicable.
2.2.2.1. Needs, goals, and values in the flexibility of chemosensory preferences
As mentioned earlier, the chemosensory system is intimately linked to ingestion, avoidance of environmental hazards and social communication. As such, it provides the organism with assistance for answering questions such as: “Shall I eat the rest of this dish that has been sitting in my fridge for two days, or should I cook something new?” “Shall I approach or stay away from this person?” “What self-image do I want to communicate by wearing a perfume during this romantic dinner?”
The aim of the following part of this section is to provide insight into how needs, goals and values are able to modify preferences regarding smells and flavors. The role of the amygdala and the orbitofrontal cortex (OBC) in such modulation is also presented.
Needs
Needs refer to a psychological entity assumed to arouse actions towards goals that a person would pursue for their satisfaction (Gendolla, 2009). For instance, the need for
nourishment is intimately connected with the goal of food intake. Preferences regarding such a goal depend notably on both the quality and quantity of previously ingested food. Regarding the quantity of already ingested food, one study showed that even if chocolate was rated as pleasant at the beginning of an experiment, the more chocolate that participants consumed, the less the chocolate was rated as pleasant (Small, Zatorre, Dagher et al., 2001). There is evidence that such preferences related to food intake may be coded by the activity of the OBC cortex. Its activity is related to the representation of the affective value of smells (Rolls, Grabenhorst & Parris, 2010; Small, Bender, Veldhuizen et al., 2007) and tastes (Rolls, Critchley, Verhagen & Kadoshisa, 2010) and more generally to the affective value of stimuli, independently from their sensory modalities (see however Gottfried & Zelano, 2011). Even more generally, there is evidence that OBC activity is related to social and monetary stimuli (Grabenhorst & Rolls, 2011; Rolls & Grabenhorst, 2008). Crucially for our point here, appetite modulates OBC cortex activity, the activity of which is decreased after consumption to satiety of chocolate (Small et al., 2001), bananas (O’Doherty, Rolls, Francis et al., 2000), tomato juice or chocolate milk (Kringelbach, O’Doherty, Rolls & Andrews, 2003).
Regarding the quality of already ingested food, a phenomenon called sensory-specific satiety elicits modulation of food pleasantness. Thus, the decrease in pleasantness of the sensory properties of a food eaten to satiety (Rolls, Rolls, Rowe & Sweeney, 1981) is larger than the corresponding decrease for foods that have not been eaten. This phenomenon also applies to foods that share sensory properties of the eaten food (Rolls, Vanduijvenvoorde &
Rolls, 1984). Thus, the pleasantness of food items can be modulated by how much and which type of food has already been ingested. As appetite and food intake are by definition in constant change, preferences regarding food items appear to have a highly modulated character.
Goals
Goals are another important factor for food intake preferences. You might, for example, eat tofu rather than meat if your current goal is to keep your vegetarian co-workers happy, even if you like meat more than tofu. The impact of diet on food consumption preferences has been experimentally studied, both at the behavioral and at the neural level.
Results have shown that participants trying to control themselves chose unhealthy but tasty
food items less frequently than did participants who were not trying to control themselves (Hare, Camerer & Rangel, 2009). Moreover, in a food consumption context, the decisions of dieters, as well as non-dieters, were related to the activity of the ventromedial prefrontal cortex; this activity correlated with the expected reward that was associated with the consumption of a given food. But in contrast to the decisions of non-dieters, dieters’ decisions were also related to dorsolateral prefrontal cortex activity, which plays a role in self-control (Hare et al., 2009). The pursuit of a goal such as losing weight consequently modified food consumption preferences in dieters.
Values
Values are defined as broad motivational constructs that determine what we consider important and which goals we choose to pursue (Rohan, 2000). Does the influence of values, discussed above, also extend to the domain of chemosensory preferences? The answer is yes:
when there is a match between one’s most important values and the value symbolized by a product, the product tastes better (Allen, Gupta & Monnier, 2008). To give a concrete example, we first need to point out that the consumption of red meat is correlated with the value of social power, while vegetables and fruits symbolize the rejection of social power (Allen & Ng, 2003). This correlation has been shown in three studies that measured human values together with attitudes towards different types of foods by means of questionnaires (Allen & Ng, 2003). Experimental evidence showed that participants who reject social power evaluated a vegetarian alternative to a sausage roll as more tasty and had a higher purchase intention. This effect was independent of the food they actually tried (Allen et al., 2008).
Amygdala: A central structure in Needs-, Goals- and values-related relevance
According to Gottfried (2010), “the function of sensory systems is optimized to detect and encode behaviorally relevant events (objects) that are encountered in the real world” (p.
637). Needs, goals and values all contribute to the relevance of a given chemosensory stimulus. In other words, a smell or a flavor can have a different significance across different individuals and different contexts, depending on current needs, goals and values. This differential importance of a smell or a flavor across individuals and contexts may lead to changeable preferences towards it, as discussed earlier. In terms of neural underpinnings, the amygdala is known to be activated by odors (de Araujo, Rolls, Velazco et al., 2005; Gottfried,
2008). The amygdala has also been proposed to act as a “relevance detector” (Coppin &
Sander, in press; Pessoa, 2010; Sander, Grafman & Zalla, 2003; Sander, 2009; Sander, in press). In the olfactory domain, highly aversive olfactory (Zald & Pardo, 1997) and gustatory (Zald, Lee, Fluegel & Pardo, 1998) stimulation can be considered very relevant – and has been shown to elicit amygdala activity. However, amygdala activity is not just related to aversive chemosensory stimulations. Winston, Gottfried, Kilner and Dolan (2005) have shown that the amygdala responds to a combination of valence and intensity, which could reflect the overall relevance of a given smell. In the gustatory domain, Small, Velduizen, Felsted et al. (2008) suggested that the amygdala, together with the thalamus, may predict the meaningfulness or biological relevance of tastes and flavors.
What experimental evidence exists for this hypothesis? Amygdala activity was higher when participants were hungry compared with when they were not hungry during the visual presentation of food items (LaBar, Gitelman, Parrish et al., 2001). A similar pattern was found when participants were told to imagine being in a restaurant and choosing their favorite food from the menu (Hinton, Parkinson, Holland et al., 2004). The activation of the amygdala was also more pronounced when participants were reading food names that they particularly liked, compared with reading names of more neutral food (Arana, Parkinson, Hinton et al., 2003). By recording single neuron activity from the amygdala while participants were making purchase decisions about food items, Jenison, Rangel, Oya et al. (2011) showed that the amygdala response was linearly related to the value assigned to a given food. This representation seems to flexibly depend on the level of hunger of the participants. Both the amygdala and the OBC cortex responses to a predictive cue of the presentation of food-related smells were shown to decrease as a function of satiety (Gottfried, O’Doherty & Dolan, 2003).
Taken together, the findings suggest that amygdala activity is related to the importance of smells and flavors in a given context, and more generally, to the relevance of a stimulus or a situation (Sander et al., 2003). Note that relevance is highly flexible across time and space and can be acquired and modified. This may explain why some authors such as Köster (2002) think that it is absurd to ask people the question “Why do you like this food?”. Why we like smelling or eating something at a particular point can depend on numerous and variable factors that we are not necessarily aware of.
In the next part, we will discuss the role of learning and exposure in modulating, and even creating, preferences for smells and flavors.
2.2.2.2. The role of learning and exposure factors in the flexibility of chemosensory preferences
The role of aversive and appetitive conditioning
Olfactory learning is crucial for allowing flexibility and adaptation in a given olfactory environment (Hudson & Distel, 2002). One way to achieve learning is by means of conditioning, which has been extensively studied in the context of preference creation and modulation (e.g., De Houwer, Thomas & Baeyens, 2001). Conditioning can allow the discrimination, both perceptually and cortically, of previously non-discriminable odor enantiomers, which are mirror-image molecules (Li, Howard, Parrish & Gottfried, 2008).
Aversive and appetitive conditioning also plays an important role in preference learning and modulation, both for olfactory and gustatory stimuli. Regarding aversive conditioning, the smell of eugenol, for example, is rated as unpleasant by participants who fear going to the dentist, probably by association with potentially painful dental treatment (e.g., Robin, Alaoui-Ismaïli, Dittmar & Vernet-Maury, 1999). In the gustatory domain, the power of aversive conditioning is even more impressive: a conditioned taste aversion requires only single-trial learning and works despite a long delay between a given taste and an illness (Bernstein, 1991; Garcia, Hankins & Rusiniak, 1974). Regarding appetitive conditioning, food preference reinforcement can have two sources: so-called flavor-flavor learning and flavor-nutrient learning (Ackroff, 2008). Flavor-flavor learning consists of the increased evaluation of flavors by association with already preferred flavors (such as a sweet flavor;
e.g., Fanselow & Birk, 1982). Flavor-nutrient learning represents the increased hedonic character of flavors whose ingestion is followed by the pleasant effects of the nutrients. The cerebral structures underlying this second type of preference learning have recently been investigated in humans. The cerebral areas involved are the striatum, the amygdala and the medial OBC cortex (Fobbs, Veldhuizen, Douglas et al., 2011).
Learning includes – but is not restricted to – aversive and appetitive conditioning. The mere repetitive exposure of a smell or a flavor can be another important type of learning, as it leads to the creation or the modulation of its hedonic evaluations.
The role of exposure
The creation of chemosensory perception
Androstenone is a particularly interesting molecule for investigating the creation of olfactory perception and preference. A large part of the population cannot smell this odorant (approximately 40% of individuals cannot perceive an odor when presented with androstenone; Boyle, Lundström, Knecht et al., 2006). When not perceived, “smelling”
androstenone is rated as neither unpleasant nor pleasant. It is, however, possible to acquire the capacity to perceive androstenone after repeated exposure (e.g., Wysocki, Dorries &
Beauchamp, 1989; Coppin, Pichon, et al., 2012). Once this ability has developed, androstenone is on average perceived as unpleasant.
The mere exposure effect with olfactory stimuli
The mere exposure effect (Zajonc, 1968) has been reported to affect preferences in a variety of settings. Its effect on smells is of particular interest, because the effect of repeated exposure to smells may be somewhat complementary to the survival relevance of quick and powerful taste aversions (Delplanque, Grandjean, Chrea et al., 2008).
As in other sensory domains, a high correlation between familiarity and pleasantness has been reported in the olfactory domain (e.g., Cain & Johnson, 1978). If this is a causal relationship, then repeated exposure to a smell increases one’s preference towards it.
However, this relationship would not be true for all types of smells: an increased preference occurs only for initially neutral to positively evaluated odors, but not for negatively evaluated odors (Delplanque et al., 2008; Delplanque et al., 2012). This non-extension of the mere exposure effect to negative odors may be understood regarding the high relevance of malodors for survival. Given the importance of olfaction for ingestion behaviors (Stevenson, 2010), it would not be adaptive to start liking, and possibly to have an appetite for or to approach, something that is potentially lethal just because of repeated exposures.
This selectiveness of the mere exposure effect has also been demonstrated in the case of interpersonal evaluations, depending on the reward versus punishment associated with seeing a particular person. The more a given person was seen, the more the person was liked, but only when he or she was associated with the delivery of a reward and not when associated with a punishment (Swap, 1977).
The role of culture and social appraisal
The frequency of exposure to different kinds of smells and foods is highly dependent on one’s culture. Although common folk psychology suggests that “there is no accounting for taste,” it is in practice not uncommon for people to criticize one another’s culinary tastes when these are perceived as bizarre – particularly across cultural divides. For example, anecdotally, the British humorously call the French “frogs” – because frog’s legs are considered a delicacy in France, but not typically in Britain. Similarly, baby mouse wine (a bottle of rice wine containing the bodies of baby mice) may not sound as appealing to foreigners as it does to its consumers in China and Korea, where it is available.
In scientific research, culture has been demonstrated to be a powerful force in olfactory (Ayabe-Kanamura, Schicker, Laska et al., 1998; Ferdenzi, Schirmer, Roberts et al., in press; Moncrieff, 1966) and gustatory (Bourdieu, 1984; Wright, Nancarrow & Kwok, 2001) preferences. The smell of durian, a common fruit in Asia with a very powerful and characteristic aroma, evoked feelings of disgust in Geneva or Liverpool, while it was evaluated as mainly pleasant in Singapore (Ferdenzi et al., 2011). Directly related to culture, identity might also be an important factor in the evaluation of identity-relevant smells. Thus, the olfactory evaluation of chocolate, for which Switzerland is world famous, is modulated by accessibility to the Swiss identity in Swiss participants (Coppin, Delplanque, et al., 2012).
The concept of social appraisal probably also plays a particularly important role here.
Social appraisal (Manstead & Fischer, 2001) globally refers to social influence on the appraisal of a given stimulus or person. It has been shown to notably influence women’s preferences for men. Thus, a man being looked at by a smiling woman is going to be perceived as more attractive than is a man being looked at by a woman with a neutral facial expression (Jones, DeBruine, Little et al., 2007). Similarly, in the chemosensory domain,
